Die Überprüfung von Hypothesen über Autorabsichten

An important type of hypotheses within the interpretation of discourse and text pertains to the intentions and beliefs guiding the production of discourse and text. Such hypotheses can explain various properties of the discourse or text to be interpreted. I concentrate upon hypotheses about authors' intentions. I try to show that such hypotheses can be tested empirically, i. e. that they can be confirmed and disconfirmed with empirical data. Hypotheses about authors' intentions are singular hypotheses about specific authors and their specific circumstances, as opposed to universal hypotheses. I proceed from general aspects of testing singular hypotheses to more specific aspects. In section 1 I discuss the general question how singular hypotheses can be confirmed and disconfirmed, i. e. how they can be subjected to severe or even to critical tests. If an empirical statement follows from a hypothesis and background knowledge and the empirical statement is tested, then we have a severe test of the hypothesis. If, in addition, the hypothesis is confronted with alternative hypotheses, we have a critical test. In section 2 I analyze methods of testing hypotheses about mental states in general, in section 3 methods of testing hypotheses about authors' intentions in particular. How do we test ascriptions of beliefs and intentions in everyday life? We confront them with verbal utterances and non-verbal behaviour, using general nomological hypotheses of folk psychology within the background knowledge. In the social sciences and the humanities we proceed in the same way as in everyday life, but we do this in a more systematic manner: (1) we formulate our hypotheses more precisely; (2) we formulate implicit hypotheses explicitly; (3) we subject presupposed hypotheses about beliefs and intentions to severe and, possibly, critical tests; (4) we try to improve the nomological hypotheses within the background knowledge. With regard to hypotheses about authors' intentions there are two types of evidence particularly relevant for testing such hypotheses: (1) the conformity with the context of the text to be interpreted (in particular former drafts of the text, parallel texts, the social environment); (2) assertions of intentions made by the authors themselves. Furthermore, (3), a special role is played by the question whether hypotheses conform with charity principles within the background knowledge. I argue that charity principles have only a heuristic value, that they can