Date of Award

May 2023

Degree Type


Degree Name

Doctor of Philosophy



First Advisor

Jae Yung Song

Committee Members

Anne Pycha, Fred Eckman, Hamid Ouali


This dissertation investigates the effects of L1 transfer, cluster-markedness, and input frequency (i.e., word-frequency and cluster-frequency) on Hijazi Arabic speakers’ acquisition of English consonant clusters that are unattested in their native language. Three experiments were conducted: (1) a production task that elicited Hijazi speakers’ production of English words with consonant clusters, which explored the repair strategies they used in their attempt to produce the target non-native linguistic structures, (2) a lexical decision task that examined the degree of participants’ acceptance of unmodified (e.g., ‘group’ [gɹup]) and modified (e.g., [gʊɹup]) English tokens as real English words, and (3) an AX discrimination task that assessed participants’ ability to perceive the epenthetic vowels and to differentiate between two spoken utterances. Non-native English speakers with a low-proficiency level (N= 30) participated in all three experiments while native English speakers (N= 30) took part in the two perception tasks as a control group. The stimuli were English monosyllabic CCVC and CVCC high- and low-frequency words, with rising sonority and falling sonority clusters at one edge of the words. They were manipulated by vowel insertion that either obeys (e.g., [gʊɹup]) or disobeys (e.g., [gɪɹup]) Hijazi phonological rules.The results show that Hijazi speakers tended to use vowels reflecting their L1 phonology when repairing English clusters, and had difficulty rejecting modified words with vowels obeying their L1 rules as real English words. Less marked clusters were produced and perceived more accurately and quickly than more marked structures which were in line with the prediction suggested by cluster-markedness, while low-frequency clusters were easier to process and produce than high-frequency clusters, contrary to the proposed patterns by cluster-frequency of type and token. Moreover, clusters in high-frequency words showed better performance than those in low-frequency words. It was concluded that L1 phonotactic knowledge, cluster-markedness, and input frequency had robust effects on Hijazi speakers’ production and perception of English consonant clusters that are absent in their native language. Hijazi participants’ native language hindered their ability to perceive or produce English words in a target-like fashion. Furthermore, in both modalities, the developmental path of non-native consonant clusters mirrored the pattern that was predicted by cluster-markedness and not the one proposed by cluster-frequency. Lastly, the process of acquiring non-native clusters was on a word-by-word basis. The findings illuminate the obstacles that L2 learners encounter when acquiring non-native linguistic structures due to the complex interplay of multiple factors.

Available for download on Friday, May 30, 2025

Included in

Linguistics Commons