It is against nature that he made the most excellent creation that could ever be; for things are normally born imperfect, then grow and gather strength as they do so. He took poetry and several other sciences in their infancy and brought them to perfect, accomplished maturity. [Thus] one may call him the first and last of poets, in accordance with that fine tribute left to us by antiquity: that, having had no predecessor to imitate, he had no successor capable of imitating him.

That's my old pal Michel talking about Homer.1 He is almost completely wrong.

The Homeric Question

What he gets right is progress in the arts: early Greek sculpture was copied from the Egyptians and was so amateurish that it seems abstract, but the lack of detail simply reflects a lack of ability. The art was "born imperfect" and slowly gathered strength, eventually reaching an apex four or six centuries later with pieces like the Laocoön.

Homer appears to completely break any such notions of artistic development: the Iliad and the Odyssey emerge at the end of the 8th century with no visible tradition behind them, and until the 19th century this miraculous event was accepted at face value, elevating Homer from merely a great poet to a superhuman figure.2

But the Age of Reason rolled around and not even Homer remained unruffled. In 1795 Friedrich August Wolf published Prolegomena ad Homerum which established the basis for the Homeric Question debates over the next century (though, as the 11th edition of the Encyclopædia Britannica reminds us, his argument was so overwhelming it would not be challenged for 30 years3). Wolf was skeptical, likening the poems to an enormous ship constructed far inland, with no access to tools or water.

I find it impossible to accept the belief to which we have become accustomed: that these two works of a single genius burst forth suddenly from the darkness in all their brilliance, just as they are, with both the splendor of their parts and the many great virtues of the connected whole.

Wolf approached the problem by historicizing it, tracing the text and its historical context from Homer down to the first century.4 Like Sherlock Holmes, he systematically eliminated the impossible until only the truth remained:

  • There was no plausible occasion at which a poem of 15,000 lines would be recited.
  • There is no internal evidence in the Iliad for books or writing.5
  • If writing existed at all in Homer's time, it was primitive and utilized a limited alphabet.
  • The Greeks at that time had no access to papyrus or parchment, materials necessary to preserve a poem of this length: it must have been transmitted orally (and therefore lossily).
  • The earliest long-form writing came much later than Homer and was utilitarian/public in nature, for example Solon's laws written on wood for public display.
  • It took yet more time for writing to move from the public sphere to the artistic one.
  • The poem is full of "obvious and imperfectly fitted joints" which suggest alterations from a "later period".
  • The Alexandrian editors of the Hellenistic age felt free to edit the Homeric poems: "often removing many verses, and elsewhere adding polish where there was none".

Taken together, these arguments imply that 1) Homer did not compose anything like the Iliad as we know it, and 2) even if he somehow had done it, it was hopelessly corrupted in the process of transmission.6

Instead Wolf proposed that Homer was responsible for some small parts which were then combined and expanded by Peisistratus with the assistance of his poet friends in the middle of the 6th century, and finally polished by Aristarchus or Aristophanes 400 years later.7

Following Wolf, a cottage industry of Homeric analysis cropped up. Any line that did not suit their tastes was declared an interpolation, and repetitions were taken as proof of copying by later poets. They all agreed that the Iliad was the work of multiple poets, and that philologists could "scientifically prove"8 it. Yet no two Analysts could ever agree on which parts were Homer's and which were later additions.

Bethe thought it was the work of two poets. Theiler argued for 5 or 6. Hermann proposed that the Iliad consisted of a bunch of crap surrounding the pristine core of the original Homer, while Lachmann argued that it actually consisted of 18 different folk stories strung together, like the Finnish Kalevala. Ulrich von Wilamowitz-Moellendorff (I swear I'm not making up these names) claimed to detect three different "layers" in the Odyssey, one of which (the "old Odyssey"—books 5-14 and 17-19) had in turn been compiled from three different, even earlier poems, two of which had originally been part of other, longer poems. Wilamowitz went as far as to call the Iliad "a miserable piece of patchwork". Bryant argued the city of Troy never existed at all.

Arrayed against the Analysts were the Unitarians led by Gregor Wilhelm Nitzsch, who still believed the poems were mostly the work of a single author. Unable to counter Wolf's historical arguments, they relied on internal evidence, appealing to the poem's stylistic unity and intricate structure. How is it possible, they would ask, for a collection of different folk tales to have a coherent plot and characters, and where do all these elaborate structural correspondences come from?

The Unitarians had an anti-enlightenment aesthetic, fundamentally doubting the possibility of logical analysis of poetry. De Quincey got involved and declared all the external arguments irrelevant: "all arguments worth a straw in this matter must be derived from the internal structure of the 'Iliad.'" And in some cases the Unitarian cause was wrapped up with religious conservatism: if one accepted the analyst arguments in the case of Homer, one could hardly deny them in the case of the Old Testament.

Perhaps the most famous Unitarian was Heinrich Schliemann who was not a philologist at all but rather a globetrotting con artist and businessman. After amassing a vast fortune, Schliemann dedicated his life to finding the ruins of Troy. His trust in Homer was complete (even compared to other Unitarians, most of whom believed Homer contained no useful topographical details), and it was rewarded when he discovered Troy exactly where Homer said it would be. What he actually found was a later city on the same site as Troy and unfortunately he had no training as an archaeologist which meant that he completely destroyed the ruins, but Homer had led him to the right place.9

The battle raged on, and by the early 20th century the Analysts had the upper hand while Unitarianism was a heretical minority view. Thus in 1920 Georg Brandes could write that "save for a few uncritical people, of course, no one to-day believes that a single poet named Homer wrote either the Iliad or the Odyssey". Yet in the same year, John A. Scott, a respected philologist, published his passionately polemical The Unity of Homer, ridiculing the German Analysts and proclaiming the greatness of Homer.

And there were problems that neither side could solve, despite more than a century of intense scholarship, such as the peculiar mixture of dialects in Homer. The Iliad and the Odyssey are composed in a language that was never spoken by anyone, a blend of dialects from all across Greece. This was obviously a problem for the Unitarians, but the Analysts did not have any explanations either: the dialects were not cleanly separated (as you might expect from a combination of different poems), but jumbled together, sometimes even within a single word: a prefix from one dialect and a postfix from another. It was difficult to argue that 6th century Athenian poets wrote like this.


In the early 19th century, scientists started noticing that the orbit of Uranus did not match the predictions of Newton's laws. In 1846, the French astronomer Urbain Le Verrier proposed a solution: he hypothesized the existence of an undiscovered planet and, based on the magnitude of the perturbation, worked backwards and deduced its mass and orbit. Le Verrier sent his results by post to Johann Gottfried Galle at the Berlin Observatory. They were received on September 23. At midnight, Neptune was discovered—exactly in the position predicted. It was an absolute triumph: a bold mathematical prediction and confirmation of Newton's laws, instantly validated by the discovery of an entirely new planet.

A decade later, Le Verrier reported that the precession of the perihelion of Mercury did not match the predictions of Newton's laws. But he knew what to do, and once again estimated the mass and orbit of a hypothetical new planet which would explain the anomaly. We may deduce his confidence from the fact that he quickly named this planet—"Vulcan". (Astute readers may recall that no planet with that name can be found in our solar system.)

Astronomers failed to detect Vulcan, though they kept trying with every new eclipse up to 1908. In the face of this failure, they started coming up with increasingly absurd solutions: Asaph Hall suggested changing Newton's law of gravitation from Gm1m2/r2G m_1 m_2 / r^2 to Gm1m2/r2.00000016G m_1 m_2 / r^{2.00000016}. Others thought there was an invisible band of matter inside Mercury's orbit.

Sometimes the bottleneck to scientific progress is not data, but hypotheses. And hypotheses tend to be trickier to acquire. The problem was finally solved in 1915 when Einstein introduced general relativity, which perfectly predicted Mercury's perihelion advance without any ad hoc hypotheses or fudge factors.

The pragmatist philosopher Charles Sanders Peirce coined the term "abduction" (as opposed to "deduction" and "induction") to describe inference from effect to cause (in other words, from data to novel and successful hypotheses). As Le Verrier's example shows, it is a fickle art. The fact that abduction can be done at all, wrote Peirce, is "the most surprising of all the wonders of the universe".10

Part of what makes abduction so surprising is that scientific discovery often involves eureka moments, when the unconscious suddenly provides the solution to a previously insoluble problem. Ludwik Fleck characterized the process of discovery as going "from false assumptions and irreproducible initial experiments", through "many errors and detours" to eventually arriving at a result whose development none of the principal actors can explain. Gauss once said after a eureka moment: "I have the result, only I do not yet know how to get to it".11

Philosophers of science have generally avoided the topic entirely,12 using phrases like "ineffable", "nonrational", "intractable", and "surrounded by dense mists of romanticism".13 Popper famously wrote a book called The Logic of Scientific Discovery in which he denies any role to logic in scientific discovery.14 Einstein himself thought that "there is no way from experience to the setting up of a theory". Philosophers argue the matter should be left to psychologists, but psychologists haven't exactly taken up the torch. Neuroscientists haven't done much better. And from a computational complexity perspective, abduction is more or less impossible.

But I think there are reasons to be optimistic. First, the abductive process can at least be steered to some degree. It's not like scientists working on wallaby testicles have sudden insights about neutron stars. Second, some people clearly have greater abductive abilities than others, which at least tells us that there is something that one can have more of which improves the process. Finally, the frequent phenomenon of multiple discovery shows that abduction maybe isn't so magical after all: once the requisite foundations are in place, making the next leap is easy enough that we often get two15 people doing it independently. Skeptics ask awkward questions like: if abduction is a rational process, why haven't we discovered the rules? But just because they are hard to find doesn't mean that they do not exist.

To sum things up: your brain is a black box that "you" feed hard problems to and occasionally get answers from which "you" could never consciously discover, based on some alchemical mixture of conscious work, intuition, heuristics, thought experiments, and other impenetrable subconscious processes. Nobody knows how or why this magical meat works, nobody has even proposed a set of rules that might approximate the abduction process, and the continued upward techonomic trajectory of our civilization fundamentally depends on its continued operation.

The Solution

By the early 1900s the Homeric Question war had been raging for over a century, with no end in sight. Homerists had all the data, but what they really needed was the right hypothesis. The man to provide it was the American Milman Parry, who proposed a third alternative: neither pastiche nor unitarian composition, but the culmination of a formulaic oral tradition. The first breakthrough came with his 1928 dissertation, L'Épithète Traditionnelle dans Homère: Essai sur un problème de style Homérique16 in which Parry demonstrated that Homer composed orally, using fixed phrases and combinations called formulae. Five years later he travelled to Bosnia with Albert Lord, seeking out oral poets in places still unspoiled by literacy.17 By studying their technique and comparing it to Homer, they confirmed the oral-formulaic hypothesis.

To understand this system we have to start with a short refresher on Homer's meter, the dactylic hexameter: five dactyls (a long syllable followed by two short ones, —UU) plus a spondee (— —) or trochee (—U) at the end:

— U U, — U U, — U U, — U U, — U U, — —

Imagine a rapper freestyling a song about a war. Now imagine him doing that for 15,693 lines (or about 24 hours). Now imagine not just freestyling, but freestyling to a specific meter. Extemporaneous composition was only possible because Homer had a vast repertoire of formulae which fit onto the dactylic hexameter like lego blocks. For example, Achilles is πόδας ὠκὺς Ἀχιλλεύς (swift-footed Achilles) when Homer needs to fill seven syllables, and δῖος Ἀχιλλεύς (godlike Achilles) when he needs to fill five (and both formulae appear only at the end of the line).

Parry found two essential qualities in this system: scope, and economy. Homer had "a noun-epithet formula to meet every regularly recurring need. And what is equally striking, there is usually only one such formula". The formulae are not limited to epithets, but also cover type-scenes such as battles and speeches—explaining the repetitions which the Analysts blamed on copying by later poets. Similarly, the various small inconsistencies are simply the result of the method of composition: it's hard to keep every detail in mind when you're busy composing and singing.18

We shall find then, I think, that this failure to see the difference between written and oral verse was the greatest single obstacle to our understanding of Homer, we shall cease to be puzzled by much, we shall no longer look for much that Homer would never have thought of saying, and above all, we shall find that many, if not most of the questions we were asking, were not the right ones to ask.

Thus by abducting the right hypothesis, Parry could explain everything that had been mysterious about Homer's style. The mixture of dialects was a natural result of an old and widespread tradition. Archaisms and neologisms could exist side by side because the formulae evolved at different points in time.

The formulaic system was vast and Parry emphasized that it was impossible for any one man to put it together: this is where the element of tradition comes in. It developed by a slow processes of accumulation, a long lineage of bards across the centuries, each one adding a few phrases of his own to the great edifice of epic poetry.

We can use the development of the Greek language to trace the age of various formulae. Once a phrase was put in a particular metric place, it tended to stay there regardless of whether it continued to actually fit the meter. Early Greek featured a letter called the digamma (which represented the /w/ sound) and formulae were composed using that letter. As the language evolved the digamma disappeared and the formulae no longer fit the meter—but they stayed in the same place. This was a mortal blow to the Analysts: it would have been an extraordinary coincidence for a 6th century poet to write a metrically faulty line as if it had been caused by a letter that he had never heard of.

Just like the oral tradition preserved the ghost of a lost letter, it also preserved forgotten words: for "the great number of noun-epithet formulas in which the meaning of the epithet has been lost to us", the meaning had been lost to Homer also. And with the language it also preserved details from the world of Bronze Age Greece: wealthy Mycenae, war chariots, boar tusk helmets, etc. If the poems had continued being oral after Homer's time, the language would have continued evolving. Thus we can establish probable dates for the composition based on the neologisms.

Comically, the Germans simply pretended that Parry did not exist for a while. Harder wrote in 1942 that "no one any longer doubts that Homer could write, and wrote his poetry down".19 In the end, even though almost none of Homer's language was his own direct creation, you could say the Unitarians won: in his 2011 review of the Homeric Question, Martin West writes "we all agree these days that the Iliad and Odyssey are unified poems [...] And we all accept the implication, that this design is a design conceived by a single author, however much it may owe to earlier poems". If there were any doubt left, computer-based stylometric analysis finds "an astonishing homogeneity" in the Iliad (though not so much in the Odyssey).20

The final piece of the puzzle fell into place in the 1950s when Michael Ventris deciphered Linear B. Up to that point it was believed that the the Myceneans (the stars of Homer's story) were simply not Greek. Ventris himself was absolutely certain they were Etruscan, writing that the theory that Linear B "could be Greek is based...on a deliberate disregard for historical plausibility".21 This raised all sorts of questions about the transmission of details that predate the 13th century collapse. But since they were Greek, an oral epic tradition reaching back to that time is entirely plausible. This also provides us with an elegant explanation for the Bellerophon story: epic poetry preserved the idea of writing even though the technology had been lost for over 500 years:

"Would you be killed, o Proitos? Then murder Bellerophontes who tried to lie with me in love, though I was unwilling."
So she spoke, and anger took hold of the king at her story.
He shrank from killing him, since his heart was awed by such action,
but sent him away to Lykia, and handed him murderous symbols,
which he inscribed in a folding tablet, enough to destroy life,
and told him to show it to his wife's father, that he might perish.

And what about the transmission issues? The most popular view today (though there is no consensus) is the amanuensis theory: Homer was illiterate, but someone wrote down his song and passed it along. The example of Hesiod, whose long written works survived, shows that it's not outside the realm of possibility. Some of the difficulties Wolf pointed out are still there, and Haslam (a supporter of the amanuensis theory) says that writing down the poems would have been "an enterprise so remarkable that it is hard to credit". But the alternatives seem yet more doubtful.

In the 18-19th centuries many scholars, in whose minds literacy and intelligence were inextricably linked, worried that illiteracy would degrade Homer's achievement.22 But it only makes it more impressive. A poem like the Iliad can only exist because of Homer's illiteracy, not despite of it. Once literacy penetrated the Greek world, oral poetry disappeared. By the 6th century the bards had been replaced by rhapsodoi, no longer creative poets but merely parrots (like Plato's Ion).

Some argue that Homer was a transitional figure, that he could write but still knew the ways of oral composition. Others think that's impossible. The idea of a literate Homer opens up all sorts of intriguing and unlikely possibilities: writing—the technology that killed oral poetry—also immortalized what used to be transient and perishable. Remarkably, what it preserved was a poem about the use of art and fame (klea andron, the famous deeds of men) to overcome mortality. Perhaps the Iliad is one big allegory for the written word. Would Homer have been conscious of the decline of oral poetry? Ultimately, the only reason we can read him today is that he happened to live in that tiny sliver of time when oral poetry was still alive and the technology existed to preserve it for the future. Who knows what great poets sang into the void across those letterless aeons.

Not only did Homer have predecessors to imitate, he was drawing on an ancient tradition that survived the Greek Dark Ages—an eternal flame passed from singer to singer across the centuries, sustaining a vague, mythic memory of a fallen civilization. A tradition that was quickly extinguished when, on a bright summer morning, a curious trader asked some Phoenicians about their scribbles.

Further Reading

  • G. S. Kirk, The Songs of Homer
  • E. R. Dodds, Homer, in M. Platnauer (ed.) Fifty Years of Classical Scholarship
  • John Wright (ed.), Essays on the "Iliad": Selected Modern Criticism
  • George Steiner & Robert Fagles (eds.), Homer: A Collection of Critical Essays

  1. 1.The "tribute left to us by antiquity" is from Velleius Paterculus 1.5: in quo hoc maximum est, quod neque ante illum, quem ipse imitaretur, neque post illum, qui eum imitari posset, inventus est.
  2. 2.Such a miraculous event really did occur a bit later, with Herodotus and Thucydides. Perhaps the moral is that history is easier than poetry.
  3. 3."The effect of Wolf’s Prolegomena was so overwhelming that, although a few protests were made at the time, the true Homeric controversy did not begin till after Wolf’s death (1824). His speculations were thoroughly in harmony with the ideas and sentiment of the time, and his historical arguments, especially his long array of testimonies to the work of Peisistratus, were hardly challenged."
  4. 4.He was mainly inspired by Eichhorn and Heyne, who approached the Bible as a historical and anthropological document with multiple authors.
  5. 5.Actually, there is: the Bellerophon story. Wolf interpreted it away.
  6. 6.There was actually nothing original in Wolf's argument. Everything he wrote had been written before by other scholars: Blackwell, Wood, Bentley, Heyne, Eichhorn, even Rousseau. Casaubon had already written about the transmission problems in 1583. But Wolf's synthesis of all the arguments, combined with a masterful presentation, made those views persuasive.
  7. 7.Wolf never really explained how these people could produce poems from a very different era, why they would insert episodes they considered immoral, or why they were content to credit Homer for their own efforts. A century after Wolf, Comparetti commented: "where and when are great poets known and proved to have been so humble?"
  8. 8.Eduard Meyer, Geschichte des Altertums
  9. 9.Schliemann also excavated in Mycenae, and his finds there (like the Mask of Agamemnon) confirmed the Homeric idea that it was an extremely wealthy place. Mycenae never recovered from the 13th century Bronze Age Collapse, so by Homer's time it was a backwater.
  10. 10.This business is additionally complicated by the issue of underdetermination but let's leave that aside for now.
  11. 11.Which of course raises the question: did his subconscious have the proof? Why not "send up" the proof as well as the answer? Or was it just a guess? In which case how does the brain generate such good guesses? Perhaps it was simply a display of superiority.
  12. 12.With the exception of Carnap's abortive attempt in The Logical Structure of the World (1928).
  13. 13.Simon, Models of Discovery
  14. 14."The initial stage, the act of conceiving or inventing a theory, seems to me neither to call for logical analysis nor to be susceptible of it."
  15. 15.Or more: Merton found two cases of 9 independent co-discoverers.
  16. 16.In case you have forgotten your high-school French, Parry outlined his argument in English in two essays which can be found in Harvard Studies in Classical Philology, vols. 41 (1930) and 43 (1932).
  17. 17.An example of the importance of contingency in scientific progress: in 1819, the Austrian Empire’s renowned Slavic scholar, Jernej Kopitar, argued in a letter to Wolf that “today there is no better match for your Homeric ‘Homerids’ than in Serbia and Bosnia”. Wolf never bothered to follow up on the lead.
  18. 18.This might seem a bit mechanical, and Parry believed that they were "purely ornamental", but later work has shown that the epithets are context-sensitive to some degree. But words were not Homer's "unit of composition" as they would be for a literate poet, he composed by combining phrases. See John Miles Foley on the context-sensitive usage of epithets.
  19. 19.Das Neue Bild der Antike, ed. Berve, p. 102.
  20. 20.Dietmar Najock, Letter-Distribution and Authorship in Early Greek Epics
  21. 21.His belief was so powerful that even after successfully deciphering the script, he dismissed the result "as a mirage and cast the idea aside". It took him several months to figure out he was actually right. Eventually, with the right hypothesis, everything fell into place: "Once I made this assumption, most of the peculiarities of the language and spelling which had puzzled me seemed to find a logical explanation, and although many of the tablets remain as incomprehensible as before, many others are suddenly beginning to make sense." And speaking of inexplicable scientific leaps, Ventris was never able to produce a narrative of the method that allowed him to decipher Linear B.
  22. 22.Dodds: "That the Homeric poems are oral compositions is of course no new idea. It was put forward by Robert Wood in 1767, and developed (as regards the original state of the poems) by Wolf in 1795. But before Milman Parry it was open to any one to deny the assertion, and many scholars dismissed it as out of the question."