Word classes: linguistic categorisation
Recall that the scope of predication of a linguistic expression is its base, and its profile is what the expression designates from within that base. We have also seen that symbolisation is the link between the phonological and semantic poles of a linguistic unit, while coding is the link between a linguistic unit and a speech event. As we saw in Chapter 14, the cognitive model views lexicon and grammar in terms of a continuum of symbolic units within the inventory rather than in terms of separable subsystems of language. Indeed, Langacker was an early pioneer in developing this view. At the open-class end of the continuum, units have rich and specific content meaning, and at the closed-class end of the continuum, units have schematic meaning. Despite broad acceptance of the distinction between open- and closed-class expressions, Langacker (1987: 18–19) cautions against viewing these as discrete categories. He argues that just as conceptual categories relating to content words have fuzzy boundaries, so do grammatical categories. This entails that certain linguistic expressions may fall at the periphery – or near the middle of the continuum. For example, the expression thing, which has been called a ‘conceptual shell’ (Schmid 2000), is an open-class word, but lacks the semantic specificity of a prototypical open-class word like cat. Langacker also points out that while the closed classes are resistant to change, they are not immune to it. In other words, the closed classes are not entirely closed. He provides the example of the Southern US expression y’all (second person plural), which has entered the ‘closed’ class of personal pronouns, an expression that has the counterpart yous in certain dialects of British English.
In contrast to the distributional approach to the characterisation of word classes (see Chapters 14 and 22), Langacker adopts the position that semantic characterisations of the major word classes are possible. Furthermore, Langacker supports the cognitive semantics model of categorisation, arguing that the formal view of category membership in terms of necessary and sufficient conditions should be abandoned in favour of a prototype model (see Chapter 8). Langacker argues that grammatical categories, like conceptual categories, display prototype effects and that a semantic characterisation of the category prototypes is therefore uncontroversial. In other words, it is only problematic to define nouns in terms of THINGS (matter) and verbs in terms of PROCESSES (action) if we assume that these rather specific semantic properties should hold for all members of the category, an idea that follows from a necessary and sufficient conditions model of categorisation. It is for this reason that a semantic characterisation of word classes is traditionally disfavoured in com parison to a structural characterisation based on morphological features and syntactic distribution.
However, the idea that prototypical nouns and verbs might have a semantic characterisation is not at the heart of Langacker’s proposal. The crux of his proposal is rather that all nouns and verbs have a ‘schematic semantic characterization’ (Langacker 2002: 60), and furthermore that these characterisations are universal. To illustrate the idea that word classes can be described in terms of schematic meaning, consider the following examples:
Although the verb love in (1a) and the noun love in (1b) might be difficult to distinguish in terms of content meaning, Langacker argues that they do encode different meanings because they encode different construals of the scene. The same argument applies to the verb destroy in (2a) and the noun destruction in (2b). As we saw in Chapter 15, construal is central to the choices that speakers make about how a scene is linguistically ‘packaged’, and this in turn explains the availability of related yet distinct constructions. For example, the nominal expressions in (1b) and (2b) involve the process of reification, which construes what Langacker calls a PROCESS (action) in terms of what he calls a THING (matter). As we will see in this chapter, construal is central to Langacker’s theory of word classes.
