Abstract Pinker presents a broad rationale about the scientific examination of art as a natural phenomenon. He explores to how the studies of language within the humanities can interact with the scientific study of language from psycholinguistics and neurolinguistics, and how the analysis of fiction can interact with cognitive and evolutionary psychology.
Full Text (5003 words)
Copyright The Committee for the Scientific Investigation of Claims of the Paranormal (SCICOP) Nov/Dec 2006
What does psycholinguistics have to do with poetics, or fiction with mental computation? Nothing, according to the academic humanities as currently practiced. So much the worse for them.
----------------------------------- My starting point is the widespread perception that the arts and humanities are in trouble. I have a collection of despondent articles with titles such as "The Decline and Fall of Literature," "Have the Humanities Collapsed?" "The Humanities at Twilight," "The Humanities Plight," "What Happened to the Humanities?" and so on. Indeed, some of the signs of the health of the humanities are not good. There has been a decline in enrollment, faculty, and resources, and, most ominously, in interest among high school students. In recent polls only 9 percent of high school students express a desire to major in the humanities. There is also a widely acknowledged sense of malaise, a resignation about a lack of progress in the humanities. One solution to this problem was tried out by Indiana University a number of years back when it hired an advertising agency to recruit more students into its humanities programs. They had a campaign called "Think for Living" with slogans such as: "Do What You Want When You Graduate or Wait Twenty Years for Your Mid-Life Crisis" "Insurance for When the Robots Take Over All The Boring Jobs" "Okay Then, Follow Your Dreams in Your Next Life" "Yeah, Like Your Parents Are So Happy" I think this shows a certain desperation, which is a regrettable state of affairs, because the humanities are indispensable to being an educated person in a democracy. First of all, our lives are shaped by ideas. Our system of law, government, our economy, our assumptions about education, childrearing, and the relation between the sexes all have a rationale that was first worked out by thinkers in what we call the humanities. Second, the arts are touchstones for our private and public discourse. Even within the sciences, for example, one can't talk about biotechnology or genetic engineering without alluding to the novel Brave New World. Third, our lives are affected by contingencies of our culture, and part of being a capable citizen of a democracy is having a cosmopolitan appreciation of other times, places, and peoples. So, given, on the one hand, the fact that the humanities are indispensable to an informed citizenry, and on the other, the fact that they seem to be in trouble, what's going wrong? One diagnosis is that the malaise of the humanities comes in part from the separation from the sciences-the famous "two cultures" of C. P. Snow-which has led to an insularity of the academic humanities from new ideas and discoveries coming from our most exciting sciences. The sciences, for the last several centuries, have been characterized by a phenomenon that was given the lovely word consilience by E. O. Wilson, although I think the ideas were best expressed earlier by the founders of evolutionary psychology, John Tooby and Leda Cosmides. The history of modern science has been a history of unification of supposedly incommensurable metaphysical realms. Perhaps Newton's greatest accomplishment was to subvert the ancient doctrine that there was a fundamental division in the universe between the supralunary sphere of the moon-supposedly governed by pristine eternal laws-and the grubby, chaotic Earth below. Newton, of course, showed that the same force that brings the apple down to the earth also keeps the moon in orbit around it. It used to be believed that there is a fundamental division between the formative past, when the planet was shaped and created, and the static present, until Charles Lyle showed that forces that we see around us such as erosion, climate, volcanoes, and earthquakes, if operating over a sufficient length of time, would have been sufficient to sculpt the landscape as we find it today. It used to be believed that living and nonliving matter occupied separate realms; that living things were constituted out of some quivering gel called protoplasm, which is utterly unlike nonliving matter. Then Friedrich Wohler showed that one could synthesize urea out of chemical compounds, and, therefore, that the stuff of life was ordinary chemicals obeying the laws of chemistry. The integration of the nonliving and living worlds was further advanced by Darwin, who showed that the ubiquitous presence of adaptation in the living world could be explained by the natural selection of replicators. Later, Mendel and Watson and Crick would show that replication itself can be understood as a physical process. The last remaining chasm in our ontology is between the biological and the cultural, with the sciences on the one hand and the humanities and social sciences on the other. The ground for optimism about the humanities is that the process of consilience will continue. We're beginning to see the glimmerings of the unifications of the natural sciences with the social sciences and humanities. There are two realms in which this is being carried out. One is the study of deep history-the use of genetics, linguistics, and archaeology to bridge the end of human biological evolution and the beginning of history, civilization, and culture. Those familiar with Jared Diamond's Guns, Germs, and Steel, for example, have some exposure to this new science, which bridges biological evolution with the beginning of recorded history. But the approach I'd like to address lies in the sciences of human nature, the bridging of the biological and the cultural not in terms of time but in terms of causation. Cognitive neuroscience, evolutionary psychology, and behavioral genetics can play a role, illuminate our culture and society. The key idea is that our culture and society are products of the human mind. They weren't laid out by a committee of Martians, but are products of the activity of the brain, which itself is a product of evolution. It is the human faculties of perception, imagination, social cognition, and emotion that artists exploit in order to achieve their effects. What we call culture is not a separate realm from the biological or the psychological. These realms emerge in an epidemiological process in which ideas, inventions, and social contracts are shared to the point that they become epidemic in a community, at which point we call them cultures. I think that the humanities and sciences are already drawing together. For example, the philosophy of mind nowadays blends into the sciences of cognition and neurobiology. Topics such as consciousness, innate ideas, the mind/body problem, imagery, private language, and epistemology cannot be carried out in a hermetically sealed philosophical discourse, but must incorporate what we know about the incarnation of all of these processes in living brains. The visual arts and the study of visual perception could mutually inform one anodier. An example is evolutionary aesthetics, and the analysis of why certain colors, forms, faces, and landscapes elicit certain afFective and cognitive responses. Or consider the cross-fertilization between jurisprudence and moral philosophy on the one hand, and moral psychology on the other. Some of our debates in ethics might be colored by basic intuitions about right and wrong that we may have inherited from our evolutionary background. Let me focus on two instances, each of which I've done some work on: how the studies of language within the humanities can interact with the scientific study of language from psycholinguistics and neurolinguistics; and how the analysis of fiction can interact with cognitive and evolutionary psychology. Let's begin with a simple problem in the science of language, the vast expressive power of language. Just by making noises with my mouth I can cause you to think an unlimited domain of ideas, everything from theories of the origin of the universe to the developments in your favorite reality show on television. The puzzle we have to explain as language scientists is how this is possible. What's behind our knack of causing each other to think an unlimited range of ideas just by making noises with our mouths? It turns out that there is not one trick behind this talent, but two. One is the trick of the memorized word, whose underlying principle was captured by Ferdinand de Saussure as an arbitrary pairing between a sound and meaning. The word duck, for example, doesn't look like a duck, walk like a duck, or quack like a duck. But 1 can use it to get you to think of a duck, because all of us at some point in our childhoods have memorized an association between that sound and that idea. That means that we have stored in our brain a linkage between that sound and that meaning. But we don't just blurt out individual words; we combine them. And that brings me to the second principle behind language: compositorial grammar, most famously worked out by the linguist Noam Chomsky, in which we combine words in such a way that the meaning of the phrase or sentence can be computed from the meanings of the individual words and the way they're arranged. In very simplified form, we have in our heads rules that say that in English a sentence is composed of a noun phrase followed by a verb phrase (the subject and the predicate). The power of compositorial grammar is that it's composed out of symbols that can be recombined in an unlimited number of ways. They plug into each other, allowing us to create and understand sentences that we had not memorized in the way that we memorize individual words. Now a test of the idea that the language faculty has a bipartite architecture is to find some area of language in which the putative word system and the putative rule system express the same ideas, but still would be psychologically and ultimately neurobiologically distinguishable. We do have such a case: what you might dimly remember from grammar classes as the difference between regular and irregular forms. A refresher: verbs in English and in many other languages come in two flavors. There are regular verbs like walkl walked, jogljogged, and kisslkissed, which form the past tense in a completely predictable way: take the verb and glue the suffix "ed" onto the end. This is an open-ended class. There are thousands of existing regular verbs in the English language, and new ones are being added to the language all the time. For example, when you first heard the verb to spam, in the sense of unwanted e-mail, you didn't have to go to the dictionary to look up its past tense form. Everyone instinctively knows that it's spammed. Likewise, snarf, mosh, flame, diss, and other neologisms are instantly inflected as snarfed, masked, flamed, and dissed without the benefit of memory. Even children do it. If you bring a child into the lab and say, '"Here is a man who knows how to wug. He did the same thing yesterday," children will fill in the blank by saying, "wugged," a form that they could not have memorized from their parents. And all children show off this talent when they go through the stage in which they make grammatical errors such as "we holded the baby rabbits" or "the alligator goed kerplunk." They couldn't have memorized such verb forms from their parents, but must have created them in their minds. The second flavor of verb is irregular-forms like bring brought, go/ went, sing/sang, sleepl slept, make!made, and flylflew. They differ from the regulars in just about every way. In contrast to the monotonous predictability of regular forms, the irregulars are idiosyncratic and quirky. The past of sink is sank, but the past of cling is not clang, but clung. The past of think is neither thank nor thunk, but thought. The past of blink is neither blank not blunkor blought, but is regular, blinked. And the irregulars form a closed class. There are about 165 irregular verbs in contemporary English and we haven't seen any recent new ones. This leads to a simple theory: irregular forms are words that you memorize in the same way that you memorize dog and duck and man. That is, you memorize the verb itself, bring, and you memorize its past tense form, brought. Each is a separate entry in memory. They are linked, because one of them is the past tense of the other, but in other regards they're simply committed to memory separately. Regular forms, in contrast, don't have to be stored, because they can generated by a rule whenever they are needed. Then you can just throw a regular form away and add the suffix to the verb the next time you need one. If this idea is correct, the brain system for word memory should be tied to irregular forms, and the brain system for online grammatical computation should be tied to tegular forms. There is some evidence that this is true. In a study that I did with Jaemin Rhee using magnetoencepholography, which measures the faint magnetic signals emanating from die brain during mental activity, we presented verbs to subjects. They had to pronounce the past tense form while their heads were in the scanner. At each time slice the apparatus estimated the location of the source of the magnetic field pattern, namely the magnetic dipole. Initially, in the first quarter of a second after seeing the verb stem, the greatest activity for both regular and irregular verbs is in the same part of the brain, in the temporal lobe, in which it's thought that memories are formed and stored. About a tenth of a second later, with regular forms, you get a second patter of activity in more anterior regions of the brain, those that are thought to be responsible for planning and computation in the online assembly of forms. Are there any implications of this scientific understanding for issues that are studied in the humanities? For one thing, there are implications for explaining the nature and history of languages as we find them today, which are products of human minds accumulated over centuries. If you look at the most frequent verbs in English in terms of number of occurrences in a million words of text, you get the following top ten list: be have da make go take come see get Interestingly, all ten of them are irregular (bel was, hovel had, do/ did, say/ said, make/ made, go/ went, and so on), a fact that frustrates learners of English as a second language who find that the verbs used in almost every sentence are those most difficult to master. Now let's look at the least frequent verbs in the language. There's no such thing as a bottom-ten list, because we find a 788-way tie for last place, last place being one in a million which, of course, is the lowest frequency you can measure in a million word corpus. But here are the first ten of those 788 verbs in alphabetical order: abate abbreviate abhor ablate abridge abrogate acclimatize acculturate admix adulterate As you can see, they are all regular: abate/ abated, abbreviate/ abbreviated, and so on. You find this in most languages-a massive correlation between frequency and irregularity. Although it might seem perverse, there is a simple explanation in terms of the psychology of language. Irregulars depend on memory, because they're idiosyncratic and have to be learned one at a time. Memory depends on frequency, the more often you hear something, the better you remember it. Any word that declines in popularity may reach a point at which it isn't heard often enough to be memorized uniformly by a generation of children. If the word is not memorized, children will default to the regular "-ed" form, converting the verb from irregular to regular for their generation and all subsequent generations. There's some evidence from philology that that's true. Old English and Middle English had more than twice as many irregular verbs as we find today. If Chaucer were here he would tell you that the past tense of cleave is clove, likewise, crawl crew, chide/ chid, gripe/grope, and so on. The common verbs in Chaucer's time stayed irregular, and the rarer ones defected to the regular side, which is exactly what you'd expect based on the properties of human memory. In addition, the regular rule got first dibs on all of the new entries into the language when English welcomed a massive influx of verbs from French after the Norman Invasion in 1066, and then from Latin during the Renaissance. As with spam, flame, diss, and so on, they got sucked into the regular camp, leading to the enormous imbalance in favor of the regulars in English today. You can actually feel this historical force acting today, in your own mind: the remaining infrequent irregulars (the ones that are at the lower end of the frequency range) sound strange to our ears. For example, if I ask you to complete the sequence "I stride. I strode. I have...," you might be tempted to say stridden, but there's something a little bit odd about it. Likewise, consider smite smote, slay/slew, bid/bade, and forsake!forsook. All have a bookish feel, and one can almost hear them slipping out of the language before our ears. One can predict that in a few hundred years they'll have the status of chide/chid and gripe/grope. This account also has implications for style. Because irregulars are dependent on memory and exposure, they can serve as a shibboleth in a consideration of what is "proper" English, the kind of thing we correct in the term papers of our undergraduates or get exercised about when we worry about the decline of the language. Many of the bones of contention are irregular forms. Data is technically the plural of datum, so one should say "The data are" rather than "The data is." There are people who get up in arms over the Disney movie Honey, I Shrunk the IGJs, which they say should have been Honey, I Shrank the Kids. There are fastidious people who get upset when people thank the alumnis. The reason that these irregular forms are such good markers of education and breeding is that you can't deduce them by a general rule. Knowledge of them crucially depends on which people you hung out with at various stages of your life. Here is another implication for style. Whenever a verb is a doublet, with both regular and irregular forms, poets, novelists, and lyricists tend to prefer irregular forms like striven, shone, slew, hove, clove. This is in part because of euphony. Irregular forms have the sound of basic words in the language, whereas regular forms sound like agglomerations of a stem and a suffix glued together by a rule. Since irregulars fit the phonological template for a basic word, they are more pleasing to the ear. In contrast, in the product of a concatenation operation, the two units can sometimes be stuck together without any consideration for euphony. In extreme form you get tongue twisters such as edited or sixths which are perfectly grammatical, but could never have passed muster as ordinary English words. Also, irregular forms are evocative. Again, because they need to be memorized, they are ded in memory and emotion to the circumstances in which they were originally learned. For most people it would be the King James Bible, for example, for forms like smote and slew. Finally, the organization of words in memory parallels some of the most commonly used poetic devices. The irregular forms are not completely idiosyncratic, but tend to fall into families of similar-sounding verbs with similar-sounding past tense forms: wear! wore, hear! bore, tear! tore, wind! wound, grind!ground, bind] bound, swing/swung, fling/flung, ring/rung. They also may be grouped by alliteration: swing/swung, stick!stuck, strike/struck, sneak!snuck. These verbs don't rhyme with each other, but they all begin with consonant clusters including "s." Sometimes they're grouped by assonance: fielI felt, dream! dreamt, keep! kept, mean! meant. You get an ele alternation in all of them. This is puzzling. Why should a list of irregular vocabulary items sound like bad poetry when you read them in sequence? Why would the devices used by poets also characterize the contents of the mental dictionary, just sitting there in no particular order? The basic idea-first articulated by Paul Kiparsky, and by Roman Jakobson before him-comes from the fundamental principle of linguistics that words and phrases are not just linear sequences of sounds but are mentally represented in hierarchical tree structures. Poetic and rhetorical language often repeats one of the basic structures of language: exacdy those structures that linguists say are the ones in which words are stored in memory. Linguists represent words in terms of a split between an onset and a rime. When you repeat a rime you get a rhyme. When onsets are repeated, this gives rise to the phenomenon of alliteration. Assonance arises from the repetition of what linguists call the vocalic nucleus namely, the vowel in the middle. And indeed, at a higher level when surface or deep syntactic phrase structures are repeated you get the phenomenon known as structural parallelism, much used in poetry and rhetoric. The general moral for consilience is that in this area it's not clear where the science leaves off and the poetics begins. I would be hard-pressed to say that part of this investigation belongs in one pile of bricks and mortar at a university and another part belongs in another pile. Let me mention another case study, the psychology of fiction. Again, I'll start with a puzzle: in all cultures people lose themselves in stories. They expend time and energy thinking about people and events that don't exist. One might ask, from a Darwinian perspective, why would a mind have evolved that would seemingly waste time and energy thinking about hypothetical worlds? One hypothesis is that fiction does have an adaptive function, similar to what philosophers call a thought experiment. In a work of fiction the author places a character in a hypothetical situation, a world with at least some similarity to our world. The protagonist is given a goal and pursues it in the face of obstacles, and the author plays out the consequences for the reader. The reader watches what happens, takes mental notes of the outcomes of the various strategies and tactics in pursuing those goals. Since none of us can actually benefit from the experience of living all of the scenarios that we're likely to face in the future, we can benefit vicariously by seeing what would happen in a fictional version of the world. There's a rich cognitive psychology that applies to this process, much of which has been worked out by Lisa Zunshine of the University of Kentucky. First, there's the capacity for metarepresentation: representations of representations which allow us to entertain a realistic hypothetical world without confusing it with reality. For instance, we may entertain a representation of a person, while knowing that that representation does not refer to a real person in the world-a remarkable ability of the human brain. (It's been suggested that a breakdown of this ability is the main symptom in schizophrenia.) second, there is our "theory of mind" or intuitive psychology: the ability to infer beliefs and desires from characters' behavior. As Zunshine has shown, one parameter in fiction is whether the narrator requires the reader to infer thoughts and feelings from a character's behavior, or reports the contents of the character's consciousness directly. A third is the use of visual imagery-a recruitment of a cognitive process in fiction that has been explored by Elaine Scary and others. By the use of words, authors can evoke visual images and manipulate the reader's perspective, vividness of imagery, and so on. But fiction is not merely an exercise of our intellectual faculties. It engages our emotions and our social psychology. It's often been noted that the goals of fictitious characters are at least indirectly related to the ultimate goals of evolution: survival and reproduction. This is not the red-in-tooth-and-claw struggle that we're familiar with from the animal world, but the struggle for survival and reproduction in a species that's intensely cognitive and social-a species that lives by its wits, and by its social coalitions. In his delightful book The 36 Dramatic Situations, Georges Polti tries to categorize the world's literature and myths into a small number of archetypes. I found that about 80 percent of them can be characterized either as involving adversaries, or as involving tragethes of kinship or love. Examples include "mistaken jealousy," "the vengeance taken by kindred upon kindred," "the discovery of the dishonor of a loved one," or "an enemy loved." All are familiar to evolutionary psychologists. Yet if literature were nothing more than this, it would be a dreary affair. Despite claims by Polti and others that there are a finite number of plots, we also know that literature is openended, that we will never run out of plots and novels. While the summaries may be finite, fiction itself is not. One reason is that the main obstacles to a protagonist's goals are other people in pursuit of their own goals, leading to a combinatorial explosion of possibilities. The fundamental insight of modern sociobiology and evolutionary psychology-owing to Robert Trivers-it is diat every kind of social relationship involves both a confluence and a conflict of interests. These partial conflicts of interest are more cognidvely fertile than a neartotal overlap of interests (such as you find in social insects like bees and ants, which are highly genetically related and therefore make sacrifices for one another that would puzzle a human), or a total conflict of interests, a war of all against all. With some overlap and some non-overlap one has a far more fecund situation for mentally playing out the possibilities. Though I've attributed mis insight to Trivers and evolutionary psychology, the temporal sequence may be the other way around: commentators on fiction could have educated evolutionary biologists with this insight long ago. To take one example, an evolutionary biologist would look at the relationship between two spouses in the following way. The overlap comes from the fact that the Darwinian interests of a pair of spouses are identical under the following idealization: if the spouses are faithful, if they favor their own nuclear family over their blood relatives, and if they can expect to the at the same time under those circumstances, they are as, as far as evolution is concerned, one flesh, and what is good for one is good for the other. One would expect perfect empathy and perfect love. Needless to say those aren't the conditions of most human marriages. Marital conflict is expected to be triggered by infidelity, real or imagined, by in-laws, and by the investment horizon-how long a reproductive career you expect in front of you. I daresay it's likely that many works of romance find the protagonists stumbling into conflict over exactly these areas. The relationship between a parent and child likewise has some degree of overlap and some degree of non-overlap, as well as sibling and sibling, man and woman, friend and friend, individual and society. Trivers argued that even the relationship of a person to himself is not one of complete overlap, but gives rise to areas of conflict as well. It has long been known that partial conflict is a source of the enduring appeal of fiction, but not the reason why. Aristotle said that a story about two strangers who fight to the death is nowhere near as interesting as a story about two brothers who fight to the death. Polti's dramatic situations all involve partial conflicts of interest. In his book Antigones, George Steiner suggests that Antigone may be the best work of fiction ever written. His reasoning captures the hypothesis perfecdy. He says, "It has, I believe, been given to only one literary text to express all the principal constants of conflict in the condition of man. These constants are fivefold: the confrontation of man and of woman; of age and of youth; of society and of the individual; of the living and the dead; and of men and gods.... [BJecause Greek myths encode certain primary biological and social confrontations and self-perceptions in the history of man, they endure as an animate legacy in collective memory and recognition." To sum up: I have suggested that the malaise of the humanities is related to their isolation from the sciences, and that we now have an opportunity to extend the consilience enjoyed by the sciences in the past to the social sciences and the humanities. One connection is deep history, another is the sciences of human nature, in particular the sciences of mind, brain, genes, and evolution. I've given examples from language and literature. In concluding, I want to acknowledge that there is an obvious objection to applications of science to the humanities. The applications we've seen so far are bound to consider only a fraction of the richness in a work of an or scholarship. To understand a work of art requires expertise in the particulars of the work and the idiom, and not just generalities of psychology and biology. The objection is completely accurate. What I envision is not a hostile takeover of the humanities, but rather opportunities for convergence from different approaches. A better metaphor than a takeover is two teams tunneling into a mountain from opposite sides and meeting in the middle.
gripe grope chide chid
bid bade cleave cloved [Sidebar]
forsake forsook [Author Affiliation]
Steven Pinker is the Johnstone Family Professor in the Department of Psychology at Harvard University. Until 2003, he taught in the Department of Brain and Cognitive Sciences at MIT. He conducts research on language and cognition, writes for publications such as The New York Times, Time, and Slate, and is the author of six books, includingThe Language Instinct, How the Mind Works, Words and Rules, and The Blank Slate. This article is adapted from a lecture delivered at the Poetics-Cognitive Science Colloquoy in November 2005 at the Dactyl Foundation in New York City and co-sponsored by the Center for Inquiry-New York City.