The Greek alphabet has ψ (psi) and ξ (ksi) , both letters correspond to a bi-consonantal sequence. Latin has X, Cyrillic alphabet has Щ which in some languages corresponds to ʃt͡ʃ, etc.

Normally, an alphabet is phonemic (before the language changes and departs from it). Is there any motivation for a single letter representing more than one consonant? how did such letters come to be?

  • I've always wondered about the Greek digraphs ξ and ψ too. They don't seem to correspond to anything in the Phoenician alphabet.
    – jogloran
    Jun 8, 2012 at 13:04
  • 1
    I would say not many alphabets are phonemic. Even by 400 AD, a paradigm of "one sound = one letter" was a luxury. Here is a story of creating Armenian language. So there's no surprise there are clustered consonants. Jun 9, 2012 at 3:12
  • @jogloran: There's a theory that ksi has derived from the Phoenician Samekh, but its name and sound value from Šin Jun 9, 2012 at 3:25
  • 3
    There is little reason to expect historical alphabets to be phonemic. Most languages' scripts were adopted and/or adapted from those of other languages, often by people with no particular linguistic training.
    – Colin Fine
    Jun 9, 2012 at 23:42

3 Answers 3


In the case of Cyrillic Щ, it originated as just a scribal ligature of Ш and Т when writing Old Church Slavonic (where, naturally, it was pronounced /ʃtʲ/). It still does represent a cluster in some languages (e.g., Bulgarian /ʃt/, Ukrainian /ʃtʃ/), but in others has evolved into a single sound (e.g., [standard] Russian /ɕ:/).

The reason this particular letter stuck around is because it had a useful function: the cluster /ʃt'/ in Old Church Slavonic wasn't just any cluster, but it had an important morphophonemic function, being the jotated counterpart to /st/. In addition, Old Church Slavonic simply didn't have another good way of representing /ʃtʲ/, since palatalized /tʲ/ only occurred phonemically in this particular cluster; this probably encouraged more frequent use of the letter, although this is just me speculating.

I'm not sure, but I believe Greek ψ and ξ are similar, in that the clusters /ps/ and /ks/ have a special significance in the language, but I'm not very familiar with Greek.

Basically, morphophonemics can be a good motivator to have a single letter represent what's phonetically two sounds.

  • I'm attempting to learn cyrillic script now and have big trouble with Ш and it's many relatives.. THANK YOU! I might manage to remember 'em now :D
    – kaleissin
    Jun 8, 2012 at 18:17
  • This answer is now the second time I really want to ask a question "What is the difference between palatalization and iotation?" Jun 9, 2012 at 10:30
  • 1
    @hippietrail : I use "jotation" in this case to refer to a specific sound law in Common Slavic, where the "jotated" sequences *tj, *dj, *sj, *zj, *rj, *lj, *nj, *stj, *mj, *bj, *pj, and *vj (I think that's all of them) underwent palatalization, which yielded different results in different Slavic languages. When I say "jotated counterpart", I mean it alternates within a single paradigm, such as jotated OCS мьщѭ "I avenge" (Common Slavic *mьstjǫ) and non-jotated OCS мьстиши "you avenge" (Common Slavic *mьstiši).
    – voikya
    Jun 9, 2012 at 13:22
  • @kaleissin if you need help, ping me in chat!
    – Alenanno
    Jun 9, 2012 at 22:59

The Greek graphemes Ψ ('psi', [ps]) and Ξ ('xi', [ks]) represent consonant clusters in some varieties of the Greek alphabet. These clusters are represented by the same graphemes even when they are spilt by a morpheme boundary. Some sources describe these clusters as phonemes but this is not correct, they are combinations of two phonemes in all varieties of Greek.

It is not certain why the Greeks used single graphemes for these two clusters but it may be related to the fact that they are the only clusters that occur in coda) (ie. final) position in syllables. Earlier forms of the Attic alphabet lacked these graphemes and used ΦΣ and ΧΣ, respectively. While the majority of the Greek alphabet is believed to have been derived from either the Canaanite and/or Phoenician writing system, the graphemes Ψ and Ξ are have no obvious prototype in these traditions and may possibly have been derived from the South Arabian script.

While the alphabetic tradition normally has the target of representing individual phonemes, this may not be possible when borrowing an alphabet for use with a different language. Other considerations such as prosodic phenomena, syllable structure, morphophonemics, may make single graphemes for multiple phonemes a better solution than a straight mapping of segmental phonemes. And of course there are many languages which have single graphmes for phonemes, such as affricates and double articulations, that combine phones found in other languages as distinct phonemes.

  • As is says here, so many words, like the sigmatic aorist stem of many verbs, for instance, had forms ending in s that Greek developed special letters for stops followed by s. By Classical times the dental affricate Zeta had changed from /ts/ to /dz/, and thence to /z/. Velar Xi was borrowed into Latin as "X", for similar reasons, but labial Psi never made it out of Greek.
    – jlawler
    Jun 12, 2012 at 19:24
  • @jlawler Thanks very much, I had wondered if there was some morphosyntactic relevance in the occurence of these clusters (and no others) in coda position. It's nice to see that confirmed and I'm sure it was an important factor in the move to single characters to represent them. Jun 13, 2012 at 8:21

In Greek, psi and ksi are used commonly in verb conjugations.

For instance, βάφω (I paint) becomes έβαψα (I painted), θα βάψω (I will paint) and βάψε (paint!). Similarly, τρέχω (I run) becomes έτρεξα (I ran), θα τρέξω (I will run) and τρέξε (run!).

I figure these two very common consonant clusters in so many verb conjugations lent itself to being written with one character. I believe there are no other consonant cluster so prevalent in verb conjugations.

  • 2
    This is true, but it doesn't really illuminate the question (and your examples are from Modern Greek, which is a couple of thousand years too late to explain the alphabet). Latin also had desinences of the s-preterite resulting in /ks/ and /ps/, but while it adopted X (ultimately from Greek X, which in some dialects was used for /ks/ as well as for /kʰ/), it never had a letter for /ps/. Notice also that /ks/ is common in Germanic languages, but while English (and Swedish) commonly use Latin 'X' for it, German rarely does, preferring 'CHS'.
    – Colin Fine
    Mar 21, 2022 at 13:09
  • Ancient Greek uses similar verb forms with high usage of psi and ksi. I'm just providing a possible reason why the two consonant clusters are written with one letter. Like so much in linguistics, what I put forward is a sufficient reason but by no means is it a necessary reason -- Ancient Greek could well be written in hieroglyphics if you really wanted it to. So citing what happens in other languages is not necessarily relevant either. There is no iron-clad law of how consonant clusters get written.
    – Sarhanis
    Mar 21, 2022 at 22:52

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.