While Ackermann et al.'s theory is interesting, seems plausible, and may initially appear tempting, it is based on incomplete readings of several literatures. First, it is unclear why some of their arguments should only apply to the specific instances of rhythmic and prosodic control the authors discuss or why they fail to apply in other animal species. Their model assumes that enhancement of in-group cooperation and cohesion was the main driving force for the evolution of speech via the intermediate step where vocal control and rhythm production would serve as chorusing and bonding tools. A key assumption is that speech would produce rhythmic abilities as an evolutionary by-product. This scenario is in line with some empirical observations (for reviews, see Fitch Reference Fitch, Rebuschat, Rohrmeier, Hawkins and Cross2012; Geissmann Reference Geissmann, Wallin, Merker and Brown2000) and previous theoretical frameworks for the origins of music (Hagen & Bryant Reference Hagen and Bryant2003; Hagen & Hammerstein Reference Hagen and Hammerstein2009; Merker Reference Merker2000; Merker et al. Reference Merker, Madison and Eckerdal2009). However, when applied to language, Ackermann et al.'s evolutionary model does not withstand cross-species validation: Many nonhuman animals exhibit rhythmic behaviors while lacking speech. Before primate rhythmic abilities can be compared with humans' at all, more evidence regarding flexibility in vocalizations' temporal patterning (Fedurek et al. Reference Fedurek, Schel and Slocombe2013) and motor synchronization (Hattori et al. Reference Hattori, Tomonaga and Matsuzawa2013) is needed in apes (cf. (Ravignani et al. Reference Ravignani, Olivera, Gingras, Hofer, Hernández, Sonnweber and Fitch2013).

Evidence from non-primate species also seems to undermine Ackermann et al.'s model. Two bird species, both vocal learners, have been shown to entrain to steady pulses (Hasegawa et al. Reference Hasegawa, Okanoya, Hasegawa and Seki2011; Patel et al. Reference Patel, Iversen, Bregman and Schulz2009a), supporting Ackermann et al.'s model and Patel's hypothesis, whereby auditory-motor entrainment skills would be evolutionary by-products of vocal learning abilities (Patel Reference Patel2006). However, recent evidence suggests that vocal learning and rhythmic abilities might be dissociated. Sea lions, unlike seals, show no evidence of vocal learning (Janik & Slater Reference Janik, Slater, Slater, Rosenblatt, Snowdon and Milinski1997) but nonetheless can reliably synchronize their movements to a range of musical stimuli at different tempi (Cook et al. Reference Cook, Rouse, Wilson and Reichmuth2013). Humans and sea lions are both rhythmically skilled, but only humans evolved vocal learning and speech. Therefore, sea lions constitute outliers inconsistent with the prediction of Ackermann et al.'s model. This species evolved cognitive rhythmic abilities, without evolving speech. Invoking additional evolutionary forces and physiological mechanisms thus appears necessary: How can Ackermann et al.'s model be modified to avoid incorrectly predicting vocal learning in rhythmic-skilled species?

Second, Ackermann et al.'s model assumes that prosodic modulation of speech conveys mainly simple motivational-emotional information, and thus, that prosody and complex speech production had separate evolutionary histories. But evidence showing a tight connection between prosody and complex linguistic functions argues against this “double pathway” theory. Prosodic contour is influenced by syntactic constituent structure, semantic relations, phonological rhythm, pragmatic considerations, as well as by the length, complexity, and predictability of linguistic material (Wagner & Watson Reference Wagner and Watson2010). Furthermore, prosodic cues are used in childhood during acquisition of words (Christophe et al. Reference Christophe, Millotte, Bernal and Lidz2008) and grammatical constructions (Männel et al. Reference Männel, Schipke and Friederici2013), and in adulthood for syntactic processing (Christophe et al. Reference Christophe, Millotte, Bernal and Lidz2008; Kjelgaard & Speer Reference Kjelgaard and Speer1999; Langus et al. Reference Langus, Marchetto, Bion and Nespor2012; Wagner Reference Wagner2010) and word recognition (Cutler et al. Reference Cutler, Oahan and van Donselaar1997).

Contra Ackermann et al., such complex linguistic modulation of prosody seems to be a prerequisite for the acquisition and use of language, and this process is likely to be influenced by cognitive mechanisms specially modified in the human lineage. Comparative research on syntax precursors favors this hypothesis: The ability to assemble sequences of sounds into hierarchical patterns might be either human-specific, or very poorly developed in other species (Conway & Christiansen Reference Conway and Christiansen2001; ten Cate & Okanoya Reference ten Cate and Okanoya2012). Hence, developmental and comparative evidence point to a more complex cognitive integration of prosody and speech than allowed by the dual-pathway proposal of Ackermann et al. The challenge for Ackermann et al.'s theory is, therefore, to account for the modulation of prosody by human-specific cognitive functions (e.g., syntax), which are clearly not evolutionary homologues of primate emotional vocalizations controlled by the anterior cingulate cortex.

Finally, Ackermann et al. propose an ontogenetic pathway in which: (1) basal ganglia (BG) are important to generate integrated templates of orofacial and laryngeal movements during childhood, but (2) in adulthood can be retrieved from cortical areas because these motor templates become well-trained. Later in ontogeny, BG would mostly subserve the modulation of emotional prosody, and not the coordination of speech production. These claims are not supported by currently available empirical data. For instance, Ackermann et al. cite Parkinson's Disease (PD) data to support their claims that, in adults, BG lesions only impair emotional prosody. In fact, PD patients with normal cognitive functioning are more impaired in semantic fluency tasks than in phonetic fluency (Henry & Crawford Reference Henry and Crawford2004). Additionally, contra Ackermann et al., BG subserve complex syntactic and semantic processing in adults, with empirical findings consistent across PD (Dominey & Inui Reference Dominey and Inui2009; Henry & Crawford Reference Henry and Crawford2004; Lewis et al. Reference Lewis, Lapointe, Murdoch and Chenery1998), BG lesion (Kotz et al. Reference Kotz, Frisch, Cramon and Friederici2003; Teichmann et al. Reference Teichmann, Gaura, Demonet, Supiot, Delliaux, Verny, Renou, Remy and Bachoud-Levi2008; Ullman et al. Reference Ullman, Corkin, Coppola, Hickok, Growdon, Koroshetz and Pinker1997), and neuroimaging research (Friederici & Kotz Reference Friederici and Kotz2003). These data suggest that in adults the BG support multiple functions relevant to spoken language, not just simple emotional prosodic modulation.

Furthermore, contrary to the developmental pathway proposed by Ackermann et al., the acquisition of novel syntactic structures in adults depends on the medial temporal cortex, and the retrieval of syntactic templates after thorough learning mostly recruits the BG and perisylvian structures (Ullman Reference Ullman2004). This evidence shows that, contra Ackermann et al., BG are active in the retrieval of over-learnt procedures. Ackermann et al. therefore need to propose alternative explanations to reconcile child and adult data concerning the function of BG.

In conclusion, to make their model robust, Ackermann et al. must modify and refine their evolutionary and mechanistic explanations, and clarify which assumptions are necessary, and which are sufficient, for their explanatory framework to hold. Is their model robust enough to stand up to the clear, strong relationship between prosody and complex linguistic functions? How can Ackermann et al.'s model account for the complex functions of BG in adulthood? If in-group cohesion had to be achieved, why was precise vocal control specifically selected for, rather than general non-vocal rhythmic abilities? These and other questions need to be addressed if Ackermann et al.'s model is to become convincing.