In my understanding, the different languages exist in spoken form and (mostly) in written forms (what about sign languages?).

Some languages have developed a close relationship between the written form and the spoken one: if you know the rules, you know how to pronounce a word from its textual representation (actually pronouncing the word ("Jiři"?) is another story).
E.g., I know the rules for my mother tongue, Italian, and I know how to pronounce every new word that I see written, with very few ambiguities (mostly "parole sdrucciole", for which rules exist but are usually not implemented).

Other languages are less stringent in this written/spoken mapping ...

My questions:

  • Are there studies about the level of this "stringentness"?
  • How can their conclusions be summarized?
  • 6
    This is known as languages having shallow or deep orthographies. Finnish has a very shallow orthography; English has a very deep one. The concept doesn’t really apply to writing systems that are not directly sound-based (e.g., Chinese, Japanese, Mayan), and there’s so much variation of so many different types that I don’t think you can really ‘summarise’ it. Sep 13, 2023 at 10:55
  • 1
    @JanusBahsJacquet Thank you for the clarification. As it happens, I ;m just returning from lunch at the cafeteria, and looking at all those Chinese students (10-15%) I realized that my question doesn't make sense for a large fraction of world population...
    – gboffi
    Sep 13, 2023 at 11:38
  • Your “very few ambiguities” look quite substantial for those whose mother tongue isn't Italian, like me. The main ambiguities in Italian orthography are that stress placement and vowel quality (for ⟨e⟩ and ⟨o⟩) are not notated, ⟨s⟩ and ⟨z⟩ may be voiced or not, ⟨i⟩ and ⟨u⟩ may represent vowels or semivowels, and even the first two are enough to say that the Italian spelling is not quite phonetic.
    – Yellow Sky
    Sep 16, 2023 at 10:11

1 Answer 1


As far as I know, no signed language has a standard written form in general use, so we can limit the question to the relationship between writing and speech. We then start by eliminating languages because of properties of their writing system – Chinese for example because a given Chinese character does not represent pronunciation, it represents a particular word which, if you know the word, has some pronunciation(s). Chinese characters in other languages can for similar reasons be excluded from consideration, but Japanese also as kana which allows you to pronounce words that you haven't seen before.

One should also eliminate unwritten languages like Sentinelese, since no Sentinelese speakers read any language or writing system. There are other languages which are not conventionally written but whose speakers may know how to read some other language, and may therefore be able to create a writing system of sorts. This is pretty much the common case in Africa, that local languages don't have standardized writing systems but there are national languages whose orthographic principles can be applied to the local language with varying degrees of success. Such systems are typically highly variable, and literacy is correspondingly low.

The inquiry starts to make sense when we focus on standardized systems rather than spontaneous practices. To take one example, "Norwegian" has two written forms, Bokmål and Nynorsk. There is no standard pronunciation of either written form, instead, a person speaks some particular dialect like Valdrismål, and writes either in one of the two official written forms, or in an unofficial dialect-writing system (where Bokmål vi "we" is spelled and pronounced ). The lack of standard pronunciation of the written form in Norwegian is somewhat of an anomaly, compare that to British English or High German, or "Italian" (with the proviso that you are referring to Florentine Tuscan, not Lombardi, Venetian, Friulian, Sicilian, Sardinian and so on). We can say that there can be standardized rules of pronunciation for reading standardized written languages, but actual speech is usually highly variable to the point that you can't tell how a given written form is going to be pronounced in an area, without some additional study. (Example: North Saami spelling is standardized, but highly variable in its actual pronunciation, yet you can generally learn the spelling-to-speech rules for a particular "area" and know that <ŋ> is actually pronounced [ɲ] in Guovdageaidnu).

In light of the complexity of the relationship between writing and speech, the first thing that would need to be done is to more narrowly define the question. The obvious thing to focus on is standardized languages and writing systems, in major national languages with well-developed sound-based writing systems. English is the classical worst-case spelling situation in that it is hard to predict standard pronunciation from spelling, or spelling from standard pronunciation. Finnish is the really easy case. French is challenging because you will struggle to get the correct spelling given a pronunciation, though inferring the pronunciation from spelling is not hard.

There are numerous studies of the process of learning reading and writing for individual languages, but little that attempts to discern global generalizations about e.g. learning to read and write in Yoruba, Shona, Hindi, Korean, Indonesian, Turkish, German, Danish and Italian. It's hard to imagine what kind of meaningful survey one could even conduct that compares cognitive processing of Hindi, Italian, Korean and Yoruba.

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.