THE CULTURAL ORIGINS OF HUMAN COGNITION Michael Tomasello HARVARD UNIVERSITY PRESS Cambridge, Massachusetts London, England 1999 ACKNOWLEDGMENTS {Individual human beings are able to create culturally significant arti- {acts ony if they recive significant amounts of assistance from other ‘human beings and social institutlons In my ease, {was able to wete this bookwhatever is faults and however lite its cultural sig- pifeance—only because I received direct asistance fom the follow ing people an institations (and, of course, indirect assistance from all the ater people over the past 2500 years of Westen civilization ‘who have thought and seriten about the basic puzzles of human | «cognition, Financial support fora one-year leave of absence, during which the majority of the ook was writen, was provided by the Guggen- ‘heim Foundation, by Emory University (Dr. Steven Sanderson, Dean (of Emory College), and by the Max Planck Gesellschaft. Support for -y empirical research over the past decade oe so has ben provided by the Spencer Foundation, the National Science Foundation (Ani ral Behavior Section), and the National Insite of Child Health and Human Development express my deepest gratitude to all of these people and institutions. am ever hope that they feel they got their money’s worth have benefited immensely from discussions of many ofthe is: sues inthis book witha numberof friends and colleagues OF special Importance are discussions Ihave had with Philippe Rocha, Josep (Call, Malinda Carpenter, Nameera Akhtar, Gina Cont-Ramsden, lena Lieven, Tricia Strano, Holger Diesel, Nancy Budi, and ‘Ann Kruger. All ofthese people also read some or all ofa deft of the manuscript and proved extremely sel commentary. 1am also grateful to Michael Cole and Katherine Nelson, who reviewed the manuscript for Harvand University Press and provided very helpful and timely feedback as wel Capra 1 Me Toma Prt inte Ur sate Amer Ura of Cons Cate ation Dat wazaee 2005-10 Finally, would lke to thank Katharina Habedl and Anke Fstr for all of their editorial and other assistance on the home front in Leipeig and Elizabeth Kaoll and Camille Sith for al of ther eito- Fal and other assistance at Harvard University PresCONTENTS 1A Puzzle anda Hypothesis 1 2 Biological and Cultural Inheritance 13 5. Joint Atention and Cultural Learning 56 4 Linguistic Communication and Symbolic Representation 9 5 Linguistic Constructions and Event Cognition 13 6 Discourse and Representational Redescription 161 7 Cultural Cognition 201 Refrences 219 Index 241 A PUZZLE AND A HYPOTHESIS Al tert skirts of id eb eh por fied ia Somer in Africa, sometime about 6 millon years ago, rou tine evolutionary event, a population of great apes became repro- luctively isolated from is conspecifics. This new group evolved and split into stil other groups, leading eventually to several diferent species of bipedal ape ofthe genus Austalpthocs. lof these new species eventually dled out except one that survives unl about 2 zilon years ago, by which time i had changed so much that it ‘eed not usta new species designation but a new genus designe tion, Homo. Compared with i austalopthecine forebears—who were four feet al with apesized brains and no stone tols—Homo ‘was larger physically, had a larger brain, and made stone tools. Be {ore long, Homo began to tava the globe widely, although none of its enly forays out of Africa suceeded in establishing any popula tions that survives permanent Then, somewhere stil in Africa, sometime about 200000 years ‘ago, ane population of Homo began on anew aod diferent evolu tNonary trajectory. It began living in new ways in Africa and then spread out across the world, outcompeting al ethee populations of Homa and leaving descendants that are known today as Homo sapiens {oe Figre 1-1). The individual ofthis new species had a numberof ‘new physical characteristics, ineluding somewhat larger brains, but most striking were the new cognitive skills and products they created: “They began to produce a plethora of new stone tools adapted 40 specific ends with each population ofthe species eating its own fook-use “industry”—resuting eventually in some pepula- tions creating such things as computerized manufacturing processes. + They began to use symbols to communicate ad to structure ‘heir social ives, including not only Linguistic symbols but also tsi symbols inthe frm of stone carvings and cave paint ingy—resuling eventually in some populations creating such things as written language, money, mathematical notation, ond ar They began to engage in new kinds of social practices and onga- izations, including everything from the burying ofthe dead ‘ceremonially to the domestication of plants and animals—result- Ing eventually in some populations creating such things as or- ‘malized religious, governmental, educational, and commercial institutions. ‘The basic puzzle this. The 6 milion years that separates human beings from other great apesiss very short ime evolionaiy, ith rmodeen humans al chimpanzees sharing something on the order ‘0f 98 percent oftheir gonetic material—the same degree of related ‘essa that of other sister genera such as lions and tigers, hoses and 2abras, and rats and mice (King and Wilson, 197). Our problem is thus one of ime. The fc is, there simply has aot been enough time for normal proceses of biological evolution involving genetic vara- tion and natural selection to have eeated, one by one, each of the cognitive skills necessary for modern humans to invent and sine tain complex tool-use industries and technologies, complex forms of ‘symbolic communication and representation, and complex socal or ‘ganizations and institutions. An the puzzle is only magrifiod if we lake seriously current research in paleoanthropology suggesting that {@) forall but the last 2 milion years the human lineage showed nosigns of anything other than typical great ape cognitive sil and () the fist dramatic signs of speciesunique cognitive skills ‘emerged only in the Ist one-quaeter ofa milion Years with modern. “Homo sapien Foley and Lahr, 1997; Klein, 1985; Swinger and MeKle, 1996). ‘Ther s only one posible solution to this puzae. That, theses ‘only one known biological mechanism that could bring about these kinds of changes in behavior and cognition in so short a time— ‘whether that time be thought of a5 6 million, 2 million, or onequar- terof« million years, This biological mechanism is social or cult transmission, which works on time scales many orders of magnitude faster than those of organic evolution. Broadly speaking, cultural transmission isa moderately common evolutionary proces that en- ables individual organisms to save much time and effort, not to mention isk, by exploiting the already existing knowledge and sills of conspecifics, Cultural transmission includes such things a8 fledgling bids mimicking their spesies-typical song from pacent, rat pups eating only dhe foods eaten by their mothers ants lating food by following the pheromone tails of conspecifics, young chin ppanzecs leaming the tool-use practices ofthe ads around them, And man childeen acquiring the linguistic conventions of others in their sccial groups (Mundinger, 1980; Heyes and Galt, 196), Hov- ever, despite the fact that all of these processes may be grouped under the general bri of cultural transmission, the precise beha- ioral and cognitive mechanisms involved inthe different cases are ‘numerous and diverse including everything from parents eliciting fied action pattems from thee offspring to transmission of kills by Iniatve leaning and instruction —which suggests the posbity of significant subtypes of eulural tanamission processes (Tomasll 1980; 1991), One reasonable hypothesis, then, is thatthe amazing site of cognitive skills and products displayed by modern humans isthe result of some sort of speces-unigue mode or modes of cul tural ransmission, ‘The evidence that human beings do indeed have species unique modes of cultural transmission is overwhelming, Most important, the cultural traditions and artiocts of human beings accumulate ‘modifications aver time in a way that those of other animal specs do not—so-called cumulative cultural evolution. Basically none of the most complex human artists or social practices including tool symbolic communication, ad socal institutions were invented once and forall at single moment by any one indvidal ‘or group of individuals Rather, what happened was that some ini vidal or group of individuals first invented a primitive version of ‘the artifact or practice, and then some later user or users made a modification an “improvement” that others then adopted perhaps ‘without change for many generations at which point some other in- dividual or group of individuals made another mosliication, which ‘was then learned and used by others, and so an over Mistrial ne in what has sometimes been dubbed “the ratchet effec” (Tomasell, Kage, and Ratner, 1988), The proces of cumulative cltual evel tion requires not ony creative invention but also, and just as impor: tantly, faithful soil transmission that can work as a ratchet to re vent slippage backward-so that the newly invented artifact of practice preserves its new and improved form at lest somewhat ‘aittully unt a further modification or improvement comes along, Pethaps surprisingly, for many animal species it isnot the creative component, but eather the stabilizing ratchet component hats the lficult feat. Thus, many nonhuman primate individuals regularly produce intelligent behavioral innovations and novelties, but then their groupmates do not engage in the kinds of socal lamin that ‘would enable, overtime, the cultural ratchet to do its work (Kum -mer and Goodall, 1985), ‘The basic fats thus that human beings are able to poo their com nitive resources in ways that other animal species are not. Accond- ingly, Tomasello, Kruger, and Ratner (1983) distinguished human cultural Ieaming from more widespread forms of soil lesening, ‘dntiving thee basic types: imitative leaning, instructed learning, ‘nd collaborative learning. These thee types of cultural learning are made posible by a single very special form of social cognition, namely, the ability of individual organisms to understand con- specifics as beings like thensles who have intentional anc mental lives like their own. This understanding enables individuals to indus 6 Imagine themselves “in the mental shoes” of some other person, 30 that they can learn not just ro te other but Hough the athe. This understanding of others as intentional beings lke the self crucial in human cultural learning berause cultural artifacts and social pra ticesexemplied protatypcaly by the use of tools and linguistic ‘symbols invariably point beyond themselves to other outside enti ties: tools point tothe problems thy are designed to solve and ln- _uistic symbols point to the communicative situations they are de ‘signed to represent. Therefore, to socially ler the conventional use fof a tool ora symbol, children must come to understand Why, to- ‘ward what outside end, the other person is using the oo or symbol, that is to say, they must come to understand the intentional sgaii cance ofthe tol use or symbolic practice—what it “for” what ‘we, the users ofthis tool a symbol do with i Processes of cultural leaning are especially powerful forms of socal leaning because they constitute both (a) especially faith forms of cultural transmission (creating an especially powerful cul tural ratchet) and (b) especially powerful forms of soial facts for developing children because they embody the ways that previous generations of human beings in a scial group have found Fr useful to categorize and construe the world for purposes of inter personal communication. For example, in different communicative ° situations one and the same object may be construed asa dog. an an- Jnl, a pet, ora pest one and the same event may be construed 35 ‘running, moving, Hesing, or surviving: one and the same place may be construed asthe coast, the shore, the beach, othe sand—all de ‘pending on the communicative goals ofthe speaker. As the cil masters th linguistic symbols of her culture she thereby acquices| ‘the ability to adopt multiple perspectives simeltaneously on one and ‘the same perceptual situation. As perspectvally based cognitive representation, then, linguistic symbols ae based not on the ecord- Ing of ciret Sensory or motor experiences as are the cognitive rep- resentations of other animal species and human infants, but rather ‘on the ways in which individuals choose to construe things out ofa ‘number of other ways they might have constued them as embed ied inthe other available linguistic symbols that they might have chosen, but did not Linguistic symbols thus fee human cognition ‘rom the immediate perceptual situation not simply by enabling ref ference to things outside this situation (“displacement ”; Hocket, 194), but rather by enabling multiple simultaneous representations ‘oF each and every indeed ll possible, perceptual situations Later, a children become more skilful with theie native language dltional possibilities for construing things in diferent ways open, up. For example, natural languages contain cognitive resources for ‘avtoning the world into such things as events and their partii= ‘pants—who may play many and various roles in these events—and {or forming abstract categories of event and participant types, More ‘over, natural languages also contain cognitive resources for const ing whole events or situations in terms of one another, that sor creating the various kinds of analogies and metaphors that are so Important in adult cogaition—such as seeing the atom as sola s)s- tem, love as a journey, or anger as heat (Lakot, 1987; Getner and Markman, 197; see Chapter 5). Also, childeen’s growing sil of linguistic communication enable thers to participate in complex dis- «course interactions in which the explicitly symbolized perspectives of ineratants lash and so must be negotiated and resolved. These kinds of interactions may lead children to begin to construct some- ‘thing like a theory of mind of their communicative partners, and, in » some special cases of pedagogical discourse to internalize adul structions and so begin to sel-regulate and to reflect on thelr own, thinking perhaps leading to some types of metacogition and zep- resentational redesripion (Karmilot#-Smith, 1992). The interalza- tion of discourse interactions containing multiple, conflicting per spectves may even be identified with certain types of uniquely naman, calogical thinking proceses (Vygotsky, 1978), In this book—for which the foregoing may be seen as a kind of [précis attempt to spell out this general line of argumentation in Some detail. That i, my specific hypothesis is that human cognition tas the species-unique qualities it oes because + Phylogenctialy: modern human beings evolve the ability #0 iceniy” with conspecifics, which led toan understanding of them as intentional nd mental beings like the sl + Historical this enabled neve forms of cultural leaning and | _uistic symbols (and analogies ane metaphors constructed fom these symbol) and (internalize certain types of discourse i teractions into sills f metacognition representational re- description, and dlalogic thinking 1 should emphasize at the outset that my’ focus is only on the specier-unique aspects of human cognition. Of course human cogni= Hin is in large measure constituted by the kinds of things that ap- pear as chapter headings in traditional Cognitive Peychology text books: perception, memory, attention, categorization, and soon. But these are all cognitive processes that human beings share with other Primates (Tomasello and Call, 197; Tomasello, 1988), My account ‘here simply presupposes them, and then focuses in Vygotskian fash-fon on the kinds of evolutionary, historical, and ontogenetic processes that might have transformed these fundamental skills nto the special version of primate cognition that is human cognition. 1 should also emphasize that I will dal with the biological and histor- ical processes involved in the evolution of human cognition only briefly and somewhat indivectly—mainly because the event of terest took place deep in the evolutionary and historical past and out information about them is very poor (Chapter 2). On the ther hand, [will focus in some detail on human cognitive ontogeny—about "which we know a good eal through several decades fcrct obser vation and experimentation—and the processes by which human children atively exploit and make use ofboth their biological and cultural inheritances (Chapters 3-8), Unfortunately, in tosay’s intellectual climate my argument may be taken by some theorists to be an essentially genetic ane: the socia-cognitive adaptation characteristic of modern humans i a [ind of “male bullet” that diferentates human beings from other ‘primate species ut this isan erroneous view that basicaly ignores ll ofthe sciakcultral work that must be done by individuals and _groups of nividual, in oth historical and ontogenetic time, ore te uniguely human cognitive skills and product. From an histori cal perspective, a quarter ofa million years is avery long time dur ing which much may be accomplished culturally, and anyone who has spent time with young chien knows how many leaning ex petiences may take place within the course of several years—or even Several days of several hours of continuous, active engagement ‘withthe environment, Any serious inquiry into human cognition, therefore, must include some account of these historical and onto” _enetic processes, which ate enabled but nat in ny way determined by human beings’ biological adaption fora spacial form of social cognition. Ince, my-central argument in tis book is that it i these processes, not any specialized biological adaptations directly that Ihave done the actual workin creating many, if nt al ofthe most distinctive and important cognitive products and. processes ofthe species Homo sapiens. And itis worth noting inthis context that ake ing these processes seriously enables us to explain not only the uni- 2 BIOLOGICAL AND CULTURAL INHERITANCE But hres thing ao te prc of proves contin oF bina sent fin the ut decane tt pres, ‘The overarching fact of the organic world i evolution by means of natural selection. A key element inthis process i biological inheri= tance, by means of which an organism inherits the basic Bruplor of its forebears, along, with its implications for perceptual, behavioral, and cognitive functioning. But fr all mammalian species including all primate species, much of the ontogeny by’ means of which this ‘Balan comes into being takes place while the developing organism 4s interacting with its environment. The relatively long period of imamaturity in which this interaction takes pice is ofcourse a very Fishy life-history strategy, as it means that offspring are totally dependent on ane of mare parents for food anc protection fom predators for some time, The compensating advantage of along, immaturity is that it enables ontogenetic pathways that incorporate significant amounts of individual learning and cognition, svhich Typically result in moce flexible behavioral and cognitive adapta- tions. Flexible behavioral/cognitive adaptations closely attuned tothe local environment are especially useful for organisms ‘whose populations live in diverse environmental niches, or whose cenvironmental niches change relatively rapidly ever time (Bruner, 1073). versal features of uniquely human cognition—such as the creation and use of material, symbolic, and institutional artifacts with cc ‘mulated historios-but aso the particularities of particular cultures, ‘each of which has developed for Itself via these same historical and ‘ontogenetic processes a variety of culturally unique cognitive skill land products during the post several dozen millennia of human, history In some animal species, the developing organism individually ac- ‘uires information not onty fom ts physical environment but also ftom its socal envionment —oe from aspects of ite physical environ- rent that have been modified in important ways by conspecifics. For example, a alluded to above, some hie species acquire their speciestypical song by listening to the song oftheir parents, and ‘some insets ate able to find food on their first day in the external fenvironment Because they know instinctively how to follow the [pheromone trails laid down by conspecifics (Mundinger, 1980; Heyes and Gale, 1996). In its broadest definition, as used by many evolutionary biologists, this process is eallod eultueal wansmission, for cultural inheritance, and it produces cultural waditions. Recent recognition of the importance of cultural transeission for many a= tal species has led fo the creation of Dual Inheritance Theory, i ‘which the mature phenotypes of many species are seen to depend ‘on what they inherit rom thei forebears both biologically an cul= turally (Boyd and Richerson, 1985; Durham, 199). Human beings of cours, are the prototypical species for Dual In- hreritance Thooey, as normal human development depends crucially ‘on both biological and cultural inheritance. My particular claim is that in the cognitive geal the biological inhertance of humans is very much like that of other primates. Theze jst one major difer fence, and that the fact that human beings “identify” with con- specifics more deeply than clo other primates. This identification is not something mysterious, but simply the process by which the Jhuman child understands that other persons are beings like her solfin a way that inanimate objects ae not, for example—and so she sometimes tres to understand things from their point of view. During early ontogeny in a process that willbe spelled out in more lel in later chapters, the child comes to experience herself as an Intentional agent—that i, a being whose behavioral and afentional steatogies are organized by gotls—and so she automatically ses ‘other beings with whom she identities in these same terms. Later in “ontogeny, the chill comes to experience herself asa mental agent thats, being with thoughts an belies that may differ from those6 ‘of other people aswell as frm realty—and so fom that time on she ‘will ee conspecifics in these new terms, For purposes of exposition refer to this process generally a8 "understanding others as inten- onal (or mental) agents (lke the self.” This one cognitive difterence has many cascading effects because it makes possible some new and uniquely powertl forms of uleral hertance, Understanding other persons as intentional agents lke ‘the self makes possible both (a) processes of saciogeness by means ‘of which multiple individuals collaboratively crentecultral artifacts and practices with accumulated histories, and (b) processes of cul ‘ual learning and internalization by means of which developing in- slviduals lesen to use and then internalize aspects of the collabora tive products created by conspeciis. This means that most if not all of the speciessunique cognitive skills of human beings ae not «fue toa unigue biological inheritance dvclly, but ather result fom. 1 variety of historical and ontogenetic processes that are set into ‘motion by the ane uniquely human, biologically inherited, eogaitive capacity Biological Inheritance ‘Human beings are primates. They have the same basic sense organs the same basic body plan and the same basic brain plan as all other primates, Therefore, if we are attempting to characterize the evalu Hlonary bases of human cognition, we must begin with primates in ‘genera Inthe current context there are two questions of central t= portance (a) How does the cognition of primates dlr fom that of ‘other mammals? and (b) How does the cogaition of humans dite from that of other primates? My answers to these to questions il bbe based onthe resazch of Tomasello and Call (197), who provide _more detailed analyses of the elvan empirial studies and hearst cal arguments as well as a more complete st of references, It must be acknowledged at the outset, of course, that other answers to these {questions are aso possible (se, for example, Byrne, 1985, for some sliferent views) 6 ‘Manama and Primate Cognition All mammals ive in basically the same sensory-motor world of per: ‘manent objects aerayed ina representational space; primates, nclud- ing humans, have no special shils in this regaed. Moreover, many mammalian species and basically all primates cognitively represent the categorical and quantitative relations among objets as well: These cognitive skills are evidenced by their ability to do sch things as + remember “what” is“where" in their lecal environment, 4, ‘which rit age in sehich tres (at what times) + take novel detours and shortcuts in navigating through space; + fellow the visible and invisible movements of abject (Le, pass rigorously controlled Pagetin object permanence tets~ some Stage 6) + categorize objects on the basis of perceptual similarities; * understand and thus match small nmerosities of objets; + use insight in problem solving. ‘And much evidence suggests that mammals do not acquiee these lis in some behaviorstie connecting of simul and responses, of ia Some simple form of eote memory, but eather actually come #0 comprehend and cognitively represent spaces and objects and cate {ories and quantities of ebjects) in ways that enable creative infer ‘ces and insightful problem saving Similarly, all mammals lve in basically the same social word of individually recognized conspecifis and their vertical (dominance) and horizontal atfiliative) relationships, and they have the ability 0 predict the behavior of conspecifics in many situations based on a variety of ues an insights. These cognitive sis are evidenced by thei ability todo such things a= + recognize individuals in their social groups; + form disc elationships with other individuals based on such things as kinship, feiendship, and dominance rank; + predict the behavior of individuals based on such things as thee emotional state and thei direction of locomotion; + use many types of socal and communicative strategies out ‘compete groupmates for valued resources: + cooperate with conspecifics in problem-solving asks and in forming social coalitions and alliances; + engage in various forms of social leningin which they len ‘valuable things fom conspecifics ‘Aad again much evidence suggests that mammalian individuals do ‘ot at blindly socially, but rather actually comprehend and cog tively represent what they are doing when they interact with their _groupmates in these various complex ways There is one exception to ths overall cognitive similarity among -mamnals, however, and thatconcers primates‘ understanding ofre- lational categories, which manifest itself in both the social and phys ‘cal domains Inthe sccial domain, primates, but not other mammals, understand something ofthe thied-party socal elationshps that hold among ther individuals for example, they understand suc things as ‘the kinship and dominance relations that thi parties have with one another. Thus, primates are selective in choosing thei coalition part ‘ers selecting as an aly, for instance, an individ whois dominant to their potential adversary—indcatng their understanding of the ‘lative dominance ranks of these 90 individuals, They also seek ret ‘ution for attacks against themselves not just on the attacker, but lsoinsome circumstances onthe aacke’skin—in hiscase evident dng an understanding of third-party kinship elatons. And there is even some evidence thal primates understand whole categories of ‘thied-party seca lationship azoss different individuals for exam ple, many diferent instances of the relationship “mother