Exploring Substantive Bias Through Incidental Phonological Learning: Accuracy and Conscious Status of the Resultant Knowledge

In phonological learning, the substantive bias hypothesis holds that phonetically natural patterns (e.g., vowel harmony) are more learnable than unnatural ones (e.g., vowel disharmony). However, many empirical studies have failed to find evidence for this bias. One possible reason for such null results is the involvement of intentional learning. In addition, previous studies largely relied on accuracy rates as a measure of this bias and mostly assumed unconscious resultant knowledge without adequate consciousness measures. Given these two concerns, our study made two improvements to further investigate the conditions of emergence and forms of manifestation of substantive bias. First, a disguise task was included to minimize intentional learning. Second, in addition to accuracy rates, we comprehensively measured the conscious status of resultant knowledge as another potential form of manifestation of substantive bias. Results showed overall low learning rates of both vowel harmony and disharmony. Interestingly, contrary to our predictions, participants exposed to vowel disharmonic items had a higher likelihood of possessing conscious judgment knowledge than those exposed to vowel harmonic items, although the two groups were comparable in overall accuracy rates and likelihood of possessing conscious structural knowledge. We attribute this finding to the possibility that vowel disharmonic items may be perceptually more distinctive at least under a sufficiently incidental learning context. This finding may thus support substantive bias through measures of knowledge consciousness during incidental phonological learning.