Science news editors love stories about animals to which people have a strong emotional reaction. Whether the reaction is awe, empathy, envy, fear, disgust ... readers' eyes rest on these stories - and remain.

That makes sense, of course. We don't want to learn a mere hodgepodge of facts about nature, we want to learn about what we care about. One endless source of fascination is animal minds. What does it feel like to be a dog, a cat, a chimpanzee?

Consider, for example, the story of Oscar the "deathcat", who appears to have developed a sure sense (perhaps by scent) of when one of the infirm elderly residents on the floor of the nursing home where he lives will shortly die (thus ensuring that no one died absolutely alone). Had Oscar been a doctor with good diagnostic skills, he would be commended, to be sure. But Oscar fascinates serious scientists, as a recent essay about him in the New England Journal of Medicine attests, precisely because he is a cat, not a doctor.

Also, few of us want to believe that our own cat (or dog) is dumb. On the contrary, we readily accept stories of unusual powers in animals, whether they are anecdotes or reports from science labs.

But when does hope become hype?

Some have fallen for hype about animal abilities. For example, in The Design of Life, William A. Dembski and Jonathan Wells note that the famous American astronomer Carl Sagan (1934-1996) believed that chimpanzees, with proper training, could develop a human-like culture:

“What sort of culture, what kind of oral tradition would chimpanzees establish after a few hundred or a few thousand years of communal use of a complex gestural language?” He speculated that the chimpanzees would see the pioneer primatologists as gods. (- Carl Sagan, The Dragons of Eden: Speculations on the Nature of Human Intelligence (New York: Random House, 1977), 123–24. See GN pp. 4-7 on the CD.)

In fact, chimpanzees probably do not have the mental equipment to have language skills anywhere near what Sagan had hoped for. But why did the initial reports that convinced him sound so favorable? One problem was overinterpretation.

The ability to learn a large number of signs is not the same thing as the ability to learn a language whose meaning depends largely on grammar. The former achievement is sometimes found among birds as well as mammals, but the latter seems unique to humans.

The ability to string a sequence of words together does not necessarily mean awareness of grammar. The sentence "Tom shot John" does not mean the same thing as "John shot Tom," and the difference is pretty important. Overly generous assumptions were made about the extent to which apes such as Washoe and Kanzi were using grammar. When they were examined by scientists other than their trainers, they did not perform well.

Follow the evidence wherever it ...  fits our template?

In addition to our natural desire to communicate with and understand animals and to protect endangered species, another factor - unique to present-day society - can lead to hype about animal achievements. That factor is an underlying assumption - not evidence but an assumption - that human abilities are simply a souped-up version of animal abilities. Thus, Einstein's E=MC2 is understood as simply an eventual outcome of the ability of an animal to determine which pile of food is bigger.

On this view, no intelligence underlies the universe; rather, intelligence is simply an adaptation for survival and spreading one's genes. So it is assumed that humans moved gradually, and perhaps accidentally, from squabbling over bones to E=MC2.

Sagan, for example, considered himself to be a serious skeptic, but he passionately believed in this view of the universe. Thus he could easily convince himself that apes can learn and use complex languages.

Hype aside, the evidence points away from the assumption that abstract mathematics is simply the outcome of squabbles over bones. There is a gap that is simply not bridged by the studies of animal number sense, nor do available studies shed much light on the gap.

For example, consider a recent study highlighted in Nature News, in which two rhesus macaques were tested against 14 college students in simple addition:

In the battle of man versus macaque, students bested the monkeys for overall accuracy at 94% to 76%.

That doesn't sound like very much but the article goes on to point out that the sums were flashed on a screen for only half a second - to prevent the humans from using cognitive skills to solve the problems. Still, regarding the 94% to 76% figure,

Such similarities, researchers say, suggest an evolutionary continuity between basic mathematical skills in humans and other primates.

But do they? We then learn,

... the two monkeys performed thousands of trial runs before being tested. A critical next step will be to test the animals in the absence of training, Hauser adds. "This will be important," he says, "especially since the human animal comparison often breaks down because the former are tested without training and the latter with training."

Essentially, humans with no training performed significantly better than monkeys with a great deal of training even when the experiment was set up in such a way as to prevent them from using normal human methods of solving  number problems. Far from narrowing the gap between humans and other primates, the study starkly reveals it.

An interesting recent article in Monitor on Psychology (Volume 38, No. 3 March 2007) offers some insight into the debate over animal number sense, including some helpful clarifications, among them one by the University of Guelph's Hank Davis:

Animals may naturally make a distinction between more versus less, he says, but they aren’t using specific quantities to do it. For instance, a hawk might see that one field has more mice than another, but it doesn’t know that one field has 367 mice while the other only has 215.

“This is relative numerosity, not absolute number,” he says. “People would like to believe that number is a natural sense for nonhumans, but I don’t believe it. I still think it is a last resort.”

Also:

Update January 11, 2008: Here's a chapter of a book on "birds' judgment of number and quantity", which explains,

Birds can also discriminate relative differences in proportion when the stimulus quantities are either continuous areas of color or are mixtures of numerous colored items. Various models have proposed how non-verbal organisms might process such an abstract property as number. This chapter also considers ways in which the ability to assess the number of things birds encounter in their general environment may be advantageous to them in an evolutionary sense.

Mythbusters: Learning to read the popular science press critically

When reading stories about animal number sense - and animal intelligence generally - in the popular science press, note whether the story shows any of these characteristics:

Cherrypicking - that is, carefully selecting only the evidence that demonstrates one's point. This is common in reports on animal number sense or intelligence, even though the practice is discouraged in other areas of science. That is, the few animals who learn the routine satisfactorily are presented for public consideration as if they were a norm. The ones that baulked, despite all incentives, are quietly ignored. This is about as useful for estimating actual animal abilities as assuming that the average mutt shows the traits that the Mounties look for when selecting dogs to train for police work.

As law professor (emeritus) Phillip Johnson has noted,

The reductionist approach will always find some rigged test on which an animal species performs comparably to humans, since finding such comparability is the whole purpose of the enterprise.

Unnatural conditions - When animals are held captive for years by humans in a laboratory and drilled continuously in a skill, with food rewards for performance, many adapt to their situation by learning how best to get the reward. But the natural behavior of the same animal in its normal environment might be quite different. And, where evolution is concerned, it is the natural behavior in the normal environment that matters. Counting beyond a few digits, for example, is unlikely to benefit most wild animals. It is not necessarily a reflection against their intelligence if they do not readily acquire the skill.

Questionable assumptions that no one questions - Chimpanzees are assumed to be the most intelligent non-human animal because they are closest to humans genetically. Actually, many species of birds can give chimpanzees (though not humans) quite a challenge, where intelligence is concerned. In other words, the origin of intelligence may be quite other than conventional theories have assumed. The implication - that human intelligence may not have arisen in a long slow process of gradually cleverer apes - is not examined.

 Double standards in evidence - This is a very common problem in evolutionary psychology (the attempt to understand human psychology in terms of evolution from an animal ancestor). Here is an example: Dogs are much better than chimpanzees at reading human facial expressions - despite the similarity between the human and ape face, as compared with the human and dog face. Of course, that's not really a puzzle. We can account for the dog's superior performance by observing that the dog is highly motivated to understand humans, and the chimpanzee is not.

But notice the following: Had chimpanzees proved more adept than dogs, researchers would say "That's because chimpanzees are more closely related to humans!" Well in that case, given that chimpanzees are actually less adept than dogs, does anyone now propose to redraw the map of life to place dogs closer to humans?

Of course not. We know that chimpanzees are similar to humans because of the similarity of their genetics and body shape. That and that alone is the reason they are considered closely related to us. Their psychology - no matter what it is - changes nothing.

The authors of The Design of Life explain this double standard as follows:

Cognitive similarities support evolution when they are found but do nothing to undermine it when they are absent. This double standard makes clear that evolutionary psychology is not a scientific theory but rather an ideological project to undermine human uniqueness. (GN p. 5, CD)

The popular science press is an excellent way to keep up with interesting new developments, as long as critical thinking skills are used to identify hidden assumptions.