The model of the attribution process that Kelley developed between 1967 and 1973 still stands as the most thoroughgoing attempt to systematize findings about the manner in which people arrive at explanations of their own and others' behavior. But Kelley (1967) himself could only list the ways in which real explanations might deviate from his idealized model and did not try to integrate the various biases that he and Heider (1958) before him had recognized as important aspects of the explanations that people actually give. Apart from its inability to do the job for which it was designed, Kelley's model has certain fundamental internal weaknesses. It is extremely vague on crucial points. One omission that has attracted criticism is the failure to address the question of hypothesis generation.