Computational Phonology: the Delaware School
AP&M 4218, Wednesdays 12-1:30pm, Spring 2017
The immediate aim of this reading group is to get us up to speed on the computational phonology work of Jeff Heinz (Delaware, but soon-to-be at Stony Brook), Jane Chandlee (Haverford), and Adam Jardine (Rutgers). Because Chandlee and Jardine were students of Heinz at Delaware, I’ll call this “the Delaware School” of computational phonology.
An aspirational aim of this reading group is to be sufficiently equipped to understand (and perhaps to begin to tackle) Heinz’s “Hilbert Problem for Phonology”, described in more detail here. Briefly, Heinz revives the question originally and famously posed by Kiparsky (1968) in the title of his paper, “How abstract is phonology?” Heinz restates the question this way: “How distinct can underlying forms be from surface forms?” Then, in an effort to “[place] the focus on the analytical principles that would lead a scientist towards, or away from, an abstract analysis,” Heinz formulates the problem this way: “What principles determine whether an abstract analysis of the observable facts is warranted or not?”
Week 1: introduction
- Heinz, Jeffrey. 2011. Computational Phonology - Part I: Foundations and Computational Phonology - Part II: Grammars, Learning, and the Future. Language and Linguistics Compass 5.4, 140–152 and 153–168.
- Helpful reading (from Stacy): Jurafsky, Daniel, and James H. Martin. 2006. Computational Phonology. Ch. 10 of Speech and Language Processing: An introduction to natural language processing, computational linguistics, and speech recognition.
- Helpful links (from Eric M.): Why are Polynomial Time Problems Considered Tractable, and Larger Times are Not?, Machine models and complexity measures - the last three parts, Best, worst and average case, Nondeterministic algorithm - Introduction and Use sections, NP (complexity) - Introduction & Equivalence of Definitions sections.
- Rosen, Kenneth. 2012. Discrete Mathematics and Its Applications (7th ed.).
- Jurafsky, Daniel, and James H. Martin. 2006. Speech and Language Processing: An introduction to natural language processing, computational linguistics, and speech recognition (2nd ed.); draft chapters of 3rd ed. in progress available here.
Weeks 2-3: foundation
- Kaplan, Ronald, and Martin Kay. 1994. Regular models of phonological rule systems. Computational Linguistics 20, 331-378. (Commentary by Mark Liberman; commentary by Graeme Ritchie.)
Possible further reading
- Johnson, C. Douglas. 1972. Formal aspects of phonological description. The Hague: Mouton. (1-up version; 2-up version.)
- Mohri, Mehryar. 1997. Finite-State Transducers in Language and Speech Processing. Computational Linguistics 23.2, 269-311.
- Rogers, James, Jeffrey Heinz, Margaret Fero, Jeremy Hurst, Dakotah Lambert, and Sean Wibel. 2013. Cognitive and Sub-Regular Complexity. In Glyn Morrill and Mark-Jan Nederhof (eds.), Formal Grammar (Lecture Notes in Computer Science vol. 8036), 90-108.
- Heinz, Jeffrey, and William Idsardi. 2013. What Complexity Differences Reveal About Domains in Language. Topics in Cognitive Science 5.1, 111-131.
- Rogers, James, and Geoffrey K. Pullum. 2011. Aural Pattern Recognition Experiments and the Subregular Hierarchy. Journal of Logic, Language and Information, 329-342.
Notes from Week 2 (from Eric M.)
- General plan: Revisit discussion of Kaplan & Kay after getting more hands-on experience with making FSTs and reading a bit more about what motivates their use.
- Mechanical practice with FSAs:
- Chapter 10/sections 1,2 and 5 of SLP 3rd edition; there are also a handful of end-of-chapter exercises.
- Consider using FOMA as a tool...
- If desired, Chapter 2 of SLP 2nd edition has a bunch more exercises (with solutions available - see first result) and additional explanations.
- Motivational reading:
- Chapter 10 sections 3.1 (on finite state OT) and 6 (on learning) of SLP 3rd edition
- Note: The second Heinz paper ('computational phonology part II') is probably worth revisiting if these two sections are intriguing.
- Research-y question to pursue - chain shifts and FSTs:
- Do FSTs (more specifically, the structure of FST diagrams or properties of/results about transducer composition) offer any insight into why chain shifts are the way they are or how they should be analyzed?
- IIRC, someone is going to lead us through chapter 10/section 5 of SLP 3rd edition and someone is going to take the lead on walking people through FOMA next week.
Weeks 4-5: overview
Week 6: metathesis
Week 7: vowel harmony
Weeks 8-10: strictly local processes
- Chandlee, Jane, and Jeffrey Heinz. To appear. Strictly Local Phonological Processes. Linguistic Inquiry.
- Chandlee, Jane, and Steven Lindell. To appear. A logical characterization of input strictly local functions. In Jeffrey Heinz (ed.), Doing Computational Phonology.
- Chandlee, Jane, Jeffrey Heinz, and Adam Jardine. Submitted. Input Strictly Local Opaque Maps.
Possible further reading