English uses a variety of digraphs to represent sounds which lack their own letters. Some of these (such as "th" and "sh") appear in native words; others (such as "kh") only appear in loanwords.

I know "th", "ph", and "ch" appear in Latin words borrowed from Greek, to represent θ, φ, χ. What other letters/digraphs did the Romans use to represent non-native sounds? Did these vary over time? Were any of the sounds "standardized", the way "kh" in English now tends to represent a velar fricative?


1 Answer 1


In Classical times, few digraphs were used.

The letters K, Y, and Z were used for Greek words, along with the digraphs Ph, Th, Ch for aspirated stops, but other words were generally assimilated to Latin phonology. (Source: searching for the combinations below in the Packhum corpus and finding nothing.)

Z was used for more foreign sounds in Late Latin.

Jordanes' De Origine Actibusque Getarum uses Z when transcribing the sounds /ts/ and /dz/ in foreign languages, as well as on its own for /z/:

  • Tz: Emnetzur, Ultzindur, Dintzic, etc
  • Dz: Scandza, Augandzi, Tadzans, etc
  • Z: Boz, Bizzim

Other foreign sounds were still assimilated, but these particular combinations weren't utterly impossible in Latin (e.g. the compound etsi).

This tradition seems to have gotten slowly broader and broader over the next several centuries, as people with native Slavic names were mentioned in Latin texts (e.g. the 12th-century Bull of Gniezno with rz for a retroflex sibilant, ch for some sort of affricate, and z for various other Polish consonants; the 12th-century Chronica Slavorum with zc for /tʃ/).

Unfortunately, my sources are sparse, and more data from Classical or Late Latin would be appreciated: surely other writers than Jordanes mentioned foreign names.

  • First of all, a quote from Schmidt 2006: "Mommsen based his edition on the least ‘classical’ ‘Class 1’, mainly on Heidelberg, Pal. 921 (see VIII/IX; destroyed in a fire at Mommsen's house in 1880); the linguistically more standard ‘Class 3’ is now represented by the older Bobiensis Palermo, Archivio dello Stato, Cod. Basile (see VIII2, Codices Latini Antiquiores Suppl. 1741). Only the future edition by Bradley (cf. [13]) promises a stemmatically balanced text, which should provide a new basis for any investigations of language and style."
    – Alex B.
    Commented Jul 1, 2018 at 2:12
  • Secondly, Mommsen himself gives the following variants: emnetzur, emnedzur, emnezur, emmedzur, emmezur, zinnezur, emnetdzar, emnedzar just for Emnetzur only.
    – Alex B.
    Commented Jul 1, 2018 at 2:15
  • Bradly 1995 writes: "It is then possible that Mommsen has too readily dismissed the testimony of class iii manuscripts, and taken too little account of the possibility that the aberrations of a series of copyists are responsible for much of the flawed orthography and forms which characterise class i manuscripts and O" (p. 348). Also Mommsen didn't think highly of agramatus Iordanes; he calls him "scriptor infans et barbarus." See Bradly 1995 for a rather long list of errors in Mommsen.
    – Alex B.
    Commented Jul 1, 2018 at 2:20
  • @AlexB. That would make a good answer in and of itself!
    – Draconis
    Commented Jul 1, 2018 at 2:26

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.