It is now nearly fifty years since Spencer first published his celebrated essay on "The Origin and Function of Music." That essay has been elaborately assailed from many quarters; it has been objected to as insufficient from the standpoint of Æsthetic psychology, and as at variance with some of the known facts of musical history. Nevertheless Spencer, in accordance with his general intellectual habit, always clung tenaciously to his theory, and, without modifying it at all, returned to the subject in later years only in order to re-asseverate his doctrine and to repel the critical assaults that were made upon it. He had no difficulty in dealing with the counter-theory of Darwin—that music sprang from the amorous rivalry of the males in the presence of the females of certain species—for Darwin's brief excursion into the alien field of musical Æsthetic was as humorous and unprofitable as a discussion of bimetallism by Tchaikovski would have been. Then Spencer dealt with the redoubtable objections of the late Edmund Gurney and those of Dr. Wallaschek, undoubtedly scoring at times against them when they had needlessly overstated their own case, though not, it seems to me, removing the impression that they had successfully attacked the central point of his theory. Towards the very end of his days he returned yet again to the subject, in his Facts and Comments, and did me the honour to combat the brief criticism of his theory which I had put forward in my Study of Wagner, asserting that I exhibited a "confusion between the origin of a thing and the thing which originates from it," and that some of my criticisms "went far towards conceding" what I denied. I can only say that while I considered that Spencer passed over in silence certain of the stronger points I had urged against him, aiming at a merely dialectical victory here and there by interpreting my words in a different sense from that intended by me, I was still quite unconvinced, even by his later arguments, of the truth of his original theory. I shall try to show that that theory rests on a misunderstanding of the real nature of music, and on too ready an assumption of a causal connection between phenomena that are really only similar, and that it is helped out by unintentional misstatements as to some of the main factors of the problem. The question has an interest above and beyond Spencer's connection with it. The speech-theory of the origin of music has here and there been adopted as an established Æsthetic fact, and Æsthetic deductions have been made from it that must affect our views of current developments of the art. Wagner—working of course on lines of his own—contended that song is "just speech aroused to highest passion," and admiring commentators innumerable have followed him in his error. It is worth while therefore, as a contribution to a rather obscure point in musical Æsthetic, to try to demonstrate the falsity of the speech-theory, and at the same time to place over against it a theory of the origin and nature of music that squares better with the facts of history and psychology; and this is best done by examining the speech-theory in the hands of its strongest advocate.

Briefly, Spencer's theory is this: "Variations of voice are the physiological results of variations of feeling," since "all feelings ... have this common characteristic, that they are muscular stimuli." Thus according to the intensity and the quality of the feeling, the tones in which it is expressed will vary in loudness, in timbre, in pitch, in width of intervals, and in rapidity. "These vocal peculiarities, which indicate excited feeling, are those which especially distinguish song from ordinary speech." In other words, excited speech merges into recitative, and recitative in its turn merges into song; and song "originally diverged from emotional speech in a gradual, unobtrusive manner." Against this view I argued, in my Study of Wagner, that "it errs in supposing that, because song exhibits some of the characteristics of speech, the one has necessarily taken its rise from the other. The resemblances between the external characteristics of speech and those of song are only what might be expected, seeing that both are phenomena of sound, and sound can only vary in the ways indicated by Spencer.... The mere resemblance of song and speech in their most external characteristics is not a proof that one is the outcome of the other, but simply that they have certain causal phenomena in common; while the internal differences between them are greater than their resemblances." The careful reader will observe, in fact, that Spencer unconsciously sophisticates his argument from the very commencement. It is quite true that "variations of voice are the physiological results of variations of feeling"; it is also quite true that the "vocal peculiarities which indicate excited feeling"—such as loudness, high pitch, increased resonance, and so on—are more pronounced in song than in ordinary speech. But it does not at all follow from this that song took these peculiarities from speech—that speech got them first, then developed them into recitative, and then still further into song. To make a symmetrical but artificial chain of this kind is to beg the question at the outset. Spencer never put before himself the obvious alternative—"Could not, and would not, song have had all these peculiarities even if speech had never been invented? Given, that is, the capacity of men to feel emotion in varying degrees, would not a strong emotion naturally express itself in louder, more varied, more resonant tones than a weak emotion—and this even if man had as yet no language?" Spencer, in fact, simply details the characteristics of tone as the expression of feeling, and then illegitimately appropriates them, in the first place, to one order of tone, namely speech. No one would dream of disputing the physiological facts which he established in his essay with his usual patient and scrupulous accuracy. It is unquestionable that, on the whole, a loud tone in speaking and a loud tone in singing both indicate heightened feeling; and that in all the other respects enumerated by him, song and speech exhibit precisely the same characteristics. But this does not authorise us, in any way, to assert that song has "grown out of" speech. Spencer argued too hastily from a mere analogy to a cause. We are prepared to admit—to state the foregoing argument in another way—that in moments of emotional excitement the ordinary speech of men becomes more rhythmical, acquires a more pronounced timbre, and generally varies in the ways Spencer has enumerated. What we are not prepared to admit is that this is either a lower form of music or the stuff out of which music has grown. Our contention is that while the difference between speech and excited speech is one of degree only, the difference between speech and music is one not merely of degree, but of kind—we are dealing with similar physiological but widely separated psychological phenomena; and that this is true not only of modern music, as Spencer seems to admit, but of that primitive music out of which our complex modern art has grown.

Moreover, Spencer ignored the new light which modern physio-psychological research has thrown upon the question, some of which I referred to in the Study of Wagner. Stricker, in his Du Langage et de la Musique (1885), has, among a lot of statements and conclusions that need to be taken with caution, at all events made out a good case for believing that the organs of speech and the organs of song are controlled by different cerebral spheres. Wallaschek's conclusions, again, are too important to be passed over in silence by any advocate of the speech-theory. I venture to quote in full from my Wagner the passage in which I condensed Wallaschek's argument: "Further, it is now not only placed beyond dispute that the faculty of articulate speech has its distinct cerebral centre, but it has been localised in the third frontal convolution of the left hemisphere of the brain;" and Dr. Wallaschek, in a brilliant paper, has striven to show that there must be another centre that controls musical thought and speech. [39] Without going into Dr. Wallaschek's theory in detail, it may be sufficient here to note some of his facts and conclusions: (a) "the forming of concepts goes on in a different part of the brain, and the concepts travel along other channels, than the expression of the feelings and the merely automatic processes;" [40] (b) children with aphasia (i.e. destruction or disturbance of the faculty of articulate speech) are yet able to sing; [41] (c) patients with aphasia, who cannot speak connectedly upon ordinary occasions, can sometimes articulate the words when singing a song—the words being brought up into consciousness by association with the melody; [42] (d) the third left frontal convolution (which controls articulate speech) is very small in idiots and lower races, who yet are highly susceptible to music; [43] (e) the faculty of musical memory may be destroyed without disturbing the other mental faculties; [44] (f) consequently "we express ourselves and hear in quite a different manner when we sing and when we speak." [45] All this evidence Spencer ignored to the last.

Nor does it ever seem to have occurred to him to analyse the state of mind of a musician at the moment of composition, and to utilise the result thus obtained in order to throw light on the origin of music. Had he done this he would have seen the force—which his criticism of me showed he had not seen—of M. Combarieu's remark that "Mr. Spencer neglects or ignores everything that gives to the art he is studying its special and unique character; he does not appear to have realised what a musical composition is, what are the rules it obeys, what is the nature of the charm and the beauty we find in it. In short, we can bring against him a fundamental fact, in comparison with which everything else has only a quite secondary value: that is, the existence of a musical manner of thinking (une pensÉe musicale). The musician thinks with sounds, as the literary man thinks with words." [46]

Here, indeed, is the crux of the disagreement between Spencer and those who reject the speech-theory as an absolutely inadequate explanation of the origin of music. What was his criticism of this criticism? "Here," he says, "we have a striking example of the way in which an hypothesis is made to appear untenable by representing it as being something which it does not profess to be. I gave an account of the origin of music, and now I am blamed because my conception of the origin of music does not include a conception of music as fully developed. What is every process of evolution but the gradual assumption of traits which were not originally possessed?" It will be seen, I think, that Spencer quite missed the true point of M. Combarieu's objection. We do not expect that from a theory of the origin of music among primitive men one should be able to forecast all the later forms into which music has branched; but we do expect that, since evolution is a continuous process, the theory of the earlier music should not be at variance with all the main psychological features of the later music. We say to Spencer, "Take your theory, and we are unable to work it out in detail. You assert that the expression of musical thought and emotion has taken three successive forms—excited speech, recitative, and music. Well, we find it impossible to leap to this conclusion, as you have done, merely because there are certain resemblances, due to physiological causes, between speech and song. We cannot trace such a process historically—for your own sketch of the supposed historical process is demonstrably inaccurate in evidence and hasty in inference—nor can we even imagine the process psychologically. To us, there is a great psychological and Æsthetic gulf fixed between excited speech and song—not only between the speech and the song of to-day, but between the ruder speech and ruder song of primitive man. On the other hand, we have a theory that imposes no such strain, either historical or psychological, upon us. That theory is, that music arises from a peculiar set of stimuli and peculiar organs of expression of its own, with which speech not only has nothing whatever to do now, but never had anything to do, as fons et origo. Allowing for all the differences between our music and that of the savage who blows his reed and thumps his tam-tam, and for all the differences of general mental structure between him and us, we can still see that the same causes which incite us to music incited him. Now no one will for a moment contend that there is any but an infinitesimal resemblance between a Bach fugue or a Strauss symphonic poem and excited speech; neither can we perceive that there was ever any but the faintest resemblance between the causes that provoked the savage to excited speech and those that impelled him to his rude kind of music. But your theory, while it disregards the plain fact that no demonstration could deduce a Bach fugue from excited speech, and overlooks the mental elements in primitive man from which the Bach fugue could develop step by step, invites us to believe that music grew out of something with which we are unable to correlate it either now or in the most primitive times."

"But," it may be objected, "all this is pure assertion. You simply take music as it is written to-day, attribute this to something which you call a 'musical faculty,' or a 'musical manner of thinking,' and then, having invented this convenient faculty, blandly assume that it is from a similar faculty that the rude music of primitive man originated. What you have to do is to prove the existence of this musical faculty, this specifically musical way of conceiving and expressing things, that, on your assumption, is innate in the human mind, and needs no help from speech even in the earliest days of the race." Well, no one, I think, will question the existence in us, at the present day, of something that may well be called, in general terms, the musical faculty. For the musician as we now know him—and, indeed, have known him for some centuries—music is a means of emotional expression that can function without the aid of poetry or even of speech. It takes its rise from its own order of feelings; it has its own self-sufficient manner of expressing them; it tells its own story to the mind of the hearer; and neither the feeling, nor its manner of expression, nor its effect on the auditor, suggests any dependence on speech. The musician, in order to begin composition, has not to receive a preliminary stimulus either from poetry, or from any concept or sentiment that could for a moment be expressed in words. (He may, of course, set poetry to music; but on the other hand he may not; and it is the self-existing order of music we are now discussing.) The musician, under the influence of an inward stimulus of some very obscure kind, may take three or four tones—say those of the opening subject of a sonata or a fugue—and build with them a structure ordered and controlled by certain laws purely its own, having for its object the susciting in our minds of a series of feelings from which all thought of speech is absent. The musician joys in building tones together in this way; we in our turn joy both in the process of building and in the finished edifice itself.

"So far, so good," the opponent may say; "this is what music now is, as the result of a long course of evolution from its original germ. But will you assert that primitive man was impelled to his rude music by the workings of some similar faculty—that, without any reliance, even in the earliest days, upon speech, and without any intermediate stage of recitative, he produced music, bearing the same relation to his emotions as the music of Bach and Beethoven did to theirs?" Well, this is precisely what we do assert; nor do I see any difficulty in the way of the theory that primitive man came to utter himself in his rude music by the same psychological processes by which we utter ourselves in ours to-day. Speech had no more to do with the impulse to his music than it has with the impulse to ours. Spencer's theory would have it that first man spoke, then he advanced to excited speech, that this became more rhythmical and more definite and thus expanded into recitative, and that from this there emerged song "in a gradual, unobtrusive manner"—so gradual and so unobtrusive, I am afraid, that we can neither trace it emerging nor imagine it doing so. Is it not more reasonable to believe that music first came into the world when the savage took delight in any tones—those of the human voice, of a reed, or of a drum—purely as tone, and began to take a further simple delight in the relations between tones? Need we concern ourselves at all with speech, excited or torpid? Can we not begin with mere feeling venting itself in mere sound—as we know it must have done at first—and draw a line from this straight through all the music of all the world? Why should we assume that for man to express his feelings in tone he must first have invented speech, and then have developed the emotional side of this until it was able to cut itself loose and commence life on its own account, by some process that is really unimaginable? We know that feeling vents itself in sound, and that waves of feeling vent themselves in waves of sound, as may be observed in the vague crooning of an infant over its toys, or the moaning of a man in pain. This is one fundamental fact in the origin of music. Another is the indisputable fact that men, whether civilised or savage—that many animals, indeed—are susceptible to tone purely as tone; and a further fact is that the primitive organism takes pleasure in the relations between tones, as may be seen in the boy who keeps on thumping two tin cans that happen to give out different sounds. There is surely no need to insist upon the point that both tones and the relations between tones in themselves interest and charm, in a minor degree, the savage as they do us. It is from this phenomenon, I should imagine, not from excited speech, that music took its rise; and the evidence from the music of primitive tribes, upon which Spencer drew in support of his theory, does nothing to invalidate mine. In his original essay he quoted, from his own Descriptive Sociology, a number of passages relating to the song-customs of various undeveloped races. I cannot, among all these quotations, see one that suggests that the music of these people was simply a hyper-excited form of speech. On the contrary, it is clear from his own citations that their delight was in music purely as music; that their feelings spontaneously flowed, as ours do, into a system of tones and relations between tones that existed in and by and for itself, with only the same kind of dependence upon the words that is exhibited in a song by Brahms or a chorus by Handel. [47] No doubt the general course of the words controls the general course of music in some degree, as it does in our own song-writing; but there is nothing whatever to contradict the view that savage music, even as our own, springs spontaneously from a non-verbal emotion, and seeks an expression either absolutely independent of speech or only remotely influenced by it. The East African, says Spencer, "in singing, contents himself with improvising a few words without sense or rhythm, and repeats them till they nauseate." If this does not betoken a state of mind fundamentally analogous to that of the absolute musician, it is hard to say what the words mean. Plainly, what sets the East African singing, what determines that one note shall follow another, what makes him so indifferent to the sense or nonsense of the words, is simply the delight in tone as tone, in the relations of tones as relations of tones, simply the need for what he feels to vent itself in precisely that way and no other—in a word, the primitive pensÉe musicale, the primitive "musical manner of thinking." [48] Abundant evidence can be had to corroborate this, and I quoted some of it in my Wagner. "Speaking of the Iroquois, Dr. Morgan says that their war-songs are in a dead language, or, at all events, they are unable to interpret them.... Mr. Baker, too, observed the meaninglessness of the Indian song." [49] There is not much trace here of excited speech first becoming recitative and then musical song. [50]

Dr. Wallaschek's conclusions, again, as to the music of savages are as follows:—

(1) "In primitive times vocal music is not at all a union of poetry and music. We find, on the contrary, vocal music among tribes which, owing to the insufficient development of language, cannot possibly have any kind of poetry. Thus the position of vocal music is quite independent of any other art. (2) It is impossible that in these cases music arose as a direct imitation of the natural accents ready made in speech. (3) Because these texts are neither themselves a language, nor could the melody alone have been taken from a developed language, for in such case the words would have been borrowed together with the music. Entirely meaningless words simply serve to facilitate the vocalisation." Further, "another striking feature of these savage songs is the liberty with which the composer treats the grammatical structure of the sentence and the logical order of words. Thus in many of the Andamanese songs the words in their poetic form are so mutilated to suit the metre as to be hardly recognisable.... If negroes sing they keep strict time, and do not allow themselves to be hindered by any obstacle in the use of the words." Other evidence of the same kind might be adduced, from which it is quite clear that we are face to face with a phenomenon on which Spencer's theory throws no light at all. There seems to be no doubt that there is in the savage, though of course in a relatively undeveloped form, the same musical sense as in ourselves, something that has always flown directly, for its expression, to a mode of utterance of its own, compounded of tones, relations of tones, and rhythm, which is the natural language of this sense, and which never needed to pass through the intermediate stage of imitation or exaggeration of the accents of speech. [51]

Look for a moment at the two theories and their implications side by side. We know that primitive man, like the animal, [52] is susceptible to tone, sequences of tone, colour of tone, and rhythm; and that, from purely physiological causes, a number of his feelings tend to express themselves in vocal sounds. Now these are all the elements we require in order to construct modern music. The composer feels strongly, and is impelled to find an outlet for his emotions in tone. According to the line of his emotion, so to speak, is the line of his music—the pure feeling takes hold of the sounds through which alone it can utter itself, and shapes them, in form, in colour, in sequence, in intensity, after its own image. We have in primitive man, in a rude and undeveloped stage, all these elements out of which the modern music-maker builds his gorgeous palaces. According to the intensity of the emotion of the savage will be the width of the intervals of his voice, the resonance, the colour of it; according to the shade of his feeling will be the shade of his rude melody; and from the ensemble of the qualities of the sounds in which he is uttering himself will his hearers be able to guess what mood it was that animated his song. Here, then, are all the elements out of which music could grow, even if man had never learned to speak three connected words. Yet we are asked by Spencer to believe that these elements, sufficient in themselves to give birth to music, remained dormant in the human breast for untold centuries, until man had evolved a fairly elaborate system of speech—for it must be remembered that Spencer's theory presupposes not the rude and merely utilitarian speech of the man only one remove from the beast, but a comparatively highly organised language, capable of expressing connectedly a savage's thoughts about something more than his daily physical wants. Some such abstract, Æsthetic, reflective form of speech we are compelled to postulate if we are to grant the probability of music arising, as Spencer says it did, from the excited speech of man. Then, when man has slowly and painfully learned to speak, and had plenty of practice in speaking excitedly, we are invited to believe that by some mysterious process music arose, the expression of feeling in organised tone, the delight in tone qu tone, in sequences and relations qu sequences and relations. And all this time the elements out of which this organised system of sound could grow, which were innate in man from the very first, by reason of the fact that he had nerves, muscles, and vocal organs, have been doing absolutely nothing! Though they required only the stimulus of feeling to call them into being, and though they were receiving this stimulus day by day, hour by hour, they had to deny themselves for centuries upon centuries, until they could receive precisely the same kind of stimulus after man had learned to speak! Is this credible?

II

If Spencer's theory is Æsthetically and psychologically inconceivable, he is hardly happier in the pseudo-historical evidence by which he seeks to support it. His notion seems to be that all ancient music, and the Oriental and savage music of the present day, represent the art at the second or recitative stage of development—a kind of half-way house between excited speech and full-blown song. Thus the Chinese and Hindoos "seem never to have advanced" beyond recitative. "The dance-chants of savage tribes are very monotonous, and in virtue of their monotony are more nearly allied to ordinary speech than are the songs [53] of civilised races"—which is surely a quite illegitimate comparison. Again, "hence it follows that the primitive (Greek) recitative was simpler than our modern recitative, and, as such, much less remote from common speech than our own singing is." These typical quotations will serve to show how blandly Spencer assumes the very thing he has to prove. The dance-chants of savages are not as highly organised as our European songs; but does this indicate that there is not the same psychological difference between the song and the speech of the savage as there is between the song and the speech of the European? The ancient Greek music was not so complex as ours; but will Spencer be bold enough to say that a man of Athens, listening to contemporary music, did not feel under it precisely the same kind of Æsthetic pleasure as we feel when we listen to a song by Brahms or a symphony by Beethoven—a kind of pleasure different in essence and in temperature from any that can be given by speech? Did the Greek, that is, listening to Greek music, feel as I do when I listen to an eloquent preacher or an intoning Quaker, or as I do when I listen to music in the real sense of the term? Surely there can be no doubt in the matter. Setting aside the difference due to the enormous development of our art on the formal and technical side, there can be no question that the Greek took pleasure in his music qu music, not qu "recitative." [54] And as with the Greeks, so with Orientals and savages. How Spencer can imagine that Oriental music as a whole, and particularly that of China and India, has for the most part remained stationary at recitative, is a mystery to me, in face of the mass of evidence that may be had from any history of music or any collection of travels. There is, indeed, in much Oriental music, that dubiety of scale (according to our notions) which has misled unwary travellers into the belief that the native singing cannot be real music, because it is so different from ours. But nothing can be better established than the fact that melodies pure and simple, tunes written and sung merely to express that pensÉe musicale to which I have already referred, are common in the music of all Oriental nations. Spencer's statement "that the music of Eastern races is not only without harmony, but has more the character of recitative than of melody," and that "the chant of the early Greek poet was a recitative with accompaniment in unison on his four-stringed lyre," is a fair sample of the uncritical way in which he has assumed anything that would be likely to bear out his theory. His confusion of two or three distinct things by dubbing them all "recitative" is one of the main sources of his errors on this question. As for his attempt to limit harmonic music to modern Europe, I will only say, with Naumann, that wherever we have, as in the old Egyptian paintings, a representation of a concert with many instruments of various shapes and sizes, it is incredible that the performers should all have been playing the same notes. The result, of course, could not have been harmony in our acceptation of the word, for this is to a large extent dependent upon theory for its development; but it was conceivably one of the roots from which harmony could grow. And as Spencer admitted that his theory contained no explanation of harmony, that theory is obviously weakened by any fact indicating that the desire for harmony is innate in the human breast, like the love of tones, sequences of tones, and relations between tones. We must dismiss from our minds all the misleading connotations of the term "harmony," as we must with the term "recitative"; and when we do this there is ample evidence to show that the harmonic sense—the joy in hearing two tones sounded together—is as innate, and as independent of the stimulus of speech, as the melodic sense. The mere sweeping of the harp-strings during singing is not what we would call harmony; but if it does not point to a rudimentary feeling that tones in combination are more pleasurable than single tones, it is difficult to say what it does indicate. Everywhere, in truth, we come down to the really fundamental fact, that there is even in primitive man a real musical sense, independent of speech in origin, and, as far as we can see, much earlier than speech in the order of time, for man certainly expressed his feelings in pure indefinite sound long before he had learned to agree with his fellows to attach certain meanings to certain stereotyped sounds.

III

The music of savage tribes is, however, the last stronghold of Spencer; and if his theory fails to find proper support in that quarter, it can hardly resist all the weight of evidence that may be brought against it from others. Here, he says, he has Sir Hubert Parry on his side, "who adopts the view I have here re-explained and defended," and who "has in his chapter on Folk-Music exemplified the early stages of musical evolution, up from the howling chants of savages—Australians, Caribs, Polynesian cannibals, etc.—to the rude melodies of our own ancestors. I do not see how any unbiassed reader, after examining the evidence placed by him in its natural order, can refuse assent to the conclusion drawn." Well, the final refutation of Spencer can be had out of the mouth of Sir Hubert Parry himself. What Sir Hubert's own theory of the origin of music may be I do not know; but certainly neither the facts nor the arguments he has adduced in his Art of Music give any colour to the theory that music first arose as a modification of the attributes of emotional speech. Let us examine Sir Hubert Parry's evidence.

We begin at the beginning with the descending chromatic howl of the Carib which he quotes on page 49 of his book—the "howling chant" to which Spencer refers; and if, as the philosopher will have it, this represents "the early stages of musical evolution," his case has gone by the board at once. There could be no more conclusive testimony to the fact that music has its origin not in speech, but in the venting of mere vague emotion in mere vague sound; for where Spencer sees the previous influence of speech in this howl of the Carib I cannot imagine. He might as well suppose that speech antedates the howl of a dog or the roar of a lion. On what grounds does he find support for his theory here? Simply that a howl of this kind, like the song of the Omaha Indians, is distinguished by indefiniteness of intervals! "Now this," he says, "is just one of the traits to be expected if vocal music is developed out of emotional speech; since the intervals of speech, also, are indefinite." Was there ever a more palpable non sequitur? Because A has one of the characteristics of B, therefore A must have grown out of B! Here is a complete justification of my previous remark that Spencer has converted a mere likeness into a cause. The real reason for music exhibiting some of the traits of speech is that, music and speech being the expression of allied orders of feeling, and both finding voice through the same muscular apparatus, they simply cannot help having a great many features in common. But we really require something more than a demonstration that the intonations of music, being affected by the same physical organs, point to very much the same mental and physical phenomena as the intonations of speech, in order to convince us that music had its origin in speech.

Take now the further examples given by Sir Hubert Parry, and discover in them, if you can, any evidence that does not go to show that they are born directly of a primitive pensÉe musicale, without any sign of the previous intervention of speech. Written over them all, indeed, is conclusive proof that when primitive man sings, or even croons, to himself, he is unconsciously guided by a rudimentary musical sense. Savages contrive, says Sir Hubert, "little fragmentary figures of two or three notes, which they reiterate incessantly over and over again. Sometimes a single figure suffices. When they are clever enough to devise two they alternate them, but [naturally] without much sense or orderliness"; and he shows, later on, how even among savages there is a continuous growth in this primitive sense of design. Now all this is in accordance with the theory of the origin of music already advanced in this essay; and these phenomena of savage music will easily account for all the most complex modern developments of the art, which Spencer half admits his theory will not account for. Savage man, merely because he is a physical organism, expresses himself in sound. Again, merely because he is a physical and psychical organism, he takes pleasure in sounds, in successions of sounds, and in the co-relations of sounds; and, to complete the list of the elements necessary to constitute all the music that has ever been written in the world, Sir Hubert Parry shows that, even in the savage whose rude attempt at song is little more than a howl, there is a rudimentary sense of form, of balance, of design. "When little fragments of melody [55] become stereotyped," says Sir Hubert, "as they do in every savage community sufficiently advanced to perceive and remember, attempts are made to alternate and contrast them in some way; and the excitement of sympathy with an expressive cry is merged in a crudely artistic pleasure derived from the contemplation of something of the nature of a pattern." Is there any support for the speech-theory here? Is it not, indeed, an interloper pure and simple, obscuring a trail that is perfectly clear and open if left alone?

The one fact upon which Spencer always seems to rely is that the intervals of speech and the intervals of the most primitive chant are both indefinite. Even here, however, Sir Hubert Parry's book is unpropitious to him, for Sir Hubert insists on the obvious fact that indefiniteness of intervals in early music is entirely a matter of lack of instruments by which to fix the various notes of a scale. "It is extremely difficult to make sure what intervals savages intend to utter, as they are very uncertain about hitting anything like exact notes till they have advanced enough to have instruments with regular relations of notes more or less indicated upon them." To pass from an indefinite howl to a definite series of notes, when an instrument has been invented that guides the voice and fixes its tones, may be the work of a day. Wherein then comes the function of speech and recitative, which are supposed to occupy the intermediate stages of evolution between the howl and the song—for I suppose Spencer would hardly contend that man learned to speak before he learned to howl? And at what stage appears this elementary feeling for musical design which the savage exhibits? Can this be conceived to grow out of the habit of speech? If not, if it is independent of speech, if it is something that concerns itself with pure sound alone, what was it doing in all the ages when man was making sounds, but had not yet made himself a language? "The crudest efforts of savages," says Sir Hubert Parry, "throw light upon the true nature of musical design, and upon the manner in which human beings endeavoured to grapple with it." Again, "the savage state indicates a taste for design, but an incapacity for making the designs consistent and logical; in the lowest intelligent stage, the capacity for disposing short contrasting figures in an orderly and intelligent way is shown." Once more, can speech be logically conceived as playing the leading part in this long but continuous drama of evolution?

Finally, in Sir Hubert Parry's own pages Spencer could have found evidence of yet another element of pure musical enjoyment in the savage mind—none other than an incipient desire for harmony. Speaking of the rise of harmony in the Middle Ages, and of the curious device of making two wholly different tunes go together by the process of "easing off the corners and adapting the points where the cacophony was too intolerable to be endured," Sir Hubert shows the existence of this same practice among savages. "This," he says, "may seem a very surprising and even laughable way of obtaining an artistic effect, but in reality the actual practice of combining several tunes together is by no means uncommon. Several savage and semi-civilised races adopt the practice, as, for instance, the Bushmen at the lower end of the human scale, and the Javese, Siamese, Burmese, and Moors about the middle. In these cases the process usually consists of simultaneously singing or playing short and simple musical figures, such as savages habitually reiterate, with the addition in some cases of a long sort of indefinite wailing tune which goes on independently of all the rest of the performance. The Javese carry such devices to extremes, producing a kind of reckless, incoherent, instrumental counterpoint, very much like a number of people playing various tunes at once, with just sufficient feeling for some definite central principle to accommodate the jarring elements. The practice of combining tunes seems to have become universal quite suddenly, and it led very quickly to fresh developments. And it is worth noting that one of these developments was precisely the same in principle as that adopted by the Bushmen and the Javese, and other semi-savage experimenters in such things; which was to accompany the main combination of two melodies by a short musical figure which could be incessantly reiterated as an accompaniment." Phenomena like these undermine the crude and hasty inference that Orientals and savages have no notion of harmony; they prove that, as low down in the human scale as our investigations will carry us, man tries to make harmony because it pleases his musical sense. How far he succeeds depends upon other things than his mere desire.

So that, to sum up, we can dismiss speech altogether from our hypothesis of the origin of music, seeing that, while no man can represent to us either the psychological or historical processes by which music has grown or could grow out of speech, we find innate in the human organism every element out of which music can grow, independently of speech—the delight in tone, the delight in successions of tones, the delight in combinations of tones, the delight in rhythm, the delight in design. Even Spencer himself, in the chapter on "Developed Music" in his Facts and Comments, sees that these elements are sufficient to account for certain kinds of music, though his total analysis, particularly in the distinction between merely symmetrical music and poetical music, is Æsthetically incomplete and À priori. To Spencer's oft-reiterated question, "If my theory does not explain the origin of music, how else can its genesis be explained?" we may reply that his theory really explains nothing; it only asserts. It points to certain resemblances between speech and song, and then dogmatically lays it down, without an atom of proof, that the one has arisen from the other. Per contra, an analysis of primitive music shows us that in the rudest savage we have, in embryo, every element that goes to make the most complicated music of modern times—some of these elements, indeed, appearing even in animals. If we are to believe that these in themselves could not develop into music, we must have a reason why; and if we are to believe that an imitation of the accents of speech was necessary before primitive man could express what he felt in mere indefinite sound, we must have not only some proof that it ever occurred, but some demonstration of how the process is possible; for to me, at least, it is psychologically inconceivable. When Spencer says that "song emerged from speech," he is, I contend, merely using a verbal formula that conveys nothing representable to us; it is of the family of those "pseudo-ideas" upon which he himself has emptied the vials of his scorn in First Principles.