Teaching Is a Natural Cognitive Ability for Humans

Wesuggest that amultidisciplinary approach to teaching has potential to widen its scope. In that vein, we revisit our original claim that teaching is a natural cognitive ability among humans. We elaborate on three requirements for such an ability and report that, first, teaching strategies may be developmentally reliable. Findings indicate a possible normative developmental trajectory from age one year through adulthood. Second, teaching seems to be species-typical, that is, it is a universal human achievement. Third, human teaching with a theory of mind (ToM) is species-unique. Nonhuman animals may teach without a ToM. Teaching is often seen exclusively as what happens when an adult teacher teaches children in school settings. However, teaching’s reach is greater than that. Adults teach youngsters in societies where there are no schools. Regardless if there are schools or not in a particular society, young children teach each other. These and other findings about teaching led us to claim, a decade ago, that teaching is a natural cognitive ability in humans (Strauss, Ziv, & Stein, 2002). That claim was subsequently expanded by Strauss (2005). In the ensuing years, the idea has been picked up by others, especially Csibra and Gergely (2009), and has been elaborated and expanded. We revisit our claim in light of advances occurring in a number of fields that impinge on our idea. In this article, we provide a panoramic view of teaching, define teaching, characterize properties of natural cognitive abilities, discuss the claim that teaching is a natural cognitive ability with emphasis on its being developmentally reliable, and make some final remarks. 1Center for Academic Studies, Tel Aviv University 2Alqasemi Academic College of Education, Tel Aviv University Address correspondence to Sidney Strauss, School of Psychology, Center for Academic Studies, Or Yehuda, Israel or School of Education, Tel Aviv University, Tel Aviv 69978, Israel; e-mail: [email protected] or [email protected] A PANORAMIC VIEWOF TEACHING Teaching, as an area of scholarship and research, has often been viewed within the prisms of solitary fields. It has been the focus of many disciplines, yet researchers often work in domain-isolation, perhaps unaware of work carried out in other domains. For example, researchers who investigate teaching in the field of education are often unaware of research in the field of cultural evolution and vice versa. We suggest that teaching research can benefit from an integrated multidisciplinary effort. Domains that have potential to add to our basic knowledge about teaching include the ontogenetic development of human teaching (Davis-Unger & Carlson, 2008a, 2008b; Strauss et al., 2002), anthropology (Greenfield, 2004), cultural evolution (Mesoudi,Whiten,&Laland, 2006), cognitive evolution (Arbilly, Motro, Feldman, & Lotem, 2011; Shultz, Nelson, & Dunbar, 2012), nonhuman animal teaching (comparative psychology) (Thornton & Raihani, 2008, 2010), brain sciences (Battro, 2007, 2010), artificial intelligence (Dessus, Mandin, & Zampa, 2008), cognitive archeology from prehistoric periods (Goren-Inbar, 2011), psycholinguistics (Bartsch,Wright,&Estes, 2010), philosophy (Scheffler, 1965), intelligent tutoring systems (Kopp, Britt, Millis, & Graesser, 2012), computer–human interface systems, silicon–biology interfaces, and more. We believe, however, that at the core of human teaching stands cognition and its development. Teaching’s mirror image, learning, has been a major focus of the cognitive sciences. But the flip-side of learning, human teaching that occurs when a teacher intends to cause learning in others has, by and large, been flying below the radar in the cognitive and developmental sciences. We suggest that a multidisciplinary study of teachinghas potential to unlock adeepunderstanding of a fundamental aspect of human cognition. These insights can also shed a focused light on teaching that could, in turn, serve the education of our children. When scientists and scholars express interest in teaching, they often do so in an attempt to explain other important phenomena. For instance, in the field of cultural evolution, teaching is invoked when scientists explain the cumulative nature of human culture (Tomasello, Kruger, & Ratner, 1993; © 2012 The Authors 186 Journal Compilation© 2012 International Mind, Brain, and Education Society and Blackwell Publishing, Inc. Volume 6—Number 4 Sidney Strauss and Margalit Ziv Whiten, 2011). Another example comes from comparative psychology where human and nonhuman cultures are comparedandteaching isseenasoneof thedifferencesbetween the two (Whiten, Horner, & Marsall-Pescini, 2003) that might have significance for human culture versus nonhuman tradition. In contrast, teaching is stage center in our work. Instead of it being used to explain phenomena, we believe that teaching itself is in need of explanation (and description). An avenue into this is through the multidisciplinary approach we are advocating. This does not deny the significance of explaining phenomena via teaching. Itmerely shifts thefigure andground. We see both projects as complementary. DEFINITIONS OF HUMAN TEACHING Definitions of teaching are suggested in varied disciplines and are incongruent. In philosophy, Green (1964), for example, presents a teaching continuum running from training to instruction and excluding conditioning and indoctrination. In psychology (e.g., Pearson, 1989), intentionality stands at its core, and in education (e.g., Scardamalia & Bereiter, 1989) several conceptions of teaching are suggested, such as cultural transmission, training of skills, fostering natural development, and producing conceptual change. In computer sciences (e.g., Dessus et al., 2008), emphasis is placed on monitoring others’ minds and systematically communicating knowledge and feedback. Thus, teaching is a slippery concept. Yet with all of its elusiveness, at the heart of definitions of human teachingmaybe theoryofmind (ToM)(Olson&Bruner, 1996; Strauss, 1993; Strauss & Shilony, 1994). According to this definition, teaching involves an understanding of others’ mental states. Teachers understand others’ (i.e., learners’) knowledge, beliefs, desires, and so on; they recognize that there is a knowledge gap between a knowledgeable person in a certain domain and a learner who has less knowledge; they foster others’ knowledge or understanding in an attempt to reduce theknowledgegapanddoso intentionally (i.e., teachers attempt to cause learning in others’ minds). Thus, a ToM definition of human teaching refers to both the intentionality involved in teachingandtheknowledgecomponent, as follows: teaching is an intentional activity that is pursued in order to increase the knowledge (or understanding) of another who lacks knowledge, has partial knowledge or possesses a false belief (Ziv & Frye, 2004). Armed with this definition of teaching,wecannowprogress toourmajor claim that teaching is a natural cognitive ability. TEACHING AS A NATURAL COGNITIVE ABILITY What is anatural cognitive ability?According toCosmides and Tooby (n.d.), natural cognitive abilities have five properties. They (1) are complexly structured for solving a specific type of adaptive problem, (2) develop without any conscious effort and in the absence of any formal instruction, (3) are applied without any conscious awareness of their underlying logic, (4) are distinct frommore general abilities to process information or behave intelligently, and (5) reliably develop in all normal human beings. To this list, we add the following: they are (6) species-typical, that is universal, and (7) species-unique. Species typicality, or universality, has importance because it indicates that the natural cognitive ability under discussion, teaching in our case, is not restricted to certain groups but, instead, is a characteristic of all human beings. It also suggests that all humans are exposed to teaching, which is a prerequisite for development. The property that teaching is unique to human beings suggests there may be phylogenetic differences between humans and animals that make human cognitionunique.This, in turn, suggests that theremightnotbe continuity in the phylogeny of cognition. Onemuch-discussed candidate for a natural cognitive ability is human spoken language. Others are face recognition, navigation through space, and the interpretation of a threat (Cosmides & Tooby, n.d.). Each solves an important adaptive problem. We previously suggested that all seven properties hold for teaching (Strauss, 2005, 2011; Strauss et al., 2002). Due to space constraints, we won’t discuss all seven properties in detail. We briefly present our ideas about the first four and then elaborate on the final three. With regard to the first property, concerning complexity, teaching may be complexly structured for solving a specific type of adaptive problem. An adaptive problem has two main characteristics. First, it is one that has appeared during the evolutionary history of our species. Second, adaptive problems have affected the reproduction of organisms. That adaptive problem for teaching is passing on information, ideas, and procedures to others more effectively than were learning to be unassisted and in a manner that addresses great variability among learners (Premack & Premack, 1996). The complexity of human teaching is remarkable. It involves assumptions about learning (Olson & Bruner, 1996), defining goals and standards of the required change (Premack & Premack, 1996; Ziv, Solomon, & Frye, 2008), applying multiple teaching methods while adapting them to the learner (Ziv et al., 2008), providing feedback (Premack&Premack, 1996) andmore. For a fuller discussion of how teaching is complexly structured for solving problems of communication, see Strauss (2005). Regarding the second property, that teaching is learned effortlessly and without instruction, despite teaching’s complexity (Strauss, 2005), youngsters seem to learn it and spontaneously engage in teaching effortlessly, without intentional instruction. Children teach peers and adults how to play certain games (Ziv & Frye, 2004) and to perform skills such as dancing moves and building constructions with blocks. Young children are exposed to teaching, of Volume 6—Number 4 187 Teaching Is a Natural Cognitive Ability for Humans course. But we believe it is rare that they are taught how to teach. The third property is that teaching occurs without teachers being aware of its underlying logic. Research conducted on professional adult teachers’ mental models of others’ minds and of how learning occurs revealed that teachers were not aware of their models’ underlying logic (Strauss, 1993; Strauss & Shilony, 1994). Teachers could speak about what they did when teaching and they offered reasons why they did what they did. For example, teachers said that they broke up complex material so that it would be easier for their pupils to learn (Steiner, 2002; Strauss & Shilony, 1994). However, they could not clearly explain the cognitive principles underlying the procedures of reducing complexity of what was being taught nor did they speak about how that was connected to other understandings they had, such as pacing the lesson so that it does not cover material too quickly or teaching too much material. Teaching is distinct frommore general abilities to process or transmit information,which is the fourthproperty.Teaching is one ofmany forms of social learning (e.g., imitation, emulation, local stimulus enhancement contagion, etc.), and all concern the transmissionof knowledge fromone conspecific to another (Leadbeater & Chittka, 2007; Whiten, 2000). But teaching is unique in its psychological causality. At the heart of teaching is the intent to cause learning in someone else. No other kind of social learning has that at its core. In that sense, it is distinct from other abilities to process information. Developmental Reliability in Human Ontogenesis An additional criterion for teaching to be a natural cognitive ability is that it should be shown to be reliably developing. We elaborate on this criterion because it highlights the importance of including developmental studies in the domain of teaching. Two questions ought to be answered for making a determination regarding the developmental reliability of teaching. First, what does it mean that something is developmentally reliable? The answer provided by Cosmides and Tooby (n.d.) is that the ability under examination universally follows a similar developmental trajectory among children in the normative range (i.e., without serious developmental disabilities)who are exposed to a conventional social environment, no matter what that society might be. Second, what ability should we describe in order to claim that it is developmentally reliable? The answer for teaching is complex and not obvious. Strauss and Ziv (2012) noted that there may be four separate, yet intertwined, strands of cognitive abilities each of which follows a developmental course and all of which are related to teaching: (1) cognitive templates necessary for teaching but which are not specific to teaching, e.g., executive functions (Davis-Unger&Carlson, 2008b), (2) cognitive prerequisites not specific to teaching, e.g., language (Premack & Premack, 1996) and ToM (Strauss et al., 2002; Ziv & Frye, 2004), (3) cognitive prerequisites specific to teaching, such as understanding intentional and knowledge-related aspects of teaching, e.g., that teaching is an intentional act, that one can be a source of knowledge for someone else, and that teaching relies on the teacher’s beliefs about the learner’sknowledgeorunderstanding, and(4)actual teaching (Strauss et al., 2002; Ziv, Solomon, & Strauss, in press). In this article,we reviewonly the fourth developmental course, actual teaching, and restrict that to teaching strategies. This has two versions: the proto-teaching of episodic knowledge and the teaching of generalizable knowledge. The distinction between passing on episodic and generalizeable knowledge comes from work by Miller (2000) and Csibra andGergely (2009).Here-and-now information passed on in proto-teaching activities is episodic. For example, in the case where were we to ask you, the reader, what time it is, and were you to tell us, you would not have taught us the time. What you did was pass on information that is correct for that particular moment. Had we asked the same question at the same time to a friend in a different place, say Budapest, or to the same person at a different time, say 5 min later, the answers would be different. In contrast, information that is passed on in teaching activities is generic and generalizeable. That information goes beyond the here and now. For example, were a 5-year-old to ask you to teach her how to read the time on a clock, and were you to do that, you would be passing on information that is generalizeable and that goes beyond both the place and the moment. You would be teaching her. This distinction assists in portraying the developmental course of teaching. We now present the developmental story ordered by age. Nonverbal children age 1 seem to be capable of closing a knowledge gap. An experimenter ‘‘accidentally’’ pushes an object off a table in full view of a child and when the experimenter looks for it and cannot find it, she asks the child where it is (Liszkowski, Carpenter, Striano, & Tomasello, 2006; Liszkowski, Carpenter, & Tomasello, 2008). The infant points to its location. Their research indicated that infants at this age did not want the object for themselves nor did they want the adult to do something with the object. Liszkowski and his colleagues interpreted these findings as evidence for cooperative informing, via pointing, an ability that seems to be unique to humans. Here is how we interpret their findings, as they relate to teaching. First, there is a knowledge gap: the experimenter ‘‘does not know’’ where the object is and the infant does. And second, the infant points and, as a consequence, closes the knowledge gap. However, this is a case of proto-teaching because the knowledge being passed on is episodic. Now it is to the left of the table but later it might be to the right. Nevertheless, at least two parts of teachingmay be in evidence among 1-year-olds: perhaps ‘‘recognizing’’ a knowledge gap 188 Volume 6—Number 4 Sidney Strauss and Margalit Ziv and acting to close it, which may be an intentional act. Similar research was conducted by Akagi (2012). In the presence of 1-year-olds, an experimenter made amistake by trying to put a triangular-shaped object in a round hole, and the infants often pointed to the correct hole to put it in. Here, too, it might be the case that fundamental aspects of teaching are present at a very early age. The only study to assess 2 and 21/2-year-olds (Ashley and Tomasello, 1998) indicated that what they do cannot be considered teaching. However, the task they used to tap teaching was not simple. It required children to perform two actions in order to get an attractive sticker: moving a lever that moved a plate with the sticker on it in a plexiglass tube to a window and rotating a handle 180◦ which opened the window and allowed a child to remove the sticker. It is possible that a reason children at these ages did not teach was not because they couldn’t but because the task used had cognitivedemands that exceeded2 and21/2-year-old children’s information processing capabilities for teaching. More work is needed at these ages that extends Liszkowski et al.’s (2006, 2008) work. Several studies tapped children’s teaching from age 3. Children were taught how to play a game, and after they learned it, they were asked to teach a friend how to play the game so that they could play it together. For expository purposes, we now present a task used by Strauss et al. (2002). Later we use it to illustrate children’s teaching strategies. It is a board game whose purpose is to collect three flowers of different colors according to the rules of the game. Each child has a turn to roll a die that has four flowers of different colors on four of its six faces. The two remaining faces had a picture of a smile and a frown. A player can take a flower that is on the board when the upward face is a color of a flower. For example, if after rolling the die, the cube has a red flower facing up, the child who threw the die can take a red flower from the board. If the smile is the face that is up, the participant can take any color flower she wants. If a frown appears, the participant cannot take a flower from the board. A detailed description of the game is found in Strauss et al. (2002). Children age 3 engage in emergent teaching and do so despite their low performance on ToM tasks, as tested by classic false belief tasks and false belief tasks for teaching (Bensalah, Olivier, & Stefaniak, 2012; Davis-Unger&Carlson, 2008a, 2008b; Strauss et al., 2002). As teachers, 3 1/2-year-olds predominantly demonstrate how to play a game with little explanationaccompanyingthosedemonstrations.Others,who did not assess ToM, also reported that demonstration was the principal teaching strategy among 31/2-year-olds (Ashley & Tomasello, 1998; Ellis & Rogoff, 1982; Feldman, DevinSheehan, & Allen, 1976; Feshbach, 1976; Feshbach & Devor, 1969; Maynard, 2002, 2004; Wood, Wood, Ainsworth, & O’Malley, 1995). Explanations of the game’s ruleswere found among children at this age but were not the dominant teaching strategy. It was also found that when the learner errs, 3-year-olds often ignore these errors or correct them by performing themselves the required move instead of the learner. For example, in the above-mentioned board game, if the learner mistakenly picks a blue rather than a red flower, as indicated by the die, a 3-year-old teacher will typically ignore this mistaken move and will progress to throwing the die for his turn or replace the learner’s blue flower with the correct red one. This response reflects emerging awareness and monitoring of learners’ behavior. Hence, 3-year-olds recognize learners’ lack of knowledge and engage in peer teaching that relies on demonstration and entails some correction of learners’ behavior. This pattern correlates with 3-year-olds’ success in tasks that test recognition of knowledge gaps between teachers and learners (Ziv&Frye, 2004) but low performance ontasksthat testmorecomplexmentalaspectsof teaching,e.g., teachers’ beliefs about learners’ knowledge and the distinction between intentional teaching and learning by imitation (Ziv et al., 2008). Children age 5 teach by explaining the game’s rules, and that is accompanied by demonstrations (Bensalah et al., 2012; Davis-Unger & Carlson, 2008a, 2008b; Strauss et al., 2002). Children at this age also demonstrate the way the game is played, but the dominant teaching strategy is explanation. The majority of these children pass ToM tasks as tested by classic and teaching false belief tasks. It was found that when the learner errs, a 5-year-old teacher will typically repeat or remindher of a rule that has previously been explained, usually in a shortened version or by demonstration. For example, a 5-year-old teacher first explained the rule: ‘‘You have to throw thedie and if it shows red, youpick the redflower.’’ In response to the learner’s mistaken picking of a blue flower, the teacher explained and demonstrated: ‘‘Look, the dice shows red, so I pick a red flower, not blue, see?’’ Hence, 5-year-olds are better than 3-year-olds at monitoring the learning process and begin to adapt their teaching to the learner’s progress and errors. This pattern correlates with their better performance on false belief and intention tasks, as well as on teaching tasks. As for teaching tasks, 5-year-olds showed understanding of teacher’s beliefs regarding learner’s knowledge (Strauss et al., 2002; Ziv & Frye, 2004) and could distinguish between intentional teaching and learning by imitation (Ziv et al., 2008). Children age 7 engage in more systematic contingent teaching (Wood et al., 1995; Ziv et al., 2008). In this kind of scaffolded teaching, a teacher teaches a learner and as that learner’s knowledge state does or does not change (she gains more knowledge and becomes increasingly competent, or the knowledge state does not change, or the learner becomes confused andwhatwas seen as sure knowledge is now shaky), the teacher adjusts his teaching to his representation of that changing knowledge state.Wood,Wood&Middleton (1978) Volume 6—Number 4 189 Teaching Is a Natural Cognitive Ability for Humans noted that a rule for contingent teaching is: If the learner succeeds, when next intervening, offer less help. If the learner fails, take over more control when next intervening. Children age 7 were quite proficient at this. This is an on-line ToM and is an advance over the ToM found among 5-year-olds. Regarding children ages 9 and 11, Garbarino (1975) and Ludeke andHartup (1983) found teachers’ behaviors thatwere similar to those found among 7-year-olds. One exception was that 9-year-olds offered strategic advice regarding alternative choices available to the learner. In showingthat there seemstobeadevelopmental trajectory from proto-teaching to demonstration, to explanation and to contingent teaching, we presented each in general terms. Table 1 illustrates the variety and richness of teaching strategies children use over the course of development and can assist in refining analyses of children’s teaching in future studies. It shows, for example, that there are several kinds of demonstration and verbal explanations. The table does not include ages because, for themajor categories of strategies, there is a change in emphasis rather thananabsence at a certain age and then a presence at another. For example, for the tasks at hand,whichwere almost always teaching a peer how toplay a game, even the 3-year-old children explained here and there. However, it was not the dominant strategy; demonstration was. Similarly, among 5-year-olds, where explanation was the main teaching strategy, there were always demonstrations as well. We now jump to teaching among adults but note that there is surely more development concerning actual teaching between age 9 and adulthood. This is a lacuna in need of filling. Research on adults’ teaching explored the mental models about the mind and its dynamic workings which underlie how they teach. Previous studies focused on mental models of professionals regarding the workings of dynamic physical objects. An example is mental models engineers have about running a power plant. In acquiring knowledge about it, they learn a considerable amount about physics, engineering, mathematics, and so on. But when actually running the power plant, they do not necessarily consult that knowledge. Instead, they use shortcut rules they mentally constructed about how to keep the plant running as it changes dynamically, e.g., when there is an increase in pressure, reduce the temperature. These rules and their relations comprise mental models (JohnsonLaird, 1983; Norman, 1983; Seel, 2006). Studies on people’s mental models of teaching and learning initiated by Strauss and colleagues (Haim, Strauss, & Ravid, 2004; Mevorach & Strauss, 2012; Strauss, 1993, 2001, 2011; Strauss,Ravid,Magen,&Berliner, 1998; Strauss,Ravid,Zelcer, & Berliner, 1999; Strauss & Shilony, 1994) involved a move that changed the above. Instead of describing people’s mental models of dynamic physical objects in the environment, one can describe adults’ mental models of the dynamic workings of pupils’ minds when learning occurs. Unlike a power plant, nobody has seen amind. It is tacit, and there is little agreement about its structure and workings. Nevertheless, people seem to have mental models about the structure of minds and how learning takes place as a result of teaching. One can conceptualize two kinds of mental models: (1) espoused, which can be culled from the ways people speak about their teaching and (2) in-action, which can be inferred from peoples’ actual teaching. Strauss (1993, 2001, 2011) and Strauss & Shilony (1994) found that teachers have a direct transmission, source-recipient espoused mental model of the mind, how learning takes place there, and how teaching brings about learning. Knowledge is possessed by teacherswho stand outside the learners’ mind. Teaching is an engineering task of how to get the teacher’s knowledge inside children’s minds and to get it to stay there. This same espoused mental model was found for professional teachers and professionalswith the same level of university degrees who were not teachers, such as journalists. Teachers’ in-action mental model was revealed by videotaping and analyzing teachers’ teaching (Haim et al., 2004; Mevorach & Strauss, 2012; Strauss, 2011; Strauss et al., 1998; Strauss et al., 1999). Findings show that teachers’ inactionmentalmodel is comprisedof the following: (1) cognitive goals that teachers want their pupils to achieve, (2) cognitive processes that teachers think lead to these cognitive goals, (3) assumptions about how teaching in a particular way leads to these cognitive process that, in turn, lead to the cognitive goals, and (4) meta-assumptions about learning and teaching. These two mental models, the espoused and in-action, reflect adults’ fundamental folk psychology understandings about the mind and its dynamic workings when learning occurs there. Thefindings correspondwithOlson andBruner’s (1996) notion of teaching according to which teachers’ teaching reflect their assumptions and beliefs about learner’s mental abilities. They argue that the majority of adult teachers regard learners as capable of understanding and applying rules and accordingly teach mainly by explaining rules. Only a minority of teachers regard pupils as thinkers and engage them in conversations and activities in which they express and refine their opinions and ideas. The espoused mental model is common to adult teachers who teach inkindergarten throughuniversity, aswell as adults withno formal teaching training and experience. The in-action mental model is common to experienced and novice teachers who teach different subject matter and to those whose subject matter knowledge is extensive and deeply organized and those with impoverished and shallow subject matter knowledge organization. To summarize, teaching appears to be developmentally reliable. The trajectory is from proto-teaching among 1-year-olds who do not yet speak, to demonstrations among 3-year-olds, then explanations found in 5-year-olds, followed by contingent teaching among 7-year-olds and culminating 190 Volume 6—Number 4 Sidney Strauss and Margalit Ziv