Intelligence of Apes and Other Rational Beings

  • Duane M. Rumbaugh &
  • David A. Washburn
Yale University Press: 2003. 326 pp. $35

This intriguing book recounts a lifetime's effort by Duane Rumbaugh to establish an accurate and objective picture of the cognitive capacities of apes and other primates, spanning the past half-century's sea changes in our understanding of animal minds. When Rumbaugh began his career, the field was still dominated by a narrowly focused behaviourism that tended to reduce all learning to simple conditioning. But in the past 20 years, he has collaborated with Sue Savage-Rumbaugh to generate claims of cognitive richness in apes that are about as far from classical behaviourism as it is possible to imagine.

We are treated in this book to an honest and intimate account of Rumbaugh's struggle to square his early training and continuing commitment to behaviourism (“the only data available to our science are behaviors”) and to good, objective science (through the use of automated keyboards to record data, for example) with the quirky creativity and rationality that he encountered in decades of work with chimpanzees and other primates.

Examples of this cognitive complexity include both experimental results and one-off observations of unique, innovative actions. The inherent non-replicability of the latter is of course scientifically problematic, but may nevertheless be the kind of raw material with which any scientific engagement with creative rationality must deal.

An illustration comes from an attempt to use food rewards to train the chimpanzee Lana to urinate in a pan. Rumbaugh points out that, contrary to classical conditioning theory, her response to this reinforcement did not become more vigorous and stronger; instead, she came to eke out her contributions in smaller and smaller doses, thus gaining more rewards. The key observation, however, is that when she had no urine left, she approached the pan and spat into it instead. So what Lana had learned cannot be explained as operant conditioning of her urination behaviour. Her innovation can instead be described as creative, rational and intelligent, and it is one of numerous such cases that pepper the book.

Learning curve: the chimpanzee Lana has taught scientists a great deal about intelligence in apes. Credit: F. KIERNAN/YERKES CENTER

Rumbaugh's half-century scientific quest to understand the kinds of minds that lie behind such episodes may offer an education for the younger generation of researchers who can all too easily fail to appreciate the scope of the intellectual revolution that has taken place. However, Rumbaugh and Washburn have greater ambitions: to identify a larger category of complex behaviour, which they call emergents, that goes beyond operants and respondents, integrating all three into a coherent theoretical framework, 'rational behaviourism'. No succinct definition of emergents is offered, but their scope can perhaps be glossed as behaviours for which no specific reinforcement history offers an adequate explanation. These include innovative or creative actions, such as Lana's pan-filling, along with others documented through the quantitative and statistical results of numerous experiments.

Prime examples of these are experiments on 'learning set'. In any series of such experiments, subjects are faced with the same learning contingency, such as that food is hidden under the odd one of three objects, with the objects themselves being changed for each experiment. 'Learning set' refers to the capacity to grasp the overarching rule governing the series (in this case, odd-one-out) such that, over time, successful choices are made faster with each new instantiation of the task. Rumbaugh and Washburn say that this achievement must count as an emergent. Considering this and a range of other kinds of evidence, they conclude that chimpanzees, and probably the other great apes, show the greatest manifestations of emergents, in some cases achieving a qualitative superiority over the smaller-brained monkeys tested.

The book, then, makes two key claims: that our understanding of behaviour is advanced by the concept of emergents, which is in turn nested within rational behaviourism; and that apes show both quantitative and qualitative superiority over other animals in these terms.

Should we buy these claims? On the first, I am sceptical, for several reasons. For a start, it doesn't seem clear that such phenomena as learning sets are inherently more reliant on causal reasoning or other aspects of rationality than simpler kinds of associative learning. Instead, where the latter can be described as learning about first-order patterns in the world (for example, learning that one of three specific objects is the one to pick), the emergent can be described as learning about second-order patterns (that the game is always to pick the odd object among three, say).

My second reason is that contemporary cognitive learning theory seems to have assimilated many of the complexities that the authors worry about, such as the case of Lana learning not a specific behaviour, but about achieving a state of the world (fluid in pan). More bridges have been built between learning theory and the products of the cognitive revolution than the authors acknowledge, so it is a concern that contemporary learning theorists such as Nicholas Mackintosh and Anthony Dickinson, as well as integrators such as Sara Shettleworth, are absent from the bibliography.

I am more convinced of the relative achievements of the apes, as this is consistent with much other evidence. However, Rumbaugh and Washburn fail to cite a recent book — Primate Cognition by Michael Tomasello and Josep Call (Oxford University Press, 1997) — that offers a much more comprehensive overview of the kinds of experiment, such as learning set, that they highlight, but which does not find the overall evidence of qualitative ape superiority persuasive. Such alternative conclusions beg a refutation, if the case for the apes is to be made convincingly.

But perhaps the main strength of Intelligence of Apes and Other Rational Beings is the development of further technologies to probe cognition, most recently the use of computerized joystick-driven tasks. A string of experiments exploiting these advances has already thrown up some fascinating discoveries (ironically, perhaps, mostly in macaque monkeys rather than apes), and I predict that this may represent one of the most influential legacies of the work collated in this volume.