Grammatical Constructions as Relational Categories

Abstract

This paper argues that grammatical constructions, specifically argument structure constructions that determine the "who did what to whom" part of sentence meaning and how this meaning is expressed syntactically, can be considered a kind of relational category. That is, grammatical constructions are represented as the abstraction of the syntactic and semantic relations of the exemplar utterances that are expressed in that construction, and it enables the generation of novel exemplars. To support this argument, I review evidence that there are parallel behavioral patterns between how children learn relational categories generally and how they learn grammatical constructions specifically. Then, I discuss computational simulations of how grammatical constructions are abstracted from exemplar sentences using a domain-general relational cognitive architecture. Last, I review evidence from adult language processing that shows parallel behavioral patterns with expert behavior from other cognitive domains. After reviewing the evidence, I consider how to integrate this account with other theories of language development.

The generalisation of linguistic constructions in children with or without developmental language disorders

“…This could also have disentangled the influences of low variability and progressive alignment. Nevertheless, these results are in line with the theory according to which analogical comparisons lead to relational abstractions (Gentner & Hoyos, 2017), and so to construction generalisation, as constructions are considered to be a kind of relational category (Goldwater, 2017). This study also confirms the idea that high similarity prompts analogical comparison, especially when similarity is introduced from the beginning and progressively diminishes, thus favoring alignment and abstraction (Gentner & Hoyos, 2017;Gentner, Loewenstein & Hung, 2007;Loewenstein & Gentner, 2001).…”

The impact of the SMART program on cognitive and academic skills: A systematic review and meta‐analysis

“…The proposal that relational reasoning is a core feature of cognition is consistent with recent findings from within cognitive psychology (Goldwater et al, 2018), linguistics (Everaert et al, 2015; Goldwater, 2017) and education (Alexander, 2019; Goldwater & Schalk, 2016). Alexander (2019) outlined four categories of relational reasoning that appear key in approaching common tasks and tests in educational settings (analogy, anomaly, antimony and antithesis).…”

Analogy as a Catalyst for Cumulative Cultural Evolution

“…Analogy has been studied by cognitive scientists and linguists for decades, and a substantial corpus of work has demonstrated the importance of analogies for perception and categorisation (French, 1995;Hofstadter & Sander, 2013;Mitchell, 1993), relational reasoning and abstract representation (Doumas et al, 2008;Gentner & Goldin-Meadow, 2003a;Goldwater et al, 2018;Hummel & Holyoak, 1997;Levinson, 2003), learning and reasoning in children (Navarrete & Dartnell, 2017;Richland et al, 2006;Richland & Begolli, 2016), and language evolution (Gentner, 2016;Goldwater, 2017;Goldwater & Gentner, 2015) , as well as the role of language in relational reasoning (Gentner & Christie, 2008;Gentner & Goldin-Meadow, 2003b;Levinson, 2003;Loewenstein & Gentner, 2005;Lupyan, 2008) and in structuring our thinking (Gentner et al, 2001;Levinson, 2003;Lupyan & Bergen, 2016;Lupyan & Zettersten, 2020). Indeed, some argue that analogy is the 'core of our cognition' (Hofstadter, 2001;Hofstadter & Sander, 2013).…”