Symmetric predicates and their argument realization: a case study and some implications for the interaction of verbs with constructions
Symmetric verbs such as meet and marry are normally defined as predicates that have a bidirectional relation between their participants built into the meaning of the lexeme. In this paper I will examine the interaction between the different senses of the symmetric verb meet and its possible argument structures. It will be argued that a classical construction grammar account, which strives for monosemy maximisation, would not be able to account for both the systematicity in argument realization and the subtle semantic differences between the various uses of meet. Furthermore, the assumption that every verb has a lexically specified valency pattern turns out to be problematic in the case of verbs like meet. In contrast to most other English verbs meet does not easily lend itself to an analysis in which one of the possible argument realizations is basic whereas the others are derived from the basic pattern by way of verb-construction interaction.
On the basis of corpus data possible ways of deriving the senses of meet from verb and constructions will be considered. The two extreme positions – (i) the different senses of meet are derived from a monosemous verb and the construction in which it occurs, or (ii) the different senses of meet are all lexically represented, independently of each other and their possible argument structures – are not tenable. In the case of (i) one would either underdetermine the relevant meanings or be forced to assume a high number of transitive clause constructions, each with a special meaning that is suited to derive the relevant sense from a monosemous verb. In the case of (ii) one would lose generalizations concerning the systematic nature of how verb senses and constructions may or may not co-occur. I will propose a network representation of the different senses and argument realization patterns of meet which more adequately accounts for the behaviour of this verb.
More generally I will argue that the way the polysemy-monosemy dichotomy is treated in cognitive linguistics should be incorporated into the representation of verbs and constructions in Construction Grammar, thereby avoiding the assumption of an implausible number of related meanings a construction can have. A network which represents the relations between the different senses of a verb, their possible argument structure configurations and the degree to which the two are connected in the speaker’s mind appears to solve the problem of deriving the different uses of symmetric predicates. At the same time it contributes to characterizing the meaning of one of the most abstract constructions in English: the transitive clause. As is well-known, symmetric verbs demonstrate that the meaning contribution of this construction may reduce to asymmetries in perspective or topicality differences between participants. In this way they are a good test case of firstly how abstract a syntactic construction can be and secondly of how the interaction between verbal meaning and constructional meaning exactly works.