Yabancı Dil Olarak Türkçede Biçim-Sesbilgisel Değişimlerin Edinimi
Gamze HallıSözcüklerin dil edinicisinin zihinsel sözlüğünde nasıl saklandığı ve bunlara nasıl erişildiği uygulamalı dilbilimin en sık araştırılan konularından biridir. Alan yazınında sözcük işlemlemeye dair çeşitli modeller sunulmuştur. Fakat bu modeller, Türkçe gibi biçimler açısından zengin sondan eklemeli dillerdeki sözcük tanıma sürecini açıklamakta yetersiz kalmaktadır. Görsel sözcük tanıma sürecinde sözcük erişimini etkileyen birçok faktör (örn. seslem ve seslem başlangıç etkisi, dil edinim yaşı, fonolojik ve imla komşu sayısı vb.) olmasına rağmen uygulamalı dilbilim çalışmalarında sıklıkla iki ana faktöre atıfta bulunulmaktadır: a) biçimbilimsel faktörler ve b) sözcük sıklığı etkisi. Görsel sözcük tanıma sürecine ilişkin modellerin çoğunlukla tek dilli bireylerden elde edilen bulgularla sınırlı olduğu görülmektedir. Bu nedenle bu çalışmada, yabancı dil olarak Türkçe öğrenen bireylerin biçim-sesbilgisel değişimleri edinme yolları incelenmiş, çekimli sözcüklerin zihinsel sözlükte temsilinin araştırılması amaçlamıştır. Yarı-deneysel araştırma deseninin kullanıldığı bu çalışmada katılımcılara süreç-içi yöntemlerden sözcüksel karar testi uygulanmıştır. Yabancı dil olarak Türkçe öğrenen bireylerin a) dil yeterlik düzeylerinin, b) biçim-sesbilgisel değişim türünün (düzenli veya düzensiz) ve c) sözcük sıklığının (düşük veya yüksek) görsel sözcük tanıma sürecini ne yönde etkilediği araştırılmıştır.
Acquisition of Morpho-phonological Changes in Individuals Learning Turkish as a Foreign Language
Gamze HallıOne of the most frequently researched topics in the field of applied linguistics is how words are stored and accessed in a lexicon. Although various models of word processing can be found in the literature, they have failed to explain the word recognition process in agglutinative languages such as Turkish. Moreover, despite the fact that there are many factors that can affect the visual word recognition process (e.g., sound onset effect, age of language acquisition, number of phonological and orthographic neighbors), previous studies in applied linguistics have frequently referred to two main factors: a) morphological factors and b) the word frequency effect. Meanwhile, the majority of the models related to the visual word recognition process have been limited to the findings of monolingual individuals. Therefore, this study examines the acquisition of morpho-phonological changes among a sample of individuals learning Turkish as a foreign language, with specific focus on the representation of inflected words in the lexicon. For this purpose, a quasi-experimental research design was used and a lexical decision test was administered to the participants. In this regard, the effects of language proficiency levels, morphophonemic changes (regular or irregular), and word frequency (low or high) on their visual word recognition process were investigated. Based on the data obtained, it is concluded that (1) in the acquisition of morpho-phonological changes by adults with L2 Turkish, words are processed holistically at the beginner level - as a whole without separating roots and affixes – and as their proficiency level increases, the analysis of roots and affixes develops afterwards, in other words, dual mechanisms are effective in the lexical access process depending on the proficiency level. Also, (2) the words with a high number of orthographic neighbors have a positive effect only on the linguistic performance of the L2 A2 level group, and (3) the words with a high frequency of use have a positive effect only on the linguistic performance of the L2 B2 level group, while processing words with regular morpho-phonological changes.
In the study, a quasi-experimental method, one of the quantitative research designs, was used. Lexical decision task was applied in a computer lab by using the PyschoPy program (Pierce et al., 2019) to collect the data. Considering the effect of phonology on morphology, stimuli ending with /p/, /t/, /tʃ/ and /k/ sounds and changing by inflection were presented to adults learning Turkish as a foreign language. These target/critical stimuli used in the lexical decision task were prepared using the Sketch Engine (Kilgarriff et al., 2014) corpus platform out of 157.293 word tokens. The words in the corpus were divided into four groups: 1) words with high frequency and regular morpheme-phonological changes (DU-YU), 2) words with low frequency and regular morpheme-phonological changes (DU-DU), 3) words with high frequency and irregular morpheme-phonological changes (DZ-YU), and 4) words with low frequency and irregular morpho-phonological changes (DZ-DU). For the analysis of the reaction time data in the study, 3 (L1 Turkish, L2 A2 level, L2 B2 level) X 2 (regular morphophonological changes, irregular morpho-phonological changes) X 2 (high frequency and low frequency) mixed ANOVA was used for repeated measures, and Bonferroni correction was applied for post hoc multiple comparison.
The results of the study revealed that L2 A2 level participants processes the words with irregular morpho-phonological changes faster than the regular ones (p<.001). However, the control group L1 Turkish and experimental group L2 B2 level participants didn’t show any statistically significant difference. This finding was interpreted as A2 and B2 language learners of Turkish as a foreign language process regular and irregular inflected words in different ways. Among the intermediate level participants (B2 level), the reaction time averages of irregular words getting longer and becoming similar to regular inflections, and thus the difference between regular and irregular morpho-phonological change types decreases and becomes statistically meaningless. The awareness of morphological analysis begins to emerge as the language proficiency level increases. Another reason for the faster processing of words with irregular morpho-phonological changes in the L2 A2 group may be explained with orthographic neighbors effect. According to Van Heuven (2005), the increase in the number of orthographic neighbors of the words reduces latency in the visual word recognition process in foreign language acquisition, and the effect is more obvious in words with low frequency than in words with high frequency (see Andrews, 1989, 1992; Pollatsek, Perea, & Binder, 1999). In the study, the average of orthographic neighbors of the words with irregular morphophonemic changes was 5.5, and the average of the regular ones was 1.95. It was interpreted that the average difference of 4 words might have positively affected the performance of participants by activating more words in their mental lexicon. The difference between L2 A2 participants' processing of DU-DU and DZ-DU words also supports this interpretation (p<.001). The mean number of orthographic neighbors of DU-DU words is lower than that of DZ-DU words (M=1.8, M=6.2, respectively). Although L2 A2 participants might not have recognized a target word presented as DZ-DU, this word may have affected the processing process positively by activating other words that are syntactically similar to the target word, reducing the average reaction time of the target word.
In conclusion, the study revealed that depending on the language proficiency level, dual mechanisms may be active in the lexical access in visual word recognition process. At the beginning, the foreign language learners of Turkish might be using full-listing strategy, as their proficiency increases, they might show a tendency to use decomposition. Secondly, words with a high number of orthographic neighbors positively affect the performance of L2 A2 level language learners of Turkish. In other words, at the beginning, when a word is encountered, other words orthographically related to that word may be activated and facilitating the word access process. Finally, it has been observed that the hypothesis that words with a high frequency will positively affect the linguistic performance (reaction time) of the participants compared to words with a low frequency is limited only to the L2 B2 group when processing words with regular morpho-phonological changes. This result was interpreted as L2 B2 level language learners process words with high frequency as a single input in their mental lexicon, while final devoicing rule is applied to those with low frequency, consequently, they process words with high frequency faster.