I’ve been learning Greek for a little while and had always assumed that the grave accent was not to be pronounced at all, because, as my textbook literally explains, “it is there only to indicate the deletion of an otherwise acute accent”.
However I’ve read on Wikipedia that no one knows for sure how the grave accent is to be pronounced, and that some scholars believe it should be pronounced as a weakened acute pitch.
I am a bit surprised by such diverging accounts and was hoping that someone could help me figure this out. I’d like to try and record myself reading out my Greek vocab to aid memorization (as @Andriko judiciously suggested in another thread) but I just don’t know what to do with all those grave accents.
Hi, this all depends on your pronunciation system really.
I think those who use a stress system just ignore it.
If you follow a pitch system (and many do not), a good place to start would be section 31 of Probert’s New short guide to the accentuation of ancient Greek. Probert discusses briefly the history of the dispute over the pronunciation of the grave accent, its use (or absence) in papyri and manuscripts etc., and concludes cautiously that it is “likely that the grave that replaces the acute on a final syllable indicates that the syllable is to be pronounced on a pitch higher than that of the preceding unaccented syllables but lower than that of other accented syllables”.
How you apply this depends on how you pronounce pitch more generally. I follow Devine & Stephens which is tricky but very generally this is how it would come out. Take Iliad 1.3:
πολλὰς δ᾽ ἰφθίμους ψυχὰς Ἄϊδι προΐαψεν
There are three non-grave accents, which represent the pitch peaks:
The first four syllables rise progressively in pitch (straight through the first grave), stepping up pitch by pitch, to the peak at the end πολλὰς δ᾽ ἰφθί, then you have a sharp pitch drop on μους,
then you start rising progressively through the next three syllables (straight through the second grave) to the peak at the end ψυχὰς Ἄ, then you have a sharp drop on the next syllable and a lighter drop after it (ϊδι),
and so on.
I’m quite happy to say that this is very speculative and could be very wrong. It just doesn’t matter. I wouldn’t say that one system is better than another, and so if you really want to know how to pronounce it, I’d suggest you simply settle on a pronunciation system that aligns with your mode of learning Greek (solitary? in class etc.?) and go with the flow. I have never actually spoken with anyone in classics (teachers, students etc.) in real life, and so can pronounce it however I like with no damage to anyone
I guess I use a pitch system as my pitch goes up and down, a bit like in Chinese.
I didn’t check but this is probably the source used in the Wikipedia article on ancient Greek pronunciation, which states exactly that.
How come there is no drop after the first grave in πολλὰς (that’s how I would intuitively pronounce this at the moment, for want of more specific instructions)?
Just to be sure, in words like τὴν, τὰς, τοὺς , would you pronounce the accented long vowel in two stages, i.e. e-é, a-á and u-ú (as it were) where the beginning is at baseline and the lengthening goes upwards, basically like the second tone in Mandarin? I’ve heard quite a few recordings where people just pronounce the entire length of the vowel on a steady high note, which I find suspicious but I can’t really be sure what’s right and what’s not.
I agree. I don’t really mind if my pronunciation is not the proven historical one. I mainly worry about this simply because pronouncing pitch is the best way for me to memorize it, as I’m more auditive than visual, so I’d like to do it at least approximately right.
Hi, for your first question on the grave in πολλὰς, that’s the key difference between a grave and an acute: check out Devine & Stephens, who describe what happens in the pitch of the syllable after these accents.
With your second question on the accent of articles, that opens a whole separate series of problems. Non-lexical words like articles can have different accentuation (both orthographical conventions and pitch) to lexical words like πολλὰς (Devine & Stephens talk about that too).
Once again, if you stick with your use case (and follow any associated community conventions) you should be doing fine; as this is largely speculative, I suggest find your system and get back into the Greek itself!
Would you have a specific article in mind? I’ve just googled them (I have no idea who they are, I’m really just a beginner and all I have at home is my beginner textbook) and they seem to have written a very large number of articles. I assume that Prosody of Greek Speech would be an interesting book but it doesn’t seem to be available online.
I practice vocabulary using tonal accents, but it’s tough to thoroughly integrate the tones into my brain.
(1) My brain doesn’t want to consider tones as salient, so as my memory of a word fades, the first thing to fade is distinctions like acute versus circumflex. I took a semester of Chinese once, and it was the same way. The first thing I lost from my memory was the memory of what tone a word was.
(2) Unlike Chinese, ancient Greek doesn’t give us access to native speakers whom we can hear doing the tones naturally.
(3) I find that when reading an actual sentence out loud, it gets really difficult to to put it all together while also reading expressively. It’s really tough to improvise a melody like a jazz musician, so that the sentence as a whole rises and falls while the the small-scale structure of the words’ pitch accents is also present. But if you just assemble a sentence out of accented words without the large-scale structure, it sounds robotic, like you’re just ping-ponging back and forth between the two pitches.
You can, for example, find recordings of people reciting Homer online. However, many of them actually do the robotic ping-ponging, which I can’t stand to listen to, and in addition there’s the problem that different people use different pronunciation systems, so a given recording won’t necessarily be compatible with what I’m using.
I used to be fluent in Chinese, and I honestly don’t think there is any marked large-scale structure to sentences. Most of what Westerners call intonation manifests in Mandarin mostly as pitch variations when applying the standard tones, like jumping almost an entire octave up and down when pronouncing stuff like 不是 to express disbelief or disagreement or simply making everything more high-pitched, coupled with modulations in speed. I may have totally missed it if it is actually there, but I have never noticed (nor heard of) any significant larger-scale rising and falling that would apply to entire sentences (aside from the probably universal slight lowering of general pitch towards the end of a sentence).
Besides Chinese does sound robotic to many people. I wonder if ancient Greek might not have sounded just like that. Or do we have evidence to the contrary?
I felt somewhat compelled when I started learning Greek to decide upon the system I should use, for the same reason you are now looking for one. The way I see (and try to apply) accents is explained here: https://vasilestancu.ro/accent_en.html. The same website also contains some audio samples. Just like you, I need to have a consistent audio reflection of what I read if I am to memorize it.
Hi, I’ve been learning Greek for over five years and I’m also learning how to pronounce Greek correctly. I chose the Erasmus pronunciation at the beginning and pronounce the accents only as stresses. I don’t know what you choose, but I think it’s important to read texts aloud with the correct intonation, like a real text, preferably as if we were reading it to small children as a bedtime story. It’s not good to read it like a lot of people on the internet who try to pronounce many things correctly and forget the content of the text, or perhaps don’t even understand it. As bcrowell writes, I also think it makes it impossible to listen to at all.
Paveln, if you’re using Erasmian, then doesn’t it make it hard for you to use the recordings by Stratakis that use reconstructed pronunciation?
The Ellenizometha dialogs sound really nice and natural. Unfortunately my listening comprehension isn’t good enough to tell what they’re saying. If they were on youtube I’d probably try slowing them down, but there doesn’t seem to be any way to do that on their spotify-based site.
Hi bcrowell, I understand the Stratakis thing very well, because it’s almost like an Erasmian pronunciation in my opinion.
I’m learning Greek as a normal living language and my teacher and I speak Greek for two hours every week, similar to the Ellenizometha dialogues, so I understand those dialogues very well. When COVID started two and a half years ago, my teacher and I started meeting in the park because everything was closed, and then later in the shopping mall. Gradually we learned to speak quite well in those two and a half years
Thank you for the link. I must say those explanations diverge quite drastically from those in my French textbook.
I’ll try to include images here, not sure if that will work.
I don’t know if you can read French, but it says that the circumflex accent consists of starting the accented syllable with a high pitch and lowering it (vs. rising and falling pitch in your link), and that the acute is just a higher pitch on the accented vowel and you immediately go back to baseline pitch on the next one.
Vasile’s web page does say straightforwardly that “The explanations below are not the result of any rigorous linguistic study, but a mere personal opinion on how the rules found in grammar books could be applied in practice by the student.” Experts have tried to reconstruct some kind of correct or accurate intonation, but they don’t all agree. Some info:
My notes from ch. 4 of Devine, Prosody of Greek speech:
strophic means AAA song form
"nonstrophic music respects the accents while strophic music does not"
Dionysius of Halicarnassus: unclear passage probably says that the range of pitch in speech never exceeds a 5th; this is reasonable compared to modern languages
"in Greek, the turning point between falling and rising tone coincides with the word boundary"
"So the Greek grave can be viewed as a sort of compromise between the accentual High and the postaccentual Low or between the accentual High and the word final Low."
in music, the acute averages to about a wholetone, while the grave is about a third of a tone on the average (statistically significant)
"low tones as well as high tones fall as the utterance progresses due to the effects of catathesis and declination"
me: I think they're saying that the standard contour for a word is mid-high-low; the high is fairly fixed, while mid and low show more variation below that when the word is long
lexical forms of words tend to be different from forms in connected speech
When I’m practicing vocabulary words in isolation, I do pretty much what your French textbook says. But connected speech is different. The pronunciation of the grave only comes up in connected speech.
Here’s a sample of what I’ve been doing. The image shows some nonsense words and approximate notation for the musical pitches. The link below it is to a recording I made of myself.
One thing to watch out for about the French textbook presentation shown by mikomasr is that it’s shown as if they’re actually notating some kind of approximation to the pitches using real musical pitches. But they notate the interval as a perfect fifth, which IMO cannot possibly be right. If you tried to do that at a real speaking tempo, it would sound like yodeling. I assume they didn’t mean it to be construed as actual pitches, but then in that case I don’t understand why they didn’t just choose some other interval that would be more realistic.
Perhaps the extract is from J. V. Vernhes’ ἕρμαιον (I do not have the book with me now, and your reference just reminds me of this manual which I used some time ago). Just in case, there are some more detailed explanations of J. V. Vernhes in his article “JEAN-VICTOR VERNHES : QU’EST-CE QUE L’ ACCENT GREC ANCIEN?” here: https://ch.hypotheses.org/3623.
I am not in the position to recommend any pronunciation system or technique regarding the Greek accents, as I am no specialist in this field, but just a passionate student of Ancient Greek. In my opinion, the explanations given in Ancient Greek manuals are in most cases insufficient for an autodidact to really find a way to put in practice the rules and suggestions given: one needs more detailed technical explanations, ideally accompanied by audio samples, in the manner Ben just did in his post. May I have some personal observations:
It would be quite artificial (if not practically impossible) in real speech to “attack” an accented syllable in the highest tone of the pitch, unless that syllable is the one that opens a clause or an isolated word. Therefore, since the accent comes after some sounds uttered at normal tone, there must be a rising form a “zero” to some higher level; this cannot be sharp vertical, unless you use some trick like a glottal stop, or an imaginary full stop, etc.
My impression is that all languages I have come in contact with use pitch accents; I may be wrong, but this is what my ear perceives. If pure stress accent is used (more intensity on the accented syllable), that would sound to me like a machine/robot speech. By the way, I would appreciate it very much if somebody could prepare some audio sample (or indicate some reference) where the two systems are emphasized and compared, pitch vs. stress. If I read the English text below (cited from Ben’s post), it sounds to me as if it were accented like I tried, using the three types of Greek accent. I may be wrong, as English is foreign language to me, but it really sounds like that to me.
"Éxperts have triêd to reconstrùct sòme kìnd of corrèct or áccurate intonâtion, but they dôn’t âll agreé."
The three types of accent have different grades of liberty: the grave can only stand on an ultima, etc. Why is that? And why do we count the syllables backwards? I think because it is important how we finalize the uttering of a word.
The grave can only stand on an ultima because it should mark the end of the word. Only graphically? no effect on the speech? I doubt it: we can see the end of the word just as we look at it.
If the acute consumes its effect on one syllable only, why is it that the circumflex, which does the same, cannot stand on an antepenult? My guess is that the accute must extend its effect on two syllables, leaving thus enough “energy” to properly pronounce the last part of the word.
And yes, the amount of variance in pitch accent across a small island nation like Japan should sap some of our confidence in our knowledge of the pitch patterns of the ancient Greek language simply gathered by the descriptions and writing system of the much later language.
I think it’s true that languages in general (or at least the languages I’m familiar with) use pitch variations, but that isn’t the same as a tonal accent that differentiates one word from another.
I think you probably do have a point in that when people try to vary stress, they also tend to vary pitch, and vice versa, and neither speakers nor listeners really know that that’s happening at any conscious level. Listening to my recordings, I’m confident that one emphasizes pitch more, and the other emphasizes stress more, but I found it impossible to do the stress without any correlated pitch variations. I think in order to do that, I would have had to shift into a singing mode where I was intoning a single note.
I remember seeing scientific data on singers and people playing wind instruments that showed that although performers in Western traditions who used vibrato thought they were doing a pure pitch modulation, they were also doing an amplitude modulation at the same time. But that doesn’t mean that they’re the same thing, or can never be at all independent. For instance, one way to recognize when you’re being assaulted by the loathsomeness of Kenny G is that he uses a different type of vibrato that is more based on stress.
Yes, that’s the one. The pictures I took are from the first pages in the manual where he explains how to pronounce accents.
Thanks for the link! I’ll go read it right away.
one needs more detailed technical explanations, ideally accompanied by audio samples, in the manner Ben just did in his post.
Actually the funny thing is that Vernhes has done a few recordings, but they sound nothing like what he describes in the pictures I posted. So either he doesn’t have a very good ear (which is a distinct possibility for anyone well versed in classics, just like the rest of the population) or the explanations given in his manual do not reflect his actual knowledge of accents.
. It would be quite artificial (if not practically impossible) in real speech to “attack” an accented syllable in the highest tone of the pitch, unless that syllable is the one that opens a clause or an isolated word. Therefore, since the accent comes after some sounds uttered at normal tone, there must be a rising form a “zero” to some higher level; this cannot be sharp vertical, unless you use some trick like a glottal stop, or an imaginary full stop, etc.
I’m not entirely sure I understood what you’re saying, but based (again, sorry) on Mandarin, I don’t see why it would be impossible to have any syllable starting on ’top’ of the pitch curve. That’s the definition of Mandarin’s first and second tones.
“Éxperts have triêd to reconstrùct sòme kìnd of corrèct or áccurate intonâtion, but they dôn’t âll agreé.”
That’s a fun way of illustrating all this.
I too hear a circumflex on “tried” where you go up and down on the vowel. For intonation, I rather hear an acute accent but I may be completely wrong if my understanding of the acute accent is incorrect to begin with.
