Computer Science > Computation and Language
[Submitted on 15 Dec 2019]
Title:Computational Induction of Prosodic Structure
View PDFAbstract:The present study has two goals relating to the grammar of prosody, understood as the rhythms and melodies of speech. First, an overview is provided of the computable grammatical and phonetic approaches to prosody analysis which use hypothetico-deductive methods and are based on learned hermeneutic intuitions about language. Second, a proposal is presented for an inductive grounding in the physical signal, in which prosodic structure is inferred using a language-independent method from the low-frequency spectrum of the speech signal. The overview includes a discussion of computational aspects of standard generative and post-generative models, and suggestions for reformulating these to form inductive approaches. Also included is a discussion of linguistic phonetic approaches to analysis of annotations (pairs of speech unit labels with time-stamps) of recorded spoken utterances. The proposal introduces the inductive approach of Rhythm Formant Theory (RFT) and the associated Rhythm Formant Analysis (RFA) method are introduced, with the aim of completing a gap in the linguistic hypothetico-deductive cycle by grounding in a language-independent inductive procedure of speech signal analysis. The validity of the method is demonstrated and applied to rhythm patterns in read-aloud Mandarin Chinese, finding differences from English which are related to lexical and grammatical differences between the languages, as well as individual variation. The overall conclusions are (1) that normative language-to-language phonological or phonetic comparisons of rhythm, for example of Mandarin and English, are too simplistic, in view of diverse language-internal factors due to genre and style differences as well as utterance dynamics, and (2) that language-independent empirical grounding of rhythm in the physical signal is called for.
Current browse context:
cs.CL
References & Citations
Bibliographic and Citation Tools
Bibliographic Explorer (What is the Explorer?)
Litmaps (What is Litmaps?)
scite Smart Citations (What are Smart Citations?)
Code, Data and Media Associated with this Article
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub (What is DagsHub?)
Gotit.pub (What is GotitPub?)
Papers with Code (What is Papers with Code?)
ScienceCast (What is ScienceCast?)
Demos
Recommenders and Search Tools
Influence Flower (What are Influence Flowers?)
Connected Papers (What is Connected Papers?)
CORE Recommender (What is CORE?)
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.