The Developmental Origins of the Formal Structure of Kind Representations

Abstract

There exists an intellectual thread that traces back to Aristotle, can be seen in the works of many of the Enlightenment philosophers, especially the Cambridge Platonists, and extends into contemporary philosophy and empirical psychology. The core idea of this intellectual thread is that concepts like dog, hammer, river, city, and triangle possess a central form. This central form represents the properties that capture the core identity of these kind concepts, what it means to be each of these kind of things (e.g., the properties of barking for a dog, having a handle for a hammer, having a source for a river, having citizens for a city, and having three sides for a triangle). This form does not determine necessary and sufficient conditions for referring to entities in the world (e.g., we can encounter a dog that cannot bark, or a city without citizens, and still think about these entities as dogs and cities, respectively). Instead, the form alters the way we think about the properties that are part of it, licensing a range of conceptual and linguistic consequences for those properties. The present thesis extends this thread, especially more recent approaches within contemporary cognitive science (e.g., Prasada, 2000; Pustejovsky, 1986). In the first paper, two experiments investigated whether we see reflections of this form in familiar kind representations throughout conceptual development. These experiments provide evidence that by age four, children represent some properties as part of the form of familiar kind representations like dog and plate, as part of their core identity, and that these properties possess distinct conceptual consequences not licensed by other properties that are construed as true of those kinds, but which are not part of the form of those concepts (e.g., wearing a collar for a dog, being white for a plate). In the second paper, we asked whether the capacity to configure the form of a kind representation is part of the kind acquisition machinery that allows humans to acquire thousands of new kind representations throughout conceptual development, often from only a single encounter with a novel entity. These experiments demonstrate that when adults are introduced to a property of a novel kind with evidence that the property is part of the form of the representation, then they immediately infer a conceptual consequence of that form. If they are introduced to the same property with evidence that it is merely true of the kind, but not part of its form, they do not infer this consequence. These results provide evidence that one part of the generative kind acquisition machinery possessed by human adults is a capacity to represent properties as part of the form of the representation. The third paper extends this finding and provides evidence that children by age four can represent a property as part of the form of a kind representation from only a single encounter with a novel entity. Together, these papers provide further evidence that human kind representations possess a central form. They provide the first evidence that this structure is part of our early emerging kind acquisition machinery. We conclude that this structure is an important and early emerging part of the formal structure of kind representations, the structure that is intrinsic to this type of mental representation, distinguishing it from the other mental representations the mind possesses (e.g., representations of sets, representations of logical operators, representations of properties, and so on).