Spontaneous assimilation of continuous values and temporal information in causal induction.

Abstract

Existing models of causal induction primarily rely on the contingency between the presence and the absence of a causal candidate and an effect. Yet, classification of observations into these four types of covariation data may not be straightforward because (a) most causal candidates, in real life, are continuous with ambiguous, intermediate values and because (b) effects may unfold after some temporal lag, providing ambiguous contingency information. Although past studies suggested various reasons why ambiguous information may not be used during causal induction, the authors examined whether learners spontaneously use ambiguous information through a process called causal assimilation. In particular, the authors examined whether learners willingly place ambiguous observations into one of the categories relevant to the causal hypothesis, in accordance with their current causal beliefs. In Experiment 1, people's frequency estimates of contingency data reflected that information ambiguous along a continuous quantity dimension was spontaneously categorized and assimilated in a causal induction task. This assimilation process was moderated by the strength of the upheld causal hypothesis (Experiment 2), could alter the overall perception of a causal relationship (Experiment 3), and could occur over temporal sequences (Experiment 4). Keywordscausal reasoning; categorization; ambiguity interpretation; belief assimilation People appear to learn many novel, causal relations without much difficulty: They do not need intensive training to learn that flipping a switch causes a light to be turned on or that eating spoiled food causes stomachaches. Such apparently straightforward causal learning is actually impressive, given that the environment does not always present a learner with neatly organized data. In the current set of studies, we examine one of the difficulties people have to face as a result; namely, that observations may not come readily classified as evidence for or against a causal hypothesis.Let us illustrate this problem in the context of covariation-based models of causal induction. The focus of these models has been on how people use covariation data to learn causal relations (e.g., Cheng, 1997;Cheng & Novick, 1990;Collins & Shanks, 2002;Jenkins & Ward, 1965;Rescorla & Wagner, 1972;Shanks, Lopez, Darby, & Dickinson, 1996). In these models, Correspondence concerning this article should be addressed to Jessecae K. Marsh, Department of Psychology, Box 2051, Psychology Building, Texas Tech University, Lubbock, TX 79409-2051. jessecae.marsh@ttu.edu. This article was submitted as part of Jessecae K. Marsh's dissertation project. A discussion of Experiments 1 and 2 can be found in a section of the book chapter "Dynamic Interpretations of Covariation Data" in the edited book Causal Learning: Psychology, Philosophy, and Computation (2007). NIH Public Access NIH-PA Author ManuscriptNIH-PA Author Manuscript NIH-PA Author Manuscript covariation data are preclassified into four events: What happens to an effect w...

Context and Time in Causal Learning: Contingency and Mood Dependent Effects

“…However, these models hold that knowledge of time, in terms of contiguity and succession, is a starting point where causal knowledge is used to generate initial causal models or hypotheses about a given situation, with contingency information used secondarily to test causal hypotheses [37], [38] or assess the strength of the causal relationship [11], [12], [13]. Importantly, the assumption of the existence of causal knowledge can, quite naturally, explain data which associative models find difficult to incorporate.…”

The influence of perceived causation on judgments of time: an integrative review and implications for decision-making

“…But it may have implications for causal inference more generally. Marsh and Ahn (2009) noted that people sometimes may need or want to link ambiguous events as causes and effects. The authors suggested that time compression may be one mechanism that enables this to occur.…”

Causal Structure Learning in Continuous Systems

“…Given the ubiquity of such systems, continuous variables have received surprisingly little attention in the study of causal cognition. In a reanalysis of data from Marsh and Ahn (2009) and a novel experiment, Pacer and Griffiths (2011) showed that people are capable of learning individual cause-effect relationships between continuous variables. Soo and Rottman (2018) investigated causal relations in non-stationary time series, i.e., those where long term trends affect the average values of the variables in ways that obscure and complicate the causal relations between those variables.…”