Talk:Ancient Greek phonology

Editors who are interested in improving this article are encouraged to read this talk page discussion and the previous discussion at Archive 1, Archive 2, and Archive 3.

metrical evidence

"The metres used in Classical Greek poetry are based on the patterns of light" hillarious. --62.90.23.122 (talk) 12:57, 12 June 2012 (UTC)

Aristoteles

Hello,

is there a quote from Aristoteles (and a translation of that) concerning the existence of aspirated and unaspirated consonants in Ancient Greek?

Greetings HeliosX (talk) 17:58, 6 December 2012 (UTC)

That would be Poetics 20. What we are saying in the article is that Aristotle refers to the three series βδγ, πτκ, φθχ together as "aphona", which he indeed does in that passage. His description is not very precise phonetically, but his definition of "aphona" clearly implies that they are momentary, non-protracted sounds (sounds that "have no sound by themselves but only become audible together with a neighboring vowel"), and that in this they differ from "hemiphona", i.e. consonants that can be pronounced by themselves. The latter group include fricatives such as /s/, so we can infer that the nine sounds in the "aphona" group cannot have been fricatives.

Original text [1]: "Ταύτης δὲ μέρη τό τε φωνῆεν καὶ τὸ ἡμίφωνον καὶ ἄφωνον. Ἔστιν δὲ ταῦτα φωνῆεν μὲν <τὸ> ἄνευ προσβολῆς ἔχον φωνὴν ἀκουστήν, ἡμίφωνον δὲ τὸ μετὰ προσβολῆς ἔχον φωνὴν ἀκουστήν, οἷον τὸ Σ καὶ τὸ Ρ, ἄφωνον δὲ τὸ μετὰ προσβολῆς καθ᾽ αὑτὸ μὲν οὐδεμίαν ἔχον φωνήν, μετὰ δὲ (30) τῶν ἐχόντων τινὰ φωνὴν γινόμενον ἀκουστόν, οἷον τὸ Γ καὶ τὸ Δ. Ταῦτα δὲ διαφέρει σχήμασίν τε τοῦ στόματος καὶ τόποις καὶ δασύτητι καὶ ψιλότητι"

Translation: [2] "The sound I mean may be either a vowel, a semi-vowel, or a mute. A vowel is that which without impact of tongue or lip has an audible sound. A semi-vowel, that which with such impact has an audible sound, as S and R. A mute, that which with such impact has by itself no sound, but joined to a vowel sound becomes audible, as G and D. These are distinguished according to the form assumed by the mouth and the place where they are produced; according as they are aspirated or smooth, long or short;"

Fut.Perf. ☼ 18:17, 6 December 2012 (UTC)

My impression is that the ancient Greeks understood quite well the difference between an aspirated and a non-aspirated (unvoiced) consonant, but not the difference between voiced and voiceless. The notion that the voiced plosives are somehow "in the middle" between the aspirated and unaspirated voiceless plosives is strange by today's standards. Voiced plosives can be sustained for a short time until the mouth is full of air, whereas the aspiration of an aspirated plosive can be sustained as long as there is air in the lungs. In this sense the media are intermediate regarding sustainability.  Andreas  ^(T) 16:04, 24 January 2013 (UTC)

BTW, Aristotle in this section doesn't actually provide an exhaustive list of which sounds he considers to fall into each category, but we can be pretty sure about what he meant based on the treatments of other authors. A fuller account, using the same distinction between "vowels", "hemiphona" and "aphona" is given by Dionysius Thrax [3], who also clearly distinguishes between the three sets βδγ, πτκ, φθχ. Fut.Perf. ☼ 22:27, 6 December 2012 (UTC)

Thank you for your answer! Is there perhaps a translation of that text of Dionysios Thrax (I didn't find any yet)?

Greetings HeliosX (talk) 15:07, 17 January 2013 (UTC)

Hmm, I'll try my own:

"Of these [the 17 consonant letters], eight are hemiphona: ζ ξ ψ λ μ ν ρ σ. They are called "hemiphona" because they are less sonorous than vowels, like moaning or hissing sounds. The other nine are aphona. They are called thus because they sound less good than the others, just as we also say of a bad singer that he "lacks a voice". Of these, three are thin (psila): κ π τ; three are thick (dasea): θ φ χ; and three are intermediate (mesa): β γ δ. They are called "mesa" because they are thicker than the thin ones but thinner than the thick ones; thus, β is intermediate between π and φ; γ between κ and χ, and δ between θ and τ."

He then goes on to demonstrate the correspondence between the psila and the dasea, by showing how each psilon turns into the corresponding dasy when followed by an aspirated vowel, with one line of Homeric verse for each pair ("εἴπε ὁπηι" > "εἴφ’ ὅπηι", "αὐτίκα ὁ" > "αὐτίχ’ ὁ", "ἔφαθε οἱ" > "ἔφαθ’ οἱ")

Fut.Perf. ☼ 15:36, 17 January 2013 (UTC)

Thank you for your answer. But are there also some Ancient Greek grammarians, who described the sounds more into the study of sound, as to say whether it's aspirated or smooth, or if it consists actually of two sounds as Ξ?

Greetings HeliosX (talk) 15:03, 21 January 2013 (UTC)

There are quite a number of other rhetoricians and grammarians who have chapters on the sound system; I believe most of them follow pretty similar lines in their presentation (these guys tended to copy a lot from each other), but I certainly don't know all of them. I just found a somewhat more detailed treatment in Dionysius of Halicarnassus though (De compositione verborum, §14). He uses the same terminology as the others, but has more concrete phonetic detail, such as describing vowels in terms of the open–close and rounded–unrounded dimensions, and describing consonants systematically according to place and manner of articulation. As for consonant letters composed of two sounds, he (just like D. Thrax) quite explicitly distinguishes Ξ, Ψ and Z in this way. About the aphona, he is explicit about all three series beinɡ stops, i.e. produced with a full closure of the airstream followed by an openinɡ burst ("ὅταν τοῦ στόματος πιεσθέντος τότε προβαλλόμενον ἐκ τῆς ἀρτηρίας τὸ πνεῦμα λύσῃ τὸν δεσμὸν αὐτοῦ") – that's pretty much the same way we define plosives in modern linɡuistics. The way he defines the three manners of articulation is still a bit hazy, as he continues to use that odd impressionistic metaphor of ΦΘΧ being "rough" and ΠΤΚ being "smooth" and ΒΔΓ "in the middle", but he does speak of the "rough" (dasea) ones involving "the addition of aspiration" ("καὶ τὴν τοῦ πνεύματος προσθήκην"), as opposed to the "smooth" ones consisting only of "their own sound quality by itself" ("τὴν ἑαυτῶν δύναμιν [...] μόνην"). That is the closest I have found so far to a phonetic description of the distinction between aspirated and unaspirated stops. Fut.Perf. ☼ 19:25, 21 January 2013 (UTC)

Thank you for your answer! Can these classifications be interpreted that way: π [p] - β [βʰ] - φ [pʰ] as the Ancient Greek classified a sound's thickness maybe out of how "wide" it was or how much it sounded "like" a vowel.

And can you help me by the translation of this: „διπλᾶ δὲ τρία τό τε ζ καὶ τὸ ξ καὶ τὸ ψ. διπλᾶ δὲ λέγουσιν αὐτὰ ἤτοι διὰ τὸ σύνθετα εἶναι τὸ μὲν ζ διὰ τοῦ ς καὶ δ, τὸ δὲ ξ διὰ τοῦ κ καὶ ς, τὸ δὲ ψ διὰ τοῦ π καὶ ς“

Three [sounds] are double: ζ, ξ and ψ. They are called double because they are composed, ζ of ς and δ, ξ of κ and ς, and ψ of π and ς.  Andreas  ^(T) 15:33, 24 January 2013 (UTC)

Greetings HeliosX (talk) 15:08, 24 January 2013 (UTC)

Not quite sure what you mean by "[βʰ]" above – "β" in the sense of the IPA symbol β? That would mean a fricative, but the way I read it, it would directly contradict the description of all of π/β/φ as stops. Also, combining such a fricative with aspiration would be pretty much physically impossible, as far as I know. If, on the other hand, you meant IPA [bʰ], i.e. an aspirated voiced plosive, that too is, strictly speaking, not really a physical possibility as far as I know. (The sounds you have perhaps seen described that way in reconstructed Indo-European linguistics are really something different, see our article breathy voice). I also don't see anything in the texts that relates the "smooth-medium-rough" terminology to different degrees of being "like a vowel" (in fact, the distinction between "aphona" and "hemiphona" has such a meaning, but all three of "smooth-medium-rough" are always grouped among the "aphona" by these authors. Note that I may have inadvertently muddied the waters earlier by rendering "psila" and "dasea" as "slender" and "broad". I believe "smooth" and "rough" are much more correct translations.)

As for the "διπλᾶ δὲ τρία..." sentence, it means "The dipla ["double"] consonants are three in number: Ζ, Ξ, and Ψ. They are called "double" either because they are composite – Ζ being composed of Σ and Δ; Ξ of Κ and Σ; and Ψ of Π and Σ […] [or because they stand in the place of two consonants in syllable structure]." Fut.Perf. ☼ 15:39, 24 January 2013 (UTC)

Thank you for your answer, what do they say about the ΕΥ-Diphthong? (And also, had Sappho written something about pronounciation? It would then be Lesbian Greek?)

Greetings HeliosX (talk) 17:21, 31 January 2013 (UTC)

Comparison with younger/derived alphabets

For example, in Cyrillic, the letter В (ve) stands for [v], confirming that beta was pronounced as a fricative by the 9th century AD, while the new letter Б (be) was invented to note the sound [b]. Conversely, in Gothic, the letter derived from beta stands for [b], so in the 4th century AD, beta was still a plosive in Greek.

Wait a second! For a long time, β was used to transcribe /b/, /v/ and sometimes /w/ of foreign names without distinction. The emperor Valentinianus ended up as Βαλεντινιανός.* The Slavic languages have both /b/ and /v/, so the invention of Cyrillic forced a decision, and there it turns out that /v/ was a better match. But Gothic only had a single phoneme in that place, something that was [b] behind /m/ and perhaps word-initially and [β] or [v] elsewhere. This actually shows up in Gothic spelling because of a Gothic innovation, word-final devoicing, which merged this phoneme into /f/, never into /p/. So, the pronunciation of β must have been close to one of the voiced allophones of that Gothic phoneme, but not necessarily to the plosive one (which was, furthermore, comparatively rare).

I guess it's also possible that β itself had such allophones for some time.

*...which may or may not tell us something interesting about Latin, namely that v had changed from [w] to [v] sometime between him and whichever Valerius it was that ended up as Οὐαλεριός.

David Marjanović (talk) 20:36, 14 March 2013 (UTC)

I just finished reading Caragounis (1995)

...just... wow. :-(

Good, he mentions a few real problems, for instance the fact that hundreds of theologians think they're using a classical pronunciation when in fact they can't understand each other; the fact that it's often implicitly assumed that all sound changes happened, if not all at the same time, then still all after every book of the New Testament was written; and a few more.

Also, Caragounis does not claim that, say, Homer used exactly modern pronunciation. He does allow for a few sound changes, he just puts them way earlier than anyone else (and denies the rest).

He doesn't always immediately jump to conclusions either; sometimes he explains potential caveats to his own ideas.

Finally, sometimes his confusion just highlights the confusion of other people: scholars of Greek, and Indo-Europeanists, use the term "pitch accent" in a completely different way than phonologists! The latter use it for languages where the stressed syllable of every word has one of at least two phonemic tones, while the pitches of all other syllables are predictable. Ancient Greek may marginally count as one (the difference between acute and circumflex was phonemic, see below, but only on long vowels and diphthongs), but not because its stressed syllables had higher pitch than the unstressed ones. Swedish, Japanese and Shanghainese are pitch-accent languages.

However, he's a theologian, not a linguist, and it shows. It shows again and again and again. He has no clue of phonetics, not knowing there's a difference between diphthongs and vowel clusters (guess why!), and even going so far as to *headdesk* assume in his explanation of what [γ] is that [j] is literally a sequence of [γ] and [i] (you have 3 guesses about why he believes so, and the first 2 don't count). He has no clue about the existence of aspirated consonants in the world; I wonder what his English sounds like, and whether he has ever noticed. He flatly denies that vowel length was ever phonemic, failing to mention any evidence other than the claim that it's "natural" for Greek to pronounce all vowels with the same length. *headdesk* He claims the letter digamma was pronounced [v], mentioning neither evidence nor the very existence of the usual opinion that it was [w] instead (...which would totally trounce the point he's trying to make), and he does not seem to know that the Latin v was pronounced [w] well into Classical times either. He implies that Latin transcriptions of Greek are based on the very earliest spelling conventions of Greek, failing to take into account all the evidence that the Romans had no respect for foreign spellings and only wanted to indicate the pronunciation: ει turns into i, never ei, ου turns into u, never ou – indeed, in graffiti, φ changes ph to f in the 2nd century, as the article says, and this completely trounces his point. In Pompeii there's a mention of a Pilipphus; clearly, the author of that graffito, native speaker of a language (whether Latin, Oscan or both) with a /f/ but without aspirated consonants, simply had no idea where to put the aspiration and ended up putting it on the /p/ that was already "strengthened" because it was lengthened! (Sorry, forgot the citation.) He has never heard of assimilatory voicing, nor of the spelling rule that does the same for aspiration in Ancient Greek and is ignored in plenty of inscriptions, and instead implicitly takes assimilatory fricativization for granted several times... even in front of λ, which clearly never was a fricative in the history of Greek. Oops, turns out he even says "sounded voiced" (endnote 69) when he means "pronounced as a fricative"; he must have been misunderstanding everyone else, then. *headdesk* He resorts to an unevidenced claim of "physiologically easier" at one point. He *headdesk* uses Nietzsche as an example of a word with an incredible consonant cluster, having evidently no clue that the five letters tzsch stand for nothing more than [t͡ʃ]! (The z is a 16th-century affectation. – 3 of the 4 additional examples in endnote 75 only work in the few remaining non-rhotic dialects, the 4th relies on counting the phoneme /t͡s/ as 2 consonants, and all have a morpheme boundary through the cluster, except the last, which has two.) His discussion of ζ is much, much shorter than in this article, and incomparably much more simple-minded. He claims in all seriousness *headdesk* that the circumflex was always purely etymological, not knowing that there are several Ancient Greek texts in which the authors make fun of overenthusiastic rhetors who end up pronouncing one word instead of another by getting the accent wrong! In the same place, he claims that a rising-and-falling tone on a single syllable is "an impossibility in actual speech", not knowing that several languages (let's try the Kam language) have just such a thing as a phonemic tone, and that others (like Mandarin) have a falling-and-rising phonemic tone. (I'm not saying that the circumflex was ever pronounced that way. I'm just saying that Caragounis' argument is completely ignorant.) It doesn't occur to him that the Hebrew waw (even though that's what he calls it!) was ever pronounced [w]. In endnote 70, he wants to have -νδρ- pronounced [nðr], a particularly stupid hypercorrectivism of Katharévusa that has, I bet, never existed in less artificial varieties of Greek (even Spanish doesn't do such nonsense).

Given all this, it's a bit bizarre that he uses full polytonic spelling for Modern Greek... in 1995.

Speaking of spelling: when he talks about the history of the Greek alphabet, it's practically all about Athens, Athens, Athens. When he cites inscriptions, he practically never mentions where they're from, leaving all that to his sources.

The winning quote (p. 179, or 19 of the pdf): "No unexpressed sound can have objective existence in a language!" The context makes obvious that "unexpressed" means "unexpressed in writing". The statement, thus, is of such unbelievable stupidity that I don't even need to explain why.

In some cases, Wikipedia knows more about Modern Greek dialects than he does, and again it shows – painfully. In Tsakonian, ω has merged into ου, not into ο. In Pontic, IIRC, η has merged into ε, not into ι. In a whole bunch of central dialects, including Tsakonian, υ has merged into ου, not into ι, except that it remains distinct from both behind consonants that can be palatalized (so that λυ has turned into [lʲu], neither [lu] nor [lʲi]).

He pays extremely little attention to the dialect diversity of Ancient Greek, too. Oh, in endnote 111 he manages to mention their existence.

There are several examples of the all-or-nothing fallacy (if comparative linguistics or Latin transcriptions or whatever aren't absolutely 100 % reliable, they are utterly useless and must be completely ignored). Sure, sheep don't make [vi], but because descriptions of animal sounds are never completely accurate, he simply declares βῆ βῆ irrelevant! What is he, Borg? (Hint: where I come from, sheep are claimed to make mäh, because m is voiced while b is not in Upper German dialects except those southernmost ones with heavy Slovene influence.)

(One of the German examples is misspelled, with c instead of z, though not in a way that changes the pronunciation. In a German quote in an endnote, he leaves a grammatical ending off that is not silent but is clear from context. And a third quote strikes me as self-ironic, but he reads it straight and is outraged.)

He quotes "Allen's Vox Graeca" on Allen's own difficulties in distinguishing the aspirated from the unaspirated voiceless plosives in hearing and speaking, and tries to make some major point of it. Come on. What's going on here is that English lacks such a distinction, having just one series of weakly aspirated plosives instead. Learn Hindi or Thai or, probably best, Icelandic and then come back!

On p. 185 (24/25 of the pdf), he makes a complete U-turn, abandons all his arguments about ancient texts, and claims we can't know anything at all ever and therefore have to resort to the Modern Greek pronunciation. ...like... seriously? Classical Latin, let alone Old Latin, isn't Standard Italian ( = 14th century Tuscan), let alone modern Tuscan, either. (Incidentally, modern Tuscan famously has [ɸ θ x].)

In endnote 32 he talks about "a thicker, rougher b-sound". Dude, learn some phonetics and then come back, because then we might understand you.

Endnote 98: "Is it really credible that Greeks would have sounded all their circumflexed words as words expressing astonishment?" That's exactly how phonemic tone works. Learn Mandarin.

Endnote 108, finally, completely confuses spellings and sounds, and further confuses one Greek, French and Italian phoneme with two English, German and Dutch ones.

Outside of theology (I guess), he's just another half-educated crackpot with delusions of knowledge and understanding. He should be pitied, not cited as a source; and from how he cites Jannaris, Jannaris should be treated the same way! I'm not removing the citation, because I don't know if there's anybody left who makes defensible arguments that would support any of his claims. But I... strongly doubt there's anyone. Someone please look for one.

David Marjanović (talk) 00:38, 15 March 2013 (UTC)