Categories and induction in young children

Abstract

One of the primary functions of natural kind terms (e.g., tiger, gold) is to support inductive inferences. People expect members of such categories to share important, unforeseen properties, such as internal organs and genetic structure. Moreover, inductions can be made without perceptual support: even when an object does not look much like other members of its category, and even when a property is unobservable. The present work addresses how expectations about natural kinds originate. Young children, with their usual reliance on perceptual appearances and only rudimentary scientific knowledge, might not induce new information within natural kind categories. To test this possibility, category membership was pitted against perceptual similarity in an induction task. For example, children had to decide whether a shark is more likely to breathe as a tropical fish does because both are fish, or as a dolphin does because they look alike. By at least age 4, children can use categories to support inductive inferences even when category membership conflicts with appearances. Moreover, these young children have partially separated out properties that support induction within a category (e.g., means of breathing) from those that are in fact determined by perceptual appearances (such as weight). Since

Necessity and natural categories.

“…But although the similarity-coverage model is consistent with the Gelman and Markman (1986) finding, it doesn't explain the result, as the model provides no reason why indirect transmission should dominate direct transmission in this setting. On an intuitive account, the children are correctly generalizing on the basis of the fact that the property is a biologically relevant one, and brontosauruses and triceratops are in the same lower level biological category (i.e., dinosaurs), whereas rhinos and triceratops are not.…”

“…Because these predicates are unfamiliar, "they are unlikely to evoke beliefs that cause one argument to have more strength than another" (Osherson, 1990, p. 186), and they can therefore isolate the effects of categories on inductive inference. The results by Gelman and Markman (1986) and Heit and Rubinstein (1994), however, suggest that the biological flavor of these predicates leads people to assume that the predicates will generalize according to specifically biological mechanisms rather than according to overall similarity. In some cases, people may assume that these mechanisms run along the lines of taxonomically related species or superordinate categories.…”

“…Training on contrasting properties of subcategories (e.g., setters "help us find birds," whereas samoyeds "help us pull sleds") further restricted the range of the children's generalization. Somewhat older children (4-yearolds) are usually willing to generalize unfamiliar biological properties (e.g., having cold blood) by category, even when perceptual appearance is placed in direct opposition to category membership (Gelman & Markman, 1986). These children, for example, preferred generalizing the property having cold blood from one dinosaur (a pictured brontosaurus) to a second dinosaur (triceratops) over generalizing having warm blood from a rhino to the same triceratops, even though the picture of the The model computes the strength of the conclusion as a weighted average of the direct similarity between cows and bats (direct transmission route) and the similarity between cows and all mammal species (indirect transmission route).…”

Children's evaluative categories and inductive inferences within the domain of food

“…perhaps because the evaluative choices are perceived as being more highly associated with the targets than the non-category choices), it was crucial to demonstrate that children do not use these categories for other types of inferences (see Gelman, 1988;Gelman & Markman, 1986 for discussion). Therefore, two control conditions were included in Study 2.…”

Labeling Bias and Categorical Induction: Generative Aspects of Category Information.

“…It uses exclusively attribute-based similarity to account for inferential processes. The model was developed to account for category-extension tasks; however, it can be ad- 1 There are other tasks that investigate the relationship between category information and inductive judgments with visual stimuli (e.g., Gelman & Markman, 1986;Goldstone, 1994aGoldstone, , 1995Murphy & Ross, 1994;Tajfel, 1970;Tajfel & Wilkes, 1963;Wisniewski & Medin, 1994). A similar examination was introduced elsewhere (Yamauchi & Markman, 2000), so this article focuses on the aforementioned two tasks.…”