150 Evolutionary Anthropology

ARTICLES

Do Animals Have Culture?

KEVIN N. LALAND AND WILLIAM HOPPITT

Culture is probably not rare in animals, although hard experimental evidence is some form of extra-genetic inheri- lacking. The strongest case for culture is found in the species most amenable to tance sufficient to be deemed cul- experimental manipulation, rather than in nonhuman primates. Human culture is tural. Thus, central to the issue of much more likely to be cumulative than animal culture, but the reasons for this are whether animals exhibit culture is not well established. At this point, there is no reason to assume that cumulative the problem of whether it is possible culture depends critically on teaching, imitation, language, or perspective-taking. to come up with a nonarbitrary def- Currently, animals are being judged according to stricter criteria than humans. inition of culture; one that will prove useful to a substantive proportion of interested researchers. We begin by WHY IS CULTURE RARE IN pologists seem to have given up on the seeking such a definition, and in do- ANIMALS? OR IS IT? notion of culture altogether.3,4 Ironi- ing so are guided by the twin criteria cally, and perhaps benefiting from of consensus and utility. In simple The concept of culture has become blissful ignorance of the problems, bi- terms, we are looking for the kind of a quagmire for the social sciences. In- ologists and students of animal behav- definition of culture that our profes- tuitively, we all know what it is and ior are currently giving unprece- sional colleagues will recognize as feel it to be important, yet it has dented attention to the “culture” of something with which they can proven virtually impossible for social other animals.5–7 This community has work. scientists to derive a consensual defi- witnessed an explosion of interest in The first step is to ask whether nition or to find a satisfactory means animal social learning and purported such a definition would be more use- 1,2 of operationalizing it. Indeed, after animal culture6,8 exemplified by a pro- ful if it were broad and minimalist, much painstaking self-analysis and liferation of books, conferences, and reflecting the continuity between hu- soul searching, many social anthro- papers dedicated to the topic.5,9–11 mans and other animals, or whether Whether or not the culture concept a narrow and exacting definition has utility for our own species, an in- that stresses human uniqueness creasing number of researchers are might have greater utility. We sus- Kevin Laland is a Royal Society University using it to explain both diversity and pect that to most biologically Research Fellow and Reader in Biology at minded researchers the answer to the University of St. Andrews, U.K. His transgenerational continuity of be- research encompasses a range of topics havior in animals. this question is immediately appar- related to animal behavior and evolution, In the absence of a satisfactory ent. A narrow definition, say some- particularly social learning, cultural evolu- tion, and niche construction. He is the and universally accepted definition thing like, “Cultures are systems of co-author (with Gillian Brown) of Sense of culture, most answers to the ques- linguistically encoded conceptual and Nonsense. Evolutionary Perspectives tion, “Do animals have culture?” phenomena that are learned through on Human Behavior (2002) published by Oxford University Press and (with John take on a hackneyed appearance. De- teaching and imitation, socially Odling-Smee and Marc Feldman) of Niche fine culture one way and it is the transmitted within populations, and Construction. The Neglected Process in Evolution (2003) by Princeton University exclusive province of human beings; characteristic of groups of people,” Press. E-mail: [email protected] another way, and a multitude of spe- is not likely to prove particularly Will Hoppitt is a BBSRC-funded PhD stu- cies are deemed worthy of the acco- useful, at least not to the issues ad- dent in the Department of Zoology at the University of Cambridge, U.K. His re- lade. At one pole of the spectrum, dressed in this article. This is not search includes empirical work on social many social anthropologists con- only because the answer to the ques- learning in birds and mathematical mod- tion of whether or not animals have eling of social learning processes. ceive of culture as so deeply shrouded in language, politics, mo- culture is a fait accompli; by defini- rality and institutions, that it would tion, they do not. But in addition, Key words: social learning, innovation, teaching, be grossly devalued were it attrib- this denotation would act as a bar- imitation, niche construction uted to dolphins or apes. At the other rier to understanding of the evolu- extreme, biologists Charles Lums- tionary roots of culture. It would en- Evolutionary Anthropology 12:150–159 (2003) den and Edward Wilson12 concluded courage the misguided view that DOI 10.1002/evan.10111 Published online in Wiley InterScience that in excess of 10,000 species, in- humans are (uniquely) unique, (www.interscience.wiley.com). cluding some bacteria, manifest which is only trivially true since all ARTICLES Evolutionary Anthropology 151 species are unique. No light would may vary from one population to the ture.5,10 While our particular beliefs be shed on how culture came into next. Thus, culture helps to explain are of little intrinsic value, we suspect existence, nor on human’s place in both continuity within groups and that many researchers that study ani- nature. More fundamentally, it diversity between groups. mal social learning would agree with would act as a barrier to the integra- These considerations lead us to us on this point. tion of the biological and social sci- the following working definition of To return then to our question, cul- ences. Premature, over-exacting dis- culture, toward which we hope a ture is as rare or as common among tinctions jeopardize the ability to see substantial proportion of research- animals as it is defined to be. On the relationships between culture-like ers interested in cultural evolution assumption that broad definitions lead phenomena in diverse taxa. would be sympathetic. Cultures are to greater utility and consensus than do Given the primitive state of knowl- those group-typical behavior pat- narrow ones, we have defined culture edge of animal cultures and the cur- terns shared by members of a com- according to its necessary conditions. rent malaise in human culture stud- munity that rely on socially learned This definition leads us to the view that ies, we suggest that a broader and transmitted information. (Later culture is not at all rare, although hard definition is likely to be more stimu- we will discuss addition characteris- experimental evidence of cultures lating. It may prove valuable to de- tics that some researchers regard as among animals, including nonhuman bate the criteria by which cultural essential to human culture.) primates, is currently lacking. and acultural species are distin- guished and perhaps, in time, even Which Animals Have Culture? WHAT IS UNIQUE ABOUT to make distinctions between differ- According to the preceding defini- PRIMATE CULTURE? ent classes of culture. But such dis- tion, which animals have culture? tinctions, if they are going to be in- A “Brainist” Bias? There are two kinds of answers to this formed, can be based only on question. The first kind is based exclu- A slightly (but only slightly) face- empirical data. A broad definition sively on hard experimental evidence. tious answer to the question of what is encourages such data to be collect- That is, for which species do we have unique about primate culture is that ed; a narrow definition does not. A reliable scientific evidence of natural primates uniquely benefit from the bi- broad definition has the additional communities that share group-typical ased anthropocentric perspective of advantages that it provides a com- behavior patterns that are dependent investigators who themselves, without parative framework with which to on socially learned and transmitted exception, are primates. Nonhuman investigate the evolutionary history information? The answer, which will primates are probably studied more of culture and that it encourages the surprise many, is humans plus a than most other species; the findings cross-fertilization of information handful of species of birds, one or two of these studies perhaps receive and methods between the biological whales, and two species of fish. greater attention than those of other and social sciences. Naturally, the No doubt many readers will find this species; and arguably lower standards definition could also be too all-en- conclusion disturbing, while primatolo- of evidence are required for primates compassing, so that “culture” be- gists will probably be up in arms. How than other species to be awarded comes unrecognizable and research- can we attribute culture status to fish some sophisticated psychological or ers become alienated. An overly and not chimpanzees? A full explana- social attribute, including culture. broad definition, like an overly nar- tion will follow, but the short answer is This bias, we suspect, stems in part row one, would have neither utility that for chimpanzees, as for other non- from a need dating back to Darwin for nor consensus. human primates, the hard evidence researchers to demonstrate continuity By what means can we settle on a that their “cultures” are socially learned of mental abilities between humans broad and useful definition of cul- is not yet there.13–15 Neither evidence of and other animals. The easiest and ture? One approach is to synthesize group-typical behavior patterns nor a most convenient means by which the array of disparate definitions of demonstration that the species is capa- such continuity can be demonstrated human culture and isolate the crite- ble of social learning is in itself strong is to find the most sophisticated cog- ria that are deemed essential by all, evidence for culture.16 nition in our closest relatives, the or at least most, parties.1,7 What The second kind of answer to the apes, slightly less impressive abilities then, when we strip it down to its question of which animals have cul- in monkeys, and increasing levels of barest bones, is the essence of cul- ture is a best guess. If we were to say stupidity as the degree of relatedness ture? One central criterion is that which animals we believe have cul- with ourselves diminishes. Reinforced culture is built upon socially learned ture, based on our knowledge of ani- by the observation that brain size and socially transmitted informa- mal social learning, observations of crudely fits this pattern (if we ignore tion. The term “culture” does not ap- natural behavior of animals, intuition, some inconvenient outliers such as ply to inherited genetic information and the laws of probability, we would cetaceans), many of us begin our re- or the knowledge and skills that in- say that many hundreds of species of search expecting to find that nonhu- dividuals acquire on their own. A vertebrate have culture. After all, man primates will be smarter than second is that this socially transmit- many vertebrates have been shown to other animals. From an evolutionary ted information can underpin group- be capable of social learning of a sort perspective, this bias is unjustifiable, typical behavior patterns, which that theoretically could maintain cul- since convergent evolution could 152 Evolutionary Anthropology ARTICLES manifestly have favored intelligence, first experiment, a sample of individ- Ideally, both experiments would be sophisticated cognition, or culture in uals from population A are introduced carried out on the same populations, quite distinct lineages. into population B, and vice versa, ide- which would potentially provide the Indeed, the compelling assumption ally at a formative age. The observa- most compelling evidence for culture. that species with larger volumes of the tion that the introduced animals The first experiment, in combination relevant brain structures will show adopt the behavior exhibited by mem- with alternative evidence that the be- greater cognitive sophistication is still bers of the host population is incon- havior in question cannot be learned disputed17 and remains largely un- sistent with an explanation in terms of asocially, or the second experiment, in tested. There is evidence among non- genetic differences between popula- combination with alternative evi- human primates for a link between tions and consistent with an explana- dence that there are no relevant ge- brain size and behavioral flexibility, as tion reliant on some form of learning. netic differences between popula- measured by frequency of reports of In the second experiment, population tions, would also provide satisfactory innovation, social learning, and tool A is collectively removed from its en- evidence for culture. use.18 However, this finding demon- vironment and replaced in the envi- To our knowledge these experi- strates that big-brained species ex- ronment of population B, which in ments have not been carried out in a hibit more social learning than small- turn is removed and replaced in the systematic and rigorous manner in brained species, and not that the any nonhuman primates. What is former employ more sophisticated more, in many cases, including for psychological processes than do the chimpanzees, it is difficult to envisage latter. Given its expense to that they will. Not only would such In sum, it is not clear that a big produce and maintain, experiments be extraordinarily expen- brain is necessary for culture as de- sive, and present enormous logistical fined earlier. If some fish exhibit cul- all of the neural tissue challenges, but many people would re- ture, as we will argue, then presum- that primates have in gard such manipulations as unethical. ably a large brain is not a requisite Consequently, the claim that nonhu- feature. Given its expense to produce abundance must have man primate populations exhibit cul- and maintain, all of the neural tissue some important ture rests exclusively on observations that primates have in abundance must function. Yet the fact of natural populations in situ. have some important function. Yet That there are population differ- the fact remains that it may have little remains that it may ences in the behavioral repertoires of to do with culture. Most people, our- have little to do with many primates is now well docu- selves included, believe that nonhu- mented. Best known are the distinct man primates have culture in spite of culture. Most people, tool-using and other group-typical be- the scientific evidence, rather than be- havior patterns of chimpanzee popu- cause of it. ourselves included, lations.19–22 Similar behavioral diver- believe that nonhuman sity consistent with a cultural Why There Is Better Evidence primates have culture in explanation is reported in several other primates, notably orangutans23 for Culture in Fish Than in spite of the scientific and capuchins.11,24 Primates evidence, rather than Differences in the behavioral reper- What evidence is needed to demon- because of it. toires of chimpanzee populations strate that a species exhibits culture, have become, for many, the jewel in according to our definition? Research- the crown of nonhuman primate cul- ers would need to establish that the tural traditions and the prime exem- animal concerned exhibits diverse plar of animal culture. Yet, in the ab- group-typical behavior patterns in environment of population A. If the sence of the kind of experimental natural populations and, in addition, introduced populations come to ex- evidence described, the case for chim- that these behavior patterns are un- hibit the same behavior as the former panzee culture rests entirely on a derpinned by social learning. The lat- residents, this would suggest that the plausibility argument. There is clear ter criterion requires ruling out alter- behavior results from shaping to di- experimental evidence that chimpan- native explanations for the behavioral vergent ecological conditions, and zees are capable of social learning.25 diversity, including that the popula- asocial learning could not be ruled Indeed, the evidence for seemingly tions exhibit genetic differences or out. However, the observation that the more complex forms of social learn- that variant ecological conditions introduced population exhibits group- ing, such as imitation, is as good in have resulted in different patterns of typical behavior that differs signifi- this species as in any other, humans asocial learning. cantly from that exhibited by the aside.26,27 Many of the group-typical In principle, ruling out these alter- former residents would be inconsis- behavior patterns, such as termiting, natives through experimentation is tent with an explanation based on eco- appear quite complex, and research- straightforward. Two experimental logical differences between environ- ers envisage that they would be diffi- manipulations are necessary. In the ments and consistent with culture. cult to learn alone. Some behaviors, ARTICLES Evolutionary Anthropology 153 such as hand-clasp grooming,28 ap- This problem is highlighted by a re- other so-called chimpanzee cultural pear arbitrary, even functionless, and cent analysis of chimpanzee ant dip- behaviors will eventually be ques- consequently seem to many to be un- ping by Humle and Matsuzawa.29 tioned, when careful scrutiny reveals likely candidates for behavior shaped Whiten and coworkers20 singled out that variation coincides with ecologi- by ecological contingencies (see also ant dipping as a particularly compel- cal differences? Will all thirty-nine of the primate social conventions de- ling case of chimpanzee cultural vari- Whiten and coworkers’20 behaviors scribed by Perry and Mason24). In ation where the differences between eventually be whittled away? Humle their seminal paper on chimpanzee communities are not related solely to and Matsuzawa’s29 analysis implies cultures, Whiten and coworkers20 whether the behavior is present or ab- that, had the experimental manipula- claim to have found “39 different be- sent, but where two different methods tions described been employed by havior patterns, including tool usage, of ant dipping were employed. In the translocating Gombe and Tai chimps, grooming and courtship behaviors, first or “pull through” method, re- we would have found Gombe-born [that] are customary or habitual in ported as customary at Gombe, a long chimps using the direct mouthing some communities but are absent in wand is held in one hand and a ball of method in Tai and Tai-born chimps others where ecological explanations ants is wiped off with the other. In the using the pull-through method at have been discounted.” For most in- second “direct mouthing” method, Gombe (that is, chimpanzees would terested researchers, the probability is customary at Tai, a short stick is held have failed the culture test), although that at least some of the group-typical in one hand and used to collect a we will not know for sure unless this behavior patterns reported for chim- smaller number of ants, which are experiment is actually done. Let us be panzees are learned socially. Yet to transferred directly to the mouth. Like clear: Neither Humle and Matsuza- our knowledge, in not one instance is Boesch and Boesch30 and McGrew19 wa’s29 analysis nor the hypothetical there irrefutable evidence that a natu- before them, Whiten and colleagues20 translocation experiment would rule ral chimpanzee behavior is socially were satisfied that ecological differ- out a cultural explanation for ant dip- learned. ences between the Gombe and Tai ping. However, while the alternative Indeed, there are serious grounds sites could not explain this variation. ecological explanation remains plau- for being suspicious of this plausibil- Furthermore, they concluded “it is sible there will always be a reason to ity argument, and senior figures in the difficult to see how such behavior pat- doubt the claim of culture. The trans- animal social-learning community re- terns could be perpetuated by social location experiment method is supe- main unconvinced.13–15 The main learning processes simpler than imita- rior because, where such experiments problem is that it is difficult to rule tion.” Humle and Matsuzawa29 inves- have been carried out and behavior out ecological explanations for the ob- tigated ant dipping among chimpan- consistent with cultural variation has served variation in behavior. Whiten zees of Bossou, in southeastern been observed, we are not left doubt- and coworkers20 consider the possibil- Guinea, a site at which both forms of ing whether unconsidered ecological ity of false-positive reports of cultural ant dipping are observed. They found factors might explain the variation. variation resulting from an “ecologi- that both the length of the tools and While primatologists are forced to cal explanation, for which absence is the technique employed were strongly resort to circumstantial evidence for explicable because of a local ecologi- influenced by the nature of the prey. culture (illustrated by the plausibility cal feature.” However, only three of Several species of ants were preyed arguments put forward by Boesch,21 sixty-five behavior patterns were actu- on, and these differed in their density, Whiten,22 and Perry and Mason24), ally removed because an ecological aggressiveness, behavior, and in other researchers are free to employ explanation could be given. Skeptics whether they were migratory or at more direct experimental methods. might be forgiven for questioning how nest sites. In situations in which ants Helfman and Shultz31 translocated seriously this alternative was consid- were abundant, aggressive, or had se- French grunts (Haemulon flavolinea- ered. For instance, in the case of a vere bites, the first method, with its tum) between populations and found population reported as not nut-ham- long wand, was employed, and was that while those fish placed into estab- mering, a skeptic would want to know probably associated with fewer bites lished populations adopted the same not only that there are suitable nuts than the second method, while in schooling sites and migration routes and hammers present, but also that other circumstances the direct as the residents, control fish intro- there are not alternative and more mouthing method was used. Skeptics duced into regions from where the readily accessed alternative sources of would find it easy to envisage that residents had been removed did not foods with equivalent nutritive and chimpanzees could individually be adopt the behavior of former resi- energetic properties that render nuts a shaped by biting insects to use the dents. Helfman and Shultz’s study comparatively less profitable food strategy that resulted in the fewest combines the procedures for experi- source. Similarly, the skeptic would bites. ment 1 and 2 described earlier in one want to know whether populations re- Humle and Matsuzawa’s29 findings of the most elegant demonstrations of ported to not fish for ants or termites suggest that an ecological explanation animal culture to date. Their study is are exposed to exactly the same spe- for ant dipping cannot be ruled out. If not quite the same as the symmetrical cies of ants and termites as those that confidence in this flagship example of experiment suggested because only do, and that the ants and termites be- cultural variation in chimpanzees one population was introduced into have the same way in each case. must be suspended, then how many the environment of the other. None- 154 Evolutionary Anthropology ARTICLES theless, this is strong evidence for cul- males.38 Furthermore, in the labora- (the inactivation of an allele depen- ture. Warner32 showed that translo- tory it has been demonstrated that dent on the parent it was inherited cated populations of Bluehead wrasse preferences for particular male traits from) could account for this pattern. (Thalassoma bifasciatum), a Carib- can be culturally transmitted between Likewise, an ecological explanation bean coral-reef fish, did not adopt the female birds.37 However, to our could be viable if sympatric pods use same mating sites as the populations knowledge it has never been demon- the same habitat in different ways. Of they replaced, but rather maintained strated that differences in female pref- course, such behavior itself could be distinct sites long after the manipula- erence at lekking sites in the wild are cultural but, alternatively, different tions, consistent with a cultural expla- cultural. Freeberg39,40 has gone one pods could have genetic predisposi- nation. This finding is a direct appli- step further toward demonstrating tions to use a habitat in a particular cation of the experiment 2 method. culture in cowbirds (Molothrus ater). way. Rendell and Whitehead42 pro- Combined with the observation that Male cowbirds in two distinct popula- vide what we feel is stronger evidence high levels of mixing during the early tions exhibit different types of court- for culture in Humpback whales life of this species mean that reef pop- ing behavior and the females in each (Megaptera novaeangliae). During the ulations are not subject to significant population prefer to mate with males breeding season all males in a popu- genetic differentiation, this finding who court in the local fashion. Free- lation sing, at any one time, “nearly provides acceptable evidence of cul- berg found that juvenile birds of a the same song,” but this song evolves tural variation. common genetic background, when over time, changing even within a In the light of a burgeoning litera- housed with adult birds from one or breeding season. The rapidity of ture on learning and cognition in fish, other of the populations, adopted the change would seem to rule out a ge- in which it is quite apparent that the courting behavior and mating prefer- netic explanation, although demo- abilities and complexity of social be- ences of the population they were graphic changes or common external havior of this group have been under- housed with (see also Freeberg and influences might account for song estimated,33 we should hardly be sur- coworkers41). This case introduces the evolution; however, neither would prised that two species of fish should possibility of an unusual transmission seem likely to account for the ob- provide some of the best evidence for mechanism in which juvenile males’ served synchronized change in song animal cultures. A recent review of courting behavior is shaped by adult between many individuals. This evi- fish cognition33 reveals a rich experi- females and juvenile females’ prefer- dence that can be mustered for birds mentally derived databank reporting ences are shaped by the behavior of and cetaceans is at least as strong, and social intelligence (social strategies, adult males. However, we cannot en- in some cases stronger, than that for Machiavellian intelligence, coopera- tirely rule out an ecological explana- nonhuman primates. tive hunting) as well as sophisticated tion, since during the experiment each The primate-centric brainist bias is foraging skills, tool use, cognitive population was housed at a different well illustrated by a reconsideration of maps, and the construction of com- location. However, this would seem to one of the most well-known cases of plex artifacts. Fish have compara- be unlikely to account for the ob- animal social learning, that of sweet- tively small brains and are not re- served differences in behavior. In- potato washing by Japanese ma- nowned for their intelligence or deed, there is a large body of evidence caques.43 For a novel cultural behav- learning abilities. Yet there is now un- that the vocal signals of many avian ior pattern to spread through an equivocal evidence from laboratory species may be cultures (reviewed by animal population two processes are and field studies that a variety of dif- Freeberg36). However, as far as we are necessary: There must be the initial ferent species of fish are capable of aware, there is no case where an eco- inception of the behavior, which we social learning, including learning logical explanation for population dif- refer to as innovation, and its spread how to find food, which foods to eat, ferences in vocal repertoire can be to others by social learning, which is how to recognize predators, how to ruled out entirely. diffusion. Imo, the famous originator avoid predators, and perhaps even There is also some suggestive evi- of potato- and wheat-washing, has be- with which fish to mate (see Brown dence for culture in several species of come celebrated as a great innovator, and Laland34 for a review). If fish can cetaceans. Killer whales (Orcinus and therefore as an initiator of ma- do all this with small brains, then orca) show a number of pod-specific caque “culture.” For instance, E.O. maybe simple mechanisms could ac- behavior patterns, including foraging Wilson44 in Sociobiology and John count for primate behavior. specializations, migration patterns, Bonner45 in The Evolution of Culture When it comes to hard experimen- and vocal dialects. Rendell and White- in Animals both refer to Imo as “a tal evidence for culture, the only tax- head42 argue that ecological explana- genius,” while Kummer and Goodall46 onomic group that comes close to fish tions can be ruled out, since sympat- describe her as “gifted.” In the 1960s, is birds, for which the cultural creden- ric pods develop different pod-specific food washing seemed an unlikely nat- tials of song variants and mating pref- behaviors. They also consider a ge- ural behavior for monkeys. Hence, erences are reasonably well-estab- netic explanation unlikely since pods these behavior patterns were regarded lished in several species.35–37 It has are matrilineal, so that inheritance of as novel and intelligent. When the been observed that female quail behavior would have to be principally food-washing behavior spread it was (Coturnix japonica) at lekking sites from the mother. However, sex-linked described as “preculture” or “pro- prefer to mate with particular genetic traits or genomic imprinting toculture,” the clear implication being ARTICLES Evolutionary Anthropology 155 that it could be regarded as an analog, found that some fish consistently ner and, indeed, in the absence of any or perhaps even a homolog, of human solved foraging problems, not because evidence that it was her personality to culture. they were hungrier, faster swimmers, which her success can be attributed. With the benefit of hindsight, and in more active, or healthier, but because Under such circumstances, Bertha the light of accumulated scientific of some stable personality character- seems to have greater claims to the knowledge, Imo’s achievements and istic. To our knowledge there is no “genius” plaudit than does Imo. those of her troop members appear equivalent experimental evidence for To return to the question that heads more modest. It seems that food personality differences in problem- this section, it remains conceivable washing is a stable feature of ma- solving ability in any species of non- (indeed, we view it as quite likely) that caque behavior, and Imo’s innovation human primate. Many primate stud- nonhuman primates exhibit a greater involved the application of an estab- ies report differences in performance richness of cultural variation com- lished behavior pattern to novel between individuals, but none have pared with other taxa. What is striking 47 foods. Experimental studies have re- ruled out alternative explanations for about the repertoires of chimpanzees vealed that food washing is learned individual differences (for example, is the multifaceted nature of their pur- relatively easily by monkeys and can ported culture, in comparison with motivational factors). Hence, the vari- become common in a troop through which the cultural repertoires of fish ation cannot reliably be interpreted as processes other than the imitation of a and birds appear impoverished (see reflecting differences in personality. rare “creative genius.”47 There is no Boesch21 and Whiten22). Yet we see no The one study of which we are aware evidence that food washing spread reason to value weak evidence for pri- that sought such evidence, Fragaszy through Imo’s troop by imitation, mate cultures over strong evidence for teaching, or any unusually sophisti- culture in birds and fish. The best ev- cated form of social learning. Indeed, idence for culture is found in the spe- there are grounds for concern that it There is no evidence cies that are most amenable to exper- may even have been an artifact of that food washing imental manipulation. human provisioning.14 The lauding of spread through Imo’s We would not wish primatologists Imo as a genius and the use of the to forsake the study of population terms “protoculture” or “precultural” troop by imitation, variation in behavior just because to describe this case of social trans- teaching, or any they cannot prove that the differences mission were widely accepted only be- are cultural. How can primatologists cause the behavior was reported for a unusually sophisticated get around their inability to carry out primate. Would the scientific commu- form of social learning. translocation experiments? We have nity have accepted the same terminol- two suggestions. First, appropriate ogy to describe social learning in rats Indeed, there are data collection would allow the feasi- or budgerigars? grounds for concern that bility of genetic and asocial learning The members of our laboratory it may even have been explanations to be assessed and re- have carried out a detailed analysis of jected if the probability that they can the factors that underpin innovation an artifact of human account for the data is unrealistically in fish, birds, and monkeys, and found provisioning. small. Laland and Kendal54 describe that state-dependent factors (sex, size, simple statistical methods that would social rank, hunger level) account for allow alternatives to the social-learn- much of the variation. These and a ing explanation to be dismissed if re- host of other findings48–52 lead us to searchers have access to good pedi- the view that the adage, “Necessity is and Visalberghi’s53 investigation of in- gree and diffusion data (see also the mother of invention” probably ap- novation in capuchin monkeys, failed Fragaszy16). For instance, experimen- plies to most animal innovation. Cer- to find it. tal data from laboratory or captive tainly when it comes to foraging inno- Let us give the single fish that con- studies estimating the probability of vation, more often than not the sistently solved our maze problems asocial learning could also be used to innovators are those individuals with faster than others the name of Bertha. assess the probability that a particular the greatest motivation to find food. We do not yet know whether it is ap- pattern of diffusion or level of inci- To our knowledge, no consideration propriate to describe the personality dence is explicable in terms of asocial has been given to the possibility that trait that Bertha exhibits as intelli- processes. Imo was the first individual to apply gence or creativity as opposed to, say, Second, researchers could carry out food washing to sweet potatoes boldness, perseverance, or nonconfor- a careful analysis of the small number merely because she was among the mity but at least we know that there is of cases in which individuals have dis- hungriest individuals in the troop, yet some feature of Bertha’s make-up that persed, transferred between popula- this would seem a real possibility. Our makes her an effective problem solver. tions, or reintroduced into established experiments have found evidence for Gifted highbrow intellectual qualities, populations, or in which populations personality differences in the prob- as opposed to less salubrious person- have been introduced into new areas lem-solving ability of guppies, a small ality traits, have been attributed to (see Boesch21 for examples). A careful freshwater fish. Laland and Reader48 Imo in an essentially arbitrary man- analysis of these natural translocation 156 Evolutionary Anthropology ARTICLES experiments may eventually shed light ventions; and history, manifest in a server, B, at some cost or at least on primate cultures. This, after all, is stable, transgenerational, cumulative without obtaining an immediate one reason that we can be confident culture. He suggested that while chim- benefit for itself. A’s behavior that humans have culture. panzee behavioral traditions exhibit thereby encourages or punishes B’s strong evidence for universality, there behavior, or provides B with experi- was only weak evidence for uniformity ence, or sets an example for B. As a WHAT (IF ANYTHING) IS or history. Tomasello hypothesized result, B acquires knowledge or UNIQUE ABOUT HUMAN that, compared with chimpanzee “cul- learns a skill earlier in life or more CULTURE? tures,” the greater universality, unifor- efficiently than it might otherwise mity, and history of human culture are do, or that it would not learn at all. If we are to identify unique factors a manifestation of higher fidelity of in- The use of this rigorous definition of human culture, then we must com- formation transmission among hu- pare its quantifiable characteristics has lead to the view that information with purported animal culture on a mans, reflecting differences in the psy- transmission in animals is inadvertent level playing field. On what grounds chological mechanisms employed. In and not actively facilitated by the 14 13 can human and animal cultures be accord with Galef, Tomasello ar- demonstrator. Yet there are other distinguished? gues that human culture alone is reliant means by which animals actively ad- Intuitively, the sheer complexity of on teaching and imitation, together just their behavior in a manner that human behavior seems to divide us with language and perspective taking. facilitates learning in another animal, from the rest. Driving cars, building In reconsidering these hypotheses, which do not satisfy definitions of houses, and writing articles on culture we begin with a focus on teaching. We teaching. The opportunity teaching of would be orders of magnitude more felids provides one example;57 the complex than the most intricate chim- food calls of adult callitrichid mon- panzee tool-using behavior. Yet to ...we see no reason to keys,58,59 the food and alarm calls of suggest that what renders human cul- value weak evidence mother hens,60 and the stranding of ture unique is its complexity merely whale calves42 may be others (see also begs the question of why human cul- for primate cultures over the “stimulating” and “facilitating” of ture should be more complex. Com- strong evidence for infants by chimpanzee mothers de- plexity is a consequence of the fea- scribed by Boesch21). Irrespective of tures of humanity that render our culture in birds and fish. their goal or knowledge of the learn- culture unique, and not itself such a The best evidence for er’s abilities, the demonstrators ap- feature. culture is found in the pear to function as teachers in these Galef14,15 has suggested that the examples. Indeed, leading researchers similarities between animal and hu- species that are most have called for teaching to be rede- man culture may be more superficial amenable to fined in functional terms.61 There than real because animal and human seems to be a gradation of teaching- culture rest on different processes. experimental like phenomena, yet this has received Galef places emphasis on teaching manipulation. little attention from students of ani- and imitation, which he regards as mal social learning. As a result, there underpinning human culture but un- is no satisfactory classification important to animal social learning. scheme or terminology to describe an- Conversely, Galef regards purported imal teaching. Just as the reputations animal cultures as heavily dependent know of no single widely accepted of animals have suffered from the on local and stimulus enhancement, case in any nonhuman animal that treatment of teaching as an all-or- and other comparatively simple pro- satisfies contemporary definitions of nothing phenomenon, so our own cesses that mediate social learning. 56 teaching, while human children are species has benefited. Human teach- Boyd and Richerson55 make a similar taught vast amounts of cultural ing is not required to satisfy the strict argument. While these researchers knowledge. At first sight, it is difficult have not specified precisely how criteria applied to animals, and to conceive of a clearer dichotomy in teaching and imitation are deemed to “teaching” in common parlance when the intellectual abilities of animals render human culture more stable, applied to humans is a much broader and humans. Yet this dichotomy is complex, or cumulative than pur- concept altogether. It would be inter- vulnerable to criticism from both ported animal cultures, to many the esting to know what proportion of hu- hypothesis is both intuitive and com- sides. Purported instances of animal man teaching actually satisfies Caro pelling. teaching are evaluated according to and Hauser’s56 definition. Teachers Tomasello13 identified three impor- strict criteria. For example, Caro recognize the value for children of op- tant characteristics of human culture: and Hauser56 defined teaching as fol- portunity teaching, and many believe universality, the observation that some lows: that allowing children to discover and behavior patterns are exhibited by vir- An individual actor A can be said learn for themselves is frequently tually everyone in the society; unifor- to teach if it modifies its behavior more productive than more classical mity, or widespread conformity to con- only in the presence of a naı¨ve ob- forms of pedagogical instruction. ARTICLES Evolutionary Anthropology 157

(Boesch21 and Whiten22 make similar than do other forms of social learning. strators is reported in animals as di- arguments.) Heyes65 argues compellingly that the verse as guppies, pigeons, and rats.67 The same point can be made with fidelity of cultural transmission is in- Conformity may be a characteristic regard to imitation, where once again dependent of the learning processes feature of animal social learning, as it animals are being judged according to that individuals employ to acquire a is for humans. If, as Tomasello13 a stricter definition than is applied to behavior; rather, it depends on the notes, chimpanzee populations are humans. The fact that humans are ca- patterns of reinforcement that indi- characterized by individuality rather pable of imitation is no reason to re- viduals subsequently experience. As than uniformity, then it is chimpan- gard it as the sole or even principle we argued earlier, much of what in zees that are “unique.” process of information transmission common parlance is described as hu- A stronger case can be made for underlying human culture. Perhaps man imitation is more appropriately history: Much human culture is char- children learn how to get around their termed program-level imitation, yet acterized by the “rachet effect,”13 with neighborhood, where the grocery this will not necessarily have high fi- an increase in the complexity or effi- store is, and which locations are dan- delity. ciency of technology over time (see gerous by following parents around, Moreover, universality and unifor- also Alvard4). There is little direct ev- like reef fish. Perhaps humans acquire mity do not seem to be features on idence within animal populations of many skills through emulation, like which human culture and animal so- additive, incremental improvement in other primates.62 Even when they cial learning can be distinguished. behavior or technological advance. “imitate,” do humans really copy each Field evidence for purported animal Evidence for cumulative evolution in other’s exact motor patterns? We sus- culture is provided either by different tool manufacture has recently been pect that “program-level imitation,”63 behavior in two or more populations reported in New Caladonian crows where the underlying organization of of the same species (group-typical be- (Hunt and Gray, personal communi- behavior is copied, rather than the havior) or by an increase in the fre- cation). There are one or two primate surface details, is more common. quency of a novel behavior in a single behavior patterns, such as leaf clip- While we are open to the idea that population (diffusion). In the former ping and nut cracking by chimpan- teaching and imitation may be more case, unless the differences between zees, and sweet-potato washing by important to human culture than to the populations’ behavioral reper- Japanese macaques, for which a case animal social learning, we would like toires are substantive they are likely can be made for racheting,21,68,69 but to reserve judgment until humans and either to be unnoticed or to be attrib- they remain speculative and conten- other animals are evaluated according uted to random fluctuations, observer tious. to the same standards. It is conceiv- error, or ecological differences. It is However, before we congratulate able to us that a careful analysis of when behavioral variants are custom- ourselves for isolating a distinctive teaching and imitation in humans ary in one population and not in oth- and unique feature of human culture, would leave our species looking con- ers that a claim of cultural variation is it is worth reflecting on two points. siderably less distinct. most likely to be put forward. Simi- First, one can look at millions of years For Tomasello,62 teaching and imi- larly, diffusions typically are identi- of our ancestors’ lithic technologies tation are significant because they are fied only when a significant propor- and many ethnographic studies of tra- deemed to generate a higher degree of tion of the population has acquired ditional peoples without seeing evi- fidelity of transmission among hu- the novel variant. Not all diffusions dence for racheting. Oldowan and mans than is found for animals. We reach universality in animal popula- Acheulean hand-axe technologies see two problems with this argument. tions, but not all human traditions are lasted without significant changes for First, we know of no evidence that universal either. The claim for animal over one million years each.70 Cul- there is greater fidelity associated culture usually rests on the recording tural evolution in other primates with human than animal culture. Cav- of behavioral variants that are univer- could be cumulative at rates equiva- alli-Sforza and coworkers’64 survey of sal, near universal, or at least com- lent to those in early Homo or even patterns of cultural transmission re- mon within a population. preindustrial societies, and we might ported average parent-offspring corre- Mathematical analyses have found not notice. Second, the assertion that lations for habits, entertainments, and that in most circumstances where nat- animals do not exhibit racheting is sports to be as low as 0.07, 0.16 and ural selection favors reliance on social tantamount to the argument that all 0.13, respectively. Boyd and Richer- learning, conformity is also fa- nonhuman group-typical behavior son55 reported parent-offspring corre- vored.55,66 These theoretical consider- patterns could have been invented in lations ranging from 0.34 for attitudes ations lead us to expect most animal their final form by a single innovative towards feminism to 0.8–0.94 for po- social learning to be conformist, with individual. Many researchers, partic- litical party affiliation, with values of individuals commonly adopting the ularly those familiar with the complex around 0.5–0.6 being common. Can behavior of the majority. This predic- behavior of apes, find this difficult to we really be sure that humans have tion has empirical support across a believe. Yet it matters little whether higher cultural fidelity than animals? broad array of species, and social nonhuman animals exhibit no rachet- Second, to our knowledge, there is no learning in which the probability of ing, slow ratcheting, or very little ra- experimental evidence that imitation adopting a pattern of behavior in- cheting: There clearly is a fundamen- results in higher fidelity transmission creases with the proportion of demon- tal quantitative difference in the 158 Evolutionary Anthropology ARTICLES degree to which contemporary human are humans less content to satisfice other animals because humans ex- and animal cultures are cumulative. (that is, make do with a functional but hibit a greater capacity to construct If purported animal cultures do not sub-optimal solution to a problem) and regulate those conditions, that is, exhibit cumulative evolution, then than other animals? If so, why? for “niche construction.”16,77,78 Regu- why not? Tomasello’s and Galef’s em- Second, in hierarchical animal soci- latory behavior such as tracking phasis on human’s capacity for teach- eties a disproportionate amount of in- game, storing food, and building shel- ing and imitation provides one sort of novation appears to be carried out by ters damps out temporal variation in explanation, but from the preceding low-status individuals.50,51 Low-rank- conditions and resources to ensure discussion it is apparent that this has ers often occupy the periphery of the that cultural behaviors maintain their only limited support. What is more, social group, are attended to less than utility from one generation to the explanations that posit the historical high-ranking individuals, have less in- next.77,78 While the capacity for niche emergence of a special character fluence on group behavior, tend to construction is universal to living (teaching, imitation, language) as crit- move away or are behaviorally inhib- creatures, human niche construction ical to the evolution of human culture ited when approached by higher- is extraordinarily powerful, in part as do little to stimulate empirical re- ranking group members, and are a consequence of our culture. Perhaps search. These explanations betray an more vulnerable to scrounging than what is unique about human culture underlying assumption that only a are high-ranking individuals. All these is that, through niche construction, unique human characteristic can ex- characteristics would mitigate against cultural transmission has become plain unique human culture. We do the diffusion of innovations generated self-reinforcing, with transgenera- not accept this logic. Not that we by low-rankers. Conversely, in more tional culture modifying the environ- think capabilities such as teaching, egalitarian societies social learning ment in a manner than favors ever imitation, and language are unimpor- appears commonly to exhibit confor- more culture. tant to the evolution of human cul- mity, and experimental studies have ture, but rather that an emphasis on found that conformity makes it less ACKNOWLEDGMENTS them puts the cart before the horse. likely that animals will adopt novel When it comes to the issue of what is behavior.67 Human societies obvi- Research supported by a Royal So- unique about human culture, we see ously are not more egalitarian or less ciety University Research Fellowship two key questions: Why do so few an- conformist than animal societies. (In- to KNL and a BBSRC studentship to imal innovations spread and why is so deed, Henrich and McElreath66 de- WH. much human culture stable and trans- scribe conformity and prestige learn- generational? While we are not yet ing biases as characteristic of REFERENCES able to provide clear answers to these humanity.) So how are these inhibit- questions, we see no barriers to their ing effects circumvented? Does lan- 1 Durham WH. 1991. Coevolution: genes, cul- ture and human diversity. Stanford University empirical investigation, and believe guage allow innovators to advertise Press. such research would contribute sig- their inventions? Does trade give them 2 Kroeber AL, Kluckman C. 1952. Culture: A nificantly to understanding of what an incentive to do so? Do social con- critical review of concepts and definitions. Cam- makes human cultures unique. tracts reduce the likelihood of their bridge, MA. 3 Kuper A. 2000. If memes are the answer, what Reader (personal communication) exploitation? Again, there are clear is the question? In: Aunger R, editor. Darwinising found that of 606 cases of learned be- opportunities for empirical investiga- culture: the status of memetics as a science. Ox- havioral innovation in nonhuman pri- tion. ford: Oxford University Press. 4 Alvard M. 2003. The adaptive nature of cul- mates only 16% spread to a second Why is so much human culture sta- ture. Evolutionary Anthropology 12:3:136–149. individual, and an even smaller pro- ble and transgenerational? Purported 5 Heyes CM, Galef BG. 1996. Social learning in portion spread throughout the group. animal cultures are largely reliant on animals: the roots of culture. London: Academic Parallel observations, with many in- short-lived horizontal transmis- Press. 73,74 6 Shettleworth SJ. 2001. Animal cognition and novations but comparatively few dif- sion, while humans appear to ac- animal behavior. Anim Behav 61:277–286. fusions, have been made in bird pop- quire large amounts of information 7 de Waal F. 2001. The ape and the sushi mas- ulations.71 This seems to contrast with from the parental generation.64,75,76 ter. London: Penguin. human innovation, where many, al- These findings suggest that the lineage 8 Galef BG Jr, Giraldeau L-A. 2001. Social in- fluences on foraging in vertebrates: causal mech- though still by no means all, advanta- leading to Homo sapiens has been se- anisms and adaptive functions. Anim Behav 61: geous innovations spread.72 What are lected for increasing reliance on verti- 3–15. the barriers to the spread of innova- cal and oblique cultural transmission. 9 Zentall TR, Galef BG Jr. 1988. Social learning: psychological and biological perspectives. Hills- tion in animals? Theoretical analyses imply that a shift dale, NJ: Erlbraum. First, we suspect that one reason toward increased transgenerational 10 Box HO, Gibson KR, editors. 1999. Mamma- cumulative culture appears rare in an- culture reflects greater constancy in lian social learning: comparative and ecological 55 perspectives. Cambridge: Cambridge University imals is that there are comparatively the environment over time, yet there Press. few natural circumstances where it is no evidence that ecological environ- 11 Fragaszy D, Perry S, editors. n.d. Traditions would pay an individual to adopt a ments have become more constant in nonhuman primates: models and evidence. more advanced means of doing some- over the last few million years. Cul- Chicago: University of Chicago Press. In press. 12 Lumsden CJ, Wilson EO. 1981. Genes, mind thing when a simpler solution works ture may be adaptive for a larger and culture. Cambridge: Harvard University well. A prime question seems to be, range of conditions in humans than Press. ARTICLES Evolutionary Anthropology 159

13 Tomasello M. 1994. The question of chimpan- biological themes and variations. Cambridge: 58 Roush RS, Snowdon CT. 2000. Quality, quan- zee culture. In: Wrangham R, McGrew W, de Cambridge University Press. tity, distribution and audience effects on food Waal F, Heltne P, editors. Chimpanzee cultures. 36 Freeberg TM. 2000. Culture and courtship in calling in cotton-top tamarins. Ethology 106:673– Cambridge: Harvard University Press. p 301–317. vertebrates: a review of social learning and trans- 690. 14 Galef BG Jr. 1992. The question of animal mission of courtship systems and mating pat- 59 Roush RS, Snowdon CT. 2001. Food transfer culture. Hum Nat 3:157–178. terns. Behav Processes 51:177–192. and development of feeding behavior and food- 15 Galef BG Jr. n.d. Social learning: promoter or 37 White DJ, Galef BG. 2000. “Culture” in quail: associated vocalizations in cotton-top tamarins. inhibitor of innovation? In: Reader SM, Laland social influences on mate choices of female Ethology 107:415–429. KN, editors. Animal innovation. Oxford: Oxford Coturnix japonica. Anim Behav 59:975–979. 60 Nicol CJ, Pope SJ. 1996. The maternal feeding University Press. In press. 38 Kirkpatrick M. 1987. Sexual selection by fe- display of domestic hens is sensitive to perceived 16 Fragaszy D. 2003. Making space for tradi- male choice in polygynous animals. Ann Rev chick error. Anim Behav 52:767–774. tions. Evol Anthropol 12:61–70. Ecol Syst 18:43–70. 61 Shettleworth SJ. 1998. Cognition, evolution 17 Macphail EM. 1982. Brain and intelligence in 39 Freeberg TM. 1996. Assortative mating in and behavior. New York: Oxford University vertebrates. Oxford: Clarendon Press. captive cowbirds is predicted by social experi- Press. 18 Reader SM, Laland KN. 2002. Social intelli- ence. Anim Behav 52:1129–1142. 62 Tomasello M. 1990. Cultural transmission in gence, innovation and enhanced brain size in 40 Freeberg TM. 1998. The cultural transmission the tool use and communicatory signalling of primates. Proc Natl Acad Sci USA 99:4436–4441. of courtship patterns in cowbirds, Molothrus chimpanzees? In: Parker S, Gibson K, editors. 19 McGrew WC. 1992. Chimpanzee material cul- ater. Anim Behav 56:1063–1073. Chimpanzee cultures. Cambridge: Harvard Uni- ture: implications for human evolution. Cam- 41 Freeberg TM, Duncan SD, Kast TL, Enstrom versity Press. bridge: Cambridge University Press. DA. 1999. Cultural influences on female mate 63 Byrne RW. 1998. Learning by imitation: a 20 Whiten A, Goodall J, McGrew WC, Nishida T, choice: an experimental test in cowbirds (Molo- hierarchical approach. Behav Brain Sci 21:667– Reynolds V, Sugiyama Y, Tutin CEG, Wrangham thrus ater). Anim Behav 57:421–426. 721. RW, Boesch C. 1999. Cultures in chimpanzees. 42 Rendell L, Whitehead H. 2001. Culture in 64 Cavalli-Sforza LL, Feldman MW, Chen KH, Nature 399:682–685. whales and dolphins. Brain Behav Sci 24:309– Dornbusch SM. 1982. Theory and observation in 21 Boesch C. 2003. Is culture a golden barrier 382. cultural transmission. Science 218:19–27. between human and chimpanzee? Evol An- 43 Kawai M. 1965. Newly-acquired pre-cultural 65 Heyes CM. 1993. Imitation, culture and cog- thropol 12:82–91. behavior of the natural troop of Japanese mon- nition. Anim Behav 46:999–1010. 22 Whiten A, Horner V, Marshall-Pescini S. keys on Koshima islet. Primates 6:1–30. 66 Henrich and McElreath, 2003. Evolutionary 2003. Cultural panthropology. Evol Anthropol 44 Wilson EO. 1975. Sociobiology. Cambridge: Anthropology 12: 12:106–108. Harvard University Press. 67 Day RL, Macdonald T, Brown C, Laland KN, 23 van Shaik C. n.d. In: Fragaszy D, Perry S, 45 Bonner JT. 1980. The evolution of culture in Reader SM. 2001. Interactions between shoal editors. Traditions in nonhuman primates: mod- animals. Princeton: Princeton University Press. size and conformity in guppy social foraging. els and evidence. Chicago: Chicago University 46 Kummer H, Goodall J. 1985. Conditions of Anim Behav 62:917–925. Press. In press. innovative behavior in primates. Philos Trans R 68 Boesch C. 1993. Toward a new image of cul- 24 Perry S, Manson J. 2003. Traditions in mon- Soc London, Series B 308:203–214. ture in wild chimpanzees? Behav Brain Sci 16: keys. Evol Anthropol 12:71–81. 47 Visalberghi E, Fragaszy D. 1990. Do monkeys 514. 25 Whiten A, Custance DM. 1996. Studies of im- ape? In: Parker ST, Gibson KR, editors. Lan- 69 McGrew WC. n.d. Twenty lessons learned itation in chimpanzees and children. In: Galef guage and intelligence in monkeys and apes. from labors in cultural primatology. In: Joulian BG Jr, Heyes CM, editors. Social learning in an- Cambridge: Cambridge University Press. p 247– F, editor. How the chimpanzees stole culture. imals: the roots of culture. New York: Academic 273. Culture and meaning in apes, ancient humans Press. p 291–318. 48 Laland KN, Reader S. 1999. Foraging innova- and modern humans. Kluver/Plenum Press. 26 Whiten A, Custance DM, Gomez J, Teixidor P, tion in the guppy. Anim Behav 52:331–340. 70 Foley R, Lahr MM. 2003. On stony ground: Bard K. 1996. Imitative learning of artificial fruit 49 Laland KN, Reader SM. 1999. Foraging inno- lithic technology, human evolution and the emer- processing in children (Homo sapiens) and chim- vation is inversely related to competitive ability gence of culture. Evolutionary Anthropology 12: panzees (Pan troglodytes) J Comp Psychol 110:3– in male but not in female guppies. Behav Ecol 71 Lefebvre L. 2000. Feeding innovations and 14. 10:270–274. their cultural transmission in bird populations. 27 Whiten A. 1998. Imitation of the sequential 50 Reader SM, Laland KN. 2001. Primate inno- In: Heyes C, Huber L, editors. The evolution of structure of actions by chimpanzees (Pan troglo- vation: sex, age and social rank differences. Int J cognition. Cambridge: MIT Press. p 311–328. dytes) J Comp Psychol 112:270–281. Primatol 22:787–805. 72 Rogers EM. 1995. Diffusion of innovations, 28 McGrew WC, Tutin CEG. 1978. Evidence for a 51 Reader SM, Laland KN. n.d. Animal innova- 4th ed. New York: Free Press. social custom in wild chimpanzees. Man 13:234– tion: an introduction. In: Reader SM, Laland KN, 73 Galef BG Jr. 1988. Imitation in animals: his- 251. editors. Animal innovation. Oxford: Oxford Uni- tory, definition, and interpretation of data from 29 Humle T, Matsuzawa T. n.d. Ant dipping versity Press. In press. the psychological laboratory. In: Zentall TR, among the chimpanzees of Bossou, Guinea and 52 Day RL, Coe RL, Kendal JR, Laland KN. n.d. Galef BG Jr, editors. Social learning: psycholog- some comparisons with other sites. Am J Prima- Neophilia, innovation and social learning: a ical and biological perspectives. Erlbaum. p tol. In press. study of intergeneric differences in callitrichid 3–28. 30 Boesch C, Boesch H. 1990. Tool use and tool monkeys. Anim Behav. In press. 74 Laland KN, Richerson PJ, Boyd R. 1993. An- making in wild chimpanzees. Folia Primatol 54: 53 Fragaszy DM, Visalberghi E. 1990. Social pro- imal social learning: towards a new theoretical 86–99. cesses affecting the appearance of innovative be- approach. Perspect Ethol 10:249–277. 31 Helfman GS, Schultz ET. 1984. Social tradi- haviors in capuchin monkeys. Folia Primatol 54: 75 Hewlett BS, Cavalli-Sforza LL. 1986. Cultural tion of behavioral traditions in a coral reef fish. 155–165. transmission among Aka pygmies. Am Anthropol Anim Behav 32:379–384. 54 Laland KN, Kendal JR. n.d. What the models 88:922–934. 32 Warner RR. 1988. Traditionality of mating- say about animal social learning. In: Fragaszy D, 76 Aunger R. 2000. The life history of culture site preferences in a coral reef fish. Nature 335: Perry S, editors. Traditions in nonhuman pri- learning in a face-to-face society. Ethos 28:1–38. 719–721. mates: models and evidence. Chicago: Chicago 77 Laland KN, Odling-Smee J, Feldman MW. University Press. In press. 33 Bshary R, Wickler W, Fricke H. 2002. Fish 2000. Niche construction, biological evolution, cognition: a primate eye’s view. Anim Cogn 5:1– 55 Boyd R, Richerson PJ. 1985. Culture and the and cultural change. Behav Brain Sci 23:131– 13. evolutionary process. Chicago: The University of 175. Chicago Press. 34 Brown C, Laland KN. 2001. Suboski and 78 Odling-Smee J, Laland KN, Feldman MW. Templeton revisited: social learning and life 56 Caro TM, Hauser MD. 1992. Is there teaching 2003. Niche construction. The neglected process skills training for hatchery reared fish. J Fish Biol in nonhuman animals? Q Rev Biol 67:151–174. in evolution. Princeton, NJ: Princeton University 59:471–493. 57 Caro TM. 1980. Predatory behavior in domes- Press. 35 Catchpole CK, Slater PJB. 1995. Bird song: tic cat mothers. Behavior 74:128–148. © 2003 Wiley-Liss, Inc.