Academic journal article Memory & Cognition

Causal Explanation in the Face of Contradiction

Academic journal article Memory & Cognition

Causal Explanation in the Face of Contradiction

Article excerpt

Published online: 14 January 2014

© Psychonomic Society, Inc. 2014

Abstract Previous work has shown that predictions can be mediated by mechanistic beliefs. The present study shows that such mediation only occurs in the face of contradictory, and not corroborative, evidence. In four experiments, we presented participants with causal statements describing a common-cause structure (E^sub 1^[arrow left]C[arrow right]E^sub 2^). Then we informed them of the states of C and E^sub 1^ and asked them to judge the likelihood of E^sub 2^. In Experiments 1 and 2, we manipulated whether the mechanisms supporting the two effects were the same or different, and whether the evidence presented confirmed or contradicted the participants' expectations. The relation between the mechanisms only influenced predictions when evidence contradicted the expectations, but not when it was consistent. In Experiments 3 and 4, we used a common-cause structure with identical mechanisms. We manipulated the order in which predictions were made. When confirmatory predictions were made before contradictory predictions, mechanistic modulation was not observed in the confirmatory case. In contrast, the modulation was found when confirmatory predictions were made after contradictory ones. The results support the contradiction hypothesis that causal structure is revised during prediction, but only in the face of unexpected evidence.

Keywords Belief update · Causal Bayes nets · Screening-offrule · Mechanism · Contradiction · Rational analysis

In the face of new evidence, beliefs need to be updated. When those beliefs concern causal relations-how things work-a reasoner must decide whether to make a small change to the strength of his or her beliefs or to make a more radical change to the structure of his or her belief system. To illustrate, imagine that my causal model tells me to expect that the economy is going to improve, and therefore that the value of my house is going to rise. Unfortunately, its value does not rise. To accommodate the new evidence, one option is to change my probability judgments, either my judgment that the economy is going to improve or the conditional probabil- ity that my house value will rise if the economy improves. We refer to this kind of parametric property of a causal model as strength of belief. Another option is to reformulate my beliefs about what causes what. For instance, I might decide that house values only rise in neighborhoods without too many foreclosures, even when the economy is improving. This is a more structural change to belief, in that I have introduced a newvariableintomycausalmodel.1

In this article, we will argue that people change their beliefs about causal structure rather than the strength of causal rela- tions in the face of contradictory evidence. In particular, we predict that people are more likely to modify the structure of their causal model in the face of evidence that contradicts their expectations than in the face of evidence that is consistent with them. Although this behavior may appear quite rational, we will argue that it is not predicted by theories whose central concept is that we maintain beliefs in such a way that our expectations converge to the true probabilities of events (Glymour 2001; Griffiths, Kemp, and Tenenbaum 2008; Oaksford and Chater 2001).

We explored this question by asking adults to make pre- dictions on the basis of limited causal knowledge. We first briefly review causal Bayes nets theory, a normative account of how causal knowledge guides reasoning (Griffiths and Tenenbaum 2009). We also review previous studies showing that people revise their causal structure by adding mechanistic knowledge to make predictions. Then we present the contra- diction hypothesis, the idea that contradictory evidence leads to structural update using mechanistic knowledge. After pre- senting four experiments that support the contradiction hy- pothesis, we discuss the implications of the results. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.