The veracity of neonatal imitation has been debated ever since Meltzoff and Moore (Reference Meltzoff and Moore1977) first presented evidence to suggest that newborns imitate adult oral and manual gestures. In an attempt to chart the prevalence of newborn imitation and determine its relationship with later infant development (Suddendorf et al. Reference Suddendorf, Oostenbroek, Nielsen and Slaughter2013), we undertook the largest and most comprehensive study of neonatal imitation to date (Oostenbroek et al. Reference Oostenbroek, Suddendorf, Nielsen, Redshaw, Kennedy-Costantini, Davis, Clark and Slaughter2016). We tested a sample of 106 infants four times, when infants were 1, 3, 6, and 9 weeks of age. A female adult experimenter modeled nine social gestures: four facial gestures (tongue protrusion, mouth opening, happy face, and sad face), two hand gestures (index finger protrusion and grasping), and three vocal gestures (“mmm” sound, “eee” sound, and tongue clicks). Additionally, two nonsocial models were included to test alternative interpretations of neonatal imitation (tube protrusion simulating tongue protrusion, and box opening simulating mouth opening) (Jacobson, Reference Jacobson1979; Jones Reference Jones1996). Across this range of gestures, there was no sign that infants selectively imitated any of the modeled gestures.

In our analyses of infants' responses to these models (as outlined above), we replicated some previous findings in restricted subsections of the data. Specifically, when we used the common cross-target procedure of comparing infants' tongue protrusions in response to the matching model with their tongue protrusions in response to the control model of mouth opening (see Meltzoff & Moore Reference Meltzoff and Moore1977), we found that infants produced significantly more matching responses than non-matching responses at 1 and 9 weeks of age. However, this effect disappeared when we used other gestures as the control model. Because there is no a priori reason to favor mouth opening as the comparison control model over any other control model (e.g., happy face), these cross-sectional findings also do not provide evidence of newborn imitation. This suggests that failure to include adequate control conditions or to test infants across multiple time points in previous studies has resulted in the false impression that infants selectively copy tongue protrusions, thereby perpetuating the idea that newborn imitation exists.

In their target article, Keven & Akins (K&A) offer a novel explanation for why tongue protrusion in particular may be falsely identified as imitation. They argue that what others have interpreted as imitation may be spontaneous oral activity – with tongue protrusion being characteristic – arising from maturation of the infant's respiratory and digestive systems. K&A not only offer a rationale for why newborns might engage in tongue protrusion behaviors at such high rates, but also provide a compelling and consistent argument for why this gesture may decline after the first 3 months of life. We welcome this contribution because it provides another reason why previous data claiming evidence of neonatal imitation, and tongue protrusion imitation in particular, should be interpreted with caution.

In their article (sect. 7, para. 2), K&A cite our research while discussing the frequency of newborn behaviors. They state, “Of these ‘early’ stereotypies, TP/R [tongue protrusion and retraction] and MO/C [mouth opening and closing] and index finger protrusion are produced with the highest frequencies during the first week after birth” (Oostenbroek et al. Reference Oostenbroek, Suddendorf, Nielsen, Redshaw, Kennedy-Costantini, Davis, Clark and Slaughter2016). We take this opportunity to clarify that the high frequencies of these behaviors do not provide evidence of imitation. Although 1-week-old infants in our sample engaged in higher levels of tongue protrusion, mouth opening, and index finger protrusion than any of the other modeled gestures at this age, they produced these behaviors as often in response to many of the control models as they did to the corresponding matching models. K&A's alternative explanation for the “imitation” of oral gestures, coupled with our comprehensive, longitudinal data, supports the conclusion that imitation does not exist in the newborn period. Rather, its apparent expression is a result of the maturation of developmental mechanisms that may have nothing to do with social learning.

After more than 30 years of controversy over its existence in the newborn period (see Oostenbroek et al. Reference Oostenbroek, Slaughter, Nielsen and Suddendorf2013; Ray & Heyes Reference Ray and Heyes2011; Simpson et al. Reference Simpson, Murray, Paukner and Ferrari2014a; Suddendorf et al. Reference Suddendorf, Oostenbroek, Nielsen and Slaughter2013) driven largely by studies using cross-sectional designs and only tongue protrusion as the model gesture, K&A's article, along with our findings, necessitates a reframing of imitation research. These data suggest that neonatal imitation is an illusion and should encourage new perspectives and research goals. There is no longer a compelling empirical basis for the idea that human neonates imitate. Developmental scientists should reconsider the origin of imitation, and explore novel hypotheses about the functions of newborn oral behaviors, starting with K&A's well-articulated and compelling account.