9.94 The Cognitive Science of Intuitive Theories Quiz Please return your answers by email to Josh Tenenbaum (jbt@mit.edu), by February 2, 2006, to receive credit for 9.94. Please send your answers in plain text, rather than an attachment in Word, pdf or other enhanced formats. The policies for this quiz are explained on the class website: http://www.mit.edu/~jbt/994.html. We will assume that you have read those policies before proceeding any further with this quiz. The quiz consists of five questions. You need only answer three, according to the constraints indicated below. We expect you to spend around two to three hours on this quiz, enough time to put some thought into each question and write coherent answers, but not to write a whole term paper. Of course, the exact amount of time you spend is up to you. We hope you learn something from and enjoy thinking about these issues. Good luck! Please answer *either* Question 1a or Question 1b. Question 1a. Compare two kinds of "theories" that Josh Tenenbaum discussed in his lecture: (1) the visual system's theory of the world that supports lightness perception or the recovery of three-dimensional shape from two-dimensional images; (2) the more cognitive theory of how biological species and properties are structured, which supports inductive reasoning about new properties and new species. Does it make sense to call both of these "theories"? Is it valuable to think about these two systems of knowledge as being the same kind of thing? Is it valuable to think about them both as analogous to scientific theories? Take a position pro or con, but regardless of which position you argue for, introduce at least some plausible arguments for both sides. Question 1b. Intuitive psychology, as Rebecca Saxe discussed in her lecture, and intuitive physics, are probably the two best-studied domains of intuitive theorizing. Much of what cognitive scientists believe is generally true about intuitive theories come from studying one or both of these theories. How is an intuitive theory of mind like or not like an intuitive theory of physics? Describe at least two similarities, and at least one difference. If you had to choose one position, would you say they are basically the same kind of knowledge representation, or fundamentally and importantly different? Question 2. Throughout the course, you've been presented with several examples of intuitive theories, such as folk psychology (or Theory of Mind), folk physics, and folk biology. Propose a new domain about which people might have an inuitive theory, and explain why it's a plausible candidate. In particular, why is your proposed domain likely to be governed by a theory rather than a different kind of knowledge structure, like a memorized list of examples or a set of associations? Please answer *either* Question 3a or Question 3b. Question 3a. In this class you've seen examples of intuitive theories that effectively subserve predictions, guide interventions, and generate explanations. However, you've also seen cases in which theories can lead us astray. Describe two ways in which theories can lead to incorrect generalizations of some kind (e.g. incorrect predictions, inaccurate probability estimates, useless actions). Do these errors suggest a fundamental problem with theories, or are they more like "optical illusions," occasional errors that result from a generally effective procedure? Would it be possible to have theories that never lead to errors of any kind? Question 3b. In her lecture, Laura Schulz proposed a 'spiral' of cognitive development going from action to evidence to theories and back around again. Josh asked whether such a 'spiral' would converge on the truth about the world. He suggested that nothing in the spiral was sufficiently constrained to guarantee that it would, most of the time, get most things right. Do you agree or disagree? If you think the spiral will work, why? If you think it is insufficiently constrained, what other constraints might you want to add?