This page intentionally left blank part iii SITUATING COGNITIVE LINGUISTICS This page intentionally left blank chapter 20 COGNITIVE LINGUISTICS AND FUNCTIONAL LINGUISTICS jan nuyts 1. Introduction This chapter is an inquiry into how Cognitive Linguistics relates to, complements, and/or differs from other approaches within the wider field of functionally oriented linguistics (of which Cognitive Linguistics is a member as well). In order to avoid terminological confusion, I will use the notion ‘‘Functional Linguistics’’ strictly to refer to such ‘‘other functional approaches’’ only, to the exclusion of Cognitive Lin- guistics. The notion ‘‘functionally oriented linguistics’’ will serve as a cover term for all functional approaches to language, including Cognitive Linguistics. This chapter is organized as follows. In section 2, I will discuss the problem of delimiting Cognitive Linguistics on the one hand, and Functional Linguistics on the other. The subsequent sections deal with major dimensions along which one can compare Cognitive Linguistics and Functional Linguistics. Section 3 assesses the position of the two paradigms vis-a ` -vis the basic theoretical notions from which their names have been derived: ‘‘functionalism’’ and ‘‘cognition.’’ Section 4 considers how and to what extent these two paradigms deal with major dimensions of the object domain of language (structure, meaning, discourse, etc.). Section 5 briefly addresses the methods of inquiry. Section 6, finally, turns to the views on the nature and status of (linguistic) knowledge espoused in Cognitive Linguistics and Functional Linguistics, including what may be the most significant distinctive element between them, namely, the ‘‘pattern’’ versus ‘‘process’’ issue. 2. Delimiting Cognitive Linguistics and Functional Linguistics In order to uncover the differences and similarities between Cognitive Linguistics and Functional Linguistics, one obviously needs a clear notion of how these re- search fields or paradigms can be mutually delimited. It is far from easy, however, to draw an actual borderline. Cognitive Linguistics is generally considered to have a clear scientific identity, distinguishing it from other approaches in linguistics, and thus also from Functional Linguistics. The existence of the present Handbook, spe- cifically devoted to Cognitive Linguistics, may testify to this point. However, when it comes to characterizing this identity, it soon turns out that there are very few criteria (if any) that are really specific or unique to Cognitive Linguistics and allow us to oppose Cognitive Linguistics to other functional approaches. Ultimately, any delimitation of Cognitive Linguistics is bound to be, to some extent, arbitrary, or inspired by nonscientific criteria, such as social ones (see below). Consequently, opinions about where to draw the line can easily diverge. The present Handbook can again serve as an illustration, as it features several subjects that many would consider to go well beyond Cognitive Linguistics proper, into the realm of Func- tional Linguistics, while others might find that the Handbook covers too narrow a section of the entire field of Cognitive Linguistics. Here, then, is a brief characterization of how I will delimit the fields of Cog- nitive Linguistics and Functional Linguistics for the purpose of this chapter, in terms of research groups or traditions belonging to each. Cognitive Linguistics I define the field of Cognitive Linguistics at two ‘‘levels of extension.’’ At the first level, I define it fairly narrowly (more narrowly than the present Handbook), 1 confining it to what could be considered its ‘‘core’’: a body of research centering around semantic analyses of the type pioneered by Talmy (1988a, 1988b, 1996, 2000a, 2000b); Lakoff (1987; Lakoff and Johnson 1980; see also Gibbs 1994); Fau- connier (in his semantic ‘‘Mental Spaces’’ theory; 1985, 1997); Langacker (in his model of ‘‘Cognitive Grammar’’; 1987, 1991); and further found in several of the grammatical models labeled ‘‘Construction Grammar.’’ Construction Grammar is 544 jan nuyts itself a conglomerate of approaches (see Croft, this volume, chapter 18), some of which can be more or less unambiguously situated within the range of Cognitive Linguistics (e.g., the variants by Lakoff 1987; Goldberg 1995; Croft 2001). Other variants of Construction Grammar do not fall under the Cognitive Linguistics heading. This is most notably true for the version that could rightfully claim the patent on the label and the concept of Construction Grammar, namely, the ver- sion by Fillmore, Kay, and colleagues (Fillmore 1988; Fillmore, Kay, and O’Connor 1988; Fillmore et al., forthcoming), which has been ‘‘under construction’’ since the mid-eighties. This branch of Construction Grammar is currently clearly inclin- ing toward a grammar model such as Head-Driven Phrase Structure Grammar (usually abbreviated to HPSG; Pollard and Sag 1994)—not the kind of model one would readily associate with Cognitive Linguistics. 2 There is one straightforward nonscientific, geographical and thus social ele- ment which binds together Cognitive Linguistics as characterized above, namely, the fact that all the linguists mentioned have worked for at least part of their ca- reers at Californian universities in the 1980s and 1990s, mostly at the University of California at Berkeley and at San Diego. (Of course, in the mean time, PhD stu- dents of these linguists, and even some of these linguists themselves, have spread to many other places in the United States and beyond, so that the geographical cri- terion has mainly become a historical one, and its validity is weakening rapidly as time goes by.) One important consequence of this social fact is that it has facili- tated the dissemination of a number of basic views, notions, and research attitudes (which will be explicated in the following sections) among these scholars. This process has no doubt been stimulated considerably by the additional fact that quite a few of these linguists have gone through much of the same scientific develop- ments: many of them had their roots in Generative Grammar, subsequently turned to Generative Semantics, and have ultimately come to completely reject the gen- erative paradigm (Generative Semantics was a first step in their reaction to ortho- dox Generative Grammar and can be considered a stepping-stone to the creation of the Cognitive Linguistics movement). Not all of these cognitive linguists have been influenced by the same ideas and notions to the same degree, however: even among the group of cognitive linguists considerable heterogeneity remains. Furthermore, (most of) the ideas character- izing this group are not exclusive to it. On the one hand, certain elements of the cognitive linguistic views and notions have developed independently in (func- tional) linguistic circles elsewhere, even if they have not always acquired the same central status there. What characterizes Cognitive Linguistics, then, is not any of these views or notions in particular, but rather the whole cluster of them. On the other hand, since UC Berkeley and UC San Diego have been strong attraction poles for linguists from around the world, the ideas developed there have had ample chance to influence other research paradigms In a quite sizeable group of linguists, however, there has been such a strong mutual influence between the ideas and perspectives from Cognitive Linguistics and from other linguistic traditions—particularly, Functional Linguistics—that a blend of the two fields has emerged. Although many scholars in that situation cognitive linguistics and functional linguistics 545 squarely situate themselves within the field of Cognitive Linguistics (rather than within Functional Linguistics), I will call them ‘‘functional-cognitive linguists,’’ in order to distinguish their approach from that of the cognitive linguists as defined above. These functional-cognitive linguists, then, represent the second ‘‘level of extension’’ in the present definition of Cognitive Linguistics. In a way, this char- acterization already applies to Croft’s (2001) version of Construction Grammar (although it has been included in the above overview of ‘‘core’’ cognitive linguistic work), as it blends ideas from Cognitive Linguistics and language typology. The characterization certainly applies to scholars such as Kemmer (1993, 2003; Bar- low and Kemmer 2000), Verhagen (1995, 2000; Kemmer and Verhagen 1994), or Geeraerts (1989, 1993, 1997), to name just a few (many more will be mentioned later in this chapter). 3 In fact, it applies to nearly the entire ‘‘European branch’’ of Cognitive Linguistics (of which Verhagen and Geeraerts are representatives, of course). After all, whereas the North American branch of Cognitive Linguistics has to a considerable extent grown out of a negative reaction against the Generative Grammar tradition that has been prevailing on that continent, the members of the European branch have mostly been trained in one of the European, functionally oriented (structuralist or post-structuralist) linguistic schools or traditions and have adopted the ideas from Californian Cognitive Linguistics in the course of their careers, while at the same time retaining many of the basic ideas and research attitudes they have been raised with. This is possible because there is usually perfect compatibility between the ‘‘old’’ and the ‘‘new’’ ideas (see below). 4 Functional Linguistics The field of Functional Linguistics is even harder to delimit, at least in terms of scholars or research groups belonging to it, because it is much wider and much more diverse than that of Cognitive Linguistics. In fact, the only reasonable char- acterization of Functional Linguistics is in terms of all research in linguistics (and directly related fields) that adopts a functional approach to the analysis of lin- guistic phenomena (see section 3)—this would, in principle, include Cognitive Linguistics, if it were not for the terminological conventions specified in the in- troduction. In other words, unlike Cognitive Linguistics, Functional Linguistics cannot be caught in terms of some specific, regional, social, or related (e.g., his- torical) criterion, but only in terms of a general ‘‘conceptual’’ criterion, namely, a single basic research attitude shared by many researchers and research groups all over the world. Although it is nearly impossible to give an exhaustive overview, the best way to give an impression of the extension of Functional Linguistics is to present a survey of some of its major exponents. 5 This survey should obviously include a number of ‘‘schools’’ of functional grammar models, the most important of which are Systemic-Functional Grammar (as part of the wider field of Systemic Linguistics; Halliday 1994), Functional Grammar as developed in the Amsterdam tradition 546 jan nuyts (Dik 1997), and Role and Reference Grammar (Van Valin 1993)—see also Butler (2003). It should also cover a range of conceptual frameworks or traditions that have not really taken the shape of a (more or less formalized) model, but do for- mulate sets of (theoretical) principles aimed at grasping and explaining linguistic facts within different domains of the organization and functioning of language, including domains which are not covered by classical functional grammar models. These include, among many others: 6 a. In the domain of ‘‘grammar’’: the discourse-oriented syntactic work by Givo ´ n(1984, 1990, 1995) and Chafe (1994), the Columbia School (e.g., Otheguy, Stern, and Reid 2002), and the strong tradition in cross-linguistic and typological research (usually more data-oriented than theoretical) in the style of Greenberg (1966, 1978; see also Comrie 1981; Croft 1990; and the typology group at the Max Planck Institute for Evolutionary Anthropology). b. In the domain of semantics: the framework developed by Wierzbicka (1980, 1996), and also the ‘‘semantic-typological’’ work of the ‘‘cogni- tive anthropologists’’ at the Max Planck Institute for Psycholinguistics (Pederson et al. 1998), 7 work done in Slobin’s (1996) ‘‘thinking for speaking’’ framework, 8 or in the Whorfian tradition (Lucy 1992a, 1992b). c. In the domain of discourse studies: Mann and Thompson’s (1988, 1992) Rhetorical Structure Theory, the linguistically oriented branches of Conversation Analysis (see Ochs, Schegloff, and Thompson 1996), Dis- course Analysis (Sinclair and Coulthard 1975), and the framework developed by Halliday and Hasan (1976). 9 d. In the domain of diachronic studies (grammatical and semantic): the body of research focusing on grammaticalization and related seman- tic phenomena such as subjectification 10 (Heine, Claudi, and Hu ¨ nnemeyer 1991; Hopper and Traugott 1993; Bybee, Perkins, and Pagliuca 1994). 11 Let us see, then, to what extent we can pinpoint distinctive features between Cognitive Linguistics and Functional Linguistics, both in their specific research goals and practices and in their theoretical viewpoints. In line with what has been suggested above, the following survey of potential sources of divergence will not reveal any points of radical opposition. When differences do emerge, they are rarely clean-cut, but rather involve tendencies within (one of) the two paradigms, to which there are (often numerous and important) exceptions on both sides (es- pecially the ‘‘functional-cognitive linguists,’’ who systematically defy any attempts at formulating oppositions). Further, differences rarely involve real incommensura- bility: they are mainly cases of complementarity or relatively minor differences in opinion. cognitive linguistics and functional linguistics 547 3. The Basic Principles: ‘‘Functionalism’’ and ‘‘Cognition’’ Let us start with the philosophical background and first have a look at the two foundational concepts from which Functional Linguistics and Cognitive Linguis- tics have taken their names: ‘‘functionalism’’ and ‘‘cognition.’’ Do these stand for any real differences? Functionalism Functionalism in language research can be characterized as a basic method of analysis: it starts from the assumption that linguistic structure cannot be analyzed independently of the uses to which it is put (contrary to formalism, which denies the relevance of language use to understanding linguistic structure; see Nuyts 1994b). Usually, these uses are captured under the covering term of ‘‘communi- cation,’’ but this still includes a wide range of more specific dimensions and factors determining how language appears in actual contexts—from semantic ones, to do with the transmission of information (world knowledge), to interactive or social ones (often called ‘‘pragmatic’’), to do with interpersonal relationships and with the discursive nature of language use. 12 This basic orientation characterizes Func- tional Linguistics. But, as mentioned in the introduction, Cognitive Linguistics, too, is an integral subpart of the wider field of functionally oriented linguistics, as its approach to language is in line with the basic premises of linguistic function- alism (see Langacker, this volume, chapter 17, section 1). Nevertheless, there may be a difference in the way, or in the extent or intensity, to which this basic orientation is applied in both fields. Many functionalists will argue that Cognitive Linguistics only deals with certain aspects of the functionality of language: although it pays due credit to the role of purely semantic dimensions in language (specifically, dimensions pertaining to the way we conceptualize and categorize the world), it deals less extensively with the role of truly communicative dimensions (interactional and discursive features of language, the role of mutual knowledge and its effect on information structuring, interpersonal relations be- tween speaker and hearer, etc.). Of course, even within the field of Functional Linguistics itself, this kind of contrast exists, but it usually works in the opposite direction: several functionalists pay due attention to the communicative dimen- sions of language but are minimally—if at all—concerned with its conceptual se- mantic dimension. In sum, then, although differences in the conception of the functional dimen- sion and its relation to language structure no doubt trigger substantial differences between specific approaches and models, such differences do not characterize 548 jan nuyts the position of Cognitive Linguistics vis-a ` -vis Functional Linguistics per se. They rather characterize many of the differences within the field of functionally oriented language research in general. Cognition If the functionalist dimension is not a distinguishing factor, then maybe the cog- nitive dimension is. Accepting a cognitive orientation in language research means adopting a particular research goal, namely, discovering the organization and operational principles of the systems that are ‘‘implemented’’ (to use a dangerous word) in the human brain and are responsible for producing and interpreting lin- guistic behavior. There are, of course, very different views of what cognition in- volves more specifically, and, as a result, the cognitive orientation is found in a very heterogeneous set of approaches (at least as heterogeneous as the set of functional approaches). Anyway, while, in principle, one can distinguish between those who (in some way) do and those who do not take into consideration the problem of language and cognition, again, such a division does not coincide with the division between Cognitive Linguistics and Functional Linguistics. Obviously, cognition has at all times been a central and explicit concern to Cognitive Linguistics, but the same is true for quite a few functional linguists, including (to name some promi- nent examples) Givo ´ n, Chafe, Slobin, and Wierzbicka (see Chafe 1970, 1980; Givo ´ n 1979, in addition to the references given in section 2). Moreover, in the last few decades, cognition has drawn the attention of more and more members of the functional linguistic community—think only of the increasing attention for cog- nition in functional grammar models such as Dik’s Functional Grammar and Van Valin’s Role and Reference Grammar or of the renaissance of cognitive anthro- pology and its effect on, among others, linguistic typology. (Surely, Cognitive Lin- guistics has been a considerable catalyst in these.) All in all, despite different accents in Functional Linguistics and Cognitive Linguistics concerning basic philosophical principles, there is no reason to call them substantially different research paradigms, at this level of generality. 4. Domains of Language Covered A second potential source of divergences between research paradigms—and thus also between Cognitive Linguistics and Functional Linguistics—is the issue of which areas or domains of the object of inquiry are within their scope of attention. At first sight, this has more to do with practical choices (and research fashions) cognitive linguistics and functional linguistics 549 . Func- tional Linguistics, while others might find that the Handbook covers too narrow a section of the entire field of Cognitive Linguistics. Here, then, is a brief characterization of how I will delimit the. they have not always acquired the same central status there. What characterizes Cognitive Linguistics, then, is not any of these views or notions in particular, but rather the whole cluster of. from that of the cognitive linguists as defined above. These functional -cognitive linguists, then, represent the second ‘‘level of extension’’ in the present definition of Cognitive Linguistics.