The Ackermann et al. target article offers intriguing suggestions about a dual-pathway approach to evolution of vocal capabilities and language. But the approach could be enhanced with regard to its behavioral assumptions by taking into account key information on vocal development. Without this information, the article misses the opportunity to augment some of its most interesting claims with evidence that could help situate the model in a more evolutionary-developmental (evo-devo) frame (Bertossa Reference Bertossa2011).

In particular, the article does not take note of the evidence that early vocal development of humans (especially in the first 3 months) shows emergent phonatory control, rather than emergent articulatory control (Buder et al. Reference Buder, Chorna, Oller and Robinson2008; Koopmans-van Beinum & van der Stelt Reference Koopmans-van Beinum, van der Stelt, Lindblom and Zetterstrom1986; Oller Reference Oller, Yeni-Komshian, Kavanagh and Ferguson1980). Phonatory control does not necessarily imply “prosodic” control (the focus of much of the target article) in the sense that the term prosody is used in literature about mature languages. Prosody is normally meant to denote the capacity to integrate suprasegmental variations across multisyllabic strings, and the human infant in the first months produces no such syllabic strings. Instead, categories of infant vocalization that are recognized in the first few months by parents and laboratory staff cross-culturally include “protophones” (Oller Reference Oller2000), such as vowel-like “vocants” (sounds produced in the mid-pitch range of the individual infant, with “normal” phonation, the kind that typically occurs in speech), squeals (high-pitched sounds for the infant in question, often in falsetto), and growls (low-pitched or raucous sounds often in creaky voice). These typical protophones occur as phonatory events with little or no articulatory modulation, and to the limited extent that supraglottal modulation occurs, it appears to be disorganized and unpredictable at this stage. On the other hand, the protophones are easily recognized as distinguishable categories because they tend to occur in clusters of the same types (a series of squeals, for example, followed later by a series of vocants) even as early as 3 months (Kwon et al. Reference Kwon, Oller and Buder2007). The high rate of production of the protophones, along with the fact that they occur both in solitary and social circumstances (Locke Reference Locke1993; Stark Reference Stark, Yeni-Komshian, Kavanagh and Ferguson1980; Yale et al. Reference Yale, Messinger, Cobo-Lewis, Oller and Eilers1999), suggests endogenous motivation in the infant to explore and seemingly to practice vocalization, as well as to use vocalization to serve social functions. In addition, all the protophones are used by infants in expression of positive, neutral, and negative affective states (as indicated by facial affect), and these expressions are predictably related to responses of caregivers ranging from encouraging interaction in response to positive expressions, to changing the situation (or talking about the need for it) in response to negative expressions (Oller et al. Reference Oller, Buder, Ramsdell, Chorna, Warlaumont and Bakeman2013). All these properties of very early vocal development (spontaneous production, the ability to repeat sounds in clusters, vocal social interaction, and the ability to use sounds to express differing emotional states on differing occasions) in the human infant at 3 months are based on phonatory control, and all of them are foundational for language, since every aspect of human language requires flexible control of phonation.

Phonatory control takes naturally logical precedence over supraglottal control in the sequence of development, and that logical precedence is reflected in the facts of development. Phonatory categories appear in development before systematic supraglottal articulated categories such as “canonical” syllables, wherein well-formed syllables (heard as, for example, “dada” or “baba”) are produced through coordination of phonation and systematically repeatable supraglottal articulations (Koopmans-van Beinum & van der Stelt Reference Koopmans-van Beinum, van der Stelt, Lindblom and Zetterstrom1986; Oller Reference Oller, Yeni-Komshian, Kavanagh and Ferguson1980; Stark Reference Stark, Yeni-Komshian, Kavanagh and Ferguson1980). In 40 years of longitudinal research in human infant vocalization, I have never witnessed any infant developing systematic and communicative supraglottal movements and using them in practice-like play or social communication, prior to developing phonatory categories as described above.

These facts offer an opportunity to Ackermann et al. to more thoroughly incorporate developmental patterns into their expectations regarding the “monosynaptic refinement of the projections of motor cortex to the brainstem nuclei that steer laryngeal muscles.” Such refinement, it would appear, must begin to be manifest in the human infant's brain by 3 months or earlier, although by that age, the behavioral data do not suggest “prosodic” control, but something simpler – control over gross differences in phonatory pattern and pitch. The developmental picture appears to be sufficiently clear to help illuminate differences between human and nonhuman primate brain organization at maturity, by suggesting notable cross-species similarity of phonatory and articulatory control capabilities across nonhuman primates and humans early in life with much larger differences developing as time passes.

In the near future, it may also be possible for the modeling of Ackermann et al. to be enhanced by quantified direct comparisons among vocal capabilities of humans and nonhuman primates. On the one hand, studies in human infants are rapidly tying down facts regarding vocal rate (volubility) for protophones across ages and across social and non-social circumstances (Franklin et al. Reference Franklin, Warlaumont, Messinger, Bene, Iyer, Lee, Lambert and Oller2014; Goldstein et al. Reference Goldstein, Schwade and Bornstein2009; Nathani et al. Reference Nathani, Neal, Olds, Brill and Oller2001) as well as substantially improving our understanding of vocal types and their flexibility in human infants (Griebel & Oller Reference Griebel, Oller, Oller and Griebel2008; Scheiner et al. Reference Scheiner, Hammerschmidt, Jürgens and Zwirner2006; Stark et al. Reference Stark, Bernstein and Demorest1993). Similarly, considerable progress is being made on the description of both the amount of vocalization that occurs across age in nonhuman primates and the degree to which these vocalizations are used in differing contexts, the latter representing an attempt to characterize the degree to which the social functions of nonhuman calls may indeed show flexibility (Crockford & Boesch Reference Crockford and Boesch2003; Laporte & Zuberbühler Reference Laporte and Zuberbühler2010). With the recent development of a facial affect coding system for chimpanzees (Parr et al. Reference Parr, Waller and Heintz2008) modeled on the Ekman scheme for human affect (Ekman & Friesen Reference Ekman and Friesen1978), quantitative comparison of functional flexibility in vocalization across humans and chimpanzees should soon be reached. Such improvements in our quantitative understanding of development, amplified by cross-species comparisons, should fundamentally enhance the modeling of the evolution of language and the brain mechanisms that underlie it.