The target article by Goldin-Meadow & Brentari (G-M&B) emphasizes the importance of maintaining a firm distinction between sign and gesture; I suggest that this has consequences that are both helpful and unhelpful for those who wish to understand the cognitive and communicative roots of linguistic structure. These questions can be asked at multiple timescales, ranging from the earliest systems used by our hominin ancestors, to the kinds of systems that arise de novo among modern humans today.

On the helpful side, G-M&B's proposal may resolve one objection to gesture-first theories of language emergence over the course of human evolution. Under such theories, human language is thought to have achieved crucial milestones in the manual modality, with potential explanations including the greater intentional control of the manual articulators relative to the vocal tract in primates and the iconic affordances of the manual modality for accomplishing reference, and so forth. Variations on this theme figure prominently among both classical (de Condillac Reference de Condillac1798; Jastrow Reference Jastrow1886) and contemporary (Arbib Reference Arbib2005; Corballis Reference Corballis2003; Hewes Reference Hewes1973) theories of language evolution.

One important objection to such views is captured in McNeill et al. (Reference McNeill, Duncan, Cole, Gallagher and Bertenthal2008). They argued that because modern language crucially integrates codified linguistic forms (speech) with synthetic/imagistic forms (gesture), any evolutionary trajectory that fails to involve this “dual semiosis” from the very beginning ends up predicting what did not evolve (speech-only, or gesture-only) instead of what did (a tight coupling of speech+gesture). However, by demonstrating that both modes of semiosis are in fact available in the manual modality, G-M&B's target article provides a possible resolution to the concerns raised by McNeill et al. (though other objections remain).

On the unhelpful side, G-M&B's perspective on the relation between silent gesture and language is puzzling. For example, in section 4.2, they suggest that the signature of language is uniformity (adhesion to standards of form), whereas gesture is characterized by variability. They then proceed to argue that handshape must be linguistic, because it shows variability from system to system, whereas movement and location may be gestural, because they are so uniform. This is precisely the opposite of their earlier argument. Then, in the very next section, we are encouraged to believe that silent gesture “might be more appropriately called spontaneous sign” (sect. 4.3), despite later acknowledgments that silent gesture does not contain all of the properties of even homesign, let alone conventional sign languages. (I note that silent gesture may also have additional properties that homesign doesn't, and that such differences could be an important source of insight.)

I suspect that this lack of clarity stems from a tension between wanting to delineate what is gestural from what is linguistic in the face of a reality in which the lines between language and gesture are often blurry, as G-M&B periodically acknowledge. But from my perspective, and from the perspective of evolutionary linguistics, these lines should be blurry. Phenomena in synchronic analyses are often more clear in the light of diachrony, in which change happens gradually and continuously. G-M&B themselves suggest that conventional sign languages have their roots in homesign systems, perhaps also influenced by the co-speech gestures of hearing people. (I would submit that the silent gestures of hearing people may also influence a developing system in some contexts.) In my view, the ability to clearly delineate gesture from language is less important than the ability to understand how gesture becomes language.

This process-oriented view need not undercut the key insight that the manual modality contains more than one semiotic mode. But it does free us to worry less about precisely where to draw the line, as biologists gave up doing long ago regarding speciation (De Queiroz Reference De Queiroz2007; Irwin et al. Reference Irwin, Irwin and Price2001). We can instead recognize a gradient from loosely concatenated, paratactic orderings in silent gesture to grammatically constrained syntactic forms in conventional sign languages. Sandler (Reference Sandler and Nicksubmitted) presents an exemplary illustration of this approach as applied to several generations of an emerging sign language. It may also be profitable to consider whether such a progression from less language-like to more language-like also characterizes deaf homesigners as they grow from infancy to adulthood. A similarly gradual process is likely to have characterized the progression from no language to protolanguage to language, as early hominins moved from less-organized to more-organized forms of representation and communication.

With sign language and other gestural systems (e.g., homesign, silent gesture) among our best models of human language evolution, some “line blurring” may be necessary to understand the progression from nonlanguage to language. G-M&B's challenge for us is to figure out how to do so without sacrificing the important insights to be gained from cases where gesture and sign operate in different semiotic modes.