Home Page || The Calvin Bookshelf || Science Surf magazine || Table of Contents
A book by
William H. Calvin
The Throwing Madonna
Essays on the Brain
Copyright 1983, 1991 by William H. Calvin.

You may download this for personal reading but may not redistribute or archive without permission (exception: teachers should feel free to print out a chapter and photocopy it for students).

Scanned, OCR'ed, and webbed -- but NOT proofread (14 Jan 97)


When we study human language, we are approaching what some might call the "human essence," the distinctive qualities of mind that are, so far as we know, unique to man.
NOAM CHOMSKY, Language and Mind

No matter how eloquently a dog may bark, he cannot tell you that his parents were poor but honest.


The great difference between a human being and the cleverest chimpanzee is that the human being can have images of persons, times, and places that have not been personally experienced.


Social insects behave like the working parts of an immense central nervous system: The termite colony is an enormous brain on millions of legs; the individual termite is a mobile neuron. This would mean that there is such a phenomenon as collective thinking, that goes on whenever sufficient numbers of creatures are sufficiently connected to each other. It would also mean that we humans could do the same trick if we tried, and perhaps we've already done it, over and over again, in the making of language....

LEWIS THOMAS, in Discover magazine

Much language is symbolic and poets teach us to use words with special force. We may need their help in finding new ways to talk about brains.
J. Z. YOUNG, Programs of the Brain


Linguistics and the Brain's Buffer

In the beginning was the word. But by the time the second word was added to it, there was trouble. For with it came syntax....
JOHN SIMON, Paradigms Lost

I don't want to talk grammar. I want to talk like a lady.

Eliza, in GEORGE BERNARD SHAW s Pygmalion

Strange though it might seem as an introduction to the study of the spoken word, consider the visual world. Color, pattern recognition, neuroanatomy, optics, illusions, contrast, composition--there is no end of it. Each topic is a subspecialty. But the reductionist methods of neurobiology have made headway, analyzing the wiring of the retina, studying the transformations that take place at each relay station in the brain (Chapter 11 of Inside the Brain relates that story; David Hubel and Torsten Wiesel won the Nobel Prize in 1981 for their role in it). Many aspects (such as color) are still unsettled, but the neural templates (usually called "receptive fields") discovered by the reductionist approach now aid everyone's thinking.
      If vision seems like a large subject, consider language: speaking and gesturing, listening and watching, phonemes, words, syntax, rhythms and emotional colorations, writing-- there is no end to the subject, partly because it is constantly changing--indeed probably still rapidly evolving, unlike the visual system. Just as the natural sciences split off from philosophy several centuries ago, so language's traditional relation to philosophy has begun to be replaced by ties to psychology, neurophysiology, and evolutionary biology. We now think of listening to speech in somewhat the manner of cognitive psychology: that there are a series of templates, or schemata, in the brain to which sound sequences are matched, to which word groups are matched. That the correctness of a sentence (whether it has "good grammar") is a matter of goodness-of-fit to one of the schema formed when you learned the language. Our thinking about language understanding is tending toward the model established by those templates for visual patterns that proved so successful.
      This chapter, and the one that follows, will emphasize this template thinking and also another important notion: that of an increasingly large "buffer" memory. Buffers hold on to something temporarily; for example, when you dial a telephone number, the central off~ce machinery has a buffer that holds the numbers until the last one is finished, then it spits them out in one rapid burst to a distant switching computer (you can often hear this as a rapid series of tones when the call is going through). The size of the buffer is seven digits long, with larger buffers for long distance dialing. Your brain "buffers" a string of numbers which the Directory Assistance operator tells you until you can dial it. A buffer is a scratch pad, not intended to save something forever but to hang on to it long enough for some operation to be carried out.
      For some buffers (like a grocery list), order is unimportant. But for the buffers we will be considering, keeping things in the right order is terribly important (jumbled phone numbers are worthless, jumbled sentences mean something different), and we will be concerned with buffer size for much the same reasons that elementary textbooks use shorter sentences than advanced ones: more complicated propositions require longer word buffers in the brain.
      From the outside, language is studied by us all, experts that we are in the acquisition of new words, in effortlessly translating a series of sounds into a meaning and acting upon it. Children are especially attuned to acquiring new words by imitation (it was noted over two centuries ago that European children had no trouble learning the Hottentot language, even if their parents couldn't); just like parrots, the children may try to produce the sounds they hear. But unlike parrots, they also invent new words and try them out: if they learn "bad" and "dad" they may try out "bate" and "dab." Infants seem to be born with an ability to distinguish many phonemes; e.g., to tell /ba/ from /pa/, where the major difference is the repeated vibration of the vocal cords near the beginning of /ba/.
      Soon children speaking words are trying out new ways of combining them. They may say "Where it is?" instead of "Where is it?" and only gradually realize that others do not use this arrangement of subject, object and verb (except, of course, in saying "Show me where it is"). We build up a series of acceptable templates or schemata for phoneme order (words) and for word order (phrases) which allow us to make sense out of a sentence we have never encountered before, such as this one.
      Linguistics did not start from this neurobiological perspective but from the comparative study of languages. The common features of all languages led to the notion that the regularities between otherwise dissimilar languages might have something to do with biology, with features of the brain.
      The deep structure of language is defined, more or less, as that which allows us to say that two sentences are synonymous; e.g., "Mary washed John" is the same relation of actor-action-object as is "John was washed by Mary." Linguists argue about just how deep the deep structure should be removed from the surface structure of the sentences, but many of their diff~culties are related to trying to create a one-way series of processing steps--the little boxes with labels connected with arrows-- relating syntax (word order) to semantics (meaning) and to surface structure (grammar) and phonetics (e.g., inflections).
      Naturally, linguists have attempted to infer a series of steps in the production of a sentence. But there seem to be too many feedback loops, as when the meaning of the sentence (deep structure, presumably) is changed at the last moment simply by the voice rising or falling at the end of "You are happy." Which, of course can communicate a lot: Maureen Howard tells of growing up in a neighborhood "where by a mere inflection of 'Hello' you could tell that the pert young widow had lost her beau, the Montours were coming up in the world this week, the Drews had not paid their grocery bill." Indeed, from the single word "Hello" at the beginning of a phone call, you can frequently identify a unique individual in the whole world; even if you don't recognize the individual, you can still make a good estimate of the speaker's sex, age, friendliness, nationality, education or socioeconomic status, and health. If one word can convey an enormous amount of information far beyond its literal meaning, then it is hardly surprising that linguistics has trouble with simple theories for the production of word order.
      That deep structure is innate and preprogrammed is perhaps the best known of Noam Chomsky's arguments, made from the comparative study of languages and similar sources. This notion seems to have come as a revelation in those areas of the humanities whose neurological vocabulary is limited to tabula rasa, though I have yet to meet a biologist who was surprised. Linguistics is, for better or worse, common ground for the humanities and the natural sciences, one of the major places where C. P. Snow's two cultures meet. And deep structure is a major point of congruence.
      Though seldom expressed in neurological terms by its enthusiasts, innate deep structure is just another way of saying that we must seek a neurological basis of the linguistic actor-action-object equivalence relations in the anatomy and physiology of the brain. If language is a system which is superimposed upon another older sensory-motor-regulatory neural system, then we might hope to see the actor-action-object paradigm in the more primitive system.
      chimera drawing from THE CEREBRAL CODE In the throwing madonna painting, one sees actor-action-object: it depicts a mother-with-infant, a throwing motion with rock in flight, and the rabbit target. Most motions outside of simple locomotion have goals. Ballistic motions, in particular, are simply characterized by a starting point and an end point with a rapid motion in between. The actor is self: self, motion, goal. In looking around the environment, the actor may be another subject of interests: a deer approaching a salt lick. Like the difference between scalar and vector qualities in mathematics, the representation is not a simple static set of interrelationships: motion, or at least a direction, is suggested. And so one wonders whether actor-action-object is reflected in sentence structure in more than a casual way, whether subject-verb-object is an offshoot of a nonverbal paradigm.
      Spoken language output is sequential (sign language may, in contrast, express various sentence elements simultaneously). Furthermore, word order plays an important role ("John washed Mary" has a quite different meaning from "Mary washed John"). Tactical considerations in language also influence the choice of the active form "Mary washed John" over the softer passive construction "John was washed by Mary"--or the invention of an even more pronounced circumlocution.
      This emphasis on sequence tends to bias our view of language processes, as witness the sequential black boxes that give linguistic theorists such difficulties. But the premotor aspects of spoken language need not be any more sequential than the premotor aspects of movement planning: in deciding when and where to pounce, my pet cat is simultaneously taking into account all sorts of factors including her present posture, the bird's position, and strategic considerations such as whether the bird is too big to tackle (or whether I am watching). The fact that the pounce starts with extension of knee and hip joints does not mean that this aspect was planned first or last.
      The subject-verb-object (SVO) order of most English language sentences ("Borg served an ace") happens to correspond with the stated order of the actor-action-object paradigm. Is that important? The Japanese might consider it a bit of cultural imperialism because typical Japanese sentences have the verb last (SOV: "Borg an ace served"). At least in stating complicated logical propositions, there are alleged to be some advantages to the SOV order. The arguments can be heard among students having the two different kinds of hand calculators: those that use the traditional "A plus B equals" keystrokes mimicking SVO and those requiring the SOV-like reverse Polish format "A, B, plus." Or among the computer enthusiasts similarly favoring the SOV-like language called FORTH. Essentially, one can avoid clustering ambiguities--which otherwise require the liberal use of parentheses and brackets--in reverse Polish.
      In German, the verb is last only in dependent clauses: "If SOV, then SVO." In classical Arabic, the verb comes first: VSO ("Served Borg an ace"). Indeed, of the six possible permutations of the three items, all six are used as the standard in one language or another (it was once thought that OVS was an exception, but Hixkaryana is now known to use it: "An ace served Borg"). A language's choice of major word order does seem to often have implications for the minor words: for example, prepositions in Japanese are postpositions ("by bus" would come out as "bus by"). All of which would indicate that, while sentences are being assembled in the brain's buffer, there is no mandatory biological order, only that dictated by local convention and long practice. It is probably as reasonable to say that deep structure reflects actor-object-action as actor-action-object. Alas, the throwing order does not help us understand word order, at least at this level.
      It may, however, be worth pursuing word-order cues when the language is not overlearned and has a different word-order preference than the native language. I can remember being distressed by SOV when learning German: to have to read through to the last word in the clause before the action was revealed seemed like an interminable wait (and at my reading speed, it was). It led Mark Twain to comment in A Connecticut Yankee in King Arthur's Court: "Whenever the literary German dives into a sentence, that is the last you are going to see of him till he emerges on the other side of the Atlantic with his verb in his mouth."
      One can, of course, speculate endlessly on the philosophical implications of such word-order preferences for national personality types: that SOV might produce, for example, a more contemplative personality because one must suspend judgment until all the facts are in--rather than, under English-style SVO, impetuously getting the action in motion before its target is revealed.
      The brain's buffer for words is also essential when trying to fathom a long sentence with many clauses. You need to analyze each subject-verb-object group, translate it into deep structure of actor-action-object, repeat the analysis for each phrase of the sentence, and then relate them to the sentence taken as a whole, performing yet another actor-action-object deep structure job on the overall reduced sentence. (The in-group term for repeating the action over and over for improving results is "recursive," as in that tracing-over exercise from first grade when you learn cursive handwriting.) The rules of word order (grammar) have to be applied recursively, and to do that you need a big sequential buffer to store the whole sentence temporarily.
      But even 5-year-old children can accomplish that. Consider the nursery rhyme about the house that Jack built:

This is the farmer sowing the corn,
That kept the cock that crowed in the morn,
That waked the priest all shaven and shorn,
That married the man all tattered and torn,
That kissed the maiden all forlorn,
That milked the cow with the crumpled horn,
That tossed the dog
That worried the cat
That killed the rat
That ate the malt
That lay in the house that Jack built.
Each phrase is analyzed and then substituted for the "That" of the next line, simply becoming the new actor. But adult sentences have embedded phrases that can modify each of the actor-action-object triumvirate: "I believe Jack says you think you heard him."
      You need a good holding buffer, and as noted in Chapter 5, evolution has long had an increasing need for short-to medium term memory storage as social life has become more complex, as omnivore tendencies to try out new foods have expanded. But an extremely sequential memory, rather than a general gestalt memory. What could have selected for a sequential buffer? Ballistic motor sequencing is one candidate.
      The motor buffer is perhaps more easily understood when it is used for speaking a sentence. If language was planned using a motor buffer for storing the sentence, many factors could operate simultaneously and interactively upon that sentence buffer: semantics, phonetics, and transformation rules. One could be speaking while planning ahead, with the words already spoken serving as a constraint on subsequent syntax; as the Roman poet Horace said, "Once a word has been allowed to escape, it cannot be recalled." But otherwise, there might remain much freedom to create alternative sentences which all satisfy the same desired actor-action-object interrelationships. As when we change our minds in mid-sentence and complete the sentence with a new ending, but fitting it to the grammar of what had come before.
      But a motor buffer for a sensory decoding task? Strange but probably true: The learning-by-imitation theory has been extended to listening-using-(subaudible) speech to account for various phonological and physiological evidence. The motor theory of language perception and the cortical mapping evidence (discussed in Chapters 4 and 16) say that we are using some motor templates to sort phonemes into categories (telling the difference between /ba/ and /pa/ by using the premotor program for forming those phonemes), though with the muscular output suppressed at some premotor stage.
      Two aspects of motor systems have entered our discussion of language: the concept of motor templates in a sensory perception task, and the concept of a motor sequencer--in particular, a buffer into which proposed actions can be stored during "get set" and then rapidly emptied at "go."
      Yet the same buffer could hold the incoming string of phonemes, allow schemata to identify the words, then allow surface and deep structure to be analyzed--recursively, if there are embedded phrases. In other words, deep structure could communicate simultaneously with the other levels of both language reception and production.
      Though the examples given in this book tend to imply that the buffer was shaped by throwing success and that the sequential actor-action-object is a paradigm derived from throwing stones, this need not be the only deep paradigm on which language could operate. The equivalence relationship, used by the cat to know that creeping up on the bird from behind will accomplish the same actor-action-object relationship as a frontal attack, is also what we use to tell that two sentences are synonymous or that sui generis means about the same as unique or peculiar.
      Which brings up the contention of some linguists that language is a human capacity sui generis, without parallel in the animal kingdom or in other human systems. Though probably not intended in the same spirit, this always reminds me of the creationists' argument for special creation (because humans are so special, they have followed a different path of creation than mere animals). We may yet come to conclude that human verbal language is unique, but first we must try out some more mundane hypotheses. Such as evolution from gestures. Such as evolution from species-specific vocalizations. Such as the possibility that language was built atop a motor sequencing system which greatly enlarged for another "reason" (throwing range and bigger prey) and selection pressure (hunting success, as opposed to communication success). Such as the idea that recognition may consist of fitting incoming information to templates or schemata, just as syntax may correspond to how well an utterance fits with a schema of prototypical word-type sequence formed by an experience-modified innate premotor program.
      So sequence and grammar are all important--but, alas, I still cannot diagram a sentence. That is likely left over from high school. I was comparing notes with a classmate of mine who got Ds in English but still managed to become an editor of the Los Angeles Times (he still refuses to diagram sentences). We couldn't stand the diagramming exercise because we already knew what was correct from long experience with language. Taking a sentence apart into pieces seemed pointless, since we already knew what sounded correct and what didn't. It wasn't until I tried to learn German in college that I suddenly realized what grammar was all about: spotting the different parts of the sentence is quite essential if your vocabulary is shaky.
      I now suspect, having been further exposed to examples of overquantification and "physics envy" in the softer sciences, that our teacher was trying to make English grammar into a rigorous subject comparable to Euclidian geometry. But language isn't like that: it is naturally a bit loose and always evolving (just try reading something written 200 years ago in the original). There are some foundations underneath, as Chomsky noted, and they may rest on some nonlanguage specializations of the brain for another sequential task needing a big holding buffer.
      "Get set" and "throw" as a foundation stone for language? Stranger things than that have happened in evolution. Birds, for example.
1997 note: Linguists such as Derek Bickerton have been trying hard to educate me since 1982, when this essay was written. My more recent attempts can be found in chapter 5 of How Brains Think (1996) and the last chapter of The Cerebral Code (1996).
The Throwing Madonna:
Essays on the Brain
(McGraw-Hill 1983, Bantam 1991) is a group of 17 essays: The Throwing Madonna; The Lovable Cat: Mimicry Strikes Again; Woman the Toolmaker? Did Throwing Stones Lead to Bigger Brains? The Ratchets of Social Evolution; The Computer as Metaphor in Neurobiology; Last Year in Jerusalem; Computing Without Nerve Impulses; Aplysia, the Hare of the Ocean; Left Brain, Right Brain: Science or the New Phrenology? What to Do About Tic Douloureux; The Woodrow Wilson Story; Thinking Clearly About Schizophrenia; Of Cancer Pain, Magic Bullets, and Humor; Linguistics and the Brain's Buffer; Probing Language Cortex: The Second Wave; and The Creation Myth, Updated: A Scenario for Humankind. Note that my throwing theory for language origins (last 3 essays) has nothing to do with the title essay: THE THROWING MADONNA is a parody (involving maternal heartbeat sounds!) on the typically-male theories of handedness.
Many libraries have it (try the OCLC on-line listing, which cryptically shows the libraries that own a copy), and used bookstores may have either the 1983 or the 1991 edition.

Email || Home Page || The Calvin Bookshelf || Table of Contents || End notes for this chapter || Continue to read next chapter