Constructions and Frames - Volume 7, Issue 1, 2015

Adopting a corpus-based approach, lexical collocations are reconsidered from a lectal perspective. Analyzing adjective-noun collocations, it will be shown that lexical collocations are conditioned by the language settings in which they are used. These lectal constraints do not only apply to lexical collocations as a measure of lexical association but also to their potential function as a determinant of other constructions. These results argue for the inclusion of the heterogeneity of the corpus settings in empirical linguistic models and for the integration of a full-fledged lectal dimension in theoretical frameworks advocating a usage-based methodology, such as construction grammar.

In this paper I focus on MGr lίγo ‘(a) little’ in non-quantifying constructions codifying affect and argue that it has grammaticalized from a quantifier to an (inter)subjective metalinguistic hedging device: i.e., a positive politeness marker in the hedged directive construction and a marker of attitude to one’s own utterance in the hedged statement construction. Motivation for this development is provided by MGr interactional ethos and the importance attached to the friendly, informal “politeness of involvement”, a factor discussed in the context of universal tendencies in the semantics of diminutives and affective language. I argue that non-quantifying lίγo is a case of semantic change motivated by sociopragmatic considerations which are anchored to specific constructions.

My aim in this paper is to provide evidence from diachronic linguistics for the view that some predictable units are entrenched in grammar and consequently in human cognition in a way that makes them functionally and structurally equal to nonpredictable grammatical units, suggesting that these predictable units should be considered grammatical constructions on a par with the nonpredictable constructions. Frequency has usually been seen as the only possible argument speaking in favor of viewing some formally and semantically fully predictable units as grammatical constructions. However this paper presents evidence from historical linguistics, and more specifically diachronic arguments of a syntactic nature, to support the argument of frequency. I demonstrate that contrary to other Medieval Greek future periphrases, the Early Medieval Greek form [méllo (‘I am about to’) INFINITIVE (INF)] was both semantically and formally predictable. Despite this difference, [méllo INF], like the other future periphrases, seems to be highly entrenched in the cognition (and grammar) of Early Medieval Greek language users, and consequently a grammatical construction. The syntactic evidence speaking in favor of [méllo INF] as a grammatical construction consists in the following features: 1) [méllo INF] is more resistant to restructuring than many other verbal periphrases with an INF; and 2) in LBG [méllo INF] seems to have grammaticalized — as one of its uses — a specialized function embedded in subjunctives, in which méllo was semantically empty. Both these developments indicate that despite being both formally and semantically predictable, [méllo INF] was highly entrenched in EBG grammar.

This work follows the progression of a grammatical construction that unifies the Romanian accusative preposition pe with a coreferential pronominal clitic, together forming the cd-pe construction. On the basis of historical texts, it is argued that these two grammatical phenomena evolved into a clause-level construction with a dedicated semantics and pragmatics in the modern language. A corpus analysis illustrates how cd-pe won out against a competitor pe-only construction that persisted until as recently as the early 20th century, and which is still retained in some dialects and registers. The broader scope is to refocus the discussion of clitic doubling on pragmatic motivators in light of diachronic constructionalization processes, and to reflect on the nature of similar clitic doubling phenomena evident in other Romance languages in terms of the entrenchment of language-specific constructionalization processes.