NotesFAQContact Us
Search Tips
ERIC Number: ED547769
Record Type: Non-Journal
Publication Date: 2011
Pages: 112
Abstractor: As Provided
Reference Count: N/A
ISBN: 978-1-2675-1912-2
Infant Phonotactic Learning
Thatte, Victoria Anne
ProQuest LLC, Ph.D. Dissertation, University of California, Los Angeles.
For the past several decades, researchers have been investigating the stages infants go through on their way to acquiring their native language. Research into the question of the order in which, and time when, various facets of phonology are acquired has resulted in a basic timeline of development. Exploration of a second question, namely what learning mechanism infants rely on most heavily in acquiring the phonology and phonotactics of their native language, has led to the emergence of two competing approaches: infants could be using statistical induction to deduce phonotactics from the ambient language, or they could be tapping in to biases motivated by pre-phonological, domain-specific knowledge of phonetic principles. The research presented in this dissertation investigated both these questions. In the first half, an analysis of spoken corpus data confirmed that voiceless fricatives in fact appear more frequently word-initially in English than voiced fricatives, both in normal speech and in Infant-Directed Speech. Next, a series of infant experiments tested whether infants of 4.5, 6, and 8 to 10 months of age with monolingual American English language input display operative knowledge of the prevalence of voiceless fricatives in word-initial position, as evidenced by an attentional preference to the former over the latter. While no significant difference in attention time was found for the younger age groups, the 8- to 10-month-olds displayed a significant preference for the voiceless fricatives. This was interpreted as preliminary evidence in favor of the existence of either a statistical induction learning mechanism, or an a priori bias founded on inductive grounding (Hayes, 1999), as either learning mechanism would be predicated on knowledge gained through observation of properties of the input. Therefore, it would be reasonable for either one's effects to be more apparent slightly later in development, when the infant has had sufficient input and/or time to acquire this knowledge. Furthermore, the pooled data from all three age groups uncovered a significant preference for voiceless fricatives, indicating that infants may also have access to an "a priori" bias that is too weak to motivate a preference on its own, and therefore further impetus from statistical learning must be present for the infants to have a significant preference. The second half of the dissertation turned to exploring the question of which learning mechanism is dominant at the time that infants first show a strong preference for voiceless initial fricatives. This was accomplished first by exposing 8- to 10-month-old infants to only dental word-initial fricatives, of which [ð] has a higher word-initial token frequency---so that it should be preferred if the infants are using statistical induction while [?] is more in keeping with the principle of ease of articulation (Ohala, 1997)---so that it should be preferred if the infants are being guided by an unlearned bias. No significant preference was found in attention to either condition, a result which appeared to indicate that either (a) both an "a priori" bias and statistical learning are operative, but they exert equal influence, or (b) the statistical learning mechanism is stronger, but facts about the types of words making up the vast number of [ð]-initial tokens are weakening its effect in this case. The final experiment tested whether the knowledge demonstrated in the first series of experiments is generalized to the featural level, by investigating whether 8- to 10-month olds' preference for voiceless fricatives is generalized to Polish fricatives. No significant preference was found in the final experiment alone, nor when the data were pooled with those from the experiment testing infants' preference for either class of English fricatives, confirming that infants' statistically-gained knowledge is not generalized across segments. [The dissertation citations contained here are published with the permission of ProQuest LLC. Further reproduction is prohibited without permission. Copies of dissertations may be obtained by Telephone (800) 1-800-521-0600. Web page:]
ProQuest LLC. 789 East Eisenhower Parkway, P.O. Box 1346, Ann Arbor, MI 48106. Tel: 800-521-0600; Web site:
Publication Type: Dissertations/Theses - Doctoral Dissertations
Education Level: N/A
Audience: N/A
Language: English
Sponsor: N/A
Authoring Institution: N/A