"This may sound like a frivolous question," conceded a cat-owning human in the audience at this weekend's "Animal Intelligence" symposium at the National Zoo, "but if a cat catches a rat and brings it to you, would you interpret that as a gesture of generosity?"

University of Maryland psychologist William Hodos was reluctant to go quite so far. He would rather not call it "generosity," he said, but perhaps the phenomenon might be said to reflect a "momentary feeling of good will."

His listener seemed satisfied with this guarded reply. If so, it was one of the few occasions when the symposium achieved harmony--or even momentary good will--on the basic question of the weekend:

To anthropomorphize or not to anthropomorphize?

Should scientists, in other words, put themselves into the minds of animals, interpreting their behavior in terms of humanoid qualities and processes? Or should they look at animal activities with the utmost detachment and caution, describing only what can be seen or proved beyond a reasonable doubt?

Many speakers stressed the dangers of the first course, recalling the sort of sensational claims that circulated a few years ago about the language systems and startling brainpower of dolphins and apes.

The National Zoo's own dolphin expert, Sheri Lynn Gish, not only dismissed the grander notions about dolphin-talk, but said there was no reason to think that vocal signals (rather than visual or tactile ones) are the primary mode of dolphin communication.

Psychologist Bennett Galef Jr., from McMaster University in Ontario, apologized for the fact that his Norway rats were not as "intrinsically appealing" as chimps and dolphins, so he had no "cute slides" to show. But by putting his rats through controlled experiments in the laboratory, he said he had demolished one sophisticated interpretation of how animals learn and think. And the same fate, he implied, might be in store for other such theories.

If you give a Norway rat something that makes it sick, said Galef, it will quickly learn not to eat the food in which it has found the offensive substance. What's more, young rats get the word too--even if they have never eaten the poisoned food. As a result, scientists long have assumed that adult rats "teach" their young what to eat and what not to eat, or, failing that, that the young rats "learn" by observing their parents.

But Galef said his study had yielded a set of simpler causes. Most notably, he said, it turns out that the taste of the mother-rat's milk reflects the taste of the foods she has been eating, thus conditioning her young to develop her own dietary preferences.

Here as elsewhere, Galef contended that what looks like imitative behavior is really, in the words of pioneering American psychologist Edward Thorndike, "the indirect result of instinctive acts." Imitation and observation, he said, "are simply not very important learning processes--at least in non-primate species."

Galef followed--and sharply challenged--animal behaviorist Benjamin B. Beck, who had reported on his study of the seagulls of Cape Cod and their practice of dropping shells on hard surfaces to extract the fleshy occupants thereof.

Sometimes, Beck said, one favored patch of shore rock "can look like Friday night at National Airport. You can have a number of gulls circling in a holding pattern, waiting to make their drops." Things can get so hectic, according to Beck, that the gulls decide to move on to less crowded sites--perhaps a large parking lot--where no waiting is required.

Because of the distances between the capture sites at sea and the drop sites on land, along with the substantial time lapses involved, Beck speculated that the gulls retain a conceptualized map of the region in their minds, and a conceptualized representation of the end-product--the food inside the shell.

But Galef was dubious. "The implication that the animals have a mental map is really, it seems to me, in the mind of the experimenter, not in the mind of the animal," he said.

Beck agreed there was no proof that gulls conceptualize on this level, but this would be a fruitful area for further research, he said, because "internal representation promises to be a synthesizing idea in the study of tool use and animal intelligence."

"Most cases of what we call observational learning are reducible to much more elementary processes," said Beck, invoking the example of the assassin bug, a "brainless insect" that dons a portion of termite nest as camouflage and uses an injured termite as bait in order to attract other termites and devour them. This seemingly elaborate behavior is completely inborn, or "prewired," said Beck, and it is only one of many such cases.

The "bird and bee man" of Princeton University, James Gould, described one astonishing case of pre-wired sophistication: the dance-language of the honeybee. Inside the nest, the bee performs an elaborate figure-eight dance that, using the sun as a reference point, precisely conveys the location where its nest-mates will find food. Even more incredibly, said Gould, "a bee can deliver a dance without ever having observed one."

"That's a really impressive bit of behavior for a little insect we like to swat and get rid of," he said. In fact, the dance-language of the honeybee is "the second most complex language there is," according to Gould--second only to the verbal language of the human.

True, said Gould, the bees' dance-language is innate and inflexible. But flexibility is not always an asset, he said, noting when bees deviate from the inborn norm, they tend to pay a heavy price for it. On the critical first day when young bees leave their nest and begin foraging, for example, fully half of them mysteriously fail to navigate to their feeding spots and never are heard from again.

Just because humans think, said Gould, they shouldn't necessarily regard it as a desirable facility for all animals. "Thinking is a time-consuming and error-prone process which should only be used as a strategy of last resort," he said.

Higher intelligence, said Beck, evolves only in specific situations--for example, when animals begin to exploit food sources that come inside hard, imbedded coverings, or when (like early humans) they begin to range across large and physically varied environments that favor species capable of making long-range plans and adopting different feeding strategies in different seasons.

Of all the symposium's participants, only cognitive psychologist Carolyn Ristau explicitly defended the use of anthropomorphic phrases to explain animal behavior. After studying birds that lure predators away from the helpless young in their nests by feigning injury, Ristau said she thought the most useful way to explain this behavior was to say that the parent "wants" to lead the predator away. In any case, she said, "birds are not reflexive machines."

No indeed, said Hodos, who pointed out that "some of the largest bird brains would be regarded as quite typical for primates of equivalent body weight."

"Keep that in mind the next time you call someone a birdbrain," he said. "You're actually paying them a compliment."