Date: Wed, 14 Apr 2004 10:59:08 +0100 From: Eva Belke Subject: The New Psychology of Language, Volume 2
EDITOR: Tomasello, Michael TITLE: The New Psychology of Language, Volume 2 SUBTITLE: Cognitive and Functional Approaches to Language Structure PUBLISHER: Lawrence Erlbaum Associates YEAR: 2003
Eva Belke, Behavioural Brain Sciences Centre, Department of Psychology, University of Birmingham
INTRODUCTION In "The New Psychology of Language, Volume 2", nine of the leading cognitive linguists present current research topics and methods in their field. This edited book follows and extends "The New Psychology of Language, Volume I", edited by Michael Tomasello in 1998. The aim of the editor and the authors is to present their research to psycholinguists and psychologists who are interested in studying language. Many researchers who have had a general education in linguistics are aware of the classical structuralist approaches to language, as well as later generative theories on language in the Chomskyan tradition.
Cognitive linguistics has evolved from these research traditions but differs markedly from each of them. Cognitive linguists focus on spontaneous spoken speech (SSS), i.e. on speaker performance. By contrast, structuralist and generative approaches sought to study linguistic competence. Ideally, the new focus on cognitive and functional aspects of linguistic performance should yield new insights to the psychological underpinnings of language production and perception. One of the declared goals of the present volume (and volume I) is to present theoretical and methodical aspects of cognitive linguistics to psycholinguists and psychologists in order to establish a new interdisciplinary discourse between linguistics and psychology that incorporates results from cognitive linguistic research. I will review the papers presented in Volume II as a psycholinguist who has a well-founded education in linguistics but whose research topics and methods are clearly based in experimental psychology and psycholinguistics.
SUMMARY AND EVALUATION OF INDIVIDUAL PAPERS In his introduction, Tomasello gives an overview of the two most influential theoretical frameworks in linguistics, structuralism and generative linguistics. He presents the major shortcomings of these two approaches and outlines how these limitations have stimulated and motivated cognitive linguists in developing a new approach to studying language. Most importantly, cognitive linguists do not seek to study the 'Ideal Speaker' or the 'Universal Grammar' any more. Instead, they study corpora of utterances in everyday interactions to find out about the universal (cognitive) foundations of language. Specifically, cognitive linguists view language as a cognitive process which is connected and in parts even determined by the features of other cognitive domains, such as vision or motor planning. One of the objectives of the book is therefore to inform other cognitive scientists, specifically psycholinguists, about Cognitive Linguistics. For this particular audience (as well as for all other readers) the introduction provides a well-written access to Cognitive Linguistics by linking this 'new' research area to former empirical and theoretical research directions in linguistics that will be familiar to many readers.
Talmy (I: Concept Structuring Systems in Language) identifies four concept structuring systems that, as he proposes, underlie conceptual processing in general and language processing in particular. He outlines how general conceptualization principles, e.g. the organization of events/objects in space and time, are mirrored in function words of a language (here: English). Function words or closed class words transport information about how entities in the real word (as expressed in content words) are structured and relate to each other. Talmy argues that in English as well as in other languages, function words will reflect structural properties of the concept structuring domains. Eventually, this presents an attempt to (universally) establish the structural properties of the events, rather than the linguistic means to express these. A logical next step therefore entails the comparison of "the structuring system found in language with those found in other cognitive domains, such as perception, reasoning, affect, and motor control" (p. 46). This paper presents an excellent introduction to some of the key ideas in cognitive semantics. By providing carefully chosen examples and illustrations, Talmy has geared the complex subject matter well towards the target audience of the book.
DuBois (II: Discourse and Grammar) outlines regularities in argument structure, as they can be derived from quantitative analyses of spontaneous discourses He points out the similarities between the rules of the grammar of a language and the constraints that govern discourse structure and links the established regularities and constraints convincingly to aspects of information management in discourse. While he backs his arguments with quantitative data from corpus-analysis throughout the initial sections of the paper, later sections on "Constraint and Strategy" (p. 75f.) include some claims that many psychologists might consider as being too speculative. For instance, when the author claims that the constraints on discourse pragmatics "are soft, to the extent that the discourse preferences are cognitively based. It would be risky to operate always at the outer limits of cognitive capacity. Better set a routine limit [of processing load] lower than the maximum [of cognitive capacity]; under special circumstances one may then momentarily exceed this flexible limitation". Here, the author should clearly draw on the work of experimental psychologists to back his claims. There are relatively few experimental studies on discourse pragmatics, mainly because pragmatics constitutes a domain that is difficult to control and study experimentally. As DuBois points out "grammar is general; discourse varies at the will of its speakers and their topics" (p.47). This will not necessarily alienate experimental psychologists, as long as linguists are willing to re-assess the extent to which their theories can be narrowed down to a size that is manageable for (experimental) psychologists (e.g. Bock, 1996). Ideally, linguists and psychologists should jointly try to formulate hypotheses on discourse pragmatics that can be tested experimentally, working in a piecemeal fashion from the very small to the bigger picture. It is only via such a procedure of small steps that a reliable basis of empirical data can be established. In view of the potential readers of this paper, I am afraid that the author may potentially ask too much of linguistically less proficient readers. His writing style is very complex, and he includes many allusions and references to existing theories and their historic value that the less proficient reader may not be familiar with.
Kemmer (III: Human Cognition and the Elaboration of Events: Some Universal Conceptual Categories) presents a cognitive-typological approach to the conceptual and linguistic elaboration of events. She exemplifies the approach with the conceptual spectrum of transitive and intransitive verbs, specifically focusing on the status of reflexive constructions and their realization in different linguistic and conceptual systems. She illustrates how, cross-linguistically, universal conceptual foundations of linguistic phenomena can be established. In many aspects, this chapter ties in with the concept structuring systems presented by Talmy (I). With her contribution, Kemmer provides the audience with a very clear presentation of the cognitive-typological approach, carefully discussing its merits as well as its potential limitations (e.g. p. 96).
Ford, Fox, and Thompson (IV: Social Interaction and Grammar) study grammar in interactional settings, investigating grammaticization or "grammar 'at work'" (p. 119). They represent a view of grammar as a set of local regularities and linguistic routines rather than as a set of rules. The authors argue that the routines found in the language domain may be similar to routines in other cognitive domains that also involve a certain degree of procedural learning. To illustrate the role of social interaction on the (grammatical) form of utterances, they present examples from five different conversational domains, such as turn constructions and utterance repairs. While the authors convincingly present their argument, their empirical basis is necessarily bound to a very casuistic analysis of conversational corpora. As pointed out previously, this may not meet the demands that many psychologists would ask for in empirical research. A theoretical and empirical complement to the paper by Ford et al. is presented in Chapter V.
Bybee (V: Cognitive Processes in Grammaticalization) presents an approach to investigating the regularities and principles of grammaticalization as a universal aspect of language and language use. She links the principles of grammaticalization to fundamental cognitive phenomena (automatization, categorization, inferencing, and habituation to repeated stimulation), which can be observed in various domains of information processing. She argues that the cross-linguistic applicability of principles of grammaticalization and observations on grammaticalization in diachronic changes and in creole formation suggest that "the true language universals are universals of change" (p. 151). Bybee's theory is closely linked to the analyses of grammaticization in discourse analysed in Ford et al. (IV). Compared to Ford et al.'s approach, Bybee's cross-linguistic, diachronic research method seems to provide a more systematic approach to language change that allows the formulation of a clear-cut theory of grammaticalization in language change.
Van Hoek (VI: Pronouns and Point of View: Cognitive Principles of Coreference) investigates pronoun anaphora in discourse. She presents the shortcomings of syntax-only approaches to pronoun resolution (as put forward in the C-command analysis) and outlines an approach that is based on general principles and ideas of cognitive linguistics. She argues that the cognitive-pragmatic foundations of pronoun use can be outlined in terms of accessibility (new vs. old) and distance (e.g. on- stage vs. off-stage in a stage model). In this framework, however, the points of view (POV) of speaker and hearer play a central role in pronoun resolution. Van Hoek develops a reference-point model of pronoun resolution that allows for an abstract analysis of pronoun anaphora and for the identification of a number of basic underlying principles (connectivity, prominence, linear order). Specifically, the model incorporates the observations on the role of the POV in pronoun anaphora. Within this framework, many of the central notions of cognitive linguistics play a crucial role (see also chapter I), which, in turn, provides potential extensions to other areas of anaphora and cross-reference. Van Hoek guides the reader very smoothly through the different theoretical viewpoints on pronoun resolution, tying in the reference-point model with earlier stages of analysis and alternative (formalist) views on pronoun resolution. In view of the overall purpose of the book, however, I had difficulties in allocating this model within the big picture of cognitive linguistics. This may not necessarily be a shortcoming of the paper. Instead, I would like to think that the papers presented in this volume should have been interlinked by some explanatory texts that bridge the passage from one paper to the next and allow the reader to allocate each paper within the bigger picture of the cognitive linguistic approach. I will come back to this issue in the General Evaluation below.
Comrie (VII: On Explaining Language Universals) reviews why language universals might exist in the first place. He refutes the generative view of the language-faculty as being separate of other cognitive systems and as being unique. He argues that instead, language can be understood as a special case, if not even as a result of cognitive processing (and its limitations). In this framework, language universals emerge as a result of the structural limitations of the cognitive and physiological capacities of its speakers as human beings. Apart from this structural dependence, there is a (universal) functional requirement for language to convey messages unambiguously and economically. Comrie exemplifies this point in cross-linguistic data on reflexive markers that are specifically relevant for third person reflexive pronouns (e.g. himSELF) to unambiguously signal co- reference. To linguists and psychologists alike, this paper provides a well-written and concise review of explanations for the existence of language universals that ties in nicely with other chapters of the book, such as Talmy (I).
Haspelmath (VIII: The Geometry of Grammatical Meaning: Semantic Maps and Cross-Linguistic Comparison) presents the semantic-map method (also referred to as cognitive-map method) for the cross-linguistic analysis of the meaning of grammatical morphemes. Semantic maps provide a tool to describe semantic-syntactic structures within and across languages and across time. They may thus be used in studying universal semantic structures and diachronic changes, e.g. in grammaticalization (see V). Haspelmath illustrates the method of semantic maps and its applicability in a number of examples. In his conclusion, he presents a critical review of the semantic-map method, pointing out potential pitfalls of the method, such as the generality fallacy. Interestingly, he is the only author who directly addresses psychologists and makes an explicit attempt to link his work to their work. Being foremost a method paper, chapter eight ties in well with many of the papers presented in the previous chapters.
Chapter IX is a reprint (with slight modifications) of the classic paper by Fillmore, Kay, and Connor (1988): "Regularity and Idiomaticity in Grammatical Constructions: The Case of Let Alone". In this paper, Fillmore et al. laid the foundations of a view of grammar as consisting of constructions as central grammatical units. Constructions can differ markedly in size, as Fillmore et al.'s typological classification of complex idiomatic constructions exemplifies. In their paper, Fillmore et al. provide a structural, semantic and pragmatic analysis of one of the typological classes of formal idioms, using the example of the 'let alone' construction. They outline a new conception of grammar that covers the structural properties of formal idioms and complex phraseological constructions as well as, as a corollary, the rules for combining smaller grammatical constructions, e.g. nouns. As Tomasello points out in the introduction, the paper by Fillmore et al. marked an important step in the emancipation of cognitive linguistics from generative linguistics. It would have been helpful for the reader if Tomasello's annotations on this paper had appeared in direct vicinity of the paper rather than at the other end of the book. With the present arrangement of single chapters without bridging comments the local value of each paper is difficult to assess for less proficient readers.
GENERAL EVALUATION As Tomasello (1998, p. viii) points out, virtually all of "the empirical studies of syntactic structure [in psycholinguistics and developmental psycholinguistics] have been concerned with narrow-range phenomena, or else they have relied on very generic descriptions that seemingly do not depend on any particular theory of syntax". In addition, most of the studies concerned everything BUT syntax (i.e. lexical retrieval, pragmatics, or else). With the two edited volumes on "The New Psychology of Language", Tomasello and all contributors aim to make recent developments in linguistic theories (of syntax) available to psycholinguists in order to enhance cross-disciplinary research into language and cognition. While this is a noble aim and the contributors to both volumes of "The New Psychology of Language" clearly do their best to make their research and theories understandable for less proficient linguists, many psycholinguists will have trouble to get the 'big picture' of cognitive-functional linguistics from the present assembly of papers. The introduction helps considerably in allocating cognitive-functional linguistics within the framework of previous linguistic schools (see also Tomasello, 1998); however, the assembly of individual papers seems to be too eclectic to integrate the papers into this picture. For instance, some papers can be classified as methodological papers (e.g. chapters III and VIII) whereas other papers aim at giving theoretical reviews (e.g. chapters I and VII). To alleviate the eclectic character of the book, it would have been helpful to include connecting passages that link the individual contributions to each other and to the introduction. Generally, for less proficient linguists, it may be helpful to start out with a general introduction to cognitive-functional linguistics, and to read Volume I of "The New Psychology of Language" before turning to the present volume.
Tomasello (1998) suggested that many psychologists may have trouble following current directions in linguistics because, among other obstacles, "linguistic descriptions and explanations require specialized linguistic terminology that discourages outsiders". This clearly is one of the problems underlying the lack of cross- disciplinary exchange but might be accommodated by adapting the writing style and scope to less proficient audiences, as is done in the present volume. However, while reading the book I could not help but realize that the difficulties of communication between psychologists and cognitive linguists may root in a deeper, methodological rift between these two research areas. Linguistics in general and cognitive- functional linguistics in particular are largely based on a casuistic research logic; cognitive linguists study language in situ, as it is actually produced. This implies that cognitive linguists can make only limited generalizations of their observations to other speakers/languages/situations. While being aware of this limitation, linguists nevertheless often make generalizations and formulate hypotheses that they entertain until proven wrong. Psychologists may consider many of these hypotheses as very bold, given the scale and range of the underlying empirical evidence. Contrary to linguists, psychologists try to narrow down the scope of their investigations to situations under laboratory control, which allow them to study samples of speakers in well-controlled situations and to make certain, statistically founded, generalizations to other samples. To linguists, however, psychologists may often seem to be too scrupulous for their own good. Given this large gap between linguists and psychologists, my impression is that it needs much more than a communication of ideas across disciplines to enhance new fruitful lines of research. What it needs is a more detailed assessment of which research questions posed in the framework of cognitive grammar can be operationalized appropriately for psychological testing.
Language acquisition is one of the domains that has proved to be accessible for psycholinguistic investigations into cognitive grammar. Further areas should clearly be opened up but seem to be too difficult to assess experimentally at present. Crucially, in order for psychologists and linguists to jointly investigate the cognitive foundations of language a fundamental mutual understanding must be established between the disciplines. Specifically, this implies that linguists need to acquire an understanding of psychological/experimental research methods (and their limitations). Psychologists might feel that their work is misrepresented if it is presented, for instance, as having "spent decades on studying on how subjects in the narrow confines of laboratory cubby-holes parse and interpreted isolated sentences that bear little relation to either the type of clauses used in natural communication, the communicative context in which such clauses are used or the linguistic context in which clauses in natural communication are produced and interpreted" (Givon, 1998, p. 63f.). Ironical statements like these do suggest that linguists need to put more effort into understanding why psychological testing exerts so many restrictions and how, within this framework, cognitive linguistic theories can be tested. Having said all this, the present volume together with volume I clearly is a step into this direction that will hopefully be followed by further cross-disciplinary exchange.
REFERENCES Bock, K. (1996). Language production: Methods and methodologies. Psychonomic Bulletin and Review, 3, 395-421.
Givon, T. (1998). The functional approach to grammar. In: M. Tomasello (Ed.), The New Psychology of Language, Vol.1 (pp. 41-66).Mahwah, NJ: Lawrence Erlbaum Associates.
Tomasello, M. (1998). Introduction: A cognitive-functional perspective on language structure. In: M. Tomasello (Ed.), The New Psychology of Language, Vol.1 (pp. vii - xxiii).Mahwah, NJ: Lawrence Erlbaum Associates.
ABOUT THE REVIEWER:
ABOUT THE REVIEWER Dr Eva Belke is a research fellow at the Department of Psychology, University of Birmingham. Having originally studied Clinical Linguistics and Psycholinguistics for therapeutic purposes, she is now doing experimental psycholinguistic research, focussing in particular on language production processes. Further research interests include aphasic and neuropsychological disorders, specifically reading disorders, and models of healthy and impaired language processing.