English is not one of those, while German should be. Italian is one of those and French is not. So it seems that this feature does not depend on the linguistic similarity and historical relation among languages. Or am I wrong?

  • 2
    The granularity of the question is too large. French has 1-1 spelling-to-sound correspondence, but not sound-to-spelling. German is 1-1 both ways. English is not 1-1 either way. And any language can use a phonemic orthography (like Finnish, Turkish, Lushootseed, Indonesian, or Korean) and therefore guarantee a 1-1 relation. But it's not always a good idea; Chinese characters, for instance, unite languages that are mutually incomprehensible when spoken.
    – jlawler
    Commented Jun 9, 2013 at 17:03
  • What languages do you mean when you speak about Chinese characters? The different dialects of Chinese? Yes, the question is broad, but I meant to know whether is there some historical pattern in the different behaviours. Commented Jun 9, 2013 at 17:15
  • 4
    Your question is about orthography, not language. Language is spoken, universal, and evolved; orthography is none of these. Orthographies are subject to quite different forces and patterns from spoken languages, and can change overnight, unlike speech. To get some idea, check out Daniels and Bright's The World's Writing Systems.
    – jlawler
    Commented Jun 9, 2013 at 17:45
  • Armenian is known to have almost one-to-one writing system, including such difficult artifacts like palatalization. Commented Jun 10, 2013 at 12:16
  • @lawler - Regarding 1-1 spelling-to-sound correspondence in French, this is mostly right. There are some exceptions due to etymology: écho vs échoppe, schizophrène and eschatologie vs schiste and schisme. Ending consonnants can sometimes be a problem : bout ("bou"=end) vs bout ("boute"=rope on a boat, 'corde' is proscribed). There are also two "o"s in French, open and close. I am not sure the pronunciation can always be inferred from the spelling, even including spelling context. But then, some people do not properly distinguish them. I am no specialist. I was only intrigued.
    – babou
    Commented Jun 12, 2013 at 6:57

4 Answers 4


For your question to make sense, it would have to be a question about particular language-orthography pairs (and if in talking about sound-symbol correspondences we interpret "sound" to mean "phoneme", then the question is further restricted to language-alphabet pairs, excluding syllabaries and logographic orthographies). So for example, you could ask whether the Roman alphabet as used to write Parisian French shows a 1:1 phoneme:grapheme relationship. If you're interested in syllabaries, you would ask whether a 1:1 mapping existed between graphemes and phonotactically legal syllable shapes in that language (syllabaries are by definition not 1:1 with regard to phonemes).

However, a further complication is that spoken languages change over time, and orthographies do not necessarily change along with them. As such, even if a particular language-alphabet pairing has a one-to-one correspondence of phonemes to graphemes at one point in time, there is no guarantee it will remain so.

As far as global patterns go, there is no systematic relationship between the orthography used to write a language and the genealogical history of that language. For example, lots of languages all over the world from all different language families use the Roman alphabet, and in many of those cases they do so because:

  1. they had no orthography prior to contact with European anthropologists, linguists, and missionaries, and the development of their orthographies were strongly influenced by those Europeans, and/or

  2. their orthographies are very recent and are designed to make easy use of language technology (like computers).

To give another example, the Mongolian script is used to write Mongolian, Evenki, and Xibe, but in principle could be adapted to write many other languages too. On the other hand, the Mongolian language is also commonly written using a slightly modified form of the Cyrillic alphabet.

  • Why would the question not apply to syllabaries? Japanese has a few kana terms that are not 1:1. Commented Jun 9, 2013 at 18:48
  • And then there's Hangul.
    – jlawler
    Commented Jun 9, 2013 at 19:21
  • 1
    @jlawler: Hangul is an alphabet that happens to compose letters non-linearly into syllable-sized blocks, but is not a syllabary.
    – drammock
    Commented Jun 9, 2013 at 21:50
  • 1
    @drammock: Ah well in fact phonemes are just considered a useful artificial construct and they don't fit for all languages, so "speech sound" could fairly be used for arbitrary purposes unless it has a specific set meaning withint linguistics terminology. Commented Jun 10, 2013 at 8:19
  • 1
    edited answer to clarify/disambiguate some of these issues.
    – drammock
    Commented Jun 10, 2013 at 21:14

If you insist on a strict reading of "one-to-one correspondence between sound and written symbol", there are actually very few, if any, well-known languages that qualify. Italian certainly doesn't: its writing system uses the letter H in words like "ho" and "hanno" where it doesn't correspond to any sound. German also certainly does not qualify: among other things, voiceless /f/ can correspond to either the letter V (as in Vater) or the letter F (as in Fall).

Languages with contrastive stress/pitch often leave that out, but it's a bit more arguable whether that qualifies as a violation of the principle of "one-to-one correspondence between sound and written symbol", since stress/pitch could be seen as a "non-segmental" element of pronunciation. Serbian, mentioned in Jelena's answer, apparently has phonemically contrastive vowel length that is not indicated in the spelling, which interacts in some way with a phonological pitch accent system (which also seems to be left out in the spelling).

If you relax the criteria, you get more of a spectrum. It is well known that for French, it's easier to predict pronunciation from spelling than it is to predict spelling from pronunciation. English is pretty tricky in both directions, but there are significant patterns. The elements of spelling systems that have less clear correspondences to pronunciation usually tend to be older.

  • How about Japanese katakana or pinyin for Mandarin?
    – Mitch
    Commented Feb 28, 2019 at 0:01
  • 2
    @Mitch: Japanese katakana doesn't mark pitch accent. Also, it's typically only used as one part of the Japanese writing system, which is non-phonetic as a whole. In Japanese texts written exclusively in katakana, Wikipedia suggests that the ヲ would be used to write the "wo" grammatical particle, which is (usually) homophonous with オ as [o]. Commented Feb 28, 2019 at 0:19
  • And pitch accent is distinctive (that is, you'd need context to determine what pitch accent there is straight from writing) ?
    – Mitch
    Commented Feb 28, 2019 at 0:23
  • @Mitch: Pitch accent is distinctive in Tokyo-type accents of Japanese. Commented Feb 28, 2019 at 0:28

A complete 100% one-to-one correspondence between the letters and the phonemes is Serbian language. In serbian there are two scripts Cyrilic and Latin but only in Cyrilic is one letter-one phonem. We have 30 letters and therefore 30 phonemes. Vuk Stefanović Karadžić reformated language in 1814. making ideal phonemic orthography.



I'd say that Esperanto is one of those languages. I don't remember a change in pronouciation possible.

The only exception maybe being the commonplace use of letter x, but this is due to limitation of computer keyoards about accents.

  • ĉ, ĝ, ĥ, ŝ, ŭ can be more easily written cx, gx, hx, sx, ux knowing that the letter x is not exisiting at all in the normal esperanto alphabet.

Chinese PinYin:

I don't know if that count as a writting system since it is not the main one used in everyday life for chinese people. But the pinyin writting system, for transcribing pronunciation of Mandarin/Han Chinese, has a one-to-one correspondance between sounds and written form.


I was thinking about spanish: What you can read, you can pronounce it. But as a comment rightly highlighted there are much more exceptions in pronunciations than what I thought.

Exception with letter c.

  • may be pronounced [k] like letter q or [s] like letter s, according to a fixed rule on the following vowel.

Also letter l which becomes another sound when doubled into ll.

  • 1
    Spanish spelling has more serious complications like ge = je, b = v, and silent "h" in words like hacer. Commented Feb 28, 2019 at 8:40
  • @sumelic : you're right. I edit the answer accordingly. Commented Feb 28, 2019 at 8:47
  • Also, ll vs y, s vs z, and x can either make a ks or s sound. Commented Feb 9 at 5:05

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.