Constructions and Frames - Volume 17, Issue 1, 2025

Creation verbs are traditionally divided into at least two classes: verbs of creation of an entity and verbs of image creation. Such classification analyses face two important problems: how to account for the common creation semantics and how to explain the behavior of verbs which can describe either type of creation. Assuming a cognitive-semantic approach and based on Brazilian Portuguese data, this paper proposes that types of creation are different construals of a unified creation event, represented as a frame-based causal chain. Thus, the common creation semantics is defined as a causal chain, which is evoked by creation verbs in general, and the distinct properties of different classes of creation verbs follow from profiling of specific parts of that frame via argument structure constructions. Some creation verbs are lexically constrained as to how they construe the creation event, but others have the potential to construe the event in different ways.

We argue for the lexicon-grammar continuum by looking into the separability of Persian complex predicates, an open class of verb constructions for which it has been argued that the components are either separable or inseparable. We contend that separability is best described as a scalar rather than an all-or-nothing phenomenon. Our analysis of zadan ‘to hit’ within the framework of Construction Grammar yields a semantic description of zadan complex predicates as a radial category. We measure both the frequency of zadan complex predicates and their rate of separation based on corpus attestations. We explore the relationships between meaning, frequency, and separation rates and the implications of our findings for cognitive linguistics and construction grammar. We find that semantic compositionality and lower frequency are associated with a higher rate of separation. Overall, our data points towards separability as a continuum rather than a binary opposition.

The present paper investigates the similative-pretence alternating pair in Mexican Spanish (como/igual que and como si constructions) based on the analysis of 1362 instances from The Corpus del Español NOW (News on the Web) corpus. We apply a revised version of distinctive collexeme analysis to our concordance data to explore the variety of different verbs that can occur in the first slot of similative and pretence constructions. Specifically, we use (i) a new measure that distinguishes the attraction that a verb lemma exerts on a construction from the attraction that a construction exerts on a verb lemma and (ii) bootstrapping on the level of files to, for the first time, provide proper confidence intervals in the context of collostructional studies. The from-verb lemma-to-construction analysis shows that there is a significant attraction of epistemic lemmas (e.g., parecer ‘to seem’) to occur in como ‘like’, and of mistaken identity lemmas (e.g., actuar ‘to act’) to appear in como si ‘as if’ constructions. The from-construction-to-verb lemma analysis, on the other hand, demonstrates that igual que ‘like’ significantly attracts epistemic perception verbs (e.g., ver ‘to look’), and como si ‘as if’ constructions significantly attract the verb lemma sentir ‘to feel.’

In this paper, we propose a model that accounts for the change in the grammatical status of the members of the Hebrew Ultimate construction family, the only full sentences in Hebrew to undergo lexicalization, from extra-sentential elements to full-fledged words, specifically, flexible modifiers. We propose a specific type of contextual construction — the Anaphoric degree-adverb exclamative — and argue that it is this construction that enables both the grammatical change and the categorial flexibility of the newly evolved words by incorporating the members of the Ultimate construction family as subordinate clauses. Our analysis emphasizes the critical role of network links between the (various) constructions involved in this lexicalization process.

This paper documents and reviews the state of the art concerning computational models of construction grammar learning. It brings together prior work on the computational learning of form-meaning pairings, which has so far been studied in several distinct areas of research. The goal of this paper is threefold. First of all, it aims to synthesise the variety of methodologies that have been proposed to date and the results that have been obtained. Second, it aims to identify those parts of the challenge that have been successfully tackled and reveal those that require further research. Finally, it aims to provide a roadmap which can help to boost and streamline future research efforts on the computational learning of large-scale, usage-based construction grammars.