Saturday, September 09, 2006

ASK CORAX: pronouncing ancient greek

Magister,

In reading and discussing the Iliad, I'm picking up an actual aversion to using character names because I don't know how to say them correctly, and neither (it seems) does anyone else. For example,

Patroclus - Pátroklos - which I would instinctively pronounce something like PA-truh-kluss, I have heard in forms such as PE-truh-kluss, puh-TRAW-kluss, and even with a textually invisible extra schwa, puh-TRAW-kuh-luss.

I realize that none of these is probably analogous to the original Ancient Greek (if we can even say for sure how that language sounded), but I'd really like to be able to mumble something passable when I'm talking about Diomedes, Idomeneus, Andromache, etc. Any suggestions? Many thanks in advance,

Ian



SALVE DISCIPVLE OPTIME : S : V : B : E

ah my good ian, what a can of worms you've opened up here. nonetheless, i am glad you did so -- i want you to pursue your classical education up hill and down dale, and into every nook and cranny of knowledge that you can find. just so you know, though: this is the hard-core stuff you're peering into now.

the reason i call this 'a can of worms' is that you have touched on a couple of issues that cannot be disambiguated without getting pretty technical, nor without also addressing several other issues equally esoteric and difficult. that said, i'll try and keep this as streamlined -- and as lucid -- as i can. you know me well enough to know not to hesitate to ask for clarification if i have not been clear on this or that topic.

first, time out for some nomenclature.

[1] ACCENT in greek does not mean what it means in latin. rather it means something more like what is meant in modern chinese: classical greek accent was TONAL. thus [despite whatever you may hear, even out of the mouths of classics professors] those diacritical marks on a page of greek are NOT there to mark stressed syllables: they are to mark what was originally a quasi-musical tone [intervals of perhaps a musical third to a fifth].

[2] STRESS: when we say 'accent' in speaking of modern western-european languages, and even of classical latin, we are talking about the stress that falls on one or more syllable in a word. physiologically i assume this has to do with greater breath coming through the larynx and out the lips. as you can see, though, this is an accident waiting to happen, because of the confusion between classical and modern notions of 'accent.' so it's good to keep the notion of 'stress' separate from 'accent' in ancient greek.

when speaking of the stressed syllables in greek [or even latin] verse, some scholars, at least, still call this the 'ictus.' it's useful to have another word than 'accent' to talk about this, so i support the use of 'ictus' in this context. it is roughly synonymous in such a sense with 'rhythm,' but that's a mine-field of its own, in classical metrics, so best for the moment just to set the word 'rhythm' aside.

[3] QUANTITY: modern verse is ictus-based: that is to say, modern metrics depends on recognizing patterns of stressed and unstressed syllables. tennyson's

to strive, to seek, to find, and not to yield


[from his poem 'ulysses,' in fact] is a perfect iambic pentameter -- in the modern sense: five unstressed syllables alternating with five stressed. [note, just FWIW, that they are all ten of them monosyllables, and that 'not' is one of the stressed. masterful.] classical greek meter, on the other hand, was NOT ictus- or stress-based, but rather what we call quantative -- i.e. based on the QUANTITY [long or short] of each syllable. thus the first line of the iliad

andra moi ennepe, mousa, polutropon, hos mala polla


is scanned as a perfect dactylic hexameter -- in the ancient sense: six dactyls, i.e. six feet, each composed of one dactyl: one long syllable followed by two short ones. NOTE, however, that 'long' does not necessarily mean 'accented' [in either the modern or the ancient sense of that word]: i.e. it is possible for a LONG syllable to have a tonal marking or not. *** since both VOWELS and SYLLABLES can be termed 'long,' some scholars use the terms LIGHT and HEAVY to distinguish the short or long syllable when speaking of ancient greek [quantitative] verse.

[4] MELODY. as if this were not complicated enough already, add to all the above that homer's iliad and odyssey were originally SUNG to the lyre. that is, in addition to knowing where the stressed and unstressed syllables [if any] came, and where the tones rose and fell in ordinary conversation, there was also a melody mapped on to the text. i have always surmised that the melody [probably fairly primitive -- and n.b., the lyre in the 8th century BCE only had four strings] roughly followed the ups and downs of the spoken tones of greek. but there's no absolute proof of that; we don't have any reliable evidence of what homeric music was like; but i was gratified some years ago when martin west, one of the greatest living homeric scholars, published an essay in the JOURNAL OF HELLENIC STUDIES positing this theory himself. [for some inexplicable reason he does not, however, acknowledge me as the fons et origo of the idea ...]

so you see, to answer your question we are conjuring with a number of different elements here. you ask, innocently enough, how to pronounce 'patroklos'; i assume that at one level you are asking how the name was spoken in conversation. but i do want you to realize that that is [or might be] different from how it is spoken [or sung!] in homeric verse. also, that the way homer pronounced it [in conversation] might have differed from how a byzantine [or later] scholar pronounced it.

the accent marks, incidentally, were not yet written on greek words in plato's day, and certainly not in homer's [if anyone was even writing greek in homer's day -- yet another can of worms]. plato is of course aware of the differences in tonal accent, and in fact refers to the phenomenon in his CRATYLVS [399], but he is talking about the sound, not any written indication of accent. it appears that accent marks began to be added in about 200 BCE -- probably at alexandria, and very likely by the scholar aristophanes of byzantium.

by the time of socrates and plato, the living tradition of improvised oral [pre-literate] epic, i.e. what homer produced as a creative artist, was long gone; what you had instead was a professional class of re-creative artists known as RHAPSODES, who RECITED [rather than SANG] the portions of the epics that they had MEMORIZED. so already the performative tradition had morphed drastically by the fifth century BCE, the glory days of athens.

i have made a couple of recordings of classical greek verse. these are online in streaming audio at

http://www.cla.purdue.edu/academic/fll/kirby/index.htm

if you play the one of stesichorus's 'palinode,' you will hear my attempt at rendering classical greek tonally. the other excerpt, which is a bit longer, is a passage from homer [iliad 3.395-412], and in this one i despaired of trying to keep a regular hexameter rhythm going while also pronouncing the tonal accents correctly. but if homer had been induced to speak rather than to sing his verses, that is precisely what i expect he would have done. did the rhapsodes also recite tonally? i bet they still did, for the simple reason that [on the evidence of plato's cratylus] they were still speaking conversationally with tonal accents.

here are some other recordings of homer read aloud with attention to the tonal accents:

http://turdpolish.com/greek.html [despite the inelegant URL he has some smart things to say here, as well as some well-produced sound-recordings]
http://www.prosoidia.com/od1-10.html

several hundred years after plato -- certainly by the late fourth century CE, and likely a couple hundred years before that -- the shift had probably reached the point where the tonal accent was being lost in greek. what was happening at that point was that the syllables marked with [originally tonal] accent-marks were beginning to be pronounced as if they received the stress. and, moreover, new verse being composed in that era was stress-based, rather than quantitative.

i still have not answered your question, i see. here is what w. sidney allen, in his authoritative book VOX GRAECA, has to say about stress in classical greek [i assume he was not talking about its effect on the student studying it]:

the classical greek accent was, as we have seen, tonal. it is, however, improbable that greek words and sentences had no variations of stress. this has often been recognized, but there has been a tendency to assume that any such element of stress must have been connected with the high tone, since pitch is frequently an important factor in the complex phenomenon of stress-accentuation. but, for one thing, it is not necessarily high pitch that is involved in such cases -- in different languages it may be high, low, or changing pitch; and for another, stress is not conversely a necessary feature of tonal accentuation; so that it is possible for a language having a tonal accent to have also a stress-patterning that is quite unconnected with this accent.

moreover, any connection of stress with high tone seems to be ruled out by the fact that in classical greek there is no correlation of the accent with the metrical stress or 'ictus' .... we therefore assume that greek verse was recited with a stressed rhythm .... it should ... be possible, from a statistical study of various types of greek metre, to discover whether there are any strong tendencies for particular syllables in words of various quantitative patterns to coincide with the presumed ictus of the verse; if so, it may be reasonably deduced that such syllables were normally stressed in speech. from such a study the following rules seem to emerge [for words of more than one syllable]: [a] words were primarily stressed on their last heavy syllable; [b] a secondary stress fell on preceding heavy syllables if separated from the primary stress by at least one mora of quantity ... this hypothesis produces 90-95% agreement between verse-ictus and natural [prose] speech-rhythm in both the tragic trimeter and the epic hexameter
.... [[pp 120-121]]

and how, you may well ask, does one distinguish a heavy syllable from a light? here's allen again:

if a syllable contains a long vowel, it is always 'heavy' .... but if it contains a short vowel, its quantity depends upon the nature of the syllable-ending. if it ends with a vowel ['open' syllable], the syllable is 'light' ... but if it ends with a consonant ['closed' syllable], the syllable is heavy. [[p 97]]

that book was first published in 1968 [third edition 1987; my citations are from the second edition of 1974]. allen went on to advance a more fully-enunciated version of his stress-accent theory of spoken ancient greek in a book called ACCENT AND RHYTHM [1973, so my 2nd edition of VOX GRAECA would reflect that]. the whole thing boiled up a firestorm of controversy: based on what we can find in ancient grammarians who sort-of-discuss this stuff, the received wisdom was that there was no stress accent in classical greek, whether spoken or recited [i.e. prose/conversation nor verse]. but allen was one of the 'big guns' of historical linguistics at that time, so his theories could not be dismissed lightly. on the whole, i would say that most linguists have felt the need to reject allen's thesis -- very respectfully, to be sure, but reject it nonetheless. a good example would be the article by a. m. devine and l. d. stephens in the transactions of the american philological society 115 (1985) 125-152, 'stress in greek?' in this long and technical essay, they weigh allen's thesis on its merits, but decide when all is said and done that classical greek was probably a non-stress-accented language like japanese. here is the clearest passage in that essay [this from pp 146-147]:

if we believe that in ancient greek the accented vowel was not only higher pitched but also significantly louder and/or longer ceteris paribus than unaccented vowels, then the greek accent was a pitch differentiated stress like that of serbo-croat or lithuanian. in that case, the innovation of the modern greek accent is limited to the elimination of the rising/falling distinction between acute and circumflex and the sharpening of the durational distinction between accented and unaccented vowels (note in particular the elimination of long vowels in unaccented syllables). if, on the other hand, we believe that in ancient greek accented vowels were not significantly louder or longer ceteris paribus than unaccented vowels, then the ancient greek accent was a pure pitch accent comparable mutatis mutandis to that of japanese. this latter view has the better chance of being correct.

i suppose one of the most important, perhaps THE most important, lesson to be taken away from this is that not even the experts can agree. allen is a god in that subfield of classics, and devine & stephens are among the very most respected classical linguists working today; but they are diametrically at odds here -- and yet the most emphatic demurrer devine & stephens can offer is 'this latter view has the BETTER CHANCE of being correct.' nobody truly knows.

now the best argument that can be rallied [apart from allen's own] against what devine & stephens assert here is that greek is, like latin [and serbian and croatian and lithuanian] an indo-european language, whereas japanese is not. it would be very helpful if one could cite uncontroversially an indo-european language in which there is no stress accent. FRENCH is sometimes adduced as an example -- i.e. one is often taught that each syllable in a french sentence should receive an identical amount of stress -- but all you have to do is turn on any french TV or radio to see that this is science fiction. [in practice, it's better to teach one's french student that the default syllable for a stress-accent is the ultima. but that's for someone else's blog to hash out, i guess.]

a second argument might be the very difficulty of imagining a pre-literate poet improvising, on the spot, hundreds of lines of hexameter verse [in a meter, by the way, that was not native to greek -- it seems to have been imported into the greek-speaking world from the indus valley] in which he is expected sometimes to balance syllables that are long but unstressed, on one hand, against syllables that are short but stressed, on the other. [this is precisely what vergil does, incidentally, in the AENEID -- but remember that vergil is [a] composing in LATIN, where there has been a long prior tradition of stress-based verse; [b] composing in WRITING, which gives him the opportunity to linger as long as he wants [we calculate from his comments about the GEORGICS that he composed these at the rate of one hexameter line per day], and [c] experimenting [as it's been shown by scholars] with a particular technique in which the first half or so of the hexameter line deals with conscious tension between long syllables and stressed syllables, whereas the second half of the line has much greater coincidence of long and stressed. in other words, the vergilian hexameter is quite a different animal from the oral improvised verse of homer and hesiod.

we've come a very long way, i fear, for you just to be told 'there's no clear answer to your query.' if you are fuming by now, you have every right to be. let me say, though, that this also gives YOU a certain license: it means that you can pronounce 'patroklos' in whatever way pleases you best.

one way that has appealed to many who come [like you] from latin to greek, is to use the rules of latin accent: if the vowel of the penult is long, the penult is stressed; otherwise the antepenult. [remember now that we are talking just about VOWELS, not about SYLLABLES, in which a short vowel + 2 consonants = a long syllable.] as long as you remind yourself every day that that is NOT a rule in greek, you could do worse than that. moreover, it accords somewhat with what allen so gingerly advances in the theory cited above.

if you were to approach it this way, you would have the following the [tonal] accent is on the first syllable, but the stress is on the penult'. saying 'putt-ROK-loss,' with a trilled R, ought to get you somewhere in the right ball park. if you want to get fancy, pronounce the first syllable higher than the other two -- that should be the most 'accurate' for the greek of homer's day.

as for the other names you mention -- diomedes, idomeneus, andromache -- the situation is further complicated by [a] the way the ROMANS pronounced greek words, which may well have been something like what we've discussed above, and [b] the way MODERNS [in english, german, french, etc] pronounced greek [and, for that matter, latin] before the twentieth century. if you read the wonderful novella GOODBYE, MR CHIPS, or c. s. lewis's autobiography SURPRISED BY JOY, you will see references to the 'new' pronunciation of greek and latin.

why are you hearing such disparity of pronunciation amongst your colleagues? well, the pronouncing of the language overall is a difficult problem, but the whole NAME issue is particularly hopelessly mucked up. we refer to vergil's poet friend as HORACE, for example, but that's because the french call him that; we borrowed their name for him. his actual full name, in latin, was 'publius horatius flaccus' -- except in the vocative case, which was 'publi horati flacce.' his mother probably called the young horace to dinner with 'publiiiii!' -- but in adulthood, friends and acquaintances probably referred to him as horatius [the family name] or flaccus or horatius flaccus. so what should we do? insist on 'horatius'? nobody will know whom we are talking about, at least not till we explain. [and, to be consistent, it would mean calling catullus 'valerius' from now on ... etc etc etc]

so too with APOLLO: this is what the romans called this [greco-roman] god. but in greek he was called APOLLON, with the final O being an omega. so what should we do? insist on the 'correct' greek name? we will sound hopelessly pedantic. [even more than most classical scholars do.]

this is the problem we are facing with 'diomedes.' in greek, those vowels come out, more or less, 'dee-oh-MEH-dess.' that will work fine in germany. but -- if you are speaking english, to someone in the USA or UK, it will seem quite esoteric. most americans say 'dye-oh-MEE-deez.'

'idomeneus' entails a particular problem: that final diphthong, which is odd to american ears. if you can imagine the duchess of devonshire wrinkling her nose in disdain, and exclaiming 'oh!' -- that is more or less the sound of the diphthong EU in classical greek. [my first greek professor told us to say it the way choate boys say 'choate.' as an old choate boy myself, i took umbrage, but i said nothing. the next day i did come to class wearing my choate tee shirt, though.] 'idomeneus' is an unusual enough name that you can probably get away pronouncing it 'accurately' in that way -- but if you are going to strive for absolute consistency, you will have to pronounce 'zeus' that way too -- and nobody i know says anything, in english, other than 'zooss' for that.

andromache: again, the ancients would have trilled that R at least lightly. you will sound precious if you do that in english. i suggest 'ann-DROM-ma-khee,' bearing in mind that in greek, the final E was an eta, and thus more or less a longish open 'ehh.'

this should at least get you started. i should close by stressing once again that these are deep and troubled waters: not entirely uncharted, but full of sea-serpents. i've probably given you a lot more philology here than you bargained for; but you deserve nothing less. that's what i say.

some more relevant and useful URLs, if you haven't yet quenched your linguistic thirst:

greek phonology
http://en.wikipedia.org/wiki/Ancient_Greek_phonology

phonation in different cultures
http://en.wikipedia.org/wiki/Pronunciation_of_Ancient_Greek_in_teaching

mora
http://en.wikipedia.org/wiki/Mora_%28linguistics%29

erasmian pronunciation
http://abecedaria.blogspot.com/2005/10/erasmian-pronunciation.html

an impassioned plea, by william harris, for reading greek tonally
http://community.middlebury.edu/~harris/Classics/Greekaccents.html

5 comments:

Anonymous said...

Professor,

Your answer [such as it is :) ]makes me glad I'm a Linguistics major, and even so I think more than a little of it is at the periphery of my comprehension. That said, I'm glad this can of worms was opened, if only thanks to my naivete. I've taken a lot from what you've said, and I'm sure I'll do the same the next three times I read it. At any rate, now I can start dropping words like 'ictus' and referring to Horace as 'Publiiiii!' to my [hopefully] amazed [rather than nonplussed] friends. Thanks very much for taking time to swat the pronunciation fly for me and others.

Ian

Anonymous said...

So, I took a course of Ancient Greek a couple of semesters ago, and was very surprised to hear how foreign the language sounded to me, even though I speak modern Greek. As you obviously know, Ancient Greek is much more complex than modern Greek, but it uses the same alphabet, and there are many similarities in the vocabulary. I know you already covered the complexities of tonal emphasis and accents in the blog, but I'm just bringing up the topic of mere pronunciation. The Greek pronunciations I had to learn in that class made me cringe(by the way, the Greek version of the adage 'It's Greek to me' is 'It's Chinese to me', and that's just like what it sounded when we had to read passages aloud in class).

In Greece, they teach Ancient Greek with modern Greek pronunciation. So why is it that, at least in the United States, the pronunciation of certain vowels and diphthongs are so different? An example to make my point: I was taught in my Ancient Greek class that the letter upsilon was pronounced "oo", while all my life I've known it as a long "ee" sound. Also, the diphthong "oi" (omicron iota) should be pronounced as a long "ee", not "oi" as in "coin".

I asked my other professor the same question, but he retorted with something along the lines of, "They teach it wrong in Greece. We are teaching it the right way." What do you think of this discrepancy?

Αγγελος said...

Quintus Horatius Flaccus, wasn't it? His mother must have called him "Quinte!"

Αγγελος said...

To Daphne: Greeks sometime have trouble accepting this for a fact, but it is absolutely certain that the pronunciation of Greek changed very drastically in Hellenistic and Roman times, and in particular that H was a kind of long and open E, that diphthongs were pronounced more or less as spelt and that Β, Γ, Δ sounded like B, G and D (which is why the bleating of sheep is transcribed as BH, BH in Aristophanes). It is this pronunciation that the teaching of Ancient Greek outside Greece tries to approximate, both because it is closer to (ancient) reality and because it makes it much easier to learn to spell!

corax said...

for the record, aggelos is exactly right: horace's *praenomen* was QVINTVS -- vocative QVINTE. blip!