teorema Revista internacional de filosofia CONSEJO EDITORIAL
J.J. ACERO, U. Granada
J. McDowell, U. Pittsburgh tF. MONTERO, U. Valencia 1. MOSTEIUN, CSIC, Madrid
R. BENEYTO, U. Valencia
tJ.L. BLASCO, U. Valencia R. BODEl, U. Pisa M. BUNGE, U. McGill,Montreal
C.U. MOULINES,U. Munich
C. MaYA. U. Valencia C.P. OTERO, UCLA
M. CACCIARI, U. Venecia J. CORBi, U. Valencia
D.F. PEARS, U. Oxford J.L. PRADES, U. Gerena
N. CHOMSKY, MIT i'D. DAVIDSON, DC Berkeley
D. QUESADA, U. Autonoma, Barcelona i'W.v.O. QUINE, u. Harvard
J. EqIEVERRfA, CSIC, Madrid
tJ.
FERRATER MORA, Bryn Mawr
College
J, FODOR, U. Rutgers-CUNY
I. REauERA, U. Extremadura M. SABATES, U. Kansas
tJ.D. GARdA BACCA, U. Caracas
M. GARcfA-CARPINTERO, U. Barcelona A. GARciA SUAREZ, U. Oviedo C.GARcfA-1RE~JANo,U.Comprurens~
Madrid M. GARRIDO, U. Complutense, Madrid P. GOCHET, U. Lieja C. G6MEZ, CSIC.Madrid A. GOMILA., U. La Laguna S. HAA.CK,U. Miami tS. HAMPsH:fRE, U. Oxford J. HIERRO, U. Antonoma, Madrid F. JARAUTA, U. Murcia M. JIMENEZ REDONDO, U. Valencia J. de LoRENZO, U. Valladolid
M,A. QUINTANILLA, U. Salamanca V. RANTALA, U. Tampere
tM. SANCHEZ-MAZAS, U. Pais Vasco J. SANMARTiN, U. Valencia J.R. SEARLE, UC Berkeley J. SEOANE, U. Valencia G. SOLANA, U. Autonoma, Madrid E. SOSA, U. Brown P.F. STRAWSON, U. Oxford C. THIEBAUT, U. Carlos III,Madrid Ch. THIEL, U. Erlangen R. TUOMF.LA, U. Helsinki A. VALCARCBL, U. Oviedo tG.H. von WRIGHT, Academia de Finlandia
DIRECTOR
L.M. VALDES, U. Oviedo SECRETARIO
A. GARCfA RODlUGlJEZ. U. Murcia
[
i\
Ii II, ~
I
INDICE ASPECTS OF TI:IE PlllLOSOPHY OF JOHN McDoWELL DIRECTOR INVITADO: .ANGEL GARCIA RODRIGUEZ .1
A. GARCiA RODRIGUEZ, Introduction
5
ARTICULOS J. M.cDoWELL~ The Disjunctive Conception ofExperience as Material for a Transcendental Argument
19
S. VIRVlDAKIS, On Mclsowell's Conception ofthe 'Transcendental'
35
J. McDoWELL, Response to Stelios Virvidiakis
59
J. VEGA ENCABO, Appearances and Disjunctions: Empirical Authority in McDowell is Space ofReasons
63
J. McDoWELL, Response to Jesus Vega Encabo
82
J. CHURCH, Locating the Space ofReasons
85
1. McDoWELL, Response to Jennifer Church
97
C. PAGONDIOTIS, Mclrowell's Transcendental Empiricism and the Theory-Ladenness ofExperience
101
J. McDoWELL, Response to Costas Pagondiotis
115
S. GONzALEZ Reasons
ARNAL~
Non-Articulable Content and the Realm of 121
J. McDoWELL, Response to Stella Gonzalez Arnal
132
J.-L. PRADES, Variettes ofInternal Relations: Intention, Expression and Norms
137
1. McDoWELL, Response to Josep Lluls Prades
155
W. CHlLD, On Having a MeaningBefore One's Mind
161
J.McDowELL, Response to William Child
176
S. SAWYER, The Role ofObject-Dependent Contentin 181
Psychological Explanation 1. McDoWELL, Response to Sarah Sawyer
193
D. L6PEZ DE SA, Values vs. Secondary Qualities
197
1. McDOWELL, Response to Dan Lopez de Sa
211
BIBLIOGRAFIA DE JOHN McDOWELL
215
LIBROS RECIENTES DE PENSAMlENTO
225
Lilli lata (JBolledn de b. Altlelr:\teO
CatedT
39
3) We must be able to make sense of a distinguished class of experiences, in which how things are makes itself available to us in perception. 4) There exists an epistemically distinguished class of experiences, in which how things are makes itself available to us in perception. 9 Unfortunately, it is not at all clear that the fact that we cannot avoid making sense of a "distinguished" class of experiences, in which how things are makes itself availa J so far from being items that are in themselves content-free but acquire content as an addwith content on by an analogue to interpretation, arc nothing but content conceived as what, in minds, is capable of meshing with the crystalline superhard material (as it were; this language carries its pictorial character on its face) that supposedly constitutes the essence of the thinkable (see §97). Intentional states are configurations in the mental material (as it were) composing the wheels that engage with the super-rigid rails marked out in conceptual space by meanings (see §218). 3. We can appreciate the pull of such imagery in connection, first> with the regress of interpretations that Wittgenstein identifies as the basis of the paradox about rule-following. If there is to be application for the concept of acting in accord with a rule, it can seem that the mere dead object that we want to see as an expression of the rule - say a sign-post - needs to be interpreted as prescribing some actions and forbidding others. But to give an interpretation would be to make another attempt at an expression of the rule (see the last paragraph of §201). As such, the new attempt would be just as much in need of interpretathe sign-post, tion as the first candidate for being an expression of the rule to stay with the same example. It is only for a moment that some attempt to give an interpretation can seem to bridge the gulf between the dead object and its being correct, for instance, to go to the left if one is aiming to reach the destination to which the sign-post points the way. Offering to interpret the interpretation only postpones the issue. The gulf remains; the object stays dead. That could be put by saying that the regress of interpretations threatens the hardness of the logical "must", as it applies ill connection with rulefollowing: the hardness of the "must" that figures in saying something of the form "To conform to the rule expressed here, you must do such-and-such'>.
Response to Josep Lluis Prades
157
And now it can be tempting to think we can avert this threat even if we let the regress start. We purport to bring the regress to a stop by appealing to something that is an interpretation, but unlike an ordinary interpretation, in not being vulnerable to being interpreted otherwise than so as to yield the right sorting of action into what accords with the rule and what does not. Arriving at an interpretation of this supposedly hardness-preserving kind would be getting one's mental wheels locked 011 to the right rails, engaged with the right structure in the crystalline order. In §431, Wittgenstein considers a counterpart to this pseudo-solution of the problem about rule-following posed by the regress of interpretations. The counterpart is this: "There is a gulf between an order and its execution. It has to be filled by the act of understanding." Here, as in the case of the supposed gulf between an expression of a rule and an action in which one follows the rule, the right move is to deny that there is always a gulf. Of course orders, like attempts at expressing rules, can be unclear, needing to be interpreted. But, to echo §20] , there is a way of obeying an order that is not an interpretation, If we do not question the claim that there is always that gulf, we are subject to a version of the regress of interpretations. And now it can seem that an act of understanding might be able to the supposed gulf - to preserve the hardness of the "must" that in saying something of the form "To obey this order, you must do such-and-such", But this can seem to work only if we conceive the act of understanding on the model of the supposedly regress-terminating interpretation that figures in that pseudo-solution of the rule-following problem: that is, in a way that is captured by the image of something that engages with the crystalline essence ofthe thinkable. This opens naturally into §437. By this point in the text, we have seen two cases, rule-following and obedience to orders, in which there is a lively temptation to suppose the hardness of the logical "muse' can be protected only by resorting to the mythology of a crystalline structure. I think the primary target, in the discussion of intentional states that begins at §437, is a generalization of that. The idea under attack is that we can preserve the hardness of the "must" that figures in saying, for instance, "If this expectation is to be fulfilled, such-and-such is what must happen" only by conceiving expectations in terms of that mythology of engagement with a super-rigid order. 4. Kripke uses the idea of normativity to bring under a single head the relation of the meaning of a rule to action in accord with it and the relation of intention to execution. I endorse that, and Prades objects. He thinks the assimilation sets up a tension in my views. The only interpretation Prades considers, for the thesis that the relation between meaning and action is normative, is one we could explicate by elaborating Wittgenstein's remark that obeying a rule is a practice. The norms that
John McDowell
158
are brought into view by this interpretation of the thesis are the norms of a communal practice. So Prades thinks the assimilation puts at risk my entitlement to deny, as I did above (§1), that we might exploit a conception of executing an intention as a practice, in undermining the difficulty Wittgenstein considers about the idea of intentions and their execution. But as I understand Kripke's remark, the sense in which, for these purposes, it is correct to say the relation of the meaning of a rule to action is normative pertains to what is needed for generating the apparent problem about role-following, not to what Wittgenstein says in dissolving it, which can be, as Prades urges, special to that case. What generates the apparent problem about rule-following is that the idea of a rule needs a conceptual context in which we can use the idea of accord. (See §§198 and 201.) And the notion of accord is already a normative notion, in the sense that matters for Kripke's assimilation. We need to be able to distinguish action that is correct in the light of the rule from action that is not. It can be tempting to think we can make this distinction only by attaching an interpretation to a dead object. If we succumb to that temptation, we have embarked on the regress, which threatens to undermine the applicability ofthe notion of accord. The Kripkean assimilation I mean to endorse is this: the notion of accord which is itself, as I said, a normative notion in the relevant sense is needed, in a parallel way, for the viability of the idea of intention. (And similarly with intentional states in general.) Just as my understanding of the instruction "Add 2 is such that when I have reached 1000, only my writing "1002" will accord with it, so my intention to climb a certain tree is such that only my climbing that tree will accord with it. As Prades says, a primitive intention is not an intention to submit one's behaviour to a pre-existing norm. But this does not tell against the assimilation, rightly understood. It is the intention itself that is, in the relevant sense, a norm for the behaviour of its possessor. Kripke argues that meaning cannot be reductively explained in terms of dispositions..He means to be speaking of dispositions in a sense that does not allow for a parallel application of the notion of accord. In the relevant sense, a disposition is something that, in certain circumstances, results in a certain outcome. There is no sense in which the outcome is correct in the light of the disposition. Prades is right that we can understand the word "disposition" in a different sense, one that makes it appropriate to identify a certain disposition to act in some way with an intention to act in that way. With dispositions so understood, the notion of accord fits, and with it the hardness of the logical "must". But in the sense that matters for Kripke's rejection of a reduction, the relation of a disposition to its actualizations involves no logical "must", but only a cause-effect connection. And there is nothing problematic about generalizing Kripke's claim to intentional states. \Ve can reject a reduction of inH
Response to Josep Lluls Prades
159
tention to dispositions in the relevant sense, on the ground that the relation of intention to action is normative, in the sense I have tried to explain. 5. I do not think any of this conflicts with Prades's tine treatment of expressive behaviour as the proto-phenomenon of intentionality. But the priority he attaches to that topic is questionable) if we see things in the way I have been urging. On the reading I have indicated, to get straight about intentionality we would need to free ourselves from the temptation to resort to the mythology of the crystalline order wherever we find tile hardness of the logical "must", It is not clear how insisting that expressive behaviour is the protophenomenon of intentionality could help with freeing ourselves from that temptation. As far as that goes, we might accept a central role for expressive behaviour even if we were still bewitched by the mythology. We might suppose that expressive behaviour reveals cases of engagement with the crystalline order. Insisting on expressive behaviour as proto-phenomenon might help to undermine a conception of intentional states as in themselves contentfree. But on the reading 1 have sketched, that is not the primary target of Wittgenstein's re-flections about intentional states. To dislodge the mythology, nothing would serve but uncovering and discrediting the fundamental source of its attractiveness to us. And the best candidate for that is the illusion that it affords a way to protect the hardness of the logical "must" in the rule-following case. So dissolving the rulefollowing paradox has a priority that we do not put in doubt when we agree with Prades, as we must, that acting intentionally is not as such subjecting oneself to communal norms. JOHN McDOWELL
NOTE 1 Executing an intention is at least doing something. But if someone is inclined to think that is enough ground for it to be correct to say executing an intention is a practice, she should consider another case. Is expecting things, and having the expectations either satisfied or disappointed, a practice? Surely not.
teorema Vol. XXVIl, 2006, pp, 161-175
On Having a Meaning Before One's Mind Willianl Child
RESUMEN
l,En que sentido puede, segun Wittgenstein, presentarse ante una mente el significado de una palabra? Esta cuestion es abordada a la 1uz de algunos ejemplos poco discutidos. Wittgenstcinrechaza una explicacion del fenorneno de que un significado se presente ante una mente en terminos de un doble componente de (a) habilidades mas (b) experiencias conscientes que carcccn de contenido intencional intrfnseco. Pero no dice sin mas que se trata de un fen6meno basico de Ia conciencia que 110 requiere explicacion, Mas bien, hace varias observaciones positivas acerca de 10 que sucede cuando el significado se presenta ante una mente, observaciones que buscan iluminar el fenomeno. ABSTRACT
In what sense, according to Wittgenstein, can the meaning of a word come to mind? That question is considered in the light of some little-discussed examples. Wittgenstein rejects the kind of two-component view that explains the phenomenon of a meaning's coming before sorneonc's mind in terms of (a) abilities and (b) conscious experiences that lack intrinsic intentional content. But he does not say simply that a meaning's coming to mind is a basic phenomenon of consciousness that needs 110 explanation. Instead, he makes various positive remarks about what happens when a meaning eome before someone's mind, remarks that aim to illuminate the phenomenon.
I
In what sense, if any, can the meaning of a word come to mind, be present to someone's mind, or be present in consciousness? And what can we learn from Wittgenstein's writings on the topic? Wittgenstein acknowledges the existence of various phenomena that we might think of as ways in which a meaning can come before someone's mind. He insists on the experiential difference between perceiving written or spoken words in a language one does not understand, or in an unfamiliar code, and perceiving sentences in one's ordinary language [see (1974), p. 58; '(1958a), p. 214]. He discusses the phenomena of hearing a word in a particular sense [(1958a), §534] and of grasping the meaning of a word in a flash
161
162
William Child
[(1958a), §197]. And he talks about the "game" of experiencing the meaning of a word that is pronounced in isolation [(1958a), pp, 215-6]. But how should we understand what Wittgenstein says about such phenomena? According to John McDowell, the target of many of Wittgenstein's comments about mind, meaning and intentionality is a particular conception of what it is for something to "come to mind" or to be "a content of a mind".' According to this "target conception"} he writes, minds: are populated exclusively with items that, considered in themselves, do not sort things outside the mind ... into 11108e that are correct or incorrect in the light of those items.... [T]he contents of minds are items that, considered in themselves.just "stand there like a sign-post", as Wittgenstein puts it [(1992), p. 264].
So "the introspectible [is pictured] as a domain of self-containedly knowable states of affairs, only externally related to anything outside themselves" [(1991), p. 302]. Now that is exactly the picture of consciousness that is attributed to Wittgenstein by, amongst others, Crispin Wright (who is broadly sympathetic) and Michael Ayers (who is unsympathetic)." Ayers writes: [Wittgenstein assumes] that consciousness consists in nothing but a string of sensations, images, verbal images and "introspectible qualities", together with "feelings" of desire for something, of certainty, of understanding and so forth, the latter conceived of as quasi-sensations... [He assumes] that it must be pessible to describe any aspect of "what is going on in consciousness" in neutral terms, as it is intrinsically and in itself, without reference to its intentional content [(1991), p. 281].3
On this conception, what is going on in consciousness can never be, or involve, a meaning. So the phenomenon of grasping the meaning of a word in a flash, for example, will be analyzed as having two components. On the one hand, there is an experience; all internal picture or image, perhaps, or a feeling of conviction. On the other hand, there is the acquisition of a disposition or ability; an ability to use the word correctly." So the conscious state of someone who grasps the meaning of a word in a flash may be exactly like the conscious state of someone who does not; the difference between them is a matter of their abilities, not a matter of conscious awareness. McDowell argues that this two-component model misunderstands the place of the "target conception" in Wittgenstein's discussion: Wittgenstein's point is to exorcize the "target conception", not to recommend it. On the twocomponent model, what is before the mind of someone who grasps a meaning will be a bare image or feeling; the meaning of the word itself will not be present to her mind. And apart from anything else, McDowell thinks, that conflicts with Wittgenstein's stated view that there is nothing wrong with saying that, when I grasp the meaning of a word, the use of the word is in
On Having a Meaning Before One's Mind
163
some sense present; the only thing wrong is to think that the use of the word must be present "in a queer way" [(l958a)) §195]. On McDowell's view, Wittgenstein does not object to the idea that it is the meaning of the word itself that comes to mind when one grasps the meaning in a flash. His target is) rather, the "misconception ... of what it is for anything to come to mind" that is embodied in the "target conception". In summary: Once we understand what the target is, we can see that insisting that that conception of the mental cannot make room for meanings to come to mind docs not amount to denying that meanings can come to mind, And now there should be nothing against allowing meanings to come to mind: for instance, when one grasps a meaning in a flash; or differently - when one visualizes something; or - differently again - when it suddenly occurs to one that one has forgotten to mail a letter [McDowell (1991), p. 309].5
The contrast between Mcfrowell's interpretation of Wittgenstein and Wright's comes out starkly in what they say about mental images. Here is Wright: When an image, or picture, comes before my mind, it presumably cannot constitute a more explicit or substantial presence than the coming of a real physical picture before my physical eye. And when the latter happens, it is of course con· sistent with my being in full command of all manifest features of the object that I remain ignorant precisely of its intentionality of what it is a picture of. I want to say that, analogously, in the sense in which an image or mental picture can come before tile mind, its intentionality cannot [Wright (200 1b), p. 342].
McDowell responds: This strikes me as back to front. The truth is more like this: the only thing that comes before the mind, when (as we say), an image does, is its intentionalitythe image's content, what it is that is pictured. It is my wife's face that comes before my mind when I imagine my wife's face. Nothing else does, at least nothing that is relevant to the fact that I am imagining my wife's face; certainly not some inner analogue of, say, a photograph, with properties describable independently of any content it can be seen as carrying [McDowell (1998c), p, 56].
The difference between the two interpretations is clear. Which has a better claim to capture Wittgensteiri's views?
II
I want to discuss two cases where, at least at first sight, Wittgenstein does seem to accept some sort of two-component view.
William Child
164
(a) First, consider the discussion of understanding the word "cube" that starts at Philosophical Investigations § 139: When S0111eone says· the word "cube" to me, for example, I know what it means. But can the whole use of the word come before my mind, when I understand it in this way? Well, but on the other hand, isn't the meaning of the words also determined by this use? And can these ways of determining meaning conflict? Can what we grasp in a flash accord with a use, fit or fail to fit it? And how can what is Pl'Csent to us in an instant, what comes before our mind in an instant, fit a use? What really comes before our mind when we understand a word? - Isn't it something like a picture? Can't it be a picture? [(1958a), § 139]
Wittgenstein goes on to make his familiar point, that any picture, and anything that is like a picture, can be applied in numerous different ways, So it cannot be sufficient for someone's grasping the meaning of the word "cube", or understanding an instance of the word "cube" in an utterance, that she simply associates it with a picture or image of a cube: .. .the same thing can come before our minds when we hear the word and the application still be different. Has it the same meaning both times? I think we shall say not [(1958a), §140].
The discussion continues: Suppose, however, that not merely the picture of the cube, but also the method of projection comes before our mind? - How am I to imagine this? - Perhaps I see before me a schema showing the method of projection: say a picture of two cubes connected by lines of projection, - But does this really get me any further? Can't I now imagine different applications of this schema too? Well, yes, but then can't an application come before my mind? - It can: only we need to get clearer about our application of this expression [(1958a), §141],
Here we approach the crucial point. An application in the sense that Wittgenstein is talking about - a method of application - is not something that "just stands there like a signpost", susceptible of different interpretations or applications; it is internally related to its instances. Now both sides to the interpretative dispute will agree that there is a sense in which the application of a schema can come before someone's mind, The dispute is about what that sense is. The two-component view says that what it is for a method of application to come before someone's mind is for her to have some image or picture in mind, plus the disposition or ability to go on to apply it in the right way, The opposing, anti-constructivist view says that a method of application's coming before someone's mind is itself a basic kind of conscious, intentional state which is 110t to be explained, or explained away, in any other terms. What does Wittgenstein himself say? He asks:
On Having a Meaning Before One's Mind
165
Suppose I explain various methods of projection to someone 80 that he may go on to apply them; let us ask ourselves when we should say that the method that J intend comes before his mind [ibid.].
And he answers his question like this: Now clearly we accept two different kinds of criteria for this: on the one hand the picture (of whatever kind) that at some time or other comes before his mind; on the other, the application which - in the course of time - he makes of what he imagines [ibid.].
On Mcfrowell's account, the dialectic ought to go something like this. "Suppose we start with the 'target conception'. Now consider the question: can a method of application come before someone's mind? Ifwhat comes before a person's mind is something like an ordinary picture or sign-post (as the target conception insists it must be), then it is not possible for a method of application to come before anyone's mind; for an ordinary picture or sign-post is consistent with numerous different methods of application. That makes it tempting to appeal to something that is present to the mind in just the same way as an ordinary picture but which really does have a method of application built into it. But that is hopeless; there could be no such thing. Should we conclude that a method of application cannot come before anyone's mind? No. We should, instead, give up the initial assumption that what can come before the mind must be something like a sign-post or picture. Once we do that, we can accept the common-sense thought that, when someone consciously grasps a method of application, it is the method of application itself that comes before her mind." But this interpretation does not seem to capture the line of thought in the passage I have been quoting. In that passage, Wittgenstein seems to offer the two-component view in his own voice as an account of the sort of thing that does actually happen when the method that I intend comes before someone's mind. He does not seem to be saying: "An application's coming before the mind can be an occurrent phenomenon of consciousness - in a way that is sui generis and irreducible". Rather, he seems to be saying: "An application can come before someone's mind. And what happens when an application comes before someone's mind need be no more than this: a real or imagined picture comes before his mind; and he is able to apply what he sees or imagines in the appropriate way.:" (b) Second, consider this passage from Philosophical Grammar: One imagines the meaning as something which comes before OW' minds when we hear a word. What comes before our minds when we hear a word is certainly something characteristic of the meaning. But what comes before my mind is an example,
I "I
William Child
166
an application of the word. And this coming to mind doesn't really consist in a particular image's being present whenever I utter or hear the word, but in [the] fact that when I'm asked the meaning of the word, applications of the word occur to me [(1974), pp. 118-9].
Again, this passage acknowledges that there is something that we call "a meaning's coming before someone's mind". And again, it insists that a meaning's coming before someone's mind does not consist in an image's being present whenever she utters or hears the word. What does it consist in, then? An anti-constructivist will say that a meaning's coming to mind is a basic, unanalyzable phenomenon of consciousness. The two-component view, on the other hand, says that when a meaning comes before someone>s mind, no more need happen than that a particular example or application comes to her mind and that she is able to cite other applications when asked what the word means.' And in this passage, at least, Wittgenstein again seems to side with the two-component view. I have mentioned two passages that, on the face of it, seem to support the two-component interpretation. But there are many others. For example, there is the discussion of suddenly understanding how to continue a series [(l958a), §§ 151-5, §§ 179~83]: here, the conscious component is, for instance, a formula's occurring to me; the practical component is the ability to continue the series. And there is Philosophical Investigations §200, where Wittgenstein seems to envisage that the conscious states of someone who cannot play chess may exactly mimic those of someone who can - so that the difference between them is entirely a matter of differences in context and practical ability.
III In examples like those just considered, Wittgenstein acknowledges that we speak of meanings and methods of application coming before the mind. But, at least at first sight, he really does seem to be offering an account of what it is for a meaning to come to mind in terms of the presence before the mind of something like a picture or image, plus a context of actual or potential applications of that picture. So is the Ayers/Wright interpretation vindicated? Well, the starting point for that interpretation was the idea that conscious mental phenomena have no intrinsic intentional properties. But, as Ayers himself remarks, the examples do not vindicate that point but contradict it. 8 In saying what happens when a method of projection comes before someone's mind, Wittgenstein helps himself to the idea that a picture of two cubes connected by lines of projection comes before his mind. And similarly, ill the other case, he helps himself to the idea that a particular application of a word comes before my mind. But, on the face of it, the fact that I am imagin-
On Having a Meaning Before One's Mind
167
ing a picture of two cubes, or. the fact that a given application qf a word comes before my mind, is an intrinsic intentional feature of my state of awareness. So Wittgenstein does 110t seem to be working with a conception of consciousness on which what comes before someone's mind can have no intrinsic intentional content. Ayers suggests, at this point, that Wittgensteiri's account is simply muddled or internally incoherent. Wittgenstein's official view is that "what is going 011 in consciousness" can always be characterized "in neutral terms, intrinsically and in itself, without reference to its intentional content" [Ayers (1991), p. 281]. But when he comes to give an account of what goes on in consciousness, he freely appeals to items like sensations and images that are intrinsically intentionaL So his view is incoherent; what conceals the incoherence is that the fact "that sensations and images themselves must have intentional content is either implicitly denied or conveniently overlooked" [Ayers (1991), p. 281]. This interpretation seems too unsympathetic to be plausible. A more sympathetic interpretation would be that the intentionality of mental images is not simply taken for granted by Wittgenstein but is itself given a two-component explanation in terms of the presence before the mind of something non-intentional and purely phenomenal, on the one hand, and the subject's disposition or ability to use or respond to this phenomenal element in a particular way, on the other." But it seems highly implausible that Wittgenstein thought of mental images as internal pictures whose intrinsic properties are non-intentional and which acquire intentional properties only by being intended or treated in a particular way. Such a view seems hard to square, for example, with comments like this: "what makes my image of him into an image of him? Nothing of what holds for a portrait holds for the image. The question makes a mistake" [(1980a), p, 262]. For present purposes, then, I will assume that Wittgenstein takes it for granted that when one has a mental image, or when a picture comes before one's mind, one is in a conscious state with an intrinsic intentional content. If that is right, then in passages like those I have quoted, Wittgenstein is not advancing the kind of two-component view on which what comes before one's mind when, as we say, a meaning or method of application comes before someone.'s mind is, in itself, only ever a contentless quale or raw feel, But in that case, what is he doing? One of Wittgenstein's aims in such passages is evidently to argue against the idea that associating a word with a mental image can achieve something that associating a word with an ordinary picture could not achieve. As part of the argument, he suggests replacing mental images with real drawings or pictures: let us adopt the method ... of replacing [the] mental image by some outward object seen; e.g. a painted or modelled image. Then why should the written sign
t68
William Child plus this painted image be alive if the written sign alone was dead? In fact, as soon as you think of replacing the mental image by, say, a painted one, and as soon as the image thereby loses its occult character, it ceases to seem to impart any life to the sentence at all [(1958b), p, 5].10
Ayers might object that passages like this commit the very error he complains about: of treating mental images exactly like painted images and thereby denying the intrinsic intentionality that mental images have and painted images lack. But Wittgenstein's point is not that there is an exact analogy between having a mental image and looking at a red patch. The point is that, even though there is a crucial disanalogy between the two cases, an appeal to the mental image can achieve no more than an appeal to the actual sample of red in fixing or explaining the meaning of the word "red". If I associate the word "red" with an actual red patch, J am associating it with something that has no intrinsic intentionality; whereas, if I associate the word with an image of a red patch, I am associating it with something that does have intrinsic intentionality; that is the disanalogy between the two cases. But the reason why such an association cannot by itself fix the meaning of a word is essentially the same in each case; even if the association establishes that the word "red" is applicable to this particular patch, it does not answer the question, what else the word is to apply to. That problem is completely unaffected by the difference between an actual sample of red and a mental image of a red patch. This first point is compatible with the anti-constructivist view. But according to the two-component theorist, the passages we are considering have a further point, which is harder for the anti-constructivist to accommodate. What comes to my mind when I form an image of my wife's face is, as McDowell says, my wife's face. In this case, there is some definite thing a face for me to be conscious of But when I grasp the meaning of a word in a flash, what is there to come before my mind; what is there for me to be conscious of? The anti-constructivist will answer, "the meaning of the word". But, as we have seen, Wittgenstein's own answer to the question is not (or not simply) that I am conscious of the meaning of the word. Rather, he says that what I am conscious of are examples a particular application or applications of the word. The suggestion is not that Wittgenstein is attempting to explain a meaning's coming before the mind in terms of something nonintentional; for imagining, or thinking of, an application of a word is being in an intrinsically intentional state. The point is rather that, when he talks of an application or applications of the word coming before my mind, Wittgenstein is attempting to articulate or illuminate what goes on when a meaning comes before my mind, not simply treating it as something sui generis about which nothing illuminating can be said. This interpretation sits well with Wittgenstein's idea that the way to explain or communicate the meaning of a word is by giving examples [(l958a),
On Having a Meaning Before One's Mind
169
§208]. In that case, Wittgcustein is clear that there is such a thing as explainor communicating the meaning of a word. But he does not say that the phenomenon of explaining the meaning of a word is simply sui generis, or that there is nothing informative to say about it. On the contrary ~ he tells us what happens when someone explains a meaning to someone else: she gives him examples of how the word is applied; she stands ready to give more examples if necessary; and in doing this, she does not communicate less than she knows herself.
IV The discussion of the previous section suggests a revised version of the two-component view> in which the conscious component is not something that "just stands there like a signpost" but, rather, a state of awareness with an intentional content. When we describe someone as having a meaning before their mind, what is present to their mind is an application, or applications, of the word (real or imagined). In the right context (a context in which the person is able to go on and apply the word in the right way), someone who has an application before their mind counts as having the meaning of the word in mind. But, strictly speaking, there will be no experiential difference between someone who imagines a set of applications of a word and understands the word in one way and someone who imagines the set of applications and understands the word in a different way or not at all. But there are good reasons to think that this still cannot be right as an interpretation of Wittgenstein. After all, he does say that the difference between hearing a word in one sense and hearing it in another is a real difference in how the words sound. And consider again his answer to the question, when we should say that a particular method of applying a schema comes before someone's mind: clearly we accept two different kinds of criteria for this: on the one hand the picture (of whatever kind) that at some time or other comes before his mind; on the other, the application which - in the course oftime - he makes of what he imagines [(1958a)~ §141].
The two-component view reads that as an account of what it is for a method of application to come before someone's mind. But what Wittgenstein says is only that these are the two kinds of criteria for a particular method of application's coming before someone's mind. And to say that is not to say that a method of application's coming before my mind consists in, or amounts to no more than, a picture's coming before my mind plus my having the disposition or ability to apply that picture in a certain way. If these considerations
170
William Child
suggest that the two-component view is interpretatively inadequate, should we, then, adopt the anti-constructivist view and accept that there is nothing more to say than that the meaning of a word itself can be what comes before someone's mind? But in that case, why does Wittgenstein say that, when the meaning of a word comes before my mind, what comes before my mind is an application or applications of the word? One response to this interpretative dilemma would be to distinguish between different stages in the development of Wittgenstein's views. So, for example, it might be suggested that the two-component view is a fair account of his views in Philosophical Grammar or in Part I of Philosophical Investigations, and that it was only later, in Part Il of Philosophical Investigations and elsewhere, that he came to think that we really can experience a word or sentence as meaning what it does. I I Similarly, it might be observed that we need to distinguish between different kinds of case in which the meaning of a word could be said to come before someone's mind; they do not form a homogenous category. And, it might be argued, different cases will need to be treated in different ways. So the two-component view may be right for some cases; perhaps, for example, in some or all cases where I grasp the meaning of a word in a flash, I do simply acquire the ability to use the word correctly without in any sense experiencing the word as having the meaning I grasp it as having. 12 In other cases, the two-component view will be wrong. And there is more than one kind of case where the meaning of a word in some sense characterizes our experience of that word. For example, Wittgenstein explicitly warns against taking the experiences we enjoy when playing the "game" of experiencing the meaning of an isolated word as a model for our conscious awareness in everyday linguistic interaction: If a sensitive ear shows me, when I am playing this game, that 1 have now this doesn't it also show me that I often do not now that experience of the word have any experience of it in the course of talking? [(1958a), pp. 215-6J
But that is not to say that, when ordinarily hearing words "in the course of talking", I hear mere words - without hearing them as meaning what they do. Wittgenstein's point is just that, when I hear a word uttered as part of a sentence of spoken English, 1 do not normally have experiences of the same kind as those I have when pronouncing it in isolation. Another response to the dilemma is to distinguish between what is immediately before my mind and the various properties that I experience it as having. When I see or imagine my wife's face, there is 110 problem in saying both that all that comes before my mind is my wife's face and that I experience the familiarity of her face. The familiarity cannot come before my mind all by itself; there must always be something else that I am experiencing as familiar. But the familiarity really is experienced; for the face is experienced
On Having a Meaning Before One's Mind
171
as familiar." Similarly in the current case. The meaning of a word cannot come before someone's mind all by itself; there must be something else that is heard, 01' seen, or imagined. So when the meaning of a word comes before my mind, what is immediately present to my mind is the heard, seen or imagined word itself; but the perceived or imagined word is experienced differently when it is understood from how it is experienced when it is not understood (or is understood differently). If we take this view, we must give up some of the anti-constructivist's slogans: in particular, we cannot say that all that comes before someone's mind when she grasps the meaning of a word in a flash is the meaning of that word; for, on this account, consciousness of the meaning of a word will always involve consciousness of something else - the perceived or imagined word - that bears the meaning. But that does not undermine the basic anti-constructivist idea that, when the meaning of a word comes before my mind, the meaning itself is a feature of my conscious awareness. 14 The point is well taken. But we can still press the question: what is it for the meaning of a word to come before one's mind; what is involved in hearing a word in one sense rather than another? In response to such questions, the anti-constructivist is likely to make two moves. First, he will say that there is nothing more informative to say about what is involved in experiencing a word in a given sense than j ust that it involves experiencing that word in that sense; "the content of the experience just is to be described by the speeffie expression (of the experience)" [(1980a), p. 105]. He may allow that there is plenty to say about the background that is required if someone is to be equipped to hear a word in a given sense: as in many other cases, Hit is only if someone can do, has learnt, is master of, such-and-such, that it makes sense to say he has had this experience" [(1958a), p. 209]. But, he will insist, when it comes to characterizing the content of the experience itself, there is really nothing more to say. Second, he will offer a diagnosis. W11at makes us worry that there is nothing to be the content of an experience of meaning, he thinks, is our tendency to construe having such experiences on the model of having a mental image: for the content of an image is the kind of thing that can be identified by pointing to a picture; and there is nothing of that kind to be the content of an experience of meaning. But the lesson to draw is not that there is no content for such an experience to have but, instead, that we should refrain from modelling experiences of one kind on experiences of some other kind. Applying Wittgenstein's comment about a different though related case, we can say that the experience of meaning "is not queerer than any other; it simply differs in kind from those experiences which we regard as the most fundamental ones, our sense impressions for instance" [(1958a), p. 215)]. Wittgenstein does make exactly these moves. IS But as we have seen, he says something else too: that when the meaning of a word comes before my
172
William Child
mind, what comes before my mind is all example (or examples) of its application, And in saying that, he seems to be spelling out what is involved in the meaning of a word's coming before one's mind, It might be claimed that this relatively early comment (from Philosophical Grammar) is superseded by Wittgensteiu's later views. But we can in fact see the same idea at work in something he says later, something that expresses a real insight about the phenomenology of meaning. For a word to be understood in a given way, Wittgenstein thinks, is for it to be understood as having a given use - a given pattern of potential applications and potential transitions to other words. So to hear a word as having a particular meaning is to hear it as part of a pattern of potential applications and transitions: "Phrased like this, emphasized like this, heard in this way, this sentence is the first in a series in which a transition is made to these sentences, pictures, actions" [(1958a), §534]. And this is a genuinely experiential phenomenon; the word really is heard as an element in one set of potential applications rather than another. This way of fleshing out what is involved in experiencing a meaning brings out the parallel with other cases where something is experienced as having a particular continuation for example, hearing a chord as a modulation into one key rather than another, or feeling the ending of a church mode as an ending [(l958a), §§536, 535].16 And it fits well with the idea that what comes before my mind when a meaning comes to mind is an application or applications of the word: for hearing a word as part of a pattern of potential applications involves having an application or applications in mind. These comments of Wittgenstein's offer no reduction of the experience of meaning to anything else. And they will not apply to every case that can be classified as one in which the meaning of a word or words comes before someone's mind; as we have seen, the phenomena are very various. But Wittgenstein's comments do tell us something more informative than just that what comes before someone's mind when, e.g., one grasps the meaning of a word in a flash is the meaning ofthe word. And perhaps they go as far as one can to illuminate what such experiences involve whilst respecting their sui generis character, 17
University College Oxford University Oxford, OX14BH, UK E-mail: [email protected]
On Having a Meaning Before One's Mind
173
NOTES I For "come to mind" see McDowell (1991), p. 303; for "a content of a mind" see McDowell (1992), p. 264. 2 For Wright's reading, see Wright (L987), (1991), (2001b). (Wright allows that psychological items do have internal relations to the outer. But he presents these internal relations as essentially derivative; a conscious item is not in itself internally related to anything else [see (2001b), p. 343].) For Ayers's reading, see Ayers (1991), chs 22 and 31. (Ayers says explicitly that he is less interested in the views of the historical Wittgenstein than in the impact of those views on contemporary philosophy [see Ayers (1991), p. 9 for a general disclaimer]. Accordingly, he sometimes identifies the views he attacks as Wittgensteinian (or "Wittgensteinian") rather than ascribing them to Wittgenstcin himself') 3 As Ayers himself points out, there is an evident tension in a view that conceives of consciousness in terms of images but denies that conscious experience has intrinsic intentional content. For, on the face of it, mental images are intrinsically contentful; so, if consciousness contains mental images, it does contain things with an intrinsic intentional content. I will return to this point below. 4 The two-component view need not say that we can give a reductive account of this ability, in terms that do 110t presuppose the notions of meaning or understanding. The point is that we can separate the experiential component of grasping the meaning of a word in a flash from the ability or practice aspect; not that the ability or practice can itself be adequately characterized in non-normative terms. 5 Note that when McDowell talks about a meaning's coming to mind he is working with a more inclusive concept of meaning than I am. For him, "having a meaning in mind" is, roughly, being in a conscious state with an intrinsic intentional content. I am focusing on a subset ofthe cases that concern McDowell: those in which the Hn~ujstic meaning of a word or sentence comes to mind. It might be said that, though Wittgenstein does talk in §§138-9 about grasping the meaning ofa word in a flash, the subject matter of those sections is not the sudden acquisition of understanding but, rather, our knowledge of what someone means when they use a word in ordinary conversation. That is true. But what matters for present purposes is that the discussion that starts in § 138 leads on to the question, what happens when a method of application comes before someone's mind. It is Wittgenstein's discussion of that question that seems to suggest a two-component view. '7 Note that, in this passage, "an application" means a particular example. In (1958a), §141, "an application" meant a general method of application. S See the comment from Ayers (1991), p, 281 quoted in the next paragraph. (Here we return to the issue registered in note 3 above.) 9 Such a view is suggested by Wright. See Wright (2001b), p. 342. . 10 He makes the same point immediately after the passage I quoted earlier from Philosophical Investigations §§ 139~41: "can't it be clearly seen here that it is absolutely inessential for the picture to exist in his imagination rather than as a drawing or model in front of him; or again as something that he himself constructs as a model?" II No doubt Wittgenstein's views in this area developed over time. In 1929, he clearly did have a two-component view of conscious phenomena, which he later abandoned - though he did not, as far as I know, apply his 1929 view to the phe-
174
William Child
nomenology of language-use. (For his 1929 view, see the discussion of the meaning of the word "toothache" in Wittgenstein (1975), pp. 88-9.) But the precise suggestion in the text is implausible; for Wittgenstein discusses the case of hearing a word in a certain sense in Philosophical Investigations §543 and seems to treat it there as a genuinely experiential phenomenon. 12 "If the meaning has occurred to you, you know it now, and its occurring to you was simply the beginning of/mowing." [(1980a), p. 263, second emphasis minej]. That suggests that the meaning's occurring to you may simply consist in the onset of the ability to use it. (This suggestion is removed from the later version of the remark at (1958a), p. 176, where Wittgenstcin says only that knowing how to use the word "began" when it occurred to you.) 13 See e.g. (1974), p. 175 on the impossibility of separating the impression of familiarity from the impression of the face. 14 I should emphasize that McDowell himself never implies that all that comes to mind when the meaning of a word comes to mind is the meaning itself. He does say that all that comes to my mind when an image of my wife's face comes to mind is my wife's face. But he docs not suggest that the case of grasping or experiencing a meaning is exactly parallel to the case of imagining a face. Indeed, when he says elsewhere that our awareness of facts about what people are saying His an exercise of a perceptual capacity" [(1981), p. 332], he makes it clear that the capacity in question is a capacity to hear someone's meaning in his speech. 15 See (1980a), p. 105; (1980a), p. 248; (1980b), p. 573; and (1958a), pp. 175~6. 16 Indeed, Wittgenstein suggests that perceiving an aspect quite generally involves perceiving one thing as related in particular ways to others: "what I perceive in the dawning of an aspect is not a property of the object, but an internal relation between it and other objects" [(l958a), p. 212]. 17 Earlier versions of this material were presented at the 14th Interuniversity Workshop on Philosophy and Cognitive Science, Murcia (March 2004), at a conference in honour of Michael Ayers in Oxford (August 2003), at the Joint Session ofthe Aristotelian Association and the Mind Association in Kent (July 2004), and at seminars and workshops in Oxford and Bertinoro. 1 am grateful to the audiences on those occasions for many helpful comments, and especially to Michael Ayers, Darragh Byrne, John Campbell, Quassim Cassam, Jennifer Church, Anandi Hattiangadi, Robert Hopkins, John Molrowell, Oliver Pooley, Barry Smith, and Galen Strawson.
REFERENCES
AVERS, M. (1991), Locke Volume I: Epistemology, London, Routledge. McDOWELL, J. (1981), "Anti-Realism and the Epistemology of Understanding". Reprinted in McDowell (1998b), pp. 314-43. - (1991), "Intentionality and Interiority in Wittgenstein". Reprinted in McDowell (1998a), pp. 297-321. - (1992), "Meaning and Intentionality in Wittgenstein's Later Philosophy". Reprinted in McDowell (1998a), pp. 263-78. - (1 998a), Mind, Value, and Reality, Cambridge, Mass., Harvard University Press. - (1998b), Meaning> Knowledge, and Reality, Cambridge, Mass., Harvard University Press.
On Having a Meaning Before One's Mind
175
(1998c), "Response to Crispin Wright", in Wright, C., Smith, B. and Macdonald, C. (cds), Knowing Our Own Minds, Oxford, Oxford University Press, pp. 47-62. WITTGENSTElN, L. (l958a), Philosophical Investigations, 2nd edition, edited by G. E. M. Anscornbe, R Rhees and G. H. von Wright, translated by G. E. M. Anscombe, Oxford, Blackwell. (1958b), The Blue and Brown Books, Oxford, Blackwell. (1974), Philosophical Grammar, edited by R. Rhecs, translated by A. Kenny, Oxford, Blackwell. - (1975), Philosophical Remarks, edited by R. Rhees, translated by R. Hargreaves and R. White, Oxford, Blackwell. (1980a), Remarks on the Philosophy ofPsychology, Volume 1, edited by G. E. M. Anscombe and G. H. von Wright, translated by G. E. M. Anscombe, Oxford, Blackwell. - (l980b), Remarks on the Philosophy of Psychology, Volume 2, edited by G. H. von Wright and H. Nyman, translated by C. G. Luckhardt and M. A. E. Aue, Oxford,
Blackwell. C. (1987), "On Making Up One's Mind: Wittgenstein on Intention". Reprinted
WRIGHT,
-
in Wright (2001 a), pp; 116-42. (1991), "Wittgenstein's Later Philosophy of Mind: Sensation, Privacy and Intention". Reprinted in Wright (2001 a), pp. 291-318. (2001a), Rails to Infinity: Essays on Themesfrom Wittgenstein '8 Philosophical Investigations, Cambridge, Mass., Harvard University Press. (2001b), "The Problem of Self Knowledge", in Wright, C. (2001 a), pp. 319-73.
"
teorema Vol. XXVIl, 2006, pp. 176-180
Response to William Child
1. When we consider what Wittgenstein has to say about the issues Child discusses, it is important to distinguish two sorts of question. An example of the first sort is this: "What is it - what does it consist in for a meaning to come to mind, for instance when one arrives at an understanding of the principle of a number series?" Staying with that case, the second sort of question is exemplified by this: "What happens when the principle of a number series comes to mind?" Wittgenstein says plenty in response to that second question, and questions of a similar sort about different topics. A typical passage is Philosophical Investigations §151. He is considering a case in which A begins on a number series, and at a certain moment B says "Now I can go on", or "Now I understand", Assuming B's move is correctly made, what happened at that moment? (It seems harmless to add "in B's mind'") Wittgenstein gives a number of examples of what may have happened. B may have found that a certain algebraic formula yields the initial segment A has given. Or he may have hit on a series of differences between successive elements) a series he already knows how to extend. Or he may have had "the sensation "That's easy!" (compare "a feeling of relief" in §179). And so on. But we can acknowledge that Wittgenstein offers answers (at least answers that give examples) to the second question, white consistently supposing that in his view the first question is misconceived. To hold that the first question is misconceived is to hold that there is no informative answer to the question what it consists in for the principle of a series to come to mind. Of course this is not the same as holding that there is no such thing as having the principle of a series come to mind. One might put the thought by saying that what it is to have the principle of a series come to mind is sui generis and unanalysable. But it is as well to remember that this comes to nothing more than acknowledging that there is such a thing as a principle's coming to mind, while denying that there is an answer to the question what it consists in, apart from the answer that is already expressed in the question: it is the principle's coming to mind. Child gives a rich presentation of material in Wittgenstein on the lines of what I have cited from §151. (He mentions §§151-5 in passing, as something else he could have exploited.) If we firmly distinguish the two questions, this kind of thing is in principle unthreatening to the claim that what it is to have the
176
f I r
I' I:,
us
Response to William Child
177
principle of a number series come to one's mind is, in a proper working out of Wittgenstein's thinking, sui generis and unanalysable, In passages of the sort Child invokes, Wittgenstein is not answering the question what it is for the principle of a series to come to mind. From the fact that he is willing to give examples of kinds of things that happen when the principle of a series comes to mind, nothing follows about his attitude to the other question. 2. However, if having the principle of a number series come to mind is something sui generis and unanalysable, it is surely something that happens on the relevant occasions. But what Wittgenstein offers in the remarks I have been considering, in response to the question "What happens when the principle of a series comes to mind?", seems to be restricted to kinds of happening that are in themselves less, as it were, than that putative happening itself: less than having the principle of the series come to mind. One might suppose there is reason here to think he endorses what Child calls "the two-component view". According to that view, nothing that happens in someone's mind at the relevant moment can itselfbe the person's having the principle of the series come to mind. Having the principle come to mind is a composite of something that happens in the mind (which may be a different something on different occasions on which a principle comes to mind) and something else that does not have the character of an occurrence, in the mind or anywhere else, at all. Perhaps Wittgenstein's view is something on those lines. Perhaps he thinks hitting on the principle of the series is not an occurrence in the mind of someone who hits 011 the principle of the series. Of course that contrasts with thinking it is an occurrence in the mind, but one that is sui generis and unanalysable. We can sharpen this doubt by invoking §154. There Wittgenstein says: "Try not to think of understanding as a 'mental process' at all. For that is the expression which confuses you." His topic is understanding as it figures in B's exclamation "Now I understand". We would not be misinterpreting Wittgenstein's advice if we replaced "understanding" with "the onset of understanding". And for "process" (Wittgenstein's word is "Vorgang") we can substitute "occurrence" or "happening". Wittgenstein's advice is to try not to conceive the onset of understanding as something that happens in the mind at alL And that is certainly hard to reconcile with taking him to think the onset of understanding is a sui generis and unanalysable happening in the mind. 3. I think we can neutralize this doubt We should take the advice of §154 to be local to a specific context, shaped by what would be required in an answer to the other question, the question "What is it for the principle of a series to come to mind?" Once we realize that the other question is not a good question, the advice can lapse.
178
.John McDowell
Tf one thinks there is a good question of the "What is it for...?" or "What does it consist in for. .. T' form about the onset of understanding, one must think there is an informatively different specification of the phenomenon, a specification that could be used to explain what the phenomenon is. It would not be a response to the question, as it is meant, if someone said "It consists in having the principle of the series come to mind". Now suppose one asks what happens when the principle of a series comes to mind, but in asking that question one is asking for an informatively different specification of the phenomenon itself - what one would be asking for if one asked the "What is it for ... T" question. This need not be merely ignoring the difference between the two questions. If one thinks the "What is it for ... '1" question is a good one, running the questions together can be reasonable. It is natural to think what happens when the principle of the series comes to mind is that the principle of the series comes to mind. (If the principle's coming to mind is something that happens, how could that be wrong?) And if there is an explanatory redescription of what that happening is, the redescription should be just as good an answer to the "What happens when ... T" question as the original specification of the happening. So on the assumption that the "What is it for ... T" question is a good one, it should be legitimate to use the "What happens when ... 1" question in order to ask for precisely what the "What is it for ... T" question asks for. On that assumption, then, it seems sensible to ask what happens when the principle comes to mind, but with the expectation of finding an answer that would be informative about what it is for the principle to come to mind. But if we put that requirement on an answer, we exclude answers on the lines of "the onset of understanding of the principle". And now, because our assumption is wrong, we find ourselves in trouble. Nothing we can find in consciousness at the relevant time, with the restriction on what we are allowed to claim to find there that is imposed by the assumption that fixes the point of our asking our question in the first place, amounts to the principle's coming to mind. It is only the beginning of the trouble this lands us in that we are now liable to think what we are looking for the happening itself, the onset of the understanding - must be hidden behind the occurrences in consciousness to which our assumption is restricting us. And that is a palpable confusion. To paraphrase §153: "How can the happening that is the onset of understanding have been hidden, when I said 'Now I understand' because I understood?' And if I say it is hidden then how do I know what I have to look for? I am in a muddle." It is in this context that Wittgenstein issues his advice not to conceive the onset of understanding as a happening in the mind at alL But note that what leads to the trouble is not just trying to find a happening in the mind to be what the onset of understanding is, but conducting the search under the assumption that it will yield a non-trivially different characterization of that
Response to William Child
179
happening. The warning should be against the assumption. The advice not to conceive the onset of understanding as a happening in the mind is local to a context fixed by the assumption. Once we question the assumption, the advice is no longer necessary. We can innocuously find a happening in the mind to be what the onset of understanding is. It is, simply, the onset of understanding. 4. It may seem a stretch to propose that Wittgcnstein intends a reader of
§154 to arrive at that point. If someone suggested that this is one of those places where Wittgenstein has not quite achieved a way of thinking that his reflections point to, I would not resist very strenuously - though I would stand by the point that this is what is actually recommended by the considerations from which he arrives at that advice. In any case, I think we can find some reason to suppose the advice he formulates at §154 is for a temporary and local purpose. Towards the end of part I of the Investigations, Wittgenstein considers what is involved in recalling occasions like the one described in §151, when one was in a position to say "Now I understand" or "Now I can go on". He connects that topic with remembering what one was going to say on some occasion when one was interrupted. In §660 he says: "The grammar of the expression 'I was then going to say ... I is related to that of the expression 'I could then have gone on. "I ("I could then have gone on" is a past-tense counterpart of "Now I can go on'") The question arises what one can find in one's memory of such occasions. If we address this question under a restriction matching the restriction that is in force in the passages that lead up to § 154, where it governs the inquiry into what one can find in one's consciousness on the relevant occasions , what we are allowed to say one finds in one's memory is less than an onset of understanding or an intention to say something. So it comes to seem that one can answer the question what one intended to say only by interpreting scattered fragments that do not add up to the intention itself (see §635). This would be an epistemological position suitable to a two-component view of past intentions - suitable to the idea that nothing that was in one's mind at the time itself amounted to one's intention to say such-and-such. Wittgenstein firmly rejects this. In §634 he says: "1 did not choose between interpretations. I remembered that I was going to say this." And in §660: "In the one case I remember an intention I in the other I remember having understood." He is saying that the intention itself, or, in the case that concerns us, the onset of understanding itself, figures in the deliverances of memory of the relevant past occasions. That seems clearly right. And it is hard to see how it could be prevented from implying that the onset of understanding itself figured among the contents of one's consciousness at the time. It is the contents
John McDowell
180
of one's contemporary consciousness that one calls to mind when one recalls the occasion. In these passages Wittgenstein rejects the assumption that precludes the contents of memory of past occasions from being recognized as including onsets of understanding. A parallel thought is that what is responsible for the trouble of, for instance, § 153 is the assumption that precludes the COI1tents present consciousness from being recognized as including onsets of understanding - not the very idea that onsets of understanding are happenings in the mind.
of
JOHN McDOWELL
j
! j
\
teerema Vol. XXVI1, 2006, pp, 181-192
The Role of Object-Dependent Content in Psychological Explanation Sarah Sawyer
RESUMEN
Este articulo constituye una novedosa defensa de la nocion de sentido de re de McDowell. Se ha pensado que admitir sentidos de re tiene implicaciones problematicas para la explicacion psicologica de 1a accion; implicaciones sefialadas par el llamado "argumento de las dos listas", Argumentare que reconocer que, en general, los conjuntos de condiciones suficientes han de entenderse como incluyendo ausencias sirve para desannar el argumento de las dos listas, Ademas, esta defensa de los sentidos de re puede generalizarse, y servir asf para defender otras formas de antiindividualismo de ataques similares. ABSTRACf
The paper constitutes a novel defence of McDoweU's notion of a de re sense. Countenancing de re senses has been thought to have problematic implications for the psychological explanation of action implications brought out by the so-called "Two List Argument". I argue that recognising that sets of sufficient conditions must in general be understood as including absences disarms the Two List Argument. Moreover, the defence of de re senses presented generalizes and can be used to defend other forms of anti-individualism against a similar line of attack.
1. INTRODUCTION Anti-individualism about psychological kinds is the thesis that certain of a subject's mental states and events are dependent for their individuation on the subject's environment. The thesis opens up the possibility that a subject's mental state and event kinds might vary with variations in the subject's environment, even while her physical properties, including her functional properties and her physical history, all individualistically and non-intentionally described, remain constant. Hence anti-individualism opens up the possibility that physically indistinguishable subjects could nevertheless be distinguishable psychologically. One such anti-individualist theory is the strong singular thought theory (hereafter "SSTT,,)I articulated by Gareth Evans (1982) and John McDowell (1977), (1984) and (1986). At the heart of SSTT is the thesis that singular
181
~r !,
Sarah Sawyer
182
thoughts, thoughts expressed by sentences containing demonstratives and proper names, have object-dependent contents, To maintain that the content of a thought is object-dependent is to maintain that the content would not be available to be thought in the absence of the object which the thought in fact concerns, Thus the ability to think singular thoughts is taken to presuppose the existence of the particular items concerned?' 3 The account which Evans and McDowell propose is a combination of the following two considerations. First, ... the idea that there are things which we say and believe whose content cannot be faithfully represented without the reporter himself making a reference to an object in the world which those utterances and beliefs concern - so that, where there is no such object, there would be no such content available to be faithfully represented .., [Evans (1982), p. 82].
And second, the idea that there needs to be a way in which to distinguish pairs of thoughts which ascribe the same property to the same object, and that only Fregeau sense will suffice for this purpose. The result is an account of singular thoughts as object-dependent, but as containing senses, and not objects, as constituents, There are, in McDowell's terms, de re senses, SSTT has implications for the psychological explanation of action which have been thought problematic. In this paper I address a prima facie compelling argument - the Two List Argument - which draws out the supposedly problematic implications, and find it wanting."
II. THE ESSENTJALLy DE RE To a large extent Evans and McDowell's concern has been to show that a Fregean framework can accommodate singular thoughts understood as essentially de reo To countenance the essentially de re is to maintain a fundamental distinction between the de dicto and the de re in such a way that the latter is not reducible to the former.' McDowell begins his (1984) as follows. It is commonly believed that a Fregean philosophy of language and thought can represent an utterance, or a propositional attitude, as being about an object only by crediting it with a content that determines the object by specification, or at least in such a way that the content is available to be thought or expressed whether the object exists or not. To resist this restriction would be to hold out for the idea that utterances and thoughts can be essentially de re~ and that idea is supposed to be incapable of being made to fit within the framework provided by the theory of sense and Bedeutung [McDowell (1984), p. 98].
The Role ofObject-Dependent Content in Psychological Explanation
183
The reason a Fregean framework is thought unable to accommodate the essentially de re arises, according to McDowell, from a prior conception of the essentially de reo This prior conception informs a number of theories which can be grouped together loosely under the heading "the dual-component theory" (hereafter "DCT"). I take the theory articulated by Tyler Burge as representative." Burge draws two intimately connected distinctions: a semantic distinction between ascriptions of de dicta and de re thoughts, and an underlying epistemic distinction between de dicta and de re thoughts themselves. Suppose that (1) and (2) are ascriptions of a de dicta and a de re belief respectively.
(1) Ortcutt believes that someone is a spy. (2) Someone in particular is believed by Ortcutt to be a spy. On the Burgean conception the logical form of these ascriptions can be given as follows. (1') Bd (Ortcutt, [(x) Spy (x)] ) (2') {x)(Br (Ortcutt, <x>, [Spy (y)] )) "Bd" denotes the de dicta belief relation, "Br" the de re belief relation. The pointed brackets contain a bound variable representing the object presented. The square brackets are intended as corner quotes and are to be regarded as "a convenience for denoting the proposition, or component of proposition, expressed by the symbols they enclose" [Burge (1977), p. 341]. De dicta ascriptions, then, relate the subject to a complete proposition, expressed by a closed sentence; de re ascriptions relate the subject in part to an incomplete proposition, expressed by an open sentence, and in part to a res. The underlying epistemic distinction is between de dicto beliefs that are fully conceptualised, and de re beliefs, "whose correct ascription places the believer in an appropriate nonconceptual, contextual relation to objects the belief is about" [Burge (1977), p. 346]. Perception provides the paradigm example of such a nonconceptual, contextual relation. It is clear how DCT accommodates the essentially de reo But it is equally clear that it does so in a radically unFregean way. According to DCT, de re thoughts contain both a conceptual component and objects. A Fregean theory, in contrast, requires that thoughts contain nothing but senses as constituents. The distinction DCT draws between a de re thought, which is acknowledged to be object-dependent, and the content of that thought, which (at least in the relevant sense)" is not, would be anathema to the view, and Evans and McDowell see it as unmotivated. In contrast, "[ c]ountenancing de re
'~ '1 1
Sarah Sawyer
184
Fregeau senses gratifies both the natural motivations that Burge's framework represents as incompatible; it yields thoughts that are both de re and part of the thinker's cognitive world" [McDowell (1984), p. 109].8
III. DELUDED SUBJECTS McDowell may be correct to think that a Fregean framework can accommodate the essentially de re, and hence that we need not choose between DCT~s conception of the essentially de re and a Fregean view that rejects the essentially de re altogether. However, there is a prima facie compelling line of argument that tells against SSTT and favours DCT over it. In particular, SSTT bas been thought problematic because of its apparent conflict with common-sense views about the role of psychological states in the explanation of action. The argument focuses on the implications of SSTT for deluded subjects; subjects who take themselves to be entertaining a de re thought when there is no appropriately related object. In this section I layout the contrast between SSTI and DCT with regard to deluded subjects. In the following section I turn to the argument against SSTT based 011 its implications for psychological explanation. According to DCT the content of an actual thought can remain constant across various counterfactual possibilities in which the actual object is replaced by an object indiscernible from it, and across various counterfactual possibilities in which there is no object at all. The truth-conditions of the thought would differ as the object thought about was replaced or removed altogether, but that would not be in virtue of a difference in the content of the thought but rather due to a difference in context. In this sense similarity and difference in the way things seem to a subject is intimately related to similarity and difference in the content of the thought she is entertaining." In contrast, since SSrT ties the very content of a thought to the object concerned, the content of a thought, and not merely its truth-conditions, would differ across the kinds of counterfactual possibilities described, even while the differences between the possibilities remain indiscernible from the subject's point of view. Importantly, DCT allows for the possibility that a subject could be entertaining a thought even in the 'absence of an object the thought concerns. According to SSTT, in contrast, if a singular term lacks a referent, an utterance containing that singular term fails to express a thought. Consequently, a deluded subject is not only deluded about whether there is such an object, but is in addition deluded about whether she is entertaining a thought. McDowell writes,
The Role ofObject-Dependent Content in Psychological Explanation
185
... a subject may be in error about the contents of [her] own mind: [she] may think there is a ... thought at, so to speak, a certain position in [her] internal organization although there is really nothing precisely there [McDowell(1986), p. 145].
Following Evans, I will refer to such a deluded subject as entertaining a "mock thought".'? However, 1 do not intend to discuss the strength and source of the intuition that a deluded subject must have some thought before her mind. Instead I note the difference between SSTT and nCT, and turn directly to the issue of psychological explanation.
lV. THE Two LIST ARGUMENT Harold Noonan has argued that reference to strong singular thoughts is never required for the purposes of psychological explanation, and hence that either the adequate psychological explanation of action does not invoke object-dependent thoughts, or it invokes only those object-dependent thoughts whose contents are not object-dependent.!' Noonan's argument purports to establish that there are no strong singular thoughts, and hence that there are no de re senses. He writes, [w]henever an action is directed towards a concrete, contingently existing object, other than its agent, in the sense that it is intentional under a description in which there occurs a singular term denoting that object, then an adequate psychological explanation of it is available under a (possibly distinct) description ill which occurs a term denoting that object; and in this explanation the only psychological states of the agent referred to are ones which would also be present in a counterfactual situation in which the object did not exist [Noonan (1986), pp. 68-9].
Noonan's claim is supposedly established by the following kind of argument. While walking in his garden, Ralph spies the cat he believes to have killed his beloved canary. Angry Ralph lashes out and kicks the cat. If strong singular thoughts are essential for the purposes of psychological explanation, the explanation of Ralph's action must invoke his strong singular thoughts about the cat. Now imagine a counterfactual situation exactly like the actual situation except for the fact that there is no cat: Counterfactual Ralph is subject to a hallucination. Since things seem the same to Counterfactual Ralph as they do to Ralph, we can assume that Counterfactual Ralph will move in the very same way as Ralph in fact moves. That is, Counterfactual Ralph will lash out at what he takes to be the cat in question. This is where the challenge to SSTT comes into play. The challenge is to explain Counterfactual Ralph's behaviour. Noonan writes,
Sarah Sawyer
186
[tlhe [proponent of strong singular thoughts] is thus faced with a dilemma: he must either deny that the behaviour of [Ralph] in the hallucinatory situation is rationally explicable by reference to his contentful psychological states, or he must acknowledge that reference to a proper subset, X, of the thought contents available to [Ralph] suffices to explain [Counterfactual Ralph's] actions [Noonan (1993), p. 286],
The first horn is primajacie unattractive, since Counterfactual Ralph's behaviour does appear to make sense. The second horn, however, is unavailable to SSTT. For consider, in order to avoid the first horn of the dilemma, it must be assumed that Counterfactual Ralph has a set X of beliefs and desires that constitutes a sufficient reason for him to lash out, and hence explains his doing so, But Ralph is Counterfactual Ralph's physical duplicate. Consequently, according to Noonan, Ralph also has this set of beliefs and desires. He may in addition have some strong singular thoughts not shared by Counterfactual Ralph, but the set X constitutes a sufficient reason for his acting as he does, Hence, X is sufficient to explain Ralph's behaviour, and any strong singular thoughts we might have supposed Ralph to have drop out of the picture as explanatorily redundant. But if strong singular thoughts are explanatorlly redundant, there is no reason to think there are such thoughts. Consequently, SSIT is undermined. It is easy to see how DCT avoids the dilemma. According to DCT, the content of an actual thought can remain constant across various counterfactual possibilities including the possibility in which the actual object is absent. Deluded subjects are regarded as being in the same psychological state as their non-deluded counterparts precisely because the content of a de re thought is not regarded as object-dependent. Since psychological content is referenceindependent and psychological explanation invokes psychological content, psychological explanation is itself reference-independent, The problem arises for SSTT because of the claim that sentences containing empty singular terms fail to express thoughts.
V. Two ASSUMPTIONS
The Two List Argument depends upon two crucial assumptions: first, that Ralph and Counterfactual Ralph perform the same actions; and second, that because Counterfactual Ralph has a particular set of beliefs and desires X, any physical duplicate of Counterfactual Ralph also has this set ofpsychological states, 12 The assumptions are intimately related. After all, it is plausible to think that psychological theories appeal to causal roles to individuate psychological states. If it were further assumed that the psychological states of Ralph and his counterpart have the same causal role, there would be reason
The Role ofObject-Dependent Content in Psychological Explanation
187
to think they share a set of psychological states. If, on the other hand, it were assumed that Ralph and his counterpart did not share a set of psychological states, there would be reason to think their respective psychological states did not have the same causal role. Essentially, psychological states and actions are typically individuated in the same way, whether that be individualistically or anti-individualistically. What this shows is that the first of Noonan's assumptions will have little force against a proponent of SSTT. An argument from the claim that twins in general appear to behave in the same ways, since their bodies move through the same physical trajectories, to the claim that they are correctly subsumed by the same psychological theory, will be unpersuasive to an antiindividualist. The fact that twins follow the same physical trajectories through space and time will be thought irrelevant to whether they are subsumed by the same psychological laws, since the psychological properties that feature in laws do not, according to the anti-individualist, supervene locally on physical laws. Physical laws are one thing; psychological laws another. The first of Noonan's assumptions, then, is question-begging. This point has already been made by Burge who, in addition to showing the assumption to be questionbegging, argues persuasively that it is false. 13 Consequently, I leave the first assumption to one side and focus instead on the second.
VI. SUPERVENIENCE RELATIONS The plausibility of the second assumption can be brought out as follows." The anti-individualist is typically committed to the global nomological supervenience of psychological properties on physical properties. From this it follows that there is some minimal set of physical properties which suffice for a subject's possessing a given set of psychological states. Moreover, the minimal set of physical properties sufficient for Counterfactual Ralph's possession of the specific set X is also present in the veridical situation, and hence would suffice for Ralph's possession of the set X, What, then, is wrong with this line of argument? What the argument overlooks is the fact that the absence of a condition can itself be part of the minimal set of physical properties which suffices for possession of a set of psychological states. In this case it is the absence of the cat that constitutes part of the minimal set of physical properties sufficient for the possession by Counterfactual Ralph of X, and this is not duplicated in the veridical situation. Even if we assume that Counterfactual Ralph possesses a set of psychological states that suffice to explain his lashing out, and that everything in Counterfactual Ralph's situation is present in Ralph's situation, we are not entitled to infer that Ralph possesses the same set of psychological states as
188
Sarah Sawyer
Counterfactual Ralph. This is because the absence of the cat from Counterfactual Ralph's environment is itself a relevant factor. It might sound strange to think of an absence as being able to contribute to the conditions sufficient for the instantiation of a set of properties. However, it is a general truth that sufficient conditions for the instantiation of a property must involve absences. This is not an ad hoc move on the part of the externalist. 1 will return shortly to the psychological case. For the moment, here is an argument which I take to be obviously unsound, but which is analogous to the one offered above. Corundum is a mineral with chemical constitution Ah03. In its pure form corundum is completely colourless. However, both rubies and sapphires are derived from corundum and consequently share the same chemical constitution: AhO). The difference between pure corundum, rubies, and sapphires is dependent solely on the presence, or absence, of certain kinds of impurities. Thus a ruby may contain less than 1% chromium, resulting in its distinctive red colour. Similarly, a sapphire may contain traces of titanium and iron, resulting in its being blue. Now suppose there is a minimal set of physical properties sufficient for the instantiation of the property being pure corundum. On the assumed construal of minimal sets of physical properties, any situation in which this set is duplicated is a situation in which the property being pure corundum is instantiated. Consequently, if the argument were sound we should have to conclude that rubies are pure corundum, and similarly that sapphires are pure corundum. After all, the minimal set of physical properties sufficient for the instantiation of the property being pure corundum would not be allowed to include the absence of impurities, and would hence be duplicated both when the property being a ruby and when the property being a sapphire is instantiated. By the transitivity of identity we should also, it would seem, have to conclude that rubies were sapphires. One possibility would be to conclude that there is no minimal set of physical conditions sufficient for the instantiation of the property being pure corundum. There are two ways to understand this claim: (i) as the claim that there is no minimal set of positive physical conditions sufficient for the instantiation of the property being pure corundum; (ii) as the claim that there is no minimal set of physical conditions, positive and negative, sufficient for the instantiation of the property being pure corundum. Neither option would be conducive to the proponent of DCT. The former would amount to conceding the point; the latter would entail a denial of even the global supervenience of non-physical properties on physical properties. IS It would seem, then, that we must allow that the absence of certain conditions can, and often will, itself form part of the minimal set of physically sufficient conditions for the instantiation of a given property. The minimal set of physical conditions that forms the subvenient base for the property of being pure corundum includes the condition that certain impurities not be present. Similarly, the minimal set of
The Role ofObject-Dependent Content in Psychological Explanation
l 89
physical conditions that forms the subvenience base for possessing the set of psychological states X includes the condition that 110 cat be present. If this is right, of course, the argument fails: there is a relevant difference between the veridical and the non-veridical situations, and the set of conditions sufficient for Counterfactual Ralph's possessing the set X is not in fact duplicated in the actual situation. Minimal sets of sufficient conditions must generally be taken to include absences. Before closing this section I will briefly address two potential objections to my argument. The first concerns the particular example on which my claim turns, as follows." The kind ruby, and the kind sapphire are both varieties of the kind corundum. Given this fact, it might be reasonable to conclude that sapphires and rubies are indeed simply instances of pure corundum. However, this objection depends upon a particular micro-reductive theory of kinds which I take to be independently implausible." Rubies, sapphires, and instances of pure corundum are indeed best thought of as instances of distinct kinds. While the distinction may not be recognised from within chemistry, it is clearly recognised from within mineralogy. The question of how to mark out natural kinds, and of how to distinguish the natural from the non-natural kinds, is of course a difficult one. But there is every reason to think that two entities correctly thought of as of the same kind from the point of view of one science might nevertheless be correctly thought of as of distinct kinds from the point of view of another science. If I am right on this, then the kinds ruby, sapphire, and pure corundum can happily be thought of as distinct. The second, modified from a claim made by Segal, concerns the extension of the argument to the psychological case. Taking the absence of a suitable object into account in the possession of a set of psychological states, says Segal, is "out of line with psychological practice. Developmental psychologists do not usually care whether the [singular thoughts] they study are empty or not". Consequently; the position entailed by my argument according to Segal, "coherent, but implausible" [Segal (2000), p. 47]. This, however, does not obviously engage with the considerations I have put forward. First, that psychologists do not care whether the thoughts they study are empty may well show that psychologists can study thoughts without first settling all the questions relevant to the individuation conditions of those thoughts. Whether a given thought-content is individuated individualistically or anti-individualistically is plausibly not a question psychologists need settle in advance of their investigations. Second, and more importantly, the example I present concerning rubies and sapphires demonstrates that sufficient conditions for the instantiation of a natural kind must include absences. Consequently, it is illegitimate to argue from the presence of conditions sufficient in the empty case for possession of a given set of thoughts without taking absences into account to sameness of thoughts across physical twins. IS Minimal sets of sufficient conditions involve absences.
Sarah Sawyer
190
VII. CONCLUSION The Two List Argumentdepends upon two crucial assumptions. The first has already been discredited. Here I have tried to discredit the second. In conelusion, the Two List Argument cannot be used to undermine the existence of de re senses. Department ofPhilosophy University ofN ebraska-Lincoln 1010 Oldfather Hall, Lincoln, NE 68588-0321, USA E-mail: [email protected] NOTES 1 The expression "strong singular thought theory" was introduced by Blackburn in his (1984), chapter 9. 2 Existence is here to be construed atemporally. In addition, 1 do 110t mean to imply that the only constraint SSTT imposes on the ability to think singular thoughts is the existence of the objects concerned. Indeed, this constraint by itself would make little sense. See Sawyer (2001). 3 The class of singular thoughts is not identical to the class of thoughts about particulars. The thought that Quine was a great philosopher is a singular thought, presupposing the existence of Quine; the thought that the Edgar Pierce Chair of Philosophy in 1999 was a great philosopher is not a singular thought, and does not presuppose the existence of Quine. 4 There is reason to think other forms of anti-individualism face the same prima facie worry but that they too can be defended along similar lines. 5 I leave to one side those theories which reject the irreducibility thesis. 6 See Burge (1977) and (1982). For variants see Noonan (1986), (1991) and (1993), Peacocke (1981), and Segal (1989). 7 I say "in the relevant sense" since the content of an object-dependent thought construed as the DC theorist would construe it might nevertheless be object-dependent in virtue of containing externally individuated concepts which are not singular. I have in mind here anti-individualist considerations such as those put forward by Burge in his (1979a). 8 But see Burge (1979b) for an account of the tension between Frege's logical principles and his epistemic motivation for introducing the notion of sense. 9 Again, I leave to one side here issues that arise from an anti-individualist understanding of the propositional content of such de re thoughts, JO Mock thoughts are no more thoughts than toy penguins are penguins. 11 Noonan (1986), (1991) and (1993). Essentially the same argument can be found in Segal (1989). 12 Noonan does not beg the question against anti-individualism by assuming that physical duplicates are psychological duplicates. Noonan assumes the weaker thesis that physical duplicates will share a set of psychological states, and allows that non-deluded duplicates may well have other psychological states in addition.
1.
J
l \
\ \
I I
I I
The Role ofObject-Dependent Content in Psychological Explanation
191
IJ See for example Burge (t989) in which he responds to Fodor's arguments presented in Fodor (1987) chapter 2. 14 The argument I present here is a variant of an argument presented by Gabriel Segal against natural kind externalism in his (2000). The counter argument 1 present in this section draws in large measure on my (2003). For further elaborationsec my (2004). 15 In this connection, and for an extended discussion of related issues, see Corbl and Prades (2000). 1 am grateful to both authors for clarification on this point. 16 Thames to Claire Horisk for pointing out a potential objectionalong these lines. 17 Arguments against such a view abound in the literature. 18 The mistake is, I suspect, due to all illegitimateand yet commonplace tendency to shift between levels of theory: from sameness of kind at a lower level to sameness of kind at a higher level.
REFERENCES BLACKBURN, S. (1984), Spreading the Word, Oxford, Clarendon Press. BURGE, T, (1977), "Belief De Re", The Journal ofPhilosophy, voL 74, pp. 338-62. - (1979a), "Individualism and the Mental", Midwest Studies in Philosophy, vol. 4, pp.73-121 - (1979b), "Sinning Against , The Philosophical Review, vol. 88, pp, 398-432. - (1982), "Other Bodies", in Woodfield, A. (ed.), Thought and Object: Essays on Intentionality, Oxford, Clarendon Press) pp. 97-120. - (1989), "Individuation and Causation in Psychology", Pacific Philosophical Quarterly, vol. 70, pp. 303-22, CORDI, J. E. & PRADES, J. L. (2000), Minds} Causes, and Mechanisms, Oxford, Blackwell Publishers Ltd. EVANS, G, (1982), The Varieties ofReference, edited by J. McDowell, Oxford, Oxford University Press. FODOR, I.A. (1987), Psychosemantics, Cambridge, Mass., MIT Press, McDoWELL,1. (1977), "On the Sense and Reference of a Proper Name", Mind, vol, 86, pp.159-85. - (1984), "De Re Senses", in Wright, C. (ed.), Frege: Tradition and Influence, Oxford, Basil Blackwell. - (1986), "SingularThought and the Extent ofInner Space", in Pettit,P. and McDowell, 1. (eds.), Subject, Thought and Context, Oxford, Clarendon Press, pp. 137-68. NOONAN, H. W. (1986), "Russellian Thoughts and Methodological Solipsism", in Butterfield. J. (ed.), Language, Mind and Logic, Cambridge, Cambridge University Press, pp, 67-90. (1991), "Object-DependentThoughts and PsychologicalRedundancy", Analysis, vol. 51, pp, 1-9. (1993), "Object-Dependent Thoughts: A Case of Superficial Necessity but Deep Contingency", in 1. and Mele, A. (eds.), Mental Causation, Oxford, Oxford University Press, pp. 283-308, PEACOCKE, C. (1981), "Demonstrative Thought and Psychological Explanation", Synvol. 49, pp. 187-217.
192 SAWYRR~
Sarah Sawyer
S. (2001), "The Epistemic Divide", The Southern Journal of Philosophy, vol, 39, pp. 385~40J. - (2003), "Sufficient Absences", Analysis, voL 63~ pp. 202-8. (2004), "Absences, Presences and Sufficient Conditions", Analysis, vol, 64, pp. 354-7. SEGAL, G. (1989), "The Return of the Individual", Mind, vol. 98, pp. 39-57. - (2000), A Slim Book About Narrow Content, Cambridge, Mass., MIT Press.
teorema Vol. XXVII, 2006, pp. 193-196
Response to Sarah Sawyer
1. Obviously Sawyer's project is welcome to me. But I have a couple of quibbles over points of detail, and a more substantial query about her novel contribution.
2. Sawyer aims to defend the Evans-McDowell view about singular thoughts against a certain objection. She formulates the Evans-McDowell view as a combination of two theses: first, that there are object-dependent contents, and second, that if a pair of speech acts, say, with that kind of content ascribe the same property to the same object, it does not follow that the two speech acts have the same content. She represents it as an extra element in the view, over and above the second thesis as I have stated it so far, that only Fregean sense will suffice to enable us to distinguish pairs of contents related in that way. My first quibble is that I do not see this as a further claim. I think the idea ofFregean sense, as applied to the senses expressible by singular terms, just is the idea that pairs of contents related in that way may be distinct; not, as Sawyer's presentation implies, one option, perhaps among others, for enabling ourselves to make such distinctions, 3. Sawyer's target is the "Two List Argument", which purports to show that there are no object-dependent thoughts, on the ground that they are not required for the psychological explanation of behaviour. Making sense of behaviour is the very point of the conceptual apparatus that centres on the notion of thoughts. So if a supposed kind of thought never figures in the explanation of behaviour, it must be mythical. The argument turns on imagining pairs of situations related as follows. In one member of a pair, a proponent of object-dependent thoughts will want to ascribe object-dependent thoughts to a subject. In the other member of the pair, things are, as far as the subject can tell, exactly as they are in the first situation, but the corresponding object-dependent thoughts are not ascribable, because there is no suitable object; the appearance that there is such an object is a hallucination. In the example Sawyer works with, Ralph has a thought he would express by saying "That cat killed my canary", and he kicks at the cat he sees. Twin Ralph would express himself in the same way, and he makes exactly
193
194
John McDowell
matching movements with his legs. But there is 110 cat in his field of view; he thinks there is, but as a result of a hallucination. Any object-dependent thoughts we might want to attribute to Ralph cannot figure in the explanation of Twin Ralph's behaviour, since Twin Ralph does not have them, Twin Ralph's thoughts, which cannot be objectdependent (at least not in the relevant respect: they cannot be dependent on the existence of the cat that is their supposed topic), suffice to account for his lashing out with his legs in the direction of the cat that he thinks he sees. Ralph's action is his lashing out with his legs in exactly that way. So according to the argument, the non-object-dependent thoughts that we are anyway required to attribute to Twin Ralph suffice for the psychological explanation of Ralph's action, and the supposed object-dependent thoughts he is credited with by the Evans-McDowell position do no explanatory work. 4, The most obvious point at which this argument is open to question is this: the argument assumes that what suffices to explain Ralph's action under a description that also applies to Twin Ralph's action say, "trying to kick the cat that he thinks he sees" - exhausts what is needed for explaining what Ralph does. Specifications of what someone does intentionally are themselves specifications of psychological content, the content of intentions that are getting executed. If we see the cat Ralph kicks at, we are in a position to say what he does in object-dependent terms. We can say "He is trying to kick that cat". The perceived presence of the cat makes a contribution to the meaning of this utterance. And we cannot attribute to Twin Ralph the intention we can thus attribute to Ralph, or even an intention that matches it apart from involving a different cat. Reference to a particular cat is an essential element in our specification of the content of that intention of Ralph's. A matching reference to a particular cat is ex hypothesi not possible for specifying any intention of Twin Ralph's. So there is something Ralph intentionally does and Twin Ralph does not do. There is something extra to be explained in Ralph's case. Contrary to what the Two List Argument claims, there is no bar to supposing that psychological states that Ralph does 110t share with Twin Ralph might have explanatory work to do. In Noonan's version, which Sawyer discusses, it is explicit how the argument aims to exploit a certain plausible view of the identity of actions. In our example, this view of action-identity would entail that the action of Ralph's that falls under the object-dependent description we could give if we had the cat in view, "trying to kick that cat", is the very same action that falls under the non-object-dependent description, "trying to kick the cat that he thinks he sees". The second description fits Twin Ralph's action too. As before, Twin Ralph's non-object-dependent psychological states suffice to ex-
Response to Sarah Sawyer
195
plain his action. So they suffice to explain Ralph's action under the nonobject-dependent description that fits it too. But we can grant all that, and insist that it constitutes no argument for supposing that those non-object-dependent psychological states suffice to explain what Ralph does. What we are granting, if we grant all that, is that those non-abject-dependent psychological states suffice to explain Ralph's action under one of its descriptions. But that is not to grant that they suffice to explain his action the identical event, if you like - under another of its descriptions. It is a familiar point that the same event can call for different explanations depending on how it is specified. What is needed, in order to explain what Ralph does, depends on what we say Ralph does when we ask for an explanation that is, on how we describe Ralph's action when we pose the explanatory task. It does not make any difference to this if all the descriptions we might use in order to pose an explanatory task concerning Ralph's action are descriptions of a single event. The explanatory tasks are still different. 5. The considerations 1 have just rehearsed, at greater length than Sawyer does, undermine what she identifies as 111e first of two assumptions on which the Two List Argument depends, the assumption that Ralph and his twin perform the same actions. She credits showing what is wrong with the assumption to Burge, but the point is already in Evans.' Sawyer puts the point in terms of what would be required for Ralph and his twin to be subsumed ,by the same psychological laws. My second quibble is that it is tendentious to suppose that the psychological explanation that constitutes the very point of psychological concepts works by subsuming explananda under laws. In fact I think proper attention to the real-life use of those concepts leaves this thought looking quite implausible. Rejecting the Two List Argument has no need of it. I made no use of it in my sketch of the considerations that undermine the first assumption. 6. Sawyer's novelty is an attack on the second of the two assumptions she finds in the Two List Argument: the assumption that Ralph shares the psychological states that figure in the explanation of Twin Ralph's action. It is common ground that psychological properties supervene on nonpsychologically specifiable properties. Sawyer undermines an inference to the assumption she attacks, from the premise that Ralph is ex hypothesi a duplicate of Twin Ralph in all non-psychologically describable respects compatible with the difference between their situations that the argument turns on. The conclusion does not follow, because the subvenience base for a supervenient property can include absences. Sawyer that the subvenience base for Twin Ralph's relevant psychological states includes the absence of the cat. If that is right, a supervenience thesis yields no ground for supposing those psychological states are shared by Ralph.
John Mcllowell
196
Sawyer's positive point here, that subvenience bases can include absences, seems unquestionable. I think her example, about the subvenience base for the property of being pure corundum, establishes it beyond doubt But I do not find it clear that the point tells against the Two List Argument. That subvenience bases can include absences is not by itself a reason for supposing that the subvenience base for the psychological states of Twin Ralph that the Two List Argument appeals to includes an absence. And in fact that does not seem plausible. What accounts tor Twin Ralph's behaviour is, for instance, his belief that the cat that killed his canary is at a certain position in his field of view. The subvenience base for a belief with that content should consist in facts that leave it open whether or not there is a cat in the believer's field of view, not facts that include there not being a cat in the believer's point of view. The subvenience base for a belief should not include something such that, if the believer knew it, that would destroy the appearance that it is rational for him to believe what he does. So nothing that entails that a belief is false should figure in its subveniencebase. This principle seems reasonable. And it implies that the subvenience base for that belief of Twin Ralph's does not include the absence of a cat. So there is no problem in supposing that Ralph is like Twin Ralph in believing that the cat that killed his canary is at a certain position in his visual field. Of course this is not a defence of the Two List Argument As Sawyer would acknowledge, the wrongness of the first assumption is enough to display it as unconvincing anyway. But I am not persuaded that she has, as she claims, found a different way to refute the argument JOHN McDOWELL
NOTE 1 See
The Varieties ofReference (Oxford: ClarendonPress; 1982), pp. 203-4.
I
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _- L -
teorema Vol. XXV/I, 2006, pp. 197-210
Values vs. Secondary Qualities * Dan L6pez de Sa
RESUMEN
McDowell, respondiendo al argumento a partir de la "rareza" (argumentfrom queerness) de Mackie, defendio el realismo sobre los valores por analogia con las
cualidades secundarias. Se pondra de relieve una tension entre dos interpretaciones posibles de la respuesta de McDowelL De acuerdo con la primera, el realismo sobre los valores se vindicarla, efectivamente, perc a costa de no proporcionar una respuesta apropiada al argumento de Mackie. La segunda interpretacion, sin embargo, properciona una respuesta adecuada a dicho argumento, pero haciendo peligrar e1 realismo
evaluativo.
Assrascr McDowell, responding to Mackie's argument from queerness, defended realism about values by analogy to secondary qualities. A certain tension between two interpretations of McDowell's response is highlighted. According to one, realism about values would indeed be vindicated, but at the cost of failing to provide an appropriate response to Mackie's argument; whereas according to the other, McDowell does provide an adequate response, but evaluative realism is jeopardized.
John Mackie developed a famous argument "from queerness" against there actually being objective values, where an objective good would be sought by anyone who was acquainted with it, not because of any contingent fact that this person, or every person, is so constituted that he desires this end, but just because the end has to-be-pursuedness somehow built into it [Mackie (1977), p. 40J.
In his no less famous response to Mackie, John McDowell urged that not primary qualities, as Mackie supposed, but secondary qualities could provide a suitable model for real evaluative properties: [1Jt seems impossible -
at least on reflection - to take seriously the idea of something that is like a primary quality in being simply there, independently of human sensibility, but is nevertheless intrinsically (not conditionally on contingencies about human sensibility) such as to elicit some "attitude" or state of will from someone who becomes aware of it [McDowell (1985), p. 132].
197
r fl
John McDowell
196
II
Sawyer's positive point here, that subvenience bases can include absences, seems unquestionable, I think her example, about the subvenience base for the property of being pure corundum, establishes it beyond doubt. But I do not find it clear that the point tells against the Two List Argument. That subvenience bases can include absences is not by itself a reason for supposing that the subvenience base for the psychological states of Twin Ralph that the Two List Argument appeals to includes an absence. And in fact that does not seem plausible. What accounts for Twin Ralph's behaviour is, for instance, his belief that the cat that killed his canary is at a certain position in his field of view. The subvenience base for a belief with that content should consist in facts that leave it open whether or not there is a cat in the believer's field of view, not facts that include there not being a cat in the believer's point of view. The subvenience base for a belief should not include something such that, if the believer knew it, that would destroy the appearance that it is rational for him to believe what he does. So nothing that entails that a belief is false should figure in its subvenience base. This principle seems reasonable. And it implies that the subvenience base for that belief of Twin Ralph's does not include the absence of a cat. So there is no problem in supposing that Ralph is like Twin Ralph in believing that the cat that killed his canary is at a certain position in his visual field. Of course this is not a defence of the Two List Argument. As Sawyer would acknowledge, the wrongness of the first assumption is enough to display it as unconvincing anyway. But I am not persuaded that she has, as she claims, found a different way to refute the argument. JOHN McDOWELL
NOTE l
See The Varieties ofReference (Oxford: ClarendonPress, 1982), pp. 203~4.
I
_ _._-~-
---_._--_._--_.
teoreesa VoLXXVIl,2006,pp.197-2JO
Values vs. Secondary Qualities" Dan Lopez de Sa
RESUMEN
McDowell, respondiendo al argumento a partir de la "rareza" (argument from queerness) de Mackie, defendio el realismo sabre los valores por analogia con las cualidades secundarias, Se pondra de relieve una tension entre dos interpretaciones posibles de la respuesta de McDowell. De acuerdo can la primera, eJ realismo sobre los valores se vindicaria, efectivamente, pero a costa de no proporcionar una respuesta apropiada al argumento de Mackie. La segunda interpretacion, sin embargo, proporciona una respuesta adecuada a dicho argumento, pero haciendo peligrar el realismo evaluativo. ABSTRACT
McDowell, responding to Mackie's argument from queerness, defended realism about values by analogy to secondary qualities. A certain tension between two interpretations of McDowell's response is highlighted. According to one, realism about values would indeed be vindicated, but at the cost of failing to provide an appropriate response to Mackie's argument; whereas according to the other, McDowell does provide an adequate response, but evaluative realism is jeopardized.
John Mackie developed a famous argument "from queerness" against there actually being objective values, where an objective good would be sought by anyone who was acquainted with it, not because of any contingent fact that this person, or every person, is so constituted that he desires this end, but just because the end has to-be-pursuedness somehow built into it [Mackie (1977), p. 40].
In his no less famous response to Mackie, John McDowell urged that not primary qualities, as Mackie supposed, but secondary qualities could provide a suitable model for real evaluative properties: [I]t seems impossible at least on reflection - to take seriously the idea of something that is like a primary quality ill being simply there, independently of human sensibility, but is nevertheless intrinsically (not conditionally on contingencies about human sensibility) such as to elicit some "attitude" or state of will from someone who becomes aware of it [McDowell (1985), p, 132].
197
198
Dan Lopez de Sa
Shifting to a secondary-quality analogy renders irrelevant any warty about how something that is brutely there could nevertheless stand in an internal relation to some exercise of human sensibilities. Values are not brutely there - not there independently of our sensibility any more than colours are: though, as with colours, this does not prevent us from supposing that they are there independently of any particular apparent experience ofthem (McDowell (l985), p. 146].
My aim in this paper is to highlight a certain tension between two interpretations of Mcfrowell's response. According to one interpretation, realism about values would indeed be vindicated, but at the cost of failing to provide an appropriate response to Mackie's argument from queerness. According to the other, McDowell would provide an adequate response, but evaluative realism would be jeopardized. Mark Johnston (1989) introduced the notion of response-dependence with the aim of articulating an "analogist" defense of realism about values such as McDowell's: to the extent to which secondary qualities can be regarded as perhaps less than fully objective but genuinely real properties, so can evaluative properties. The notion of response-dependence has generated considerable literature during the last decade. Some philosophers have argued that the notion of a response-dependent concept over-generalizes, by also covering concepts for primary qualities, and hence fails with respect to the project for which it was introduced. Others have argued that precisely for this reason, the original characterization should be modified. They propose an account of a response-dependent property as one which essentially involves the disposition to elicit certain mental responses (in certain subjects under certain conditions). Given the notion of a response-dependent property, a distinction may be drawn between, on the one hand, those which essentially involve the disposition to elicit certain mental responses in certain subjects as they actually are under certain conditions as they actually are, and on the other, those which essentially involve the disposition to elicit certain mental responses in certain subjects whatever they are like under certain conditions whatever they are like. I will call them rigid vs. flexible response-dependent properties. This distinction will be crucial for my claim about the two contrasting and conflicting interpretations of McDowell's response. I shall argue that according to one, values and secondary qualities are rigid response-dependent properties, whereas according to the other, they are both flexible response-dependent properties. This paper is in four sections. In section I, I briefly present the notion of a response-dependent property, and the distinction between rigid vs, flexible response-dependent properties. In section II, I present the two interpretations of Mclfowell's response, which exemplify this distinction. In section III, I
Values vs. Secondary Qualities
199
present Mackie's argument from queerness and argue that although the view of values as rigid response-dependent properties does qualify as a realist proposal, it fails to respond to the argument. In section IV, the final section, I argue that the view of values as flexible response-dependent properties does not face the queerness problem, but has relativist consequences that vindicate a non-realist position about the evaluative.
1.RESPONSE-DEPENDENT PROPERTIES: RiGID VS. FLEXIBLE In his "Dispositional Theories of Value" (1989), Johnston attributes to McDowell an "analogist' response against anti-realist arguments, whose "leading idea ... has been to show that by the same standards of genuineness it would follow that colour is 110t a genuine feature of surfaces" [Johnston (1989), p. 139], and introduced the notion of response-dependence as a means of stating the relevant analogy: The most plausible, if highly way of taking the analogy is this: evaluational concepts, like secondary quality concepts as understood by the analogists, are "response-dependent" concepts [Johnston (1989), p. 144].
Since then response-dependence has usually been characterized by means of conditions on certain biconditionals. Let us say a response-dependence biconditional (rd biconditional for short) for a (predicate signifying a certain) property F is a substantial biconditional of the form x is f iff x has the disposition to produce In subjects S the mental response R under conditions C; or the form x is f iff subjects S have the disposition to issue the x-directed mental response R under conditions C, where "is f" signifies F, and "substantial" is there to avoid "whatever-ittakes" specifications of either S, R or C. (One such "whatever-it-takes" specification of, say, subjects S would be "those subjects, whatever they are like, such that something is disposed to produce in them responses R under conditions C iff it is F." Mutatis mutandis for the responses and the conditions.) Johnston's own characterization of a response-dependent concept required only that there was one such biconditional (for a predicate expressing
200
Dan Lopez de Sa
it) holding a priori. As I mentioned, some philosophers, including Frank Jackson and Phillip Pettit (2002), have argued that this original characterizaHan over-generalizes, by also covering concepts for primary qualities. Very briefly, the key element in the arguments is this: regardless of the primary VB. secondary nature of the signified property, there will be descriptive material associated with the predicate, playing at least a reference-fixing-role which, in the cases at hand, will easily involve the relevant responses. Statements of them can be cashed in the form of I'd biconditionals, which for familiar Kripkean reasons will hold a priori. Furthermore, given the possibility of rigidifying on the relevant subjects and conditions, the notion would over-generalize in the same way if one further requires that the rd biconditionals hold necessarily as well as a priori. Some philosophers, including Manuel Garcia-Carpintero (2002) and Ralph Wedgwood (1998),1 have argued that precisely for this reason, the original characterization of "response-dependent" should be modified. Dwelling on the ideas of Kit Fine on essence (1994), they have independently proposed an account of a response-dependent property as one which essentially involves the disposition to elicit certain mental responses (in certain subjects under certain conditions). That Socrates belongs to singleton Socrates holds necessarily, but not in virtue of the nature of Socrates (but, presumably, of the set). That Plato is distinct from Aristotle again holds necessarily, but not in virtue of the nature of Plato (but, presumably of both Plato and Aristotle). Similarly, in the case of primary, fully objective, properties, the (perhaps rigidified) rd biconditionals might hold necessarily, but not in virtue of the nature of the properties. And when they do hold in virtue of the nature of the property, the property is response-dependent, as in the case of secondary qualities. In other words: (RD) A property F is response-dependent iff there is an rd biconditional for (a predicate signifying) it which holds a priori and in virtue of the nature of F. rd biconditionals, as characterized so far, may contain rigidifying devices. Let us say that a specification of the subjects in an rd biconditional is rigid iff the relevant predicate involved in the specification is rigid, andflexible otherwise. So take for instance "human who fails no discrimination test passed by other human subjects." This is not, as it stands, a rigid specification. For take the relevant predicate "is a human who fails no discrimination test passed by other human subjects" and suppose that in the actual world, it is true (even if knowable only a posteriori) that being a human who fails no discrimination test passed by other human subjects is being a human with a perceptual apparatus meeting condition A. Now consider a counterfactual situation in which, due to whatever reason you might think of, humans who
Values vs. Secondary Qualities
201
fail no discrimination test passed by other human subjects are those with a perceptual apparatus meeting the different condition B. Now intuitively, it is this other property of being a human with a perceptual apparatus meeting condition B which would be relevant for evaluating sentences containing "is a human who fails no discrimination test passed by other human subjects') with respect to this other world. But then "is a human who fails no discrimination test passed by other human subjects" is not a rigid predicate. Its relevant rigidification, which can be put as something like "is a human who fail no discrimination test passed by other human subjects, as they actually are» leads, nonetheless, to a rigid specification of the subjects, of the sort of "humans who fail no discrimination test passed by other human subjects) as they actually are.,,2 An rd biconditional is rigid iff it involves a rigid specification of the subjects, and it isjlexible otherwise. Now we can draw the distinction that will be crucial in what follows: A response-dependent property is rigid iff the rd biconditionals for it holding in virtue of its nature are rigid. A response-dependent property is flexible iff there is a flexible rd biconditional for it holding in virtue of its nature. Any response-dependent property is rigid or flexible but not both. Rigid response-dependent properties are dispositions to produce in certain (rigidly specified) subjects certain responses under certain (rigidly specified) conditions; flexible response-dependent properties, by contrast, are properties whose extensions, in each possible world w, are those things which have in w the disposition to produce in certain subjects, as specified with respect to w, the relevant response under certain conditions, as specified with respect to w. Suppose that "is f" signifies" a response-dependent property F, with response R, and suppose that Sand C are relevant flexible specifications of subjects and conditions, and S@ and C@ their relevant rigidifications, and that the only relevant rd biconditionals are (R) x is fiffx is disposed to produce in S@ the response R under conditions C@. (F) x is f iff x is disposed to produce in S the response R under conditions C. Both are, we may suppose, true with respect to the actual world and, we may also suppose, knowable a priori. But the following asymmetry arises: (abstracting now from issues about essence vs. necessity) their metaphysical status co-varies with the nature ofF as stated in
Dan Lopez de Sa
202
F is a rigid response-dependent property iff (R) is necessary (i.e., iff (F) is contingent);
F is a flexible response-dependent property iff (F) is necessary (i.e., iff (R) is contingent). This provides a way of testing whether "Is f" signifies a rigid or a flexible response-dependent property, and based just on a priori considerations. The recipe, very abstractly put, is this: consider what could be a counterexample of the necessity of the relevant statement on the assumption that the predicate signifies one particular kind of property, neutrally described. 1 will refer to them as target situations. Then check how these should be intuitively described, with respect to the relevant predicate, and conclude accordingly."
II. Two INTERPRETATIONS OF McDoWELL'S RESPONSE As we are about to see, McDowell's views concerning values and secondary qualities in his response to Mackie can be stated as asserting their response-dependence, in the general sense characterized by (RD). Thus there are two interpretations of his response, corresponding to the rigid and the flexible variety of response-dependence. The discussion of the difference between them in the subsequent sections will, I hope, also vindicate the relevance of the distinction. According to McOowell, \
a secondary quality is a property the ascription of which to an object is not adequately understood except as true) if it is true, in virtue of the object's disposition to present acertain sort of perceptual appearance: specifically, an appearance characterizable by using a word for the property itself to say how the object perceptually appears. Thus an object's being red is understood as something that obtains in virtue of the object's being such as (in certain circumstances) to look, precisely) red [McDowell (1985), p. 133].
This is to assert that secondary qualities are response-dependent properties in our general sense of (RD), provided that being a property the ascription of which to an object is not adequately understood except as true, if it is true, in virtue of the objects disposition to respond in a certain way amounts to being a property such that it is in virtue of its nature that objects to which it is (truly) ascribed do have the disposition to respond in a certain way.
Values vs. Secondary Qualities
203
I think that the crucial further claims that McDowell makes in his response confirm this attribution, and stating them also helps to illustrate the general notion of a response-dependent property further. In my own words: if His red" signifies a response-dependent property, being red, then commonsensical predications of it, of the sort "this rock is red," are: (i) evaluable as true or false (for the most part); (ii) some of them indeed true; (iii) some of them indeed lrnowably true. Furthermore: (iv) being red is subjective in the sense that it depends on the responses as entailed by its response-dependence; but (v) being red is not subjective in the sense of making all occurrences of the responses automatically correct. (i)-(v) hold provided that being red is response-dependent, regardless of whether it is a rigid or a flexible response-dependent property. Let us label the view according to which secondary qualities and values are rigid response-dependent properties disposittonalism, and the view according to which they are flexible response-dependent propertiesflexibilism. [ can now reformulate my main aim in this paper as that of highlighting a tension between the dispositionalist and the flexibilist interpretation of Mclzowell's response to Mackie's argument from queerness, to which now I turn;
III. THE RIGID CASE:
THE PROBLEM WITH THE PRACTICALITY OF THE EVALUATIVE
As I understand it, Mackie's argument from queerness aims to establish an incompatibility between values being real properties and what is sometimes called the "practicality of the evaluative," that values, whatever they are, are "internally" connected to motivation, i.e. have "to-be-pursuedncss" somehow built into them: An objective good would be sought by anyone who was acquainted with it, not because of any contingent fact that this person, or every person, is so constituted that he desires this end, but just because the end has to-be-pursuedness somehow built into it [Mackie (1977), p. 40].
I propose to state this internalist claim about values thus: (I) It is necessary and a priori that: If something is good, then we would desire it (under appropriate reflective conditions, weakness of will and the like aside). Several remarks are in order. First, this is an internalist claim about values themselves, and not an internalist claim about, say, evaluative judgments,
204
Dan Lopez de Sa
judgments to the effect that certain things. are good or not, evaluations for short, to which people sometimes also refer with the same label. The connection between the two internalisms is, at best, complex, partly because the second is indeed a family of quite different claims. In any case, Mackie is clearly concerned with the "action-guiding" character that values themselves should have, which according to him would make them "queer," if they were real. Second, it might be argued that the necessary (and a priori knowable) charactel' of the connection falls short of capturing what was behind the traditional idea of values and motivation being "internally" connected, perhaps the notion of essence should be invoked here instead. This might be right, but it does not affect the present discussion, as the necessity of a statement is at least clearly a consequence of its holding in virtue of the nature of some entity. Third, strong as the connection between values and motivation is held to be, it can not be absolute, as the frequent cases of weakness of will, to name just one of the most famous ones, illustrate. This is what justifies a parenthetical clause like that in (1), obviously being taken with a pinch of salt. I do not mean to suggest that there might not be important difficulties in the vicinity, in so far as a sensible and satisfactory explicit formulation of internalism about values is concerned (see, among others, Johnson (1999) and subsequent discussion). Fortunately, a specification at the level of elaboration already provided will suffice, I think, for the considerations to come. Fourth, last but not least, I am interpreting Mackie's "to-be-pursuedness" as requiring that values would be desired (under certain conditions), and not (merely) that they should be, or that desiring them would be appropriate. r think this is indeed a fair interpretation: The need for an argument of this sort [the argument from queerness - DLdS] can be brought out by reflection on Hume's argument that "reason" ... can never be an "influencing motive of the will" [Mackie (1977), p. 40].
(According to some, in my view plausible but controversial, views on the issue, the "prescriptive" claims would follow, at least in some central cases, from the corresponding "descriptive" ones.) If the reality of values were modeled by fully objective primary qualities, then internalist claims of the (1) sort would clearly be false. What is crucial for me here is to argue that, for essentially the same kind of reason, claims like (1) would still be false even if the reality of values were modeled by the, in some sense subjective, secondary qualities, if they are understood as rigid response-dependent properties. For the sake of vividness, let me focus on one particular, somehow Lewisian,' dispositional proposal about values. According to this view, the following holds a priori and in virtue of the nature of goodness:
Values vs. Secondary Qualities
(L@)
205
X is good iff we, as we actually are, are disposed to value x under appropriately reflective conditions;
where valuing is the favorable attitude of desiring to desire, "appropriately reflective conditions" are spelled out as those of the fullest possible imaginative acquaintance, and "we" refers to a population consisting of the speaker and those relevantly like him. (We can assume that to be relevantly like a given subject is to be disposed, with respect to valuing the relevant thing in question in the relevant conditions, exactly how the subject is disposed.) Thus understood "we" turns out to be aflexlble characterization of a group of subjects. It "is relevantly like me" actually picks out the property of being relevantly the way I am actually. But I could be otherwise, and in particular my disposition to value particular things could be very different from what it actually is. But then, with respect to those worlds in which I am suitably different, "is relevantly like me" signifies the property of being relevantly the way I would be in those situations. ("We, as we actually are," is the rigidification of "we," as understood here, and hence a rigid specification.) Dispositionalism about values, of the considered Salt, cannot account for the truth of (I). The reason is straightforward: provided that dispositions to value particular things are obviously contingent, the view does entail that the followingflexible biconditional is (if true) merely contingent: (L) x is good iff we are disposed to value x under appropriately reflective conditions. But any counterexample to the necessity of (L) is such that the embedded conditional in (I) is false with respect to it. Hence the embedded conditional (I) is not necessary, and hence (1) is false. The argument easily with respect to any dispositional account of values, as it only depends on the tension between the contingency of the relevant dispositions to elicit the "evaluative" responses VB. the necessity of the practicality requirement. (See Holland (2001) for further discussion.) This consideration, even if sound, quite obviously, fails to refute dispositional accounts (or evaluative realism in general)," Evaluative objectivists, claiming that evaluative properties fully objective and not response-dependent, typically argue explicitly against anything along the lines of (1) and go externalist about values, by holding something like:
(E) It is contingent and a priori that: If something is good, then we would desire it (under appropriate reflective conditions, weakness of wilt and the like aside). Or, equivalently,
),
Dan Lopez de Sa
206
(I@) It is necessary and a priori that: If something is good, then we, as we actually are, would desire it (under appropriate reflective conditions, weakness of will and the like aside). Here is what David Brink says: [T]he internalist cannot rest content with the extensional claim that everyone is in fact motivated (by what is morally good]. Any externalist could claim that The internalist about motives claims that it is true in virtue of the concept of morality that [moral goodness] necessarily motivates. According to the internalist, then, it must be conceptually impossible for someone to [know that something is morally good] and remain unmoved [Brink (1986), pp. 29-30].7
That requires, of course, rejecting Mackie's point at the beginning of this section concerning values being practical. It is not part of my aim to argue against externalism here: it suffices to observe that according to the present interpretation of McDowell's response, although values certainly are real though in some sense subjective properties, Mackie's argument is still in force. Shifting to a secondary-quality analogy, as understood now, does not render irrelevant worries about how values stand in an internal relation to some exercise of human sensibilities.
IV. THE FLEXIBLE CASE: EVALUATIVE REALISM VS. EVALUATIVE RELATIVISM
Flexibilism about values, by contrast, straightforwardly accounts for the truth of internalist claims of the (1) sort. With respect to the particular Lewisian proposal, flexibilism would hold that the following holds a priori and in virtue of the nature of goodness: (L) x is good iff we are disposed to value x under appropriately reflective conditions.
As Lewis himself states, something like (1) is indeed a consequence of the proposal: If something is a value, and if someone is of the appropriate "we", and if he is in ideal conditions, then it follows that he will value it. And if he values it, and if he desires as he desires to desire, then he will desire it [Lewis (1989), p. 72].
Values vs. Secondary Qualities
207
The problem now is that flexibilism has relativistic consequences and hence falls short of constituting a realist position about values. The main idea behind relativism, [ take it, is that there are "essentially contestable" claims, in the domain in question. Following Crispin Wright (1992), one might say that relativism has it that it is conceivable that there are (irremovable) faultless divergences, in the domain, not constituted by anyone's being in the error in judging something false. With a little more detail: let us say that two subjects diverge in their judgments with respect to a given sentence-type iff they master it, one of them has a judgment she could express, in an ordinary situation, by uttering a token of that sentence with its conventional meaning, whereas the other has a judgment she could express, in an ordinary situation, by uttering a token of the negation of that sentence with its conventional meaning. So assume that 1'm tired but you're 110t: we both diverge in our judgments concerning "I'm tired" in the intended sense. Let us focus on the particular case of simple sentences: predications of a given predicate, i.e., the completion of the predicate by a singular term (or a singular definite descriptionj.f Let us say that this divergence in judgments with respect to a simple predication of a given predicate is irremovable iff it is not explainable in terms of (0 the contextdependence of the singular term (or description) expression-type; (ii) vagueness or other kinds of indeterminacies; or (iii) facts independent of the subjects in question. To illustrate: our previous divergence in our judgments concerning "I'm tired" is excluded by (i); our possible divergence in our judgments concerning "Fifl is pink'), provided Fill is a borderline case of pinkness, is excluded by (ii); and a divergence in judgments with "water covers more than half the Earth's surface" between me and my counterfactual self in a situation in which the Earth is almost dry is excluded by (iii). Finally, let us say that one such (irremovable) divergence is faultless if no-one is being thereby in error ofjudging something false. I propose to characterize relativism with respect to a domain thus: (R) Relativism concerning a predicate holds iff it is conceivable that two subjects irremovably faultlessly diverge in their judgments concerning a simple predication of the predicate. As suggested, I think that (R) captures well enough the intuition the traditional antirealist tries to exploit, according to which certain claims in a discourse are essentially contestable. (Relativism so characterized consists in there being conceivable irremovable divergences. This does not entail that those divergences actually occur. But it does not preclude it either. Factual relativism concerning a given predicate can be seen as the claim that the sort of irremovable divergences whose conceivability establishes relativism concerning it actually occur.)
208
Dan Lopez de Sa
Now flexible response-dependence does entail relativism, so conceived, and hence, in particular, that values are flexible response-dependent properties vindicates evaluative relativism. The argument for the claim that if a predicate signifies a flexible response-dependent property, then relativism concerning it, in the sense of (R), follows, is indeed quite direct: any target situation whose proper intuitive description favors the flexible responsedependence of its signification contains the materials for a suitably irremovable divergence. 9 In SUfi, the second interpretation of McDowell's response can be claimed to satisfactorily respond to Mackie's argument from queerness, but it does entail a form of relativism about values. One might think that this is indeed the right interpretation. After all, in the discussed paper McDowell does say: "I can see no reason why we should not regard the contentiousness [of values] as ineliminable" [McDowell (1985), p, 145, fn. 46]. What about his explicit statements concerning the reality and genuineness of values? They would certainly be jeopardized, provided that realism is understood in a sufficiently exigent sense so as to be incompatible with relativism (and hence flexibility) although not requiring full objectivity (and hence counting dispositions to elicit mental responses as real). This is indeed, I think, a sensible sense, and it is the one I have taken for granted in this paper. But, sensible or not, it is certainly not the only conceivable sense. My guess is that McDowell might require less than this for reality, so that values qualify as genuine features in virtue of the relevant evaluative predications being truth-apt (against non-cognitivism), and some of them being true (against error-theorismj'" - and knowably so (against skepticism). This last remaining question seems to me to be merely about words and, in a way, temperament
LOGOS - Grup de Recerca en Logica, Llenguatge i Cognicio Universitat de Barcelona
Arche -AHRC Research Centrefor the Philosophy ofLogic, Language, Mathematics and Mind University of St Andrews St Andrews, Fife KY16 9AL, Scotland E-mail: [email protected]
NOTES
~ Related material was presented at the LOGOS GRG (Barcelona 2003) and at the VI Taller d'Investigacio en Filosofia (Tarragona 2004). I would like to thank the
Values vs. Secondary Qualities
209
audiences present at these occasions, as well as participants in the XIV InterUniversity Workshop on Philosophy and Cognitive Science: John McDowell (Murcia 2004), for fruitful discussion. 1 am particularly indebted to Nestor Casado, Josep Corbt, Esa Dfaz-Leon, Manuel Garcla-Carpintero, John McDowell, Kevin Mulligan, and Josep L. Prades for helpful suggestions and objections. The research presented in this article is supported in the framework of the European Science Foundation EUROCORES programme "The Origin of Man, Language and Languages", the research project BFF2003-08335-C03-03 (Spanish Government), the research group 2001SGROOOl8 (Catalan Government), and my postdoctoral grant EX2004-1159 (Spanish Government), and I would like to express my gratitude to these institutions. 1 And, more recently, also Mark Johnston (2004). 2 I am assuming, in common with Kripke (1980), and many others in discussions on philosophy of mind, philosophy of science or metaethics, that the notion of rigidity might be extended to be applicable to predicates, roughly along the lines of: a predicate is rigid iff it signifies the same property in all relevant worlds. Proposals like this have recently received criticisms, among which: that it would trivialize, making all predicates trivially rigid, and that in any case it would over-generalize, counting as rigid some predicates that do not signify natural properties/kinds. 1 try to respond to these criticisms, respectively, in my unpublished "Rigidity for Predicates and the Trivialization Problem" and "Predicates Rigidly Signifying the 'Unnatural. '" In the latter I also argue that the relevant simple predicates like those that will concern us here, "is red," "is funny;' "is good" and the like are, nonetheless, rigid. Given this ( will speak of them signifying properties, without rclativizing such talk to worlds. 3 See previous footnote. 4 This section contains numerous (here pertinent) simplifying assumptions, the removal of which requires substantial elaboration. This is done in the first part of Lopez de Sa (2003). S For details, see Lewis (1989) and Lopez de Sa (2003). My view is that Lewis himself, however, would probably favor - as I would also do - the flexibilist position to be considered in the next section. 6 In my own view, it does crucially contribute to the case against realism about values, when placed in an appropriate and broader context. I hope to elaborate on this elsewhere. 7 In the original passage, instead of the inserted claims about moral values Brink makes claims about moral considerations and judgments, but I take it that he would certainly concur with what I say about properties and facts. 8 The reason for so doing is the following: a relativism that could be attacked merely by pointing to (the obviously not "essentially contestable") cases of "if that is good then it is good" is not worth considering. Silly as the observation may be, I think drawing attention to it serves to dissolve most of the usual claims which have it that relativism is, somehow, "self-refuting:" statements of the semantic features of the relevant predicates, which eventually entail that all simple predications are "essentially contestable," need not be themselves "essentially contestable." 9 For further details see again Lopez de Sa (2003). An evaluative relativism ofthis sort, it is often said, contradicts a basic platitude regarding conversations concerning the evaluative: ordinary participants are committed to regard utterances of "that is good" and "that is not good" as (literally) contradicting each other. Here is what Wright says:
210
Dan Lopez de Sa
"If [Indexical RelativismJ were right, there would be an analogy between disputes of inclinations and the 'dispute' between one who says