The Mental Lexicon - Current Issue

This study aims to model lexical attrition in L2 mental lexicon for Chinese EFL learners using a network science technique. To this end, we constructed a large lexical network with 5746 English words with free association data collected from Chinese EFL college students. The attrition process was modeled by removing connections progressively from the network, leaving words less connected or isolated. Further, we controlled the order in which connections underwent attrition to explore whether order of attrition affected the attrition process. The results showed that: (1) The modeled L2 lexicon was a complex network, whose structure was different from random networks in terms of connectedness, average path lengths, clustering coefficient, power-law degree distribution; (2) Attrition events weakened the structure of L2 mental lexicon by increasingly reducing the number of interconnected words, increasing path lengths between words, and diminishing overall network efficiency, but the process was not linear; (3) The order in which attrition events occur affects the attrition process; specifically, attrition starting from peripheral connections dismantled the lexical network more slowly when compared with random and central-to-peripheral orderings. These findings add to our understanding of the cognitive organization of words in the mind, and provide fresh insight into lexical attrition.

Speakers of multiple languages must store the respective lexical items efficiently to enable correct access. Importantly, all items must be linked to semantic information and world knowledge. One prominent model of the mental lexicon of late bilinguals is the Revised Hierarchical Model (Kroll & Stewart, 1994), which postulates bidirectional but asymmetrical connections between separate stores for L1 (native language) and L2 (second/foreign language), and a shared conceptual store. Using German native speakers with advanced English proficiency, Experiment 1 largely confirmed model predictions regarding different preferred mental routes and processing times depending on translation direction. Moreover, the original design was extended by including abstract stimuli and picture naming in L2. A series of additional measures, such as proficiency and age of acquisition, served to specify the language experience of the participants and made it possible to compare the results with a group of non-native speakers of German (Experiment 2). Interestingly, the results suggest that the model also applies to two or more non-native languages, potentially influenced by the experimental and environmental language context.

Polysemy has attracted much interdisciplinary interest in recent times. Recent discussions in psycholinguistics focus on the different processing profiles of polysemous and homonymous words, and on how to explain such different profiles. Much current research assumes that while homonymous meanings are stored in different lexical entries in the mental lexicon, polysemous senses relate to just one lexical representation, be this a list of senses or a core meaning formed by features common to all the different senses. However, there is growing skepticism towards such a one-representation hypothesis. After differentiating regular and irregular polysemies along several dimensions (not only in terms of sense representation, but also in terms of sources, acquisition and word class distribution), this paper argues that the variants of the one representation model can meet some of the challenges that have been raised. However, there are further challenges that have not yet been considered. On the one hand, nested polysemies (senses generated on the basis of iterations of metonymies or metaphors) put some pressure on the idea that senses of irregular polysemies share some set of features. On the other hand, sharing some features that could constitute a core meaning may not be sufficient for entering in co-activation patterns. In sum, while the paper defends the one-representation hypothesis in the light of recent skepticism, it also calls for further research and an eventual reformulation of the hypothesis.

Measures of orthographic typicality have long been studied as predictors of lexical access. The best-known orthographic typicality measure is orthographic neighbourhood size (Coltheart’s N or ON), the number of words that are one letter different, by substitution, from the target word. A more recent related measure of orthographic typicality is orthographic Levenshtein distance 20 (OLD20), the average Levenshtein orthographic edit distance of a target word from its 20 closest neighbours (Yarkoni, Balota, and Yap, 2008). Both measures have been implicated in lexical access. In this paper, we propose and assess a family of measures of word form similarity we call orthographic uncertainty. These measures are based on Shannon entropy (Shannon, 1948), which has a long history of being considered psychologically relevant. Orthographic uncertainty measures are superior to ON and OLD20 at predicting lexical decision and naming reaction times and accuracies. They are also superior to the older measures insofar as they are naturally tied to the widely-accepted quantification using Shannon Entropy of the psychological functions of familiarity, uncertainty, learnability, and representational and computational efficiency.