Historically, English did not have standardized spelling; see e.g. this paragraph from the Washington Post:

At one point, English speakers lived in a world without standardized spelling. According to the Linguistic Atlas of Late Medieval English, in the late Middle Ages a word such as “through” could have as many as 500 variant forms, from recognizable formulations such as “thurgh” and “thorough” to more inventive combinations such as “orowe,” “drowg,” “trghug” and “trowffe.” There were pecuniary reasons to use inventive spellings: Lawyers’ clerks were often paid by the inch, and they added superfluous letters to words to pad the bill. Typesetters, on the other hand, might spell the same word several different ways in the same text to save space. Source

The article goes on to explain that there were attempts to standardize English spelling from as early as the fifteenth century. Nowadays English spelling is almost entirely standardized (within versions like British English, American English etc.) except for very few words, and even in those cases (e.g. focused, focussed) writers aim to be consistent within a given document. The same is true for the other modern languages I'm aware of.

Are there any modern languages which are written down but without standardized spelling, as English was in the paragraph above? A user of such a language could validly write down a sentence in multiple ways which could be correctly read and understood as the same sentence (without one version being 'better writing', 'more correct' or anything similar), and the same word might be spelled in different ways within a text.

If there are no, or vanishingly few, such modern languages, what factors lead to this being the case? I can imagine some causes (especially technology and the desirability of computer search functions, but this is a very recent consideration) but I am interested in a wider global-historical linguistic view.

  • 4
    Paradoxically, but it is the advent of computers and Internet that's exposed how little standardized even the major world languages are. When you enter a non-linguistic forum it looks like few of the ones there ever went to school. If a word doesn't become a nearly cryptogram like “drowg,” “trghug” and “trowffe” of your example, every fancy spelling will do nowadays.
    – Yellow Sky
    Commented Jun 6, 2022 at 22:39
  • 15
    I'm not confident enough to give an answer, but I think Swiss german (Zurich region) might fit that bill. We never really learn how to write in swiss german, we only learn how to write in german. Poems and songs written in swiss german tend to use different spellings, and everyone just writes it a bit different (which makes it hard to read). At the same time, these different words are regional enough to be able to locate a town or village just by the sound of words (there's even an app for that). But perhaps there is some spelling dictionary, nobody I know knows about that though.
    – Katai
    Commented Jun 7, 2022 at 8:32
  • 3
    btw: "through" is already pretty inventive spelling! Commented Jun 7, 2022 at 9:01
  • 3
    It is more likely that it was spelled "ðrowg" and not "drowg" and the character was erroneously substituted. Commented Jun 7, 2022 at 20:44
  • 4
    @EvilSnack England had a standing army from 1645 and a navy well before that, but English was still far from standardized at that point! (I'm aware you weren't being totally literal there.)
    – dbmag9
    Commented Jun 8, 2022 at 13:55

9 Answers 9


First, let's define our terms. Spelling is "standardized" if there's some authority that people listen to on the topic. This can be a government agency, like the Académie Française, or a private entity, like Webster's dictionary or Microsoft Word's spell-checker. (Even now, English spelling has several competing standards: is it "thru" or "through"? "worshiped" or "worshipped"? "color" or "colour"?)

With this definition, I'd say a good percentage of languages lack this standardization. In sub-Saharan Africa, for example, there are far more languages than language-regulating groups. New phones may come with Swahili spell-checkers, but as far as I know, nobody's created such a thing for Kikongo or Kituba, and while linguists might use letters like ɛ and ɔ combined with various tone diacritics, most keyboards can't type those and people will find other ways to make do (e.g. using è for ɛ in French-influenced areas).

In other places, we see a similar lack of standardization for indigenous and minority languages. While Standard Italian has official rules for spelling, there's less standardization around, say, Veneto. If your dialect has a /θ/ phoneme, the rules of Standard Italian won't tell you how to write it. Or if you're a speaker of Oneida, which until recently was much more commonly taught in homes than in schools, you may never have learned the orthographic standards linguists have invented. Even if a standard technically exists, it doesn't mean much if the native speakers aren't using it.

  • 14
    @Tristan While true for many dictionaries, I wouldn't exactly describe Webster's as "descriptive" — many of the differences between Commonwealth and US spelling are because he deliberately decided to unilaterally make changes in his early dictionaries to force a divergence between the languages, as part of his nationalistic view that America was inherently "superior" to European nations… Fortunately, some of his attempts like "Masheen" ("Machine"), "Soop" ("Soup"), "Ake" ("Ache") or "Beleev" ("Believe") were rejected by the public. Noah Webster very much ran it as a "standardising agency". Commented Jun 7, 2022 at 9:30
  • 8
    @Chronocidal whilst that was certainly the case of Webster's original dictionary, the modern iteration is descriptive. Even his edition received some criticism for its descriptive inclusion of proscribed words like "ain't"
    – Tristan
    Commented Jun 7, 2022 at 9:56
  • 6
    @Tristan It’s still a standardizing agency to some extent, it just operates completely differently from how organizations like the Académie Française work. The dictionary is descriptive of a de facto standard that it effectively perpetuates, as opposed to the Académie Française formally creating a de jure standard. Commented Jun 7, 2022 at 12:20
  • 4
    @phoog wiktionary says its the norm in the US
    – Tristan
    Commented Jun 7, 2022 at 14:18
  • 10
    I particularly like the fact that the spelling of "standardised" isn't! Or rather there's a choice of standards.
    – Chris H
    Commented Jun 7, 2022 at 14:31

As Katai has pointed out in the comments Swiss German is a modern langauge that has no standardised spelling. In "Mundartwörterbüchern"(=Dialect dictionaries) one can see two different ideas of how to spell. Either a phonetic way or one that is close to the Standard German spelling. Either way it is not standardised and in day to day life people tend to write whatever they are comfortable with ("Nach Gefühl" = By instinct).

This (in part) is why Swiss German is not an official language of Switzerland. People are taught Standard German in school and official correspondence is always in Standard German.

  • 4
    I can attest that Swiss german has no standardized spelling by any means. (a) There is no standardizing body that is recognized officially (by the government) or by the public. (b) Event if you'd consider those Mundartwörterbücher as standardizing body, no-one here recognizes them as such (it's very rare that someone "corrects" my spelling and no-one ever referred me to the spelling in one of those books). We only have them as gifts to foreign friends and family that find the sound of our language funny ☺️. Commented Jun 7, 2022 at 13:19
  • 5
    This can easily and rightfully expanded to include all dialects of German. But especially the indeed different languages of northern Germany, including parts of the Netherlands, which have even different entire Wikipedia sections, with namespaces like frr, fy, nds, nds-nl, for the most significant northern parts alone, but these at best only starting to advertise the need for standardised spelling, having noe. Commented Jun 7, 2022 at 23:46
  • 1
    @LаngLаngС I'd call that a stretch. Those are spoken only by a minority in those regions so yes they are indeed just a dialect or a mixture of two languages. Schwyzerdütsch on the other hand is the language spoken by the german speaking parts of switzerland with its own dialects.
    – SirHawrk
    Commented Jun 8, 2022 at 5:58
  • 1
    Moreover in ZH canton even the walking signs erected by the canton can be spotted spelling the same thing different ways on different signs ("Degenreid" being a local to ZH example).
    – abligh
    Commented Jun 8, 2022 at 6:11
  • 6
    @SirHawrk What does only being spoken by a minority in those regions have to do with being a language or not? Frisian is well established as an independent language or languages, being closer to English than Dutch, for example.
    – prosfilaes
    Commented Jun 8, 2022 at 15:33

It depends on what you mean by "standardized", also "are written down". In Logoori (a Bantu language of Kenya), there are multiple observed spelling practices, but they can generally be related to some specific standard: the first Bible translation of the '20s, the 1967 Bible translation, a booklet promulgated by the "Orthography Committee" which had some relation to the colonial government, and the "Swahili model" which is to "make it look like Swahili". In each case, a spelling choice is made by reference to a standard (either a specific book, the rules of a book, or the rules of spelling of another language) – there are at least 4 standards.

I cannot think of a clear contemporary case where a language is in fact systematically written, but where people randomly make up all of the spelling conventions. I would exclude many languages of Tanzania (and elsewhere) which are only un-systematically written, e.g. a Matumbi speaker might write a letter in Matumbi to another Matumbi speaker, but even then they will follow the "Swahili standard" where you ignore features not present in Swahili and otherwise write /b/ as <b> as you do in Swahili.

Any human language can in principle be written down, but many languages are not in fact habitually written down – Sentinelese is not written down (at all, as far as we know). Without a much narrower definition of "standardized", I think there is a contradiction between "is written" (not just potentially of sporadically) and "follows no standard".


Serbo-Croatian. The Serbo-Croatian language only has the convention that writing is phonetic and so you "write the way you speak". That is, each letter corresponds to precisely one sound and each sound corresponds to precisely one letter.

Combined with regional speech variations, this results in all regional accents being represented directly in writing. So people from different cities who have slight differences in how they pronounce words carry over those same differences to their writing. For example, the adverb meaning "beautifully" can be spelled in at least the following three different ways by different people: "lepo", "ljepo", "lijepo".

  • 3
    I do not see how that is free spelling. Especially the specific examples are just writing how you indeed speak in the dialect. Why would you write something else than what you say? It would not make sense - provided there is a clear correspondence howto pronounce a written word - unlike English. You can of course have a standard dialect and prefer that for the written communication, but if you indeed speak differently, than writing what you actually speak does not make your orthography non-standardized. Commented Jun 7, 2022 at 12:20
  • 2
    For example, in Czech I can write what I say in the standard language: "Dnes je mi lépe." or what I say in Common Czech "Dnes je mi líp." or in some regional or historical dialect "Dnes je(st) mi lépěji.". But in all these cases the spelling is clear, it just follows the actual sounds pronounced using some standard rules. Commented Jun 7, 2022 at 12:23
  • 2
    This is the case for languages that actually directly map the orthography to what is pronounced. In English people often utter very different things in different dialects and write it the same way. Then they have to learn the spellings for individual words at school. This is mostly not needed for languages that have reasonably clear mapping between the sound and the orthography (with exceptions, the i/y or mě/mně are homophonous and tricky for spelling - based on etymology - in Czech). Then you can write some differences you otherwise could not. Commented Jun 7, 2022 at 12:28
  • 3
    @VladimirFГероямслава but many of the supposed irregular spellings of English in centuries past reflect the same phenomenon. Some people actually pronounced "through" as "thurgh," for example (note German durch). I would add for those who insist that spelling conforms precisely to pronunciation that the spelling of Hrvatska ought to be Hrvacka and that, while there are some dialects that do not maintain the distinction between ć and č I have not seen this reflected in spelling. The variability in the vowels is better seen as multiple standards rather than a lack of standardization.
    – phoog
    Commented Jun 7, 2022 at 13:17
  • 6
    For what it's worth, I never received any formal training in the Serbo-Croatian language (picked it up by exposure and then used it to study a STEM subject in university), so I don't know the official stance on whether it's standardised or not. I am just familiar with the "vibes", and seeing many of my professors in university spell the same words in different ways felt to me like the language is not very standardised.
    – Mash
    Commented Jun 7, 2022 at 14:03

Occitan has two main accepted orthographies, the Provencal standard and Languedocien. In the Middle Ages the situation was similar to English and French of the era with a wide variation in regional spellings, and personal idiolects mixing different forms.

There has been some standardisation based on different spellings. The Provencal form was used across a broader region during the revival of the late 1800s but now is only used in Provence itself, and the standard 'classic' spellings are based on the dialect of Languedoc with a strong influence of medieval spelling.

Other sub dialects exist but they tend to be separate and have either internally consistent spelling or a couple of accepted variants for each word. The borders of different dialectical features were mapped pretty thoroughly in the 19th C and that killed a lot of the personal idiolects. Gascon remains visibly different, often includes archaic spellings and has more accepted variants, but is still far from being free-spelt.

If a language is reasonably well used (in written form), there will have been some attempts at standardisation. Languages considered purely 'dialects' - like the examples other people have given, or Franco-Provencal, are less likely to have this. They are also less likely to have extensive written literature from the last few centuries and so on. A possible exception is Scots English.

The other point is that many people will use a variable spoken dialect but will write in a standardised from - you can blame universal schooling for this. In IPA the different English dialects are wildly different but everyone uses classic spelling (except in error). For first languages, or languages with pretensions to be one, there will be only more and more standardisation since 1800ish. Few places would approve a curriculum where you can use multiple spellings of the same word, and word processing will only add nails to the coffin.

  • 2
    Why is Scots an exception? I've looked for books written in Scots, and there's a handful of modern works, and some medieval works, but in 1700-2000, Scots was used as color, very rarely as literature. Robert Burns is somewhat exceptional, but on the other hand, shows the rule; as Wikipedia says, much of his writing was in Scottish English rather than Scots to be readable to an English audience.
    – prosfilaes
    Commented Jun 8, 2022 at 15:47

Yiddish... if you call it a language.

  • 12
    What else would it be?
    – Robert C.
    Commented Jun 7, 2022 at 19:21
  • 8
    So like English then?
    – Robert C.
    Commented Jun 8, 2022 at 0:05
  • 5
    There are standardized spellings for Yiddish, notably the one suggested by YIVO. Then, of course, it is one of many languages that has multiple varieties and used in more than one country.
    – Roger V.
    Commented Jun 8, 2022 at 13:20
  • 2
    @RobertC. - it looks like you don't speak Yiddish... An American will understand an Australian or a South African, despite the huge geographical distance between those places and despite the huge amount of speakers in different areas who bring local influence. I do understand Lithuanian Yiddish to certain degree but I understand almost no Chassidic Yiddish that originates in Ukraine. Those two countries are quite close to each other. Amount of speakers is tiny. And I grew up in Ukraine and speak Hebrew, German, Ukrainian, Russian - all the languages that constitute the mixture, called Yiddish. Commented Jun 9, 2022 at 20:49
  • 3
    @dotancohen - Yiddish is actually used mostly to communicate. At least in my town. Commented Jun 9, 2022 at 21:10

The Scots language has a variety of orthographic standards based upon historical dialects, with standardization attempts ongoing among Scots-speaking communities. Since there is already a substantial, established written Scots literary canon, it is questionable whether these efforts are likely to be successful. At the time English was being standardized, it was still developing its literary canon.


Western Abanaki in Northern New England and parts of Québec has no standardized spelling; however, work is currently being done to more or less standardize it. The reason is simply because until fairly recently in history, it was a spoken language only. The word for 'friend' (literally "my friend") for example can be written as either 'nidôba' or 'nid8ba'. The word for large/big can be either 'geci', 'kechi', 'kchi' - the 'e' is the 'schwa' sound and it's not always written.


I see folks pointing out that, to answer productively, we need to define our terms, such as "what do we mean by standardized, and what do we mean by written down".

I think there's another question we must ask to clarify the problem space for this question-and-answer thread:

  • What is the point of writing? What is the use case or purpose?

This might seem silly, or overly broad. Specifically what I'm driving at is, are we talking about "spelling" to capture the sound of the word, or to capture the meaning?

My background is working in and with the Japanese language, and to a related but lesser extent, Chinese.

Written Chinese is widely described as logographic, where individual graphemes convey meaning more than sound. This allows for a very flexible relationship between text and speech, to the extent that (at least some) Chinese poems from 2,000 years ago still make sense and still work as poetry, despite substantial changes in spoken Chinese during the intervening years.

That flexibility allows folks in China to read and write across linguistic boundaries, such that someone who grew up speaking only Mandarin can still carry on a written correspondence with a Cantonese speaker. This flexibility also means that, to a more limited extent, folks from China can still get around in Japan, and vice versa, despite Chinese and Japanese being completely different languages, with wholly unrelated grammars, and large chunks of unrelated vocabulary. In effect, the written language becomes a shared context separate from the spoken language.

For alphabetic / abjadic / abugidic writing systems, meanwhile, the individual graphemes record sound more than meaning. This kind of default tighter coupling to phonetics necessarily leads to one of two approaches in writing, which we see reflected to some degree in all of the examples brought up so far in this thread:

  • Writers focus on the spoken forms, and spellings change as pronunciations change -- be it due to the passage of time, or to differences in sociolect. ** We see this in English up through at least Middle English. Chaucer's knyght was pronounced basically as written, as /kniːçt/.
  • Writers opt instead to focus on the written forms, and spellings remain the same even as pronunciations change. ** We see this in English since some time after Shakespeare. Chaucer's /kniːçt/ became /naɪt/ instead, but the written form preserved the now-silent ⟨k⟩ and ⟨gh⟩.

English might be a bit of an outlier for alphabetic languages, where the individual graphemes have lost their phonetic values when appearing in specific contexts. One could argue that the meaningful graphemic unit has shifted in some cases from the letter to the entire word -- perhaps making English's oddball spelling closer in some respects to logographic Chinese, and similarly allowing for mutually intelligible written correspondence even across mutually unintelligible differences in pronunciation -- becoming a shared context separate from the spoken language.

To loop back to the OP's question:

Are there any modern languages which are written down but without standardized spelling?

We also need to define what do we mean when we say "language"? Are we referring solely to the written form, where such exists? Or are we referring to the spoken form? Many of the spelling variations mentioned by earlier posters in this thread arise from variations in pronunciation. Where do we draw the line between "language" and "dialect"? The "variations" seen here might well be internally consistent enough to call "standardized", and are only cast as "variations" due to the sociopolitical status of different speech communities within the shared polity that contains them.

Just within the US, any attempt at expressing vernacular speech in writing would be cast as a "variation" or "non-standard". Just look at web postings for a smattering of these.

From this perspective (the indefinability of "language" versus "dialect", and what is "standard"), the answer to the OP's question is arguably, "all of them". 😄

  • English spelling is famously not great at reflecting pronunciation, but I'm not sure how much of an "outlier" it is, compared to say, French, where almost every word ends in a silent letter, or Tibetan, from which we get the word yeti, that's spelled in its native writing system using the letters for gyadred.
    – dan04
    Commented Jun 9, 2022 at 22:22
  • 1
    @dan04, by "outlier", my thought is that written English diverges from spoken English not necessarily so much in the spelling, as in the way that a standardly-spelled word might have half a dozen different phonetic realizations depending on local lect. Take the first-person singular pronoun "I", for example — this shows up commonly as /aɪ/, but also as something like /ɛi/, /ɑe/, or /aː/. The spelling ⟨I⟩ maps to the concept of "first-person singular pronoun" more than it does to a specific sound. Commented Jun 9, 2022 at 22:33
  • In that "spelling ↔ concept" 1-to-1 matching, and in that "spelling ↔ pronunciation" 1-to-many matching, English orthography struck me as a bit more like Chinese than (what I'm aware of for) other alphabetic languages. Commented Jun 9, 2022 at 22:42

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.