Some months ago, I received an e-mail from a coursemate of mine, Ylenia, who asked me if I had ever thought about how we manage to recognize a “core meaning” in a polysemous word (thank you Yle for raising the point). Actually, she had almost finished her M.A. thesis (which she successfully defended on February 23), in which she carried out study of polysemy adopting a computational approach firmly rooted in the formal semantics tradition (see Pustejovsky 1995). From such an approach, polysemy is seen as an epiphenomenon, roughly “emerging from monosemy: a single relatively abstract meaning from which other sences (…) are derived on the basis of context, speaker intention, recognition of that intention by the hearer, and so on.” (Evans and Green 2006: 329). I replied to her message underlining the fact that the existence of a “core meaning” common to all the senses of a polysemous word should not be taken for granted. Indeed, from a cognitive / constructionist perspective, polysemy is approached in a rather different way, which I am going to introduce in the following few lines, partly based on a section of the M.A. thesis I am going to defend on April.
Several proponents of the Cognitive Linguistics framework (see e.g. Langacker 1987; Taylor 2003) adopt the so-called network model of polysemy, according to which the senses of a unit constitute the nodes of a network, linked horizontally by relations of similarity, and vertically by a sort of “subcase” relation1 (another model which is also widespread in the Cognitive Linguistics community is Lakoff’s 1987 radial category model, which is compatible with the network model). The lowest nodes of the network can be seen as the most specific senses of a unit, while the highest node can be seen as the most abstract representation of the meaning of that unit. Consistent with the usage-based theory of language acquisition (see e.g. Tomasello 2003; Goldberg 2006), specific senses of the item arise first, coming to light via entrenchment due to high token frequency; subsequently, the more schematic representation arises as a result of an operation of abstraction performed by the language user. It is not the case that the most abstract representation is necessarily invoked in the understanding of contextual variants; on the contrary, using a computer-related metaphor, we may say that each specific sense of a unit, in context, can be accessed randomly rather than sequentially by the language user, without having to examine all the possible senses of the unit2. From a constructionist point of view, we may use the notion of “core meaning” in the discussion of certain groups of verbs, but in a sense which is very different from Pustejovsky’s. Let us consider the case of Italian caused-motion constructions with force-exertion verbs, i.e. sentence-level constructions which express actions whereby an entity causes another entity’s movement through space by means of an act of force transmission. An active caused-motion construction will have the form illustrated in (1), the meaning captured in (2), and the argument structure depicted in (3) below:
(1) Subj V Obj Oblpath.
(2) X causes Y to move Zpath.
(3) CAUSE-MOVE (causer patient path)
Force-exertion verbs have at least one sense which describes the following situation: an agent participant exerting force on a patient: lanciare (“to throw”), gettare (“to toss”), spingere (“to push”), tirare (“to pull”), sollevare (“to lift up”), etc. With this sense, these verbs can appear in a caused-motion construction. We can label these senses of these verbs “lanciare1”, “gettare1”, “tirare1”, etc. They can be said to share a “core meaning”, which is the fact that all of them denote an act of force-exertion on the part of an agent whereby a patient is forced to move in the space. This core meaning may be labeled “Caused-Motion Action”. It captures the common property shared by a particular sense of a range of verbs. Nevertheless, it has nothing to do with the belief that a polysemous word is endowed with an abstract, underspecified meaning, with its various senses arising as the result of the application of lexical generative devices, as claimed by Pustejovsky and collaborators.
Evans, Vyvyan, and Melanie Green. 2006. Cognitive Linguistics. An Introduction. Edinburgh: Edinburgh University Press.
Goldberg, Adele E. 2006. Constructions at Work. On the Nature of Generalizations in Language. New York: Oxford University Press.
Klein, Devorah E., and Gregory L. Murphy. (2001). The Representation of Polysemous Words. Journal of Memory and Language 45: 259-282.
Langacker, Ronald W. 1987. Foundations of Cognitive Grammar. Vol. 1. Theoretical Prerequisites. Stanford, CA: Stanford University Press.
Pustejovsky, James. 1995. The Generative Lexicon. Cambridge, MA: MIT Press.
Taylor, John R. 2003 (3rd ed.). Linguistic Categorization. Prototypes in Linguistic Theory. New York: Oxford University Press.
Tomasello, Michael. 2003. Constructing a Language. A Usage-based Theory of Language Acquisition. Cambridge, MA: Harvard University Press.
Tuggy, David. (2007). Schematicity. In D. Geeraerts and H. Cuyckens (eds.), The Oxford Handbook of Cognitive Linguistics. Oxford and New York: Oxford University Press. 82-116.