认知语言学与二语

SKY Journal of Linguistics 21 (2008), 125–153 Noriko Matsumoto Bridges between Cognitive Linguistics and Second Language Pedagogy: The Case of Corpora and Their Potential Abstract Cognitive linguistics offers a way out of the dilemma between helpful, productive linguistics and helpless, unproductive linguistics in second language pedagogy. This paper applies cognitive linguistics insights to grammatical instruction of the verb find and its complementation in communicative activities, searching for descriptively adequate, intuitively acceptable, and easily accessible accounts of how the verb find functions and how widely various uses of the verb find are systematically related to one another. This paper also claims that the potential of learner corpora and the concept of entrenchment in cognitive linguistics make a positive contribution to grammatical instruction. In second language pedagogy, a data-driven analysis on the basis of both a learner corpus and a native speaker corpus is essential to explain the concept of entrenchment. Consequently, this paper shows some significant results in the data-driven analysis with respect to not only complementation patterns of the verb find that Japanese-speaking learners of English use, but also those that native speakers of English use. 1. Introduction Second language pedagogy is a truly multidisciplinary endeavor, because a matter of pedagogy is a matter of linguistics as well as a matter of acquisition. Every linguist does recognize how essential every experience and substantial knowledge in the area of language pedagogy is. However, the issue as to whether or not much research that we linguists have carried out is helpful or productive has sharply divided the linguists. Generative linguistics has focused on pure scholarship for its own sake. The only motivation is a desire to understand language much better. This tradition is most clearly represented by Noam Chomsky, who denies that linguistics has, can have, or indeed should have any relevance to language teaching NORIKO MATSUMOTO 126 (see e.g., Olson et al. 1991). On the other hand, cognitive linguistics has claimed that the practical benefits are partly evident, because any major innovation in linguistic theory is bound, sooner or later, to have an impact on the teaching of grammar in foreign language pedagogy (see e.g. Achard & Niemeier 2004; Langacker 2001; Radden & Dirven 2007; Taylor 1993). One of the motivations is a desire to improve language teaching at school to some extent. The aim of this paper is to defend the latter idea, which implies that linguistics contributes substantially to language teaching, although this paper will not of course indicate that every part of academic research has a clear payoff in terms of practical benefits. Specifically, this paper will apply cognitive linguistic insights to grammatical instruction of the verb find and its complementation in communicative activities on the basis of the corpus-based approach, that is to say, the usage-based approach. 2. Cognitive linguistics In order to provide my approach in this paper with an appropriate context, it is necessary first to discuss cognitive linguistics. Cognitive linguistics originally emerged in the 1970s and arose out of dissatisfaction with dominant formal approaches to language at that time. Some researchers such as Fillmore (1975), Lakoff & Thompson (1975), and Rosch (1975) rejected the dominant ideas that syntax is separate from other aspects of language, and that language is separate from cognition. Moreover, cognitive linguistics has always been strongly influenced by theories and findings from the other cognitive sciences, particularly cognitive psychology and Gestalt psychology. Cognitive linguistics therefore acknowledges that language is part of, dependent on, and influenced by human cognition, including human perception and categorization, and that language develops and changes through human interaction and experiences in the world (see e.g., Fillmore 1975; Lakoff 1987; Langacker 1987; Talmy 2000a, 2000b). Cognitive linguistics practice could be roughly divided into two main areas of research: cognitive semantics and cognitive grammar. Although the study of cognitive semantics and the one of cognitive grammar are occasionally separate in practice, their domains of inquiry are tightly linked. Cognitive semantics is concerned with modeling the human mind as much as it is concerned with investigating linguistic semantics. Cognitive semantics is not a single unified framework. However, Evans et COGNITIVE LINGUISTICS AND SECOND LANGUAGE PEDAGOGY 127 al. (2007) point out that there are four guiding principles characterizing cognitive semantics: (i) conceptual structure is embodied; (ii) semantic structure is conceptual structure; (iii) semantic representation is encyclopedic; and (iv) semantic construction is conceptualization. Some significant theories and approaches in cognitive semantics best exemplify the four guiding principles.1 In this paper, one of the significant theories and approaches in cognitive semantics, cognitive lexical semantics approach is important. Cognitive lexical semantics takes the position that lexical items are conceptual categories. A word represents a category of distinct related meanings. In particular, Lakoff (1987) argues that a lexical item represents a type of complex category, which he calls a radical category. Therefore, word meanings are stored in the mental lexicon as highly complex structured categories of meanings. Cognitive grammar is concerned with modeling the language system rather than the nature of mind itself. This means that meaning is central to cognitive grammar. Cognitive grammar assumes cognitive semantics and builds a model of grammar which is consistent with the assumptions and findings of research in cognitive semantics. In addition to this, the two guiding principles of cognitive grammar are (i) the symbolic thesis, and (ii) the usage-based thesis. The symbolic thesis holds that the fundamental unit of grammar is a form-meaning pairing, that is to say, a symbolic unit. All linguistic forms, from single morphemes to words, phrases, idioms, clauses, and sentences, contribute to and express meaning. The usage-based thesis is primarily concerned with the characterization of language as it is spoken and understood, as well as with the dynamics of its use.2 Langacker (1987: 494) states that the usage-based thesis constitutes a 1 The significant theories and approaches include image-schema theory, encyclopedic semantics approach, categorization and Idealized Cognitive Models (ICMs) approach, cognitive lexical semantics approach, conceptual metaphor theory, conceptual meto- nymy approach, Mental Spaces theory, and conceptual blending theory. 2 The usage-based thesis is central not only to cognitive grammar but also to language acquisition which takes a cognitive linguistic perspective. Tomasello (2000a: 237–238) argues that usage-based models constitute strong theoretical frameworks for the description of child language acquisition because they do not demand that a child’s grammar be identical to the adult system. In a usage-based model, the goal of child language acquisition research is to characterize the steps by which the child’s inventory of conventionalized units comes to resemble the adult’s. It predominantly involves the investigation of the development of the cognitive abilities that allow children to eventually master the adult system. NORIKO MATSUMOTO 128 non-reductive approach to linguistic structure. The goal of a usage-based thesis is to depict the complexity of language use. 3. Applying cognitive grammar to pedagogical grammar Cognitive linguistics claims that the learner’s interlanguage resembles a child’s grammar where it is composed of an assortment of eclectic constructions at various levels of systematicity, abstraction, and productivity. Second language learners are attempting to master the specific array of symbolic units that represents the linguistic conventions of the target language. In a developing second language system, the target units are in direct competition with the native ones because they both represent alternative ways of construing the same reality. Second language learning can be viewed as a gradual process by which the target system gains increasingly more differentiation and autonomy from the native one, because mental experience must be organized so as to conform to the conceptual structures symbolized by the available symbolic units. Thus, it should be emphasized here that learning a foreign language will involve not only learning the forms of the language but also simultaneously learning the conceptual structures associated with these forms. The nature and purpose of pedagogical grammar requires that it focus on learning problems. The function of pedagogical grammar is to promote the learner’s insight into the foreign language system. In essence, promoting the learner’s insight means reducing the perceived arbitrariness of the foreign language system. For this reason, it is not enough to merely inform the learner that a particular element belongs to a given formal category. Also, it is not enough to merely state that such-and-such an expression is grammatically correct while other word formations are grammatically incorrect. Especially, grammatical instruction in communicative methodologies has been at the core of the pedagogical literature. The central concern is the perceived discrepancy between communicative competence and accuracy in language use. Most teachers agree that communicative activities should constitute the most important focus of the foreign language lesson, and that students’ enthusiasm to express themselves should not be stifled by undue expectations placed upon accurate grammatical production. Accordingly, the grammatical instruction in communicative approaches constitutes one of the hardest pedagogical challenges that foreign language teachers, especially non-native teachers face, because obviously the nature of grammatical COGNITIVE LINGUISTICS AND SECOND LANGUAGE PEDAGOGY 129 instruction depends critically on each teacher’s view of the nature of rules and overall organizations of the target system, as well as his/her beliefs about the specificity of grammatical knowledge to language acquisition. However, the potential of learner corpora in a data-driven learning approach explicitly makes a positive, objective contribution to such grammatical instruction. Learner corpora are strongly related to the usage-based thesis. Learner corpora could be applied to pedagogical material in at least three different ways: (i) they can help to decide what features should be particularly emphasized in teaching or even lead to the introduction of so far neglected elements; (ii) results from learner corpus studies can give indications on how to teach certain features; and (iii) results on developmental sequences can help to determine in what order language features should be taught. In other words, the more direct and probably more important way is to use a learner corpus to identify what is particularly difficult for a certain group of learners. The more indirect and more problematic way is to derive insights about second language acquisition from learner corpus analyses and to draw implications for teaching and possibly textbook writing from these insights. In a corpus-based approach, that is to say, a usage-based approach to language, the concept of entrenchment is well known. ‘Entrenchment pertains to how frequently a structure has been involved and thus to the thoroughness of its mastery and the ease of its subsequent activation’ (Langacker 1991: 45). Entrenchment is interrelated with input. Specifically, entrenchment can be identified by an adjustment of the connection weights and can be brought about by the occurrence of a specific pattern of activation which renders more likely the occurrence of the same or a similar pattern. In cognitive linguistics, linguistic constructions are seen as being abstracted from usage events by the reinforcement of recurring commonalities. In second language acquisition, the role of entrenchment has been widely accepted as one of the most decisive factors in acquiring a second language.3 It is thus significant to pay attention to individual occurrences of linguistic items and at the same time to collect numerical data about types of construction. In second language pedagogy, a data-driven analysis on the basis of both a learner 3 Tomasello (2000b: 70) points out that an important aspect of first language learning is some form of imitative learning and that ‘it is also important that children seem to have special difficulties in going beyond what they have heard when they have heard it multiple times, that is, when it is entrenched.’ NORIKO MATSUMOTO 130 corpus and a native speaker corpus is essential to explain the concept of entrenchment. This paper focuses on both complementation patterns of the verb find that Japanese-speaking learners of English use and those that native speakers of English use, and shows some significant results in the data-driven analysis on the basis of both one learner speaking corpus and one native speakers’ speaking corpus. Obviously, native speaker corpora are indeed useful for the improvement of language teaching. They are useful mainly because they can reveal what native speakers of the language in question typically write or say either in general or in a certain situation better than native speaker intuition. In deciding what content we should teach, we teachers not only need to focus on patterns revealed in the data shown in native-speaker corpora as showing target frequencies, but also need to focus on the data shown in learner corpora as showing learning gaps and relative stages in mastery. For second language teaching, nevertheless, it is not only essential to know what native speakers typically write or say, but also what the typical difficulties of the learners of a certain language, or rather of certain groups of learners of a certain language are. 4. The verb find The reason why this paper focuses on the verb find is two-fold. One is that every Japanese-speaking learner of English knows the verb find. The other is that most of Japanese-speaking learners of English are not fully aware of the fact that the familiar verb find can take various types of complements, which gives rise to a situation where there are remarkably few Japanese-speaking learners of English who have a good command of find. This paper will defuse such an uncomfortable situation by focusing on the complementation patterns of find on the basis of both one Japanese-speaking learner corpus and one native speaker corpus. In order to give my approach an appropriate context, it is necessary to explain the various complementation patterns of the verb find syntactically and semantically. Roughly speaking, find can take nine syntactic patterns, as in (1). (1) a. find + NP b. find + NP + NP c. find + that-COMP COGNITIVE LINGUISTICS AND SECOND LANGUAGE PEDAGOGY 131 d. find + that-deleted-COMP e. find + wh-COMP f. find + NP + ADJ g. find + NP + to be h. be found (passive) i. find out Semantically, find is mainly divided into ten areas, as in (2)–(11), but the various meanings of find are not always easy to keep apart. (2) GET BY SEARCHING a. I can’t find the car keys. b. Can you find me my bag? c. The child was eventually found safe and well. (3) SEE BY CHANCE a. Look what I’ve found! b. I didn’t expect to come home and find him gone. (4) DISCOVER STATE OF SOMEONE/SOMETHING a. She woke to find a man by her bed. b. He tried the door and found it unlocked. c. She looked at her glass and was amazed to find it was empty. (5) DO SOMETHING WITHOUT MEANING TO a. She woke up and found herself in a hospital bed. b. We came home and found him asleep on the sofa. c. I was disappointed to find that they had left already. d. He found he was shivering. (6) LEARN SOMETHING BY STUDY a. I managed to find a solution to the problem. b. Can you find me a hotel? c. His study found that married men and women had similar spending patterns. (7) THINK/FEEL a. Will Gary and Gail find happiness together? b. She finds it a strain to meet new people. NORIKO MATSUMOTO 132 c. She found the work very dull. (8) EXPERIENCE a. You might find that his work improves now he’s at a new school. b. I find people are often surprised at how little it costs. c. We found the beds very comfortable. d. I found the people to be charming and very friendly. (9) EXIST IN A PLACE You’ll find this style of architecture all over the town. (10) GET ENOUGH MONEY/TIME ETC He’s struggling to find the time, the support, and the resource to do all this. (11) IN A COURT OF LAW: to make an official decision in a court of law The jury found him guilty of manslaughter. The ten semantic areas correlate in interesting ways with the syntactic patterns where find can occur, as shown in Table 1. Table 1. Semantic areas correlating with syntactic patterns where find can occur. NP NP+NP that-COMP that-deleted NP+ADJ NP+to be 1. GET BY SEARCHING ○ ○ ○ 2. SEE BY CHANCE ○ ○ 3. DISCOVER STATE OF SOMETHING/SOMEONE ○ ○ ○ 4. DO SOMETHING WITHOUT MEANING TO ○ ○ ○ ○ 5. LEARN SOMETHING BY STUDY ○ ○ ○ 6. THINK/FEEL ○ ○ ○ 7. EXPERIENCE ○ ○ ○ ○ 8. EXIST IN A PLACE ○ 9. GET ENOUGH MONEY/TIME ETC. ○ 10. IN A COURT OF LAW ○ In this section, it is pivotal to demonstrate the complexities of the complementation patterns of find, that is to say, the ones of various find-constructions. A construction is defined here as follows: A construction constitutes a conventional form-meaning pairing. Within the functional and the cognitive paradigm, it is generally accepted that if one COGNITIVE LINGUISTICS AND SECOND LANGUAGE PEDAGOGY 133 verb can be followed by more than one type of complement, there must be semantic differences among the sentences with different pragmatic effects. Most previous studies have attempted to derive each semantic characterization from the different type of complement associated with it. Borkin (1973, 1984), for example, provides examples of such differences, as in (12). (12) a. I find that this chair is uncomfortable. b. I find this chair to be uncomfortable. c. I find this chair uncomfortable. (Borkin 1973: 46) Each sentence in (12) has the same propositions that the chair is uncomfortable; however, the differences among them are closely linked to ‘whether or not a complement represents a fact based on experience or, rather, describes the experience itself’ (Borkin 1984: 79). According to Borkin (1984), (12a) might be used for a judgment based on indirect evidence through asking people or learning the results of consumer reaction tests, but (12c) implies that I myself actually sit on the chair and directly experience the discomfort. (12b) might be used in either circumstance. Verspoor (2000), agreeing with Borkin, modifies her explanation. To explicate the distinctions among (12), she applies the concept of level of consciousness, which Edelman (1989) demonstrates, to the linguistic analysis of English complementation, as summarized in Table 2. Table 2. Level of consciousness with respect to E