Difficult languages and easy languages, part 2
« previous post | next post »
On March 4, 2017, I posted on "Difficult languages and easy languages". The response was overwhelming — there were 151 comments.
First of all, I want to thank everyone who participated in this survey. The large number of respondents who contributed their thoughtful appraisals means that the results do carry a certain degree of significance.
Considering the fact that tabulating the results was a rather daunting, time-consuming task, I was not able to post them as quickly as I had hoped. The main reason that I was able to finish the work at all is simple: although Cathay Pacific has wonderful service, they do not have Wi-Fi, at least not on the planes I flew to and from Hong Kong in late April of 2017. Consequently, during the nearly 30 hours of my flights back and forth across the Pacific to review the Translation Department of the Chinese University of Hong Kong, I was able to concentrate on recording the figures on the pages of the survey I had printed out and brought with me. Further delays since then were the result of the press of teaching and mentoring, writing blogs and newsletters and articles and books…. Finally, on Memorial Day, May 27, 2019, I was at last able to type up the results (the tabulations were almost lost when my backpack got soaked in a rainstorm two years ago; fortunately, the pages on which they were written were buried deep inside, so they were not destroyed — that would have been the obliteration of weeks of work).
I had originally stated that I wanted to limit the lists that people sent in to six languages so as to keep the project within manageable size. I am grateful to everyone who sent in their ranked list of languages, from Easy to Difficulty, and resisted the temptation to go much beyond six languages.
I realize that this survey cannot possibly yield scientifically rigorous and reproducible results for the determination of easy and difficult languages. There are many reasons why this is so. For one thing, there are many variables, some of which are poorly defined. What do we mean by "difficult" and "easy"? Are we talking about written language or spoken language? I had emphasized at the very beginning that I find spoken Mandarin to be quite easy (for me!) to learn, but written Chinese is extremely challenging. A commenter on the o.p. said, "Written Japanese is not hard; it just needs a lot of memorization." I don't know about others, but for me, who am always pressed for time, one of the definitions of "hard" with regard to language learning is "needs lots of (time for) memorization".
Then there's the problem of the order in which one studies a succession of languages. Especially if they are somehow related, subsequently learned languages may be more readily acquired than the first foreign languages one learns. The age at which one learns a second language is also relevant, since the capacity for learning a second language is greater the younger one is. The environment in which one learns a second language (total immersion? academic / classroom setting?) and the pedagogy applied (emphasis on vocabulary? grammar? memorization? communicative competence?) are also operative.
Another factor that should be taken into account are the differential learning abilities of various individuals with regard to different aspects of language.
In the following list, the two numbers are: average score (level of difficulty, with "1" being easiest and "6" being most difficult) / total votes cast for a particular language).
Here we go:
EASY
Haitian Creole 1 / 2
Malayalam 1 / 2
Afrikaans 1 / 1
Maori 1 / 1
Papiamento 1 / 1
Swahili 1 / 1
Tagalog 1 / 1
Tetun 1 / 1
Tok Pisin 1 / 1
Esperanto 1.25 / 8
Malay / Indonesian 1.3 / 3
Catalan 1.5 / 4
Spanish 1.7 / 61
Dutch 1.81 / 11
Italian 1.93 / 16
Hindi-Urdu 2 / 3
Klingon 2 / 2
American Sign Language 2 / 1
Classical Nahuatl 2 / 1
Hawaiian 2 / 1
Lojban 2 / 1
Marathi 2 / 1
French 2.23 / 65
Nepali 2.3 / 3
English 2.45 / 24
Norwegian 2.5 / 4
German 3 / 64
Swedish 3 / 9
Portuguese 3 / 7
Danish 3 / 4
BCS (Bosnian / Croatian / Serbian) 3 / 2
Koine Greek 3 / 2
Albanian 3 / 1
Bantu 3 / 1
Chichewa 3 / 1
Dari (Tajik) 3 / 1
Farsi 3 / 1
Old Norse 3 / 1
Sinhala 3 / 1
Telugu 3 / 1
Yiddish 3 / 1
Latin 3.17 / 23
Mandarin (spoken) 3.5 / 25
Breton 3.5 / 2
Icelandic 3.5 / 2
Modern Greek 3.5 / 2
Korean 3.71 / 7
Hebrew 3.75 / 16
Welsh 3.75 / 4
Mongolian. 4 / 1
New Testament Greek 4 / 1
Pali 4 / 1
Persian 4 / 1
Tamil 4 / 1
Japanese 4.1 / 20
Irish 4.12 / 8
Russian 4.14 / 35
Cantonese (spoken) 4.25 / 4
Thai 4.33 / 3
Finnish 4.5 / 2
Scots 4.5 / 2
Ancient Greek 4.6 / 13
Polish 4.66 / 3
Arabic 4.69 / 13
Czech 4.75 / 4
Turkish 4.77 / 9
Georgian 5 / 3
Estonian 5 / 1
Old Javanese 5 / 1
Pashto 5 / 1
Sámi (northern) 5 / 1
Ubykh 5 / 1
Chinese (written) 5.11 / 9
Sanskrit 5.18 / 11
Old Irish 5.33 / 3
Hungarian 5.33 / 3
Vietnamese 5.83 / 6
Literary Sinitic 6 / 2
Navajo 6 / 2
Armenian 6 / 1
Burmese 6 / 1
chiShona 6 / 1
Crow (Plains) 6 / 1
Maltese 6 / 1
DIFFICULT
Compare our rankings with those of the Foreign Service Institute (FSI) — it is remarkable how closely the two tally with each other:
"Language Difficulty Ranking":
| Category I: 23-24 weeks (575-600 hours) Languages closely related to English |
|
|---|---|
| Afrikaans Danish Dutch French Italian |
Norwegian Portuguese Romanian Spanish Swedish |
| Category II: 30 weeks (750 hours) Languages similar to English |
|
| German | |
| Category III: 36 weeks (900 hours) Languages with linguistic and/or cultural differences from English |
|
| Indonesian Malaysian |
Swahili |
| Category IV: 44 weeks (1100 hours) Languages with significant linguistic and/or cultural differences from English |
|
| Albanian Amharic Armenian Azerbaijani Bengali Bosnian Bulgarian Burmese Croatian Czech *Estonian *Finnish *Georgian Greek Hebrew Hindi *Hungarian Icelandic Khmer Lao Latvian |
Lithuanian Macedonian *Mongolian Nepali Pashto Persian (Dari, Farsi, Tajik) Polish Russian Serbian Sinhala Slovak Slovenian Tagalog *Thai Turkish Ukrainian Urdu Uzbek *Vietnamese Xhosa Zulu |
| Category V: 88 weeks (2200 hours) Languages which are exceptionally difficult for native English speakers |
|
| Arabic Cantonese (Chinese) Mandarin (Chinese) |
*Japanese Korean |
The second commenter* to the FSI thread notes: "Chinese isn’t as hard as people think it is and although the writing system IS a nightmare, the grammar couldn’t be any simpler." I couldn't agree more. If you ask me, "What's the easiest language you've ever learned?", I will immediately say "spoken Mandarin". If you ask me, "What's the most difficult language you've ever tried to learn, I will answer without hesitation, "Literary Sinitic / Classical Chinese". I will write a separate post to explain why that is so.
[*Their entire comment, though opinionated, is also insightful on languages such as Japanese, Italian, German, Turkish, and Arabic.]
It is noteworthy that most languages fall in a range of around 3, the midpoint on the scale of difficulty. Also worth pointing out is that creoles consistently score among the easiest languages to learn.
Update: Here are the FSI's Language Difficulty Rankings in map form.
Victor Mair said,
May 28, 2019 @ 12:30 pm
A late addition by Jason M, who was not able to respond to the original post in early March, 2017:
Somehow missed the easy-hard languages survey before it closed so will post here because I have been waiting for something like this for years (making me doubly incredulous I missed the original survey…might have been posted when I was largely off the grid in Mexico). I have told many people that Mandarin (spoken) is by far the easiest language to learn to say stuff in — not necessarily understand stuff said to you — that I have ever studied. My wife told me, though, that to native English speakers, that makes me sound like some kind of arrogant SOB. When she gets out of the shower, I will show her this proof that I am not the only one!
Mine:
EASY
1) Mandarin (speaking it yourself)
2) Swedish
3) French
4) Russian
5) Czech
6) Telugu
HARD
I am sure the long survey thread that is now closed has posts discussing nuances of language easiness-difficulty spectrum. I will post mine here, because that thread is now closed. First, there is both a spoken vs. written divide (I have only really given Chinese characters serious consideration because I realized knowing a few key ones helps me parse Mandarin TV with subtitles) as you mention, and a hearing-speaking divide. Mandarin is far easier to speak than comprehend orally because of the limited phonemes resulting in your brain having to parsing and weighting their sequence of occurrence as much as actual occurrence. French is far easier to speak (accent aside) and read than to understand spoken to you (limited phonemes plus aggressive liaisons of words hindering meaning extraction). Another nuance is the easy to learn to communicate vs. hard to speak like a native. English is a great example of that phenomenon being simple for basic communication but, with quirky grammar and preposition usage, very hard to master. A final obvious nuance: one's native language (Bulgarians is far easier for Russians than English speakers)
My unofficial list clusters languages at more or less similar levels so I can cram in more of them for fun. It splits languages where some of the nuances above makes for a split in ease of learning, and it is based on my personal experience (native English and childhood Russian speaker with subsequent forgetting of Russian and then formal schooling in Russian and French, other languages studied for travel and for speaking with immigrants):
1) Chinese (spoken)
2) Swedish, Spanish, Italian (basic oral communication, reading), Portuguese (all but oral comprehension at least in Brazil)
3) French, mastering speaking/writing in Italian, understanding spoken Portuguese, Bulgarian (basic communication for a Russian speaker), reading Japanese hiragana or katakana
4) German, Russian (for me with extensive Russian background still very hard), Hindi (basic reading/writing/oral), other related Northern Indian Indo-Aryan languages
5) Czech (definitely much harder than Russian with grammar even worse and fewer loan words from western languages, Latin alphabet infuriates rather than helps)
6) Written/mastery of Chinese, Japanese, Telugu (probably other Dravidian languages, too), Vietnamese
I suspect if I worked more in Japanese, the basic spoken aspect would move towards easier, but, in addition to the kanji problem, the problem with Japanese understanding is that the phoneme conveying meaning is buried in many syllables detailing relationship/politeness/respect/sex of speaker. Then there is the word order also. Telugu is horrible in every aspect I've worked on. Resources are few, English, Sanskrit-derived, Hindi words are thrown in randomly by different speakers, like Japanese many extra syllables that are irrelevant to meaning, and Google translate never seems to work.
Some of my impressions are colored by short trips where, for example, the Portuguese I could read reasonably well did nothing to help me understand what anyone was saying for the week I spent in Brazil, even if I could make myself understood speaking sans problem.
Ash said,
May 28, 2019 @ 1:32 pm
1. It seems like the level of difficulty is highly tied into the learner's native language and languages they've already studied and/or mastered. Spanish is easy for a native English speaker, but difficult for a native Mandarin speaker. Cantonese is relatively easy for a native Mandarin speaker, but difficult for a native English speaker. Without having looked into it, I would think that's the reason for the differences between the results here and the FSI levels.
2. I would heartily disagree that spoken Mandarin is easy.
Tones:
Unless you have a highly skilled teacher — which would very likely be a non-native speaker that has mastered tones (and there are very few of those), learning tones is going to be super difficult.
"Lack of grammar":
Beyond that, the "lack of grammar" is a hindrance, not a help. There is nothing to grab onto in a sentence. If there is a 4 syllable combination that you don't understand, it could be 4 one-syllable words, 2 two-syllable words, 1 one-syllable and 1 three-syllable word, etc. As a native English speaker, I learned Dutch to fluency and German to semi-fluency before even starting on Mandarin. It probably took 5 to 6 times the amount of work to gain fluency in Mandarin.
Just from the Spanish I took as a kid, I can recognize verbs, adjectives and nouns even though I don't speak Spanish. No such luck with Mandarin (until after having spent a gazillion hours on it).
A good way to test this theory is:
Open a speaking-only Mandarin school and a speaking-only Spanish school. Put a bunch of English speakers into both and come back 6 months later to see their progress. I money would be on the Spanish learners.
Neil Kubler said,
May 28, 2019 @ 1:56 pm
When I worked at FSI (the U.S. State Department training institution that came up with Categories I to V that Professor Mair describes above), we on 2 or 3 occasions had to curtail students' training assignments from 88 to 44 weeks due to illness on the part of the officers at post whom they were being trained to replace. To help them attain the highest possible level in speaking in the shortest amount of time, we transcribed into Pinyin all the Chinese characters in their speaking and reading textbooks and materials (including newspaper readers, so they could discuss international news, etc.). The students were able in half the normal amount of time to reach the regular training goal of S-3 in speaking (though obviously not in reading, which must be in the official Chinese writing system, which currently is still Chinese characters). This experiment seemed to demonstrate, though admittedly with a very small sample, that if Chinese were romanized beginning tomorrow (which it could be, if written more or less as spoken), then training time for native speakers of English could be cut IN HALF!
cameron said,
May 28, 2019 @ 2:07 pm
That Persian ended up with the same average difficulty ranking as Tamil, and only a slightly lower ranking than Japanese or Thai, is a good indicator of how small sample sizes skew the results.
I didn't participate in the original survey, but I can assure you that learning Persian is much easier for a native English speaker than Tamil, Thai, or Japanese.
Scott P. said,
May 28, 2019 @ 3:39 pm
Note that "Koine Greek" and "New Testament Greek" are the same thing, though listed seperately.
GeorgeW said,
May 28, 2019 @ 3:45 pm
Comedy is easy. Arabic is hard.
David Marjanović said,
May 28, 2019 @ 6:18 pm
As tone languages go, Mandarin seems to be one of the easiest ones, because it has only one level tone. Cantonese has three that really differ only in pitch. And where Mandarin has just one rising tone, Cantonese has two that also differ only in pitch. I find that scary to imagine. (Haven't actually tried to learn Cantonese.)
The grammar is so different from Standard Average European that it takes a while to wrap your mind around it. For instance, there are no relative pronouns – things like "the one who" are handled completely differently. "And" seems like a single concept to us, or at most like two, but in Mandarin it's something like ten. At first glance Mandarin appears to have both pre- and postpositions, but then it turns out the prepositions are really verbs while the postpositions are really nouns. The copula is not a verb, which explains some peculiarities of its usage and lack thereof. And while there's no grammatical gender, the sheer number of classifiers puts another burden on the memory.
Chris Button said,
May 28, 2019 @ 7:54 pm
I find pitch distinctions more challenging than contour tones too since they're so context specific (i.e. the pitch level is only distinct relative to what comes before or after). I think the best way to teach both kinds is by making the learner acutely aware of how they use pitch and contour distinctions for intonation purposes in their own native language and then just show them how to apply such things lexically.
Gwen Katz said,
May 28, 2019 @ 10:51 pm
Small sample sizes are definitely skewing some of the top and bottom numbers–for instance, Tagalog rated 1 when FSI rates it IV. I suspect the less common languages were more likely to be listed by polyglots who would tend to find them less difficult.
Another factor is that the languages were only ranked, not rated, which skews the results for people who learned languages that were all easy or all difficult. For instance, if someone learned French, Spanish, and Italian, those would be rated 1, 2, and 3 (or 1, 3, and 5?) even if they were all easy to learn.
Chandra said,
May 29, 2019 @ 12:06 am
Interesting to see these results, regardless of how scientifically accurate they may or may not be. Thank you for the time you put into tabulating this!
Jim said,
May 29, 2019 @ 12:32 am
I'm always suspicious of such surveys because they are so subjective in nature. That FSI survey is based entirely on the perceptions of native-English American speakers, for example.
And another thing: in the list, does 'Scots' refer to the language parallel to English which started developing as an independent variety after about the 8th c. AD, or to the Celtic language kin to Irish and Manx, which should be referred to as 'Gaelic' (the two terms are too often used interchangeably)? Surely the level of difficulty would be very different in each case.
astrange said,
May 29, 2019 @ 2:18 am
Scots is Scots, not Scottish Gaelic. It's very useful because you can use it to read Scottish Twitter.
Philip Taylor said,
May 29, 2019 @ 3:37 am
"Scots is […] very useful because you can use it to read Scottish Twitter".
And without at least a passing acquaintance with it, much of the subtlety of Oor Wullie and The Broons will pass straight over one's head …
Rodger C said,
May 29, 2019 @ 8:38 am
Returning to what Scott P. pointed out, why is "New Testament Greek" so much harder than "Koine Greek" (the same thing)? I suspect that those who call it by the latter name tend to be divinity students who aren't in their field because of language ability.
Ex Tex said,
May 29, 2019 @ 10:52 am
On the topic of two labels for the same language, why is Afrikaans listed separately from Dutch, whenFlemish is not similarly included?
The said,
May 29, 2019 @ 11:31 am
As a wannabe hyperglot, I found your articles (both parts) and the resulting commentaries fascinating. I consider myself around FSI 4.0 in French, German, and Italian, and have about a two dozen other languages that I have mastered at various levels, from being thrown in jail for blasphemy to survival to "hey, he's not bad." I have been cracking my head against Japanese for 30 years with relatively little success; I'm convinced at this point the only way I could even get to a reasonable comfort level would be to live there for a year or so. But I've also dabbled in other languages such as spoken Mandarin, Albanian, Lingala, Hebrew, Egyptian Arabic, Bulgarian, Romanian, Modern Greek, Korean, Farsi, Croatian, Slovenian, Serbian, and the Scandinavian triplets, all of which I found neither especially challenging nor terribly odd. Turkish I studied as an exercise in self-mortification – the word order and agglutination is a bear. At any rate, I thoroughly enjoyed your writings about a subject so dear to my heart.
Coby Lubliner said,
May 29, 2019 @ 4:31 pm
It was surprising to see "Arabic" (whatever that means) listed as considerably easier than Maltese, considering that Maltese is an Arabic language and has, for Westerners, the additional advantage of a very regular Latin script.
Andrew Usher said,
May 29, 2019 @ 10:34 pm
It would have avoided some of these questions if the languages were ranked only relatively, not absolutely. As I understand it the responders did do so, and the correct (computerised, hopefully!) algorithm could try to extract a univeral ranking through optimisation, though the sample size might still not be large enough. It would certainly, even if imperfect, address anomalies among languages listed only by one or a few.
This could be a response to many of the comments above including the last one. I myself would not venture any ranking, being monolingual.
k_over_hbarc at yahoo dot com
Vanya said,
May 30, 2019 @ 5:17 am
@Cameron, the odd thing about Professor Mair's table is that "Persian" is apparently significantly more difficult than "Farsi".
Vanya said,
May 30, 2019 @ 5:24 am
" Turkish I studied as an exercise in self-mortification – the word order and agglutination is a bear."
I think Turkish is much easier than its reputation. The orthography is fairly simple, it has far fewer grammatical irregularities than Russian or English, and the pronunciation is not that difficult for an English speaker compared to Chinese, Arabic or Polish. The 20th century language reforms may have sheared off a lot of the interesting grammatical quirks and poetic vocabulary but made the language much easier for non-native speakers to acquire.
Rodger C said,
May 30, 2019 @ 8:56 am
*the former name
David Marjanović said,
May 30, 2019 @ 1:00 pm
Standard Dutch and Standard Flemish are about as similar as the various standard Englishes, if not more so. Afrikaans, on the other hand, has a substantially different grammar (not to mention more superficial differences).
On the other hand, Persian/Farsi, Dari, and Tajik seem to differ mostly in the pronunciations of their vowels; and even in that respect, the first two are more similar to each other, yet the last two are listed together.
Scott P. said,
May 30, 2019 @ 2:47 pm
I think Turkish is much easier than its reputation. The orthography is fairly simple, it has far fewer grammatical irregularities than Russian or English, and the pronunciation is not that difficult for an English speaker compared to Chinese, Arabic or Polish. The 20th century language reforms may have sheared off a lot of the interesting grammatical quirks and poetic vocabulary but made the language much easier for non-native speakers to acquire.
My brother, an expert linguist fluent in German and Russian and proficient in a half-dozen other languages, one opined that he thought Turkish was the easiest language he had encountered.
Toni said,
May 31, 2019 @ 2:24 am
I suspect that learning a modern tongue before, or in tandem with, learning its ancestor is the best course of action for any aspiring philologist. I found my ability with Latin was enhanced after I had studied French. Similarly, I started studying Mandarin with the intention of wanting to get to meaty literary texts, and eventually delved into Literary Chinese as well as Cantonese out of curiousity. Learning these different Chineses at the same time, having a diachronic view of the language, was very interesting and kept my motivation up. I found it's easier to grasp the polysemy of Chinese morphemes in the modern tongues; once you've got a grasp on them, moving onto Classical Chinese removes some of the vocabulary burden and lets you focus on the frustrations of grammar.
That being said I count myself lucky as having every advantage. My mother's side of the family is Chinese, so I knew where to go. I started going to casual introductory Mandarin classes at my diaspora community center and studied texts at home with my mother and grandmother. The first year was very difficult, but after that, I could parse sentences and almost read with the Pleco dictionary. Then I moved on to the advanced class downstairs, which was really a social thing consisting of elderly Cantonese women wanting to dust up their Mandarin while getting to see each other and drink tea. The class was accepting of code-switching and test-free. I wouldn't wish jumping straight into Literary Chinese on any human being, what a cruel and bizarre punishment that would be.
While my mother is Chinese, my father is Egyptian. Literary Arabic, despite being exposed to the Egyptian spoken dialect during childhood (but never really acquiring it fully) has caused me enormous suffering. I find the written language ambiguous, a pain to search in a dictionary and impossible to parse. I like to imagine that if I had the same advantages I had in learning Chinese, that of a patient tutor, good introductory books and a sort of immersive environment, I could make some headway.
Alas. Perhaps one day.
Andrew Usher said,
May 31, 2019 @ 7:47 am
Really, classical languages shouldn't be compared to modern ones, it's not the classical languages are easy, but the same standard of proficiency – ability to actually communicate with native speakers, without which one can't be said to 'know' a modern language – is different.
I'm disappointed no one responded to (or understood ?) my point about better ranking languages; I do think that would have some value, and I don't like, either, subjective evaluations of this or that language.
PeterL said,
May 31, 2019 @ 2:39 pm
For listening comprehension, I wonder how much difficulty is due to early training of the brain's processing of sounds and syllables (and I don't mean just tones or specific sounds, but even how syllables/mora are distinguished). I seem to recall a study of bilingual French/English speakers in Montreal, who were equally fluent in both languages but who took slightly more milliseconds when processing French vs English sounds, depending on which language they had used first.
For Japanese, I suspect that native speakers process each CV syllable as a single item ("ka") and a long vowel as two items CV-V ("ka"+"a"); but I (native English speaker) hear C-V and C-VV ("k"+"a"; "k"+"ā"). So, when I listen to Japanese, it seems to be very fast — I should process "ka" as a single item but instead process it as two items "k"+"a"), and I also have difficulty distinguishing short and long vowels. [There's also the pitch accent vs stress accent … I find Ōsaka accent easier to understand than Tōkyō]
As for learning English, Japanese native speakers have a well-known difficult distinguishing "l" and "r"; but also other sounds such as English short "a" and "u" (e.g. "rag" vs "rug").
Fluent listening comprehension requires quick disambiguation; and when your brain processes sounds differently from how a native speaker's brain works, that makes a language "difficult".