Academic journal article Memory & Cognition

Feature Integration in Natural Language Concepts

Academic journal article Memory & Cognition

Feature Integration in Natural Language Concepts

Article excerpt

Two experiments measured the joint influence of three key sets of semantic features on the frequency with which artifacts (Experiment 1) or plants and creatures (Experiment 2) were categorized in familiar categories. For artifacts, current function outweighed both originally intended function and current appearance. For biological kinds, appearance and behavior, an inner biological function, and appearance and behavior of offspring all had similarly strong effects on categorization. The data were analyzed to determine whether an independent cue model or an interactive model best accounted for how the effects of the three feature sets combined. Feature integration was found to be additive for artifacts but interactive for biological kinds. In keeping with this, membership in contrasting artifact categories tended to be superadditive, indicating overlapping categories, whereas for biological kinds, it was subadditive, indicating conceptual gaps between categories. It is argued that the results underline a key domain difference between artifact and biological concepts.

The primary focus of this article is the probabilistic nature of people's categorization of the world around them and the way in which multiple sources of information are integrated to arrive at categorization decisions. A classic study by McCloskey and Glucksberg (1978) demonstrated that across a range of semantic categories, there was both disagreement and inconsistency in classification. Asked to decide whether a pencil is a tool, some people say "yes" and some say "no." When asked the same question again some weeks later, as many as 30% may change their answer. This uncertainty in categorization may even affect experts, as in the case in which animal behaviorists were found to have no clear concept of behavior (Levitis, Lidicker, & Freund, 2009).

In part, uncertainty in categorization can be attributed to the multiple dimensions that need to be integrated in arriving at a decision. Most natural categories, such as chair or apple, are represented with multiple features. A chair has a characteristic appearance, is made of certain materials, can be used in particular ways, and is created by certain processes. An apple has a visual appearance, taste, and texture and has internal biological processes and causal relations to the apple tree on which it grew, the farmer who chose to plant it, and the potential apple trees that may grow from its seeds. The issue to be addressed here is how these different features are combined in order to arrive at a categorization decision. Suppose that a fruit had the appearance of an apple but was picked from a pear tree. How would people resolve this contradictory evidence in deciding whether it is an apple or whether it is a pear? Or consider the lighthouse/bell tower in the small port of Collioure in southwest France (see Figure 1). Beginning in medieval times as a lighthouse to mark the entrance of the port, in the 18th century, with the demise of the port, it was turned into a bell tower for a chapel. How do people decide in this case what kind of thing it is? Is it still a lighthouse? Is it now a bell tower? When a Tuscan-style cupola was added in the 19th century, giving it a more traditional bell tower appearance, did this change its status further (see http://fr.wikipedia.org/wiki/ Collioure)?

The aim of our experiments was to investigate how people integrate conflicting information in order to classify objects of this kind. In particular, we were interested in whether each feature has a constant additive effect, independent of the others, or whether the effect of one feature varies as a function of the presence or absence of others.

The integration of information in the mind is a broad issue (Anderson, 1981), which has been investigated in a wide range of situations, ranging from perceptual judgments (Ernst & Banks, 2002) to real-life decision making (Dhami & Harries, 2001). In relation to categorization, a number of different models of feature integration have been proposed (e. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.