Most Difficult Language

Is it possible to quantify translation? Is it possible to assess if one text is more difficult than another? Is it possible to claim that one language is more difficult than another language? The whole idea of quantifiability in language and linguistics has often been shunned and thought to be impossible, however, the advent of corpus linguistics and statistical approaches to language has shown us that there is much that statistics can give us. Certain aspects of language are quantifiable and probabilistically-motivated.

A question that is often put by non-linguists is whether one language is more difficult than another. Interestingly, this question is very rarely posited by linguists and sadly has not been taken up by statistical linguistics. Are linguists afraid to ask this question or do they believe it is redundant? Perhaps this question is marked by a form of linguistic political correctness and it is not ‘right’ to ask whether one language is more difficult than another. A ‘better’ question might be how one language differs from another. Yet we are all aware of the fact that some languages are more difficult than others. Obviously, there are a number of factors at play here, such as one’s mother tongue, how many languages one already knows and the relation of the new language to the mother tongue as well as our own individual predisposition for learning languages. But still, with all these factors taken into consideration, certain languages are easier or more difficult than others.

Opinions are varied with Mandarin (Chinese), Arabic, Polish, Basque and Xhosa seen as the most difficult. But when attempting to answer this question we find that it is in fact redundant with there being as many answers as there are people asking the question. Yet where statistical linguistics fears to tread, second language learning boldly goes. It is fascinating that some language teaching institutions have tried to approach this answer quantitatively. The US Military (as well as other US institutions) use a system which divides languages into four groups from I to IV. French, Italian and Dutch can be found within Group I; German in Group II; Polish, Thai and Hebrew in Group III whereas Arabic, Chinese, Japanese and Korean fall into the Group IV bracket. The methodology behind this system is both simple yet extremely practical with languages in each successive group requiring a greater number of teaching hours than those in the preceding group.

It is interesting how practical demands often create the lines of demarcation that linguists (and translators) are often so loathe to draw. The same, of course, can be said of different types of text which the translator faces. Some believe medical or legal texts to be the most difficult, some maintain that poetry is a much more arduous task but in the thick of the translation jungle and within translation agencies the distinction is much less elegant. Initial quotes by translation agencies are often undertaken based on the distinction of general vs. specialised, express vs. non-express translation. It is only when such (theoretical) distinctions are pushed up against the cold, hard truth of practice do these distinctions come good.


38 thoughts on “Most Difficult Language

  1. I think somebody from Instituto Cervantes once told me that they devised their in-country course syllabuses according to students’ mother tongue – in Italy it took three years to reach the same level that in Japan took ten or so.

    1. The fact is that english is an easy language. Not only because is widely used, but because it doesnt have as many grammar rules as castillian or “spanish” for a more common way to call it (north american way).

  2. The problem is that there’s no objective standard of difficulty against which to measure languages so we’re left looking for secondary effects of difficulty and …. there are none. If some language were objectively more difficult you’d expect that children would take longer to learn it but no such language has been found. Discounting individual outliers, very roughly, kids start talking at about 2 achieve fluency around 4-5 (and may not gain mastery of some kinds of difficulty till around 12).

    IIRC Hjelmslev’s theory of language as a structured network of oppositional minimal units contains a potential model for establishing relative difficulty of languages but no one can (or wants to?) operationalize it.

    The only practical levels of difficulty in L2 learning, have to take the native language(s) of the learner into account.
    The US Military categories are clearly for native English speakers and would make no sense for Russian speakers, for whom Polish would be a Group I language. Japanese and Korean are at most Group 2 languages for each other.

  3. Michael,
    There is a lot to be said for glossematics. Linguists seem to have the adequate tools to do the job but have not even tried putting them to use. Of course, as you mentioned, the speaker’s L1 invariably affects the acquisition of the L2. English is the world language that it is and so (unfortunately for other languages), L2 research focuses primarily on English.

  4. Actually, a fair amount research has been done in this–I’m surprised people in the field of linguistics and languages aren’t aware of this.

    The most commonly cited learning rates cited (for native speakers of English) is the FSI (Foreign Service Institute) scale, where languages of the world are divided into groups. Group 1 languages take 8 weeks (or 240 instructional contact hours) to achieve “elementary proficiency” (these terms are defined precisely by FSI), or 24 weeks (720 hours) to achieve “working proficiency.” This scale is designed specifically for application in L2 learning environments. (More information here:

    The groups, from easiest to hardest (in terms of time) for native speakers of English, are:

    Group 1: French, German, Indonesian, Italian, Portuguese, Romanian, Spanish, Swahili, Danish, Norwegian, Swedish

    Group 2: Bulgarian, Burmese, Greek, Hindi, Persian, Urdu

    Group 3: Amharic, Cambodian, Czech, Finnish, Hebrew, Hungarian, Lao, Polish, Russian, Serbo-Croatian, Thai, Turkish, Vietnamese

    Group 4: Arabic, Chinese, Japanese, Korean

    Thus, it will typically take a native speaker of English about 44 weeks to achieve “working proficiency” in, say, Japanese, while it will typically take only 24 weeks to achieve the same level in French.

  5. Does this list include literacy?

    I’m astounded that Burmese would be in group 2 if learners have to learn a really confusing looking script with very counter-intuitive etymological spellings.

    If it’s just spoken Burmese then maybe, except it’s got some weird phonology and is typologically strong head final with the usual complications that that creates for English speakers. I would have thought that spoken Cambodian would be easier (lots of hard to distinguish vowels but no tones and SVO word order.

    IME the language that’s the farthest conceptual distance from English is maybe Navajo (which being spoken inside the US wouldn’t be on the FSI list).

  6. Burmese is a strange one here but then again, the same could be said for Bulgarian vs. Serbo-Croatian. The FSI is practically-motivated but that does not necessarily mean that the groupings are 100% spot on. Thanks for the link, Masked Translator.

  7. Assuming you are a native-speaker of English, those groups in Masked Translator’s reply (5.) are a good guideline. However, there is one crucial factor missing. Some people learn more than one language from any given group.

    I personally, was brought up with some Dutch as a child. This automatically made German easier to learn at school, as the two languages are closely related. An even closer relative to Dutch is Afrikaans, which is a doddle to read if you know Dutch, but harder verbally / orally.

    When I learnt Swedish at university, it required very little effort to get a reading knowledge of both Danish and Norwegian. However, to this day (some 30 years after starting to learn Swedish) I still cannot understand spoken Danish, although I can read the newspaper with ease.

    And so on. It very much depends where you are coming from. Not all native-speakers of English start at Square One with every new language they try to learn. And there are four distinct skills to be measured when you learn a language: speaking, understanding, reading and writing. You may be able to read a novel in a language, but are utterly tongue-tied if you try to speak it.

  8. Eric,
    Yes, the starting point is invariably the most important and so any categorisation of this kind will always run into problems. In relation to what you mentioned, there are many people, for example, who can read six to eight languages, but not speak a word of any of them. Thus, we are dealing with initial linguistic input (what languages are known prior to education) and what is meant by competence.

  9. I had a boost in my upbringing by having access to the Dutch language (my mum). My mum was a one-off in the area we lived in Yorkshire.

    But what about all those thousands (and I mean thousands) of second-generation Poles in Britain? Those people whose parents arrived in Britain in about 1945. They will have learnt Polish in the days when Brits didn’t think some weird East European lingo was important. Those same people, now in their fifties, can do something to help the next generation of Polish immigrants, integrate – if the latter community wants to stay in Britain, not go back to Poland.

    I too can read six to eight languages. But the vital thing to grasp is that I can read them at different levels of understanding. I can only speak two languages really well: Dutch and Swedish.

    Most British people do not understand that saying ten pat phrases is not the same as being able to read the newspaper in any given language. And being able to take part in a conversation in any language requires quite different skills to the passive activity of reading.

  10. Very true regarding the Poles in Britain. Their linguistic competence is of course very different from the average Brit.
    Good point you also make about the British. Their view of competence is skewed and unlike most people’s view of competence. As we all know, most of the world’s inhabitants are bilingual (and many trilingual), take for instance the inhabitants of India.

  11. Group 3: Amharic, Cambodian, Czech, Finnish, Hebrew, Hungarian, Lao, Polish, Russian, Serbo-Croatian, Thai, Turkish, Vietnamese…. Hi,hi…and I once thought I could learn Polish!! But I did realize Polish could communicate with those speaking Serbo-Croatian and not with Slovenes s (my mother tongue) and I still wonder why, why could we comunicate with those speaking Serbo-Croatian and not with those speaking Polish. DOes it mean that my friend´s friend is my “enemy”?

  12. Interesting point, Ljudmila, very interesting. If we take this line of thinking to its logical conclusion we run into problems. For example, Polish is more difficult than Serbian is more difficult than Slovenian etc. Obviously, this works well from a logical point of view but the results are absurd.

  13. As I implied in a previous posting, these “groups” (as in Posting 5, above) have been invented by some cerebral desk clerk who has probably learnt one foreign language and thinks all the rest “difficult”.

    Many Brits have allowed themselves to be tricked into thinking that they can learn a language in three months or from a set of LPs or CDs, without taking ability and knowledge of other languages into account.

    When you read Group 3: Amharic, Cambodian, Czech, Finnish, Hebrew, Hungarian, Lao, Polish, Russian, Serbo-Croatian, Thai, Turkish, Vietnamese

    how on Earth do people compare the difficulty of learning languages where grammar, alphabets and diacritic marks, culture and knowledge of other languages all play their role? Group 3 is a fantasy, conjured up by a linguist who has totted up a few grammatical factors and left it at that.

    To answer Ljudmila, the Slav languages are divided up into, I think, three groups: eastern, western and southern. Obviously it is easier to understand or learn a language within your own group, but certain factors of sound and grammar will come into play.

    For instance, Polish resembles Slovak more than it does Czech. Ukrainian and Belarusian both look like something halfway between Russian and Polish (and the geography bears this out too). And (Ljudmila will know this better) the ex-Yugoslavian languages are fairly close to one another. And given the fact that older people will have probably had to learn Serbo-Croat when Yugoslavia was one country, surely Slovenians will have had quite a lot of exposure to Serbo-Croat. But why Poles can cope with Serbo-Croat and not with Slovene is a mystery to me too.

  14. Eric,
    Very succinct comments which one cannot help but agree with. No doubt, these groups HAVE ‘been invented by some cerebral clerk’ which in a sense makes the whole discussion pointless. However, what can linguists do to counter this? What can linguists do to counter the literally thousands of products that go onto the market claiming to be able to ‘teach you a language in three months’? Probably very little but awareness of this problem is most certainly a step forward in the right direction.

  15. Actually the FSI rankings are based on practical experience of how long it has taken them to teach Americans the languages in question. There might be some atypical results (where exceptionally good or untalented teachers or students give unexpected results) but overall I can understand the logic.

    Or course languages not taught by the FSI aren’t included and the results might be a little different even for Non-US English speakers (and have little to no relevance for anyone else).

    Bulgarian makes sense as being easier than Polish for English speakers. All else being equal, English speakers will find lots of articles and lots of tenses without much nominal inflection much easier than no-articles, not so many tenses and lots and lots of nominal inflection especially when paired with lots of morphophonemic changes (which Bulgarian also has relatively little of).

  16. Michael,
    Yes, the FSI rankings make sense for English (or more strictly American) speakers, although there do seem to be ‘holes’ in it, but as you stated, the list does not take into account all languages and it is of course from the point of view of those taking part in FSI courses.

  17. Reply to Michael Farris, Message 16:

    Surely, any educationalist in the USA must realise that a country virtually built up from immigrant stock must have very different starting points for all the different immigrant communities.

    In this context, “Americans” don’t exist. If you try to teach a Hispanic Spanish speaker Portuguese, this cannot be compared with trying to teach that same person Finnish. Yet an Estonian immigrant to the States would find Finnish a doddle, but Portuguese hard. And so on. The huge flaw with these gradings is that they assume that every language learner is a native-speaker of English with an average middle-class education.

    Could you explain the logical and empirical base on which the FSI bases its groupings?

    Bulgarian may be easier, from the similarity of noun endings, simplified if you compare with Polish, but a language is also vocabulary and connotations, historical, political, cultural and social. That is why, in some respects, a Finn and a Swede understand one another better than a Finn and an Estonian. The fact that Estonia was part of the Soviet Union, with a Soviet mentality for 50 years, has everything to do with this.

    Linguists tend to divorce languages totally from their historical and social context and concentrate solely on linguistic factors. This creates fantasies, such as the fact that Hungarian is supposed to resemble Finnish. Sure, it does in the bookish world of linguists, with their derivations and morphology, but the two are NOT mutually intelligible, by a long chalk. I know, I have studied Finnish, Estonian and some Hungarian. I bet you the FSI linguists have never done much language learning that involves mutual intelligibility and common vocabulary. They concentrate on the endings of words and a bit of syntax.

    There are no holes but chasms in the groups.

  18. “In this context, “Americans” don’t exist”

    I disagree. Culturally and linguistically I’m American. I would also argue that the great majority of those born and raised in the US have common linguistic and cultural characteristics that separate them from others (including other English speakers, even Canadians). That some of us also have other characteristics is irrelevant.

    I’m not interested in defending the FSI rankings to a great degree, merely pointing out that they’re based on empirical evidence and came out of long experience in teaching the languages in question to Americans (as opposed to pre-formed ideas of what would be easy or hard to teach).

    I would imagine that the rankings are based on learners who were raised monolingually in (US) English. Of course other possible home languages or second languages would have to be controlled for but in establishing the American baseline you don’t take those into consideration.

  19. i think that polish is the easiest language on the world,very logical:) double NEGATION gives negation. no confirmation.
    polish “r” is easier to say than english,,thr” i cant tell nothing about gramatic conjugation declension.boredom but english a an and the are boring too we seldom use them)
    anyone who wants to learn polish is in the better situation than people from poland(majority of the pole are ashamed of speak english when englishmen are by.. we start stutterin or turnin mute :)and deaf and we like when foreigner chat with us polish 9then we havnt to try to say something english(the foreigner speak polish well i havnt heard badlly (for example) strange when i translate english text everything(words are noy declineted) straight taken from dictionary ,,,,well pronucation is problem also.. when you come to poland dont speak(sometimes when people (from abroad) say i love poland we ponder whether they spoke polish or not.pozdrawiam.(polish song must be very good to want to listen to them(maybe therefor we write english songs. it is dreadfull,strange but polish is so that

  20. What none of these surveys/studies takes into account is the difficulty of aural comprehension of a given language. For instance, Lithuanian is an extremely coplex language grammatically, but is much easier to comprehend aurally than, say, Spanish, in which many sounds are nearly identical and, given the rapidity with which it is usually spoken, much harder to develop an ear for.

  21. Not sure about your ideas, Lindquist. The claim that that Spanish has more similar sounds seem far-fetched. Such claims need to be backed up with hard evidence.

  22. For me, Spanish is easier than French even though these are both Romance languages, grammaticaly very similar.

    what’s more, i have learned French for 3 years, and Spanish only 2. Even Catalan language (also Romance lang.), which I’ve never practiced, is more understandable for me than French. Why? I don’t really know. Maybe Spanish ‘sounds’ better…

  23. Written and spoken French are essentially two different languages. Either one seems interesting enough to learn, but learning both as a single language is …. not easy.

    Written and spoken Spanish are basically one language. It’s not that Spanish is easy (in many respects it’s theoretically more difficult than French) but written and spoken Spanish reinforce each other in a way that written and spoken French don’t.

  24. I’ve heard a lot of people here saying English is the most dificult language, and I’ll have to disagree. I find English pretty easy to be honost, most of the grama rules seems pretty consistant, and even though I have some trouble saying “th” or “r” in the middle of words, it is pretty easy to pronounce, too. And I have to think about when it is “is”, “are” or “am”, sometimes. In my language, any of those is just “er” :b
    I’ve heard that Chineese is actually not so difficult, the only challenge is to memorize the charecters,but that there is no gramma. A friend of mine is studied it for a while.

  25. I did realize Polish could communicate with those speaking Serbo-Croatian and not with Slovenes s (my mother tongue) and I still wonder why, why could we comunicate with those speaking Serbo-Croatian and not with those speaking Polish. DOes it mean that my friend´s friend is my “enemy”?

    Perhaps it has something to do with White Croatia being in Kraków area and White Serbia in Poznań area? Poles might have common vocabulary with Croats and Serbs that Slovenes don’t share.

    I wonder if it’s really so easy for a Russian person to learn Polish as it was claimed here. Personally I found French much easier than Russian. Russian grammar and spelling are easy to comprehend, there’s almost nothing to learn in fact, but the huge number of fiend words is hell. It’s like some freaking alternative universe where nearly every word means exactly what it’s not supposed to.

  26. “White Croatia” and “White Serbia” are seen as pan-Slavic appellations rather than historical former regions inhabited by the Croats and Serbs, although this cannot be discounted. The reason might simply be that certain languages are more similar to each other. This might be lexical, grammatical or even phonological. Take the mutual intelligibility (or not) of Norwegian-Danish-Swedish and Czech-Slovakian-Polish, for example.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s