This paper argues that social representations cannot be used as independent variables in causal explanations of social behaviour. It is shown that the structure of investigations often follows a causally explanatory design despite explicit statements to the contrary by the researchers. This fact is analyzed with three investigations. It is argued that verbal data used to assess the contents of a representation as independent variable are logically equivalent to data obtained from the “dependent” overt behaviour. Therefore these two kinds of (...) data must be seen as two illustrations of the same representational contents. The researchers’preference for using verbal data to assess the independent variable and the tendency to introduce a causal relationship between representation and behaviour is shown to result from misplacing folk-beliefs. Folk-beliefs about intentional causality, it is shown, pertain to the same level as other beliefs about the world on the part of the subjects. Hence they are part of the folk-representation itself and must be treated as such; their use in scientific accounts of the belief-action relationship is not implied by data on rational belief systems. It is suggested to conceptualize social representations as integral units of beliefs and action which may be used to explain causally subsequent contingent social events. The function of folk-beliefs in intentional causality for the self-concept of rational people and for social accountability is discussed. (shrink)