Teaching involves bidirectional communication (Strauss et al. Reference Strauss, Calero and Sigman2014), which is inherently social. Learning, the complement of teaching, has been studied extensively in the cognitive sciences. As important as teaching is for humans, it has been largely neglected in the cognitive sciences. Kline's target article is a beginning attempt to redress this problem.

To explain and describe teaching, Kline casts her net wide to include phylogeny, cultural anthropology, and nonhuman animal teaching. But as important as it is to include these domains, there is a need to have an even broader study of teaching. Other cognitive areas could include the brain sciences (Hasson et al. Reference Hasson, Ghazanfar, Galantucci, Garrod and Keysers2012; Holper et al. Reference Holper, Goldin, Shalom, Battro, Wolf and Sigman2013), artificial intelligence (Dessus et al. Reference Dessus, Mandin, Zampa, Tazi and Zreik2008), psycholinguistics (Bartsch et al. Reference Bartsch, Wright and Estes2010), philosophy (Scheffler Reference Scheffler1965), intelligent tutoring systems (Kopp et al. Reference Kopp, Britt, Millis and Graesser2012), cognitive archeology (Morgan et al. Reference Morgan, Uomini, Rendell, Chouinard-Thuly, Street, Lewis, Cross, Evans, Kearney, de la Torre, Whiten and Laland2015) and computer–human interface systems in robotics (Vollmer et al. Reference Vollmer, Muhlig, Steil, Pitsch, Fritsch, Rohlfing and Wrede2014) among others. (See Strauss [Reference Strauss, Kornhaber and Winner2013] for a call for such a comprehensive, multidisciplinary view.) Part of this enlarged view of teaching includes cognitive development. In that domain, Strauss et al. (Reference Strauss, Ziv and Stein2002) were the first to suggest that teaching may be a natural cognitive ability for humans. Of the support that can be harnessed to bolster this claim, here we slightly elaborate only one: teaching is developmentally reliable among humans. The nub of the argument is that teaching is complex and opaque, and that, although toddlers are exposed to teaching, they are not taught to teach. Nonetheless, teaching appears early (Strauss Reference Strauss, Pillemer and White2005).

Developmental research conducted in different industrialized countries with different tasks indicates convergent paths of teaching strategies (Strauss & Ziv Reference Strauss and Ziv2012). Strategies proceed from teaching precursors in infancy, to teaching via demonstration among 3-year-olds, to explanation and contingent teaching among 5-year-olds. Strauss and Ziv (Reference Strauss and Ziv2012) constructed a taxonomy of teaching strategies culled from research with children from industrialized societies.

In contrast, there is a dearth of developmental research on teaching in non-industrialized societies. We propose that there is room for such studies. The result will likely be an extension of Strauss and Ziv's (Reference Strauss and Ziv2012) taxonomy, which will help us better understand the scope of human teaching.

In the target article, Kline proposes a framework based on a functional approach for integrating three research traditions regarding teaching: mentalistic, culture-bound, and functional. However, it may be that by contrasting these traditions and determining the different kinds of teaching they identify, they could yield a more fine-grained taxonomy. For example, Frye and Ziv (Reference Frye, Ziv, Homer and Tamis-LeMonda2005) argued that functional definitions cannot identify failed teaching as teaching because they incorporate changes in the recipient's (learner's) behavior as a part of the definition.

In the mentalistic approach, teaching is identified by its goal, so failed teaching is still teaching even if the goal is not fulfilled. This example illustrates that mental states are not just epiphenomena in teaching but make possible different actions. It also shows that if mental states are not considered, then entire categories of teaching will be absent from our taxonomies and not studied.

We suggest that incorporating mentalistic and culture-bound definitions of teaching could be beneficial for understanding human teaching. What is at stake here is the nature of the relations between the cultural view and the mentalistic, cognitive substrate. This question is akin to issues in the debate between adherents of cultural niches (Boyd et al. Reference Boyd, Richerson and Henrich2011) and cognitive niches (Pinker Reference Pinker2010).

As Kline notes, modern industrialized-type teaching has been found to be rare in non-Western small-scale societies. Some ethnographers (Lancy Reference Lancy2010; Paradise & Rogoff Reference Paradise and Rogoff2009) have noted little vertical teaching (parent to offspring); however, the application of evolutionary theory to cultural transmission in non-industrialized cultures predicts vertical and oblique (non-family adult to child) teaching. Both were found when adults reported who gets taught by whom, for what tasks and at what age (Kline et al. Reference Kline, Boyd and Henrich2013; see also Hewlett et al. Reference Hewlett, Fouts, Boyette and Hewlett2011).

It is possible that children come into the world with similar cognitive abilities that allow them to teach in culturally relevant ways. This claim is neutral to the ways in which children learn teaching behaviors, such as Bayesian learning or Chomskian-type universals and acquisition devices. If, as we claim, there are universal human teaching abilities, an explanation is needed for how they get expressed differently in various societies. There are two main kinds of explanations.

First, we might expect similar cognitive origins and developmental trajectories of teaching among children from different societies. Regarding origins, research on Western preverbal infants indicates that they correct others' mistakes and even anticipate errors and act to ward them off (Knudsen & Liszkowski Reference Knudsen and Liszkowski2012a; Reference Knudsen and Liszkowski2012b). These are at the foundation of teaching. It would be of interest to see if infants from, say, small-scale subsistence and hunting and gathering societies show the same abilities. If so, we could argue that at their inception, cognitive prerequisites for teaching are not cultural. Nevertheless, teaching is not uniform across cultures. A developmental route from common infant precursors of teaching to diverging teaching in adults in various cultures could be that teaching develops along a similar pathway until it parts ways at a certain juncture, in directions that are culturally based. This would suggest that cultural variations of teaching may need a certain developmental level of cognitive substrates before these variations appear.

Second, infants from modern industrialized and small-scale societies may be cognitively different from the earliest points when cognitive prerequisites for teaching can be detected. This possibility suggests that teaching's cognitive precursors may be cultural in nature even before actual teaching appears.

Choosing between these two alternative explanations is an empirical question. Neither of these alternatives has been studied experimentally, and we believe they should be.

In summary: We propose that a more comprehensive multidisciplinary approach to teaching is in order. It could include cognitive development. We argue that teaching may be a natural cognitive ability among humans. If so, we would need to explain relations between a possibly common cognitive beginning point in infancy and cultural variations of teaching.