Two problems are discussed: Can and should psychological concepts be defined, and can and should they be organized in an axiomatic system? I point out that definitions in terms of physiological or behavioral measures are strictly impossible because any particular measure can mean anything, whereas phenomenological definitions always point to antecedents and consequents. I then point out that definitions of antecedents and consequents can be given either in terms of causes or in terms of reasons, and that causes and reasons (...) belong to two incompatible languages. Causes exist independently of persons and reasons exist only for persons. Only the latter belong to psychology, as I see it. The working of the system of psycho-logic is illustrated by an analysis of the conditions of trust. I conclude that psychological terms should be defined by means of semantic primitives, and that one should try to organize them in an axiomatic system. 2012 APA, all rights reserved). (shrink)