May 02, 2007

Ineffable apes

Look at me, baby! News from the animal communication front: Chimps and bonobos have arbitrariness of the sign, at least sort of, at least with respect to some gestures.

From the New Scientist article:

By observing captive groups of bonobos and separate groups of captive chimps, Pollick and de Waal identified 31 gestures and 18 facial or vocal signals made by the apes, and recorded the context in which they were used. It turns out that the facial and vocal signals were practically the same in both species, but the same gesture was used in different contexts both between and within species.

For example, the vocal signal "bared-teeth scream" signals fear in chimps and bonobos, but the signal "reach out up" -- where an animal stretches out an arm, palm upwards -- has different meanings. It may translate as begging for food or as begging for support from a friend, says de Waal. "The open hand gesture is also used after fights between two individuals to beg for approach and contact during a reconciliation. So the gesture is versatile, but the meaning depends on context."

In other words, chimps and bonobos seem to have gestural homophones--one symbol with two or more meanings.1 The authors, Amy S. Pollick and Frans B.M. de Waal, of the Yerkes National Primate Research Center at Emory University, find this suggestive in thinking about the evolution of language: Perhaps the earliest symbolic communications were gestures, and the symbolic use of vocal signals came later? In other words, perhaps this supports the gestural hypothesis of human language origins?

A colleague of theirs at the Yerkes Primate Research center agrees:

The openness of the hand-gesture system among chimps and bonobos "is consistent with the idea that the early hominid communications system was gesture based and that vocal communication came later," said William Hopkins, a Yerkes researcher not involved in the study. "The speech system is a very recent adaptation in hominids."

But are the gestures really symbolic? Maybe they're just used as an enhancement of the vocal/facial signal (like the gestures accompanying speech among users of spoken languages). Are they really independent communicative units on their own (like the symbols of signed languages)?

In fact, it seems to me that the 'flexibility' of (at least some) gestures is consistent with the notion that they are general attention-getting devices. On this interpretation, they show up in multiple contexts precisely because they are not symbolic. A gesture might be saying, 'Pay attention to my vocal /facial signal!' rather than, say, denoting 'Help!' in one context and 'Can I have some of that?' in another. Flexibility of context is the opposite of symbolic communication, in a way.

Figure 3 Consider Fig. 3 from the article, where the reliability of the correlation between particular contexts and particular gestures and vocal/facial signals is reported. The gestures that show the least reliability are hand/arm motions in various directions -- 'reach out side', for example. These gestures happen to be the ones which intuitively (speaking as a communicative primate myself) have the least symbolic content.

If these gestures are actually signal amplifiers, rather than symbols, then we wouldn't expect them to show up by themselves, in the absence of the 'real' signals coming through in another modality (another communicative medium, like sound). They would always appear in combination. This seems like the $64,000 question to me, if use of symbols is what is at issue. (Marc Hauser, of {Hauser, Chomsky, Fitch} is quoted in the NY Times article as wondering the same thing).

Indeed, in the article itself, the authors introduce the research with remarks on the pervasiveness of mulitmodality in the communications systems of many different species, and its amplifying effect:

Gestures [in chimps and bonobos] are rarely produced in the absence of other communicative signals, such as facial expressions and vocalizations. Multimodal communication has been appreciated in humans for several decades (43) and is becoming increasingly important in the study of animal communication (44). Multimodal signaling occurs across taxa, from snapping shrimps to spiders and birds, and in all contexts, although those related to courtship and mating are best documented (A.S.P., unpublished work). This communication strategy can have a variety of functions, including amplification and modulation of signal meaning. Combined with the graded facial/vocal signals typical of the apes (46), gestural flexibility has the advantage over the more stereotyped signaling by monkeys that it permits greater communicative complexity.

So, you get multimodal signaling across all taxa; makes its existence in chimps and humans seem less special. But you don't get it in monkeys, which makes it seem more special, within our lineage, anyway. Looking for clues about whether the gestural signals were always in a multimodal context, it turns out there is a difference between bonobos and chimps in this regard:

No significant difference was found in the proportion of signals that was gestural versus facial/vocal, but chimpanzees did combine these two signal classes relatively more often than did bonobos.

So there are at least some non-mulitmodal gestures in the data, enough to run stats on.

Interestingly, although chimps used more multimodal signals, multimodal signals were significantly more effective than gesture alone in eliciting a response in bonobos, but, remarkably, not in chimps. Among chimps, the other chimps responded about 67-68 percent of the time to any gestural signal, whether it was embedded in a multimodal context or not. Among bonobos, on the other hand, a gesture alone elicited a response at about that same rate, 67 percent-ish, but if a bonobo gestured and made some facial or vocal signal along with the gesture, they got a response a whopping 83 percent of the time! And yet the bonobos were less likely to produce mulitmodal signals than the chimps! What's up with that?

The interest of this inverse correlation does not escape the authors, who write in their discussion,

That this contrast between multimodal and single modality utterances held for bonobos only is interesting given that multimodal combinations are less common in bonobos. Could the relative scarcity of multimodal signaling in bonobos relate to a more deliberate combination of gestures with other forms of communication, perhaps in an attempt to add critical information to the message instead of merely amplifying it?

The whole question of amplification is complicated by the methodology. The authors wanted to have good criteria for which signals (gestural or otherwise) to count as communicative, so they only counted gestures made at the start of a 'social interaction'... which one individual approached another and attempted to engage the recipient with a communication signal. The two individuals may have been in proximity before, but without observable interaction. Signals were not included in the analysis, therefore, if they occurred in the middle or toward the end of an ongoing interaction.

So all of the communicative contexts examined were attention-getting contexts; this may muddy the waters for sorting out the precursor-to-symbolic-communication hypothesis from the signal-amplification hypothesis, since amplification is presumably useful in attention-getting.

Even if the gestures are more attention-getters than symbols, though, it's definitely interesting that apes, but not monkeys, use them. Silly monkeys -- it couldn't be more obvious that you should wave your hands to get attention! Working to get someone's attention, though, presupposes that you know they have an attention to get -- presupposes a theory of mind. And that is likely a sine qua non for language.

Maybe the reason that bonobos use multimodal signals more selectively than chimps do is not that they are trying to 'add critical information' to the signal. Perhaps it's because they have a better theory of mind than chimps do, and hence they've got a better grasp of how to deploy the whole 'Look at me!' thing.

Update: Here are Mr Verb's thoughts on the study. Also, Mark Liberman adds:

The chimp gesture meaning-variation may not be that different from the well-known vervet monkey variation in alarm-call meaning. They have three alarm calls, colloquially known to human researchers as the "eagle alarm call", the "leopard alarm call", and the "snake alarm call". But really the "eagle alarm call" is a "predator in the air" alarm, and baby vervets need to learn which large birds in each area are actually dangerous; e.g. not storks, not egrets, but yes to kites or various species of eagles, depending on the birds that live in the vicinity. Ditto for the "leopard alarm", which in the West Indies means "feral dog".

So like the open-hand gesture, the alarm calls have a basic core that is constant, along with referential details that can vary quite a lot, not only with context, but also as a result of learning from the local history of usage.

You can make the local differences seem to go away by re-naming the calls: "avian predator alarm", "mammalian predator alarm", etc. But that wouldn't recognize the fact that the different local populations of vervets have actually learned different associations for the calls, in terms of when to make them and in terms of when NOT to make them.

Like the basic hand gesture, the calls themselves are invariant across space, and presumably are genetically programmed. It's the details of their interpretation in context that are locally "arbitrary"

Um, comments?

1 Anyone want to coin a term for this? Maybe there already is one. Homogests? Ick.

Posted by Heidi Harley at May 2, 2007 03:02 AM