Doctoral Dissertation Research: Role of Prior Knowledge in Consolidation of Novel Phonotactic Patterns for Speech Production

Project: Research project

Project Details

Description

As any adult language learner can attest, no language is learned in a single day. Benefits of consolidation—an improvement in what you can recall across a period of sleep or resting while awake—have been well-documented for language learning, with much work examining its help- fulness for word learning (see Dumay and Gaskell, 2007 for review), but relatively less focus on linguistic pattern learning. A consolidation period is not only helpful but indeed required for adults to learn how to produce certain sequential generalizations on the shapes of possible syllables and words in a new language (Gaskell et al., 2014), called phonotactic patterns (here, phonotactics). The complex phonotactics requiring consolidation before they can be learned are those relating multiple sounds within a syllable—e.g., “if the vowel is [ɪ], then [s] can only appear at the end of the syllable.” Such patterns are common across the world’s languages. The present work will clarify the role of prior knowledge in consolidating these second-order phonotactics.

According to McClelland et al. (1995), consolidation integrates new knowledge with old to ensure old knowledge is not overwritten by new. Lewis and Durrant (2011) propose that new and old memories are replayed simultaneously during consolidation, with information shared across new and old memories more strongly activated than that from only a single source. Applying these theories to phonotactic learning, we hypothesize that prior knowledge can be more readily leveraged when new phonotactic patterns are similar to previously-learned ones.

To test this, we will perform three multi-day production experiments manipulating the prior knowledge learners have about new sequential patterns. In Exp. 1, the source of this prior knowledge is our participants’ native language. We expect structurally similar phonotactics to benefit more from consolidation. In Exp. 2, we test the effect of pre-trained prior phonotactic knowledge, predicting that participants learning phonotactics similar to pre-trained ones should consolidate these new phonotactics more effectively. Exp. 3 replicates Exp. 1 in a sequential action domain where participants do not have structured prior knowledge (button-pressing), to check that Exp. 1’s anticipated results are not due to differences in ease of planning between conditions.
StatusActive
Effective start/end date9/1/212/28/23

Funding

  • National Science Foundation (BCS-2116802)

Fingerprint

Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.