Be the wapple of my eye: predicting the sensorimotor pattern of novel words from language-based representations.

Embodied theories of cognition propose that conceptual representations are grounded in sensorimotor experience. However, sensorimotor experiences can be activated even for referents that we have never directly experienced, such as abstract concepts. In the present study, we investigate whether language can predict sensorimotor experiences for unfamiliar novel words, for which a referent is by definition unavailable. Indeed, recent studies provide evidence for activation in semantic memory elicited by nonwords, which have long been considered devoid of any meaning. Here, we question whether this activation could also imply that nonwords evoke experiential features across various sensorimotor modalities. To investigate this, we trained a model aimed at predicting the sensorimotor patterns elicited by nonwords on the basis of their estimated semantic activation. Such mapping model was trained on a set of 39,707 English words, for which semantic vector representations were obtained using fastText, a Distributional Semantic Model that can also be used to approximate the meaning of nonwords via their sub-lexical units. Sensorimotor norms were obtained from the Lancaster Sensorimotor Norms (Lynott et al., 2020). Then, eleven separate ridge regression models were trained to predict individual ratings across six perceptual modalities (vision, touch, hearing, smell, taste, and interoception) and five action effectors (mouth/throat, hand/arm, foot/leg, head, and torso). Sensorimotor ratings for 27,136 nonwords, retrieved from the British Lexicon Project (BLP , Keuleers et al., 2012), were predicted by feeding the model their corresponding fastText embeddings. Moving from these predictions, perceptual strength measures (i.e., Maximum Perceptual Strength and Minkowski 3 distance) were computed for each nonword and considered as predictors of rejection times to the nonwords in the BLP lexical decision task. Results show that Minkowski 3 significantly predicted responses, implying that nonwords with higher estimated sensorimotor strength take longer to be rejected. This finding suggests that nonwords can be associated to experiential traces via the distributional properties of their sub-lexical units. We provide a first data-driven insight into the possible sensorimotor features of nonwords, challenging the assumption whereby they lack semantic content and suggesting instead that they can evoke experiential features. Further research is needed to investigate the mechanisms through which nonwords may be grounded, in order to understand which source of information may play a role when assigning perceptual features to novel words.