Popular Science Monthly/Volume 28/March 1886/A Thinking Machine
| A THINKING MACHINE. |
By GRANT ALLEN.
"THINGS marvelous there are many," says the Attic dramatist, "but among them all naught moves more truly marvelous than man." And, indeed, when one begins seriously to think it over, there is no machine in all the world one-half, nay one-millionth part, so extraordinary in its mode of action as the human brain. Minutely constructed, inscrutable in all its cranks and wheels, composed of numberless cells and batteries, all connected together by microscopically tiny telegraphic wires, and so designed (whether by superior intelligence or evolutionary art) that every portion of it answers sympathetically to some fact or energy of the external universe—the human brain defies the clumsy analysis of our carving-knife anatomists, and remains to this day a great unknown and almost unmapped region, the terra incognita of modern physiology. If you look into any one of the ordinary human machines, with its spokes and cogs, its springs and levers, you can see at once (at least, if you have a spark of native mechanical intelligence within you) how its various portions are meant to run together, and what is the result, the actual work, to be ultimately got out of it. But not the profoundest microscopist, not the acutest psychologist, not the most learned physiologist on earth could possibly say, by inspecting a given little bit of the central nervous mechanism of humanity, why the excitation of this or that fragment of gray matter should give rise to the picture of a brown umbrella or the emotion of jealousy, why it should rather be connected with the comprehension of a mathematical problem than with the consciousness of pain or the memory of a gray-haired, military-looking gentleman whom we met three years ago at an hotel at Biarritz.
Merely to state these possible alternatives of the stimulation of a portion of the brain is sufficient to bring up vividly into view the enormous and almost inconceivable complexity of that wonderful natural mechanism. Imagine for a moment a machine so delicate that it is capable of yielding us the sensation of a strawberry-ice, the æsthetic delight of a beautiful picture, the intellectual perception of the equality of the angles at the base of an isosceles triangle, the recollection of what we all said and did the day we went for that picnic to the Dolgelly waterfalls, the vague and inconsistent dissolving views of a disturbed dream, the pain of toothache, and the delight at meeting once more an old friend who has returned from India. The very mention of such a complicated machinery, let alone the difficulty of its possession of consciousness, is enough to make the notion thus nakedly stated seem wild and absurd. Yet there the machine actually is, to answer bodily for its own possibility. You can not cavil at the accomplished fact. It may be inconceivable, but at any rate it exists. Logic may demolish it; ridicule may explode it; metaphysics may explain it away; but, in spite of them all, it continues still imperturbably to be, and to perform the thousand-and-one incredible functions which argument conclusively and triumphantly demonstrates it can never compass. Call it materialism or what else you like, experimental physiology has now calmly demonstrated the irrefragable fact that on the brain, and on each of its parts, depends the whole of what we are and what we feel, what we see and what we suffer, what we believe and what we imagine. Everything that in our inmost souls we think of as Us, apart from that mere external burden, our body, is summed up in the functions and activity of a single marvelous and inscrutable organism, our human brain.
But, though physiology can tell us very little as yet about how the brain does its work, it can nevertheless tell us something; and late researches have made such a difference in our way of looking at its mode of activity, and have so upset many current and very crudely materialistic errors, that it may perhaps be worth while briefly to state, in popular and comprehensible language, how the organ of thought envisages itself in actual working process to the most advanced among our modern physiological psychologists.
Let us begin first with the old-fashioned and, as we now believe, essentially mistaken view—the view which found its fullest and most grotesque outcome in the spurious science of so-called phrenology, but which still lingers on, more or less carefully disguised, among the "localizations" and "specific energies" of many respectable modern authorities.
According to this superficial view, overtly expressed or implicitly suggested in different cases, each cell and ganglion and twist of the brain had a special function and purpose of its own to subserve, and answered to a single special element of sensation or perception, intellect or emotion. In a certain little round mass of brain-matter, in the part of the head devoted to language (if we push the theory to its extreme conclusion), must have been localized the one word "dog"; in the next little mass must have been localized "horse"; in the next, "camel," in the next again, "elephant," and so on ad infinitum. Here, a particular cell and fiber were intrusted with the memory of the visible orange; there, another similar little nervous element had to do with the recollection of the audible note C flat in the middle octave of a cottage piano. Thus reduced to its naked terms, of course, the theory sounds almost too obviously gross and ridiculous; but something like it, not quite so vividly realized or pushed so far into minute detail, was held not only by the old-fashioned phrenologists, but also by many modern and far more physiological mental philosophers.
When we come to look the question in the face, however, the mere number of cells and fibers in the human brain, immense as it undoubtedly is, would surely never suffice for the almost infinite variety of perceptions and facts with which our memory alone (not to mention any other mental faculty) is so abundantly stored. Suppose, for example, we take merely the human beings, living or extinct, with whose names or personalities we are more or less fully acquainted, and try to give a cell or a fiber or a ganglion to each; how many cells or fibers of ganglia would be left unappropriated at the end of the enumeration for all the rest of animate or inanimate nature, and all the other facts or sensations with which we are perfectly familiar, to say nothing of emotions, volitions, pleasures, pains, and all the other minor elements of our complex being! Let us begin, by way of experiment, with Greek history alone, and try to distribute one separate nerve-element apiece to Solon and Periander, to Themistocles and Aristides, to Herodotus and Thucydides, to Zeuxis and Pheidias, to Socrates and Plato, to Æschylus and Sophocles, to Aristides and Alexander, and so on straight through down to the very days of the Byzantine Empire. Then let us begin afresh over again, and give a cell all round to the noble Romans of our happy school-days, Romulus and Remus (myth or reality matters little for our present purpose), the seven kings and the ten decemvirs, the Curtius who leaped into the gulf and the Scævola who burned his hand off in the Etruscan fire, those terrible Scipios and those grim Gracchi, our enemy Horace with his friend Mæcenas, and so down through all the Cæsars to the second Romulus again, pretty much where we originally started. Once more, apply the same thing to English history, and allot a single brain-element apiece to everybody we can remember from Cerdic of Wessex to Queen Victoria, from Cædmon the poet, through Chaucer, Shakespeare, Milton, and Pope, to Tennyson, Swinburne, and Oscar Wilde—a cell each for all the statesmen, priests, fighters, writers, thinkers, doers, and miscellaneous nobodies whom we can possibly recall from the limbo of forgetfulness, from the days when Hengist and Horsa (alas! more myths) drove their symmetrical three keels ashore at Ebbsfleet, to the events recorded for our present edification in this evening's newspaper. (And observe in passing that, out of deference to advanced Teutonic scholarship, I have simply flung away Caractacus and Boadicea, Carausias and Allectus, and all the other vague and vaguely-remembered personalities of the earlier British and Romano-British history.) Why, by the time we had got through our historic personages alone, we should have but a very scanty remnant of places for the thousands and thousands of living individuals with whom each one of us must have come in contact, and each of whom seems to occupy a separate niche or distinct pigeonhole in the endless archives of the particular memory.
And this is only a single small department of the possibly memorable, a mere specimen category out of an innumerable collection that might equally well have been adduced in evidence. Take the animal world, for example—the creatures themselves, and not their names— and look at the diversity of cats and dogs, goats and sheep, beetles and butterflies, soles and shrimps, that even the ordinary unlearned man knows and recognizes, and mostly remembers. Narrow the question down to dogs alone, and still you get the same result. Consider the St. Bernards and the mastiffs, the pugs and the bull-dogs, the black-and-tans and the King Charlies, the sheep-dogs and the deerhounds, the shivering little Italian greyhounds and the long dachshunds that you buy by the yard. Every one of these and countless others has got to have its cell all to itself in the classificatory department of the human brain, and I suppose another cell for its name in the portion specially devoted to language also. Add to these the plants, flowers, fruits, roots, and other well-known vegetable products whose names are familiar to almost everybody, and what a total you have got at once 1 A good botanist, to take a more specific case, knows (in addition to a stock of general knowledge about equivalent on the average to anybody else's) the names and natures of hundreds and thousands of distinct plants, to say nothing about innumerable small peculiarities of stem, and leaf, and flower, and seed in every species and variety among them all. No, the mere bare weight of dead fact with which everybody's memory is stored and laden defies the possibility of reckoning and pigeon-holing. Make your separate dockets ever so tiny, reduce them all to their smallest dimensions, and yet there will not be room for all of them in the human brain. The more we think on it, the more will the wonder grow that one small head can carry all that the merest infant knows.
And now observe once more in turn a still greater and more fatal difficulty. I have spoken throughout, after the manner of men, as though each separate object, or word, or idea had a clearly defined and limited individuality, and that it could be distinctly located and circumscribed by itself in a single solitary isolated cell of the nervous mechanism. But in reality the very terms I have been obliged to use in describing the matter have themselves contained the implicit condemnation of this crude, hard, and impossible materialistic conception. For no idea and no word is, as a matter of fact, so rigidly one and indivisible, like the French Republic. Take for example once more our old friend "dog," and let us confine our attention just now to the word alone, not to the ideas connoted by it. Dog is not one word: it is a whole group and set of words. There is, first of all, the audible sound, dog, as it falls upon our ears when spoken by another. That is to say, there is, imprimis, dog auditory. Secondly, there is the muscular effort, dog, as it frames itself upon our own lips and vocal organs when we say it aloud to another person. That is to say, there is, secundo, dog pronounceable. Thirdly, there is the written or printed word, dog, DOG, in capitals or minuscules, script, or Roman, or italic, as we recognize it visibly when seen with our eyes in book or letter. That is to say, there is, tertio, dog legible. Now, it is quite clear that each of these three distinct dogs is made up of separate elements, and can not possibly be regarded as being located in a single cell or fiber alone. Dog auditory is made up of the audible consonantal sound D, the audible vowel-sound aŭ or ŏ (unhappily we have no universally recognized phonetic system), and the other audible consonantal sound G hard; in that precise order of sequence and no other. Dog pronounceable is made up of an effort of breath against tongue and teeth, producing the soft dental sound D, followed by an unimpeded vocalized breath, producing the audible vowel-sound aŭ or ŏ, and closed by a stoppage of the tongue against the roof of the mouth, producing the soft palatal G. Finally, dog legible, in print at least, is composed of the separate symbols D and O and G, or d and o and g, or d and o and g. Yet all these distinct and unlike dogs would be unhesitatingly classed by most people under the head of language, and be located by phrenologists, with their clumsy lumping glibness, in the imaginary "bump" thereto assigned, or by more modern physiologists (whose excellent scientific work I should be the last to undervalue) in the particular convolution of the left hemisphere found to be diseased in many cases of "atactic aphasia," or loss of speech.
How infinitely more complex and varied, then, is the idea of dog, for which all these heard, spoken, written, or printed dogs are but so many rough and incomplete symbols! For the idea of dog comprises the head thereof, and the tail, the four legs, the eyes, the mouth, the nose, the neck, the body, the toes, the hair, the bark, the bite, the canine teeth that inflict it, and all the other known and remembered peculiarities of perfect doghood as ideally realizable. If we are to assign peradventure a special tract in the brain to the concept dog, it must be clear at once that that tract will be itself a very large and much subdivided region. For it must include all the separate visible attributes of the dog in general; and also it must contain as sub-species in subordination to it every kind of known dog, not only those already enumerated, but also the Eskimau dog, the Pomeranian, the French poodle, the turnspit, the Australian dingo, the Cuban bloodhound, the Gordon setter, and so forth, through every other form of dog the particular possessor of that individual brain has ever seen, cognized, or heard of. Is it not clear that, on the hypothesis of such definite and distinct localization, dog-tract alone ought to monopolize a region about one sixth as big every way as our whole assignable provision of brain-surface?
Moreover, about this point we seem to be getting ourselves into a sad muddle. For we have next to remember our own private dog, Grip, let us call him, or if you prefer it. Prince or Ponto. Now, I suppose, his name, viewed as a name, will be localized in the language department of our particular brain, and will there be arranged under the general heading of proper names, division dog-names. But there must be some intimate cross-connection between the cell or cells representing the audible and pronounceable name Grip, or the letters O, R, I, P, and the cell or cells which have to do with the idea dog, and also, I imagine, with the name dog: for both the word (Grip is intimately connected in my mind with the words "my dog," and the idea drip is intimately connected in that same humble empirical subjectivity with the idea of dog in general. In fact, I can't think of Grip without thinking at once of his visible appearance, his personal name, and his essential dogginess of name and nature. Grip is to me a symbol, primarily, of some dog or other, and secondarily or more particularly of my dog. But whether Grip and Ponto are arranged and pigeon-holed in cells next door to one another, as being both by name dogs; or whether one is arranged under G, as in a dictionary, and the other under P (just after Pontius, for example, and just before Pontus Euxinus, both of which form distinct component elements of my verbal memory), I can not imagine. At each step in the effort to realize this wooden sort of localization, is it not clear that we are sinking deeper and deeper into a bottomless slough of utter inconceivability?
Once more (and this shall be my last attempt to point out the absurdity of the extreme cell-theory), what are we to make of the case of a man who knows more than one language? Take for example the word chien. Here, in one direction, all the associations and connections of idea are exactly the same as in the word dog. If I happen to be speaking English I say, "It's a dog"; if I happen to be speaking French, I say, "C'est un chien," and in both cases with just about the same idea in my mind. The picture called up by the one word is exactly the same, in most respects, as the picture called up by the other. Yet not precisely. If I write Paris, so, the notion immediately aroused in the reader's mind is that of a white and glaring brand-new city across the Channel where we all go to waste our hard-earned money at periodical intervals. But if in the preceding line I had happened to talk of Priam and Helen, the idea called up by that self-same combination of one capital letter and four small ones would have been a wholly different one, of an idyllic shepherd, as in Tennyson's "Enone," or of a handsome scamp as in (Homer's) "Iliad." If I write "baker," everybody knows I mean the man who supplies hot rolls for breakfast; but if I write "Baker," everybody is aware that I allude to Sir Samuel or to his brother the Pasha. Now, this alternative possibility is even worse in the case of chien. For, if I am talking French, the sight of a particular animal which usually calls up to my lips the word "dog," calls up instead the totally different word chien. And if the subject in hand is philology, while dog immediately suggests to me the curious practical falling out of our language of the primitive word hund, hound, now only applied to a special class of dogs, and the substitution for it of a Scandinavian and Dutch root not found in Anglo-Saxon, chien immediately suggests to me its ultimate derivation from its original canis, and the habitual change of c before a into ch in the passage of words into French from Latin. By this time, I think the reader (with his usual acuteness) will begin to perceive into what a hopeless network of cross connections and crooked combinations we have managed to get ourselves in our search after the definitely localizable.
How, then, does the mechanism of the brain really act? I believe the true answer to this question is the one most fully given by M. Ribot and never yet completely accepted by English psychologists. It acts, for the most part, as a whole; or at least, even the simplest idea or mental act of any sort is a complex of processes involving the most enormously varied brain-elements. Instead of dog being located somewhere in one particular cell of the brain, dog is an idea, audible, visible, legible, pronounceable, requiring for different modes of its perception or production the co-operation of an enormous number of separate cells, fibers, and ganglia.
Let us take an illustration from a kindred case. How clumsy and awkward a supposition it would be if we were to imagine there was a muscle of dancing, and a muscle of walking, and a muscle of rowing, and a muscle of cricketing, and a muscle for the special practice of the noble art of lawn-tennis! Dancing is not a single act; it is a complex series of co-ordinated movements, implying for its proper performance the action of almost all the muscles of the body in different proportions, and in relatively fixed amounts and manners. Even a waltz is complicated enough; but when we come to a quadrille or a set of lancers, everybody can see at once that the figure consists of so many steps forward and so many back; of a bow here, and a twirl there; of hands now extended both together, and now held out one at a time in rapid succession; and so forth, throughout all the long and complicated series. A quadrille, in short, is not a name for one act, for a single movement of a single muscle, but for many acts of the whole organism, all arranged in a fixed sequence.
It is just the same with the simplest act of mental perception. Orange, for example, is not the name of a single impression; it is the name of a vast complex of impressions, all or most of which are present to consciousness in the actuality whenever we see an orange, and a great many of which are present in the idea whenever we remember or think of an orange. It is the name of a rather soft, yellow fruit, round in shape, with a thick rind, white inside, and possessing a characteristic taste and odor; a fruit divisible into several angular, juicy segments, with cells inside, and with pips of a recognized size and shape—and so forth, ad infinitum. In the act of perceiving an orange we exercise a number of separate nerves of sight, smell, taste, and feeling, and their connected organs in the brain as well. In the act of thinking about or remembering an orange we exercise more faintly a considerable number of these nerves and central organs, though not, of course, all distinctly, or all together; otherwise, our mental picture of an orange would be as vivid and all-embracing as the sight of the actual orange itself.
Now, the name orange calls up more or less definitely the picture of several among these separate qualities. But it doesn't call them all up; indeed, the word in itself may not perhaps call up any of them. For instance, in the phrase, the Prince of Orange, where identical symbols meet the eye, I don't think of the fruit at all; I think, according to circumstances and context, either of William III of blessed memory, or of the eldest son of the present King of the Netherlands, whose memory (in Paris especially) is somewhat more doubtful. An Orangeman and an orange-woman are not, as one might innocently imagine, correlative terms. Even without this accidental ambiguity, derived from the name of the town of Orange on the Rhone, the word orange need not necessarily connote anything more than the color by itself; as when we say that Miss Terry's dress was a deep yellow or almost orange. Nay, when we actually mean the fruit in person, not the tree, flower, or color, the picture called up will be different according to the nature of the phrase in which the word occurs. For, if I am talking about ordering dessert, the picture in my mind is that of five yellow fruits, piled up pyramid-wise on a tall center-dish; whereas, if I am talking to a botanical friend, my impression is rather that of a cross-section through a succulent fruit (known technically as a hesperidium), and displaying a certain familiar arrangement of cells, dissepiments, placentas, and seeds. In short, the word orange, instead of being a single unity, localizable in a single ganglion, represents a vast complex, of which now these elements are uppermost in consciousness and now those, but which seems to demand for its full realization an immense co-operation of very diverse and numerous brain-organs.
Every thought, even the simplest, involves for its production the united or associated action of a vast mass of separate brain-cells and separate brain-fibers. One thought differs from another dynamically rather than statically. It differs as running differs from dancing—not because different muscles are employed, but because the same muscles are employed in a different manner.
Trains of thought are therefore like a quadrille. One set of exercises is followed by another, which it at once suggests or sets in motion.
Of course, I do not mean to deny that every cell and fiber in the brain has its own particular use and function, any more than I would deny that each particular muscle in the body is intended to pull a particular bone or to move a particular definite organ. But what I do mean is that each such separate function is really elementary or analytical: its object is to assist in forming a conception or idea, not to contain, as it were, a whole conception ready made. Chinese symbols stand each for an entire word, and it takes thousands of them to make up a language; alphabetical letters stand each, not for a word, but for an elementary sound or component of a word, and twenty-six of them do (very badly, it is true) for all the needs of our mother English. Just so, each cell or fiber in the brain does not stand for a particular word or a particular idea, but for some element of sensation or memory or feeling that goes to make up the special word or idea in question. Horse is made up of five letters, or of four phonetic sounds; it is made up also of a certain form and size and color and mode of motion; and when we speak of it all these elements are more or less vaguely present to our consciousness, coalescing into a sort of indefinite picture, and calling up one another more or less symbolically.
This theory at first sight seems to make the explanation of memory far more difficult and abstruse than formerly. For on the old hypothesis (never perhaps fully pushed to its extreme in realizable thought by any sensible person) it seemed easy enough to say that every act of perception and every fact learned was the establishment of a line of communication between two or more distinct cells or ganglia in the brain, and that the communication, once fairly established, persisted pretty constantly ever afterward. I am told "Shakespeare was born at Stratford-on-Avon"; and forthwith, cell Shakespeare (or Shakspere, or Shakspear, etc.) has a line run from it to cell birth and cell Stratford on-Avon (a pretty complex one indeed, this last), which line remains from that day forward permeable to any similar exercise of nervous energy. This method is undeniably simple, neat, and effective. But, setting aside the difficulty of realizing that any one tract of the brain can possibly hold our whole vast mental picture of Shakespeare or of Stratford-on-Avon (especially if we have ever read the one or visited the other), there is the grotesque difficulty of the innumerable lines and cross-connections of association. A central telephone station would be the merest child's-play to it. For even so simple a word and idea as gooseberry is capable of arousing an infinite number of ideas and emotions. It may lead us at once to the old garden in the home of our childhood, or to the gooseberry-fool we ate yesterday; it may suggest the notion of playing gooseberry, or the big gooseberry of the newspaper paragraph; it may lead to etymological dissertation on its derivation from gooseberry, allied to north country grosers and French groseille, or it may summon up visions of bad champagne, incidentally leading to "The Vicar of Wakefield," and the famous wine manufactured only by Mrs. Primrose. In fact, I have no hesitation at all in expressing my private opinion that, if the chart of the brain were at all like what most people imagine it to be, the associations of the word gooseberry alone would suffice to give good and solid employment to every fiber, cell, and convolution it anywhere possesses.
On the other hand, if we regard the brain as mainly dynamical, as an organism capable of very varied combinations of action, we can easily see, not only how memory becomes possible, but also how such infinite variations of association are rendered conceivable. For if every thought or perception is, as it were, an organized tremor in a vast group of diverse nerve-elements, often indeed in almost all together, it is simple enough to understand how these tremors may fall into regular rhythms, may excite one another in regular successions, may got habitual, just as the steps do in dancing, or the movements of the hand in writing a familiar and well-remembered formula—for example, in signing one's name. Here, in this immense and minutely organized workshop, we have a constant succession of motions in wheels and gearing, so arranged that each motion may be communicated in a thousand directions, and what is apparently a single impetus may call up the most diverse and extraordinary results. But, in reality, the impetus is not single: for, when we are thinking of horse in one way, we have a certain fixed form of movement called up; while, if we are thinking of it in another way, the form called up, though analogous in many respects, is far indeed from being identical. When I write "nice," you think of something or other vaguely pleasant; but, when I write "Nice," the very pronunciation is altered into something very like "niece," and the picture that rises before your mind is the very definite one of the Promenade des Anglais, with its long line of white villas and stunted palm-trees, bounded by the blue horizon of the Mediterranean and the beautiful slopes of. the coast toward Villefranche. It is just the same with the apples and the oranges. The elements of the picture vary incessantly; and while one combination now suggests one association, another combination another time suggests a second. The elements join together in an infinite variety of ways, and so a finite number of cells and fibers enable us to build up all the wealth of thought, just as twenty-six tiny symbols allow us to express all the wonderful conceptions of Milton and all the beautiful ideas of Shelley. There are only fifty-two cards in a pack, it is true, but no two games of whist ever yet played, in all probability, were absolutely identical.
To sum it all up: it is the brain as a whole that thinks, and feels, and desires, and imagines, just as it is the body as a whole that walks, and swims, and digs, and dances. To locate, say, the faculty of language in a particular convolution of a particular hemisphere is almost as absurd, it seems to me, as to locate, say, the faculty of writing in the last joint of the right forefinger. Convolution and forefinger may be absolutely essential or indispensable for the proper performance of speech or writing; but to say that is not to say that the function in question is there localized. The brain as a whole is the organ of mind, but there is no organ for the word Canonbury or for the proper perception of a Mrs. Pollock geranium.—Gentleman's Magazine.