Language, usage and cognition
Citation
Contents
Author
Context
linguistics and philosophy have tended to analyze language without reference to usage and experience. Chomsky’ (competence and performance) and Saussure’s (langue and parole)
Thesis
grammar is as an emergent system consisting of fluid categories and dynamic constraints that are in principle always changing under the influence of general cognitive and communicative pressures of language use. grammar consists of emergent form-function units, or ‘fluid constructions’ (Goldberg 2006), which are related to each other by probabilistic links that are determined by their similarity and cooccurrence in usage. Since constructions involve the same cross-modal associations of form and meaning as words and morphemes, they are subject to the same cognitive processes as lexical expressions. Domain-general processes are cognitive mechanisms that are relevant not only for language but also for other cognitive phenomena such as vision and thought.
Methodology
Key Terms
rich memory (Ch. 2) - the storage of detailed information from experience
chunking (Ch. 3) - the formation of sequential units through repetition or practice
analogy (Ch. 4) - the mapping of an existing structural pattern onto a novel instance
categorization - the most general cognitive mechanism that accounts for the identification of individual tokens through comparison with previously established categories
cross-modal association - the cognitive capacity to link form and meaning
Criticisms and Questions
Notes
In the usage based approach, there is no (innate) language faculty, and grammar is grounded in general cognitive processes that are involved not only in language but also in nonlinguistic cognitive activities (e.g. visual and auditory perception, nonlinguistic memory, joint attention, and reasoning).
There is evidence that ‘perspective taking’ and ‘mind reading’ are fundamental capacities of the human mind that influence the use and development of language in specific ways.
Thesis 2: The synchronic analysis of grammar cannot be separated from the analysis of diachronic change
the book emphasizes the importance of diachronic processes for the analysis of grammar. grammaticalization has ‘demystified’ the mental view of grammar that underlies the generative approach and has provided strong evidence for the assumption that grammatical patterns are shaped by general cognitive processes.
Thesis 3: Frequency is an important determinant of language change
The preserving effect of token frequency results from the interaction between the strength of linguistic representations in memory and the power of analogy.
Thesis 4: Linguistic categories are based on concrete tokens.
grammatical categories and constructions are based on concrete tokens or exemplars. every linguistic expression in usage contributes to the representation and development of linguistic categories and constructions. The theory rests on the assumption that language users are endowed with a very rich memory system that allows them to store large amounts of information, which may even survive (in memory) if this information is subsumed under a generalization.
"memory is cheap and computation is costly."
Thesis 5: Syntactic structure is lexically specific.
In the generativist theory of grammar, syntactic structure is abstract and independent of concrete lexical expressions; grammar and lexicon are strictly separated in this approach. But in the usage-based model of grammar, syntactic structure is lexically specific: most grammatical constructions are associated with specific lexical items, which is of course a consequence of the fact that grammar emerges from our experience with concrete tokens.
Thesis 6: Grammatical categories are gradient.
grammaticalization usually consists of a sequence of many small changes whereby an existing construction is gradually transformed into a new one (especially Ch. 6). Since the source construction and other intermediate constructions are not generally discarded in this process, they often constitute a chain of related grammatical patterns with fuzzy boundaries between them.
Thesis 7: Constituent structure is determined by chunking.
Thesis 8: The meaning of grammatical markers and constructions is polysemous.
grammatical meaning does not involve binary semantic oppositions, as proposed by Jakobson, but reflects the language users’ experience with particular situations. Since our experience of the world is open-ended, the meaning of linguistic expressions cannot be adequately analyzed by means of a restricted set of semantic features; rather, what is needed is a dynamic theory of meaning, in which the semantic features of linguistic expressions are determined by their use in different situations and contexts. Since linguistic expressions are never tied to one particular situation—that is, they are always used in multiple situations and contexts—they are usually polysemous.
Thesis 9: Linguistic productivity involves analogy rather than rules.
If linguistic categories are gradient and lexically specific, linguistic productivity cannot be analyzed in terms of traditional rules. In generative linguistics, grammar is commonly characterized as a closed, deductive system that has been analyzed in terms of formal language theory (cf. Chomsky 1957). On this view, grammar involves discrete symbols and categorical rules that allow for no exceptions. In the usage-based analysis of linguistic productivity, traditional rules have been abandoned and replaced by schemas (Bybee 1995, Langacker 2008). A schema is a grammatical template, or abstract construction, that has evolved through generalization over concrete tokens. Since there are usually multiple schemas that a speaker can potentially use to produce (or comprehend) a novel utterance in a particular situation, linguistic productivity can be defined as the process whereby language users select a specific schema from a set of alternatives (Langacker 2008). The choice (or selection) of a schema is determined by a cognitive process that B and others have characterized by the notion of analogy, which in turn is influenced by two factors: (i) type frequency and (ii) similarity (Ch. 4).
Thesis 10: Language universals are dynamic and statistical rather than static and absolute.