Documentation for the Grammar Matrix Customization Case Library


This document explains how to fill out the Case page of the Grammar Matrix Customization questionnaire and presents background information on the Case library of the Grammar Matrix Customization System (Bender et al., 2002; Bender and Flickinger, 2005; Bender et al., 2010). General instructions on using the questionnaire can be found here.

Citing the Case Library

The standard reference for the Case Library and its implementations is Drellishak 2009. The full reference and .bib entry can be found here.


The case library allows the user to specify the range of case values used in the grammar (if any), and the general type of the language's system for marking core cases. On the Lexicon page, each verb type can have an argument structure specified. The options available there depend partly on the answers to the questions on the Case page, but there is always the option of defining a verb class that uses a different case pattern than those provided on the basis of the selection for core case marking.

The Grammar Matrix customization system provides you with nine pre-defined case system options covering the most commonly used case systems, as well as one option allowing you to define additional cases. Following Dixon's terminology (Dixon 1994), cases are discussed in terms of the roles of the arguments: A (agent of a transitive verb), O (patient or object of a transitive verb), and S (subject of an intransitive verb).

The next four options are the subtypes of the split ergativity case systems, which are neither nominative-accusative, nor ergative-absolutive. These case systems can be conditioned by the following factors: (1). semantic nature of the main verb, (2). semantic nature of the core NPs, (3). tense, aspect, or mood of the clause, and (4). grammatical status of the clause (Dixon 1994:70). Please note that the latter two (case conditioned on tense/aspect/mood and case conditioned on grammatical status of the clause) receive a similar analysis, described in more detail in Drellishak 2009, Section 3.2.

The values you assign to the CASE feature will be available to use later in the customization system, especially the Morphology and Lexicon pages. The answers you provide on this page will determine the values available on the Lexicon page for the CASE feature. CASE will also appear as one of the features of the lexical rule types on Morphology page.

Further information about the options on the Case page is provided in in Drellishak 2009, Section 3.3.


While some language do not have a case system, other languages that do use cases rely on them heavily. The case library of the Grammar Matrix Customization system was developed to cover case systems that are most commonly used across languages, providing user with nine pre-defined options in the questionnaire. To provide the user with more flexibility in coverage of the less commonly used case systems or of the quirky case, the Grammar Matrix Customization system also allows user to define additional case values instead of just relying on the pre-defined analyses. For more information on the analyses of case systems please refer to the Analyses section below.


After you define a case system for your language, your starter grammar will include the feature [ CASE case ], with possible values based on the choices you make on the Case customization page.

Below is a snippet of code related to case from choices file for a language with nominative-accusative case system:


This can be compared to code from the same section of the choices file for a language that does not use a case system:


If no further specifications for case is defined on the Lexicon or Morphology pages of the Grammar Matrix customization system, the case hierarchy will be flat, i.e. all cases will be direct subtypes of the feature-value case.

case := avm.
nom := case.
acc := case.
gen := case.
dat := case.
abl := case.

A more elaborate hierarchy can be formed by indicating common underspecifications for lexical items or morphemes. For instance, the Latin word tempus can either be nominative or accusative, but not genitive, dative or ablative. If you define tempus in the lexicon with nom and acc as case values, the hierarchy will included an underspecified form for case that is either nominative or accusative:

case := avm.
nom-or-acc := case.
nom := nom-or-acc.
acc := nom-or-acc.
gen := case.
dat := case.
abl := case.

Following Pollard and Sag (1994), CASE is a feature of the syntactic HEAD. The cases of both arguments (subject and object) are specified on the verb lexical types. Grammar Matrix provides the following case-marking strategies in the Lexicon section: (1). marking morphologically on nouns, determiners, or (2). marking of whole NPs via case-marking adpositions, or (3). both (mixed marking).

For the detailed analyses of each of the case-marking systems covered by Grammar Matrix Customization system please refer to Drellishak 2009, Section 3.2.

Upcoming Work

Although Case library of Grammar Matrix Customization system is already quite flexible, providing user with nine pre-defined case systems and allowing user to create additional cases, it could be extended for a broader coverage of case and its interactions with other grammatical phenomena.

One of the limitations of the Grammar Matrix is its lack of coverage of the more complex argument-marking cases, e.g. some fine interactions between case-marking and verb forms. Adding support for these interactions between case and verb forms (or between case and other parts of grammar) would provide a broader coverage of the grammatical phenomena described in these sections, as well as make the interaction between these sections of Grammar Matrix Customization system more dynamic.

Another current limitation of Grammar Matrix customization system is lack of coverage of the syntactic ergativity. Although syntactic ergativity is an inter-clausal phenomenon and absence of its coverage does not yet affect grammars created by Grammar Matrix Customization system, it would be beneficial to implement it in the future. Providing support for syntactic ergativity would also aid in the enhancements to the sections on relative clauses, control, and binding, which can interact with syntactic ergativity. This could be achieved through the integration of the HPSG analysis of syntactic ergativity that has already been developed by Manning and Sag (1995).


Drellishak, S. (2009). Widespread but Not Universal: Improving the Typological Coverage of the Grammar Matrix. PhD thesis, University of Washington.

Dixon, R. M. W. (1994). Ergativity. Cambridge: Cambridge University Press.

Levin, L. S., & Simpson, J. (1981). Quirky Case and Lexical Representations of Icelandic Verbs. Chicago Linguistic Society 17, 185–196.

Manning, C. D., & Sag, I. (1995). Dissociations between Argument Structure and Grammatical Relations. In Lexical and Constructional Aspects of Linguistic Explanation, pages 63–78, CSLI.

MatrixDoc/Case (last edited 2013-10-01 16:58:49 by EmilyBender)

(The DELPH-IN infrastructure is hosted at the University of Oslo)