The Mental Lexicon - Volume 7, Issue 3, 2012

This study investigates whether lexical knowledge in the first language (L1) of late Turkish-Dutch bilinguals becomes less accessible for the production of fluent speech and in controlled experimental tasks as a result of extended stay in the Netherlands. It is also considered to what degree extra-linguistic factors can account for this phenomenon. Data are collected from the first generation Turkish migrants (n = 52) and from a monolingual reference group in Turkey (n = 52) via a lexical naming task, a free speech task and a sociolinguistic background questionnaire. The results show that the bilingual group is indistinguishable from the monolinguals on the experimental task. However, in the free speech task, they not only are significantly more disfluent than the monolinguals but also make significantly less use of diverse, in particular low-frequency, vocabulary. Overall, the results signal that bilinguals were outperformed by the monolinguals in spontaneous language production but not on a controlled task. We interpret this finding to indicate a decrease of automaticity in the access to linguistic knowledge which impedes the rapid integration of information from all linguistic levels. Further analyses with respect to the relations between the L1 change and nonlinguistic factors are discussed within the Activation Threshold Hypothesis (ATH).

Do speakers possess universal linguistic restrictions on the sound structure of their language? We examine this question by investigating the restrictions on onset clusters (e.g., bl in block). Cross-linguistic comparisons suggest that certain onset clusters are universally preferred: Onsets like bn are preferred to bd, which, in turn, are preferred to lb. In four experiments, we demonstrate that such preferences constrain onset identification by Spanish speakers: the worst formed the onset, the more likely its misidentification. Onset structure, however, determines not only the rate of disyllabic recoding but also its type. While better-formed onsets of rising sonority are repaired epenthetically (e.g., bnif→benif), worse-formed onsets are recoded prothetically (e.g., lbif→elbif), and the choice of repair (epenthesis vs. prothesis) is modulated by linguistic experience. These findings suggest that speakers possess broad linguistic restrictions that extend to structures unattested in their language, but the response to such putatively universal pressures is experience-dependent.

This study compares antonymous relations between lemmata (such as soft/softer/softest and hard/harder/hardest), words (soft and hard) and word senses (for example, the sense of soft indicating a yielding surface and that of hard describing an unyielding surface). In agglomerative cluster analyses of data from the British National Corpus, specific antonymous adjective senses are found to cluster more tightly and neatly than either antonymous words or lemmata. Moreover, when pairs such as soft and hard co-occur in the corpus, the co-occurring senses are typically instances of antonymous senses that cluster together. This evidence from clustering and co-occurrence suggests that antonymy operates primarily at the level of the word sense, rather than the word or the lemma.

Studies of vowel mutability have shown that it is easier to change a nonword (e.g., /tibl/) into a real word by changing a vowel (/tebl/) than by changing a consonant (/fibl/). All previous studies have used auditory materials, suggesting that the effect is a spoken language phenomenon. We conducted two studies with print materials, one in English and one in Spanish. Both showed clear vowel mutability effects, suggesting that vowel mutability is a more a general phenomenon. Vowel mutability is also shown to be one of many phenomena in which vowels and consonants show asymmetrical effects. Implications for models of auditory and visual word recognition are discussed.

This study investigates tense morphology in agrammatic aphasia and the predictions of two accounts on processing of regular and irregular verbs: the Dual Mechanism model, that is, for aphasic data, the Declarative/Procedural model, and the Single Mechanism approach. The production of regular, irregular and mixed verbs in the present, simple past and past participle (present perfect) was tested in German by means of a sentence completion task with a group of seven speakers with agrammatic aphasia. The results show a difference between regular verbs and irregular verbs. Mixed verbs were equally difficult as irregular verbs. A frequency effect was found for irregular verbs but not for regular and mixed verbs. A significant difference among the correctness scores for present tense and simple past forms was found. Simple past and past participle were significantly more difficult than present tense. Error types were characterized by pure infinitive responses and time reference errors. Neither of the above accounts is sufficient to explain these results. Correctness scores and error patterns for mixed verbs suggest that such minor lexical patterns can be useful in finding new evidence in the debate on morphological processing. The findings also highlight time reference as well as language specific characteristics need to be taken into consideration.