A closer look at communication among our closest relatives

I am writing this while conducting fieldwork in Zambia, with only very limited access to the internet. Therefore, I could not read already existing posts and the corresponding responses and discussions, and some of the issues I will refer to might have been raised by others already. However, as a researcher interested in the gestural and facial communication of great apes, I want to offer some comments and facts from a comparative perspective on human communication and language evolution.

As a more general comment, I specifically liked the way Thom Scott-Phillips navigates the reader through this book by providing definitions of different terms, particularly of those often causing confusion when being used by scholars of different disciplines, as well as summaries of the most important facts of each chapter together with a brief outlook about what the reader can expect in the following chapter. Furthermore, although I am interested in potential precursors of human language in other primates, thus favoring a continuous approach to language evolution, I agree that researchers interested in language evolution sometimes compare “pears and apples”, since the behaviors of interest are not correctly (or too broadly) defined. Sometimes there is a tendency to focus on similarities between humans and other primates, while at the same time, differences might be neglected. However, I will point out later why I do not agree with some of the conclusions Thom Scott-Phillips draws from comparative research, with special focus on his chapter 4, dedicated to the origins of ostensive communication.

The nature of ostensive signal

Thom Scott-Phillips defines ostensive signals as “signals that express communicative intentions, and hence informative intentions” (p. 9). While the informative intention represents what the signaler wants to communicate, the communicative intention conveys the information that the signaler wants to communicate. He uses the example of tiling a cup (“I want more coffee”) plus establishing eye-contact (“I want to communicate my intention”) and then comments that “typically, both types of intention are expressed in one and the same behavior, such as the tilt of the coffee cup”. Based on this example, I am not sure about the nature of ostensive signals: On the one hand, Thom Scott-Phillips suggests that both intentions are expressed by one behavior (tilting the cup), but then refers to the importance of eye-contact to signal the communicative intention to a specific recipient and to make clear that one intends to communicate. Without this signal, Thom Scott-Phillips continues, the recipient would not realize that “tilting the cup” was directed at him.

Why am I so picky about this? If we are not clear about what exactly ostensive signals are and which forms they can take in humans (visual gestures? facial expressions? eye gaze? tactile communication? body postures/movements? vocalizations?), then it is also not clear what exactly we are looking for in nonhuman great apes (or other primates) and how we should interpret findings from comparative research. To give an example: Chimpanzee males use a “penis offer” to communicate that they want to copulate with a female. To make sure that the female perceives this visual behavior, the males combine it with a “leaf-clipping” behavior, which causes a sound to attract the female’s attention. Thus, while the “penis offer” refers to the “what”, the “leaf-clip” conveys the intention to communicate. One of course could argue that the males only want to change the female’s behavior, but not her mental representations. Furthermore, I am aware that this example only covers the expression, but not the recognition of ostensive-inferential communication (see p. 86), and that particularly in the gestural (in contrast to vocal) modality, we still know very little about great apes’ understanding of others’ communicative (and informative) intentions (in contrast to vocalizations, as highlighted in the book). Still, I think it is important to point out which types of ostensive signals we would expect to observe in other primates. It is also important to define clearly why we would consider some behaviors, but not others as potential ostensive signals, but not others. Gestures, and some vocalizations, are considered, but not facial expressions, which are barely mentioned throughout the book. Facial expressions might not be used intentionally, as are gestures and some vocalizations; however, despite the common notion that facial expressions might merely express internal, affective states and are not intentionally used, this aspect has never been studied systematically by applying the suggested criteria for intentional use to facial communication in nonhuman primates.

Imagine a chimpanzee approaching another one and using an “arm raise” gesture (ritualized from actual hitting), but combining this gesture with a play face to signal that he wants to play and not attack. The gesture is directed at a specific individual to signal his communicative intent, and the play face informs the recipient about the nature of this approach, since the very same gesture is also used in aggressive contexts. Again, one could argue that it is still unclear whether the signaler is meant to change the other’s behavior instead of mental representations, and that facial expressions are merely expressions of affective states instead of intentionally produced signals. Taken together, what I would like to emphasize is that we have to define clearly which forms ostensive signals can take in other species.

Origins of ostensive communication (Chapter 4)

This particular chapter, from my point of view, does not always represent our current knowledge about primate communication and the corresponding socio-cognitive skills in an appropriate way. While reading it, I got the impression that Thom Scott-Phillips postulated a potential fundamental difference between the communication of humans and other primates, which he then confirms by selecting the corresponding (but not always representative) studies, while neglecting others.

Furthermore, what I missed here was a more in-depth discussion of the function of so-called “attention-getting gestures”, which are the tactile and auditory gestures that chimpanzees seem to use to attract the attention of a not-attending recipient. Thus, it could have been of special interest to discuss whether nonhuman primates use attention-getters to manipulate the attentional states (and thus visual perspectives) of others (conspecifics or humans), thus signaling communicative intentions, and whether signalers combine attention-getters with subsequent specific intention movements, particularly since existing studies draw different conclusions. Furthermore, I think the conclusions drawn from section 4.5 on the mindreading skills of nonhuman primates are rather problematic. The conclusion that “there is little evidence that chimpanzees have command of the type and extent of mental metarepresentations that have been identified as cognitive pre-requisites for ostensive communication” is simply premature. Thom Scott-Phillips even acknowledges that it is difficult to develop a paradigm that is appropriate and ecologically valid for nonhuman primates and that such studies might be very difficult to conduct because of methodological reasons. I do not want to argue that the socio-cognitive skills of nonhuman primates are most likely as complex as those of humans and that potential differences are only a matter of quantity, not quality. However— with apologies for this almost trite remark—, “the absence of evidence is not the evidence of absence”. In this context it is also not helpful to refer to the possibility that negative findings have been most likely not published. Currently we have to acknowledge that we still know comparably little about the mindreading skills of nonhuman great apes. The point I want to emphasize here is that we simply don’t know (yet) the mindreading skills of nonhuman primates to the extent they have been studied in humans.


  • comment-avatar
    Dan Sperber 4 July 2015 (19:58)

    I found Katja Liebal comments highly relevant, in part, maybe, because I interpreted them as being more constructive than critical (whether or not they were so intended).

    It shouldn’t be controversial that humans are capable of indicating to their audience both that they intend to communicate something and what it is that they intend to communicate. Moreover they can do this in such a way that their indicating that they intend to communicate helps their audience recognize what it is that they intend to communicate. If I happen to stare sadly at my empty glass, my host may notice this and refill my glass, but this is not a case of ostensive communication. If I establish eye contact with my host and then stare sadly at my glass, this is a case of ostensive communication that my host should take as a request for a refill. The two aspects of an act of ostensive communication—informing the audience of one’s communicative intention, and informing her of one’s informative intention—may be both carried out by the communicative act as a whole (as in most of ordinary verbal communication) or there may be a distinct ostensive cue or signal that informs the audience of the communicator communicative intention. This is what happens, for instance, in the kind of adult-infant pedagogic interaction studied by Csibra and Gergely. The fact that attracting attention, for instance by establishing eye contact, and conveying a content may be merged or be two distinct aspects of the overall communicative act is a well-known fact and not at all a problem in the study of ostensive communication. This possibility is highly relevant when we look at communication in primates, but it is not something that diminishes the relevance of the ostensive communication framework to ask comparative questions, or it shouldn’t be.

    Do other animals and in particular other primates also engage in ostensive communication? There is no conclusive evidence that they do but there is at least evidence relevant to asking the question. The case of chimpanzee penis offer that Katja describes has long been my favourite, with the leaf-clipping looking very much like an ostensive signal and the interpretation of the display of the erect penis, to which the leaf-clipping sound serves to attract attention, as indeed an offer being, favoured by the leaf-clip. Of course, this is not enough to conclude that genuine ostensive communication is taking place: a behavioural explanation without mindreading, let alone, higher order mindreading, is as Katja points out, possible.

    Before being able to profitably classify penis offers and comparable cases as fully-fledged / rudimentary / incipient ostensive communication, or as nothing of the sort, it might be more profitable to concentrate on these “attention-getting gestures” Katja mentions. Ostension itself is an elaborate form of attention manipulation and, from an evolutionary point of view, this is where one should look for possible precursors, I believe.

    In any case, if the description of typical human communication as ostensive is correct, it is clear that humans are quite unique in using it as much and as richly as they do, whether there is something of the sort in our closest relatives or not. In fact, if there is ostension in other primates, or just if their attention-getting gestures are related to ostension, this would be hugely interesting both from a comparative and evolutionary point of view, and it might help us better understand forms of interaction in humans infants, children, and adults that are not quite ostensive, but close.

  • comment-avatar
    Thom Scott-Phillips 5 July 2015 (15:26)

    Katja needn’t have apologised for her poor internet connection. Her expertise on non-human primate communication ensures that, despite not being able to follow the discussion so far, her comments still make a unique and substantial contribution (as Dan indicates in his own response above).

    Katja begins by asking for some clarification about the nature of ostensive signals, so let me try to provide some. Ostensive signals are those that express communicative (and hence informative) intentions. We can, in principle, do this with any behaviour at all, including all those that Katja suggests: gestures, facial expressions, eye gaze, touch, body posture, movement, vocalisations, and indeed anything else. So, to pick up on the example Katja raises, it is possible, in principle, for ostension to be expressed only with a tilt of the coffee cup. However, it is sometimes more appropriate to combine two behaviours in the one signal, and so eye contact is add to the tilt of the coffee cup (this could be because, say, the visual environment is noisy, or because there is a politeness norm that forbids just gesturing to waitresses in an impersonal manner). The two behaviours are produced as two parts of the same signal (which expresses both communicative and informative intentions). It is also possible, as Dan suggests, for two behaviours to serve these two functions independently (i.e. one expresses the communicative intention, the other expresses the informative intention), although I’d suggest that this is relatively rare.

    How, then, to interpret the example Katja offers, of a chimpanzee combining a “penis offer” with an attention-grabbing behaviour like leaf-clipping? These two behaviours have functions that are, as Katya notes, superficially similar to the functions of informative and communicative intentions, respectively. The penis offer is the what, and the leaf-clipping is the that. However, this functional similarity does not imply a cognitively similarity. Informative intentions are intentions to change others’ representations of the world. An intention to simply change behaviour is not the same thing, and, as Katja and Dan both point out, it is quite possible that one or both of these behaviours is focused on behaviour rather than mental states. This is why one especially good way to test for an informative intention is to experimentally dissociate the behavioural outcome from the intended change in mental representations (SOM, p.87). I don’t know how one might do this in the case of the chimpanzee penis offer (?!), but until somebody does so, and shows that the behaviour is driven by an intention to change mental states, we cannot conclude that the penis offer is an expression of an informative intention. The story for leaf-clipping is similar. If this is an expression of a communicative intention, we need evidence that it is driven by intentions not simply to change behaviour, but rather to encourage others to mentally recognise that the chimpanzee has an informative intention. Again, we don’t have this sort of evidence for chimpanzees. If we were to acquire such evidence, I would happily change SOM’s conclusion that non-human primate communication is likely not ostensive. This would be true regardless of modality.

    Katja concludes her comments with a complaint that SOM’s conclusion about chimpanzee mindreading are premature. As she rightly points out, absence of evidence is not the evidence of absence. I fear, however, that there has been a slight overinterpretation here. Nowhere do I claim that it has been shown that chimpanzees do not have command of the type and extent of mental metarepresentations in question. I claim only that “there is little evidence” for such a conclusion. I get the impression that, at present, Katja would agree.