The Mating Mind: How Sexual Choice Shaped the Evolution of Human Nature (56 page)

children are word-sponges. By contrast, even the brightest chimps in ape language experiments require at least 20 to 40 exposures before they learn the meaning of a visual sign. One has to do the equivalent of saying "bassoon, bassoon, bassoon" over and over until it loses all meaning to a human and acquires one for the chimp. In humans, word meaning appears to be stored in special brain areas, and damage to these areas through injury or stroke produces vocabulary deficits.

Is a vocabulary of 60,000 words excessive? Most of it is not used very often. The most frequent 100 words account for about 60 percent of all conversation; the most frequent 4,000 words account for about 98 percent of conversation. This sort of "power law" distribution is common: the 100 most successful movie actors probably account for 70 percent of all money paid to all actors; the 100 most popular Internet sites probably handle a similar proportion of Internet traffic; and so forth. It is not surprising that vocabulary use follows a power law, but it is surprising that our average vocabulary is so large, given how rarely we use most of the words that we know. It could easily have been that just 40 words account for 98 percent of speech (as it does for many two-year-olds), instead of 4,000 (as it does for most adults). As it is, any of the words we know is likely to be used on average about once in every million words we speak. When was the last time you actually spoke the word "cerulean"? Why do we bother to learn so many rare words that have practically the same meanings as common words, if language evolved to be practical?

To see whether our large vocabularies evolved as ornamental luxuries, we can compare them with artificial languages and "pidgin" languages specifically created for pragmatic communication. Artificial languages can work with very small vocabularies. In the 1920s, the Oxford philosopher I. A. Richards and collaborator C. K. Ogden developed a stripped-down English vocabulary of just 850 words that they called Basic English. Their motive was to promote international peace and understanding by making it easier for non-native speakers to acquire a minimal, functional version of English, which they recognized as the emerging planetary language.

Basic English works with ordinary English grammar. Despite it having a vocabulary only 1 per cent as large as normal, Richards wrote that "it is possible to say in Basic English anything needed for the general purposes of everyday existence—in business, trade, industry, science, medical work— and in all the arts of living, in all the exchanges of knowledge, desires, beliefs, opinions, and news which are the chief work of a language." Indeed, Richards wrote this passage using Basic English. Richards and Ogden also found that they could easily define any other English word using just the Basic vocabulary: their
General Basic English Dictionary
did this for 20,000 non-Basic words. Basic is really quite simple: it gets by with just 18 verbs, which Richards called his "willing, serviceable little workers . . . less impressive than the more literary verbs, but handier and safer." Basic is not quite as compact as ordinary English—it takes perhaps 20 percent more words to state a given idea—but it is vastly easier to learn, and easier to understand by a wider range of people. A slightly expanded Basic even works for expressing scientific ideas: the Basic Scientific Library series in the 1930s included introductory textbooks on astronomy and biology.

Like Basic English, "pidgin" languages illustrate how useful even small vocabularies can be. Pidgins arise when people speaking mutually unintelligible languages are thrown together in a situation, such as a slave plantation, that forces some means of communication to develop. Most pidgins have small vocabularies, like Basic English, and minimal grammar. Yet they suffice for trade, cooperative work, and ordinary survival functions. However, children brought up learning a small-vocabulary pidgin tend to transform it into a larger-vocabulary "creole," which is a full-sized language. Language researchers take "creolization" as evidence that small-vocabulary pidgins must have been insufficient for pragmatic communication in some respect. But that implies that all complexity must be due to pragmatic demands. A different view is possible: perhaps Creoles, like language itself, arose as better verbal ornaments and better indicators of verbal intelligence.

If Basic English and pidgins allow people to communicate, trade, cooperate, and live together using very small vocabularies, why do all mature, natural human languages have a hundred times as many words? An analogy to bird song may be useful here. Most bird song evolves under sexual selection through mate choice. Most birds produce a fairly small repertoire of courtship songs, but in a few bird species, such as marsh warblers and nightingales, the number of distinct songs seems to have undergone some sort of explosive evolution, resulting in repertoires of over a thousand distinct songs. In these species repertoire size itself became a criterion for mate choice, with males who sing more songs being perceived as more attractive. Above-average repertoires may work as reliable indicators of a bird's age, learning ability, intelligence, brain size, brain efficiency, or general fitness. Males with larger repertoires appear to sire healthier offspring, suggesting that repertoire size may be an indicator of heritable fitness.

Although particular bird songs do not have any meaning, their overall repertoire size does; it indicates heritable fitness. Human words do have meaning, but perhaps our overall vocabulary size has the same meaning as their song repertoires. A large vocabulary may be a good fitness indicator. Large vocabularies may have been favored in mate choice, and may have evolved through sexual selection.

Obviously, vocabulary size differs enormously between people, so it could be a useful cue in mate choice. The American Scholastic Achievement Test includes plenty of vocabulary questions because vocabulary knowledge varies enough to be a reasonable indicator of intelligence and general learning ability. Evidence shows that vocabulary size is at least 60 percent genetically heritable, and has about an 80 percent correlation with general intelligence. (The correlation with intelligence is not 100 percent, of course—people with Williams syndrome, for example, have lower than average general intelligence, but delight in unusual words such as "diplodocus," and develop fairly large vocabularies.)

Since words are learned, it may seem odd that overall vocabulary size should be heritable, but that is what behavior-genetic studies find. Identical twins reared apart (who have the same genes but different family environments) correlate about 75 percent for their vocabulary size. By contrast, the environmental effect of parenting accounts for only a small proportion of the variation in the vocabulary size of children, and just about 0 percent of the variation in adult vocabulary size. If you have a large vocabulary, that is because your parents gave you genes for learning lots of words quickly, not because they happened to teach you lots of words. Actually, most of vocabulary's heritability is carried by the link between vocabulary learning ability and general intelligence, which in turn is highly heritable.

This link between vocabulary and intelligence may extend all the way to biological fitness. Perhaps general intelligence itself, or what intelligence researchers call "the
g
factor," is a fitness indicator. One study has shown that intelligence correlates about 20 percent with body symmetry, which is a known fitness indicator. Thus, vocabulary size could indirectly advertise fitness. Our ancestors would have benefited by favoring sexual partners with large vocabularies. If vocabulary was a criterion for mate choice, they would also have benefited by evolving larger vocabularies, just as peacocks evolved larger tails.

Few will admit to—or even be aware of—a sexual preference for a large vocabulary. It would be unusual to see a personal advertisement that ran "Single female seeking man who knows fifty thousand useless synonyms." However, couples in long-term relationships tend to have vocabularies of similar sizes, and the strength of this assortative mating for vocabulary size is higher than for most other traits. Although one may not consciously prefer a date who uses "azure" instead of "blue," one may shudder if a date uses "azure" as if it meant teal, mauve, or vermilion.

So how would one display a large vocabulary size in courtship? Consider vocabulary as an intelligence-indicator. We know from intelligence-test research that there tend to be minimum IQ thresholds for producing and comprehending certain words.

According to the widely used WAIS-R intelligence test, for example, English-speaking adults with an IQ of 80 typically know the words "fabric," "enormous," and "conceal," but not the words "sentence," "consume," or "commerce." IQ 90 speakers typically know "sentence," "consume," and "commerce," but not "designate," "ponder," or "reluctant." If you are flirting with someone, and they say they would like to "consume" your body in a passionate embrace, but they do not understand when you say you are "reluctant," you can probably infer they have an IQ between 80 and 90. We make these sorts of inferences quite automatically and unconsciously, of course.

We may not realize that we use vocabulary as an intelligence-indicator. Yet, what we do not admit, wise nannies may understand. In the film
Mary Poppins
, the song "Supercalifragilisticexpialidocious" celebrated the power of unusual words to advertise intelligence, attract mates, and make friends with maharajas.

Near the end of the song, Mary suggested that when you cannot find the right word to express your thoughts, the "super" word can fill the gap. However, she also warned that its life-changing power must be used with caution. At that point, her back-up drummer interjected a personal example: he once uttered the word to his girlfriend, and it led straight from his verbal courtship to their marriage. Mary's song captures a key feature of the verbal courtship theory—words can work as reliable indicators of intelligence (and articulation ability), even when, like birdsong, they have no meaning whatsoever.

To test this verbal courtship theory of vocabulary properly, we would have to find out much more about human verbal behavior than language researchers know at present. We don't know the size of typical ancestral or tribal vocabularies. We don't know whether people use more impressively obscure words during courtship. We don't know whether large vocabularies are valued directly in human mate choice. We don't know how vocabulary sizes correlate with brain size, physical health, physical attractiveness, fertility, or general fitness. Sex differences in the distribution

of vocabulary sizes are rarely reported in the scientific literature (though they are perfectly well known to the Educational Testing Service that administers the SAT).

Words appear to have evolved for symbolic reference. This appears to set them apart from other forms of animal signaling. My point in this section has been that words can also evolve as indicators. The servicable little vocabularies of Basic English and pidgins suggest that we learn and display many more words than we really need to communicate: our huge vocabularies make no sense as pragmatic adaptations for survival. Human vocabulary size may have evolved through the same sexual selection process that favored enormous song repertoires in some bird species. But whereas only male birds sing, both men and women use large vocabularies during courtship, because courtship and choice are mutual, and because unusual words work as reliable displays only if their meanings are understood.

Why Do Women Have Higher Verbal Ability
than Men, if Language Was Sexually Selected?

When sex differences do show up in human mental abilities, women typically show higher average verbal ability, while men show higher average spatial and mathematical ability For example, women comprehend more words on average, and this sex difference accounts for almost 5 percent of the individual variation in vocabulary size. But sexual selection normally predicts that males evolve larger ornaments. If language evolved as a sexual ornament, it seems that males should have much higher average verbal abilities. Is this a fatal problem?

The standard predictions of sexual selection are hard to apply because language is used for both speaking and listening—both verbal display and the judgment of verbal displays by others. Normally, sexual selection makes males better display-producers and females better display-discriminators. Peacocks can grow bigger tails, but peahens may be better at seeing and judging tails. Most tests of human verbal abilities are tests of language comprehension, not tests of language production. Given a strict

male-display, female-choice mating system, we should expect female superiority in language comprehension and male superiority in language production.

For example, females should recognize more words, but males should use a larger proportion of their vocabulary in courtship, biasing their speech towards rarer, more exotic words. In this simple picture, more women might understand what "azure" really means (so they can accurately judge male word use), but more men might actually speak the word "azure" in conversation (even if they think it means "vermilion"). Standard vocabulary tests measure only comprehension of word meaning, not the ability to produce impressive synonyms during courtship. Reading comprehension questions are more common than creative writing tests. Women are faster readers and buy more books, but most books are written by men.

But the male-display, female-choice system is not an accurate model of human conversation anyway. Throughout this book I have stressed the importance of mutual mate choice in human evolution. Human courtship means, above all, men and women talking to one another. It is not restricted to men standing up and pouring forth a stream-of-consciousness verbal display to anyone who will listen. Such male verbal broadcasts can be observed in churches, parliaments, and scientific conferences, but human speech is typically more private and more interactive. The interactiveness of conversation makes terms like signaler and receiver problematic. All humans are both. As with other mental abilities, mutual display and mutual choice tend to produce sexual equality in the display ability