In English for example, the "ch" sound (as in China) is sometimes written as /t͡ʃ/, other times as /ʧ/ or simply as /tʃ/. Similarly, I have seen the German "tz" (e.g. Katze) transcribed as /t͡s/, /ʦ/ and /ts/.

Is there an "official" recommendation on which option should be considered correct?

  • 3
    I can highly recommend this book for Unicode topics in linguistics. Sep 19, 2021 at 9:27

3 Answers 3


"tʃ", "t͡ʃ", and "t͜ʃ" are the only representations of the affricate currently sanctioned by the International Phonetic Association. The ligature symbols "ʧ", "ʦ", etc. were removed from the alphabet at the 1989 Kiel Convention.

The arches in "t͡ʃ, t͜ʃ" are called tie bars. Note the IPA chart says "Affricates and double articulations can be represented by two symbols joined by a tie bar if necessary" (emphasis added). They are essential in languages where an affricate phoneme and a homorganic stop–fricative sequence frequently contrast, such as Polish, but otherwise they are omitted more often than not. I do not know of any dictionary of English or German that makes use of them.

  • Did I get it right? The tie bar is only for representing a single consonant of Affricate or double articulation? Apr 17 at 16:05
  • 1
    What is a single consonant is context-dependent. There's nothing in the articulations that indicates [ts] should be regarded as a single consonant in German and not in English, but that is nonetheless the case because of the distribution and history of such sounds in each language. Tie bars may be used to denote affricate or double articulation, but just because a sequence of two consonants lacks one doesn't automatically mean it's not an affricate or doubly articulated. It just means the author didn't denote it.
    – Nardog
    Apr 18 at 7:14
  • Tie bar and Affricate reminds me of 'Tashdid or Shaddah' with the symbol: (ــّ) which is two same consonants that are considered one consonant with high stress (in pronunciation) but two consonant in linguistic analyzing or poetry. I found it mostly in Arabic and Persian languages. Or 'Idgham' which is two consonants with very close pronunciation that are merged together in speaking (representing one consonant) but still two consonant in analyzing. I was wonder if they have any relation to Tie bar or Affricate. Languages are great topic to study. Apr 19 at 12:42

In theory, the difference between /t͡ʃ/ with a tiebar and /tʃ/ without is that the former represents a single unit and the latter represents two units. This is sometimes important for theoretical reasons (you might want to treat /t͡ʃ/ as a unit because it lets you make some phonological generalizations) and more rarely for practical ones (Polish contrasts affricates against stop-fricative clusters: /ˈt͡ʂɨsta/ "clean", /ˈtʂɨsta/ "300").

In practice, most people don't bother to write the tie bar unless it's important (as in Polish). So I've seen both /t͡ʃ/ and /tʃ/ in different transcriptions of English, without any meaningful implication.

/ʧ/, the single character, is obsolete and no longer part of the IPA. But it would mean the same thing as /t͡ʃ/ with a tie bar: a single affricate unit rather than a stop followed by a fricative.

  • 3
    Wonderful! Allow me to expand. Polish trzysta "300" continues PSlav. *tri- akin to three from *treyes, where rz is a digraph, but czysty "clean" continues PSlav. *čìstъ akin to Latvian šķîsts "thin, clear, clean, vel sim. apparently from PIE *skeyd- (according to wiktionary).¹ So, it looks like they come from very different clusters, except that both changes seem to be conditioned by front(ed semi-)vowels. See palatization and assybilation: mega, maioris; *cansmn, carmen; *kaput, chapel; hio, she; *hwaz, who; indeed Sino-, China. Polish maintains the theory in practice.
    – vectory
    Sep 17, 2021 at 9:49
  • 1: I would further compare Ger. gescheit "clever; proper", origin ultimately uncertain, and I remark doubts about s-mobile, satemization, glottalization, laryngeals and the nostratic-like hypothesis (cf. @anixx) that PIE *H correspond with PSem *t, noting that PSem *š reads */ts/ (or */t͜s/?). Sanity check (cf. WT): Semitic "3" is somewhat uncertain, *ṯalāṯ- or *śalāṯ?! Further extra-familiar comparison notwithstanding, and despite a chance of wanderwords, I suggest to compare En cutsy (from Latin, from *h₂eḱ- (“sharp”), according to opinio communitis [sounds like an illness]) JSYK
    – vectory
    Sep 17, 2021 at 11:28
  • I am pretty sure the research has shown that in Polish we actually have /t͡ʂ/ and not/t͡ʃ/, contrary to Czech. Sep 17, 2021 at 18:48
  • I think it would be useful to mention or explain the difference between // and [] here! :) Sep 19, 2021 at 23:58

The 2015 IPA chart is freely available from the Association's website. It does not include affricate characters in the form of ligatures (such as ʧ and ʦ); it does include a note stating "Affricates and double articulations can be represented by two symbols joined by a tie bar if necessary", giving the examples "t͜s" and "k͡p". The implication is that if not necessary, affricates and double articulations may be represented by two symbols in sequence without a tie bar, like your examples of ts and tʃ.

There may be more detail in the 1999 Handbook; since it is not free, I have not read it. Wikipedia states without a source that the ligatures for affricates "are no longer standard IPA".

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.