The grammatical relevance of thematic roles

Table of contents

1. Introduction

2. Government and Binding Theory
2.1. Hypothesis

3. Discussion
3.1. ө-roles
3.1.1. Case grammar

4. The grammatical relevance of thematic roles
4.1. Ergativity and ө-roles – empirical data

5. Conclusion

Bibliography

1. Introduction

Thematic roles are a phenomenon trying to explain the more specific relations between a predicate and its respective arguments. This is especially important in the field of language acquisition, or, more generally stated, for the process of understanding. It is acknowledged that if you know the grammar of a language you really know this language. Since grammar takes a closer look at units of speech such as sentences, and smaller chunks of sentences, one has to come up with some strategies for interpreting these chunks and the sort of relation they have with each other in order to know grammar. Theta theory is one of these strategies, focusing on “who does what to whom in a clause”^[1]. In short: a predicate is able to assign certain thematic roles to its respective arguments - we say that predicates have a thematic structure. The part of grammar which regulates the assignment of these roles is called theta theory. As an issue of linguistic interest, thematic roles have gained importance during the last two decades:

“Roles were first introduced into generative grammar by Fillmore in 1968, then taken up by Gruber, and more recently by Jackendorff […] The semantic content of the thematic roles themselves, though, was not closely examined at the time […] Today, the status of thematic roles is different. With Chomsky`s recent acceptance of thematic roles and by his explicit references to Fillmore`s and Gruber`s theories, thematic roles have become popular as the basis of deriving syntactic structures, as can be observed in the literature.”^[2]

Theta theory as part of the Government and Binding Theory will be discussed in the following chapters. First, the theory has to be localized in the field of grammar. Then, there follows a brief introduction to famous linguist Noam Chomsky`s current theory of principles and parameters. After these introductory sections the main issues will be addressed, for example the grammatical relevance of thematic roles, and also the question as to which extent theta marking is grammatically realized. There is no overall agreement on this question. Instead, various authors have given different answers to it, with some of them denying the visibility of thematic roles in grammar altogether^[3], ascribing linking processes to alternating mechanisms of grammar. On the other end of the line the meaning of a verb is said to completely determine its configuration of thematic roles, which in turn “determine the grammatical structures in which the verb may appear and the morphological forms it relates to.”^[4] But what part of grammar is it exactly in which to place the concept of thematic roles?

1.1. Thematic roles in grammar

First of all, it is a part of lexical semantics, and we can say that it is a part which poses much more questions than others. The lexical items that lexical semantics is concerned with form the basic expressions which composite meaning. As in semantics in general, one distinguishes three major approaches to meaning^[5]:

- To ask how words refer to entities (denotational approach)
- To ask how words are represented in the mental lexicon (mentalistic approach)
- To ask how words are used in the appropriate context (pragmatic approach)

Theta theory is about the denotational approach, and the discussion about the grammatical relevance will soon lead to a discussion of the respective entities and their representation. Theta roles are a traditional problem in lexical semantics. The reason is that with its overlapping between syntax and semantics, theta theory in a way questions the modular model of grammar. This is called the modularity problem, with syntax being one module of grammar, and semantics another. But there are also theories claiming that actually there is an interaction going on between these modules during the processing of thematic information. Jackendorff`s thematic theory would be an example for this interactive or parallel approach:

“The interactive position makes the prediction that there should be circumstances in which parsing decisions are affected by thematic, semantic, and even discourse factors. Over recent years evidence has accumulated for this position.”^[6]

Now, modularity is part of the build up of Noam Chomsky`s GB theory, from which theta theory as also described by Liliane Haegemann derives. Here one of the most important assumptions is the so-called Projection Principle, stating that lexical information is syntactically represented. This idea raises questions about the amount of influence thematic, and thus also semantic information has on syntactic structure building. As Yael Ravin writes:

“One of the central issues in modern linguistics has been the relationship between syntax (or grammar) and semantics (or meaning).”^[7]

But in order to step into the discussion, a quick look at the principles of GB is necessary, since the here proclaimed relevance of thematic roles is based on the concepts developed in this theoretic framework.

2. Government and Binding Theory

This theory has been first described by famous linguist Noam Chomsky in his Lectures of Government and Binding (LGB). It is on the one hand a further development of his former concept of transformational grammar, but also a serious departure from this concept, because these lectures had some substantial revisions in them:

“GB is based squarely upon two ideas. First, the grammars of all languages are embedded in a universal grammar, conceived as a set of universal principles applying equally to the grammar of every language. Second, within universal grammar, the grammars of particular languages may differ only in small and specified respects. These possible variations are conceived of as parameters, and the idea is that the grammar of any single language will be characterized by the use of a particular setting for each one of these parameters…”^[8]

Chomsky talks about three different levels of representation, there is the former deep structure (now called D-structure), a more superficial structure (S-structure), and finally the level of Logical form (LF). GB also distinguishes several modules, which are called the components of grammar. These are: Government Theory, Binding Theory, X-bar Theory, Control Theory, Case Theory, and, finally, Theta Theory. The lectures were also published as a book, and here Chomsky gives a short introduction to theta-role assignment and introduces the central notion of a Theta-Criterion:

“Idioms apart, each position satisfying the subcategorization features of the lexical head of a construction is a ө-position; in the terminology of X-bar theory, each complement position is a ө-position…Some ө-positions are filled by arguments; all complements of a head are ө-positions apart from examples restructured by idiom rules… the subject is a ө-position where a ө-role is determined for it. A reasonable criterion of adequacy for LF is (4):

(4) Each argument bears one and only one ө-role, and each ө-role is assigned to one and only one argument .”^[9]

In GB, theta theory is the module which deals with the valency requirements of verbs. By valency we understand “the property possessed by a predicate of requiring a certain number of argument noun phrases or clauses.”^[10] The proper distribution of ө-roles, according to GB principles, is mediated chiefly by the Projection Principle and the Theta-Criterion. But in how far do these premises hold true when comparing different languages? What about ergative languages? And is it possible to define the respective theta roles the arguments take in a universal way, thus giving an exact number of them, while also showing their visibility in grammar/in syntactic realities? Do we then, finally, “need” ө-roles in grammar at all? These are questions not answered in the Pisa Lectures, and critics have found all sorts of other complaints, but nevertheless “GB has for years been far the most influential and widely-practised theory of grammar in existence.”^[11]

[...]

^[1] Huddlestone, 19.

^[2] Ravin, 3-5.

^[3] For example linguists like Dowty/Bouchard, in: Härtl, 94/95.

^[4] Ravin, 4.

^[5] In: Lexicon of Linguistics, http://tristram.let.uu.n./Uil-OTS/Lexikon/index.html.

^[6] John N. Williams, 11-12.

^[7] Ravin, 1.

^[8] Trask 1999, 108.

^[9] Chomsky, 35/36.

^[10] Huddleston, 368.

^[11] Trask 1999, 109.