MEi:CogSci Conferences, MEi:CogSci Conference 2011, Ljubljana

Font Size: 
Implicit learning of language rules: Influence of the constraints on vowel occurrences
Tena Gnjatovic, Maja Peretic

Last modified: 2011-06-08

Abstract


Vowel harmony is cross-linguistically a common phenomenon cross-linguistically, which constrains the appearance of vowels in a word depending on their quality. Since the possible occurrence of vowels is limited to a certain set, vowel harmony produces a certain level of redundancy. The aim of this research is to see whether vowel harmony facilitates the implicit learning of structured sequences containing both vowels and consonants. The experiment will be conducted on two groups of participants: native speakers of Croatian and native speakers of Hungarian. Their task will be to evaluate the grammaticality of structured phoneme sequences in the test phase, based on their similarity to those in the training phase.
During the training phase, the participants are presented with 70 stimuli, which are pseudowords following the CVCiVCVCi pattern, i.e. the second consonant in a pseudoword is always the same as the final consonant. Additionally, half of the pseudowords observe the front-back vowel harmony rule, so that all the vowels are either front or back. In the other half of the stimuli both front and back vowels are combined freely, i.e. there are no constraints on vowel occurrence. The participants are instructed to read the pseudowords carefully, and to left-click on the mouse if they find the letter ‘l’ in the stimulus, right-click if they do not. This task is used as distraction so as to prevent the participants from recognizing and explicitly formulating the CVCiVCVCi rule.
The test phase consists of 120 pseudowords, divided into four different categories, based on the possible combinations of two features: front-back vowel harmony and the CVCiVCVCi rule. Again, the mouse is used as a response device, and the participants’ task is to left-click if the pseudoword seems familiar with respect to the stimuli in the training phase, and to right-click in case it does not. Since the only rule consistently applied throughout the training phase was the CVCiVCVCi rule, the correct response would be to left-click after every stimulus of the same structure in the test phase, regardless of the vowel harmony. Response accuracy and reaction time is measured, in order to see whether the participants’ responses will be in any way influenced by the constraints on vowel occurrence.
The hypothesis is that vowel harmony, due to its increased regularity effect, should facilitate the recognition of the phoneme sequence structure presented in the training phase. Therefore, differences are expected primarily in the reaction time to different stimuli categories, but also between the two groups of participants, since native speakers of Hungarian are exposed to vowel harmony from the very beginning of language learning.

References:
LaCross, A. (2010) Regularity and vowel harmony: Statistical learning and rule abstraction. LSA Annual Meeting Extended Abstracts [Online].
Finley, S. and Badecker, W. (2008) Analytic Biases for Vowel Harmony Languages. Proceedings of the 27th West Coast Conference on Formal Linguistics, ed. Natasha Abner and Jason Bishop, 168-176. Somerville, MA: Cascadilla Proceedings Project.