Portal de Periódicos da CAPES

The prediction and explanation of criminal violence

1995; Elsevier BV; Volume: 18; Issue: 2 Linguagem: Inglês

10.1016/0160-2527(95)00001-x

1873-6386

Vernon L. Quinsey,

Mental Health and Psychiatry

Scientific explanations of behavior involve a specification of both antecedent events or initial conditions and general laws that describe how each situation causes the succeeding one. Thus, laws acting on the current state of a system endlessly produce new states (Simon, 1992). To logical positivists, explanations are deductions about particular phenomena made from universal laws. Thus, in a formal sense, predictions and explanations differ only in that, in the former, future events instead of past events are at issue (cf. Blackburn, 1993). However, prediction can also be based purely upon induction; in this case, prediction does not entail explanation. For example, we can predict on inductive grounds that the sun will rise in the morning without knowing anything about the celestial mechanics that can explain its apparent rise. Similarly, we can predict on inductive grounds that someone who has committed many assaults on persons will commit another (Walker, Hammond, & Steer, 1967), but not be able to explain why. Explanations, therefore, involve specifying a causal mechanism, whereas predictions may be based on pure induction. Intellectually satisfying explanations must involve a causal mechanism that can be imagined to produce the effect. Theoretical models are explanations of this kind in which the initial conditions and general laws are quantified. An illuminating discussion of these issues has been provided by Einhorn and Hogarth (1986), who argue that scientists and lay people use similar systematic rules for assessing the causal status of variables. A variable is likely to be perceived as the cause of some effect to the degree to which it (a) stands out as a difference among the variety of other potential causes, (b) covaries with the effect, (c) precedes the effect, (d) is contiguous with the effect in time and space, (e) is similar to the effect in terms of duration, magnitude, and physical characteristics, (f) can be linked to the effect by a causal chain or mechanical