Lexical development and masked orthographic priming in the second language

Thumbnail Image


Publication or External Link





The fuzzy lexical representations (FLR) hypothesis proposes that form encoding of words in a second language (L2) is often fuzzy, and this concerns both phonological and orthographic representations. FLR occur because of difficulties in encoding of L2 word forms as well as insufficient L2 experience. The FLR hypothesis also suggests that fuzzy L2 orthographic representations are the reason for the weak lexical competition for orthographic neighbor prime-target pairs in the L2 that has been observed in previous research (e.g., Jiang, 2021). However, this hypothesis also assumes that as orthographic representations become robust along with learners’ L2 experience, L2 words are eventually able to take part in lexical competition just like first language (L1) words. The current study tests these hypotheses using the individual-differences measures of the quality (orthographic precision) and the quantity (vocabulary size) of orthographic representations. At the same time, this study explores the relationship between sound perception (word and phoneme identification) of nonnative contrasts (e.g., the /l/-/ɹ/ contrast for Korean L2 learners of English), phonolexical encoding, and form facilitation for minimal pairs with these contrasts.

A masked priming LDT was employed, in which minimal pairs with a nonnative phonological contrast (e.g., read-LEAD) and minimal pairs without a confusing phonological contrast (e.g., dear-TEAR) were used as the prime and target. Before the experiment, it was predicted that low-proficiency L2 speakers would show significant form facilitation under all prime conditions. On the other hand, medium-proficiency L2 speakers were expected to show evidence of emerging lexical competition (a null priming effect) for prime-target pairs without a difficult phonological contrast (e.g., dear-TEAR), although they would still show form facilitation for minimal pairs with a nonnative phonological contrast (e.g., read-LEAD). The facilitation for the latter pairs was predicted to occur because of less successful orthographic encoding of these pairs caused by fuzzy phonological representations of L2 words with difficult phonological contrasts. It was further expected that high-proficiency L2 speakers would show a nativelike pattern of form priming across all the prime conditions.

Thirty L1 speakers and 90 L2 learners of English with a wide range of L2 proficiency were recruited for the experiment. In auditory word and phoneme identification tasks, L2 speakers showed less accurate identification of the /l/-/ɹ/ contrast compared to L1 speakers indicating that they indeed had problems in accurate sound perception and/or phonological categorization of the nonnative contrast as had been predicted. 

In the masked priming LDT, L1 speakers showed a null priming effect across the prime conditions. L2 speakers showed significant form priming for words with the /l/-/ɹ/ contrast but not for other words without a difficult contrast. When form priming in each L2 participant group was examined separately, low- and medium-proficiency L2 speakers showed significant facilitation for pairs with the /l/-/ɹ/ contrast, but high-proficiency L2 speakers showed a null priming effect for these pairs as L1 speakers did. This finding supports the prediction of the current study. At the same time, the influence of global proficiency, as measured by a cloze test, on the orthographic form priming was statistically non-significant. Furthermore, form facilitation for prime-target pairs without a confusing contrast (e.g., dear-TEAR) was not significant even in low-proficiency L2 participant groups.

Through a series of investigations on the relationships between the form priming found in L2 speakers and their performance on individual-differences measures (spelling, vocabulary, word identification and phoneme identification tasks), the present study discovered that form facilitation was significantly modulated by L2 speakers’ orthographic precision (spelling scores). Moreover, it was found that the influence of orthographic precision on the form facilitation was more prominent for words that were more difficult for accurate phonological encoding, and as a consequence, orthographic encoding (i.e., minimal pairs with the /l/-/ɹ/ contrast) than others without a confusing contrast. These findings support the FLR hypothesis which argues for the role of the quality of orthographic representations in lexical competition between orthographic neighbors. The role of vocabulary size (vocabulary scores) was also found for four-letter stimuli indicating that the development of the size of the mental lexicon also affects lexical competition. On the other hand, no modulating role was observed of accurate word or phoneme identification of nonnative contrasts in form priming for minimal pairs with these contrasts.

Based on these findings, this study suggests that (1) the orthographic form facilitation discovered at initial stages of L2 lexical development is due to fuzzy L2 orthographic representations. In addition, it claims that (2) as L2 speakers establish a larger and more precise L2 lexicon, L2 words can take part in lexical competition just as L1 words do. It also proposes that (3) the establishment of precise orthographic (or phonological) representations of L2 words with a confusing phonological contrast is more challenging than those without a difficult contrast. (4) Finally, although the observed weak effect of sound perception on form priming seems to indicate no systematic relationship between the development of phonological categorization ability and the form facilitation for these words, the present study contends that it may be premature to draw a conclusion about the role of phonolexical representations involving a nonnative contrast in orthographic representations. Indeed, the results may be due to methodological limitations of the word and phoneme identification tasks as a measure of the quality of phonological representations.