In all the textbooks I have seen, the list of Ancient Greek diphthongs are more or less the following:

  • αι: /ai̯/
    • ᾳ: /aːi̯/
  • αυ: /au̯/
    • ᾱυ /aːu̯/
  • ει: /ei̯/ > /eː/
    • ῃ: /ɛːi̯/
  • ευ: /eu̯/
    • ηυ: /ɛːu̯/
  • οι: /oi̯/
    • ῳ: /ɔːi̯/
  • ου: /ou̯/ > /uː/
    • ωυ: /ɔːu̯/
  • υι: /yi̯/ > /yː/

(IPA from Wikipedia: ‘Ancient Greek Phonology’: ‘Diphthongs’. List from the same article and compared with Clyde Pharr / John Wright: Homeric Greek : A Book for Beginners. In Hansen and Quinn’s Greek : An Intensive Course, they specifically state (p. 4) that what Pharr/Wright list as improper diphthongs (long vowel plus short vowel, pronounced as diphthongs in Homeric time (§ 506)) are pronounced as monophthongs by the time of Ancient Greek.*)

From this list I notice that ι cannot be initial in a diphthong. Pharr/Wright notes that ι, υ and ϝ are semi-vowels. Latin, though, included diphthongs with both initial /i/ and /u/, such as Iᴠ́ʟɪᴠs [ˈi̯uːlʲiʊs̠] and ᴠɪ́ɴᴠᴍ [ˈu̯iːnʊ̃ˑ]. (The prior even includes i for both the diphthong and the syllabic pure vowel.)

Why then is it that initial /i-/ cannot create a diphthong in Ancient Greek? Did it at some point exist and get monophthongised? I am sorry that I cannot provide any more background for my question, but my knowledge of Greek is horribly poor.

* I understand that Ancient Greek is considered the time period following Homeric Greek; if this is incorrect, I will be very happy to be educated on the matter.

  • I think most authorities state that some long diphthongs were viable during most of Ancient/Classical Greek. According to Sydney Allen, the changes to monophthongs started at least by the 4th century BC (with ῃ) but were not complete until the 2nd century BC during Koine times. Τhe change can be seen in the different treatment in Latin borrowings of Greek words including ᾠδή. Early borrowings (tragoedia and comoedia) show the diphthong. Later borrowings (rhapsodia and melodia) do not. Commented Apr 7, 2022 at 17:38

1 Answer 1


Latin had a phonemic /j/ and /w/ separate from the vowels /i/ and /u/, hence contrasts like Julius vs Iulus. The short answer is that "Ancient" Greek—as in Classical Attic—simply didn't. Therefore initial iota is always /i/, never /j/: Ἰάνθη has a sequence of two vowels like Iulus, not a consonant and a vowel like Julius.

At one point, the ancestor of Classical Attic did have these sounds. /w/ survived quite a long time in some dialects (though not Attic), to the point that it has a well-recognized letter to represent it: ϝ. And indeed, scansion suggests that Homer (or more precisely one of the people involved in composing and redacting the Homeric epics) pronounced it in words like ϝέπος /wépos/, later ἔπος. /j/ didn't survive to the alphabetic stage, but we see it attested in Mycenaean: Linear B jo-qi /jókkʷi/ for Classical ὅτι. But Attic lost both of these sounds fairly early, and thus has no initial semivowels.

But what about the diphthongs like αι and αυ, you ask? Well, phonologically, they seem to have been treated as single segments rather than sequences of two segments. The second half of αι wasn't affected by the loss of /j/, for example, and the second half of αυ wasn't affected by the loss of /w/—or, for that matter, by the shift of /u/ to /y/. It seems best to analyze these diphthongs as single atomic units rather than sequences of two vowels, or of a vowel and a semivowel.

  • 2
    This comment does appear to have a very strong answer within it, though in the obviously condensed form of a comment, it is—let us agree—quite compact. May I suggest you elaborate on this and provide us with another answer? The more the merrier, after all!
    – Canned Man
    Commented Apr 5, 2022 at 17:46
  • 4
    @Unbrutal_Russian Sure, you can analyze it as a conditioned change rather than an actual segmental difference. But then how do you deal with αϋ αϊ etc? There needs to be some difference between αι and αϊ on the segmental level—so you need a difference between /ai/ and /aj/ either way. And at that point I don't see a major difference between analyzing /aj/ as a single unit and analyzing /j/ as a special phoneme that can only ever appear after /a/, /o/, and /y/.
    – Draconis
    Commented Apr 5, 2022 at 18:35
  • 2
    @Unbrutal_Russian If you analyze αυ as /av/ (short vowel + consonant), how do you account for the fact that it can be circumflexed?
    – TKR
    Commented Apr 6, 2022 at 8:42
  • 3
    @Unbrutal_Russian It's true that you can describe the facts by saying that Greek has consonants /w j/ that have a very limited distribution and differ from all other consonants in that they can bear tone. But I'm curious why you prefer that to the simpler conventional analysis of these sequences as diphthongs.
    – TKR
    Commented Apr 6, 2022 at 13:27
  • 2
    @Unbrutal_Russian Fair enough, but presumably that analysis would extend to the unstressed instances too.
    – TKR
    Commented Apr 6, 2022 at 16:56

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.