Production vs Perception: The Role of Individuality in Usage-Based Grammar Induction

Abstract

This paper asks whether a distinction between production-based and perception-based grammar induction influences either (i) the growth curve of grammars and lexicons or (ii) the similarity between representations learned from independent sub-sets of a corpus. A productionbased model is trained on the usage of a single individual, thus simulating the grammatical knowledge of a single speaker. A perception-based model is trained on an aggregation of many individuals, thus simulating grammatical generalizations learned from exposure to many different speakers. To ensure robustness, the experiments are replicated across two registers of written English, with four additional registers reserved as a control. A set of three computational experiments shows that production-based grammars are significantly different from perception-based grammars across all conditions, with a steeper growth curve that can be explained by substantial inter-individual grammatical differences.

Syntactic variation across the grammar: modelling a complex adaptive system

“…CxG is a usage-based approach to syntax which, in practical terms, means that more item-specific constructions are learned first and then generalized into more schematic constructions (Nevens et al, 2022;Doumen et al, 2023). The grammar learning algorithm used in this paper is taken from previous work (Dunn, 2017;Dunn, 2018b;Dunn, 2019a;Dunn and Nini, 2021;Dunn and Tayyar Madabushi, 2021;Dunn, 2022), with the grammar trained from the same register as the dialectal data (tweets). Rather than describe the computational details of this line of work, this section instead analyzes constructions within the grammar as examples of the kinds of features used to model syntactic variation.…”

Exploring the Constructicon: Linguistic Analysis of a Computational CxG

“…An external metric evaluates and compares constructicons using their performance when applied to a specific prediction task. Recent work has focused on the use of computational CxG for modelling individual differences (Dunn and Nini, 2021), register variation (Dunn and Tayyar Madabushi, 2021), and population-based dialectal differences (Dunn, 2018a(Dunn, , 2019cDunn and Wong, 2022). Because CxG is a usage-based paradigm, the definition of a construction that is referenced above depends on both entrenchment and idiomaticity.…”