Abstract
This paper argues that pitch accent patterns of two-member Sino-Japanese compounds, hitherto considered unpredictable, can be strongly predicted by positing gradiently-valued accent features in the input, in the framework of Gradient Symbolic Computation "GSC", (Smolensky and Goldrick 2016). A simple machine-learning algorithm finds accent-affecting propensities = activations that collectively work for a set of compounds with frequently-occurring morphemes from the NHK corpus. I show that gradient input representations are needed to explain these kinds of phenomena. Examining a set of examples in which switches of morpheme order can change the accent pattern in ways that prosody cannot account for, I show that such phenomena can be explained by GSC but not by systems that have discrete-valued inputs and weighted, lexically-indexed constraints, thus providing evidence in favour of the GSC framework.
Highlights
Predicting the pitch-accent patterns of two-member Sino-Japanese compounds presents an analytic challenge (Kawahara, 2015:460)
The foregoing analysis shows that the pitch accent patterns of two-member Sino-Japanese words exhibit subtle, gradient tendencies, where the surfacing of accent depends not just on prosodic factors but is lexically influenced by underlying properties of both M1 and M2, where neither of M1 or M2 alone can singlehandedly determine where and if accent surfaces
The facts of Sino-Japanese pitch accent provide support for models such as GSC with partially-activated input features, given the argument in §9 above that models that only allow discrete inputs and lexically-indexed constraints cannot capture all of the data
Summary
Predicting the pitch-accent patterns of two-member Sino-Japanese compounds presents an analytic challenge (Kawahara, 2015:460) Both morphemes ( M1 and M2), show general accenting tendencies, I show, using a corpus of compounds from the NHK Accent Dictionary, that deriving their accent patterns without lexically listing each compound accent requires gradient feature values, as provided by the Gradient Symbolic Computation framework. A simple machine-learning algorithm finds accent-affecting propensities = activations that collectively work for a set of compounds with frequently-occurring morphemes from the NHK corpus. This analysis provides evidence that gradient input representations are needed to explain these kinds of phenomena. I show why the GSC framework can capture these data but discrete inputs with lexically-indexed constraints cannot, even if constraints are weighted. §10 gives details of the learning algorithm that was tested through computer simulation in order to discover input values and constraint weights that can account for the data. §11 explains how cross-validation was used to test how well a learner might predict unseen forms from exposure to partial data. §12 discusses how some unwanted types of coalescence might be ruled out. §13 concludes
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.