Coercion: a case of saturation
Résumé
This paper investigates the notion of coercion from the perspectives of Construction Grammar and Relevance Theory, and proposes a reanalysis of the process in terms of saturation. Construction grammarians have paid particular attention to cases of compositionality that involve a semantic incompatibility between a lexeme and the construction in which it is used, as in the sentences in (1) to (2). These examples are said to illustrate a process of coercion (Goldberg 1995:159, Michaelis 2004:2, Yoon 2012:2) whereby the semantics of each construction (X IS THE NEW Y in (1) and the WAY construction in (2)) is mapped onto that of the lexeme (referred to as the override principle, Michaelis 2004:25). This observation has essentially provided support for the construction status of argument structures. This status, however, is generally not recognized within the pragmatics literature. Instead of referring to coercion, for instance, relevance theorists tend to analyse the interpretation of Sampras and Houdinied in terms of ‘free’ pragmatic enrichment. The re-interpretation of the lexemes is argued to be pragmatically motivated by the need to arrive at an optimally relevant interpretation.
(1) Federer is the new Sampras. (Wilson, 2004:3)
(2) He Houdinied his way out of the closet. (Wilson, 2004:4)
These two perspectives seem to provide contradictory analyses of the observed phenomenon of coercion. However, the aim of this paper is to question this apparent opposition and to argue that a better understanding of this notion can be arrived at by combining insights from the two frameworks. On the one hand, I will argue that it is necessary to treat coercion not as a systemic mechanism but as a context-dependant process in which speakers are involved (Lauwers and Willems, 2011:1224). At the same time, contra Ziegeler (2007), it will be shown that coercion is not a superfluous notion that can be better captured solely on the basis of pragmatic principles (Gonzálvez-García, 2011:1316). The construction status of argument structures has received empirical support (cf. Bencini and Goldberg, 2000:648). It is the semantics of these constructions that triggers coercion and the re-interpretation of the lexeme found in them.
Finally, I will argue that the process of coercion is in fact best described in terms of saturation (Recanati 1989:304, Carston 2009:50). The notion of saturation makes it possible to capture both the linguistic origin of the process involved (as argued in Construction Grammar) and the non-linguistic nature of the actual re-interpretation (as argued in Relevance Theory). In the sentences in (1) and (2), re-interpretation of the lexemes is first triggered by the semantics (and idiosyncratic properties) of the constructions used, but occurs in accordance with the principle of relevance.