We welcome Goldin-Meadow & Brentari's (G-M&B's) view that communication mode influences the structure of language, and we suggest that it might spread its influence to cognitive processes beyond language.

Accumulating evidence indicates that, within the population of deaf peoples, there are differences with respect to the frequency of use of a signed mode of communication. The preference for a signed or an oral modality to communicate might expand to language-related tasks apparently not directly linked to social communication – for example, on the processing of written language and visual word recognition. Barca et al. (Reference Barca, Pezzulo, Castrataro, Rinaldi and Caselli2013) compared the performance of young adults with different communication modes (i.e., deaf individuals with a preference for signed language, and deaf individuals using an oral modality and with less or no competence in signed language) in a visual lexical decision task. The lexicality decision task is an established paradigm in psycholinguistics and consists of the speeded categorization (with key-press responses) of visually presented real words and nonwords. It is typically used to assess the ease of access and retrieval of stored lexical information from memory, with the basic finding of faster and more accurate processing of words than of nonwords and pseudowords (Balota & Chumbley Reference Balota and Chumbley1984; Barca & Pezzulo Reference Barca and Pezzulo2012; Reference Barca and Pezzulo2015; Ratcliff et al. Reference Ratcliff, Gomez and McKoon2004).

In recruiting participants for the study, Barca et al. (Reference Barca, Pezzulo, Castrataro, Rinaldi and Caselli2013) put forth an effort in controlling for individual factors heavily affecting performance in linguistic and cognitive tasks, such as the age of deafness diagnosis, the degree of hearing impairment, the age of exposure to sign language, and the preferred language modality. Deaf individuals were either those deaf individuals who communicate using mainly sign language, which they learned since infancy (before 3 years of age) in a natural environment, or deaf with a preference for spoken language (learned via formal instruction) and were poorly proficient in sign language. Results illustrate that different recognition strategies are in play between these groups, because the lexicality effect was present only in deaf individuals using mainly sign language to communicate. In a subgroup of participants (Napolitano et al. Reference Napolitano, Andellini, Cannatà, Randisi, Bernardi, Castrataro, Pezzulo, Rinaldi, Caselli and Barca2014), communication mode appears also to shape functional connectivity of cerebral networks related to language and cognitive processes (Laird et al. Reference Laird, Fox, Eickhoff, Turner, Ray, McKay and Fox2011; Smith et al. Reference Smith, Fox, Miller, Glahn, Fox, Mackay and Beckmann2009), with increased activity in intrinsic connectivity among deaf signers compared to deaf individuals using oral language in an auditory network and a fronto-parietal network.

These findings converge within an embodied framework that sees action and language networks as deeply interconnected (Barca et al. Reference Barca, Cornelissen, Simpson, Urooj, Woods and Ellis2011; Barca & Pezzulo Reference Barca and Pezzulo2012; Pezzulo et al. Reference Pezzulo, Barca and D'Ausilio2014; Pulvermüller et al. Reference Pulvermüller, Huss, Kherif, Moscoso del Prado Martin, Hauk and Shtyrov2006; Pulvermüller & Fadiga Reference Pulvermüller and Fadiga2010; Willems & Hagoort Reference Willems and Hagoort2007). In such a perspective, words are not represented or processed in amodal format. The same neuronal networks are in play supporting perception and action performance (Barsalou Reference Barsalou1999) – which would also explain why, for example, the processing of action verbs like “lick,” “pick,” or “kick” activates motor regions that mediate the execution of the corresponding actions (Hauk et al. Reference Hauk, Johnsrude and Pulvermüller2004; Meteyard et al. Reference Meteyard, Zokaei, Bahrami and Vigliocco2008). This implies that language modality (oral or signed) deeply influences the way linguistic representations are formed and processed. Specifically, the use of a sign language since infancy might shape a person's motor and language neural circuits resulting in specific visuo-motor representations of words that are different from (and perhaps richer than) those of oral speakers. For example, creating “signed-based representations” that include the movements of the hands and other body parts, which in turn would enhance the involvement of cerebral regions related to the coding of motor acts and praxis information (e.g., inferior parietal lobule) (Corina et al. Reference Corina, Chiu, Knapp, Greenwald, San Jose-Robertson and Braun2007; Pobric et al. Reference Pobric, Jefferies and Lambon Ralph2010). Such visuo-motor representations might be elicited automatically during online linguistic processing or rehearsal and overlap with the perceptual-motor processes required to execute the task (in the case of Barca et al. Reference Barca, Pezzulo, Castrataro, Rinaldi and Caselli2013 study, the hand movements required to click a response button), thus potentially producing interference or facilitation effects.

From the point of view of a predictive coding and active inference framework, language makes use of generative models to predict how linguistic stimuli unfold in time: for instance, sequences of speech sounds and the proprioceptive states and motor commands required to utter them, or sequences of eye movements required to read a text (Donnarumma et al. Reference Donnarumma, Costantini, Ambrosini, Friston and Pezzulo2017; Friston & Frith Reference Friston and Frith2015; Lupyan & Clark Reference Lupyan and Clark2015; Pezzulo et al. Reference Pezzulo, Rigoli and Friston2015). In signers, such generative models might predict the unfolding of hand movements and other (perceptual and proprioceptive) content associated with signs and gesture, or the eye movements required to track them – and this is why these latter visuo-motor predictions would be automatically elicited during language processing and rehearsal. What's more, given the established role of language in modulating various aspects of (nonlinguistic) cognition (Carruthers Reference Carruthers2002; Spelke Reference Spelke, Gentner and Goldin-Meadow2003; Vygotsky Reference Vygotsky1962), the fact that “pragmatically rich” visuo-motor content becomes integrated within linguistic representations (e.g., word representations) can have profound effects on deaf signers’ cognitive processing at large. For example, if the online processing (or internal rehearsal) of linguistic stimuli elicits visuo-motor representations (e.g., of hand movements), these can become part and parcel of (say) the categories or the episodic memories that are formed in the meantime. The embodied/motor cognition literature highlights the importance of perceptual and motor representations in higher order cognitive processes such as categorization, memory, and planning (Barsalou Reference Barsalou1999; Jeannerod Reference Jeannerod2006; Pezzulo Reference Pezzulo2012; Pezzulo et al. Reference Pezzulo, Barca, Lamberti Bocconi and Borghi2010; Reference Pezzulo, Barsalou, Cangelosi, Fischer, McRae and Spivey2011; Reference Pezzulo, Candidi, Haris and Barca2013). Analogously, “signed-based” visuo-motor content originally associated with language processing can potentially extend well beyond it and affect a wide range of cognitive processes. This hypothesis remains to be tested empirically.

Language is tightly connected to the perceptual and motor processes involved in the expression and recognition of communicative signs. As “language directly interfaces with the mental representations that are used in perception and action” (Lupyan & Bergen Reference Lupyan and Bergen2016, p. 7), the extensive use of a signed-based communication learned through social interaction in a natural environment might have a broad impact on shaping mental representations. In the case of signed language, the motor actions that are deeply tied with communication mainly comprise upper limb and hand movements (expression) and visuo-spatial processes of gesture recognition and speech reading (recognition). As this repertoire of motor acts can be automatically enacted during language processing or rehearsal, it can affect signers’ linguistic skills and, potentially, cognitive processing at large.