Chapter 1: Who is the Smartest of Them All?
The question almost everyone asks when the matter of animal minds comes up is: How smart are they? It is a question that occupied early researchers, too, at the dawn of experimental comparative psychology a century ago. Because intelligence seemed to be a quantifiable, testable parameter, studying an animal's performance on problem-solving tasks seemed the most promising avenue to exploring their mental processes.
Ranking species according to their relative intelligence fit in well, too, with many popular -- though often terribly distorted -- conceptions about Darwin's theory that were current in the late nineteenth century, and which have not yet altogether vanished. One of the most enduring misperceptions about evolution is that life represents a sort of chain of progress from inferior to superior forms. A particularly muddled version of this embodies the additional notion that progress up the chain of evolutionary advancement corresponds to the steps that the most advanced forms of life (i.e., humans) follow in the course of development from infancy to adulthood. This theory that "ontogeny recapitulates phylogeny" has no credence whatsoever in modern biology. Yet it still pops up regularly, and certainly most people tend to think about evolution in these terms, as a stepladder along which species can be ranked as higher or lower.
George Romanes, the great nineteenth-century collector of animal anecdotes, constructed an elaborate chart of the relative rankings of animals' stages of "mental development" that is a paradigm of this sort of thinking. To make the picture of a stepladder of mental ranks even neater, he equated each step with a corresponding age of mental development in humans. Thus a portion of his ranking chart looked like this, with the rank number in the first column and the corresponding stage of human development in the last:
28. Indefinite morality. -- Anthropoid Apes and Dog. -- 15 months.
27. Use of tools. -- Monkeys, Cat, and Elephant. -- 12 months.
26. Understanding of mechanisms. -- Carnivores, Rodents, and Ruminants. -- 10 months.
25. Recognition of pictures, Understanding of words, Dreaming. -- Birds. -- 8 months.
24. Communication of ideas. -- Hymenoptera. [Bees, Ants] -- 5 months.
23. Recognition of persons. -- Reptiles and Cephalopods. -- 15 months.
22. Reason. -- Higher Crustaceans. -- 14 months.
21. Association by similarity. -- Fish. -- 14 weeks.
18. Primary instincts. -- Larvae of Insects. -- 3 weeks.
17. Memory. -- Echinodermata [Starfish, etc.] -- 1 weeks.
7. Non-nervous adjustments. -- Unicellular organisms. -- Embryo.
3. Protoplasmic movements. -- Protoplasmic organisms. -- Ovum and Spermatozoa.
ANIMAL IQ
There are two basic flaws with this approach. One is the fundamental notion that some species are more highly "evolved" than others. We might naturally think of monkeys occupying a position higher on the phylogenetic scale than cats, and cats higher than rats. Yet the fact is that primates, carnivores, and rodents all diverged from a common ancestor at the same time. They are all equally "evolved." The branching tree of evolution has not just one culmination, but millions of culminations -- represented in every living species on earth today. Each is a brilliant success at what it does. The idea that fish, now stuck at level 21, are trying with all their might to ascend to level 22 is, from an evolutionary point of view, nonsense. Fish are adapted by virtue of millions of years of evolution to their particular, special evolutionary niche. They have had just as long to evolve as we have. It is not as if they are mere instances of incomplete evolution, the culmination of which is man (or Nordic man, perhaps).
The other problem is the implicit assumption that intelligence is something that can be measured on a linear scale -- and a scale where humans equal 100. Of course, by explicitly defining his stages of mental evolution in the animal kingdom in terms of the stages of mental development in human infants, Romanes was bound to end up with a purely anthropocentric definition for his rankings of intelligence. The "Properties of Intellectual Development" he those to list (morality, use of tools, recognition of persons, etc.) all have a distinctly self-centered air about them. How would one fit the navigational abilities of pigeons or the web weaving of spiders or the nest building of bowerbirds or the food caching of nutcrackers into Romanes's scheme? One wouldn't.
Some modern attempts to define intelligence in universal terms do not fare much better. Most common definitions of intelligence emphasize flexibility, creativity, and recognition of underlying patterns and overarching concepts. Some researchers, such as Steven Pinker of M.I.T., define intelligence in a more restrictive way that would seem to rule animals out of the running altogether: Pinker says intelligence is an ability to figure out how things work in order to overcome obstacles.
But animals, which as we shall see do not show any notable ability to figure out how things truly work, nonetheless show great facility at accomplishing things by acting on information they receive from the environment. They make decisions that are flexible and often appropriate. Unconsciously operating algorithms in the animal mind (ours included) produce what we would not hesitate to call intelligence were we to see a robot do it. Coordinating the movement of four legs over uneven ground while avoiding obstacles is a sophisticated computational task. We do not normally think of such automatic tasks as part of "intelligence" but why not? From a purely computational viewpoint, the sort of unconscious thought that permits an animal to recognize a predator and take evasive action could certainly involve far more brain power than distinguishing a group of three items from a group of four items. Is the former just dumb reflex and only the latter intelligence?
Another huge problem in attempting an honest, "zero-based" assessment of animal intelligence is the bias and assumptions built into the many tests that have been devised to measure it. Intelligence tests for humans have long been criticized for being culturally biased. Many of the early IQ tests used in the United States in particular suffered from this fault; they claimed to measure "native intelligence," but included questions that really measured nothing so much as familiarity with the middle-class American culture of the middle-class American psychologists who devised the tests. For example, tests given to Army recruits in World War I featured such questions as, "Washington is to Adams as first is to..." Other questions required the examinees to draw in missing parts on a series of pictures -- a stamp on an addressed envelope, a net on a tennis court, a filament in an electric light bulb, a horn on a wind-up phonograph, a trigger on a pistol, strings on a violin. Not surprisingly, many recent immigrants to America did not score very highly on the exam and were rated as "morons" or "feebleminded." Probably not very many of them played tennis, either.
But differences between human cultures pale in comparison to differences between animal species. Animals differ in temperament, perceptual abilities, motivation, social behavior, all of which affect their performance on tests we might devise for them.
WHO'S SMARTER: THE SHEEPDOG OR THE SHEEP?
Sheep have a reputation for being dumb. Border collies have a reputation for being smart. But both of these impressions may say more about our underlying prejudice than their underlying intelligence.
Much of what impresses us about dogs, after all, is their obedience to us. To put it in a slightly cynical fashion, we say a dog or a horse is smart when it does what we want it to. But many disobedient
