Pickering & Garrod (P&G) have set out an argument for a possible integration of language production and comprehension processes in adults. However, charting how these two processes come to be set up during development is critical in attempts to understand their subsequent integration. An obvious problem when considering production/comprehension processes within a developmental framework is that language production lags behind language comprehension. So, to reiterate P&G's question, can silent naming use the production system, if no word production has yet emerged? Furthermore, the dynamic and nonmonotonic changes in the development of children's motor systems, as well as the specificity of motor-language links during these stages, need to be integrated into any such model. The need to examine developmental evidence in building this theory is therefore critical.

Numerous links between production and comprehension occur during language learning in childhood. These links have been reported in developmental studies assessing gesture production and language comprehension (Bates & Dick Reference Bates and Dick2002; Iverson & Thelen Reference Iverson and Thelen1999). From an atypical development standpoint, there is also a greater incidence of co-morbid motor coordination and planning difficulties in children with language impairment (Iverson & Braddock Reference Iverson and Braddock2011). In support of P&G's model, there are studies specifically demonstrating how auditory perception/language comprehension can also affect speech motor performance in childhood. For instance, perceptual ability can influence the learning of motor gestures. Seemingly due to the complexity of articulation, affricates are produced later in development for English-speaking children. By contrast, for Putonghua-speaking children, affricates are acquired very early, probably due to their salience within the language (Dodd & McIntosh Reference Dodd and McIntosh2010). Furthermore, for higher cognitive-linguistic demands as compared to lower ones, speech motor variability also increases (reviewed in Goffman Reference Goffman, Maassen and van Lieshout2010), indicating that the production processes are influenced by comprehension/perceptual processes. Indeed, a catalyst to changes in motor control may be vocabulary increases (Green & Nip Reference Green, Nip, Maassen and van Lieshout2010). Speech motor variability can also act as an index of learning; kinematic analyses of motor movements reveal that children receiving training for articulatory disorders produce different motor gestures associated with phonetic categories, which were imperceptible at the acoustic level (Gibbon Reference Gibbon1999).

However, it is important to note that these links do not extend to all motor skills, and in particular, not to gross motor ability as assessed by locomotion or play (Bates et al. Reference Bates, Benigni, Bretherton, Camaioni and Volterra1979). In longitudinal studies of infants, using parental reports, two orthogonal factors of language comprehension and production seemed to exist (Bates et al. Reference Bates, Bretherton and Snyder1988). Alcock and Krawczyk (Reference Alcock and Krawczyk2010) specifically investigated the link between oral motor control, and other motor and language abilities in 21-month-olds, and concluded that oral motor control was significantly associated with the grammatical complexity of utterances and with language production ability overall. They found no relationship between overall motor control and language comprehension ability at this age. In our study with school age children, oral motor control was linked to the production of novel words, but did not predict individual differences in the comprehension of syntactically complex sentences (Krishnan et al., in press). This evidence may appear contradictory, but taking a developmental perspective may provide some explanation. First, there may be specific motor behaviours that provide an opportunity to acquire and practise skills necessary for language. For example, rhythmic hand banging peaks around 28 weeks of age, and this is also when children start to produce reduplicated babbling (Iverson Reference Iverson2010). And, when rhythmic banging is delayed, as is the case in the neurodevelopmental disorder Williams syndrome, babbling and subsequent comprehension and production are also delayed (Masataka Reference Masataka2001). Understanding the specific skills that are likely to cause changes in behaviour during a particular time-window may therefore be necessary. P&G's model fails to provide an account of how comprehension/production processes for learning language might be integrated within specific skills over developmental time.

The second factor that must be considered in a model integrating comprehension/production processes is the nonmonotonicity of these developmental trajectories, which are consistent across children. For example, rhythmic banging is low in pre-babbling children, increases sharply as infants start to babble, and then declines as infants become experienced at babbling (Iverson Reference Iverson2010). Similar nonmonotonic trajectories are seen across other speech motor skills, for instance, in the variability of lip and jaw movements (Smith & Zelaznik Reference Smith and Zelaznik2004) or for the coordination of upper and lower lip movements (Green et al. Reference Green, Moore, Higashikawa and Steeve2000). The combination of gestures and language during development may have a similar nonmonotonicity, as event-related potential (ERP) evidence suggests children infants younger than 20 months interpret symbolic gestures and words similarly, but that gestures and words take on divergent communicative roles when infants are 26 months old (Sheehan et al. Reference Sheehan, Namy and Mills2007). Although P&G suggest that experience may be important for learning inverse-forward model pairing, their model lacks explanations of how these kinds of trajectories might arise, how trajectories change with time, and what kind of input may be necessary for change. For example, these trajectories may arise due to some combination of the changes in contextual support that are needed while a skill is learnt, or the neural changes that occur during development.

Therefore, in the model that P&G outline, I agree that integrating knowledge about the production processes may help us understand more about language comprehension, but this would be possible only if the specificity of production-comprehension links and the developmental timing of their occurrence are taken into account.