Table 1.

Summary of concrete versus abstract word processing models, with their basic tenets, predictions, and supporting research.

Theory	Basic tenets	Predictions regarding concrete versus abstract word processing	Empirical support for predictions
Dual Coding Theory (Paivio, 1971)	• Concrete words are represented by linguistic and imagistic codes; abstract words are only represented by a linguistic code.	• Concrete words should be processed faster than abstract words.	Reviewed e.g., Paivio (1991)
Context Availability Theory (Schwanenflugel and Shoben, 1983)	• Concrete words are associated with stronger and denser associations to contextual information compared to abstract words.	• Concrete words should be processed faster when presented in isolation. • There should be no difference between concrete and abstract word RTs when context is provided.	Reviewed, e.g., Schwanenflugel (1991)
Qualitatively Different Representational Hypothesis (Crutch and Warrington, 2005)	• Concrete words are primarily organized by semantic similarity (i.e., same category, similar features), whereas abstract words are primarily organized by semantic association (i.e., shared linguistic context or ‘real life’ associations).	• When processing concrete words, similarity-based connections are identified faster than association-based connections • When processing abstract words, association-based connections are identified faster than similarity-based connections	Crutch et al. (2009)
Perceptual Symbol Systems (Barsalou, 1999)	• Both concrete and abstract word processing involves simulation of sensorimotor experiences (i.e., perceptual symbols) associated with a given concept. • Concrete and abstract words differ in the content of these simulations. Introspective, social, and event knowledge is central to abstract simulations, and object knowledge is central to concrete simulations.	• Human generated properties for concrete and abstract concepts will vary in content. • Concrete words should elicit primarily object-related properties, while abstract words should elicit introspective, social, and event-related properties	Barsalou and Wiemer-Hastings (2005) Wiemer-Hastings and Xu (2005)
Hub-and-Spoke Model (Rogers et al., 2004; Lambon Ralph et al., 2007; Patterson et al., 2007)	• The anterior temporal lobes bilaterally serve as a central amodal hub for semantic knowledge by integrating knowledge from amodal cortical areas	• Damage to the anterior temporal lobes should impair knowledge for both concrete and abstract words	Pobric et al. (2007, 2009), Hoffman and Lambon Ralph (2011)
Theory of Embodied Abstract Semantics (Vigliocco et al., 2009)	• Both concrete and abstract words are composed of embodied/experiential (sensorimotor, affective) and linguistic associative information. Concrete words are primarily composed of sensorimotor information. Abstract words are primarily composed of emotional and linguistic information.	• When concrete and abstract words are controlled for sensorimotor information, there should be an advantage for abstract words. Affective associations should account for this abstract word advantage.	Kousta et al. (2011)