The Many Types of 'Theme' in English: their semantic systems and their functional syntax

In the 1960s Halliday made a number of innovative proposals about the concept of ‘theme’ which have been highly influential - both within Systemic Functional Linguistics (SFL) and in linguistics in general (e.g. in Quirk et al 1985). These proposals were modified in minor respects in Halliday 1985 & 1994, but in recent years scholars - including many working in SFL - have increasingly pointed out various problems. The present paper describes an alternative SFL approach to ‘theme’ in English that is based on the same principles, that is both comprehensive and pragmatic, and that resolves these problems. My starting assumption is that ‘thematic’ meanings are meanings between which a user of a language chooses, just as we choose between meanings of ‘transitivity’, ‘mood’ and so on. While the focus here is on the system networks for the various types of ‘theme’, I adopt an explicitly ‘trinocular’ approach. That is, in deciding how the system networks for the various types of ‘theme’ should be, I draw on evidence that is (1) ‘from above’ (the ‘discourse purposes’ that the various ‘thematic’ constructions may serve); (2) ‘at its own level’ (the relations of the various systems of choices in ‘thematic’ meanings to each other and to other systems, especially those in ‘transitivity’); and (3) ‘from below’ (the functional syntax that realizes the ‘thematic’ choices). This fresh approach to the data from first principles leads to the recognition of seven or eight distinct types of ‘theme’ (depending how we count them), together with several minor variants. The systems for the major types are summarized in one diagram (Figure 2). These include new syntactic and semantic analyses of the structures known in formal grammars as ‘cleft’, ‘pseudo-cleft’, ‘extraposition’ and ‘left-dislocation’, but I present the case for replacing these labels by explicitly functional ones. This new SFL description of ‘theme’ provides for five types (some quite frequent, some very rare) that are not covered in either Halliday 1994 or Matthiessen 1995, and provides analyses for other types that are different from - and simpler than - those of Halliday. In the presentation of each type of ‘theme’ the ‘trinocular’ approach is expressed though the fact that each has sub-sections on ‘form and meaning’ and ‘discourse purposes’. The paper also contains supplementary sections on (a) why we should not treat elements of the clause that always occur early - such as and and because -- as types of ‘theme’; (b) some apparently borderline cases of ‘theme’ in ‘reporting’ clauses; and (c) the fact that (‘unmarked’) Subject Themes are frequently covert - and that they need to be shown in the analysis. Taking this position has major consequences when analyzing texts, especially in the so-called ‘pro-drop’ languages such as Spanish. I provide analyses of the functional syntax of the key examples throughout, and I also analyze selected examples at the level of meaning. The picture of the discourse purposes served by the various types of theme is less complete, but I describe work in the COMMUNAL Project which models the ‘higher’ decision-making that affects what is required in the corresponding semantic system networks for ‘theme’. By the end of the paper it will be clear that the various types of ‘theme’ recognized in this comprehensive treatment cover a very wide range of meanings, and that they serve an even wider range of purposes in discourse - so raising the question of how far it is still helpful to think in terms of a single, ’unified’ concept of ‘theme’.