Sensitivity to Purely Taxonomic Associations among Young Children

The goal of the present study was to examine an aspect of semantic development. Previous research investigated whether different types of relations between entities influence the degree to which children perceive them as similar. The purpose of the present study was to compliment this research by investigating whether the number of relations between entities influences their perceived similarity. The results showed that preschool-aged children perceive entities related along both single and multiple dimensions as more similar than unrelated entities, and entities related along multiple dimensions as more similar than those related along a single dimension. Therefore, by four years of age children are sensitive to the number of relations linking entities. Taxonomic Association Sensitivity 3 Introduction To function in the world, one must develop knowledge, a foundation of information with which one can formulate thoughts, actions, and behaviors. Semantic knowledge, an understanding of concepts and their meanings, is a fundamental aspect of cognitive development. As children grow and develop, their ability to observe and understand objects and events increases, their experiences in the world helping to establish a foundation of knowledge (e.g., Fisher, Godwin, Matlen, & Unger, in press; Tversky, 1985). It has been suggested that as children develop, they go from using perceptual features as a means of understanding similarities between stimuli to making inferences based upon previous knowledge, expediting the process of observation and understanding as semantic knowledge grows (Tversky, 1985). Knowledge of concepts ultimately allows people to go from simply observing the world around them to making inferences based on prior experience and knowledge, increasing mental efficiency and efficacy. As children observe and learn about the world around them, they build associations among concepts. A large body of previous research has investigated children’s sensitivity to different types of associations between concepts. Taxonomic and thematic associations have been identified as two of the fundamental forms of relation. Taxonomic relations pair together members of similar kind with shared properties, like “dog” and “lion” both being mammals (e.g., Blaye, Bernard-Peyron, Paour, & Bonthoux, 2006). Thematic relations, on the other hand, group together stimuli based on whether entities appear together or are associated with the same environment (e.g. “turkey” and “cow” are animals that are both commonly found on farms) (e.g., Tversky, 1985; Blaye et Taxonomic Association Sensitivity 4 al., 2006). Though concepts that are thematically related may share few to no perceptual properties, they are united by having a complementary relation such as occupying the same space (Lucariello, Kyratzis, & Nelson, 1992). As such, thematic associations are more rooted in ad hoc observation, while taxonomic associations rely more on acquired knowledge. The primary focus in the field of investigating children’s understating of relationships has been to determine whether the types of relationships to which children are sensitive changes with age. A number of studies have demonstrated that children’s preference for taxonomic relations over other types of relations increases with age (e.g., Blaye et al., 2006; Lin & Murphy, 2001; Nguyen, 2007; Tversky, 1985). For example, Tversky (1985) conducted a study with children three, four, six, and eight years of age. This study used a match to sample paradigm in which children were asked to judge which of two test items should be grouped with a target item, and then to justify their selection. The test items included an option that was perceptually related to the target, and an option that was taxonomically related to the target. For instance, children might be presented with “snake” as the target, and be asked to choose whether it goes with a “hose” (perceptual relationship choice) or a “lizard” (taxonomic relationship choice). The results showed that children younger than six years of age, were at chance at choosing taxonomic or perceptual matches; however, starting at age six children showed a preference for taxonomic relations. Children’s ability to justify their choices (both perceptual and taxonomic) also increased markedly between three and eight years of age. Therefore, it appears that preferences for linking entities according to taxonomic relationships increase with age. This increased preference for taxonomic relationships is accompanied by Taxonomic Association Sensitivity 5 increasingly rich representations of taxonomic relations, as revealed by the more detailed and appropriate justifications for taxonomic choices that children provided with increasing age. However, most studies in the literature have used forced choice tasks to assess children’s lack of preference for taxonomic versus thematic relations. Although these prior studies shed light on children’s sensitivity to different types of relationships, entities in the real world are often linked by multiple relationships simultaneously. For instance, lions and tigers are both taxonomically and thematically related to each other. However, research to date has not investigated children’s sensitivity to the number of relationships between entities. Moreover, the match to sample paradigm used in most prior studies is inappropriate for investigating this question. Specifically, this task only allows participants to indicate whether a given test item is or is not related to the target. To determine children’s sensitivity to the number of relationships between entities, participants must be able to provide judgments of relatedness on a graded scale. The purpose of the present study is to investigate the influence of number of relationships between entities on children’s judgments of similarity between entities, using a paradigm that affords a graded measurement of perceived similarity. The Semantic Space Task, created in the Cognitive Development Lab at Carnegie Mellon University, removes the forced-choice component from methodology (Fisher, Godwin, Matlen, and Unger (in press) and makes it possible to collect graded judgments of semantic similarity. The Semantic Space task is presented to participants as a game called the Help Zibbo Game, in order to assess how children organize semantic knowledge about entities based on relationships between them. Researchers presented Taxonomic Association Sensitivity 6 each participant with two neutral-colored blocks, each representing a member of an animal dyad. The experimenter would place one square on the game board, to serve as a target animal, the block introduced as an animal (e.g., a crocodile). The experimenter then handed the second block to the child while posing the question: “If the crocodile goes here, where do you think the alligator should go?” Each participant was instructed to put animals of the same kind close together on a 9 x 9 square grid game board. The experimenter removed the two blocks from the game board after each trial and used the blocks to represent a different dyad of animals on the next trial. Participants’ responses were scored as the distance between the two blocks. The goal of the Fisher et al. (in press) study was to examine whether children distinguish between highly semantically similar dyads and dissimilar dyads. Therefore, in that study participants were presented with dyads of animals that were related both thematically and taxonomically (e.g., lamb-sheep). The goal of the present study is to examine whether preschool-age children are sensitive to the number of relations linking different concepts. Therefore, in the present study participants were presented with dyads of animals that either were related only taxonomically or both taxonomically and