Volume 49, Issue 6 - New Reflections on Coercion

Whether words can be coerced by constructions into new uses is determined in part by semantic sensicality and statistical preemption. But other factors are also at play. Experimental results reported here suggest that speakers are more confident that a target coinage is acceptable to the degree that attested instances cover the semantic space that includes the target coinage. The relevance of coverage is supported by combined effects of type frequency and variability of attested instances [Experiments 1a–1b], and an expected interaction between similarity and variability [Experiment 3]. Similarity to an attested instance is also found to play a role: speakers are more confident of a target coinage when the coinage is more similar to an attested instance [Experiment 3]. Experiment 2 provides a manipulation check that indicates that participants are in fact basing their confidence ratings on the perceived productivity of constructions. The results reported here lend support to the idea that the productivity of constructions depends on general properties of induction.

This article investigates the role of coercion and contextual background information in influencing acceptability judgments. Focusing on data from English resultative constructions I propose a usage-based constructional approach that accounts for the factors that allow for a verb's conventionalized argument structure specifications to leak, thereby allowing otherwise unacceptable nonconventionalized utterances such as ??Ed hammered the metal safe to be judged acceptable by means of coercion. By putting less emphasis on independently existing meaningful constructions I argue that frame-semantic information at the level of lexical units (mini-constructions) can be used effectively to link semantic information to syntactic information to arrive at both lower-level and higher-level constructional descriptions for both decoding and encoding purposes.

The present article challenges the claim that the notion of coercion, as understood in Construction Grammar(s) (henceforth CxG), is superfluous and can simply be accounted for in the light of metaphoric and metonymic extension in general (Ziegeler, Journal of Pragmatics 39: 990–1028, 2007, Constructions and Frames 2: 33–73, 2010). Specifically, coercion will be argued to be of pivotal importance to account for the semantico-pragmatic and discourse-functional properties of the family of the English subjective-transitive construction, understood as a web of interrelated form-function pairings (Gonzálvez-García, Language Sciences 31: 663–723, 2009b). Two substantive broad-scale generalizations emerge from the theoretical discussion as well as the cursory bottom-up, data-driven analysis on which this article draws. First, although the phenomenon of the interpretation of slot-determined meaning may involve metaphor and metonymy, a descriptively and explanatorily adequate account of the kinds of type-shifts allowable in the family of the subjective-transitive construction in English requires us to posit a variety of lower-level (verb-class or verb-specific) subconstructions, which vitiate a general-purpose metonymy-based account. Crucially, these constructions do not appear to be amenable to an explanation in terms of metaphor or metonymic extension alone, thus pointing to the inevitability of retaining coercion as well as constructions as free-standingtheoretical entities. Second, it is proposed that the phenomenon of coercion should be understood in terms of a continuum, that is, as a scale representing the extent to which a type of coercion is entrenched and conventionalized (Langacker, Cognitive Linguistics 20: 167–176, 2009) rather than as an allor-none property leading to a full matching or, by contrast, a semantic conflict between specific lexical items and constructions. Thus, what actually needs to be refined is the notion of the (degree of ) semantic (in-)compatibility between lexical items and construction meaning. At a higher level of granularity, it is further suggested that the notion of coercion be treated in terms of a continuum at a vertical or intra-constructional level, that is, within items of the same paradigm as well as a horizontal or inter-constructional level, among different instances, including lower-level configurations of the same construction.

Fundamental to narrative is the ability to indicate what events overlapped what other events. Crucial to this ability is stativization: only stative situations can include (as opposed to being included within ) reference time. But how exactly are stative type-shifts effected, and what does it mean for an event to be “turned into” a state? There are two purported paths to stativity: use of a compositional type-shifting device, as exemplified by (i), and coercion, the creation of resolvable semantic conflict between a combinatoric pattern and an openclass word, as exemplified by (ii): (i) The House is voting on the legislation. (ii) We talk on the phone every Sunday What is the trigger for stative coercion in (ii)? According to de Swart (Natural language and linguistic theory 16: 347–385, 1998, Coercion in a cross-linguistic theory of aspect, CSLI Publications, 2003) and others, it would be the iterative adverbial every Saturday . But iteration is insufficient to secure stativity, as, e.g., activity verbs, which also consist of repeated actions, denote events rather than states. The stativity of progressive sentences is likewise mysterious under the standard analysis: while the progressive is typically said to highlight the pre-culmination portion of an event representation (Parsons, Events in the semantics of English, MIT Press, 1990; Langacker, Foundations of cognitive grammar, Stanford University Press, 1987; Smith, The parameter of aspect, Kluwer, 1997), that portion is presumably an activity rather than a state. To resolve these paradoxes, I propose a selection-based model of two kinds of stative type-shifts: those effected by derivational means (e.g., the progressive construction) and those effected by inflectional means (e.g., the English present tense). I then provide a formal representation of the relevant constructions using conventions of Sign Based Construction Grammar (SBCG) (Sag, Language 6: 486–545, 2010, Michaelis forthcoming). According to this model, stativizing constructions both denote states and select states in the Aktionsart representations of verbs with which they combine. This model relies on the existence of rests, periods of stasis entailed by the Aktionsart representations of dynamic verbs. I argue that by viewing stative typeshifts as the products of construction-verb combination, we can explain: (i) the relationship between a verb's input and output representations, (ii) how tense inflections affect the Aktionsart representations of verbs with which they combine and (iii) the functions of the so-called relative past tense . In addition, I argue, that if we view stative type shifts as functions of constructions, rather than, say, the products of semantic rules, we can better understand differences in the combinatoric potential of a given tense inflection both across languages and over historic time.

In this article, I address the general mechanisms of argument selection in language, and specifically the role played by type coercion operations in linguistic theory. The view presented here is that compositional mechanisms must be open to more information than just the base type of an expression. Such “compositional flexibility” in the type that accommodates to the one selected by the verb has been the topic of considerable research recently. In this article, I position how such coercion mechanisms operate in the larger context of general selectional strategies in the grammar. I first outline my assumptions regarding the general theory of type selection, adopting a model of Generative Lexicon Theory (GL). I distinguish between two strategies at work in language, type matching and type coercion, both treated as strong compositional mechanisms. A library of possible coercion operations is defined, as well as apparent constraints on their application in language. I conclude with a discussion of corpus-based experiments based on the formal models of coercion and selection presented here.