As this semester’s first special guest of the SALT project, Gaja Jarosz from UMass Amherst is visiting us next week.
On Wednesday she will talk about the Nature versus Nurture debate and Sonority Sequencing in Polish and on Friday she will give the semester’s first CASTL Colloquium talk on how hidden generalizations are learned.
Don’t miss out on this!
Wednesday, 14.09.2022, 10:15-12:00, Breiviklia: M-025
Gaja Jarosz: Sonority Sequencing in Polish: An Ongoing Nature v Nurture Debate
Friday, 16.09.2022, 12:15-14:00, SV- & HUM-building E 0104
Gaja Jarosz: Learning Hidden Generalizations
Below you find an abstract for each talk.
See you all next week!
Martin
Sonority Sequencing in Polish: An Ongoing Nature v Nurture Debate
Recent work on phonological learning has questioned the traditional view that innate principles guide and constrain language development in children and explain universal properties cross-linguistically. In this talk I focus on a particular universal, the Sonority Sequencing Principle (SSP), which governs preferences among sequences of consonants syllable-initially. Experimental evidence indicates that English, Mandarin, and Korean speakers exhibit sensitivity to the SSP even for consonant sequences that never occur syllable-initially in those languages (such as [nb] vs. [bn] in English). There is disagreement regarding the implications of this finding. Berent et al. (2007) argue that these results can only be explained with reference to an innate principle; however, Daland et. al (2011) show that computational models capable of inferring statistical generalizations over sound classes can detect evidence for these preferences based on related patterns in the language input (and therefore no reference to innate principles is required). Building on these studies, I argue that English is the wrong test case: it does not differentiate predictions of these two hypotheses. I examine learning of syllable structure phonotactics in Polish, a language with very different sonority sequencing patterns from English. Polish provides a crucial test case because the lexical statistics contradict the SSP, at least in part. I review developmental evidence indicating that children acquiring Polish are nonetheless sensitive to the SSP, producing larger sonority rises more accurately in spontaneous production (Jarosz 2017). I then present results from two experiments investigating adult Polish native speakers’ phonotactic knowledge (Jarosz & Rysling 2017, in prep). The findings indicate that Polish native speakers’ phonotactic preferences are sensitive to the SSP and that this SSP sensitivity is not predicted by the computational models that succeeded for languages like English, Mandarin, and Korean. These results suggest a crucial role of an inherent bias or a constraint on generalization from the input. However, in a recent twist, Nelson (2022) developed a new family of phonotactic learning models and showed that some of these models do succeed in predicting Polish speakers’ generalizations about sonority sequencing. I discuss implications of these modeling results for future work.
Learning Hidden Generalizations
Language acquisition proceeds on the basis of incomplete, ambiguous linguistic input, and one source of this ambiguity is hidden phonological structure. Due to recent developments in computational modeling of phonological learning, there now exist numerous approaches for learning of various kinds of hidden phonological structure from incomplete, unlabeled, and noisy data. These computational models make it possible to connect the full representational richness of phonological theory with noisy, ambiguous corpus data representative of language learners’ linguistic experience to make detailed and experimentally testable predictions about language learning and generalization. In this talk, I briefly review these computational developments and then discuss two ongoing projects that utilize these mutually-informing connections between computation, phonological theory, and experimental data to test hypotheses about the abstract representations that underlie phonological knowledge.
