Longitudinal Analysis of Early Semantic Networks

Abstract

Analyses of adult semantic networks suggest a learning mechanism involving preferential attachment: A word is more likely to enter the lexicon the more connected the known words to which it is related. We introduce and test two alternative growth principles: preferential acquisition—words enter the lexicon not because they are related to well-connected words, but because they connect well to other words in the learning environment—and the lure of the associates—new words are favored in proportion to their connections with known words. We tested these alternative principles using longitudinal analyses of developing networks of 130 nouns children learn prior to the age of 30 months. We tested both networks with links between words represented by features and networks with links represented by associations. The feature networks did not predict age of acquisition using any growth model. The associative networks grew by preferential acquisition, with the best model incorporating word frequency, number of phonological neighbors, and connectedness of the new word to words in the learning environment, as operationalized by connectedness to words typically acquired by the age of 30 months.

Quantifying the structure of free association networks across the life span.

“…In particular, they offer an inroad for understanding the relative stability of word meanings over time. The prominent use of the University of South Florida free association norms (Nelson, McEvoy, & Schreiber, 2004) in cognitive modeling work has been extremely productive (e.g., Griffiths, Steyvers, & Firl, 2007;Hills et al, 2009b). Our results suggest that ageappropriate norms may further enhance this productivity.…”

“…Inferences from aggregated association norms are generally the rule in cognitive psychology and they have been highly successful at predicting behavior (e.g., Griffiths et al, 2007;Hills et al, 2009b). Nonetheless, corroborating inferences about individual change are naturally limited by the difficulty of acquiring longitudinal data from individuals.…”

“…One of the advantages of network analysis is that it allows researchers to study phenomena at scales that range from individual words (i.e., nodes) to the entire lexicon (i.e., the network). Network analyses have been widely used to study the large scale structures of associative networks (De Deyne, Navarro, & Storms, 2013;De Deyne & Storms, 2008a;Steyvers & Tenenbaum, 2005), the formation of categories in toddlers (Hills, Maouene, Maouene, Sheya, & Smith, 2009a), and the trajectories of network development during early life (Bilson, Yoshida, Tran, Woods, & Hills, 2015;Hills et al, 2009b).…”

First learned words are not forgotten: Age-of-acquisition effects in the tip-of-the-tongue experience

“…This raises the question of how vocabulary is acquired during childhood. Hills and collaborators have characterized vocabulary acquisition by exploiting the structure of growing networks (Hills, J. Maouene, Riordan, & Smith, 2010;Hills, M. Maouene, J. Maouene, Sheya, & Smith, 2009). The authors argued that vocabulary acquisition is mainly determined by the principle of preferential acquisition, which states that the order of acquisition of words is driven by the connectivity of the words in the learning environment.…”

Real-world visual statistics and infants' first-learned object names

“…When these co-occurrences, in space (same scene) and/or time (adjacent scenes), are represented as networks, they typically exhibit a scale-free or small-world structure, showing clusters of inter-related items as well as sparser links among the clusters [82][83][84]. The structure of these small-world networks, in turn, are relevant to how new items are added to the network with the addition of new items predicted by their connectivity to already required items (see [85]). Past research shows that by the time children know 50 words, the network structure of their vocabularies shows the characteristic properties of smallworld networks [86].…”