Is there a language that has a complete one-to-one correspondence between the graphemes (letters) and the phonemes of the language?

In other words, is there a language that is 100% ideally phonemic?


17 Answers 17


Finnish is the usual exemplar for that.

Many recent alphabetizations, like those of Native American languages (Lushootseed is one example), are still phonemic in the sense that the spoken language hasn't had time yet to change away from the phonemic system it had when the alphabet was developed. Or in other cases, where there are no native speakers any more, all language learning is based on the alphabetic representation, by necessity.

Of course, there are plenty of non-alphabetic writing systems with little or no useful correspondence between phonemes and graphemes. Since phonemics is an alphabetic representation system, it can't be put into 1-1 correspondence with an abjad or an abugida, let alone with a lexically-based system like Chinese.

  • 4
    @jlawler I disagree with part of your second para. Abjads and abugidas are still based on phonology, but represent syllables rather than individual phonemes, with vowels being underspecified. Where they do represent the whole syllable they can be very shallow orthographies. Commented May 20, 2012 at 14:05
  • 3
    True. Sanskrit devanagari is an abugida and the phonemes are famously legible. However, Sanskrit is dead and therefore can't change; and the relationship between alphabetic phonemes and devanagari symbols is hardly 1-to-1.
    – jlawler
    Commented May 20, 2012 at 14:58
  • 1
    It's still spoken, just as Latin is, but there is no native speech community. Some gurus have tried to raise their children as native Sanskrit speakers, but if it were still alive, it would change. Sanskrit doesn't change.
    – jlawler
    Commented May 21, 2012 at 17:29
  • 9
    Even Finnish doesn't have a perfectly phonemic orthography: it doesn't write the underlying segment at the end of certain words which surfaces as either glottal stop (before an initial vowel in the next word) or as consonant lengthening (of an initial consonant in the next word).
    – TKR
    Commented Sep 28, 2014 at 16:52
  • 3
    @TKR It also has an entire grapheme (d), which doesn't correspond to a single phoneme in more than a few select dialects, Helsinki Finnish being of course one of them. In most other dialects, the weak-grade variant of /t/ conditionally maps to several different phonemes, just as is the case for /k/. Commented Sep 20, 2015 at 11:11

There are many languages that have true phonemic orthographies. An example is Wajarri (Pama-Nyungan family, Western Australia), which has a one-to-one correspondence between graphemes and phonemes.

As Wajarri has no voicing contrast but uses the "voiced" series to represent the plosives potential ambiguity between the velar nasal /ŋ/ and the sequence alveolar nasal + velar stop /nɡ/ is dealt with by placing a full stop between the two letters in the latter case i.e. ng vs. n.g. Many other Australian languages also have highly phonemic orthographies. Some other languages said to have phonemic orthographies are: Zambian languages and Austronesian languages (such as Malay and Maori).

Some European languages with very shallow orthographies are: Finnish, Croatian, Serbian. According to Seifart Croatian and Serbian are particularly shallow, even representing allomorphy according to the surface realisation, unlike Finnish which, while very shallow, does not show allomorphy.

Many of the languages with phonemic orthographies have comparatively little morphophonological variation and have had their orthographies developed relatively recently. As languages change, orthographies also must change or they will become gradually less phonemic. Languages with complex morphophonologies cannot be represented with a shallow orthography, though it's clear that English orthography is much more complex than it need be.

  • 5
    Doesn't Serbo-Croatian have distinctive pitch accent, which isn't represented at all in the orthography? Commented May 23, 2012 at 23:20
  • 2
    @Mechanicalsnail You're quite right, the standard orthography does not show pitch-accent and so is not perfectly phonemic. Having said that, there are not many minimal pairs for the pitch-accent and there is a standard set of diacritics that are typically used in dictionaries and are (I believe) understood by speakers. Commented May 24, 2012 at 2:50
  • 3
    Also, Malay orthography doesn't distinguish the phonemes /e/ and /ə/, and doesn't distinguish hiatus between adjacent vowels. Commented Jan 26, 2013 at 23:37
  • 1
    Hi @CJDennis, no language now spoken had a written form until relatively recently (leaving aside languages such as classical latin, ancient greek, etc), the last 1000 years or so. And most of the languages that are now written only came to be so within the last 100 years. Anyway, the OP made no mention of excluding recently developed writing systems. Commented Oct 5, 2015 at 22:24
  • 1
    I am no linguist but Croatian is my first language and it is so phonemic that I used to find it a bit odd that other languages had different spelling and pronounciation when I first started learning them. Diacritics are used always, written and pronounced, they sound very different and can be distinguished easily when pronounced. Commented May 17, 2019 at 19:12

I'm generally skeptical of any claim of a perfectly ideally phonemic script for any natural non-engineered in-use language. I don't know very much about the potential languages suggested as answers (except Hindi), but even if there were such a language I wouldn't imagine that the ideally phonemic status would last very long as dialects develop and as the language changes over time.

If a language has few speakers, then they start actively trying to preserve it. In this case, you might expect to see less (natural) change and variation. So if this language had a commonly-used ideally phonemic script, it might stay that way until the language finally died out...

By the way, I'd also be skeptical of any claims of an ideally phonemic script coming from native speakers of that language. Not to say that the claim is necessarily false, but there's clearly potential for a huge bias to be a confounding factor. On the other hand, I would imagine that languages/scripts for which this claim is made might have more transparent orthographies than, say, English; native speakers of other languages tend to come to English and (rightfully) complain about how obtuse the grapheme-to-phoneme mapping is, but then somehow imagine that theirs is perfect.

In actuality, they've just become accustomed to the idiosyncrasies of their own language script and don't see them anymore. (This happens in English too, but the orthography is so obviously opaque that you'd be hard-pressed to find a native English speaker who thinks it's a perfect mapping.) For example, I've often heard the claim that Hindi has a perfectly transparent orthography, but this is plainly false (and yes, I speak it, but not as a native speaker). Schwa deletion is the biggest and most obvious example of a mismatch between the spoken and written counterparts of a word or name.

If you'd like something empirical, take a look at Learning Pronunciation Dictionaries: Language Complexity and Word Selection Strategies. They look at, among other things, the number of letter-to-sound rules required for a language and how it changes with vocabulary size.

  • 2
    In many languages the spelling changes as the pronunciation changes. There were recent spelling reforms in Russian, German and other languages.
    – Anixx
    Commented Oct 19, 2016 at 7:41
  • Even without reforms you just change the way how you write the particular word. You can write the word differently in different dialects. Commented Oct 19, 2016 at 11:19
  • 2
    I second that bias is a thing. Many germans I talked to considered german orthography as highly phonemic, which isn't really, it's just not as much irregular as English, but it's more than Portuguese or Italian for example. Commented Apr 27, 2019 at 12:52
  • 1
    @Anixx Off the top of my head I can’t name any change in the 1996 reform that adjusted a non-phonemic spelling to a more phonemic one; most changes were between letters that would be pronounced in the same way in a given word.
    – Jan
    Commented Nov 13, 2019 at 9:16
  • This link is broken: Learning Pronunciation Dictionaries: Language Complexity and Word Selection Strategies. Commented May 13, 2020 at 0:51

Abugidas are going to be more phonetic than other systems. Look at Indian languages. Still, there are some languages there that have ambiguity in pronouncing letters based on position. E.g. Tamil pronunciation of a letter as voiced or unvoiced depends on position of the letter as well as some conventions. Kannada on the other hand is 100% phonetic. No all people speaking Kannada might know how to pronounce those letters that were borrowed from Sanskrit, but the writing system is definitely phonetic.

  • 4
    I think you mean "phonemic", rather than "phonetic"? Commented May 27, 2012 at 23:39
  • 1
    Virtually no one distinguishes between "ಋ" and "ರು" or "ರಿ" however. In fact, there literally may be one hundred people in the entire state of Karnataka who can explain the difference between "ಋ" and "ರು" or "ರಿ" (maybe not that few, but a similarly low number). Therefore, in practice, "ಋಷಿ" sounds the same as "ರುಷಿ" or even "ರುಶಿ."
    – user67444
    Commented Aug 10, 2015 at 21:55
  • 2
    There’s nothing inherently phonemic or phonetic about abugidas. The archetypal abugida, for example, the Amharic fidäl, is exceedingly non-phonemic, completely ignoring several phonemic distinctions (geminate consonants and /ə/ vs no vowel), so the same three fidäl ግብር can in theory be read gəbr, ɡəbər, ɡəbbər, gəbərr, gəbbərr, ɡəbrə, ɡəbərə, ɡəbbərə, gəbərrə or ɡəbbərrə. Commented Jul 19, 2020 at 21:47

As far as I'm aware Georgian (https://en.wikipedia.org/wiki/Georgian_scripts) is 100% Ideally phonetic language. Being pretty small and neglected countries, people often forget or don't know about the history of one of the oldest civilization on earth with its own unique alphabet and language. Every letter, every word is pronounced as it is written in Georgian. There are no exceptions. What you write is what you say. No different pronunciations of letters. To me as Georgian, it was always very unusual for me to see all these other languages with their very weird sets of rules (from my point of view). While the grammar is very difficult (It is extremely rare for someone who isn't native you speak Georgian correctly, even after years of experience), reading is very easy. What you see, is what you say. I've tried to find other languages like Georgian in this regard, while some languages come close, I found nothing that is 100% phonetic.



The Hawaiʻian language is known for its very small consonant inventory: there are only eight generally-accepted consonant phonemes, each with its own letter in the orthography (P, K, H, M, N, L, W, and ʻ).

The ten* vowels are also represented unambiguously in the orthography. Each short vowel (A, E, I, O, U) is given its own letter, while the long vowels are marked with macra (Ā, Ē, Ī, Ō, Ū).

Thus, each grapheme designates exactly one phoneme, and each phoneme is designated by exactly one grapheme.

On the flip-side, this doesn't mean it's always clear how a word should be pronounced. One side effect of the small consonant inventory is a LOT of free variation; free allophones of /k/ include [t], [s], [d], [z], [ts], [c], [tʃ], [ʃ], [g], and [x], for instance. So when place names and loanwords are borrowed into other languages, they tend to be written more phonetically than phonemically.

* (This number is not entirely agreed-upon. It comes down to whether or not diphthongs and geminated vowels are phonemes. Since diphthongs are still formed productively, some linguists consider there to be only five vowels in the underlying representation. Others count the long vowels as separate, giving ten. A third camp categorizes each diphthong as its own phoneme, for a total of 25. Luckily, these different interpretations agree that the orthography is phonemic; they disagree on whether or not digraphs are necessary.)

  • What do you mean with it not being always clear how a word should be pronounced? You say those many consonantal varieties are in free variation, which I understand as meaning you can pronounce them in any of those ways... if they are allophones constrained by context, then I think it's not valid to say they are in free variation (but the spelling would still be phonemic if the right allophones can be picked by deterministic rules).
    – LjL
    Commented May 27, 2019 at 18:02
  • 4
    @LjL It is truly free variation, to my understanding: it's valid to pronounce the word in any of those ways. My meaning with "should" is more "an English-speaker will often hear native Hawai'ian-speakers speaking in ways that don't seem to line up with the phonemic spelling", just because speakers of most languages won't expect [d] and [k] to be in free variation.
    – Draconis
    Commented May 27, 2019 at 18:46

Lojban is designed to have a one-to-one mapping: audio-visual isomorphism

  • 6
    Lojban is an artificial language, so it's a bit off topic for this list. Commented Jul 4, 2012 at 4:30

Why do folks in all walks of life neglect Hawaii? If you count the okina; a glottal stop, represented by a apostrophe, and the kahako (macron), and a few diphthongs, the Hawaiian language is highly phonetic. Some say perfectly phonetic.

Me ka ha'aha'a (with humility) La'i

  • 4
    Welcome to Linguistics.SE! This answer would benefit from more details and authoritative sources Commented Oct 18, 2016 at 20:54

Esperanto. Its writing system is phonemic in relation to the commonly used pronunciation, including the pronunciation used by native speakers. Moreover, for some speakers the script can be considered phonetic, as they try to avoid allophones and pronounce e.g. banko as [banko] and not [baŋko], even though the latter form is not generally considered to be incorrect.

  • That's a constructed language so it doesn't count Commented Jan 9 at 3:58


Here are some corollaries from graphemics (or grammatology if you prefer), which absolutely is a proper field of linguistics for some of the reasons listed:

  • Writing came after speech phylogenetically (i.e. historically).
  • Writing comes after speech ontogenetically (i.e. individually).
  • Writing was made to represent speech.
  • Writing does not represent speech.
  • Writing and speech correspondent with each other.
  • Writing can transcribe speech.
  • Written language is not transcribed spoken language.
  • Written language evolves in and by itself.
  • Written language follows some rules that don’t apply to spoken language – and vice versa.
  • Written language and spoken language influence one another.
  • Written language evolves slower than spoken language.
  • Written languages may share the same script.
  • Written language could use any script.
  • Written language cannot be regulated by a central authority, but orthographies may.
  • A writing system needs at least one of each: script, orthography, language.
  • An orthography is a set of restrictions to and exceptions from the graphotactics of a written language; mostly it’s the collection of choices of one of several valid systematic alternatives.

It follows that it’s neither possible nor the goal of a writing system to be a dependent 1:1 mapping of phonemes (or morphemes) to graphemes. It wouldn’t be “ideal” in any way.

  • I don't understand your argument. Sure it isn't possible to have a perfect phonetic script, but phonemic? Why wouldn't it be? Many languages have very few phonemes.
    – curiousdannii
    Commented Jun 27, 2014 at 10:33
  • 2
    @curiousdannii, have you ever seen 2 linguists agree on the exact number of phonemes of any given language? The number also isn’t relevant at all. When a language has very few phonemes it can actually make sense for its writing system to correlate many graphemes with a single phoneme, because this may help to disambiguate homophones, since words more often appear isolated in writing than in speech. Take Chinese and Japanese, though very different languages: their writing systems benefit greatly from sinograms (which in bulk are neither logograms nor ideograms but rather syllabograms).
    – Crissov
    Commented Jun 28, 2014 at 13:59
  • 2
    The real counter arguments, however, are language change and dialects.
    – Crissov
    Commented Jun 28, 2014 at 14:01

Serbian language is built on phonemic priciples

Serbian is practically the only European standard language with complete synchronic digraphia, using both Cyrillic and Latin alphabets; speakers read the two scripts equally well. The Serbian Cyrillic alphabet was devised in 1814 by Serbian linguist Vuk Karadžić, who created the alphabet on phonemic principles.

Reference: Serbian Language Wiki

  • 10
    Serbian orthography isn't perfectly phonemic: in some cases it is morphological (bratski instead of bracki), and is always so between words (s bratom instead of z bratom). The biggest obstacle is that the standard accentuation and length differences aren't recorded in the writing, so that for example pȁra, pàra, pȃrā and pȃra are almost always written simply as para, with parâ as a reasonably well understood and sometimes used substitute for the third.
    – user54748
    Commented Sep 18, 2015 at 2:46
  • 3
    great stuff @user54748, maybe you can further extend the answer?
    – maljukan
    Commented Sep 18, 2015 at 12:22

A 100% match would be extremely difficult - every language contains exceptions to the usual pronunciation rules (not only from loan words and often used foreign words), so many fine nuances may only appear in a few words (often because it's difficult to pronounce 'correctly' due to surrounding letters and stops).

Not to mention that a lot of differences in pronunciation are situation based and usually not written down in the words - like raising the pitch in a question, for instance.

A good phonemic/phonetic writing system would basically have to include all of IPA and then some more for the cases where even linguists aren't sure, which also means a lot of redundant letters and a very difficult to learn writing system. And with all the stresses, pitches, length variations and so on, it would still sound mechanical if we fed it to a speech synthesizer, and there would still be errors when a computer wrote down speech, even with the best possible hardware and under laboratory conditions.

A phonemic writing system which works with the way locals pronounce their words and foreign words would reduce the complexity quite a bit, but at the expense of being unable to pronounce foreign words correctly. Even this reduced set would be quite large (Expressed mathematically: half of infinity is still infinity).

Simplifications are accordingly in order. Things which would change a word's spelling for a kind of pronunciation which is situational can often be left out. In western languages, we can usually leave out pitch, while a tonal language needs it as there are lots of words where the tone makes the difference quite literally. In other languages, there are probably other things which we can do with punctuation or such or which are usually clear from the context.

Also, simplifying sounds which natives usually don't distinguish clearly or where there might be slight differences depending on region can make the writing system more useful.

Some languages have often no difference between voiced and unvoiced versions of a sound - German with s, o and a few others, for instance. When the rules of the language make it clear where to use which version, even that should be ok.

For larger differences in pronunciation, I'd prefer different letters personally, but some linguists seem to find it more important to make words spelled the same even across minor region differences. Iirc, there was a letter introduced into cyrillic in some country which would allow for two different pronunciations of the same words where there was a common difference in pronouncing a part and therefore an inconsistent spelling of those words.

English speakers seem to find it more important to stick to the roots of words and ignore the actual pronunciation in their spelling. Similiar to French, though in slightly different ways. I personally would prefer new words to be added where the difference in pronunciation has become too big. So I'd add "taff", "nyu", "teibel" and many more... :)


Macedonian and Serbian are completely phonetic languages. I am Macedonian, therefore will give examples from it.

The Macedonian language has 31 letters, and each of them is a represented with a well defined and distinguishing sound. The official alphabet is Cyrillic, and today, 24th of May is the day of Saints Cyril and Methodius, who wrote the first Glagolitic alphabet.


| English   | Macedonian  | Cyrillic|
| letter    | bukva       | буква   |
| alphabet  | azbuka      | азбука  |
| Cyril     | Kiril       | Кирил   |
| Methodius | Metodij     | Методиј |
| George    | Gjorgji     | Ѓорѓи   |
| love      | ljubov      | љубов   |

The first 4 examples are quite simple.

The 4th one, George in Macedonian is Gjorgji, where Gj is one letter in Cyrilic ѓ.

The 5th one, love in Macedonian is ljubov, where lj in Cyrillic is one letter љ

Bulgarian is also close to 100% phonetic, but it has a couple of letters that are pronounced like two sounds, i.e. я (/ja/, /jɐ/, /a/ or /ɐ/), щ(/ʃt/), ю(/ju/, /jo/, /u/ or /o/), and also in order to make the sound J, like in John, you need 2 letters (Дж), so John = Джон, while in Macedonian and Serbian is Џон.

There are some Romanesque language that are close to 100% phonetic, like Spanish and Italian, but they have some rules where one letter may sound different depending of its position in the word, and sibling letters.

Examples: The g letter has different pronunciation in Jorge /ˈxoɾxe/ and Gato /ˈɡa.tu/ in Spanish.

The c letter has different pronunciation in both its occurrences in Calcio /ˈkaltʃo/ in Italian.


Sinhala (used in Sri Lanka) is a fully phonetic language. Sinhala language


Traditional (orthodox) Judaism prescribes, to some extent, the pronunciation of the Hebrew in prayers and the communal ritual reading of the Bible. Therefore, even though there are distinct phonemes with the same orthography (letters + Tiberian pointing), such as quiescent vs. mobile schwa, dagesh lene vs. dagesh forte, and broad vs. small qamatz, there are a number of recent publications (Bibles and prayer books) that distinguish visually all such pairs, thereby effecting a writing system that marks as different all distinct phonemes in the speech (albeit deliberate) of the books' audience.

  • 1
    @hippietrail e.g. books.google.com/…
    – msh210
    Commented Jun 29, 2012 at 15:49
  • 1
    That book doesn't differentiate dagesh lene from forte AFAIK.
    – Double AA
    Commented Apr 25, 2013 at 17:57
  • 1
    There are several different liturgical pronunciations of Hebrew. In the Yemenite one, for instance, all letters (both with and without dagesh) have a different sound. Whether that was the case as well in Ancient Hebrew is still a matter of debate.
    – Joe Pineda
    Commented Nov 27, 2013 at 13:33
  • 1
    European versions of Hebrew (whether Sephardi or Ashkenazi) fail to distinguish several pairs of sounds which were originally distinct), so in those renderings there are multiple letters for the same phoneme.
    – Colin Fine
    Commented Sep 28, 2014 at 16:00
  • 2
    There is no 1 to 1 correspondence in Hebrew/ Two letters for /t/, a lot of different chartacters for the same vowels etc.
    – Anixx
    Commented Sep 29, 2014 at 6:15

I was searching for phonetic languages and realized most people don’t know Turkish is a phonetic language. If you know how the read the letters correctly, you can write anything you hear and read any written word correctly. It is probably more phonetic than Spanish. (For example in spanish the letter “c” is pronounced differently depending on the next letter, but in Turkish all the letters sound the same no matter which letter comes before or after it. Also, French is definitely not phonetic. Piquenique (French): Piknik (Turkish) while we are trying to learn how some English words are pronounced we just write them phonetically and everyone can read it :D for example: How are you? (Hav ar yu?)

  • You might not know where a yumuşak ge should be inserted, though. Silent letters are always a problem for non-natives.
    – jlawler
    Commented Jun 5, 2022 at 16:29
  • While Turkish is comparatively phonetic, it's not fully. E, G, Ğ, K and L all change pronounciation based on their position. Commented Jan 9 at 3:54
  • Ke ([ce]) is not the same as Ka ([ka]) and Ge ([gye]) is not the same as Ga ([ga]). There are other irregularities too such as â and a sounding the same although â is used to distinguish palatalized and unpalatalized sounds However, you could just write it as Kyar instead of Kâr Commented Jan 9 at 4:18

Hindi (I haven't come across any exceptions in my life, and I'm a native speaker)

Hindi is the Official language of India and so is English

The Devangari script employed by Hindi contains both vowels (10) and consonants (40) and is characterized by bars on top of the symbols. Hindi is highly phonetic; i.e. the pronunciation of new words can be reliably predicted from their written form. This is in strong contrast to English, with the result that Hindi learners may struggle with English spelling. Conversely, they may mispronounce words that they first encounter in writing.

The phonetic correctness of Hindi has originally lead to the famous "Indian" accent, since Indians try to pronounce every single vowel in a certain specific way.

The high number of alphabets used in the script ensure the phonetic correctness of the language.

The famous English word "Mercedes" if pronounced has three different sounds, denoted by 'e'. Hindi has three different vowels for these the three sounds

Not just Hindi but many other Indian languages are very phonetically accurate (also some from the surrounding nations), especially those that use the Devanagari script and its sisters (like Marathi)

You can visit the following site to know more about the phonetics : https://ielanguages.com/hindi-pronunciation.html

  • 3
    How about व, which might be pronounced [ʋ], [w], or [v] depending on context? What about schwa-deletion: राम is pronounced [ra:m] instead of [ra:mɐ]. हरकत is [harkat] but सरकना is [sarakna] even though both contain "रक". Hindi/Devanagari is very close to phonetic, but not perfectly so. Commented Apr 23, 2019 at 20:46
  • 1
    Hindi is already mentioned in a few of the answers below. The comments on them seem not to agree that its orthography is phonemic
    – b a
    Commented Apr 24, 2019 at 12:03
  • 2
    @MarkBeadles There is only "v", in Hindi language. The letter "w" does not have an equivalent in the language. But, does that mean it is phonetically incorrect? No Commented Apr 24, 2019 at 15:50
  • 1
    @ThePerson how do you address the schwa deletion mentioned by Mark Beadles? I have very little experience with Hindi, but I immediately noticed that when I looked at it. Are those always predictable from immediate rules?
    – LjL
    Commented May 27, 2019 at 18:10
  • 2
    Yeah schwa deletion is not predictable easily, it is still an open problem. Hindi is definitely not orthographically transcript. I have a recent paper on the topic using a machine learning approach that is under review, and we achieved about 98% accuracy on schwa deletion on our dataset (which is still not perfect).
    – Aryaman
    Commented Feb 12, 2020 at 3:09

Not the answer you're looking for? Browse other questions tagged or ask your own question.