Volume 73, Issue 3 - Special Issue: Describing English Constructions; Issue Editors: Barthe Bloom and Thomas Herbst

This article revisits Charles Fillmore’s vision of the lexicon as the control center of grammar, advancing the claim that idioms are ordinary linguistic objects. Within Sign-Based Construction Grammar (SBCG), idiomatic expressions – whether fixed phrases, idiomatically combining forms, or formal idioms – are modeled as typed feature structures licensed by lexical entries and constructions. This unified architecture dissolves the traditional boundary between words and rules; in it, idioms occupy distinct points along a continuum of lexical fixity rather than standing outside the grammar. Through analyses of fixed expressions like in the know , idiomatically combining forms such as causative have , and formal idioms like the Split Interrogative, the paper demonstrates how SBCG integrates idiomaticity into a general theory of sign combination, fulfilling Fillmore’s goal of a grammar grounded in the structure of the lexicon.

Constructional approaches view argument structure constructions as grammatical templates with open slots to be filled by verbs and their arguments, organized in an inheritance network. This paper argues that this view is no longer in tune with the field’s move towards dynamic usage-based models and proposes two frame shifts. First, rather than a rigid hierarchy of templates and fillers, linguistic structure is best modeled as a network of interdependent constraints that hold each other in balance. Second, rather than organizing the grammar network taxonomically, according to what constructions are , we can organize it according to their interactions – what they do . The approach is illustrated for English argument structure and related constructions in Fluid Construction Grammar, a computational platform for developing and testing construction grammars. The result is a constructional framework that preserves the insights of earlier models while generalizing them to all forms of constructional interaction.

This paper presents a usage-based, constructional analysis of the so-called BIG MESS construction in Present Day English (e.g. how big a mess , that big a deal , too strong a word ). Data from the Corpus of Contemporary American English (COCA) is presented in order to analyze the frequency, genre distribution and lexical biases of seven subtypes ( how , however , so , as , too , this , that ). The paper also zooms in on the feature of ‘discontinuous modification’ by a clause (e.g. too ADJ qualitative a N sg to -Clause, as ADJ qualitative a N sg as- Clause). The main aim is to sketch the constructional network of the BIG MESS subtypes, discussing the respective form-meaning pairings with their differences in usage and their parent and peer relations. By using collostructional and configural frequency analyses, it is shown that some of the constructional types are formally and functionally very different from each other, which is why their relatedness is called into question. It is argued that not all subtypes form a tight constructional family but that some of them are distant relatives at best. Additionally, the existence of an abstract ‘mothernode’, i.e. [ADV intensification ADJ qualitative a N sg ] Cx will be discussed.

This article provides an account of the different types of information that are used to characterize constructions in the approach taken in Herbst, Thomas, and Thomas Hoffmann. 2024. A Construction Grammar of the English Language. CASA – A Constructionist Approach to Syntactic Analysis . Amsterdam, Philadelphia: Benjamins. and the CASA|Con constructiCon. The most important types of information are those that characterize the construction as such (type C) and those that show the relation of a construction to others in the constructional network (type N). Special attention is given to collo-profiles and problems of identification of constructions.