This hypothesis, now more than a century old, would, if it rested on good evidence, explain how a single version of the various lays came to be accepted and received as authorised. The Greek world, by the theory, had only in various places various sets of incoherent chants orally current on the Wrath of The public was everywhere a public of listeners, who heard the lays sung on rare occasions at feasts and fairs, or whenever a strolling rhapsodist took up his pitch, for a day or two, at a street corner. There was, by the theory, no reading public for the Homeric poetry. But, by the time of Pisistratus, a reading public was coming into existence. The tyrant had the poems collected, edited, arranged into a continuous narrative, primarily for the purpose of regulating the recitals at the Panathenaic festival. When once they were written, copies were made, and the rest of Hellas adopted these for their public purposes.

On a small scale we have a case analogous. The old songs of Scotland existed, with the airs, partly in human memory, partly in scattered broadsheets. The airs were good, but the words were often silly, more often they were Fescennine – "more dirt than wit." Burns rewrote the words, which were published in handsome volumes, with the old airs, or with these airs altered, and his became the authorised versions, while the ancient anonymous chants were almost entirely forgotten.

The parallel is fairly close, but there are points of difference. Burns was a great lyric poet, whereas we hear of no great epic poet in the age of Pisistratus. The old words which Burns's songs superseded were wretched doggerel; not such were the ancient Greek heroic lays. The old Scottish songs had no sacred historic character; they did not contain the history of the various towns and districts of Scotland. The heroic lays of Greece were believed, on the other hand, to be a kind of Domesday book of ancient principalities, and cities, and worshipped heroes. Thus it was much easier for a great poet like Burns to supersede with his songs a mass of unconsidered "sculdudery" old lays, in which no man or set of men had any interest, than for a mere editor, in the age of Pisistratus, to supersede a set of lays cherished, in one shape or another, by every State in Greece. This holds good, even if, prior to Pisistratus, there existed in Greece no written texts of Homer, and no reading public, a point which we shall show reasons for declining to concede.

The theory of the edition of Pisistratus, if it rested on valid evidence, would explain "how a single version of the poems came to be accepted," namely, because the poem was now written for the first time, and oral versions fell out of memory. But it would not, of course, explain how, before Pisistratus, during four or five centuries of change, the new poets and reciters, throughout the Greek world, each adding such fresh verses as he pleased, and often introducing such modern details of life as he pleased, kept up the harmony of the Homeric picture of life, and character, and law, as far as it confessedly exists.

To take a single instance: the poems never allude to the personal armorial bearings of the heroes. They are unknown to or unnamed by Homer, but are very familiar on the shields in seventh century and sixth century vases, and AEschylus introduces them with great poetic effect in {blank space}. How did late continuators, familiar with the serpents, lions, bulls' heads, crabs, doves, and so forth, on the contemporary shields, keep such picturesque and attractive details out of their new rhapsodies? In mediaeval France, we shall show, the epics (eleventh to thirteenth centuries) deal with Charlemagne and his peers of the eighth century A.D. But they provide these heroes with the armorial bearings which came in during the eleventh to twelfth century A.D. The late Homeric rhapsodists avoided such tempting anachronisms.

Wolf's theory, then, explains "how a single version came to be accepted." It was the first WRITTEN version; the others died out, like the old Scots orally repeated songs, when Burns published new words to the airs. But Wolf's theory does not explain the harmony of the picture of life, the absence of post-Homeric ideas and ways of living, in the first written version, which, practically, is our own version.

In 1892 (COMPANION TO THE Iliad) Mr. Leaf adopted a different theory, the hypothesis of a Homeric "school" "which busied itself with the tradition of the Homeric poetry," for there must have been some central authority to preserve the text intact when it could not be preserved in writing. Were there no such body to maintain a fixed standard, the poems must have ended by varying indefinitely, according to the caprice of their various reciters. This is perfectly obvious.

Such a school could keep an eye on anachronisms and excise them; in fact, the Maori priests, in an infinitely more barbarous state of society, had such schools for the preservation of their ancient hymns in purity. The older priests "insisted on a critical and verbatim rehearsal of all the ancient lore." Proceedings were sanctioned by human sacrifices and many mystic rites. We are not told that new poems were produced and criticised; it does not appear that this was the case. Pupils attended from three to five years, and then qualified as priests or tohunga {Footnote: White, THE Ancient HISTORY OF THE Maori, VOL. i. pp. 8-13.}. Suppose that the Asiatic Greeks, like the Maoris and Zuñis, had Poetic Colleges of a sacred kind, admitting new poets, and keeping them up to the antique standard in all respects. If this were so, the relative rarity of "anachronisms" and of modernisms in language in the Homeric poems is explained. But Mr. Leaf has now entirely and with a light heart abandoned his theory of a school, which is unsupported by evidence, he says.'

"The great problem," he writes, "for those who maintain the gradual growth of the poems by a process of crystallisation has been to understand how a single version came to be accepted, where many rival versions must, from the necessity of the case, have once existed side by side. The assumption of a school or guild of singers has been made," and Mr. Leaf, in 1892, made the assumption himself: "as some such hypothesis we are bound to make in order to explain the possibility of any theory" (1892). {Footnote: COMPANION TO THE Iliad, pp. 20, 21.}

But now (1900) he says, after mentioning "the assumption of a school or guild of singers," that "the rare mention of {Greek: Homeridai} in Chios gives no support to this hypothesis, which lacks any other confirmation." {Footnote: Iliad, vol. i. xviii. p. xix.} He therefore now adopts the Wolfian hypothesis that "an official copy of Homer was made in Athens at the time of Solon or Pisistratus," from the rhapsodies existing in the memory of reciters. {Footnote: Iliad, vol. i. p. xix.} But Mr. Leaf had previously said {Footnote: COMPANION TO THE Iliad, p. 190.} that "the legend which connects his" (Pisistratus's) "name with the Homeric poems is itself probably only conjectural, and of late date." Now the evidence for Pisistratus which, in 1892, he thought "conjectural and of late date," seems to him a sufficient basis for an hypothesis of a Pisistratean editor of the Iliad, while the evidence for an Homeric school which appeared to him good enough for an hypothesis in 1892 is rejected as worthless, though, in each case, the evidence itself remains just what it used to be.

This is not very satisfactory, and the Pisistratean hypothesis is much less useful to a theorist than the former hypothesis of an Homeric school, for the Pisistratean hypothesis cannot explain the harmony of the characters and the details in the Iliad, nor the absence of such glaring anachronisms as the Cyclic poets made, nor the general "pre-Odyssean" character of the language and grammar. By the Pisistratean hypothesis there was not, what Mr. Leaf in 1892 justly deemed essential, a school "to maintain a fixed standard," throughout the changes of four centuries, and against the caprice of many generations of fresh reciters and irresponsible poets. The hypothesis of a school was really that which, of the two, best explained the facts, and there is no more valid evidence for the first making and writing out of our Iliad under Pisistratus than for the existence of a Homeric school.

The evidence for the Iliad edited for Pisistratus is examined in a Note at the close of this chapter. Meanwhile Mr. Leaf now revives Wolf's old theory to account for the fact that somehow "a single version" (of the Homeric poems) "came to be accepted." His present theory, if admitted, does account for the acceptation of a single version of the poems, the first standard written version, but fails to explain how "the caprice of the different reciters" (as he says) did not wander into every variety of anachronism in detail and in diction, thus producing a chaos which no editor of about 540 A.D. could force into its present uniformity.

Such an editor is now postulated by Mr. Leaf. If his editor's edition, as being written, was accepted by Greece, then we "understand how a single version came to be accepted." But we do not understand how the editor could possibly introduce a harmony which could only have characterised his materials, as Mr. Leaf has justly remarked, if there was an Homeric school "to maintain a fixed standard." But now such harmony in the picture of life as exists in the poems is left without any explanation. We have now, by the theory, a crowd of rhapsodists, many generations of uncontrolled wandering men, who, for several centuries,

"Rave, recite, and madden through the land,"

with no written texts, and with no "fixed body to maintain a standard." Such men would certainly not adhere strictly to a stereotyped early tradition: that we cannot expect from them.

Again, no editor of about 540 B.C. could possibly bring harmony of manners, customs, and diction into such of their recitals as he took down in writing.

Let us think out the supposed editor's situation. During three centuries nine generations of strollers have worked their will on one ancient short poem, The Wrath of Achilles. This is, in itself, an unexampled fact. Poets turn to new topics; they do not, as a rule, for centuries embroider one single situation out of the myriads which heroic legend affords. Strolling reciters are the least careful of men, each would recite in the language and grammar of his day, and introduce the newly evolved words and idioms, the new and fashionable manners, costume, and weapons of his time. When war chariots became obsolete, he would bring in cavalry; when there was no Over-Lord, he would not trouble himself to maintain correctly the character and situation of Agamemnon. He would speak of coined money, in cases of buying and selling; his European geography would often be wrong; he would not ignore the Ionian cities of Asia; most weapons would be of iron, not bronze, in his lays. Ionian religious ideas could not possibly be excluded, nor changes in customary law, civil and criminal. Yet, we think, none of these things occurs in Homer.

The editor of the theory had to correct all these anachronisms and discrepancies. What a task in an uncritical age! The editor's materials would be the lays known to such strollers as happened to be gathered, in Athens, perhaps at the Panathenaic festival. The répertoire of each stroller would vary indefinitely from those of all the others. One man knew this chant, as modified or made by himself; other men knew others, equally unsatisfactory.

The editor must first have written down from recitation all the passages that he could collect. Then he was obliged to construct a narrative sequence containing a plot, which he fashioned by a process of selection and rejection; and then he had to combine passages, alter them, add as much as he thought fit, remove anachronisms, remove discrepancies, accidentally bring in fresh discrepancies (as always happens), weave transitional passages, look with an antiquarian eye after the too manifest modernisms in language and manners, and so produce the {blank space}. That, in the sixth century B.C., any man undertook such a task, and succeeded so well as to impose on Aristotle and all the later Greek critics, appears to be a theory that could only occur to a modern man of letters, who is thinking of the literary conditions of his own time. The editor was doing, and doing infinitely better, what Lönnrot, in the nineteenth century, tried in vain to achieve for the Finnish Kalewala. {Footnote: See Comparetti, The Kalewala.}

Centuries later than Pisistratus, in a critical age, Apollonius Rhodius set about writing an epic of the Homeric times. We know how entirely he failed, on all hands, to restore the manner of Homer. The editor of 540 B.C. was a more scientific man. Can any one who sets before himself the nature of the editor's task believe in him and it? To the master-less floating jellyfish of old poems and new, Mr. Leaf supposes that "but small and unimportant additions were made after the end of the eighth century or thereabouts," especially as "the creative and imaginative forces of the Ionian race turned to other forms of expression," to lyrics and to philosophic poems. But the able Pisistratean editor, after all, we find, introduced quantities of new matter into the poems – in the middle of the sixth century; that kind of industry, then, did not cease towards the end of the eighth century, as we have been told. On the other hand, as we shall learn, the editor contributed to the Iliad, among other things, Nestor's descriptions of his youthful adventures, for the purpose of flattering Nestor's descendant, the tyrant Pisistratus of Athens.

One hypothesis, the theory of an Homeric school – which would answer our question, "How was the harmony of the picture of life in remote ages preserved in poems composed in several succeeding ages, and in totally altered conditions of life?" – Mr. Leaf, as we know, rejects. We might suggest, again, that there were written texts handed down from an early period, and preserved in new copies from generation to generation. Mr. Leaf states his doubt that there were any such texts. "The poems were all this time handed down orally only by tradition among the singers (sic), who used to wander over Greece reciting them at popular festivals. Writing was indeed known through the whole period of epic development" (some four centuries at least), "but it is in the highest degree unlikely that it was ever employed to form a standard text of the Epic or ANY part of it. There can hardly have been any standard text; at best there was a continuous tradition of those parts of the poems which were especially popular, and the knowledge of which was a valuable asset to the professional reciter."

Now we would not contend for the existence of any {blank space} text much before 600 B.C., and I understand Mr. Leaf not to deny, now, that there may have been texts of the ODYSSEY and Iliad before, say, 600-540 B.C. If cities and reciters had any ancient texts, then texts existed, though not "standard" texts: and by this means the harmony of thought, character, and detail in the poems might be preserved. We do not think that it is "in the highest degree unlikely" that there were no texts. Is this one of the many points on which every savant must rely on his own sense of what is "likely"? To this essential point, the almost certain existence of written texts, we return in our conclusion.

What we have to account for is not only the relative lack of anachronisms in poems supposed to have been made through a period of at least four hundred years, but also the harmony of the CHARACTERS