Predictive and diagnostic learning within causal models: Asymmetries in cue competition.

Abstract

Several researchers have recently claimed that higher order types of learning, such as categorization and causal induction, can be reduced to lower order associative learning. These claims are based in part on reports of cue competition in higher order learning, apparently analogous to blocking in classical conditioning. Three experiments are reported in which subjects had to learn to respond on the basis of cues that were defined either as possible causes of a common effect (predictive learning) or as possible effects of a common cause (diagnostic learning). The results indicate that diagnostic and predictive reasoning, far from being identical as predicted by associationistic models, are not even symmetrical. Although cue competition occurs among multiple possible causes during predictive learning, multiple possible effects need not compete during diagnostic learning. The results favor a causal-model theory.

Causal models and the acquisition of category structure.

“…In that way possible content effects could be maximally controlled because the postulated causal models are manipulated by instructions. Waldmann and Holyoak (1992) presented an experiment (Experiment 3) that met these criteria but that did not manipulate category structure. In the present Experiment 4 we replicated the design of Experiments 1 and 3 using materials that controlled for content domain.…”

“…One response to the present results, as well as to other recent evidence that causal directionality influences learning (Waldmann & Holyoak, 1992), is to develop network models in which the causal interpretation of events guides the assignment of information to layers of the network. That is, cues that are interpreted as "causes" could be assigned to the "input" layer, and cues that are interpreted as "effects" could be assigned to the "output" layer, regardless of the temporal order in which the cues are overtly presented (see van Hamme, Kao, & Wasserman, 1993).…”

“…A fundamental assumption of causal-model theory is that, regardless of the temporal order in which we receive causal information, the underlying mental representation of the situation honors the cause-toeffect direction. The preference of people to learn in the cause-to-effect direction has been demonstrated in a number of studies (e.g., Eddy, 1982;Tversky & Kahneman, 1980;Waldmann & Holyoak, 1992; see also Einhorn & Hogarth, 1986).…”

An intelligent problem solving environment for designing explanation models and for diagnostic reasoning in probabilistic domains

“…Reviews of published case studies in the domain of environmental medicine support this hypothesis. More generally, there is empirical evidence that qualitative reasoning by (in)dependencies as supported by Bayesian networks (like for example "explaining away") corresponds closely to human reasoning patterns (Henrion, 1987;Pearl, 1993;Waldmann & Holyoak, 1992). Alternative approaches.…”

Learning of contingent relationships

“…Waldmann and his collaborators (e.g., Waldmann, 2000;Waldmann & Holyoak, 1992) have argued that judgments depend on the set of assumptions about the nature of the cues and outcomes-the causal model-adopted by the observer. One such model stipulates that multiple causes interact both when they precede and when they follow a common effect.…”