Ambivalence, i.e., the simultaneous holding of negative and positive evaluations toward the same object, is an empirically well-documented phenomenon and an important aspect of ordinary experience. However, it has not received sufficient philosophical attention. This essay accomplishes two aims: first, a comprehensive and empirically informed account of ambivalence is provided; second, the rationality of ambivalence in practical and nonpractical contexts is defended.

It can be assumed that value judgements, which are needed to judge what is ‘good’ or ‘better’ and what is ‘bad’ or ‘worse’, are involved in every decision-making process. The theoretical understanding and analysis of value judgements is, therefore, important in the context of bioethics, for example, to be able to ethically assess real decision-making processes in biomedical practice and make recommendations for improvements. However, real decision-making processes and the value judgements inherent in them must first be investigated empirically (‘empirical (...) bioethics’). For this to succeed, what exactly a ‘value judgement’ is and of what components it might consist must initially be theoretically clarified. A corresponding conceptual model can then support or even enable empirical data collection and analysis and, above all, subsequent ethical analysis and evaluation. This paper, therefore, presents a value judgement model with its theoretical derivation. It also illustrates its application in an interview study of decision-making between animal experimentation and alternative methods in the context of biomedical research. Though the model itself can be theoretically deepened and extended, the application of the model works in general and helps to uncover what value judgements can enter into decision-making. However, the empirical methods, for example, qualitative interviews, can also be better oriented towards eliciting value judgements (as understood according to the model). Further applications of the model to other topics or by means of other empirical methods are conceivable. (shrink)