Volume 28, Issue 3 - The Theta System – an overview

This paper presents an overview of a larger project in progress on the concepts interface. In part, it is based on the findings in Reinhart (2000), where several of the problems are discussed in greater details. However, many aspects of the system have been further developed, or changed, since that manuscript.

When we consider a new theory and we wish to uncover its merits and shortcomings, it is necessary to take into account the data that support both the theory as a whole and each individual assumption that makes up the theory and to compare it with alternative theories and assumptions that have similar descriptive and explanatory goals. This is what I will do in the following pages with Reinhart's “The Theta System”.

Introduction Reinhart (this volume, henceforth abbreviated RTS) defines Theta Roles in terms of two features, [c] and [m], and their binary values [+], [−]. Feature [c] covers the concept of causation, feature [m] is associated with the concept of motivation/mental state, or, more basically, with animateness (in view of the fact that non-animate entities don't bear motivations). Argument alternations (called ‘arity’ operations in RTS) are viewed as output of strictly lexical operations: whenever a given verb varies between different argument patterns, one of them has to be analyzed as basic and the others as lexically derived ( Lexicon Uniformity Principle ). RTS focusses on one type of arity operation: the reduction of arguments. It accounts for the reduction of internal arguments (reflexivation operation The barber shaved Max → Max shaved ) as well as for the reduction of external arguments ( Max opened the door → The door opened ) and is furthermore concerned with EXPERIENCER-verb-alternations ( Something (THEME) worries Lucie (EXPERIENCER) → Lucie (EXPERIENCER) worries about something (SUBJECT MATTER) ). The languages under consideration in RTS include English, Hebrew and Dutch. These are different enough to support universal claims (more or less implicit in RTS). In the following remarks, I will present data from German, where this adds further evidence to the PROs and CONs of RTS.

Reinhart's ‘The Theta System’ is part of a series of publications (Reinhart 1997, Reinhart 2000, Reinhart and Siloni to appear) dealing with the lexicon as the interface of the computational system and the conceptual-intentional interface. In this paper, Reinhart focuses on Theta Theory, i.e. a complex consisting of Theta roles as feature bundles; Arity manipulation rules; Mapping procedures. It is work in progress, which means that Reinhart changes her positions in these papers, but for argument sake, and for the sake of the reader, I will limit myself to positions taken by Reinhart in this volume. Reinhart sketches the general contours of the Theta System, necessarily succinct, and limits the empirical domain in this paper to experiencer constructions. In this respect too, I will take over the limitations set out by the paper.

1. Introduction Reinhart's paper outlines her view on the structure of the Theta system. She considers the Theta system as “the central system of the system of concepts”. The Theta system outputs lexical entries of a certain arity with specified theta roles and information on the merging properties of the arguments. The marking procedures that determine the merging properties refer to the formal features that define the theta roles.

0. Introduction In order to be legible to the CS, thematic roles in the ‘Theta-System’ (Reinhart, this volume) are formally coded in terms of two binary features: +/−c (cause change) and +/−m (mental state).

Tanya Reinhart's “The Theta System” is mainly meant to explain the linking behaviour and the particular binding properties of experiencing verbs. Following Pesetsky (1995), Reinhart argues that verbs like worry possess three arguments: a cause , an experiencer and a subject matter role. Backwards bound anaphors are enabled only if it is the subject matter which is realized externally – it originates VP-internally and its trace is c-commanded by the experiencer : (1) His i health worries every patient i . (Reinhart's example (42a)) If the subject is construed as the cause , it merges directly externally, and the c-command requirement is not met. Bound anaphors are not licensed in this case: (2) ??His i doctor worried every patient i . (Reinhart's example (65b))

The theta system, which Reinhart (this volume) views as the central part of the systems of concepts, includes (i) lexical entries, which are coded concepts; (ii) derivational operations, which are operations that apply to concepts to derive related entries; (iii) mapping principles, which instruct the computational system (syntax) with regard to the order of merging of arguments. The central part of the systems of concepts is, in fact, the component traditionally labeled (mental) lexicon. Reinhart assumes that the lexicon is an operative component of grammar; the theta system is a theoretical model for the lexical component.

Tanya Reinhart's “Theta System” (henceforth: TS) is yet another version of what we call linking theories, i.e., theories that formulate generalizations about how verbs project their semantic arguments into syntactic constituent structure. The last decades of linguistic research have seen many proposals of this kind. The TS is particularly designed to fit into a minimalist theory of syntax, but, as far as I can see, there is nothing particularly minimalist in it, apart from details of technical execution. Neither is TS the only possible linking theory within minimalism.