Reading compound words by adult Korean-English bilinguals

Abstract

The present study investigated how Korean-English bilinguals read compound words in Korean L1 and English L2. Two masked priming lexical decision experiments were conducted, in which written compound words (e.g., ‘bedroom’) and monomorphemic words with a compound-like structure (e.g., ‘hammock’) served as visual targets and were preceded by constituent visual primes (e.g., ‘room’, ‘mock’). In Experiment 1, within-language prime-target pairs (Korean constituent-Korean compound word), and in Experiment 2, cross-language prime-target pairs (Korean L1 constituent-English L2 compound word) were presented. In Experiment 2, the time course of cross-language activation was examined by manipulating the prime duration (36 ms, 48 ms vs. 100 ms). Experiment 1 showed that Korean compound words are processed based on the morpheme unit rather than the salient visual syllable form. In Experiment 2, when reading English L2 compound words, L1 morphological and phonological information are both activated in the early stage (36 ms prime duration), regardless of semantic relatedness. In the later stage (48 and 100 ms prime durations), L1 morphological activation is constrained by semantic relatedness. Shared phonological (form) information without morphological relatedness between L1 and L2 did not facilitate L2 complex word recognition at all stages. Taken together, these findings suggest that bilingual readers are more sensitive to morphological information than form information in both L1 and L2. There is a quick cross-language activation of L1 morphemic information in reading L2 complex words.

Keywords:

• Compound words
• Korean-English bilinguals
• Cross-language activation

This work was supported by Summer Research Fellowship and Ann Wylie Dissertation Fellowship from the University of Maryland to the first author.

This work was supported by Summer Research Fellowship and Ann Wylie Dissertation Fellowship from the University of Maryland to the first author.

Notes

¹ The Boston Naming Test contains 60 pictures, arranged in increment levels of difficulty. Participants were asked to say the name of the object in each picture. The English C-test included five passages, and each passage was chosen from Encyclopedia Britannica, Practice and progress, readers choice and Developing reading skills in which 25 words were incomplete by deleting two-thirds or half of the words or leaving only the first letter in each passage. Participants were asked to restore the missing letters.

² In Roberts, Garcia, Desrochers, and Hernandez (2002), the average scores for the English Boston Naming Test in fluent Spanish-English bilingual adults (0.71) and fluent French-English bilingual adults (0.66) were significantly below the mean score of the English-speaking adults (0.85). In Babaii and Ansary (2001), the average score for the English C-test was 0.52 in Persian-English bilingual adults who had the upper-intermediate level of English proficiency (119 junior students majoring in English translation and literature at Khayam University (Mashhad)). Based on the proficiency test results from the previous studies, we assume that the level of English proficiency of participants in the present study is at an advanced level.