The predecessor to the modern Greek alphabet was the Phoenician alphabet, which had four "sibilant" letters:

  • 𐤆 zayin /z/
  • 𐤎 samekh /s/
  • 𐤑 ṣade /ṣ/
  • 𐤔 šin /š/

According to Jeffery, these turned into the four Greek letters Ϻ Ζ Ξ Σ — but the names, forms, sound values, and alphabetical order got all mixed up in the process. So for example, Σ took its name and sound value from 𐤎, but its shape and place in the alphabet from 𐤔.

Others have suggested that this theory is wrong, and indeed it would be rather odd for all the forms and sound values to get mixed up for no reason. But I've never seen any alternative explanation.

So: what other theories are there on where the Greek sibilant letters came from? And what evidence is there either way (either for or against Jeffery's theory)?

  • Who is Jeffery?
    – vectory
    16 hours ago
  • @vectory Lilian Hamilton Jeffery, The Local Scripts of Archaic Greece
    – Draconis
    13 hours ago

3 Answers 3


Krahmalkov, in his Phoenician-Punic Grammar, suggests (based on data from later Punic) that the "traditional" sound values OP mentions are wrong. Instead:

  • Zayin was pronounced [zd] or [dz], depending on dialect. It was definitely some sort of affricate, because inscriptions often add an extra aleph before words starting with zayin, indicating an epenthetic vowel. The old Phoenician form looked something like .
    • So the sound, form, and alphabet position all line up nicely with Ζ.
  • Šin was originally some unknown fricative, but it had merged into [s] before 500 BCE (the "sibboleth merger"). Inscriptions consistently confuse šin and samekh, and Roman-era puns confirm that it was pronounced the same as Latin s. The old Phoenician form looked something like W.
    • So the sound, form, and alphabet position all line up nicely with Σ.
  • Ṣadhe's pronunciation varied significantly over time, but the earliest available evidence indicates [ts]: Plautus transcribes it inconsistently as either t or ss, and Punic inscriptions in the Latin alphabet sometimes use a T superposed on an S to represent it. The old Phoenician form looked something like 𐌌.
    • So the form and alphabet position line up nicely with Ϻ.
  • Samekh was almost certainly [s]: in later Punic, when the four sounds merged entirely, samekh is the letter most often used by mistake. The old Phoenician form looked something like .
    • So the form and alphabet position line up nicely with Ξ.

The only problems now are the sounds of san and xi, and the names of the letters.

The sound problem is far from insurmountable. Nobody's quite sure what sound san originally had, since no Greek dialect distinguished it from sigma (each dialect generally chose one or the other for /s/ and ignored the other); Etruscan distinguished sigma and san but it inherited them through Greek so it's not too informative. And xi seems to be a case of the Greeks taking a redundant letter and co-opting it for their own purposes, like how ʕayin [ʕ] became Ο [o] and he [h] became Ε [e].

The name problem is harder, and I haven't found any solid explanation for it. Sigma could be a -ma noun from sízō "to hiss" (so literally "a hissing"), xi (like psi) was renamed for its new purpose, and zēta could be by analogy with the following ēta and thēta. But this is all speculation on my part, and for san I have no idea.

TL;DR: Jeffery's assumption of mixing up all the sounds and signs isn't necessary if we reconstruct Punic-esque values for the sibilant phonemes.

  • The name problem is hard, but not insurmountably so, I think. Zēta as you say is a very plausible case of contamination. Sigma becomes a little easier if we remember that g before m sometimes or always stood for a nasal: siŋ-ma could be from sin with an added dummy syllable to give the trochaic shape of most of the other letter names. San conversely would be a clipping. Some ad hoc assumptions are needed but not as many as in the confusion scenario.
    – TKR
    May 15, 2019 at 21:22
  • @TKR Was gamma a nasal before nasals? I thought it was only a nasal before velars.
    – Draconis
    May 15, 2019 at 21:43
  • Also before some or all nasals, see e.g. Sturtevant 64f.
    – TKR
    May 15, 2019 at 22:47
  • Thanks a lot for summarizing Krahmalkov’s research! I promised to do it but there’s too much going on at work now.
    – Alex B.
    May 16, 2019 at 22:04

perhaps not an answer but rather a comment - I just found my notes, based on Woodard 2010/2014.

Adaptation: Greek-Phoenician bilingual speakers (scribes or mercenaries literate in the syllabic Cypriot orthographic tradition), “proceeding with intentionality and arbitrariness” (p. 35).

Consonant clusters with /s/ in Phoenician: three affricates, /dz/ (grapheme: zayin), /ts/ (grapheme: samek), /ts’/ (grapheme: sade).

Consonant clusters with /s/ in Ancient Greek:

enter image description here

The Cypriot syllabary: V, CV are allowed, two CCV (ksa, kse). “It would appear that the value of zeta is somewhat schizophrenic” (Woodard 1997: 162).

*zd > dz happened prior to the alphabetic period (see p. 162 for evidence) In literary Lesbian (Sappho, Alcaeus) zeta is used for *dy; sigma+delta is used for zd.

Woodward 1997: 187"

“The Phoenician character zayin was utilized by the Greek adapters to represent the sound sequence [zd] (the sound of zeta). Syllabic characters with the consonantal value [zd-] existed in the Cypriot script (probably ultimately a feature inherited from the Mycenaean syllabary); consequently, provision was made by these Cypriot adapters for expressing such a sequence with an alphabetic symbol. Semitic zayin represented a voiced fricative [z], but in Cypriot Phoenician this sound appears to have been some sort of "double consonant," thus rendering zayin particularly suggestive for representing the Greek sequence [zd]. Without positing such a syllabic Cypriot background for the alphabetic character zeta, the use of a single alphabetic symbol to represent [z] + [d] (the value consistent with linguistic evidence) is enigmatic.”

  • 1
    Do you know what the evidence is for reconstructing affricates in Phoenician? It seems unusual that it would have sibilant affricates but no sibilant fricatives (except shin).
    – TKR
    Apr 16, 2019 at 20:36
  • 1
    @TKR: As described in Wikipedia, it seems that shin would be reconstructed as [s] in an affricate-heavy reconstruction, which fits as far as I know with the affricate-heavy hypothesis for (the original state of) Semitic in general (discussed in the Wikipedia article for Proto-Semitic)
    – Asteroides
    Apr 17, 2019 at 2:41
  • @TKR e.g. Krahmalkov, Charles R. "Chapter Two: The Alphabet, orthography and phonology". In A Phoenician-Punic Grammar, (Leiden, The Netherlands: Brill, 2001)doi: doi.org/10.1163/9789004294202_003
    – Alex B.
    Apr 17, 2019 at 3:38
  • @AlexB. Thanks -- I can't access the full text, unfortunately.
    – TKR
    Apr 18, 2019 at 0:51
  • 1
    @TKR I found a copy; Krahmalkov says zayin was /zd/ or /dz/, samekh was /s/, tzade was /ts/, and shin was an unknown sibilant that merged into /s/ by Classical Phoenician times. In later Punic they all merged into /s~z/.
    – Draconis
    May 15, 2019 at 16:08

So for example, Σ took its name and sound value from 𐤎, but its shape and place in the alphabet from 𐤔.

Following your own insightful explanation of connected cursive writing deriving Cyrilic З from Ζ versus Numeral 3 from "a Brahmi glyph with three lines, similar to Chinese 三" (linguistics.SE: Is the Cyrillic letter 'Z' the same as the number 3?), you should be able to see that a similar principle may well apply to Σ << 𐤎 especially in those ancient forms which have more edges than 𐤔 shin.

Presumably, this was enough justification to commit spelling pronunciations to any remotely similar shape, potentially causing a merger.

So: what other theories are there on where the Greek sibilant letters came from? And what evidence is there either way (either for or against Jeffery's theory)?

It's not quite clear why that question should be limited to san, zeta, xi and sigma. There is no reason to excempt psi, sampi and for that matter digamma, seeing that *sw- (as in Epic ἑθεν, Aeolic ϝέθεν) must have been originally sibilant. It is quite simple then, based on visual impression, to compare digamma ϝ to Cypriot Syllabary's 𐠞 derived from Linear A, and to Half-Eta elsewhere (as in ἕθεν). "The fact that it [phi] is not used in Cretan and Theran writing might suggest it was not added to their alphabetic sequences." (Astoreca 2021:129)

I don't know if u heard that Spanish and Greek sound alike (Why do Spanish and Greek have such a similar phonology?), but one thing they have in common is what would be in my experience called "lisp". That makes it fairly easy to compare a prototype of Sigma and Digamma.

Quoting here because some might find "lisp" a kind of racist slur:

European Spanish famously "lisps" soft c's as /θ/. Other languages use c for /ts/, /dʒ/, /c/, etc. [@dan04, Why do so many loan words have a different pronunciations of letters like X and Q (among others)?)

NB: Wikipedia, "It may be that phi originated as the letter qoppa (Ϙ, ϙ)" [citation needed].

That's a typological parallel between the two. Following the aforesaid it obviously applied to Myc. i-qo / Pho. 𐤎𐤎 though the details are hazy.

  • 2
    Come on, man. When did *sw- disappear relative to the introduction of the alphabet? What was the phonetic value of 𐠞? When was the digamma first used and when the half-eta, and why wouldn't you look to the digamma's obvious source (Phoenician waw)? Why do you think the Modern Greek retracted [s̠] (its "lisp") has anything to do with the Spanish [θ], what basis do you have for projecting it to Archaic Greece, and how is it relevant to anything? Surely even you have to realise this entire answer is nonsense.
    – Cairnarvon
    8 hours ago
  • @Cairnarvon, sorry, you must have heard me saying that *sw disappeared relatively late. I was saying the writing existed relatively early, based on my redconstruction. Sihler 1995:§171 confirms that *sw- "evidently passed through a stage of being pronounced as voiceless resonants". The Digamma's obvious source, according to Wikipedia, is "in Crete" [citation needed]. As you see, the question is not when, but where, and "in Greek" is a racist stereotype (I have seen "Semitic" used as if it was a language more than once today, smh). Aha ja, nonsenf :D as is your quick shtick
    – vectory
    8 hours ago

Your Answer

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

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