Intuitive theories as grammars for causal inference

By Josh Tenenbaum, T. Griffiths, Sourabh A. Niyogi. 2007

105 citations recorded by the source corpus.

This chapter considers a set of questions at the interface of the study of intuitive theories, causal knowledge, and problems of inductive inference. By an intuitive theory, we mean a cognitive structure that in some important ways is analogous to a scientific theory. It is becoming broadly recognized that intuitive theories play essential roles in organizing our most basic knowledge of the world, particularly for causal structures in physical, biological, psychological or social domains (Atran, 1995; Carey, 1985a; Kelley, 1973; McCloskey, 1983; Murphy & Medin, 1985; Nichols & Stich, 2003). A principal function of intuitive theories in these domains is to support the learning of new causal knowledge: generating and constraining people’s hypotheses about possible causal relations, highlighting variables, actions and observations likely to be informative about those hypotheses, and guiding people’s interpretation of the data they observe (Ahn & Kalish, 2000; Pazzani, 1987; Pazzani, Dyer & Flowers, 1986; Waldmann, 1996). Leading accounts of cognitive development argue for the importance of intuitive theories in children’s mental lives and frame the major transitions of cognitive development as instances of theory change (Carey, 1985a; Gopnik & Meltzoff, 1997; Inagaki & Hatano 2002; Wellman & Gelman, 1992). Here we attempt to lay out some prospects for understanding the structure, function, and acquisition of intuitive theories from a rational computational perspective. From this viewpoint, theory-like representations are not just a convenient way of summarizing certain aspects of human knowledge. They provide crucial foundations for successful learning and reasoning, and we want to understand how they do so. With this goal in mind, we focus on

Subfields: Artificial Intelligence, Cultural Studies, Developmental and Educational Psychology

View Paper: https://www.semanticscholar.org/paper/102a4fece64958f03f055ed002358fea9b6fab7f

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Cultural Studies

In Developmental and Educational Psychology

Citations in the corpus

Darwin's mistake: Explaining the discontinuity between human and nonhuman minds (2008, unknown): “These tacit systems of higher-order relations at various levels of generality modulate how human subjects judge and discover novel relations within those domains by a process akin to analogical inference (Goldvarg & Johnson-Laird 2001; Lee & Holyoak 2007; Lien & Cheng 2000; Tenenbaum et al. 2007).”
Goal-directed decision making as probabilistic inference: a computational framework and potential neural correlates. (2012, background): “…centering on probabilistic inference over generative models have figured in numerous other realms, including language (Chater & Manning, 2006; Xu & Tenenbaum, 2007), memory (Hemmer & Steyvers, 2009), conceptual knowledge (Chater & Oaksford, 2008; Griffiths, Steyvers, & Tenenbaum, 2007),…”
Learning a theory of causality. (2011, background): “We have previously proposed that intuitive theories— systems of abstract concepts and laws relating them—can be represented in a “language of thought” which includes aspects of probability and logic (Kemp, Goodman, & Tenenbaum, 2008; Tenenbaum, Griffiths, & Niyogi, 2007; Goodman, Tenenbaum, Griffi”
Intuitive Theories (2020, background): “Probabilistic graphical models do not generalize well beyond the task that they were built for (cf. Gerstenberg & Goodman, 2012; Goodman et al., 2015; Tenenbaum et al., 2007).”
Bayesian models of cognition (2008, unknown): “…graphical models, using probabilistic forms of generative grammars, predicate logic, or other structured representations (Griﬃths, 2005; Griﬃths & Tenenbaum, 2007b; Mansinghka, Kemp, Tenenbaum, & Griﬃths, 2006; Tenenbaum et al., 2006; Tenenbaum, Griﬃths, & Niyogi, 2007; Tenenbaum & Niyogi, 2003).”
Theory-based causal induction. (2009, background): “Following our definition of theories as hypothesis space generators, these higher level theories should generate hypothesis spaces that consist of different theories of particular causal systems (Tenenbaum et al., 2007).”
On the nature and origin of intuitive theories : learning, physics and psychology (2015, background): “Previous applications of grammar-based hierarchical Bayesian models have shown how, given sufficient evidence and a suitable theory grammar, an ideal Bayesian learner can identify appropriate theories in domains such as causality [72, 63], kinship and other social structures [87], and intuitive biol”
Bayesian decision theory in sensorimotor control. (2006, unknown): “Novel Bayesian approaches have started to be able to describe how people make causal inference [ 56 ], a skill that people are particularly good at. In general, Bayesian inference on complicated real world problems are often still proving prohibitively hard from a computational perspective.”