Topic, Focus and Local Relevance

Summary

T. Reinhart (80) has claimed that the naive identification of topic information with old information and of focus information with new information is inconsistent.

I shall try to reconstruct the old versus new information approach in such a way that no inconsistency can arise.

My topic information will be what Chomsky (71) calls presupposition. The focus information of a sentence will be the material implication between its topic information and its content (its truth-conditions). The focus-information is ‘the meaning’ of the sentencequa focus-structure.

The usefulness of these concepts is illustrated because they are needed to show claims like the following.

• 1. Different focus assignment does not affect the truth conditions of a sentence, but it is responsible for the fact that the same sentence expresses different assertions, if it has different foci.

• 2. Topichood is one essential notion for the definition of the concept of the ‘local relevance’ of one utterance for another.

• 3. The relation between answers and questions should be described in terms of local relevance. Such an approach will provide a much more general account of the relation of being an answer to a question than commonly found in the literature.

• 4. Negative sentences with different foci express diffferent assertions although they have the same truth-conditions. It makes a difference whether the negation belongs to the focus or not. The phenomena to be discussed can’t be explained by differences in the scope of the negation.

• 5. We are able to explain the difference between questions like ‘Who called, JOHN or MARY?’ or ‘Did JOHN call or did MARY call?’ on the one hand and the corresponding alternative question without focus, ‘Did John or Mary call?’ on the other hand.

The approach developed in this paper belongs to truth-conditional semantics. It is assumed that propositions determine truth-conditions. But we argue for something more: Propositions should be regarded as structured entities. They should at least have a topic-focus structure.

Acknowledgements. I profited from discussions with Irene Heim, Ruth Manor, Tanya Reinhart and Wolfgang Klein. Last but not least I wish to thank our secretary Sonnya Borho for patiently typing the manuscript. Thanks also to Josef Baier, Wolfgang Skonefeld, Christa Hauenschild. Dieter Wunderlich and Ede Zimmermann, who commented on an earlier version.