This article is from the Greece FAQ, by Nikolaos (Nick) C. Fotis, firstname.lastname@example.org with numerous contributions by others.
From: email@example.com (Dimitrios FILIPPOU)
[ Regarding Allen's book ; added R. Wallace's <firstname.lastname@example.org> and
Stavros Macrakis' <email@example.com> remarks to this message. R.Wallace's
comments are prefaced with RW>, while Stavros' are prefaced with
SM>; I hope these aren't too hard to follow -- nfotis ]:
SM> Below, some comments on your notes. But the basic questions you don't
SM> address are: why would the ancients bother to invent six different
SM> ways of writing the sound "i"? And two different ways of writing "e"
SM> or "o"? And sometimes doubling consonants, and sometimes not? And
SM> how do you explain the structure of ancient poetry without referring
SM> to syllable quantity, which depends on vowel length?
SM> The other issues (pronunciation of gamma as hard g or as gh, etc.) are
SM> less important, because they don't change the STRUCTURE of the system.
SM> In fact, I think it would make sense -- at least for teaching in
SM> Greece -- to preserve the modern pronunciation for them. Keeping
SM> distinct pronunciations for the diphthongs and long vowels (eta,
SM> omega), on the other hand, would preserve the ancient structure and
SM> seems important.
First, let's see some *facts*:
1. The system of (Ancient and Modern) Greek writting -- as
we know it today -- has been developped by the Alexan-
drian and (mostly) by the Byzantine grammarians. For
example, it is the Byzantines who introduced the small
Greek letters around the 9th c. AD.
RW> This is true, but the writing system is immaterial. There is a good deal
RW> of inscription material from the 5th and 4th centuries BC, and arguments
RW> from orthography are based on that.
SM> Although the system of diacritics (accents, breathings, iota
SM> subscript) was introduced by the Alexandrines, the consonants and
SM> vowels were around long before that! As for small letters, I don't
SM> see how that affects pronunciation.
2. There's not much -- if any at all -- difference between
the Byzantine (after, say, 4-5th c. AD) pronounciation
and the Modern Greek pronounciation. According to Allen
himself, changes from the Ancient Greek pronounciation
(i.e., Allen's version) to the Byzantine/Modern Greek
pronounciation may have come as early as in the first
century of Roman occupation of Greece (2nd c. BC).
RW> Quite right. Spelling mistakes in 2nd century AD papyri seem to show that
RW> substantial changes in pronunciation had taken place, while spelling (as
RW> often) remained more conservative. I vividly remember the first time I
RW> was faced with the text of a papyrus letter from this period, written by
RW> a young man who clearly had not been paying attention to his teachers at
RW> school! It didn't look like Greek at all! Then I pronounced it in the
RW> Modern Greek way, and it all became clear. But the fact that the pattern
RW> of variation in spelling is quite different from that of the 4th or 5th
RW> centuries is evidence that pronunciation had changed (as you would
RW> expect it to over such a period).
3. How the Ancient Greeks (here, we are talking about the
Attic dialect, 5-4th c. BC) were pronouncing certain
letters, diphthongs, etc. is and -- I think -- will
remain an unsolved problem.
SM> Of course, the _exact_ pronunciation will never be known, but there is
SM> lots of evidence to help us get a good idea. Modern pronunciation is
SM> one kind of evidence.
On the one hand, we have the bleating of the sheep in
Aristophanes which is written as:
beta-eta (w. acc. circ.) -- beta -eta (w. acc. circ.)
In Modern Greek pronounciation this reads: "vee-vee",
when common logic suggests that it should be read as
"bebe". Therefore, Allen recommends that "beta = `b'
as in `book'"; and "eta ~= epsilon".
RW> There is actually more to the argument than this. Latin, for example,
RW> transliterates beta as a B, and epsilon as an E.
SM> True, we expect sheep to say "be be" and not "vi vi", but there is a
SM> lot of other evidence for these pronunciations. When you say
SM> "therefore", it's as though this is the only evidence!
On the other hand, we have the oracle of Delphoi to the
Athenians, who could not understand whether it meant
that they would suffer from famine ("limos") or from
plague ("loimos") the first year of the Peloponnesian
War (the Athenians' confusion is quoted by Thucydides).
This confusion can be understood only if the Athenians
iota = omicron-iota
as Modern Greeks do. But Allen suggests: "NO! omicron
-iota = o-ee (i.e., a "true" diphthong). (Allen discus-
ses this notorious quote of Thucydides, but, I don't
remember his points -- I returned the damned book :-( )
RW> No. The story requires the pronunciations to be similar, but not
Some other *observations*:
a. Allen accepts the Byzantine/Modern Greek pronounciation
of the accents on the basis of "we don't know enough about
the melodic accent reading of the Ancient Greeks".
SM> I don't think he "accepts" the modern pronunciation as being a good
SM> reconstruction; he simply recommends using it to simplify things,
SM> since the tonal system is not fully understood, and adds a lot of
SM> difficulty to teaching the language. Given what we know of the
SM> Ancient Greek tonal system, by the way, it is more like the Japanese
SM> or Serbian systems than it is like the Chinese system. Foreigners
SM> have trouble learning the Japanese and Serbian systems, and in fact
SM> usually "hear" the tones in those languages as stress patterns at
SM> first. Given that there are no speakers of Ancient Greek, it would
SM> seem unproductive to spend a lot of time teaching this. Conversely,
SM> teaching modern pronunciation would NOT help travellers in Greece make
SM> themselves understood!
Could not this apply also in the way the Greek letters are read?
I.e., once we don't know for sure how the Ancient Greeks
were reading certain vowels, consonants, combinations of
letters, etc., why don't we stick to the closest relative,
the Byzantine/Modern Greek pronounciation?
RW> Some people regard this as a good argument. At least Ancient Greek
RW> pronounced as Modern Greek does sound as if it might be a real language!
RW> The argument against is that the modern pronunciation makes nonsense of
RW> Ancient Greek poetry, and loses much of the sound-play in any ancient
RW> text. Personally, I find this objection compelling, but it is possible
RW> to take a different view. But this is just a question of pedagogic
RW> convenience, and doesn't contribute to the question of how the language
RW> was pronounced. I think Allen is right about accents. It is certain that
RW> the ancient accents were pitch accents (as in Chinese) rather than
RW> stress accents; we know a good deal in theory about how they were
RW> pronounved (the musical interval over which the voice moved on a
RW> particular syllable and so on) but all actual attempts to put it into
RW> pracitice seem unconvincing to me. And students have quite enough hassle
RW> learning the language as it is!
SM> Of course, we don't know "for sure", but we do have a lot of good
SM> evidence, including borrowings, related languages, and the internal
SM> structure of the language and the orthography.
b. If we adopt Allen's recommendations certain sounds will
be excluded from the Ancient Greek pronounciation. Even
if Allen is right in saying that most likely "beta =
English `b'", I find it hard to believe that the Ancient
Greek (more precicely, the Athenians) had not ANY of the
following soft (e.g., fricative) sounds in their
v --> Modern Greek "beta"
y (as in English `young') --> M.G. "gamma"
th (as in English `there') --> M.G. "delta"
th (a in English `theatre') --> M.G. "theta"
All languages that I'm familiar with (Modern Greek,
English and French) have at least some of the above
sounds. Why not Ancient Greek?
SM> !! Actually, the (modern) "gamma" (the gh sound before a/o/u, not the
SM> y sound before i/e), "theta", and "dhelta" sounds are UNCOMMON in the
SM> world's languages. For instance, none of Italian, Japanese, French,
SM> Turkish, Serbian, German, or Hawaiian has any of them.
c. If we adopt Allen's recommendations, we get a pronouncia-
tion full of hiatuses ("hasmodies"). My poor ear suffers
when I try to read loudly by Allen's system words such
aiphnidiazomai (= I get surprised)
chairekakos (= malicious), etc.
SM> Why would you expect Ancient Greek to sound good to your ear? Latin
SM> pronounced according to the historical pronunciation sounds strange to
SM> Italians, too.
as it suffers when I hear my colleagues talking about
"k-eye" (and they mean "chi" = `hee'), or "ps-eye" (and
they mean "psi" = `psee') in Maths.
RW> Yes of course, what do you expect? And the pronunciation of Chaucerian
RW> English sounds weird to me! But we will all agree that Mathematicians
RW> pronounce Greek in a barbarous way!
SM> These are of course incorrect pronunciations according to Allen.
SM> Something like "k-eye" is the pronunciation of "kai" (and), not of the
SM> letter chi.
d. Allen makes a direct attack in the Preface of the latest
edition of his monography, on another Swedish (?) scholar
who dares to say that Attic Greek was pronounced almost
the same as Byzantine/Modern Greek from the 4th c. BC.
This attack -- it's just a dismissal of the Sewde's
position w/out much justification -- has really surprised
me. (I'm not used to such scholar stabbings in the
Prefaces of books!)
In conclusion, I believe that anyone who wants to learn Ancient
Greek, he should better learn to pronounce it the way Byzantines
did and (Modern) Greeks do. In this way, he/she will be
learning at least 50% of the Modern Greek language as well!
RW> If I were teaching a Greek, I might agree. The principal objection to
RW> believing that the modern pronunciation is basically the same as the
RW> ancient pronunciation (apart from the inherent plausibilty of any
RW> language remaining unchanged in pronunciation for two and a half
RW> thousand years, through a period when we know that accentuation,
RW> grammatical structure, and vocabulary did change substantially) is that
RW> it assumes that when the ancients adopted the alphabet they chose a
RW> system which was by no means phonetic (i.e. there are several ways of
RW> representing the same sound). In other words, the ancient greeks were
RW> dotty, which I am unwilling to accept. It is surely more likely that
RW> they initially adopted a system where there was a more or less
RW> one-to-one correspondence between letters and sounds, and that gradually
RW> pronunciation changed while orthography remained the same (as indeed it
RW> has done in many languages, including English and French), leading to
RW> these poor kids in the 2nd century AD getting all their spellings wrong.
RW> That, I think, is where the evidence, but we will always be guessing.
SM> Pronunciation is probably the easiest thing to learn about Modern
SM> Greek if you know Ancient Greek. (Although of course too many
SM> foreigners don't bother!)
PS. I repeat: I'm not a classicist neither a linguist! Just an
"boring/bored" engineer ....:-)