Concreteness and imageability differentially predict judgments of manual and visual similarity

How we understand language is fundamentally shaped by our interactions with the world around us, according to modern theories of cognitive embodiment and predictive inference in perception. Sensorimotor experiences associated with specific words are reflected in those words’ levels of concreteness and imageability, psycholinguistic constructs that are meant to capture the sensorimotor tangibility of word referents. Here, we investigate the relationship between word concreteness, imageability, and judgments of manual and visual similarity. Rather than examining features of individual words, we examined responses to word pairs consisting of German object nouns. Participants rated both the manual and visual similarity of each object word pair’s referents. The ratings were correlated with intra-pair concreteness and imageability distances, based on norms previously acquired using a supervised learning algorithm (Köper & Schulte im Walde, 2016). Smaller concreteness distances between words in each pair correlated exclusively with higher manual similarity ratings, and smaller imageability distances correlated only with higher visual similarity ratings. This dissociation provides new evidence that concreteness and imageability can be thought of as distinct – albeit somewhat related – features associated with different modalities. Whereas concreteness distance was more closely linked with manual object interactions, imageability distance was found to be more strongly associated with visual similarity. These findings may be of use to researchers designing stimulus sets for studies involving word pairs, such as linguistic priming or translation studies.