We disagree with two of Rendell and Whitehead's assertions. Culturemay be an ancestral characteristic of terrestrial cetaceanancestors; not derived via marine variability, modern cetaceanmobility, or any living cetacean social structure. Furthermore,evidence for vocal behavior as culture, social stability, andcognitive ability, is richer in birds than Rendell and Whiteheadportray and comparable to that of cetaceans and primates.

The use of anecdotes is not a viable research strategy to studyanimal culture. Social learning processes can often be documentedwith careful quantitative analyses of observational data.Unfortunately, suggestions that killer whales engage in teaching areentirely based on subjective interpretations of qualitativeobservations. Thus, “evidence” of teaching in killer whales cannotbe used to argue for the occurrence of culture in marinemammals.

Rendell and Whitehead have drawn attention to some striking cetaceanbehaviour patterns. However, the claims for “culture” are premature.Weak examples of cetacean social learning do not, in sum, providestrong evidence for culture. Other terms, such as social learning,vocal learning, imitation, and tradition may be applied in somecases without resorting to more complex and controversial terms.

The target article provides stimulating evidence for culture incetaceans but does not provide a coherent theoretical framework. Weargue that a complex, adaptive systems approach not only can providesuch a framework but also can contribute advanced data analysis andsimulation methods. For humpback whale songs, we suggest theframework of “small-world networks” to model the observedspatio-temporal dynamics.

We agree with Rendell and Whitehead that cetaceans acquire knowledgefrom caretakers and peers, and that a clear understanding of thisprocess can provide insight into the evolution of mammaliancognition. The passive observational methods they advocate, however,are inadequate for determining what cetaceans know. Only byexperimentally investigating the cognition of cetaceans can we hopeto understand what they learn through social interactions.

Results of genetic analyses show that social groups of female andimmature sperm whales are comprised of multiple matrilines asevidenced by the presence of multiple mitochondrial (maternallyinherited) control region haplotypes. These data suggest: (1) asocial environment in which the transmission of culturalinformation, such as vocal dialects, is more likely to be horizontalor oblique rather than strictly vertical (mother-offspring) and (2)lead us to question the data presented to support gene-culturecoevolution.

Rendall and Whitehead overstate the weak evidence for social learningin cetaceans as a group, including the current evidence for vocallearning in killer whales. Ethnographic techniques exist to testgenetic explanations of killer whale calling behavior, andadditional captive experiments are feasible. Without such tests,descriptions of learning could be considered pseudo-scientific, adhoc auxiliary assumptions of an untested theory.

Defining culture as social learning means that culture is present inmany birds and mammals, suggesting that cetacean culture is not sospecial and does not require special explanation. Contrary to theirown claims, Rendell and Whitehead present culture as having variantforms in different species, and these forms seem inconsistentlyapplied and compared across species.

The types of cetacean cultural behavior patterns described (primarilyfood-related and communication-related) reflect a very differentresearch focus than that found in primatology, where dietaryvariation and food processing is emphasized and other potentially“cultural” patterns have (until recently) been relatively neglected.The lack of behavioral research in all but a few cetacean species isalso notable, as it mirrors a bias in primatology towards only a fewgenera.

Monkeys and apes, inhabiting variable environments and subjected toK-selection, exhibit cultural behavior transmitted horizontally andvertically, like cetaceans. Behaviors enhancing better health andnutrition, predator avoidance, or mate selection, can affectdifferential reproduction.Furthermore, dominance hierarchies andsocial status not only affect the transmission and acceptance of newbehaviors but they may also affect genetic inheritance.

We argue that the function of human culture is to clarify what peoplevalue. Consequently, nothing in cetacean behavior (or any otheranimal's behavior) comes remotely close to this aspect of humanculture. This does not mean that the traditions observed incetaceans are uninteresting, but rather, that we need to understandwhy they are so different from our own.

Social learning and imitation is central to culture in cetaceans. Thetraining technology used with cetaceans facilitates reinforcingimitation of one dolphin's behavior by another; the same technology,now widely used by pet owners, can lead to imitative learning insuch unlikely species as dogs and horses. A capacity for imitation,and thus for cultural learning, may exist in many species.

Assessments of cetacean behavior has been hampered by popularmisconceptions and mythic imagery. Rendell and Whitehead arguepersuasively for accepting the idea of cetacean culture. Approachesto “culture,” however, must resist positivist approaches thatreaffirm the observable. Culture is also comprised of “lacunae” whenorganisms choose to resist or avoid behavioral patterns.

Sociality may not be a defining feature of social learning. Complexsocial systems have been predicted to favour the evolution of sociallearning, but the evidence for this relationship is weak. In birds,only one study supports the hypothesis that social learning is anadaptive specialisation to social living. In nonhuman primates,social group size and social learning frequency are not correlated.Though cetaceans may prove an exception, they provide a useful groupwith which to test these ideas.

Rendell & Whitehead's minimalist definition of culture does notallow for the important gaps between cetaceans and humans. A morecomplete analysis reveals important discontinuities that may be moreinstructive for comparative purposes than the continuitiesemphasized by the authors.

Rendell and Whitehead state that marine ecosystems are more variablethan terrestrial ecosystems over time scales of months or longer.The marine environment is actually less variable than theterrestrial at these shorter time scales, and probably equallyvariable over centuries. This issue is important when consideringclaims that environmental variability affects benefits of sociallearning versus individual learning and genetics.

While cetaceans clearly show social learning in a wide variety ofcontexts, to label this as “culture” hides more than it reveals: weneed a taxonomy of culture to tease apart the differences ratherthan hiding them in a catch-all category.

While supporting the claim for culture in cetaceans, I suggest thatRendell and Whitehead's argument is potentially incomplete if basedsolely on ethnographic evidence. The notion of cetacean culture canalso be explored experimentally. I hope to complement the authors'assertion by discussing dolphin brain and behavioural researchfindings, thus contributing to a more holistic argument for culturein dolphins.

Rendell and Whitehead adopt a weak definition of culture to allow lowstandards of evidence for marine mammals, but they do not adequatelyrule out genetic factors or individual versus social learning. Theythen use these low standards to argue that some whales have uniquecultures only matched by humans. It would have been more helpful tospecify data gaps and suggest critical tests.

Recent evidence suggests imitation is more developed in somecetaceans than the authors imply. Apart from apes, only dolphinshave so far shown a grasp of what it is to imitate; moreoverdolphins ape humans more clearly than do apes. Why have suchabilities not been associated with the kind of progressive culturalcomplexity characteristic of humans?