Empirical Applications of Type-Logical Categorial Grammar

Empirical Applications of Type-Logical Categorial Grammar (LSA Institute 2015)


Final assignment

Final assignment (due midnight Thursday 7/30)

grammar file for the final assignment

lexicon.pl with the abbreviation for ptv(X)


Sample solution for the final assignment

grammar file containing sample solutions

Comments on the final assignment

(1) There are (at least) two solutions for this question. One is to follow the strategy described in the Day 2 slides (p.32) and slant the quantifier alone in the right category. test(final1) gives this solution (and (5) is the relevant derivation). The other approach that one of you has come up with is to slant down the whole shared string "John gave a book" in the category S/(VP\VP)/PP. test(final1alt) gives this solution (and (2) is the relevant derivation).

(2) The point of this example was to show that the S|NP|NP analysis of parasitic scope extends straightforwardly to the RNR case. So, all you need to do is to modify the syntactic category for thesame in the lexicon to S|(S/NP)|NP and define a conjunction entry for 'sums' of S/NP type meanings. The proof strategy is explained in the Kubota 2015 paper (see p.26-27). final(test2) gives the solution and (3) is the relevant derivation.


Instructors

Yusuke Kubota (lastname.firstname.fn@u.tsukuba.ac.jp), University of Tsukuba, Ohio State University

Robert Levine (lastname.1@osu.edu), Ohio State University


Course website

http://www.u.tsukuba.ac.jp/~kubota.yusuke.fn/lsa/lsa-institute2015.html (this page)


Course description

The goal of this course is to familiarize students with Type-Logical Categorial Grammar (TLCG) as a framework that provides a new perspective on the syntax/semantics interface of natural language. While TLCG has so far been mostly studied in its connection to mathematical logic, our course emphasizes its value to the working linguist as a framework for characterizing linguistic generalizations, especially in empirical domains that have been widely regarded as problematic or even intractable in the mainstream linguistic literature, including both transformational and nontransformational approaches. To this end, we introduce a new version of TLCG which synthsizes two strands of research in the logical tradition, one old (Lambek calculus) and the other relatively new ('Lambda grammars').


Times

Second session

Monday: 3:10pm-5:00pm

Thursday: 3:10pm-5:00pm

Room

Saieh Hall for Economics 242


Schedule, readings and slides

Week 1: Basics

Week 2: Advanced topics


Final assignment, grading

We'll ask you to complete one of the following two tasks that you choose:

  1. Do some moderately challenging derivations that we assign (this will be given at some point during the second week).

  2. Develop a toy grammar in the LinearOne parser, gradually adding lexical entries as we extend empirical coverage, and demonstrate how it handles some moderately challenging empirical phenomenon that the student chooses him/herself (this can be either one of the topics we cover in the course or something else).

Some notes on recommended learning strategy:

Some suggested topics for the final assignment (if you choose the parser implementation option):

Note: These are somewhat advanced and some of them require additional reading. Since these tasks are challenging, we'll give you an A even if you don't come up with a completely satisfactory solution, as long as your fragment contains some thoughtful initial attempts at implementing these analyses (which could be extended to a more complete analysis with more work).

  1. Implement the 'affix-hopping' analysis of auxiliaries from Greg Kobele's course on MG and show how it interacts with coordination (e.g. John kicked and opened the box, John has been reading and will be discussing the book). [You don't need to implement the affix realization part. Just pretend that a morphological component handles this appropriately.]

  2. Read the Winter/Zwarts article on incorporating event variables in ACG. Implement their analysis in the parser.

  3. Implement the Jacobson 2007 analysis of Principle B effects. [For convenience, treat the presupposition on verb lexical entries she introduces as entailments.]

  4. Implement the Jacobson analysis of ACD discussed in our paper on pseudogapping. You should be able to do this by combining our pseudogapping analysis and the Muskens analysis of filler-gap dependency (Exercise B from Day 2).

  5. Implement the Morrill analysis of pied-piping described in Carpenter 1997, pp.364-367.

  6. Implement the Barker 2013 analysis of sluicing. [Read Kubota 2014 first for how to translate this analysis to the Hybrid TLCG setup.]


Related/background materials

Lambek calculus, TLCG

ACG, Lambda grammar, Linear Categorial Grammar

Integrating event semantics with the CG syntax-semantics interface (ACG, in particular)

Hybrid TLCG

See YK's publications page for other papers.

CG analyses of some linguistic phenomena

Key linguistic literature