Montague Grammar Induction

We propose a computational model for inducing full-fledged combinatory categorial grammars from behavioral data. This model contrasts with prior computational models of selection in representing syntactic and semantic types as structured (rather than atomic) objects, enabling direct interpretation of the modeling results relative to standard formal frameworks. We investigate the grammar our model induces when fit to a lexicon-scale acceptability judgment dataset – Mega Acceptability – focusing in particular on the types our model assigns to clausal complements and the predicates that select them.

Formal Basis of a Language Universal

Steedman (2020) proposes as a formal universal of natural language grammar that grammatical permutations of the kind that have given rise to transformational rules are limited to a class known to mathematicians and computer scientists as the “separable” permutations. This class of permutations is exactly the class that can be expressed in combinatory categorial grammars (CCGs). The excluded non-separable permutations do in fact seem to be absent in a number of studies of crosslinguistic variation in word order in nominal and verbal constructions. The number of permutations that are separable grows in the number n of lexical elements in the construction as the Large Schröder Number Sn−1. Because that number grows much more slowly than the n! number of all permutations, this generalization is also of considerable practical interest for computational applications such as parsing and machine translation. The present article examines the mathematical and computational origins of this restriction, and the reason it is exactly captured in CCG without the imposition of any further constraints.

Categorial Grammar

The term “categorial grammar” refers to a variety of approaches to syntax and semantics in which expressions are categorized by recursively defined types and in which grammatical structure is the projection of the properties of the lexical types of words. In the earliest forms of categorical grammar types are functional/implicational and interact by the logical rule of Modus Ponens. In categorial grammar there are two traditions: the logical tradition that grew out of the work of Joachim Lambek, and the combinatory tradition associated with the work of Mark Steedman. The logical approach employs methods from mathematical logic and situates categorial grammars in the context of substructural logic. The combinatory approach emphasizes practical applicability to natural language processing and situates categorial grammars within extended rewriting systems. The logical tradition interprets the history of categorial grammar as comprising evolution and generalization of basic functional/implicational types into a rich categorial logic suited to the characterization of the syntax and semantics of natural language which is at once logical, formal, computational, and mathematical, reaching a level of formal explicitness not achieved in other grammar formalisms. This is the interpretation of the field that is being made in this article. This research has been partially supported by MINICO project TIN2017–89244-R. Thanks to Stepan Kuznetsov, Oriol Valentín and Sylvain Salvati for comments and suggestions. All errors and shortcomings are the author’s own.

Non-linear Second Order Abstract Categorial Grammars and Deletion

We prove that non-linear second order Abstract Categorial Grammars(2ACGs) are equivalent to non-deleting 2ACGs. We prove this resultfirst by using the intersection types discipline. Then we explainhow coherence spaces can yield the same result. This result showsthat restricting the Montagovian approach to natural languagesemantics to use only $\L I$-terms has no impact in terms of thedefinable syntax/semantics relations.