**Course:** TThu 09:30-10:50, Compling lab

**Instructor:** Jeffrey Heinz, jeffrey.heinz@stonybrook.edu

**Office Hours: Mondays 2-5pm and by appointment**

- Class Canceled. Office Hour by Zoom to catch up and ask questions if needed.

- We discussed ISL functions, ISL transductions, and QF(succ,pred) functions.
- Here is the draft chapter of “Logical Perspectives on Strictly Local Transformations” by Chandlee and Lindell.

- Using this handout, we formally introduced two learning criteria: identification in the limit from positive data and PAC.
- We discussed compared the two frameworks and thought about what choices in a learning criterion make it easier or harder.
- Supporting reading material: “Computational Theories of Learning and Developmental Psycholinguistics” (Heinz 2016)
- Please read “Strict Locality and Phonological Maps” (Chandlee and Heinz 2018) for Thursday.
- Also, we had some discussion about computable and noncomputable real numbers. This recent summary from the ACM reviews some of the issues. This paper “How real are real numbers” is a bit more in depth.

We continued our discussion of learning problems, and developed one related to learning strictly k-local languages. Those techniques have been generalized and developed in the papers String Extension Learning and Learning with Lattice-Structured Hypothesis Spaces. See also Typology emerges from simplicity in representations and learning.

On November 7, Jack will present Computing and classifying reduplication with 2-way finite-state transducers. Then Anton will present Computational universals in linguistic theory: Using recursive programs for phonological analysis.

- We discussed related work (references). Students were asked to review those papers and think about which ones they would be interested in for class.
- We began discussing how to formulate learning problems. Learning handout 1

- Jake finished his presentation.
- Alina presented on Chapter 19 Representations of Gradual Oppositions.

**FALL BREAK – NO CLASS**

- M presented Chapter 17 on Correspondence Theory.
- Jake began presenting on Chapter 16 on Primitive Constraints for Nonlexical Stress Patterns.
- Jake will finish presenting next Thursday before Alina’s presentation.

- Sarah presented Chapter 15 on Syllable Struture and the SSP.
- Logan presented Chapter 18 on Phonetically-based Phonology.
- Next Thursday Oct 15, Alina will present Chapter 19 Representations of Gradual Oppositions.

- Shigeto’s presented Chapter on Zigula and Shambaa.
- Elizabeth presented Chapter on Tibetan.

**Short Paper Assignment**(for students enrolled in 3 credits): Please turn in no later than Tuesday October 22, 2024 (in class). Please print your assignment and hand it in to me.For this assignment, choose a morpho-phonological data set and provide an analysis of it. Your analysis should formalize the analysis using the logical transductions we have studied in class. Chapters 7 through 14 in Part 2 of DCP/DPC are samples of the kind of short paper expected. It is not necessary to compare your formalization to some other framework. It is necessary to be clear about the representations you are positing in the underlying and surface forms.

You

**may not choose**one of the datasets in chapters 7 through 14 (unless you are proposing a radically different analysis.) You**may choose**a dataset you have previously analyzed using a different set of formal tools (such as with rules or OT). For example, datasets from Phonology 1 or 2 can be used, or even a dataset from an undergraduate phonology class or introductory textbook. You may choose a dataset you have collected, but such original research is not the purpose of this assignment, and may be more appropriate for a final project.A collection of tex files providing macros and examples for your convenience. I commpiled the pdf in there with the command

`latexmk dolatian-russian.tex`

.Ola presented Chapter 9 on Lamba.

Geonhee presented Chapter 10 on Turkish.

We developed a plan for next week. Updated DCP/DPC text here

- Oct 8: Presentation on chapter 15 Syllabls by Sarah
- Oct 8: Presentation on chapter 16 Stress by Jacob
- Oct 10: Presentation on chapter 18 Phonetically-grounded representations by Logan
- Oct 10: Presentation on chapter 17 Correspondence Theory by M

- A few more words were said about suffix substitution closure, subsequence closure, and tier-based successor signatures.
- We then discussed BUFIA, using slides from a minicourse a few years back.
- Next week, please read chapters 9-12 in DCP/DPC and support your classmates as they present!

- We finished discussing chapter 5.
- On Thursday we will review the BUFIA learning algorithm in this paper.

- We discussed section 5.1 and 5.2.
- We made a plan for the next few weeks.
- Oct 1: Presentation on chapter 9 Lamba by Ola
- Oct 1: Presentation on chapter 10 Turkish by Geonhee
- Oct 3: Presentation on chapter 11 Tibetan by Elizabeth
- Oct 3: Presentation on chapter 12 Zigula and Shambaa by Shigeto

- For Tuesday, please finish reading chapter 5.

- We finished discussing Chapter 3.
- Upated DCP/DPC Parts 1 and 2

- We discussed logical transductions as they can be applied to phonological processes.
- We got through the deletion example in Chapter 3.

- We finished discussing Chapter 2.
- If you are interested in attending the Rutgers workshop on subregular phoonology, please let me know this week.
- For Thursday, please finish read up to Chapter 3.3 in Part 1 of DCP/DPC.

- We discussed the Chapter up to 2.6. In particular, how the successor model of strings, with and without features, can be combined with FO logic to provides a way to define well-formedness constraints. The representational primitives can be thought of as claims regarding the psychologically real primitives. They are the aspects of the representation that can be “looked up” for free.
- For Tuesday, please finish reading chapter 2 in Part 1 of DCP/DPC. (Note it has been updated
**again**to fix more errors).

- We discussed Chapter 1 and began discussion of models of strings and FO logic in Chapter 2.
- For Thursday, please read up to section 2.6 (p.43) in Part 1 of DCP/DPC. (Note it has been updated to fix some errors).

- We finished the Phonology Forum talk. Next Tuesday we discuss Chapter 1.

- We went over the syllabus.
- As an overview of the class, I presented a talk I gave last week at the Phonology Forum sponsored by the Phonological Society of Japan.
- Here is Part 1 of “Doing Computational Phonology”/“Doing Phonology, Computationally”.
- For Thursday, please read Chapter 1 of Part 1.