Imagery, Context Availabilty, Contextual Constraint and Abstractness
Katja Wiemer-Hastings, Jan Krug, Xu Xu · eScholarship (California Digital Library) · 2001
Imagery, Context Availability, Contextual Constraint and Abstractness Katja Wiemer-Hastings (K ATJA @ NIU . EDU ) Jan Krug ( JKDKRUG @ YAHOO . COM ) Xu Xu ( Z 015504@ STUDENTS . NIU . EDU ) Department of Psychology, Northern Illinois University, DeKalb, IL 60115, USA Abstract A constraint-based theory of abstractness was investigated according to which abstractness of entities is a function of (i) perceptual observability and (ii) characteristics of contextual constraints. Participants performed ratings of context availability, imagery, and abstractness for 36 nouns that varied in abstractness and familiarity. The ratings were used to compare the predictions of abstractness ratings by context availability, dual coding theory and the constraint-based approach outlined in this paper. We found that only constraints explain variation of perceived abstractness for abstract concepts, whereas context availability and imagery are good predictors of the dichotomous distinction of concrete-abstract, and of variations of concreteness for concrete concepts only. A second study shows that introspection-based constraints are most critical for abstractness ratings. Implications are discussed. Abstractness Every-day communication is pervaded by references to abstract entities, such as explanation, regret, and intention. Typically, we think of an entity as abstract when it cannot be perceived. However, there are no clear-cut criteria for what makes entities abstract or concrete. Several theoretical approaches exist to predicting perceived abstractness. This paper compares three theories: dual-coding theory, context-availability theory, and our approach, called the contextual constraint theory. We propose that perceived abstractness depends on two factors. First, entities are abstract or concrete, depending on whether they are physical in nature (i.e., perceivable through vision, touch, etc.). Second, within these groups, abstractness varies according to more specific types of information. Together, we call this the two-factor model of abstractness. We will start by reviewing the plausibility of the dichotomy of abstract and concrete, as proposed by the first factor. The remainder of this paper will address the factors underlying abstractness variation within the groups of abstract versus concrete entities. Abstract and Continuum? Concrete: Dichotomy or Concrete and abstract nouns are commonly defined by reference to perceivability: Concrete entities are considered to be physical entities with characteristic shapes, parts, materials, etc., whereas abstract entities lack physical attributes (e.g., Crystal, 1995). The first proposed factor follows this broad distinction. Some entities challenge the notion of a dichotomy of abstract and concrete entities. Examples for entities that cannot clearly be classified as abstract or concrete are government, officer, or anger. A government is abstract in that we cannot really point to who or what it is, but it is also concrete in that it involves a number of specific, concrete entities, such as people, buildings, and particular locations. Officer is a social agent term, referring to concrete individuals with characteristics defined by a particular social role or profession. Their roles are not obvious characteristics, but are inferred from more complex information, such as behavior patterns in specific situations. Finally, emotion terms such as anger are a special group of entities. Emotions can be perceived within individuals who experience them. Outwardly, we can perceive emotion through nonverbal and verbal behavior. Still, emotions are qualitatively different from concrete entities such as cups and office chairs. In fact, they have been proposed to constitute a distinct group from both concrete and abstract entities (Altarriba, Bauer, & Benvenuto, 1999). The alternative view suggested by these challenges is a contiuum view, according to which all entities vary in concreteness, and the distinction of abstract versus concrete is an oversimplification. A simple way to test both views is to ask people to rate the concreteness of a large sample of entities, including abstract and concrete ones. If concreteness is one dimension and all entities vary along this dimension, then ratings should be distributed pretty evenly across the entire scale. In contrast, if abstract and concrete entities were two distinct classes of entities, ratings should fall into two clusters. There would be a lot of entities rated as abstract, versus a lot of entities rated as concrete. That is, the distribution of