Cognitive approaches to grammar
Cognitive Grammar
Discussion
One of the features setting Cognitive Grammar apart from other cognitive models is the contention that notions such as “subject” and “object” can be defined schematically, i.e. independently of specific forms (or realizations) in a way that encompasses all their uses across all languages. Indeed, such a position may appear controversial (especially in the light of typologically-oriented research, see the discussion of Radical Construction Grammar). Cognitive Grammar claims for example that there should be analyzed as a subject in a sentence like There are a lot of people in the lift (i.e. a presentational there-construction). There, in the case at hand, is categorized as a subject because it refers to an abstract setting which is used as a first trajector, i.e. an initial reference point for accessing the described situation, or a spotlight of prominence (see also Langacker’s 2005b analysis of it-subjects). Observe that, under this analysis, the trajector (there) refers to an entity larger than that evoked by the landmark (a lot of people). That is, the schematic definition of subject does not necessarily pick out smaller entities as trajectors (cf. also the setting-subject construction The garden is swarming with bees, where the trajector the garden refers to a location and the smaller and moving entities, the bees, are the clause-level landmark). Related to this is the question of whether the analysis of presentational there constructions poses any problems for a sentence such as Down the hill came a strange procession (an inversion construction, see Chen 2003). Is down the hill also to be categorized as a subject? Down the hill, like there, refers to a setting, although a concrete one, and, like there, can be taken to be the starting point for accessing the configuration depicted by the clause (see Langacker 2005b). Still, down the hill is categorized as a landmark in Cognitive Grammar.
To be sure, the Cognitive Grammar analysis of the notion of subject defies traditional analyses. Traditionally, subject identification is based on morphological and/or syntactic criteria (e.g. verb-agreement and word-order, respectively, in English). Consequently, distributional facts point to both there and the postcopular noun phrase (e.g. a lot of people) as having subject status. The morphological criterion (at least in Standard English, cf. colloquial There’s a lot of people in the lift where a lot of people does not agree with the copula) identifies a lot of people as a subject. The syntactic criterion (cf. Are there a lot of people in the lift?) forces us to conclude that there is also a subject. In the case of the inversion construction, by contrast, distributional facts would lead us to identify a strange procession as the subject.
At the very least, one can conclude that the Cognitive Grammar analysis is conceptual, not grammatical (in the traditional sense). That is, the Cognitive Grammar analysis is not necessarily meant to pick out those “subjects” identified by formal criteria. After all, distributional evidence simply points to the fact that some elements have a special status compared to others. The deeper and more interesting question is why this is so. Cognitive Grammar contends that such special (formal) behavior is symptomatic of a specific cognitive ability (i.e. trajector/landmark alignment).
The word class analysis proposed by Cognitive Grammar may also appear controversial. As is pointed out by Taylor (2002: 563), (acquisitionally) word classes emerge as abstractions out of specific constructions (see also Tomasello 2003). The question is: is it possible to reconcile a usage-based perspective with the construction-independent characterization summarized in Figure 1 in Word classes and grammatical functions in Cognitive Grammar? At present, the question cannot be answered satisfactorily because, among other things, the psychological reality of some of the notions invoked by the Cognitive Grammar analysis has not been proved. This is the case of summary and sequential scanning, whose existence (and relevance to word class categorization) has not been validated experimentally yet (see e.g. Broccias and Hollmann 2005, Duffley 2005; but see also Matlock 2004 and Matlock et al. 2005 for some psycholinguistic experiments involving the notion of scanning).
Nevertheless, the Cognitive Grammar approach remains very innovative because it strives to link (traditional) grammatical categories with cognitive processes. The fact that certain schemas (e.g. noun, adjective, subject, etc.) emerge out of specific usage events at all should ultimately be related to our cognitive abilities. Of course, the details of what abilities are involved, how they interact with one another and the extent to which they define cross-linguistically valid schemas remain to be established.
