Multiplex model of mental lexicon reveals explosive learning in humans

Abstract

Word similarities affect language acquisition and use in a multi-relational way barely accounted for in the literature. We propose a multiplex network representation of this mental lexicon of word similarities as a natural framework for investigating large-scale cognitive patterns. Our representation accounts for semantic, taxonomic, and phonological interactions and it identifies a cluster of words which are used with greater frequency, are identified, memorised, and learned more easily, and have more meanings than expected at random. This cluster emerges around age 7 through an explosive transition not reproduced by null models. We relate this explosive emergence to polysemy – redundancy in word meanings. Results indicate that the word cluster acts as a core for the lexicon, increasing both lexical navigability and robustness to linguistic degradation. Our findings provide quantitative confirmation of existing conjectures about core structure in the mental lexicon and the importance of integrating multi-relational word-word interactions in psycholinguistic frameworks.

Network Growth Modeling to Capture Individual Lexical Learning

“…While phonological network structure varies greatly from semantic network structure and network model results have shown differences in acquisition in relation to phonological networks [20,33,34], both of these individual network representations are useful in predictive models of lexical acquisition. In the future, we hope to jointly consider the effects of these levels of analysis in predictive modeling and as a means to understand the process of language development, possibly by building ensemble models within this framework or by extending this approach to multiplex representations [35,36].…”

Applying network theory to fables: complexity in Slovene belles-lettres for different age groups

“…Those findings were later confirmed and expanded for different languages and network construction techniques [27][28][29][30][31]. Besides syntactic and co-occurrence based networks, also conceptual and semantic networks of words were constructed with the aim to provide a more cognitive-oriented point of view [25,[32][33][34]. Notably, those networks were also found to express the previously mentioned complex topological features, such as power-law behaviour and a very small average shortest path lengths [33,35,36], although they were found to differ from syntactic networks in their hierarchical properties [37].…”

Forma mentis networks quantify crucial differences in STEM perception between students and experts

“…a dictionary, but it rather is a dynamical system optimized for cognitive computing which stores and processes individual concepts together with their associated linguistic data, e.g. semantic overlap in meaning (Steyvers & Tenenbaum, 2005), phonological similarities (Vitevitch, 2008;Vitevitch, Siew, & Castro, 2018), syntactic relationships between word categories (Stella, Beckage, Brede, & De Domenico, 2018). Psycholinguistic evidence has shown that the associative structure of the mental lexicon influences language processes such as word learning (Hills & Siew, 2018;Stella, 2019;Stella, Beckage, & Brede, 2017) and processing (De Deyne, Navarro, & Storms, 2013;De Deyne, Navarro, Perfors, Brysbaert, & Storms, 2018;Kenett, Levi, Anaki, & Faust, 2017;Steyvers & Tenenbaum, 2005).…”

“…Hence, free associations are largely free from any specific semantic definition (e.g., synonyms). Previous work (Stella et al, 2017;Stella, Beckage, Brede, & De Domenico, 2018) has shown that free associations partially overlap with other semantic word-word similarities such as synonyms (i.e., two words sharing the same meaning in a given context) or generalisations (i.e., a concept being a special type of another word) but also display a small overlap with phonological similarities among words (e.g., when pronunciations differ in one phoneme).…”