In writing about causality, I feel like a lamb going out among wolves. The philosophers have argued for centuries about what causation means and when we are justified in making causal attributions. Modern physics has, in effect, eliminated causation and replaced it by sets of constraint equations, although there are forever efforts to introduce causation back into physical analyses (e.g., DeKleer & Brown, 1984). While cognitive psychologists
busied themselves elsewhere, social psychologists have developed a sophisticated body of data and theory about causal attribution in person perception (e.g., Hilton & Slugoski, 1986; Jaspars, Hewstone, & Fincham, 1983; Kelley, 1973, 1983; Orvis, Cunningham, & Kelley, 1975). Developmental psychologists seem to have been the only ones to have given much thought to the issue of causal attribution in cognition, led, as always, by Piaget (1974). For a long time, the one major piece of research relating to adult perception of causation in a nonsocial setting has been the classic work of Michotte (1946). Causal relationships do play an important part in modern theories of text processing (e.g., Trabasso, Secco, & van den Broek, 1984). Here, the emphasis is how a causal rule such as "lightning causes fire" is used to understand a text. However, there has been little concern in cognitive psychology with how such rules are inferred in the first place. This is the principal concern of the present chapter.