Volume 21, Issue 3

The debate over how best to characterize inflectional morphology has been couched largely in terms of the “dual-mechanism” approach described in Pinker (Words and rules: the ingredients of language, Basic Books, 1999) versus “single-mechanism” connectionist approaches derived from Rumelhart and McClelland (On learning past tenses of English verbs, MIT, 1986). There are, however, other single-mechanism approaches. The exemplar-based or analogical models of Daelemans et al. (TimBL: Tilburg Memory-Based Learner, version 4.3 reference guide, ILK, 2002) and Skousen (Analogical modeling of language, Kluwer Academic, 1989) also model inflectional usage accurately within a single-mechanism. The most striking theoretical claim peculiar to these purely analogical models is that they do not posit any resident linguistic generalizations for processing language. Instead they process new instances of usage by comparing them systematically to remembered instances of previous usage. Based on a comparison of their Minimal Generalization Model with an adaptation of Nosofsky's (Journal of Mathematical Psychology 34: 393–418, 1990) Generalized Context Model, Albright and Hayes (Cognition 90: 119–161, 2003) argue that such purely analogical models are intrinsically inadequate for modeling the English past tense. This paper shows, however, that Skousen's (Analogical modeling of language, Kluwer Academic, 1989) Analogical Model performs as well as the Minimal Generalization Model. The implications of these results for cognitive linguistics are discussed.

Differences between languages in count (e.g., cup) and mass (e.g., rice) nouns have been shown to impact cognition, but few studies have directly examined how the morphology associated with count and mass constructions is acquired and linked to differences in meaning. Two experiments examined the relation between English and Spanish plural morphology and the interpretation of nouns as individuated objects. In Experiment 1, English- and Spanish-speaking children and adults participated in two tasks. One task examined how participants produced plurals for nouns. Results from this task indicated that both language groups make a distinction in their use of plural morphology for count and mass nouns between 5 and 7 years of age. However, that morphological distinction was stronger and occurred earlier among English speakers. A second task examined whether the count (but not the mass) nouns in each language denoted individuated objects. Speakers of both languages tended to treat all nouns as referring to individuated objects at 5 years of age. Beginning at 7 years of age, English speakers made a reliable distinction between the referents of mass and count nouns. Speakers of Spanish however, treated both types of nouns as referring to individuated objects in this task throughout development. Experiment 2 examined the interpretation of nouns by adult speakers of both languages using a different task, and the results offer converging evidence that Spanish speakers are more likely than English speakers to interpret nouns as referring to individuated objects. The reason for the difference between English and Spanish is discussed in terms of the proportion of count and mass nouns acquired early in development and as a function of developing concepts.

This paper argues that in the context of the copular complement construction ( est possible que ‘is possible that’ for example), French possesses two impersonal constructions respectively introduced by il ‘it’ and ça ( c' ) ‘this’. This analysis runs counter to most syntactic accounts which structurally distinguish impersonals ( il ) from dislocated ( ça ) constructions. Two arguments are proposed in defense of the two impersonals analysis. First, following the Cognitive Linguistics tradition, it shows that il should not be considered a meaningless dummy but a referential (albeit general) expression. Secondly, a comparison with ceci ‘this’, a pronoun with an unquestionable cataphoric sense, reveals that ça cannot be considered a cataphoric pronoun, and that its meaning within the context of the copular complement construction is very close to il's . Consequently, the syntactic competition between these two pronouns reflects their conceptual overlap, and their distribution in discourse is motivated by their semantic differences. Beyond its relevance to the understanding of the il/ça distribution, the analysis presented in this paper not only argues in favor of a broader account of impersonals than syntactic accounts generally advocate, but also provides a way of constraining the kinds of constructions which receive the impersonal label.

This research analyses the semantics of Spanish sentences with causal or adversative connectives using force dynamics, emotional valence and subjectivity parameters. Participants were given stimulus sentences, each followed by a connective, and were asked to generate meaningful continuation sentences. For each stimulus sentence, four versions were offered, differing only in the connective (two causal and two adversative). Participants' responses were encoded using a set of variables related to force dynamics, emotional valence, subjectivity, complexity and continuity. A discriminant analysis of the data found two main statistical functions. The Continuity-Discontinuity function (polarity) discriminated between causal and adversative sentences: force dynamics and emotional valences tended to be continuous in sentences linked by a causal connective and discontinuous in sentences with an adversative connective. The Internal-External function, orthogonal to continuity, discriminated between specific connectives: sentences with “porque” (because) and “aunque” (although) usually expressed internal or volitional events, and sentences with “puesto que” (given that) and “a pesar de que” (in spite of) were associated with external or factual social events.

This article approaches common topics in the diachronic literature on modal categories from the perspective of adjectives. It thus expands on what has been found for the better studied category of modal auxiliaries as regards sources of modal meaning and pathways of change. Most importantly, it proposes two new pathways from premodal to (dynamic) modal meaning, one followed by essential and vital , and one followed by crucial and critical . It also shows that in the four cases the development of dynamic meaning depends on the emergence of two semantic properties, viz. relationality and potentiality. Finally, this study makes it clear that the mechanisms driving the various semantic changes are not new, but rather have proved useful in explaining a varied set of developments. For the final semantic extension of the adjectives from dynamic to deontic meaning, for instance, the process of subjectification (Traugott, Language 65: 31–55, 1989) will be invoked.

The present paper examines the variation between finite and infinitive complements of the Portuguese perception/causation verbs ver (‘see’), ouvir (‘hear’), sentir (‘feel’), deixar (‘let’) and fazer (‘make’) from a cognitive grammar perspective. It is argued that the distribution of the structures main verb + finite /infinitive complement can be explained by iconicity and subjectification . The hypothesis is put forward that the structure perception verb + infinitive complement designates direct physical perception, while the structure perception verb + finite complement designates an inferential relation between the main verb and the complement event. In addition, the structure causation verb + infinitive complement designates direct causation, whereas causation verb + finite complement designates an indirect causation with inferential features. Further, it is claimed that the inferential and conceptually more complex character found in the structure main verb + finite complement represents a prime example of subjectification. Being so, it is argued that Portuguese complementation provides a remarkable connection between iconicity and subjectification.