Applied Linguistics, volume 31, issue 2 , 2010

APPLIN-31(2)Cover.qxd 3/30/10 5:54 PM Page 1 APPLI ED LI NGUISTICS ISSN 0142-6001 (PRINT) ISSN 1477-450X (ONLINE) ...

Author: Oxford University

31 downloads 1240 Views 2MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form

DOWNLOAD PDF

APPLIN-31(2)Cover.qxd

3/30/10

5:54 PM

Page 1

APPLI ED LI NGUISTICS

ISSN 0142-6001 (PRINT) ISSN 1477-450X (ONLINE)

Applied Linguistics

Volume 31 Number 2 May 2010


Published in cooperation with AAAL American Association for Applied Linguistics AILA International Association of Applied Linguistics BAAL British Association for Applied Linguistics

OXFORD

www.applij.oxfordjournals.org


3/30/10

5:54 PM

Page 2

EDITORS

NOTES TO CONTRIBUTORS

Ken Hyland, Director, Centre for Applied English Studies, KK Leung Building, The University of Hong Kong, Pokfulam Road, Hong Kong Jane Zuengler, Nancy C. Hoefs Professor of English, University of Wisconsin-Madison, 6103 Helen C. White 600 North Park Street Madison, WI, 53706 USA Assistant to Jane Zuengler: Heather Carroll, University of Wisconsin-Madison

Articles submitted to Applied Linguistics should represent outstanding scholarship and make original contributions to the field. The Editors will assume that an article submitted for their consideration has not previously been published and is not being considered for publication elsewhere, either in the submitted form or in a modified version. Articles must be written in English and not include libelous or defamatory material. Manuscripts accepted for publication must not exceed 8,000 words including all material for publication in the print version of the article, except for the abstract, which should be no longer than 175 words. Additional material can be made available in the online version of the article. Such additions will be indexed in the print copy.

REVIEWS AND FORUM EDITOR Stef Slembrouck, Professor of English Linguistics and Discourse Analysis, Universiteit Gent, Vakgroep Engels, Rozier 44, B-9000 Gent, Belgium. <[email protected]> Assistant to Stef Slembrouck: Tine Defour, Universiteit Gent

ADVISORY BOARD Guy Cook, British Association for Applied Linguistics Aneta Pavlenko, American Association for Applied Linguistics Martin Bygate, International Association for Applied Linguistics Huw Price, Oxford University Press

EDITORIAL PANEL Karin Aronsson, Linko¨ping University David Block, London University Institute of Education Jan Blommaert, University of Jyva¨skyla¨ Deborah Cameron, University of Oxford Lynne Cameron, Open University (BAAL Representative) Tracey Derwing, University of Alberta Zoltań Do¨rnyei, University of Nottingham Patricia Duff, University of British Columbia Diana Eades, University of New England, Australia ZhaoHong Han, Columbia University (AAAL representative) Gabriele Kasper, University of Hawai’i at Manoa Claire Kramsch, University of California at Berkeley Angel Lin, City University Hong Kong Janet Maybin, Open University, UK Tim McNamara, University of Melbourne Junko Mori, University of Wisconsin-Madison Greg Myers, Lancaster University Susanne Niemeier, University Koblenz-Landau (AILA Representative) Lourdes Ortega, University of Hawai’i at Manoa Alastair Pennycook, University of Technology, Sydney Ben Rampton, King’s College, University of London Steven Ross, Kwansei Gakuin University Alison Sealey, University of Birmingham Antonella Sorace, University of Edinburgh Lionel Wee, National University of Singapore Applied Linguistics is published five times a year in February, May, July, September and December by Oxford University Press, Oxford, UK. Annual subscription price is £254/US$457/E381. Applied Linguistics is distributed by Mercury International, 365 Blair Road, Avenel, NJ 07001, USA. Periodicals postage paid at Rahway, NJ and at additional entry points. US Postmaster: send address changes to Applied Linguistics (ISSN 0142-6001), c/o Mercury International, 365 Blair Road, Avenel, NJ 07001, USA. # Oxford University Press 2010 All rights reserved; no part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise without prior written permission of the Publishers, or a licence permitting restricted copying issued in the UK by the Copyright Licensing Agency Ltd, 90 Tottenham Court Road, London W1P 9HE, or in the USA by the Copyright Clearance Center, 222 Rosewood Drive, Danvers, MA 01923. Typeset by Glyph International, Bangalore, India Printed by Bell and Bain Ltd, Glasgow

Applied Linguistics operates a double-blind peer review process. To facilitate this process, authors are requested to ensure that all submissions, whether first or revised versions, are anonymized. Authors’ names and institutional affiliations should appear only on a detachable cover sheet. Submitted manuscripts will not normally be returned. Forum pieces are usually reviewed by the journal Editors and are not sent for external review. Items for the Forum section are normally 2,000 words long. Contributions to the Forum section and offers to review book publications should be addressed to the Forum and Reviews Editor. For more detailed guidelines, see our website http://www.oxfordjournals.org/applij/for_authors/index.html

PROOFS Proofs will be sent to the author for correction, and should be returned to Oxford University Press by the deadline given.

OFFPRINTS On publication of the relevant issue, if a completed offprint form has been received stating gratis offprints are requested, 25 offprints of an article, forum piece or book review will be sent to the authors free of charge. Orders from the UK will be subject to a 17.5% VAT charge. For orders from elsewhere in the EU you or your institution should account for VAT by way of a reverse charge. Please provide us with your or your institution’s VAT number.

COPYRIGHT Acceptance of an author’s copyright material is on the understanding that it has been assigned to the Oxford University Press subject to the following conditions. Authors are free to use their articles in subsequent publications written or edited by themselves, provided that acknowledgement is made of Applied Linguistics as the place of original publication. Except for brief extracts the Oxford University Press will not give permission to a third party to reproduce material from an article unless two months have elapsed without response from the authors after the relevant application has been made to them. It is the responsibility of the author to obtain permission to reproduce extracts, figures, or tables from other works.

Applied Linguistics Journal online The full text of Applied Linguistics is available online to journal subscribers. Online access has a number of advantages: . quality PDFs ensure articles look the same as the print original and are easy to print out . access is easy—all you need is your subscription number or institutional IP address (see below) . online access is available ahead of print publication—so view while you await your print version! . access the text wherever you are (or from any part of your institution network if you have a library subscription) . perform searches by word or author across the full text of the articles of any part of the journal . download articles whenever you choose—you will be able to access past online issues as long as you have a current subscription . free sample copy available online . fully searchable abstracts/titles going back to volume 1 . Table of Contents email alerting service. The print version will continue to be available as previously. Institutions may choose to subscribe to the print edition only, online only, or both. Individual subscribers automatically receive both.

CONTRIBUTORS There is no need for contributors to format their articles any differently; online files are produced automatically from the final page proofs of the journal. However, if you know that an item in your list of references is available online, please supply the URL. If you have your own website, you are welcome to include the URL with your contact address in your biodata.

ADVANCE ACCESS Applied Linguistics now has Advance Access articles. These are papers that have been copyedited and typeset but not yet paginated for inclusion in an issue of the journal. More information, including how to cite Advance Access papers, can be found online at http://www.applij.oxfordjournals.org.

Applied Linguistics Subscription Information

A subscription to Applied Linguistics comprises 5 issues. Annual Subscription Rate (Volume 31, 5 issues, 2010) Institutional. Print edition and site-wide online access: £254.00/US$457.00/E381.00; Print edition only: £233.00/US$419.00/E350.00; Site-wide online access only: £212.00/US$382.00/E318.00.

Oxford University Press, Great Clarendon Street, Oxford OX2 6DP, UK. Email: [email protected]. Tel (and answerphone outside normal working hours): +44 (0)1865 353907. Fax: + 44 (0)1865 353485. In Japan, please contact: Journals Customer Services, Oxford Journals, Oxford University Press, Tokyo, 4-5-10-8F Shiba, Minato-ku, Tokyo 108-8386, Japan. Tel: þ81 3 5444 5858. Fax: þ81 3 3454 2929.

Personal. Print edition and individual online access: £82.00/US$164.00/E123.00.

subscribe to applied linguistics

Please note: US$ rate applies to US & Canada, Euros applies to Europe, UK£ applies to UK and Rest of World.

For new subscriptions and recent single issues only. Current subscribers will automatically receive a renewal form.

Prices include postage by surface mail, or for subscribers in the USA and Canada by airfreight, or in India, Japan, Australia and New Zealand, by Air Speeded Post. Airmail rates are available on request. There are other subscription rates available for members of AAAL, BAAL, AILA, and LSA, for a complete listing please visit www.applij.oxfordjournals.org/subscriptions. Full prepayment, in the correct currency, is required for all orders. Orders are regarded as firm and payments are not refundable. Subscriptions are accepted and entered on a complete volume basis. Claims cannot be considered more than FOUR months after publication or date of order, whichever is later. All subscriptions in Canada are subject to GST. Subscriptions in the EU may be subject to European VAT. If registered, please supply details to avoid unnecessary charges. For subscriptions that include online versions, a proportion of the subscription price may be subject to UK VAT. Personal rate subscriptions are only available if payment is made by personal cheque or credit card and delivery is to a private address. The current year and two previous years’ issues are available from Oxford Journals. Previous volumes can be obtained from the Periodicals Service Company at http://www.periodicals. com/oxford.html or Periodicals Service Company, 11 Main Street, Germantown, NY 12526, USA. Email: [email protected]. Tel: +1 (518) 537 4700. Fax: +1 (518) 537 5899. For further information, please contact: Journals Customer Service Department,

Please complete the form below and return it to: Journal Customer Service Department (please see above). Please record my subscription to Applied Linguistics, starting with Volume__________ (Subscriptions start with the March issue and can be accepted for complete volumes only.) Please send me the following single issue(s) Volume_________ Issue_________ Name (BLOCK CAPITALS please) _________________________________________ Adresss __________________________________ _________________________________________ _________________________________________ City _____________________________________ Country _________________________________ Postcode _________________________________ I enclose the correct payment of (see rates above): £/US/E __________________________________ Please debit my credit card: American Express / Mastercard / Visa (delete as appropriate) Card number: __|__|__|__|__|__|__|__|__|__|__|__|__|__|__|__|__ Expiry date: |__|__|__| Signature ________________________________ œ Please tick this box if you do NOT wish to receive details of related products and services of OUP and other companies that we think may be of interest.

Aims Applied Linguistics publishes research into language with relevance to real-world problems. The journal is keen to help make connections between fields, theories, research methods, and scholarly discourses, and welcomes contributions which critically reflect on current practices in applied linguistic research. It promotes scholarly and scientific discussion of issues that unite or divide scholars in applied linguistics. It is less interested in the ad hoc solution of particular problems and more interested in the handling of problems in a principled way by reference to theoretical studies. Applied linguistics is viewed not only as the relation between theory and practice, but also as the study of language and language-related problems in specific situations in which people use and learn languages. Within this framework the journal welcomes contributions in such areas of current enquiry as: bilingualism and multilingualism; computer-mediated communication; conversation analysis; corpus linguistics; critical discourse analysis; deaf linguistics; discourse analysis and pragmatics; first and additional language learning, teaching, and use; forensic linguistics; language assessment; language planning and policies; language for special purposes; lexicography; literacies; multimodal communication; rhetoric and stylistics; and translation. The journal welcomes both reports of original research and conceptual articles. The Journal’s Forum section is intended to enhance debate between authors and the wider community of applied linguists (see Editorial in 22/1) and affords a quicker turnaround time for short pieces. Forum pieces are typically responses to a published article, a shorter research note or report, or a commentary on research issues or professional practices. The Journal also contains a Reviews section. Applied Linguistics is covered by the following abstracting/indexing services: Bibliographie Linguistique/Linguistic Bibliography, BLonline, British Education Index, Current Index to Journals in Education, ERIC (Education Resources Information Centre), International Bibliography of the Social Sciences, ISI: Social Sciences Citation Index, Research Alert, Current Contents/Social and Behavioral Sciences, Social Scisearch, Sociological Abstracts: Language and Linguistics Behaviour Abstracts, Language Teaching, MLA Directory of Periodicals, MLA International Bibliography, PsycINFO, Sociological Abstracts, Zeitschrift fu¨r Germanistische Linguistik.

ADVERTISING Inquiries about advertising should be sent to Linda Hann, Oxford Journals Advertising, 60 Upper Broadmoor Road, Crowthorne, RG45 7DE, UK. Email: [email protected]. Tel/Fax: +44 (0)1344 779945.

PERMISSIONS For information on how to request permissions to reproduce articles/information from this journal, please visit www.oxfordjournals.org/permissions.

DISCLAIMER Statements of fact and opinion in the articles in Applied Linguistics are those of the respective authors and contributors and not of Applied Linguistics or Oxford University Press. Neither Oxford University Press nor Applied Linguistics make any representation, express or implied, in respect of the accuracy of the material in this journal and cannot accept any legal responsibility or liability for any errors or omissions that may be made. The reader should make his/her own evaluation as to the appropriateness or otherwise of any experimental technique described.


3/30/10

5:54 PM

Page 2

EDITORS

NOTES TO CONTRIBUTORS

Ken Hyland, Director, Centre for Applied English Studies, KK Leung Building, The University of Hong Kong, Pokfulam Road, Hong Kong Jane Zuengler, Nancy C. Hoefs Professor of English, University of Wisconsin-Madison, 6103 Helen C. White 600 North Park Street Madison, WI, 53706 USA Assistant to Jane Zuengler: Heather Carroll, University of Wisconsin-Madison

Articles submitted to Applied Linguistics should represent outstanding scholarship and make original contributions to the field. The Editors will assume that an article submitted for their consideration has not previously been published and is not being considered for publication elsewhere, either in the submitted form or in a modified version. Articles must be written in English and not include libelous or defamatory material. Manuscripts accepted for publication must not exceed 8,000 words including all material for publication in the print version of the article, except for the abstract, which should be no longer than 175 words. Additional material can be made available in the online version of the article. Such additions will be indexed in the print copy.

REVIEWS AND FORUM EDITOR Stef Slembrouck, Professor of English Linguistics and Discourse Analysis, Universiteit Gent, Vakgroep Engels, Rozier 44, B-9000 Gent, Belgium. <[email protected]> Assistant to Stef Slembrouck: Tine Defour, Universiteit Gent

ADVISORY BOARD Guy Cook, British Association for Applied Linguistics Aneta Pavlenko, American Association for Applied Linguistics Martin Bygate, International Association for Applied Linguistics Huw Price, Oxford University Press

EDITORIAL PANEL Karin Aronsson, Linko¨ping University David Block, London University Institute of Education Jan Blommaert, University of Jyva¨skyla¨ Deborah Cameron, University of Oxford Lynne Cameron, Open University (BAAL Representative) Tracey Derwing, University of Alberta Zoltań Do¨rnyei, University of Nottingham Patricia Duff, University of British Columbia Diana Eades, University of New England, Australia ZhaoHong Han, Columbia University (AAAL representative) Gabriele Kasper, University of Hawai’i at Manoa Claire Kramsch, University of California at Berkeley Angel Lin, City University Hong Kong Janet Maybin, Open University, UK Tim McNamara, University of Melbourne Junko Mori, University of Wisconsin-Madison Greg Myers, Lancaster University Susanne Niemeier, University Koblenz-Landau (AILA Representative) Lourdes Ortega, University of Hawai’i at Manoa Alastair Pennycook, University of Technology, Sydney Ben Rampton, King’s College, University of London Steven Ross, Kwansei Gakuin University Alison Sealey, University of Birmingham Antonella Sorace, University of Edinburgh Lionel Wee, National University of Singapore Applied Linguistics is published five times a year in February, May, July, September and December by Oxford University Press, Oxford, UK. Annual subscription price is £254/US$457/E381. Applied Linguistics is distributed by Mercury International, 365 Blair Road, Avenel, NJ 07001, USA. Periodicals postage paid at Rahway, NJ and at additional entry points. US Postmaster: send address changes to Applied Linguistics (ISSN 0142-6001), c/o Mercury International, 365 Blair Road, Avenel, NJ 07001, USA. # Oxford University Press 2010 All rights reserved; no part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise without prior written permission of the Publishers, or a licence permitting restricted copying issued in the UK by the Copyright Licensing Agency Ltd, 90 Tottenham Court Road, London W1P 9HE, or in the USA by the Copyright Clearance Center, 222 Rosewood Drive, Danvers, MA 01923. Typeset by Glyph International, Bangalore, India Printed by Bell and Bain Ltd, Glasgow

Applied Linguistics operates a double-blind peer review process. To facilitate this process, authors are requested to ensure that all submissions, whether first or revised versions, are anonymized. Authors’ names and institutional affiliations should appear only on a detachable cover sheet. Submitted manuscripts will not normally be returned. Forum pieces are usually reviewed by the journal Editors and are not sent for external review. Items for the Forum section are normally 2,000 words long. Contributions to the Forum section and offers to review book publications should be addressed to the Forum and Reviews Editor. For more detailed guidelines, see our website http://www.oxfordjournals.org/applij/for_authors/index.html

PROOFS Proofs will be sent to the author for correction, and should be returned to Oxford University Press by the deadline given.

OFFPRINTS On publication of the relevant issue, if a completed offprint form has been received stating gratis offprints are requested, 25 offprints of an article, forum piece or book review will be sent to the authors free of charge. Orders from the UK will be subject to a 17.5% VAT charge. For orders from elsewhere in the EU you or your institution should account for VAT by way of a reverse charge. Please provide us with your or your institution’s VAT number.

COPYRIGHT Acceptance of an author’s copyright material is on the understanding that it has been assigned to the Oxford University Press subject to the following conditions. Authors are free to use their articles in subsequent publications written or edited by themselves, provided that acknowledgement is made of Applied Linguistics as the place of original publication. Except for brief extracts the Oxford University Press will not give permission to a third party to reproduce material from an article unless two months have elapsed without response from the authors after the relevant application has been made to them. It is the responsibility of the author to obtain permission to reproduce extracts, figures, or tables from other works.


3/30/10

5:54 PM

Page 1

APPLI ED LI NGUISTICS

ISSN 0142-6001 (PRINT) ISSN 1477-450X (ONLINE)

Applied Linguistics



Published in cooperation with AAAL American Association for Applied Linguistics AILA International Association of Applied Linguistics BAAL British Association for Applied Linguistics

OXFORD

www.applij.oxfordjournals.org

APPLIED LINGUISTICS Volume 31 Number 2 May 2010 CONTENTS Articles Sex/Gender, Language and the New Biologism DEBORAH CAMERON The Contribution of Written Corrective Feedback to Language Development: A Ten Month Investigation JOHN BITCHENER and UTE KNOCH Probabilities and Surprises: A Realist Approach to Identifying Linguistic and Social Patterns, with Reference to an Oral History Corpus ALISON SEALEY Lexical Diversity in Writing and Speaking Task Performances GUOXING YU Discourse Particles in Corpus Data and Textbooks: The Case of Well PHOENIX W. Y. LAM An Extended Positioning Analysis of a Pre-Service Teacher’s Better Life Small Story GARY BARKHUIZEN

173

193

215 236 260

282

FORUM Input Quality Matters: Some Comments on Input Type and Age-Effects in Adult SLA JASON ROTHMAN and PEDRO GUIJARRO-FUENTES

301

REVIEWS Hannes Kniffka: Working in Language and Law: a German perspective HELEN KELLY-HOLMES

307

Nancy H. Hornberger (ed.): Can Schools Save Indigenous Languages? Policy and Practice on Four Continents JULIA SALLABANK E. Alcoń Soler and Alicia Martıńez Flor (eds): Investigating Pragmatics in Foreign Language Learning, Teaching and Testing GILA A. SCHAUER

309

313

Jeff Siegel: The Emergence of Pidgin and Creole Languages ANDREI A. AVRAM

316

NOTES ON CONTRIBUTORS

320

Readers During 2008/2009

323

Applied Linguistics: 31/2: 173–192 ß Oxford University Press 2009 doi:10.1093/applin/amp022 Advance Access published on 31 May 2009

Sex/Gender, Language and the New Biologism DEBORAH CAMERON Oxford University, UK

INTRODUCTION This article considers some ideas about language and sex/gender,1 which are currently influential in both expert and popular discourse. Advocates of what I am calling ‘the new biologism’ (the term is discussed further below) contest the belief that male–female behavioural differences are effects of social and cultural processes. They argue instead that many such differences are biologically based, produced by evolutionary processes which have led the two sexes to differ not only in the obvious physical ways, but also in their cognitive abilities, their psychological dispositions, and consequently their habitual ways of behaving. This is not just an argument about language, but language features prominently among the phenomena to which it is applied. Here I will consider the new biologism’s claims by way of a critical examination of the arguments and the evidence. The recent resurgence of biologism in debates on sex/gender is one manifestation of a larger ‘Darwinian turn’ in the study of human behaviour, thought, and culture. The study of language and gender has not been directly affected by this development: though theoretical overviews of the field (e.g. Bing and Bergvall 1996; Bergvall 1999; McElhinny 2003) often do discuss the issue of biological sex, the point is usually to distinguish sex from gender, and to reaffirm the primacy of the latter. This research community, in short, remains committed to socio-cultural approaches.2 But in the larger academic/ scientific community, the new biologism is gaining ground. High-profile commentators like Steven Pinker (2002) cite new discoveries in rapidly advancing fields such as genetics and neuroscience to argue that biologism is now the ‘cutting edge’ approach to sex/gender, while socio-cultural approaches are

Downloaded from applij.oxfordjournals.org by guest on December 31, 2010

In recent years there has been a striking shift in both academic and popular discourse on the subject of male–female differences. It is increasingly common for biological explanations to be proposed for differences that had previously been treated by most investigators as effects of socio-cultural factors. This article critically examines the arguments as they apply to the specific case of male– female differences in linguistic behaviour. It concludes that the relevant linguistic research evidence does not on balance support the new biologism; that evidence is more adequately accounted for using the socio-cultural approaches which most linguistic researchers favour.

174 SEX/GENDER, LANGUAGE AND THE NEW BIOLOGISM


outmoded and discredited, maintained more for reasons of ‘political correctness’ than because the evidence supports them. Applied linguists do not dispute the general principle that scientists must be willing to follow the evidence. But in the spirit of science we should also be willing to ask what the evidence actually is, and whether it really leads where it is said to lead. In the case of language and sex/gender, how well do the new biologism’s substantive claims fit with the findings of linguistic research? How compelling is the argument that those findings are more convincingly explained in biological than socio-cultural terms? For applied linguists with their focus on language in the ‘real world’ there are reasons to broach these questions that go beyond the purely academic. The new biologism’s influence is not felt only in the academy: its arguments about sex/gender now dominate popular folklinguistic discourse, and are increasingly being taken up in the spheres of policy and professional practice. In the US, for instance, claims about innate differences between boys and girls are currently fuelling a campaign for single-sex classrooms, the idea being not just to teach girls and boys separately, but to educate them differently, with curriculum content and teaching methods tailored to suit what are alleged to be their distinctive intellectual capacities and learning styles (Gurian et al. 2001; Rivers and Barnett 2007). Supporters of this move make much of the idea that girls are innately endowed with superior verbal abilities. Boys, they say, are disadvantaged by the emphasis modern schooling places on language and literacy; they need both different language pedagogy and a less languagecentred curriculum. Elsewhere, though these ideas may not be driving educational policy, they are influencing attitudes in ways that have the potential to affect outcomes. One recent Australian study (Carr and Pauwels 2006) found both teachers and pupils invoking biological sex-differences and the notion of ‘brain sex’ to explain why girls were good at language subjects and boys were not. If we accept that educational attainment can be affected by societal expectations, and particularly by the expectations of teachers, then we might well have concerns about the current popularity of what amounts to a biologically based deficit model of boys’ verbal abilities. That model is no less in need of critical interrogation than the other verbal deficit claims (for instance, about working-class non-standard dialects, Black Englishes and ‘semilingualism’) to which linguists have turned their attention in the past. By now, readers may well be asking precisely what—or who—I mean by ‘the new biologism’. I should acknowledge that this is not a label which some particular group of scholars uses to identify itself. Rather it is an umbrella-term which I have chosen to apply to an interdisciplinary and somewhat heterogeneous collection of current scholarly enterprises, on the basis that they share certain fundamental assumptions and preoccupations, and conduct their arguments within certain parameters. It is those assumptions, preoccupations and parameters which I take to define ‘the new biologism’ as what some theorists might call a ‘discursive field’. Within that field, an important

D. CAMERON

175

shared assumption is that the most powerful explanations of the way humans think, feel, and act are those which appeal to the principles of evolutionary theory. Probably the most important source for explanatory stories of this kind—and as such, an important reference-point for the new biologism—is evolutionary psychology (hereafter ‘EP’). While few of the scholars whose work I discuss below would describe themselves as evolutionary psychologists, all of them make use of EP’s Darwinian narrative. At this point, then, I will turn to EP, setting out some of its principles and going on to consider how these are applied to the subject of language and male–female differences.

Evolutionary psychology has been described as a new ‘science of human nature’. Its central claim is that the human mind, like the human body, is a product of Darwinian evolution. On that basis it contests what its adherents represent as the hegemonic position among social scientists and laypeople, that the mind at birth is a ‘blank slate’, waiting to be inscribed by experience and socialization. Rather what humans have is an ‘adapted mind’ (Barkow et al. 1992), predisposed to develop in particular ways that proved advantageous for survival among our early human ancestors. The word ‘predisposed’ is used advisedly here: popular writers may use terms like ‘(hard) wired’ and ‘(genetically) programmed’, but most reputable scholars disclaim the strong determinism that language implies (e.g. Dunbar et al. 2007: 6–7). While such disclaimers are not unproblematic (Rose 2005), EP should not be crudely characterized as an approach that attributes everything to ‘nature’ (in the form of genes) while giving no weight to ‘nurture’ (environmental/cultural influences). On the contrary, it seeks to move beyond that dichotomy by arguing that the adaptations which produced human nature resulted from the demands made on humans by their physical and social environment interacting with biological mechanisms. The biological mechanism which is central to this story is natural selection, the process whereby genetic traits which confer a survival advantage (‘survival’ in modern evolutionary theory meaning the survival of an organism’s genes when they are passed on to offspring) gradually spread through the population. But what traits are advantageous depends on the conditions in which the organism must survive. For EP, the conditions which are relevant for our understanding of human nature are not the ones most humans live in now, but the ones in which our species evolved many millennia ago. We modern humans inherit our genes from ancestral peoples who roamed the plains of Africa in small bands of hunter-gatherers: our nature is therefore shaped by the requirements of that way of life. This general reasoning underpins the argument that cognitive and psychological differences between men and women (including the specifically linguistic ones which I will consider in more detail below) are not cultural


EVOLUTIONARY PSYCHOLOGY



epiphenomena, but part of our genetic inheritance. It is assumed that among early humans male and female roles were sharply differentiated: women gathered and men hunted, women nurtured infants and men fought other men. This made it adaptive for the two sexes to have different skills and traits (e.g. for women to be more empathetic and men more aggressive). In addition, the sexes are and always will be differentiated by their reproductive biology. For men, whose essential contribution to the reproductive process is fertilizing the ovum, the number of offspring who carry their genes is closely related to the frequency with which they mate. They are therefore advantaged by traits which enable them to maximize their mating opportunities by either persuading women to choose them over other men (e.g. because they are ‘better providers’) or more directly overcoming their rivals. Women, on the other hand, can only produce a limited number of offspring: for them what matters most is ensuring those they do produce survive. They are therefore advantaged by traits like empathy and nurturance which enable them to select mates judiciously and care effectively for dependent children. EP’s evocation of the lives of early humans is necessarily extrapolated from very limited direct evidence. On some questions (e.g. about our ancestors’ physiques and their diet, their tools and their visual art), conclusions may be drawn from the evidence of fossil remains and preserved artifacts. There are other questions, however, on which that evidence sheds no light, because they concern forms of behaviour which do not leave material traces. Questions about our pre-literate ancestors’ ways of using language are clearly in this category, as are questions about sexual behaviour and parenting practices. The stories EP constructs about these matters must therefore be based on more indirect evidence. The approach for which EP is best known involves identifying some trait or behaviour-pattern which is widely attested in modern human populations: on the assumption that its current prevalence (especially if it is prevalent in otherwise unrelated cultures) is most parsimoniously explained by assuming it is inherited from ancestral humans, the investigator then sets out to construct an explanation of what made it adaptive for our ancestors. It is this way of proceeding which has led critics to charge EP with telling ‘Just So Stories’. However, relevant evidence may also come from genetics, from the study of the brain and of human developmental processes, from research on nonhuman primates and from anthropological work with modern hunter-gatherer peoples, who are held to provide the closest directly observable parallel to ancestral human groups. In practice, discussions of language and sex/gender difference tend to rely on two main sources of evidence. One is the laboratory-based work of neuro- and psycholinguists investigating male–female differences in verbal ability and/or the functional organization of language in the brain. The other is research— often conducted originally by socio- or applied linguists and linguistic anthropologists—on naturally occurring verbal behaviour among men and women in

D. CAMERON

177

modern speech communities. Together, these bodies of work are taken to yield two large-scale generalizations: (a) That one sex (in most versions of the argument females, but in some versions males) is innately endowed with superior verbal abilities and a greater predisposition towards verbal communication. (b) That the two sexes differ in their typical modes of verbal interaction: men favour more competitive speech styles and genres, while women are more co-operative, empathetic, and nurturant.

SEX/GENDER AND THE EVOLUTION OF LANGUAGE There is a long history of scholarly speculation on the origins of language, and questions of sex/gender have often featured in these discussions. For instance, a once-popular story about what made the possession of a language faculty sufficiently valuable to humans to offset the cost of the large brain needed to accommodate it suggested that it enabled men to co-ordinate joint activities such as hunting and warfare more efficiently. But although there are still some theorists (e.g. Crow 1998, 2005; Miller 2000) who believe (for other reasons) that language originated in men, others now prefer accounts which suggest that it developed first in women (Knight 1990; Dunbar 1996; Joseph 2000). This preference often goes along with the view that the main adaptive value of language related to the important function (given that humans are social animals) of creating and maintaining group cohesion. Robin Dunbar (1996) argues that linguistic exchange (more specifically, the exchange of social information about community members’ activities, relationships, status, and moral worth) served the same purposes grooming serves for non-human primates. As changing environmental conditions made it desirable for the size of early human groups to increase, one-to-one physical grooming became unsustainably time-consuming. Language permitted social information to be exchanged, and relationships negotiated, in larger groups and


These modern generalizations are adduced as supporting evidence for the larger theory that the most reproductively successful females among our ancestors had highly developed social (including communication) skills, an innate disposition to nurture and the ability to co-operate with others, while the most successful males were effective competitors who relied more on practical and non-verbal skills. At the same time, that theory is said to offer the deepest and most powerful explanation for the modern findings. In the following discussion, I will suggest that both the generalizations themselves and the conclusions drawn from them are questionable: they are based on a very selective reading of the evidence, and in some instances also on linguistically naı¨ve or tendentious interpretations of it. I will begin, though, by considering two accounts of language and sex/gender differences which have been proposed by evolutionary scholars, and placing these in the larger context of ongoing debates about the evolution of language itself.



while engaged in other tasks. In Dunbar’s view, this account favours the hypothesis that women rather than men were the prime movers. Among non-human primates, social networks are generally centred on females. Dunbar also believes, as do other theorists who place similar emphasis on social networking/group cohesion, that his account gains support from the superior verbal skills and greater disposition to communicate which have frequently been attributed to modern human females (Halpern 1992; Kimura 1999). Scholars who take this view of language evolution have offered various proposals about how female verbal superiority evolved. One focuses on the sexual division of labour that is assumed to have prevailed among our ancestors. Joseph (2000) suggests that the habitual engagement of early human females in child-rearing, gathering, and the fashioning of domestic objects contributed to the functional evolution of the language areas of the brain, whereas men’s engagement in hunting blocked that development (gatherers and mothers can, as Joseph puts it, ‘chatter’ while they work or care for infants, but hunters must remain silent if they are not to scare off their prey). A related suggestion is that hunting developed men’s visual–spatial skills, one side-effect of which was to leave them with less neural capacity to devote to linguistic functions. But there are problems with this line of argument. Apart from the fact that modern anthropological research with hunter-gatherer peoples (Dahlberg 1981; Lee and Daly 1999) suggests that ‘men hunt, women gather’ is a considerable oversimplification, the argument relies heavily on the notion that modern women are more talkative than men; and despite being common folk-wisdom, this idea is not supported by any reputable evidence. Systematic studies using a variety of methods and measures overwhelmingly contradict it, showing that in informal peer-interaction there is typically no sex/gender difference, while in formal and status-marked situations it is most commonly men who talk more (James and Drakich 1993; Mehl et al. 2007). A similar problem arises with the claim that women favour co-operative and emotionally nurturant styles of speech whereas men are competitive and status-oriented. Unlike the female loquacity thesis, this generalization has had some support from linguistic scholarship: it was championed by the ‘difference’ current of language and gender research which was influential (though never uncontested) during the 1980s (Maltz and Borker 1982; Sheldon 1990; Tannen 1990). Difference researchers proposed that adult linguistic gender-norms reflect habits acquired in childhood, when girls and boys socialize mainly with peers of their own sex. Boys form large, hierarchical groups with leaders and followers; girls’ groups are smaller, with a looser and more egalitarian structure. Boys play competitive games with complex rules; girls play simpler games where everyone gets a turn. These differing group structures and activities teach group members to use language in different ways. Boys acquire a competitive, status-oriented communication style: they learn to argue, boast, criticize, give, and receive orders. Meanwhile, girls

D. CAMERON

179


acquire a more co-operative and supportive style: they learn to agree, praise, empathize, make suggestions and resolve disputes. These observations are not incompatible with an explanation in terms of innate characteristics, though that inference was not made by the original researchers. The problem, however, is that more recent research has cast doubt on the empirical validity of the difference generalizations, particularly about girls. Marjorie Harness Goodwin (2006) spent three years regularly videotaping interactions among a group of pre-adolescent girls in Los Angeles and observed that they did all the things girls had been said by difference researchers not to do. Their group had an internal hierarchy (with the lowest-status participant, a peripheral individual dubbed the ‘tag-along girl’, subjected to regular bullying by the others). When they played hopscotch or jump-rope, their goal was to win or to outshine their peers, and they frequently argued about the rules. They also engaged in regular boasting—about their skills, their possessions and the relative wealth of their families—and in issuing orders, to one another and to boys. Though Goodwin’s study stands out for the quantity and quality of the data it is based on, it is not the only case in which researchers have observed pre-adolescent and adolescent girls behaving in these ways (Eckert 1996; Baxter 2006). So far I have concentrated on the view that language evolved primarily to facilitate the (co-operative and putatively female-centred) activity of social networking. There is, however, an alternative account which emphasizes the advantages of language as a tool for courtship—a more competitive activity in which it is assumed that males took the lead. A key difference between the two accounts is that while the first centres on the mechanism of natural selection (where selected traits are those which make their possessors better adapted to survive in the prevailing environmental conditions) the second allots a central role to sexual selection, the concept which explains why traits may be selected despite conferring no obvious survival advantage. Anomalies such as the peacock’s tail (which appears to militate against survival by making the bird more conspicuous to predators) can be explained by positing that they made individuals who possessed them either more attractive to the other sex, or more successful in competing with their own sex for access to the other. This enabled them to mate more often, and so produce more offspring with the same characteristic. Sexually selected characteristics (e.g. large horns or elaborate plumage) are often found in males, and this reflects the fact that in many species it is males who do the courting while the role of females is to choose among potential mates. Peacocks, for instance, engage in ‘lekking’—ritually displaying themselves in areas frequented by peahens. Some scholars think that language fulfils analogous functions among humans. Geoffrey Miller (1999, 2000), for instance, argues that human languages are much more elaborate than they need to be to serve purely communicative purposes. This can be explained by hypothesizing that speaking served the purpose of displaying the (male) speaker’s reproductive fitness. Dunbar (1996), who believes that language evolved primarily to facilitate


EXPLAINING SEX/GENDER VARIATION: THE LIMITATIONS OF BIOLOGISM A basic assumption made by all parties to the debate is that the phenomena they discuss—instances of sex/gender-related variation in linguistic behaviour—can in principle be explained in terms of inherited biological traits. Yet as Derek Bickerton (2006) points out in a response to Locke and Bogin, this entails conflating what are arguably two different things: language itself and the uses to which it may be put. Few linguists dispute that there is a biological basis for the mental faculty which enables all developmentally normal humans to produce grammatical speech, but many would join Bickerton in questioning whether such applications of that faculty as gossiping


social networking, agrees that it may also have developed a secondary function as a means for men to advertise themselves to women. There is an obvious tension between this account and the orthodox wisdom which holds that women are the more talkative and verbally skilful sex. If the courtship story is correct, then men would logically have had more to gain than women from traits like loquacity, articulacy and fluency. Furthermore, on the principle that modern humans inherit the genes of the most successful reproducers among their ancestors, this male verbal advantage should still be observable. Supporters of the sexual selection account must therefore deal with the objection that the evidence points in the other direction. That objection is addressed in an article by John Locke and Barry Bogin (2006), who raise doubts about the strength of the modern psycholinguistic evidence for female verbal superiority, pointing out that meta-analytic studies have found the overall effect of sex/gender to be slight (Hyde and Linn 1988; Hyde 2005). More radically, though, these scholars question whether modern psycholinguistic findings are relevant. What bearing, they ask, can studies measuring performance on structured tests (usually requiring subjects to respond to a written stimulus under classroom-like conditions) have on debates about the innate abilities of our unschooled and pre-literate ancestors? More pertinent, they suggest, are the findings of ethnographic research in nonliterate and vernacular cultures, especially regarding oral public performance (a human analogue of lekking). These findings lead to quite different conclusions about which sex has more advanced linguistic abilities: ‘anthropological research . . . reveals that performative applications of language in the form of speech and voice consistently favor males’ (Locke and Bogin 2006: 270, emphasis in original). This internal debate among evolutionary scientists is revealing in two ways. Most obviously, it suggests a lack of consensus on some fairly fundamental issues. Yet it is also striking how much agreement there is on some of the underlying assumptions which frame the debate—assumptions which, from a linguist’s perspective, are far from self-evident.

D. CAMERON

181


or telling stories are themselves part of our genetic endowment. If they are not, though, then it is surely a category mistake to propose an evolutionary explanation for them. Why do scholars apparently assume that findings about sex/gender differences in the use of language are in principle susceptible to this kind of explanation? Are they suggesting there may be a gene for storytelling or a part of the brain dedicated to gossip? In most cases, no; but there is, perhaps, a clue to their reasoning in the argument about which sex is endowed with superior verbal skills. The significance accorded to this issue by both camps, even though their positions are opposed, suggests they share the assumption that observed differences in men’s and women’s linguistic behaviour must be expressions of underlying differences in the two groups’ verbal abilities. This might enable them to answer Bickerton by saying that what their account directly explains is not the behavioural differences themselves, but the innate cognitive difference which is assumed to give rise to them. But there are pressing objections to the idea that surface variation in the use of language must arise from ‘deeper’ differences in verbal ability. Students of linguistic variation do not deny that some individual language-users may be more skilful than others, but they do not link larger-scale sociolinguistic patterns to the relative verbal abilities of different groups within the population: distributional analysis suggests that the important links are to social factors such as speakers’ relative power and status, their differential access to resources like education and jobs, and their varying involvement in local networks and communities of practice (Eckert and McConnell-Ginet 1992; Holmes and Meyerhoff 1999). Evolutionary discussions of sex/gender-linked sociolinguistic patterns persistently gloss over these factors, preferring to explain variation in terms of innate cognitive differences; but their arguments are often flawed by their dependence on linguistically dubious premises. A case in point is the way much of the literature on language evolution and cognitive sex-differences treats the classical variationist generalization (now superseded, but I will leave that aside until later) that women’s speech is typically closer to the standard or prestige norm than men’s. This finding is repeatedly interpreted as evidence of women’s more advanced verbal skills. For instance, in a paragraph listing various measures on which women have been found to perform better than men in tests, Simon Baron-Cohen (2003: 60) includes the information that ‘[women’s] utterances show standard grammatical structure and correct pronunciation more often’. Doreen Kimura (1999: 91) notes that ‘girls speak more grammatically (than boys of the same age)’: later she adds that this difference is also observed in adults, which confirms that ‘speak(ing) more grammatically’ must be a reference to prescriptively defined ‘correctness’ rather than, say, the frequency of developmental errors. Evidently these scholars take the production of standard or ‘correct’ rather than non-standard or ‘incorrect’ variants as an indicator of verbal ability. For linguists, however, that is an obvious non sequitur, since judgements of correctness are based on social rather than linguistic



criteria. There is no reason to suppose that the speaker who says ‘isn’t’ or ‘haven’t’ is showing a higher level of verbal ability than the one who says ‘aint’. The problem is compounded by writers’ apparent unawareness that the sociolinguistic pattern they are discussing is not exclusively linked to sex/ gender: it also appears when researchers compare speakers from different socioeconomic classes and (in some contexts) ethnic groups. If this variation arose from innate differences in verbal ability, then presumably it would be logical to conclude that such differences exist not only between men and women, but also between middle-class and working-class speakers or white and non-white ones. It is not only linguists who would find that conclusion unpalatable. Most new biologists emphasize that they do not share the preoccupations with race and class which brought some of their predecessors into disrepute. Such obsessions, they say, were scientifically as well as politically dubious: they rested on a failure to appreciate that differences which are socially highly salient, like those between racial/ethnic groups, may nevertheless be, in Steven Pinker’s phrase, ‘biologically minor and haphazard’ (2002: 144). Male–female differences by contrast are ‘major and systematic’, making sex/gender a special case where biological explanations are legitimate. But even Pinker might find it implausible to suggest that the same linguistic pattern has a biological explanation when correlated with sex, but a cultural explanation when correlated with other demographic variables. Above I noted that the ‘women’s speech is more standard’ generalization no longer accurately represents the consensus view among variationist linguists. After reviewing the accumulated evidence, Labov (1990) formulated a set of principles drawing attention to what he has dubbed the ‘gender paradox’: in many speech communities the behaviour of women shows two opposing tendencies. On some variables women do adhere more closely than men to linguistic norms, but on others they adhere to those norms less closely. This has led variationists to reject accounts which explain women’s more ‘correct’ use of language as a consequence of their innately superior verbal abilities, since such accounts leave half the data unexplained. An alternative account which addresses this issue has been proposed by Jack Chambers (1995). He suggests that women’s superior verbal abilities are manifested not in more ‘correct’ speech, but in greater overall stylistic flexibility (i.e. women use a larger subset of the styles, genres and varieties which make up their community’s verbal repertoire). But while this might account for the patterns most commonly observed in modern western societies, it does not fit so well with the findings of cross-cultural and historical research. Linguistic anthropologists report that in many traditional non-western cultures it is men who command a wider range of styles, and who dominate or monopolize the most culturally prestigious ones (Sherzer 1987). Historical sociolinguists have made similar observations about pre-modern western communities (Nevalainen and Raumolin-Brunberg 2003).

D. CAMERON

183

SOCIO-CULTURAL APPROACHES TO SEX/GENDER VARIATION There is not just one undifferentiated ‘socio-cultural approach’ to the study of language and gender, but here I am less interested in the theoretical differences that exist within the socio-cultural camp (Note 2) than in the principles which are generally accepted by contemporary language and gender scholars. These go beyond the simple preference for socio-cultural over biological explanations of male–female differences. Most researchers also agree in rejecting the ‘essentialist’ assumption that there are characteristics (whether biologically based or socially produced) which all men or all women axiomatically share. While some take anti-essentialism to more extreme lengths than others, there has been a general retreat from the goal of making global generalizations about the linguistic behaviour of men and women. That was often the goal of early feminist research on language (though those researchers also favoured sociocultural rather than biological approaches), but today even scholars with no strong theoretical objection to it must recognize that it is in tension with the now copious empirical evidence showing that sex/gender-linked sociolinguistic patterns are variable in time and space. Accordingly, most researchers now eschew the old quest for universalizing statements, and follow instead Eckert and McConnell-Ginet’s (1992) injunction to ‘look locally’—to examine the relationship between gender and language-use in specific communities and contexts. The assumption is that similarities and differences between men and women can be related to the particularities of local social arrangements (for instance, people’s occupations, social networks, power relations, levels of literacy, rates of exogamy, beliefs about identity, etc.), but the argument is not that those arrangements determine speakers’ linguistic behaviour: rather they present men and women with


These observations might lead us to question not merely the arguments for female verbal superiority, but whether any account centred on innate cognitive traits can explain what needs to be explained about sex/gender-linked sociolinguistic patterns, given the empirical evidence that those patterns vary both across cultures and within them, and that they can change significantly over time. Accounts in which the behaviour of men and women arises from traits which have allegedly been part of their make-up since prehistory do not explain why the same array of male and female traits should produce such different behaviour-patterns in different times and places. Below I will expand on this argument, that the accounts proposed by the new biologism do not offer the best explanation for the empirical evidence concerning sex/gender variation, by comparing them with alternative accounts proposed by researchers who adopt a socio-cultural approach. I will begin by outlining some of the general principles which inform that approach among contemporary language and gender scholars.



particular constraints and opportunities which are part of the context for their linguistic behaviour. To illustrate how this approach works in practice, let us return to the issue of why women are found to be the more advanced users of prestigious language forms in some cases, but not in others. This variation, as already noted, presents difficulties for the argument that the use of prestige forms is connected to verbal ability. It can, however, be more convincingly related to the access women do or do not have, depending on social conditions, to the educational and professional institutions in which prestige forms are generally acquired. For example, the historical sociolinguists Terttu Nevalainen and Helena Raumolin-Brunberg (2003) found in a study of early English personal correspondence that men were more advanced than women in the adoption of high-status grammatical variants—the opposite of what would be expected now. As they point out, though, women in Tudor and Stuart England did not have the same opportunities as their present-day counterparts to acquire prestigious forms of language: even those of high social rank had very limited access to education, and none at all to the learned professions. Though in the west these restrictions have now disappeared, there are other societies where they continue to apply, and where consequently it remains the case that superordinate languages and varieties are preferentially associated with men (see e.g. Sadiqi 2003 on the case of Morocco). Even in societies where access to linguistic resources is not restricted by sex/ gender, other factors may still produce differences in men’s and women’s uptake of those resources. This point is relevant to the concern that currently exists in many western countries about the ‘gender gap’ in educational achievement—a tendency for boys’ attainment levels to lag behind those of girls, which is particularly marked in relation to language and literacy skills. Though this is often attributed to sex-differences in verbal ability, linguistic research suggests that other factors may be more significant. Some of these have to do with the salience of gender as a dimension of personal identity among children and adolescents: for many boys there is a conflict between the kind of behaviour that promotes success in school and the kind that is judged by peers to be appropriately masculine (Carr and Pauwels 2006). In addition, though, sociolinguistic patterns can often be related to the work people do, or in the case of school pupils, the work they expect to do in future; and in many cases, gender-segregated labour markets produce differing opportunities and aspirations for the two sexes. For instance, many of the ‘pink collar’ jobs available to non-elite women (e.g. in the customer service, clerical and secretarial sectors) require more standard language competence and higher levels of literacy than the ‘blue collar’ jobs open to men from the same social stratum: the latter may be able to earn as much or more in sectors like construction, which do not place a premium on language skills. Where these conditions apply, boys may invest less than girls in developing their language and literacy skills because they calculate that their prospects are less dependent on that investment. Conversely, in communities and social strata where the same conditions

D. CAMERON

185


do not apply, the behaviour of boys and girls will reflect that. In societies like the UK, the male underachievement pattern is far less marked among socially elite pupils—presumably because these pupils aspire to enter professions where advanced qualifications and high levels of literacy are indispensable for both sexes. These socially based accounts of sex/gender variation are, I would argue, preferable to the competing biological explanations on two main grounds. First, they are better able to deal with the empirical evidence showing that the relevant sociolinguistic patterns are not uniform cross-culturally and historically. An account which relates women’s more advanced use of prestige language forms to their innate verbal abilities has nothing to say about cases like Morocco where prestige forms are associated with men. By contrast, an account which relates prestige usage to speakers’ access to/uptake of educational and professional opportunities, and then links the sex/gender patterns observed in particular societies to the way sex/gender affects access and uptake in those societies, offers a single explanation for all the relevant empirical observations. Second, the socio-cultural approach offers a less reductive kind of explanation. The biological account by implication treats all kinds of linguistic behaviour as the natural expression of cognitive traits embodied (or ‘embrained’) in individuals; the socio-cultural one treats behaviour as the outcome of calculations and choices which, though ultimately made by individual language-users, arise within and are affected by a larger social context. The new biologists might object here that in fact their accounts give a central role to individual agency and choice—most notably mate-choice, which is treated by Darwinians as a key influence on women’s behaviour. The biological imperative to maximize reproductive success leads women to make various strategic choices with the goal of securing mates who will be ‘good providers’. In modern social conditions, one way for women to achieve that goal is through hypergamy, marriage to a man of higher social status. Robin Dunbar has argued that women’s desire to ‘marry up’ and thus advance the interests of their future children may be the underlying motivation for the sociolinguistic pattern whereby women tend to use more standard pronunciation: as he puts it, ‘It pays girls to develop an all-purpose accent that allows them to move easily up the social scale when the opportunity arises’ (1996: 185). This particular proposal is vitiated by its dependence on a gross misrepresentation of the relevant sociolinguistic pattern. Where it exists, the ‘more standard’ pronunciation of women is a matter of small differences in the frequency of particular variants (e.g. how often a /t/ sound is realized as a glottal stop); it is not the case that lower-class women speak with a non-localized accent which makes them perceptually indistinguishable from their social superordinates. However, even if Dunbar has chosen a poor illustration, the general point that women’s marital ambitions may influence their linguistic choices is one with which many linguists would concur. The question



is whether the choices women make are best explained in terms of global biological imperatives. In her classic study of language shift in an Austrian village, Susan Gal (1978) found that one key factor in the shift from Hungarian to German was the behaviour of young bilingual women whose preference for German reflected an aspiration to marry German-speaking workers rather than peasant men from their own community. However, the women’s calculations did not seem to be primarily about which men would be better providers for their future offspring. Arguably, the peasant men had advantages in this regard, since they (and later their sons) would inherit land. The women, however, were more concerned about their own quality of life: they saw marrying outside the community as a way of escaping the peasant wife’s traditional double burden of domestic and agricultural labour. Similarly self-interested motives are suggested by Ingrid Piller and Kimie Takahashi’s study (2006) of a group of Japanese women who had invested significant resources in acquiring English because of their desire to contract relationships with Anglophone men. Once again, it is not obvious that these women’s preferences reflected the ‘good provider’ principle (Japan has both higher levels of job-security for men and lower rates of divorce than Anglo-Saxon cultures). Rather they were based on a perception of Anglophone men as more glamorous than Japanese, and less traditional in their attitudes to women. In all the cases mentioned so far, it is evident that the constraints negotiated by women (e.g. restricted educational access, gender-segregated labour markets, economic dependence on marriage) are connected not merely to gender difference but also to gender hierarchy (albeit different societies exhibit this in varying degrees and forms). Socio-cultural approaches treat power relations between men and women in a given community as potentially an important influence on their linguistic behaviour. The new biologism, on the other hand, tends either to overlook the effects of male dominance or else to reinterpret them as manifestations of innate difference. Once again, it may be argued that the resulting explanations of behaviour are reductive and inadequate. A phenomenon which many researchers have related to power differences is the sex/gender-linked patterning of linguistic politeness behaviour. Evolutionary scholars, however, often refer to these researchers’ findings while discarding their explanations. Thus Dunbar’s (1996) discussion of Penelope Brown’s (1980) study of politeness behaviour in a Mayan community treats her finding that women are ‘more polite’ (i.e. use higher frequencies of certain politeness features) as yet another piece of evidence for their superior verbal abilities. In Brown’s own analysis, by contrast, the issue is not ability, it is motivation; and the motivation arises from inequality. In this community women’s subordination is extreme, and non-compliance with men’s demands carries a high risk of physical punishment. Brown suggests that this explains why women’s greater politeness is not just a global, undifferentiated phenomenon: their preference for using deferential forms of politeness to men, and more solidary forms to other women, becomes

D. CAMERON

187

SEX, LANGUAGE AND THE BRAIN Many discussions of sex, language and the brain focus on the phenomenon of lateralization, the specialization of functions to one or other of the brain’s two hemispheres. Language is (in right-handed individuals) predominantly a left-hemisphere function. But there has long been interest in the possibility that the functional organization of the brain, including its organization for language, may be different for males and females, with males (on average, and other things—like hand-preference—being equal) showing a greater degree of lateralization. Evidence bearing on this question comes from neuroimaging studies, experiments using the Wada technique, and clinical observations of brain-injured patients. A related area of investigation concerns the effect of hormones. It has been suggested that high levels of testosterone (a hormone which is found in both sexes, but normally at significantly higher concentrations in males) produce earlier and more rapid growth on the right side, including the right hemisphere


understandable if we view politeness as a strategy they employ to deal with problems that arise from their social position. With men it is rational to try to disarm potential threats by displaying a submissive or non-provocative attitude; with other women it is rational to try to form protective alliances by displaying solidarity and mutual regard. Men are ‘less polite’ not because they cannot use these strategies, but because in most situations they feel no need to. The same pattern would not be expected in societies where male–female relationships are organized differently: indeed, there are societies where the most elaborately polite linguistic styles are dominated or monopolized by men while verbal aggression or impoliteness is associated with women (Keenan 1974; Kulick 1993). Above I argued that the new biologism’s accounts of sex/gender differences in linguistic behaviour are flawed by their reliance on generalizations which the evidence does not support (such as the belief that women talk more than men), and/or assumptions which are linguistically ill-founded (such as the equation of ‘correctness’ with verbal skill). Here I have drawn attention to other problems, such as inattentiveness to cross-cultural and historical evidence, and failure to consider (or in some cases even to acknowledge) competing interpretations of data. The most pertinent objection to these accounts is not that they offend feminist and other ‘politically correct’ sensibilities; it is that they do not meet the scientific criterion of being able to account fully and convincingly for the evidence. But as I noted in the introduction, there is another body of evidence to which the new biologism appeals, drawn from the study of language, cognition and the brain. Findings about what popularizers have dubbed ‘brain sex’ are often invoked, in both scholarly and non-scholarly sources, as if they constituted irrefutable evidence for the reality and significance of biologically based sex-differences. How far, though, is that interpretation justified?


the widely held assumption that there are major differences in the degree of brain lateralization of function between men and women is debatable. Even where sex differences in lateralization clearly exist, no convincing case has yet been made that they influence cognitive function. In this area of research it is possible that future advances will yield firmer conclusions (whereas we would need to solve the problem of time-travel to get much further with the study of early human linguistic behaviour). But there will still be questions about how far neuroscientific findings license the broader claims they are often invoked to support. How, that is, do we get from specific statements about the performance of men and women on measures X, Y, and Z (as demonstrated in laboratory tasks designed so that investigators can isolate a particular skill, observe it systematically and measure it reliably), to much more general assertions about women’s superior communication skills when interacting with others in everyday life? Kimura is typical of many commentators on this subject in treating the connection as obvious. ‘The impression many people have’, she remarks (1999: 91), ‘that women generally possess better skill with words than men do, probably arises from differences apparent between very young girls and boys’. She then lists some differences for which there is experimental evidence: girls have larger working vocabularies, are better spellers, superior readers, and do better on ‘tests requiring them to generate words with particular limitations


of the brain, and that this has implications for our understanding of both normal male–female differences and the aetiology of certain developmental abnormalities (e.g. Geschwind and Galaburda 1985). Scholars who favour this argument relate it to an observation whose centrality to some evolutionary accounts has already been discussed: that in comparison to most females, most males have less well-developed verbal abilities, but better-developed visual–spatial skills. If the testosterone theory is correct it may explain this, since linguistic functions are typically concentrated in the left hemisphere and visual-spatial functions in the right one. However, its correctness remains a matter of dispute (e.g. Crow 1998), as indeed does the assertion that females have more advanced verbal skills (see above). On lateralization too, many questions remain unresolved. Recent advances in the study of the brain (especially the advent of new imaging techniques) have produced a wealth of new data, but it has not necessarily become easier to draw clear conclusions from it: studies have proliferated, but their findings are very mixed. Supporters of the testosterone theory can appeal to studies in which hemispheric dominance for language was found to be stronger in males; but sceptics can equally point to studies using the same tasks which found no sex-differences (Frost et al. 1999; Knecht et al. 2000). Even Doreen Kimura, an influential supporter of the general thesis that there are innate cognitive differences between the sexes, has urged caution when interpreting this body of evidence (1999: 181):

D. CAMERON

189

CONCLUSION The new biologism is part of a larger ‘Darwinian turn’, in the light of which linguists (among others) are now frequently urged to re-examine their assumptions about the nature and behaviour of human beings. I do not dispute the importance of paying attention to developments in other fields which may have implications for our own. But I do think there is a need for caution about the rhetorical claims which are often made by and for the new biologism. In this discussion I have focused specifically on its claims about sex/gender differences in linguistic behaviour, and I have argued that its approach to this topic is not preferable on scientific criteria to the socio-cultural approaches favoured by most current language and gender researchers. I have tried to show that the latter explain more of the relevant empirical evidence with less resort to assumptions and speculations for which there is no evidential basis. I have also drawn attention to the selective and misleading way in which linguistic research evidence is often used by supporters of the new biologism. When these scholars accuse others of being unwilling to follow the evidence where it leads (i.e. to the conclusion that male–female differences are biologically based), they are open to the charge of throwing stones from a glass house. Scientific accountability is not a one-way street: if linguists must take account of new knowledge about genes, brains and evolution, then Darwinians must equally be held accountable to the knowledge our research has produced about language-use and linguistic variation. It is right that we should learn from other scientists’ work; but that need not and should not mean devaluing our own. Conflict of interest statement. None declared.


on them (such as words containing particular letters)’, though they are ‘not necessarily more fluent in narrative output’ (1999: 91). But on reflection it is hard to see how most of these variables could be directly involved in the everyday impressionistic judgements which Kimura suggests they may explain. Spelling and reading skills are not displayed in conversation, and most conversations do not require participants to display more than a small part of their working vocabulary, let alone ‘generate words with particular limitations on them’. The only variable mentioned which does seem plausible as a basis for real-world judgements of skill is narrative fluency, which is also the only variable on which Kimura says there is no clear evidence of female superiority. Does she take the variables which have shown sex/gender differences in the lab to be legitimate proxies for other skills that might be observable outside it, and if so what skills would those be? Just placing one set of observations (many people have the impression that. . . ) adjacent to another set (laboratory tests have shown that. . . ) does not magically make one into the explanation of the other.


NOTES 1

is influenced by postmodernist theories of identity and subjectivity, or alternatively by ethnomethodological theories of the construction of social reality. The distinctive feature of this approach is that it denies gender any ontological reality (here I paraphrase the postmodernist philosopher Judith Butler)—in other words, proposes that ‘men’ and ‘women’ are effects of our discourses and practices rather than entities which exist independently of those discourses and practices. For the most radical ‘gender relativists’, an account which invokes external social structures, hierarchies or norms to explain gendered behaviour is not ‘social constructionist’, because it requires the analyst to posit that gender or some aspect thereof does have a pre- or extra-discursive reality (this position is labelled ‘gender realism’). I am a ‘gender realist’, and I do not accept that this position is not also ‘social constructionist’ in the classical Beauvoirian sense. However, since the meaning of ‘social constructionism’ is now contested, and for some people the term does exclude non-relativist or non-postmodern approaches, I have tried to pre-empt any misunderstanding by using the term ‘socio-cultural approach’. I intend this to function simply as a generic label for the nonbiologically based approaches to sex/ gender which this article contrasts with the new biologism.

REFERENCES Barkow, J., L. Cosmides, and J. Tooby (eds). 1992. The Adapted Mind: Evolutionary Psychology

and the Generation of Culture. Oxford University Press.


2

The slightly cumbersome formula ‘sex/ gender’ is used here and (where appropriate) below in an effort to avoid misrepresenting the competing positions in this debate, or pre-empting the argument between them. The term ‘gender’ (meaning ‘the social condition of being a man/woman’) is generally preferred to ‘sex’ in social scientific and/or feminist research, but the new biologism argues that many or most socalled ‘gender differences’ are really effects of biological sex. While it will be clear that I disagree with the new biologism, I have tried to avoid using terminology which inherently favours one position. Some readers may wonder why I refer throughout this article to ‘socio-cultural approaches’ rather than ‘social constructionism’. In my own usage, in fact, these expressions would be more or less interchangeable. Until quite recently, the description of feminist theory or scholarship as ‘social constructionist’ implied only that the theorist or scholar accepted Simone de Beauvoir’s foundational insight (first aired in 1949) that ‘one is not born a woman, one becomes one’— in other words that men, women and the differences between them are the products of society and culture, not biology. Since the 1990s, however, the term ‘social constructionism’ has undergone narrowing: some feminist scholars now use it to refer specifically and exclusively to one particular variant of social constructionism, which

D. CAMERON

Belief Systems. Berkeley Women and Language Group, pp. 183–90. Eckert, P. and S. McConnell-Ginet. 1992. ‘Think practically and look locally: Language and gender as community-based practice,’ Annual Review of Anthropology 12: 461–490. Frost, J. A., J. R. Binder, J. A. Springer, T. A. Hammeke, P. S. Bellgowan, S. M. Rao, et al. 1999. ‘Language processing is strongly left lateralized in both sexes: Evidence from functional MRI,’ Brain 122: 199–208. Gal, S. 1978. ‘Peasant men can’t get wives: language and sex roles in a bilingual community,’ Language in Society 7/1: 1–17. Geschwind, N. and A. M. Galaburda. 1985. ‘Cerebral lateralization. Biological mechanisms, associations and pathology: A hypothesis and a program for research,’ Archives of Neurology 42: 428–654. Goodwin, M. H. 2006. The Hidden Life of Girls: Games of Stance, Status and Exclusion. Blackwell. Gurian, M., P. Henley, and T. Trueman. 2001. Boys and Girls Learn Differently! Wiley. Halpern, D. F. 1992. Sex Differences in Cognitive Abilities. Lawrence Erlbaum. Holmes, J. and M. Meyerhoff. 1999. ‘The community of practice: theories and methodologies in language and gender research,’ Language in Society 28: 173–183. Hyde, J. S. 2005. ‘The gender similarities hypothesis,’ American Psychologist 60/6: 581–592. Hyde, J. S. and M. Linn. 1988. ‘Gender differences in verbal ability: A meta-analysis,’ Psychological Bulletin 104: 53–69. James, D. and J. Drakich. 1993. ‘Understanding gender differences in amount of talk’ in D. Jannen (ed.): Gender and Conversational Interaction. Oxford University Press, pp. 281–312. Joseph, R. 2000. ‘The evolution of sex differences in language, sexuality and visualspatial skills,’ Archives of Sexual Behavior 29/1: 35–66. Keenan, E. O. 1974. ‘Norm makers, norm breakers: Some uses of speech by women in a Malagasy community’ in R. Bauman and J. Sherzer (eds): Explorations in the Ethnography of Speaking. Cambridge University Press, pp. 125–43. Kimura, D. 1999. Sex and Cognition. MIT Press. Knecht, S., M. Deppe, B. Draeger, L. Bobe, H. Lohmann, E.-B. Ringelstein, and


Baron-Cohen, S. 2003. The Essential Difference: Men, Women and the Extreme Male Brain. Allen Lane. Baxter, J. 2006. ‘ ‘‘Do we have to agree with her?’’ How high school girls negotiate leadership in public contexts’ in J. Baxter (ed.): Speaking Out: The Female Voice in Public Contexts. Palgrave Macmillan, pp. 159–78. Bergvall, V. L. 1999. ‘Toward a comprehensive theory of language and gender,’ Language in Society 28: 273–293. Bickerton, D. 2006. ‘Language-use, not language, is what develops in childhood and adolescence,’ Behavioral and Brain Sciences 29: 280–281. Bing, J. M. and V. L. Bergvall. 1996. ‘The question of questions: beyond binary thinking’ in V. L. Bergvall, J. M. Bing, and A. F. Freed (eds): Rethinking Language and Gender Research: Theory and Practice. Longman, pp. 1–30. Brown, P. 1980. ‘How and why are women more polite? Some evidence from a Mayan community’ in S. McConnell-Ginet, R. Borker, and N. Furman (eds): Women and Language in Literature and Society. Praeger, pp. 111–36. Carr, J. and A. Pauwels. 2006. Boys and Foreign Language Learning: Real Boys Don’t Do Languages. Palgrave Macmillan. Chambers, J. K. 1995. Sociolinguistic Theory. Blackwell. Crow, T. 1998. ‘Sexual selection, timing and the descent of man: A theory of the genetic origins of language,’ Cahiers de Psychologie Cognitive/Current Psychology of Cognition 17/6: 1079–1114. Crow, T. 2005. ‘Who forgot Paul Broca? The origins of language as test case for speciation theory,’ Journal of Linguistics 41: 133–156. Dahlberg, F. (ed.) 1981. Woman the Gatherer. Yale University Press. Dunbar, R. 1996. Grooming, Gossip and the Evolution of Language. Faber. Dunbar, R., L. Barrett, and J. Lycett. 2007. Evolutionary Psychology, a Beginner’s Guide: Human Behaviour, Evolution and the Mind. Oneworld. Eckert, P. 1996. ‘Vowels and nail polish: the emergence of linguistic style in the preadolescent heterosexual marketplace’ in N. Warner, J. Ahlers, L. Bilmes, M. Oliver, S. Wertheim, and M. Chen (eds): Gender and

191


Miller, G. F. 1999. ‘Sexual selection for cultural displays’ in R. Dunbar, C. Knight, and C. Power (eds): The Evolution of Culture. Edinburgh University Press, pp. 71–91. Miller, G. F. 2000. The Mating Mind: How Sexual Choice Shaped the Evolution of Human Nature. Doubleday. Nevalainen, T. and H. Raumolin-Brumberg. 2003. Historical Sociolinguistics. Longman. Piller, I. and K. Takahashi. 2006. ‘A passion for English: Desire and the language market’ in A. Pavlenko (ed.): Bilingual Minds: Emotional Experience, Expression and Representation. Multilingual Matters, pp. 59–83. Pinker, S. 2002. The Blank Slate: The Modern Denial of Human Nature. Allen Lane. Rivers, C. and R. C. Barnett. 2007. ‘The difference myth,’ Boston Globe, October 28. Rose, S. 2005. The 21st Century Brain: Expanding, Manipulating and Mending the Mind. Vintage. Sadiqi, F. 2003. Women, Gender and Language in Morocco. Brill. Sheldon, A. 1990. ‘Pickle fights: gendered talk in pre-school disputes,’ Discourse Processes 13: 5–31. Sherzer, J. 1987. ‘A diversity of voices: women’s and men’s speech in ethnographic perspective’ in S. Phillips, S. Steele, and C. Tanz (eds): Language, Gender and Sex in Cross-Cultural Perspective. Cambridge University Press, pp. 95–120. Tannen, D. 1990. You Just Don’t Understand: Men and Women in Conversation. Morrow.


H. Henningsen. 2000. ‘Language lateralization in healthy right-handers,’ Brain 123: 74–81. Knight, C. 1990. Blood Relations: Menstruation and the Origins of Culture. Yale University Press. Kulick, D. 1993. ‘Speaking as a woman: structure and gender in domestic arguments in a Papua New Guinea village,’ Cultural Anthropology 8/4: 510–541. Labov, W. 1990. ‘The intersection of sex and social class in the course of linguistic change,’ Language Variation and Change 2: 205–254. Lee, R. and R. Daly (eds) 1999. The Cambridge Encyclopaedia of Hunters and Gatherers. Cambridge University Press. Locke, J. and B. Bogin. 2006. ‘Language and life-history: A new perspective on the evolution and development of human language,’ Behavioral and Brain Sciences 29: 259–280. Maltz, D. and R. Borker. 1982. ‘A cultural approach to male-female misunderstanding’ in J. J. Gumperz (ed.): Language and Social Identity. Cambridge University Press, pp. 196–216. McElhinny, B. 2003. ‘Gender in sociolinguistics and anthropology’ in J. Holmes and M. Meyerhoff (eds): The Handbook of Language and Gender. Blackwell, pp. 21–42. Mehl, M. R., S. Vazire, N. Ramı´rez-Esparza, R. B. Slatcher and J. W. Pennebaker. 2007. ‘Are women really more talkative than men?,’ Science July 6: 82.

Applied Linguistics: 31/2: 193–214 ß Oxford University Press 2009 doi:10.1093/applin/amp016 Advance Access published on 20 May 2009

The Contribution of Written Corrective Feedback to Language Development: A Ten Month Investigation 1

JOHN BITCHENER and 2UTE KNOCH

1

AUT University, Auckland, New Zealand, 2University of Melbourne, Australia

INTRODUCTION The call for longitudinal research into the efficacy of written corrective feedback (WCF) can be traced back to the debate between Truscott and Ferris in the mid- to late 1990s. In 1996, Truscott claimed that error correction in ESL (English as a second language) writing programmes should be abandoned because it is ineffective and harmful. He claimed (i) that there was no research evidence to support the view that it ever helps student writers; (ii) that, as typically practised, it overlooks SLA (second language acquisition) insights about how different aspects of language are acquired; and (iii) that practical problems related to how teachers provide WCF and how students receive it to make a futile endeavour. However, in his 1999 response to Ferris (1999), he admitted that future research should look for, and might well find, some special, narrow uses of correction which may have value. He also claimed that it is harmful because it diverts time and energy away from more productive aspects of writing instruction. However, Ferris (1999) pointed out that the claims were premature because the body of evidence he presented was too limited and because there were too many methodological flaws in the design and analysis of the published studies. She also explained that short-term investigations involving text revision reveal improvement in accuracy as a result of WCF and that students believe it helps them improve their writing.


The call for longitudinal evidence on the efficacy of written corrective feedback (WCF) for ESL (English as a second language) writers has been made repeatedly since Truscott (1996) claimed that it is ineffective, harmful, and should therefore be abandoned. This article discusses some of the theoretical issues raised against the practice, outlines the status of recent empirical evidence and presents a 10-month study of the effects of WCF on two functional uses of the English article system given to 52 low-intermediate ESL students in Auckland, New Zealand. Assigned to four groups (direct corrective feedback, written, and oral meta-linguistic explanation; direct corrective feedback and written meta-linguistic explanation; direct corrective feedback only; the control group), the students produced five pieces of writing (pre-test, immediate post-test, and three delayed post-tests). Each of the treatment groups outperformed the control group on all post-tests and no difference in effectiveness was found between the three treatment groups.

194 THE CONTRIBUTION OF WCF TO LANGUAGE DEVELOPMENT

THEORETICAL ISSUES RAISED BY TRUSCOTT From a theoretical perspective, Truscott (1996) claimed that there are a number of reasons for not expecting WCF to be effective. Agreeing with well-established SLA claims about the gradual and complex nature of the acquisition process, he built his first argument against the value of WCF around the belief that a simple transfer of information (in the form of WCF) cannot be expected to work. However, this stance does not take into account the fact that learners who notice the difference between target-like input (be it oral or WCF) and their non target-like output are able to modify it as targetlike output. There is sufficient evidence in both the oral SLA literature (Schmidt and Frota 1986; Swain 2005) and the text revision literature (Ferris 2004, 2006) to confirm that this can occur. Moreover, in recent WCF studies (to be discussed in the following section), it has been reported that learners are able to apply the feedback given to them on targeted linguistic forms/structures when writing new texts over at least a 6-month period (Bitchener et al. 2005; Sheen 2007; Bitchener 2008; Bitchener and Knoch 2008a, 2008b, 2009). However, further research is required to see what enduring effect WCF might have beyond this period of time, thereby determining whether there is a role for the practice beyond text revision and short-term development.


However, she acknowledged that some of Truscott’s reasoning was compelling, particularly with regard to some of the theoretical issues he raised and to some of the practical problems he outlined. Ferris concluded her response with a call for further research into questions about the efficacy and provision of WCF on ESL student writing. While agreeing with this agenda, Truscott (1999) insisted and continues to insist (Truscott 2004, 2007) that until clear evidence is produced that WCF is effective and it must be assumed that it is a misguided practice. To date, Truscott’s theoretical arguments against WCF have received limited response in the literature. On the other hand, the call by both Truscott and Ferris for more empirical evidence on the long-term benefits of WCF has resulted in a growing number of published studies. However, the extent to which it has an enduring effect beyond a 2-month period has yet to be explored. A range of studies have investigated the relative effects of direct and indirect feedback options but the results from these studies have varied so much that further investigations are required before conclusions can be made. To address these needs, this article (i) responds to some of the theoretical issues identified by Truscott; (ii) surveys the current status of empirical evidence (on the efficacy of WCF and on the relative effect of different feedback options); and (iii) presents the findings of a 10-month investigation into the effectiveness of providing WCF on two targeted linguistic error categories, measured by the level of accuracy achieved in the writing of new texts.

J. BITCHENER and U. KNOCH

195


Truscott further argues that no single form of correction can be expected to help learners acquire knowledge of all linguistic forms/structures because the acquisition of syntax, morphology, and lexis require an understanding not only of form but also of meaning and use in relation to other words and other parts of the language system. It is important that future research agendas include investigations of the extent to which WCF, and different types of WCF, can facilitate the acquisition of different linguistic forms/structures. To date, there is research evidence to suggest that WCF can be helpful in the acquisition of simple, rule governed forms/structures like aspects of the English article system and the past simple tense (Bitchener et al. 2005; Bitchener 2008). Concerning the effectiveness of WCF in other linguistic areas, further research is required to see if there are limitations on its wider effectiveness. Truscott suggests that WCF cannot be expected to facilitate syntactic knowledge because the domain is not characterized by a collection of discreet items that can be learned one by one. Given the positive effects that have been observed in oral corrective feedback studies, for example, with syntactic structures like question forms (Mackey and Philp 1998; Mackey and Oliver 2002; Mackey et al. 2002) and the use of dative constructions (McDonough 2006), it is possible that further WCF research may also reveal positive effects for these and other forms/structures. Moreover, it may well be the case that WCF proves to be more effective as a result of the additional time and space ESL writers have to consider and process it. The second reason that Truscott presents against the practice of WCF concerns the feasibility of providing WCF at a time when the learner is ‘ready’ (Piennemann 1998) to acquire a particular form/structure. Because the acquisition of some linguistic forms and structures has been shown to follow a natural order (Clahsen et al. 1983), it is understandable that WCF may not be effective if it is provided at a time that is inconsistent with that order. However, if teachers take into account a learner’s current stage of development when determining their areas of focus, the potential would always exist for it to be effective. Concerning this, Truscott questions the feasibility of providing learners with WCF at a time that coincides with their readiness. However, this does not mean that the task is impossible. It is possible, for example, in both individual and small group contexts for WCF to be effectively provided on one or two targeted forms/structures that learners and teachers identify as being repeatedly problematic and that learners agree should be targeted for an agreed period of time. Truscott continues his argument by suggesting that WCF is unlikely to be effective if it is provided on too comprehensive a range of error categories. While the comprehensive approach may not prevent learners from accurately revising a number of errors in a single text, it may not enable learners to demonstrate equally positive improvements in their writing of new texts over time. On the other hand, Truscott also discounts a more selective approach (whereby a few error categories are consistently corrected over a period of time), claiming that there is evidence to suggest its ineffectiveness. However, he refers to only one study


EMPIRICAL EVIDENCE IN SUPPORT OF WCF As well as presenting theoretical reasons to suggest that WCF is unlikely to be effective, Truscott claimed that several empirical studies (Semke 1984; Robb et al. 1986; Kepner 1991; Sheppard 1992) had demonstrated that there


(Hendrickson 1981). Before such a claim can be made, evidence from single study investigations that compare the effect of WCF on targeted and untargeted categories is needed. Just as positive findings have been reported in a number of oral corrective feedback studies where single error categories have been targeted (Doughty and Varela 1998; Han 2002; Iwashita 2003; Lyster 2004), so it may also be the case with similar WCF investigations. The third argument that Truscott presents against the practice of WCF is that any learning that results from the practice is likely to be only ‘pseudolearning’, described by Truscott (1996: 345) as ‘a superficial and possibly transient form of knowledge’. The distinction between knowledge of language and knowledge about language is one that has been made elsewhere in the literature. Krashen (1982, 1985, 1994, 1999) distinguished between acquisition and learning while DeKeyser (2004), Ellis (2004), and others have distinguished between implicit unconscious procedural knowledge and explicit declarative knowledge. Truscott argues that, at best, WCF may have some limited value for developing meta-linguistic knowledge or explicit declarative knowledge and, therefore, may be useful for editing purposes. The value of WCF for editing purposes has been empirically demonstrated (Ferris 2004, 2006) and, in the case of some error categories (see next section), shown to also have value for language development. However, the durability of these effects over longer periods of time has yet to be investigated. The extent to which learners draw upon explicit or implicit knowledge during the writing process is not known. It is likely, however, that they draw upon both explicit and implicit linguistic knowledge. At certain times during the writing process, they are likely to reflect upon the form or structure that should be used, and, in such situations, they are most likely to draw upon their explicit, meta-linguistic knowledge, whereas, at other times, the decision about which form or structure to use will be made without conscious reflection. Whether the WCF that is used to edit a single text (Ferris 2003) or write a new text over time (Sheen 2007; Bitchener 2008; Bitchener and Knoch 2008a, 2008b, 2009) is stored in the learner’s memory as explicit declarative knowledge or as implicit procedural knowledge has yet to be investigated. However, there is a growing longitudinal research base to show that ESL writers make use of the WCF that they are given when it comes to the writing of new texts over time. This demonstrates that improvements in accuracy can result from WCF. Thus, it is incorrect to suggest that, in such cases, only pseudo-learning has taken place. Improved accuracy in the writing of new texts over time demonstrates more than a superficial and transient level of knowledge. The following section outlines the status of such evidence.


197

is evidence to support an abandonment of the practice. Given the design and analytical shortcomings of these studies (Ferris 2003; Guenette 2007; Sheen 2007; Bitchener and Knoch 2008a), such conclusions need to be read with caution. Since making this claim, an expanding research base has been considering (i) whether or not WCF can help learners improve the accuracy of their writing and thereby facilitate the acquisition process and (ii) the relative effectiveness of different types of WCF.

The effectiveness of WCF

The relative effectiveness of different types of WCF Assuming that WCF is effective in helping learners improve the accuracy of their writing and in facilitating the acquisition process, a range of studies have investigated whether certain types of WCF or combinations of different types


A number of studies conclude that WCF is effective in helping ESL students improve the accuracy of their writing. However, seven of these (Lalande 1982; Ferris 1995, 1997, 2006; Chandler 2000; Ferris and Helt 2000; Ferris et al. 2000) were designed without a control group so there is no way of knowing whether or not the reported improvements in accuracy were actually a result of WCF. Of those that did compare groups of students who received WCF and those who did not, eight (Fathman and Whalley 1990; Ashwell 2000; Ferris and Roberts 2001; Sheen 2007; Bitchener 2008; Bitchener and Knoch 2008a, 2008b, 2009) report that it had a positive effect on accuracy. Studies by Ashwell (2000), Fathman and Whalley (1990), and Ferris and Roberts (2001) reported improved accuracy in text revisions while those by Bitchener (2008), Bitchener and Knoch (2008a, 2008b, 2009), and Sheen (2007) report improved accuracy in the writing of new texts over time. On the other hand, two other studies (Kepner 1991; Polio et al. 1998) that included a control group reported that WCF was not effective in helping students improve the accuracy of their writing. However, it needs to be realized that both of these studies contained design flaws so each should be read with this in mind. Kepner’s study had no pre-test measurement, no control over the length of the journal entries, no control over the texts that were written out of class, and a range of analytical flaws. Less problematic was the study by Polio et al. where different instruments were included in the post-test. Other studies (Lalande 1982; Chandler 2000; Ferris 2006) that have not included a control group have also claimed that WCF facilitates improvements in accuracy but, without a non-feedback group, they can only be regarded as possibly predictive of effectiveness. Further research that includes a control group, that tests the efficacy of WCF in new pieces of writing and that measures the level of retention over more extensive periods of time is needed if conclusions about the acquisition potential of WCF, as opposed to the text revision potential, are to be made.



are more effective than others. These studies have most often categorized feedback as either direct (explicit) or indirect (implicit). Direct corrective feedback has most often been defined as the provision of the correct linguistic form or structure by the teacher to the student above or near the linguistic error. It may include the crossing out of an unnecessary word/ phrase/morpheme, the insertion of a missing word/phrase/morpheme, or the provision of the correct form or structure. More recently, direct corrective feedback has also included written meta-linguistic explanation (the provision of grammar rules and examples of correct usage). Sometimes oral metalinguistic explanation (e.g. in the form of class discussion or one-on-one conferences) has also been included in this category. On the other hand, indirect corrective feedback is that which indicates that in some way an error has been made, but correction is not supplied. This may be provided in one of four ways: underlining or circling an error; recording in the margin the number of errors in a given line; or using a code to show where an error has occurred and what type of error it is. Rather than the teacher providing an explicit correction, students are left to resolve and correct the problem that has been drawn to their attention. Studies that have investigated the relative merits of these approaches have tended to be grouped according to those that have compared (i) direct and indirect types of WCF; (ii) different types of indirect feedback; and (iii) different types of direct feedback. In studies that have compared direct and indirect types, two (Lalande 1982) have reported an advantage for indirect feedback, two (Semke 1984; Robb et al. 1986) have reported no difference between the two approaches, and one (Chandler 2003) has reported positive findings for both direct and indirect feedback. Clearly, firm conclusions cannot be made from these conflicting results. In comparison, findings from some oral corrective feedback studies in SLA research point to an advantage for direct over indirect corrective feedback (Carroll and Swain 1993; Nagata 1993; Muranoi 2000; Carroll 2001; Havranek and Cesnik 2003; Rosa and Leow 2004; Ellis et al. 2006), but there are others (Dekeyser 1993; Kim and Mathes 2001; Leeman 2003) that claim the opposite. As well as comparing direct and indirect approaches, several other studies (Robb et al. 1986; Ferris et al. 2000; Ferris and Roberts 2001) have investigated the relative effectiveness of different types of indirect feedback (coded and uncoded), but none has found any difference between the two options. Less attention has been given to a comparison of different types of direct feedback. Bitchener et al. (2005) compared the effect of different direct feedback combinations typically practised in advanced proficiency classroom settings: direct correction plus oral meta-linguistic explanation in the form of five minute oneon-one conferences; direct error correction; no corrective feedback. They found that group 1 (direct error correction and oral meta-linguistic explanation) outperformed group 2 (direct error correction) and the control group for the past simple tense and the definite article but not for prepositions.


199

THE STUDY Aims Two aims informed the design of this study. The first was to add to the growing body of research investigating the extent to which WCF results in improved accuracy in four new pieces of writing over a 10-month period and, as a result, to determine the role that WCF might play in long-term acquisition. The second aim was to investigate whether there is a differential effect on accuracy for three different WCF options: (i) error correction plus written and oral meta-linguistic explanation; (ii) error correction plus written meta-linguistic


They suggest that the addition of oral meta-linguistic explanation may have been the crucial factor in facilitating increased accuracy. Bitchener (2008) investigated the effectiveness of other direct feedback combinations (direct error correction with written and oral meta-linguistic explanation in the form of a 30-minute classroom lesson; direct error correction with written meta-linguistic explanation; direct error correction; no corrective feedback) on only two functional uses of the English article system (indefinite ‘a’ for first mention and definite ‘the’ for subsequent mentions) with 75 low intermediate, international visa students. Groups 1 and 3 outperformed the control group while group 2 (direct error correction and written meta-linguistic explanation) only just failed to do so. When the study was extended (Bitchener and Knoch 2008a) to include 144 international visa and migrant students, no difference was observed between the same three treatment combinations. It is possible that the larger sample size eliminated the difference in effect between group 2 and the other two treatment groups in the first study by Bitchener (2008). Sheen (2007) investigated the relative effect of two types of direct feedback (error correction and written meta-linguistic explanation) with 91 intermediate ESL learners and found that those who received WCF outperformed those who received no feedback. She also found no difference between the two feedback options in her immediate post-test, but an advantage for written meta-linguistic explanation over direct error correction in the delayed posttest conducted 2 months later. Considering four studies, it is clear that focused WCF was more effective than no feedback in bringing about improvements in the accuracy of ESL learners. Whether or not there is an advantage for meta-linguistic explanation over error correction alone for some forms/structures has yet to be confirmed. On the one hand, Sheen (2007) found an advantage for meta-linguistic explanation but this only became evident 2 months later in her delayed post-test. On the other hand, Bitchener and Knoch (2008a, 2008b) found no advantage for those who received meta-linguistic explanation after a similar 2-month period. Further research that addresses these factors over time may enable firmer conclusions to be drawn.


explanation; and (iii) error correction. Two research questions were therefore framed to investigate these aims: 1 Does accuracy in the use of two functions of the English article system improve over a 10-month period as a result of WCF? 2 Does accuracy in the use of these two functions of the English article system vary according to the WCF options provided?

Design

Context The study was conducted in the English Language Department of a university in Auckland, New Zealand. Students who were new to the university were assigned to a proficiency level after taking a standardized grammar test, a writing test, and a one-on-one interview while students who had previously been studying at a lower proficiency level were placed in the low-intermediate level on the basis of earlier competency-based assessments. The English Language Department describes its approach to the teaching of English as communicative and gives an equal focus to reading, writing, speaking, and listening. Grammatical aspects of the language are also explicitly taught in the programme. The two functions targeted in the study are taught at elementary level but the majority of participants entered the department’s classes at low-intermediate level so there was no guarantee that all participants would have received prior instruction in the use of the article functions.

Participants Most of the students were migrants who had settled in New Zealand within 18 months of commencing study at the low-intermediate level. The students (19 males and 33 females) were predominantly from East Asian countries: Korea (15 per cent), Japan (11 per cent), P.R. China (18 per cent). Other


Accuracy in using two functions of the English article system was measured over a 10-month period by means of a pre-test–post-test design (a pre-test at the beginning of the 10-month period and post-tests after 2 weeks, 2, 6 and 10 months). Fifty-two low-intermediate ESL learners were arbitrarily assigned to one of four groups: group 1 received direct error correction above each targeted error as well as written and oral meta-linguistic explanation; group 2 received direct error correction and written meta-linguistic explanation; group 3 received direct error correction; group 4 was the control group and therefore did not receive WCF. Each group comprised of 13 students. The study examined the effect of these particular feedback options (variable combinations) as they are well-established practices used by ESL classroom teachers. In adopting this aspect of the design, we acknowledge the need for researchers also to examine the relative effects of single feedback variables.


201

countries represented were Vietnam, Yemen, Russia, Switzerland, Saudi Arabia, Chile, Brazil, Serbia, Turkey, Somalia, Romania, Iran, Sri Lanka, India, and Indonesia. The average age of the students was 31.7 years. The majority (78 per cent) claimed to have had formal instruction though their length of earlier study varied across a 7-year period. Essentially equal numbers of participants in the four groups came from backgrounds where articles are included in their L1.

Target structures

Treatment Group 1 received direct error correction above each functional error as well as written and oral meta-linguistic explanation. Direct error correction involved placing a tick or check above correct uses of the two functions, correcting incorrect uses with ‘a’ or ‘the’ above each error, and inserting ‘a’ or ‘the’ where they were omitted but required because of the learner’s noun choice. In the following example, it can be seen that the learner’s first reference to the wind-break was not preceded by the required inclusion of ‘a’ for first mention and in the second reference to the same noun by the required inclusion of ‘the’ for a subsequent mention of the noun: Two people are sitting inside windbreak and are looking over the top of windbreak.


Compared with earlier studies on the value of WCF (Ferris 2002, 2003, 2006), where sometimes as many as 15 linguistic forms and structures had been examined, this study investigated the effect of targeting two functional uses of the English article system: the referential indefinite article ‘a’ for referring to something the first time (first mention) and the referential definite article ‘the’ for referring to something already mentioned (subsequent mention). The decision to limit the focus was based on the positive results that have been found in SLA studies where intensive oral corrective feedback has successfully targeted a single linguistic category (e.g. Carroll and Swain 1993; Muranoi 2000; Ellis et al. 2006 amongst others). The chosen functions were targeted because students across English language proficiency levels (including particularly those at a low-intermediate level of proficiency) experience difficulty in the use of the English article system (Master 1995; Butler 2002; Ferris 2002, 2006; Bitchener et al. 2005). For example, they may experience difficulty deciding whether an article is required and, if it is required, whether it should be the definite or indefinite article. Accuracy in the use of these functions in the pre-test revealed a mean score of 59.41 per cent, thereby indicating that students at a low-intermediate level have only a partial mastery of the functions.


Because there are occasions when the definite article is required for referring to something for the first time (e.g. uniquely identifiable referents like ‘Pass the butter, please’) or for referring to mass nouns (e.g. ‘People were buying ice-cream at the kiosk’), WCF was not provided on such occasions. A post hoc analysis of 25 per cent of the analysed texts revealed that only 8 per cent of occasions in which articles were used were non-targeted uses. The written meta-linguistic explanation included a simple explanation of the two targeted functional uses of the definite and indefinite articles together with an example of their use. Attached to their pre-test pieces of writing, the students received the following explanation and illustration.

Example A man and a woman were sitting opposite to me. The man was British but I think the woman was Australian. Attention was drawn to this information at each point in the student’s writing where such errors were made. An asterisk referred the students to the meta-linguistic explanations. Oral meta-linguistic explanation took the form of a 30-minute mini-lesson. During this lesson, the researcher explained the meta-linguistic information (rules and example) attached to the students’ returned text. Additional examples were illustrated on the whiteboard and discussed with the class. The students were then given a short ‘controlled practice’ exercise (Appendix 1) and asked to fill the gap in each sentence with ‘a’, ‘the’, or neither. They were given 5 minutes to complete the exercise. The lesson concluded with a plenary discussion of the answers. Group 2 received direct error correction above each functional error and written meta-linguistic explanation. Group 3 only received direct error correction above their errors. Group 4 did not receive feedback. Feedback was only provided in the treatment session that took place a week after the pretest, that is, on the same day as the immediate post-test.

Instruments Each of the five pieces of writing required a description of what was happening in a picture of a social gathering (a beach; a picnic; a campsite; a family celebration, and a sporting event). Picture descriptions were chosen because the range of people, objects, and activities illustrated had the potential to create opportunities for the use of both English article functions. Although, it was acknowledged that they would be able to avoid such uses, if they were uncertain about which was appropriate, and choose other determiners such as ‘one’, ‘two’, ‘this’, ‘that’, it was believed that this would not happen or be possible


1 Use ‘a’ when referring to something for the first time. 2 Use ‘the’ when referring to something that has already been mentioned.


203

Procedure One week prior to the pre-test, and in accordance with the requirements of the university’s Ethics Committee, the students and the teachers (in separate sessions) in the four classes were provided with information sheets about the study and were given the opportunity to ask questions before signing a participant consent form. Because of the amount of marking required for each class, the data collection took place at different times during the year. On day one, the pre-test was administered. One week later, the treatment (WCF) was provided. For group 1, the immediate post-test was completed after the students had been given 5 minutes to consider the error corrections and the written meta-linguistic explanation and had received the 30-minute lesson (oral meta-linguistic explanation). For group 2, the immediate post-test was completed after the students had been given 5 minutes to consider the error corrections and the written meta-linguistic explanation. For group 3, the immediate post-test was completed after the students had been given 5 minutes to consider the error corrections. For group 4, the immediate post-test was completed as soon as the uncorrected pre-test piece of writing had been returned. The immediate post-test for all groups was returned 1 week after it had been written. Corrective feedback was not provided on this occasion. The first delayed post-test was administered in week eight. The students were not told when the researcher would be returning to the class. The reason for this was to eliminate the possibility of any student studying the feedback that they had earlier been given prior to the test. The researcher did not want the students to be primed in any way beforehand. The first delayed post-test


in all linguistic environments. In each of the five writing tasks, no student made fewer than six uses of ‘a’ and ‘the’. Pictures were also chosen because, in ESL classrooms, they are seen as tasks that approximate authentic communication activities. Because the students were at a low-intermediate level of proficiency, some of the key vocabulary items (concrete nouns) were provided around the margins of each picture with arrows pointing to the relevant person, object, or activity. It was decided that this would lower the anxiety level for the students if unknown key words were provided. Additionally, they were allowed to use dictionaries and ask the researcher for a particular word if necessary. Thirty minutes was given for the writing of each description. One of the criticisms that has been levelled against earlier studies that have used a range of instruments and writing conditions concerns the reliability and validity of the data (Ferris 2004; Guenette 2007). Heeding the advice that has been offered so that design flaws such as these are not repeated, this study collected data from genres and writing conditions that were identical across the five writing occasions.


was returned to the students 1 week later. The second delayed post-test occurred after 6 months and the final delayed post-test occurred after 10 months.

Analysis

RESULTS Table 1 shows the descriptive statistics for the mean test scores for the three treatment groups and the control group at the five different testing periods. Figure 1 provides a visual representation of the mean percentages for the five testing periods for each group. As can be seen, whilst the four groups were very similar at Time 1 (the pre-test), the three treatment groups increased

Table 1: Descriptive statistics for mean test scores by group and testing period Group

N Time 1 Mean SD

1. CF, written, oral 2. CF, written 3. CF 4. Control

Time 2

Time 3

Time 4

Time 5

Mean SD

Mean SD

Mean SD

Mean SD

13 62.46 11.98 89.23

8.22 79.61 10.29 84.31 12.33 88.38

9.53

13 55.31 20.08 83.77 10.36 78.46 11.04 81.69 10.93 88.77 8.55 13 59.69 20.45 79.15 14.27 79.84 13.74 78.31 13.41 81.46 13.90 13 60.17 17.58 67.08 21.45 56.62 22.29 62.46 18.97 58.92 16.16


Uses of the targeted features were first identified and corrected for each text on each of the five testing occasions. For the texts of students in group 4, the control group, this was done on a photocopy of their writing so that they did not receive the targeted feedback. The same process occurred for all groups with each of the delayed post-tests. Accuracy on each occasion was calculated as a percentage of correct usage. For example, in any one script, three correct uses of the targeted features from 10 obligatory occasions meant a 30 per cent accuracy rate. Inter-rater reliability calculations with a trained research colleague revealed a 95 per cent agreement on the identification of targeted errors and a 98 per cent agreement on the assignment of errors to the targeted categories. Descriptive statistics for the pre-test and the four post-tests were calculated separately for the four groups. Because, no statistically significant differences on the pre-test scores were found, a two-way repeated measure ANOVA was chosen to address the research questions. One-way ANOVAs with Tukey’s post hoc pair-wise comparisons were used to isolate the exact points in time where differences between the groups occurred.


205

Effectiveness of feedback types over time 90 85 80 CF, written and oral CF, written CF

75 70 65

Control

60 55 Time 1

Time 2

Time 3

Time 4

Time 5

Figure 1: Mean percentage for five testing periods for each group

Table 2: Two-way ANOVA analysis Source Between subjects WCF type Within subjects Time Time X WCF type

df

F

p

3

6.21

0.001

4 12

37.21 5.95

Applied Linguistics, volume 31, issue 2 , 2010

Applied Linguistics, Volume 31, issue 1, 2010

Applied Linguistics, Volume 31, issue 5, 2010

Applied Linguistics, Volume 31, Number 3, 2010

OECD Economic Outlook, Volume 2010 Issue 2

Applied Linguistics Volume 1, No.2, 1980

Applied Linguistics, Volum1 , Issue 1, 1980

Explorations in Applied Linguistics 2 (v. 2)

ADVANCES IN APPLIED MICROBIOLOGY VOL 31, Volume 31

Advances in Applied Mechanics, Volume 31

sports medicine Volume 40 Issue 2 February 2010