Newest math whiz - Sarah the chimp

Two ethologists at the University of Pennsylvania have added a new element to the ongoing scientific controversy over the intelligence of animals. They believe they have demonstrated primitive mathematical ability in a chimpanzee.

Or as David Premack and Guy Woodruff state it, ''The results (of the experiment) reveal the presence of simple 'proportion' and 'number' concepts in a nonhuman primate.'' The test animal, in other words, seems to grasp the essence of such fractions as one-fourth, one-half, or three-fourths and of numbers such as 1, 2, 3, or 4.

Five chimps were tested - four juveniles and an adult, Sarah. The juveniles did poorly, showing no clear sense of number or proportion. However, Sarah, a veteran of language experiments, did spectacularly well. She made few mistakes, performing significantly better than would be expected were she merely matching objects at random.

In the tests, the chimps were asked to pick which of two alternative objects matched a given test object. The only common property would be proportion or number. A container half full of liquid would have to be matched with half of a round wooden disk, for example. Alternatively, three containers of liquid would be matched with a box containing three blocks of wood. In all these tests, Sarah seemed to have a clear sense of the difference between, say, one-half and three-fourths, or between the numbers three and four. She did not try to match a half-full container with a wooden disk that was three-quarters whole.

Researchers Premack and Woodruff speculate that Sarah used a complex reasoning process to distinguish between the concepts of ''the part'' and ''the whole'' and recognized these as meaningful catagories. This possibility, they say, ''is in keeping not only with previous results showing her inferential ability with other quantitative properties, but also more recent data from our laboratory showing explicitly that she is capable of analogical reasoning.''

This research follows in the tradition of more than a decade of work with chimpanzees and gorillas in which several scientists, including Premack, have taught some of these apes to use different forms of sign or symbol language. In some cases, this has been the common American Sign Language. In other experiments, including those using the chimpanzee Sarah with which Premack works , a simple language using symbols which may be generated with a computer is used. Such work has led to claims of demonstrating language capacity in apes.

The language experiments have their critics who contend that the experimenters overinterpret their results in claiming chimps show true language ability. Such critics assert that the apes are merely imitating their human trainers or showing a conditioned behavior. Harvard University psychologist B. F. Skinner and his colleagues Robert Epstein and Robert P. Lanza, for example, have duplicated some of the ape results with pigeons. Mr. Skinner has suggested that the supposed language capacity of apes may be no more than an example of a classical conditioned response.

It is against the background of such criticism that Premack and Woodruff have carried out their experiments with the mathematical concepts of number and proportion. As they explain in describing their work in the journal Nature, they tried hard to rule out any possibility that the chimps could take their cue from the human experimenters or merely go through a learned routine.

At this writing, it is too early to tell how this work will stand up as others try to repeat the experiment. So far, the most that can be said is that the Pennsylvania ethologists believe they have demonstrated a kind of primitive mathematical conceptualization in one chimpanzee.