How does semantic richness impact the processing of word meaning across different languages?

Background

Semantic richness refers to how certain words can be more easily identified and responded to in visual word recognition tasks, due to them having richer semantic representations (Pexman et al., 2008). The "richness" of a word can be measured for example by the number of contexts the word appears in, the number of differing neighbouring words or concepts that may occur, or the number of features represented by the word. Different languages vary in how they represent meaning, even for words that may appear to have equivalent translations. This has contributed to research showing that native speakers of languages that differ in semantic density and richness, may accordingly react to and assess words differently in word association and judgement tasks (Aceves & Evans, 2024).

Research Questions / Hypotheses

The key questions this research aimed to examine were:

1. Does processing the meaning of an individual word differ between languages due to language-specific semantic richness measures?
2. Does semantic processing of the relations between words vary between languages due to language-specific semantic richness measures?
3. Do these relationships hold across a variety of different semantic richness measures?

Participants

Inclusion criteria for the study included being a native speaker in one of the specified languages of interest. 214 participants completed the study through REP. After data cleaning processes, data from 187 REP participants was analysed. The exclusion criteria during data cleaning comprised successful completion of both tasks and non-random responses.

Methods

The study utilised reliably cross-translated words across multiple languages as stimuli, with semantic richness measures for each word sourced from the Small World of Words Project (De Deyne et al., 2019). Participants completed two tasks in their native language: a semantic categorisation task evaluating individual word processing, and a semantic relatedness task evaluating the processing of relationships between words. For both tasks, response time and choice were recorded.

Results

Various mixed effects models were created to compare participant data, to gauge the presence of cross-lingual differences in response time and choice distribution. Additional variables such as word concreteness and frequency were also modelled. Results found significant cross-lingual differences in reaction time for both tasks, as well as response distribution in the semantic relatedness task, with the latter result significantly influenced by semantic richness. However, SR was not found to significantly affect reaction times in either task. Significant effects were also observed for non-linear word concreteness, as well as language-interaction terms.

Implications

These findings suggest nuanced and intriguing language-specific processes exist that affect semantic decision-making. Such processes appear to modulate the extent to which semantic richness influences response time and choice during decision-making, suggesting that semantic richness alone may not fully account for observed cross-lingual differences.