This example occurs in Iliad 1.33. In running speech, when there are no pauses between words, I'm able to articulate this as "edeisend ho." However, I would imagine (possibly just because I'm ignorant) that part of the definition of a "word" in linguistics would be that one can choose to pause between words without making one's speech incorrect. Personally, I am not able to articulate "edeisen dho." All the same issues seem to arise with τέ as with δέ.

My guess would be that one of the following is true:

  1. The elision only exists in running speech, so a speaker who chose to pause after ἔδεισεν would have to say "ἔδεισεν δέ ὁ γέρων." This messes up the meter, so when actually performing this line, one could introduce a principal caesura at this point (not that it's logical), but it would then be necessary to add a "pick-up" syllable after the pause.

  2. Greeks could pronounce word-initial consonant clusters that I can't pronounce, and this is one of them.

  3. The postpositions τέ and δέ are very strongly bound to the word before it, like "'s" in "Bob's." Writing a space in between is a spelling convention, but it really acts more like the ζε in Ἀθήναζε. You just can't pause after ἔδεισεν.

  4. Even if we don't pause between words, the aspiration following this elision is dropped, so although we still spell the word ὁ, really this phrase is properly pronounced "edeisendogeron."

  5. Similar to 4, but the pronunciation becomes something else, maybe θο γέρων.

Re #5, searching on the text of the Iliad shows that although it's quite common to have stuff like τ' ἀνδρῶν and τε ἀρνῶν, one never sees τ' followed by a vowel with rough breathing, which would presumably have resulted in a θ pronunciation.

Do we have any way of knowing which of these, if any, is true? Since δέ and τέ are extinct, I guess we can't tell based on modern speech. Since they're postpositions, there are no cases where they're the first word of a sentence. Searching through the text of the Iliad, I also don't find any cases where δέ or τέ is the first word on a line (which would seem to support possibility 3 above, since it was probably normal to pause at the end of lines).

As a side note, it seems that δ and τ are never word-final in Greek, except in κάδ, a synonym for κατά, and this seems to occur only in "κὰδ δέ."

2 Answers 2


Phonetically, there is little doubt that this sequence would have been pronounced [-endho-].

To take your possible answers in reverse order:

5 can be ruled out because if this was the pronunciation it would presumably be spelled as such, θ᾽ ὁ. Note that this aspiration is in fact what happens to τε before ὁ and is thus written (θ᾽ ὁ = τε ὁ), but this never happens with δέ, though the Greeks were usually pretty good about writing things phonetically when they could.

4 can be ruled out for the same reason -- it isn't written ὀ but ὁ. Here things are a bit trickier, though, because the East Ionic dialect, which was the main basis for the Homeric dialect, was psilotic -- that is, it lost the sound [h] completely, so that the definite article we know as ὁ was always pronounced [o]. However, the text of Homer as we have it, though it contains occasional examples of psilosis, is not generally psilotic; this is shown for example by the above-mentioned elision τε ὁ -> θ᾽ ὁ, since with psilosis this would instead be elided to τ᾽ ὁ (or more correctly τ᾽ ὀ). If an East Ionic-speaking bard were reading this verse, he would indeed have pronounced [-endo-]. But in the Attic recension as we have it, the aspiration was clearly there. (That said, it may well have sometimes dropped out in normal speech, as weak sounds often do. But it would be there for someone who was enunciating clearly.)

Your other three options have to do with the question of word boundaries or of where a Greek speaker would pause in this sequence, if they had to. This is harder to answer both because of scantness of evidence and because there is no straightforward definition of what a "word" is. One thing to note is that δέ has its own accent, which means it's at least somewhat prosodically independent of what precedes it; in this respect it actually differs from τε, which is an enclitic, i.e. is part of the same accentual unit or "prosodic word" as what precedes it.

If a Greek wanted to say the words ἔδεισεν δ᾽ ὁ γέρων but introduce one pause, where would they do it? I think they would most likely remove the elision and say ἔδεισεν δὲ | ὁ γέρων. This is because both the options ἔδεισεν | δ᾽ ὁ γέρων and ἔδεισεν δ᾽ | ὁ γέρων both contain sequences that conflict with the phonotactics of Greek words: there's no Greek word that ends in [nd] or begins with [dh]. But it's hard to be sure. It may well be that a detailed analysis of metrical patterns involving postpositives, or of musical settings, would shed some light on this question; I suspect there's some relevant information in Devine and Stephens' The Prosody of Greek Speech, which I don't have handy.


Turns out this was me just being stupid as the key assumption that I state doesn't hold, i.e. the pronunciation of consonant's has remained the same between ancient and modern Greek.

Thanks to the commenters for setting me straight. Just ignore everything below.

(Not deleting in case someone else ever arrives here with the same assumptions)

Greek here but not a proper scholar of Ancient Greek. If it helps the h in ho isn't really pronounced, nothing like the ch in loch.

As is presented, the most natural place to pause for aspiration seems after ἔδεισεν and I wouldn't separate δ' ὁ. In this case, δ' ὁ would sound a bit like the beginning of the but with a different vowel. Please note that this is a completely different sound to θ (which is the sound in theatre, thespian, theory etc.)

δε is not completely extinct, still used following μεν for a phrase that is similar to the structure "on one hand (μεν), on the other hand (δε)", but of course as a single syllable word it has lost its accent marks in modern Greek. Other usages of δε in modern Greek is as a variation of δεν which negates (like "not"), as a way to connect another element to the previously reported usually adding some emphasis (similar somewhat to "furthermore" but different syntax) and can also work to add emphasis when coupled with conjunctions (can't readily think of an equivalent in English right now, apologies). With the exception of the use of negation (variation on δε) all the other uses stem from Ancient Greek and sound somewhat formal and maybe even old-fashioned to modern Greek speakers.

Further, the consonant sounds, to the best of my knowledge, have remained exactly the same from ancient to modern Greek, which is what I base on my educated guess.

That said, if needed, it wouldn't seem completely out of place if a speaker just removed the elision.

Hope this all helps somewhat. :)

  • 1
    The languages that do have a H have a very different sound in comparison to the CH in loch. Modern greek may be missing the former so perhaps H sounds like a CH to you, but even in English H in humble is very diferent from CH in loch. In German hundert vs. dach. There are voiced and voiceless variants, but I think they are normally allophones. For example, Russians tens to pronounce ch where h is in order in foreign languages. Dec 10, 2021 at 11:38
  • 2
    Further, the consonant sounds, to the best of my knowledge, have remained exactly the same from ancient to modern Greek, which is what I base on my educated guess. No, they've changed quite a bit.
    – user3597
    Dec 10, 2021 at 14:07
  • 1
    In Modern Greek δ does indeed sound like the th in the, but this wasn't the case in Ancient Greek, where it was pronounced [d].
    – TKR
    Dec 10, 2021 at 18:32
  • 1
    @TKR and Ben Crowell: thanks for setting me straight.
    – gkampolis
    Dec 14, 2021 at 15:54
  • @VladimirF that's very interesting! To my ear, I don't hear much of a difference between huble and loch (other than people tending to "stretch" the ch in loch a bit)
    – gkampolis
    Dec 14, 2021 at 15:54

Your Answer

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