Gilbert Harman

In this article, we provide a tutorial overview of some aspects of statistical learning theory, which also goes by other names such as statistical pattern recognition, nonparametric classification and estimation, and supervised learning. We focus on the problem of two-class pattern classification for various reasons. This problem is rich enough to capture many of the interesting aspects that are present in the cases of more than two classes and in the problem of estimation, and many of the results can be extended to these cases. Focusing on two-class pattern classification simplifies our discussion, and yet it is directly applicable to a wide range of practical settings

Solomon argues that, although recent research in social psychology has important implications for business ethics, it does not undermine an approach that stresses virtue ethics. However, he underestimates the empirical threat to virtue ethics, and his a priori claim that empirical research cannot overturn our ordinary moral psychology is overstated. His appeal to seemingly obvious differences in character traits between people simply illustrates the fundamental attribution error. His suggestion that the Milgram and Darley and Batson experiments have to do with such character traits as obedience and punctuality cannot help to explain the relevant differences in the way people behave in different situations. His appeal to personality theory fails, because, as an intellectual academic discipline, personality theory is in shambles, mainly because it has been concerned with conceptions of personality rather than with what is true about personality. Solomon’s rejection of Doris’s claims about the fragmentation of character is at odds with the received view in social psychology. Finally, he is mistaken to think that rejecting virtue ethics implies rejecting free will and moral responsibility.

Any acceptable account of moral epistemology must accord with the following points. (1) Different people acquire seemingly very different moralities. (2) All normal people acquire a moral sense, whether or not they are given explicit moral instruction. Language resembles morality in these ways. There is considerable evidence from linguistics for linguistic universals. This suggests that (3) despite the first point, there are moral universals. If so, it might be possible to develop a moral epistemology that is analogous to the theory of universal grammar in linguistics. In what follows, I will try to sketch what might be involved in such a moral epistemology.

Philosophers sometimes approach meaning metaphorically, for example, by speaking of “grasping” meanings, as if understanding consists in getting mental hands around something.1 Philosophers say that a theory of meaning should be a theory about the meanings that people assign to expressions in their language, that to understand other people requires identifying the meanings they associate with what they are saying, and that to translate an expression of another language into your own is to find an expression in your language with the same meaning as the expression in the other language.