I thought about saying something about this (“Homer’s great masterpieces, The Iliad and The Odyssey, have been dated to around 762 BCE by new research based on the statistical modelling of language evolution”) back when it was in the news, but frankly it made me tired. Fortunately, Memiyawanzi had a rant about it that expressed what was in my soul:

I really don’t know what to say. Bayesian phylogenetics applied to raw lexical data gives tenuous results at its good, bizarre BBC headlines like ‘English Language originated in Turkey‘ at its bad, and now, can be used as a terrible, terrible replacement for traditional textual criticism and philology at a bar-lowering new ugly for the mindless glottogonic speculation that is increasingly being made in this area by researchers in genetics with little to no actual historical linguistics training.

And as eoforholt says in the comment thread: “It’s worse than the press release makes out. In the article itself, they say that their method on its own actually yields a 95% confidence interval 61-1351 BC, a 1290 year range. It’s only when they weight their model earlier (to take into account Herodotos’ mention of Homer) that they get the range discussed in the press release.” I will, of course, be interested in responses from people who know more about this stuff than I do; my skepticism meter sometimes goes higher than is entirely warranted.


  1. I’m glad you posted this, if only so I could reread Mattitiyahu’s wonderful phrase “crimes against humanities.”
    Some fifty-five years ago, when my wife was a teenager and mailed in song requests on postcards to her local radio station under the name of “Tigerlily”, her favorite DJ was suddenly taken off the air without a word of explanation. About a month later, she saw a brief article in the local newspaper saying he had been convicted of “crimes against nature”. She had never heard this expression before, and supposed that he must have been caught doing something like cutting down trees in a state park.

  2. -What’s the age of this dinosaur?
    -75 million and 12 years!
    – Wow! How can you be so sure of this dating?
    – When I joined this museum twelve years ago, director said it was 75 million years old.

  3. I looked at the article. All they do is use glottochronology (the fancy Bayesian kind) to compare Homeric Greek with Modern Greek and Hittite to estimate an age. Fine.
    To their credit, they do give error estimates, which the traditional glottochronologists didn’t. So by their estimate Homeric Greek date quite certainly was between the first and 14th centuries BC, and probably between, oh, 500 and 1100 BC (I’m eyeballing the graph). Fine. Thus far the method agrees with traditional methods, though it’s not as precise.
    But then, what’s this 707 BC, or with Herodotos, 762 BC? That’s where the mean of the distribution is. Why not July 11th, 762 BC, 11 minutes after suppertime? You could estimate the mean of the distribution to any number of significant digits, and it makes for a better press release than “scientists figure out what other people figured out before, much quicker but not as precisely.” Newspapers use too much precision all the time (“Go about 1609 meters down the road, you can’t miss it.”) Supposed statisticians don’t have an excuse for doing that.

  4. Some dates and scenarios for “Homer” (i.e., Iliad and Odyssey) by some eminent late 20th to early 21st century (CE) Homeric scholars (the dates below are all BCE, of course):
    C.J. Ruijgh – late 9th century in
    G.S. Kirk – Iliad composed orally mid-8th century; transmitted orally until mid-7th century, when reduced to writing.
    Richard Janko (and others) – mid to late 8th century; oral dictated text.
    M.L. West – Iliad composed in writing by a single individual working in an oral tradition over many years in the mid-7th century, in Asia Minor.
    Minna Skafte Jensen – text dictated and written around 525 after a long period of oral evolution.
    Greg Nagy – text “stablilized” late 6th century; continuing fluidity in oral performances until second century, when the text represented by the medieval ms. tradition became more or less fixed.
    Take your pick.
    All we can be certain about is that the Iliad and the Odyssey are the products of a long tradition of oral, composed-in-performance poetry, possibly dating back to Indo-European times but almost certainly to the Mycenean era (pre-1200 roughly); and that written texts of the Homeric poems, reflecting “Ionic” Greek spoken in Asia Minor or possibly the island of Euboea (but including material from other dialects and earlier Greek) were in circulation by the end of the 6th century, but everything else is speculation based on very slender evidence.
    At this point, just about every theory that could possibly be articulated has been advanced by someone. None of these theories is wholly satisfying and none has managed to command universal assent among scholars working in this field, beyond the fact that the poems are the product of an oral tradition. These questions have played themselves out.
    Here’s what I think is about the most that can be said: The Iliad “is likely to be the result of extremely complicated processes involving both orality and writing, which we can no longer reconstruct.” A.C. Cassio, “Early editions of the Greek epics and Homeric textual criticism”,in “Omero tremila anni dopo” (Rome 2002), 114.
    (But still, speculation is irresistible.)

  5. We know (or think) that the oral tradition was very conservative and preserved words and grammatical forms from very early times (some argue that there are verses that can be traced back to as early as 1600 BCE), but at the same time introduced more “contemporary” words and grammatical forms (leaving open the question “contemporary with what?”). The medieval manuscript tradition (the “paradosis”) even includes grammatical forms that are alleged to date no later than the 4th century BCE. It’s argued that these represent very late spelling changes that were made substantially after the poems were reduced to writing in order to normalize the text for readers. (M.L. West’s Teubneriana edition of the Iliad, published around 2000, eliminates these “late” forms, reconstructing more archaic forms that aren’t found in any manuscripts, even in papyrus fragments dating from the 3rd century BCE to the 6th or 7th centuries CE–this is a questionable practice in my view.) So I suspect that to arrive at the 762 BCE date you have to engage in some manipulation of the evidence of the paradosis.

  6. I should have noted that M.L. West is a very great Greek scholar who has made enormous contributions to the study of archaic and classical Greek literature, and whose knowledge of the ancient Greek language is unrivalled today. I don’t agree with many things he has written about the Iliad–or rather, I’m skeptical, and I’m not alone in my skepticism–but I have to confess that he knows the poems much more better than I ever will.

  7. How to tell it’s time to cut down on the pop culture and get back to the classics: when your first thought on reading the title “Dating Homer” is that the post will have something to do with Marge Simpson reminiscing about her high school days.

  8. I have to confess I was not incognizant of that possible interpretation when I titled my post; it gave me a secret smile.

  9. befuggled says

    I’m just surprised that none of the spam I’ve seen on the comments for this post are for skeezy dating sites. (Although I imagine you’ve deleted more spam than I’ve seen.)

  10. There’s nothing really wrong with the analysis, except for the “I’m in need of a headline” accuracy of 762 B.C. Since it’s a Bayesian estimate (i.e., an estimate that depends on the estimator’s estimate of the extent of his/her own ignorance), anyone who knows more about Homer than the authors is fully entitled to do better.

  11. MattF, this analysis in and of itself is actually of no value. If it’s a crude stepping stone on the way to developing something actually useful, then I guess that’s something – but this study itself is so broad in its predicted time span (having basically a 1 in 4 chance of covering the estimated date of Homer by randomly guessing) that I’m not sure it even suggests that the method has any real hope of ever being refined to an actually useful level.

  12. It might bear repeating that:
    “All that mathematical modelling reveals is the consequences of your assumptions. Change the assumptions and a model yields different results (except in Climate Science of course).”

  13. I came to the same conclusion as Nelson: if the new dating method gives results that are hopelessly less precise than the old one (1290 year range v. 400 years for the scholars Bill W. cited) haven’t the scientists in Reading just successfully disproved their method?

  14. I can’t understand why they are only doing this with these three data points. It would make more sense to repeat the method with different ‘bracketing’ languages to see if they yield different results.
    If they yield different results, then the method is at best as good as the choice of ‘bracketing’ languages. If they yield similar results, maybe with some problems with some poorly-dated splits, then the method is at least consistent with itself.
    For the later language, various stages of Koine Greek and medieval Greek would make as much sense as modern Greek.
    For the earlier language-split, although Linear B doesn’t preserve poetry or literature, it would be ideal because it’s relatively datable. The split between the Anatolian family and the rest of the Indo-European families is incredibly controversial. The later splits aren’t all as controversial, so using an Indo-Iranian language and/or an Italo-Celtic language for the earlier language-splits might work better than using Hittite.

  15. “although Linear B doesn’t preserve poetry or literature, it would be ideal”
    I think the problems with using Linear B would be that (1) the attested vocabulary is limited very limited–largely to property that can be taxed, with few verbs and not many of the words that glottochronologists like to focus on–and (2) the interpretation of many words is still somewhat controversial, and the script provides very incomplete information about the phonological realization of words that are attested.
    One other problem with the method itself as applied to the Homeric poems is that, as is well known to people who have actually bothered to engage with the linguistic aspects, the Homeric language was never a spoken language–it’s an artificial language that that evolved over a long period of time, perhaps as long as 1,000 years, in the specific context of oral performances and incorporates “fossilized” words and grammatical forms from at least a good part of its history. Knowing that, any attempt to pinpoint a specific date for the language with no consideration of historical context seems futile.

Speak Your Mind