PERSPECTIVES
ON PSYCHOLOGISM EDITED BY
MARK A. NOTTURNO
E.]. BRILL LEIDEN • NEW YORK • K0BENHAVN • KOLN 1989
PSYCHO...
13 downloads
713 Views
84MB Size
Report
This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form
PERSPECTIVES
ON PSYCHOLOGISM EDITED BY
MARK A. NOTTURNO
E.]. BRILL LEIDEN • NEW YORK • K0BENHAVN • KOLN 1989
PSYCHOLOGISM J.N. MOHANTY The word 'psychologism' has been used, in the brief history of just about a century, pejoratively, and not descriptively. Like many such words 'capitalism' in socialist countries and 'socialist' in western democracies labeling a philosopher's idea as involving psychologism has been taken as tantamount to saying why it should be rejected. Brentano, in exasperation, complains that 'psychologism' is "a recently fashionable term, at the sound of· which many an ingenuous philosopher, like many orthodox Catholics at 'Modernism', makes the sign of the cross, as though the devil himself lurked in the word. I If this is the case, then let us not be swayed by the pejorative connotation the word has acquired and let us make an effort to understand and save the phenomenon on which psychologistic doctrines are based without falling prey to some of their ruinous excesses. There is also something odd about the complaint of psychologism. The very same people who sought to demolish psychologism have themselves been accused of it. Thus Kant is the first to have sought to avoid psychologism (ind~ the introduction of this word by J.E. Erdmann2 was to highlight the contrast with 'criticism') and yet, not long after Erdmann, Windelband insisted that Kant's criticism was in fact dependent on his psychological theories.3 As is well-known, Husserl the critic par excellence of psychologism, suffered the same fate. Not even Frege escaped this fate: his insistence on the assertion sign has been read as introducing psychologistic elements into logic- not to speak of Philip Kitcher's recent attempt to find a psychological account of knowledge in Frege.4 Less known perhaps is Carnap's case: he attacked psychologism,S and is then attacked by Popper as being psychologistic.6 What, in view of this cross fIre, is the prospect that we can ever be free from psychologistic elements in our thinking? But what after all is psychologism? And why should it be so easily taken for granted that psychologism does indeed need to be gotten rid of] I will not pretend to add more clarification to a theme to which Frege and Husserl devoted so much of their energy. But no less a thinker than Frege seems to have been confused. In his review of Husserl's Philosophie der Arithmetik, Frege accused the psychologistic logician of reducing everything to subjective ideas.7 Now, note that reducing everything to subjective ideas amounts to what is known as subjective idealism, and that is not eo ipso psychologism. The following may suffice to show that psychologism does not necessarily amount to fI
2
J.N. MOHANTY
subjective idealism. A psychologistic philosopher may believe, consistently, that there are indeed mind-independent, objective realities. He may even further hold that we do have knowledge of this objective reality. What he must hold is that such knowledge is made possible not only through the structure of the reality that is known but also through the structure of the mind that knows. In so far as knowledge is made possible by the structure of the mind, we cannot sufficiently explicate the foundations of knowledge save by explicating the way the human mind works. It is not, then, surprising that Kant, whose theory of knowledge, as transcendental philosophy, wanted to set itself off from all psychologism, remained nonetheless under the constant threat of psychologism. For the line that divides the transcendental foundation from a psychologistic foundation is so thin that its transgression may go unnoticed. Psychologism, then, is not an ontological thesis. It is an epistemological thesis, which traces back all epistemological questions to some aspects of psychology. It need not have to hold, however, that everything is nothing but mental representations. Once we have thus separated the ontological question, we can certainly go on to ask: what is it that the psychologistic thinker holds which makes his position both interesting and disturbing? There is a trivial assertion to the effect that knowing is a mental experience, that all knowledge whatsoever is merely a mental performance and accomplishment- from which it appears to follow that only a science of mind - psychology, to be sure - can provide the foundational expectations of knowledge. (One may likewise argue that since all knowledge is expressed in language, a study of language should also have that foundational status! I do not intend to make this last remark only as an aside. My ulterior motive is to emphasize that when these two trivialities supplement each other, a more interesting thesis is likely to emerge. But more of this later.) But the psychologistic consequence is taken to follow from a trivial assertion because of a certain equivocation in the use of the word 'psychological'. 'Psychological' may mean the mental as such, or it may mean the mental as thematized in the science of psychology, or it may apply - as when one speaks of psychological laws - to the propositions stating laws or lawlike connections within the psychological discourse. To claim then that because all knowledge is a mental performance or accomplishment, the foundations of knowledge must be explicated in psychological tenns may mean either: (1) that all knowledge must ultimately rest upon the mental as such, Le., as lived experience; or (2) that all knowledge must ultimately be analyzable into those components, such as ideas or presentations which empiricist psychology took mental life to consist in; or, finally, (3) that we must look at the laws of mental life for the source of the laws of our more developed logical and epistemic performances. Dilthey's psychologism was of the frrst sort, Mill's of the third sort; whereas the psychologism Frege and Husserl were criticizing was either or both of the second and third sorts. There is obviously a reason
------------------------------------------------------------------------------
PSYCHOLOGISM
3
why the second and the third sorts are put together: the sorts of psychological laws one takes into account depend upon the sorts of elements into which mental life is decomposed and whose law-like connections the psychology under consideration is formulating. Of the Diltheyan sort of foundationalism, I will only say this much: while it may be true that any act of knowing or thinking must be a lived experience on the part of the knower or the thinker, that lived experience, by itself, has no explanatory value in so far as the knowledge or the thought - whether the specific performances or the generic - is concerned. The inner lived experience is, if that theory be true, only a necessary accompaniment of, for example, an act of judging, but it does not give us any insight into the structure of judgment or into the conditions of its possibility. Such a psychologism is harmless, it gives necessary but not sufficient conditions of our cognitive concepts. If we did not have this inner lived experience of what we are doing (judging, etc.), we might not have had the appropriate concepts (of judging, e.g.), but simple inner experience does not by itself enable us to have the right concepts. Thus when Lipps says "nur das Erleben kann der Quell sein, aus welchem der Logiker schopft," the point is just a hannless triviality, if true.8 Let us now consider the second sense of 'psychological'. The mental as thematized in empirical, introspectionist, associationist psychology becomes a private, discrete, event concerned with other such events by the laws of association. It is in this sense that Frege and Husserl found psychologism unacceptable for obvious reasons. How can the necessary truths of logic, they asked, be founded upon the more or less probable, empirical generalizations of such psychologies? But while they were right in calling into question psychologism in this sense, it is not obvious that the only remaining option is an epistemological theory for which thinking is a grasping of thoughts and knowing a grasping of whatever is known. I shall make myself clearer. The issue is not that of Platonism, which is an ontological question: are there abstract entities? The issue is epistemological: are the fundamental principles of knowledge, or of thought, grounded, in some sense, in the nature of human mind? I agree with the Frege-Husserl critique that psychology - in the sense of the British empiricists, associationist psychology - cannot play this foundational role. It is no use to return to the immediacy of Diltheyan lived experience (or, the Brentanian inner perception)., The opponent of psychologism would want to strip the mental of all contents. If the mental is reduced to the mere 'grasping', the bare act with no content, if all alleged content is expelled out into the world and the mental left to be the bare Sartrean nothingnesspsychologism cannot even fmd a foothold from which to get off the ground. Psychologism knew only one sort of content within the mental: this is what Locke, Berkeley, and Hume, the British empiricists, called sensations, impressions, or ideas (Frege called them Vorstellungen). These contents and the laws that weld them together are not the sorts of things that could provide the basis for higher order cognitive achievements. Rightly, therefore,
4
J.N. MOHANTY
the anti-psychologists rejected any such attempt, but many of them left the psychologistic concept of the mental precisely as it was. This is where the crux of the matter lies. What, in fact, the anti-psychologistic thinkers did is a poor compromise. The received picture of mental life as consisting of incurably private particulars was not iL~lf challenged. That picture was left unchanged; however, it has no role to play in thinking and other cognitive achievements. Let us call these latter the 'noetic acts'. The noetic acts were construed as bare grasping, bare recognition- having no contents of their own and no structure. What is grasped, asserted, recognized falls on the other side of our acts, in fact belongs to the world- a real or unreal entity, a concrete physical object or an abstract entity such as a number, for example. Being bare acts with no content or structure of their own, little can be said or done about them. This compromise left the logician to do his work, and the empirical psychologist his. They are assigned different domains, neatly marked off from each oth~r; but it provided, in my view, a poor philosophy of mind and a poor philosophy of logic. My goal is to bring them closer, to ground philosophy of logic and epistemology in a philosophy of mind- without having to court the ruinous consequences of psychologism. This goal can be sought to be achieved in a number of ways, as follows: 1. Make the rules that mental operations follow important, but the acts and operations themselves dispensable. The rules would have a logical structure, the acts with their subjectivity and privacy would be irrelevant for foundational purposes. This move is easily derivable from the Kantian. This is in fact what the Marburg neoKantians, notably Cassirer did. 2. The preceding move has also a strange similarity with the cognitivist reading of Husserl that Hubert Dreyfus gives. If the Husserlian noema is a set of rules and determines what the object must have to be by way of conforming to these rules, we get rid of the subjectivistic consequences of psychologism, and are able to make a rapproacJunent with a different sort of psychology, namely, with a 'computational theory of mind'. 3. Keep the mental act, but give it a structure. This can be done by assigning a content (not an object) to an act, and by making sure that the content is not a subjective, private particular but rather a structure that numerically different acts, performed by different individuals, may share in common, and so somewhat universal-like. This is done by Husserl in his thesis about the act as having meaning, a Sinn, or noesis-noerna correlation. 4. Finally, raise a psychological conception of the mental to a transcendental level- a strategy adopted by transcendental philosophers from Kant to Husserl. There are certainly other possibilities, and I have not tried to be exhaustive. For my present purpose, I shall argue that a good philosophy of mind needs to combine the above four principles - the truths in each of them, that is to say - and, given the resulting, enriched picture of mental life, the fundamental insight of an enlightened psychologism can be
PSYCHOLOGISM
5
preserved and reconciled with the undoubtedly valid point of the anti-psychological thinkers. First, some brief comments on each of the four: (1) Cassirer insisted that the essential content of the Kantian philosophy is not the relation of the world to the ego, but rather its relation to "the legality (Gesetzlichkeit) and the logical structure of experience. "9 His argument rests on the consideration that the concrete ego, the '1', is the subject as much of true knowledge as of error. But since the Kantian philosophy is concerned with the conditions of the possibility of truth, we have to be satisfied with a theoretically normative autonomy of the logical principles which need no further philosophical grounding. Now, the principles of the Kantian Grundgesetze which underlie physics, are, for one thing, reached by a process of 'transcendental argument' starting from the fact of Newtonian physics and its concepts of space, time, etc. If you start with a different physical theory, you would reach a set of Grundgesetze different from the Kantian. Thus the validity of the set of principles chosen as foundational is relative to the already taken-for-granted validity of a theory. So far so good. But there is a different level of questioning: namely, how are these principles themselves constituted? By what sort of mental operation? As meaning-structures, how are they possible? Recall the Humean question regarding the principle of causality! (Correlatively, how do we understand such principles? Through what sort of mental operationssuch as counting, imagining, phantasizing possibilities, etc.?) (2) If mental life, in its cognitive operations, consists in entertaining Fregean Sinne, and if a Sinn consists in a set of rules prescribed for whatever is to be the referent, one may - as Dreyfus suggests - appropriate a Frege-Husserl theory into a computational theory of mind. According to Dreyfus, Husserl thought of the noemata or Sinne as meaningful complex formal structures, since even without the digital computer to supply a model for his intuitions, he thought of the noema as a "strict rule for possible synthesis. "10 And if a computational theory is a mathematical theory about how the mind operates, to ground knowledge on such a theory, even if psychologistic, would hardly generate the consequences which Frege and Husserl drew from psychologism. But there is a deceptive gain: we are to explain, how logical thinking and cognition are possible. We answer by inserting these abilities into the most basic stratum of mind. The mind is like a computing machine. Our conceptual gain is very little. The psychological theory that is used, itself presupposes formal logic and mathematics, and so cannot provide for their foundation- not to speak of the narrowly syntactical idea of computation that the theory has to work with. This last point incidentally shows the untenability of Dreyfus' interpretation of the Husserlian noerna; for Husserl the rules are also semantic as well as syntactical. (3) The third way out appeals to the thesis of intentionality- not the simple Brentano thesis that an act is directed towards an object, but the enriched Brentano-Husserl thesis to the effect that every act has its correlative Sinn, so that we have always a concrete meaningful act- a Sinn
6
J.N. MOHANTY
structure. Every act is performed by some one, at a point of time or enduring through a temporal stretch, has a certain act-quality (Le., is either a believing, doubting, imagining, remembering, judging, etc.), and has a content (or Sinn) which is the structure shareable by other acts perfonned by other individuals. Given such a conception of mental or noetic acts, there can be a perfectly good eidetic psychology which would discover universally valid laws by which acts of a certain quality and having certain structure and temporal location would intentionally imply acts of some other sorts. Besides, one can simply abstract the Sinne from their imbeddedness in acts, and relate them to other Sinne by familiar logical relations. Such an eidetic psychology - as Brentano well saw - can provide foundations for logic and cognition is a certain sense- without entailing psychologism in a perjorative sense. One way of doing this is as follows: Even if we leave out of consideration, as irrelevant, the historical factual nature of research or any reference to the psychological/biographical antecedents, there still remains another sort of psychical locution which is involved in an essential manner in theoretical inquiry. One may say, for example: "Supposing A, is it an open question whether B holds good or not? Is B a valid consequence or supposition?" Or: "If one is certain that A, one cannot have any more doubt that B." Or: "We do not know if B, but from the already developed theory we have 'reasons' to believe that most probably B:' These locutions concern the context of justification, and eventually refer to certain sorts of relations between appropriate mental acts having specified contents. 11 Long before Brentano, the Indian logicians developed a theory of inference as a theory of eidetic, rule-governed, psychology of inferential cognition. Take the following time-worn instance. One sees smoke on a distant mountain. This leads one to remember the rule (previously learnt) 'Wherever there is smoke, there is fIfe', which one recollects as having been instantiated in cases such as the familiar stove in the kitchen. It is now recognized that this column of smoke is a mark of fife in accordance with the rule just remembered. At this point, if there is no hindrance, the person would infer: 'There is a fife on the mountain'. What we have in this rough account is a sequence of psychological events: a perception, a remembrance, a recognition, leading finally to an inferential cognition. These events belong to one and the same person, and are individuated both by ownership and temporal position. How can any such temporal sequence yield a logical rule? We can do that by: (i) replacing the particular cognizer by a variable and universal quantifier over it; (ii) retaining the appropriate relations of succession, but doing away with the actual temporal positions; (iii) identifying the cognitions involved by their contents and relative temporal positions; and (iv) requiring that all
PSYCHOLOGISM
7
cognitions figuring in the rule must have one and the same knower. Then we get a rule such as the following: For any knower 5, if 5 has a perceptual cognition Fx, and then remembers the rule 'Wherever F, there G', as instantiated in the uncontroversial case 0, and then perceives in x the same F as before but this time as figuring in the remembered rule, then 5 will have an inferential cognition of the form 'Ox' provided there is no relevant hindrance. 12
This is a rough account of how such an eidetic psychology of cognition would proceed. (4) Now as to the last Le., the transcendental turn. I shall try to present its nature in several steps. First, let us bear in mind the fact that we do not have a concept of the mental that is not an interpretation of it in the light of a theory: this is true as much of the view that it is an inner, cognitive, computational structure as of the view that the mental is the bodily of a certain sort. When the mental is taken for granted as being the psychological, one is interpreting it with a familiar psychological theory. One and the same mental act may be conceived as a private particular, as a psychological experience occurring within the interiority of an ego, or under any other description. It is not beside the point that Frege regarded psychologism - as Hans Sluga has shown - as an ally of naturalism. 13 The next step would consist in suspending these, or any such, interpretive frameworks. Now that we recognize that what passed for an ontologically self-evident thesis is in fact an interpretation, let us set them aside and salvage what is an indispensable phenomenal constraint which these interpretations had, in any case, to reckon with. These are, to my mind, the intentionality of the mental life: its directedness towards the world, its having a content or a structure, which is but the way the world is presented to it; its Sinn which is but the world's 'mode of presentation'. Couldn't the act's Sinn be a causal consequence of the world's acting upon the mind? If this were so, the 'transcendental' move would be stopped. But such a naturalism is not any longer open, now that we have set it aside as another of those interpretive frameworks. It is important, for my purpose, to emphasize that that was not an arbitrary decision. A naturalistic psychology simply cannot deliver the goods. The point has been well fonnulated by Fodor: 1.. a naturalistic psychology requires law-like relations between an organism and an object in its environment when one is thinking about the other, but for this one needs a description of the object such that the causal connection obtains in virtue of its satisfying that description. But such a description is not available until all the other sciences are complete. So let us begin with whatever description of the object the thinker has in mind- the Sinn or the noema. There is no access to the object per se save by means of such a Sinn. Once, then, the naturalistic framework: is 'put under brackets', one can still go on to talk the mentalistic language, but no longer in the sense of
8
J.N. MOHANTY
psychological theories. The mental acts are now taken precisely as they are experienced/performed, with their Sinn-structure, and as thus referring to their objects. One then begins to see that the act with its Sinn and the object as so described, Le., under such and such description, are but necessary correlates. Keeping this move in mind, we can take a look at such mentalistic assertions as these: Cantor: "Unter einer 'Mannigfaltigkeit' oder 'Menge' verstehe ich niimlich allgemein jedes Viele, welches sich als Eines denken lasst... "IS Or even Hilbert: "Die Grundidee meiner Beweistheorie ist nichts anderes als die Tiitigkeit unseres Verstandes zu beschreiben, ein Protokoll aber die Regeln auj'zunehmen, noch denen unser Denken tatsiichlich verjiihrt... "16 Or Brouwer: "This neo-intuitionism considers the intuition of two-oneness (the fundamental phenonmenon of the human intellect) as the basal intuition of mathematics which creates not only the numbers one and two, but also all finite ordinal numbers..."17 Or, following Brouwer, Heyting: "Mathematical objects are by their very nature dependent on human thought Their existence is guaranteed only insofar as they can be determined by thought"18 We find in these a mentalistic locution that intends to be 'transcendental', rather than psychologistic. Or, perhaps, like the Kantian locution of synthesis, it belongs to 'transcendental psychology', an Ersatz psychology, and so involves a sort of 'psychologism' that does not rob our cognitive accomplishments of their objectivity and intersubjectivity. What, iIi effect, I am proposing is briefly: what psychologism and anti-psychologism have in common is a certain conception of psychological discourse. It is an interpretation of the mental as a merely private particular which makes psychologism and so also its opponent's stance possible in the fIrst place. What we need to do is recognize this as an interpretation. A genuine overcoming of psychologism requires, not a rejection of this, or of any other, interpretive framework, but, fust of all, a recognition that an interpretive framework should not be construed as an ontology. The framework or frameworks need to be subjected to the famed epocM, and their origin - historical and genetic - sought for. The positive view that has guided this lecture, not defended within it, is that the logical and the mental cannot be radically sundered completely from each other- that our thoughts and theories are products of our acts of thinking, that in the long run all meaningfulness must lie in our being able to perform such operations. To delineate then the constitution of the logical, we need to look at the noetic operations that go into it. Such a constructivism, if psychologistic, should rather be called a sort of transcendental psychologism. The radical anti-psychologistic thinker is wrong in reifying thoughts into things, the psychologistic thinker erred in reducing thoughts into subjective ideas. Each saw one side of the truth: the noetic act and the objective Sinn are but two sides of a structure, separable only by abstraction.
PSYCHOLOGISM
9
Why is it that in spite of these precautions 'psychologism' is a trap into which one always tends to fall? By 'refuting' psychologism - as by refuting 'relativism' - one cannot altogether escape it. It is only by seeing through it, by following it as far as possible, by appropriating the truth in it, that we can hope to overcome it But the line that separates the eidetic from the particular, the transcendental from the empirical, is so thin, that 'the illusion of psychologism' will persist. Aren't such things as concepts, propositions, thoughts, and theories produced by the mind, in the mind and not out in the world. All depends upon how one construes 'production', 'being in', and ' minds'. I have suggested some constructions, but nothing guarantees that a transcendental philosopher will not fall into the trap. Basically, as Seebohm has recently insisted, it is due to the paradox of subjectivity which is both in the world and of the world or, as Foucault has put it, "an empirico-transcendental doublet" There are not two different sets of mental acts: one empirical, the other transcendental. One and the same act is both, depending upon how one looks at it, and what function one assigns to it
10 ENDNOTES 1. Franz Brentano, Psychologie von Empirischen Standpunkt (Leipzig: Duncker & Humblot, 1874), Vol. 2, p. 179 (my translation). 2. See: J.E. Erdmann, Grundriss der Geschichle der Philosophie, Vols. I and IT, (Berlin: W. Hertz, 1870); Vol. I, p. 636. Cf., Carl Stumpf, "Psychologie und Erkenntnistheorie," ABHANDLUNGEN DER AKADEMIE DER WISSENSCHAFfEN, Band XIX, II Abteihmg, p. 468. 3. Cf., IbilL, p. 468, fn. 1. 4. See: Philip Kitcher, "Frege's Epistemology," PIllLOSOPlllCAL REVIEW, LXXXXVIll, (1979), PP. 235-262. 5. See: Rudolf Camap, Logical Foundations of Probability, 2nd 00. (Chicago: University, 1971), pp. 39-47. 6. See: Karl Popper, The Logic of Scientifze Discovery (New York: Basic Books, 1959), p. 95. 7. See: Gottlob Frege, "Rezension von E. Husser!, Philosophie der Arithmetik," ZEITSCHRIFf FOR PIDLOSOPIDE UNO PIDLOSOPffiSCHE KRITIK, 103, (1894), pp. 313-332. 8. Theodor Lipps, "Zur 'Psychologie' Wld 'Philosophie'," in Psychologisclu! Untersuchungen, Band II, Heft 1, (Leipzig: Verlag von W. Engelmann, 1912), pp. 1-29, esp. 11. 9. Ernst Cassirer, Dos Erkenntnisproblem in der Philosophie und Wissenschaft der neueren Zeit, Band 2, 2nd ed. (Berlin: Bruno Casirer, 1922), p. 662. 10. Dreyfus quotes this from Hussed on p. 11 of his "Introduction" to: Hubert L. Dreyfus, 00., Husserl, Intentionality and Cognitive Science (Cambridge: MIT,
1984). 11. See: Edmund Husserl, Einleitung in die Logik und Erkenntnistheorie, 00., U. Melle, Husserliana, Band XXIV (The Hague: M. Nijhoff, 1984), chapter 4, esp. 25. 12. For more on this, see: IN. Mohanty, "Psychologism in Indian Logical Theory," in B.K. Matilal and J.L. Shaw, oos., Analytic Philosophy in Comparative Perspective (Dordrecht: Reidel, 1985), pp. 203-211. 13. See: Hans Sluga, Goulob Frege (London: Routledge and Kegan Paul, 1980). 14. Cf., e.g., Jerry Fodor, "Methodological Solipsism considered as a Research Strategy in Cognitive Psychology," reprinted in Dreyfus, 00., Husserl, Intentionality
and Cognitive Science. 15. Georg Cantor, Grundlagen emer allgemeinen Mannigfaltigkeitslehre, (Leipzig:
B.G. Teubner, 1883), p. 204. 16. David Hilbert, Die Grundlagen der Mathematik (Leipzig: B.G. Teubner, 1928), excerpt in o. Becker, 00., Grundlagen der Mathematik in geschichtlicher Entwicklung (Freiburg/Munchen: Verlag Karl Alber, 1954), p. 383. 17. L.E.I Brouwer, ''Intuitionism and Formalism," in Paul Benacerraf and Hilary Putnam, oos., Philosophy of Mathematics, Selected Readings, 2nd 00. (Cambridge: Cambridge University, 1983), esp. p. 80. 18. A. Heyting, ''The Intuitionist FOWldations of Mathematics," in Benacerraf and Putnam, eds., Philosophy of Mathematics, esp. p. 53.
THE MORE DANGEROUS DISEASE: TRANSCENDENTAL PSYCHOLOGISM, ANTHROPOLOGISM AND HISTORISM mOMAS M. SEEBOHM I. Introduction Hussed's arguments against psychologism l in the "Prolegomena" tell us that psychologism is simply a kind of relativism and that relativism can be refuted because it is self-contradictory. It has been said that the argument is spurious because the argument against relativism is spurious. 2 Indeed, the argument was later restated by Husser! himsele According to psychologism, empirical psychology as a science is the final arbitrator in all questions concerning validity claims of knowledge. Husser! argued that psychologism implies relativism. Relativism denies that scientific knowledge has universal objective validity. Hence psychologism denies the objective validity of its own judgments concerning validity claims of knowledge. Furthermore, already in the "Prolegomena" Husser! had pointed out that there is one type of relativism which cannot be refuted, viz., the individual relativism which claims the relativity of validity claims not for the species but for each individual.4 In connection with Husserl's later critique of Dilthey and world-view philosophy, another type of relativism entered in, namely, historism. 5 Husserl did not offer arguments against historism. He only emphasized the devastating consequences. It is obvious that consequentialist arguments cannot be directed against historism. Like individual relativism, historism is not committed to a belief in the objective validity of scientific knowledge. In general, it can be said that consequentialist arguments have some plausibility only with respect to a specific type of psychologism and not with respect to what Husser! later called 'anthropologism'. In the later writings Husser! considered the transcendental phenomenological reduction as the only remedy against psychologism in the broader sense. Historism is the main topic of this paper and is taken up in sections two and three. In the first section it is necessary, however, to discuss the different types of psychologism and/or anthropologism mentioned by Husserl in order to determine historism's systematic place among them.
12
THOMAS M. SEEBOHM
ll. Different Types of Psychologism The main topic of the Prolegomena is naturalistic psychologism, i.e., the psychologism which is developed in a psychology understood as a natural science. Since psychologism so understood reduces psychic phenomena to physical phenomena, Husserl calls this type of psychologism 'naturalism'. The relativism connected with it is the 'relativism of the species' which is also called 'anthropologism'. The term 'naturalism' also indicates that this psychologism is a counterpart to historism. The "Prolegomena" mentions other types of psychologism, fIrst of all transcendental psychologism. 6 With this tenn Husserl refers to Kant's attempted justification of the objective validity of principles a priori through the assumption of a transcendental subject. The frrst edition of the Logical Investigations characterized phenomenology as 'descriptive psychology'. The psychologists of the time frequently claimed that descriptive psychology is the main tool of epistemology.' Already in 1903 Husserl rejected this view. 8 He claimed though this was not made clear enough in the rust edition - that the descriptions in the Logical Investigations do not refer to empirical persons and psychic facts. Rather, as phenomenological they refer to the a priori essential structures of the experience in which logical objects are given. A description of inner experience and structures of intentionality is thus not a naturalistic psychology, although it can still be understood as an empirical discipline. Twenty years later Husserl wrote that the Logical Investigations, though not an empirical descriptive psychology, still represents a subtle psychologism. Indeed, any attempt to justify validity claims by means of a turn to the subject without explicitly perfonning the transcendental phenomenological reduction represents a transcendental psychologism of some sort.10 If in investigations of the structures of experience on1y the eidetic reduction were performed, then the investigation would not be phenomenology proper but rather eidetic psychology. Since in this case there would be no transcendental phenomenological reduction, the being of the subject would still be a being in the world. Thus eidetic psychology is later characterized as 'transcendental psychologism' or 'anthropologism'.11 It ought to be kept in mind that in comparison with the "Prolegomena" there is a significant shift in the meaning of these terms. When Husserl uses 'transcendental psychologism' against Kant in his later writings, he does not point to a mistake in the theory of the a priori, but to the constructions Kant had to use because he did not know the reduction. 12 'Anthropologism' no longer means that 'relativism of the (biological) species' which is the result of naturalistic psychologism. It is rather a term applied universally to all positions in which the subject, because of the missing reduction, is in the last instance understood as a being in the world, even when - as in the case of Kant - it is a world behind the world of experience, Le., a metaphysical world. Thus, transcendental psychologism has now to be
THE MORE DANGEROUS DISEASE
13
distinguished as a more universal and subtle mistake than logical psychologism. What is called 'transcendental psychologism' in the Logical Investigations is thus later characterized as a type of 'logical psychologism' is by no means the same as the transcendental psychologism mentioned in the Formal and Transcendental Logic. 13 In his critical remarks about Heidegger, Husserl tells us that though Heidegger offers an intentional psychology, he rejects the reduction and therefore does not have a real understanding of the constitution of the object and reality. His philosophy, like the philosophy of his forerunners Dilthey and Scheler, is thus an anthropologism connected to an objectivism and naturalism. From these comments it is obvious that the tenn 'naturalism' no longer refers to the natural sciences in any sense. Rather, in its most general sense it signifies the thesis that the existence of the world is not bracketed because the reduction is not perfonned. 14 The relation between descriptive and eidetic psychology, including their tendency toward either transcendental psychologism or transcendental phenomenology is by no means a negative relation of mutual exclusion. On the contrary, Husserl in his later writings asserts again and again that all the contents of eidetic psychology's descriptions are also without restriction valid in transcendental phenomenology. The only difference is the attitude with which these contents are considered. In eidetic psychology they are still considered with the mundane attitude. In transcendental phenomenology they are considered with the transcendental attitude, which is the result of the reduction. Furthermore, in addition to the Cartesian path there is a path to the transcendental phenomenological reduction via psychology. All that is necessary is a thorough critical phenomenological reflection on the presuppositions of the descriptions of eidetic psychology itself, i.e., an intentional analysis of the cognitive activities of eidetic psychology itself. The paragraphs which immediately precede the explication of this interplay of psychology and transcendental phenomenology in the Crisis 1s offer the formula which in part elucidates the principle of this explication. It is the formula of the paradox of subjectivity: On the one hand, the world and all its contents are given as the correlate of the consciousness which conceives them. On the other hand, we can think of the being of consciousness only as a being in the world The second part of the paradox explains why anthropologism is a necessary part of the self-apprehension of subjectivity. The 'where' of its being can be apprehended only in the world. This, in turn, explains why such a self-apprehension is necessarily infested with relativism. To be in the world is to be in one specific context in the world. Whatever is in the world in a context is in its being determined by this context But under this presupposition it must be admitted that the way in which consciousness has the world as its correlate is detennined by the specific context, and, therefore, is different in different contexts. The additional puzzle is that the way in which this determination is thought depends upon the basic structme of the world-view and the different ways in which determinations can be thought in different world-views.
md
14
THOMAS M. SEEBOHM
The paradox of subjectivity is helpful in some respects but leads to some other difficulties in the theory of the reduction. Given the paradox, the reduction can be understood as an arbitrary decision in favor of the frrst part of the paradox. It is not obvious that in the reduction a full account of the paradox can be given, and thus it is not obvious that the overcoming of relativism is more than a decision to deny the motives for accepting relativistic positions. 16 This question will not be discussed or answered in this paper, but an answer would presuppose the thorough analysis of the different types of anthropologism that is the topic of this paper. It follows from what has been said that historism is a kind of anthropologism or transcendental psychologism. As such, it is the counterpart of naturalistic psychologism or naturalism in the sense of Husserl's early writings. Some hints for a further analysis can be gained from Husserl's judgments about Dilthey. The judgment is negative and critical in connection with the early critique of historism. 17 It is positive in later writings. Dilthey's descriptive psychology is there praised as a forerunner of a phenomenological psychology.18 Husserl' s theory of the constitution of the cultural world seems to be heavily influenced by Dilthey's.19 Most interesting for the present purpose, however, is that Husserl in manuscripts20 sketched out the idea of a path to the transcendental phenomenological reduction via the human sciences. According to Husserl the human sciences in their research employ a specific epocM. If this epocM is radicalized, it leads into the phenomenological epocM and with it into the reduction. A universal human science under a radicalized epoc he has as its correlate the contents of all cultures, past and present, and is, hence, coextensive with transcendental phenomenology. A universal human science in Dilthey's sense is necessarily a universal historical human science. The conception of history which emerged in the nineteenth century in the historical school and which culminated in Dilthey belongs to a universal human science. It is, however, this conception of history out of which historism was developed. There is, hence a parallel: the universal human science is related to historism on the one hand and to transcendental phenomenology on the other in the same way that descriptive eidetic psychology is related either to transcendental psychology or tnmscendental phenomenology. Both are instances of the paradox of subjectivity. The world in which consciousness has its being is in the frrst case the 'historical world' of the universal human science, and the contexts are the historical contexts of different cultures and epochs. An analysis of historism presupposes, therefore, a study of the genesis of the historical world which is the correlate of historical research in the late nineteenth and early twentieth centuries. Not very much can be gained from Husserl in this respect In connection with studies on the origin of geometry, he made some possibly helpful remarks about the significance of writing for all higher activities of the intellect,21 but he never reflected on either the specific methods of the human sciences and history, their correlate, viz., the historical world, or the genesis of both. This task would
--------------------------------
------------
THE MORE DANGEROUS DISEASE
15
be parallel to Husserl's analysis of the genesis of science out of the
II
I
prescientific realm and by no means - as the following considerations will show - included in or dependent on the genesis of science. Following Dilthey22 one can assume that the genesis of the conception of history of the universal human science is closely connected to, and to a very high degree identical with, the development of hermeneutics qua methodology of the interpretation of texts. It presupposes the historical-philological method. Hermeneutics as a methodology and art has its own genesis. The second section of this paper will point out that this genesis is guided by a certain necessity. The analysis of historism is given the paradox of subjectivity - itself a necessarily generated mcxle of self-apprehension of consciousness. Two remarks about the mode of presentation are in place. Husserl presented his description of the genesis of science in the Crisis in the form of a narrative. Section II of this paper uses the same procedure for the explication of the genesis of what will be called 'hermeneutical consciousness'. The task of narrative is, however, to reveal a necessity of the development. Two characteristics of this 'necessity' can be mentioned. (1) It should become obvious that a certain level N of the development can be reached only if the preceding levels are given in the order in which they occurred. (2) Given a certain level, and if a reflection which is guided by a certain interest in the unity of the literary tradition takes place, only a limited - perhaps only one - possibility is left. Whether it actually takes place or not is, however, a contingency. Those who are uncomfortable with the assumption of a material a priori may claim that it cannot have the character of an absolute objective validity because such a 'feeling of necessity' is at best a result of the situation of those who reflect upon the development. Indeed, one conclusion of this paper will be that no 'refutation' of such historistic assumptions is possible. Secondly, readers will be frustrated because the paper talks about rules and methods without explicating them. I have given such explications elsewhere. 23 What the paper analyzes are reflections about a methodological practice that had hitherto been taken for granted. The main question is how such reflections changed the principles and not what kind of rules could be derived for practice from the principles.
ITI. The Genesis of the Historical World View The expression 'type of hermeneutical consciousness' will refer for the purposes of this essay to the specific structure of the understanding of the Ipast, the tradition, and its significance, a structure of understanding which is Ilcharacteristic of a type of consciousness. The term 'hermeneutical' has a 'Imethodological meaning in the old conception of henneneutics, but in this \essay the term will have broader use. Its use will also, however, be
16
THOMAS M. SEEBOHM
narrower than the more recent meaning in philosophy according to which consciousness, Dasein, is throughout hermeneutical. Given the definition which has been chosen here, it can be said that historism is a very complex type of hermeneutical consciousness which has itself a 'history' in which less complex types of hermeneutical consciousness can be found. The analysis of this genesis is the task of this section. The mediation of the past into the present via traces, i.e., fixed life-expressions, is a necessary component of hermeneutical consciousness in general. Fixed life expressions are tools, buildings, objects used in cultsbut also memorized and habitualized patterns of actions, including all kinds of speech acts. 24 If written discourse is not among the ensemble of fixed life-expressions, the hermeneutical consciousness and the culture connected with it is often characterized as the culture of the 'savage mind'- a term which should be used without all the pejorative connotations which it sometimes has had. It is, however, doubtful, whether the past can have for such a hermeneutical consciousness the meaning of a realm with identifiable objects.~ The following analysis will restrict itself, therefore, to the types of hermeneutical consciousness which relate to the past first of all via written discourse. This restriction does not imply: (a) that the historical world-view itself is restricted to this domain; or (b) that the paradox of subjectivity can be fully understood without referring to structures which belong in the most archaic forms of hermeneutical consciousness. Expositions of the term 'hermeneutics' begin very often with an interpretation of the use of 'hermeneuein' in Plato's lon.7f, However, in the classical Attic period there was no 'art of interpretation', and even when it was developed in the Hellenistic period. ' hermeneia' was not a technical term connected with it A more appropriate anticipation of this art is rather to be found in the various hints in Plato's Epinomis. Those who with the aid of divination practice the ' hermeneutike techne' to interpret the poets, the laws and signs and oracles are there called the ' exegetai'.XI Methodical rules for such interpreting were not known at Plato's time. Even a grammar was still missing and with it the rules of 'grammatical exegesis', which presuppose the art of grammar. By contrast. the later employment of this term indicates that what was at stake was already the interpretation of written discourse. However, the art of grammar was not only the art of writing correctly and of analysing a text according to given grammatical rules. It also included all rules and methodical viewpoints for the exegesis of texts. The mastering and practicing of the art of grammar led to philology. Philology meant for the Hellenistic age - as well as for the later humanistic tradition - universal erudition, i.e., the results of the practice of the art of grammar. It included all wisdom which the tradition had to offer: divine revelation, poetry, philosophy, the sciences etc. 28 Since the art of grammar also included the art of criticism in all fields, philology was also able to answer the question 'what is truth' in all fields. Finally, the true philologists also had to be a masters in the field of rhetoric, and the true rhetorician had to be a perfect philologist According to Quintilian and
THE MORE DANGEROUS DISEASE
17
others, rhetoric is only by implication the art of delivering effective speeches. Essentially it is the art of applying the knowledge supplied by philological education to newly given concrete situations. In this sense it is also the art of the invention of new knowledge. 29 Three main characteristics of this philological-rhetorical type of hermeneutical consciousness must be emphasized for the purposes of this investigation. (1) The literary tradition represents for it a harmonious unity of a system of truths, values, duties and beauty, an ideal of humanity - the same which later guided the humanists of the Renaissance - which is of significance for any present time and which provides the framework for the education of future generations. (2) Interpretation, explication, and application are for this type of hermeneutical consciousness an inseparable unity. (3) It has a conception of history which is quite different from the one which provides the medium of modern historism.30 In the early periods of classical antiquity, 'historia' meant the investigation of facts of all kinds, and later it was a term also for a literary genre, namely, the one in which reports about the results of such investigations were given. There was in the art of grammar a level which corresponded precisely to this meaning of 'historia', namely the 'explanation of facts and words' in the so-called historical exegesis. In the nineteenth century it was still recognized as a relevant activity of the henneneutical method.31 Later on, 'facts' were often understood primarily as deeds. The Romans coined the well-known expression 'res gestae', things done. The word 'factum' corresponds in Latin to this narrower interpretation of 'fact'. Seen from the rhetorical point of view mentioned above, such reports about res gestae were used as examples of deeds and acts, which could be applied to present situations. It has been shown that this concept of 'historia' still dominated the early Christian literature. 32 Christianity introduced some changes in hermeneutical consciousness (which will be discussed immediately), but it cannot - as is usually thought - be credited with invention of an idea which was later of great significance for the development of hermeneutical consciousness, namely, the idea of universal history. The genesis of this idea has other roots. Roman historians collected histories about the deeds of the Romans in chronological order. But Rome turned out to be much more than one polis among others. Its fate became the fate of the known world. It was Vergil who created the idea of a development which had a goal-directed history leading to the Roman empire as an empire of peace. This idea had an impact on the work of Christian chronologists and historians much later (circa 500 A.D.) and only after a serious change within the hermeneutical consciousness had taken place in the New Testament. The criticism of the Hellenistic philosophers who said that philology is only a collection of facts and as such has no access to truth and universality was a repetition of Aristotle's description of 'historia' .33 It was repeated again and again but had no real impact on the development of hermeneutical consciousness. The critique of the skeptics 'against the
18
THOMAS M. SEEBOHM
grammarians', however, was an expression of a change. They pointed out that 'the art of grammar is circular in all of its rules- an early preview of the hermeneutical circle. Of even greater significance was their downgrading of the tradition as a whole into a source 'diaphonies', contmdictions. The skeptical method showed that the tradition was no longer the general source of wisdom and truth but rather throughout contradictory.34 The skepticism of this type has nothing in common with the skepticism of historism. The fonner type of skepticism is ahistorical, and its attitude is the counterpart of the attitude of the philologist It was, fwthermore, not only a philosophical position but also an expression of an experience of henneneutical consciousness. One had not to be a skeptic in order to know that there were contmdictions in the tradition. The most significant diaphony was the contradiction of the conception of reason and virtue developed in philosophy since the classical Attic period with both the myths of the poets/prophets and archaic lawgiving. In order to save the unity of the tradition, the Stoics developed the method which was later called 'higher' hermeneutics, i.e., allegorical interpretation. The myths were said to have not only a literal meaning, but also a higher meaning in which the truths of reason are revealed in the fonn of an allegory. The Stoics also invented a theory of culture which explained why those who created the myth had to tell the truth in allegories.3' A modification of the new technique of interpretation and its justification was later adopted by Jewish Hellenistic philosophers for the interpretation of the Old Testament36 The problem of diaphonies created in a literary tradition swfaced also in the New Testament. In Luke 24:27 we read that Jesus interpreted (dihermeneusen) Moses and the Prophets in the light of his own deeds and teachings. This principle of the typological interpretation became a general principle of the interpretation of the Old Testament. The present, later period of a tradition is the archetype which contains more truth (sensus plenior) than the earlier period, or ektype. The ektype can be fully understood only in the light of the archetype. The hermeneutical consciousness of the early Christian church was in every respect 'critical'. It was no longer possible to gloss over the diaphonies that emerged between its own and other traditions. The rabbinic tmdition of interpreting the Old Testament had frrst to be 'cut off as a false tradition. In the next step the gnosis, i.e., the attempt to harmonize the Hellenistic tradition with the Christian tradition via allegorical interpretation, had to be 'cut off' as false from the tradition of the true patristic literature. In general, other literary traditions including the schools of the philosophers along with theological teachings not accepted by church- were considered as sactae, heresies cut off from the true catholic tradition.37 The demand of the church that 'the worldly power prosecute what was cut off from its tradition is an immediate consequence of the type of hermeneutical consciousness represented by the early church. It is a type which is not
the
THE MORE DANGEROUS DISEASE
19
restricted to the Christian tradition. With modifications it can be found elsewhere, e.g., in the development of Islam. Even a type of hermeneutical consciousness which tries to save its unity by continually cutting off false traditions creates in itself diaphonies- if only because the excluded ideas reenter the tradition in disguise. The problem of the diaphonies occurs now in: (1) the repeated attempts to prove the concordance of the tradition with itself against the apparent sic et non, Le., the systematically collected contradictions; and (2) the revival of higher hermeneutics and allegorical interpretation, though the church condemned all attempts to use such techniques to harmonize the New Testament with elements of the Hellenistic tradition and admitted only the typological interpretation of which it had to apply the methods of higher interpretation to its own tradition. The tension between the literal meaning of the Scriptures and the dogmatic teachings, moral teachings, and institutional structure of the church was obvious. Allegorical interpretation of the New Testament was necessary for the justification of the dogmas. Other types of higher interpretation - the moral and the anogogical - were added. 38 Together with this new type of hermeneutical consciousness a new concept of history developed Its basic structure is given with Augustin's distinction between the sacred history and the history of the gentiles. 39 In the Byznatine monk chronicles- bu't also in the self-understanding of the Holy Roman Empire the sacred history has its 'worldly' aspect in its connection with Rome.40 The sacred history is in its structure not a historical development in the modem sense. It is - to use Heidegger's term - an ekstatic history in which anything has significance not 'in itself but rather with respect to the crucial events of the Creation, the Fall, the Incarnation of the Word, and the coming of the Kingdom of God. Though it is 'universal', it has nothing in common with the nineteenth century historians' idea of a 'universal history'. The history of the gentiles also exhibits no development or any significance for itself. The old conception of history dominates it. Each story told serves as an example of either God's punishment of evil deeds or reward for virtuous deeds, and this is the only reason for remembering such stories. The modem conception of history depends upon a radical turn in hermeneutical consciousness, but this turn was prepared by certain intellectual developments during the Middle Ages. Thomas Aquinas had already recommended restricting higher hermeneutics to the cases in which 'the Scriptures themselves, at the level of their liteml sense, ask for such interpretations.41 Likewise the Franciscans, who dominated the movement of the devotio moderna, favoured the literal sense.42 Nevertheless, it must be recognized that Luther's principle scriptura sancta est sui ipsius interpres initiated a revolution in hermeneutical consciousness. If the Scriptures can be understood out of themselves, then such an understanding can be used as the basis of a critique of the tradition of the interpretation of the Scriptures. Since this tradition had - as a unity of interpretation and application continuously developed new contents which cannot be found in the
20
THOMAS M. SEEBOHM
Scriptures, the critique was a cntlque of the tradition in principle. The tradition itself was the result of an ongoing distinction between the true and the many false traditions. The new principle turns this distinction against the 'true' tradition itself to the extent to which it is possible to declare that this tradition is 'false' .43 The situation created by this turn was not stable. An interpretation of the Scriptures which turned against the old tradition of interpretation necessarily created its own tradition. The Reformation soon had its own 'fathers'. The underlying principle of the new type of hermeneutic consciousness implies, however, the possibility that it can be turned against this new tradition. Seen from a formal point of view, this procedure can be iterated indefinitely and was in fact iterated by the steadily increasing number of Protestant sects. Thus a situation was created which could be used for the development of a counterargument. The Council of Trent decided that no understanding of the Scriptures without the mediation of tradition is possible. The Council argued that, to begin with, a tradition of translation is necessary. Hence the question had to arise: which tradition is the true tradition? The Council claimed that an answer could not be found in an interpretation of the Scriptures, but rather had to be found in the nature of the tradition. The answer was: the tradition which can prove that it is connected without interruption to the Scriptures is the one which is true and which alone can deliver a true interpretation of the Scriptures" The task of the Protestant theologians was to prove against this formal argument that it is indeed possible to find criteria and methodical rules which allow us to understand the Scriptures without, and if necessary, against the tradition. The principle they developed was the fIrst canon of hermeneutics. This canon says that the text can be understood as it was understood by the contemporary addressee. The claim that a text has to be understood according to the original intention of the author is another and, if properly understood, by no means 'psychological' - fonnulation of this canon. The methodologically pure formulation would be: the text has to be understood out of its own context and not out of the context of the interpreter. The Latin fonnulation is sensus non est inferrendus sed efferendus.44 With respect to the Scriptures, the understanding of the contemporary addressee could be reconstructed with the aid of the art of grammar and its resulting philosophy. The practice of this art had survived as a subculture during the Middle Ages and surfaced again in the humanism of the Renaissance. 4s The Renaissance as well involved a hermeneutical consciousness which declared a tradition of interpretation to be false and recommended an immediate return to the sources. However, because the texts in question were not texts which grounded the basic system of truths, values, and noons, the Renaissance - in comparison with the Reformation was in itself of only secondary significance.
THE MORE DANGEROUS DISEASE
21
The alliance of the Reformation with the Renaissance had serious consequences. It is certainly true that the Protestant theologians and humanists did not separate interpretation, explication, and application.46 It is also true that humanists and some - though not all! - Protestant theologians propagated an ideal of erudition and education in which a harmonious unity of the literary traditions of Christianity and classical antiquity was presupposed. But it is also true that the further development showed with logical necessity that such assumptions were not compatible with a rigorous application of the fast canon. Although Protestant Theologians employed the art of grammar, they did not - like the Hellenistic and humanistic philologists - regard its result, wisdom or philologia, as the truth. Rather, they regarded a specific text, the New Testament, as the source of truth, and they used the art of grammar as well as philologia as the instrument to understand it out of its own context. A few other texts within this context could be considered as sources of truth as well, but some - especially those which occurred within the pagan and rabbinic traditions - had to be considered as false. Consequently, the texts emerging in these latter traditions had to lack any application. Thus for these texts the unity of interpretation and explication with application was dissolved. The method of understanding a text out of its own context is, as such, neutral with respect to the question of application. But it does demand a lot of work. Indeed, the method can stretch indefmitely through the universe of texts and contexts and the meanings they reveal. The interest of the interpreter determines the selection of groups of texts to be subjected to methodical interpretation. This interest and the texts selected can be deliberately changed. In the last instance, this interest of the interpreter is determined out of his/her context. But that does not undermine the methodical attitude as such. Furthermore, it should not be forgotten that the interest can be a pure professional interest, and fmally that it can be mere curiosity about what strange things people believed and thought to be true. The rigorous application of the method showed the diaphonies in the literary tradition in all their harshness. That the ' spirit of the Greeks' and the 'spirit of Christianity' are irreconcilable was later the experience of Hegel, HOlderlin, and their contemporaries. Indeed, a new level of the hermeneutical consciousness' experience of the thoroughly contradictory character of the literary tradition is the presupposition of Hegel's attempt to think the whole of the development of mind or spirit as a movement through contradictions.47 However, for the newly developed methodology in henneneutics, the attempt was soon itself a text which has to be understood out of its context48 The philologist and, as will be shown in the next section, the histo~ practiced what Husser} later called a universal epocM in which the question of truth, value judgments and application were bracketed. What was and is at stake from the professional point of view is the reconstruction of the original meaning of a text in its context This reconstruction has to
22
THOMAS M. SEEBOHM
eliminate the non-understanding and misunderstanding which arises in any interpretive tradition that does not employ such an epoche. The professional attitude is - in this sense which is defmed by the first canon - 'objective' and 'contemplative'.49 It registers the truth claims made in texts, but it does not, as such, share them. It understands and withholds any judgment about agreement or disagreement. The given account of the genesis of this type of hermeneutical consciousness entails the rejection of some currently fashionable opinions about the nature of the philological-historical method developed by the hermeneuticists of the nineteenth century. It is claimed that the ideal of objectivity in interpretation depends upon the Cartesian idea of scientific method as a method of discovering truth, and it is then objected that this idea does not fit the real task of interpretation. The Cartesian idea of method is indeed a verificationist idea, but for the methodologists of the nineteenth century the method was not a method of discovering truth, but a method of avoiding error.so The hermeneuticists of the nineteenth century obviously had an idea of method which would today be called 'falsificationist'. Furthermore, they knew the limits of the professional methodological attitude. They were well aware that with respect to works of genius - what we today call 'eminent texts' - the question of truth cannot be separated from the task of interpreting them.~l It is sometimes also asserted that the philological-historical method and its claim for objectivity borrows something form the principles determining the methods of the natuml sciences. However, on the basis of the narrative given above, it can be said that, on the contrary, the idea that the truth of the Scriptures can be understood outside of and against the tradition provided a major justification of the attempts to gain a similar access to nature. This does not mean that the methods in the natural and human sciences have no common roots,S2 but these roots go much fwilier back. It also does not mean that a certain alliance between historism and naturalism is impossible- such an alliance will be considered in the next section. But the opinion that scientism or naturalism is a genetic presupposition of the modern conception of history and historism cannot be maintained. Finally, it is frequently claimed that the Enlightenment's prejudice against prejudice is responsible for the development of the hermeneutical method of the nineteenth century. However, it was the above delineated turn of the hermeneutical consciousness at the time of the Reformation which made the Enlightenment's prejudice against prejudice and tradition possible- and with it the development of the hermeneutical method.
IV. The Modern Conception of History and Historism The philological-historical method has as its primary objects the texts and their meanings. The meaning of a text is to be understood out of its own
THE MORE DANGEROUS DISEASE
23
context This implies that every text is in some sense determined by its context It is possible to have a fairly precise concept of what a context is. There is a quasi-temporal order in the process of determination of any context A text is determined by the texts to which it refers explicitly or implicitly, and in the realm of its efficient history it determines the texts which refer to it. Thus contexts are webs and dimensions of developments. Already from this point of view, in which only the interpretation of meaning of texts is the concern, it can be said that the method is intrinsically historical. The historical connectednessS3 of texts is not restricted to historical textual connectedness. (I mention in passing that the use of language is another connectedness which has as well its historical dimension. 54) The most relevant historical dimension is, of course, given in texts with reports about events happening in their present context or in the context of their past. But even texts which do not report events at all always report one event: their being written by one or more known or unknown authors. Thus texts themselves are fixed life-expressions which belong to a second order context of events and deeds, a context which is reported by some texts. It is this specific second order context in which the author occurs. Though it is possible to neglect the author at the level of historical textual connectedness, he cannot be banished from this second order level of the historical effective connectedness to which some texts refer and to which all texts belong. History written under the modem conception of history has the character of a construct produced on the basis of a complex web of fixed life-expressions and knowledge of the natural environment in which they have been produced. Historical writings are explicationsss of a certain epoch or development in a culture at a temporal and/or geographical distance from the explicating historian.56 There are two ways of constructing such writings. One way involves an immediate application in the explication. It presupposes the assumption of the wholeness of the historical development, which in some sense culminates in the present or in a historical period that will supposedly follow the present. Hence, each period receives its significance from the outside. Dozens of such secularized universal histories of salvation or damnation have been created since the beginning of the nineteenth century.57 They are themselves an object for a reflective hermeneutical consciousness. Since they constitute conflicting traditions of interpreting history as a whole, the history of the modern literary tradition repeats the Hellenistic pattern in which an increase of diaphonies leads natumlly to skepticism. The other approach rejects from the outset such 'metaphysical' interpretations of history.sa It is grounded in an application of the first canon to historical epochs. Every culture has to be understood out of itseJ.F9 and has its significance primarily for itself, and not for us as mediated through a world historical process in which it has a certain function. Therefore, the attitude of the historian is objective, contemplative, and tolerant with respect
24
THOMAS M. SEEBOHM
to foreign systems of values and truth claims. But that means that the historian also has a relativistic point of view. Each value system and system of truth claims is regarded as having validity only with respect to its historical context Historical positivism and historical relativism thus belong together and are both constitutive for historism. The level of the historical connectedness of deeds and events is presented to the present not only through texts but also through fixed life expressions of other kinds such as buildings, tools, cult objects, works of art, etc. These 'silent' traces or witnesses are the object of archeological hermeneutics. Such traces are primarily those which can be understood as being connected with 'inscription' that are the products of intentional acts of authors. But such traces are always connected with traces from the unintentional animalic life of the authors and with traces of the natural environment to which this life as animalic life belongs. Efficient history is a web of events in an environment. In this web these events are related as causes and effects, which are analyzed in historical explanations. The difficult question of the logical structure of historical explanations cannot be discussed here.60 It must suffice to note that they are similar to the explanations which we give in the realm of the life-world and are strongly analogous to the explanations given by courts of justice. (fhe original meaning of 'causa' was 'guilt'.) Furthermore, actions can here be explained by referring to 'motives' which can be understood, and the understanding implies here the understanding of values and nonns as well. However, the fIrst canon requires that such an understanding be exercised from the historist's point of view. Only the norms and values which belong to the context of an action are valid viewpoints for explanations. An explanation is falsified if it can be shown that it uses viewpoints foreign to the context Because historical explication is not restricted to the hermeneutical interpretation of meaning, historism can be combined with naturalism. One possible mode of combining historism and naturalism is to graft scientific explanations onto historical explanations in the same way in which they can be grafted upon explanations in our life-world. Historians have always exploited this possibility not only to explain actions, but also to secure data concerning traces. History reaches down to the level of cultures of the 'savage mind' which have not left traces in written discourse. (Indeed, in our and other literary traditions there are histories and even traditions of cultural contacts with such cultures.) The task of understanding such cultures out of their own context is most difficult The impact of scientifically secured traces and explanations of such cultures increases with the temporal distance. The investigations finally lead back to a grey zone in which there are fewer and fewer fIXed life-expressions or traces which can be distinguished as in some sense 'human' from those traces left by animalic life. The hermeneutical consciousness which is generated by this combination of historism and naturalism views the past of humanity and its significance for the present and future. However, with the growing temporal
THE MORE DANGEROUS DISEASE
25
distance it is less and less chamcteristic for the 'human condition' in geneml. Science has thus moved mankind out of the center of the universe. Historism combined with naturalism teaches: (8) that mankind is 8 very brief episode in the history of nature; (b) that the character of mankind shows 8 high degree of plasticity; and, as 8 conclusion (c) that this character will develop in the future - if it survives at all - into something which is quite different from what it is now. However, there is 8 puzzle connected with this type of hermeneutical consciousness, 8 puzzle which does not invalidate this consciousness but rather indicates bow it stands under the paradox of subjectivity. Science, seen from the viewpoint of history, has had its own history and its own historical epochs with diaphonical bUth claims. Moreover, especially in the light of the historico-naturalistic type of hermeneutical consciousness, it is not reasonable to expect that in two thousand years, science - if there is something like that at all - will be in principle similar to our science. We cannot even expect that our science will be regarded as something better than myth. Thus, the attempts in the second half of the twentieth century to develop 8 theory of science on the basis of the history of science have had a consequence which surprised its developers, namely, that the claim for the validity of scientific methods, 'paradigms', came to be regarded as historically relative.'· However, the practitioners of 8 paradigm in any historical epoch cannot know this paradigm as historically relative unless they adopt 8 hermeneutical consciousness which has the first canon of hermeneutics as its guiding principle. The puzzle is that the historist contradicts himself: he claims on the one hand that all truth claims have validity only in their contexts, but also on the other hand that historical investigations of such truth claims and their contexts have some objective validity, i.e., that the fust canon and all methodical moves following from it create some type of unrestricted, context-free, objective validity. Gadamer has exploited this puzzle to cast serious doubts on the possibility of historical objectivity.'2 He claims that the modern conception of philology and history somehow implies that the interpreter has attained a standpoint 'outside of history' and can reflect at a neutral level of pure contemplation. However, in order to refute the possibility of any such standpoint, Gadamer attacks the fust canon in the form of the principle of the 'original intention of the author' and the 'understanding of the contemporary addressee' .'3 In its place he offers the thesis that the validity claims for an interpretation depend on the tradition in which the interpretation is made. He claims that every interpretation of the past is itself valid only for the tradition in which it is given. From these claims it follows that there can be no difference between correct and incorrect interpretations in geneml. Interpretations are always only different.64 Gadamer's thesis that no interpretation can sepamte itself from its tradition has a prima facie family resemblance to the position of the Council of Trent The difference, however, is that the Council of Trent gave a criterion of the we tradition. Gadamer gives no such criterion. Thus, his
26
THOMAS M. SEEBOHM
abandonment of the first canon only leads to a radicalized historism for which historical relativism is itself a position that has only historical validity. However, no practitioner of such a radicalized historism can even have historically objective knowledge of historical relativism as historically relative unless he or she adopts a hermeneutical consciousness based on the ftrSt canon. Furthermore, if such a practitioner of radicalized historism simply decides not to attempt to obtain historically objective knowledge of historical relativism as historically relative, then this decision must itself be considered as a disguised renewed longing for ahistoricaJ objective knowledge. The ftrSt canon is thus the presupposition not only of historism's objectivism, but also of its relativism. As explicated above, without the ftrSt canon there is only the mutual exclusive relationship of traditions to be false- with the well known political consequences included. The attempt to overcome historism's objectivism through the rejection of the ftrSt canon eliminates the possibility of both relativism and the tolerance of other traditions. Conversely, the retention of this tolerance presupposes the retention not only of historism's relativism, but also of its objectivism, i.e., it presupposes a hermeneutical consciousness which adopts the fast canon. Just as Gadamer's attempt to exploit this puzzle to cast doubt on historical objectivity is self-defeating, so any utilization of this puzzle which attempts to 'refute' the fast canon, and so historical relativism, on behalf of an ahistorical objectivism must also fail. To repeat: such an ahistorical objectivism had as its product intolerant traditions. The price for the tolerance that the fast canon brings is a historical, and so historically relative, objectivism. What has been said should by no means be understood as a refutation of radicalized historism or effective-historical consciousness." As long as it does not deny historical objectivism it cannot be refuted. But the relation of the two intertwined levels of historism is itself in need of an explication. This explication can be given with the aid of the paradox of subjectivity. Indeed in this specific instance of the paradox the 'world' is the 'historical world', the world of history, the world given through an open multiplicitly of past and possible future consciousness which is under the universal epocM of the human sciences as Husserl characterized it, but such a consciousness is the hermeneutical consciousness of the first level of historism. As such, it is nothing but the choice of the ftrSt part of the paradox of subjectivity and therefore has the close affmity to the transcendental-phenomenological reduction which Husserl ascribed to it. The historism of the second level - the radicalized historism of historicity and effective-historical consciousness - is only a reduction of the historist's view to the second part of the paradox. It asserts that the historist's point of view can be given only as a part of the historical world itself and is, therefore, determined by its own context and its own tradition. The task of this paper has been to explicate how historism is itself a specific and most radical instance of the paradox of subjectivity. Only a few hints about the consequences of this explication can be given. The fast
THE MORE DANGEROUS DISEASE
27
consequence is pragmatic. Since the paradox of subjectivity belongs to the structure of the self-apprehension of consciousness, all research which is in some sense concerned about subjectivity and consciousness is self-referential and hence infested with the paradox. It is pointless to exploit this paradox to develop so-called 'refutations' of naturalism, psychologism, and historism. What has been characterized as 'the historico-naturalistic approach' has collected most of the knowledge that has resulted from mankind's experience of itself and its environment- though this knowledge is by no means ordered in any kind of 'system' which modem philosophers from Descartes to Hegel wanted. This approach is also open in the highest degree to new experiences. Any putatively competing henneneutical consciousness most likely already belongs to a stage of the history of this approach. To go back to earlier stages would involve both a scientific and moral regression. The second consequence concerns the possibility of a prima philosophia, a frrst philosophy. Some may be tempted to try to overcome this paradox through a return to or renewal of 'dialectical interpretation' . Intensive studies of this potion have convinced me that its only outcome is an ordering of phenomena on the basis of logical presuppositions which are given the present state of the art - unsatisfactory in every respect. Rather, it is the late writings of Husser! which clearly determine what phenomenological research has to investigate in order to achieve a better understanding of the self-giveness of consciousness, namely, passive synthesis and especially the problems of temporality and intersubjectivity in order to formulate a new transcendental aesthetic.
28 ENDNOTES James H. Wilkinson gave highly valuable advice concerning the presentation of the structure and style on the material in this essay. 1. See: Edmund Husserl, "Prolegomena to Pure Logic," in Logical Investigations (Halle: M. Niemeyer, 1900), Vol. L 34-38. In order to refer to all English or Gennan edition of Hussed's writings I cite - wherever this is possible the paragraph number. I also give references to the volwnes of Husserliana. For the "Prolegomena" this reference is Husserliana, XVllI, 1975. If it is necessary to refer to pages, the reference will be given to an English translation. 2. See: Dagfinn F011esdal, Husserl und Frege (Oslo: I Kommisjon hos Aschehoug, 1958), pp. 36 ff. Cf., IN. Mohanty, Husserl and Frege (Bloomington: Indiana University, 1982), pp. 20 ff. 3. Cf., EdmlDld Husserl, "Philosophy as a Rigorous Science," in Q. Lauer, 00., Phenomenology and the Crisis of Philosophy (New York: Harper & Row, 1965), especially pp. 92-105. 4. See: Hussed, "Prolegomena," 35. 5. See: HusserI, "Philosophy as a Rigorous Science," p. 129. Lauer and others often translate 'Historismus' as 'historicism'. I follow Karl Popper, The Poverty of Historicism (London: Routlege & Kegan Paul, 1960) in distinguishing between historicism and historism. Popper understands by historicism a theory - like those of Hegel and Marx - which gives an account of universal history as a whole. From the historist's 'positivistic' point of view - as well as from Popper's - historicism would be rejected as bad metaphysics. Cf., note 59 below. 6. See: HusserI, "Prolegomena," 28, n. 38, 58. The German word is 'Transzendentalpsychologismus' . 7. Cr., H. Spiegelberg, The Phenomenological Movement (The Hague: Martinus Nijhoff, 1960), Vol. I, especially p. 57 (about Stumpf). 8. See: Edmund HusserI, "Preface to the Prolegomena," in Edmund HusserI, Logical Investigations, Vol. I, ed. IN. Findlay (London: Routledge & Kegan Paul, 1970), p. 47. 9. See: EdmlBld Husserl, Formal and Transcendental Logic. Husserliana XVII, 56,99. 10. For an account with more references, see: Thomas M. Seebohm, Die Bedingungen der Moglichkeit der TranszendenJalphilosophie (Bonn: Bouvier, 1962), pp. 156-159. 11. Husserl's notes about Heidegger are published in: Alwin Diemer, Edmund Husserl. Versuch einer systematischen Darstellung seiner Phanomenologie (Meisenheim: Hain, 1956), pp. 29-30. 12. Cr., Seebohm.., Bedingungen, sec. 6-8 for references. 13. See: Husserl, Formal and Transcendental Logic, 56,99. 14. See: Diemer, Edmund Husserl, pp. 29-30. 15. See: Husserl, Phenomenology and the Crisis of Philosophy, 53, 55. 16. For a criticism of this kin~ cf., Gerhard Funke, "Practical Reason in Kant and Husser!," in T.M. Seebohm and J.1. Kockelmans, eds., Kant and Phenomenology (Washington: Center for Advanced Research in Phenomenology and University Press of America, 1984), pp. 1-29. 17. See: Husserl, "Philosophy as a Rigorous Science," 122 ff. 18. See: Edmund HusserI, Phiinom£nologische Psychologie. Husserliana. IX (The Hague: Martinus Nijhoff, 1968), "Introduction," 1,2, Beilage IJI.
THE MORE DANGEROUS DISEASE
29
19. See: Edmund Husser!, ldi!en II, Husserliana, W (The Hague: Martinus ·Nijhoff, 1952), Dritter Abschnitt, "Die Konstitution dec geistigen Welt" 20. See: Ibid., Beilage IV, 311-315; Beilage xn, 11-12. See also: Edmund Husserl, Erne Philosophie, Vol~ II, HusserliaM, VII (The Hague: Martinus Nijhoff, 1959), Beilage XXVI, 458; and EdmlDld Husserl, Phiinomenologische Psychologie, Beilage V, 376-379. 21. See: Jacques Derrida, Edmu.nd Husserl's Origin of Geometry: An Introduction, trans. J.P. Leavey (New York: Nicolas Hays, 1977). 22. Cf., Thomas M. Seebohm, ''Boeckh and Dilthey," in 1. N. Mohanty, ed, Phenomenology and tlu! Human Sciences (The Hague: Martinus Nijhoff, 1979). Cf., Wilhelm Dilthey, Selected Writings, ed. H.P. Rickman (Cambridge: Cambridge University, 1976), especially pp. 247-263, ''The Development of Hermeneutics." 23. See: Thomas M. Seebohm, Zw Kritilc di!r hermeneutischen Vernunft (Bonn: Bouvier, 1972). Cf., also the essay mentioned in the prior footnote. 24. Cf., Dilthey, Selected Writings, especially pp. 247-263.
25. See: Thomas M. Seebohm, "The Significance of Written Discourse for Henneneutics," in 1.1. Pilotta, eel., Interpersonal Communication (Washington: The Center for Advanced Research in Phenomenology and U.P. of America, 1982), pp. 141-160. 26. Plato, Ion, 534c. 21. Plato, Epinomis, 957c. 28. Cf., Gradmann, "Grammatik," in August Friedrich von Pauly and Georg Wissowa, eds., Realencyclopijdie der Idassischen Altertwnswissenschoften Vol. VII (Stuttgart: Druckenmuller, 1912). Material for the following considerations is also taken from: Heimich Kuch, PhiloioglU. Untersuchlmgen eines Wortes von seinen ersten Anfdngen in der Tradition bis ZIU ersten iberlieferten lexikalischen Festlegung (Berlin: Akademie-Verlag, 1965. 29. See: Quintilian, The Institutio Oratoria of Quintilian, Vol. I, ed. H.E. Butler (London: Loeb, 1960), chapter 1 of book 2; and Ibid., Vol. IV, chapter 1 of book
12.
30. GA. Press, The Development of 1M Idea of History in Antiquity, (Kingston: McGill-Queen's University, 1982). Its discussion of 'mstoria' is the counterpart to the discussion of 'philologia' in the above cited book of Kuch. 31. August Boec~ Enzyklopiidie IU&d Methodenlehre der Philogischen Wissenschoften, ed. Bratuscheck (Leipzig: B.G. Teubner, 1877), pp. 93 ff.; translated in abridged form by 101m Paul Pritchard as On Interpretation and Criticism (Norman: University of Oklahoma, 1968), pp. 76 ff. Cf., also Thomas Bitt, Hermeneutik IU&d Kritik, Handbuch der Klassischen Altertwnswissenschoften, ed. I. MUller (MuncheD, C.H. Beck), Vol. I., Abtlg. 3. 32. Press, TIu! Development of the Idea of History in Antiquity, pp. 123-124 in the conclusion. 33. Kuch, Philologus, pp. 75, 100 ff. 34. See: Sextus Empiricus, Against the Professors, 00. R.G. Bury (London: Loeb, 1949), Vol. IV. See also: Sextus Empiricus, Outlines of Pyrrhonism (London: Loeb, 1955), Vol. I, especially chs. 4-6. 35. Cf., Karl Reinhard, Poseidonios (M1lnchen: C.H. Beck, 1921); and Leo Strauss, PhUosophie und Gesetz (Berlin: Schocken, 1935), pp. 87 ff., esp. p. 118. The second level myth tells us that in the beginning all people were good and followed the will of God. Later with the increased number of men evil spread out, and finally only very few just people were left. To these God gave the art of writing and they wrote down what is good and true. But they could not address the
30
THOMAS M. SEEBOHM
unjust and uneducated rabble directly. They had to tell the truth in the disguise of myths in order to be understood. 36. Eusebius of Caesarea, Praeparatio Evangelica, Lib. XIII, caput 11, 12, Migne PO XXI, pp. 1095 ff. reports that Aristobul created the new version of the myth according to which the time of Noah and his children was the time of old when all people were just The last just men were the prophets, and the Greeks learned all their wisdom from the prophets or from the Egyptians, who in tum learned it from the prophets. This tradition was still alive in the eighteenth century. Cf., Johann Jakob Brucker, Historia Critica philisophiae a mundi incunabilis ad nostram ~tam dedu£ta, Vol. I (Leipzig: Breitkipf), pp. 1741-44. 37. For a formal defInition of 'hairesis', cf., Johannes Darnascenus, Dialectica, Migne PO vI. 94, 657. It is already the leading principle of Eusebius of Caesarea, Historia Ecclesiae, Migne PO XX. 38. The new crisis announces itself as early as Abelard's four rules for the treatment of contradictions, the sic et non in the tradition. Cf., M. Grabmann, Geschicte der Scholastischen Methode (Freiburg im Breisgaie: Herdersche Verlagshandlung, 1911), Vol. II, pp. 200 ff. The main source for the new types of higher interpretation is Thomas Aquinas, Summa Theologiae, Pars I, Quaestio I, Art. 10. Cf., E. v. Dobschtitz, Vom vierfochen Schriftsinn. Geschichte einer Theorie, Beitrage ZlU Kirchengeschichte (Leipzig, 1921), pp. 1-13. 39. See: Press, The Develop11U!nt of the Idea of History in Antiquity, Section V, pp. 89-119 for more material on this topic. 40. It is the conception of the 'Holy Roman Empire' first developed in the Byzantine chronicles written by monks after Justini~ e.g., by Georgios Hannatolos. 41. See: Thomas Aquinas, Summa Theologiae, Pars I, Quaestio 1, Art 10. 42. See, for instance: Nikolaus de Lyra, Biblia sacra cum glossa ordinaria ... et Postilia Lirani Franciscani. Opera et studio Theologorum Duacensium, T. IV, Antverpriae apud Joannem Meurium, Anno M DC XXXIV. 43. For the following passage, cf., Dilthey, Selected Writings. 44. About the sources, cf., Emilio Betti, Allgemeine Auslegungslehre als Methodik der Geisteswissenschaften (fUbingen: Mohr, 1967), (abridged German version of the Italian original Teoria genoale della Interpretazione (Milano: Giuffre, 1955). For a discussion of the justifications of the validity of this canon, cf., the material mentioned in notes 25, 23 above. 45. Cf., E.R. Curtius, Ewopiiische Literalur und Lateinisches Mittelalter, 3rd ed. (Munich: Bern, 1962), pp. 47 ff. about Martianus Capella. More material is given in: Kuch, Philologus, pp. 128 ff. 46. Cf., Hans-Georg Gadamer, Truth and Method (New York: Seabury, 1975), p. 162 about Chladenius. 47. Hegel's Phenomenology of the Mind must be understood as a significant document of a stage in the development of hermeneutical consciousness in the sense in which this term has been introduced in this essay. 48. See: Dilthey, Selected Writings, pp. 229, 253. Cf., Gadarner, Truth and Method, pp. 144, 173-187 about the 'dialectic' between Hegel and the so-called 'historical school' in Germany. 49. See: Gadarner, Truth and Met~ p. 187. 50. Cf., Boeckh, Enzyklopiidie IUId Methodlehre der Philogischen Wissenschaften, pp. 52 ff., 43 ff. (Pritchard 35 ff.) 51. See: Boeckh, Enzyldopiidie IUId Methodenlehre tIer Philogischen Wissenschaften, p. 241. Cf., Seebohm, "Boeckh and Dilthey," p. 24 about page 157 of the English translation by Pritchard
THE MORE DANGEROUS DISEASE
31
52. The common root is medieval Averroism which: (1) is the consequence of a cultural contact; (2) represented an Aristotelian naturalism; and fmally (3) developed the theory of a double truth- the truth of faith which depends on will and the scientific truth which depends on reason. This Latin version does not correspond to the original Arabic and Jewish tradition in which 'faith' was for the rabble and 'science' for the few wise men. Cf., note 35 above about the roots of this version. 53. The concept stems from Dilthey; see, for example: Diithy, Selected Writings,
p.229. 54. About the 'lower levels' of hermeneutics, cf., Seebohm, "Boeckh and Dilthey," pp. 91-92, 97-98. 55. Recent literature restricts itself in many cases to the discussion of interpretation and application. Traditional methodical hermeneutics and its sources reveal a more complex pattern in which 'explication' plays a significant role, cf., Ibid., p. 90. 56. Cf., Dilthey, Selected Writings, p. 230. 57. Hegel as well as Marx - and nowadays Heidegger as well as Derrida - give us accounts of history as a whole and its meaning. 58. Positivistic historism's use of the word 'metaphysical' is taken over from Kant's concept of special metaphysics and means an acco\Dlt of history as a whole, i.e., what was called 'historicism' by Popper, cf., note 5 above. In this sense, Derrida is 'metaphysical'. 59. See: Gadamer, Truth and Met~ p. 178 about Ranke. 60. See: Thomas M. Seebohm, "Historische Kausalerklirung," in Gunter Posch, ed, Kausalitiit, Neue Texte (Stuttgart: Reclam, 1981), pp. 260-288. 61. Thomas S. Kuhn, The Structure of Scientific Revolution (Chicago: Chicago University, 1962) faced the problem of relativism in the theory of science after his attempt to ground theory of science in historical research! 62. Namely: Gadamer, Truth and Method. 63. See: Ibid., pp. 336, 356. 64. See: Ibid., p. 264. 65. See: Ibid., p. 305.
ENTAILMENTS FROM 'NATURALISM = PHENOMENOLOGY' JOSEPH MARGOLIS
I. Introduction We intend, here, to support a heterodox thesis and, by extension from it, a whole series of further heterodoxies. The thesis maintains that 'naturalism = phenomenology', and the argument in its favor claims that it is well-nigh ineluctable. What follows from it is quite remarkable, so its defense should be examined with some care. The relationship between naturalism and phenomenology has, of course, been memorably mapped by Edmund Husserl in a way completely opposed to the equation given, mapped also with a sense of genuine foreboding that would have regarded its defense as little more than the mad courting of conceptual disaster. Nevertheless, the equation is defensible- and strategically important; and, it may be claimed, the entire movement of contemporary Western philosophy is effectively committed (reasonably committed) to its support, including the work of a good number of phenomenologists. If it were read in a suitably comprehensive way - as addressing the entire nmge of first-order knowledge: roughly, knowledge of the real world that includes physical nature and human minds and human culture; or, possibly, what is true, where what is 'true' may not (on a theory) accord with what is merely 'real', may, instead, address another 'world' entirely (the 'worlds' of logic or geometry or arithmetic or the 'meal' world Husserl dabbled in 1) - then a very large number of powerful theorems may be shown to follow directly from our equation. At the risk of excessive compression, then, we may effect an important economy.. The argument to be considered goes more or less as follows. If naturalism = phenomenology, then (1) ontological and epistemological questions cannot be disjoined, (2) frrst- and second-order cognitive aptitudes cannot be disjoined, and (3) all forms of cognitive privilege foundationalism or logocentrism - must be abandoned;2 and if (1), (2), and (3) hold, then (4) knowledge of reality and truth cannot be disjoined.
'NATURALISM
=PHENOMENOLOGY'
33
ll. First-order and Second-order Powers If we admit that the real world is cognizable as it is independent of human inquiry, or if we admit that human agents are capable of knowing what is true of that world or of a world altogether independent of their minds and of the real or natural world itself, or if we admit that human agents are capable of knowledge of the real structures and properties of a world they themselves constitute, that is nevertheless not merely an artifact of some particular inquiry or flight of imagination - these alternatives being hardly equivalent to one another - then there are only a limited number of conceptual strategies by which to ensure, in part or whole, that human agents can achieve the knowledge they would thus pursue. That is, there are only a limited number of strategies if the ones required are supposed to be cognitively sufficient in first-order terms. Call any such strategy privileged, in the sense that a second-order assurance of such frrst-order knowledge is taken (somehow) to be entailed in or abstractable from a given ftrst-order cognitive competence. When the source of assurance lies in a reflexively detectable feature of fIrSt-order experience or cogniscent reports or the like, the legitimating argument may be said to be foundationalist; and when it does not depend on such a feature but can be counted on nevertheless, the legitimating argument may be said to be logocentric. 3 The logocentric may, then, be said to include the foundationalist commitment and more. Foundationalist and logocentric arguments may be as arbitrary as you please, but in principle they suppose that our fIrSt-order inquiries somehow provide a basis for our second-order assurance; they suppose that, distributively, fIrSt-order claims of determinate sorts either ensure knowledge directly or provide the best, or at least a viable and privileged, source of or constraint on the body of knowledge. W.V. Quine, for example, in the very process of repudiating the pertinence of intentional complications with regard to the prospects of any rigorous empirical science,4 clearly proceeds in a logocentric (but not in a foundationalist) way- in a way, in fact, that has dominated most of Anglo-American theorizing about knowledge and science for nearly forty years. It would certainly not suit Quine to acknowledge only that his own regimented brand of extensionalism was entitled to an epistemological inning. Quine clearly intends to discharge intentionality as utterly irrelevant epistemologically: "If we are limning the true and ultimate structure of reality," he says - and he means what he says - "the canonical scheme for us is the austere scheme that knows no quotation but direct quotation and no propositional attitudes but only the physical constitution and behavior of organisms."' It is possible that Quine is registering here an altogether undefended prejudice. But it is difficult to believe that reading in the face of his sustained interest in the puzzles of knowledge and science, his persistent efforts over an entire career to entrench his extensionalism, and the ve adhered to his apparent seriousness with which his man allow
34
JOSEPH MARGOLIS
instruction. For the moment, our interest does not concern such maneuvers and countermaneuvers. It rests rather with the plain fact that the requisite fast-order competence is taken to be a natural capacity of the members of human communities: to form, codify, discipline, enlarge, confmn, even improve those relatively systematic collections of claims that they are pleased to treat as the core of the various sciences. Our concern, we may say, is almost nosological. The would-be legitimation of fust-order knowledge (or of somces of knowledge) is, trivially, a second-order matter. ·When the required assurance is taken to instantiate a specifically cognitive aptitude of its own, legitimation is (as remarked) either foundationalist or logocentric. When it is not - that is, when the legitimation of frrst-order cognitive competence is itself denied any special competence (when it is denied cognitive 'privilege', when it neither applies distributively nor relies on fast-order evidence when applied holistically) - it functions (we may say) in the pragmatist manner. It is conceivable (but indefensible) to favor [rrst-order cognitive competence and to deny the pertinence or eligibility of second-order legitimation. For such a maneuver either installs, logocentrically again, some fIrst-order aptitude (and so violates the pretended repudiation of second-order arguments) or it abandons every rntional concern regarding the compared relevance and power of competing would-be resources. This second policy (segregating frrst- and second-order inquiries) has also come to be called pragmatism- largely through the recent efforts of Richard Rorty. Pragmatism in this second sense calls for an end to traditional philosophy, on the grounds that any legitimating inquiry cannot fail to be logocentric. Pragmatism in the frrst sense - in denying privilege but pursuing matters of legitimation - cannot but tend in the direction of holism, internalism, historicism, relativism, and an emphasis on the role of praxis. (We need not return to the details of that story.') What concerns us rather is that both logocentric and pragmatist forms of legitimation are construed naturalistically, that is, as involving no cognitive powers other than those entailed or embedded or exercised in, or assignable to, the native capacities of human investigators. (They involve no cognitively greater titans or gods, no revelations.) Needless to say, the first-order powers admitted as cognitively apt on pragmatist grounds are apt only in the sense of salience, not of privilegethat is, only in the sense of what, provisionally, perspectivally, reflexively, in a way internal to the very achievement to be accounted for, appear to be the most promising candidates (for the time being) for the explananda required. They are subject to revision or replacement for all sorts of reasons having to do with how salience itself may be altered over the range of historically shifting experience and within the equilibrative, diachronically deployed pressures of trying to match our picture of the real world and our picture of our cognizing powers apt for producing that picture. The importance of admitting salience is just that it precludes privilege and acknowledges the profound transience and contingent stability of what, in
'NATURALISM
=PHENOMENOLOGY'
35
terms of the implicit consensus of actual societies pursuing inquiries of high discipline, appears to them to be their science and cognitive power. In short, the validity of our second-order legitimations is an artifact of the relative stability of our informal saliences, which are themselves, quite frankly, affected by our own ongoing efforts at science and legitimation. So doubts regarding the force of legitimation cannot escape applying with equal relevance to the ftrSt-order standing of our scientific accomplishments. By the same token, endorsing the latter as at least sufficiently reliable and promising so that we may press on with our fIrst-order work is tantamount to endorsing the legitimacy of legitimation. In a word, legitimation need not (and does not here) presuppose privilege or risk its constitutive and regulative function in abandoning privilege. Once we see matters in this light, it proves surprisingly easy to demonstrate that there is no principled division or disjunction of an epistemically pertinent sort between what may be legitimated as naturalism and as phenomenology. This goes entirely contrary, of course, to the main theme of Husserl's master project, which, in a very real sense, could be said to be focused on ensuring the most profound disjunction possible between naturalism and phenomenology. To press the 'correction' against Husserl is to side with a good many who have been perceptibly influenced by Husserl's own attempt to correct naturalism's pretensions of privilege. It is to side, to a considerable extent, with the programmatic intentions of Heidegger, Merleau-Ponty, Sartre, Derrida, and (stretching things some) Gadamer- insofar as those theorists mean to pursue phenomenological corrections of a naive naturalism; but it is also to insist on the obvious convergence (not necessarily intended) between their (quite varied) programs and such others (also eschewing logocentric advantage to one degree or another) that have evolved, for example, from American pragmatist sources (from Peirce and Dewey and James and Quine), from Continental Hegelian and Marxist and Frankfurt Critical sources, and, of course, from Nietzsche and his most recent adherents (Derrida and Foucault, in particular). For the essential criticism made of Husserl - by this time, almost banal though it is assuredly still effective is simply that, in correcting the pretension of privilege among the naturalists, Husserl failed to admit (indeed, resisted to a preposterous extent) the plain fact that the validity of that correction could not but extend to the further (cognate) pretensions of phenomenology as well. Husserl apparently 'did not understand' what he was doing in demonstrating that the naturalists did not understand the limitations of their own undertaking.7 This is surely the neatly focused message of Merleau-Ponty's 'adjustment' of Husserl's theme. (We may allow it to stand proxy for the convergent intent of the rather different figures just mentioned.) The perceiving mind [says Merleau-Ponty] is an incarnated mind I have tried, first of all, to re-establish the roots of the mind in its body and in its world, going against doctrines which treat perception as a simple result of the action of
36
JOSEPH MARGOLIS external things on OlD' body as well as against those which insist on the autonomy of consciousness [that is against empiricism and rationalism as forms of naturalism or objectivism]. These philosophies commonly forget - in favor of a pure exteriority or of a pure interiority - the insertion of the mind in corporality, the ambiguous relation which we entertain with our body and, correlatively, with p:rceived things.'
It is a message reinforced in the opening lines of Phenomenology of
Perception: What is phenomenology? ... Phenomenology is the study of essences; and according to it, all problems amount to rmding defmitions of essences: the essence of perception, or the essence of consciousness, for example. But phenomenology is also a philosophy which puts essences back into existence, and does not expect to arrive at an understand.ing of man and the world from any starting point other than that of their 'facticity'. It is a transcendental philosophy which places in abeyance the assertions arising out of the natural attitude, the better to understand them; but it is also a philosophy for which the world is always 'already there' before reflection begins- as an inalienable presence; and all its efforts are concentrated upon re-achieving a direct and primitive contact with the world, and endowing that contact with a philosophical status.'
Merleau-Ponty's intended corrective exhibits the necessary tact and subtlety of functioning as a corrective within the pale of Husserl's own deeply admired endeavor. It may be instructive, therefore - for a reason which will gradually become clear - to juxtapose without elaborate preparation a rather famous if somewhat primitive (but characteristic) remark of John Dewey's (already refined, in 1938, from earlier speculations going back before the tum of the century) that clearly converge (innocently) with the deeper intent of Merleau-Ponty's pronouncement but without having to work through the unfortunate disjunction Husserl imposed on his followers. It shows at a stroke the obvious, the marvelously simple, sense (within the American pragmatist movement) in which naturalism could be said to be 'phenomenologically' constrained (natively) without ever having to pass through an official correction. In context, the passage in question was intended by Dewey to concede the preformational and incarnate biological and social world within which human inquiry cannot but proceed, the horizonal and contingent nature of its every effo~ the place it affords withal for a viable logic and science, and (perhaps most important) the respect in which standard (otherwise logocentric) oppositions between cognizing subject and cognized world are already treated by Dewey as no more than abstractly posited within what (mythically) precedes such oppositions. The linkage with Heidegger's myth of Sein and Dasein is plain enough; it supplies the reason for linking Dewey and Merleau-Ponty as well. tO The remark itself is drawn from Dewey's Logic:
'NATURALISM = PHENOMENOLOGY'
37
Inquiry is the controlled or directed transformation of an indeterminate situation into one that is so determinate in its constituent distinctions and relations as to convert the elements of the original situation into a unified whole. 11
Dewey's essential point - one must bear in mind how early relative to phenomenology it was made - is that "the unsettled or indetenninate situation might have been called a problematic situation [and that, existentially, it is] precognitive."12 The full import of these remarks remains still to be drawn out: they may actually be a little surprising. But for the moment it is perhaps enough to exhibit the unintended convergence between a 'phenomenologized' naturalism and a 'natumlized' phenomenology. For, by the partly mythic, partly naturalized device of the 'problematic situation', Dewey genuinely intends to risk the fixity of every would-be structure of an 'independent' world (including whatever, of would-be inquiring agents, is similarly 'independent' of or prior to their explicit inquiry); and, in doing that, he has surely avoided and completely bypassed the unnecessary traffic of the whole of official phenomenology. Clearly, Merleau-Ponty means to offset the idealist and solipsistic possibilities that, though never intended by Husserl, invariably beckon from some point deep within Husserl's theory. Nevertheless, it is Merleau-Ponty himself, who, 'adjusting' Husserl, still insists - ambiguously - that "the world is always 'already there' before reflection begins-- as an inalienable presence": so that all the efforts of philosophy "are concentrated upon re-achieving a direct and primitive contact with the world, and endowing that contact with a philosophical [that is, with an epistemically pertinent] status." Merleau-Ponty may have meant this in the "pragmatist" sense (or in a sense very much like it- holistically and mythically). That certainly would not be an unfair reading of a related, late manuscript entry: ttWe will not admit a preconstituted world, a logic, except for having seen them arise from our experience of brute being, which is as it were the umbilical cord of our knowledge and the source of meaning for US."13 SO construed, the passage represents Merleau-Ponty's thoroughly un-privileged (mythic) speculation about the conceptual connection between the 'presence' of an unnamed and unnameable brute world - 'there' prior to our objectifying inquiries - and the valid work of those same inquiries, also (ambiguously now, as the activity of 'mobile bodies') precognitively 'located' in that same brute world But the question still nags, whether and to what extent the 'lived body', the 'embodied subject', the original percipient source of science, is, as the very center of consciousness (not in the idealist's manner) also and for 'that reason the 'source' of the cognizable world that surrounds it "The whole universe of science," Merleau-Ponty remarks, "is built upon the world as directly experienced, and if we want to subject science itself to rigorous scrutiny and arrive at a precise assessment of its meaning and scope, we must begin by reawakening the basic experience of the world of which science is the second-order expression." "I am [he adds] the absolute
38
JOSEPH MARGOLIS
source, my existence does not stem from my antecedents, from my physical and social environmenl..[it is through that original] consciousness, through which from the outset a world forms itself round me and begins to exist for me. To return to things themselves is to return to that world which precedes knowledge, of which knowledge always speaks, and in relation to which every scientific schematization is an abstract and derivative language.... tt14 Here, it is difficult to avoid the conclusion that, in an impossibly gymnastic way, the 'second-order' pronouncements of science are somehow to be reconciled (in a cognitively pertinent way) with the precognitive but (seemingly) distributed encounters of 'primary' perception. Otherwise, why say that science 'speaks' of the things of the 'world which precedes knowledge', and why mention the semiotized 'relation' holding between the two? It is true that Merteau-Ponty insists that "our relation to the [originally perceived] world is not that of a thinker to an object of thought, tt and yet he explicitly says (and means) that the analysis of primary perception is essential to psychology (possibly to all the sciences, though, here, he hedges - inconsistently - under questioning).1' There is reason to believe that Merleau-Ponty did not resolve this difficulty- could not (and, in a way, did not want to). FOf, in insisting on the continuity of science and philosophy and on the continuity of science and primary perception, he characterizes science as a second-order coding and systematization of whatever is given (albeit precognitively) in primary perception; and that makes no sense unless what is thus given (in a 'fust-order' respect) bears in a cognitively pertinent (recoverable) way on our science. It is partly for this reason that he assigns intentionality to the 'lived body' functioning below the level of explicit consciousness. He opposes innatism, but already in the primary biology of the human organism Merleau-Ponty finds a need for a somewhat equivocal vocabulary that is not yet fully 'mental' (or behavioral in the mentally informed sense) but that still requires a mode of perception and experience that may serve to fix the 'frrst-order' intentional life that discursive consciousness and science 'refer' to in their 'second-order' way. So he says, at the very close of The Structure of Behavior, "The natural 'thing', the organism, the behavior of others and my own behavior exist only by their meaning; but this meaning which springs forth in them is not yet a Kantian object; the intentional life which constitutes them is not yet a representation; and the 'comprehension' which gives access to them is not yet an intellection. "16 This may serve to clarify the essential equivocation (not inadvertence) of one of Merleau-Ponty's characteristic pronouncements: "Our own body is in the world as the heart is in the organism: it keeps the visible spectacle constantly alive, it breathes life into it and sustains it inwardly, and with it forms a system. "17 Nevertheless, the existentialized phenomenology is meant to support an opposition to every form of cognitive privilege-- what we have unceremoniously associated with Dewey's seemingly thinner theme. The irony is that it is Dewey's pronouncement that is the clearer of the two and that helps us to extricate Merleau-Ponty from his own dilemma: here,
'NATURALISM
=PHENOMENOLOGY'
39
naturalism hurries to the rescue of phenomenology. The intentionality of the lived body need not be compromised; it must, however, either be raised to a mythic pronouncement (like Heidegger's or Foucault's) or else it must descend to the pronouncements of a tempered naturalism 'critically' informed by that myth (as in Dewey's solution). But in favoring the better alternative, we must not devalue or defuse Merleau-Ponty's equivocation. On the one hand, he asserts that "my existence does not stem from my anteeendents, from my physical and social environment" and, on the other, that "Our own body is in the world as the heart is in the organism"; on the one hand, he affirms our presence in a "world which precedes knowledge" and, on the other, speaking of his notorious glass cube, that it is "by conceiving my body itself as a mobile object [in that world] that I am able to to interpret perceptual appearance and construct the cube as it truly is."IB Viewed in the manner favored, phenomenology is a naturalism shorn of its own logocentric pretensions and attentive (especially) to the preformative conditions under which all inquiry proceeds; and, similarly viewed, naturalism is simply phenomenologized. The intended reconciliation is meant to be comprehensive: there is nothing left that, in principle (though not, of course, in determinate detail), needs to be added in order to ensure the viability of the frrst- and second-order epistemic concerns with which we began. Deconstruction, therefore, is merely the ultimately attenuated instruction of any such reconciliation: that is, the instruction that every particular such reconciliation is. subject to the future supplement of a now-unfathomable future such adjustment; that the conceptual relation between the one and the other is and must remain opaque yet effective; that there is no sense at all to pretending to be able to project mtional sunnises of continuity, progress, verisimilitude, or totalizing based on cognitive sources external to the contingent work of just such surmises; and that every such surmise is, if internal to such work, fatally 'infected' by an absence of privilege. In effect, these considerations mark the almost unnoticed common error of such varied projects as those of Peirce, Quine, Chisholm, Popper, Gadamer, Habennas, Apel- and Husser!.1'
III. Constraints on Phenomenology Now, then, what are the benefits of all this? Husserl had resisted the simple conclusion that naturalism = phenomenology, for the double reason that he (rightly) saw that naturalism could claim no cognitive privilege and that he (wrongly) supposed that phenomenology could. There are really three essential strategies on which Hussed and Merleau-Ponty diverge in this regard (taking Merleau-Ponty as proxy for all the reconciling maneuvers intended by our partial list of favorable movements and figures provided just above). These are centered on: (1) cognitive 'voices' or sources; (2)
\intentiOnal or propositional content; and (3) dependencY__Ol'_~()Il~_gency
on.
40
JOSEPH MARGOLIS
preformational processes. A glance at the passages cited from Merleau-Ponty confrrms that, for Merleau-Ponty, it is the same incarnated and 'enworlded' mind that pursues natumlized perceptual truths and processes them phenomenologically; and that that dual enterprise is 'factically' encumbered by the same Lebenswelt within which humans live and exercise these abilities. Merleau-Ponty, therefore, opposes any principled or privileged disjunction of cognitive voices between the naturalist and the phenomenologist (1), although, as we have seen, he introduces a further, uneasy complication. He also opposes any pretense that phenomenological inquiry can escape the constructive role of the existential preconditions of 'primary' or 'lived perception' (3).20 Furthennore, once his stand on (1) and (3) is declared, it becomes clear that any further distinction regarding the special content of this or that inquiry can neither override the constraints just admitted regarding (1) and (3) nor escape whatever strictures are there imposed. Hence, for Merleau-Ponty, any distinction regarding (2) - even the phenomenological pursuit of essences - will be entirely neutral as regards the disjunction between naturalism and phenomenology. (To be sure, naturalism in Dewey's sense cannot be characterized as even interested in Merleau-Ponty's revision of the Husserlian pursuit of essences, the skillful work of eidetic variations. But the point of linking Dewey and Merleau-Ponty is not to erase such differences but to provide a fair sense in which the special work of the one could be quite easily fitted to the conceptual orientation of the other. In any case, the point of distinction (2) is just that it raises no epistemological issue of its own that would affect our fmdings regarding (1) and (3).) Husserl implicitly opposes all three of Merleau-Ponty's stands and for the same reason. This may be seen, without actually explicating Husserl's entire theory, by juxtaposing two brief remarks of his on the role of the Lebenswelt. First, Husserl says, "Things, objects (always understood purely in the sense of the life-world), are 'given' as being valid for us in each case (in some mode or other of ontic certainty) but in principle only in such a way that we are conscious of them as things or objects within the world-horizon. Each one is something, 'something of the world of which we are constantly conscious as a horizon."21 But then, secondly, he says: Now, how can the pregivenness of the life-world become a universal subject of investigation in its own right? Clearly, only through' a total change of the natural attitude, such that we no longer live, as heretofore, as human beings within natural existence, constantly effecting the validity of the pregiven world; rather, we must constantly deny ourselves this. Only in this way can we arrive at the transformed and novel subject of investigation, "pregivenness of the world as such": the world purely and exclusively as - and in respect of how - it has meaning and ontic validity, and continually attains these in new forms, in our conscious life. Only thus can we study what the world is as the ground-validity for natural life, with all its projects and undertakings, an~ correlatively, what natural life and its subjectivity ultimately are, i.e., pmely as the subjectivity which functions here in affecting validity. The life which effects world-validity in
'NATURALISM
= PHENOMENOLOGY'
41
natural world-life does not permit of being studied from the attitude of natmal world-life. What is required, th~ is a total transfonnation of [cognitive] attitude, a completely unique, universal epoche.22
Segregating the cognitive voices of naturalism and phenomenology (1) and declaring the one but not the other subject to the preformative influence of the Lebenswelt (3), Husserl moves effortlessly to affirm that the pertinent cognitive pronouncements of particular (naturalistic and phenomenological) sciences are of profoundly different epistemic sorts (2). Husserl's position, then~ depends entirely on his insistence on the cognitive privilege of transcendental phenomenology. He agrees (in effect) with Merleau-Ponty that the epistemic fortunes of (2) depend entirely on the resolution of the matters sorted under (1) and (3); but, for that very reason, Husserl and Merleau-Ponty part company. This bears directly, of course, on Husserl's fundamental criticism of Galileo's geometry and on his own attempt to ground geometry in some transcendentally apodictic, 'primary' or "original meaning-giving achievement which, as idealization practiced on the original ground of all theoretical and practical life - the immediately intuited world (and here especially the empirically intuited world of bodies) - resulted in the geometrical ideal constructions. tt23 It may be, as Merleau-Ponty enthusiastically affmns, that Husserl was seeking "a way between psychology and philosophy- a mode of thinking... which would be neither eternal and without root in the present nor a mere event destined to be replaced by another event tomorrow, and consequently deprived of any intrinsic value." But the characterization is more apt autobiographically. Hence, the following manifesto, possibly one of the most explicit Merleau-Ponty ever offered, points to the essential division between the two: One may say indeed that psychological knowledge is reflection [Merleau-Ponty had just announced that 'Reflection is historicity'] but that it is at the same time an experience. According to the phenomenologist (Hussed), it is a 'material a priori'. Psychological reflection is a 'constatation' (a finding). Its task is to discover the meaning of behavior through an effective contact with my own behavior and that of others. Phenomenological psychology is therefore a search for the essence, or meaning, but not apart from the facts. Finally this essence is accessible only in and through the individual situation in which it appears. When pushed to the limit, eidetic psychology becomes analytic-existential. 2A
So seen, the 'primacy' of perception is meant to account (holistically) for the (mythic) origination of a phenomenologized naturalism within the space of an existential, precognitive presence in the world, and to permit and enable an (internalist) recovery of whatever, reflectively, can appear as the (existentially) thus-encumbered essences or meanings within the flux of experience. Ultimately, therefore, Merleau-Ponty refuses Husserl's "total change of the natural attitude."
42
JOSEPH MARGOLIS
What, however, is an unexpected bonus resulting from this otherwise curious review of the convergence and opposition of naturalism and phenomenology is the illumination (and resolution) of Husserl's famous diatribe against psychologism. Given (1)-(3) and the divergence between Husserl and Merleau-Ponty (and the mild promise of Dewey's seemingly thinner naturalism), the puzzle of psychologism cannot but prove shallower (in one sense) than Husserl supposed and incapable (in another) of a resolution favorable to his own extreme philosophical program (or to Frege's). Furthermore, even that profound issue proves, quite transparently, to be only one among a large family of expanding questions that the reconciliation of naturalism and phenomenology entails. These matters should prove instructive. Before turning to them, we need to remind ourselves that, in Husserl, one fmds all the principal forms of cognitive privilege, both foundationalist and logocentric. These include Husserl's reliance on: (a) the apodictic (self-disclosure yielding unconditional cognitive certainty); (b) the originary (assurance regarding the ultimate source on which correspondence, essences, the epistemic link between the possible and the actual and between the necessary and the contingent depend); and (c) the totalized (achievement of the conceptual closure within which every considered possibility is related to the inclusive system of all such possibilities). Ultimately, for Husserl, these three are only different aspects of one and the same transcendental source of epistemic assurance.2S It makes no difference that Husserl concedes (in a sense) the provisional, step-by-step, 'infinite' exercise of what is required by his own notion of transcendental phenomenology. No such concession ever blunted, for Husserl, the full accessibility and necessity of the transcendental leap itself: Instead of this universal abstention in individual steps [in effect, the naturalized approximation of the epoche Husserl requires], a completely different sort of universal epoche if possible [he insists], namely, one which puts out of action, with one blow, the total performance running through the whole of natural world-life and through the whole network (whether concealed or open) of validities- precisely that total performance which, as the coherent 'natural attitude', makes up 'simple' 'straightforward' on-going life.... An attitude is arrived at which is above the pregivenness of the validity of the world, above the infmite complex whereby, in concealment, the world's validities are always founded on other validities, above the whole manifold but synthetically unified flow in which the world has and forever attains its content of meaning and its ontic validity.26
Through the 'liberation' thus effected, one can and does (that is, "the philosopher [now] situated above his own natural being and above the natural world"27 can and does) discover "the universal, absolutely self-enclosed and absolutely self-sufficient correlation between the world itself and world-consciousness. "28 It is only this extraordinary great gasp of Husserl's that makes one uncertain of just how much (later)
'NATURALISM
=PHENOMENOLOGY'
43
phenomenologists are willing to admit regarding the convergence between naturalism and phenomenology. Also, to retreat from Husserl's extreme position (to suggest even that Husserl himself relented) is to make utter nonsense of his own anti-psychologism. If we assume the force of our original equation, such a strategy would completely subvert the entire purpose of the transcendental function of Russerl's phenomenology. That is why, to risk a sly conjecture, the good-humored simplicity of Dewey's pragmatism (itself a de-privileged, de-Iogocentric, precipitate of Hegelianized 'phenomenology') is enjoying such an inning at the present moment (The recovery of William James was rightly bound to follow.~ We are not here concerned with the vagaries of why Husserl - or, in his own way, Merleau-Ponty - obscures the convergence indicated. The important thing is to draw out its lesson. This surely includes at least the following: (i) that naturalism un-phenomenologized cannot fail to be logocentric; (ii) that phenomenology un-naturalized has no epistemic relevance whatsoever, and where it supposes it does it cannot fail to be foundationalist; and (iii) that, given (i) and (ii) - in effect, the symbiosis of frrst- and second-order knowledge - all science, the fruits of all acknowledged inquiry, cannot fail to be psychologized (psychologistic). Admittedly, these are rather cryptic fmdings, but they have a surprising force. For one thing, (i) signifies that any frrst-order inquiry that posits an epistemically pertinent relationship between cognizing subject and cognized object or world, that is itself not construed holistically in the pragmatist manner as a precognitive artifact of the structuring power of the Lebenswelt within which we come to occupy such a cognizing role, posits to that extent an initial or privileged transparency or correspondence or similarly favorable preharmony in virtue of which such frrst-order work succeeds as it does. This is, in fact, just the critical theme that Merleau-Ponty and Dewey explicitly enunciate and unwittingly share: in Merleau-Ponty, the 'relation' between 'primary' and 'reflective' experience; in Dewey, that between affmning the 'problematic situation' and rejecting the 'spectator' theory of perception and knowledge.30 It shows not only: (i-a) that any adequate epistemology must take the form of a realist/idealist symbiosis, but also (I-b) that that symbiosis must be phenomenologically encumbered. Kant ultimately fails to escape the logocentric predicament, though he does take the fIrst step; he fails to question the conceptual preconditions for the transcendental arguments he sought to refme over an entire lifetime. It is an irony that Husserl takes the second step but converts it at once into a new Cartesianism; Husserl simply places the Transcendental Ego beyond the life-worldly encumbmnces of the natural, the empirical ego. What (ii) signifies, then is that phenomenology has no other function but to experiment, within the life-worldly constraints of naturalistic science, with the conceptual variability of whatever saliently appear as the encumbering limitations, distortions, prejudices, interests, habits and the like that (precognitively) affect the cognized horizons, invariances, regularities, universalities, necessities, essences, laws, rules, theories, principles,
44
JOSEPH MARGOLIS
categories, concepts, and systems of concepts with which we organize the body of our particular disciplines. (The point may be usefully intruded, here, that Bertrand Russell's demonstration of the fatal paradox of Frege's original logicist account of arithmetic - the rightful mate and source of Husserl's anti-psychologism - decisively demonstrates that, even in the world of numbers [Frege's 'third realm'], one cannot count of the required measure of self-evidence both Frege and Husserl needed for their respective versions of anti-psychologism.'l) It takes but a step, therefore, to see that, once the reciprocity of (i) and (ii) is acknowledged, there is no principled second-order difference to be made out in legitimating the 'critical' first-order functioning of Hegelian, Marxist, Husserlian, Frankfurt Critical, Nietzschean, Wittgensteinian, and pragmatist accounts of the preformative constitution of the subject/object relation: there are only those fIrst-order differences and their (now) de-privileged second-order mates. (Those differences, are, of course, well worth considering.) The reciprocity of (i) and (ii) also signifies that there is no science or serious inquiry except as it is the work of a community of actual inquiring agents (selves) who address their cognizing powers to this or that sector of the world. To paraphrase Kant, naturalism without phenomenology is blind, and phenomenology without naturalism is empty. The reflexive analysis of the prefonnative conditions constituting any inquiry (as the inquiry it is) need be no more blind or empty than the inquiry it purports to examine: indeed, it is that inquiry, seen through the reciprocity of (i) and (ii). (On the argument here advanced, anything less or more pretends to a form of privilege.) What needs to be remembered are: (ii-a) that cognizing subjects (selves) are preformed by their environing world but in ways we can only guess at - reflexively, distributively, always 'prejudicially' (in Gadamer's sense'2J within the holist space of what we take our world to be; (ii-b) that what we take our world to be is, reciprocally, an artifact of what we take ourselves to be as the cognitively apt creatures that we are; and (ii-c) that both of these sorts of conjecture are assignable only to cognizing subjects, who regard themselves as addressing an objective world accessible to their native powers. Since, however, our frrst- and second-order reflections are symbiotically linked - are intentional in that sense - and since, given (i) and (ii), self and world are similarly symbiotized, we may reasonable draw out another element of the dawning lesson: (iv) that science and the legitimation of science cannot be conceptually disjoined. What we regard as the fruitful prospects and determinate promise of our first-order science are, ineluctably, the causal consequence and more of our second-order reflections on the nature of self and world." There you have the precise point of worrying the deep equivocation of Merleau-Ponty's insistence on the 'primacy' of perception and the 'relation' between science and the 'brute' world. There you have, also, the clue to the essential pathos of Imre Lakatos's doomed effort to form, in fIrst-order terms, a linear measure of the inherent promise of scientific research
'NATURALISM = PHENOMENOLOGY'
45
programs scrupulously compared across and within any interval of historical work.'" Kuhn's and Feyerabend's discontinuities and incommensurabilities the one meant mildly enough, the other more radically - completely subvert Lakatos's dream, merely by invoking (in their respectively thin ways) the phenomenological (or critical) theme. It may then prove a further economy to add, as a consequence of these last remarks, another strand of the intended lesson drawn from the convergence of naturalism and phenomenology: (v) that (cognizing) self and (cognized) world are never more than historicized referents - possibly discontinously, even incommensurably, certainly relativistically, posited referents - that cannot but be symbiotically linked. That is, what the nature of self and world is and what the cognitive relationship between them is are artifacts of salient truth-claims that, on a theory powered by the very ability to make such claims, supposes every such conjecture to be prefonnatively constituted under diachronically shifting conditions about which we must also conjecture only from the vantage of the other. In a plain sense, therefore, critique is no more than the first-order inquiry (or its second-order legitimation) into the historically prefonnative conditions of any putatively objective, distributed truth-claims engaging cognizing subjects and cognized world- itself similarly affected. Accordingly, psychologism is the second-order theory (or its narrow rust-order application to logic, arithmetic, geometry, transcendental phenomenology) that the legitimation of all such claims must forego any and all forms of foundationalism and logocentrism. It is easy to see that, however unlikely it may seem, critique and psychologism effectively require the same commitment: simply that naturalism = phenomenology (in the sense supplied). Reviewed in terms of these distinctions, deconstruction is nothing but a negative idiom - almost a via negativa - by which (i)-(v) are affmned or at least favorably featured. It is perhaps the most attenuated version possible of a purely (necessarily) mythologized instruction regarding the preformational contingencies affecting the emergence of cognizing subject and cognized world In what may well be one of Jacques Derrida's most felicitous formulations, the 'critical', interventionist function of deconstruction is unmistakable acknowledged: "The incision of deconstruction," he says, "which is not a voluntary decision or an absolute beginning, does not take place just anywhere, or in an absolute elsewhere. An incision, precisely, it can be made only according to lines of force and forces of rupture that are localizable in the discourse to be deconstructed."3S It effectively explains Derrida's resistance to any straightforward Marxist historicism (in his Positions), where historicism (in the French context) is taken by Derrida to conceal some logocentric privilege. But, in doing that, it also demonstrates that deconstruction is the mythologized unity of all fonns of discourse tempered in accord with (i)-(ii). In this sense, deconstruction catches up the hisroricizing theme common to Hegel, Marx, Nietzsche, Husserl, Heidegger, Merleau-Ponty, Lukacs, Gadamer, Habermas, Derrida, Foucault (and Dewey). But the theme itself (in the hands of just
46
JOSEPH MARGOLIS
these theorists) takes a naturalistic form (either logocentrically or not), a phenomenological or critical form (again, either logocentrically or not), and (on occasion) a deconstructive form (one that is purely mythological or ultimately converges with that of the others). Nietzsche's 'will to power' is, undoubtedly, the purest version of an exclusively deconstructive theme. Denida's 'differance' is merely an extraordinarily abstract adjustment of Nietzsche's theme fitted to (and intended to combat) recent preoccupations with structuralism, (Husserlian) phenomenology, Marxism, and historicismall of which are clearly prone to logocentric excess. Foucault's 'power/knowledge' is also a formula derived from Nietzsche but deliberately employed (with consummate skill) naturalistically, phenomenologically (that is, critically), and deconstructively at one and the same time.36 (It is, in fact, just this complexity that has baffled Foucault's commentators and encouraged the simpler but altogether inadequate picture of the straightforward social critic and activist. 37 Furthennore, it cannot (and certainly should not) be denied that Marx's notion of praxis plays a similar double role (critical and deconstruetive) that, within the often opposed phenomenological tradition, is best exemplified by Heidegger's tale of Sein and Dasein (and Derrida's differance)- though, of course, with an entirely different message on its critical side. The double role involves the mythological genesis of symbiotized subjects and objects - cogniscent, subcognitivety mobile (in Merteau-Ponty's sense), technologically active (in Heidegger's) pmxically effective (in Marx's) - and the critically oriented, distributive review and assessment of frrst-order life and work and inquiry informed by the particular second-order legitimating themes Marx happens to favor. Understandably, such a second-order choice will be installed, rhetorically, in the very idiom of the deconstructive mythology preferred, if indeed a deconstructive theme is actually prepared. Clearly, it would not be easy to demonstrate that Marx does work deconstructively, but it is certainly not an unreasonable suggestion. In fact, in a particularly important passage in The German Ideology, in which he specifically criticizes Ludwig Feuerbach's conception of praxis, Marx observes: He does not see how the sensuous world arolDld him is not a thing given direct from all eternity, remaining ever the same, but the product of industry and of the state of society; and, indeed, in the sense that it is an historical product, the result of the activity of a whole succession of generations, each standing on the shoulders of the preceding one, developing its industry and its intercourse, modifying its social system according to the changed needs. Even the objects of the simplest 'sensuous certainty' are only given him through social development, industry, and commercial intercourse.3I
This is certainly a passage that reminds one of the treatment of related themes in the thin naturalism of Thomas Kuhn, the anti-naturalistic phenomenology of Husser!, and the Nietzsehean genealogies of Foucault. It
'NATURALISM = PHENOMENOLOGY'
47
may well be that, without acknowledging the 'deconstructive' side of Marx's account of praxis, we risk failing to offset the familiar charge of incoherence so often leveled against him. That charge, tantamount to that of the sociology of knowledge, cannot stand once the subject/object relationship is treated in the historicized and praxicalized way Marx does,,39 Once one concedes incommensurabilities at the conceptual or cognitive level, and once one construes these as historically and praxically generat~ then the usual dilelnma invoked - either incoherence (assuming the privilege of a cognitively fixed stance) or vicious regress (the self-referential stigma of construing knowledge as ideology) - cannot fail to dissolve. The sociological thesis need not, of course, be restricted to the Marxist idiom: it could be deconstructive although Marxist; and it could, as with Foucault, be Nietzschean as well. But we must break off abruptly here, if we are ever to collect compendiously all the issues broached.
IV. The Ineluctability of Psychologism In drawing out the lesson of the convergence of naturalism and phenomenology, we mentioned but did not actually pursue a constituent item of our earlier tally- namely (iii) that science cannot but be psychologized (psychologistic). The meaning of the key term is a source of considerable quarrel. It could, for example, merely mean what it usually means in the naturalistic tradition, as in Quine's Millian conception of logic. On that view, logic, the study of the 'laws of thought', is a sub-study of the psychology of thinking. On Husserl's and Brentano's view, a grasp of the laws or of the necessary, self-evident constraints of logic may well require reflexive attention to the processes of thought in which they are embedded; but the laws are not merely the contingent regularities, or idealizations drawn from the regularities, of those processes. They are not psychologistic in that sense, though they need not, for that reason alone, entail the irrelevance of psychological process: what is reflexively presented (to consciousness, say) may well be (initially) psychological, but what is self-evident or apodictic in what is thus presented may (on an argument) depend on other grounds (psychologically embedded) that may be brought to bear. This is, in fact, Roderick Chisholm's view, partially based on a reading of Brentano's view;40 it could also have accommodated Husserl's view (contrary to Chisholm's intent) if we were (as we are not) prepared favorably to segregate what we had sorted in an earlier tally as (1) and (3), the difference between distinct cognitive sources or 'voices' (in Hussert's account, the empirical ego and the Transcendental Ego) and the affmnation and denial, respectively, of one or the other's being preformatively encumbered by the Lebenswelt. On Chisholm's reading of Brentano - and in his own name - a "person's self-presenting properties...are such that he can be absolutely certain that they are all had by one and the same thing-
48
JOSEPH MARGOLIS
namely, himself. "41 On Chisholm's (and Brentano's) view, phenomena or appearances - the 'elements of consciousness' - are intentional: "a phenomenon or appearance is an appearance to something or to some one"; hence, "there is something else to which [a phenomenon, even the phenomenon of being that to which a phenomenon appears] appears only as a phenomenon, and....this something else exists in itself and is apprehended as such."42 The apodictic, therefore, on Brentano's and Husserl's (mature) views, is drawn (on both accounts) from psychological phenomena, is not equivalently or identically construed by them, and is viewed on neither account as psychologistic. As Chisholm summarizes the matter elsewhere, the psychologistic "conception of the truths of logic, if it were tenable [which he holds it is not], would reduce 'reason' as a source of knowledge, to our 'inner consciousness'- Le., reduce [what he earlier identified as] (4) to (3)," namely: 3) 'Inner consciousness', or the apprehension of our own states of mind- for example, our awareness of our own sensations, of our beliefs and desires, of how we feel, of what we are undertaking to do; 4) Reason, as the source of our a priori knowledge of necessity- our knowledge, for example, of some of the truths of logic or mathematics.43
Chisholm's (and of course Brentano's) view would be construed psychologistically by Husser/. It is not so construed by Chisholm himself, apparently because the distinction between his (3) and (4) entails a distinction between any phenomenal content presented to a cognizing self and that self, and because the certitude assigned to the awareness of what is thus presented depends on the selfs reflexive awareness and not on (the mere awareness of) the psychological content of what is presented. On an interesting reading of Frege's anti-psychologism, Mark Notturno holds that all forms of psychologism that Frege opposed (Notturno actually manages to fonnulate four distinct fonns) are committed to the thesis that "truth is... dependent upon the judging subject. "« On Fregean grounds, then, truth is mind-independent: "A third realm [a realm of things other than spatio-temporal physical entities and other than mental ideas] must be recognized. ... The thought, for example, which we expressed in the Pythagorean theorem is timelessly true, true independently of whether anyone takes it to be true. It needs no bearer. It is not true for the fIrst time when it is discovered, but is like a planet which, already before anyone has seen it, has been in interaction with other planets. "4S Husserl would not be satisfied with Frege's fonnulation - though of course Frege's criticism of Hussert's Philosophy of Arithmetic decisively influenced his own version of anti-psychologism - simply because Frege did not pursue (whereas Husserl required, for his transcendental phenomenology) a disjunctive distinction (as we have already seen) between the empirical ego and the Transcencental Ego. So Hussert's criticism of Chisholm (and Brentano) would have depended on Chisholm's c~~~~!1& _~~ _J~9- _"Yllile __
'NATURALISM = PHENOMENOLOGY'
49
preserving apodicticity. The attack on psychologism is, then, neither an attack on the mere ineliminability of psychological sources of the materials on which a priori, apodictic truths are alleged to depend nor a mere affrrmation of the a priori or apodictic as such. This is often not appreciated. Furthennore, as Chisholm goes on to say, to grasp the absolute certainty of a person's self-presenting properties "is the closest [one] comes - and can come - to apprehending himself directly. But this awareness that there is something having the properties in question is what constitutes our basis, at any time, for all the other things that we may be said to know at that time."46 Unfortunately, if the phenomenological or critical theme of preformation is admitted, and if the very nature of the cognizing self is subject, in praxical and historicist terms, to diachronically changing preformational forces, then Chisholm's (and Brentano's) line of argument in effect, Descartes's - cannot fail to be inadequate for any body of science sustained beyond the specious present; and in the specious present, it would be pointless to affirm or deny the claim. The conceptual necessity of assigning the mere 'possession' of conscious states to a cogniscent self (Strawson's rejection of 'no ownership'47) is altogether independent of the apodictic standing of anything therein distributively presented. So there is a fair sense in which Chisholm's recovery of the apodictic is itself subject to the contingencies of the Lebenswelt and, for that reason, is ultimately psychologistic. The quick conclusion suggests itself at once: Frege's 'third realm' is an entirely arbitrary posit, insufficiently motivated philosophically, question-begging; and Husserl's transcendental source of apodicticity is itself simply subject, contrary to his own argument, to the prefonnative forces of the Lebenswelt. If so, then there is no escape from psychologism. More than that, one cannot fail to see that the vindication of psychologism is essentially the same as the vindication of the sociology-of-knowledge thesis. Both depend on a clear grasp of the force of historicist and praxicalist concessions within the tenns of our original equation. The naturalistic, therefore psychologistic, view is explicitly favored, confessedly under Quine's direct influence, by Gilbert Harman. Psychologism, on Harman's view, maintains that "the valid principles of inference are those principles in accordance with which the mind works,ft that is, the principles involved in "the working of the mind when nothing goes wrong: how it works ideally:'48 Harman's notion is that induction is the psychological process of ftinference to the best explanatory statement" yielding, in real-life terms, what may be called 'knowledge of the world'. On that view, there is no 'inductive logic': "Deductive logic is the only logic there is." Induction is inference - mental processing - directed (according to Harman's solution of the problem of induction) "to the best total explanatory account" "Deductive arguments," he adds, "are not inferences but are explanatory conclusions that can increase the coherence of one's view": "There are neither inductive arguments nor deductive inferences. There are only deductive arguments and inductive inferences."49
50
JOSEPH MARGOLIS
For Harman, therefore, "reasoning is a mental process," "mental states and processes are functionally defmed," reasoning itself is modeled by programs of inference the function of which lies "in giving us knowledge," and deductive arguments assign "an abstract structure consisting of certain propositions as premises, others as conclusions, perhaps others as intermediate steps" by which the successful functioning of actual processes of reasoning can be explained- the mapping of the reasoning process can be fitted to what appear to be the ideal formal conditions of success.'O Both because deduction is explanatory of induction and because there are no independent sources of apodicticity that Harman would or could acknowledge, the theory is thoroughly psychologistic. It is worth noting how extremely casual and straightforward is Hannan's reference to 'knowledge of the world', to how the mind works 'ideally' to achieve knowledge, to how deductive logic serves as a theory of the structure of inferential processes yielding in real-life circumstances 'the best total explanatory account' of the supposed events of the world. From Husserl's view, psychologism confuses and conflates the processes of the thinking of empirical egos or natural minds and the 'processing' of certain a priori 'pure ideal truths' (such as 2+2=4 or the law of non-contradiction) "in whose meanings not the least is said about the spatio-temporal, factual world"- whose truth holds "irrespective of whether there is a world or not ,,51 Quine, of course, had rendered the notions of logical necessity and apodicticity inherently problematic for naturalists: by rejecting, in the 'Two Dogmas' paper, any principled analytic-synthetic distinction. Hannan effectively follows him in this, for he abandons the explicit use of the expression 'logical necessity' and insists, without further explanation, on treating 'truth-conditional structures' in a way relativized to given natural languages- but so as not to preclude different such languages from exhibiting the same such structures. Harman nowhere discusses the conditions WIder which such structures are suitably 'idealized'. He merely claims universal scope for a suitable theory: "If all obvious implications could not be accounted for by means of a fmite list of axioms, smoething would be wrong with the theory of logical form. "52 Michael Dummett hints at an additional way of characterizing psychologism, but he does not carry the analysis out sufficiently. The Frege of "Der Gedanke," he says, "launches a renewed assault on psychologism, i.e. the intrusion of appeal to mental processes in the analysis of sense. tt He adds: "in doing so, [Frege] produces his most WIcharacteristic piece of writing: for, in the process, he for once essays a criticism of the idealist thesis that we are aware only of our own ideas, and hence have no ground for believing in the existence of a world external to us. US) The point at stake seems to be just that the domain of sense [Sinn] is not (whatever it is) the same as the domain of sensory experience, of psychological processes, or even of the 'outer' world of normal experience and empirical science. But, as we have already seen, that distinction (our consideration (2» would affect absolutely nothing regarding the apodictic assurances with
'NATURALISM = PHENOMENOLOGY'
51
which Sinn could be examined. Dummett has nothing really to say about that; and, in fact, his own (self-styled) anti-realist insistence on decidability undermines, if it can be construed as relevant at all, any Fregean or Husserlian sources of assunmce.54 Unless we actually link this theme - the specification of the intentional content or 'objects' of logic or arithmetic, what logic or arithmetic is 'a!x>ut' - with some further, nunified theory of the cognizable nature of the world or reality within which such 'objects' may be found, the very question of psychologism simply does not arise at all. In an odd sense, then, on Dummett's line of argument, in spite of what Harman openly declares, Harman does not actually opt for a psychologistic account of logic- in the narrow sense in which he confines logic to certain formal structures that thinking may instantiate in the mental processes of inference. It is only when he subsumes (as, of course, he does) this paired distinction under his more general naturalism that he rightly claims to be advocating psychologism. By a curious reversal, then, Frege remains a psychologist manqUe (for all one knows), in the sense that (like Dummett after him) Frege neglects to explain what the nature of the privileged cognitive competence of man is by which the truths of the 'third realm' may be assuredly grasped." The short conclusion stares us in the face: that either psychologism is quite trivially avoided (by making 100 much of our consideration (2); or it is altogether unavoidable, once the equivalence of natunilism and phenomenology is conceded, once we concede the reciprocity of (i) and (ii) of the tally of the previous section. Item (v) of that same tally now leads us to a further consolidation. For, the point of (v) is just that what is posited as the symbiotized subject and object of cognitive inquiry is an artifact of the historically preformative conditions that we posit in some present interval of just such inquiry. But that, as we saw, is exactly what Marx intended, in explicating (against Feuerbach) his own notion of praxis. Exactly the same lesson could be drawn from what, on its 'critical' side, Heidegger and Foucault intend by the use of their respective myths of Sein and Dasein and ,power!knowledge' . Surprising as it may seem, therefore, the issues of critique and psychologism are essentially the same. By 'critique' or 'critical' philosophy - adhering to (i) and (ii) - one merely means the legitimation of fIrst-order accounts of the preformational conditions under which any well-defmed frrst-order science or inquiry (physics, mathematics, sociology, medicine, law) is said to be so constituted. Phenomenology, then, is simply a particular kind of genemlized second-order theory of such preformational forces- as are, also, Hegelian, Marxist, pragmatist, Nietzsehean, Frankfurt Critical and similar undertakings. Here, then, for sheer convenience, also as a memento of the needlessness of an entire philosophical history, also for the sake of a nice irony, we are deliberately collapsing the distinction between 'phenomenology' and 'critique'- that is, as far as their generic second-order function is concerned, though oot, of course, in terms of their actual fIrst-order claims or in terms of their particular (un-privileged)
52
JOSEPH MARGOLIS
second-order policies or in terms of the agonistic roles particular champions may have pursued vis-a-vis one another. One might well say that Karl Mannheim's thesis of the sociology of knowledge, extended, generalized, and (thereby considerably) altered from Marx's original critique of ideology is the affmnative analogue, applied in the human sciences, to what (usually without explicit attention to historicist and praxical considerations) is criticized in the anti-psychologist's attack on psychologism. Mannheim speaks of "the characteristics and composition of the total structure of the mind of this epoch or of this group.ttS(i There is no way to make such specifications coherent without either retreating to a privileged access to a fixed and independent reality (the anti-psychologist's view) or to a frank admission of a serially historicized, preformative passage of cognitive orientations within the horizon of which the pluralized, relativized, potentially incommensurable structures "of the mind of this epoch or of this group" could possibly be said to be detected (the praxicalist's and the critiquer's view). That the latter maneuver places in jeopardy the entire question of what we should mean by objectivity goes without saying: it is in a sense just what Mannheim inherits from Max Weber's unresolved obsession with fixing the objectivity of sociology under the joint conditions of the 'interested' status of presently active human agents and the historical conditions forming those same interests. S7 On the argument here advanced, the retreat to privilege has been cut off and the advance toward a recovered objectivity remains problematic. Be that as it may, we surely see the sense in which no other options lie at hand, and we surely see both the irresistibility and the force of afftrrning that naturalism = phenomenology.
53 ENDNOTES 1. See: Edmund Husserl, ''The Task and the Significance of the Logical Investigations," trans. I.N. Mohanty, in Readings on Edmund Husserl's Logical Investigations, ed. IN. Mohanty (The Hague: Martinus Nijhoff, 1977), 11, 198-199. See also: Gottlob Frege, ''The Thought: A Logical Inquiry," trans. A.M. and M. Quinton, in Essays on Frege, ed. E.D. Klemke (Urbana: University of Illinois, 1968). 2. Many of the detailed analyses and arguments on which the support of our equation depends - suitably independently - are provided in: Joseph Margolis, Pragmatism withou.t Foundations: Reconciling Realism and Relativism (Oxford: Basil Blackwell, 1986). 3. The term 'logocentrism' was coined by Derrida, partly at least in criticism of Husser!. We are using it here more generally. The term appears in: Jacques Derrida, Of Grammatology, trans. Gayatri Chakravorti Spivak (Baltimore: The Jolms Hopkins University, 1976)- see, particularly, ''Translator's Preface." See also: Jacques Derrida, "The Supplement of Origin," in Speech and Phenomena and Other Essays on Husserl's Theory of Signs, trans. David B. Allison (Evanston: Northwestern University Press, 1973). 4. W.V. Quine, Word and Object (Cambridge: MIT, 1960), p. 219. 5. Quine, Word and Object, p. 221. 6. See further: Richard Rorty, Philosophy and the Mirror of Nature (Princeton: Princeton University, 1979); and Richard Rorty, Consequences of Pragmatism (Minneapolis: University of Minnesota, 1982). The discussion is aired in Pragmatism without Foundations and constitutes an important point of contrast affecting the present program. 7. Husserl's own intention and his usage of 'naturalism' and 'phenomenology' are conveniently given in: Edmund HusserI, Phenomenology and the Crisis of Philosophy, trans. Quentin Lauer (New York: Harper & Row, 1965). 8. Maurice Merleau-Ponty, "An Unpublished Text by Maurice Merleau-Ponty: A Perspective of His Work," trans. Arleen B. Dallery, in The Primacy of Perception, ed James M. Edie (Evanston: Northwestern University, 1964), pp. 3-4. 9. Maurice Merleau-Ponty, "Preface," Phenomenology of Perception, trans. Colin Smith (London: Routledge and Kegan Paul, 1962), p. vii, italics added. 10. This, indeed, is very close to the point of Rorty's deconstruetive reading of Dewey (and Heidegger). See: Richard Rorty, "Overcoming the Tradition: Dewey and Heidegger," in Consequences of Pragmatism. 11. John Dewey, Logic; The TJu!ory of Inquiry (New York: Henry Holt, 1938), pp. 104-105. The term 'situation' is somewhat clarified on pp. 66-67. In a class on Dewey's logic that I attended at Columbia University, Ernest Nagel made a great deal of (that is, fO\B1d utterly baffling, utterly untenable) the notion that it was the 'objective situation' that was 'indeterminate'. Nagel preferred assigning the intended indetenninacy, bafflement, uncertainty and the like to the inquiring subject But, of course, Dewey was profoundly opposed (avant la lettre) to the 'logocentric' implications of the simple, relatively constant relationship between subject and object. For an early anticipation of the doctrine of the Logic, actually directed against Peirce, see: John Dewey, 'The Superstition of Necessity," in John Dewey, The Early Works. 1882-1898 Vol. 4, 00. 10 Ann Boydston et ale (Carbondale, TIL: Southern Illinois University, 1972); and John Dewey, ''The Reflex Are Concept in Psychology," in Vol. 5 of the same series. See also: R.W. Sleeper, The Necessity of Pragmatism (New Haven: Yale University, 1986); Sandra B. Rosenthal, Speculative Pragmatism (Amherst: University of Massachusetts, 1986); and Joseph Margolis,
54
JOSEPH MARGOLIS
''The Relevance of Dewey's Epistemology," in New Studies in the Philosophy of John Dewey, ed Steven M. Calm (Hanover, N.H.: University Press of New England, 1977). 12. Dewey, Logic, p. 107 (in the context of pp. 105-107). 13. Mawice Merleau-Ponty, "Preobjective Being: The Solipsist World," in The Visible and the Invisible, trans. Alphonso Lingis (Evanston: Northwestern University, 1968), p. 157. 14. Merleau-Ponty, Phenomenology of Perception, pp. viii-ix. Only the expression 'speaks' was italicized in the original. Cf., also, pp. 201-206. 15. Maurice Merleau-Ponty, ''The Primacy of Perception and Its Philosophical Consequences," trans. James M. Edie, in The Primacy of Perception, p. 12. Cf., p. 38, where Merleau-Ponty answers a question posed by a certain M. Cesario 16. Maurice Merleau-Ponty, The Structure of Behavior, trans. Alden L. Fisher (Boston: Beacon Press, 1963), p. 224. Cf., also, pp. 170-176, 184. 17. Merleau-Ponty, Phenomenology of Perception, p. 203, italics added. 18. Merleau-Ponty, Phenomenology of Perception, p. 203. 19. The details are provided in Pragmatism without Foundations. The issue regarding Husserl is developed further in Joseph Margolis, Science without Unity; Reconciling the Human and Natural Sciences (Oxford: Basil Blackwell, 1987), Ch. 2- and here. 20. Merleau-Ponty's distinction between first-order and second-order considerations is captured (in reverse order) by the neat (but difficult) formula: "The knowledge of a truth is substituted for the experience of an immediate reality." Merleau-Ponty, The Structure of Behavior, p. 176. 21. Ednllmd Husserl, The Crisis of European Sciences and Transcendental Phenomenology, trans. David Carr (Evanston: Northwestern University, 1970), p. 143. 22. Ibid., p. 148. 23. Ibid., p. 49. See also: Appendix IV, ''The Origin of Geometry" and pp. 68-69. 24. Maurice Merleau-Ponty, "Phenomenology and the Sciences of Man," trans. John Wild, in The Primacy of Perception, pp. 92, 95. 25. A conveniently brief, sympathetic account of (what amounts to) these elements in Husserl is given in: Timothy 1. Stapleto~ Husserl and Heidegger: The Question of a Phenomenological Beginning (Albany: SUNY, 1983). 26. Husserl, The Crisis of European Sciences and Transcendental Phenomenology, p.150. 27. Ibid., p. 152. 28. Ibid., p. 151. 29. The point is very briefly noted in: Gerald E. Myers, William James; His Life and Thought (New Haven: Yale University, 1986), pp. 490n35. Myers does give a number of leads, however, about the relationship between James' pragmatism and phenomenology proper: see particularly pp. 490n35, 504n30. There is also an intriguing, somewhat impressionistic chapter that specifically associates Dewey and James and Merleau-Ponty in: John J. McDennott, Streams of Experience: Reflections of the History and Philosophy of American CultlUe (Amherst: University of Massachusetts, 1986). The sub-title of that chapter, Ch. 9, is actually "A Phenomenology of Relations in an American Philosophical Vein." 30. The clearest linkage between these two themes, in Dewey, is the following: ''The new realism [Dewey's] fmds...that thinking (including all the operations of discovery and testing as they might be set forth in an inductive logic) is a mere
'NATURALISM = PHENOMENOLOGY'
55
psychological preliminary, utterly irrelevant to any conclusions regarding the nature of objects known. The thesis of the following essays is that thinking is instrumental to a control of the environment, a control effected through acts which would not be undertaken without the prior resolution of a complex sitlUllion into assured elements and an accompanying projection of possibilities- without, that is to say, thinking." John Dewey, Essays in Experimental Logic (New York: Dover, n.d), p. 30, italics added. The formula is clearly as phenomenological as it is naturalistic. In a similar vein, Dewey opposes the assumption "that knowledge has a uniquely privileged position as a mode of access to reality in comparison with other modes of experience." John Dewey, The Quest for Certainty (New York: Minton, Balch, 1929), p. 106. I have consulted, here: Arthur E. Murphy, "Dewey's Epistemology and Metaphysics," in The Philosophy of John Dewey, 2nd 00., ed. Paul Arthur Schilpp (New York: Tudor, 1951). 31. The point is effectively pressed in: Mark Amadeus Notturno, Objectivity, Rationality and the Third Realm: Justifl.Cation and the Grounds of Psychologism; A Study of Frege and Popper (Dordrecht: Martinus Nijhoff, 1985), Ch. 6. 32. See: Hans-Georg Gadamer, Truth and Method, trans. from 2nd ed. Garrett Barden and John Cumming (New York: Seabury, 1975). 33. This is the central issue explored in Science without Unity, Pt One. The argument here sketched amounts to a reductio of Rorty's' thesis in Philosophy and the Mirror of Nature. 34. See : Imre Lakatos, "Falsification and the Methodology of Scientific Research Programs," in Imre Lakatos, Philosophical Papers, Vol. 1, ed. John Worrall and Gregory Currie (Cambridge: Cambridge University, 1978). 35. Jacques Derrida, Positions, trans. Alan Bass (Chicago: University of Chicago, 1981), p. 82. I have italicized the final phrase. 36. See: Michel Foucault, ''Truth and Power," in Power/Knowledge: Selected Interviews and Other Writings 1972-1977, 00. Colin Gordon, trans. Colin Gordon et ala (New York: Pantheon, 1980). See also: Michel Foucault, "Nietzsche, Genealogy, History," in Language, Counter-Memory, Practice; Selected Essays and Interviews, ed. Donald F. Bouchard, trans. Donald F. Bouchard and Sherry Simon (Ithaca: Cornell University, 1977). The most single impressive example of Foucault's applicaJion of his (deconstructive) theme at the level of frrst-order naturalistic and phenomenological (critical) history may be found in: Michel Foucault, Discipline and Punish; the Birth of the Prison, trans. Alan Sheriden (New York: Vintage, 1977). But see also: Michel Foucault, The History of Sexuality, Vol. 1, trans. Robert Hurley (New York: Random House, 1978). 37. See, for example: Mark Poster, Foucault, Marxism and History; Mode of Production versus Mode of Information (Cambridge: Polity, 1984). 38. Karl Marx, The German Ideology, in Karl Marx: Selected Writings, 00. David McLellan (Oxford: Oxford University, 1977), p. 174. Nicholas Lobkowicz notes the passage but makes very little of it; see: Nicholas Lobkowicz, Theory and Practice; History of a Concept from Aristotle to Marx (Notre Dame: University of Notre Dame, 1967), Ch. 25. Michel Henry is much clearer about the import of the passage. As Henry remarks: "As long as Marx remained Feuerbachian and understood reality as sensuous reality, sensuous representations had the meaning of reaching this reality in itself; images, dreams, religion were its 'mere representations'. But when reality is defmed by praxis, the ontological meaning of original representaJion, of the representation which presents being as an object and which in this presentation gives it as it is in itself, is lost.... The representaJion which reveals itself to be unequal to its ontological claims, which, as it gives us external
56
JOSEPH MARGOLIS
being, cannot give us being but is only a mere representation of it, this representation is ' consciousness' ." This Henry takes to be "the key concept to the whole Marxian interpretation of ideology." See: Michel Henry, Marx; A Philosophy of Human Reality, trans. Kathleen McLaughlin (Bloomington: Indiana University, 1983), p. 161, also, further, the whole of Cbs. 4-5. For an explicit comparison of Marx's efforts and deconstruction, see: Michael Ryan, Marxism and Deconstruction (Baltimore: Johns Hopkins University, 1982), Cbs. 2-3. It is interesting to compare here, on Heidegger: Reiner SchUrmann, Heidegger on Being and Acting: From Principles to Anarchy, trans. Christine-Marie Gros and the author (Bloomington: Indiana University, 1987). The 'soft' convergence between Marx's view and the views, already examined, of Merleau-Ponty and Dewey is clear enough. 39. For a recent, somewhat bland, sketch of the issue, see: Anthony Giddens, Central Problems in Social Theory; Action, Structure and Contradiction in Social Analysis (Berkeley: University of California, 1979), Ch. 5. See also: Karl Mannheim, Ideology and Utopia (New York: Harcourt, Brace and World, 1936); and Karl Mannheim, Essays on the Sociology of Knowledge (London: Routledge and Kegan Paul, 1952). 40. Roderick M. Chisholm, "Brentano's Descriptive Psychology," in The Philosophy of Brenlano, 00. Linda L. McAlister (Atlantic Highlands, N.J.: Humanities, 1976). 41. Roderick M. Chisholm, The First Person; An Essay on Reference and Intentionality (Minneapolis: University of Minnesota, 1981), p. 89. This needs to be read in the context of Chisholm's discussion of the whole of Ch. 7. See also: Franz Brentano, "On the Unity of Consciousness," in Psychology from an Empirical Standpoint, 00. Oskar Kraus; English ed. Linda L. McAlister, trans. Antos C. Rancurello et ale (New York: Humanities, (1973), especially p. 160 (cited by Chisholm). 42. Chisholm, "Brentano's Descriptive Psychology," p. 98. 43. Roderick M. Chisholm, The Foundations of Knowing (Minneapolis: University of Minnesota.. 1982), pp. 114, 119; cf., also, p. 155. 44. Nottwno, Objectivity, Rationality, and The Third Realm, p. 56. Cf. the rest of
Ch.3. 45. Frege, ''The Thought: A Logical Inquiry," pp. 523-524. 46. Chisholm, The First Person, p. 89. 47. See: P.F. StrawsOD, Individuals (London: Methuen, 1959). Cf. Margolis, Science withowl Unity, Ch. 3. 48. Gilbert Harman, Thought (Princeton: Princeton University, 1973), pp. 18-19. Cf., "Preface," p. viii. 49. Ibid., pp. 20-21, 162, 168, 172, italics added. The theme is quite Millan in spirit See: John Stuart Mill, Logic, 00. JM. Robson (Toronto: University of Toronto, 1973). 50. Harm~ Thought, pp. 43-44, 46-47m 48, 164; cf. Ch. 5, sec. 2. 51. Husser!, ''The Task and Significance of the Logical Investigations," pp. 198-199. 52. Harm~ Thought, p. 76, in the context of Ch. 5. 53. Michael Dummett, Frege: Philosophy of Language (New York: Harper and Row, 1973), p. 659. 54. See: Michael Dummett, Truth and Other Enigmas (Cambridge: Harvard University, 1978). 55. See: Dummett, Frege: Philosophy of Language, pp. 155-159; Hans Sluga, Gottlob Frege (London: Routledge and Kegan Paul, 1980); N~~~~_l!'Y~~tivJry~
'NATURALISM = PHENOMENOLOGY'
57
Rationality and the Third Realm; I.N. Mohanty, Hu,sserl and Frege (Bloomington: Indiana University, 1982), p. 43 and Ch. 2; and the charge made in Isaac Levi, The Enterprise of Knowledge: An Essay on Knowledge, Credal Probability, and Chance (Cambridge: MIT, 1980), Ch. 18: ''The Curse of Frege." Levi, it may be no~ treats Frege's curse as the imposition of a disjunctive choice between psychologism and anti-psychologism and claims to have beglDl to offer a third "alternative to suffering from its tyranny" (p. 428). But it is not clear that there can be a genuine alternative, and it is certainly not clear how Levi wishes us to understand his own option as such an alternative. On the contnuy, at least in terms of the distinctions here advanced, it seems to be a form of psychologism itself- that is, an option that could not demonstrate more than a naturalistic form of objectivity. NotturDo presses the objection in a reasonable way, quite apart from his quarrel with Levi regarding Popper's objection to psychologism; see~ Notturno, Objectivity, Rationality, and the Third Realm, pp. 216-217. A recent collection of new essays on Frege confrrms, at least implicitly, Frege's neglect of the epistemological issue. See: Crispin Wright, ed, Frege; Tradition and Influence (Oxford: Basil Blackwell, 1984), particularly the papers by Peter Carruthers, Harold Noonan, Bob Hale, Gregory Currie, and John Skorupski. 56. Mannheim, Ideology and Utopia, p. 56, italics added See also: Marx, The German Ideology, passim. 57. See further: Peter Hamilton, Knowledge and Social Structure,· An Introduction to the Classical Argument in the Sociology of Knowledge (London: Routledge and Kegan Paul, 1974); Gunter W. Remmling, The Sociology of Karl Mannheim (New York: Hwnanities, 1975); and Ted Benton, Philosophical Foundations of the Three Sociologies (London: Routledge and Kegan Paul, 1977). Unfortunately, the issue of the sociology of knowledge has not advanced very far beyond the repetition of the original puzzle.
PSYCHOLOGISM AND THE PRESCRIPTIVE FUNCTION OF LOGIC HERMAN PHILIPSE
I. Introduction Psychologism, as a view on the nature and epistemology of logic, is rejected by most modern philosophers. But the grounds for its dismissal may vary, and the question which grounds are the correct ones is not altogether immaterial. For the answer to this question often betrays the conception of logic one holds oneself. Husserl and Frege are usually praised for having finally refuted the doctrine of psychologism. Sometimes, however, reasons for doing so are attributed to them which they themselves would have thought to be misleading. In this paper I shall discuss an example of such a reason. A common view of the matter is expressed by Herbert Feigl in the following observation: Ever since Frege's and Hussed's devastating critiques of psychologism, philosophers should know better than to attempt to reduce normative to factual categories. It is one thing to describe the actual regularities of thought or language; it is an entirely different sort of thing to state the rules to which thinking or speaking ought to conform. l
Similar quotations can easily be found. So G.Radnitzky writes: Thanks largely to the pioneering work of Frege and Husserl, psychologism in logic and metamathematics is largely a thing of the past: the attempt to reduce the norms of logic to laws of thought is now merely a historical curiosity. 2
According to these authors, it seems, Husserl and Frege considered psychologism as a special case of the naturalistic fallacy, as an attempt to deduce ought from is. Such an interpretation rests on the assumption that they would have conceived logic as an essentially normative or prescriptive discipline. In the following pages it is argued that Husserl explicitly (and Frege implicitly) rejected such a view. Neither Husserl nor Frege conceived of logic as an essentially nonnative discipline. As a consequence they
THE PRESCRIPTIVE FUNCTION OF LOGIC
59
would have considered the idea that psychologism is incorrect because it commits the naturalistic fallacy as fundamentally mistaken. The 'fork' they use in combatting psychologism did not consist of the dichotomy between the factual and the normative, but, in the case of Husserl of the distinction between the factual and the ideal or, in Frege, the distinction between the subjective and the objective or between the objectively real and the objectively non-real. To make this historical point is, of course, not to say that Husserl and Frege were right, and I shall briefly discuss to what extent they were.
II. Theoretical Laws and Practical Prescriptions Of course, neither Frege nor Husserl denied the possibility of logical prescriptions or norms. But they both thought that the norms of logic are somehow derived from non-normative laws, the theoretical laws of pure logic. These theoretical laws would express an is rather than an ought. On the basis of such a view, one clearly cannot refute psychologism by saying that it commits the naturalistic fallacy, for one presupposes that it is legitimate to derive logical norms from non-normative propositions, Le. the laws of pure logic. What justifies this idea of the nature of logic? Husserl's Logical Investigations are on this point more revealing than the work of Frege. It is an integral part of this picture of logic that in general normative disciplines stand in need of theoretical foundations, a thesis Husserl tries to justify by his analysis of normative propositions in Chapter Two of the frrst volume of the Logical Investigations, the "Prolegomena zur reinen Logik. tt3 I shall fIrst summarize and discuss this analysis, and argue later that Frege must have held a similar view. Husserl's analysis of nonnative propositions serves several purposes. It not only purports to settle the question, much debated at the end of the nineteenth century, whether logic is a theoretical or a practical discipline,4 it also enables l-Iusserl to defme more precisely what psychologism is. In fact, he conceives both psychologism and his own conception of logic as 'pure' logic as two distinct answers to one and the same question, a question which he thinks of as being neutral in relation to all possible positions in the philosophy of logic. To elucidate the nature of norms, so Husserl seems to presuppose, is to define nonns in terms of something else. The customary observation that nonns express what ought to be the case (a Seinsollen) whereas theoretical laws express what is the case, is not sufficiently precise. A reduction of nonns to commands leaves many occurrences of ought unexplained, for a command presupposes an authority who issues it, and often there is no such authority.s Rather, Husserl attempts to define norms in terms of value judgments. Apparently inspired by syllogistic logic, he enlists four 'forms'
60
HERMAN PHILIPSE
of normative propositions, accompanied by the value-judgments in terms of which they may be defined: 1) 'An A should be B' = 2) 'An A should not be B' == 3) 'An A may be (is allowed) 4) 'An A need not to be B' =
=
'An A which is not B is a bad A', or 'Only an A which is a B is a good A' 'An A which is a B is a bad A', or 'Only an A which is not B is a good A' ,An A which is B is not, for that reason, a bad A' 'An A which is not B is not, for that reason, a bad A'
Apart from slight modifications to account for the distinction between necessary and sufficient conditions for being good or bad, Husserl is convinced that this list of the 'forms' of normative propositions is exhaustive.' Clearly, the four forms constitute the deontic Square of Opposition. The expressions 'good' and 'bad', as used in the forms of the value-judgments which are, according to Husserl, equivalent to forms of nonnative propositions, have a general meaning. 'Good' in the analysing expression means: what is somehow valuable. But in the substitution-instances of these fonns, 'good' and 'bad' will acquire the particular meaning determined by a certain evaluation (Werthaltung) which is presupposed, so Husserl thinks, by the corresponding form. Accordingly, in some value-judgments 'good' will mean useful, in others agreeable, in others beautiful, etc. From this Husserl infers that each norm presupposes an evaluation, by which specific notions of good and bad are defmed for a certain domain of objects. Such an evaluative definition may be of two kinds. If we define, for example, 'good' as what is a means to pleasure and 'bad' as what causes pain, the objects of the domain can be ordered along a scale, and good and bad are relative or polar opposites. If however, we define 'good' as deductively valid and 'bad' as deductively invalid, good and bad are absolute opposites.' HusserI's definition of normative propositions incorporates the idea that all forms in this way presuppose an evaluation: In relation to a general Wlderlying valuation, and the content of the corresponding pair of value-predicates detennined by it, every proposition is said to be 'normative' that states a necessary, or a sufficient, or a necessary and sufficient condition for having such a predicate.a
Moreover, this idea of an 'underlying valuation' or an evaluative definition enables Husserl to explain what the unity of a nonnative science or discipline consists in. A normative science contains all the noons which presuppose one and the same normative definition. Fundamental to such a
-------------------------------------
THE PRESCRIPTIVE FUNCTION OF LOGIC
61
nonnative science, so Husserl says, is the norm which states that the objects of the domain should posses the characteristics of the defmed Good (in the highest degree, if the evaluative defmition leaves room for degrees of goodness). He calls this norm the basic norm (Grundnorm) of the nonnative discipline. The answer to the question what differentiates the nonns belonging to a normative science from each other is" implied by Husserl's definition of normative propositions. Assuming certain evaluative definitions of 'good' and 'bad', we shall want to know what circumstances or properties of objects belonging to the domain of the defmition guarantee, contribute to, or exclude goodness or badness in the defined sense. In short, we shall want to know what are the necessary and sufficient conditions for something's being good. If, for example, one accepts, for a certain domain of objects, the hedonistic definition of 'good' as what is a means to pleasure, the question will be what circumstances or properties are required for a particular thing, event, or action to be a means to pleasure. Accordingly, there will be in relation to each kind of object of the domain a number of norms stating what these circumstances or properties are. Having discovered that salted soup in general tastes better than soup without salt, we shall affIrm the norm 'One should put salt in the soup'. The equivalent valuation reads: 'Only a soup which is salted is a good soup'. By substituting the hedonistic defmiens for 'good', we get: 'Only a soup which is salted is a means to pleasure'. For Husserl, the possibility of these substitutions shows that each norm contains a purely theoretical propositional content, which can be stated without normative or evaluative overtones. We may abstract the theoretical content from each particular norm by fIrst substituting for the norm the value-judgment which is, according to Husserl, equivalent to it, and then substituting the defmiens of the evaluative expression for that expression. Husserl affmns this conclusion of his analysis with complete generality: each normative proposition of the form 'An A should be B' contains a purely theoretical proposition of the form 'Only an A which is B has the properties C, where 'c' represents the descriptive content of the relevant evaluative predicate 'good'.9 This would hold mutatis mutandis for all forms of normative propositions: all noons incorporate a theoretical content, which states that certain conditions are necessary or sufficient for something having the properties C specified in the relevant normative definition. Conversely, every purely theoretical proposition may be transformed into a nonn if we assume an evaluative attitude towards the properties C. The nonn can be deduced from the theoretical proposition to the effect that the presence or absence of certain conditions is necessary or sufficient for something to be C, together with the basic nonn that the objects of the domain should have the properties C. Hussed summarizes his analysis by saying that each normative science has one or more theoretical sciences as its basis.
62
HERMAN PHILIPSE
IlL Psychologism and Pure Logic This idea of the theoretical basis of nonnative propositions is then used by Husserl to clarify the discussion between the proponents of psychologism and of pure logic, a discussion which, during the last half of the nineteenth century, suffered from great confusion. According to Husserl, a large part of the muddle was due to the fact that champions of pure logic like Drobisch and Bergmann grounded their case mainly on the argument that logic is a nonnative discipline. To this, the psychologistic logicians would reply that they did not deny the nonnative or practical character of logic, but only contended that the practical science of logic is, as it were, a technology based on theoretical psychology. This way of putting the problem suggests that the core of the debate between psychologism and the proponents of pure logic is the issue whether logic is a normative discipline. But both parties affmned this, so that one was tempted to conclude that the dispute is merely verbal. In Chapter One of the "Prolegomena," Husserl discusses the idea of logic as a normative discipline in order to clear the ground for a more correct construction of the dispute. He argues that this conception of logic is so obviously justified that it cannot be at stake in the conflict about psychologism. The central question is rather, so he says at the beginning of Chapter Two, whether the defmition of logic as a normative or practical science captures the essential nature of logic, i.e., whether logic is nothing else than a nonnative discipline, the theoretical basis of which is to be found in other, familiar and established sciences like psychology, or rather a distinct theoretical and non-normative science itself. IO In the remainder of the "Prolegomena," Husserl attacks the fonner and defends the latter view. The theoretical content of the nOnDS for correct deduction belongs not to psychology but to 'pure logic'; its validity is a priori and does not, like the laws of psychology, depend on facts. The issue is not whether logic is a nonnative discipline, but what kind of science provides normative logic with its theoretical basis. And the mistake of psychologism is not that it tries to deduce ought from is, for psychologism might assume, apart from psychology, a basic norm as a premise for the deduction of logical prescriptions. Its mistake is that it conceives this is as a factual is, and thus makes the norms of logic dependent on facts. II We are now in the position to see clearly what is correct and what is misleading in the interpretation of Husserl's position given by Feigl. The way in which Husserl reconstructed the issue of psychologism implies a certain view of psychologism itself. Psychologism does not necessarily identify the description of regularities of thought with stating the rules to which thinking ought to confonn, as Feigl suggests. Rather, it distinguishes the two and then affmns a relation between them, viz. that the former is the theoretical basis of the latter. Nor does psychologism necessarily 'reduce nonnative to factual categories'. If Husserl's account of nonns is correct, we can define normative (or evaluative) categories in non-normative terms.
THE PRESCRIPTIVE FUNCTION OF LOGIC
63
But these defmitions have a partial character: they merely express the descriptive content of the defmed expression. As we shall see, Husserl himself was aware of the special nature of his evaluative defmitions. Accordingly, the mistake of psychologism is not that it defines evaluative categories in descriptive terms. Its error rather consists in thinking that the defining expression of the evaluative definition relevant to logic contains factual categories. Although its categories are non-normative, they are not factual. In this sense the mistake of psychologism did consist in the attempt to reduce nonnative to factual categories. 12 Husserl's conception of psychologism and its corollary, the rejection of the argument that psychologism is incorrect because it commits the general error of deducing ought from is, both depend upon his analysis of nonnative propositions. This is equally true for Husserl's conception of pure logic. What champions of pure logic like Kant, Drobisch, Bergmann or Herbart were confusedly aiming at, so Husserl claims, is the view he himself defends in the "Prolegomena" and which he expresses as follows: ...That it is the true sense of our supposed pme logic to be an abstract theoretical discipline providing a basis for a technology just as the previously mentioned disciplines do, its technology being logic in the ordinary, practical sense. 13
Accordingly, psychologism and the conception of pure logic Husserl proposes have an important element in common. They both conceive nonnative logic - or rather logic as a practical discipline (Kunstlehre)14 - as a kind of technology, founded on one or more sciences. For Husserl, the apodictic validity of the norms of logic is to be explained by the idea that the theoretical basis of these norms is to be found in a non-empirical or 'pure' science, which is a priori and has a theoretical unity of its own, irrespective of the unity of normative logic which depends on its underlying nonnative definition. As the comparison of (nonnative) logic with a technology might be seriously misleading, it is important to examine more closely Husserl's analysis of normative propositions. Is it in general correct to say that each nonnative proposition states a necessary, or a sufficient, or a necessary and a sufficient condition for having the characteristics 'c' which are good or bad according to an underlying valuation or evaluative definition? And if this is not true in general, does Husserl's analysis perhaps apply to logical nonns, notwithstanding its incorrectness as a general analysis of normative propositions? Let me fIrst soothe possible qualms concerning Husserl's claim that in a nonnative defmition words like 'good' or 'bad' are defmed by means of defining expressions which do not have any normative or evaluative connotation. The hedonist who defmes 'good' as what is a means to pleasure does not try, as G.E. Moore thought, to define by some kind of real defmition the universal essence of goodness. Rather, the words 'good'
64
HERMAN PHILIPSE
and 'bad' express a positive or negative evaluation (Werthaltung) in relation to certain possible characteristics of objects of some kind. Such an evaluation is reflected by a normative definition in Husserl's sense. The only general connotation of the word 'good' is that something good is valued positively. But in most contexts, the word 'good' has a definite descriptive meaning as well; there are criteria or standards saying what characteristics some F should have in order to be a good F. And these characteristics are listed in a specific normative definition of the word 'good'. Husserl himself stresses that such a definition is not a definition in the sense of the 'usual logical concept of defmition'.1S What he probably means is that the evaluative connotation of the definiendum 'good' is absent from the defming expression, so that the defmed expression and the defining expression are not altogether equivalent A more serious objection to Husserl's analysis of norms concerns the generality of his reduction of normative propositions to value judgments and of the idea that a normative science is based on a normative definition. Husserl assumes that in general it is possible to state the normative definition underlying a normative discipline without adverting to the specific nonns of that discipline. For the content of these norms is in principle independent of the normative definition; it belongs to a theoretical science, and the reduction of norms to value-judgments would be circular if it were otherwise. But often this is not }X>ssible. If we define, for instance, a 'good Christian' as someone who sincerely tries to practice the Ten Commandments and the rules given in the Sermon on the Mount, then the properties of 'C' , which according to Husserl would constitute the descriptive content of the Christian concept of good, logically certain descriptions of all the nonns of the relevant normative discipline, Le. of Christian morality. In such a case, there is no room for a theoretical science which provides the norms with theoretical contents of the form 'Only an A which is B has the properties C' and the like, for the properties C are defined as the properties which an A has only if it is B. So Husserl is wrong to claim complete generality for his reductive analysis of normative propositions. One might argue that this objection is relevant to the case of logic as well. Husserl nowhere tries to define what is 'good from the logical point of view', but he makes it clear that normative logic essentially consists of the norms for valid deductions. 16 Could one not say that what is good from a logical point of view is to infer correctly, and that to infer correctly is just to follow the norms for valid inference? But such an argument would be confused. For although we certainly infer correctly if we follow the rules of deductive logic, this does not imply that it is impossible to give a definition of deductive validity, that is, a definition of what is good from a logical point of view, without invoking all particular nonns of deductive logic. In general, a deductive argument is said to be valid if, assuming the truth of all the premises, it is conceptualIy or logically impossible that the conclusion is false. And confining ourselves to those arguments the
THE PRESCRIPTIVE FUNCTION OF LOGIC
65
(in)validity of which can be brought to light by fonnal analysis, we say that an argument is valid if one of its logical forms has no substitution instances with true premises and a false conclusion, and invalid if its specific logical fonn has at least one substitution instance with true premises and a false conclusion. As these definitions do not contain any particular norm of logic, the objection we are discussing does not invalidate Husserl's analysis of logical norms, although it refutes his general analysis of normative propositions. There are, however, other difficulties in Husserl's analysis of normative propositions as applied to logical norms. In order to detect them it is necessary to carry through the application, a thing Husserl does not try to do. The Grundnorm of deductive logic will be the prescription that 'all deductive arguments should be valid'. Grundnormen (basic norms) in general must be an exception to Husserl's analysis of norms in terms of value-judgments. For if one translates the Grundnorm 'all deductive arguments should be valid' into the value-judgment 'Only a deductive argument which is valid is a good deductive argument', one gets the tautology that 'Only a deductive argument which is valid is a valid deductive argument', because 'good' in relation to deductive argument just means valid. As a consequence, the ' should' in basic norms is irreducible, and the only way to elucidate the basic norm is to substitute the defming expression of the relevant 'good' for it: 'all deductive arguments should be such, that if their premises are all true, it is conceptually impossible that their conclusion is false'. Husserl seems to have perceived that basic norms are an exception to his analysis of norms, for he says: The basic norm is the correlate of the definition of 'good' and 'bad' in the sense in question. It tells us on what basic standard or basic value all normativization must be conduc~ and does not therefore represent a normative proposition in the strict sense. l1
The principal difficulties, however, concern Husserl's assimilation of normative logic to a technology which is based on a theoretical science. Apart from the Grundnorm, all specific norms of logic of, say, the form 'An A should be B' would contain a theoretical proposition of the form 'Only an A which is B is such that if all premises are true, then necessarily the conclusion is true as well'. But how are we supposed to specify the class A? Husserl seems to assume that this is possible without mentioning the properties B or C (C is the 'descriptive' content of 'deductively valid', Le. the property that if all premises are true, then necessarily the conclusion is true). A frrst objection against this assumption would be, that there is, in the case of logic, an internal relation between the concepts A and C. If we specify the class A as the class of deductive arguments, we are already using the property C. One might even say that all the elements of the class A will have the property C. For surely a deductive argument is precisely an argument such that, if its premises are true, then necessarily its conclusion
66
HERMAN PIDLIPSE
is true as well. Consequently, there is no room for a property B which is a necessary condition for A's to be C, and the comparison of logic with a technology turns out to be altogether wrong. This version of the objection, however, is too strong. Although there is certainly an internal relation between the properties A and C, we nevertheless speak of 'invalid deductive arguments', so that it must be possible to specify a class A such that not all A's are C. This will be done by means of the distinction between a claim and its correctness: a deductive argument is an argument whose conclusion is claimed to follow from its premises with necessity. Only if the claim is correct, the argument really has the property C, and now we might try to fmd out under what conditions it is correct, that is, to fmd the properties B. Of course the assimilation of logic to a technology is misleading still. For in the theoretical content of technological rules, the properties A and C are typically not internally related18. A problem concerning the possible candidates for property B is solved by Husserl himself. One might think that a property is a candidate for B only if its presence would be a (necessary or sufficient) condition for any A to be C. But there are no specific requirements which deductive arguments of all kinds have to meet in order to be valid, even though one can sometimes formulate geneml requirements for the validity of a subset of deductive arguments (e.g., having defined syllogisms as arguments with two premises and a conclusion of the S-P form, all of which may be general or particular and affrrmative or negative, we require that a syllogism must contain exactly three terms, that the middle term must be distributed in at least one premiss, that at least one of the premises is affrrmative, etc). Interpreting logic as a theory of science, that is, of deductive theories, Husserl states in section 68 of the "Prolegomena" that a given theory is not required to conform to all 'conditions for formal validity' at the same time. Rather, the laws of logic constitute a 'fund' from which, as Husserl says, a valid theory draws 'the ideal grounds of its essential validity':' And in section 11 he accordingly specifies the form of the rules of the theory of science as follows: "Every (soi-disant) methodical procedure of the form M 1 (or M2...) is a correct one."2C) Quotations of the former kind draw our attention to the main way in which Husserl's comparison of logic with a theoretically backed technology is misleading. Like Frege, Husserl was anxious to rescue the 'objectivity' of logic from the danger of subjectivism and scepticism presented by the psychologistic doctrine. As the psychologistic conception of logic, at least according to Husserl, was the view 'that normative logic is a technology based on theoretical psychology, Husserl saw no other way of explaining the objective validity of logic than by substituting another theoretical science, pure logic, for psychology. Pure logic is conceived very much like psychology, although it is said to be different at crucial points. Its laws are about a domain of objects and are verified by a special kind of perception of these objects (namely, categorical intuition).21 Whereas in psychology the
THE PRESCRIPTIVE FUNCTION OF LOGIC
67
geneml law is said to 'explain' a particular fact, the laws of pure logic 'justify' their instances.22 However, pure logic is not a science of facts; its objects are 'ideal', they exist independently of space and time and independently of our knowledge of them. As far as pure logic is concerned with deduction, these objects are ideal meanings, Platonic Essences, and not the mental acts of meaning-intention, which are supposed to be their instances. 23 Although the purely logical laws of deduction are theoretical laws (in contradistinction to nonnative) and are about ideal meanings (propositions, concepts), they are purely conceptual as welL The laws of deduction "have their whole foundation in the 'sense', the 'essence' or the 'content', of the concepts of Truth, Proposition, Object, Property, Relation...etc. "24 And Husserl interprets conceptual laws as laws about concepts, concepts being Platonic Essences, which we can know by 'eidetic intuition'. Thus, Husserl's Platonism is in various ways 'the main foundation of pure logic and of epistemology'.25 There clearly is an intimate connection between Husserl's analysis of nonnative propositions and his conception of logic.. The norms of logic belong to a technology. They have a theoretical content which in itself is not normative. This theoretical content can be expressed in the proposition of pure logic, and the objective validity or truth of these propositions is due to their correspondence with certain objects, the ideal objects of pure logic. Thus, pure logic is like any other theoretical science, except for the nature of its laws: they are necessary, a priori, and not contingent. Within the framework of such a picture, the difference between the laws of logic and the laws of physics or psychology can be accounted for only in one way: by postulating a difference between the respective objects of these sciences. So Husserl affmns that the objects of logic are not facts, like the objects of the empirical sciences. They are ideal objects, existing outside space and time. The fundamental mistake of psychologism was, that it confused judgments as mental acts in time with judgments as ideal entities or propositions, and laws about facts with ideal laws. 26 Such a picture of logic is, however, hardly more satisfactory than psychologism. It is an explanation of 'the obscure by the more obscure. What is the nature of 'ideal objects'? To say that ideal propositions are the 'meanings' of declarative sentences, or the 'ideal species' of certain mental acts does not help us very much.27 And it will not do to persuade the reluctant reader to accept the existence of ideal objects by minimalizing the claim, saying that he should take it at frrst as an indication that propositions of logic and mathematics are valid. 28 If so, the claim is trivial, whereas it was meant to be a non-trivial explanation of the validity of such propositions. Elsewhere, Husserl says that it is out of the question that one may prove the existence of universals without solving the problem how we can know them or how we can become acquainted with them29 But there is no serious attempt to solve this problem in the Logical Investigations,30 and Husserl's endeavor to prove the similarity between perception and categorical intuition, necessary for giving some plausibility to the idea that
68
HERMAN PHILIPSE
we can 'perceive' supm-sensible entities, is rejected in the second edition.31 Finally, Husserl's later theory of 'eidetic variation' does not support the thesis that we can 'intuit' ideal entities, nor is it a correct description of the procedures by which we test the validity of proposed laws of logic. 32 Husserl's picture of logic, so we might say with the later Wittgenstein, is a product of misleading features of our grammar. Admittedly, we sometimes assert 'that arithmetic is about numbers and the theory of propositional logic is about propositions, that certain formulae are 'true' or 'correct'. But this about is different from the sense in which descriptions are about, and true of, the described object We can always imagine that an object is not as the description says it is, but we cannot imagine that a proposition is not either true or false: it would simply not be a proposition. The traditional law of the excluded middle is rather like an implicit definition of what propositions are. It is not a descriptive truth, but the corollary of a rule for the use of the expression 'proposition', a rule by which we delimit the range of applicability of traditional (two-valued) logic. And a law of logic like '({P~.p)~' (or: 'for all propositions p, q: (~q).p)~') is not true in the sense that it somehow corresponds to a reality which we would have to 'perceive' in order to test the law. Calling a formula like '({P~.p)~' a law of logic means that the formula is a tautology, that whatever propositions we substitute for 'p' and 'q', we get the result 'true' if we carry out the operations indicated by the definitions of the logical constants. These tautologies may be called 'laws of logic' because there is an equivalence between tautologies of a certain form and valid inference-schemata. An inference schema is valid if it has no substitution instances whose premises are true and whose conclusion is false. The definition of the material implication shows that a conditional whose antecedent consists of the conjunction of the premises of a valid schema and whose consequent is its conclusion must be a tautology: the only case in which the conditional is false is the case in which the consequent is false and the antecedent is true, but this case is excluded if the conditional corresponds to a valid inference schema in the way explained. Now inference schemata are certainly not descriptions of some reality, nor are they expressions of general laws about propositions. A correct schema is rather a rule for constructing valid arguments. Consequently, it is a mistake to interpret the laws of logic, which are nothing but reformulations of such schemata, as laws about some independent reality. As I said, there is a sense in which one might say that something 'corresponds to' a law of logic like the Modus Ponens. The law has the fOffil of an implication. It corresponds to an inference-schema because its antecedent consists of the conjunction of the premise-forms of the schema and its consequent is identical with the conclusion-form. This 'correspondence' is clearly not the kind of correspondence that a true description bears to the reality of which it is a description. Now of course the inference-schema itself 'corresponds to' something else as well: It
THE PRESCRIPTIVE FUNCTION OF LOGIC
69
corresponds to arguments in ordinary language. The schematic letters 'p' and 'q' correspond in a certain way to statements and the logical constants correspond in quite another way to expressions like 'if...then' and 'not'. To interpret these relations of 'corresponding to' in terms of the correspondence between a description and reality would be equally mistaken. The correspondence between the logical constants and the relevant expressions of common language, for instance, is rather a correspondence between the two sets of rules for the use of these respective signs. This correspondence is governed on the one hand by a requirement of adequacy, which has to guarantee that the logical system may be used in evaluating arguments in common language, and on the other hand by requirements of the system, like truth-functionality and simplicity. The tension between these two kinds of requirement admits of various solutions. One logical system may be further removed from common language than another, but easier to operate. However, although in a sense a formal system is a better instrument for assessing the validity of arguments the closer the rules for the use of its constants correspond to the rules of common language, it would be completely misguided to think that this correspondence is the correspondence of truth, that one system of logic is 'true' or 'nearer to the truth than other systems'. Finally, expressions like 'if...then', 'and', 'or', and 'not' do not ,correspond to' anything. Contrary to what Husserl thought, they are not referring expressions and we do not need a special kind of intuition, categorical intuition, to grasp the aspects of states of affairs to which they presumably refer. These considerations show still another way why the dualistic conception of logic as a normative science and a purely theoretical discipline which provides the former with its theoretical basis, is misleading. This conception presupposes that there is a basic norm telling us to reason correctly, which is completely separated from, and external to, the theoretical content of the nonns of logic. Conversely, the laws of pure logic in themselves would be free from any normative connotation. Now it is of course possible to express the laws of logic without explicitly using normative or evaluative words, e.g. the Modus Ponendo Ponens as '«P~.p)~'. And one might fonnulate a corresponding logical norm, e.g., 'it is permissible to infer from premises of the form 'p~' and 'p' a conclusion of the form 'q". But this does not prove that the technology-conception of logic is correct The validity of the formula of the propositional calculus is in fact nothing but the necessary outcome of a certain way of cOlTlbining schematic letters, brackets and logical constants: it is exhaustively explained by the rules for the use of these signs. These rules correspond in various ways to the rules of use for the expressions of common languages. Rules of language, however, posses an inherent 'normative import': if one wants to use the expression 'if...then', one has to use it correctly. Otherwise, what one says just does not make sense. Here it is artificial to dism.guish between a basic nonn (to use language correctly) and a theoretical content specifying the
70
HERMAN PHILIPSE
means which we may use in order to confonn to this norm. Moreover, in the process of learning a language we eo ipso learn to reason correctly. For to reason correctly is nothing but to apply the rules of language in an argumentative context, especially the rules for the use of expressions like 'all', 'some', 'if...then', 'and', etc. As the rules for correct reasoning are on par with the rules of language, we do not need a separate and external basic norm which prescribes that we should reason correctly. This norm is inherent in the use of language. So we come to the conclusion that Husserl's application of his analysis of nonnative propositions to logic, his assimilation of logic to a technology, and his contention that there is a purely theoretical logic, whose non-nonnative propositions are 'true' of an ideal world and provide normative logic with its 'theoretical content', is fundamentally misleading.
IV. Husserl and Frege: a coda In one of the first comparative studies on Husserl and Frege, Dagfmn came to the conclusion that Husserl's conception of 'the normative character of logic is different from that of Frege, and that Husser! improved on Frege by his clear distinction between normative and theoretical sciences.33 If what I have argued is correct, Husserl at most 'improved' on Frege in the wrong direction. His 'clear distinction' between normative and pure, theoretical logic was the product of a mistaken application of the distinction between a technology and its theoretical basis in logic. But the historical part of F~llesdal's conclusion is no more correct than its philosophical aspect. In fact, Frege, like Husserl, embraced the conception of logic as a theoretical science, constituting the basis of the norms for valid thinking. F~llesdal's opinion to the contrary rests on a mistaken interpretation of pages XV-XVII of the "Vorwort" to Frege's Grundgesetze der Arithmetik. According to F~llesdal, Frege considered the normative character of logic to be a decisive argument against psychologism;34 and asserted that the laws of logic are essentially prescriptive.35 In fact, Frege states on page XV of the "Vorwort" that the laws of logic should be prescriptions for thinking. But F~llesdal's interpretation misses the point (as well as the letter of the text: Frege says that the laws of logic should be prscriptions!) of this pasage by absttaeting from its argumentative context. Frege is attacking here a specific argument in favour of psychologism, the argument that logic is a part of psychology because the laws of logic are laws of thought (Denkgesetze). Frege's rejoinder is that the argument contains a fallacy of ambiguity. If one takes 'law' in the sense of prescription, it is true that the laws of logic are the 'laws of thought', for they prescribe according to Frege how we should think if we want to attain the truth. But in this sense of 'law', it is not true that the laws of thought belong to psychology. If it were, even the F~llesdal
THE PRESCRIPTIVE FUNCTION OF LOGIC
71
laws of physics would belong to psychology, for "each law which says what is the case may be taken as a prescription, viz. to think in accordance with it, and is in this sense a law of thought." The only special claim of the laws of logic to the title of 'laws of thought' is that they are the most general laws, so that they would prescribe how one should think in all domains of thinking. If, however, one takes the word 'law' in the descriptive sense, it is simply false to say that the laws of logic are laws of thought, and the argument in favour of psychologism collapses for this reason.36 One cannot read into this passage any support for the claim that the laws of logic are essentially normative according to Frege. On the contrary, the laws of logic are classified as 'laws which state what is', which, like all such laws, should be taken as prescriptions for our thinking. Frege indicates only one difference between logical and other descriptive laws: the logical laws are the most general laws there are.37 Moreover, the laws of logic are Gesetze des Wahrseins and not psychologische Gesetze des Furwahrhaltens. As a proposition is true or not true independently of our believing it to be true, the laws of being-true are not psychological laws, but "boundry-stones, set in an eternal foundation.... And because they are such, they set the nonns for our thinking, if it wants to attain the truth." The comparison of the laws of logic with the rules of grammar is explicitly rejected.38 F011esdal was of course right that Husserl proposed an elabomte analysis of the distinction between normative and theoretical sciences and the interrelations between them, Whereas Frege was not interested in the issue. But Frege probably would have accepted HusserI's view on this point, so that it might be used to elucidate what Frege meant by his rather lapidary remarks on the
subject.
72 ENDNOTES This is a revised version of a paper read at the University of Oxford during Trinity term 1984. I thank Professor M.A.E Durnmet (New College), Dr. P.M.S. Hacker (St. John's College), and Dr. B. Smith (University of Manchester and Erlangen) for their stimulating comments on an earlier version.
1. Herbert Feigl, "Physicalism, Unity of Science and the Foundation of Psychology," in P.A. Schipp, ed., The Philosophy of Rudolph Carnap (La Salle: Open Court, 1963), p. 250 2. G. Radnitzky, "Popperian Philosophy of Science as an Antidote Against Relativism," in R.S. Cohen et al., eds., Essays in Memory of Imre Lakatos (Dordrecht: Reidel, 1976), p. 505. 3. See: Edmund Husserl, Logische Untersuchungen, Band I, "Prolegomena zur reinen Lokik" (Halle: M. Niemeyer, 19(0). As the considerable differences between the first edition (1900) and the second edition (1913), on which all later editions are based, are irrelevant for the subject of this paper, I will quote from the fifth edition (fllbingen: M. Niemeyer, 1968). For the English reader I add references to the pages of Logical Investigations, trans. IN. Findlay (New York: Humanities Press 1970). 4. 5. 6. 7. 8. 9.
Cf., Hussed, Logical Investigations, Vol. I, sec. 3. See: Ibid., sec. 14. See: Ibid., Vol. L pp. 83-84; Logische Untersuchungen, Band I, p. 43. Cf., Ibid., Vol. I, sec. 14. Ibid., Vol. I, sec. 14, p. 84; Logische Untersuchungen, Band I, p. 44. See: Ibid., Vol. I, sec. 16. 10. See: Ibid., Vol. I, sec. 13; cf., Vol. I, sec. 41. 11. Husserl tries to give a reductio ad absurdum of psychologism by arguing, e.g., that if it were true, factual mistakes in reasoning would falsify logical laws, it would be possible to deny logical laws without contradiction, the laws of logic would be vague and merely probable, psychological research would be relevant to logic, and that, in short, psychologism is a kind of skepticism in the strict sense, Le. a theory which denies or questions certain conditions for the possibility of theories in general. Cf., Ibid., Vol. I, Ch. 4-7 12. So psychologism is incorrect, not because it overlooked the normative character of logic, but because it assumed that a factual science can be the theoretical basis of logical norms. The crucial argument against psychologism is not the normative character of logic, but the 'ideal' character of the theoretical basis of normative logic. Cf., Ibid., Vol. I, sec. 43; Logische Untersuchungen, Band I, pp. 164-165. Cf., Ibid., Vol. II, "Investigation L" sec. 32. 13. Ibid., Vol. I, sec. 13, p. 80; Logische Untersuchungen, Band I, p. 38. 14. Husserl distinguishes between normative disciplines in general and practical disciplines or technologies (Kunstlehren). A practical discipline is a normative discipline the Grundnorm of which states that we should aim at perfonning certain actions in a certain manner. Cf., Ibid., Vol. I, sec. 11, 15. 15. See: Ibid., Vol. I, sec. 14 in [uaem. 16. Pure logic, however, comprises much more than the theoretical basis of the norms for valid deduction. The principle of unity of a theoretical science is essentially independent and different from the principle of unity of a theoretical science. Pure logic, in Husserl's sense, embraces all the 'formal' or 'categorical' sciences. Cf., Ibid., Vol. I, sec. 14, 16, 62-64. 17. Ibid, Vol. I, sec. 14, pp. 85-86; Logische Untersuchlmgen, Band ~_p--"-_~~~-------G
THE PRESCRIPTIVE FUNCTION OF LOGIC
73
18. For example, in 'Only a soup which is salted is a means to pleasme', there is no internal relation between the concept of soup and the concept of being a means to pleasme. I asswne, in this objection, that it is impossible to define what arguments are, or, in casu, what deductive arguments are, without using the concept of (deductive) validity. Iri other words, the concept of an argument belongs to pragmatics. An argument essentially embodies a claim to be correct (inductively or deductively); it is not just an ordered set of propositions and statements. 19. See: Ibid., Vol. I, p. 239; Logische Untersuchungen, Band I, p. 246. 20. Ibid., Vol. I, p. 71; Logische UntersucJumgen, Band I, p. 27. 21. Cf., Ibid., Vol. n, ''Investigation VI." 22. Cf., note 19 above. 23. Cf., Ibid., Vol. I, sec. 36, 39, 40, 42, 65; cf., also, Vol. n, "Investigation I," sec. 29. Husser! empahatically warns the reader against the idea that ideal entities are 'more perfect' than factual entities, that 'the ideal' is essentially normative, or that 'ideal species' are somehow the most perfect instances of themselves, a confusion which lies at the root of many objections against Platonism. Cf., Ibid., Vol. II, "Investigation I," sec. 32. For Husser!'s conception of ideal meanings in the Logical Investigations, see especially: Vol. II, ''Investigation 1," Ch. ill, IV. 24. Ibid., Vol. I, p. 144; Logische Untersuchwagen, Band I, p. 122. Cf., Vol. I, sec. 23, 36, 39, 40, 42, 43, 50, 66. 25. See: Ibid., Vol. IT, ''Introduction'' to ''Investigation IT." Cf., however, Edmund Husser! Indeen Zil einer reinen Phii.nomenologie IlfId phiinomenologischen Philosophie I (Halle: M. Niemeyer, 1913), sec. 5. 26. Cf., Ibid., Vol. I, sec. 44-48. 27. Cf., Ibid., Vol. IT, ''Investigation L" ch. 4. In his review of M. Pahigyi, Der Streit der Psychologisten wuJ. Formalisten in tIer modernen Logile, ZEITSCHRIFf FOR PSYCHOLOGIE UNO PHYSIOLOGIE DER SINNESORGANIE, 31 (1903), p. 290, Husserl tells us that the puzzling character of Bolzano's doctrine of Slitze an sich vanished as soon as he realized that Siitze an sich are just what we call the meaning of a statement, and that these meanings are ideal species of certain mental acts. The review is reprinted in volume xxn of the Hwsserliana, ed. and trans. Dallas Willard as "A reply to a Critic of My Refutation of Logical Psychologism," THE PERSONAUST, 53 (1972), pp. 5-13. 28. See: Husserl, Logical Investigations, Vol. II, "Investigation 1," sec. 31, in finem. 29. See: Ibid., Vol. IT, ''Investigation 11," sec. 7, p. 351; Logische Untersuchungen, Band II, pp. 122-123. 30. Cr., Ibid., Vol. IT "Investigation II," sec. 14 and ''Investigation VI," sec. 52. 31. This is the theory of categorical representation of chapter 7 of the ''Investigation VI." It is rejected in the "Vorwort" to the second edition of ''Investigation VI," Ibid., Vol. IT, p. 663; Logische Untersuchungen, Band IT, p. V. 32. As I argued elsewhere. Cf., Hennan Philipse, De FllfIdering van de logica is Husserls 'Logische Untersuchungen' (Leiden, 1983), eh. 1.7. 33. Dagfinn F011esdal, Husserl und Frege. Ein Beitrag ZIlT Beleuchtung der Entsteehung tIer phiinomenologischen Philosophie (Oslo: I Kommisjon hos Aschehoug, 1958), p. 49: "Husserl \D1terscheidet sich aber deutlieh von Frege in seiner Auffassung von dem nonnativen Charakter der Logik. Es scheint als ob er durch seine klare Trennung zwischen theoretischen Wissenschaften \Dld nonnativen Wissenschaften bier weiter gekommen ist a1s Frege." 34. See: Ibid., p. 46.
74
HERMAN PHILIPSE
35. See: Ibid., p. 45. The passage in F~llesdal to which I am refering read as follows: "B usserl fmdet, im Gege1JSlJlZ Zil Frege, dass tkr normative Charakter der Logik kein enlscheidendes Argument gegen den Psychologismu,s sein kann." Ibid., p. 46, and "In seiner Kritik des Psychologismu,s in ' GTlUIdgesetze der AritJunetik' legt Frege grosses Gewicht QIJ{ den norma/wen ChmakJer der Logik.... Ihre Gesetze sind vorschreibend, und Q44Sschliesslich dies, behauptet Frege (Gg, XV)." Ibid., p. 45. 36. The relevant passage reads as follows: "Dass die logischen Gesetze Richtschnluen fiir das Denken sein sollen ZUI' E"eichlmg der Wahrheit, wird zwar vorweg allgemein zllgegeben; aber es geriith 7UIJ' Zil Leicht in Vergessenhe.it. Der Doppelsinn des Wortes ,Gesetz' ist hier verhiingnisvoll. In dem einen Sinne besagt es, was ist, in tkm andern schreibt es vor, was sein soli. NUl' in diesem Sinne kOnnen die logischen Gesetze Denkgesetze genannt werden, indem sie festsetzen, wie gedacht werden soli. J edes Geselz, das besagt, was ist, kann aufgefasst werden als vorschreibend, es solie Un EinJdange damiJ gedacht werden, UJU.l es ist also in dem Sinne ein Denkgesetz. Das gilt von den geometrischen und physikalischen nicht minder als von den logischen. Diese verdienen den Namen ' Denkgesetze' nur dann mit mehr Recht, wenn damit gesagt sein soil, dass sie die allgemeinsten sind, die ilberall da vorschreiben, wie gedacht werden soli, wo iiberhaupt gedacht wird." Oottlob Frege, Grundgestze der AritJunetik, begriffschriftlich abgeleitet (Jena: Hennann Pohle, 1893), Band I, p. XV). 37. Cf., Gottlob Frege, Die Grundlagen der AritJunetik, eine Logisch-mathematische Untersuchung uber den Begrijf der 7Ahl (Breslau: Wilhelm Koehner, 1884), sec. 14. 38. See: Frege, Grundgesetze, Band I, p. XVI.
FREGE'S ANTI-PSYCHOLOGISM G.P. BAKER and P.M.S. HACKER
I. Introduction One celebrated aspect of Frege's philosophy of logic was his insistence on the need to purge everything psychological from the clarification of the fundamental concepts of logic and from the elucidation of the primitive terms of his logical notation (which he called 'concept-script'). He inveighed against "the irruption of psychology into logic," drawing attention to the extent of the devastation caused by it and diagnosing the germs of this "widespread philosophical disease."1 He proposed to "reject all distinctions that are made from a purely psychological point of view."2 Indeed, he made it one of his fundamental principles "always to separate sharply the psychological from the logical, the subjective from the objective. "3 These pronouncements are the leitmotifs of Frege's antipsychologism. Our task here is to identify the targets of his criticisms, to ,/' clarify the scope and force of his counterarguments and to assess his achievements in this philosophical campaign. The established tradition was to view inference as the primary subjectmatter of logic. Inferences were thought to be sequences of judgments (propositions, thoughts), and judgments to be built up out of concepts or ideas. Hence, treatises on logic standardly began with discussions of concepts, continued with material on the composition of judgments, and culminated in cataloguing laws of thought (and also standard fallacies). The notions of concepts, judgments, and inferences were the basic concepts in the philosophical discussion of logic. The general nature of logic was taken to depend on the nature of these entities. Frege himself was educated in this tradition, and in the broad outline he accepted most of it He made some important improvements. In particular, he held that the soundness of an inference depends only on what is asserted, the thoughts that are put forward as being true. The term 'judgment' (or 'assertion') may be used to designate either what is judged (or asserted) or the act of making a judgment (or assertion). Hence, Frege introduced the term 'judgeblecontent' to make clear that he held logic to concern the objects, not the acts, of judgment. Similarly, he complained that traditional logic had too narrow a notion of concepts. Although judgeable-contents are composed of ideas, concepts are not building-blocks given independently of judgments, but rather concepts must always be precipitated out of judgments. With these two provisos, Frege adhered to the traditional concept of logic. He
76
G.P. BAKER & P.M.S. HACKER
thought a judgeable-content to be a "combination of ideas. ft4 These ideas, or unjudgeable-concepts, were subdivided into objects and things or concepts and functions. s Likewise, he thought that inferences are concatenations of judgments. To infer is "to make a judgment because we are cognizant of other truths as providing justification for it"' In his view, the task of logic is to test in the most reliable manner the validity of chains of reasoning,7 and its goal is to set up the laws of valid inference.8 Frege added the controversial' (and mistaken) idea that "only a thought recognized as true can be made the premise of an inference,ttlO thereby rejecting the idea that indirect proofs involve drawing conclusions from false thoughts. ll Frege's geneml conception of logic is consonant with the traditional view that it is the nonnative science of concepts, judgments, and inferences. Rules of deduction he held to be grounded in truths about concepts and judgments. Frege had a further inclination to side with tradition in philosophy of logic. The laws of thought, or laws of truth, have standardly been taken to be necessary truths, established as absolutely certain, and known a priori. The most elementary laws such as the Law of Identity ('Whatever is, is') or the Law of Non-contradiction ('Nothing both has and lacks a given property') are conceived as self-evident first principles of an a priori science pamllel to Euclidean geometry, and more complicated laws are displayed as the theorems of an axiomatic system developed out of these initial seeds. Frege always subscribed to this conception of logic, and he strove to erect on self-evident axioms a system of propositions of logic which would encompass all forms of inference acknowledged in advanced mathematics. Like most philosophers and mathematicians, Frege was inclined to link this conception of the objectivity of the truths of logic with a Platonist conception of their subject-matter. He was disposed from the outset to view concepts and judgments as abstract entities and to suppose that the propositions of logic are made true by the properties and relations of these denizens of a suprasensible realm. Just as Euclidean geometry is commonly thought to describe relations among mathematical entities (nwnbers, operations, and functions), SO Frege thought of logic as a science about special logical entities, their properties and relations. His conception of what these entities are gradually evolved; he had not 'discovered' the two truth-values or arrived at his later conception of concepts and thoughts when he wrote Begriffsschrift. Nonetheless, his allegiance to a Platonist picture of the subject-matter of logic was the constant companion of his conception of the objectivity of logic. These traditional ideas are part of the framework in which Frege's philosophy of logic evolved. A crucial consequence is that, from his point of view, the burden of proof must always lie on any philosopher who challenges the conventional wisdom. By his lights, to rebut all serious objections is to vindicate the Platonist conception of logic; nothing further can be required. In this respect modem philosophers may be expected to fmd his arguments defective. For the 'postulation of abstract entities' is now viewed as a prima facie intellectual crime, and hence an advocate of any
FREGE'S ANTI·PSYCHOLOGISM
77
fonn of Platonism must discharge the task of proving that no other more economical philosophical explanation is available. Times have changed. So too have the implications of the word 'Platonism'. Indeed, it may promote misunderstanding to classify Frege's conception of logic as a form of Platonism. This gives rise to expectations that his work does not fulfill. That in nun may entail the verdict 'that there is a gulf between his platitudes about concepts or judgments and his claims that concepts and objects are real but non-sensible entities, to be discovered and investigated by the eye of the mind or by the 'logical faculty'. It seems certain that Frege would have seen no gulf here, and the lacuna now detected in his arguments was not one that he was in a position to make good. It would be an interpretative blunder to treat this shortcoming as evidence of foolishness or philosophical ineptitude on Frege's part. But equally, it would be foolish to set it aside as a simple blunder which should be ignored on the grounds of charity. For what Frege did not see to be problematic has decisive importance for understanding his own positions and his criticisms of alternatives. Misconceptions about the framework of his thinking have vitiated most expositions of his anti-psychologism and most assessments of its worth. What Frege considered to be the most serious challenge to his conception of logic came from a form of radical empiricism that flourished in Gennany in the nineteenth century. No doubt this emerged as the antithesis to post-Kantian idealism. It drew inspiration from classical empiricism, but it moved off in a novel direction. It is commonly held to have originated in the writings of Benecke and Fries. They initiated the so-called 'deepening of logic by psychology'12 and appropriated the label 'psychologism' for their ideas. Frege took the definitive expositions of this stream of thought to be E. Husserl's Philosophy of Arithmetic (1893) and B. Erdmann's Logic (1892), and these texts were the primary targets of his tirades against psychological logicians and the incursion of psychology into logic. Qosely allied arguments occur in the writing of Mill and James, and much of the thrust of this radical empiricism persisted in Mach and Schlick. Frege was undoubtedly right to see in this fashionable and influential psychologism the major contempomry threat to the acceptance of his own point of view about logic. The psychological logicians claimed two special insights that called for a new conception of the nature of logic. First, while agreeing that the / subject-matter of logic is concepts, judgments, and inferences, they held that v these entities are all psychological or mental, and hence subjective. Words stand for ideas, all of which are derived from experience by a mental process of abstraction. To understand a word is to have the appropriate idea or image. Understanding an assertion is a matter of combining the ideas associated with its constituent words. A judgment is a combination of ideas (or a complex idea), and an inference is a transition from judgments to another judgment. Since ideas are held to be mental entities, the judgments of which they are the parts must also be mental entities, and so too must
78
G.P. BAKER & P.M.S. HACKER
be the inferences of which judgments are the parts. According to the psychological logicians, the subject-matter of the truths of logic belongs to the domain of the mental and the subjective. The laws of thought are propositions made true by facts about psychological entities. Secondly, the truths of logic were held not merely to be about mental entities, but also to have the status of empirical genemlizations. Psychological logicians claimed that these propositions are not necessary truths, that they rest on inductions based on observations of inner sense or introspection, and hence that they are a posteriori truths. According to this view, the nature of the truths of logic has been misunderstood by most philosophers through the ages. It is alleged to be an illusion of reason to suppose that logic (and mathematics) embodies any genuine non-empirical knowledge. The radical empiricism of the psychological logicians is committed to the thesis that all knowledge is empirical, and it saved the claim that there is knowledge in logic and mathematics by alleging that these so-called a priori sciences actually consist of empirical generalizations. In short, the laws of logic were held to be psychological statements, either natural laws of human thinking or else descriptions of thought-patterns that are, perhaps, variable between cultures or through history. This form of Empiricism or naturalism made a radical break with traditional philosophical thinking about logic. Frege made it his business to challenge the central theses in this cluster of ideas. These are the overt targets of his anti-psychologism. This must be borne in mind when expounding his arguments and assessing their forcefulness. The thinking of the psychological logicians had affinities with classical empiricism in the works of Locke, Berkeley, and Home. But the criticisms of these ideas lay outside the scope of Frege's reflections. 13 Likewise, nineteenth century psychologism had more tenuous affinities with idealism as developed by Kant, Hegel, Schopenhauer, Bradley, etc. But it would be as misguided to look in Frege's writing for refutations of their leading ideas as it would be to characterize idealism as the principle target of Frege's philosophy of logic. I" One is apt to misunderstand his anti-psychologism altogether if one loses sight of its specific target Despite a few asides about the wider implications of his critical comments, the intended scope of his invective against psychological logicians was narrow. His purpose ~as to demolish the pillars of their conception of the nature of logic. In this campaign he did not take it to be necessary to probe the roots of their misconceptions. He thought it sufficient to demonstrate the absurdity of their conclusions and the lack of cogency for their arguments. In particular, Frege argued that psychologism made it unintelligible that there could be an intersubjective science of logic (because of the acknowledged impossibility of sharing ideas). He offered arguments to refute the claim that all ideas are derived by abstraction from experience (focusing for this purpose on number-words and their use in count-statements). And he pointed out that it is a fallacy to infer from the fact that thinking is a mental activity that the objects of thought (propositions or judgments) are mental entities. To establish these points was, in his view, to construct a
FREGE'S ANTI-PSYCHOLOGISM
79
definitive case against psychologism. 15 Nothing more ambitious was required, no deeper exploration of the concepts of inference, thinking, assertion, concepts, concept formation, the normativity of logic, etc. It is an open question how far Frege's purpose-specific remarks about these concepts may make a significant contribution to philosophy of logic and philosophy of mind. He engaged in polemical skirmishes with the psychological logicians in the hope of eliminating an intellectual nuisance, and hence there should be no presumption that his anti-psychologism is a set of eternal verities laying the foundation for philosophical wisdom in the twentieth century.
ll. Frege's Influence and Alleged Achievements Though parochial in intent and origin, Frege's anti-psychologism developed into a protagonist in the drama of modern philosophy of logic. His influence has been immense and his authority often cited in support of campaigns to keep logic uncontaminated by psychological consideration. In part this depends on two factors that have nothing to do with the force of his case. The fIrst is that many of his remarks about purifying logic of everything psychological are colorful and quotable. They are apt to serve as revolutionary slogans and battle cries. The second is that he invented the \/ predicate calculus and thereby solved the riddle of the formalization of inferences involving multiple generality. The prestige of this intellectual achievement rubs off on the philosophical reasoning which accompanies his axiomatization of the truths of logic. Indeed, it is tempting to suppose that Frege's solid achievement in formal logic guarantees the value of his whole philosophy of logic since the foundations of the predicate calculus must be his conception of inferences, judgments (judgeable concepts or thoughts), concepts and concept-formation, the sense/reference distinction, etc. In particular, his rejection of psychologism and his espousal of Platonism in logic are given an imprimatur. There is little wonder in the fact that his anti-psychologism has had more far-reaching influence than he could have foreseen. Though these intrinsic features of his work contributed to the impact of his criticisms of psychologism, extrinsic factors were even more important. Russell and Wittgenstein, the two seminal figures in the development of philosophy of logic in the period 1900-30, both noted and endorsed Frege's anti-psychologism. In his frrst letter to Frege, Russell noted that ttl fmd myself in full accord with you on all main points, especially in your rejection of any psychological element in logic... ."16 For the next decade Russell insisted that propositions must not be taken to be mental entities;' that propositions (or complexes) stand in objective logical relations to one another, and that psychological concepts (e.g., of assertion) must be distinguished from logical ones. Rigorous exclusion of psychology from
80
G.P. BAKER & P.M.S. HACKER
logic was a cornerstone of the Tractatus too. Philosophy of logic has nothing whatever to do with the psychological investigation of thoughtprocesses which entangled many earlier philosophers and blurred the boundary between philosophy and the natural sciences.18 Indeed, Wittgenstein even accused Frege and the authors of Principia Mathematica of having failed to draw the bounds of logic narrowly enough. In particular, he declared that the assertion-sign is logically quite without significance; it merely indicates that these philosophers hold propositions marked with this symbol to be true,19 and that is irrelevant for logic since valid inferences may be made even from false propositions.20 The sole concern of philosophy of logic is to clarify the essential nature of the proposition, the general form of a proposition that can be apprehended once and for all in advance of any experience, and to investigate the forms of elementary propositions, or catalogue simple objects.21 Considerations about the workings of the human mind have no place in these enquiries; the stimulus for this idea and the radical nature of Wittgenstein's anti-psychologism were undoubtedly among his debts to the great works of Frege. Frege's influence on Russell and Wittgenstein multiplied many-fold the historical impact of his criticisms of the intrusion of psychology into logic. For these authors played pivotal roles both in the development of formal logic and in the growth of conventionalism which dominated philosophy of logic for two decades. Principia and the Tractatus were canonical texts among the logicians who built the metalogic of the predicate calculus in the '208 and '30s. The anti-psychologism of these texts coincided with the inclination of such philosopher-mathematicians as Ramsey, GOdel, Church, Tarski, and Carnap to think of themselves as making objective discoveries, and their technical successes reinforced this philosophical point of view. At the same time, the philosophy of logic which the Vienna circle extracted from the Tractatus and brandished as 'conventionalism' seemed to vindicate hard-headed empiricism without recourse to treating the truths of logic and mathematics as laws about mental entities. By holding out the vision of the objectivity of logic without abstract objects, logical empiricism forestalled any resurgence of psychologism among most philosophers who were inclined in temperament towards empiricism.22 These two streams of thought, one formal, the other philosophical, have made it commonplace among philosophers in this century to contrast logic with psychology. Students are indoctrinated with the views that the business of logic is the investigation of the relation of logical consequence among statements; that this depends on having a semantic theory for a language which makes clear what the truth-conditions of any well-formed formula are; and hence that the science of logic belongs to the study of language. From the beginning, undergraduates are taught that validity is a property of ordered sets of sentences or Connulae, not a property of sequences of thoughts or beliefs or of mental processes of reasoning. It is now an uphill battle for a philosopher to argue that psychology or even philosophy of mind has any
FREGE'S ANTI-PSYCHOLOGISM
81
proper place in philosophy of logic. Frege's philosophical heirs have had a total victory in the campaign which he initiated. On the other han~ as the ripples of his anti-psychologism have spread, the perceived content of his initial criticisms has become more and more diffuse and vacuous. Logicians in 1987 are not trying to fend off the thoughts of the particular psychological logicians that Frege attacked. Most of these ideas are neither still in circulation nor regarded as live options. Consequently, slogans borrowed from his writing are now put to uses that he did not intend or even foresee. This evolutionary process is visible from the beginning. Russell too was eager to establish that propositions and relations are not mental entities, but real, objective things which confront us in experience; and he too emphasized the importance of distinguishing the truth of a proposition from its being taken to be true. But the target of his attack was not the naturalist psychologism of early Husserl and Erdmann; it was the very different idealism associated in England with the work of Bradley and Green. Hence, Russell conscripted Frege into an alien crusade \/ against idealism. This was the frrst step down the road of divorcing Frege's anti-psychologism from his supporting philosophical arguments and presenting particular criticisms of specific philosophical reasoning as verites eter1U!lles of great generality, the foundation-stones of sound philosophy of logic. If one then subjects Frege's arguments to careful scrutiny, one fmds that his reasoned case is far too weak to support the sweeping theses listed among his achievements. Should one condemn Frege for this defect? Would it not promote a better under standing to expound the content of his anti-psychologism by reference to his arguments and to examine the assumptions on which his reasoning actually rested? What cannot be disputed is Frege's role as a catalyst in the process of achieving consensus among" philosophers that psychology must be banished from logic. What is in dispute is what his solid achievements were, what were his own decisive contributions to this evolution in philosophy of logic. Even his admirers cannot deny that there are some serious defects in Frege's case. The most obvious one is his employment of the tenns 'psychology' and 'psychological'. Officially these terms connote subjectivity. Frege tended to explain what he claimed to be subjective in terms of mental imagery and the association of ideas. He also held that what is subjective is incommunicable. These notions have odd consequences when combined with his using 'psychological' to label whatever he thought to be irrelevant to logic. The difference in coloring between 'and' and 'but', as well as the difference in word-orqer between 'No fISh are mammals' and 'No mammals are fish' and the difference in modality between 'Possibly every event has a cause' and 'Necessarily every event has a cause' are mistakenly held to belong to psychology and implausibly explained by reference to mental mechanisms.23 At best one can excuse 'these blunders as the by-products of inattention to the details of what is non-logical since Frege's positive identifICation of what is irrelevant to logic was both sound and sharp.24 Overextension of the domain of 'psychology' is matched by misguided
82
G.P. BAKER & P.M.S. HACKER
application of the terms 'logic' and 'logical'. This converse mistake underlies Frege's introduction of the judgment-stroke or the assertion-sign into his concept-script. He held that inferences could only be drawn from propositions asserted to be true.~ Consequently he treated a thought's being asserted as a logical property which had to be symbolized in a logical notation designed to express everything relevant to determining the validity of inferences. Once it is acknowledged that hypotheses or even false judgments have logical consequences, the assertion-sign represents an intrusion of psychology into logic. Frege's assumption that what is subjective is incommunicable is also clearly misguided. He admitted the possibility of expressing judgments about subjective objects (about aches and pains, feelings and emotions, even mental images). Once possessed of the sense/reference distinction he had a ready account of the possibility of objective thoughts about subjective objects, since the senses of every ,/ expression (including designations of mental entities and concept-words ascribing their properties and relations) are held to be objective. 26 Nonetheless, he still affirmed that there is no possibility of communication about what is subjective,27 and he envisaged subjective thoughts that could not be grasped by anyone other than the person expressing them. 2lI Frege here fell into an unresolved muddle. To these sins one might add the fallacy of inferring from the untenability of psychologism about judgments and concepts to the correctness of a Platonistic conception of these logical entities.29 Even if Frege was in no position to discern this error, one has the v right to object that there is manifestly an excluded middle! His case against the psychological logicians and his arguments for his own antithetical conception are by no means fault-free. His alleged achievements, however, are widely held to be substantial, to be decisively established and independent of these demerits, and hence to outweigh the flaws in his anti-psychologism. What were these accomplishments? Different commentators have credited many insights to Frege's account He frrst exposed the conflation of reasons with causes which is integral to psychologism in 10gic.30 He proved that communicability is essential to thoughts, Le., that every thought can be grasped in principle by anybody.31 His claim that the sense of expressions are objective implies the view that what senses speakers attach to words can be ascertained from their behavior,32 hence too that whether different persons understand words in the same way or not can always in principle be discovered.33 Frege perceived that assertoric force (which distinguishes making a judgment from merely expressing a thought) is a logical, not a psychological, feature of utterances; it is essential to a proper understanding of rules of inference, and its representation by the assertion-sign made clear for the ftrst time why in the argument pattern modus ponens (P, and if p, then q, SO q) the conclusion is not already asserted in asserting the conditions for the premise.'" He gave a defmitive criticism of the empiricists' theory of concept acquisition by abstracting from sense-experience.3S He demonstrated that meaning (or understanding an expression) has nothing to do with having
FREGE'S ANTI·PSYCHOLOGISM
83
mental images, pointing out the logical gap between an image and the employment of a word" This argument is held to be Frege's most important contribution to general philosophy, an anticipation of Wittgenstein's private language argument in the Philosophical Investigations.'n Frege's own theory of sense "gave the fIrst plausible account in the history of philosophy of what it is to grasp a thought or to understand a sentence as expressing one";31 thus the positive counterpart of his anti-psychologism was the construction of the fll'St satisfactory theory of understanding a language." Within the compass of his polemics against psychologism there are thought to be the seeds of many of the most important developments of the next hundred years in philosophy of logic and philosophy of language. The grandiose claims made on Frege's behalf envelope his arguments in anachronistic controversies and obscure a clear view of the strengths and weaknesses of his case against the psychological logicians. We shall try to start afresh from a sober statement of his reasoning and work our way gradually towards a proper assessment of his achievements. It will become clear then that many of the insights attributed to him are in fact inconsistent with the main guidelines of his thinking.
HI. Laws of Thought The primary target of Frege's anti-psychologism was the conception of the troths of logic or tile laws of thought which was advocated by Erdmann. The psychological logicians of whom he was the representative held logical laws to be descriptions of patterns of human thinking. Such fundamental principles as the Law of Identity and the Law of Non-contradiction are justified by inductive reasoning from data gathered by introspection or psychological experiments. Philosophers are held to have been mistaken in viewing these laws of thought as necessary truths or self-evident axioms grounding a deductively structured a priori science. They are empirical genenilizations about the workings of the human mind Two different conceptions of these genenilizations were available to radical empiricists. One took them to be the natural laws of mental phenomena; on this view they would be independent of time and culture just as the laws of Newtonian mechanics were thought to hold universally at all times and places. The other view saw them to be historical and anthropological / generalizations; accordingly, nothing could in principle rule out the possibility that laws of thought were subject to evolutionary development or varied in different parts of the world. Insofar as rules for constructing sound arguments and avoiding fallacies were treated by psychological logicians, they were considered to be technical noons grounded in observed empirical regularities. On one view they are analogous to directives about heating rooms which are based on physical laws; the guarantee of the utility of
84
L-'
G.P. BAKER & P.M.S. HACKER
modus ponens and the disutility of affmning the consequent must be the psychological impossibility of believing any proposition and its negation simultaneously. Alternatively, rules of deductive inference might be compared with dietary recommendations based on medical knowledge; on this view, observing these principles in thinking would promote the achievement and preservation of mens sana, while violating them would produce cognitive disorder and confusion. This whole conception of logic challenged the traditional consensus about the nature of the propositions of logic, and consequently it also conflicted with the framework of Frege's logical investigations. The notion that the laws of thought are natural laws governing psychological phenomena had many advocates. Sigwart championed this conception, declaring basic logical propositions to be laws of the functioning of human thought which must hold for all thinking beings endowed with the same nature as human beings.40 Boole shared this point of view. He sought a logical symbolism "the laws of whose combinations should be founded upon the laws of the mental processes which they represent,"4! and his aim in The Laws of Thnughl was "to investigate the fundamental laws of those operations of the mind by which reasoning is performed...and upon this foundation to establish the science of Logic."42 By contrast with these later psychological logicians, Erdmann adopted a more radical relativism (later called 'anthropologism' by Husserl). Although the laws of thought govern all human thinking, in his view, human nature might change, and in these conditions the laws of thought would change too. He held that we must concede the possibility of thinking essentially different from our own and the intelligibility of the hypothesis that the laws of thought might undergo evolutionary development: ...logical laws only hold within the limits of our thinking, with out our being able to guarantee that this thinking might not alter in character. For it is possible that such a transformation should occur, whether affecting all or some of these laws, since they are not all analytically derivable from one of them. It is irrelevant that this possibility is unsupported by the deliverances of our self-consciousness regarding om thinking. Though nothing presages its actualization, it remains a possibility. We can only take om thought as it now is, and are not in a position to fetter its future character to its present one.4]
The necessities of thought are thus viewed sub specie humanitatis; they are not unconditional necessities (as most logicians have thought), but rather generalizations conditional upon the actual (and in principle mutable) nature of human thinking. To be sure, we cannot think what nonlogical thought would be like, but this constraint lies within the empirical nature of our minds, not in the nature of the object of thought: ...We cannot help admitting that all the properties whose contradictories we cannot envisage in thought are only necessary if we presuppose the character of om thought, as definitively given in our experience: they are not absolutely
FREGE'S ANTI-PSYCHOLOGISM
85
necessary, or necessary in all possible conditions. On this view our logical principles retain their necessity for our thinking, but this necessity is not seen as absollUe, bill as hypothetil:al. We cannot help assenting to them- such is the nature of our presentation and thinking. They are universally valid, provided our thinking remains the same. They are necessary, since to think means for us to presuppose them, as long, that is, as they express the essence of our thinking."
Frege launched a frontal attack on Erdmann's anthropologism. He objected that this conception subordinated logic to psychology, transforming the science of the laws of thought to the cataloguing of regularities in human thinking. It misrepresented the laws of logic as contingent truths; it depicted them as subjective and potentially variable in content; and it eclipsed their regulative or normative role in the search for truth. At the same time, Frege embarked on a more indirect attack. He tried to trace the roots of these misconceptions to more general confusions about the nature of judgments and concepts, the objectivity of truth, and the process of concept-formation. Much of the strength (and weakness) of his anti-psychologism is apparent only when one follows up those ramifying lines of reasoning. Our initial enterprise will be to expound and evaluate his direct observations about the nature of the laws of logic. There are two main points in Frege's criticisms. The fIrst is the charge that the psychological logicians do not do justice to the normativity of the c..-/ laws of thought His argument now invites misunderstanding. It focuses not on rules of inference (as one might now expect), but rather on the status of the truths of logic such as the Law of Identity and the Dictum de Omni ('Whatever is true of all is true of each'). Psychologism represents these as descriptions of patterns of human thinking, often as psychological laws displaying the nature of the human mind. Frege held this to be inconsistent with the commonplace that logic ranks as a normative science alongside ethics and aesthetics!S The laws of thought must be acknowledged to be _ guiding principles for thought in the attainment of truth; they prescribe or ./ stipulate how one oug1l1 to think if one is to think at all.46 This feature differentiates these basic truths from descriptive generalizations: "The laws in accordance with which we actually draw inferences are not to be identifIed with the laws of valid inferences; otherwise we could never draw a wrong inference."47 Frege accused the psychological logicians of obscuring the fact that logic is concerned with truth just as ethics is with goodness / and aesthetics with beaUty.41 The normativity of logic is not held to be incompatible with representing the science of logic as a system of truths; on the contrary, both ethics and aesthetics, though clearly regulative, were thought to consist of general truths too (e.g., 'Theft is wrong'). What compromises normativity is conceiving of the generalizations of logic as being empirical descriptions, hence as answerable to the nature of something external (the human mind) and consequently liable to experiential refutation. Frege's conception was that this defect could be remedied only by acknowledging the laws of logic to be a priori truths. In this crucial
86
G.P. BAKER & P.M.S. HACKER
respect, he held, they are comparable to moral and aesthetic principles, and this is what he meant to convey in reiterating that logic is a nonnative science. Frege's second main criticism is that the psychological logicians distort ./' the nature of the laws of thought by treating them as contingent on human nature and hence (in Erdmann's version) as potentially variable. Frege held /' that the underlying mistake was a conflation of a proposition's being true with its being taken to be true. Truth, he stressed, is wholly independent of what is believed to be true; no matter how many people take something to be true, no matter what view is destined to be held in the scientific millennium, there is no contradiction in this proposition's being false and in its contradictory's being true.49 Truth cannot be reduced to being taken to be true. That fundamental point, in Frege's view, demolishes the claim that logical laws are contingent, empirical generalizations. The laws of logic are laws of truth, not laws of takings-to-be-true.~ Since truth is an objective property of a thought altogether independent of human thinking and opinion, and since truth is an invariant or immutable property of a thought, the laws of truth must be wholly unrelated to the nature of the human mind and as eternal as truth itself. They must be acknowledged to be necessary truths. The fundamental propositions of logic, the axioms from which all the truths of logic flow, are self-evident truths which are apprehended by the eye of the mind or the faculty of reason, not by inductive arguments based on empirical observations. They are completely general propositions which are known a priori to apply to all judgments and concepts. Frege summarized this view in a famous passage: H being lrue is thus independent of being acknowledged by somebody or other, then the laws of lruth are not psychological laws: they are boundary stones set in an eternal foundation, which our thought can overflow but never displace.'·
These two criticisms dovetailed together in Frege's rejection of the relativism conspicuous in Erdmann's anthropologistic conception of logic. Frege conceded that we might encounter beings who thought in ways flatly contradictory to our own. Such beings would be capable of bringing off judgments contradicting our laws of logic, e.g., judging an object to be different from itself.'2 The psychological descriptions of their patterns of thinking would differ dramatically from the empirical generalizations which describe our thinking. The psychological logician will not merely register these differences and explain them as manifestations of a different mental constitution; he will also conclude that they have one logic, whereas we have another one. Just as ethical relativists deny that there is an Archimedean point from which to evaluate and choose between alternative moral systems, so Erdmann maintained that these beings have laws of thought which are just as legitimate or valid as our own. Frege held this logical relativism to be doubly erroneous. It conflicts with the normativity of logic. Although we can imagine beings who reject the Law of Identity,
FREGE'S ANTI-PSYCHOLOGISM
87
we cannot suppose that these beings are right in so doing. A law of truth is a universal nonn of reasoning which prescribes how judgments are to be made, no matter where, when, or by whom. 55 Once we distinguish between the prescriptive laws of truth and the descriptive generalizations about how and under what conditions thinkers take things to be true, it must be an open question whether thinker's inferences are sound, i.e., whether these laws of taking-to-be-true coincide or not with the laws of truth. The laws of truth, in Frege's view, are the single universal criterion of correctness of inferences. Equally, the laws of truth are in principle timeless and immutable, self-evident truths certified by reason itself. One steps away from logic as soon as one considers human nature and external circumstances as detenninants of judgments; such considerations never bear--on something's being true, but only on our taking it to be true. 54 Like truth itself, the laws of truth are eternal and objective, independent of the nature of the judging subject. Erdmann's relativism in logic conflicts with the unconditional necessity of the laws of logic. Frege epitomized his objection by claiming that, on encountering beings who departed from the Law of Identity, we should say not 'Here we have different logical laws', but rather..-----/ 'Here we have a hitherto unknown type of madness'." The contentions which Frege opposed to Erdmann's theses about logic are undoubtedly correct, though unexciting. The propositions of logic are paradigms of necessary truths; it is nonsense to call them contingent generalizations. It is equally absurd to describe them as empirical statements which are confinned by observations of the functioning of the human mind; one does indeed step outside logic (or outside the bounds of sense) in offering any psychological data as justification for any proposition of logic. Finally, Frege was right to emphasize the normativity of logic; the propositions of logic are obviously relevant to the discrimination of valid and fallacious inference in virtue of somehow setting the standard of what is correct All these points are sound even if their restatement has little value in an era in which naturalist psychologism is not a serious disease among writers on philosophy of logic. Frege's own philosophical explanations of these commonplaces do not, however, advance our understanding of the nature of logic. He did nothing substantial to clarify the normativity of the propositions of logic. He apprehended that it is nonsense to conceive of empirical propositions as v regulative since they are responsible to facts; hence he mimicked the empiricist's strategy of treating ethical propositions as a priori truths and stressed that the source of logic is reason, not experience. But this manoeuvre leaves the regulative roles of logic and moral principles equally obscure. In fact, all that it accomplishes is to secure those propositions against the possibility of empirical disconflfJllation. That nonsense is excluded by classifying them as a priori truths. But Frege thought that the propositions of logic described relations among logical entities (the two truth values True and False, and the first- and second-level functions designated by propositional connectives and quantifiers). Does this not make -../
88
G.P. BAKER & P.M.S. HACKER
these proposlbons answerable to something external? What is held to compromise the nonnativity of the propositions of logic and ethics is the possibility of their being refuted by matters of fact But if the basic truths of logic are grounded in apprehension of relations among abstract entities, the possibility of their being refuted seems to re-emerge. How can one dismiss the possibility that the eyes of the mind might be subject to hallucination or that fresh "logical experience" might compel a revision to the fundamental truths of logic?" As conceived by Frege, a priori propositions seem to be as remote from regulative principles as empirical genenilizations are! His case against Erdmann's relativism seems equally shallow. Dropping from consideration Erdmann's claims about the source of knowledge of logical truths in psychological observations, we are left with the problem of dealing with the philosophical claims that there might be different and incomparable systems of logic (like different systems of morals). Frege evidently held that the laws of truth are unique and detenninate, and he would have denied the possibility of a genuine alternative logic Gust as he opposed the intelligibility of the supposition of non-Euclidean geometries). But did he indicate a procedure for adjudicating between alternative logics (e.g., between 'classical' and Intuitionist logic)? Or outline a method for settling whether his codification of the laws of truth is correct? According to his own conception of logic, these questions are intelligible and hence require answers. A determined relativist could argue that there is a discrepancy between what is truly self-evident and what Frege took to be the fundamental axioms of logic. Frege's claim that the laws of truth are standards of correctness of inferences gives no reason for preferring his own logical principles to every set of competitors, and therefore his opposition to alternative logics, by his own lights, must be judged to be mere dogmatism. These philosophical defects in Frege's position are symptomatic of three misconceptions that lie at the very foundations of his notion of the nature of logic. The fIrst two of them he shared with the psychological logicians whom he attacked. First, it was commonplace to distinguish within the science of logic between necessary truths ('Everything is what it is', 'Whatever is true of each is true of all', etc) and rules of reasoning (both rules for sound reasoning such as the syllogism in Barbara ('All A's are B's; all B's are C's; so all A's are C's') and rules for excluding fallacies (such as affmning the consequent». A similar distinction was drawn in moral philosophy between ethical truths ('Murder is evil') and rules of right conduct ('Thou shalt not murder'). That there are conceptual connections between the two kinds of truths in these cases is indicated by the traditional classification of logic and ethics as normative sciences, and equally by the persistence of disputes whether logic should be mnked as the science or the art of right reasoning. Since logical truths were thought to be substantial, informative propositions which can stand to one another in the relation of entailment, the only available conception of the connection between these and the rules
FREGE'S ANTI·PSYCHOLOGISM
89
of logical inference was that of the relation of a natural law and technical rules of mechanical or biological engineering. Rules of inference are held to stand to the truths of logic just as rules for positioning radiators and excluding draugh~ stand to the physics of heat-generation and transfer, or perhaps as dietary principles stand to generalizations about human metabolism and pathology. In all these cases, particular rules are justified as the most effective means for achieving certain goals, and these goals can be understood and described independently of the rules advocated for attaining them. Erdmann thought these parallels to be exact because he held that logical truths have the status of psychological laws. Frege would have seen only an analogy" since he denied that the truths of logic are empirical. Nonetheless, he conceived of rules of inference as norms which specify the most efficient means to achieve the goal of attaining truth; the goal can be characterized independently of these norms, and the norms themselves are justified by the objective logical relations between judgments which are codified in the propositions of logic of the Basic Laws of Arit1unetic. This conception of logic has two basic defec~. First, it depic~ the concept of bUth as independent of the concept of valid inference. In fact, however, that a judgment can be derived by certain transfonnation-rules from propositions known to be true often serves as a criterion for holding this judgment to be true. Truth is not related to valid inference as health is to diet, but rather as checkmate is to chess. Secondly, Frege's picture of rules of inference obscures just what distinguishes the propositions of logic from all other propositions. By casting these propositions in the same role as laws of physics, it hides from view the fact that they have a completely different function in speech. As Wittgenstein later discerned, they are tautologies, propositions which say nothing all.SI Hence, they cannot serve as grounds justifying technical norms which regulate human thinking; rather, the conceptual connections between tautologies and rules of deductive inference is that to acknowledge that a pattern of propositions exemplifies a rule of inference (e.g., that the scheme 'p, if P then q, so q' is a rule of inference) is to acknowledge that the corresponding conditional expresses a tautology (e.g., that the proposition 'if p and if p then q, then q' or 'p.p~.~.q' is a tautology)." Frege's unquestioned assumption that the laws of logic are (/ significant judgments (propositions with sense) forced upon him the misconception that rules of inference are technical nonns answerable to independent bUths, and this shared misunderstanding underlay his disagreement with Erdmann. Secondly, both Frege and Erdmann thought that the truths of logic must be distinguished from other significant generalizations in terms of their subject-matter. Indeed, both agree that these propositions were statements about concepts, judgments, and inferences; they held that logical laws describe relations among concepts or judgments, thereby fannulating constraints to be respected by any acceptable rules of inference. Disagreement arose from the further clarification of this common ground. Erdmann held concepts and judgments to be mental entities, and the laws
90
G.P. BAKER & P.M.S. HACKER
describing their inter-relations to be genenilizations belonging to empirical psychology. Frege, by contrast, thought concepts and judgments to be imperceptible but real entities in a logical realm, and the laws of truth to be a priori generalizations grounded in deductions from self-evident axioms. The most serious defects in their understanding of logic arose from the point not in dispute between Frege and Erdmann. First, both went disastrously astray in trying to distinguish logical propositions from others in virtue of their content; it is the function or use of the propositions which is diagnostic. Indeed, the propositions of logic, being tautologies, have no content at all. Hence, their function cannot be to say something distinctive (whether about the mind or about the realm of logical entities). The topic neutrality and universal applicability of the propositions of logic are both compromised by the very idea of their having a specific subject-matter. (As Wittgenstein later argued, Frege's view that logical operators designate special logical functions or concepts is incoherent.~ The true function of logical truths is apparent in the reflection that to show that a fonnula is a tautology is to exhibit an internal relation among propositions (e.g., proving that p.p~.--+.q is a tautology is equivalent to showing that q follows from p and p--+q). Secondly, Frege and Erdmann had a corresponding misconception of the role of proofs within an axiomatization of the propositions of logic. They thought that these proofs established the truth of one significant proposition from the previously established truth of other significant propositions. In fact, these proofs demonstrate that a formula is a tautology by deriving it by a sequence of authorized transfonnations from other fonnulas that are tautologies.61 To apprehend this function of derivations within axiomatizations of logical truths is to undennine Frege's case for supposing that the propositions of logic are significant formulae having content or sense. Both Frege and Erdmann lost their grip on the nature of logic before their detailed reasoning even began! Thirdly, although Erdmann (and Husserl) had only a vague and distorted grasp of the connection between the concept of logic and the concept of thinking, Frege went further off course by ignoring this conceptual connection altogether. This is evident in his concession (to Erdmann) that we might encounter beings who think in ways incompatible with our laws of thought: such beings might bring off judgments contradicting the Law of Identity. Frege noted that this would be "a hitherto unknown type of madness." But, as Wittgenstein later objected,62 "he never said what this 'madness' would really be like." What would it be to "bring off a judgment contradicting the Law of Identity"? Frege did not describe what the imaginary beings must do to be judged to think: contra-logically. If they said 'a~', would not we rightly conclude that either they misunderstood ':1-' or meant something other 'than identity by this symbol? Could they think: or even try to think that this book is not identical with itself! Could they express what they think? How, if 'a~' either expresses nothing or fonnulates a different thought? The attempt to describe a case of thinking
FREGE'S ANTI-PSYCHOLOGISM
91
contrary to the Law of Identity cannot so much as get started. It is akin to the attempt to describe what it would be to castle in dmughts or to score a touchdown in bridge. Frege made a mistake in presuming that the concept of thinking can be divorced from the concept of the laws of logic. What we call rules of deductive inference partly defme what we call 'thinking', 'speaking', 'arguing', 'saying the same thing', 'contradicting oneself, etc. Though the laws of logic might be called expressions of our thinking habits Gust as the rules of chess show how we play chess), they are also expressions of our habits of thinking Gust as the rules of chess show what we call 'playing chess')." Frege was wrong to imagine that beings who regularly violated the basic laws of logic in speaking can be said to engage in what we call 'thinking'. The impossibility here is not one that might be circumvented by exercises, by striving every morning to conceive six impossible things before breakfast. It is a grammatical impossibility comparable to playing chess while castling through check, moving rooks diagonally, advancing pawns to the ninth rank, etc. Frege distorted the concept of thinking by divorcing it from the concept of logic. (And thereby he deprived himself of the possibility of arriving at an explanation of the absurdity of imagining alternative logics.) The weaknesses which are apparent in Frege's criticisms of Erdmann's conception of logic lie at the very foundation of his own philosophy of logic. To eradicate them would be to throw his entire clarification of the nature of logic back into the melting pot. Moreover, in discarding the idea that the propositions of logic have content or sense, Frege would have had to abandon his fundamental idea that function/argument analysis is the key to the treatment of propositional connectives and quantifiers.
IV. The Objects of Judgment Despite apparent agreement on taking the subject-matter of logic to be concepts and judgments, the psychological logicians fell into the misconception that the laws of logic are psychological generalizations about human thinking because, in Frege's view,they mistook concepts and judgments for mental entities. Consequently Frege's task of rooting out their errors demanded a clarification of the nature of concepts and judgments and a critical examination of the arguments purporting to show that concepts and judgments are mental objects. His primary concern was to establish a proper understanding of what judgments are since each rule of deductive inference lays down a norm for the derivation of a judgment from other judgments. Like many other logicians, Frege drew a distinction between ~ mental acts of making judgments and the objects of which these acts are directed. As an instance of this kind of act he might cite his judging on 1st January 1879 that the square of the hypothenuse of a right-angled triangle is equal to the sum of the squares of the other two sides; in this case what he
92
G.P. BAKER & P.M.S. HACKER
judged, Le., the object of his judgment, would have been the Pythagorean Theorem (i.e., that the square). It is the objects of judgment (canonically he originally called fonnulated in indirect statements) that ./ 'judgeable-contents' and later rechristened 'thoughts' (sed the claim that singular judgments could be about any objects whatever, irrespective of whether these objects are concrete or not, or whether they are subjective or objective. The sentences 'Jupiter has four moons', '2 is a prime number', and 'My idea of Bucephalus is not very vivid' express judgments about different kinds of objects. Frege's official explanation of this diversity in what judgments may be about turns on his analysis of judgments into functions and arguments. Although every judgment is an abstract object, it can be the value of a function for an argument which is not an imperceptible, objective, non-spatio temporal object. For it is a fundamental feature of the concept of a function that neither its argument nor a function itself is a part of the value of the
FREGE'S ANTI·PSYCHOLOGISM
107
function for this argwnenL This is evident from the simplest case of mathematical functions: it would be absurd to claim that either the number 2 or the fimction x2 is a part of the number 4 on the ground that 4 is the square of 2. Focussing on this point removes any tension from claiming that the planet Jupiter or my idea of Bucephalus can be the argument of a function (concept) whose value is an abstract object Gudgment). One simply abjures the thesis that what a singular judgment is about (an object) is a part of the judgment itself. It seems integral to this project of replacing subject-predicate analysis in logic by function/argument decomposition that part/whole analysis be dropped. Unfortunately Frege found it impossible to make a clean break with this tradition, and the result was a steady trickle of confused theses and fallacious arguments. Originally he called judgeable-contents 'combinations of ideas', 12S and he described objects, concepts, and relations as parts of judgeable-contents. l26 This seems bewildering. How can a concrete object, say Priam's house, be part of an abstract object, viz., the judgment that Priam's house was built of wood? If Frege had drawn no consequences from this discourse about parts and wholes, the obscurity here would have been hannless. But it had the potential to lead his own thinking astray. In particular, it led him to overstate the thesis that he offered as the antithesis to the psychologicians, claim that all judgments are about ideas. He saw that their claim stemmed from the notion that a judgment is a psychological entity, since the subject of a judgment must have the same subjective status as the judgment of which it is a part. l27 The corrective is to acknowledge that "the content of a judgment is something objective, the same for everybody, and as far as it is concerned it is neither here nor there what ideas men have when they grasp it."I28 But adding "What is here being said of the content as a whole applies also to the parts which we can distinguish within it," Frege succumbed to the temptation laid up in part/whole terminology and contradicted his own contention that judgments may be made about subjective objects. Similar, more pernicious confusion persisted after Frege introduced the sense/reference distinction and divided judgeable-contents into truth-values and thoughts. At the level of reference, he analyzed judgments into objects and concepts; every proposition presents a truth-value (the reference of the proposition) as the value of a function (concept) for certain arguments. At frrst, Frege retained talk of parts, stating that the objects and concepts designated by the constituents of a proposition are parts of the truth-value which the whole proposition designates; but he added the caveat that 'part' here must be understood in a special sense. l29 Later he decided that it is confusing to speak of 'parts' in this context. 130 At the level of sense, Frege also decomposed what a sentence expresses (a thought) into function and argument(s); indeed, this decomposition is precisely parallel to the function/argument analysis of the corresponding truth-value, and therefore a fonnula in concept-script serves a double function. 131 Here too Frege retained the tenninology of 'part' and 'whole'. He claimed that the senses
108
G.P. BAKER & P.M.S. HACKER
of the parts of formulae in concept-script are parts of the thought expressed by the whole formula. l32 This led him to draw two conclusions inconsistent with the function/argument analysis of thoughts. First, he argued that Mont Blanc, with all its rocks and snowfields, could not be a part of his thought that Mont Blanc is higher than 4000m, and therefore he concluded that the sense of 'Mont Blanc', which is a part of his thought, must be distinct from the reference of this name. l33 This argument rests on the principle that no concrete object (no mountain) can be part of an abstract object (a thought). This would be a fallacious argument (parallel to concluding that Mont Blanc cannot be part of the True, which is the reference of 'Mont Blanc is higher than 4000m') unless the term 'part' is taken to exclude taking the mountain Mont Blanc to be the argument of a function whose value is a thought. Second, Frege eventually adopted the notion that a thought is composed out of thought-building-blocks (Gedankenbauseine), and he offered this thesis as an explanation of how it is possible to grasp new thoughts. l34 This conception depends on the literal truth of the claim that the parts of a sentence have senses which are parts of the thought expressed by the whole sentence. It is, however, incompatible with the guiding ideas of function/argument analysis of thoughts. Neither a function nor its argument are parts of the value of the function for this argument (save accidentally); and any representation of an object (a thought) as the value of a function for an argument is logically equipollent with any other representation of the same object as the value of another function for other arguments Gust as 4 is no more essentially the value of x2 for the argument 2 than it is the value of x + y for the pair of arguments (3,1». Frege never fully emancipated himself from the influence of the part/whole analysis of judgments which he had diagnosed as the root of the psychological logicians' confused thesis that all judgments must be about ideas. This failure manifests the extreme difficulty of extirpating deep philosophical misconceptions.
VI. Concepts and Concept Formation. According to the psychological logicians, it is an immediate consequence of the thesis that judgments are complex ideas that all concepts are ideas. For concepts are parts of judgments, and therefore the claim that judgments are subjective or psychological entities implies that their proper parts have the same status. Indeed, the predicate of every judgment is held to be a concept, while the subject of some judgments are also concepts (as in the judgment that all men are mortal).13S So every judgment contains at least one concept as a constituent or part, and this must be an idea. Although the term 'idea' ('Vorstellung') and 'concept' ('Begriff) do not have sharply contrasted uses or even consistent implications in the writings of philosophers and psychologists, Frege followed the psychological
FREGE'S ANTI-PSYCHOLOGISM
109
logicians in employing the term 'idea' to signify an inner mental picture. l36 Hence, their thesis that concepts are ideas is equivalent to the claim that what concept-words signify are mental images. In this jargon, the psychological logicians maintained that the content ('Inhalt') of a concept-word, or the sense ('Sinn') of such a word, is always a mental image. In the judgment expressed by 'All men are mortal', both the tenns 'men' and 'mortal' are concept-words which stand for ideas, according to the psychological logicians, and therefore the claim that this proposition expresses a relation between two concepts boils down to the thesis that it states a relation between two mental images. Likewise, the judgment expressed by 'Socrates is mortal' states that the properties of the image associated with the concept-word 'man' are all properties of the image associated with the name 'Socrates' .137 This use of 'idea' as a semi-technical term stresses that what are properly called ideas are essentially mental in nature, and this subjectivity of ideas is understood to carry the implication that "one person's ideas are his and no one else's."13I Frege linked this conception of concepts with two other fundamental theses, both of which were accepted and even espoused by the psychological logicians themselves: ' i) Judgments are held to be put together of a stock of independently grasped, ready-made or pre-existing ideas. l39 This is an immediate consequence of seeing judgments as constructed out of subjects and predicates, and it is clearly implied in calling judgments 'complex ideas' or 'combinations of ideas'.140 ii) The fundamental form of concept-fonnation is held to be abstraction from experience. Concepts may be broken down into simpler concepts by analysis, namely by means of definitions per genus et differentiam, Le., by specifying characteristic marks (lrlerkmale). The atoms left after this mental chemistry will be "internal images arising from memories of sense impressions which I have had and acts, both internal and external, which I have performed."141 These images are held to be fonned from particular experiences by selective inattention to their specific individuating characteristics. To arrive at the concept 'green', one must start from having perceptions of leaves, blades of grass, yew trees, etc.; one abstracts from (or disregards) features other than color (e.g., shape, size, texture) and thereby obtains an idea associated with ' green, (perhaps a composite picture, perhaps a picture of a particular shade of green which is used to represent the whole range of shades of green). The idea of a concept may differ only in degree from the idea of an object, and more and more schematic or abstract concepts can be generated "by making one characteristic mark after another disappear."142 All concepts are held to be derived, immediately or indirectly, by this process of abstraction. Having appropriate experiences, whether perceptual or introspective, is the necessary condition for fonning simple concepts ("A congenitally blind person cannot have concepts of colors"), and the process of concept-formation is independent of, indeed
110
G.P. BAKER & P.M.S. HACKER
presupposed by, making judgments. The psychological logicians inherited this doctrine of 'abstI'actionism' from classical empiricism. Frege argued that claiming all concepts to be ideas is just as abswd as claiming all singular judgments to be about ideas. His reasoning was identical in both cases. The underlying premise is manifestly false. It is not the case that judgments are ideas. Hence, even if concepts were parts of judgments, it would not follow that concepts must be ideas (or indeed any other species of subjective or psychological entity). Acknowledging the objectivity of judgments pulls the rug out from under the psychological logicians. After introducing the sense/reference distinction, Frege had to modify both his account of what the psychological logicians claimed and his rebuttal of their view. It is the sense of a sentence (a thought) which they mistakenly took to be something mental, indeed to be a complex idea. Hence, it must be the senses of proper names and of concept-words, i.e., what Frege conceived as the parts of thoughts, which they held to be ideas. (Since Frege used the term 'concept' for the reference of a concept-word, not its sense, their thesis cannot now be formulated by claiming that all concepts are ideas.) Therefore, the refutation of the psychological logicians' conception of concepts turns on proving that the senses of some conceptwords are not mental images. Frege proceeded to argue in just this way, distinguishing the sense of a sign from the associated idea. 14! But, in any case, the only argument available to 'the psychological logicians for concluding that these thought-eonstituents are uniformly ideas must start from the premise that all thoughts are subjective entities, and this claim is itself absurd. Consequently there is no prima facie case in the need of rebuttal. Once again Frege wins on the verdict nolo contendere. Once again it is crucial to note that this reasoning is a coda on a reductio ad absurdum. The underlying conclusion was the thesis that not all thoughts or judgments are ideas. Consequently, in Frege's view, t)lere is no support for the claim that all concepts (or the senses of all concept-words) are ideas. On the contrary, it is evident that some concepts (and the senses and references of some concept-words) are not ideas. For there would be a manifest inconsistency in conjoining the objectivity of the judgment that Socrates is mortal with the claim that one part of this judgment (the concept mortal) is an idea (hence subjective). The concept mortal must be as objective as the object (Socrates) which is the subject of this singular judgmenL For a similar reason, Frege later held that the sense of the concept-word 'mortal', like the sense of the proper name 'Socrates't must be objective; otherwise the thought expressed by the sentence 'Socrates is mortal' could not be objective since the senses of its constituents are parts of the sense of the whole sentence. l44 Consequently, Frege concluded that the psychological logicians fell into absurdity by claiming that all concepts are ideas. Both before and after introducing the sense/reference distinction Frege took care to formulate his posi~o~_~_!!t~_~I!~__of _ ~_ !1Oiyersal_
FREGE'S ANTI-PSYCHOLOGISM
111
generalization. Thus he claimed that the content of a concept-word need not may like may this celebrated early passage:
be an idea, but he did not exclude the possibility that it might be. "It be that every word calls up some sort of idea in us, even a word 'only'; but this idea need not correspond to the content of the word; it be quite different in different men. "145 The same caution informs
That we can form no idea of its content is therefore no reason for denying all meaning to a word....We are only imposed on by the opposite view because we will, when asking for the meaning of a word, consider it in isolation, which leads us to accept the idea as the meaning. Accordingly, any word for which we can find no corresponding mental picture appears to have no content But we ought always to keep before our eyes a complete proposition. Only in a proposition have the words really a meaning. It may be that mental pictures float before us all the while, but these need not correspond to the logical elements in the judgment 146
In early texts which consider the thesis that the content of a concept-word (a concept) must be an idea, Frege avoided asserting that no concept could be an idea or mental image. He manifested the same circumspection in some later discussions of the senses of concept-words. In approaching the distinction between ideas and the parts of thoughts corresponding to concept-words, he remarked that: The idea is subjective: one man's idea is not that of another....This constitutes an essential distinction between the idea and a sign's sense, which may be the common property of many people.... 147
Indeed, he added, there remains a difference between the mode of of an idea and of a sense with a word: two men "are not prevented from grasping the same sense; but they cannot have the same idea."148 Frege's objection seems always the same: If the sense of every concept-word were an idea, then in principle no thought could be shared and hence any inter-subjective science would be impossible. The conclusion to be drawn is that the premise must be false, i.e., that the senses of some concept-words are not ideas. This does not exclude the possibility that the senses of some concept-words are subjective, say perhaps that the sense of 'red' cannot be ascertained exactly by any other person when this adjective is applied to phenomenal colors within the sphere of a single consciousness. 149 On the other hand, Frege sometimes veered towards a stronger thesis. This is not evident in his criticisms of the psychological logicians, but it is conspicuous in his explanations of the conception of a concept on which he rested his logical analysis of inference. In early writings he identified concepts with functions whose values are always judgeable-contents. Though he did not argue the point, it is difficult to see how an idea (mental image) could be identified with such a function (i.e., how an idea could be a law conn~tion
112
G.P. BAKER & P.M.S. HACKER
for correlating each argument with a judgeable-content). Similarly, in later writings, he identified the sense of a concept-word with a function whose values are always thoughts. These functions (laws of correlation) must be as objective as their arguments (e.g., the senses of proper names of objective objects) and their values (objective thoughts), In his view, this alone would prevent the identification of such a function with any mental image. 150 This strand of thinking inclined Frege towards advancing the claims that the sense of a concept-word cannot be an idea: Certainly it is possible in making not the sense of the sentence. It and the same sense of play of logically irrelevant accompanying take for the proper object of their
a judgment...a play of may be observed that ideas can be wholly phenomenon that [the study.151
ideas occurs; but that is with the same sentence different and it is this psychological] logicians
Anyone who hears 'the word 'horse' and \D1derstands it will probably straightaway have a picture of a horse in his mind. This picture, however, is not to be confused with the sense of the word 'horse' .152
These claims rest on considerations supplementary to Frege's central reductio of the position of the psychological logicians, and the two strands of his thinking were never reconciled. In fact, he did offer supplementary arguments to support his case against the psychological logicians' conception of concepts, and some of these arguments suggest that the content or sense of a concept-word cannot be a mental image. i) In early writings Frege stressed that concepts must generally be objective entities, unlike ideas. In his view: Sensations are absolutely no concern of arithmetic. No more are mental pictures, formed from the amalgamated traces of earlier sense-impressions. All these phases of consciousness are characteristically fluctuating and indefinite, in strong contrast to the defmiteness and fixity of the concepts and objects of mathematics.153
To identify concepts with ideas would remove the objectivity of judgments and the timelessness of the truths of mathematics. For a generalization such as 'Every equilateral triangle is equiangular' expresses a relation between concepts; hence, if concepts were all ideas, it would state that two ideas in the speaker's mind stand in a certain (subjective) relation to each other,tS4 and this relation might depend on the phosphorus content of the human brain, and hence vary from person to person or be weeded out by natural selection in the struggle for existence. 155 Similarly, the proposition 'The number three falls under the concept of a prime number' is an objective truth, independent of the contents of my mind or others' minds and not subject to gradual change or historical development. l56 In short, the objectivity of mathematical judgments entails the objectivity of mathematical
FREGE'S ANTI·PSYCHOLOGISM
113
concepts. But the possibility that concepts applicable to subjective objects might be ideas is not excluded. ii) In early writings Frege also concluded that mental images are generally distinct from and irrelevant to the content of concept-words. This tmns on his limitations of conceptual content to what the cogency of inference depends on: Time and time again we are led by our thought beyond dte scope of om imagination without thereby forfeiting the support we need for our inferences.... That we can form no idea of its content is dterefore no reason for denying all meaning [content] to a word lSI
At this stage Frege did not advance the stronger conclusion that mental images have no legitimate role in determining the validity of any inference. Later he made this move in segregating sense from coloring (i.e., thoughts from their psychological trappings). For logic only sense matters; mental images, held to be variable and impossible to pin down precisely by inter-personal comparisons, are always allocated to coloring and thought to contribute only a poetic effect. Hence, he concluded that the sense of a concept-word cannot be an idea. iii) Frege queried the intelligibility of the claim that any concept is an idea. In expressing a thought by uttering the proposition 'This blade of grass is green', one predicates the concept green of an object. A psychological logician would claim that here my idea of green is asserted. of some thing (my idea of this blade of grass). But, Frege objected, it is altogether opaque how an idea can be asserted of anything. lSI A parallel argument applies to the sense of a concept-word, which a psychological logician must treat as an ingredient idea in the composition of a thought: Thoughts are fundamentally different from ideas (in the psychological sense). The idea of a red rose is something different from the thought that this rose is red. Associate ideas or run them together as we may, we shall still fmish up with an idea and never with something which could be true. lS9
By heaping associations on associations, we may arrive at a complex-idea,
but this would no more be a thought: l60 ...than an automaton, however cunningly contrived, is a living being. Put something together out of parts that are inanimate and you still have something inanimate. Combine ideas and you still have an idea and the most varied and elaborate associations can make no difference.16l
Frege here diagnosed a fundamental confusion in the psychological logicians' calling thought 'complex ideas'. For they must also apply the label 'complex idea' to the sense of a concept-expression (e.g., 'x is a red rose') and the sense of a concept-word which can be analyzed into characteristic marks (e.g., 'x is a triangle'). This obliterates the crucial
114
G.P. BAKER & P.M.S. HACKER
distinction between the expressions of thoughts and all other expressions, and it generates an insoluble problem about the unity of thoughts (viz., explaining why some combinations of ideas fonnulate judgments while others do not).162 iv) Frege criticized various aspects of the doctrine that concepts are fonned by abstraction from experience. He rejected the strong thesis that all concepts are derived in this way. This would make it impossible to form a concept under which nothing fell (e.g., the concept unicorn or even prime number grater than jOOO). But there is nothing amiss with empty concepts (or even with concepts that can be proved to be empty,163 and the fonnulation of such concepts is a matter of arriving at them, not by direct abstraction, but by starting from defining characteristics. l64 Since classical empiricists and the psychological logicians typically conceded the possibility of introducing complex ideas by analytic defmitions, Frege's weak claim that not all concepts are derived by abstraction would raise little opposition. He never advanced the strong thesis that no concepts can be formed by abstraction from experience. 165 On the contrary, he regularly acknowledged the legitimacy of this form of concept-formation for many concepts. In particular, he held that we arrived by abstraction at concepts of properties of external things (e.g., concepts of color, weight and hardness l66) and of internal things (e.g., the concept red which is applicable only to the content of one's own consciousnessl ' ) . He never argued that abstraetionism is a wholly bankrupt account of concept-formation, and he never raised some of the standard difficulties that it must confront (e.g., differentiating the relational concept larger than from the converse concept smaller than).I68 But Frege did defend a thesis of intermediate strength: there are some concepts the formation of which defies explanation by abstraction from experience. In particular, he argued that the concepts of nwnbers as used in count-statements (e.g., 'Jupiter has four moons') cannot be so explained. Classical empiricists and the psychological logicians took these adjectival occurrences of numerals to signify properties of objects, and hence they explained how successive operations of abstraction generate the concept four out of the experience of seeing four cats of various breed, size and color. l69 Frege reasoned that this explanation was acceptable only if it applied to every use of numerals to answer the question 'How many.•.are there?'; hence that it must apply to the number zero. But it is manifestly absurd to try to form an idea of zero visible stars by starting from perceiving a property of something external. Seeing a sky entirely overcast with clouds would give nothing corresponding to the word 'star' or to 'zero', and a mental image of an overcast sky would be equally remote from an articulated complex idea of zero visible starS. I70 Frege offered fwther arguments to deny that any idea of number can be arrived at by selective inattention to features of what we experience. 171 In his view, this short coming is irremediable. Numerals used in count-statements do not signify properties of objects; rather, they function as parts of expressions predicating (second-level) properties of concepts (or fIrst-level properties).
FREGE'S ANTI·PSYCHOLOGISM
115
The distinction between first- and second-level concepts is one of the crucial logical distinctions annihilated by the psychological logicians' identification of concepts with ideas, and hence it is no wonder that the fonoation of the second-level concepts deployed in count-statements cannot be explained by abstraction from the experienced properties of objects. (It is an additional demerit of the abstractionist theory of concept-formation that it blurs the logical distinction between concepts and objects by presenting it as a matter of the degree of specificity in the itemizing of something's characteristics.)172 Frege's criticism did expose serious flaws in the conception of concepts as ideas in spite of the fact that he shared with his adversaries some basic misconceptions 'about the nature of ideas. He was surely correct in claiming that the psychological logicians obfuscated important logical distinctions by taking proper names, concept-words, and sentences alike to signify ideas. Serious confusion is invited by calling both thoughts and analyzable concepts 'complex ideas'. Frege was also right to exclude consideration of mental images from assessments of the validity of inferences, hence to separate the senses of the expressions from the ideas associated with words. Finally, he deserves credit for clarifying the grammar of generalizations (both existential and universal) and of count-statements by describing them as predicating properties of properties and hence as containing second-level concepts,173 and he rightly noted that the notion that all concepts are ideas is a major obstacle to achieving this insight. On the other hand, Frege was light-years away from offering a general criticism of the doctrine of concept-formation by abstraction. It would be wholly erroneous to suggest that he distinguished the sense of a concept-word from the associated mental image because he anticipated Wittgenstein's insight that it is nonsensical to suppose that mastery of the entire use of a concept-word somehow flows from correlating this expression with something static like a picture. Moreover, though he recognized a crucial logical difference between proper names and conceptwords, he did not clarify it by examining in detail how these expressions are used and explained in everyday linguistic practice, but sought to explain it by pointing to a metaphysical divide between the entities allegedly signified by these expressions, Le., the dichotomy between concepts and objects which is held to be founded deep in the nature of things. With the wisdom of hindsight, these are decisive weaknesses in his philosophical investigation of concepts. Frege was not content merely to expose the psychological logicians' misconception of concepts; he offered in its place an allegedly proper logical notion of a concept purged of all psychological excrescences. Like his positive explanation of the nature of judgments or thoughts, this explanation of what concepts really are generates further philosophical perplexities. In his early writings he arrived at the 'discovery' that concepts belong to the genus of functions. 174 In fact, a concept (in a logical sense) just is a
116
G.P. BAKER & P.M.S. HACKER
function whose value is always a judgeable-content. (What he called a 'function' in Begriffsschrift is identical with what he called a 'concept' in the Foundations of Arithmetic.) This had two important corollaries. First, it involved an unorthodox extension of the application of the term 'concept'. In particular, he categorized logical expressions, both propositional connectives such as 'and'. 'or', 'ir, 'not', and quantifiers such as 'some', 'all', as concept-words (respectively of flfSt and second-level); by ttadition they had been called syncategoremata and contrasted with concepts. He also called any expression derived from a logical formula (or sentence) by functional absttaetion the name of a concept; e.g., by regarding 'Socrates' as replaceable by other expressions in the sentence 'If Socmtes is a man, he is mortal', Frege arrived at the concept expressed by the sentence-schema 'If x is a man, x is mortal'. Hence, what he acknowledged to be concepts were not restricted to what general words like 'man' and 'mortal' signify; and the judgment that all men are mortal can be analyzed by stating that a (complex) concept has the property of holding for every argument (rather than necessarily stating that a pair of concepts stand in the relation of subordination). These differences, though subtle, are of the last importance. The second corollary is the surprising claim that concept-formation (in the logical sense) must always start from judgeable-content. 17s There is no possibility of arriving at a concept in any other way, a fortiori not by any process of absttaction applied to the data of experience. 176 This idea informed one of his three fundamental logical principles: viz., "never to ask for the meaning of a word in isolation, but only in the context of a proposition. "ITI This early positive conception of a concept must be less incoherent than the conception of a judgeable-content on which it is built For, in Frege's view, the idea that a concept is a function cannot be separated from the thesis that there are objects to serve as the values of these functions, and the objectivity of concepts presupposes that these objects be objective, i.e., numerically the same objects confronting us all like the SUD. If this background makes no sense, then his description of concepts as functions whose values are judgeable-contents is itself nonsensical. In addition, his extension of the word 'concept' to include the content of logical expressions has the consequence that logic has the status of an a priori science about special logical entities; but the idea that logical propositions have a distinctive subject-matter is inconsistent with recognizing them to be tautologies which say nothing at all. After making the sense/reference distinction, Frege confined the term 'concept' (in its logical sense) to the reference of a concept-word. He defined a concept as a function whose value is always a truth value. 17! Any word or expression which is held to stand for such a function is called a concept-word, and hence one again he extended the application of the term 'concept' to cover whatever is signified by a logical expression or any sentence-schema with schematic letters. Moreover, his usage of 'concept' was now incompatible with standard practice among logicians, for he held e
FREGE'S ANTI·PSYCHOLOGISM
117
that a function must be completely defined by determining its value for every argument, and this has the consequence that the extension of any concept can be calculated from the natw'e of the concept itself! 179 The sense of a concept-word Frege thought to be a function whose value is always a thought This notion is rendered incoherent by the nonsensical presupposition that a thought is an objective object, numerically the same for different thinkers. It also demands that logical expressions have senses, and hence too that the propositions of logic express genuine thoughts distinct from one another. Consequently, the positive doctrine that Frege later juxtaposed to the psychological logicians' conception of concepts was no real improvement on his earlier one. In his criticisms of the psychological logicians' conception of concepts, Frege appeared at his philosophical best His arguments do put a finger on subtle but fundamental confusions in his opponents' thinking, and the inconsistencies in his own position are so deeply hidden that they still pass unremarked. It would be a mistake to suppose that he could be defended against serious criticism by exploiting the idea that he introduced a special semi-technical concept of a concept. Though advantageous for sidestepping objections, This would impose on him the burden to give an intelligible and consistent explanation of what he meant to be understood by 'concept', and this task is one that he did not discharge. lao Moreover, if he had succeeded in explaining his own technical notion of a concept, he would thereby deprive most of his philosophical remarks of their point. In claiming 'that logical laws are generalizations about concepts and judgments, that ' All men are mortal' expresses subordination of concepts, that 'there are prime numbers' states that the concept of a prime number has a (second-level) property, or that count-statements predicate properties of concepts, did he not intend his remarks to be understood by philosophers and logicians independently of his giving an idiosyncratic defmition of 'concept'? And would he not have imposed on himself the burden of proving that any of the expressions standardly called 'concept-words' signify what he called ,concepts'? Indeed, it is arguably unintelligible to claim that any expression stands for a function whose value is always the True or t~ False. Finally, would applying his semi-technical notion of a concept in interpreting the psychological logicians' thesis that concepts are ideas not deprive his arguments of any real targets in the thinking of Erdmann and the early Husserl? For would it be credible to take them to have held functions whose values are always judgeable-conmDts (or thoughts) are to be \identified with mental images? To fall back on a semi-technical notion of a lconcept might purchase advantage in a few local skirmishes with critics, but it would lose the overall war quite decisively.
118
G.P. BAKER & P.M.S. HACKER
VII. Frege's Legacy We must now try to give a synopsis of Frege's polemic against the intrusion of psychology into logic and to evaluate his philosophical achievement in this domain with his general philosophy of logic. The focal point of his reflections was the nature of the propositions of logic, and the ramifying chains of his arguments should always be traced back to this source. This methodological principle for interpreting his antipsychologism should always be borne in mind. Having pursued his reasoning through the laws of thought down to objects and concepts, we are ourselves perhaps in danger of losing sight of the fact that our more recent investigations are subordinated to the clarification of the nature of logic. The load-bearing members of his thinking are his fundamental thoughts about logical laws, not miscellaneous and independent ideas about judgments, concepts, and objects. A second principle of interpretation is that Frege's anti-psychologistic theses must be understood as directed at very specific targets and as supported by his arguments. His arguments are, in a sense, ad homines, not contra mundum. It is ridiculous to look in his writings for a systematic refutation of idealism in all its forms. He did not undertake any examination of representative idealism in classical British empiricism, of Kant's transcendental idealism, or of the forms of Hegelian idealism that concerned Moore and Russell; nor did he embark on the refutation of scepticism about the existence of an objective domain of perceptible objects. His target was a doctrine about the nature of the truths of logic and arithmetic which was championed by Erdmann and the early Husserl. His anti-psychologism is intended to be a refutation of their case that these so-called 'necessary truths' are descriptions of contingent relations among psychological phenomena. To this misconception about logic Frege opposed the strong thesis that logical laws are significant propositions stating generalizations about special logical concepts and that they can all be deduced rigorously from an initial set of axioms stating self-evident a priori truths about entities in a logical realm. But the rest of Frege's case against the psychological logicians has the form of disproving their mistaken genemlization on which their conception of logic rested. Here, as we have emphasized, each argument is a reductio ad absurdum, and therefore Frege's antitheses to the psychological logicians' position are prima facie weak. Without supplementary evidence from his writings each antithesis should be read as a negation of a generalization, not as the assertion of a contrary genemlization. Moreover, the content of each of his counter-claims must also be related to the supporting argument. There is a strong temptation to separate his theses from the argumentative context in which they are set. But to inflate the conclusions of his reasoning in this way invites the criticism that his supporting arguments are fallacious and the scope of his reflections 100 narrow. It even opens the way to convicting him of internal
FREGE'S ANTI·PSYCHOLOGISM
119
inconsistency. To ascribe to him the central insight of Wittgenstein's arguments against the possibility of a private language would knock the props out from his argument against the psychological logicians, for it would demolish his premise that ideas, being subjective, have properties and relations which can be exactly ascertained only by their owners and it would equally render unintelligible his drawing a distinction between qualitative and numerical identity for both thoughts and mental images. Strengthening or decontextualizing Frege's anti-psychologistic theses in the name of charity has the consequence of diminishing his solid philosophical achievement Bearing in mind these two fundamental principles of interpretation, we must at last try to evaluate the achievements of Frege's anti-psychologism. It should be perspicuous that most of the general claims made on his behalf are extravagant and groundless. He did not undertake, let alone accomplish a comprehensive examination and refutation of idealism in all its forms. Far from anticipating Wittgenstein's private language argument, his reasoning against the psychological logicians depended on applying the distinction between numerical and qualitative identity to thoughts and ideas (where it is not intelligible). His distinction between content or sense and mental images (ideas) twned solely on the role of symbols in reasoning and calculation, and hence it did not amount to a refutation of imagism as a theory of meaning. Both his ambitions and the scope of his arguments were strictly limited. It is also clear that his arguments are shaped by presuppositions that he shared with his opponents and that many of these unexamined ideas are dubious at best, nonsensical at worst. His notion of subjectivity combined the quite separate strands of ownership, identification-dependence, and epistemic privacy, and he arguably fell into incoherence in affirming that it is impossible in principle to ascertain whether any judgment about something subjective is fully understood by others. He mistakenly applied the contrast between numerical and qualitative identity to ideas and thoughts, both in arguing that the ideas of different persons cannot be numerically identical, but only exactly similar, and in asserting that the thoughts of different persons may be qualitatively identical; discourse about 'inner states' lacks this conceptual articulation which Frege projected upon it from discourse about perceptible objects. He wrongly considered thinking to be a subjective mental act, the exact character of which cannot be known to anyone other than the subject himself; this was linked to a defective grasp of the concepts of having a thought, expressing a thought, making an assertion, judging something to be true, etc., since in all these cases he ignored the fact that there are behavioml criteria for performing these acts. He never shook off the influence of part/whole analysis of judgable-contents or thoughts, and this resulted in fallacious reasoning and conclusions inconsistent with his adherence to function/argument analysis in logic. All of these points must weigh heavily in the balance when we consider the value
120
G.P. BAKER & P.M.S. HACKER
of his anti-psychologism as timeless answers to eternal philosophical questions. Since the clarification of the nature of logic was his overriding purpose, by far the most fundamental weaknesses of his position are mistaken propositions about logic itself. One is the idea that the propositions of logic (the Law of Identity, the Law of Excluded Middle, etc) are significant truths distinguished from all other propositions by their subject-matter. Allied with this idea is the conception of rules of inference (modus ponens, the syllogism in Barbara, universal instantiation, etc) as technical norms grounded in the truths of logic and subservient to the independent goal of truth. Both these general presuppositions Frege shared with the psychological logicians. Wittgenstein exposed the incoherence of this conception of logic in making clear that the propositions of logic are tautologies. This provides the basis for a more radical criticism of the psychological logicians' view of the nature of logic, but at the same time it shipwrecks Frege's 'Platonist' antithesis. All these reflections suggest that Frege's anti-psychologism is not a seam of pure gold. But might there not be nuggets of gold scattered in his writings, nuggets worth collecting up today and transporting to metropolitan philosophy? This seems mistaken in principle. It is unclear to what extent it makes sense to speak of his ideas independently of his reasoning and its presuppositions. But his thinking was infused with a Cartesian mythology about the realm of the psychological and a Platonist my'thology about the realm of the logical. This suggests that only somebody who shares a large measure of these Cartesian and Platonist mythologies should now find any seeds of the Tree of Knowledge scattered in his anti-psychologistic polemics. Should we then conclude that Frege's anti-psychologism is completely worthless? This would be an hysterical overreaction. He did develop telling criticisms of the influential conception of logic championed by Erdmann and the early Husser!. He argued cogently and forcefully within the limitations of a shared framework of thinking, and his arguments, fortified by his prestige as a logician, carried 'the day against 'this particular form of empiricism. His anti-psychologistic crusade affords an instructive example of how to umavel the particular knots of a specific philosophical misunderstanding. It is this skill which every philosopher must cultivate. To the extent that it is capable of transfer from one problem to another, the careful study of Frege's anti-psychologism might contribute indirectly to the elimination of misunderstandings and confusions now prevalent about the nature of logic, language, and thought.
121 ENDNOTES· • With the exception of references to section numbers, all references to Frege's publications are to pages in the original pagination. The original texts of Frege's articles are reprinted in: Gottlob Frege, Kleine Scltriften, ed. I Angelelli (Hildeshiem: G. Olms, 1967). English translations of these articles are available in: Gottlob Frege, CoUected Papers on MathemiJtics, Logic, and Philosophy, ed. B. McGuinness (Oxford: Basil Blackwell, 1984). 1. Gottlob Frege, "Review of Dr. E. Husserl's Philosophy of AritJunaic," trans. from ZEITSCHRIFT FOR PIULOSOPHIE UNO PIULOSOPIDSCHE KRITIK, 103 (1894): 313-332, p. 332. 2. Gottlob Frege, Posthumous Writings, ed. H. Hermes, F. Kambartel, F. Kaulbach, trans. P. Long, R. White (Oxford: Basil Blackwell, 1979), p. 142. 3. Gottlob Frege, The FOIUIdations of AritluMtic, a logico-mathematical enqu.iry into the concept of 1IW1Iber, 2nd revised edition, trans. I.L. Austin (Oxford: Basil Blackwell, 1959), from Die GrlUldlagen der ArithmetiJc, eine logisch mathematische Untersu,chung iber den Begriff der Zahl (Breslau: W. Koebner, 1884), p. x. 4. Gottlob . Frege, Concept#UJl Notation, a formula language of pu.re thought modelled upon the formula language of aritJunaic, in Gottlob Frege, Concept#UJl Notation and related articles, ed. and trans. Terrell Ward Bynum (Oxford: Clarendon, 1972), from Begriffsschrift, eilU! der aritJunaischen nachgebildete Formelsprache des reinen DenJcens (Halle: L. Nebert, 1879), sec. 2. 5. See: Ibid., sec. 8-10. See also: Frege, The FolUldations of Arithmetic, p. x, sec. 70. It is disputed whether the notion of a thing (Ding) in Conceptual Notation (Begriffsschrift) is identical with the notion of an object (Gegenstand) in the FolUldations of AritluMtic, and equally whether there is a parallel equivalence between the notion of a function (Funktion) and that of a concept (Begriff). A documented case in favour of these two equivalences is presented in: G.P. Baker and P.M.S. Hacker, "Dummett's Dig: Looking-Glass Archaeology," THE PIULOSOPIDCAL QUARTERLY, Vol. 37 (1987): 86-92. 6. Frege, Postluunolu Writings, p. 3. 7. Cf., Frege, Conceptual Notation, "Preface". 8. See: Frege, PosthUlllOllS Writings, p. 3. 9. Hugo Dingler objected to this thesis in correspondence with Frege. See: Gottlob Frege, Philosophical and MathemoJical Correspondence, 00. G. Gabriel, H. Hennes, F. Kambartel, C. Thiel, A. Veraart, abridged for the English edition by B. McGuinness, trans. H. Kaal (Oxford: Basil Blackwell, 1980), from Wissenschaftlicher Briefwechsel, 00. G. Gabriel, H. Hermes, F. Kambartel, C. Thiel, A. Veraart (Hamburg: Felix Meiner, 1976), p. 18. 10. Frege, Posthumous Writings, p. 261. 11. See: Ibid., pp. 244ff. See also: Gottlob Frege, "Negation," trans. of "Die Vemeinung: Eine logische Untersuchung," BEITRAGE ZUR PHll.DSOPHIE DES DEUTSCHEN IDEAUSMUS, I (1918): 143-157, p. 145f. 12. See: Frege, Posthumous Writings, pp. 5, 142. 13. Frege had little knowledge of the ideas of the British empiricists, and his discussions of their views about nwnbers are based on an anthology, not on his reading the primary texts in full. 14. This is a major theme of Hans Siuga. See: Hans Sluga, Gottlob Frege (London: Routledge &. Kegan Paul, 1980).
122
G.P. BAKER & P.M.S. HACKER
15. It is not absurd to call Frege's arguments decisive, provided that the burden of proof is judged to lie on the advocates of the claim that concepts and judgments are mental or subjective entities. 16. Frege, Philosophical and Mathematical Correspondence, p. 130. 17. Russell eventually changed his mind on this point and adopted the view that propositions are composed of images. See: Bertrand Russell, Logic and Knowledge, ed. R.C. Marsh (London: George Allen &. Unwin, 1956), pp. 283-320. See also: Bertrand Russell, The Analysis 01 Mind (London: George Allen &. Unwin, 1921). 18. Cf., Ludwig Wittgenstein, Tractatus Logico-Philosophicus, trans. D.F. Pears, B.F. McGuinness (London: Routledge and Kegan Paul, 1961), pt 4.1121. 19. Ibid. 20. Ibid., pt 4.023. 21. Ibid., pt 5.55ff. 22. Russell was a notable exception; he elaborated a radically psychologistic conception of logic in defense of empiricism. See: Bertrand Russell, An Inquiry into Meaning and Truth (London: George Allen &, Unwin, 1940). 23. See: Frege, ConceptuoJ Notation, sec. 3-4. See also: Frege, 'The Thought: A Logical Inquiry," trans. of "Der Gedank:e: Eine logische Untersuchung," BEITRAGE ZUR pmLOSOPHIE DES DEUTSCHEN IDEAUSMUS, I (1918): 58-77, p. 63f. 24. This point of view is advocated by Michael Dummett. See: Michael Dummett, Frege: Philosophy of Language (London: Duckworth, 1973), pp. 88f. 25. See: Frege, PosthU11lOKS Writings, pp. 3, 261. See also: Frege, Philosophical and Mathematical CO"espoNUnce, p. 17f. 26. See: Gottlob Frege, "On Sense and Reference," trans. of "Ober Sinn und Bedeutung," ZEITSCHRIFT FOR PHILOSOPHIE UNO pmLOSOPlllSCHE KRITIK, 100 (1892): 25-50, pp. 29f. 27. See: Frege, "The Thoug~" pp. 68f. 28. Ibid., p. 66. 29. Cf., David Bell, Frege's Theory of Judgement (Oxford: Clarendon, 1979), p. 74. 30. E.H.W. Kluge, The Metaphysics of Goulob Frege (The Hague: Martinus Nijhoff, 1980), p. 16. 31. Michael Dummett, The Interpretation of Frege's Philosophy (London: Duckworth, 1981), p. 121. 32. Ibid., p. 54. 33. Ibid., p. 52. 34. P.T. Geach, "Assertion," THE PHILOSOPIDCAL REVIEW, Vol. 74 (1965): 449-465. 35. Dummett, Frege: Philosophy of Language, pp. 158ft. 36. Ibid., pp. 158 and 641. 37. Ibid., pp. 637f. 38. Michael Dumme~ "An Unsuccessful Dig," THE PIDLOSOPlllCAL QUARTERLY, Vol. 34 (1984), p. 391. 39. Dummett, Frege: Philosophy 01 Language, p. 684. 40. Cf., Edmund Hussed, Logical Investigations, trans. IN. Findlay (London: Routledge & Kegan Paul, 1970), Vol. L pp. 147ft. 41. George Boole, The MathemtJlical Analysis of Logic (Oxford: Basil Blackwell, 1948), p. 5. 42. George Boole, The Laws of Thought (Mineola: Dover, 1953), Ch. I, sec. 1. 43. B. Erdmann, Logik, quoted in translation in: Hussed, Logical Investigations, Vol. I, p. 162.
FREGE'S ANTI·PSYCHOLOGISM
123
44. Ibid., p. 162f. 45. See: Frege, Posthumous Writings, p. 128. 46. See: Gottlob Frege, GrlUldgesetze der Arithmetilc, Begriffsschriftlich abgeleitet, Band I and IT (Jena: H. Pohle, 1893 and 1903), Band I, p. xv. A translation of Frege's ''Introduction'' and Vol. I, sec. 1-52 are given in: The Basic Laws of Arithmetic: Exposition of the System, 00. and trans. Montgomery Furth (Berkeley: University of California, 1964). Translations of Vol. IT, sec. 56-67, 86-137, 139-144, 146-147, and Frege's "Appendix" are included in: Translations from the Philosophical Writings of Gottlob Frege, 00. and trans. Peter Geach, Max Black (Oxford: Basil Blackwell, 1960). 47. Frege, Posthumous Writings, p. 4. 48. Ibid., pp. 4, 128. 49. See: Frege, Gru.ndgesetze der Arithmetik, Band I, p. xv. See also: Frege, Posthumous Writings, pp. 132f. 50. See: Frege, Grundgesetze der Arilhmetik, Band I, p. xvi. 51. Ibid., Band I, pp. xvif. 53. See: Ibid., Band I, p. xvii. 54. See: Ibid. 55. See: Ibid., Band I, p. xvi. 56. Russell's demonstration that a contradiction could be derived from Axiom V of Grundgesetze led Frege to confess that this basic law of his axiomatization of the propositions of logic had always seemed to him to be lacking in the self-evidence essential to its serving as a proper foundation for the laws of truth. See: Ibid., Band II, p. 253). 57. Indeed, he emphasized the differences between these two kinds of laws. See: Ibid., Band I, p. xv. 58. See: Wittgenstein, Tractatus Logico-Philosophicus, pts. 6.1-6.113. 59. See: Ibid., pts. 6.1201. Cf., Cora Diamond, Wittgenstein's Lectures on the Foundations of Mathematics Cambridge, 1939: From the Notes of R.G. Bosanquet, Norman Malcolm, Rush Rhees, and Yorick Smythies (Hassocks, Sx.: Harvester, 1975), pp.277ff. 60. See: Diamond, Wittgenstem's Lectures on the Foundations of Mathematics, pp. 279ff. 61. See: Wittgenstein, Tractatus, pts. 6.126, 6.1263. 62. Ludwig Wittgenstein, Remarks on the Foundations of Mathematics, third 00., ed. G.H. von Wright, R. Rhees, G.EM. Anscombe (Oxford: Basil Blackwell, 1978), p.95. 63. Ibid., p. 89. 64. Frege, Posthumous Writings, p. 7. 65. See: Ibid., p. 198. See also: Frege, "Negation," pp. 151f. 66. See: Frege, Posthumous Writings, p. 253. 67. Ibid. 68. See: Ibid., pp. 132f. See also: Frege, Grundgesetze der Arithmetik, Band I, p. xv. 69. See: Frege, Grundgesetze tIer Arithmeti.k, Band I, pp. xv if; cf., Frege, Posthumous Writings, pp. 134f. 70. See: Frege, ''The Thought," p. 67. 71. See: Ibid., pp. 67f.; cf., Frege, Posthumous Writings, pp. 3f; and Frege, Philosophical and Mathematical Correspontknce, p. 67. 72. Frege, The Foundations of Arithmetic, sec. 27. 73. Frege, ''The Thought," p. 67.
124
G.P. BAKER & P.M.S. HACKER
74. Frege, "On Sense and Reference," p. 30. 75. See: Frege, Posthumous Writings, p. 251. 76. Ibid., p. 133. 77. Ibid., p. 251. 78. Ibid., p. 133. 79. See: Frege, ''The Thought," p. 68. 80. Frege, Posthumous Writings, p. 133. 81. See: Frege, ''The Thought," p. 69. 82. Frege, Posthumous Writings, p. 134. 83. See: Ibid., p. 132. 84. Ibid., p. 134. 85. See: Ibid., p. 132. 86. See: Frege, ''The Thought," p. 66. 87. See: Frege, "On Sense and Reference," p. 29. 88. Frege, Posthumous Writings, p. 137. 89. Frege, "Negation," p. 146. 90. See: Frege, ''The Thought," p. 69. 91. This idea was later elaborated by M. Schlick. See: M. Schlick, "Form and Content," in Gesammelte Au/salze (Vienna: Gerold, 1938), pp. 151ff. 92. Frege, "On Sense and Reference," p. 30. 93. See: Ibid.; cf., Frege, "The Thought," p. 69; and Frege, "Review of Dr. E. Husserl's Philosophy of Arithmetic," p. 325. 94. Frege, Posthumous Writings, p. 145. 95. Ibid., p. 148. 96. Ibid., p. 7. 97. Ibid., p. 137. 98. See: Frege, The FolUldations of Arithmetic, sec. 61. 99. See: Frege, Posthumous Writings, p. 148. 100. See: Ibid., p. 135; cf., p. 251. 101. See: Ibid., p. 138. 102. Frege, ''The Thought," p. 77. 103. Frege, Posthumous Writings, p. 145. 104. See: Gottlob Frege, "On the Law of Inertia," trans. of "Ober das Trigheitsgesetz," ZEITSCHRIFr FOR PHILOSOPHIE UND PHILOSOPHISCHE KRITIK, 98 (1891): 145-161, p. 160. 105. Ludwig Wittgenstein, Philosophical Investigations, 00. G.EM. Anscombe, R. Rhees, trans. G.EM. Anscombe (Oxford: Basil Blackwell, 1953), sec. 253f., 35Off., 377ff. 106. See: Frege, Posthumous Writings, p. 135. 107. See: Frege, Grundgesetze der Arithmetik, Band I, sec. 5. 108. Cf., Frege, ''The Thought," p. 68. 109. See: Frege, Posthwnous Writings, pp. 143f. According to the doctrine of the syllogism on which the psychological logicians erected their philosophy of logic, judgments contained further elements that need not be taken to be ideas, namely syncategoremeata (signified by 'all', 'some', 'no', and 'not') and the copula (allegedly indicating assertion). 110. See: Frege, Philosophical and Mathematical Correspondence, pp. 68, 100f. 111. See: Frege, Conceptual Notation, sec. 11. See also: Frege, Grundgesetze der Arithmetik, Band L sec. 8. 112. See: Frege, Grundgesetze der Arithmetik, Band I, p. xxiv. 113. See: Ibid.
FREGE'S ANTI-PSYCHOLOGISM
125
114. See: Ibid. See also: Frege, The FOIUIdotions of Arithmetic, p. x. 115. Frege, GrUNlgesetze der Arithtnetik, Band L p. xxi.
116. Ibid. 117. Frege, "Review of Dr. E. Husser!'s Philosophy of Arithmaic," p. 318. Frege quipped tha~ on their view, no judgment could be about the Moon because this would require that the Moon be part of a state of consciousness. ''But would not.. the Moon sit a bit heavy on the stomach of one's state of consciousness?" Ibid., 316. 118. See: Frege, T~ FolUlllations of Arithmetic, sec. 27n., 61. 119. Frege, ''The Thought," p. 66. 120. See: Frege, "On Sense and Reference," p. 29f. 121. Frege, GrUNlgesetze der Arithmetik, Band L p. xix. 122. Consequently, he left himself a task in philosophy of arithmetic, namely to prove that numbers are not ideas. This required separate arguments. See: Frege, The Foundations of Arithmetic, sec. 26f. 123. Frege, GrJU&dgesetze der Arithmetik, Band L p. xxiii. 124. Frege, Posthumous Writings, p. 144. 125. See: Frege, Conceptuol Notation, sec. 2. 126. See: Ibid., sec. 2n. See also: Frege, Posthumous Writings, p. 105. 127. See: Frege, Posthumous Writings, p. 143f. 128. Ibid., p. 105 129. See: Frege, "On Sense and Reference," p. 35f. 130. See: Frege, Posthumous Writings, p. 255. 131. See: Frege, Grundgesetze der Arithmetik, Band I, sec. 32. 132. See: Ibid. 133. See: Frege, Philosophical and Mathemmical Correspondence, p. 163. 134. See: Frege, Posthumous Writings, p. 243. 135. Arguably, as we noted, the subjects of all judgments are concepts, namely if the subject of a singular judgment (e.g., that Socrates is mortal) is taken to be an individual concept 136. See: Frege, "On the Law of Inertia," p. 160. See also: Frege, "On Sense and Reference," p. 29. 137. Cf., Frege, "Review of Dr. E. Husserl's Philosophy of Arithmetic," p. 316. 138. Frege, "On the Law of !nertia," p. 160. 139. Cf., Frege, Posthumous Writings, p. 17. 140. Frege himself continued to employ this traditional terminology in early writings (See: Frege, Conceptlllll Notation, sec. 2. See also: Frege, Posthumous Writings, p. 16), although he officially denied this implication. This idiom invites misunderstanding; it is probably linked with his insistent inclination to combine the doctrine of fimction/argument analysis in logic with the idea of decomposing judgeable-contents and thoughts into parts; and this may have contributed to the emergence of his late doctrine that a thought must be built up out of thought-building-blocks (Gedanken bausteine) which correspond to the words or phrases of the sentence that expresses this thought See: Frege, Posthumous Writings, p-pe 225, 243. 141. Frege, "On Sense and Reference," p. 29. 142. Frege, "Review of Dr. E. Husser!'s Philosophy of Arithmetic," p. 316. 143. See: Frege, "On Sense and Reference," p. 29. 144. See: Frege, Philosophical and Mathemmical Co"espondence, p. 80; Frege, GrlUldgesetze der AritluMtik, Band I, sec. 32; and Frege, "On Sense and Reference," p.29.
126
G.P. BAKER & P.M.S. HACKER
145. Frege, The FolUUlations of Arithmetic, sec. 59. 146. Ibid., sec. 60. 147. Frege, "On Sense and Reference," p. 29. 148. Ibid., pp. 29f. 149. Cf., Frege, ''The Thought," p. 67. 150. It does not follow, however, that the sense of a concept-word cannot be subjective. On the contrary, if there are singular thoughts limited to the sphere of a single consciousness, then the concept-words occurring in the expression of such thoughts would have senses as subjective as the value of these sense-functions (and perhaps references as subjective as the arguments of these concepts). It is disputed whether Frege treated either the content or the sense of concept-words as functions. Evidence that he made both these identifications is marshalled in: G.P. Baker and PM.S. Hacker, Frege: Logical Excavations (Oxford: Basil Blackwell, 1984), pp. 170ff., 322ff. Some criticisms are rebutted in: G.P. Baker and PM.S. Hacker, "Dummett's Dig: Looking-glass Archeology," pp. 88ff. 151. Frege, Gru.ndgesetze der Arithmetik, Band I, pp. xxif. 152. Frege, PosthU11lOUS Writings, p. 139. 153. Frege, The FolUUlations of Arithmetic, pp. vf. 154. See: Ibid., sec. 47. 155. Cf., Ibid., pp. v if. 156. See: Frege, "On the Law of Inertia," p. 158. 157. Frege, The FolUUlations of Arithmetic, sec. 60. 158. See: Frege, Gru.ndgesetze der Arithmetik, Band I, p. xxi. 159. Frege, Posthumous Writings, p. 131. 160. This objection ignores the fact that the psychological logicians would probably require every judgment to contain additional elements (namely, the copula and syncategoremata indicating 'quality' and 'quantity'). 161. Frege, Posthumous Writings, p. 145. 162. Hwne ignored the distinction in rmding no difference between the complex idea of a horse and the complex idea that horses exist (See: David Hume, A Treatise of Hwnan Nature, ed. L.A. Selby-Bigge (Oxford: Clarendon, 1978), "Appendix"), and Russell (attempting to update Hume with the addition of the sophisticated logical machinery of the predicate calculus) found difficulty in distinguishing a proposition from the mere assemblage of its components. Cf., Bertrand Russell, The Principles of Mathematics (London: George Allen & Unwin, 1903), pp. 47ff. 163. See: Frege, The FolUUlations of Arithmetic, sec. 53. 164. See: Frege, The FolUUlations of Arithmetic, sec. 49. 165. Never? He apparently argued this strong thesis in one early \D1published paper. ''1 allow the formation of concepts to proceed only from judgments....And so instead of putting a judgment together out of an individual as subject and an already previously formed concept as predicate, we do the opposite and arrive at a concept by splitting up a judgeable-content" (Frege, Posthumous Writings, 160. Doubt might be entertained about whether this thesis is coherent in this completely general form. It might be suggested that Frege had in mind only the complex concepts required for this treatment of generality as a second-level concept (cf., Dummett, Frege: Philosophy of LanglUlg~, pp. 27ff; and Dmnmett, The Interpretation of Frege's Philosophy, pp. 292ft). 166. See: Frege, The FolUUlations of Arit~tic, sec. 45. 167. See: Frege, ''The Thought," 1?~§1~_------------ - - - - - - - - - - - - - - - - - - - - -
FREGE'S ANTI-PSYCHOLOGISM
127
168. Russell worried about this problem and tried to develop a satisfactory solution to it (See: Bertrand Russell, Theory of Knowledge: the 1913 Manuscript, -ed. E.R. Eames (London: George Allen & Unwin, 1984), pp. 85ff). The problem is claimed to be an insuperable obstacle to the general doctrine of 'abstractionism' (See: P.T. Geach. Mental Acts (London: Routledge & Kegan Paul, 1957), pp. 320. 169. Cf., Frege, "Review of Dr. E. Husserl's Philosophy of Arithmetic," pp.
316t
I
170. Cf., Frege, The Foundations of Arithmetic, sec. 58. 171. See: Frege, "Review of Dr. E. Husserl's Philosophy of Arithmetic," pp. 317, 324f. 172. Cf., Ibid., p. 316. 173. These claims mix insights with confusions. Frege obscured the difference between logical operations and material functions (concepts), and thereby misrepresented the propositions of logic as comprising the science of special logical objects (cf, Wittgenste~ Tractatus, pts. 5.2-5.254, 6.1111). 174. Cf., Frege, Posthumous Writings, p. 184. 175. See: Ibid., p. 16f. 176. Cf., Frege, The Foundations of Arithmetic, sec. 45. 177. Ibid., p. x. 178. See: Frege, Grundgesetze der Arithmetik, Band I, sec. 3. 179. In earlier writings, when he conceived of the value of a function as a judgeable-content (not as a truth-value), he did not hold this general thesis; on the contrary, he thought that only in special cases could the extension of a concept be calculated from the intrinsic nature of the concept itself (See: Ibid., sec. 53). The change is crucial, but Frege did not call attention to it 180. This accusation is substantiated in: G.P. Baker and P.M.S. Hacker, Frege: Logical Excavations, pp. 173ff., 252ff.
ANTI-PSYCHOLOGISM AND SCEPTICISM: FREGE, DESCARTES, AND WITTGENSTEIN PHILIP DWYER
I. Introduction Chapter 19 of Michael Dummett's Frege, Philosophy of Language, is entitled ttFrege's Place in the History of Philosophy." Dummett assigns Frege his place largely vis d vis Descartes: Before Descartes, it can hardly be said that anyone part of philosophy was recognized as being fundamental to all the rest: the Cartesian revolution consisted in giving this role to the theory of knowledge. Descartes made the question, 'What do we know, and what justifies our claim to this knowledge?' the starting point of all philosophy: and despite the conflicting views of the various schools, it was accepted as the starting point for more than two centuries. Frege's basic achievement lay in the fact that he totally ignored the Cartesian tradition, and was able, posthumously, to impose his different perspective on other philosophers of the analytic tradition.1
For Frege, according to Dummett, it is not the theory of knowledge, but the theory of meaning which is the fundamental part of philosophy: Because philosophy has, as its rust, if not its only task, the analysis of meanings, and because, the deeper such an analysis goes, the more it is dependent upon a correct general account of meaning, a model for what the lDlderstanding of an expression consists in, the theory of meaning, which is the search for such a model, is the fOWldation of all philosophy, and not epistemology as Descartes misled us into believing. Frege's greatness consists, in the rust place, in having perceived this. 2
I too believe that Frege's place in the history of philosophy is best worked out in connection with Descartes. However, contrary to Dummett, I believe the right account of the matler concerns Frege's positive relation to Descartes, not his lack of a relation to Descartes. The relation between Frege and Descartes is many-sided, but the aspect on which I want to concentrate concerns the connection, in the thought of both philosophers, between scepticism and a certain kind of 'anti-psychologism': the connection between scepticism and anti-psychologism is exploited by both philosophers and thereby constitutes a connection between the philosophers themselves.
ANTI·PSYCHOLOGISM AND SCEPTICISM
129
Furthennore, the connection between Frege and Descartes on the score of scepticism and anti-psychologism is mediated by a third philosopher, viz., Wittgenstein. I want to show, inter alia, that it is no accident that the considerations Wittgenstein advances against an explicitly Fregean scepticism and anti-psychologism in Remarks on the Foundations of Mathematics and Philosophical Investigations recur in On Certainty, where the target is, at least implicitly, Descartes. The story of this philosophical menage a trois, as I set it out, is parenthesized by two digressions. At the end I offer some general comments on scepticism and so-called 'methodological scepticism'. And in what follows immediately, I wish to give an indication of the exaggeration of Dummett's claim that Frege totally ignored the Cartesian tradition.
II. Frege and Descartes Given the similarities between the philosophies of Frege and Descartes to be listed shortly, the statement that Frege totally ignored the Cartesian tradition, in any broad sense of the expression, is simply wrong. Dummett, however, no doubt means 'the Cartesian tradition' in the narrow sense of a tradition concerning what he calls 'the starting point of all philosophy'. It is neither indisputable that Descartes did in fact take epistemology to be the starting-point or 'fundamental' part of philosophy rather than, say, metaphysics, or philosophical theology, or philosophy of mind, nor clear just how these are to be precisely distinguished.3 There is no less a problem, such as it is, as to what Frege took to be the starting point in philosophy, and Dummett's claim that Frege took it to be 'the theory of meaning' in no way clarifies the matter. As was seen in the second quotation in our introduction, Dummett says that the theory of meaning is the search for a model of what the understanding of an expre~ion consists in. The construction of a 'model of understanding' would normally be construed as a problem for either or both the philosophy of mind or/and the theory of knowledge. Taking the latter interpretation for example, Dummett says: ...the complex phrase on which attention needs to be concentrated is 'knowing the meaning of... ': a theory of meaning is a theory of understanding. What we have to give an account of is what a person knows when he knows what a word or expression means, that is, when he lDlderstands it4
Here a 'theory of understanding' is surely some kind of theory of knowledge, and hence, a 'theory of meaning' is too. No recourse can be made here to a distinction between knowledge of meaning and knowledge of the world. Such a distinction is quite incompatible with the truth-conditions semantics which Dummett attributes to Frege: "to grasp the sense of a sentence is, in general, to know the conditions under which that
130
PHILIP DWYER
sentence is true and the conditions under which it is false."S The distinction between the theory of knowledge and the theory of meaning also seems to be undermined by Dummett's other remarks on Frege's notion of 'sense', a central notion, presumably, of his theory of meaning. Dummett says, "For Frege, sense is a cognitive notion: it is introduced by him in the first place to resolve a problem about the cognitive value, or information content, of true identity-statements."' In Frege, Dummett says that "it is quite wrong to think of Frege's theory of meaning as one where meaning is quite divorced from knowledge. "7 It is, in any case, not clear how seriously Dummett really means his doctrine of the starting point in philosophy. After using it as a basis for characterizing the last three hundred years of philosophy, Dummett seems to abandon it in short order. We are fll'St told that whatever part of philosophy is "fundamental to all the rest" is the "starting point for the whole subject "8 The starting point is that which "must be settled before anything else can be said."' Frege held, we are told, "that the theory of meaning is the fundamental part of philosophy which underlies all the others, "10 that "he starts from meaning by taking the theory of meaning as the only part of philosophy whose results do not depend on those of any other part, but which underlies all the rest."11 But then shortly we are told that: To say that the theory of meaning is the f01Dldation of philosophy is not to say that nothing else can be done until the main problems in the theory of meaning are resolved On the contrary, progress can be made on problems 'that arise in advance of our having a satisfactory theory of meaning, even when these are problems in the context of which the notion of meaning seems to be naturally invoked. 12
Not only is it now possible to settle some issues before we have a theory of meaning, but in some cases "it is a condition of advance in the dispute that the notion of meaning be bracketed;"l3 "in disputes of this kind, the notion of meaning must be bracketed before any progress can be made. "14 But this is not all. In this revised doctrine of the starting point, not only may we settle some issues before we have a theory of meaning; and not only, in some cases, must we settle some issues before invoking the notion of meaning, in some yet further cases, settling issues in one area may actually contribute to settling issues in the theory of meaning: It is no objection to the fundamental character of the theory of meaning that sometimes, as in the cases cited, an advance in resolving a question that arises within some other particular branch of philosophy may be made before the corresponding general question in the theory of meaning has been settled, and may, indeed, be a substantial contribution towards settling that question. IS
The question of what Frege took to be fundamental in philosophy is best construed as a question about what he took to be a very important issue in philosophy, not what he took its 'starting point' to be. Certainly he took
ANTI-PSYCHOLOGISM AND SCEPTICISM
131
epistemological issues to be fundamental, and in particular the issue of the justification of knowledge. Speculating on the effect of his The Foundations of Arithmetic, Frege says, "Someone or other, perhaps, will take this opportunity to examine afresh the principles of his theory of knowledge. n16 This would be a strange comment to come from someone who, according to Dummett, never gave us his views on knowledge. 17 Frege is clearly in the epistemological mainstream of Western philosophy. But beyond this, if his views are to be characterized as anything other than ,Fregean" they would have to be characterized as Cartesian. The following is a short list of the points of contact between Descartes' and Frege's views. (a) Both Descartes and Frege have a great concern for science - a discipline or systematic body of knowledge - taking it to be what epistemology is primarily in aid of. Descartes' Discourse on the Method of rightly conducting the Reason and seeking for Truth in the Sciences was intended as an introduction to essays on meteorology, optics, and geometry. Throughout the Discourse he refers to his method as the 'foundations of my physics'.18 In the Meditations he speaks of his doubting what were hitherto 'foundations for the sciences' ,19 while his method is "a certain Method for the resolution of difficulties of every kind in the sciences,n~ the most geneml difficulty being "to arrive at any certainty in the sciences."21 Frege's Begriffschrift opens with the sentence "In apprehending a scientific truth we pass, as a rule, through various stages of certitude. "22 The most 'secure foundation' for a proposition is provided by pure logic- "on those laws upon which all knowledge rests."23 The Begrif{schrift will have application to mathematics, Frege says, and could easily be extended to geometry. "The transition to the pure theory of motion and then to mechanics and physics could follow at this point 1124 His sample propositions in explaining his notation are not of 'the eat is on the mat' variety, but e.g., "Let a be the circumstance that the piece of iron E becomes magnetized, b the circumstance that a galvanic current flows through the wire D ete.,"2S or "the circumstance that hydrogen is lighter than carbon dioxide."26 This concern with scientific truth dominated Frege's views on logic throughout his career. Logic has the property 'true' as its object of study;27 what is true are 'thoughts', the paradigm of which is a natural law;28 the truth which is the concern of logic is "that sort of truth which it is the aim of science to discern";29 "The goal of scientific endeavor is truth."30 Descartes also speaks of his method of rightly conducting the reason in seeking truth in the sciences as his "principle rules of logic. "31 (b) Both Frege and Descartes uphold an ideal of rigour in science, whether in the broad sense of 'knowledge' or in the narrow sense of a science such as arithmetic, and take Euclid as their model (as do other Rationalists, most notably Spinoza). Descartes says that "Those long chains of reasoning, simple as they are, of which geometricians make use in order to arrive at the most difficult demonstrations, had caused me to imagine that all those things which fall under the cognizance of man might very likely
132
PHILIP DWYER
be mutually related in the same fashion. "32 In the preface to The Basic Laws of Arithmetic, Frege introduces his method with these words: "The ideal of a strictly scientific method in mathematics which I have here attempted to realize, and which might indeed be named. after Euclid, I should like to describe as follows."33 Connected with this is an ideal of 'gaplessness' in proofs. His Begriffschrift was invented to meet the demand that every gap or "jump must be barred from our deductions."34 For Frege, it is not good enough to say, in the middle of a proof, that a certain transition is merely evident, given what has come before: "if it is a matter of gaining insight into the nature of this 'being evident', this procedure does not suffice; we must put down all of the intermediate steps, that the full light of consciousness may fall upon them. "35 Gaplessness serves to answer an epistemological question: "Because there are no gaps in the chains of inference, every ,axiom', every 'assumption', 'hypothesis', or whatever you wish to call it, upon which a proof is based is brought to light; and in this way we gain a basis upon which to judge the epistemological nature of the law that is proved. "36 This same requirement of gaplessness as the ground of true certainty, is to be found among the rules set out in Descartes' Regulae and the Discourse on Method. In the Regulae, sounding very Fregeau, Descartes says: Often people who attempt to reduce a conclusion too quickly and from remote principles do not trace the whole chaia, of intermediate conclusions with sufficient accuracy to prevent them from passing over many steps without due consideration. But it is certain that wherever the smallest link is left out the chain is broken and the whole of the certainty of the conclusion falls to the ground.37
(c) The epistemological ideal of gaplessness leads Descartes and Frege to similarities in their psychological accounts of gaplessness, Le., in their philosophies of mind. Descartes' frrst principles cannot be proved, they "are given by intuition alone," while "the remote conclusions are furnished by deduction. "31 This parallels Frege's distinction in his 'strictly scientific method in mathematics' between axioms and rules of inference on the one hand, and proved propositions on the other. The latter are not be grasped "in one single act of understanding, "39 whereas axioms are 'self-evident'.40 So for Frege and Descartes there is a psychological criterion for distinguishing logical self-evidence (Frege) or intuition (Descartes) from deduction. Frege employs a further criterion ofaxiomaticity and a rather Cartesian one at that, viz. indubitability: "The axioms are truths as are the theorems, but they are truths for which no proof can be given in our system, and for which no proof is needed. It follows from this that there are no false axioms, and that we cannot accept a thought as an axiom if we are in doubt about its bUth; for it is either false and not an axiom, or it is true but stands in need of proof and hence is not an axiom."41 Axioms (and
ANTI·PSYCHOLOGISM AND SCEPTICISM
133
laws of logic) are then clear and distinct, but what is clear and distinct to one man, another man (or Martian) may doubL42 (d) There are other similarities in the philosophies of mind of Frege and Descartes. In Descartes' official doctrine there are basically two sorts of functioning of the mind: the understanding and the will. The understanding is said to include "sense-perception, imagining, and conceiving things that are purely intelligible,"43 but sense-perception is very much the poor here, set off to one side. The utility of Descartes' method, he says, is 'that "it sets out for us a very simple way in which the mind may detach itself from the senses. tt44 The operation of the will, as opposed to the understanding, consists most importantly in judgment, i.e., assertion or denial of something.45 Frege likewise distinguishes sense perception from what he calls 'conceptual thinking', and these two are again distinguished from judgment. In The Foundations of Arithmetic Frege fmnly segregates what we come to know "through the medium of the senses" from "the fundamentally different question" of what is "given directly to our reason."46 It is "the exeICise of these higher intellectual powers that distinguish men from brutes.,,47 In the late article "The SOUICes of Knowledge in Mathematics and Mathematical Natural Sciences," Frege distinguishes three such SOUICes: (1) sense perception; (2) the logical source of knowledge; (3) the geometrical and temporal source. Section A of the article deals with peICeption and is entitled "Sense Illusions": "We must be careful not to overestimate the value of sense perception, for without the other sources of knowledge, which protect us from being deceived, we could hardly get anywhere with it 1t48 Descartes of course had a similarly sceptical opinion of perception and also viewed the correction of deceptive sense perception as an achievement of the intellect.49 According to Frege, "the logical source of knowledge, which is wholly inside us" is 'thinking'.~ The doctrine that judgment is the affIrmation of a thought, whereas thinking is the mere grasping or entertaining of a thought, occurs frrst in the Begriffschrift in his explanation of the notational difference between the content stroke and the judgment
stroke.51 (e) Both Descartes and Frege view language, as well as perception, as something that contaminates 'pure thought'. The Begriffschrift, 'a formula language for pure thought', is a contribution to one of the main tasks of philosophy, viz., "to break the domination of the word over the human spirt by laying bare the misconceptions that through the use of language almost unavoidably arise concerning the relations between concepts...."52 Frege links his Begriffschrift with Leibniz's project of an ars characteristica.53 Descartes too envisioned such a universal language "for acquiring true scientific knowledge....The greatest advantage of such a language would be the assistance it would give to men's judgment, representing matters so clearly that it would be almost impossible to go wrong. As it is, almost all our words have confused meanings, and men's minds are so accustomed to them that there is hardly anything that 'they can perfectly understand. "54
134
PHILIP DWYER
Frege holds that "There is no contradiction in supposing there to exist beings that can grasp the same thought as we do without needing to clad it in a form that can be perceived by the senses. "ss For the contaminating effects of language are of a piece with its perceptual character: "Something by itself not perceptible by sense, the thought, is presented to the reader and I must be content with that - wrapped up in a perceptible linguistic fonn. The pictorial aspect of language presents difficulties. The sensible always breaks in and makes expressions pictorial and so improper. So one fights against language, and I am compelled to occupy myself with language although it is not my proper concern here."5C§ In The Foundations of Arithmetic Frege speaks of "achieving knowledge of a concept in its pure fonn, in stripping off the irrelevant accretions which veil it from the eyes of the mind." (geistige Auge)S7 Here he echoes Descartes' account of the clear and distinct perception of ideas and thoughts as "looking with the eyes of the mind"S' or "mental vision. ns' (t) Descartes speaks of apprehending 'ideas'fiO while Frege speaks of comprehending 'thoughts'."l Since Frege made so much of the difference between ideas (Vorstellungen) as subjective psychological entities and thoughts as abstract objective entities, this may seem to constitute a major difference between the two philosophers. However, from Begriffschrift to 'Dec Gedanke' Frege is, at any rate, right in Locke-step with Descartes in believing that there are such (mental) things as ideas or latterly 'sense data', e.g., that to see is to have 'visual ideas', to hear is to have 'auditory ideas', to feel is to have 'motor ideas', etc.62 His eventual objection to ideas in favour of thoughts was that ideas, as 'mental images' or as 'sense impressions', are private63 and, as psychological entities, have no internal relation to truth. 64 Descartes, on the other hand, sometimes denies that 'ideas' are images or sense impressions,6S and in one of their chief incarnations they are propositional in nature; they are what is apprehended by the understanding and affmned or denied by the will in making a judgment" Moreover, the gain in objectivity which Frege thought was achieved by twning token thoughts in persons' minds into type thoughts in a platonic realm, is, in the end, vitiated by the fact that ultimately the public thoughts are connected with the public signs (names and sentences) by a private process of association." (g) Whatever the differences between Frege and Descartes on the count of ideas versus thoughts, Descartes was not averse to recognizing mind-independent platonic objects. In the interview with Bunnan he says: All the demonstrations of mathematics are concerned with true beings and objects; and so the whole object of mathematics, and whatever mathematics considers in it, is a true and real being, and has a true and real nature, no less than the objects of physics. The only difference is that physics considers its object not only as a true and real being but also as actual, ~ as such existing, whereas mathematics considers its objects only as possible, not actually existing in space, but still able to exist.'"
ANTI-PSYCHOLOGISM AND SCEPTICISM
135
This could have been written by Frege. He too distinguishes "what is handleable or spatial or actual" from what is "objective," or "self-subsistent," "but not in space," such as numbers and thoughts.69 The above is by no means a full account of the links between the thought of Descartes and that of Frege, but it should serve to establish that Frege is fmnly rooted in the traditions of Cartesian rationalism, against which the later philosophy of Wittgenstein is, in good part, directed.
III. An Ambiguity about 'Psychologism' Scepticism plays an explicit and central role in Descartes' philosophy, and anti-psychologism plays an explicit and central role in Frege's philosophy. However, Descartes' employment of sceptical arguments implicitly relies on a kind of anti-psychologism, while Frege's anti-psychologism gives expression to a kind of scepticism. The connection between scepticism and anti-psychologism provides a further link between Descartes and Frege. It is with this latter link in particular that Wittgenstein enters the discussion. However, before setting this out, something must be said, by way of avoiding confusions, about the expression 'psychologism'. Dummett, in the essay "Frege's Philosophy" speaks of Frege's "strong attack on what he called 'psychologism'- the thesis that an account of the meanings of words must be given in terms of the mental processes which they arouse in speaker or hearer or which are involved in acquiring a grasp of their sense. "'0 Dummett goes on to link this hereditarily with "Wittgenstein's dictum that 'the meaning is the use'."'1 It is true that Frege argues against taking "as the meanings of words mental pictures or acts of the individual mind,",2 and asserts that to do so is to offend against the frrst of his 'three fundamental principles', viz., "always to separate sharply the psychological from the logical, the subjective from the objective."'3 It is also true that Wittgenstein's dictum about meaning as use is partly directed against the view that the meanings of words are mental images.'" Nevertheless, there is something very misleading in presenting Frege's attack in psychologism in these terms. Psychologism for Frege is first and foremost a thesis about the nature of logic, viz., that logical laws are really psychological laws. The "corrupting incursion of psychology into logic" is a matter of taking logical laws to be 'laws of thought' in the sense of asserting what is, rather than what ought to be.75 The very expression 'law of thought': ...seduces us into supposing that these laws govern thinking in the same say as laws of nature govern events in the external world. In that case they can be nothing but laws of psychology: for thinking is a mental process. And if logic were concerned with these psychological laws, it would be a part of psychology."
136
PHILIP DWYER
For Frege, logic concerns the 'laws of truth', psychology 'the laws of taking to be true'.T1 Here Frege may be punning on the German word for perception- 'Wahrnehmung' (I have not seen the original German). The latter mayor may not be in accord with the laws of truth.." Only if they are, can an inference be justified. The issue of the justification for making a judgment is the central motive for Frege's sharp distinction between psychology and logic." Just as logical laws are not be confused with psychological laws, the objects of arithmetic - numbers - are not be to confused with the objects of psychology- 'ideas': "If number were an idea then arithmetic would be psychology."80 Likewise, it is in connection with Frege's ontology of absttaet objects - which in effect constitutes the heart of his theory of meaning, such as it is - that the issue of meaning as a mental image comes in. In his discussion of psychologism in The Basic Laws of Arithmetic, Frege says: Surveying the whole question, it seems to me that the source of the dispute lies in a difference in our conceptions of what is true. For me, what is true is something objective and independent of the judging subject; for psychological logicians it is not.11
'What is true' for Frege, and objectively so, independently of a conscious subject, are thoughts. This, of course, rules out mental images as constituting the meaning or sense or thought behind words. But the 'subjective' for Frege includes more than mental images. Psychologism also includes any reference to 'muscular sensations', or the 'content of the brain', or human evolution.12 Not only is "the origin and evolution of ideas" ruled out as irrelevant, but also "the history either of our knowledge of concepts or of the meanings of words."83 In short, 'the subjective' includes any reference to human beings. Certainly Wittgenstein's thesis that meaning is use would be no less psychologistic for Frege than would be a mental image theory of meaning. Frege would be no less horrified at something going under the name 'anthropologism' - as Dummett describes Wittgenstein's philosophy of mathematicsM - than he would at any so-called psychologism. Indeed, as the above quote about the objectivity and independence of truth shows, even the milder forms of 'anti-realism' enunciated by DummetfS would be only so much psychologism for Frege. The only alternative to platonism, for Frege, is some form of psychologism: logic and mathematics, truth and meaning, are either person-dependent or they are not. There is more to the misleadingness of positively linking Frege's attack on psychologism with Wittgenstein than that indicated above. John McDowell writes of "the hostility to psychologism voiced by Frege in the philosophy of logic and language, and extended into the philosophy of mind by Wittgenstein."" Certainly one thinks of the philosophy of mind in
ANTI·PSYCHOLOGISM AND SCEPTICISM
137
connection with Wittgenstein being hostile to 'psychologism'. But Wittgenstein's hostility to psychologism should not be spoken of in the same breath with Frege's hostility to psychologism. Wittgenstein's anti-psychologism' is directed against mental process mongering; for example, against the thesis that thinking, as such, is a mental process. 87 But Frege was a great believer in mental processes, especially, as we've seen, with regard to thinking as a mental process. It is precisely his ontological doctrine of the thought which requires his Cartesian view of the mind: "To the grasping of thoughts there must then correspond a special mental capacity, the power of thinking."- It is then left for an 'act of judgment' to acknowledge the truth of the thought." To the extent that Wittgenstein's dictum 'meaning is use' is anti-psychologistic it is meant to undennine the view that 'meaning' - the verb - designates a mental process. For Frege this is just what meaning, as an activity, is, Le., a mental process of 'intending', or 'associating', or 'connecting'90 signs with thoughts. Again, for Frege, thinking as such involves the mind's intentional powers: "Although the thought does not belong with the contents of consciousness, there must be something in his consciousness that is aimed at the thought "91 Wittgenstein's dictum about meaning and use in directed against both the views that either mental processes or Fregean thoughts give meaning to signs.92 Crispin Wright says that one important advance of Wittgenstein's philosophy of mathematics over that of the intuitionists is this: "the case for repudiating a platonist conception of mathematical truth is no longer muddled with a psychologistic conception of mathematical understanding. "93 This passage (indexed under 'psychologism. Wittgenstein's ant-realism free from ') is ambiguous. It is precisely the case presented for platonism that is muddled with a psychologistic conception of mathematical understanding, insofar as 'psychologism' is what Wittgenstein attacked. Consider, for example, the following from Wittgenstein's Remarks on the Foundations of
Mathematics: But still, I must only infer what really follows! -Is this supposed to mean: only what follows, going by the rules of inference; or is it supposed to mean: only what follows, going by such rules of inference as somehow agree with some (sort of) reality? Here what is before our minds in a vague way is that this reality is something very abstract, very general, and very rigid. Logic is a kind of ultra-physics, the description of the 'logical structure' of the world, which we perceive through a kind of ultra-experience (with the understanding e.g.).M
On the other hand, Wittgenstein's philosophy of mathematics is riddled with 'psychologism' as Frege understood it. Indeed, Wittgenstein rubs Frege's nose in it
138
PHILIP DWYER
IV. Wittgensteinian Psychologism Remarks such 88, "For mathematics is after all an anthropological phenomenon"" seem specifically designed to set Frege spinning in his grave. When Wittgenstein says, "We shall see contradiction in a quite different light if we look at its occmrence and its consequences anthropologically...,"" we can hear Frege saying, 'Indeed we shall!'- but this concurrence would be only superficial. Wittgenstein goes on to explain that viewing contradiction 'anthropologically' means considering how it influences 'language-games', and, as he says in On Certainty, "everything descriptive of a language-game is part of logic. "fJ7 Given that a language-game is language and the actions into which it is woven,98 this view of logic would be poison for Frege. To repeat, there are only two alternatives concerning the nature of logic and mathematics: objectivity or person-dependence. This in turn must cash out as the alternative between some form of platonism and some fonn of psychologism. To the extent that Wittgenstein's rule-following considerations advanced in Philosophical Investigations and Remarks on the Foundations of Mathematics are directed against platonism, they constitute what Frege would call 'pschologism'. That Wittgenstein implicitly had Frege in mind in his discussions of rule-following could be gathered from section 80 of Frege's The Foundations of Arithmetic, entitled 'Following (in a series) is objective'. But Wittgenstein explicitly has Frege in mind in the inevitable discussion of psychologism that attends the anti-platonistic discussion of rule-following and logical compulsion in part I of Remarks on the Foundations of Mathematics. In outline, part 1 of Remarks on the Foundations of Mathematics goes as follows: Sections 1-5 concern how a role determines what steps are taken, with the upshot that it is not a mental act of meaning which determines the steps in advance (3), but rather the importance of mathematics in our life's activities that gives rise to the 'peculiar inexorability of mathematics' (5). Sections 6-23 concern logical inference and contain various antiFregean/Russellian, i.e., anti-platonist points, in particular: that the idea that rules of inference agree with some sort of person or subject-independent reality ('logic as ultra-physics') is an illusion (8); that logical inference is a transformation of signs, and that the reality it accords with is a convention or a use (9); that when one proposition follows from another, it is not that it must, but that it does: "we perform this transition" (12); it does not 'follow' in a platonic realm, short of an actual person's actual inference (21). Sections 24-112 concern proof, calculation and experiment, and mathematical belief. Sections 113-142 treat of logical compulsion, and it is here that reference to Frege on psychologism is made. Section 113 contains the dialogue on rule-following that ends with the refrain of the Kripke-style sceptic: "However many rules you give me- I give a rule that justifies my employment of your rules. "99 Section 116 says that ~__~~~_J.!f_jnfereD~e_---
ANTI-PSYCHOLOGISM AND SCEPTICISM
139
can be said to compel us in the same sense as other laws of human society.ulo Section 118 directly addresses the issue of psychologism: "It looked at first as if these considerations were meant to show that 'what seems to be a logical compulsion is in reality only a psychological one'only here the question arose: am I acquainted with both kinds of compulsion then?!" In the final paragraph of this section, Wittgenstein writes: Now we talk of the 'inexorability' of logic; and think of the laws of logic as inexorable, as still more inexorable than the laws of nature.... [Rather] There correspond to our laws of logic very general facts of daily experience. They are the ones that make it possible for us to keep demonstrating those laws in a very simple way (with ink on paper for example). They are to be compared with facts that make measurement with a yardstick easy and useful. This suggests the use of precisely these laws of inference, and now it is we that are inexorable in applying these laws" 01
This line is precisely what Frege stood against in his attack on psychologism. Part I, sections 119-130 of Remarks on the Foundations of Mathematics discuss the picture of the 'logical machine' which cranks out propositions which 'really' do follow. These sections go with Part V, section 20: "If calculating looks to us like the action of a machine, it is the human being doing the calculation that is the machine." Sections 131-133 make reference to Frege's anti-psychologism. Section 131 goes: "The laws of logic are indeed the expression of 'thinking habits' but also of the habit of thinking. That is to say they can be said to shew: how human beings think, and also what human beings call 'thinking' ."102 Section 132 has some fun with one of Frege's examples of 'a law about what men take for true', and then in sections 133-4, Wittgenstein writes: The propositions of logic are 'laws of thought', 'because they bring out the essence of human thinking'- to put it more correctly: because they bring out or shew, the essence, the technique, of thinking. They shew what thinking is and also shew kinds of thinking. Logic, it may be said, shews us what we understand by 'proposition' and by 'language'.
The direction of Wittgenstein's discussion of the laws of logic as 'laws of thought' is clearly counter to that of Frege's discussion, In the midst of these 'psychologistic' remarks, come three sections, 135-7, on error and a deceptive demon. We shall return to these. Following this in section 142, comes the remark that fmds a place in the Investigations (145): "What we are supplying are really remarks on the natural history of man...." It goes without saying that for Frege, the last thing remarks on the foundations of mathematics should be is remarks on the natural history of man. The following sections 143-51 deal with a case of 'fictitious natural history', that of the people who calculate the price of
140
PHILIP DWYER
wood in a peculiar way. In section 152 Wittgenstein writes "Frege says in the preface to the Grundgesetze der Arithmetic: ' ...here we have a hitherto unknown kind of insanity'- but he never said what this 'insanity' would really be like." (Frege was responding to the question: "But what if beings were even found whose laws of thought flatly contradicted ours and therefore frequently led to contrary results even in practice?ttl(3) In the section that follows, Wittgenstein suggests that this 'insanity' amounts to a cultural difference. Section 154, immediately following, contains the frrst remarks in Remarks on the Foundations of Mathematics on the requirement for surveyability of proofs (more about which later). Section 155 gives a further fictitious natural history and ends with the marginal note: "Are our laws of inference eternal and immutable?" Wittgenstein's question, which alludes to Frege's claim for the "unconditional and eternal validity" of the laws of logic in his anti-psychologistic diatribe in The Basic Laws of Arithmetic,l04 is answered in the immediately following section. Isn't it like this: so long as one thinks it can't be otherwise, one draws logical conclusions. This presumably means: so long as such and such is not brought into question at all. The steps which are not brought in question are logical inferences. But the reason why they are not brought into question is not that they certainly correspond to the 'truth' - or something of the sort, - no, it is just this that is called 'thinking', 'speaking', 'inferring" 'arguing'. There is not any question at all here of some correspondence between what is said and reality: rather is logic antecedent to any such correspondence; in the same sense, that is, as that in which the establishment of a method of measurement is antecedent to the correctness or incorrectness of a statement of length. lOS
This passage brings out the point that if one rejects a platonic realm in accordance with which we 'correctly' speak, think, infer, etc., then one is left only with speaking, thinking, inferring, as such, i.e., as paradigms, being the foundation or presupposition of correct speaking, thinking, inferring. And this, on Frege's view, is simply psychologism. Sections 157-8 deal with paradigms of sign sequences as guiding inference. These sections are thus connected with a remark in Part VII, section 66: "Logical inference is a transition that is justified if it follows a particular paradigm and its rightness is not dependent on anything else. ftlO6 This again is directed against Frege's platonistic view of logical laws, and would be for Frege, not psychologism, but something just as bad: formalism. Part I of Remarks on the Foundations of Mathematics contains one more strikingly anti-Fregean remark: "The mathematician is an inventor not a discoverer. 1m. Contrast Frege in The Foundations of Arithmetic: "...even the mathematician cannot create at will, anymore than the geographer can; he too can only discover what is there and give it a name. 't108 Remarks on the Foundations of Mathematics Part I ends with section 171 where the 'logical fI
ANTI-PSYCHOLOGISM AND SCEPTICISM
141
must' is compared to various (other) Gestalt phenomena (more about which later).
v. 'Two-Kinds' Theses Wittgenstein then is clearly what Frege would characterize as a 'psychological logician'. However, Wittgenstein must be distinguished from those Frege attacked, at least insofar as Wittgenstein had the benefit of Frege's thoughts on psychology and logic, but clearly was not swayed. Why was he unswayed? Frege's anti-psychologism arises in terms of a distinction which, being a commonplace of Western philosophy, he takes for granted, and which bears on his notion of the relation of logic to psychology. The distinction is roughly, that between the logical and the empirical, between a priori science whose laws are 'general laws' that are necessarily and eternally true, and a posteriori science whose laws are not 'general', but include an 'appeal to facts', i.e., to that which is empirical and contingent. 109 Logic and arithmetic are of the first kind, psychology is of the second The distinction between the logical and the empirical is mirrored in distinctions between the ideal and the actual, between the way things are 'in principle' and the way they are 'in fact', between the way things are 'in theory' and the way they are 'in practice', between the way things are 'in themselves' and the way they are 'for us', and between the 'possible' and the 'actual'. The flfSt one hundred and some sections of Philosophical Investigations presents a concentrated attack on a platonist view of logic which portrays logic, properly symbolized, as an ideal form of language of which actual languages only approximate. Wittgenstein's historical targets here are Frege, Russell, and his own Tractatus. From these sources comes the notion of the logically perfect language which is "of the purest crystal" in contrast to the "empirical cloudiness and uncertainty" of ordinary language.110 ("The proposition and the word that logic deals with are supposed to be something pure and clear-cut," in contrast with "what are ordinarily called 'propositions', 'words', 'signs'. "lll) On the model that Wittgenstein latterly attacks, logic is conceived of as presenting language as it is 'ideally' or 'in theory' or 'in principle'- it is "a logic for a vacuum,"112 ignoring the friction and air resistance which besets actual language 'in fact' or 'in practice'. We have already cited Wittgenstein's talk in Remarks on the Foundations of Mathematics of the conception of logic as an ideal kind of physics - an 'ultra-physics' - and this goes with his critique, in Investigations 89, of logic as "something sublime" relative to the emprical investigation of the empirical world Frege repeats this pattern as regards the relation of logic to psychology. Logic presents, as it were, the ideal a priori laws of thought, or laws of an ideal thinker, while psychology presents the empirical laws of thought, the laws of actual thinking. As Frege puts it in "Thoughts": "Neither logic nor
142
PHILIP DWYER
mathematics has the task of investigating minds and contents of consciousness owned by individual men. Their task could perhaps be represented as the investigation of the mind; of the mind, not of minds.ttll3 The logical/empirical distinction leads to a 'two-kinds' thesis about anything which exemplifies the logical/empirical distinction. Thus, insofar as the logical/empirical distinction is exemplified by the logical/psychological distinction, there ar;e two kinds of 'mind', two kinds of thinking, speaking, inferring, or rule-following: there is logical thinking or inferring, as it were, ideal inferring or inference-in-itself, and psychological inferring, what actual people actually do; ideal or logical rule-following, objectively and in itself, and actual rule-following as it is carried out in actual cases, etc. The ideal version transcends the actual version, as theory transcends practice, as noumena transcend phenomena, as the 'in principle' transcends the 'in fact', and the logical transcends the empirical. It is any such two-kinds thesis that Wittgenstein rejects as regards language and other meaning-related or intentional phenomena. "The philosophy of logic,tt Wittgenstein says, ttspeaks of sentences and words in exactly the sense in which we speak of them in ordinary life when we say e.g. 'Here is a Chinese sentence', or 'No, that only looks like writing; it is actually just an ornament' and so on. We are talking about the spatial and temporal phenomenon of language, not about some non-spatial, non-temporal phantasm. ttl14 Wittgenstein's later philosophy consists of several ,antittanscendence', or following Dummett, 'anti-realist' arguments, to the effect that, for some given intentional phenomenon, there is no way that phenomenon is 'in itself or 'in principle' transcending the way it is 'for us' or 'in fact'. For example, taking Wittgenstein's rule-following considerations llS as comprising an anti-transcendence argument, their upshot is that a rule does not transcend its application (meaning does not transcend use). There is no abstract, objective rule-in-itself transcending what Frege would call the 'extraneous empirical, psychological matter'116 or our actual application or 'following' of a rule. One might put this in tenns of a two-kinds thesis this way: there are not two kinds of rule-following- logical or 'theoretical' rule-following which, as it were, the rules themselves accomplish, versus psychological or 'practical' rule-following, the anthropological phenomenon managed by us down here on earth. Rules and rule-following are all practice. That there is 'following', that there are actions which we all agree upon as cases of following', produces the illusion 'that there are rules independent of our actioo-in-agreement, and which guide our action and thereby produce the agreement. But roles are variously interpretable and when it comes to acting on a rule there will be no interpretation apart from my action. To put the point another way, insofar as a rule requires application, there can be no rule for that, no rule for the application of a rule: in acting on a rule, nothing further can guide me- I must simply act. Now as far as any rules (i.e., expressions of roles) go, you may act one way and I another- but for the most part this doesn't happen. Our
ANTI-PSYCHOLOGISM AND SCEPTICISM
143
widespread action-in-agreement or 'form of life' is logically primitive. It is not founded on rules, but rather the other way round: there are 'rules' because there is 'following'. The practice is the intentional phenomenon, the abode of meaning, and not some transcendent non-spatial, non-temporal rule. Intentional phenomena are exhausted by their actual manifestation. There is no language or sense in itself (nor 'simpies'),117 nor definitions (in themselves),118 or in principle, or in theory, versus language for us, or in fact, or in practice (Le., the language-game), and no understanding, or thinking, or inferring in themselves, or in principle, or in theory, versus these things as they actually occur for us, in fact, in practice, Le., again, in the language-game. Such is the message of Wittgenstein's rule-following considerations as an anti-transcendence or anti-realist argument. Wittgenstein's considerations about rule-following are, however, only part of the extended anti-transcendence argument that he pursues in undermining a two-kinds thesis about meaning, which is, to repeat, a thesis that meaning and related intentional phenomena are one way in themselves or in theory transcending the way they are for us. A further central component in Wittgenstein's account of language and meaning might best be called his 'Gestalt considerations', the upshot of which is that linguistic understanding is largely perceptual and that meaning is phenomenal or perception-andunderstanding-dependenl That is to say, there is no way that meaning or sense is - no story to tell about the nature of meaning - apart from the way meaning, or the things that have meaning, are for us perceiving subjects. There is no objective 'meaning-in-itself' transcending' meaning-for-us': meaning is essentially subject or person-dependent. Space is lacking here to document and discuss in detail this surprisingly under-discussed aspect of Wittgenstein's later philosophy.119 His discussion of music, gestures, and physiognomy in connection with language and meaning is neither incidental nor infrequent (see any of his later works, except perhaps On Certainty). Particularly important is the connection he draws between 'seeing-as' and rule-following!20 As Wittgenstein's treatment of language is of a piece with his treatment of mathematics, it is not surprising that Gestalt themes - shape, pattern, physiognomy, music, seeing an aspect, notation and signs - run throughout Remarks on the Foundations of Mathematics. Wittgenstein's claim that surveyability has to do with the essence (Wesen) 121 of proof, and that ttif you have a proof pattern that cannot be taken in, and by a change in notation you tum it into one that can, then you are producing a proof, where there was none beforettl22 is an instance of Wittgenstein's Gestalt approach and its anti-transcendence consequences. It is the claim that proof, as a matter of sense or meaning, is phenomenal, Le., understandingdependent (perception-dependent). Surveyability is just the kind of thing that Frege would pass off as an extraneous, psychological matter, and Wittgenstein addresses the issue in these terms, as he continues Remarks on the Foundations of Mathematics, Part nI, section 2:
144
PHILIP DWYER
Now let us image a proof for a Russellian proposition stating an addition like 'a+b=c', consisting of a few thousand signs. You will say: Seeing whether this proof is correct or not is a purely external difficulty, of no mathematical interest. (One man takes in easily what someone else takes in with difficulty or not at all, etc. etc.) The assumption is that the defmitions serve merely to abbreviate the expression for the convenience of the calculator; whereas they were part of the calculation. By their aid expressions are produced which could not have been produced without them: 23
It would be quite idle here to make a distinction between a logical 'could' and a psychological or otherwise non-logical 'could' in 'expressions which could not have been produced without the definitions'. There are not two kinds of proof - the proof in itself or in theory, and the proof for us in practice - nor two kinds of understanding of a proof- that achieved by individual men, and that achieved by Frege's 'the mind'. Insofar as surveyability is of the essence of proof and, more generally, insofar as proof is subject-dependent, Wittgenstein could just as well speak of the 'logical', as of the 'psychological disadvantages' of the Russellian symbolism,l24 i.e., insofar as logic is supposed to be the essence of proof.
VI. Anti.Psychologism and Scepticism The two-kinds thesis is the basis of Frege's anti-psychologism. It is also the traditional basis of scepticism, as in Descartes. The link that the two-kinds thesis provides between anti-psychologism and scepticism is such that Frege's anti-psychologism is intimately tied to a sceptical element in his thought This is why, in the midst of the discussion of Frege and psychologism in Remarks on the Foundations of Mathematics, there should occur the sequence of three remarks concerning scepticism, Le., doubt and error. The frrst two of the remarks run as follows: Imagine the following queer possibility: we have always gone wrong up to now in multipying 12 x 12. True, it is unintelligible how this can have happened. So everything worked out in this way is wrong! - But what does it matter? It does not matter at all! - And in that case there must be something wrong in our idea of the truth and falsity of arithmetical propositions. But the~ is it impossible for me to have gone wrong in calculation? And what if a devil deceives me, so that I keep on overlooking something however often I go over the swn step by step? So that if I were to awake from the enchantment I should say: ''Why was I blind?" - But what difference does it make for me to 'assume' this? I might say: "Yes to be sure, the calculation is wrong- but that is how I calculate. And this is what I now call adding, and this 'the sum of these two numbers'. "12S
ANTI-PSYCHOLOGISM AND SCEPTICISM
145
One's immediate reaction to Remarks on the Foundations of Mathematics I, 135 is 'Why doesn't it matter if we've always gone wrong in multiplying 12 x 12?' The sense of 'not mattering' is that given in Part I, section 136 about the possibility of a deceptive demon, viz. that such a hypothesis 'makes no difference'. (Again at Remarks on the Foundations of Mathematics ill, 21, Wittgenstein says: "One might say: When it can be said: 'Even if a demon had deceived us, still everything would be all right', then the prank he wanted to play on us has simply failed on its purpose.") In what sense does the hypothesis of an evil genius make no difference? In the remark following Part I sections 135-6, Wittgenstein asks us to imagine someone "bewitched" so that in counting by threes, every third object gets counted over again as the fIrst object in the next group of three; so that foUT 'goups of three', plus two, equal ten- or, 4x3+2=10. Wittgenstein suggests that such a calculation could be applied quite happily- "he takes three nuts four times over, and then 2 more, and he divides them among 10 people and each gets one nut; for he shares them out in a way corresponding to the loops of the calculation, and as often as he gives someone a second nut, it disappears. "126 So here the person's 'bewitchment', as we see it, makes no difference in the sense of making no difference to the practice with, or application of, the calculation- in the event, everything is, so to say, copacetic. But Wittgenstein's general point is not simply that there could be cases where bewitchment or deception, as it happens, make no 'practical difference'. The general point is that the only difference that something can make in mathematics, is a practical difference. The only 'differences' in mathematics are practical differences; to make no difference to a pmctice is to make no difference at all. The practice, with its paradigmatic status (this is what I now call 'adding' and this 'the sum of these two numbers'), detennines the sense of what is called 'adding' or 'the sum of these numbers' (as well as what is called 'wrong', 'correct', 'mistake', etc.) and it is not some practice-independent sense which detennines, or perhaps fails to detennine the practice. There is only practice, only application, with no 'theory' backing up the practice. Again, this pragmatism concerning mathematics is the moral of the rule-following considemtions as an anti-tmnscendence argument. To say that there is no theory tmnscending the pmctice is to say, in Fregean tenns, that there is no 'logic' transcending the 'psychology'. But the hypothesis of the evil demon depends on this distinction. If the notion of a theoretical realm which tmnscends the practical realm is idle, then so too is this particular exploitation of that notion: the theoretical possibility of error which tmnscends what is recognized as an actual error in practice. The notion is idle, it makes no difference at all. We might put the point this way: the demon is incapable of deceiving us about the practice- without doubt it exists also if he deceives us, and let him deceive us as much as he will, he can never cause it to be nothing or illegitimate so long as we think that it is legitimate. So long as the pmctice
146
PHILIP DWYER
exists at all, even as a dream, we have paradigms and so the means of judging or recognizing error and validity. The senses of 'wrong', 'correct', ,mistake', etc. are internally tied to our practice, and our practice to paradigms. And the judging of error and right is ever more practice, precedent, and paradigm. In its practice-cum-recognition-dependence error and right, like (and as) the practice, are autonomous. (Begone Satan!) The claim then is that error, like other intentional or meaning-related phenomena, is phenomenal or recognition-dependenl But its phenomenality is not just a matter of its being constituted by and exhausted in our practices. Mathematics and language are perceptual or Gestalt phenomena. Contrary to Frege, notation is not incidental to the possibility of thoughts, and in language and mathematics the intellect is a power of the eyes and ears, hand and tongue. Perception and understanding are one, and recognition-dependence is perception-dependence. There is no error 'in itself' apart from error 'for us', apart from what we can see as error. In part III of Remarks on the Foundations of Mathematics, Wittgenstein writes: Can we be certain that there are no abysses now that we do not see? But suppose I were to say: The abysses in a calculus are not there if I do not see them! Is no demon deceiving us at present? Well, if he is, it doesn't matter. What the eye doesn't see the heart doesn't grieve over!27
The kind of error at stake in mathematics is, of course, contradiction. Doubt about there being a contradiction lurking in mathematics is what motivates Frege's project of giving mathematics a fmn foundation in logic. It is this doubt, as opposed to, so to speak, a 'naively realistic' certainty about mathematics, that constitutes the sceptical element in Frege's thought, and links him to Cartesian scepticism. Near the end of Part III of Remarks on the Foundations of Mathematics, in section 87, Wittgenstein makes the comment about looking at contradiction 'anthropologically', i.e., in terms of language-games. But in the immediately following section he can be seen to be looking at contradiction anthropologically in another sense, when he offers the following as a possible reply to the suggestion that without a consistency proof one cannot be certain that a calculus is free of contradiction: "If my conception of the calculus should sometime alter; if its aspect should alter because of some context that I cannot see now- then we'll talk some more about it...! do not see the possibility of a conttadiction. Any more than you - as it seems - see the possibility of there being one in your consistency proof. "128 Again, the suggestion is that conuadiction is perception-dependent and this is reinforced by the comparison, in section 89 which follows, between a proof and a tune. In Part III, section 78, prior to the comment about abysses in the calculus being phenomenal, Wittgenstein compares a mistake in our rules to a mistake in a musical theme, in how it sounds.__~~~~ _ ~~,!~e~ipg
ANTI-PSYCHOLOGISM AND SCEPTICISM
147
phenomenality, he says of a game whose roles make it pointless, that it stopped being a game only when this was pointed out. In general it may be said of a contradiction that, being a matter of language, it is a matter of meaning, and meaning is always phenomenal. In section 90, which ends Part Ill, Wittgenstein says: I have not yet made the role of miscalculation clear. The role of the proposition: 'I must have miscalculated'. It is really the key to any lDlderstanding of the 'fOlDldatiOns' of mathematics.
One significant thing about 'I must have miscalculated' is that it is said only of a recognized error. And the way that the phenomenality or recognition-dependence of error relates to the 'foundations' of mathematics is this. To the extent that error is recognition-dependent, there is no such thing as 'error-in-itseIr. There are not two kinds of error, nor two kinds of correctness or consistency- the kind we recognize or acknowledge versus Frege's or God's kind Likewise there are not two kinds of error-ridden or error-free inference. There is only actual (hwnan) inference. Thus of the employment of a consistency proof, Wittgenstein earlier says, "It is - I should like to say - for practical, not for theoretical purposes that the disorder is avoided. "129 He goes on to say that no consistency proof will make it "certain that people will never want to calculate differently," "that we shall never want to look at reality differently": "It is not the eternal correctness of the calculus that is supposed to be assured, but only so to speak, the temporal."l30 Wittgenstein's notion of the 'eternal' here would be no less psychologistic for Frege than the notion of the temporal, but Wittgenstein's point is clear: what, at bottom, is at issue regarding correctness in mathematics is actual ways of thinking, actual practicesthat's all there is. In Part VII of Remarks on the Foundations of Mathematics, sections 11-15, Wittgenstein looks at contradiction anthropologically in the sense of considering people with a different calculus than ours. The discussion centers around the notion of 'a hidden contradiction'. This is an important notion for Frege; 131 one of those threads which serves to intertwine his platonism, logicism, and anti-psychologism; and, of course, it is an expression of his scepticism. In Philosophical Investigations, 125, Wittgenstein compares contradiction in mathematics to a game which turns out other than expected when we follow the rules, as he does here in Remarks on the Foundations of Mathematics Vil, 13, and as he did earlier in Remarks on the Foundations of Mathematics III, 78. In the Philosophical Investigations he concludes with the remark: "The civil status of a contradiction, or its status in civil life: there is the philosophical problem. "132 The 'civil status' of a contradiction is its phenomenal status, "for," as he goes on to say in Philosophical Investigations, 126, "what is hidden, for example, is of no interest to us." Indeed, insofar as a contradiction is phenomenal, it makes no sense to speak of its being 'hidden'. Anymore
148
PHILIP DWYER
than it would make sense to speak of a 'hidden consistency' in a set of rules, for consistency is no less phenomenal (it is a Gestalt quality). This is brought out by Wittgenstein's comparison of the consistency of a calculus with the consistency of various musical themes. l33 There is no such thing as 'consistency-in-itself'; what 'consistency' is, is not established in advance of particular cases. 134 The phenomenality of contradiction, and of mathematics generally, is brought out in another sense with this remark: "If the contradiction is so well hidden that no one notices it, why shouldn't we call what we do now 'proper calculation'?"135 This is to say, 'proper calculation' is not something transcending our actual calculation when perceived to be proper. Such calculation is not any the less "mathematics in the fullest sense"l36 for being without some 'guarantee' that we shall never come across a contradiction. Moreover, whatever is proposed as insurance against contradiction will be no less in need of such insurance itself. Thus the deceptive demon meets up with the good angel: "Up to now a good angel has preserved us from going this way! Well, what more do you want? One might say, I believe: a good angel will always be necessary, whatever you do."137 While still in Part VIII of Remarks on the Foundations of Mathematics, Wittgenstein quotes his interlocutor: "We want, not just a fairly trustworthy, but an absolutely trustworthy calculus. Mathematics must be absolute.'''138 Again in Part VIII, 36, the interlocutor exclaims, '''But surely this isn't ideal certainty!'." With this notion of the absolutely trustworthy, of ideal certainty, we get, as it were, the subliming of our whole account of knowledge, certainty, and doubt. l39 In other words, we get a two-kinds thesis of certainty and doubt, relying on the 10gicaJ/psychological distinction. In the introduction to The Foundations of Arithmetic, shortly following his initial differentiation of logic from 'psychology', Frege says the following concerning definitions: Yet it must still be borne in mind that the rigolD' of the proof remains an illusion. even though no link be missing in the chain of our deductions, so long as the definitions are justified only as an afterthought, by our failure to come across any contradiction. By these methods we shall never have achieved more than an empirical certainty, and we must really face the possibility that we may still in the end encounter a contradiction which brings the whole edifice down in ruins. l40
The important notion here for understanding the relation between antipsychologism and scepticism is that of "an empirical certainty." The Begriffschrift opens with the remark that "in apprehending a scientific truth we pass, as a rule, through various degrees of certitude," and goes on to distinguish induction from logical proof. 141 This distinction is correlated with that between how we arrive at a proposition and how it may be justified,142 which is the basis for Frege's distinction between psychology and logic. In The Foundations of Arithmetic, induction is distinguished from
ANTI·PSYCHOLOGISM AND SCEPTICISM
149
logical proof which places "the bUth of a proposition beyond all doubt,nlH and there Frege speaks of induction as itself "a psychological phenomenon."144 The distinction between induction and deduction exemplifies the two kinds of certainty: empirical or psychological or practical certainty versus logical, absolute certainty. 'Scepticism' as I am concerned with it is something that arises in connection with a traditional epistemological project going under the name 'the search for certainty'. Or again, scepticism starts with the question 'Is there anything of which we can be certain?' As such, it depends on the above type of two-kinds thesis of certainty and doubt. For the natural response to any sceptical proposal that it is not certain that, e.g., here is a hand,!45 is: 'But I am certain that here is a hand'. And the immediate sceptical response to this is: 'Oh that's just a psychological certainty. Admittedly there is no psychological or practical doubt. but there is a theoretical or metaphysical doubt about it'. So Descartes proceeds. The 'supreme' kind of doubt which is 'metaphysical' and 'hyperbolical' is what is required to investigate bUths that are 'metaphysically certain' .146 This 'metaphysical mode of knowing' contrasts with "the moral mode of knowing, which suffices for the regulation of life."147 What is known by the latter is a matter of "long and familiar CUStom,"14' of "blind impulse,nl49 rather than reason. The practical certainties yielded, though sufficient, and necessary, for everyday life, are nevertheless "opinions in some measure doubtful. "150 They are psychologically certain, but not metaphysically certain. Wittgenstein does not accept this two-kinds thesis of certainty and doubt. In On Certainty he writes: The statement 'I know that here is a hand' may then be continued: 'for it is my hand that 1 am looking at'. Then a reasonable man will not doubt that 1 know. -Nor will the idealist; rather he will say that he was not dealing with the practical doubt which is being dismissed, but there is a further doubt behind that one. -That this is an illusion has to be shewn in a different way.151
The notions of a transcendent doubt beyond practical doubt, and of an absolute certainty beyond mere 'empirical' certainty are applications of the general notion of a logical realm which transcends the 'psychological' realm as the ideal transcends the actual. U2 In the Philosophical Investigations, (84) Wittgenstein makes the point that the logical possibility of a doubt, i.e., the possibility of imagining a doubt about such and such, does not mean that one actually doubts such and such. The possibility of a doubt does not make something doubtful. The sceptic, relying on the two-kinds thesis, will say: 'Oh no, not doubtful for you, but still doubtful'. If something is doubtful, even though no one doubts it, then it must be doubtful in itself. It is this latter notion that is an illusion. There are not two kinds of doubt. Doubt is actual doubt, i.e., doubts actually had by actual people. There is no doubt or doubtfulness in itself.
150
PHILIP DWYER
The possibility of doubt is not a kind of doubt. However, on the two-kinds thesis it is. In the 'actual world', the realm of logical possibility is just that- the realm of possibilities. But in the logical realm 'itself' the realm of possibilities sublimed - possibilities are the actualities. In the logical realm, a doubt that is logically possible is actual, while certainty metaphysical certainty - is impossible if doubt is logically possible, i.e., metaphysically actual. This is the realm of doubt in itself. To the exclamation 'But this isn't ideal certainty', Wittgenstein replies, in Remarks on the Foundations of Mathematics, with the question "-Ideal for what purpose?" The point of bringing in the question of purpose is to challenge the idea that there is a realm of certainty and doubt in themselves, leaving us out of account The question also refers us to Philosophical Investigations 87-8, where Wittgenstein criticizes Frege's notion of ideally exact concepts, explanations, and rules, as opposed to the actual concepts, explanations, and rules in our ordinary language. It is consideration of this which brings up the problem: "In what sense is logic something sublime?"153 The two-kinds thesis about rules goes with the two-kinds thesis about doubt and certainty, and leads to the illusion that a methodological scepticism is required if 'real certainty' - certainty in itself - is to be attained. Here Wittgenstein addresses indifferently Fregeans and Cartesians: It may easily look as if every doubt merely revealed an existing gap in the foundations; so that secure lDlderstanding is only possible if we fast doubt everything that can be doubted, and then remove all these doubts. 1St
The possibility of a doubt does not reveal an existing doubt in the alleged metaphysical foundations of our actual understanding of rules. There are no such transcendent foundations; our actual understanding is self-sufficient. Against the above suggestion of a possible doubt constituting an actual gap in the 'foundations', Wittgenstein says, comparing a rule to a signpost, "The sign-post is in order,- if, under nonnal circumstances, it fulfills its purpose."155 Secure understanding - certainty - is achieved so long as we actually have no 'practical' doubts, i.e., so long as 'the rules fulfil their purpose, i.e., so long as they serve our purposes. There is no purposetranscendent order which a rule may 'ideally' attain to (ideals themselves are strictly relative to purposeS),I56 and no practice-transcendent certainty to which our knowledge of a rule may attain. Philosophical Investigations, 87 is echoed in On Certainty: What I need to show is that a doubt is not necessary even when it is possible. That the possibility of the language-game doosn't depend on everything being doubted that can be doubted. (This is connected with the role of contradiction in mathematics.)I'"
I take the fIlSt two sentences to mean: the logical possibility of a doubt doesn't yield a doubt, for the purposes of establishing the language-game.
ANTI-PSYCHOLOGISM AND SCEPTICISM
151
That is to say, possible doubtfulness isn't a kind of doubtfulness. There are not two kinds of doubtfulness and certainty. The only doubtfulness is actual, practical doubtfulness, the only certainty actual, practical certainty. For Descartes, that for which it is logically impossible to be mistaken is 'metaphysically' more certain than that for which it is logically possible to be mistaken. But 'metaphysically more certain' does not mean 'more certain'. For one's certainty to be grounded on the law of contradiction does not necessarily mean that it is more certain than one that isn't so grounded. 'Metaphysically more certain' does not mean 'more certain' because 'metaphysical certainty' is not a special kind of certainty vis vis ordinary practical certainty. The connection between On Certainty, 392 and contradiction in mathematics is this. Contradiction being phenomenal, there is no contradiction-in-itself transcending what we actually recognize as a contradiction. Correlatively, there is no doubtfulness-in-itself transcending what is actually doubtful for us or to us. The notion of a theoretical doubtfulness transcending what we actually hold to be doubtful is as idle as the notion of a hidden contradiction. Indeed, the latter is a version of the fonner. Curiously, Descartes argues in the very same way when it is suggested that his 'clear and distinct perceptions' may nevertheless be doubtful. In The Reply to the Second Set of Objections he says:
a
To begin with, directly we drink that we rightly perceive something, we spontaneously persuade ourselves that it is true. Further, if this conviction is so strong that we have no reason to doubt concerning that of the truth of which we have persuaded ourselves, there is nothing more to enquire about; we have here all the certainty that can reasonably be desired. What is it to us, though perchance someone feigns that that, of the truth of which we are so frrmly persuaded, appears false to God or to an Angel, and hence is, absolutely speaking, false? What heed do we pay to that absolute falsity, when we by no means believe that it exists or even suspect its existence? We have asswned a conviction so strong that nothing can remove it, and this persuasion is clearly the same as perfect certitude.Ut
This is exactly how I should argue against the sceptic concerning the certainty of, e.g., the Moorean 'Here is a hand'. The sceptic I should then be arguing against would, of course, be Descartes. All the arguments he used above against the notion of a certainty transcending the certainty of his 'clear and distinct perceptions', I use against the notion of a certainty transcending the certainty of 'Here is a hand'a clear and distinct perception if ever there was one. What is it to me- though perchance Descartes feigns that that, of the truth of which I am so frrmly persuaded, appears false to God or an Angel or a sceptic, and hence is, absolutely speaking, false? What heed should I pay to that absolute falsity when I by no means believe that it exists or even suspect its existence? I have
152
PHILIP DWYER
assumed a convlctton so strong that nothing can remove it, and this persuasion is clearly the same as perfect certitude. Descartes will ask how I know that I am awake and not dreaming. The answer to the question 'How does one know one is awake?' is, I believe, 'By being awake'. If this is deemed circular, it is no more so than the answer Descartes must give to the question, 'How do you know your perceptions are clear and distinct?'I59 Alternatively, one may say that there are no criteria by which we distinguish being awake from dreaming, but that does not mean we can't and don't distinguish them. We certainly do: 60 It will be said that the point about dreaming or the evil demon is simply that a mistake about my hand is logically possible. But again, that a mistake is logically possible - that 'I was mistaken about my hand' is compatible with the laws of logic - does not mean that possibly I am mistaken (anymore than that it is logically possible for a woman to turn into a pillar of salt means that a woman could tum into a pillar of salt). The logical possibility of error in no way provides grounds for doubt, especially in the particular case, as with 'Here is a hand'. Finally, it will be said, 'The question is one about justified certainty'. But my certainty in 'Here is a hand' is justified. The only way it could fail to be is if there were two kinds of justification, one of which - 'metaphysical justification' - was the only true kind. But inasmuch as there is no such thing as metaphysical certainty (as distinct from certainty), metaphysical justification can only be justification, i.e., a version of what the notion of 'metaphysical justification' presupposes, viz., ordinary justification. And there is no question that by these basic and not unrigorous standards of justification, certainty in 'Here is a hand' is justified, and more than justified. There is really only the short way with the sceptic. He says he wants a case of something certain. We say: 'Nothing is easier: "here is a hand"'. And like Descartes above, we say "there is nothing more to enquire about; we have all the certainty that can reasonably be desired." This should be the end of the discussion, if what is sought is something that is certain. The discussion goes on, of course, because of the two-kinds thesis of certainty, which allows for a so-called 'methodological scepticism'. But on the line we have advanced, 'methodological scepticism' can be nothing other than full-blooded scepticism- for there are not two kinds of scepticism. Thus anyone who finds methodological scepticism tenable is, if unwittingly, a sceptic. And there are many such.
VIL Unwitting Sceptics The hallmark of scepticism is that it offends against commonsense or 'vulgar thought'. The vulgar hold that 'Here is a hand' is certain; the sceptic holds that it is dubious. So-called methodological scepticism,
ANTI·PSYCHOLOGISM AND SCEPTICISM
153
however it may be redeemed in the end, or whatever it may come up against that is immune to 'methodological doubt' is no less offensive to common sense than is 'outright scepticism'. Indeed, it is outright scepticism. There are not two kinds of doubt. Either 'Here is a hand' is doubtful or it is not doubtful. If it is not doubtful the game is over: we have something certain. If it is doubtful the game is over:- scepticism wins. Methodological scepticism gets its running start when it is thought that there are two kinds of doubt and certainty. Methodological doubt will be 'metaphysical doubt'. But since there are not two kinds of doubt, if metaphysical doubt is doubt at all, then it is full-blooded practical doubt To say, '1 act as though it is certain, even though it is (theoretically, ultimately, strictly speaking, etc.) doubtful, does not mitigate one's scepticism. One can't be a 'theoretical sceptic' but a 'practical believer'. H one claims to have a 'theoretical doubt', then one doubts- however one acts. Again, the only question concerning, e.g. 'Here is a hand', is whether or not it is doubtful. To say, regarding this particular case (i.e., with someone's hand stuck in your face): 'This is doubtful', is to be a full-fledged practicing sceptic. For Descartes to say, indeed insist, that propositions such as the Moorean 'Here is a hand' are "not entirely certain and indubitable,"161 are "opinions in some measure doubtful,"I62 are "merely probable,"163 and in general for Descartes to say "You can reasonably doubt of all things, the knowledge of which comes to you by the senses alone"l64 is for Descartes to be a sceptic and nothing less. Similarly, any expositor of Descartes who does not immediately pull him up short at the stage where he introduces his method of doubt, that is, anyone who accepts the project of methodological doubt, whatever else in Descartes he may reject, must be deemed a sceptic. This includes in my experience, most expositors of Descartes. A good example is Bernard Williams in his Descartes: The Project of Pure Enquiry. In his discussion of Descartes' method, Williams quotes the passage from the Discourse on Method where Descartes says, "so far as practical life is concerned, it is sometimes necessary to follow opinions which one knows to be very uncertain, just as though they were indubitable...but because I wanted to devote myself solely to the search for truth, I thought that I should reject, as though it were absolutely false, everything in which I could imagine the slightest doubt, so as to see whether after that anything remained in my belief which was entirely indubitable,"165 and the passage from the "First Meditation" which begins, "Reason persuades me already that I should withhold assent no less carefully from things which are not clearly certain and indubitable, as from things which are evidently false. "166 The things which Descartes says one 'knows to be very uncertain', or which at least are 'not clearly certain and indubitable' include such things as 'the Moorean 'Here is a hand'. Williams makes no objection to calling such things uncertain. His only objection, at this point, is that Descartes' strategy for 'fmding the truth', "is surely far from obvious: we constantly want the truth about various matters, but hardly ever demand the
154
PHILIP DWYER
indubitable. nl67 But this is surely far from obvious. Insofar as we want the truth about various matters, we certainly don't expect to get something that is open to doubt, and for the most part, we don't. There is no need to demand the indubitable; it is the nonnal thing. Having asked the whereabouts of the cat, we are told that it is on the mat If there are any doubts about this, they can be defmitively resolved, but under normal circumstances, the infonnation is absolutely indubitable. But not, apparently, for Williams. He goes on to say that Descartes' sttategy "is to aim for certainty by rejecting the doubtful,"I. it "is to preserve, out of his present beliefs, any that are genuinely certain, and the way to do that is to set aside the ones that are not."I69 If, as seems to be the case, 'The cat is on the mat' or 'Here is a hand' are, for Williams, not genuinely certain, then Williams is a sceptic. For again, there are not two kinds of genuine certainty. Nor, again, are there two kinds of genuine doubt Williams says that there is no question "of the hyperbolical doubt playing any rational role within ordinary life."170 But there are not two kinds of rational doubt, nor in geneml, two kinds or ways of being rational. In his book Descartes Against the Skeptics, E.M. Curley says that, for Descartes, indubitability "is a matter of not having reasonable grounds for doubt, "1'71 and that for Descartes, "the ground of doubt is required to be not entirely frivolous."172 But the most striking thing about Descartes' or any sceptic's grounds for doubt - that his senses have once deceived him, that he may be dreaming, that there may be an evil genius - is that they are not reasonable and are entirely frivolous. There are not two kinds of reasonableness and frivolity. This is why, in On Certainty, Wittgenstein constantly poses the issue about doubt in terms of what 'the reasonable man' believes. 173 Curley speaks of the constraints on the classical sceptics "if their position was to have plausibility at all."174 But Curley, going along as he does with Descartes, has given up all right to demand plausibility from anyone. There are not two kinds of plausibility. In On Certainty, Wittgenstein asks, "Can I give the supposition that I have ever 1x~n on the moon serious consideration at all?"17S The notion of seriousness is really the key notion in any discussion of scepticism- that is, in any serious discussion. For again there are not two kinds of seriousness: 'practical seriousness' and 'philosophical seriousness'.176 The words 'serious', 'reasonable't 'frivolous', 'plausible', not to mention ,doubtful' and 'certain', are ordinary words of English, whose meanings do not change at the whim of a philosopher. (Certainly, William's capitalization - thus, 'Doubt'ITI - does nothing to alter the ordinary meaning of the word, Le., does nothing to establish that there is another kind of doubt than everyday 'practical' doubt) There is only one way to be serious about something, and that is, so to say, to be practically serious. If it is admitted that the 'doubt' about 'Here is a hand' is not 'practically' serious, then it is admitted that it is not serious at all, in which case it needn't be taken seriously at all. There is then no sceptical problem, as indeed there
ANTI·PSYCHOLOGISM AND SCEPTICISM
155
isn't. If the doubt about 'Here is a hand' is at all serious, then the only serious response is to get the man help!"
156 ENDNOTES 1.
Michael Dummett, Frege, Philosophy of Language (London: Duckworth,
1973), pp. 666-667. 2. Ibid., p. 669. 3. See, for example: Rene Descartes, Principles of Philosophy, in The Philosophical Worb of Descartes Vols. I and fi, trans. E.S. Haldane and G.R.T. Ross (Cambridge: Cambridge University, 1968), Vol. I, p. 211. 4. Dummett, Frege, Philosophy of LangllQ.ge, p. 92. 5. Ibid., p. 5. 6. Michael Dummett, Truth and Other Enigmas (London: Duckworth, 1978), p. 420. 7. Dummett, Frege, Philosophy of Language, pp. 589-590. 8. Ibid., p. 666. 9. Ibid., p. 667. 10. Ibid., p. 669. 11. Ibid. 12. Ibid., p. 672. 13. Ibid., p. 674. 14. Ibid., p. 675. 15. Ibid., pp. 675-676. 16. Gottlob Frege, Th£ Foundations of Arithmetic, trans. I.L. Austin (Oxford: Basil Blackwell, 1953), p. xi. 17. See: Dwnmett, Frege, Philosophy of Language, p. 665. 18. See: Rene Descartes, Discourse on the Method of rightly conducting the Reason and seeking for Truth in the Sciences, in The Philosophical Works of Descartes, Vol. I, pp. 123, 127. 19. See: Rene Descartes, Meditations on First Philosophy, in Ibid., p. 140. 20. Ibid., p. 135. 21. Ibid., p. 148. 22. Oottlob Frege, Begriffschrift, in Frege and GOdel. Two Fundamental Texts in Mathematical Logic, 00. I. van Heijenoort (Cambridge: Harvard University, 1970), p.
5. 23. Ibid. 24. Ibid., p. 7. 25. Ibid., p. 35. 26. Ibid., p. 23. 27. See: Gottlob Frege, Posthumous Writings, ed. H. Hennes, F. Kambartel, F. Kaulbach, with 'the assistance of o. Gabriel and W. Rodding, trans. P. Long and R. White (Chicago: University of Chicago, 1970), p. 3. See also: Oottlob Frege, Logical Investigations, trans. P.T. Oesch and R. Stoothoff (Oxford: Basil Blackwell, 1977), p. 1. 28. See: Gottlob Frege, Posthumous Writings, pp. 133, 258-262. See also: Gottlob Frege, Logical Investigations, p. 25. 29. Ibid., p. 2. See also: Gottlob Frege, ''The Aim of the Begriffschrift," in Conceptual Notation and Related Articles, 00. and trans. Terrel Ward Bynum (Oxford: Clarendon, 1972), p. 95.; and Gottlob Frege, The Foundations of Aritlunetic, p. 99. 30. Gottlob Frege, Posthumous Writings, p. 2. 31. Descartes, The Philosophical Worb of Descartes, Vol. I, p. 212. 32. Ibid., Vol. I, p. 92. Cf., Vol. I, p. 14; and Vol. IT, pp. 52-59.
ANTI-PSYCHOLOGISM AND SCEPTICISM
157
33. Gottlob Frege, The Basic Laws of Arithmetic, 00. and trans. Montgomery Furth (Los Angeles: University of California, 1964), p. 2. 34. Frege, The FOIUUlations of Arithmetic, p. 103. 35. Frege, The Basic Laws of Arithml!tic, pp. 4-5. 36. Ibid., p. 3. 37. Descartes, The Philosophical Works of Descartes, Vol. I, p. 20. 38. Ibid., p. 8. 39. Frege, The Basic Laws of Arithml!tic, p. 4. 40. See: Frege, PosthlUnOlLf Writings, pp. 205, 210; Frege, The Foundations of Arithmetic, p. 102; and Frege, The Basic Laws of Arithmetic, p. 121. 41. Frege, PosthlUnOlLf Writings, p. 205. 42. This lDlavoidable psychological dimension of the epistemological discussion is one which Frege is never comfortable with and ultimately, for all his explicit anti-psychologis~ never squarely faces. Thus the following: ''The question why and with what right we acknowledge a law of logic to be true, logic can answer only by reducing it to another law of logic. Where that is not possible, logic can give no answer. If we step away from logic, we may say: we are compelled to make judgments by our own natme and by external circumstances; and if we do so, we cannot reject this law- of Identity for example; we must acknowledge it unless we wish to reduce OlD' thought to confusion and finally renounce all thought whatever. I shall neither dispute nor support this view; I shall merely remark that what we have here is not a logical consequence. What is given is not a reason for something's being true, but for our taking it to be true." Gottlob Frege, The Basic Laws of Arithmetic, p. 15. 43. Descartes, The Philosophical Works of Descartes, Vol. I, p. 232. 44. Ibid., p. 140. See: p. 144, and see also the famous discussion of the wax in the "Second Meditation." 45. See: Ibid., p. 174. 46. Frege, The Foundations of Arithml!tic, p. 115. See also: pp. 36, 38; and BynWll, ed., ConceptlllJ1 Notation and Related Articles, pp. 84-86. 41. Frege, The FOIUUlations of Arithmetic, p. 42. Cf., Descartes, The Philosophical Work\- of Descartes, Vol. I, p. 156. 48. Frege, Posthwnous Writings, p. 261. 49. See: Descartes, Thi! Philosophical Worh of Descartes, Vol. L p. 161, Vol. fl,
p.253. 50. Frege, PosthltU1lOus Writings, pp. 269-270. 51. Frege, Begriffschrift, p. 11. See also: Gottlob Frege, Translations from the Philosophical Writings of Gottlob Frege, trans. P. Geach and M. Black (Oxford: Basil Blackwell, 1966), p. 64; Frege, The Basic Laws of Arithmetic, p. 31; Frege, Posthumous Writings, p. 139; and Frege, Logical Investigations, p. 1. 52. Frege, Begriffschrift, p. 7. 53. See: Ibid., p. 6. 54. R.ene Descartes, Philosophical Letters, 00. and trans. Anthony Kenny (Oxford: Clarendon, 1970), p. 6. See also: Descartes, The Philosophical Works of Descartes, '/oL I, pp. 155, 252. 55. Frege, Posthumous Writings, p. 269. 56. Frege, Logical Investigations, p. 13, n. 4. See also: Frege, Posthumous Writings, p. 270. 51. Frege, The FOIUUlations of Arithmetic, p. vii. 58. Kenny, Anthony, DescQTtes: A Study of his Philosophy (New York: Random House, 1968), p. 125.
158
PHILIP DWYER
59. Descartes, The Philosophical Worb of Descartes, Vol. I, p. 174. 60. See: Ibid. 61. See, for example: Frege, Translations from the Philosophical Writings of Gottlob Frege, p. 64. 62. See: Frege, Posthumous Writings, p. 144. 63. See: Frege, The Foundations of Arithmetic, p. 37; and Frege, Translations from the Philosophical Writings of Gottlob Frege, pp. 60, 79. 64. See: Frege, Logical Investigations, p. 5. 65. See: Descartes, The Philosophical Works of Descartes, Vol. IT, pp. 67, 70, 217. 66. See: Ibid., Vol. I, pp. 174-175. 67. See:· Frege, Logical Investigations, p. 12. Cf., Locke's account of why one must give one's words 'a secret reference', in: Jolm Locke, An Essay Concerning Human Understanding, 00. P.H. Nidditch (New York: Oxford University, 1975), Book 2, ch. 41. 68. Curley, EM., Descartes Against the Skeptics (Cambridge, Harvard University, 1978), p. 149. See also: Kenny, Descartes, ch. 6. 69. See: Frege, The Foundations of Arithmetic, pp. 35, 68-73. See also: Frege, The Basic Laws of Arithmetic, p. 16; and Frege, Logical Investigations, p. 17. 10. Dummett, Truth and Otlu!r Enigmas, p. 88. 71. Ibid., p. 94. 72. Frege, The Foundations of Arithmetic, p. x. 73. Ibid. 74. See: Ludwig Wittgenstein, The Blue and Brown Books (Basil Blackwell, 1969), pp. 3-4. 75. See: Frege, TIu! Basic Laws of Arithmetic, p. 12. 76. Ibid., pp. 12-13. 77. See: Ibid., p. 13. See also: Frege, Logical Investigations, p. 2. 78. See: Frege, The Basic Laws of Arithmetic, p. 14. 79. See: Frege, Begriffschrift, p. 15; Frege, The Foundations of Arithmetic, pp. ix, 3; Frege, Postluunous Writings, p. 3; and Frege, Logical Investigations, p. 2. 80. Frege, The Foundations of Arithmetic, p. 37; see also: pp. 105, 115. 81. Frege, The Basic Laws of Arithmetic, p. 15. 82. See: Frege, The Foundations of Arithmetic, pp. v-vii. 83. Ibid., p. vii. 84. See: Dummett, Truth and Other Enigmas, p. 182. 85. See: Ibid., pp. xl, 211, 225, 248. 86. John McDowell, "Engaging with the Essential," TIMES LITERARY SUPPLEMENT, 16 Jan. 1981, p. 62. Cf., D. Bell, "Review of K. T. Fann, Ludwig Wittgenstein," IDSTORY AND PIDLOSOPHY OF LOGIC, 1, (1980), p. 236. 87. See: Ludwig Wittgenstein, Zettel, ed. G.E.M. Anscombe and G.H. von Wright, trans. G.E.M. Anscombe (Oxford: Basil Blackwell, 1967), pp. 92-130. 88. Frege, Logical Investigations, pp. 25-26. 89. See: Ibid., p. 7. 90. See: Frege, Translations from the Philosophical Writings of Gottlob Frege, pp. 201-202, 216. 91. Frege, Logical Investigations, p. 26. 92. Ludwig Wittgenstein, Philosophical Grammar, ed. Rush Rhees, trans. Anthony Kermy (Oxford: Basil Blackwell, 1974), p. 107. 93. Crispin Wright, Wittenstein on the Foundations of Mathematics (Cambridge: Harvard University, 1980), p. 252.
ANTI-PSYCHOLOGISM AND SCEPTICISM
159
94. Ludwig Wittgenste~ Remarks on the Folmdations of Mathematics, 3rd ed., cd. G.H. von Wright, R. Rhees, G.E.M. Anscombe, trans. G.EM. Anscombe (Oxford: Basil Blackwell, 1978) Part I, 8.
95. Ibid., Part VH, 33. 96. Ibid., Part ill, 87. 97. Ludwig Wittgenstein, On Certainty, ed. G.E.M. Anscombe and G.H. von Wright, trans. Denis Paul and G.EM. Anscombe (New York: Harper and Row, 1969), 56. Cf., 375; and Wittgenstein, Remarks on the Folmdations of Mathematics, Part VH, 20, 30, 35. 98. See: Ludwig Wittgenstein, Philosophical Investigations, 3rd ed., trans. G.E.M. Anscombe (Oxford: Basil Blackwell, 1968), 7. 99. See: Saul Kripke, Wittgenstein on Rules and Private Language (Cambridge: Harvard University, 1982). 100. Contrast: Frege, Logical Investigations, p. 1; Frege, Translations from the Philosophical Writings of Gottlob Frege, p. 206. 101. Cf., Wittgenste~ ReTNJTks on the Foundations of Mathematics, Part vn, 67. 102. Cf., Wittgenstein, Zettel, 309. 103. Frege, The Basic Laws of Arithmetic, p. 14. 104. See: Ibid. 105. Cf., Wittgenstein, Philosophical Investigations, pp. 241-242. 106. Cf., Wittgenstein, Remarks on the Foundations of Mathematics, Part I, 9. 107. Ibid., Part I, 168. 108. Frege, The Folmdations of Arithmetic, p. 108. 109. See: Frege, Begriffschrift, p. 5; Frege, The Folmdations of Arithmetic, p. 4; Frege, The Basic Laws of Arithm£tic, pp. 13-14. 110. See: Wittgenste~ Philosophical Investigations, 97. Cf., 107-108. 111. Ibid., lOS. 112. Ibid., 81. 113. Frege, Logical Investigations, p. 25. 114. Wittgenste~ Philosophical Investigations, 108. 115. See, inter alia: Ibid., 138-242; and Wittgenstein, Remorks on the Foundations of Mathematics, Parts I, VI, VIll. 116. Cf., Frege, The Foundations of Arithmetic, p. 93. 117. See: Wittgenstein, Philosophical Investigations, 48. 118. See: Ibid., 28. 119. Such documentation and detailed discussion is in my D.Phil. thesis, Sense and Subjectivity: A Study of Wittgenstein and Merleau-Ponty, Oxford University, 1983. 120. See, especially: Wittgenste~ Philosophical Investigations, 73, 74, p. 54 n., 185, 228; Wittgenstein, Remarks on the Folmdations of Mathematics, Part VI, 43, 44, Part VH, 47, 60; Wittgenstein, Zenel, 276-7; and Ludwig Wittgenste~ Remarks on the Philosophy of Psychology, Vol. I, ed. Anscombe and G.H. von Wright, trans. G.E.M. Anscombe, Vol. n, ed. G.H. von Wright and Heikki Nyman, trans. C.G. Luckhardt and M.A.E. Aue (Oxford: Basil Blackwell, 1980), Vol. I 344, 505, 874, 875, 882, Vol. 387-424. 121. See: Wittgenstein, Remarlcs on the Folmdations of Mathematics, Part ill, 43.
n
122. Ibid., Part
rn,
2.
123. Cf., Frege on defmitions in: Frege, The Basic Laws of Arithmetic, p. 2. Cf., also, Wittgenstein, Philosophical Investigations, 377" 124. See: Wittgenstein, Remarlcs on the Foundations of Mathematics, Part ill, 25. 125. Ibid., Part I, 135, 136.
160
PHILIP DWYER
126. Ibid., Part L 137. 127. Ibid., Part ill, 78. 128. Ibid., Part m, 88. 129. Ibid., Part IlL 83. 130. Ibid., Part m, 84. 131. See: Frege, The Foundations of Arithmetic, pp. ix, 106, 108. 132. Wittgenstein, Philosophical Investigations, 125. 133. See: Wittgenstein, Re1NJTu on the Foundations of Mathematics, Part VH, 370. 134. Cf., Wittgenstein, Remarks on the Foundations of Mathematics, VH, 15; and Wittgenstem, Remarks on the Philosophy of Psychology, Vol. H, 401. 135. Wittgenstein, Remarks on the Foundations of Mathematics, Part vn, 15. 136. See: Ibid., Part VH, 11. 137. Ibid., Part VII, 16. Cf., Part ill, 88, Part VI, 6; and Wittgenstein, Philosophical Investigations, 84. This is related to the point mentioned above in note 42. A similarly related point in COlUlection with Frege is raised by Wittgenstein at the beginning of Philosophical Grammar: "In attacking the fonnalist conception of arithmetic, Frege says more or less this: these petty explanations of the signs are idle once we u.nderstand the signs. Understanding would be something like seeing a picture from which all rules followed, or a picture that makes them all clear. But Frege does not seem to see that such a picture would itself be another sign, or a calculus to explain the written one to us." Wittgenstein, Philosophical Grammar, p. 40. 138. Wittgenstein, Remarks on the Foundations of Mathematics, Part VIT, 13. 139. See: Wittgenstein, Remarks on the Philosophy of Psychology, Vol. II, 289. 140. Frege, The Foundations of Arithmetic, p. ix. 141. Frege, Begriffschrift, p. 1. 142. See: Ibid. 143. Frege, The Foundations of Arithmetic, p. 2. 144. See: Ibid., p. 4. 145. See: G.E. Moore, "A Proof of the External World," in G.E. Moore, Philosophical Papers (London: Allen and Unwin, 1959). 146. See: Descartes, The Philosophical Works of Descartes, Vol. IT, pp. 266, 279. 147. Ibid., Vol. fi, p. 266. 148. See: Ibid., Vol. L p. 148. 149. See: Ibid., Vol. I, pp. 188-189. 150. See: Ibid., Vol. I, p. 148. 151. Wittgenstein, On Certainty, 19. 152. On Wittgenstein's view this notion of 'transcendent certainty' is also the result of a certain philosophy of mind: "Forget this transcendent certainty, which is connected with your concept of spirit." Wittgenstein, On Certainty, 47. For a striking confinnation of Wittgenstein's diagnosis, see: Descartes, Principles of Philosophy, Part I, principle XII, in The Philosophical Works of Descartes, Vol. I, p.
223. 153. Wittgenstein, Philosophical Investigations, 89.
154. Ibid., 87. 155. Ibid. 156. Ibid., 88. 157. Wittgenstein, On Certainty, 392. 158. Descartes, The Philosophical Works of Descartes, Vol. 159. See: Ibid., Vol. 1, p. 237.
n,
p. 41.
ANTI-PSYCHOLOGISM AND SCEPTICISM
161
160. This is not, of course. to concede that in dreams we are, as such, mistaken or deceived about anything. One may be mistaken about being chased by a monster when awake, but in a dream there is no mistake about it- unless one goes on to dream that it was a mistake. The distinction between being awake and dreaming is absolutely fundamental, md in any everyday account of error the relevant context is always explicit. Also. dreams are not 'propositional'; they are neither true nor false. When I 'dream that...', I don't dream about something true or false. Cf., On
Certainty, 676. 161. See: Descartes, The Philosophical Worb of Descartes, Vol. I, p. 145. 162. See: Ibid.• Vol. L p. 148. 163. See: Ibid.• Vol. L pp. 145. 149. 218. 220-221; Vol. IL pp. 205-206, 266. 164. Ibid., Vol. L p. 315. 165. Bernard Williams, Descartes: The Project of PlUe Enquiry (Harmondsworth: Penguin, 1978), pp. 34-35. 166. Ibid., p. 35. 167. Ibid., p. 36.
168. Ibid. 169. Ibid., p. 49. 110. Ibid., p. 61. 171. Curley, Descartes Against the Skeptics, p. 112; see also: pp. 85, 119, 106. 172. Ibid., p. 105. 173. See, for example: Wittgenstein, On Certainty, 19, 220, 254, 323, 327. 114. See: Curley, Descartes Against the Sh!ptics, p. 89; see also: p. 116. 115. Wittgenstein, On Certainty, 226. 176. Cf., Descartes, The Philosophical Works of Descartes, Vol. IL p. 206. 177. See: Williams, Descartes: TN! Project of PlUe Enquiry, pp. 47ff. 178. I would have thought that this is what lies behind 'Moore's paradox' (statements of the form '1 believe that P and it isn't so; cf., Philosophical Investigalions, p. 191), as well.as the stories of Moore, mouth agape and eyes agog, responding to a sceptical or otherwise outrageous assertion with: 'Do you really believe that? I, For again, a person can't have it both ways: if he doesn't really believe it, then there is no further issue; the context of 'doing philosophy' doesn't somehow preserve the situation as one of serious assertion; if he does really believe it, then Moore's response is the only serious one- the context of 'doing philosophy' does not exempt anyone from the normal responses to some given assertion. I say I would have thought this lies behind this Mooreana, but for his essay 'Four Fonns of Scepticism' where he prepares and then wolfs down whole a two-kinds thesis of doubt. See especially p. 199 in Philosophical Papers. Here Moore makes assertions with no argument, which assertions moreover, in my view, are just contradictions- e.g.: ''but from the fact that he sincerely believes that it is doubtful whether he is [sitting], it certainly does not follow that he doubts whether he is." Here at any rate, Moore is no respecter of ordinary language.
FRENCH EPISTEMOLOGY AND THE FREGEAN DILEMMA M.A. NOTIURNO
I. Introduction Psychologism is the doctrine that empirical1 psychology provides the foundations for each of the other special sciences, Le., that each of the special sciences can ultimately be derived from and justified by the results of empirical psychology. In the late nineteenth century, Gottlob Frege criticized psychologism in logic and mathematics, arguing that it results in subjectivism, idealism, relativism, scepticism, and solipsism- each of which displays a cavalier disrespect for truth and the objectivity of scientific knowledge. 2 Ever since Frege's critique, many philosophers have felt compelled to choose between psychologism and subjective knowledge on the one hand, and anti-psychologism and objective knowledge on the other. 3 Ordinarily, such a choice would not be too difficult to make. For when it comes to knowledge, philosophers almost universally prefer the objective kind.4 But when it comes to Frege, preferring the objective kind means believing in the possibility of a priori valid knowledge and in the existence of a third realm of eternal and immutable truths that serve as the objects of such knowledge.' And this, to put the matter bluntly, has become increasingly difficult to swallow. In my view, the psychologism/anti-psychologism debate can be traced to residual problems in the justified true belief theory of knowledge and revolves, ultimately, around the question 'What is the cognitive source that guarantees the truth of foundational beliefs?' Faced with this question, psychologistic philosophers traditionally answered 'a posteriori experience', and anti-psychologistic philosophers 'a priori reason'. But here, it is important to recognize that Frege's critique of psychologism presupposed a framework of Kantian epistemological categories that no longer seems tenable. In light of the great conceptual revolutions of the twentieth century, Kant's distinction between a priori knowledge that is strictly universal, necessary, and apodeictically certain and a posteriori knowledge that is not - let alone his distinction between analytic and synthetic statements - seems romantically naive. Whether it is due to the decline of Newtonian Mechanics and Euclidean Geometry or to the discovery of Russell's Paradox, to the incompleteness theorems of GOdeI, Church, and Turing or to
FRENCH EPISTEMOLOGY
163
Wittgenstein's attack on the law of non-contradiction- whatever the cause, we simply no longer have the faith in the possibility of a priori valid knowledge that Frege once had. But despite this loss of faith, philosophers who have been influenced by Frege still feel the need to provide some form of objective justification for their beliefs or, at the very least, for their believing them. The result is that philosophers who have been influenced by Frege feel caught in a Fregean Dilemma: they can choose either anti-psychologism and infallibly certain knowledge, or psychologism and no real knowledge at all. The Fregean Dilemma arises from the recognition that neither a priori reason nor a posteriori experience suffices to guarantee objective knowledge. A priori reason ensures the objectivity of justification. But it presupposes an infallibility that we do not seem to have. A posteriori experience does not presuppose infallibility. But it fails to provide for the objectivity of justification. Be this as it may, the force of the Fregean Dilemma depends upon the acceptance of what I will here call 'French Epistemology'- the Cartesian methodological doctrine that a statement should be regarded as false, or at least that it cannot be rationally accepted as true, unless it is either 'clear and distinct' (and hence, indubitably true) or derived as a logical consequence from other statements that are themselves clear and distinct. Now this, stated thus briefly, is both cryptic and in need of explication. But I mention it here at the outset because most of the attempts to resolve the Fregean Dilemma, Le., to salvage objectivity without appealing to a priori valid knowledge, have maintained French Epistemology by weakening their criteria for justification. Inductivist attempts, on the one hand, have reinterpreted the justification of statements as something other than logically sound argument- thereby ignoring Frege's warning that in lieu of a general principle of induction, "induction becomes nothing more than a psychological phenomenon, a procedure which induces men to believe in the truth of a proposition, without affording the slightest justification for so believing.'" Fideist attempts, on the other hand, have generally reduced justification at the foundational level to some form of creedal commitment- typically to a linguistic community, a scientific paradigm, or a form of life. But these attempts have done so, I will argue, at the cost of blurring the intuitive distinctions among being-true, being-believed-to-be-true, and being-justified-as-true. In so doing, they have issued, however unconsciously and contrary to expressed intentions, in subjectivist theories of knowledge. Contrary to these attempts, my own approach is to maintain: (1) that there are sharp distinctions between being-true, being-believed-to-be-true, and being-justified-as-true; (2) that there is an equally sharp distinction between the justification of a statement and the justification for believing that statement; (3) that a statement is justified as true if-and-only-if it can be shown to be the consequence of a logically sound argument; (4) that the justification of a statement is a1 ways a justification from assumptions that
164
M.A. NOTTURNO
are not, themselves, justified; but (5) that one need not justify a statement in order to rationally accept it as true. Taken together, these points constitute a repudiation of French Epistemology and the justified true belief theory of knowledge in favor of an attitude toward cognition that I have, for lack of a better name, called 'Epistemological Buddhism'.' In so doing, they also constitute a repudiation of such time-honored epistemological values as conviction and creedal commitment In what follows, I will try to make these intuitions explicit.
II. The Role of Justification Elsewhere,8 I have identified four specific anti-psychologistic theses that can be found in Frege's philosophical writings: (1) logical anli-psychologism, or the thesis that logic is an a priori normative and not an a posteriori descriptive (or natural) science; (2) linguistic anti-psychologism, or the thesis that meanings, or senses (including thoughts), are third-realm entities and not mental entities or spatio-temporal objects; (3) mathematical anti-psychologism, or the thesis that mathematical objects are third-realm entities· and not mental entities or spatio-temporal objects; and (4) epistemological anti-psychologism, or the thesis that questions of justification are distinct from questions of discovery and that questions of discovery are not pertinent to the evaluation of the truth of a theory. There, I argued that these four theses are transcendental consequences of Frege's assumptions that truth is objective and independent of the knowing subject, i.e., that being-true is distinct from being-believed-to-be-true, and that truth is the object of human knowledge. I do not wish to deny any of this here. But thus said, it is important to emphasize that the psychologism/anti-psychologism debate, like all debates concerning the foundations of knowledge, is fIrst and foremost a debate about justification. Specifically, it is a debate about what counts as a justification, about what sorts of things can justify wha~ and about what constitutes the grounds or ultimate justification of our knowledge. Insofar as this is concerned, Frege's complaint against psychologism was that it encourages us to regard as justifications arguments and 'proofs' that really are not. Psychologism does this, according to Frege, because it purports to justify the laws of logic and mathematics with empirical observations- the latter being, by nature, far too particular, contingent, subjective, and uncertain to underwrite the reputed universality, necessity, objectivity, and certainty of the former. Contrary to psychologism, Frege maintained that the integrity of justification depends upon the existence of a third realm of eternal and immutable general principles "which themselves neither need nor admit of proof" and which can be apprehended a priori without recourse to empirical observation. Here, epistemological and logical anti-psychologism maintain (respectively) that justification should be modelled on mathematical
FRENCH EPISTEMOLOGY
165
proof from a priori principles, Le., on deductively sound arguments, and that logically sound arguments suffice to guamntee the truth of whatever they justify. Linguistic anti-psychologism then adds a corollary concerning the objectivity of meaning- a corollary that many regard as a necessary prerequisite for the validity (and hence soundness) of deductive arguments. to But if the psycbologism/anti-psychologism debate is a debate about justification, it would be 100 naive to regard it as a straightforward disagreement in which the use of 'justification" by psychologistic and anti-psychologistic philosophers is clearly the same. On the contrary, it is precisely what is meant by 'justification' that is at the heart of the issue. But nor would it be right to regard the dispute as merely verbal. Defming what anti-psychologistic and psychologistic philosophers mean by 'justification' will not, in and of itself, resolve the debate. For the issue, of course, is not what philosophers do mean by 'justification', but what they should. In what immediately follows, I will argue that this debate is a consequence of residual problems in the traditional justified true belief theory of knowledge- specifically, that it is the consequence of problems regarding the ambiguity of 'belief' and the foundations of justification. Getting clear about these problems will not, in and of itself, resolve either the Fregean dilemma or the psychologism/anti-psychologism debate. But it will, I hope delineate the pitfalls that any such resolution should avoid.
A. The Ambiguity of 'Belief If we conceive of knowledge as justified bUe belief, 'then the questions 'that should immediately arise (but in fact rarely do) are "What are we trying to justify when we are trying to justify true beliefl" and "What are we claiming is justified when we claim to have justified true beliefl" The obvious answer to these questions is 'a belief' But here, 'belief' is importantly ambiguous. It may refer either to the content of belief (what is believed), or to the process of belief (the believing itself). That is, it may refer to a statement (or object) that someone believes, or to a person's (or subject's) act of believing an object (or statement). I do not wish to suggest that there is anything problematic or improper about either of these senses of 'belief'. Nor do I wish to suggest that either one of these senses should, in and of itself, be preferred over the other. It is, however, essential to note that this 8IIlbiguity of 'belief' gives rise to two quite different 'justified true belief' theories of knowledge- theories that differ with regard to: (a) what sort of thing is justified in justified true belief; (b) what sort of thing can justify it; and (c) what sort of thing counts as a justification of it According to the first (henceforth 'objectivist') theory, what we are trying to justify when we are trying to justify true belief and what we are claiming is justified when we 'claim to have justified true belief is an object that someone believes, i.e., a statement According to the objectivist theory, 0
166
M.A. NOTTURNO
only true statements can be known, and in order for someone to know a true statement, he must both believe it and be justified in his belief. But since it is, according to the objectivist theory, a statement that is being justified, and since justification, according to the objectivist theory, consists in showing that a statement is true, justification, itself, is construed as a logical relationship between that statement and whatever justifies it (henceforth, its 'justifiers'). Here, the objectivist theory maintains that statements can be justifted only by other statements, and that a statement is justified when-and-only-when it is shown to be the logical conclusion of other statements that are, themselves, already known to be true. For these reasons, justification, according to the objectivist theory, takes the fonn of a logically sound argument According to the second (henceforth 'subjectivist') theory, what we are trying to justify when we try to justify true belief and what we are claiming is justified when we claim to have justified true belief is not a statement (or object) that someone believes, but someone's act of believing a statement The subjectivist theory, like the objectivist theory, maintains that only true statements can be known, and that in order for someone to know a true statement, he must both believe it and be justified in his belief. But since it is, according to this theory, not a statement, but someone's act of believing a statement that is being justified, justification is, itself, not a logical relationship between statements and other statements, i.e., not a relationship dealing with logoi or words, but a psychological relationship between a subject's act of believing a statement and whatever might serve as the justifiers of that act, i.e., a relationship in which at least one of the terms is a mind, psyche, or psychological experience. Here, such justifiers need not be construed to be statements (logical things). Psychological experiences, e.g., sense perceptions, sensations, emotions, and acts of believing, may serve as well. Moreover, if psychological experiences are the things to be justified, and if psychological experiences can serve as their justifiers, then justification cannot take the form of a logical argumentnot, at least, unless psychological experiences can serve as the premisses and conclusions of logical arguments. Now 'psychologism' and 'anti-psychologism' are evaluative names for, respectively, the subjectivist and objectivist justified true belief theories of knowledge. They are, in other words, evaluative names for epistemologies whose primary concerns are, respectively, the justification of the subjects' knowing and the justification of the objects known. Insofar as this is concerned, 'psychologism' is used pejoratively - to denegrate subjectivist theories of knowledge - and 'anti-psychologism' is used amelioratively- to reaffirm the epistemological value of objectivity. Here, anti-psychologistic philosophers insist that objective knowledge is the knowledge of truth and that the sort of justification pertinent to objective knowledge is the justification of statements- for these are the sorts of things that are true or false. Any concern with the justification of the subjects who believe these statements will, accordingly, be denigrated as psychologism. Such a concern,
FRENCH EPISTEMOLOGY
167
according to anti-psychologistic philosophers, is either beneath the dignity of philosophy - to the extent to which it subordinates Truth to Belief and Reality to Appearance - or downright pernicious- to the extent to which it confuses being-ttue with being-believed.. to-be-true or maintains that the justification of believers is all that really matters. Now it is, I think, undoubtedly the case that epistemology's censure of psychologism has masked the need for subjective justification. For regardless of whether or not our statements can, in the [mal analysis, be objectively justified, we do want our doctors to be properly trained, our lawyers to be conversant with the law, our engineers to be accurate in their calculations, and, what is more, we want them to be held responsible for their actions when they are not or when they act contrary to accepted procedures. II It is, moreover, possible that the same sort of thing that justifies an object that is believed can, in one way or another, justify a subject who believes it, e.g., that an argument that justifies a statement also justifies the person who produces or becomes aware of it. But it is also possible that a person may be justified in believing a statement that is not justified (or justifIable) at all. And it is, in part, for this reason -that we insist that the subject who believes and the object that he believes are entirely different things. Mistaking the one for the other can only engender confusion. In particular, it engenders the confusion of mistaking the justification of the one for a justification of the other. A justification of a statement that some person believes is not necessarily a justification of that person's believing it. Conversely, a justification of a person's believing a statement is not necessarily a justification of that statement People may, by virtue of their appeal to the authority of Einstein, be justified in believing that E=MC2• But it is not the authority of Einstein that justifies the statement that E=MC2. (One might rather suppose that it is the 'justification' of E=MC2 that justifies Einstein's authority!) And relatively few who are 'justified' in believing that E=MC2 are able to reproduce the arguments and considerations used to 'justify' it (That is why they appeal to his authority!) On the other hand, a person may be able to produce (or reproduce) an argument that justifies a statement that he believes without ever understanding how or why that argument works. But all of this simply underscores the fact that the subject who believes and that object that he believes are entirely different things. Now the psychologism/anti-psychologism debate can be traced to the ambiguity of 'belier as follows. Modem epistemology (epistemology since Descartes) has been almost exclusively justificationist in nature. But while there has been a strong consensus that knowledge is justified true belief, there has, until very recently, 12 been a general failure to recognize the ambiguity of 'belier or to distinguish the two theories of knowledge that stem from it. The result is that epistemologists have generally conflated the justification of a statement with the justification of someone's act of believing that statement. I ' And they have typically done so in one of two ways. Psychologistic epistemologists have generally taken the justification
168
M.A. NOTTURNO
for a subject's believing a statement to be the justification of that statement itself. And anti-psychologistic epistemologists, for their part, have generally maintained that no one is ever justified in believing a statement unless he can give a logical proof of it. 14 In this way, failure to recognize the ambiguity of 'belieft has proved detrimental to the subjectivist and objectivist theories alike. For it either makes the justification of the object too simple, or it makes the justification of the subject too hard. It is, perhaps, ironic that the same epistemologists who conflate the objectivist and subjectivist theories of knowledge genemlly denigrate psychologism and claim to be dissatisfied with anything short of objective knowledge. This, perhaps, is why Husserl regarded psychologism as being the result of a confusion- something which, "to think it out to the end, is already to have given it up."u Still, someone might suppose that if a true statement could be justified as the conclusion of a logical argument, and if someone who believed that statement knew of that argument and was able to produce it of his own accord, and understand it, then the justification of the statement he believed might also serve as a justification for his act of believing it. Indeed, one might even insist that a subject is justified in believing a statement if-and-only-if that subject can produce and understand a justification of the statement he believes. In that case, there would still be a difference between the subject who believes and the object that he believes. But the ambiguity of 'belief and the distinction between objective and subjective knowledge would have little practical consequence. For in that case, one could not have the one without having the other. It might be thought that epistemology can safeguard objective knowledge by following either of these paths, Le., by maintaining the objectivist justified true belief theory of knowledge or by making the justification of a statement a criterion for justifiably believing it. Indeed, the second path might seem especially attractive, since it joins psychology and epistemology instead of alienating them. Each of these paths, however, leads directly to the second problem with the justified true belief theory of knowledge- a problem that infects both the subjectivist and objectivist versions of the theory and which would, independent of the ambiguity of 'belief, give rise to the psychologism/anti-psychologism debate.
B. The Grounds of Justification suppose that a person has recognized the ambiguity of 'belief and has either: (a) adopted the objectivist justified true belief theory of knowledge; or (b) stipulated that a person's act of believing a statement is justified if-and-only-if he is able to produce and understand a justification of that statement In either case, and regardless of whether we are justifying an object or a subject, that person will have knowledge of a statement if-and-only-if he is able to produce and understand a justification of that
FRENCH EPISTEMOLOGY
169
statement Thus, it might be thought that here are two ways to avoid the threat of psychologism. But this, unfortunately, is not the case. For even if we adopt the objectivist justified true belief theory of knowledge, there is still the possibility that psychological experiences may enter into our 'justifications' of statements. In fact, there is good reason to believe that psychological experiences must enter into our justifICation of statements. Let me explain. Earlier we said that justification, according to the objectivist theory, assumes the form of a logical argument and that a statement is justified if-and-only-if it can be derived as the logical consequence of other statements that are themselves already known to be true. It is important to note that this objectivist theory of justification contains two planks: (a) the justification of a statement assumes the form of a logical argument (statements can be justifIed only by other statements); and (b) the premisses of that argument must themselves be justified as true (the justifiers of a statement must already be known to be true). But it is, in this context, also important to note the rationale for (b). For it is the introduction of (b) into the objectivist theory of justification that results in the second problem with the justified true belief theory of knowledge. Why is it necessary for the justification of a statement that the justifiers of that statement are already known to be true? Well, why does a statement need to be justified in the fast place? Ordinarily, we ask for a justification of a statement only when there is some reason to doubt the truth of that statement and we do not want to accept that statement dogmatically or as a matter of faith. Now reason to doubt the truth of a statement may range from its inconsistency with itself, to its inconsistency with the way things seem, to the mere logical possibility that it is false. But if there is no reason at all to doubt the bUth of a statement or if someone is willing to accept that statement as dogma, then there would be little point in asking for or attempting to give a justification of it. JustifICation is thus required to avoid both doubt and dogma. But if this is the reason why a justification is required, then any proposed justification that appeals to either doubt or dogma must be considered unacceptable. Now suppose that a person P believes a statement S and that P is also aware of an argument A according to which S is the logical consequence of a set of statements T. That is, suppose somebody believes something and is able to produce an argument to 'justify' it. Does A justify S? Here, one might assume that A justifies S only if A is valid, Le., only if S is true in every case in which all the members of T are. For if S could be false despite the truth of all of the members of T, then it would be difficult to see how A could or why A should count as a justification of S. Any justification of S would, presumably, suffice to show that S is true. But if A is invalid, then it does not, by definition, suffICe to show that S is true. (A might, in that case, count as a justification of P's believing S, but we are here assuming the objectivist justified true belief theory of knowledge.) Still, it is elementary that a false statement may be the logical consequence
170
M.A. NOTIURNO
of a valid argument if one or more of the premisses of that argument are also false. 50 even if A is valid, P's demonstration that 5 is the logical consequence of T does not suffice to justify 5. P must know that all of the members of T are true. For if P is doubtful of 5 but also doubtful of one or more of T, then any appeal to T to justify 5 is little more than a charade. 5 may, indeed, follow logically from T, but that is small comfort unless we already have knowledge of T. IT P is willing to accept the dubious T in order to justify the dubious S, then he might as well have accepted the dubious S to begin with. Here, the rationale for (b) is clear. Without (b) justification ultimately reduces to dogmatism. For these reasons, T cannot serve as a justifier of S unless T is already known to be bUe. Hence, the rationale for (b) and, hence, 'the second problem with the justified bUe belief theory of knowledge. The second problem with the justified true belief theory of knowledge is that it leads, without modification, immediately to infinite regress (and hence to no justification at all). This is very easy to see. Suppose again that a person P believes a statement 5 and that P is also aware of an argument A according to which 5 is the logical consequence of a set of statements T. Does A justify 5? We have already seen that A does not justify Sunless T is already known to be true. But the objectivist justified true belief theory of knowledge tells us that in order to count as knowledge, T must itself be an instance of justified true belief, i.e., P must be aware of some argument A * according to which T is the logical consequence of a statement or set of statements T*. Does A * justify T? As we have just seen, T* must itself be known to be true if it is to serve as a justifier of T. Hence, T* must also be an instance of justified true belief and P must be aware of some argument A ** according to which T* is the logical consequence of a statement or set of statements T**. And so on. We have shown that the objectivist requirements (a) that justification assume the form of a logical argument and (b) that the premisses of that argument must themselves be known lead immediately to infmite regress. In order for a statement to be justified, the statements that justify it must be justified. But we have also shown that (b) is essential to the justified true belief theory of knowledge. If T is no better known than S, then the use of T to 'justify' S is a charade. Here, the lesson to be learned concerns the nature, as opposed to the possibility, of justification. The point is not that justification is impossible (in any but the special cases where: (1) justification is demanded of foundational statements; or (2) justification is demanded of all statements). The point is that every justification or proof is a justification or proof from assumptions. That is, the form and purpose of a justification is to show that a certain statement (action, belief, or what-have-you) is in accordance with and a consequence of certain assumptions. While we may indeed ask for a justification of those assumptions, it makes no sense at all to ask for the justification of foundational assumptions- Of, a fortiori, of all assumptions. But this, again,
FRENCH EPISTEMOLOGY
171
is not because justification is impossible. It is because the very nature of justification presupposes the existence of statements that go unjustified. So if an objectivist wants to maintain that knowledge is justified true belief but avoid infinite regress and charade, then he seems to have only one option open. 16 He must deny that statements can be justified only by other statements. In other words, the objectivist must commit himself to at least two of the three planks of the subjectivist theory of justification, or else admit that justification, and hence knowledge, is impossible. In order to avoid both infinite regress and charade, the objectivist must maintain: (1) that the justifiers of statements need not themselves be statements; and (2) that justification need not assume the form of a logical argument. Objectivists have traditionally sought to do this by: (a) distinguishing between immediate and mediate ways of acquiring knowledge, i.e., between intuition and inference; (b) developing theories regarding infallible sources of immediate knowledge; and (c) construing such infallible sources of immediate knowledge as the ultimate justifiers of statements, Le., as the foundations of knowledge. Here, the project has been to identify a set of statements, the foundations, the truth of whose members is guaranteed not by their valid derivation from other statements, but by their intuitive derivation from allegedly infallible sources. Such foundational statements are alleged to be special instances of knowledge. They do not admit of logical justification (that's why they are fountiational), but nor do they need to be logically justified. They do not need to be logically justified because their truth is (alleged to be) guaranteed as a result of their intuition from an infallible source. Thus guaranteed, these foundational statements can serve as the ground for all subsequent logical justification. Be this as it may, this concern with infallible sources of knowledge has typically resulted in psychologism of one form or another. For these cognitive sources have typically been construed to be faculties of the mind like sensibility, understanding, or reason. And thus construed, the ultimate justifiers of statements tum out to be psychological experiences. We have been discussing the role of justification in the psychologism/ anti-psychologism debate and, especially, two of the residual problems with the justified true belief theory of knowledge that have given rise to this debate. The problem posed by the ambiguity of 'belief is serious and the equivocations that have resulted from it are responsible for many of the disputes in the history of epistemology. One can still see its influence today in the prevalent belief that inductive inference and the so-called 'logic of confmnation' can provide justification for objective knowledge. "Even observation statements reporting a million white swans and no non-white swans fail to justify the statement that all swans are white," insists the deductivist "But surely, one would would be an irrational sceptic to believe otherwise," replies the inductivist. And so the debate continues. But as soon as one distinguishes between the justification of the statement and the justification of someone's believing it, it is easy to see that both the deductivist and the inductivist are righl The evidence cited may very well
172
M.A. NOTTURNO
justify someone's believing that all swans are white- despite the fact that it does not justify the statement Be this as it may, the questions regarding the foundations of justification pose even greater problems for the objectivist theory of knowledge. If the grounding of justification in cognitive sources is a necessary prerequisite for the possibility of justification, and if such cognitive somces are ultimately psychological in nature, then the objectivist justified true belief theory of knowledge seems an impossible dream. For in that case, the objectivist justified true belief theory of knowledge leads either to psychologistic justification (justification by psychological experience), or to no justification at all. If epistemology insists upon a role for justification, then the ensuing psychologistic justification results in at least one of three different forms of subjectivism: (1) the subjectivism of confusing the justification of the 'knowing' subject with the justification of what the subject 'knows'; (2) the subjectivism of regarding what the subject 'knows' as being justified by the subject's private experiences; or (3) the subjectivism of equating evidence that is good enough to convince some subject with evidence that suffices to justify a statement. And if epistemology rejects a role for justification, then it is left to explain what objective knowledge is, how it is possible, and how it differs from mere opinion, faith, and other forms of subjective belief. Still, one might suppose that if infallible sources of knowledge truly do exist, then a psychologism of the sort that we have been discussing would be innocuous. For if our reason or sensibility is constructed in such a way as to intuit infallible knowledge, then whatever knowledge it yields might as well be objective- the epistemological value of objectivity resulting not from the sanctity of objects, but from the bias and fallibility of subjects. Now I do not want to deny that the existence of infallible cognitive sources would render the sort of psychologism that we have been discussing innocuous. But the fact remains that the great conceptual revolutions of the twentieth century have placed all claims to infallible cognitive sources in doubt. Of course, the mere fact that we might doubt the claim that a cognitive source is infallible does not, in and of itself, refute the infallibility of that source. If the Pope speaks infallibly when he speaks ex cathedra, then he speaks infallibly regardless of whether or not Luther agrees. But when doubt regarding the infallibility of cognitive sources arises in the context of French Epistemology, it prevents us from rationally accepting a statement on the basis of intuition and, in so doing, from accepting it on the basis of inference as well. And this, as I have said, is what gives rise to the Fregean Dilemma.
FRENCH EPISTEMOLOGY
173
m. French Epistemology Some people remember Descartes as a sceptic- a little bit of irony since Descartes was, according to any literal interpretation of his work, one of the greatest opponents of scepticism in the history of Western Philosophy. It is true that Descartes raised the questions 'What can we know?' and 'How can we know it?' in his Meditations on First Philosophy. And it is true that he there resolved to doubt whatever he could doubt without contradiction in order to answer them. But people who remember Descartes as a sceptic apparently remerrlber only his "First Meditation." It is important to remind them that Descartes' appeal to doubt is purely methodological, Le., that it is not the conclusion to a philosophical investigation, but a method adopted to conduct one. For while Descartes ends his "First Meditation" wondering whether there is anything that is not subject to doubt, he ends his sixth in wonder at how very little there is that is. And along the way he 'proves', beyond the shadow of a doubt, that he exists, that God exists too, that God is not an evil deceiver, that we can never make a mistake so long as we use our God-given Intellect properly, that material objects exist in essentially the way in which we perceive them, and a host of other propositions that might be regarded as even more controversial. The epistemological engine for all this 'proof is Descartes' general principle, announced at the beginning of the third meditation, that everything he clearly and distinctly perceives is true. For this reason, people who do not remember Descartes as a sceptic often remember him as being not sceptical enough. It is, however, difficult to determine whether and to what extent Descartes intended his proofs of the existence of God to be taken literally. For in discussing his own motives for writing, Descartes tells us 'that "it is well to omit things that perhaps would yield a profit to those who are living, when it is one's purpose to do other things that yields even more profit to our posterity."17 And the presence of several 'ironies' in the Meditations suggest that they should not be taken at face value. Thus, Descartes begins his "Letter of Dedication" by saying that the existence of God should be demonstrated by the aid of philosophy, as opposed to theology, since "no unbeliever seems capable of being persuaded of any religion or even any moral virtue, unless these two are f1J'St proven to him by natural reason."18 But he ends his "Letter of :Dedication" by requesting the patronage of the Faculty of Sacred Theology of Paris, since he fears that many people will not be able to follow his arguments and is confident that those who do not will -.Submit to their authority. In that same letter, Descartes explicitly criticizes the argument "that God's existence is to be believed in because it is taught in the Holy Scriptures, and...that the Holy Scriptures are to be believed because they have God as their source"20 as circular, and warns that such an argument "cannot be proposed to unbelievers because they would judge it to be a circle."21 But in his "Third Meditation," Descartes produces a 'prooft whose circularity is both notorious and but slightly less
I'
174
M.A. NOTTURNO
obvious.22 According to Descartes' 'proof', we can know that God exists and is not a deceiver, because anything we clearly and distinctly perceive is true. And we can know that anything that we clearly and distinctly perceive is bUe, because God exists and is not a deceiver.23 It is difficult to detennine whether Descartes was aware of these ironies, much less whether he intended them to be significant. But we do know that Descartes lived in an era in which a man could be executed for expressing philosophical views contrary to those of the Roman Catholic Church,24 that he was aware of the persecution of Galileo,2S and that he suppressed publication of his own La Monde for fear of suffering the same fate. 26 In light of this knowledge, Descartes' statement that it is well for a writer to omit things when doing so yields greater profit to posterity acquires new significance. And given this significance, Descartes' reminder that the Lateran Council under Leo X had condemned those philosophers who maintain that the immortality of the soul can be held on faith alone and had "explicitly enjoined Christian philosophers to refute their arguments and to use all their abilities to make the truth known"27 has a distinctly ominous ring. One cannot help but wonder how the Latemn Council would have reacted were such 'refutations' not forthcoming. For these reasons alone, we should not ignore the possibility that Descartes intended his Meditations on First Philosophy to be taken not literally, but as an attempt at indirect communication. According to this hypothesis, the intelligent reader would quickly recognize that Descartes' arguments do not suffice to prove the existence of God. But the intelligent reader would also recognize the 'ironies' mentioned above and would conclude that Descartes was too intelligent not to have intended them. Noting that Descartes had introduced his 'proofs' for the existence of God as being "such that I believe that there is no other way by which human ingenuity can find better ones, "28 the intelligent reader would then dmw the obvious conclusion. Here, the obvious conclusion would be that the Meditations is an ironical work whose true message is that the existence of God cannot be proved on rational grounds, and must be accepted on faith if it is to be accepted at all. In that case, one could no longer appeal to God as a divine foundation for science, and since no other infallible foundation is forthcoming, it would always be possible that our scientific knowledge, including our knowledge of logic and mathematics,29 is but a grand deception and another bit of irony. Now I mention all this not because I intend to defend an ironical interpretation of Descartes, but because I do not wish to lay at Descartes' grave charges that any 'intelligent' reader would attribute to misinterpretation. Again, it is very possible that Descartes intended his Meditations to be interpreted ironically, that he believed that it is impossible to prove the existence of God by rational argument, and that he did not believe that scientifIC knowledge could lay claim to the sort of absolute certainty that a divine foundation would provide. Such an ironical interpretation need not conclude that Descartes is a sceptic. For the point of Cartesian irony might be that divine foundations and absolute certainty are
FRENCH EPISTEMOLOGY
175
far 100 much to demand of knowledge in 'the fIrSt place. But an ironical interpretation should conclude that Descartes intended his general principle to be recognized as wishful thinking, and that the subjective certainty that we have in an idea, no matter how 'clear and distinct', is no sure criterion of its truth. Be this as it may, I will, in what follows, assume that Descartes intended his Meditations to be interpreted literally. For the fact remains that most commentators have interpreted the Meditations literally, and, interpreted literally, the Meditations has exerted an astounding influence upon the history of Western Philosophy. One may debate whether Descartes discovered epistemology or whether he discovered the mind. But it is uncontroversial that Descartes discovered Modem Philosophy. Descartes discovered Modem Philosophy by insisting upon the priority of epistemology and by introducing two methodological principles that have influenced the development of epistemology ever since. The fast of these principles is well-known as Descartes' 'methodological scepticism'. It is his resolve to call into question or doubt the truth of any belief accepted uncritically on the basis of authority, tradition, or consensus. The second of these principles, often conflated with the fast, is what I call 'French Epistemology'. It is Descartes' resolve to regard as false any belief that is neither clear and distinct (and hence indubitably true) nor justified as a logical consequence of other beliefs that are themselves clear and distinct 30 Taken together, these two principles constitute the Cartesian programmea programme which, in its attempt to derive objective certainty from methodological doubt, has dominated western philosophy for the past three hundred years. That the Cartesian Programme has dominated western philosophy for the past three hundred years can be seen in the fact that rational knowledge has, during that period, been construed almost exclusively in terms of justified true belief. Ever since Descartes, philosophers who have talked about rational knowledge have talked about the justification of true belief. In so doing, they have disagreed notoriously regarding the foundations and criteria of justification- whether rational knowledge rests upon a priori or a posteriori intuition, and whether rational knowledge requires deductive or inductive inference. They have, moreover, disagreed regarding what cOWlts as a priori and a posteriori intuition and regarding what cOWlts as deductive and inductive inference. But few philosophers in the modem period have denied that rational knowledge is justified bUe belief or that a statement must be justified before it can be rationally accepted as true. And this is significant. For the past three hundred years, cognitivists who have asserted the possibility of rational knowledge and sceptics who have denied the possibility of rational knowledge have asserted and denied not merely that such-and-such a belief is true, but that such-and-such a belief can be justified. But here, it is important to distinguish sharply between methodological scepticism and French Epistemology. Methodological scepticism asserts that we should always question or call into doubt beliefs that are not indubitably
176
M.A. NOTTURNO
true, that we should always ask for a rational justification of beliefs that do not seem indubitably true, and that we should always remember that no matter how familiar or generally accepted, a belief that is neither indubitably true nor rationally justified may be false. French Epistemology, on the other hand, asserts that a statement should be regarded as false, or at least as rationally unacceptable, unless it is indubitably true or justified by other statements that are indubitably true. Methodological scepticism maintains that one should question or doubt any statement that is not justified. French epistemology, on the other hand, maintains that one should not accept as true any statement that is not justified. It is essential to realize that these are entirely different injunctions- that it is possible to call into question or doubt the truth of statements that one simultaneously accepts as bue, and hence that it is possible to simultaneously adopt methodological scepticism and repudiate French Epistemology. I call the principle that a statement should be regarded as false until proven true 'French Epistemology' partly because Descartes introduced it and Descartes was French, but primarily because the principle has a potential analogue in a French judicial principle according to which "issues of reasonable doubt in criminal cases may" contrary to Anglo-American custom "be resolved by evidence and argument bearing on prior convictions of the accused, his general behavior, even his family history. "31 But the directive that an accused must be found guilty beyond a reasonable doubt has its origin in the presumption of innocence, the principle that an accused is to be presumed innocent until proven guilty (beyond a reasonable doubt), and the two are often taken to be practically synonymous, if not legally equivalenL 32 The upshot is that an accused's prior criminal record, general behavior, or family history might, if sufficiently notorious, practically reverse the presumption of innocence. In such a case, the accused would still be presumed innocent until proven guilty, and he would still have to be proven guilty beyond a reasonable doubt But the evidence regarding his prior criminal record, etc., would be considered a sufficient rebuttal of the evidence provided by the presumption of innocence, and might, thereby, render a mere accusation a 'proof beyond reasonable doubt'. In such a case, the prosecution would not need to prove that the defendant is guilty; the defendant would need to prove that he is not. Now it seems as if something like this is at play in Descartes' resolution to doubt the veracity of his sense perceptions. Explicitly, Descartes doubts the veracity of his senses because he "noticed that they sometimes deceived me. And it is a mark of prudence never to trust wholly in those things which have once deceived us. "33 But in resolving to regard all of his former beliefs as false until proven true, Descartes in effect places the burden of proof in any epistemological proceeding upon the statement that is claimed to be true. Here, a doubter need not prove that a statement is false. Rather, those claiming to know that statement must prove that it is noL Once asserted, a cognitive claim must be proved true beyond a reasonable doubt, or be judged false.
FRENCH EPISTEMOLOGY
177
Here, what turns on the bmden of proof is not whether we should call into question or doubt statements that are accepted on the basis of
consensus, tradition, or authority. What turns on the burden of proof is whethez we should regard our inability to justify a statement as indicative of that statement's rational unacceptability. H we adopt French Epistemology, then an unproved statement should not be accepted as true, no matter how obvious or apparent its uuth may seem. In that case, any argument to the effect that a statement is not or cannot be justified is, in effect., an argument that that statement should be regarded as false-- or at least that it should not be accepted as true. And it is for this reason that the mere logical possibility of an omnipotent, evil deceiver caused Descartes so much worry. In the context of the "First Meditation," the mere possibility of an omnipotent, evil deceiver emerges as a reason to doubt even the most obvious and certain of our beliefs- those beliefs that could not be shaken by the possibilities that our senses may be deceiving us or that we may be dreaming. If it is possible that an omnipotent, evil deceiver exists, then it is possible that we are deceived even about the existence of extended things, size, and place,'" and "every time [we] add two and three or count the sides of a square.":u But here, it is important to remember that Descartes never believed that an omnipotent, evil deceiver actually exists,36 or even that the existence of such a deceiver is likely. Explicitly, Descartes regarded his "long-standing opinions" concerning the existence of extended things and the nature of mathematical relations as "highly probable, so that it is much more consonant with reason to believe them than to deny them."37 But what Descartes did believe (at least in the context of the "First Meditation tt ) is that the existence of an omnipotent, evil deceiver is a logical possibility. And here, it is not methodological scepticism but the burden of proof as defined by French Epistemology that led Descartes to regard this logical possibility as sufficient reason to "suppose not a supremely good God, the source of truth, but rather an evil genius, as clever and deceitful as he is powerful, who has directed his entire effort to misleading me. "38 If we adopt French Epistemology, then any argument to the effect that statements cannot be justified from indubitably true premisses is an argument that statements (including this one) cannnot be mtionally accepted as true. But if we adopt an epistemology that protects the presumption of innocence - an epistemology that simultaneously insists that we regard a statement or theory as ttue until proven false and refuses to accept 'evidence' regarding the history of that theory or its sow-ce as proof - then a statement may be rationally accepted as true without justification, provided, of course, that is does not seem false. In that case, we might even admit that statements cannot be justified and, admitting that statements cannot be justified, resolve to perpetually entertain the question whether or not they are true. And here, it is easy to see that an allegiance to methodological scepticism is entirely compatible with the repudiation of French Epistemology- that we may question whether or not a statement is true, but assume that it is unless or until we have reason to think otherwise.
178
M.A. NOTTURNO
French Epistemology can, in the context of our story, be seen as an attempt to unite the objectivist and subjectivist justified true belief theories of knowledge via a seemingly innocuous psychologism that would make the justification of a statement a criterion for its justifiable acceptance. Insofar as this is concerned, adherence to French Epistemology made methodological sense so long as philosophers believed in something like an objective criterion of truth. So long as philosophers believed that the truth of a statement could be guaranteed by its being clear and distinct, or a priori, or a posteriori, or even uttered ex cathedra by the Pope, the rational acceptability of a belief could reasonably be predicated upon our ability to justify that belief- either logically, via argument, or extra-logically, by tracing it logically to an infallible source. But once the infallibility of cognitive sources is itself called into question, the mtionale for French Epistemology seems to lose its methodological appeal. Instead of functioning to insure that statements are rationally accepted only if they are troe, French Epistemology now functions to preclude the rational acceptance of statements- regardless of whether or not they are true. Adherence to French Epistemology in lieu of an objective criterion of troth does not make the mtional acceptance of a statement difficult. Adherence to French Epistemology in lieu of an objective criterion of truth makes the rational acceptance of a statement impossible. It makes the rational acceptance of a statement impossible because any logical justification or proof of a statement is, as we have seen, predicated upon some prior rational acceptance of the statement(s) that justify or prove it. The French Cognitivist may, of course, claim that the prior acceptance of such justifying statement(s) is itself rational because such justifying statement(s) are themselves logically justified by other statements. But the French Sceptic will undoubtedly counter by questioning the rational acceptance of those statements. And, as everyone since Aristotle39 has known, this sort of regress cannot continue infinitely. The French Cognitivist may stop the regress at any point, accepting statements without logical justification, and claiming that these are statements whose 'justification' consists in their derivation from a privileged cognitive source. But if he does, then the French Sceptic will undoubtedly question whether and why that sowce should be regarded as privileged and whe'ther and why the acceptance of statements derived from that source - rneditately or immediately - should be regarded as rational. Descartes, if we interpret him literally, did regard the Intellect as privileged, and the clarity and distincbless of his ideas as a sure criterion of their truth. But the clarity and distinctness of ideas has not fared well as a criterion of troth, partly because too many people claimed clarity and distincbless for too many contradictory ideas, and partly because it proved too difficult to detennine whe'ther and when any given idea is clear and distinct Nor have subsequent attempts to ground our knowledge on an objective criterion of truth fared any better. Thus, David Home, who attempted to found our knowledge on sense impressions, and who regarded
FRENCH EPISTEMOLOGY
179
any chain of reasoning that did not ground itself in such impressions as "entirely without foundation,"40 was forced to admit that a science based solely on sense impressions carmot determine whether anything other than those impressions exist and carmot offer our scientifIC laws any foundation other than custom and habit. And Immanuel Kant, who attempted to found scientific certainty on the universal and necessary categories of the Understanding, and who offered a transcendental deduction of those categories from the very possibility of experience, failed to anticipate that .the very sciences that he claimed to be apodeictically certain would later be regarded as false. Now Frege believed that the general principles, or axioms, that he regarded as necessary for justification but which "neither need nor admit of proof'41 could be apprehended, a priori, by the Reason, or what he called "the logical source of knowledge. ,t42 Frege, moreover, regarded such a priori intuition as infallible- writing that "an a priori error is thus as complete a nonsense as, say, a blue concept,"43 and that "we cannot accept a thought as an axiom if we are in doubt of its truth; for it is either false and hence not an axiom, or it is true but stands in need of proof and hence is not an axiom."44 That Frege did not regard this appeal to a priori intuition and the indubitability of statements as either subjectivist or psychologistic has struck some philosophers as strange.43 But it is most likely due to his construa1s: (a) of psychology as an empirical science; (b) of the psychological source of knowledge as sense perception; (c) of statements as entities of an eternal and immutable third realm; and (d) of a priori knowledge as strictly universal. necessary, apodeictically certain, and, hence, objective.46 But even Frege realized that the a priori apprehension of third realm objects would be impossible without relating those objects to a mind, and that this relationship between mind and object might be thought to compromise the objectivity of knowledge. Frege did try to explain why the apprehension of thoughts does not compromise their objectivity- saying that apprehension affects only the 'inessential properties' of the thought.47 But what Frege failed to realize was that his Basic Law V of The Fou.ndations of Arithmetic, which as an axiom was claimed to be both true and indubitable, leads, in conjunction with the other axioms of his system, directly to Russell's Paradox. It is the accumulated weight of these and other failed attempts to articulate infallible foundations that have assured many contemporary philosophers that infallible cognitive sources do not exist. This, in away, is unfortunate. For an adherence to methodological scepticism would caution doubt concerning exactly what such failed attempts mean. And here, the fact that some contemporary philosophers regard this conclusion as justified on inductive grounds is as ironic as it is unwarranted. The non-existence of infallible cognitive sources is not justified on inductive grounds. It is not justified at all. Not unless our failure to fmd a cure for cancer means that no such cure exists.4I I do not know whether infallible cognitive sources exist or not, but I think it is reasonably clear that the philosopher who
180
M.A. NOTTURNO
thinks their non-existence is proven by induction has mistaken disappointment for proof. Nor should this contemporary denial of infallible sources of knowledge be confused with either a denial of the possibility of knowledge or a rejection of French Epistemology. While Descartes understood the sceptic to be one who denies 'the possibility of absolute certainty - 'probable guesses' not qualifying as knowledge, no matter how 'consonant with reason' they may be - few contemporary philosophers regard themselves as sceptics. And while most contemporary philosophers acknowledge that 'proofs' are always contingent upon foundations and foundations are never absolute, they nonetheless continue to presuppose that a statement needs to be justified, in one way or another, before it can be rationally accepted as true. Still, what such failed attempts do mean is that any future claims regarding infallible sources should, on Cartesian grounds, be subjected to doubt- it being wiser not to trust entirely to any thing by which we have once been deceived. It is this doubt regarding the infallibility of cognitive sources, as opposed to any certainty regarding their non-existence, that spells an end to the Cartesian Progmmme. The Cartesian Programme attempted to derive objective certainty from methodological doubt by showing that there exists some ' Archimedean point' that cannot, in the final analysis, be rationally doubted after all. The discovery of such an 'Archimedean point' would, in Descartes' words, allow us to do 'great things'. And if there really were some Archimedean statements that were rationally indubitable, then there would be little problem in saying that those statements can be rationally accepted without justification and that they provide the foundation for the rational acceptance of all other statements. In that case, the French epistemologist could maintain his demand for justification- the privilege of foundational statements being explained by their difference, by the fact of their rational indubitability. But in lieu of foundational statements that are rationally indubitable, French epistemology forces us into the Fregean Dilemma. We must either characterize any acceptance of beliefs that differ from a certain set as irrational, or deny the possibility of rational belief at all. Here, the prevailing contemporary attempts to found knowledge on convention or commitment, be it to a linguistic community or to a form of life, suffers from a kind of bad faith. For the myth of the foundation is like every other myth: its effectiveness as a myth depends upon its being taken for the truth. If it is really true that Archimedean points are a dime a dozen, then it may no longer be possible to do 'great things'. But if doubt regarding the infallibility of cognitive sources spells an end to the project of deriving objective certainty from methodological doubt, it need not prevent us from rationally accepting statements as true- not, at least, if we distinguish sharply between methodological scepticism and French Epistemology. In what follows, I will suggest that an epistemological approach that maintains an adherence to methodological scepticism but repudiates French Epistemology offers the best resolution to the Fregean Dilemma.
FRENCH EPISTEMOLOGY
181
IV. Epistemological Buddhism Earlier I said that the force of the Fregean Dilemma depends upon the acceptance of French Epistemology. There, I characterized the Fregean Dilemma as a crisis for objectivity, i.e., the problem being to salvage objective knowledge without appealing to infallible cognitive somces. But in tracing the Fregean Dilemma to French Epistemology, the terms of our discussion have changed. What was originally characterized as a crisis for obje.ctivity has now been characterized as a crisis for rationality, the problem being to salvage the rational acceptability of statements without appealing to infallible cognitive sources. Here, I hope it is obvious that if the Fregean Dilemma poses a crisis for objectivity then it also poses one for rationality- and for the same Kantian considerations that characterize what is objective as what is binding on all rational beings. Hence, Frege defended classical logic as an objective criterion for rationality, and hence Frege attacked the psychologistic tendency to regard non-classical logics as valid for those that hold them as "a heretofore unknown type of madness."·' Elsewhere50 I have argued that twentieth century philosophy is in the process of an epistemological paradigm shift in which the idea of knowledge as apodeicticly certain (what I have called 'EP1 epistemology') is gradually being replaced with the idea of knowledge as inherently fallible (what I have called 'EP2 epistemology'). There I argued that that shift in our understanding of 'knowledge' necessitates a systematically related shift in our understanding of other key epistemological teons like ' objective' , 'rational', ,scepticism', and the like. To the extent to which rational knowledge has been construed as justified true belief, this paradigm shift can be understood as an attempt to salvage rational knowledge from scepticism.. Still, the epistemological paradigm shift from EPI to EP2 has genemlly involved a weakening of the standards for justification. Where EPI sceptics had argued that: 1. Rational knowledge is justified true belief. 2. JustifICation guarantees the truth of what it justifies. 3. Logical justification guarantees the truth of what it justifies, but this guarantee is, ultimately, contingent upon the truth of premises which cannot be logically justified. 4. Extra-logical 'justification' provides no guarantee for the truth of what it 'justifies'. Therefore: S. Rational knowledge is impossible.
EP2 cognitivists have generally replied by denying that justification guarantees the truth of what it justifIeS. This weakening of the concept of justification has occurred in one of two ways. EP2 inductivists have recognized inductive inference as providing logical justification for general statements- thereby denying 2 and weakening their concept of logical
182
M.A. NOTTURNO
justification. EP2 fideists have argued that the validity of general statements follows naturalistically from our commitment to their use in defining the behavior of a linguistic community- thereby denying 4 and weakening their concept of extra-logical justification. These movements sometimes go by the name of 'pragmatism', but they seem decidedly at odds with the pragmatism of Peirce, who always fought against what he called 'the frxation of belief'.51 I have already indicated that I regard each of these tendencies as a misguided form of psychologism. By weakening standards for justification, inductivism and fideism encourage us to regard as settled issues and disputes that really are not. While there is nothing in the nature of things to prevent such a weakening of standards, my own sense is that most human beings are, in their quest for certainty and security, already too inclined toward a parochial closed-mindedness and fiXity of belief which is, or should be, the purpose of epistemology to destroy. Still, the motivation for weakening the standards for justification is understandable: we all believe that we are rational; and we have all grown wary of foundations. IT there really exists a plurality of 'foundations', and no one 'foundation' is, in any non-questioning begging sense, more justified than another, then revising our concept of justification may seem our best bet to salvage rationality. The revision, however, is deceptive. It maintains a foundationalist theory of rationality without the benefit of a rational foundation. It is, moreover, both socially dangerous and scientifically sterile. For it leads, in the fmal analysis, to the same sort of unquestioning pre-judgment that has historically stifled social integration and scientific growth, and that has traditionally been the burden of rationality to combat Indeed, it not only leads to such unquestioning pre-judgment, it seems to vindicate and endorse it For these reasons, I have written elsewhere 52 that all attempts to justify a statement as true or as false are in the same epistemological boat and that rational judgments of rationality are, within a justificationalist programme, possible only given the acceptance of foundational principles (goals, values, etc.) and are made only with reference to the foundational principles accepted. All attempts to justify a statement as true or as false are in the same epistemological boat because their success or failure ultimately depends upon the acceptance or rejection of some statement(s) without justification. Rational judgments of rationality are possible within a justificationalist programme only given the acceptance of foundational principles (goals, values, etc.) and are made only with reference to the fOWldational principles accepted because the judgment that a statement (belief, action, etc.,) is rational is, within a justificationalist programme, ultimately the judgment that that statement (belief, action, etc.) is, in one way or another, justified. But all this presupposes our distinction between the objectivist and subjectivist justified true belief theories of knowledge and is, therefore, very easily misunderstood. Hence, Tom Settle write~~: ------------
FRENCH EPISTEMOLOGY
183
Having abandoned justificationism - roughly, the idea that there can be absolute certainty as a result of the proper use of reason - Nottumo thinks that all attempts to justify theories as true or false are in the same boat. But this presumes that justifying is an all or nothing affair and I am not willing to give up the word 'justify' to the absolutists, any more than I was willing to give up the word 'true' to the conventionalists. There is an ancient history to the notion of justifying as showing to some appropriate court that someone is innocent of some charge. I dunk this is a good cue for us. Justification, in my view, can best be seen as the act of satisfying an appropriate body of people that some belief or action is what a person should hold or do, or may hold or do. Life is full of it, though which body of people we are trying to satisfy, and what the standards are that we have to meet to do so, vary enormously.53
But as we have already seen, satisfying an appropriate body of people that a statement is true, and justifying that statement as true are two entirely different things. Satisfying an appropriate body of people that some belief or action is what a person should hold or do may count as a justification of that person's holding that belief or doing that action. And an act of justifying a statement may, simultaneously, be an act of satisfying an appropriate body of people that some belief or action is what a person should hold or do, or may hold or do. But to see justification as this, and nothing more, is to adopt what I earlier called a 'subjectivist' justified true belief theory of knowledge. But apart from this, I do not equate justificationism with the idea that there can be absolute certainty as a result of the proper use of reason (though this is what Descartes, interpreted literally, and Frege seem to have had in mind). What I mean by 'justificationism' is nothing more nor less than what I have here called 'French Epistemology' - the idea that a statement must be justified - either logically or extra-logically- in order for it to be rationally accepted as true. In my Objectivity, Rationality, and the Third Realm I sketch a quite different approach to rationality- one that maintains a suspicion of objective foundations and purported justifications without either weakening the standards for justification or denying the possibility of rational knowledge. There, I suggest that the rational man in the twentieth century
can: ...best be described as a sort of epistemological Buddhist- one deeply committed to the quest for truth, ready to entertain competing theories, and even willing to apply them in practice, but always wary of fonning deep attachments to any specific claim to truth. Such a man may be characterized as having beliefs. But what the epistemological Buddhist lacks is the commitment to belief.54
Epistemological Buddhism assumes the loss of objective or justified certainty as the epistemological problem context of the twentieth century. It acknowledges that human action in the face of WlCertainty poses deep problems for both individuals and institutions. But it argues that these problems cannot be resolved by either the proliferation of 'objective'
184
M.A. NOTTURNO
standards (what might be called 'the cult of objectivism') or the denial th our actions are, in the final analysis, either meaningful or important (what might be called 'the cult of nihilism'). Rather, Epistemological Buddhism criticizes each of these contempomry tendencies as forms of self-delusion tantamount to what Sartre might have called 'epistemological bad faith' that function to illicitly reclaim the certainty and secwity they initially deny. Epistemological Buddhism also criticizes some of the consequences of the contempomry pragmatism mentioned above. One such consequence is the transformation of judgments into calculations, a transformation which camouflages the essential subjectivity of the decision-making process. Such transformation and camouflage, according to Epistemological Buddhism, are but further forms of epistemological self-delusion. Rather, Epistemological Buddhism maintains: (1) that our decisions require judgments, not calculations (that's why they are decisions); (2) that these judgments are grounded in subjectivity (that's why they are judgments); (3) that such judgments are inherently fallible (that's the lesson of the twentieth century); but (4) that such judgments are meaningful and important (because our decisions do have their consequences). In maintaining these points, Epistemological Buddhism rejects both the traditional construal of knowledge as justified true belief and its Cartesian corollary that we have here called 'French Epistemology'. But instead of embracing scepticism or the irrational, Epistemological Buddhism argues that foundational statements cannot be justified and that our rational acceptance of such statements should thus not be construed as contingent upon the quality of their justification. Contrary to French Epistemology, Epistemological Buddhism maintains that it is rational to regard a statement as true, unless or until one has good reason to believe it false. In so doing, Epistemological Buddhism maintains that we have no recourse but to trial and error (whether we wish to acknowledge it or not), and that the deep problems posed by acting in uncertainty can only be resolved by confronting our equally deep fears of error and change. Simply put, the resolution of these problems requires the acceptance of error and change as essential components of learning - as events that may be painful, but that may also lead to growth - and not, in any event, as events to be avoided, hidden, and denied at all costs. Here, Epistemological Buddhism advocates that the best way in which one can search for truth is by: (1) remaining methodologically sceptical of all claims to truth; and (2) thinking oneself into as many opposing perspectives as he possibly can. For it is only in this way that the Epistemological Buddhist will unqerstand the strength of each of the opposing candidates for truth. And without such understanding, neither genuine agreement nor genuine disagreement is possible. Here I do not want to be misunderstood I believe that quite literally any statement can be 'justified' logically provided that one is willing to accept without logical justification the statements necessary to justify it. But I also believe that the idea that we have access to an infallible source of knowledge that can give extra-logical justification for our cognitive claims
FRENCH EPISTEMOLOGY
185
is simply implausible. Insofar as this is concerned, I am, like Frege, opposed to psychologism because it encourages us to regard as justified statements and thec:ries that really are not. But I am also opposed to Frege's anti-psychologism- and for precisely the same reason! But despite all this, I agree with Settle" that the caricature sceptic or beginning philosopher, who refuses to be convinced by any argument provides no reason for someone to doubt his own beliefs- no more than someone's ability to convince this caricature sceptic of his beliefs would provide those 'beliefs with justifICation. But let's not leave the caricature sceptic just yet For the caricature sceptic raises a much deeper point, a point that is much more difficult to resolve, and that is ultimately at the very root of the problem with French Epistemology. Exactly why is the caricature sceptic a caricature? Presumably, because his scepticism is not real, because it is merely a stance, because he doesn't really have a different belief, but only a way with words and a logical obstinancy that renders him obnoxious to the more rermed epistemological sentiments around him. The caricature sceptic is not serious, he is merely cantankerous. Better to recognize him as such than to let him disturb the satisfaction and solidarity of Settle's appropriate body of people. Now I would not for a moment either doubt or deny that there are caricature sceptics every bit as disingenuous and cantankerous as Settle describes. But I also 'think that it is too simple to presume that every sceptic whom we find impossible to convince is such a caricature- and that it is too difficult to determine whether and when some particular sceptic is not. The caricature sceptic who refuses to be convinced by any argument gives us no reason to doubt our own beliefs. But the real sceptic who is not convinced by OUT arguments does. The caricature sceptic gives us no reason to doubt our own beliefs because our judgment that a given sceptic is a caricature is, simultaneously, the judgment that his beliefs are really no different from our own. But the real sceptic is an entirely different matter. The real sceptic gives us reason to doubt our own beliefs- not because of French Epistemology, but because of what Charles Peirce used to call 'the social impulse': namely, the recognition that other people think differently, and the thought that is "apt to occur...in some saner moment that their opinions are quite as good as [our] own."" It is this recognition that other people think differently and this thought that might occur in some saner moment that is at the root of the problem with French Epistemology. If there really were some statements that everyone believed, then there would be little opposition to saying that those statements can be rationally accepted without justification and that they provide the foundation for the rational acceptance of all other statements. In that case, the French epistemologist could defend his demand for justification- the privilege of foundational statements being explained by their difference, by the fact that no one believes them to be false. And in that case, the 'sceptic' who complained that the foundations were unjustified could be ignored as a
186
M.A. NOTTURNO
caricature, and would, being a caricature, give us no reason to doubt our own beliefs. But in lieu of foundational statements that everyone believes, French Epistemology once again forces us into the Fregean Dilemma. We must either characterize any acceptance of beliefs that differ from a certain set as irrational, or deny the possibility of rational belief at all.
v. Farewell to Commitment I have characterized Epistemological Buddhism as an anti-foundationist epistemology that does not reduce rational knowledge and rationality to creedal commitment and group consensus. But some contemporary philosophers faced with the decline of foundationalism, have characterized rational knowledge - and rationality itself - as presupposing some form of commitment- either to a form of life, or a linguistic community, or a scientific paradigm. Creedal commitment, according to these philosophers, is not simply something that knowers have, it is something that they must have in order to carry out their cognitive endeavors. Hence, Thomas Kuhn, in construing so-called 'normal science' as a rational endeavor, writes that: Men whose research is based on shared paradigms are committed to the same rules and standards for scientific practice. That commitment and the apparent consensus it produces are prerequisites for normal science, Le., for the genesis and continuation of a particular research tradition.57
And Stanley Fish, in denying the possibility of theory and value neutrality, writes that ...knowledge and conviction do not depend on such a neutrality, but on a commitment to the perspective from which one speaks, a commibnent one cannot possibly be without"
I too believe that a rational person is committed to something. But I also think that these attempts to link rationality to creedal commitment ultimately define rationality in terms of group identity- identity with the group that believes such-and-such. And this consensus approach to rationality is, I think, the very anti-thesis to the rational tradition. The contemporary concern with commibnent can be seen as an enlightened albeit fallacious move in response to the Fregean Dilemma. Frege, as we have already seen, maintained a sharp and intuitive distinction between a statement's being-true and its being-believed-to-be-true. According to Frege, someone's belief that a statement is true no more entails that statement's truth than a statement's truth entails that someone believes it. For this reason, Frege maintained an equally sharp distinction between knowledge and belief. Knowledge may be a species of belief, but it demands a stronger justification than belief alone can ever supply.S9 Thus, - - - - - - - - - - - - - - - - - - - - - - - - -
FRENCH EPISTEMOLOGY
187
Frege maintained that the justification for knowledge must be extrapsychological. But just because Frege also believed that such justification is possible, he argued that it is possible for humans to directly apprehend truth through extra-psychological means. Now psychologistic philosophers, like Hwne and Mill, denied that extm-psychological apprehension of truth is humanly possible. But according to contemporary psychologistic philosophers, there is no psychologically neutral perspective from which humans can objectively view the world. And just because there is no psychologically neutral perspective from which humans can objectively view the world, contemporary psychologistic philosophers maintain that the sort of justification that Frege envisioned is impossible as well. Here, psychologistic philosophers collapse the traditional distinction between knowledge and belief. If knowledge differs from belief at all, it does so by its degree of conviction and not by its quality of justification. But having collapsed the traditional distinction between knowledge and belief, psychologistic philosophers also feel obliged to deny the intuitive distinction between a statement's being-true and its being-believed-tobe-true. They feel obliged to do this because they recognize (believe?) that any assertion that a statement is true ultimately reports a belief that that statement is true. Without access to extra-psychological justification, being-true is equated with being-believed-to-be-true, and the plurality of belief with a plurality of truth. Viewed from 'within the context of this debate, the move to commitment seems enlightened in that it denies human access to any psychologically neutral perspective (or infallible cognitive source) from which to assess the validity of cognitive claims. As Fish puts it: ...facts can only be known by persons, and persons are always situated in some institutional context; therefore facts are always context relative and do not have a form independent of the structure of interest within which they emerge into noticeability. It is not simply that supra-contextual facts are unavailable to us, but that the very notion of a supra-contextual fact makes no sense since something that had no relation to the concerns of a particular hwnan situation would not be a fact and therefore could not be known by a human agent What that something might be, God only knows.-
Fish's rhetoric is apt. For it is, on the anti-psychologistic view, as if God has a library in which he keeps an encyclopedia, a dictionary, and a bible. In these books God has inscribed the laws of science, the meanings of words, and the imperatives of ethics on a medium more durable than stone. Here, the aim of cognitive inquiry is to find the key to God's library. But here, Fish's move is to deny that God's library exists. Now it is just this denial of the existence of God's library that I consider fallacious. It may well be true that facts can be known only by persons and that persons are always situated in some institutional context. But it does not follow that facts are always context relative and do not have a form independent of the
188
M.A. NOTTURNO
structure of interest within which they emerge into noticeability. Rather, the most that follows is that our knowledge of facts is always context-relative and interest-dependent. Simply put, it is not the notion of supra-contextual facts, but the notion of supra-contextual albeit human knowledge of such facts that is nonsensical. Here, Fish is right Exactly what such supra-contextual facts might be, God only knows. But contrary to Fish, God's library may very well exist And it may exist despite the 'fact' that the key to God's library is locked within the library itself. For despite this 'fact', the notion of supra-contextual facts does make sense. For our knowledge of such facts would bear a relation to particular human situations- if only we knew what God only knows! Still, the Epistemological Buddhism that I advocate is a pluralist position, and Fish believes that pluralism is incoherent. Fish, moreover, seems to believe this because it is psychologically impossible for a person to simultaneously believe in the truth of inconsistent statements. It is, moreover, in this context that Fish says that knowledge and conviction do not depend upon the existence of a neutral perspective from which to describe the world, but upon a commitment to the perspective from which one actually speaks. In so doing, Fish appeals to commitment to insure the possibility of knowledge and conviction. Now commitment and conviction are time-honored epistemological virtues. But they are also the virtues of the very epistemological tradition that Fish rejects, Le., they are the virtues of an anti-psychologistic tradition that maintains the possibility of an objective knowledge free from all perspectival ambiguity. Within this tradition, one acquired knowledge by giving objective and rational justification for a belief. And within this tradition, commitment and conviction were thought to follow as psychological consequences of believing that the justification that one had given for that belief was, indeed, objective and rational. Of course, the conviction and knowledge of which Fish speaks cannot require such objective and rational justification. So objective and rational justification is not, in Fish's view, a prerequisite for knowledge and conviction. But in lieu of such objective and rational justification, Fish says that it is one's commitment to a perspective that makes knowledge and conviction possible. For it is only given such commitment that one can maintain conviction that his 'knowledge' is true. But since Fish denies the possibility of objective and rational justification, it is difficult to see what this knowledge and conviction could amount to other than 'the psychological fact that one happens to believe that a particular proposition is true. For just this reason, claims to knowledge and conviction (as opposed to belief) decline, and should decline, as soon as one recognizes that the perspective from which he speaks is not the only perspective from which one can speak. For as soon as one denies the possibility of objective and rational justification, the possibility of a human knowledge qualitatively different from human belief collapses. One could, given this recognition, maintain his claim to knowledge and conviction by believing that the perspective from which he
FRENCH EPISTEMOLOGY
189
speaks is the justifIably correct perspective, the perspective from which he should, in some objectively justified sense, speak. But Fish does not believe this. And neither do I. Nevertheless, Fish does say that commitment to the perspective from which one speaks is something that one cannot possibly be without. And It is with this latter statement that I take issue. Fish, of course, is not alone in this position. Many contemporary philosophers fail to distinguish between belief and commitment. Others, aware that such a distinction is sometimes maintained, deny that one can properly be drawn. For these philosophers, having a belief, performing an action, or speaking from a perspective is tantamount to being committed to that belief, action, or perspective. This denial of a distinction between belief and commitment sometimes proceeds from the psychologistic tendency to extend one's creedal obligation from the particular statement that one believes to all of its logical consequences. Here, the simple belief that P is true incurs 'commitment' not only to P, but to the infinity of statements that can be deduced from P as well. But here, the confusion is not really between belief and commitment at all. Rather, it is between creedal commitment and logical commitment. In practice, our interest in the logical consequences of P lies"' with what one should believe (if he in fact believes that P) or with what one should not believe (if he in fact doubts one or more of the logical consequences of P). And it is, in practice, only because one can be logically 'committed' to a statement that he does not believe (let alone to one that he is committed to believe) that this exercise in drawing inferences has much point. But the denial of a distinction between belief and commitment is more often bolstered by an underlying allegiance to behaviorism. And it is, with this support, more difficult to combat. For the behaviorist strategy is to determine one's ~liefs from one's actions. And since 'the behaviorist recognizes no observable differences between performing an action and being committed to that action, he likewise recognizes no difference between believing that P is true and being (creedally) committed to P. I know of no logical argument that would rationally compel the committed behaviorist to recognize a real distinction between belief and commiunent. This, however, is not because arguments do not exist, but because the 'facts' to which such argwnents appeal cannot exist as facts within 'the behaviorist framework. To argue that behavior can be interpreted as action only with reference to intentions or that commitment to belief requires a mental act over and above the belief itself will not suffice. For it is just this existence of intentions and mental acts that the committed behaviorist is committed to deny. This, nonetheless, is precisely what I will suggest- with the hope of persuading those who believe in behaviorism, but are not yet quite committOO to it Is believing that P the same thing as being committed to P? What could that mean? The sense of tcommitment' most appropriate for this discussion is that of obligation, dedication, or resolution. It is the sense of 'commitment' in which one makes a promise (of one sort or another) to
190
M.A. NOTTURNO
whatever it is that he commits himself. It is in this sense that we speak of the commitment of marriage, or of the commitment undertaken when one signs a contract Promises, of course, can be broken. But it would make no sense to speak: of breaking a promise unless it were possible to make one. And it would make no sense to speak of making a promise unless making a promise differed, in some way, from not making one. But here, what differentiates making a promise from not making one is nothing more nor less than the promising itself, i.e., it is nothing more nor less than the obligation one assumes and the resolution one makes to fulftll the intention of that promise. Still, there is nothing in one's overt behavior (save, perhaps, the linguistic behavior of saying 'I promise', though one may always make a promise to oneself) that marks a distinction between promising and not promising. Two people can behave in a way indistinguishable from that of married couples without ever undertaking the marriage commitment. And two parties can engage in the activity of give and take characteristic of contractural relationships without ever entering into a contract. Still, there is a difference between making a promise and not making one. This difference, however, lies not in one's overt behavior, but in his mental actions. It lies in the obligation that one assumes and the resolution that one makes to fulftll the intention of the promise. But what sort of promise can we make and what sort of obligation can we owe to a belief or a perspective? If we claim to be committed to a belief or perspective, what are we promising to do? What the marriage vows add to the conjugal behavior of two people is simply the promise or resolution not to alter that behavior. And the same holds true for any contractural relationship. For this reason commitments are undertaken primarily to add an element of security to the relationship in question. But this element of security amounts to nothing more nor less than the resolution one makes and the obligation one assumes to remain within it. Insofar as this is concerned, believing that P is true differs from the commitment to believe that P is true in that the latter, but not the former, involves the promise to continue to believe that P. Simply put, commitment to belief involves the resolution not to change one's mind in relation to that belief. As such, commitment to a belief incurs the obligation to defend that belief (if only privately) against all criticisms oc, at least, not to consider any such criticisms seriously. In this way, creedal commitment is analogous to the marriage commitment For in committing oneself to a belief, one in effect promises to love, honor, and obey that belief and to forsake all other beliefs in order to do so. Analogously, commitment to a foundational perspective involves a resolution to continue to view the world from that perspective. As such, it is tantamount to a promise not to shift one's perspective or change one's point of view. It seems clear to me that people actually do commit themselves to beliefs and perspectives in just this way. But it seems far less clear that such creedal commitment is an appropriate response to the loss of certainty.
FRENCH EPISTEMOLOGY
191
Having a belief is not a voluntary matter. One can no more decide to have a belief than one can decide not to see what his eyes show. Rather, we believe a statement because that statement seems true. In this way, believing that P is true and P seeming true are one and the same thing. But if belief is not a voluntary matter, then there can be no guarantee that what seems true today will continue to seem true tomorrow. At least, there can be no such guarantee so long as we continue to entertain what others believe as serious possibilities. For so long as we regard what others believe as serious possibilities, it is a serious possibility that what seems true to them will suddenly seem true to us. But it is just here that the element of decision becomes crucial. One cannot decide not to see what his eyes show. But one can decide to close his eyes, or tum them in a different direction so as not to allow his eyes to show what he dres not want to see. Analogously, one cannot decide not to believe what seems true. But one can decide not to consider any opposing beliefs as possibly true. One can do this either by deciding not to infonn oneself of such opposing beliefs or by prejudging all such beliefs as mistaken. Hence, if it is possible to commit oneself to a belief at all, it is only because one can, and does, decide not to entertain any opposing beliefs as serious possibilities. But in this way, creedal commitment has the effect of stifling cognitive growth. Creedal commitment stifles cognitive growth by requiring that one close his mind to the consideration of opposing beliefs or perspectives as serious possibilities. In closing his mind to the consideration of opposing beliefs as serious possibilities, one may achieve a sort of confidence and secwity. He can 'rest assured' in subjective certainty. It is, however, just this 'rest' which precludes the possibility of growth. But Fish believes that creedal commitment is not only an appropriate response to objective uncertainty, but a necessary one to insure knowledge and conviction. For if it is truly impossible to justify any foundational perspective in a non-circular way, then it is only creedal commitment that allows one to rest assured in his conviction that a given statement is true. But even were truth purely dependent upon one's perspective, such commitment and conviction would seem to preclude the cognitive growth that occurs through an appreciation of the richness and strength of opposing perspectives- an appreciation that only occurs when one entertains such perspectives as serious possibilities. Were truth truly dependent upon perspective, then one might rather strive to apprehend as much truth as possible by viewing the world from as many perspectives as possible. Now pluralists like myself believe that it is impossible to objectively justify any foundational belief as true. But we do not thereby believe that objective truth does not exist. Nor do we believe that a statement's being-true is equivalent to, or entailed by, someone's belief that that statement is true. Again, we believe that God's library may very well exist And we believe this despite our denial that any person ever has access to a neutral perspective from which to read what might be written there, Le., from which a statement could be objectively justified as true. Moreover, we do
192
M.A. NOTIURNO
not only believe that objective truth very possibly exists, we also believe that the discovery of objective truth should be the regulative ideal for all rational inquiry. And we believe this precisely because we believe that it is possible to be right or wrong in our belief that a statement is bUe or false even in the absence of an objective justification that that statement is true or false. While we are, in truth, not committed to this or any other belief, we are committed to the continual and relentless search for truth (despite the fact that we also acknowledge the possibility (but do not believe) that bUth does not exist!) And in the face of our denial of the possibility of objective justification, this commitment to the search for truth also commits us to the activity of considering all beliefs and foundational perspectives as if they were true. It is this commitment, rather than the commitment to any belief, that differentiates Fish's position from my own. But Fish considers pluralism, and its attendant 'critical generosity', to be incoherent. Hence, Fish writes that tt •••generosity is at once the cornerstone of the pluralist ideology and the source of its confusion tt: A pluralist wants simultaneously to make room for everyone's point of view here the motto might be Wayne Booth's commandment to pluralists, "Give your neighbor's monism a fair shake" - and to insist nevertheless that one point of view is superior to all others.61
But the consistent pluralist does give his neighbor's monism a fair shake. He gives his neighbor's monism a fair shake by admitting that monism is self-consistent and can account for all the pertinent data. In short, he acknowledges that monism is possibly true. But far from insisting that his point of view is superior to all others, the consistent pluralist simply maintains that his neighbor's monism is his neighbor's, and not his own, point of view. In so doing, the consistent pluralist pronounces his pluralism as his own point of view and claims no greater authority for it than that. The consistent pluralist does, of course, assert his belief that the monist's position is false. But it would be strange to interpret this assertion as indicating that pluralism is inconsistent or that the pluralist considers his own point of view superior to all others. For all that such superiority would amoWlt to is that the plura1ist believes that pluralism is true. And what else would one expect the consistent pluralist to believe? But what the consistent pluralist does not believe, and what his pluralism is at pains to deny, is that his position has any greater justification or authority than the monist's. This, however, is exactly what distinguishes the consistent pluralist from the monist For while the consistent pluralist and the monist both believe that their respective positions are bUe, the monist asserts that his position admits of some stronger justification than all others. And it is this belief that his position is justified as true, as opposed to the belief that it is merely true, that permits the monist to believe that his point of view is superior to all others.
FRENCH EPISTEMOLOGY
193
Here, the Epistemological Buddhism that I advocate contains elements of both the psychologistic and anti-psychologistic pictures outlined above. First of all, I agree with the psychologistic philosophers that there is no psychologically neutral perspective from which humans can directly apprehend what is true. And since I agree that such direct apprehension of truth is impossible, I believe 'that the sort of extra-psychological justification that philosophers like Descartes and Frege envisioned is impossible as well. H what we claim to know differs from what we believe, it does so by the degree of our conviction and not by the quality of its justification. But I also think that it is a dangerous practice to confuse what we claim to know with what is justiftably true-- even if such 'justification' is acknowledged to come only by psychological commitment.62 When it comes to knowledge, even knowledge of mathematics ,and logic, the most that we ever have to go on in any cognitve claim is our own best guess. And so long as we remain open to 'the best guesses of others, there is nothing in our own best guess to be particularly embarrassed about But this does not mean that we should claim that our own best guess is anything more than it actually is. In particular, we should always remember that our own best guess might be mistaken. And just because our own best guess might be mistaken, I maintain 'that being true is utterly distinct from being believed to be true. For these reasons, I maintain that commitment to belief is an inappropriate response to the loss of justification. The certainty of belief that results from such commitment or conviction might allow one to rest assured in those beliefs. But if one is certain of or assured in his beliefs, then he lacks the primary motivation for considering the beliefs of others as serious possibilities. Simply put, he lacks the recognition that his own beliefs might be mistaken. Again, I do not wish to be misunderstood. It may indeed be true that the ability to view an object presupposes that the viewer occupies some perspective (taken in its broadest sense). And to that extent, it seems undeniable that the perspective that one occupies at any given moment is his own, i.e., it is the perspective that he occupies. If this is all that Fish means to say, then I readily agree- one can never be without his own perspective. But it does not follow from this that a viewer needs to always occupy the same perspective or, even less, that he need be committed to the perspective that he happens at the moment to occupy. Consider, for example, Wittgenstein's famous figme of the duck-rabbit From one perspective, this figure appears as a duck, from another as a rabbit. And it is always possible that, from some third perspective, the figure appears as something else- perhaps as lines on a sheet of paper with no apparent content Seeing this figure as a duck or as a rabbit requires that one occupies a certain perspective. And some people may be so 'committed' to their perspective that they are only able to see this figure as a duck and others may be so 'committed' to their perspective that 'they are only able to see the figure as a rabbit. Now I readily acknowledge that it may be psychologically impossible for a viewer to see this figure simultaneously as
194
M.A. NOTTURNO
both duck and rabbit Still, as soon as one has made the perspectival leap that permits him to see the figure now as duck, now as rabbit, the question of commitment seems ridiculous. I can well imagine that someone who has made this perspectival leap might 'commit' himself to seeing Wittgenstein's figure always as a mbbit But it is harder to imagine why he would do so. Well, that's not exactly true. I can understand that someone might make such a 'commitment' if, for example, he found the ambiguity in Wittgenstein's figure psychologically disturbing. But the suffocation that results from sticking one's head that far in the sand is, from my perspective, more disturbing still. But if the consistent pluralist does not believe that his own position has any greater justification than any other, then why does he bother to assert his position at all? And why should he bother to engage in rational criticism of the views of others? Here, part of the pluralist's motivation derives from the 'fact' that he believes that his own position is true and hopes to persuade others to adopt it. Again, there is no contradiction here. For the consistent pluralist believes not that no position is true, but that no position can be objectively justified as true. Nor is the pluralist's position an unassailable article of faith. For the simplest way to convince a pluralist that he is wrong in his pluralism is to produce a statement that is objectively justified as true! But another part of the pluralist's motivation derives from the 'facts' that he wants to understand the world around him and that included in what he wants to understand are the various ways in which others understand it This motivation, moreover, goes well beyond mere intellectual curiosity. Since the consistent pluralist commits himself to no belief, he is always able to consider the views of others as serious possibilities. As such, he is always able to alter his own beliefs without trauma when any such possibility suddenly appears more serious than his own. And here, the pluralist's attempt to view the world from the perspectives of those whose beliefs differ from his own functions as an important frrst step toward the achievement of such an understanding. Such adherence lnay well consititute critical generosity. But if it does, then we must also recognize that the critic's generosity extends as much to himself as it does to those he criticizes. For in his exercise of such critical generosity, the critic takes the frrst step toward increasing his own understanding and with it increases the possibilities for his own cognitive growth. And it is, in this way, that critical generosity promotes cognitive growth. I realize that my vision of rationality may be unsettling for some. For the decision to regard the opposing beliefs of others as serious possibilities will decidedly not issue in the sort of commitment and conviction that would allow one to 'rest assured' in his beliefs. But if Epistemological Buddhism entails a loss of confidence, it need not result in despair. For in the Epistemological Buddhist's view, the security 'that results from such commibnent and conviction is neither a virtue nor an aid in the search for bUlb.
195
ENDNOTES 1. It is important to emphasize that the opponents of psychologism generally understood 'psychology' to refer to an empirical science. This explains why an anti-psychologist like Frege could appeal to self-evidence in his appraisal of axioms. But it also lDulerscores the fact that psychologism is a cousin to other attempts to 'naturalize' epistemology, e.g., sociologism, biologism, etc. 2. Frege's critique of psychologism is scattered throughout his writings. The most important sources are: Gottlob Frege, The Basic Laws of Arithmetic, ed. and trans. Montgomery Furth (Los Angeles: University of Califomi~ 1967), especially Frege's "Introduction"; Gottlob Frege, The Foundations of Arithmetic, trans. IL. Austin (Evanston: Northwestern University, 1968); Gottlob Frege, "Review of Dr. E. Russerl's Philosophy of Arithetic, trans. E. K1ug~ MIND LXXXI (July, 1972); and Gottlob Freg~ ''The Thought: A Logical Inquiry," trans. A.M. and Marcelle Quinton, MIND LXV (1956). For further discussion of the epistemological nature of Frege's critique of psychologism, see: M.A. Nottumo, Objectivity, Rationality, and the Third Realm (Boston: Martinus Nijhoff, 1985). 3. Isaac Levi, I believe, has something similar in mind when he diagnoses someone as suffering from the "curse of Frege" if he "submits to the polarization and chooses either in favor of method and against psychologism, sociologism, and historicism or chooses against method and in favor of psycho10gism, sociologism, and historicism." (Issac Levi, The Enterprise of Knowledge (Cambridge: MIT, 1980), p. 428) 4. This preference for objective knowledge undoubtedly accounts for the prevalent fear of psychologism, a fear that has made "psychologism" a philosophical term of ill-repute. As Brentano put it, "Psychologism [is a word which] when it is spoken many a pious philosopher, ... crosses himself as though the devil himself were in it." (Franz Brentano, Psychology from an Empirical Standpoint, eel. Linda L. McAlister, trans. Antos C. Rancurello, D.B. Terrel, and Linda L. McAlister (New York: Humanities, 1973), p. 306. Here the point to be made is that philosophers, by and large, do not claim to be psychologistic or to be proponents of psychologisrn; they are accused of it 5. For further discussion, see: No ttumo, Objectivity, Rationality, and the Third Realm, especially chapter 6. 6. Frege, The Foundations of Arithmetic, p. 4e. 7. See: M.A. Nottumo, "Critical Generosity or Cognitive Growth?" METAPIDLOSOPHY (1989); and Nottumo, Objectivity, Rationality, and the Third Realm, chapter 11. 8. See: Ibid., chapter 3. 9. Frege, The Foundations of Arithmetic, p. 4e. 10. The argument is simple. The validity of a deductive argument can be determined only contingent upon the condition that each of its signs retain the same meaning throughout the argwnent Freg~ in fact, insists upon this condition in introducing the signs to be used in his conceptual notation. See: Gottlob Frege, Begriffsschrift, a formula language, modeled upon that of arithmetic, for pure thought, in Frege and GOdeI, 00. Jean van Heijenoort (Cambridge: Harvard University, 1970), p. 11. But it is difficult to see how the meanings of signs could be the same unless meanings were immutable objects. 11. I am indebted to Tom Settle for helping me to appreciate the importance of subjective justification: it is at the heart of every malpractice suit 12. Karl Popper is, to the best of my knowledge, the fust to have noticed these two different justified true belief theories of knowledge. See: Karl R. Popper,
196
M.A. NOTTURNO
Objective Knowledge (Oxford: Oxford University, 1972), chapters 3, 4, especially pp. 108-109.
13. This conflation is, I think, what is ultimately at the root of the so-called. 'Gettier paradoxes'. See: Edmund Gettier, "Is Justified True Belief Knowledge?" ANALYSIS, 23 (1963). 14. Even Frege seemed to believe this, as is evident from this criticism that without a general principle of induction, "induction becomes nothing more than a psychological phenomenon, a procedure which induces men to believe in the truth of a proposition, withollJ affording the slightest justification for so believing." (Frege, The Foundations of Arithmetic, p. 4e, my italics) 15. Edmund Husserl, Logical Investigations, 2 vols., trans. IN. Findlay (New York: Humanities, 1970), p. 111. 16. Assuming, of course, that the objectivist intends to avoid scepticism. 17. Rene Descartes, Discourse on Method, trans. Donald A. Cress (Cambridge: Hackett, 1980), p. 35. 18. Rene Descartes, Meditations on First Philosophy, translated by Donald A. Cress (Cambridge: Hackett, 1980), p. 45. 19. See: Ibid., p. 47. There is a double irony here. On the one hand, Descartes seems to suggest that the atheists, who refuse to submit to the authority of the Holy Scriptures, will nonetheless submit to the authority of the theologians at the Sorbonne. On the other hand, the attempt to prove the existence of God by natural reason would itself seem to preclude any appeal to authority. 20. Ibid., p. 45. 21. Ibid. 22. Indeed, the circularity of Descartes' proof is so well-known that it is often referred to as 'The Cartesian Circle'. 23. These, however, are only two of the more blatant 'ironies' to be found in the Meditations. Descartes' arguments for the existence of God are replete with apparent paradoxes and logical tensions. There is, first and foremost, his self-gratuitous use of the 'light of nature', a use which seems inexplicable in light of his resolution to doubt whatever he can doubt without contradiction. Descartes, of course, drew a sharp distinction between nature and the light of nature: When I say in this meditation that I have been taught so by nature, I \D1derstand only that I am driven by a spontaneous impulse to believing this position. and not that some light of nature shows me it is true. These two positions are at considerable odds with one another. For whatever this light of nature shows me - for example, that from the fact that I doubt, it follows that I am and so on cannot in any way be doubtful, because there can be no other faculty in which I may trust as much as the light of nature that could teach which of these positions are not true. (Ibid., p. 70) But even were we to assume that the light of nature exists and is the most trustworthy of Descartes' faculties, being most trustworthy and being in no way doubtful are two different things. Still, I call Descartes' use of the light of nature ,self-gratuitous' primarily because Descartes so often appeals to it in order to legitimate his acceptance of ideas which he needs to prove the existence of God, but which would otherwise seem to be dubitable, e.g., that the cause of an idea must have at least as much formal reality as the idea has objective reality. See: Ibid., p. 71. But aside from this, there is an obvious tension between his claims: (1) that his idea of infInite Substance is "the most clear and distinct of all ideas" (Ibid., p. 74) and (2) that "the nature of the infinite is such that it is not comprehended by me,
FRENCH EPISTEMOLOGY
197
who am fmite." (Ibid) Descartes says that these two claims are "not inconsistent". But it is difficult to see why they are not. There is, moreover, another tension in his argument that his idea of the infmite is not compounded from his ideas of negation and the finite. Descartes says that this cannot be the case since he could not know that he is finite if he did not have the idea of something more perfect. See: Ibid. But being infinite and being more perfect are obviously two different things. Finally, there is a disturbing disanalogy in his reasoning concerning the respective causes of his ideas of God and material objects. On the one hand, Descartes maintains that he cannot be the cause of his idea of Infinite Substance (God), because he is not infmite. See: Ibid. (Were Descartes infinite, then he would have enough formal reality to cause the existence of his idea of infinite substance, and would, as a consequence, not need to conclude that some infinite substance outside himself must exist in order to cause the existence of that idea.) On the other han~ he maintains that he may well be the cause of his ideas of material objects, despite the fact fact that he is not material. Here, Descartes claims that it is possible for him to be the emi1ll!nt cause of his ideas of material substances despite the fact that he is not material, because being material is only a mode of a substance, and Descartes too is a substance. See: Ibid., pp. 73-74. So one wonders why being infmite does not count as a mode of a substance. This disanalogy is significant. Descartes' proof of the existence of God would be UIDlecessary, could he be independently certain that material objects exist And Descartes could be independently certain that material material objects exist, would he not regard himself as a possible cause of 'their ideas. 24. One need only recall Giordano Bruno, who was burnt at the stake in 1600 because of his 'heretical' philosophical views. 25. See: Ibid., p. vii. 26. See: Ibid. 27. Ibid., p. 46. 28. Descartes, Meditalions on First Philosophy, p. 46. 29. It is, of course, only the possibility of an Evil Demon that allows Descartes to doubt the obvious truths of mathematics and logic. 30. Of course, Descartes' resolution to regard a statement as false until proven true is neither a positive nor a mature doctrine of Cartesian philosophy. It is, rather, a heuristic principle that Descartes introduces toward the end of the "First Meditation" in order to facilitate his doubt by insming that there be, as he puts it, an "equal weight of }X'ejudice on both sides." (Ibid., p. 60) Thus understood, it is intended primarily to guard against Descartes' habit of assenting to and believing in those "long-standing opinions" that ''keep coming back again and again, almost against [his] will." (Ibid) Descartes' positive and mature doctrine (assuming, of course, that Descartes intended the Meditations to be interpreted literally) is announced in the 'Fourth Meditation', where Descartes suggests not that we should regard a statement as false until it is proven true, but that we should suspend our judgment on a statement entirely, avoiding all "probable guesses" until we have a proof one way or the other: Although }X'obable guesses might lead me in one direction, all it takes to move me to assent to the very opposite is the knowledge that they are merely guesswork, not certain and indubitable proofs. These last few days I have ample experience of this point, since everything that I had once believed to be as true as it possibly could be, I have now presumed to be utterly false, for the sole reason that I noticed that I could one way or another raise doubts about it
198
M.A. NOTTURNO
But if I hold off from making a judgment when I do not perceive with sufficient clarity and distinctness what is in fact true, I clearly would be acting properly and would not be deceived. But were I to make an assertion or a denial, then I would not be using my freedom properly. If I tum in the direction that happens to be false, I am plainly deceived. But if I should embrace either alternative, and in so doing happen upon the truth by accident, I would still not be without fault, for it is manifest by the light of nature that the intellect's perception must always precede the will's being determined. (Ibid., p. 82) But Descartes' positive and mature doctrine is, from our perspective, just as problematic as his heuristic principle. 31. Rene David and Henry P. de Vries, The French Legal System (New York: Oceana, 1958), p. 76-77. 32. Thus, Stephen in his History of the Criminal Law writes that: "The presumption of irmocence is otherwise stated by saying the prisoner is entitled to the benefit of every reasonable doubt" (Francis A. Coffin, et al., Plffs. in Err., v. United States, U.S. REPORTS, Vol. 39 (4 March 1895) p. 493) The presumption of innocence, however, is more properly regarded as part of the evidence that an accused is innocent. And the apparently widespread tendency to identify the two in effect conflates an instrument of proof with the stale of mind that mayor may not result in an individual from that proof. For further discussion, see: Ibid. pp. 432-464. 33. Descartes, Meditalions on First Philosophy, p. 57. 34. This, I take it, allows Descartes to doubt the apparent truths of physics. 35. Ibid., p. 59. 36. On the contrary, Descartes supposes the existence of an evil genius, in the context of 'the "First Meditation," precisely because his long-standing belief in the existence of a supremely good God is so strong that it is difficult for him to doubt. See: Ibid., p. 60. 37. Ibid. 38. Ibid. 39. Aristotle put the point succinctly in: Aristotle, Analytica posteriora, in Richard McKeon, ed., The Basic Works of Aristotle (New York: Random House, 1941) p. 110: All insttuction given or received by way of argument proceeds from pre-existent knowledge. 40. David Hume, An Enquiry Concerning Human Understanding, 00. Eric Steinberg (Indianapolis: Hackett, 1977), p. 30. This, incidentally, is a beautiful statement of the psychologistic argument: In a word, if we proceed not upon some fact, present to the memory or senses, OlD' reasonings would be merely hypothetical; and however the particular links might be connected with each other, the whole chain of inferences would have nothing to support it, nor could we ever, by its means, arrive at the knowledge of any real existence. If I ask. why you believe any particular matter of fact, which you relate, you must tell me some reason; and this reason will be some other fact, connected with it But as you cannot proceed after this manner, in in[mitum, you must at last terminate in some fact, which is present to your memory or senses; or must allow that your belief is entirely without fomdation. 41. Gotdob Frege, The FOIIJIdoJions of Arithmetic, p. 4e. 42. See: Goulob Frege, "Somees of Knowledge of Mathematics and the mathematical natural Sciences," in Gottlob Frege, PosthumolU Writings, trans. Peter Long and Roger White (Chicago: The University of Chicago, 1979), pp. 266-274. 43. Frege, The FOIIJIdoJions of Arithmetic, p. 4e.
FRENCH EPISTEMOLOGY
199
44. Frege, Posthumous Writings; p. 205. 45. See, for example: Susan Haack, Deviant Logic (New York: Cambridge University, 1974), p. 29. 46. For further discussion of these }X>ints, see: Nottumo, Objectivity, Rationality, and the Third Realm. 47. See: Frege, '7he Thought: A Logical Inquiry." 48. Of course, there is a sense of 'exists' in which a cure for cancer does not exist- it is the sense in which such a cure does not exist because we do not know of one. But we do not need induction for that 49. Frege, The Basic Laws of Arithmetic, p. 14. 50. See: Nottumo, Objectivity, Rationality, and the Third Realm. 51. See: Charles S. Peirce, ''The Fixation of Belief," in Justus Buchler, ed., Philosophical Writings of Peirce (New York: Dover, 1955). 52. Nottumo, Objectivity, Rationality, and the Third Realm, p. 214. 53. Tom Settle, "A Complex Theory of Truth," in M.A. Notturno, ed., Perspectives on Psychologism (Leiden: EJ. Brill, 1989). 54. Nottumo, Objectivity, Rationality, and tlu! Third Realm, p. 195. 55. See: Settle, "A Complex Theory of Truth." 56. Peirce, ''The Fixation of Belief," p. 12. 57. Thomas S. Kuhn, The Structure of ScientifIC Revolutions (Chicago: University of Chicago, 1962), p. 11. 58. Stanley Fish, ''Interpretation and the Pluralist Vision," TEXAS lAW REVIEW, Vol. 60, No.3 (March, 1982), p. 501. 59. Frege insists that being true is distinct from being believed to be true. As we have seen, the act of believing is different from the object that is believed. But this, of course, is not all that Frege's slogan is meant to convey. It is also intended to say that mere belief in the truth of a pro}X>sition does not consititute a justification of that proposition, that one may believe that a statement is true and be mistaken in that belief. There are, however, difficulties with this fonnulation. It is easy to give examples of something that is false, but is, nonetheless, believed by someone to be true. The problem, however, is that every example of something that is true is, simultaneously, an example of something that we believe to be true. For how else could we propose it as an example of something that is true? 60. Fish, ''Interpretation and the Pluralist Vision," p. 497. 61. Ibid., p. 501. 62. No matter how circwnspect one is in acknowledging that his psychologistically 'justified' belief is tentative and subject to revision, those that follow are likely to be less circumspect, will tend to ignore these disclaimers, and, with time, regard the belief as true come what may.
SOCIOLOGISM JOHN HUND
I. Introduction 'Sociologism' is one of those ugly neologisms that may eventually fmd its way into the pages of the Q.E.D. The reason is that the word names a concept that is fast becoming pivotal in Anglo-American analytical philosophy. The conception I am referring to is the result of at least two streams in contemporary philosophy that seem to be coming together towards something that ' sociologism' seems best to name. What these two streams are to be called presents something of a difficulty. Let us provisionally call them 'anti-psychologism' and 'social realism', or what may be called 'social Platonism' in the philosophy of mathematics. In this note I propose to discuss two radically different versions of sociologism that are taking hold of, or influencing, philosophy today. One is an empiricist version that I attribute to Wittgenstein and the Wittgensteinians, but especially to Bloor. 1 The other is a social realist version that I attribute to such thinkers as Popper, but especially to Durkheim. In order to set the ground for my discussion I will have to describe and name two versions of psychologism that empiricist philosophers have had a habit of falling into: the fIrst is the attempt to reduce abstract, social or cultural phenomena - of varying degrees of specificity, from the most determinate to the most diffuse - to mental phenomena found in the minds of concrete individuals. The second psychologism, which has often been a reaction against the fast, is the attempt to reduce abstract cultural phenomena to patterns of behavior, often sanctioning behavior, that can be detected in the visible actions of individuals that an external observer could record. I will call these two psychologisms 'mentalism-psychologism' and 'behaviorism-psychologism' to distinguish them, respectively, from those two slippery impostors, mentalism and behaviorism. What is of greatest interest is not what distinguishes these two species of psychologism, but what it is they have in common. What they seem to have in common is an empiricist foundation that generates a nominalist impulse to reduce abstract phenomena2 to things (mental an, where (i) S is a non-empty set, - is a function from S into S, and n one from S x S into S such that AI'. For any A and B in S, A n B = B n A, A2'. For any A, B, and C in S, A n (B n C) = (A n B) n C, AJ'. For any A, B, and C in S, A n -B = C n -C if, and only if,
AnB=A, A4'. For any A and B in S, if A = B, then -A = -B, and Aj'. For any A, B, and C in S, if A = B, then A n C = B n C, and (ii) P is a function from S into R such that B1. For any A in S, 0 ~ P(A), B2. For any A in S, P(A n A) = 1, and B3. For any A, B, and C in S, if A n B = C n -C, then P(-(-A n -B» = P(A) + P(B).
Illustration 3: Let S consist of the two sets (2) and (a), a whatever you please. Then, as in Illustration 1, 0 and {a) are each other's complements, o is the intersection of 0 and {a}, P(0) must equal 0, and P({a}) equal 1. Illustration 4: Let S consist of some atomic statement or other A, the negation B (= -B) of any statement B in S, and the conjunction B n C (= B & C) of any statement B and any statement C in S. Then S partitions into four so-called equivalence classes: SI consisting of A and all its logical equivalents in S, S2 consisting of the negations of the members of St, S3 consisting of A n -A and all its logical equivalents in S, and S4 consisting
POPPER'S PROBABILITY THEORY
345
of the negations of the members of S3. For any statement A, in S3 P(A3) must equal 0, for any statement At in S. P(At) equal 1, and for any statement Al in 51 and statement A,. in 52 P(AI ) and P(AJ equal any two reals in the interval [0,1] whose sum is 1. Note by the way that the set S in a Quadruple , where S is a non-empty set, - a function from S into S, n a function from S x S into S, and P a function from S into R such that E1. E2. E3. E4. E5. E6.
For any A, B, and C in S, P(A n B,C) 5. P(A,C), For any A and BinS, P(A,A) = P(B,B), For any A and BinS, if P(C,B) :F- P(B,B) for some C in S, then P(-A,B) = P(B,B) - P(A,B), For any A, B, and C in S, P(A n B,C) = P(A~ n C) x P(B,C), For any A, B, and C in S, if P(A,D) = P(B,D) for every D in S, then P(C,A) = P(C,B), and For at least one A, one B, one C, and one DinS, P(A,B) :F- P(C,D).
Like Popper's 1955 account of an absolute probability field, the foregoing account of a relative one is a tour de force. But again Popper's determination not to use the Lower Bound Law as a constraint, his use in
POPPER'S PROBABILITY THEORY
355
constraints E2 and E3 of 'P(B,B)' instead of '1', and a further concern of his which I discuss below make for constraints that are perhaps not as "pointed" as could be.:n In any event many prefer these constraints, which in the frrst six cases were already known to Popper and in the seventh soon occurred to him: Fl. F2. F3. F4 F5. F6. F7.
For any A and BinS, 0 So P(A,B), For any A in S, P(A,A) = 1, For any A and B in S, if P(C,B) 1 for some C in S, then P(-A,B) = 1 - P(A,B), = E4= For any A, B, and C in S, P(A n B,C) = P(A,B n C) x P(B,C), For any A, B, and C in S, P(A n B,C) So P(B n A,C), For any A, B, and C in S, P(A,B n C) So P(A,C n B), and For at least one A and one BinS, P(A,B) 1.
*
*
Constraints Fl-F2 and F4 are of course familiar from Renyi's account; constraint F3 follows from (but, note, is not equivalent to) Renyi's constraint D3; constraints F5-F6 follow from constraint AI' by the Substitution Law for relative probability functions; and constraint F7, also familiar from Renyi's account, is a simplification of constraint E6. As indicated earlier, Popper went on to prove in Appendix *v of The Logic of Scientific Discovery that, given this definition of '=': A = B
=cIf
P(A,C) = P(B,C) for every C in S,
postulates for Boolean algebra equivalent to AI' -Aj' and the Substitution Law for relative probability fWlctions follow from constraints E1-E6 (so, from their equivalents FI-F7 as well), and hence that Boolean algebra, far from being a prerequisite to relative probability theory, can in point of fact be viewed as an outgrowth of it. The postulates he utilized for the occasion were postulates of Huntington's in "New Sets of Postulates for the Algebra of Logic," lRANSACTIONS OF THE AMERICAN MATHEMATICAL MONTHLY, vol. 35, 1933, pp. 274-304. He also noted that the relation designated by '=' is reflexive, symmetrical, and transitive- hence, an equivalence relation and, more specifically, one of substitutional equivalence. Having established a point critical to him, the "autonomy" from Boolean algebra of relative probability theory, Popper further insisted that the constraints placed on P be "autonomously independent," i.e., independent in the presence as well as the absence of Boolean constraints such as Ai' -AS' . Constraints F5-F6 are of course not autonomously independent: as just noted, they follow from the Boolean constraint AI' by the Substitution Law for relative probability functions, and hence are in Popper's eyes vestiges of Boolean algebra. He may have overlooked, however, that certain accounts of Boolean algebra use the inclusion relation c in lieu of the identity one =,
356
HUGHES LEBLANC
and that his own constraint El would follow by dint of this Substitution Law for relative probability functions:
If A c B, then P(B,C) =5 P(A,C) from the constraints placed in such accounts on the two functions - and n. There may well be some substitute for constraint E1 that would make for an "autonomously independent" as well as "autonomous" account of Popper's relative probability functions, but some of us already alerted by Popper to the autonomy of probability theory from Boolean algebra do value the simplicity and directness of constraints FI-F7. 32 As remarked three paragraphs back, Popper's constraint F3, though following from Renyi's constraint D3, is not equivalent to and hence is weaker than it. So, all of Renyi's relative probability functions figure among Popper's, but not vice-versa. Indeed, Renyi's functions are those and only those of Popper's that meet this extra constraint: (1) If P(A,B) = 1 for every A in S, then B = C n -C for some C in S, or equivalently this one: (2) If P(A,B) = 1 for every B in S, then A -(C n -C) for some C in
S.33
=
Because of (2), a statement A of our language L proves under Renyi's account to be truth-functionally true (i.e., again, a tautology) if, and only if, P(A,B) = 1 for every statement B of L and any relative probability function P. Under Popper's account, on the other hand, A proves to be truth-functionally true if, and only if, P(A,B) = 1 for every statement B of L and every relative probability function P. (1) is also of interest. A statement A of L is sometimes called P-absurd if P(B,A) = 1 for every statement B of L, i.e., if - to phrase it in belief-theoretic terms - should you believe A you'd believe anything. Under Renyi's account only truth-functional falsehoods are P-absurd; under Popper's, on the other hand, truth-functionally indeterminate statements of L (Le., statements of L that are neither truth-functionally true nor bUth-functionally false) may also be P-absurd. Popper's constraint E6 and, more pointedly, constraint F7 reassure us, however, that there exists for each relative probability function of Popper's at least one statement of L that is not P-absurd. Note, by the way, that under Popper's account the state descriptions of L in (n = 1,2,3,...) have as their probabilities relative to -(B & -B) (B here any statement of L) non-negative reals whose sum is 1; under Renyi's they also have as their probabilities relative to -(B & -B) non-negative reals whose sum is 1, but because of (1) they themselves cannot be P-absurd for any P; and, when Renyi's functions meet the extra constraint D6, the state-descriptions in question have as their probabilities
POPPER'S PROBABILITY THEORY
357
relative to -(B & -B) only non-zero reals whose sum is I (and they themselves cannot be P-absurd f(X' any P). It was R6nyi's stated aim to construct relative probability functions that relativized Kolmogorov's absolute ones. This required, as R6nyi puts it on p. 286 of "On a New Axiomatic Theory of Probability," that the latter be "special cases" of the f where F has but one member: following a suggestion of Rosenbloom's on p. 10 of The Elements of Mathematical Logic, let E be {a}, F be {{a}}, {-a} be {a} itself, {a} n {a} be as expected {a}, and P({a}) be 1, and the trick is done. To rule out oddities of this sort, Mendelson in Introduction to Mathematical Logic (Monterey: Wordsworth & Brooks, 1987), p. 8, further demands of a field that it have at least two members. Kolmogorov need not do so: because of constraint V, F is sure to contain 0 as well as (the non-empty) E. See Notes 17, 27, and 28 for more on this point Incidentally, A n A = A by constraint ..43 and A n -A = A n -A, and hence A n -A = B n B by constraint A3 again. So, constraint V could read, more compactly, For any A and B in F, if A n B = -A n ~ then P(-(-A n -B» = P(A) + P(B). But writing 'A n B = C n -C' rather than 'A n B = A n -A' (or, for that matter, 'A n B = B n -B') forestalls the occasional question: "Is A (B) the same set in A n -A (B n -B) as in A n B?" 8. Kolmogorov's choice of the letter 'E' to denote that set was surely suggested to him by the Gennan word 'Ereignisse' for events. 9. Possibly misled by Kolmogorov's phrasing of things, many alas have inferred that it does. Kolmogorov might have prevented the misunderstanding by postponing until Section 2 mention of elementary and random events. 10. For an elementary exposition of these matters see: Kemeny, Snell, and Thompson, Introduction to Finite Malh£matics (Englewood. Cliffs: Prentice-Hall, 1956), cbs. 3 and 5. The set consisting of all, and only, the state-descriptions in which a statement A of L holds is called there the truth set of A. Note that when E in a Kolmogorov quintuple <E,F,-,n,P> consists of statements and hence F of sets of statements, these sets must be construed disjunctively rather than conjunctively-{(...(AI V AJ V...) V Au} being of course the disjunctive construal of the set of statements {AI ,A2, •••,Au} and {(...(A1 & AJ & ...) & Au} its conjunctive one. Disj\D1ctively construed, the set consisting of all the state-descriptions of L would be of absolute probability 1 as in Kolmogorov's constraint V; conj1.U1ctively construed,
364
HUGHES LEBLANC
on the other hand, it would be of absolute probability 0, and so would be the truth set of any statement of L. For more on state-descriptions see p. 366. 11. Popper would say fomwl or uninterpreted where I say abstract. But, as formulated on pp. 341-342 or pp. 342-343, Kolmogorov's account is formal in the cmrent sense of the word, and it is interpreted only in that the members of F and hence arguments of P have to be sets. So, 'abstract' (or, possibly, 'generalized') would seem the more fitting appellation for the present account 12. Kolmogorov understood by = in constraint I the identity relation. Under that construal the SldJstitKlion Law holds by the very definition of a function: taking each of its arguments into exactly one real, P must obviously take B into the same real into which it takes A if B is the same as A. However, when = in constraint I is lDlderstood as a mere equivalence relation, the SldJstitution Law must be derived from the constraints on P or must be listed among those constraints. See p. 345 and pp. 346-349 for more on this matter. 13. Constraint C3 is of course tantamO\D1t to For any A and BinS, P(A n B) = P(B (1 A). As for For any A, B, and C in S, P(A (1 (B (1 C» = P«A (1 B) n C) and For any A in S, P(A (1 A) = P(A), they follow from constraints C3-C4 and constraints Cl and CS, respectively. 14. Some logicians, on the other hand, have paid considerable attention to the account, as was noted on earlier and will be documented in Part N of the paper. 15. Proof of constraint Cl - given constraints Cl' and C6 - will be found on p. 348; and, more generally, proof that - given constraints C3-C6 - constraints Cl-C2 and Cl' -C2' are equivalent will be found in my "Popper's 1955 Axiomatization of Absolute Probability," PACIFIC PIDLOSOPIDCAL QUARTERLY, vol. 63, 1982, pp. 133-145. Robert Stalnaker in "Probability and Conditionals," PHILOSOPHY OF SCIENCE, vol. 37, 1970, pp. 64-80, uses constraint Cl' in lieu of constraint Cl and this special case of constraint C6: C2". For any A in S, P(A) = 1 - P(A) in lieu of constraint C2. Proof of constraint C2" given constraints Cl' -C2' and C3-C6 will be found on pp. 348-350. 16. See: Popper, ''Two Autonomous Axiom Systems," pp. 53-54. 17. In ''Two Autonomous Axiom Systems," p. 53, note 1, Popper indicates that constraint C2 "may be split into two": C2.l. For any A in S there is a BinS such that P(B) ~ P(A) and P(A (1 B) = P(A) x P(B), and C22. There is an A in S and a BinS such that P(A) P(B). Constraint C2.1, however, is hardly more intuitive than constraint C2; and, in my opinion at any rate, constraint C22 does not really belong in an "autonomous" set of constraints on the function P since it is equivalent to a constraint on the set S, to wit: C22'. There is an A in S and a BinS such that A B. 18. In the case of C6 the reader should use B3, the Substitution Law, and these two consequences of Al' -AS': (A (1 B) (1 (A n -B) = C (1 -C and -(-(A (1 B» (1 -(A (1 -B» = A. 19. It follows from Rescher's example that '&' is not extensional in probability contexts. For further details on the probabilistic extensionality of '-' and non-extensionality of '&' and the quantifier operator 'V', see my "A New Semantics
*"
*"
POPPER'S PROBABILITY THEORY
365
for First-Order Logic, M-any-Valued and Mostly Intensional," TOPOl, vol. 3, 1984, pp.52-62. 20. Lemma 1 is of course Popper's constraint CI, and Lemma 5 Stalnaker's constraint C2". 21. The theorem in question is 431 on p. 234, one I shall again appeal to. It says that if A is provable in L, then P.(A) = 1. But by the Weak Completeness Theorem A is provable in L if truth-functional true. The subscript's' signals that P here takes statements as its arguments. 22. The theorem in question is 4.30 on p. 233. 23. Gotten from what precedes by frrst paraphrasing 'A == B' as '(A:::> B) &. (B :::> A)' and 'A :::> B' as '-(A &. -B)', and then writing '-A' for '-A' and 'A n B' for 'A &. B'. 24. Peter Roeper has pointed out to me that my defmition of '=' can be simplified to read A = B =. P(A n -B) + P(-A n B) = 0, ~ hence, A = B =. P(-(A n -B» x P(-(-A n B» = 1. It is of course because I borrow from "Alternatives to Standard First-Order Semantics" that the present proof of Theorem 3 is relatively brief. Any reader wanting a self-contained proof of the theorem should use Roeper's simplification of my definition of ':'. 25. Some mathematicians require that = in the constraints in question be construed as identity; many others - Rosenbloom among them - allow = to be an equivalence relation. To quote (in slightly edited form) from p. 9 of his book: "Here the relation "=" is taken to be part of the known syntax language. The only properties of this relation which will be used are Tl.l.l (i.e. A = A), Tl.1.2 (i.e. If A = B, then B = A), Tl.1.3 (i.e. If A = B and B = C, then A = C», and their consequences in conj\Dlction with AI' -AS'. Hence we could alternatively take "=" as an undefined term and postulate Tl.l.1-Tl.l.3. A relation satisfying the latter conditions is called an equivalence relation." The alternative suggested in the third sentence is one I cannot pursue here. 26. Statement A &. B is logically equivalent to statement B &. A, but is it the same statement as B &. A as constraint AI' would require? Or, to pass on to constraint B3, could A &. B possibly be the same statement as C &. -C for atomic A, B, and C? 27. In his 1955 paper Renyi places no restriction on G: he merely notes that in view of the contradiction reproduced in the text (2J does not belong to G, a curious handling of the matter. Restrictions ill and IV are from his "Sur les Espaces Simples des Probabilites Conditionnelles," ANNALES DE L'INSTITUT POINCARE, N Serle, Section Bl, 1964, pp. 3-21. Since F contains 0 but G is a non-empty subset of F that does not, F is sure to have at least two members. 28. This constraint: D3'. For any A and BinS, if B :I: C n -C for some C in S, then P(-A,B) = 1 P(A,B), equivalent to Renyi's constraint D3, is sometimes used in place of D3. It brings Renyi's account closer to that of Popper's on p. 354 and that of von Wright's in The Logical Problem of Induction, second revised edition (New York: Macmillan, 1957). To preserve consistency von Wright requires that A in constraint D2 rather than B in constraint D3' be distinct from 0. Few have followed him in this. Constraints DI-D4 are compatible with S having but one member. Hence D5, a constraint which will prove of considerable interest in Part IV.
366
HUGHES LEBLANC
29. Carnap's constraints in the Continwun book are a statement-theoretic version of the Substitution Law four lines further in the main text, a variant of D3, D4, and this (needlessly strong) version of D1: o So P(A,B) S. 1. The constraints are compatible with P having 0 as its one and only value, a slip I reported to Camap shortly after publication of the book. To mend matters, I suggested adoption of this further constraint: If -(B & -A) is truth-functionally true, then P(A,B) = 1, to which Carnap assented. Adoption of Renyi's constraint D2 would have done as well, but I do not recall either of us realizing it at the time. Camap's constraints in the Logical FolUJdations book are the Substitution Law, the version above of D1, the same variant of D3, pillS in effect these two constraints: H A is truth-functionally true, then P(A,B) = 1 and H A is a state-description and B is a truth-functional truth of L, then P(A,B) > o. It is of course the last of Camap's constraints that is equivalent to D6, the way this constraint: If A is a state-description of L, then P(A) > 0, would be equivalent to C7'. 30. For some reason 1956 is given by many, Popper included, as the publication year of Mace's book. Yet my copy of it says "First Published in 1957" on the back of the title page. 31. The concern in question, we shall see, denies him use of F5-F6 below as constraints on P, and ironically enough should have denied him use of E1 as well. 32. For further information on the matter of autonomy, see my paper "The 'autonomy' of probability theory: Notes on Kolmogorov, Renyi, and Popper." 33. More exactly, Renyi's functions are those only among Popper's that happen to be defUled on fields (in the identity sense) and meet, say, constraint (1). By not requiring of his set S that it be a field (in this sense) Popper acknowledges relative as well as absolute probability functions that more traditional writers such as Kolmogorov and Renyi would not 34. Given a more general algebra than Boole's the result can be made to suit all first-order statements, hence the logic of qlUJ11tiflCationai as well as truth-functional derivation. 35. At least for customers who want their logic a byproduct of absolute probability theory. The result from my "Popper's 1955 Axiomatization of Absolute Probability" is generalized in Section 5 of "Alternatives to Standard First-Order Semantics" to cover all first-order statements. 36. The result hinges on P(A,B) amoWlting by constraint F4 to P(A ("\ B,V)IP(B,V) when P(B,V) -¢ o. 37. I lDlderstand of course by glb{P(S'):S' is a finite subset of S} the greatest lower bound (or infunwn) of the set consisting of each P(S') such that S' is a finite subset of S. Since L has only denwnerably many statements, any set of statements of L is sure to be denumerable. The defmition of P(S) here and that of P(S,B) were recommended to me by Kent Bendall. The second defmition is equivalent to one by Field in "Logic, Meaning, md Conceptual Role," THE JOURNAL OF PHILOSOPHY, vol. 74, 1977, pp. 379-409. 38. The proofs in question will be found in: William H. Harper, "A Conditional Belief Semantics for Free Quantification Logic with Identity." ESSAYS IN EPISTEMOUXJY AND SEMANTICS (New York: Haven, 1983), pp. 79-94; Field, "Logic, Meaning, and Conceptual Role"; Hugues Leblanc, "Probabilistic Semantics
POPPER'S PROBABILITY THEORY
367
for First-Order Logic," ZEITSCHRIFr FUR MATHEMATISCHE LOGIK UND GRUNDLAGEN DER MATHEMATI~ vol. 25, 1979, pp. 497-509; and Section 5 of Hugues Leblanc, "Alternatives to Standard First-Order Semantics." A Soundness and Completeness Proof already appeared in Harper's doctoral dissertation, COunlerfactlUlls and Representations of Rational Belief, University of Rochester, 1974, but it concerned conditional logic rather than standard logic. 39. The inference here is from a statement of the form '('v'x)(F(x) ::> p)' to one of the form '(3x)F(x) ::> p'. 40. The reader was prepared on p. 356 for the occurrence of 'any' rather than 'every' in these variant results. 41. For a last comment on Carnap's and Renyi's probability fimctions see note 42, where I exploit a point made in the next paragraph of lhe main text. 42. In a letter of 1979 Popper remarked that probabilistic semantics takes up some suggestions which he had published in an Appendix to The Logic of Scientific Discovery, suggestions which he originally intended to develop himself but never found time for. Note that since (i) the sets on which Camap and Renyi defme their respective probability fimctions are fields and hence meet constraints Al' -AS' and (ii) with '-', '&', and ':=' substituting for '-', 'n', and '=', respectively, constraints Al' -A5' allow derivation of the formula A:= -(C &-C) for any truth-functional truth A of L, constraints Al' -AS' deliver by dint of this defmition: A is truth-fimctionally true =df A := -(C & -C), all the truth-functional truths of L. So, Camap's and Renyi's probability fields make for two accounts of truth-fimctional truth, the foregoing one which exploits only constraints Al' -AS' and the one on p. 361 which exploits constraints Bl-B4 or Dl-D4 as well as constraints Al' -AS'. The result robs the latter account of much of its point. Why bother doing probabilistic semantics when a Boolean account of truth-functional truth can be had from constraints Al' -A5' alone? In point of fact, why bother at all with probability functions and the constraints they must meet to count as Camap functions or Renyi ones? Popper's relative probability fields make for jlUt one 8CC01D1t of truth-functional truth as he saw, and so do his absolute probability fields as others saw. 43. See: Hugues Leblanc and Peter Roeper, "Absolute Probability Functions: A Recursive and Autonomous Account," The Tasks of Contemporary Philosophy (Vienna: Holder-Pichler-Tempsky, 1986), pp. 29-49. For formulations and uses of a probabilitic semantics that employ probability relations rather than probability functions, see: C.G. Morgan, "Weak Comparative Probabilistic Semantics," ZErrSCHRIFf FUR MATHEMATISCHE LOGI~ Vol. 30, (1984), pp. 199-212; and John Serembus, Absolute Comparative Probabilistic Semantics, doctoral dissertation, Temple University, 1987.
PHYSIOLOGICAL FOUNDATIONS OF OUR KNOWLEDGE OF THE MATHEMATICAL UNIVERSE J.N. Hattiangadi
I. INTRODUCTION Hume suggested that human knowledge is based on custom or habit, and therefore proposed to found all knowledge on a science of human nature.! It seems that the teon 'psychologism' was only applied to this doctrine in the nineteenth century, by Fries and Beneke.2 At that time there was also being developed an account of all human knowledge in terms of socially developed categories, by Hegel, for example, in a relativised version of Kant's point of view, which we may call 'sociologism'.3 The two crucial areas for sociologism and psychologism to investigate are the philosophy of mathematics and the empirical basis of knowledge, as I will describe these below.4 The point of view that I will outline will not try to reduce everything to biology, but will in crucial places be 'biologistic'. Such an approach, which becomes possible only after Darwin, Mendel, and the development of modem physiology, does, I think, what psychologism and sociologism could only promise. Having agreed to write on these subjects, I had prepared a paper which carefully skirted around the issues and avoided the topic of the philosophy of mathematics as much as possible. The philosophy of mathematics is an intimidating subject. Formal metamathematics, combined with the abstruse philosophical arguments current in the field, makes it difficult to study. And, to make matters worse, I found that I could not endorse any of the philosophies of mathematics that I was willing to discuss. One night I could not fall asleep, because I discovered that for the last ten years, since certain investigations of mine regarding the empirical basis of knowledge had been completed, I did have a philosophy of mathematics, without knowing that I did. Thinking about why it was that I did not know what should have been clear to me, I realized that I was unable to articulate my belief even to myself because it is a well known belief which is easily ridiculed. On the following day, I was still surprised to hear myself admit that I am a Pythagorean, even though I had already decided that I am one, and, in fact
PHYSIOLOGICAL FOUNDATIONS
369
have been one for over ten years.' I will have to postpone for another time the analysis of the extraordinary situation where one subscribes to beliefs concerning a subject which he holds to be of central importance and to which his thoughts tum regularly, without being aware of those beliefs. This paper is written to explain why I am a Pythagorean, and not a modernist, where philosophy of mathematics is concerned What I have to say, however, is neither very complicated nor as fonnal as those studies which commonly fall under the heading 'philosophy of mathematics'. I do, moreover, state my views with far more confidence than I feelconfidence, that is, not in the truth of what I have to say, which I do have, but in my ability to marshall and present the facts in a field that I have always held in much awe. The modified Pythagoreanism which I shall try to defend may be summarized as the doctrine that the universe has a mathematical structure. Geometrical and arithmetical truths, I believe, are not truths about intuition. Nor are they analytical. They are objective truths (or more usually falsehoods) about the physical universe, just as this is intended in the writings of Galileo or Descartes or Newton. When Descartes maintained, for example, that matter is essentially extension, which is space, he meant this quite literally. For Descartes, as for GallIeo before him, and for Newton after him, the world that the physicist describes is a mathematical structure, and its fundamental laws are mathematical laws of change, or motion. I am a Pythagorean neither in claiming that Pythagoras is right, nor in claiming that Gallleo is right, regarding the details of the mathematical structure of the Universe. On the contrary, their views have been superseded by the modem statistical conception of the universe. Rather, my Pythagorean claim is merely part of what Pythagoras proposed, which Galileo rediscovered, that the world itself has a mathematical structure- even if we are not sure which mathematics exactly it is that describes our world. But whichever mathematics it is, as we improve our understanding of the world since GaIileo's timet 'the features of the universe seem to be more certainly mathematical than ever, however uncertain we become about its details. This thesis about mathematics will be presented in three different categories, each of which has one central thesis. First of all there is a historical thesis in which I will attempt to fonnulate my view more clearly by contrasting it with other views from the philosophy of mathematics, and show how it came about that the correct view was abandoned in favor of one or other of these views. Because I am here primarily concerned to introduce my claims in the second and third theses, I make reference to a number of historical opinions - which I take to be facts - without citing texts or elaborating upon the various details of the arguments in their support. The facts alluded to are extremely well known, and were I to be thorough, this paper would be less accessible. So I shall leave for another occasion the full description of sources and state the thesis as follows. The scientific conception of the physical world from Galileo to the time of Newton and Locke was one in which the world is analyzed as an objective
370
J.N. HATIIANGADI
mathematical structure. Leibniz proposed a new intetpretation of mathematics which we may call 'idealist', in one good sense of that word. There followed immediately after this an empiricist critique of the Newtonian conception of the world, a critique from which 'the scientific view never fully recovered. Leibniz's beliefs seem to have triumphed on the question of the interpretation of mathematics. But my own investigations into empiricism have led me to revaluate the empirical basis of knowledge, which brings me to the second thesis. This thesis is about foundations, about how one fmds out about the world at all. 6 It is this thesis which provided the central motivation for writing this paper, for this is the thesis to which the intriguing phrase 'physiological foundations' refers. From a modem biological and physiological perspective, our perceptual appamtus cannot be what the British empiricists had thought it was, namely, a passive way of receiving impressions of the world. Rather, perception is closely related to success in action, and it is through our ability to do things that we come to know the world. This is the true foundation of human knowledge, and particularly of our knowledge of the mathematical universe. The third thesis is a description of mathematics, and our knowledge of it. According to 'this thesis, the best conception of mathematics is the one which was abandoned in the eighteenth century. The universe is mathematical. The modern twist is that we know this through our successful physical actions. This view has neither the formal elegance nor the philosophical subtlety which characterizes the so-called 'philosophies of mathematics' which I attempt to discredit in my historical thesis. But, for all its simple-mindedness, I believe that what I say is true, which is my only excuse for saying it at all.
ll. Some Historical Notes My historical thesis is this: all the issues which arise in the philosophy of mathematics are the result of a clash between a certain model for empirical knowledge, on the one hand, and the scientific conception of the world as a mathematics of motion, on the other. The clash arose because Newton claimed, with considerable merit, that his physics, though a 'mechanics' in the sense that Galileo had envisaged, was founded on experiments, or was 'empirical' .7 Interpreting empiricism as a kind of perceptual awareness, which is a model of knowledge of great antiquity, Locke tried to distinguish between the primary (mainly mathematical) properties that belong to things properly and those secondary properties that create impressions in the mind of the perceiver" This Buddhist conception of knowledge as a kind of mental awareness of what is the case, when modified to become empiricism, was the undoing of Newtonian philosophy, at least as a united system. Of course, Newton's physics continued to have widespread influence. But what came to be called 'Newtonian Physics' bore less and
PHYSIOLOGICAL FOUNDATIONS
371
less resemblance to Newton's original thought on the subject, and, for that matter, to those of Galileo, Descartes, Huygens, or Locke. The sharp contrast between philosophy and science that has since prevailed is but one part of the attempt to pull Newton's teeth. Indeed, a study of all the different philosophers who have called themselves 'Newtonians" and the views that they have held, would yield an amusing story. But none of these stories is stranger than that of the philosopher, David Home, who undennined the entire basis of Newtonian physics with his sceptical arguments, and still called himself a Newtonian; and, it may be argued, rightly so! Arriving at this conclusion requires an interesting analysis of eighteenth century thought. For this I am grateful to Dr. Michael Haynes who has allowed me to see his unpublished writings on the subject, to which I cannot do justice in the brief sketch which follows regarding Home, the Newtonian. Haynes contends, quite astutely, that the history of philosophy of the eighteenth century cannot be studied without paying due attention to the multifaceted dispute between Newton and Leibniz. Apart from the differential and integral calculus which they both claimed to have discovered, they seemed to disagree on nearly everything else of importance-- or is it that what has come to be important to us is determined by what they could not agree upon? Newton was a materialist and an empiricist. He believed in an immanent God, in absolute space and time, and thought that logic, the tool of scholastics, was a sterile instrument. Leibniz was an idealist and a rationalist. He believed in a transcendent God, in relative space and time, and thought all truths derivable from logic. Although eventually everybody accepted Newton's system of the world as the correct scientific system, there were many who were convinced of the truth of certain other things that Leibniz maintained, which may be called things of a 'philosophical sort', such as that individuals are the only substances, or that space and time are relative. Dr. Haynes suggests that it is against this background that we must interpret Berkeley's critique of Newton's theory of gravity, of fluxions, his rejection of matter, and his ontology of a community of spirits and their percepts, as a vindication of Leibnizian thought As a critic, Berkeley turns empiricism itself against Newton, claiming that we could not know the mathematical world which lies just beyond our perceptual ken if indeed all that we can rely upon is our sense experience. While Berkeley denied the theory of matter, and the absolute reality of space and time, he did not deny the existence of objects of common sense, or of spirits, or of God. The power of Berkeley's argument as a critique of Newton's view is undeniable. If we fmd out about the world solely from our experience, how could we possibly fmd out that it is unlike our experience in any respect whatsoever? Newton's claim to have described a mechanics whose features went beyond the content of our experience, but which was nevertheless a product of induction from experience, seems to be a claim that undermines itself. On the basis of this critique, Leibniz's monads come off very well.
372
J.N. HA'ITIANGADI
It was Hume's discovery that if we apply Berkeley's critical method in a thoroughgoing way then it necessarily undermines not only matter and the reality of space and time, but also common sense objects, spirits, and even God. Hume's argument takes the form of a tu quoque: he shows that no one can do any better than Newton. He is a Newtonian in the strange sense that since there are no good reasons for accepting any philosophical point of view, he is free to adopt the Newtonian point of view without rational
ground. The rise of modem philosophy of mathematics, then, is a retreat from mathematical theories of motion, a cautious reinterpretation designed to make mechanics appear to be other than what it is. Both the great opinions on mathematics later in the eighteenth century, namely those of Hume and Kant, are Leibniz-like even while they claim to be Newtonian. Hume suggests that mathematics is empty, whereas Kant asserts that it is a feature of how the world must be conceived, which may or may not be as Newton describes -it. One makes Newton's claim innocuous, the other less than objectively true. It seems to have been settled that we cannot live in a world where change itself has a mathematical structure, quite independently of language or of perceptions. Kant's suggestion is that the world appears to us as Newtonian because this is the only way that objects can be perceived, or even contemplated. Kant thought that mathematical truths are universal because they are culled from a distilled set of experiences- arithmetic from the pure intuition of time, as he called this distilled experience of temporal phenomena, and geometry from the pure intuition of space. The idea that we necessarily contemplate space in one particular way was undermined ,by the discovery of alternatives. But the discovery of Non-Euclidean geometries was met by the claim that there is a certain study of a more abstract sort called 'projective geometry' which specified the only possible geometries, of which Kant's was one. This was the defense initially proposed by Bertrand Russell.' Later, Russell abandoned this Kantian view, noting that the kind of geometry that Einstein proposed for use in the general theory of relativity in 1916 was of a sort that he had suggested was impossible in his youthful defense of Kant lO So apart from details, he came to believe that there was nothing in his youthful defense of Kant which was correct Russell's ideas regarding the foundations of mathematics changed dramatically when he discovered modem logic, as Frege's also did, independently. 11 Russell and Whitehead later endeavored to derive mathematics from logic, together with a theory of classes. This seriously undermined the claim that arithmetic is derived from the pure intuition of time and geometry from that of space!2 Neither seems to be necessary for the derivations in Principia Mathematica. Today, two defenses of Kantian thought remain today, neither of which is satisfactory. The [rrst of these attempts to limit what is legitimate in mathematics, while the second proposes to give an expanded version of intuitions. Brouwer proposed that the intuitions which lay at the basis of
PHYSIOLOGICAL FOUNDATIONS
373
mathematics are those concerning provability .13 On this basis it is possible to say that certain of the derivations used by mathematicians are satisfactory, but not others. Equating mathematical truth with what is provable, in a manner that conforms with intuitionism, leads to the familiar trichotomy of intuitionistic logic, according to which there are true statements (which are proved), false statements (which are disproved) and statements which are neither true nor false (which are the undecided statements). No doubt, the attempt to derive more from fewer assumptions leads to interesting mathematical derivations. If a mathematician should say that this is what he wishes to do, there is no reason why he should not But as a philosophy of mathematics it is unsatisfactory because it does not describe why some of that mathematics which is not constructivist, or not properly founded according to this criterion, works, whereas some of it does not Consequently, it is not satisfactory as a philosophy of mathematics, however nice it may be as a way of doing mathematics. Another way to defend intuition is Poincare's, who suggested that a most abstract study of space is possible, without regard to any restrictions except those that all spaces must have. This is the science of topology. Of the several different ways in which one could study spaces, he conjectured all would give equivalent results. This turned out not to be true when, in 1960, John Milnor showed that in the case of the seven sphere there are divergences. Later it was shown that the axiom of choice and the continuum hypothesis are independent of set theory, suggesting alternative set theories, and creating difficulties for any form of intuitionism that hopes to take all of mathematics into account. Since neither attempt described here can explain all of mathematics, we must conclude that mathematics is not founded on intuition, though it is interesting to note that some of it can be so founded. I take Formalism to be a constructive attempt employing the techniques of metamathematics to found all of mathematics on intuitively satisfactory principles. As was shown independently by Tarski and GMel, a metalanguage which is adequate to this task is richer than the object language. 14 This effectively disposes of the Formalist approach of doing more with less through a process of metalinguistic ascent Once Hilbert's Programme, as this ascent is called, has been abandoned, there is little plausibility in regarding mathematics as a formalism, since we cannot thereby show why the formal structures a mathematician studies, such as number, are useful in such wide domains as accounting, crystallography, and music. To say that a mathematician studies abstract structures, or formalisms, is true some of the time, but cannot be all the truth. There is some truth to the claim that the structures of pure mathematics are merely formal, as we shall see. It is the dependance upon intuitions that I take to be the difficulty for Formalism. Most philosophers in this century have been logicists. They have held that Home was basically right, that mathematics is empty, although their reason for believing this is better than Hume's, namely that mathematics is
374
J.N. HATIIANGADI
derivable from logic. That logic must have a theory of classes added to it in order to derive mathematics, and that this theory is beset by antinomies, can be met with the hope that they are, or at least will be, resolved. But logicism still fails because GOdel showed, using a quasi-antinomy, that if every arithmetical truth is provable in Russell's logical system (understood to include a theory of classes or sets) then arithmetic is not consistent. GMel's theorems on the incompleteness of arithmetic show that logicism cannot be true, because the assumption that arithmetical truth and provability are equivalent is not a doctrine that is consistent with arithmetic. is Platonism has recently gained in popularity in the philosophy of mathematics, mostly by default, and fmds favour among many mathematicians. 16 Plato believed that mathematics was an understanding of the eternally immutable reality which underlies the comparatively unreal world of change. Galileo's theory of a mathematics of change or of motion itself is inconsistent with such a conception of mathematics, postulating, as it does, an unchanging universe. 17 Indeed, the very possibility of a mechanics is ruled out in Plato's conception of reality. Yet, most of those who claim to be Platonists today would balk at the doctrine of the Timaeus that the world is constructed out of mathematical elements- the line and the circle. It seems impossible to modern Platonists that the real world of stuff should be made up out of these ideal elements. This implies that the doctrine of mathematical objects is about objects in the less than real world of mind. One may, indeed study number as an 'abstract object' in the same way that Arthur Conan Doyle studied Sherlock Holmes, or as Agatha Christie studied Miss. Marple's character. If we are to take mathematicians at their word regarding the infmitesimal, we fmd that it did exist once. It ceased to exist after the discovery of the epsilon delta method. But it exists again now that we have the model theory of Abraham Robinson. Mathematical objects seem to depend for their existence on mathematicians. The usual description of such objects is to call them fictions. It seems that modem Platonism would suggest then Lltat abstract mathematical objects are fictions. (It seems that the word 'abstract' is only a euphemism for the word 'fictional'.) But if this is what mathematics is all about, then it is astonishing that it can be applied to reality in so many ways by so many sciences in the last three centuries. Why novels, poetry, and short stories do not tell us as much as mathematical physics about physical reality (as even some Romantics concede) is a mystery. Fictionalism in mathematics, which is also miscalled 'Platonism', is a comfortably irresponsible way for mathematicians to evade addressing questions concerning their reasons for doing what they do. They claim the same reason that a romantic claims for being creative- namely, self-expression. How miraculously convenient for us all that the whims of Gauss, Bolyai, Lobachevsky, Riemann and others produced non-Euclidean geometries which turned out to be relevant to empirical physics in the case of relativity theory! How remarkably useful that group theory, another
PHYSIOLOGICAL FOUNDATIONS
375
mathematical 'fiction', has such profound consequences for modem particle theory! Pure mathematics is far too closely bound up with physics (often referred to as 'applied mathematics' when not experimental) and has too much prescience to be dismissed as a mere fiction. It is not that good mathematicians do not do exactly what they like, but Guher that what they like is a product of mathematical maturity: their likes and dislikes have become implicit judgments of hewistic worth. The usefulness of mathematical techniques in investigating the world can only be the result of a close relationship between the 'Platonic' objects of mathematics, and the constitution of the physical universe. It seems to me, therefore, that Platonism, or fictionalism, is a doctrine which is false, but whose virtue lies in allowing mathematicians to avoid philosophical questions which they, I suspect rightly, cannot be bothered to answer. 18 There is another point of view which has had some influence among philosophers, and is close to my own, namely that mathematics is just another empirical science like physics, chemistry, or biology. The most famous exponent of this tradition was John Stuart Mill. Of course, his point of view is based on a theory of induction which does not work for any of the sciences, so there is nothing special about its failure with mathematics as such. 19 In the current philosophical scene Quine holds a similar view, though without appealing to a theory of induction. Quine suggests that all our views are empirical, including those of logic and mathematics, because all our views may be modified in the face of empirical difficulty, if we wish. But if logical truth is empirical, then our acquiescence to logical consequence is a contingent decision, in accordance with the deduction theorem.. What he does not give us any indication of is why, in that case, we bother to use argwnents at all, since it all boils down in the end merely to what we are inclined to accept Consequently, Quine's point of view is unsatisfactory as it stands,20 because it does not show us why mathematics is done as it is. The only empiricist philosopher of mathematics who has investigated the issue to find out why mathematicians use proofs at all was Imre Lakatos. He suggested that in informal mathematics proofs are used to make better use of information gained from refutations. He offers as an example how Cauchy proved Euler's conjecture about polyhedra, Le., that the number of edges minus the number of vertices plus the number of faces is equal to two. Euler's conjecture was nevertheless found to have numerous counterexamples. Lakatos suggests that proofs that do not prove can still improve our knowledge by incorpomting the guilty lemma (or premise) into the conjectme. (The proof would still apply to all those objects which were not disqualified by the lemma, for example.) But Lakatos' views are incomplete and unsatisfactory. Formal mathematics, presumably, does prove for the sake of proof. How is 'this to be understood from the point of view of the empiricist? Furthermore, the method of lemma-incorporation is inconsistent with the point emphasized by Quine, namely, that in any
376
J.N. HATTIANGADI
derivation with a false conclusion it is not logically possible to detennine which premise is the 'guilty lemma'. Taken by themselves, Lakatos' views do not work. Added to those of Quine, they are not consistent. So having already noted the inadequacy of Quine's view by itself regarding mathematics, it seems that empiricism about logic and mathematics, however much in vogue, cannot square with the way in which mathematics is done. To summarize these historical notes: The philosophy of mathematics is always described nobly, as a search for foundations, a way of making the results of mathematics more certain. But it is rare that a mathematician fmds that she has to turn to philosophy to be sure that what she says is correct By and large mathematicians do famously without the help of philosophy. The search for the foundations of mathematics, if I am right, is a disguised flight from the discoveries of mathematical physics. This is for the very good reason that it is difficult to state how we can know that mathematical physics can be literally true if it is based on experiment and observation. We must go back and see how to defend Newton's Pythagorean metaphysics from Berkeley's critique. If we can successfully do that, then perhaps we can avoid the necessity of finding that mathematics must be either subjective, that is to say based on intuitions, or innocuous, that is to saY' empty. We must tum to the empirical foundations of science.
III. Foundations of Our Knowledge Examining the Newtonian point of view, we find that it requires that we fmd an answer to the problem of knowledge. We cannot abandon the claim that mathematical physics is empirical, because nothing since Newton has shown us that experiments do not tell us about the world, or that there is another and better way to find out about it. What is the mistake that led scientists in the eighteenth century to think that the gap between the knower and the known is unbridgeable if mathematics is believed to be about the real world? It is clear to me that the error lay in thinking that our knowledge of the world comes from perceptions, as Locke proposed. A survey of modern Darwinian theory would suggest that perceptions do not play the role that was assigned to them in the tradition of British Empiricism.21 The British Empiricist tradition attempts to trace the source of all knowledge to the content of our perceptions, calling them impressions or ideas. It seems beyond doubt that we fmd out much about the world by means of our perceptions. But if we study the physiology and adaptive value of our perceptual apparatus, we derive quite a different estimate of what can and cannot be learned about the world from our percepts. Our perceptual apparatus is derived from that of our animal ancestors. According to our understanding of the course of evolution in the last four billion years or so, the fonns of life which have survived are those which
PHYSIOLOGICAL FOUNDATIONS
377
were better able than their nearest cousins to adapt to the particular environmental niches which were available to be exploited. In these contexts, their perceptual apparatus certainly played an important part in their adaptation. For unlike plant life, animals are unable to absorb energy directly from sunlight Consequently, locomotion and the ability to track down food are vital features in the survival of any of these forms of life. Furthennore, the particular form of perception found in any species of ·animal is closely related to its foraging activity. Another important fact about animals is that they feed not only on plants but occasionally on each other. Those species which specialize in any form of hunting have evolved complex tracking systems for prey, whereas the prey have developed complex ways of hiding from predators, warding them off or fleeing from them. In a great many cases forms of perception among prey are well developed to avoid predatory attacks. Another important aspect of animal evolution is the proliferation of mechanisms which allow the sexes to make contact in order to procreate. Because animals depend on plant material and other life forms for their energy, they have to disperse to find their food. It becomes necessary, therefore, to use their perceptual apparatus to track down other members of the species, and to pick out members of the other sex, for the purposes of procreation. Many animals possess various forms of display to attract the attention of the opposite sex and to warn away members of its own. Among herds of species which intermingle, such as the ruminants of the African rift valley, for example, the different species are clearly marked, allowing for easy recognition. The perceptual apparatus, in all these cases, is well tuned to pick out important and relevant aspects of the environment on which its survival and propagation depends. In this account I have stressed the close relationship which exists between the perceptions of an organism and the characteristic actions that it perfonns in order to manoeuvre successfully around all those things which affect its ability to survive and propagate its own. It is this close relationship between perception and action in all living forms which throws doubt on the model of perception that is found in eighteenth century British Empiricism. Had God made us in order to read His or Her book of nature, as part of a larger plan, perhaps, we would no doubt have had a perceptual apparatus designed to fmd out how things truly are, at least to the extent necessary for our salvation. Adopting as we do a Darwinian explanation of our origins, we must not suppose that the content of our perceptions is veridical.. All we can conclude is that our perceptions allow us to act moderately successfully in various circumstances. As we shall see shortly, it is most important to know those circumstances in which we can act successfully. Life among mammals is never easy. During most of the history of mammmalia, most individuals have been on the verge of starvation. Nearly all the species that have come to be have become extinct, leaving but a small percentage behind. In this respect mammals are quite like other life
378
J.N. HATTlANGADI
foons. To imagine in these circumstances that they could have diverted energy to the formation of a veridical perceptual system is highly unlikely. The most that such species could hope to evolve would be some perceptual apparatus as adequate as, or slightly superior to, the apparatus of its nearest rivals in responding to the opportunities and the dangers in its environment22 On the other hand, in its own normal environnlent where it is most successful, an organj~m could not be systematically deceived to such an extent that it responds inappropriately to its environment in a systematic way. Students of animal behavior have adopted a model, which was found to be of great value in physiology, to describe how organisms perceive the world as part of their adaptation. This is the 'trigger' model of action and perception. It represents organisms as repositories of stored action patterns which are by and large characteristic of the species. These action patterns are elicited by simple cues in the environment, and allow the organism to respond appropriately to them. The model of organism being assumed is a trifle mechanical. I do not want to enter into the discussion here whether such a model is good enough to capture what we mean by 'mind'. My own view is that if it does, it cannot do so without several mechanisms which do not conform to the robot-like model that I am considering. My claim here is not that human, or for that matter animal, consciousness can be reduced to the 'trigger' model of action. It is rather the claim that even if it is not sufficient to explain our perceptual apparatus, the trigger model is necessary for its understanding. We have evolved from creatures which were robot-like. We preserve a great many of the characteristic physiological arrangements of our ancestors. Even if we have developed an organ called the 'mind' which, according to some, escapes description, our physiology is still basically that of apes. It is nevertheless an important part of my thesis that our knowledge is not part of the mind, but is diffused more generally in our physiognomy. With the trigger model, an organism never perceives the whole environment, which is far too complex to be grasped by any being which is not omniscient. Rather, what the organism perceives is a simple feature of the environment which the perceptual apparatus is designed to pick out, and this acts as a trigger which releases a complex action pattern. A hungry bird of prey is coasting along lazily in the sky, or so it seems to us, until it spots a tiny rodent moving a quarter of a mile away, whereupon it swoops in a precipitous dive, landing on top of the rodent- which is itself moving. Grasping the rodent in its talons, with the same motion, it disappears into the sky. Its dive is extraordinary, because if it is a few inches 100 long, it will destroy the bird, and if a few inches too short, or off its mark, the bird goes hungry. We do not believe that the mind of a hawk is capable of calculating its flight path, but we do recognize the extraordinary skill and adaptation exhibited by this feat The bird responds to a simple cue which elicits the dive and which consists of a response to a
PHYSIOLOGICAL FOUNDATIONS
379
characteristic movement of the rodent. As it dives, it approaches its prey by manoeuvring adroitly in response to simple aerodynamic cues. We would search in vain among the percepts of the bird for all the information which it must presuppose in its flight to successfully hunt a field mouse. Human beings are built on the same general principles. Driving on a highway with a colleague, discussing a philosophical point, one responds to the movement of other cars and to the signs as they pass by with extraordinary skill even for a relatively unskilled driver like me. 23 A slight twist of the steering wheel, or a wrong pedal pushed, could lead to a serious accident. What one perceives on the road acts only to elicit the complex learned action patterns called 'driving' that are stored in one. An illustration of the robot-like elicitation of actions is when one is sent out to pick up groceries, and fmds oneself, bewildered, in one's office at the University, having driven there automatically along the route to which he has become daily accustomed. The extremely complex actions which we perform are so often appropriate to the circumstances not because we are omniscient, but merely because the simple cues that elicit our actions are accurate as indicators of the environment. In extremely complex and fast moving environments we can learn to pick out just those transitory features in our perceptual field which allow us to deal adequately with our tasks. When we adequately respond to a complex environment E of a particular kind by an equally complex action A, the existence of a predilection in us to respond to a simple cue S which elicits the appropriate action A may be described as an implicit knowledge of the statement 'Whenever S obtains, E obtains'. In action we exhibit a propositional attitude which, if true, may be described as the knowledge that is expressed in such a statement of universal form. Because environments are complex, we can be mistaken about them. The farther we get from our natural habitat the less trustworthy are our estimates of the environment. We are consequently much more confident in or near familiar surroundings than we are far away from them.24 Our percepts, then, give us no propositional knowledge of the world, except to the extent that they too are involved in skillful actions. Our understanding of the world comes to us not in our percepts, but with them. We know about our world to the extent that the propositional presuppositions of our actions, presuppositions of the form 'whenever S obtains E obtains, in familiar surroundings', are true. H we look at all the remarkable skills we possess, then all of them attest to a rich general knowledge of local conditions. Our knowledge of the world comes with simple cues regarding when we may and when we may not trust our presuppositions, but even these cues work better in our natural habitat than they do in environments far from our own. When Bacon suggested that we experiment to fmd out about the world, he suggested this to facilitate perceptual knowledge of forms or natures. As it happens, the experimental tradition has given us new practical knowledge of
380
J.N. HATTIANGADI
the world which far exceeds anything which can be represented in our senses, or was imagined by Bacon. The connection between science and technology is primarily through the establishment of new experimental skills which are found useful, and not in the application of discovered truth, as Bacon had hoped. Technologists have to find uses for routines that are found trustworthy in the pursuit of science. The reason why an experiment can be relied upon to test a theory is not because it leads us to perceive something veridical, but rather because, as a routine performance, we can rely upon the knowledge embodied in that skill. To the extent that we can learn anything about the world from them we must reinterpret even simple observations as skills, or as elementary experiments. From this biological perspective it is not at all necessary that the world appear to us as it truly is. The cue which allows us to act appropriately is merely a device, not a representation of the world. But in those organisms which monitor their environments constantly (like us) such perceptions are part of an ovemll pictorial space which must bear some structural resemblance to the space of the environment if the cues are reliable. 25 But seeing is not believing. We believe what we must believe in order to do what we do when we do it successfully. In a sentence, my claim is that we rely upon routines which are well tried. If these routines are perceptual we may call them 'observations'. If motor, then 'actions'. And if intellectual, then 'procedures'. But in all cases what we rely upon are well tried routines, or well honed skills. The gap which Berkeley and Hume found between our perceptions and the world led to two hundred years of scepticism about Galileo's, Descartes' and Newton's conception of the mathematical world for a reason which was at least partially sound: no matter how deeply we inquire into the content of our perceptions we will not fmd there any feature which goes beyond the perceptual, suggesting the 'primary' qualities that GaIileo proposed. We do, however, know much about the world that goes beyond its perceptible features, though the knowledge is restricted to genernl features of local conditions in familiar habitats. It seems that we need not give up the scientific conception of a real mathematics of motion on the basis of the attack from the empiricist critique. All we need to do is to reinterpret empiricism as the experimental tradition, and we will have reinstated both of the claims of the seventeenth centwy physicists: that the world is mathematical, and that our knowledge of it is founded on contact with it in experiments.
IV. On Mathematics and Our Knowledge Thereof We live in a mathematical world, and we are mathematical creatures in it. Our spatial physiological structure and our ability to get about in the world show us to be of a piece with the world. Being adapted to our local environment we exhibit in our successful and routine actions an implicit
PHYSIOLOGICAL FOUNDATIONS
381
understanding of the mathematical sbUcture of the local niche. It is this which is the foundation of all knowledge that we possess concerning mathematics. There are three subsidiary theses that I will defend in this section: First of all, that mathematics and axiomatics are not the same thing. Logical truth, I claim, is either subjective or empty, where mathematical truth is neither. Axiomatics is a logical marshalling of mathematics, which gives mathematics an air of triviality, not because the content is trivial, but because the fonn is trivial, being logical. Secondly, I suggest that prior and familiar mathematical skills are the 'laboratory' in which the conjectures of pure mathematics are empirically tested. Since mathematical skills are eventually grounded in our ability to manoeuvre in the physical world, pure mathematics is just as much 'about this world' as is 'applied mathematics', though the latter is a foundation for the fonner. Thirdly I examine the amazing fact that so many mathematical skills are interrelated. To believe, as is common, that this is a sign of logical connection merely confuses logic and mathematics. Rather, this should be taken as a remarkable fact about the world in which we live- its mathematics infuses and stamps itself on so many of our skills, in much the same way that it manifests itself in the symmetries of a crystal. Consequently many of our computational skills can also be seen as manifestations of aspects of the world that we inhabit Our primary understanding of the world comes to us through the routines that we can successfully perform. Jean Piaget got this point exactly right, though he came to it while attempting to develop a Kantian conception of our spatial intuitions among children as they develop their understanding. But if we study how a frog tracks a fly, or a rattlesnake a rodent, we find that perceptual systems have very different a priori presuppositions which make organisms better or worse in their adaptation to their environment We conclude that the mathematical knowledge that any life form possesses cannot be true generally, but is a general understanding merely of its local habitat or territory or its ecological niche or that aspect of its environment which it successfully exploits. But this kind of implicit mathematical knowledge that we possess as part of our adaptation to our local niche does not constitute our knowledge of mathematics as we ordinarily understand it. Mathematics is a subject that we can smdy in a University. If all organisms have the basic implicit knowledge of some aspects of the mechanical universe, it does not follow that they are all mathematicians. It is only among humans that we fmd mathematics journals, or a dedicated group of worker-humans whose main product is new mathematical results. To understand the relationship between the- physiological foundations of mathematics and the discipline of mathematics, we must make an important distinction that is frequently neglected. Mathematics as a statement of facts, such as the ratio of the sides of a right triangle, or the fact that certain series of numbers add in a certain
382
J.N. HATTIANGADI
way, was known in antiquity as geometry and arithmetic long before Euclid axiomatized geometry. Such was the high regard in which Euclid's axiomatization came to be held, that over time the word 'mathematical' has come to have as one of its meanings 'axiomatical'. But in fact mathematics is a subject of much greater antiquity; Euclid's axiomatization of geometry was a comparative latecomer to the scene. The axiomatization of arithmetic was still later accomplished, by Dedekind and also by Peano, not so many years prior to this century. For these reasons I would like to distinguish between mathematical fact and mathematical theorem. For the sake of this discussion let us call a theorem a derivation in an axiomatic system which mayor may not be mathematical and mayor may not be factual, whereas what I call mathematics will be certain facts of geometry, arithmetic, and certain other new subjects such as the differential calculus or the calculus of probability, whether or not they come in an axiomatic fonn, but which are algorithmically related to the earlier mathematical subjects. What is characteristic of mathematical facts is not a nature that they possess, but that they originate in the intellectual traditions of geometry and arithmetic. I will leave the task of determining the true nature of mathematics to those of you who insist that our subjects have natures.26 If we glance at the mathematical sciences in antiquity, we find that they are both closely related to practical action. Land mensuration and keeping accounts are two examples of their use. Looked at thus, they are not much different from our knowledge of the properties of medicinal roots for use in alleviating suffering. Such lists, when truthful, may also be described as knowledge. They are equally closely related to practical action. But if we compare the two kinds of knowledge or purported knowledge, there is a striking difference between them. Knowledge of trees and roots seems to be relatively discrete or atomic in structure, whereas the knowledge of mathematics is holistic, in the sense that its facts turn out to be interconnected in a very remarkable way. This interconnection must not be confused with the logical connection of statements. Any knowledge which we possess in an explicit form is expressed in statements. Statements can be related to each other by the relation of entailment, or of contradiction. In this respect, any statement about the world is interrelated with a denumerable number of other statements in a trivial way. If facts about trees and roots and their uses are taken as an axiomatic basis, for example, we do not think that it is surprising if from them we could derive, by the use of rules of inference, a wide variety of complex statements. While all knowledge can be axiomatic in this sense, there is another sense in which our knowledge of trees and roots in classical systems of medicine is not like mathematics: if we study our knowledge of roots and plants we would be most surprised to fmd that the knowledge of one kind of plant and that of another were interrelated in such a way that if we had the first we could work out the second by an algorithm.
PHYSIOLOGICAL FOUNDATIONS
383
We have a great many skills but all of them are different from each other, and where we know explicitly what must be the case for some skill to be appropriate to a situation, this knowledge in tum is usually a discrete piece of knowledge. Mathematics, it turns out, is unlike that What is surprising about apparently discrete mathematical statements presupposed by discrete mathematical skills is the way in which they are nevertheless interrelated, without any axiomatic regimentation. The relation of number to space is algorithmic, and not axiomatic. Because theoretical mathematics is stated in language, it can be axiomatized, of course, but this might only serve to teach us even more about the algorithms. Logical derivations themselves may be studied mathematically. Logic has its own algorithm, but that does not take away my point. From a logicist's point of view it is only to be expected that mathematical truths are so beautifully interrelated, because all mathematical truths are regarded as empty, and therefore trivially interrelated. 27 But if GOdel is correct in that the truth of a mathematical statement is not the same as its provability, then the interrelatedness of mathematical truth as opposed to any other kind of truth is one that deserves more study. The fll'St recorded intimation of the unity of all mathematics was the Pythagorean doctrine, according to which all is number. However wild his doctrine may seem when applied to ethics and the advisability of eating beans, it gives an acceptable account of the workings of geometry. It is precisely because counting, and the objects presupposed by our skills governing counting, give us an insight into shapes that the reduction of geometry to arithmetic was such an important event in the history of ideas. Were it not for this success, the discovery of the irrationality of the square root of the number two would not have excited remark in antiquity. It is the interrelation of mathematical techniques which gives rise to the study of pure mathematics. Pure mathematics is the study of pure technique, where every technique is grounded eventually in an established skill. The knowledge of mathematical theory which results from the study of pure mathematics is not, however, the basis on which our mathematical skills are founded, but quite the contrary. Because our abstract mathematical knowledge is holistic, we are tempted to think that we understand how to do what we do with number or figure only because we have that knowledge. But a look at the history of mathematics tells us that this is not so. Pythagoms had quite an elaborate theory of number, which was known as the knowledge of the odd and the even. But in retrospect it surprises us that Pythagoras did not have the concept of zero. If we look at any modem definitions of the operation of addition or multiplication we will find that those operations are considered indefinable unless we are able to say that there is a number M such that any number N added to M gives the number N, and such that any number P multiplied by M gives us the number M. Since Pythagoras did not have the concept of zero we might conclude that he did not know how to perform the basic operations of addition, subtraction, multiplication, and division. But to do so
384
J.N. HATTIANGADI
would be a mistake. The ability shown by ancient Greek mathematicians with number series or Diophantine equations leaves no doubt of their ability to add and multiply. It is just that mathematical theory was not as well developed. A similar case can be made about the discovery of the differential calculus by Newton and Leibniz and its relation to the theory that came to be adopted later, called 'analysis'. As a general rule it seems to be the case that our skills at manipulating number and figure far exceeds our theory. In order of richness, therefore, we have the following: knowledge implicit in mathematical skills > explicit knowledge of mathematical truths > mathematical theorems. (Or: Techniques > theories > theorems.) Although this is true at any given time, there is a certain relationship between explicit knowledge and skills that is important We are organisms that are adapted to our niches as well or as poorly as other life forms. We are remarkably similar physiologically to our nearest species, the chimpanzee. We may ask ourselves what value our relatively untrustworthy explicit knowledge can be to us if a chimp gets by without it. If we are adapted to the world in the fIrst place, what is the advantage that we reap from the search for truth? What is the practical value of uncertain knowledge which can only be founded on practice, when we have the practice all along? The answer is this. Our skills work well only in our niches. Our adaptation is only a little bit better than that of our competitors. It is not perfect. These facts of Darwinian theory tell us that any form of life that experiments considerably with changing habitats must change its skills in order to survive. Among human beings, the use of explicit knowledge from a biological perspective lies simply in its potential as a modifier of our skills. An example outside of mathematics will be revealing. The clearest example of 'the relationship between explicit knowledge and natuml ability is to be found in the relationship of coaching to sport. If Ben Johnson is by far the fastest sprinter in the world, why would he need a coach? If a coach knows more about running than Ben Johnson, why is the latter the fastest sprinter? And if not, how can the coach have anything of use to offer? IT we study an athlete in training, however, we find that to remain on top of his fonn or to train for a major event, the natural ability of the athlete must be modified in little ways for improvement Starting out of the block, for example, is a learned technique, and not entirely natuml. Small pointers from a coach can often lead to significant improvements in the competitive ability of a sportsman. If this were not so, the world's fastest sprinter could not learn from anybody. That he can learn shows that a theoritician, who is a poor sprinter, may yet, as a theoritician, teach another to sprint better than he otherwise could. Explicit knowledge may not be able to tell us why Ben Johnson does what he does so well, at least not completely, but it may nevertheless help him to do what he does better.The role of explicit knowledge lies in the improvement of skills. In mathematics, this is most evident in learning new techniques to do old tricks more easily, or more reliably, or even just differently. Thus, in
PHYSIOLOGICAL FOUNDATIONS
385
mathematics, there is this odd relationship between manipulative skills with symbols and the theories of mathematics: the theories help give us more and better skills, while, as a direct result the skills, becoming richer, afford more scope for the study of theory. An example will illustrate this. Multiplication is an old technique that is well known. We can easily understand what a square or a cube of a number is in terms of multiplication, and thus also the general concept of a number to the power n. We can equally learn, using our old notion of division, that if a raised to the power n is divided by a, the result will be a to the power n-l. Generally, if we divide a to the power n by a to the power m, then we get a to the power n-m. From this we can conclude that a to the power negative m is nothing more than dividing one by the number a to the power m. As a consequence we can derive that a to the power zero (or n-n) is 1. All of a sudden we have a new operation called exponentiation, which we can practice, and fmd 'that it gives us all sorts of wonderful new techniques. Logarithms, for example, cannot be understood until we have a clear gmsp of the technique of handling exponents. These in tum give us a new way to calculate what are otherwise very tedious results. (Here, in fact, lies the clue to the importance of good notation in ma'thematics-- it supports good mathematical habits. But once we have a technique of this kind we can investigate it as an interesting object in its own right. What would happen, if, for example, we were to raise two to aleph zero? Such questions can be asked and answered only because in our old domain of natural numbers, in our well established niche, as it were, we can handle exponents with great facility. We are able to ask if, for example, we called the nUlllber obtained by raising two to the aleph zero by the name 'aleph one', and two raised to that number 'aleph two' and so on, whether there are cardinals which are strictly more than one such number and strictly less than the next higher cardinal in the series defined by exponentiation. If we were unsure of our operations of set theory, or of exponents, or of our ability to check whether one of Cantor's claims are true or not, then we would be unable to ask such a question, nor would we have the slightest clue as to how it may be answered, or how we may determine that the question is not answerable. As we increase our mathematical skills, our mathematical theory becomes that much less descriptive of the knowledge that we implicitly possess in the perfonnance. In general, then, our mathematical theory is always much less infonnative than necessary to understand our mathematical skills, but it is nevertheless good enough to improve our skills here and there. When we come finally to deductive methods, we have a new dimension in interrelating theories. But the fact that pure mathematics can be axiomatized does not make it unique in this regard. Spinoza's ethics was also axiomatic in its fonn, as was Newton's physics.2I It is interesting that logic has its own algorithm, which may be generated out of set theory by the metamathematical techniques devised by Tarski, for example. This does not show that mathematics is specially related to logic, but only that
386
J.N. HATTIANGADI
mathematical logic is a kind of mathematics. That is, to use a phrase beloved of logicists, merely a tautology. The astonishing interrelatedness of mathematical facts did suggest that there is a close relationship between mathematics and axiomatics, which is basically Leibniz's hunch. The development of proof procedures in the nineteenth century makes this point of view even more attractive. It is not at all swprising that with the development of rigor and of axiomatic technique in the nineteenth century it became the next step at the turn of 'the century to derive mathematics from logic. Once we accept the implication of GOdel's theorem for the incompleteness of the axioms of arithmetic, however, we can no longer accept logicism. We need not give up our appreciation of proof procedures as algorithm, merely as foundations. The success of Russell and Whitehead, such as it was, in deriving as much as they did from as little as they started with deserves a remark. It seems to me that their success must be attributed to the fact that they took mathematics to be its axiomatics, and derived the axiomatics from logic (including a theory of classes, of course). 29 It is revealing, though, that Russell remarked that mathematics only began, properly speaking, with George Boole. Of course it is only about then that axiomatics and mathematical rigor came into its own. It is not swprising that Russell thought that mathematics began about then, because rigorous axiomatics only begins about then.. For a logicist there is no more to mathematics than its axiomatization. We must recognize that the remarkable unity of mathematical knowledge arises only from the fact that Pythagoms and Galileo accidentally hit upon about space, time, and motion. Because we reflect a mathematical universe, its knowledge is implicit in so many of our skills, in different ways. Therefore, we find remarkable correlations between mathematical knowledge in different fields. Our physiology manifests the mathematics of the universe as surely as a crystal of common salt exhibits symmetries which reflect the geometry of the space in which we live. Our physiology exhibits 'the dynamics even more explicitly in our skills.. So completely do we exhibit the dynamics of the world, that every skill we possess, including even the very specialized linguistic ones, exhibit the mathematical structures which physicists and molecular biologists are groping to find On this construal of mathematics, we must assume that we are basically able to do any mathematics at all because we have the skills necessary to survive in our niche. It follows that the relationship between pure and applied mathematics is not what it is ordinarily assumed to be. If I am right, it is applied mathematics which is basic, and forms the testing ground for theories in the domain of so-called 'pure mathemabcs'. Pure mathematics is the science of the pure technique of manipulating symbols, whereas the mathematical reality that we know in our basic skills foons the experimental laboratory in which the theories of pure mathematics must eventually either work or fail. For this reason I would count myself also as an empiricist where mathematics is concerned, because all the sciences use
PHYSIOLOGICAL FOUNDATIONS
387
well established routines to test hypotheses, and mathematics is no exception to this rule.
I should add that logical truth, unlike mathematics, seems to me be empty. The skills on which logic relies are our skills with our explicit language, and not those skills which we need in our general dealings with the world. Logical truth yields us a device to handle our own statements better. but in the end it tells us nothing about what is the case. It seems that logical truth does depend on the logical or grammatical particles of our language. Against recent tradition in the philosophy of mathematics. therefore, I must endorse the judgment of Galileo, Descartes, Huygens, Newton, and Locke. One concludes that the logic beloved of scholastics, on the one hand, and mathematics as envisioned by Galileo to describe motion itself, on the other, must fall on opposite sides of the great seventeenth century divide: logic is either subjective in the sense of being a human product, or uninformative about the world, whereas mathematics is neither. Mathematics has given us an insight into the nature of the universe, which distinguishes the rise of modern science, and for better or worse, gives us the possibility of modem technology.30
388 ENDNOTES 1. See: David Hmne, Treatise on Human Natwe, ed. Selby-Bigge (Oxford: Oxford University, 1968). 2. See: Nicola Abbagnano, "Psychologism," trans. Nino Langiulli, in The Encyclopedia of Philosophy, ed. Paul Edwards (London: Collier-Macmillan, 1967), Vol. 5-6, pp. 520-1. 3. The expression 'sociologism' is rarely used to describe one's own philosophy, but has a certain vogue today to describe a widespread hypothesis, to the effect that every feature of human knowledge has its origin in some feature of hwnan society- be it a social category of thought, as in Durkheim's school, a politico-economic factor as in one of the schools following Marx, a 'structure' as in another school, or in a nondescript sociology as in the school from Edinburgh which calls itself the 'strong programme'. I do not know who coined the phrase, though I first came across it in: Karl R. Popper, The Poverty of Historicism (London: Routledge and Kegan Paul, 1957). 4. As Mark Notturno has identified, the two classic sources which discuss psychologism critically are Frege's (and, derivatively, Husserl's) writings concerning the nature of logical and arithmetical truth, and Popper's investigations of the empirical basis. See: MA. Notturno, Objectivity, Rationality and the Third Realm: Justification and the Grounds of Psychologism (Boston: Martinus Nijhoff, 1985). 5. Mr. Avshalom Adam, who studies conventionalism in astronomy and in philosophy with me accused me in a heated argwnent of being a Pythagorean, for which I am grateful. I had always imagined that point of view to be ridiculous, perhaps due to the influence of Bertrand Russell's views on mathematics, but in my defense against Adam's spirited attack I discovered that in all these years my excitement in discussing Galileo, Descartes, and Newton stemmed very often from my agreement with their 'Pythagoreanism' as I interpret it, which, because I subconsciously believed it to be indefensible today, I did not maintain. This experience has led me to wonder whether many historians of science who passionately expound and defend a certain philosopher or scientist might benefit from asking whether in fact their love of history is sometimes only another aspect of their unwillingness to defend a point of view which they secretly share with their hero. Along these lines I suspect that something like my view is implicit in the writings of Alexandre Koyre, who must therefore be credited with it, if not Edmund Husserl. See: Alexandre Koyre, From the Closed World to the In{miJe Universe (Baltimore: Johns Hopkins, 1957) which is a more accessible account than his classic Galileo Studies, trans. S. Drake from fro of 1939 (Atlantic Highlands: Humanities Press, 1978). As for Edmund Husserl, The Crisis of European Sciences and Transcendental Phenomenology, trans. David Carr from german orig. of 1937 (Evanston: Northwestern University, 1970), it seems to prepare the way for the work of Husserl's student Koyre though Husserl himself is writing this as an introduction to his phenomenology, with its obvious Kantian overtones. Koyre therefore deserves much of the credit for such a courageous stand on mathematics in the heyday of Russell's logicism. 6. See: "Empiricism from a Biological Perspective" in: J.N. Hattiangadi, How is Language Possible? (La Salle: Open Court, 1987), in which I propose the central thesis of the second sort in this paper. It may also be found in: IN. Hattiangadi, "Knowing That, and How," METHODOLCXJY AND SCIENCE, Vol. 17, no. 3, (1984).
PHYSIOLOGICAL FOUNDATIONS
389
7. Newton's claim is to be f01.Dld in the "General Scholium" at the end of Book ill of Isaac Newton, Mathematical Principles of Natural Philosophy, ed. F. Cajori, trans. A. Motte from Latin orig. of 1687 (Berkeley: Berkeley, 1968). 8. Locke was closely allied with Newton's point of view in the eighteenth century, though it is something of a guess that Newton agreed with Locke regarding the importance of percepts as the source of knowledge. Newton does not write much about these matters, but it seems to me that had he not agreed with Locke, this would have been noted in the Leibni7lClarke Correspondence. See: H.G. Alexander, 00., The LeibnizlClarlce Corresp~nce, from the dispute dating from about 1705 (Manchestel': Manchester University, 1956). 9. See: Bertrand Russell, An Essay on the Foundations of Geometry (New York: Simon and Schuster, 1959). originally, 1897. 10. See: Bertrand Russell, My Philosophical Development (New York: Simon and Schuster, 1959), in which Russell describes his earlier work as false in every respect except on matters of detail. It is the abandonment of Kant's philosophy which leads Russell to reexamine the arguments of Berkeley and Hurne in: Bertrand Russell, The Problems of Philosophy, Home University of Modem Knowledge Series (London: Thornton Butterworth, 1936), originally 1912, in which are laid down the central issues which were to exercise his followers, who dominated philosophy in the years following, sometimes in agreement with him and sometimes qualifying his point of view. 11. See: Bertrand Russell, Principles of Mathematics (London: Allen and Unwin, 1956), originally 1903. 12. The derivation is accomplished in: Bertrand Russell and Alfred North Whitehead, Principia Mathematica (Cambridge: Cambridge University, 1910-13), in 3 vols. 13. See: L.EJ. Brouwer, ''Inaugural Lecture, 14 Oct. 1912," trans. A. Dresden, BULLETIN OF THE AMERICAN MATHEMATICAL SOCIETY, 20, 1914. A satisfying interpretation of modem intuitionism is to be found in: Karl R. Popper, Objective Knowledge (Oxford: Oxford University, 1972), pp. 128-140. Popper shows the interesting relationship between Brouwer's thought and that of Kant See also: L.E.J. Brouwer, "Philosophy and Mathematics," PROCEEDINGS OF THE TENTH INTERNATIONAL CONGRESS OF PHll.OSOPHY, Vol. 1; and Bertrand Russell,
My Philosophical Development. 14. See: Alfred Tarski, "Concept of Truth in Formalized Languages," in Alfred Tarski, Logic, Semantics, MetQ11&Qthematics, trans. Woodger (Oxford: Oxford University, 1956), orig. 1931. Tarski showed that any metalanguage rich enough to define the semantics of an object language satisfactorily is richer than the object language. 15. For an excellent survey of the weaknesses of logicism, see: Alan Musgrave, "Logicism Revisited,.. in THE BRITISH JOURNAL FOR THE PIDLOSOPHY OF SCIENCE, Vol. 28 (2), 1977. 16. Most mathematicians have adopted this view, and so, too, it seems did Kurt G&lel. See: Hao Wang, From Mathematics to Philosophy (London: Routledge and Kegan Paul, 1974). 17. On this point, I confess, I depart from the views of Koyre and Husserl, who seem to regard the scientific revolution as a fulfillment of a Platonic
488
TAKASHI YAGISAWA
full account of the speaker-hearer's competence." Chomsky, Aspects of the Theory of Syntax, p. 6. The 'creative aspect' pertains to language use, hence performance, and as such it is not supposed to be accounted for by a universal grammar or any part of linguistics proper. Chomsky is free from this confusion in: Chomsky, Rules and Representations, pp. 76-79, 222-223. 18. Ibid., p. 4. 19. This is a purer version of Chomsky's thesis asserting the 'creative aspect'. See: Chomsky, Aspects of the Theory of Syntax, p. 6. Unlike Chomsky's thesis, it is primarily syntactic- it avoids mentioning 'thoughts' - and scarcely pragmatic - it avoids talk of 'reacting appropriately in an indefmite range of new situations'. 20. See: Noam Chomsky, 'The Case Against B.F. Skinner," NEW YORK REVIEW OF BOOKS, December 30, 1971; and Daniel C. Dennett, "Skinner Skinn~" in Daniel C. Dennett, Brain Storms (Montgomery: Bradford, 1978), pp. 53-70. 21. I follow Chomsky in suggesting that if 'belief' and 'knowledge' are inappropriate notions here, stipulate 'cognate' to be the appropriate notion and substitute it for 'belief and 'knowledge'. See: Chomsky, Rules and Representations, pp. 69-73, 92-102. 22. Unless the psychology in question is a kind which requires no experiment, e.g., purely introspective psychology. 23. See: Ibid., pp. 217-219, also pp. 117-127. 24. See: W.V. Quine, "On What There is," in W.V. Quine, From a Logical Point of View (New York: Harper and Row, 1953), pp. 1-19. See also: W.V. Quine, Word and Object (Cambridge: MIT, 1960), pp. 200-206, 211, 244. 25. See: David Lewis, COlU1lerfactuals (Cambridge: Harvard University, 1973), p. 87. 26. See: Terrance Parsons, "Entities Without Identity," in IE. Tomberlin, 00., Philosophical Perspectives, 1, (Ridgeview, 1987), pp. 1-19. See also: J.1. Katz, "An Outline of Platonist Grammar," in Katz, ed., The Philosophy of Linguistics, p. 180, for his criticism of Chomsky on this point 27. See: Jerry A. Fodor, "Some Notes on What Linguistics is About," in Ned Block, 00., Readings in Philosophy of Psychology, VoL 2 (Cambridge: Harvard University, 1981), pp. 197-207. 28. The Wrong View says, "(a) that there is a specifiable data base for linguistic theories; (b) that this data base can be specified antecedent to theory construction; (c) that the empirical content of linguistic theories consists of what they say about the data base; and...(d) that the data base for linguistics consists of the intuitions (about grammaticality, ambiguity, and so on) that informants produce (or would produce....)" Ibid., p. 198. 29. Ibid., p. 199. 30. Ibid. 31. Ibid., p. 207n 2. 32. The Right View is also stronger than psychologism in its commitment to internal representations. But this is not relevant to our concerns. 33. However, Fodor does slip into saying, "...the internal representation of the grammar (or, equivalently for these purposes, the internally represented grammar)..., talking and understanding the language normally involve exploiting the internally represented grammar," Ibid., p. 199, thus obliterating the distinction between grammar and its internal representation, which is crucial to the controversy at issue. 34. Ibid., p. 203.
PSYCHOLOGISM IN LINGUISTICS
489
35. Ibid., p. 200-201. 36. See: Ibid., p. 200. 37. Katz, The Philosophy of Lingwistics, pp. 201-202. 38. Ibid., p. 177. 39. Ibid., pp. 177-178. 40. James Higgenbotham attempts such an argument in: James Higgenbo~ "Is Grammar Psychological'!" in Leigh Cauman, et al., eds., How Many Questions? (Cambridge: Hackett, 1983), pp. 170-179. As Katz observes in: Katz, Language and Other Abstract Objects, pp. 179-18011 18, it is not convincing. 41. Stephen Stich, "Grammar, Psychology, and Indeterminacy," THE JOURNAL OF PHILOSOPHY, 79 (1972), pp. 799-818. 42 There is an argwneDt which is being assumed here: The meanings of 'walk' and 'move' are such that for any x, if 'walk' is true of x, 'move' is true of x. But for any x, 'walk' is true of x if and only if x walks, and 'move' is we of x if and only if x moves. Therefore, The meanings of 'walk' and 'move' are such that for any x, if x walks, x moves, that is, any walking person is moving. 43. See: Kripke, Naming and Necessity, for the notion of nomigid designator.
POEMS, PAINTINGS, AND INTENTIONS GREGORY CURRIE II. Psychologism Revived 'Psychologism' has turned out to be a confusingly adaptive tenn. But one kind of mistake that commonly goes by that name is the running together of questions about the origin of some intellectual, or more generally cultural product with questions about its appraisal. Thus Frege, the great opponent of psychologism in logic, insists that we "distinguish between the grounds that justify a conviction and the causes that actually produce it." I When William Wimsatt and Monroe Beardsley identified what they famously called the 'intentional fallacy' they did not describe it as a species of psychologism, but they well might have. Early on they illustrate the fallacy by imagining a critic who says: "In order to judge the poet's performance, we must know what he intended." And they reply: "to insist on the designing intellect as a cause is not to grant the design or intention as a standard by which to judge the worth of the poet's perfonnance.. "2 Wimsatt and Beardsley's paper had the emotional attractions, and the intellectual drawbacks, of a radical manifesto. While it portrayed the opposition as (at best) confused by a woolly minded attachment to 'the spirit of poetry', its sweeping denunciation and urgent message disguised the gaps and ambiguities in the argument To consider just one obvious defect: their reply to the intentionalist that I quoted above illustrates their own commiunent to a fallacy. For we may hold that intentions (of certain kinds) are relevant to a decision about the value of the work without holding that these or any other intentions dictate the nonns of criticism. But Wimsatt and Beardsley have been roughly handled by their critics already, and I don't wish to explore the details of their argument any further. 3 Anyway, by the standards of today, their essay seems impeccably conservative in its defense of objective criticism. If the idea of 'the text' has taken a beating in recent critical writing, it may in part be in reaction to some of the implausibilities in the critical objectivism that Wimsatt and Beardsley defended. Not that currently fashionable schools of criticism have gone back to the author; it is the reader whose mental states now take center stage. In this essay I want to defend the critical importance of authorial (and more generally artistic) intention. I want to enumerate a number of ways in which the artist's intention is an intrinsic feature of the work. Intrinsic in the sense that in the absence of that intention, the work would be different
POEMS, PAINTINGS, AND INTENTIONS
491
in some aesthetically important way. And this relation between intention and work is not to be thought of as causal, but logical. It is not possible for the work to have these aesthetic features in the absence of these intentions. What I should like to present is a general theory which maps and explains the exact nature of the relation between work, intention, and critical judgment Unfortunately, I do not have such a theory (though I've tried make a start on it elsewhere4). There are those, of course, who think that there can be no systematic theorizing in this area. Frank Cioffi for example, in rejecting the intentional fallacy, pr~nts a budget of examples where authorial intention seems pretty clearly" to play a legitimate role in judgments about the work. But his aim is to undermine a general thesis, not to put another in its place. 5 I think Cioffi's examples are telling, and that we can hope to take the matter further. Science has sometimes been thought of as a three level activity, comprised of particular observations, inductive generalizations, and unifying theoretical laws. Without taking this picture too seriously, I think we can use it to illustrate the present strategy. Here I want to do some work at level two; to formulate a set of relatively disconnected hypotheses about the ways in which the artist's intentions bear upon our understanding and evaluation of the work. If we can achieve generality at this level, there doesn't seem to be any reason in principle why we should not eventually achieve a deep theoretical unification of these generalizations. Art may not be quite the motley that Cioffi, in Wittgensteinian mode, supposes it to be. But there is no straight rule of induction at work here; the generalities I want to offer cannot be obtained by asceptic inference from uncontroversial data. Perhaps one needs to see a number of issues in aesthetic theory in the way I see them in order to find my generalizations convincing. I put them forward as a double challenge: to integrate them into some deeper theoretical framework, or to show that they are not true to the data culled from our experience of particular works. My theses are psychologistic - nol, I hope, in the sense that they conflate psychological causes with evaluative standards - but in that they affirm the inseparability of work and intention. Here I must sharply distinguish my position from one sometimes adopted in order to defuse the issue raised by Wimsatt and Beardsley: that an author's intentions can be regarded as fully public objects, embodied in the text. I can make no sense of this idea of 'embodied intentions' unless it is a misleading way of saying that the text contains evidence for some hypothesis about what the author's intentions are. Behavior and the results of behavior, like a text or a painted canvas, do not embody intentions. Intentions bear to the artist's product the same kind of relation as that of the theoretical entities of science to the observable world- belief in them is warranted by inference to the best explanation rather than by direct acquaintance. So when I argue, as I shall, that intentions of various kinds are constitutive of the art work, I shall be arguing that there are things constitutive of a work of art that do not belong to the public domain. So my position is psychologistic in a further sense; it is based on the idea that the mental is essentially interior. Frege himself
492
GREGORY CURRIE
would not have recognized this as a species of psychologism, since this is exactly his own view of the mental.' But I note a modern tendency to use the term in exactly this sense. Thus, Ned Block defends a view, which he calls 'psychologism', that two systems could be identical in actual and potential behavior, while only one of them is intelligent; for intelligence is a function, not of behavioral dispositions, but of 'internal information processing'.' And John McDowell has characterized psychologism as the view that "the significance of others' utterances is a subject for guesswork and speculation as to how things are in a private sphere concealed behind their behavior."' The doctrine that McDowell characterizes - with a view to refuting it, of course - seems to me roughly correct.
ll. Intentions and Text One of the things that Wimsatt and Beardsley argued for in their original essay is that the object of critical scrutiny is not the artist's intentions, but the text of the work. I want to argue ftrst that there is no dichotomy between text and intention, for the text of the work is the text the author in/ended it to have. In saying this, I do not commit myself to the proposition that if the artist merely intended the text to be better than it turned out to be, then there is a non-visible text that is the real text of the work, better than the written text. What determines the text of the work is the author's lexically specific intentions. The author has a lexically specific intention when he intends to write a certain word or sequence of words, spelt in a certain way. When the author has such an intention, his act of writing may fail to embody that intention; he may, through oversight or because he is a bad speller, write something that does not correctly mirror his intentions. If he does, the text he inscribes will deviate in some way from the text of the work. Our practice in the treatment of texts seems to conform to this line of thought Editors routinely correct the misspellings of authors; sometimes, no doubt, referring their corrections back to him for authorization. But where the author is not available for comment such corrections are still frequently made. What principles govern the alteration of such misspellings? One might suppose that a change is judged allowable simply if it amounts to a correction of spelling- bringing the text into line with correct spelling conventions. Corrections beyond this minimum may be made, but they would be regarded as a dubious practice. But we see, on reflection, that a spelling change is allowable only if there is reason to think the change brought the spelling of the text into line with the spelling the author intended, and not allowable merely on the grounds that the change amounts to a correction in the above sense. A copy editor who 'corrected' the spelling in an anthology of poems by e.e. cummings would not keep his job long. Clearly, the reason we would deplore such a change is that we have
POEMS, PAINTINGS, AND INTENTIONS
493
reason to believe that cummings intended his poems to be spelt in a non-standard way. Faced with a text that shows no signs of having been produced with such an intention, we correct the occasional misspelling with a good conscience. But when doubts about the author's intentions creep in, we hesitate. It might be argued that whatever our unreflective practices may be, the principle I offer as a rationalization of them has an unacceptable result: that where the author has false beliefs about how a word is spell, those beliefs will feed into intentions that will set his mistakes in concrete, so to speak. If the author intends to spell 'cat' 'c-a-t-t' because he thinks that is how it is correctly spelt, then the intentionally determined text will have all occurrences of 'cat' misspell. And this would be a gratuitous multiplication of error. But in these cases of error based on ignorance the author has conflicting intentions. He intends to spell 'cat' correctly, and because he believes that it is spelt 'catt' he intends to spell :t like that. Clearly, the intention to spell the word correctly dominates the intention to spell it any particular way. If the error was pointed out to the author, he would abandon the intention to spell it that way. So we let the dominant intention be the determinant of the text This shows that we cannot understand the phrase 'lexically sPecific intentions' to apply narrowly to intentions to write a certain series of letters in a certain order; to do so would be to rob ourselves of the opportunity to solve the kind of problem we are considering here by allowing the intention to spell the world correctly to dominate. Lexically specific intentions are intentions to write words spelt in way F, where 'way F' might be a rigid or a non-rigid designator of letter sequences. 'c-a-t-t' rigidly designates a letter sequence, while 'the correct sequence of letters for spelling "cat"' designates different letter sequences in different worlds. On the other hand, merely intending to spell the word the way it would be spelt by someone who was a bad speller would not count as a lexically specific intention; there is no unique way that a bad speller would spell 'cat', and 'the way a bad speller would spell "cat"' is an improper description. Notice that we are not inclined to say anything comparable about painting. We do sometimes alter the appearance of a painting in order to restore it to its original state (though it is often questionable as to whether that is actually the effect of the alteration), but it would be regarded as indefensible to alter a painting because you thought that you were bringing it into line with the painter's intentions about how it was to look (rather than, as in the case of successful restoration, bringing it into line with how it did previously look). The only legitimate revision of a painting is that done by the artist, when the revision will count as a stage in the production of a yet to be completed work. If our intuitions about what practices are legitimate are based on anything more than prejudice (and a widespread but wholly ungrounded prejudice in this area would certainly need some explaining), it seems there must be a fundamental asymmetry between the
494
GREGORY CURRIE
aesthetic' condition of painting and the aesthetic condition of writing. In what does this asymmetry consist? One rather superficial difference (superficial in this context anyway) between literature and painting is 'that there is nothing comparable to a misspelling in painting, because there is nothing comparable to an alphabet in painting.' There can be artistic mistakes, but they are of the same kind as the choice of an infelicitous metaphor. This has the effect of making it extremely difficult for anyone to attribute to the painter an intention that is specific enough for us to be able to say with certainty what alteration to the painting would count as bringing the picture into line with that intention. We might be able to say that the painter intended that shade of blue to be darker, or that figure a bit further to the left. But who can say exactly how much darker or further to the left he intended it to be? Indeed, we would strongly suspect that there simply are no answers to these questions. Since colors are continuously variable in hue and shapes continuously variable in position, it is very hard to imagine a painter intending to produce exactly that hue and no other, or exactly that positioning of the figure and no other. But nothing is easier than to intend to write this letter rather than some other. But suppose the artist was capable of a purely mental detennination of the exact pictorial structure of his projected work (exact down to the threshold of discriminable difference), yet failed to produce that exact pictorial structure on his canvas. Would the mentally conceived structure constitute the real pictorial structure of the work, to which the appearance of the canvas was but a poor approximation? I suppose Croce and Collingwood would have said so. But they would have been wrong. For the painting stands in relation to the artist's intention in a quite different way from the way the misspelling stands to the author's intentions. The painting itself is not just a register of the author's choices about the visual pattern he wants to select; the artistic performance that we judge does not consist just in choosing a structure of lines and colors- it consists in presenting such a structure using those means that the medium of painting allows. The act of painting is part of the artist's performance that we judge when we judge the quality of the work, for the artist must employ artistic skills in embodying his plan for the work on canvas. But the writer's act of inscribing words is no part of what we judge when we judge the work. The act of inscription is achievable without the employment of artistic skills, and if the writer uses such skills - say, by employing the skills of a calligrapher - they are skills irrelevant to the kind of work he is producing. Tom Jones would not be a better novel if Fielding had had more beautiful handwriting. The endpoint of the author's artistic activity is the point at which his indexically specific intentions are determinate. This explains why we might correct an author's manuscript, but never a painter's canvas. In correcting the author's manuscript, we try to make clear some aspect of his perfonnance that the manuscript obscures; we try to set the record straight. But to 'correct' the painter's canvas by overpainting, or
POEMS, PAINTINGS, AND INTENTIONS
495
whatever, would be to obscure some aspect of the painter's performance that the canvas previously made clear; it would be to falsify the record. To sum up. In literature, the execution of the work consists in deciding what words to use. Whether this is done antecedently to writing or in the very act of writing, it is the decision that counts towards the performance, not the transcription of the words. But in the visual arts, there is no such thing as a set of decisions that determine the geometrical structure of the work. That is determined in the act of painting, and that is the stage of execution.
III. Intention and Fiction One very important class of literary works is the class of fictional works. But what, exactly, distinguishes fiction from non-fiction? Again, the answer is to be found in a certain constitutive intention of the author- constitutive because fictional status is constitutive of the work. Imagine a possible world in which Henry Fielding wrote a book called 'Tom Jones' in which he detailed, with truthful intent, the life of someone he was acquainted with. Would that be a world in which Fielding wrote Tom Jones, the work with which we are familiar- a world in which that very work just happened not to be fiction? Hardly. The non-fictional story called 'Tom Jones' would, even if word for word identical with the text of Tom Jones, be different from it in aesthetically significant respects. 'Tom Jones' might be good or bad biography, but the relevant criteria with which to judge it would be different from those we would use to judge Tom Jones. Sameness of text might be necessary for trans-world identification of works, but it is certainly not sufficient. 10 To see the text alone as constitutive of the work is a New-Critical dogma we shall have to abandon. If works can share the same text while only one of them is fiction (as I have just supposed), it is plain that the fictional status of a work cannot supervene on any or all of its textual properties. It must then be something external to the text that determines fictionality, and here there are but two plausible candidates: audience practices of regarding the work in a certain way, and authorial intentions. But the fIrst is unsatisfactory; audience practices can vary over time, but a fictional work surely does not cease to be fiction if audience reaction to it changes. Elsewhere I have argued that fiction is constituted by a certain kind of communicative intention on the part of the author- a (Gricean) intention to get the audience to take the attitude of make-believe towards the story that is told. ll On this view, the act of producing fiction ('fiction-making', as Kendall Walton calls itl~ is a communicative act of the same kind as those acts, sometimes called 'illocutionary acts', amongst which are asserting and requesting. These are acts which have in common the familiar reflexive intention; 'the intention to achieve a certain effect via the recognition of that very intention, and which are differentiated one from another in terms of
496
GREGORY CURRIE
the specific effect that it is intended that they will achieve. This view stands in sharp contrast to the view that fiction-making involves some kind of pretence on the part of the author. Thus, it is said that the author of fiction pretends to be asserting what he says, that he pretends to have knowledge of the events he describes. 13 This seems to me a highly implausible view. That is, it is implausible to say that an author who types out his story and sends it off to the press is pretending to do anything. Perhaps camp fife story tellers can be said to take on a pretended persona for the duration of their tale- the persona of one who knows the truth of which he speaks. But these days, acts of authorship are rarely undertaken in so picturesque a setting. This implausible theory may be the product of a confused perception of something closer to the truth: that readers of fiction pretend that the story is related to them by someone who knows the facts of which he speaks, and they perceive that this is intended by the author to be their pretence. I" But, of course, to encourage pretence in an audience one need not engage in pretence oneself.
IV. Intention and the Story I have said that the text of the work depends upon the author's lexically specific intentions, and that the question of whether the work is fiction or not depends upon the author's communicative intentions. But to show that the traffic is not all one way along this street, I want now to argue that the story that is told by a fictional work is not a function of the author's intentions. This will seem paradoxical only if we assume that the structure of the story - what is, as we say, 'true in the story' - depends wholly on the text; for then, by the transitivity of 'detennines' we would get the result that the author's lexically specific intentions determine the story. But clearly elements of a fictional story are left undetermined by the text There are things true in the stories about Sherlock Holmes that are not explicit in the text, and not entailed by anything that is explicit in the text It is possible to fmd quite bizarre interpretations of these stories that are formally consistent with the text One might suppose that Watson is a lunatic or a liar who never met anyone called 'Holmes' and that the story is his mvings or invention. One might explain Holmes' superior powers of inference and observation on the hypothesis that he is a supernatural being in disguise. IS But it is intuitively obvious that these things are not true in the stories; Watson is by and large a reliable reporter of the facts of Holmes' cases, and Holmes himself is a human being, if an unusual one. Clearly, what is true in the stories is not just a matter of what is said in the text, but of what can be inferred on the basis of the text together with certain collateral information. Exactly what this collateral information is and how these inferences are to be made is a matter of some controversy. 16 But one thing seems clear: we cannot say that what is true in the story is what the author intended to be true in the story. Conan Doyle might have had quite strange
POEMS, PAINTINGS, AND INTENTIONS
497
intentions concerning what is to be taken as true in his story, he might have had strange beliefs about, f(»" instance, extraterrestrial beings who visit us, and he might have believed that these beliefs were widely shared by members of his community. He might have thought that it would be as obvious to most people as it was to him that the likely explanation of Holmes' superi(»" mental powers was that he was an alien being, and for this reason he did not make this part of the story explicit. But if we discovered that all this was indeed true of Conan Doyle's beliefs and intentions, I do not think it would give us any reason to revise our views about what is true in his stories. Whatever is legitimate collateral information about what is true in a story, extra-textual revelations about the author's intentions in this area are not relevant information. The reason has something to do with a point already made: that a judgment of the work is a judgment of the author's achievement. Merely for the author to intend that the story be interpreted in a certain way constitutes no achievement on his part. But it is part of his artistic achievement to write in such a way that his intentions in this area be put into effect- that readers, if they are well provided with relevant collateral information and are attentive to the text, will be likely to interpret the story in the way intended.
V. Intention and Representation I want to turn finally to an example of constitutive intention in painting. (What I say here would apply, I think, equally well to sculpture and other forms of visual representation.) My claim is that it is a necessary condition for a painting to depict something that the artist intended it to depict that thing. I am not claiming, on the other hand. that it is a sufficient condition; an artist may intend to depict something and fail, for various reasons, to be successful in carrying out his intention. Nor am I claiming that we can go any way towards explaining what it is for a picture to depict in terms of what it is for an artist to intend to depict. On the contrary, it seems likely that we shall have to explain intending to depict in terms of depicting. My claim is just that in order for there to be depiction, there must be a depictive intention. Why does depiction presuppose intention? Because only intention has the fmely tuned capacity to connect marks on canvas with the objects in the world that we regard those marks as depicting, and only intention has the kind of logical structure that will allow us to depict the non-existent (as we are inclined to say; more on this in a moment). This might be doubted on the grounds that language has the same discriminatory/intentionalistic features, without being subject to the same intentionalistic constraints that I claim for pictures. In order for a word-token to refer to a, it is not required that the speaker intend to refer to a by means of that word. Of course, this is not to deny that linguistic reference depends ultimately on, and in
498
GREGORY CURRIE
complex ways on, speakers' intentions; rather, it is to say that the conventionality of language loosens its ties with intention. But it is this element of conventionality that distinguishes the linguistic case from the pictorial one. Pictorial representation is not conventional in the way that linguistic representation is. What a linguistic expression refers to is a matter of there being a convention in the relevant community to the effect that the word has that reference. Pictures, at least as we commonly use them, are not governed by any similar convention. There are no conventions to the effect that certain patterns of shape and color depict certain objects or kinds of objects. It makes nonsense of the history of art to suppose that a regular practice of depicting a certain object by means of a certain kind of picture entrenches to the point where a picture of the kind depicts that object whatever the artist intends it to depict Pictorial depiction is more like the kind of pre-linguistic communication imagined by Bennett and other Gricean-minded philosophers, where an 'on-off' gesture or sound can be counted on to get across the speaker's intentions, because of its iconic relation to the state of affairs represented. 17 There are, of course, regularities of practice in painting that we tend to call conventions, and there may be some temptation to suppose that these 'conventions' can fall sufficiently hard on an artist's activity to tip the balance against his intentions. Imagine an artist who has seen plenty of Pietas, but does not know what they represent If he makes a work in the same form, surely it represents the Virgin and the dead Christ regardless of his intentions. Two different cases need to be distinguished here. In one, the artist has no ideas one way or the other about what is represented, but he intends to represent whoever is usually represented in works of this kind. His action is like the reference borrowing action of one who uses the name 'Napoleon', not knowing to whom it refers, but intending to refer to whoever is referred to by users back along the reference borrowing chain. Such a claim is no counterexample to the thesis that representation depends upon the artist's intentions. In the other case, the artist has intentions about who is represented, but his intentions are deviant with respect to the community; he intends to represent someone other than the person normally represented in such works. It seems clear that such a work can represent what the artist intends it to represenL Imagine a tasteless representation in Pieta style entitled 'Jackie and J.F.K., Dallas, 1963'. We may object to the high-jacking of a sacred imagine, but the reason we do is surely that a work which ought to represent one thing is being used to represent another. Without a title or some other authoritative indication of the artist's intention we would probably take the work as orthodox in its representational properties. But that just shows we can be wrong about what is represented. We are familiar in the philosophy of language with the distinction between what a speaker means and what a speaker says, or - what is more to the point here - between speaker's reference and semantic reference. Thus, philosophers have wondered how and in exactly what sense a speaker
POEMS, PAINTINGS, AND INTENTIONS
499
can say 'The man drinking champagne is happy tonight' and refer to someone who is not, in fact, drinking champagne. II But examples of the Pieta variety show, I think, that there is no comparable distinction to be made for pictorial depiction. There is no sense in which the Pieta-like representation itself depicts (or, in 'the tendentious language of semiotics, 'denotes') Mary and Christ, while the artist depicts something else. The difference between language and painting is that language is essentially conventional while painting is nOl The use of signs on an ad hoc basis, each chosen for the needs of the moment, is not the use of a language. A language involves signs recmrently used to signify a given state of affairs, where there is common knowledge between language users of the use to which a sign is to be put. Nothing precludes pictures being used in this way, but a regular practice of pictorial depiction can emerge without there being such conventions for the use of signs. Words and sentences cannot mean things without there being conventions of meaning, but pictures can depict without conventions of depiction. And that is, by and large, the condition of pictures in our artistic culture. There are no conventions that determine that a given picture depicts a given thing. Since depicting might have been conventional, we can sensibly ask why it is not. To see the answer, we need to take a closer look at the notion of a convention. According to David Lewis, conventions are regularities of behavior which are conformed to because they are solutions to co-ordination problems. I' For example, the speaker and his audience want to co-ordinate their actions; the speaker has to choose some means of expressing his thought (a sign) and the audience has to associate the sign with the thought expressed. It does not matter very much what sign is used; what matters is that speaker and audience associate the same sign with the same thought. Once communication has been achieved with the use of a given sign, it is likely that the same sign will be used again for the same purpose. For speaker and audience will expect each to use the strategy that succeeded before, and this expectation gives them both a reason for so using it themselves. A convention is born. Now something similar can be said about the situation of painter and audience. The painter wants to represent something and the public wants to know what is represented. But several things militate against ta'le development of conventions to solve this problem. First, with high probability, painter and public can co-ordinate their actions without convention. For the development of a realistic style of pictorial depiction has enabled painters to count on the ability of the audience to know what is represented simply on the basis of their perception of visual properties of the picture. Of course, this will work only some times and for some audiences. But the artist can also exploit pre-existent conventions of language in order to give his work a title and thereby indicate what it represents. Secondly, conventions here would in many cases be pointless, because painters often work without any background of regularity in depicting that thing, and without any expectation that such a regularity will
500
GREGORY CURRIE
develop. Thirdly, conventions would be costly in their imposition of constraints on the artist who wants to be free to depict in whatever manner he chooses. I have remarked that in our culture regularities do sometimes emerge in connection with the depiction of particular things or kinds of things, but these regularities are not happily described as conventions in the present sense. Thus, The Saints are regularly depicted as having halos, but this is not a convention. They are thus depicted because of certain views about the nature and/or appearance of saints that people hold or did hold. Thus, there were (what were taken to be) good reasons for so depicting The Saints, and hence, so depicting them was not an arbitrary choice. But a convention proper is arbitrary. In some periods, the Virgin is regularly depicted in clothes of a certain color. But again, this is not a convention; it is regularly adhered to because it was judged to have certain aesthetic or devotional effects, not because it solved a co-ordination problem. Pictorial style is often described by art critics and historians as conventional. Cubist painters depict objects in a way systematically different from the way, say, Impressionist painters depict objects. But the elements of Cubist style are not conventional in our sense, for they are non-arbitrary; Cubist painters paint objects that way because they prefer, for aesthetic reasons, painting that way to any other. Neither are the elements of style solutions to co-ordination problems. Once a painter paints that way, his or her audience have the problem of interpreting the work, and they may do this partly by assimilating the elements of the style. The style creates a communicative problem rather than solving one. Perhaps what sounds like a bold thesis - that there are no conventions in painting - is merely a restrictive stipulation about the meaning of 'convention'. But we must remember what is at stake here. I don't claim that what art historians have called 'conventions' in painting don't exist, and I don't dispute their right to so call them. But I do insist that we distinguish conventions in this weak sense from conventions of a kind sufficient to generate a language. Since the question before is whether depiction in painting is in any sense linguistic, it is proper to concentrate our attention on conventions of the kind that Lewis has analyzed. Turning away from conventions as an alternative to intentions, it might be said that what is represented is determined ultimately by causal facts of a brute kind; that, for instance, it was the Duke of Wellington who sat before Goya; that light reflected from the surface of the Duke and impinged upon Goya's eyes as he painted. And this is what makes it the case that the Duke is represented. We need not assume that Goya' s actions as he painted were those of an automaton, that his act of painting was not controlled by intentions. It is just that no intention he had was relevant to determining the identity of what is represented. Goya might have thought that it was Napoleon sitting in front of him, intending that it be Napoleon represented. But as long as the object in his line of sight was Wellington and not Napoleon, it is Wellington and not Napoleon that is represented.
POEMS, PAINTINGS, AND INTENTIONS
501
Such uncompromising causalism can hardly explain the representation of non-existents (since one cannot be causally related to a non-existent), nor the fact that a picture may represent St. John even though it was an artist's model that stood before the painter. These cases are much better handled in tenns of the artist's intention. What about cases involving the kind of mistake just alluded to- when an artist thinks he is painting one person or scene, but is in fact painting another? These too can be explained in terms of the artist's intention. For the artist might be mistaken about the identity of the man before him, or about the location in which he paints, but he still intends to represent that man (the one in front of him) or the scene here abouts. In these cases, the artist has conflicting intentions. He intends to represent that man, and to represent the Duke, but that man is not the Duke. I suggest that the former intention, a demonstrative intention as we may call it, has priority in determining what is represented. For the having of a demonstrative intention signifies a peculiarly close relation between the artist and his subject; it is the kind of intention that he can have only when he is in close proximity to the subject. 211 In order to represent something, the artist need not have a demonstrative intention to represent that thing- he need not have a de re intention at all. The artist can depict Smith's murderer without knowing who Smith's murderer is, carrying out his work on the basis of the police evidence alone. 21 But where there is a demonstrative intention it will be decisive for representation. Where there is no demonstrative intention, representation can, in unfavorable circumstances, become ambiguous or indeterminate. Imagine an artist whose knowledge of the Bible is hazy. He intends to paint Joseph, the man who was both the father of Jesus and the owner of the tomb in which Jesus was laid. He is confused as between Joseph of Nazareth and Joseph of Arimethea, and nothing else he knows tells in favor of the one over the other. The picture itself does not unambiguously depict an incident in the life of either man. Here there is no reason for saying it is one Joseph rather than the other that is depicted. An intentional account of representation does not dispense with causation. Causation plays its part in determining representation because it plays a part in constraining representational intention. Whatever the causal constraints on depiction are (and what they are is controversial), they are just the causal constraints on intending. Thus, the object of the artist's demonstrative intention (and hence the very identity of that intention) will be determined by his causal proximity to that object. Similarly, if there cannot be a depiction of anything that is not causally connected with the picture, the reason will be that no one can intend anything concerning an entity from which they are causally disconnected.
502
GREGORY CURRIE
VI. Conclusions I have argued that the very identity of a work of art depends upon certain quite specific intentions that the artist or author must have.. These intentions are interior states of the artist that can at best be inferred from the public object that the artist has left us - text or canvas - together with whatever relevant historical and psychological knowledge we may have. These inferences can vary from the near trivial (in which case they will be made by almost everyone under almost all circumstances of information) to the highly complex (in which case they will require a great deal of art historical knowledge and are likely to be controversial). If I am right about this, we shall have to rethink some of our assumptions about the 'public' nature of artworks.
503 ENDNOTES 1. Gottlob Frege, Nachgelassene Schriften, ed H. Hermes, F. Kambartel, and F. Kaulbach (Hambmg: Felix Meiner, 1969), p. 159. On Frege and psychologism, see: Gregory Currie, "Frege and Popper: Two enemies of Psychologism," in K. Gavroglu, ed, Proceedings of the Thessaloniki Conference (Dordrecht: Reidel, 1988). It is interesting to note that Frege, who did so much to rid logic and mathematics of unwanted psychologism, was sceptical about the possibility of objective criticism in the arts. He thought that a work of art is nothing but 'a structme of ideas within us'; that the beautiful is just what seems to individual persons to be beautiful. See: Gottlob Frege, Nachgelassene Schriften, p. 144. Here, Frege is close to endorsing another 'fallacy' announced by Wimsatt and Beardsley: the affective fallacy. 2. William Wimsatt and Monroe Beardsley, 'The Intentional Fallacy," SEWANEE REVIEW, 54 (1946), pp. 468-88; reprinted in D. Newton-de Molina, ed, On Literary Intentions (Edinbmgh: Edinburgh University, 1976), p. 1. The first quotation above appears in quotation marks in the original text, but is unattributed. Italics in the original. 3. The volume cited in note 2 is a good somce of critical reactions to Wimsatt and Beardsley's original statement 4. See: Gregory Currie, An Ontology of Art (London: Macmillan, 1988). 5. See: Frank Cioffi, "Intention and Interpretation in Criticism," PROCEEDINGS OF THE ARISTOTEUAN SOCIETY, LXIV (1963-4), pp. 85-106; reprinted in Newton-de Molina, ed, On Literary Intentions. 6. See, for example, Frege's notorious discussion of mental images in: Gottlob Frege, "On Sense and Reference"; reprinted as "On Sense and Meaning" in Gottlob Frege, Collected Papers on Mathematics, Logic and Philosophy, ed. B.F. McGuiness (Oxford: Basil Blackwell, 1984), especially p. 160. 7. See: Ned Block, "Psychologism and Behaviourism," PHILOSOPHICAL REVIEW, 60 (1981), pp. 5-43. 8. John McDowell, 'The Epistemology of Understanding," in 1. Bouveresse and H. Parret, eds., Meaning and Understanding (Berlin: de Grayter, 1981) p. 225. 9. See: Nelson Goodman, Languages of Art (Indianapolis: Hackett, 1976), cbs. 3-5. 10. See: Currie, An Ontology of Art, sec. 39. 11. See: Gregory Currie, The Nature of Fiction (Forthcoming), ch. 1. 12. See: Kendall Walton, "Fiction, Fiction-Making and Styles of Fictionality," PHILOSOPHY AND UTERATURE, 7 (1983), pp. 78-88. But the argument of Walton's paper is deeply at odds with my own conclusions; see: Currie, The Nature of Fiction, ch. I, sec. 4. 13. This view is an extremely popular one. See, for example: John Searle, 'The Logical Structme of Fictional Discomse," in John Searle, Expression and Meaning (Cambridge: Cambridge University, 1979) for a version of it 14. Closer to the truth but not true, I think. For pretence and make-believe are not the same. See: Currie, The Nature of Fiction, ch. I, sec. 7. 15. One rather dreary modem day Holmes patische takes this line. 16. For competing views, see: David Lewis, ''Truth in Fiction," AMERICAN PHILOSOPHICAL QUARTERLY, 15 (1978), pp. 37-46; see also: Currie, The Nature of Fiction, ch. 2. 17. See: Jonathan Bennett, Linguistic Behaviour (Cambridge: Cambridge University, 1976). For the reasons given above and further on, I reject all 'semiotic' theories of pictorial depiction as based on a quite fallacious assimilation of depiction
504
GREGORY CURRIE
to linguistic representation. The best known semiotic theory is, of course, contained in: Goodman, Languages of Art. 18. See: Keith Donnellan, "Reference and Definite Descriptions," PIDLOSOPIDCAL REVIEW, LXXV, pp. 183-205; see also: Saul Kripke, "Speaker's Reference and Semantic Reference," MIDWEST STUDIES IN PIDLOSOPHY, Vol. 2, (1977). 19. See: David K. Lewis, Convention (Cambridge: Harvard University, 1969). 20. Compare this with the case of conflicting lexicographic intentions discussed in section II of this paper. 21. Thus, for a picture to depict a, the artist must have the intention to depict the CI», for some CI» where the CI» is an a, or he must have a singular intention- to depict Wellington, or thai man.