Mutual intelligibility

« previous post | next post »


Assuming no prior, formal study of or contact with the opposite language in a given pair (i.e., one is coming at these languages completely cold), roughly what degree (percentage) of intelligibility would exist between the spoken forms of the languages in the list below?  Naturally, you are not expected to comment on all of these pairs, but knowledgeable assessment of any of the pairs would be both valuable and appreciated.  Feel free to add any other pairs not listed, or to combine a language from any of the given pairs with a language from any other pair.  Unless otherwise noted, the languages listed are the national standards.  If the name of a city or region is given, the reference is to the language spoken in that area.

This is a very long list, but it is only representative of the types of relationships that interest me.  Again, I warmly welcome other relevant pairs.

Forgive me for the fact that there is no particular order to this list.  I just typed out whatever languages came pouring into my head, though the pairs listed are all items that I have thought about before in terms of mutual intelligibility.

N.B.:  Naturally, I don't expect everyone to attempt to answer all or even any of the questions I have posed.  My main purpose in giving this quiz is to encourage people to think about the nature and reality of degrees of intelligibility between forms of speech.

1. Russian — Ukrainian

2. Hindi — Urdu

3. Hebrew — Arabic

4. Japanese — Mandarin

5. Osaka — Tokyo

6. Turkish — Persian

7. Persian — Arabic

8. Egyptian Arabic — Lebanese / Syrian / Jordanian Arabic

9. Finnish — Estonian

10. Finnish — Hungarian

11. Finnish — Swedish

12. Bengali — Hindi

13. Nepali — Bengali

14. Hindi — Nepali

15. Marathi — Hindi

16. Serbian — Croatian

17. Czech — Russian

18. Czech — Slovakian

19. Tamil — Telegu

20. Malay — Indonesian

21. Pashto — Persian

22. French — Italian

23. French — Spanish

24. French — Portuguese

25. Spanish — Portuguese

26. Spanish — Italian

27. Spanish — Catalan

28. Madrid — Mexico

29. Mexico — Argentina

30. Mexico — Cuba

31. Madrid — Puerto Rico

32. Korean — Japanese

33. Korean — Mandarin

34. Vietnamese — Lao

35. Vietnamese — Cantonese

36. Toishan — Hong Kong

37. Gujarati — Bengali

38. Gujarati — Hindi

39. Swahili — Arabic

40. Arabic — Berber

41. Somali — Amharic

42. Tigrinya — Amharic

43. Setswana — Swahili

44. Kinyarwanda — Kirundi

45. Malagasy — Seychellois Creole

46. Seychellois Creole — Mauritian Creole

47. Afrikaans — Dutch

48. Afrikaans — English

49. Sesotho — Xhosa

50. Swedish — Danish

51. Swedish — Norwegian

52. Swedish — Lithuanian

53. Swedish — French

56. Swedish — Dutch

57. Dutch — Flemish

57. Danish — Norwegian

58. Dutch — Norwegian

59. Dutch — German

60. Vienna — Berlin

61. English — Norwegian

62. English — German

63. English — Dutch

64. English — Swedish

65. English — French

66. English — Italian

67. English — Spanish

68. English — Portuguese

69. Mandarin — Sichuan (rural)

70. Mandarin — Gansu (rural)

71. Mandarin — Ürümchi (local Mandarin)

72. Uyghur — Turkish

73. Kazakh — Uyghur

74. Kazakh — Kyrgyz

75. Mandarin — Shanghai

75. Mandarin — Cantonese

76. Mandarin — Hokkien / Taiwanese

77. Albanian — Armenian

78. Latvian — Lithuanian

79. Estonian — Hungarian

80. Thai — Zhuang

81. Thai — Khmer

82. Tibetan — Mandarin

83. Lhasa — Amdo

84. Sherpa — Bhutanese

85. Rai — Limbu

86. Gurkha — Newar

87. Tibetan — Burmese

88. Ubykh — Georgian

89. Azerbaijani — Chechen

90. Ossetian — Georgian

91. Mongolian — Turkish

92. Mongolian — Korean

93. Quechua — Guarani

94. Belarusian — Russian

95. Belarusian — Bulgarian

96. Romanian — Italian

97. Basque — Spanish

98. Basque — Catalan

99. Romanian — Romani

100. Chippewa — Muscogee (Creek) / Miccosukee (mutually unintelligible Seminole languages)

101. Wenzhouese — Mandarin (late addition!)

(Forgive me for any duplicate pairs on this list, also for omitting any languages that might be of interest to you — I regret, for example, that I have not included pairs involving Scots, Irish [Gaeilge], and Welsh — so feel free to comment on any pairs that you think are revealing or intriguing.)

The amount of intelligibility between the members of these pairs has a direct implication for the question of language versus dialect that we have addressed so often on Language Log, e.g., "Is Cantonese a language, or a personification of the devil?", with references to earlier posts.

In each of the above cases, I pose these questions:

a. are the members of these pairs separate languages?

b. is one of the members of these pairs a dialect of the other member?

c. are the two members of the pairs dialects of some other language?

d. are the two members of the pairs languages belonging to the same group?

e. are the two members of the pairs languages that belong to different groups?

Of course, the boundary between dialect and language can never be absolute and sharply defined, but if someone does not understand a single word of what someone else is saying, the other person must be speaking a separate language from hers, not a dialect of her language.  There is zero intelligibility between them.

I also recognize that it would be well-nigh impossible to design a test instrument capable of yielding scientifically verifiable percentages of intelligibility.  So, in that sense, an assessment of intelligibility is probably always going to be somewhat impressionistic.  In other words, people can honestly say that they understand all / about half / a third / a quarter / nothing of what someone else says.  Nonetheless, such impressions of levels of comprehension are meaningful.  Understanding everything someone else says and understanding nothing or half of what someone else says are very different phenomena.

N.B.:  My Chinese in-laws (Mandarin speakers from Shandong, with an extended stay in Sichuan) understood 0% of Cantonese and only a couple of expressions in Taiwanese, after having spent 30 years on the island.

Some relevant posts:

"Mutual Intelligibility of Sinitic Languages"

"Zazaki: a West Iranian language"

"Spoken Hong Kong Cantonese and written Cantonese"

"The Economist explains:  How a dialect differs from a language"

Comments here.

"Languages and dialects:  Of dialects, armies and navies"

"Varieties of Chinese"


"25 Tibetan languages?"

"How to Forget Your Mother Tongue and Remember Your National Language"

See also:

"Army, navy, dialect, topolect, language:  once more, not with feeling, but with reason"

A comment to:

"Linguistic diversity in Greater Tibet"



  1. Theophylact said,

    May 28, 2014 @ 8:03 am

    I would add the pairs Yiddish – German, Yiddish – Hebrew, Ladino – Spanish, and Ladino – Hebrew.

    Serbian – Croatian, Moldovan – Romanian, and Hindi – Urdu are all (as I understand it) politically different languages, enforced by distinctive script (Cyrillic – Roman for the first two pairs, and Devanāgarī – Arabic for the third).

  2. Paviel said,

    May 28, 2014 @ 8:07 am

    Belarusian – Ukrainian, much better than Belarusian and Ukrainian are intelligible by Russians. Belarusian – Polish may be added as well.

  3. Barrie England said,

    May 28, 2014 @ 8:10 am

    Another possible pair: Hochdeutsch and Schwiizerdütsch

  4. Paolo said,

    May 28, 2014 @ 8:13 am

    Are you also considering language order in each language pair? Generally speaking, Italians seems to be able to understand Spanish better than Spaniards Italian.

  5. Janne said,

    May 28, 2014 @ 8:13 am


    All but complete mutual understanding, except for the occasional local expression or saying that hasn't spread.

    a. no.
    b. Yes. Both are dialects, although Tokyo forms the base of standard Japanese.
    c-e. no.


    From Swedish to finnish: Zero, save for the occasional loneword with close pronounciation. I guess Finns will tend to understand Swedish better, simply because Swedish is a minority language and you'd be hard pressed to find anybody with no prior exposure.

    a. Different languages.
    b-d. No.
    e. Yes, according to what I've learned.


    According to Japanese native speaker, zero understanding.
    a. Different languages.
    b-c. No.
    d-e. Don't know


    30-80% understanding for a Swede listening to Danish. Comprehension depends a great deal on the pronunciation, on the degree of difficulty of the subject, and on the place of origin of the Swedish listener (a northern Swede will have a much harder time understanding anything at all).

    a. Different languages.
    b-c. No.
    d. Yes
    e. no.


    50-80% perhaps, for reasons similar to Swedish-Danish above. Norwegian is actually more difficult (notable when you try to read it) but the pronunciation is easier to follow.

    a. Different languages.
    b-c. No.
    d. Yes
    e. no.


    0% I don't understand a word.

    a. Different languages.
    b-c. No.
    d. no.
    e. no.


    0%-30% Dutch is frustrating; it feels as if I _should_ understand it, but meaning mostly (though not completely) eludes me.

    a. Different languages.
    b-c. No.
    d. no.
    e. no.

  6. jfruh said,

    May 28, 2014 @ 8:16 am

    Another interesting pair would be English-Jamaican Patois (or some other English-derived Caribbean creole). I watched "The Harder They Come" and had to turn the subtitles from the DVD on; it was my first actual first-hand experience with a language that was partially, but not completely, mutually intelligible with mine.

  7. Nick H said,

    May 28, 2014 @ 8:17 am

    I've heard before that it's easier for Portuguese speakers to understand Spanish speakers than vice versa because Portuguese has a more complex phonology (I think the person said in the vowel system) – but I wouldn't know that from personal experience. My friend, who speaks Spanish fluently, says she understands Portuguese speakers pretty well, and I know she can understand written Portuguese almost completely. I also think I've read that Portuguese and Spanish have something like 90% lexical similarity.

    I know that Galician, which is spoken in Galicia in Spain, is supposed to be completely intelligible with Portuguese. I'm pretty sure that in the Spanish legal system, Galician is considered its own language, along with Basque, Catalan, and possibly Valenciano – in Spain's laws that mandate bilingual education in certain regions (Galicia, Basque Country, Catalunya, Valencia, and maybe other regions too?) – but I don't know if legal status is the definitive indicator of whether a (what should I call it without designating language or dialect? polylect? speech pattern?) is its own language.

    I actually think that Valenciano is considered a dialect of Catalan, and I'm certain that these two are mutually intelligible.

  8. Martin J Ball said,

    May 28, 2014 @ 8:19 am

    Welsh-Breton would be an interesting pair….

  9. Fernando said,

    May 28, 2014 @ 8:19 am

    I grew up in Brazil, where it is "common knowledge" that Brazilians can understand spoken Spanish, but that the converse is not always true. I find that I can also understand bits of Italian, but not 100%.

  10. Jeff DeMarco said,

    May 28, 2014 @ 8:28 am

    I would add Guarani – Spanish, as they are both official languages of Paraguay.

  11. Aniko said,

    May 28, 2014 @ 8:29 am

    Finnish – Hungarian

    As a Hungarian speaker, I do not understand a word of Finnish. However when hearing Finnish spoken from a slight distance, so I cannot make out the words, it _sounds_ a lot like Hungarian. It's quite a disturbing feeling actually – as if I should understand it, but I don't.

    So a,b,c – certainly very different languages, not dialects. d, e – no idea.

  12. AG said,

    May 28, 2014 @ 8:30 am

    Yiddish is/was nearly identical to German, in my opinion. The Yiddish terms I've come across in reading are 90% extremely clear German cognates, pronounced just slightly differently than normal, and 10% Hebrew loanwords. When I've encountered Yiddish speakers in the flesh (mostly at the factory outlets in Woodbury NY), I can almost always understand their comments in passing.

    I think the extent of the linguistic separation between Yiddish and German has been exaggerated by different writing systems and history.

    However, while I am (as an English speaker with some knowledge of Old English, German, etc.) able to read texts in Swedish, Danish and Dutch, I cannot follow any of them when spoken, at all. They all sound like a record played backwards to me.

  13. Peter said,

    May 28, 2014 @ 8:48 am

    I'm a native dutch speaker from Belgium, i live a stonetrow from the Dutch border but also in biking distance of the German one. I'm not a language student, so my opinion is purely anecdotal.

    47: Dutch-Afrikaans

    Afrikaans feels like a very ancient dutch, it shares many words but meanings differ and the grammar is quite different. I feel i can make out less then 20% of a south African conversation, though my exposure to the language is slim.

    I would say they are separate languages now, but Afrikaans clearly is an old dialect that has meandered. I would say they belong in the same group.

    57: Dutch-Flemish

    Being belgian i speak flemish, the nuance is very small but grows larger the further you go from the border. In general the intelligibility should be 100% with exceptions made for regional dialects that are hard to understand except for locals. It is the same language but it has regional differences not great enough to make it impossible to understand less then 95%. Same language.

    59: Dutch-German

    Although they may sound similar to the ear I could barely understand german at all before i was schooled in it. And even still my knowledge is limited. The only think that would help me decipher german would be the similarities with the local language Limburgish which is quite similar in ways. Dutch-German: <10% Dutch-Limburgish:70% Limburgisch-German:30% (?)

    NB: Only my grand parents really talk Limburgish, the younger generation generally can't. I do understand it flawlessly though, this would not be true for people who didn't grow up hearing it. It's a shame I never really picked it up myself.

    German and dutch are clearly different languages of the same group, Limburgish is a dialect of dutch with heavy German influences.

    63: English-Dutch

    Due to exposure to the English language the avarage Belgian child will have some sort of understanding of the language. Be it only a few words. I do remember having serious trouble with the language before schooling so i would place the intelligibility at lower then 5%. Very different languages from different groups.

    56: Swedish-Dutch

    I can't understand Swedish, the sound is familiar but i can not make much meaning out of it. Some similarities exist but my knowledge is insufficient to place them. My understanding as a dutch speaker would be near 0%.


    My exposure to Dutch and Flemish are daily, Afrikaans from a friend and some from television. I tried to read a book in Afrikaans once but failed horribly. Exposure to german, television, school, every few weeks in the wild. English, daily trough my schooling (Uni) and the internet. Swedish, sporadically on television, once Danish on holiday, similar story.

  14. KevinM said,

    May 28, 2014 @ 8:53 am

    Interestingly, when "The Harder They Come" was shown in US theatres back in the day, it had subtitles for the first half only. After that, the audience no longer needed the crutch. To my American ears, they judged it about right.

  15. David C said,

    May 28, 2014 @ 9:03 am

    Regarding Swiss German, it is quite striking how many Swiss find it a matter of personal pride to be able to understand whichever local dialect they are confronted with, and no matter how different it is from their own speech. Of course, everyone agrees that the dialect in Wallis is impenetrable.

  16. Irina said,

    May 28, 2014 @ 9:20 am

    I grew up bilingual in Dutch and English, so I can't tell if understanding a language other than my own is based on Dutch or on English! But I can say something about Dutch/Flemish:

    a. It's mostly political: southern dialects of Dutch are indistinguishable from northern dialects of Flemish just over the border. Many Flemings call what they're speaking "Nederlands" (Dutch) anyway. There are some Dutch-only and some Flemish-only words, and borrowing on both sides.

    b/c. They might both be dialects of some uber-Dutch in which Afrikaans also has a claim to membership. Neither Dutch nor Flemish is unified enough to call it a single dialect– my own idiolect is close to "official" standard Dutch, and very far from Flemish, but that's mostly because I was born in the town (Haarlem) that's commonly taken as the standard.

    d/e n/a.

  17. Irina said,

    May 28, 2014 @ 9:23 am

    Oh, and I can understand about 60-99% of Flemish depending on how broad the speaker's dialect is and whether they're making an effort to be understood.

    Flemish speakers are sometimes subtitled on Dutch TV, I think (I haven't watched TV for years so I'm not sure). Possibly also vice versa.

  18. Kellen Parker said,

    May 28, 2014 @ 9:26 am

    I made this comment when I was living in Korea, but many people didn't believe me. I'd said that if you were to go from a background where you knew any 2 Chinese languages to then move to Korea, you'd be set in a very short amount of time. I have a friend, a Mandarin and Cantonese speaker, who picked it up very quickly thanks to the benefit of Sinitic vocabulary in Korea. Friends I mentioned this to scoffed at me. However now, as I'm spending a few hours each week learning Hakka, I'm even more convinced. It's astounding how much previously-unknown Hakka I can completely guess at with only Mandarin and Korean to base it on. It does require the smallest bit of knowledge about irregular correspondences (e.g. some MC -t > -l in Korean and some MC -k > -t in Hakka), but I think even without, I could put a Meizhou or Huizhou/Hailu Hakka speaker in Seoul today and in a week they'd be fine.

    All that said, objectively quantifying mutual intelligibility is a mess, so I still feel it's not really all that useful of a measure.

    "101. Wenzhouese — Mandarin (late addition!)"

    Now you're just trying to stir up trouble. 天不怕,地不怕…

  19. Matt said,

    May 28, 2014 @ 9:31 am

    I have this recurring argument with a Dutch colleague about Swiss German: is it a dialect or a language? Seeing how a fair few Germans find e.g. Bernese Swiss German unintelligible and how tenses and cases work differently, I always argue that there's enough justification to call it a language – which she disagrees with vehemently. Not that I've ever got much of a reason out of her…

  20. GH said,

    May 28, 2014 @ 9:34 am

    Of course, the boundary between dialect and language can never be absolute and sharply defined, but if someone does not understand a single word of what someone else is saying, the other person must be speaking a separate language from hers, not a dialect of her language. There is zero intelligibility between them.

    Is that true in general? Obviously, comprehension is very dependent on circumstances (speed, clarity, subject matter, context cues, etc.) as well as the listener herself (even if we rule out exposure to that particular language/dialect, more general experience in making sense of hard-to-understand speech and a variety of dialects will surely make a difference). It doesn't seem at all implausible to me that you could have reasonable experimental setups where a listener might fail to understand even a single word of someone speaking in a strong dialect of the same language.

  21. S Frankel said,

    May 28, 2014 @ 9:49 am

    @kevinm – Mutual intelligibility varies a lot with exposure. Supposedly, when talkies started coming from Britain to America in the 1930s, American audiences had difficultly with Received Pronunciation that they don't now. I don't know about this, but I do know that when I was learning Swedish in Stockholm, I had a great deal of trouble understanding people from the south of the country at first, and even more trouble understanding Norwegian until I'd watched a bunch of Norwegian tv with Swedish subtitles. Many Swedish friends have similar reports – that Norwegian is difficult for them at first but gets much easier with exposure.

    Swedes who make an effort can understand Danish pretty well, but most don't bother to make the effort unless they're either in Denmark or southern Sweden.

    As for Welsh and Breton, exposure counts here, too. My experience with native speakers of Welsh is that they can't understand Breton at all on first exposure but the written language looks familiar, and if they listen to the language for a while, especially with translations, then they can often get an idea of what's going on.

  22. Karen said,

    May 28, 2014 @ 9:52 am

    As a second language French speaker (since age 5 but a bit rusty now, first language English), I find I can mostly understand the kind of Italian that a parent would speak to a child at a playground. I can't understand the Italian that two adults speak to each other. Presumably a first language French speaker would do better.

  23. Paul said,

    May 28, 2014 @ 10:11 am

    I am a second language Hungarian speaker, and I know a few L1 Finns. I understand 0% of Finnish. They report to understand 0% of my Hungarian.

    [Another interesting pair might be Bulgarian-Macedonian.]

  24. Cheryl Rofer said,

    May 28, 2014 @ 10:14 am

    9. Finnish – Estonian. Estonian friends tell me that it is easier for Estonians to understand Finnish than the other way around. I am learning Estonian, can understand perhaps a third of what I hear, and am beginning to be able to understand Finnish. My impression is that Estonian is somewhat simplified and Finnish is somewhat archaic.

    79. Estonian – Hungarian. A graduate student who was with me in Tallinn, doing research, and who is fluent in Hungarian said, after a few days, "Estonian is just like Hungarian, but the words are different." The vocabularies have no overlap.

    I will leave your questions to the linguists.

  25. MaryKaye said,

    May 28, 2014 @ 10:17 am

    American English speaker who studied French and Russian in school and has taught abroad on several occasions:

    In Finland, Portugal and Brazil I understood *nothing*–I couldn't even parse words most of the time. I actually did better in Japan due to some vocabulary (food, martial arts, anime–try asking directions with that!) In Sweden, Holland, Germany and Belgium I could clearly catch words and guess at some of them. Holland was probably the easiest. To my disappointment, no one spoke Afrikaans to me while I was in South Africa.

    I no longer speak Russian but when I'm in the Czech Republic I can sound out Russian signs and parse them better than Czech ones. Czech was fairly opaque but at least I heard it as words.

    I was homesick in Finland and hung out at a Chinese (I think Mandarin) restaurant because I found the staff conversations homelike and familiar, even though I speak no Mandarin–it sounded like Seattle!

  26. Victor Mair said,

    May 28, 2014 @ 10:31 am

    Sure, Kellen, I can pick up a few Sino-Korean words when I listen in on Koreans talking to each other, just as I can with Flemish, Swedish, Dutch, Portuguese, Italian, Russian, etc., where there are some cognates with or borrowings from / in English, but do you really understand what they're discussing if they're talking at normal speed? Ditto for Shanghainese, Sichuanese (in the countryside; my wife [who grew up in Chengdu till age 11, couldn't understand a word of what the Leshan people were saying when we went back about twenty years ago]), Cantonese (I've studied it for several years but still can only grasp very small bits and pieces of normal conversation)…. Could you summarize the gist of a Korean conversation just on the basis of a few words that you pick out here and there?

  27. Ben said,

    May 28, 2014 @ 10:44 am

    Turkish – Persian: 0% intelligibility
    a. yes
    b. no
    c. no
    d. no
    e. yes

    Uyghur — Turkish: <5% intelligibility
    a. yes
    b. no
    c. no
    d. yes
    e. no

    Now if you'd asked about Turkish – Azeri, Turkish – Turkmen or Uyghur – Uzbek, the answers would be more interesting!

  28. Eric P Smith said,

    May 28, 2014 @ 10:59 am

    The only pair I am competent to comment on is Scots and English. Broad Scots is at one end of a bipolar linguistic continuum, with Scottish Standard English at the other. Speakers of Broad Scots are exposed to enough Standard English to understand it 100%. Coming at Standard English and Broad Scots completely cold, I would guess that the mutual intelligibility is about 70% – 80%.

    It seems to me that Scots has been diluted by English so much over the centuries that even the broadest Scots spoken nowadays is a dialect of English. Those who classify Scots as a distinct language are (again, as it seems to me) making a political point rather than a linguistic one.

  29. Ben Artin said,

    May 28, 2014 @ 10:59 am

    Serbian and Croatian have minor vocabulary differences, on par with American/British English vocabulary differences. They are grammatically nearly identical; I can think of only one grammatical construct that differs between the two[*]. Of course, they use different scripts, but that is irrelevant for your question.

    If your question is about *interactive* spoken comprehension (that is, if you allow for the possibility that one participant can say "wait what" and the other can rephrase, etc), then they would have 100% mutual comprehension.

    If it's about non-interactive spoken comprehension (for example, watching the news in the other language), then they would have near-100% mutual comprehension for any non-trivial spoken pieces (because the few differences between the languages can usually be contextually resolved).

    Source: Croatian is my first language, I grew up in Croatia during the breakup of Yugoslavia so I spent more time than was good for me being taught how important it is that Croatian is different from Serbian (spoiler alert: it isn't), and I was a language nerd even as a kid.

    [*] Croatian construction for "Did you ___" is "Jesi li ___" (using the long variant of "to be" as auxiliary, "jesi"), whereas Serbian construction for the same is "Da li si ___" (using the short variant of "to be" as auxiliary, "si", together with yes interjection, "da"). The same difference applies in other persons and numbers.

  30. Robert Coren said,

    May 28, 2014 @ 11:00 am

    Danish-Norwegian: Many years ago, I was a counselor at a boys' camp in upstate New York whose population included a number of boys from foreign countries. One day I encountered the Danish camper and the Norwegian camper engaged in conversation in an unintelligible-to-me Scandinavian language, and I asked them whether they were speaking Danish or Norwegian, and was told that each was speaking his native language.

  31. julie lee said,

    May 28, 2014 @ 11:08 am

    Kellen Parker: Great to know that one can pick up Korean easily if one knows two Chinese topolects. I know Mandarin and Cantonese and am eager to learn Korean. (I only know the similarity of Korean and Chinese words from Korean restaurant menu items.)

  32. Paul Clapham said,

    May 28, 2014 @ 11:21 am

    Several years ago I was on a walking tour through the mountains of Lesotho. One of our guides, who was from the Eastern Cape and spoke Xhosa, tried to talk to some of the local residents, who would have been seSotho speakers. He came back and said "I can't understand them".

  33. Victor Mair said,

    May 28, 2014 @ 11:30 am

    @julie lee

    It's not that easy. See my comment to Kellen.

  34. Victor Mair said,

    May 28, 2014 @ 11:32 am


    From the Online Etymology Dictionary:


    1570s, "form of speech of a region or group," from Middle French dialecte, from Latin dialectus "local language, way of speaking, conversation," from Greek dialektos "talk, conversation, speech;" also "the language of a country, dialect," from dialegesthai "converse with each other," from dia- "across, between" (see dia-) + legein "speak" (see lecture (n.)).


    From Wiktionary:


    From Middle French dialecte, from Latin dialectos, dialectus, from Ancient Greek διάλεκτος (diálektos, “conversation, the language of a country or a place or a nation, the local idiom which derives from a dominant language”), from διαλέγομαι (dialégomai, “I participate in a dialogue”), from διά (diá, “inter, through”) + λέγω (légō, “I speak”).

    (linguistics) A variety of a language (specifically, often a spoken variety) that is characteristic of a particular area, community or group, often with relatively minor differences in vocabulary, style, spelling and pronunciation.

    The difference between a language and a dialect is not always clear, but it is generally considered that people who speak different dialects can understand each other, while people who speak different languages cannot.


    From the American Heritage Dictionary, as given in The Free Dictionary online:

    [French dialecte, from Old French, from Latin dialectus, form of speech, from Greek dialektos, speech, from dialegesthai, to discourse, use a dialect : dia-, between, over; see dia- + legesthai, middle voice of legein, to speak; see leg- in Indo-European roots.]


    From Webster's online:


    Middle French dialecte, from Latin dialectus, from Greek dialektos conversation, dialect, from dialegesthai to converse — more at dialogue
    First Known Use: 1577


    From Collin's English Dictionary:


    C16: from Latin dialectus, from Greek dialektos speech, dialect, discourse, from dialegesthai to converse, from legein to talk, speak


    From OED:


    From French dialecte (16th cent. in Hatzfeld & Darmesteter), or from Latin dialectus, Greek διάλεκτος discourse, conversation, way of speaking, language of a country or district, from διαλέγεσθαι to discourse, converse, from δια- through, across + λέγειν to speak.


  35. Oskar Sigvardsson said,

    May 28, 2014 @ 11:38 am

    Robert Coren: that is incredibly common when Scandinavians speak to each other. For a Swede, Danish can be hard to understand without exposure, but if the other person just speaks slowly enough and occasionally clarifies a word, everything works just fine. Norwegian is even easier. If there's ever real trouble, we just switch temporarily to English.

    If you want to see examples of Scandinavians speaking to each other each in their own language, watch any Danish television show which features one or more Swedish characters (which is most of them). For instance, you can look up clips from "Broen" ("The Bridge") on YouTube which is show with two cops as main characters, one Swedish and one Danish. They both communicate just fine speaking in their native tongues.

  36. phspaelti said,

    May 28, 2014 @ 11:39 am

    Albanian — Armenian?
    Surely Albanian — Serbian would be more relevant. (I assume mutual intelligibility would still be close to 0).
    More interesting would be Slowenian — Croatian (or Slowenian — Serbian).

  37. GH said,

    May 28, 2014 @ 11:41 am

    The etymological fallacy, from the Language Log? Really?

    In any case, I was not disputing that mutual comprehension is one component of distinguishing dialects from languages, merely that "if someone does not understand a single word of what someone else is saying, the other person must be speaking a separate language from hers" (my emphasis).

  38. GH said,

    May 28, 2014 @ 11:42 am

    (Last post @Victor Mair)

  39. Piyush said,

    May 28, 2014 @ 12:20 pm

    Hindi and Urdu have exactly the same grammar, but rather different technical vocabularies. As such, while I (native Hindi speaker, with no formal Urdu training) would have no trouble holding a conversation with a native Urdu speaker, I would struggle to understand,say, a significant fraction of Ghalib's poetry without the aid of a dictionary. As such, I would say that the mutual intelligibility between Hindi and Urdu can range from about 50-60% to 100%, depending upon the subjects being talked about.

    For Hindi-Bengali, I would think the fraction would be about 50%, but I would expect it to be so high only if conversation is made at an artificially slow speed.

  40. Steve L said,

    May 28, 2014 @ 12:29 pm

    Russian – Ukrainian

    a. Yes
    Almost all Ukrainians understand Russian but Russian intelligibility of Ukrainian is much lower, depending on whether it's a text or speech can be anywhere between 50-70% I would say.
    b. No
    c. Yes, Eastern Slavic

  41. Lucy Kemnitzer said,

    May 28, 2014 @ 12:55 pm

    Czech-Slovakian is rather like Serbian-Croatian: all differences are small and mutual intelligibility is high. I would honestly say they are dialects of the same language.

    I am a casual student of Czech with no fluency whatever, but I find that my very small vocabulary and grammar knowledge is almost as useful in comprehending every Western and Southern Slavic language I have encountered. Probably as I approached greater fluency in Czech, my ability to suss out things in the other languages would not increase proportionately.

  42. Ellen K. said,

    May 28, 2014 @ 12:56 pm

    Regarding Scots (mentioned in the comments), I (American) recently listened to a bit of Scots and I found it at times to sound just like English, and other times to not be comprehensible. I suspect, though, that if it wasn't for already having familiarity with English spoken with a Scottish accent, that I would have had a much harder time understanding Scots.

  43. Ray Girvan said,

    May 28, 2014 @ 1:23 pm

    @Aniko: Finnish – Hungarian … disturbing feeling

    That's exactly how I feel about Dutch. Near-zero intelligibility, but when you overhear someone at the next table, it sounds like English with a faintly American accent …. but then you're baffled by their saying stuff like "They crake the nervous wiltin', and lately over-slugs the mender fine."

  44. David Eddyshaw said,

    May 28, 2014 @ 2:42 pm

    Welsh-Breton intelligibility is low; stories of onion-bedecked striped-jerseyed beret-wearing cycling Bretons being readily comprehensible in Wales are, alas, mythical.

    Historically at least Cornish was closer to Breton, unsurprisingly. I don't know if the hardy enthusiasts trying to revive Cornish can understand much Breton.

  45. JCD said,

    May 28, 2014 @ 2:43 pm

    Latvian – Lithuanian

    Low mutual intelligibility. Some words are similar enough to be understood by a speaker of the other language, and there are lots of grammatical similarities, but in general not even close to being similar enough for speakers of one language to understand speakers of the other.

    a. Yes

    b. No

    c. No (they are the only two languages, apart from a few with almost no speakers in their group)

    d. Yes, the Baltic group

    e. No

  46. Felix K said,

    May 28, 2014 @ 2:46 pm

    I find it rather challenging to actually take the point of view of someone who has had no prior exposure at all. My estimates therefore tend towards the lower side. Note that for the case of Pashtu-Persian, mutual intelligibility is possible only through a number of common nouns (from Persian and Arabic), not through other classes of words. In all cases of course, it depends a lot on the topics discussed.
    Hindi — Urdu 60-100% Everyday speech nearly always 100%
    Persian — Arabic 5%
    Nepali — Bengali 20%
    Hindi — Nepali 20% It is probably easier for Nepali speakers to understand Hindi than vice versa.
    Pashto — Persian 5%
    French — Spanish 30%
    Dutch — German 30%
    English — German 5%
    English — Dutch 5%
    English — French 0%
    English — Spanish 0%

  47. Jonathan Badger said,

    May 28, 2014 @ 3:46 pm

    The problem with distinguishing dialects from languages is much like the biologist's problem of distinguishing varieties from separate species. It seems so obvious and objective at first, but as you look into it, it is clear that in many cases the distinction is arbitrary and made for simple human convenience or political reasons.

  48. Alex said,

    May 28, 2014 @ 3:51 pm

    So, before I do my list I'd like to note that for the English/Romance language pairs, the difference is far more marked for basic vocabulary than for multisyllabic words that entered English during the Renaissance and Enlightenment. You will find easy cognates for Spanish words like "el experimento," and not so many for words like "el pan." Paradoxically, this makes scientific discourse more mutually intelligible than basic conversation. If you have a big enough English vocabulary, you can find advanced vocabulary cognates for basic words, too, though. So, ok, "comer" doesn't resemble "to eat" very much, but it seems like it could be related to "comestible." "Encontrar" isn't a cognate to "to find," but it matches "to encounter." It doesn't really work very well to reverse the process, though, I suppose because English basic vocab is Germanic.

    93 Quechua-Warani Low inter-intelligibility, maybe 10%, mostly based on loanwords from Spanish.

    61-68 English-various European languages. Not counting loan words from English. Ranging from a low of 5% (Portuguese, Spanish, Italian), slightly higher for French thanks to William, and a bit higher still for germanic languages (maybe 15%) Dutch is the highest at about 20%, like English with a bad German accent, but not mutually intelligible enough for much communication.

    25-27 Spanish with other neighboring languages that might really be dialects. Catalan at least doesn't have the army necessary to qualify as a language :) About 85% intelligible if people talk slow.

    28-29 Spanish. Same language, 98% intelligible, and speakers are largely aware of the major differences. Some lexical items will trip you up– tortillas aren't the same thing in Madrid and Mexico City.

    30-31 Mexican and Iberian Spanish with Caribbean Spanish. Not as mutually intelligible as the preceding pairings, but not as hard as Portuguese. Say, 90%, maybe a bit lower for the Madrid-PR pair. Still the same language, especially for educated Carribbean speakers.

    The pair I'd like to have seen: Southern Quechua-Aymara. That's a fun one because there is quite a lot of mutual intelligibility (maybe 20% or more) for languages that have different structures and are apparently unrelated. It's just a heck of a lot of loan words (from the two languages and from Spanish) and similar phonology.

  49. Rubrick said,

    May 28, 2014 @ 4:16 pm

    I think an interesting pair (apologies if it's already among the many comments) is American English / Caribbean English. I've had virtually no success trying to understand conversations between English-speaking Caribbean islanders (though of course they're quite capable of making themselves understood when speaking to Americans).

  50. Catsidhe said,

    May 28, 2014 @ 4:48 pm

    English – Frisian
    English – Scots
    Irish (Gaeilge) – Scottish Gaelic (Gàidhlig)
    Irish – Manx
    Irish – Welsh
    North Welsh – South Welsh

    And for purely hypothetical giggles:
    Old Irish – Old Welsh
    Primitive Irish – Gaulish
    Primitive Irish – Celtiberian

  51. Tim Friese said,

    May 28, 2014 @ 5:35 pm

    3. Hebrew-Arabic: 10% tops. There are plenty of shared roots but the sound correspondences are often quite opaque to people without linguistic training. For example, Hebrew /za'hav/ is cognate to Classical Arabic (and some dialects) /'ðahab/ but that's about as good as you can do, most cognates are even more distant, and many words lack cognates.

    c. Well, Arabic is itself a gigantic cluster of dialects/languages d. Yes, Semitic languages, and they share a lot of basic grammar and roots

    8. Egyptian Arabic-Syrian Arabic: 50% This gets hard. People with 0 exposure would not understand negation, question words, and a lot of basic vocabulary (bread, thing, to want, to enter, etc). Plural and gender agreement can be different, and Lebanese and Syrian change some verbs beginning with /bn-/ to /mn-/ making them look like participles instead of conjugated verbs.

    I personally experienced this gap having moved to Egypt speaking quite good Syrian Arabic. Those Egyptians who did not have background in communicating with Syrians had problems communicating with me for even the most basic tasks (a kilo of eggplant, turn here, etc.).

    a. By Romance language standards, they would be different languages, though speakers don't think of them this way.

    25. Spanish-Portuguese: 80% It is often said that Portuguese speakers readily understand Spanish speakers but the reverse takes some time due to major sound changes, especially in Brazilian Portuguese.

    a. I think Spanish and Portuguese are borderline cases. If they were spoken by peoples of the same national identity, they would certainly be called one language.

    26. Spanish-Italian: 50% While having quite distinct accents, I found them more intelligible than I expected; when I was a tourist in Italy I mainly spoke Spanish and was able to accomplish basic needs without too much fuss.

  52. Myrosia said,

    May 28, 2014 @ 5:37 pm

    Re: Russian-Ukrainian, I'd disagree with Steve. I would say the mutual intelligibility for spoken language is about 50% in both directions.

    I am a native Ukrainian speaker born in the Soviet Union. Like me, most of the current Ukrainian speakers have been exposed to Russian from an early age. So they are not good test population. But I met Ukrainian diaspora in the US for whom Ukrainian is first language, but who weren't exposed to Russian, being born pre-1939 in Poland-owned parts of Ukraine. I would say that they have as much trouble understanding Russian as my Russian-born friends have trouble understanding me when I try to quote something in Ukrainian.

    There is also an interesting distinction between "intelligible" and "able to speak". My elderly relatives who were born pre-WWII in Poland were all able to understand Russian on TV, but weren't really able to speak it.

    I would also suspect that Belarussian-Ukrainian seem to be more mutually intelligible because the speakers of both languages tend to know Russian and this helps. I have observed this phenomenon with Polish. When I visited Warsaw for a summer school, people from Ukraine, including native Ukrainian and native Russian speakers, seemed to understand it a lot better than people from Russia. My theory has been that people exposed to two Slavic languages (Ukrainian and Russian) did better than those exposed to only one (Russian).

  53. exackerly said,

    May 28, 2014 @ 5:50 pm

    I hope you do intelligibility within English dialects as well. I heard a telephone interview with a Scot the other day and I could hardly understand a word. I'm standard American.

  54. Zora said,

    May 28, 2014 @ 5:51 pm

    Polynesian languages. The one I speak: Tongan. The one I studied briefly in college and to a great extent forgot: Hawaiian. The one I occasionally hear: Samoan.

    Not immediately mutually intelligible, but large overlap in vocabularies (ignoring sound shifts). Similar grammars. Pronunciation differences. (I had a hard time not pronouncing Hawaiian like Tongan.) Easy to learn another Polynesian language if you know one. Listening to one you don't know is frustrating, because you ALMOST get it.

  55. Milan said,

    May 28, 2014 @ 5:58 pm

    Low German — Dutch – 85% (written); 40% (spoken)
    I'm a native speaker of both Standard German and Low German, but Dutch is certainly closer to Low German than to High German, so I think that's were my comprehension comes from. I'm generally I able to understand written Dutch, recognizing ca. 85% of the vocabulary, and many grammatical categories, including number and diminutive for nouns as well as tense and mood for verbs. Yet in spoken form I am only able to understand isolated words. However a friend of mine was able to comprehend natural speech after a stay of a month in the Netherlands, without ever receiving any training.
    I'm pretty sure that, hadn't Dutch developed an independent written form, and were it more closely tight with the German states before German unification (that of Bismarck), it might have come to be seen as a German dialect in public consciousness, yet linguistics would probably always have insisted on it's status as a language, as it did for Frisian.

  56. Victor Mair said,

    May 28, 2014 @ 5:59 pm

    From an old German friend who grew up speaking several European languages as a child, but moved to America about 50 years ago:

    I don't quite know how to respond to your quiz as you outlined; instead I personally feel that Dutch, Flemish, Afrikaans and Yiddish are some form of German dialect, although Dutch and Yiddish I always have to "sound out" to comprehend due to their very strange spelling.

    Your pairing French-Portuguese is curious as I see little connection but . . . I cannot "hear" Portuguese, which to me sounds more like Russian, but I read it as Spanish (ignoring grammar & spelling). In the course of my work, I discovered that Portuguese, even those not highly educated, seem to understand French. There was a large, extremely unobtrusive Portuguese community in San Jose where only the kids spoke English. Dealing with the adults, I communicated in French while I had them respond to me in writing which I then read as Spanish to get their gist. This may sound a bit strange and round-about but it worked just great.

    Sad to say, my Spanish and French have mostly gone by the wayside for lack of usage, something which is also threatening my German since Dad is gone. I actually lost my (Castilian) Spanish when I came to California and I ran afoul of Mexican, which to me is completely unintelligible — like listening to a sewing machine.

  57. Sneha Narayan said,

    May 28, 2014 @ 6:20 pm

    Hindi – Urdu

    I would probably say these are over 95% similar if you're looking only at speech. The script is totally different, and there are a few words that are specific to one language rather than the other, but the grammar is exactly the same. If someone played me clips of someone speaking these languages, I wouldn't be able to tell them apart (perhaps a very well versed speaker in either language would be better able to notice dialectical differences).

    Tamil – Telegu

    These are quite different. I would say less than 10% mutually intelligible. I speak Tamil fluently, and have listened in confusion when people speak Telegu around me. If I listen to a person speaking Telegu for five minutes, I may be able to guess what topic they are talking about if I tried pretty hard to identify a few cognates.

  58. Eric Prendergast said,

    May 28, 2014 @ 6:22 pm

    Linguist who works in the Balkans, non-native speaker of Albanian, Macedonian, and Czech, I have studied Bosnian/Serbian/Croatian, Bulgarian, and Romani. No language-learning experience with Armenian or Belarusian.

    16. Serbian — Croatian

    Perfect, 100% intelligibility. This is, however, an politically and socially sensitive thing to say, and there are Serbs or Croats who will argue that Serbian and Croatian are not so thoroughly intelligible as that.

    On occasion, especially just after the Yugoslav Wars, translators were used between Serbian speakers and Croatian speakers in official proceedings. This, I am confident in saying, was political spectacle.

    Standard Serbian and Standard Croatian are, without question, different. They are written differently, they have different lexical bases (especially in the realm of official vocabulary, educated terms, and internationalisms), and there are some grammatical differences in the standard registers (these tend to dissipate in the spoken registers).

    Given all that, I have never, not once, in my time studying in the Balkans, seen a Serbian speaker and a Croatian speaker have even the slightest trouble engaging in rapid, fluid, sophisticated, and effortless conversation. They do, however, have a number of lexical differences (think things along the lines of 'pram' vs. 'stroller' in the UK vs. US) and these lexical differences are a matter of some sensitivity.

    The question of whether Serbian and Croatian are separate languages is fraught with complex, and tragically blood-soaked history. It is, frankly, a stupid exercise to engage in determining and much, much too much screaming and yelling and slur-slinging has occurred in the name of 'scientifically proving' a definitive answer to a subjective question. My own Serbian and Croatian colleagues are so sick of the subject they could vomit.

    Linguists in the United States who study in the region tend to refer to Bosnian, Serbian, and Croatian as Bosnian/Croatian/Serbian or 'BCS', and treat them together for most major linguistic research questions, while making sure to treat the matter of their separate standardization and establishment as separate official languages with all the sensitivity and respect appropriate for an issue over which people, quite literally, died.

    18. Czech — Slovakian

    80%-100% for speakers who reached adulthood before the dissolution of Czechoslovakia due to the mandatory use of both Czech and Slovak in all public media.

    50%-75% for speakers who have reached adulthood after the Velvet Divorce. Older Czechs complain, sometimes rather bitterly, about the trouble that younger speakers of Czech have in understanding Slovak, because this was not at all a problem for older generations. Familiarity, rather than any active study, is the major factor. The absence of Slovak in the Czech public sphere (with the exception of pop music) has sharply curtailed untrained speakers of Czech in their ability to understand Slovaks. Slovakia, as a smaller country with a somewhat less developed literary and high culture scene (this is rapidly changing), is still fairly awash with Czech, and so Slovaks apparently have a much easier time understanding Czech, but I can't confirm this from first hand knowledge.

    77. Albanian — Armenian

    Absolutely zero intelligibility.

    Albanian – Serbian was mentioned above in the comments: the commenter was correct. Mutual intelligibility is close to zero. There are, however, some shared lexical items, both borrowings between each other and borrowings from other languages like Turkish and internationalisms. Speakers who were not taught one (almost exclusively Albanians who were required to learn Serbo-Croatian as part of Yugoslav dominance) are generally unaware that the two languages share these lexical items.

    99. Romanian — Romani

    0–10%, almost entirely due to lexical borrowings from Romanian into Romani that would be recognizable to a Romanian speaker. There are no monolingual Romani speakers beyond childhood, so this is a hypothetical that cannot be tested in the reverse direction. Romani grammar, however, is so thoroughly different from Romanian grammar, and the basic lexicon of Romani is so divergent, and the borrowings from Romanian are adopted to Romani in ways that are distinctive enough, that intelligibility is very low if the speaker of Romani in question is not speaking a para-Romani variety. The question is complicated by the matter of what variety of Romani you are discussing: Vlax dialects developed in intensive contact with Romanian and thus have much more in common. Balkan Romani dialects (those further south), in contrast, would have effectively zero intelligibility. Romani speakers in Macedonia frequently engage in lexical codeswitching from the dominant local language (Albanian, Macedonian, or Turkish) and can adjust this degree of codeswitching up or down according to context, so it's difficult to give an exact degree of intelligibility. But if Roma who speak Romani do not want to allow themselves to be understood by those around them, this is usually very easy.

    95. Belarusian – Bulgarian

    20% or so. Mostly pan-Slavic roots. Bulgarians would have an easier time understanding Belarusians that vice-versa. Lexical base of the two languages are very similar, but the grammatical structure of the two is very different.

  59. Mark F said,

    May 28, 2014 @ 6:37 pm

    I'd agree with Felix K on:

    English — German 5%
    English — Dutch 5%
    English — French 0%
    English — Spanish 0%

    and add

    English — Norwegian 5%
    English — Swedish 5%
    English — Italian 0%
    English — Portuguese 0%

    I myself can understand Spanish and Italian fairly well, but my observation of my fellow Americans who did not grow up in the Southwest is that Spanish is completely incomprehensible except for si, gracias, etc.

    On the other hand, I've never studied a word of any of the Scandinavian languages and I find myself stunned to be able to pick out whole (short) sentences when watching films from that region.

  60. Rebecca said,

    May 28, 2014 @ 6:46 pm

    I'd put Norwegian-Swedish at 80% or more. I'm not a native speaker of either, but after studying Norwegian a couple of semesters, I could understand Swedish quite well during a summer in Finnland and Sweden. Basically, if I knew it in Norwegian, I could understand it in Swedish.

  61. Darkwhite said,

    May 28, 2014 @ 6:54 pm

    For a different observation – trivial, but maybe not entirely obvious: As a Norwegian, French and Japanese are both so strange that intelligibility is essentially zero. French shares a fair bit of vocabulary, in theory, but even those few words tend to drown in noise.

    Despite this, Japanese has a much steeper learning curve, probably because French and Norwegian have very similar grammar (noun genders, noun plurals, verb tenses, etc). Two languages which are both entirely intelligible at first listen might still be very different in how easy they are to pick up.

    For the data:

    Norwegian-Swedish: 95% (feels disfluent at first, occasional strange words, strenuous to parse until you grow accustomed over a few hours)
    Norwegian-Danish: 95% (as Swedish; tends to be harder to parse, otherwise more similar)
    Norwegian-German: 50% (you can pick out the general topic and understand fragments; you learn to understand German at the 80% level by casual exposure)
    Norwegian-Dutch: ?? (Personally, I would say somewhere around 75%, but it seems like I'm using my knowledge of German, English and Norwegian simultaneously, knowing just Norwegian it would probably place at 20%)
    Norwegian English: 25% (Much like German, except there's less common vocabulary)

  62. Pflaumbaum said,

    May 28, 2014 @ 7:13 pm

    The Romanians I've asked say that they can largely understand standard Italian, but aren't understood by Italians in turn.

    Phonologically they're very close indeed, the biggest difference probably being the presence of two central vowels in Romanian. But Romanian has plenty of Slavic loan-words, which may be what makes comprehension harder for an Italian (if indeed it is harder).

    An Italian waitress once asked me which dialect I was speaking when she overheard me talking Romanian to my wife's grandma, so it obviously sounded very close to Italian to her.

    Regarding Serbian/Croatian, I once spoke to a guy with a Serbian mother and Crostian father, and he said that though they're almost identical dialects, the subtle differences between them are now exaggerated and extended in a deliberate way. He also said that speaking one or the other was a political statement, so that for someone like him who felt both a Serb and a Croat, even speaking his own language in public had become a fraught and painful thing.

    Hmm, that may have been the single most anecdotal post ever on LL.

  63. TheStrawMan said,

    May 28, 2014 @ 7:14 pm

    Tokyo – Osaka = 99% (everyone in Japan leanrs to speak standard "Tokyo" dialect; Osaka dialect is very popular and spoken on TV by entertainers, etc., so most are familiar with it too)
    Japanese – Mandarin = 0% (but would be maybe 25% if written communication were involved)
    Japanese – Korean = 0.1% (the two languages share a few isolated words, but probably not enough for any kind of communication)

  64. Nanani said,

    May 28, 2014 @ 8:11 pm

    How about Montréal-Paris?
    As a French-Canadian I can say that the higher the register, the higher the mutual intelligibility, except for some technical vocabulary based on e.g., different laws. On the low end, though, intelligibility falls apart completely.

    For example, I can watch news from France and understand what is being said, but an unscripted talk show is much more difficult.
    Likewise, a friend of mine from France could watch scripted TV shows from Québec but found outtakes and the like incomprehensible.

    Overall, I'd say 75% mutually intelligible on average.

    a. No, both are French.
    b. I don't know how to answer that.
    c. Yes, Parisian and Montréalais are dialects of French.
    d. They aren't distinct languages, really.
    e. No.

  65. Michael Dunn said,

    May 28, 2014 @ 8:32 pm

    Turkish-Persian: the situation would have been different pre-Ataturk I think, as Ottoman had far more Persian loanwords than Modern Turkish. Persian was used at the Ottoman court; Ataturk sought to purify the language of Persian (and Arabic) loanwords.

  66. maidhc said,

    May 28, 2014 @ 8:32 pm

    Irish and Scottish Gaelic are pretty close. There are no native speakers of Manx left, but the last few were recorded quite a bit. To my ear those recordings sound very similar to Ulster Irish. Manx has a very strange spelling system that obscures the relationship.

    Welsh, Breton and Cornish are totally different. However, if you sit down with a dictionary 5-10% of the words are obviously related.

    amzer aimsir (weather)
    glas glas (blue)
    mab mac (son)
    mat maith (good)
    berr bearr (short, shorten)
    gwir fíor (true)
    skuizh sgìth (SG tired)

    It's the other 90% of the vocabulary looking completely different that causes problems.

  67. Dan Lufkin said,

    May 28, 2014 @ 8:58 pm

    Modern Swedish has an overlay of German vocabulary that many Swedish speakers learn to avoid when talking to Danes and Norwegians. Swedish fråga / German Frage — Swedish dialect spörja / Norwegian spørje, for instance. All three countries have a semester of "neighbor languages" in middle school plus exposure on TV and radio. Of course, there are places where Norwegian itself isn't mutually intelligible (Nynorsk & Bokmål) or where sounding anything but local will get you a blank look. Generally, though, Scandinavians are pretty adaptable and one can get along OK everywhere.

    Before WWII, coastal dialects in Essex and Friesland were said to be mutually intelligible among fishermen. In the late 50s I perambulated from Hamburg to Rotterdam and found it hard to isolate any language boundaries, aside from newspapers and radio, etc.

  68. Michael Dunn said,

    May 28, 2014 @ 8:59 pm

    Of course the whole discussion is supposed to be based on the premise "Assuming no prior, formal study of or contact with the opposite language in a given pair (i.e., one is coming at these languages completely cold)." In real life, comprehensibility may depend on at least some inevitable contact. On Syrian Arabic-Egyptian Arabic, Tim Friese is right about the obstacles to comprehension. But the popularity of Egyptian films in Syria and Syrian TV serials in Egypt means that most people don't come to it completely cold. (Margaret Omar years ago published a very useful little item called "Egyptian and Levantine Arabic, A Comparative Study," which is a good guide to the speaker of one to deal with the other.
    Tim is similarly right about the underlying difficulties of incomprehensibility of Hebrew and Arabic. Outside of Israeli Arabs and Palestinians in the Occupied Territories, who often have to speak it, the only Arabs I know who understand any Hebrew are scholars or intelligence analysts. Of course Modern Israeli has a lot of Arabic loanwords from foods to expressions to obscenities, an many Israelis know some Arabic, but of course that wasn't what you were asking here.

  69. S Frankel said,

    May 28, 2014 @ 10:16 pm

    In response to Catsidhe:

    Irish (Gaeilge) – Scottish Gaelic (Gàidhlig)

    Depends on the dialect. Somebody from SW Ireland (Dingle, e.g.) would have a lot of trouble with someone from NE Scotland. The various Gaelic dialects within the "same" language are different enough to impede communication. Nowadays, there's a standard Irish to refer to, and most of the non-Hebredian dialects of Scottish Gaelic are gone. Based on what my Irish teacher once said, I would guess 60% between a modern Irish and modern S Gaelic speaker..

    Irish – Manx
    NW (Ulster) Irish with Manx, at least 60%, probably more. SW Irish, much less

    Irish – Welsh
    5%. There are obvious cognates, many English loan words, and a lot of the grammar is familiar, but only snippets are understandable without study. Maybe like English and German.

    North Welsh – South Welsh
    Close to 100%. 30 or 40 years ago, this wasn't the case, but everyone has had exposure to the different varieties through mass media, and some of the local dialects are losing their most distinctive, or outlandish, features. (There's no one recognized standard for the entire language, though, and learners definitely have to chose one variety or another.) It used to be that speakers from different parts of Wales found it easier to communicate in English. The differences between N and S Welsh are maybe on the order of the differences between (standard) American and British English.

    And for purely hypothetical giggles:
    Old Irish – Old Welsh

    This one is easy: 0%. Old Irish is a real knucklebuster; the verbs don't resemble anything in any other IE, or maybe any other human, language. They're so complicated that I wonder how much of the written language that we have was simply invented (i.e., a stylization based on the spoken language).

  70. Piyush said,

    May 28, 2014 @ 10:38 pm

    @Sneha Narayan

    If someone played me clips of someone speaking these languages, I wouldn't be able to tell them apart (perhaps a very well versed speaker in either language would be better able to notice dialectical differences).

    In fact, telling Hindi and Urdu "apart" is a bit of an undefined problem: every valid Hindi sentence is also a valid Urdu sentence and vice versa. The only difference is the relative frequency of Sanskrit vs Persian loan words.

  71. Piyush said,

    May 28, 2014 @ 10:40 pm

    Of course, my last comment holds only for the khari boli dialect of Hindi. With dialects like Awadhi (which happens to be the language of some of the best known poetic works in Hindi), telling the difference from Urdu would be trivial.

  72. Mark F. said,

    May 28, 2014 @ 10:45 pm

    I think 5% is an overestimate for German or Dutch as heard by an American. Even the close cognates are hard to pick out of the stream of words.

  73. Vasha said,

    May 28, 2014 @ 10:56 pm

    In line with what orher commenters have said about Hindi and Urdu, I recently asked a native speaker of Urdu how much spoken Hindi he understood, and he said he rarely ever encountered anything he couldn't understand in that language (never learned to read its written form however). But his case isn't relevant to the original question's emphasis on encountering the other language "cold". He said that media from India is fairly often viewed/listened to in Pakistan, which gave him a chance to acclimate to Hindi and learn any number of vocabulary items and idioms.

  74. Victoria Simmons said,

    May 28, 2014 @ 11:13 pm

    2. Hindi — Urdu — I would perhaps classify as dialects of the same language, with different writing systems, except that I understand that the Islamic cultural world of Urdu has resulted in many more borrowings from Persian and Arabic, so that mutual intelligibility is difficult.

    3. Hebrew — Arabic — Different languages.

    4. Japanese — Mandarin — Different languages from different families.

    6. Turkish — Persian — Different languages from different families, with only lexical borrowings.

    7. Persian — Arabic — Different languages from different families, with lexical borrowings in both directions.

    10. Finnish — Hungarian — Same family, but different languages, no mutual intelligibility.

    11. Finnish — Swedish — There are lexical borrowings of Swedish into Finnish, but they are from different language families, and i don't think mutually intelligible at all.

    16. Serbian — Croatian — I have always understood them to be mutually intelligible and separated mainly by different writing systems.

    21. Pashto — Persian — If by Persian you mean Farsi, both Farsi and Pashto are Persian languages, but I don't think they are mutually intelligible.

    22. French — Italian — Same branch of same family, but not mutually intelligible. The same true for the other pairs in the Romance family. I think Spanish and Italian are closer than either is to French.

    32. Korean — Japanese — Different languages.

    33. Korean — Mandarin — Different languages.

    39. Swahili — Arabic — I think an Arabic speaker might be able to understand words, but the grammar is quite different.

    47. Afrikaans — Dutch — I think that Afrikaans is pretty close to Dutch.

    48. Afrikaans — English — Not mutually intelligible

    50. Swedish — Danish/Norwegian — Different languages, but with some intelligibiliy

    52. Swedish — Lithuanian — Different languages, different branches of the same family.

    53. Swedish — French — Different languages.

    56. Swedish — Dutch — Different languages, but Swedish is closer to Dutch than to French.

    57. Dutch — Flemish — Dialects of the same language, possibly not very mutually intelligible.

    57. Danish — Norwegian — Dialects of the same language.

    59. Dutch — German — Many similarities, but different languages, not mutually intelligible

    60. Vienna — Berlin — Dialects, or even just local varieties

    61. English — Norwegian/German/Dutch/Swedish — Different languages. English is closest to Dutch, closer still to Frisian, but still not mutually intelligible.

    65. English — French/Italian/Spanish/Portuguese. Different languages, with no mutual intelligibility, despite cognates and borrowings.

    77. Albanian — Armenian — Different languages. Albanian is an interesting case. I've always understood it to have so many borrowings from other languages that it is hard to recognize the lexicon as Indo-European.

    90. Ossetian — Georgian Different languages, with Ossetian a Persian language, I think, and Georgian one of the Caucasian family.

    91. Mongolian — Turkish — Different languages, different families.

    92. Mongolian — Korean — Different languages.

    96. Romanian — Italian — Both Romance languages, but not mutually intelligible. I think Romanian has many Slavic borrowings.

    97. Basque — Spanish/Catalan I think Spanish and Catalan almost relate as dialects, but Basque is a different family.

    99. Romanian — Romani If you mean Romani as spoken by the Rom, I think it is an Indo-Aryan language and a completely different language from Romanian, which is a romance language with a lot of Slavic borrowings.

    Breton and Welsh are closely related, so that supposedly Breton onion sellers used to be understood by the Welsh. But I think the degree of intelligibility has been exaggerated.

    Scottish Gaelic and Irish are closely related, but I wouldn't call them mutually intelligible. The spelling differs, as does pronunciation, and there are many differences common phrases.

    What about the everyday languages of Judaism: Yiddish and German, Ladino and Spanish, even the dead Yevanic and Greek?

  75. Paul said,

    May 28, 2014 @ 11:36 pm

    I find it interesting that speakers of related and yet unintelligible languages often express sentiments like "I don't understand it but it feels as if I should." This was first said to me by a Finn regarding Hungarian twenty years ago, and I've heard it over and over since then.

  76. Fil said,

    May 29, 2014 @ 1:58 am

    Native Prague Czech and Australian English bilingual in his twenties here.

    17) Czech — Russian: There is minimal intelligibility without relevant instruction. I have no training in Russian and cannot understand spoken Russian at all, though I can sometimes puzzle fragments of written text out (the writing system barrier notwithstanding, of course). I'd say the rate of intelligibility hovers at around 5-15%; it is roughly the same, I would say, as the level of intelligibility between English and Swedish.

    18) Czech — Slovakian: They're widely seen as separate languages by the Czechs and Slovaks, due to the consequences of the late 19th century 'National Revival' process proceeding independently in both the Czech and Slovak lands. Linguistically, I consider them to be a very clear-cut example of a pluricentric language. Mutual intelligibility hovers at around the 90% mark both ways, with a very slight advantage held by the Slovaks in understanding Czech. This advantage is exposure-related and would be completely erased by one evening's worth of conversation, I should think. The intelligibility rate holds for the standard dialects; the rate may decrease for more distant regional dialects, particularly far-eastern Slovak dialects (transitional into Rusyn and Ukrainian). I keep hearing that younger Czechs struggle with Slovak but I think this is utter hogwash; Slovak is sufficiently comprehensible to all Czechs that TV news items about Slovakia are often presented by Slovak speakers in Slovak without any subtitles or translation.

  77. Jim Breen said,

    May 29, 2014 @ 3:11 am

    I'm glad Oskar Sigvardsson mentioned "The Bridge", with the two main characters speaking Danish and Swedish respectively. A good example. A program in Swedish/Norwegian would be just as successful, if not more so.

  78. TJL said,

    May 29, 2014 @ 3:50 am

    I'll just confirm what others have already suggested: as far as the Finnish-Swedish pair is concerned, it is extremely unlikely that an L1 Finnish speaker would have "no prior, formal study of or contact with the opposite language". Swedish is an official language in Finland and studying it (not necessarily learning it…) is mandatory.

  79. Ben Hemmens said,

    May 29, 2014 @ 4:53 am

    Dutch-German I would be more optimistic and say up to about 40-50%. Once you get the more surprising of the vowel sounds sorted out (e.g. ui as in huis and a as in van).

    But then I am an English speaker who has learned German, including plenty of practice listening in to different dialects, which probably gives me an unfair advantage.
    a. separate languages, yes
    b. one a dialect of the other, no
    c. both dialects of third language, no
    d. belonging to same group, yes
    e. belong to different groups, no

    Lëztebuergesch is in contrast generally classified as a dialect of German and that's my impression from listening to speakers, too.

  80. JQ said,

    May 29, 2014 @ 4:59 am

    With regards to southern topolects of Chinese vs Korean, I would say 0% intelligibility for spoken, unless a Korean was enunciating a single Sino-Korean word slowly with the intent of being understood by a Cantonese speaker, as I did at school with Korean friends when I was 10…

    After about an hour learning Hangul and playing with a Hanja converter, and reading the first chapter of a Korean language textbook, I find that I can decipher written Sino-Korean in most cases, not that this helps in any way with understanding the actual language.

  81. Bill W said,

    May 29, 2014 @ 5:27 am

    Shouldn't Azerbaijani be paired with Turkish? They are historically closely related languages, and I've heard that there is a high degree of mutual intelligibility. Chechen is a Caucasian language not at all related to either.

    Tigre and Tigrinya may be mutually intelligible. Amharic, although genetically related to both, has some unique features; I wouldn't expect a very high level of mutual intelligibility.

    Somali — Amharic: completely different languages.

  82. Stephen J said,

    May 29, 2014 @ 5:28 am

    My Bosniak friend refers to his language as "Nash" (ours) and suggests speakers of Serbian and Croatian do the same…

    As a second language learner of Portuguese, I would say my reading comprehension of Spanish is about 80%, and Italian around 40%. As a second language learner of German, Dutch, Afrikaans and the Scandinavian languages would be about 20% reading comprehension, spoken about 10%.

    As a New Zealander, you could add Maori – Cook Islands Maori, Maori-Tahitian, Samoan-Tongan and similar pairs of other Polynesian languages. When I was little comparing the parallel translations of these languages on official documents was a source of great fascination to me.

  83. tetri_tolia said,

    May 29, 2014 @ 5:29 am

    Interesting choices (Ubykh and Osuri?) to compare Georgian (sorry, but not in the slightest part of any "Caucasian family") to. Within Georgian itself there are quite some variances from the standard Kartli-Kakheti literary language — I never met anybody from Pshaveti or Khevsureti, but I've heard that especially the ones who never leave the mountains can be quite difficult for regular Georgians to grasp. I lived in a western region which was considered to have quite a unique dialect, which most Georgians seem to consider comical rather than incomprehensible. I did have the occasional realization, travelling to other parts of the country, that I had learnt by accident the Gurian for some common object rather than the standard Georgian, occasioning great merriment.

    Then of course there is Megruli/Lazuri and Svan, which are called "ena" (tongue/language) in Georgian and often referred to as "dialects" by Western linguists. I suppose this may be justified a bit considering that they are the only other members of the Kartvelian family and except for a few Laz they all consider themselves, without any complication, ethnic Georgians — not to mention that they are all bilingual with standard Georgian, being educated in it and using it to communicate with the rest of their countrymen (probably less true for some of the Svans living in the more remote mountains). I don't understand either of them, except for a few words of Megruli I picked up, and Georgians tell me that if you were never exposed to them as a child it's impossible to tell what's going on, but if you maybe live in Samegrelo or Svaneti for a year or two you will be able to speak fluently.

    Oddly enough, the non-standard Georgian dialects are written down quite a bit in Georgian literature, as the mountain culture and dialect holds a lot of romantic/nationalistic connotation for most Georgians — e.g. the work of Vazha-Pshavela — but almost every peculiar region has its own bard, like Nodar Dumbadze for the Guruli. However, Svan and Megruli are very rarely written. I remember Megrelians telling me that Megruli was not a written language, and when I showed them a book of Megruli poetry that I found (using Georgian characters, plus one letter picked out of an old Georgian alphabet to represent a sound that exists in Megruli and not Georgian) they would exclaim at how odd and funny it looked. ("But it can't be written down — it isn't done.") In fact there is a Megruli Wikipedia, referred to as "Marguluri" which is the Megruli/Mingrelian name for their language, so evidently somebody somewhere is doing some writing.

    As for the actual languages V.M. mentions, I don't know what could have possessed him to be curious about them. It's a bit impossible to test intelligibility of Ubykh with anything anymore, and I'm fairly certain that it never had any significant degree of language contact with Georgian. Probably it had some degree of mutual intelligibility with Apkhaz, but mutual intelligibility between Apkhaz and Georgian would probably be limited to a handful of shared Russian loanwords. Osuri would be more readily understood by an Armenian than a Georgian, as the Persian influence was a great deal more extensive in Armenian, and certainly more readily by an Iranian than anyone else, but even then I would doubt that there would be significant intelligibility. I don't understand a word of Osuri, and I doubt that they, absent cultural exposure, would be able to understand any Georgian.

    tl;dr — nobody understands anybody else in the Caucasus, too many damn mountains in the way for anybody to have needed to worry about that for most of the past millennia. Mostly we all resort to a kind of barbarous dog-Russian in the unlikely event we have to deal with someone who isn't from where we are.

  84. Alon Lischinsky said,

    May 29, 2014 @ 5:42 am

    3. Hebrew — Arabic: perhaps 5% intelligibility (the odd cognate word, but sound change makes even that usually unreliable; of course, that's assuming MSA, not any of the spoken varieties, which are even more remote). Answers: a=yes, b=no, c=no, d=yes, e=no

    11. Finnish — Swedish: perhaps 1% intelligibility coming from Swedish (a few borrowings, e.g., ‘pojke’←‘poika’). Answers: a=yes, b=no, c=no, d=no, e=yes

    23. French — Spanish: perhaps 10% intelligibility coming from Spanish (the odd cognate word, but sound change makes even that usually unreliable, and the European Wanderwörter). Answers: a=yes, b=no, c=no, d=yes, e=no

    25. Spanish — Portuguese: perhaps 50% intelligibility coming from Spanish. The phonological inventory seems odd, but cognates are usually transparent and the underlying grammar is substantially identical. (It's doubtful, however, that any speakers of my native Rioplatense dialect ever come to Portuguese completely cold, given widespread exposure to Brazilian TV and music). Answers: a=yes, b=no, c=no, d=yes, e=no

    26. Spanish — Italian: perhaps 80% intelligibility coming from (Rioplatense) Spanish. False friends can be treacherous, but comprehension is almost assured. (One has to thank the double whammy of the strong influence of L1 Italian speakers in Rioplatense, and the Latinisation of Spanish in the 16th century, which restored quite a few phonological correspondences that had been obscured by history). Answers: a=yes, b=no, c=no, d=yes, e=no

    27. Spanish — Catalan: perhaps 50% intelligibility coming from (Rioplatense) Spanish.The phonological inventory seems odd, but cognates are usually transparent, the underlying grammar is very similar (though less than Portuguese), and both languages have a history of mutual borrowings that simplifies things. Answers: a=yes, b=no, c=no, d=yes, e=no

    29. Mexico — Argentina: perhaps 98% intelligibility coming from Rioplatense Spanish. The intonation patterns and some of the vowel realisations seem unusual, but an accent and some colloquial lexis do not a language distinction make. Answers: a=no, b=no, c=yes, d=no, e=no

    50. Swedish — Danish: perhaps 20% intelligibility coming from Swedish, better if the Dane in question is older or not from Copenhagen. Both grammar and lexis have very clear similarities, but sound change in rigsdansk has made things very difficult. Answers: a=yes, b=no, c=no, d=yes, e=no

    51. Swedish — Norwegian: perhaps 60% intelligibility coming from Swedish, better for Østfold and other transitional dialects. Both Swedish and Norwegian have heavy regional variation, though, so it makes little sense to ask this in the abstract. Answers: a=yes, b=no, c=no, d=yes, e=no

    53. Swedish — French: perhaps 2% intelligibility coming from Swedish (borrowings, especially Wanderwörter). Answers: a=yes, b=no, c=no, d=no, e=yes

    62. English — German: <10% intelligibility coming from English (the odd cognate word, but sound change makes even that usually unreliable, and the European Wanderwörter). Answers: a=yes, b=no, c=no, d=yes, e=no

    63. English — Dutch: <10% intelligibility coming from English. Answers: a=yes, b=no, c=no, d=yes, e=no

    (continued below)

  85. Alon Lischinsky said,

    May 29, 2014 @ 5:44 am

    64. English — Swedish: <15% intelligibility coming from English. Old Norse borrowings and similar simplifications of the Germanic verbal patterns help compared with German or Dutch, but there's no way to maintain a cross-language conversation. Answers: a=yes, b=no, c=no, d=yes, e=no

    67. English — Spanish: perhaps 5% intelligibility coming from Spanish (borrowings, either direct or via French, and Wanderwörter). The phonology seems absolutely chaotic at first: why aren't there any stable vowels? Answers: a=yes, b=no, c=no, d=no, e=yes

    93. Quechua — Guarani: 0% intelligibility. The only words that have even a passing resemblance are common borrowings (e.g., Guarani ‘kavaju’ and Quechua ‘kawallu’, both from Spanish). Answers: a=yes, b=no, c=no, d=no, e=yes

    97. Basque — Spanish: 0% intelligibility. The only words that have even a passing resemblance are borrowings, either direct or from Latin. Answers: a=yes, b=no, c=no, d=no, e=yes

    98. Basque — Catalan: 0% intelligibility. Answers: a=yes, b=no, c=no, d=no, e=yes

  86. Faldone said,

    May 29, 2014 @ 6:05 am

    I'd add Navajo – Apache. And maybe to throw in some geographical propinquity with Hopi, giving us

    Navajo – Hopi
    Apache – Hopi

    added to the mix.

  87. maidhc said,

    May 29, 2014 @ 6:09 am

    Following on with a comparison of Irish and Scots Gaelic, native speakers I've talked to from both sides have said they don't have much trouble. I've met some Scottish Gaelic speakers who said that back in the old days when there wasn't much Gàidhlig on the radio they used to listen to Radio na Gaeltachta (Irish radio) without too much difficulty. And some Irish native speakers I've talked to have said they could converse with Gàidhlig speakers without too much trouble.

    The main difference is in the vowels, while the consonants are much the same. Some standard usages in one are unusual dialect words in the other. Also there are "false friends". But overall there would be a high degree of comprehension, similar perhaps to Swedish/Norwegian.

  88. Alexandru Pănoiu said,

    May 29, 2014 @ 6:23 am

    96. Romanian — Italian: I am a Romanian, and my personal experience is that standard Italian is quite easily intelligible, say 50 to 80% depending on the subject. As an example, in the bad old days of the 1980s, when the Socialist Republic of Romania had become almost bankrupt and there was nothing to watch on national TV, a very large number of people in Bucharest followed La piovra (The Octopus) on Bulgarian TV. It is not hard for a Romanian to learn Italian "by ear": Wikipedia says that on 1 January 2013 there were more than 1 million Romanians living in Italy.
    a) They are separate languages. b) None of them is a dialect of the other. (But, in the second half of the 19th century there was a feeble attempt by some Romanian men of letters to spread the idea that Romanian was indeed a "rustic" Italian dialect, and that Romanians should abandon the effort to modernize their language and just start using standard Italian.) c) They are not dialects of another language. d) and e) They are both Romance languages; while sometimes they are both classified as belonging to the same subgroup of Romance languages, more usually Italian is classified as Western Romance and Romanian as Eastern Romance.

    99. Romanian — Romani: Not mutually intelligible at all.
    a) Separate languages. b) Not dialects of each other. c) Not dialects of another language. d) and e) As far as I know all that can be said is that they are both Indo-European.

  89. tnv said,

    May 29, 2014 @ 8:07 am

    Native speaker of Russian, having acquired English, Spanish, French and some Hebrew. Trained as a linguist.

    Russian – Ukrainian – 50-60%. It gets easier with practice.

    Yes, they are separate languages but belong to the same group.

    Russian – Belarusian – I think 60-70%, but I haven't had much experience.

    I do not have enough experience to say, but I would side with separate languages.

    Czech – Russian – 30-40%. I did know a Russian and a Czech woman who would meet on the playground tending their respective charges, and would hold conversations, though they were likely quite simplified.

    Yes, they are separate languages but belong to the same group.

    Spanish/French – Portuguese – 50% or so. It is easier in writing, and it got easier once I got the sound correspondences.

    Yes, they are definitely separate languages, but belong to the same group.

    Concerning Swedish/Norwegian: I was flying SAS a couple of years ago and the flight attendant told me that he, a Norwegian, and his Swedish coworker were just speaking each their own language to each other.

  90. tnv said,

    May 29, 2014 @ 8:09 am

    Other relevant pairs: Zulu – Zimbabwean Ndebele, and Zulu – Xhosa.

  91. Victor Mair said,

    May 29, 2014 @ 8:20 am

    As a fluent speaker of Mandarin, and knowing a bit of Cantonese and Tawanese and Shanghainese and other varieties of Sinitic, I have to say that my comprehension of Korean conversations is virtually nil. When you're listening to other people talk, you don't have time to figure out sound correspondences, and even if you did it for a few terms, you'd be so far behind in following the conversation that you'd never catch up, PLUS — as a speaker of Mandarin and / or other Sinitic languages — you wouldn't have any command of Korean grammar, syntax, idioms, and so forth. Grabbing hold of a few Sino-Korean lexical items that you can pick out of a Korean conversation is light years away from understanding the whole conversation, or even a tiny portion of it.

  92. Matt Anderson said,

    May 29, 2014 @ 8:26 am

    I used to eat regularly at a Uyghur restaurant in Beijing, and it was common for Muslim traders from other countries to eat there. If I was there, and someone from, say, Sri Lanka came in, I would sometimes speak to them in English and then translate (into Mandarin) for them. But if Turkish traders came in, they would just speak in Turkish, and the people in the restaurant would respond in Uyghur, and it surprised me that this seemed to work.

    When I asked, they replied that if everyone spoke slowly and loudly, they could understand about 60% of Turkish, which I think must be a giant overstatement (no one at the restaurant had ever studied the language, and, like Ben says above, I always thought the two languages were not really mutually intelligible), but in any case they didn't seem to have major problems taking orders & making small talk. They did say, though, that Turks had a much harder time understanding Uyghur than vice versa.

  93. Ross King said,

    May 29, 2014 @ 8:41 am

    I'm afraid I must join the chorus of Kellen Parker's friends scoffing at his assertion that one can readily 'pick up' Hakka on the basis of a knowledge of Sino-Korean (or vice versa). Sure, there are clear similarities between Hakka readings of Chinese characters and sinovocabulary shared across the former sinographic cosmopolis. But to then claim the ability to understand spoken conversation beyond random, fleeting snippets of the odd word would be sheer fantasy, and rather like claiming that all the Kazakhs and Uzbeks from the auls can immediately pick up Russian because they know a few words like kommunizm, internatsionalizatsiia and traktor when they get to the big city.

  94. Bill W said,

    May 29, 2014 @ 8:51 am

    Tagalog – Cebuano?

    Chamorro – Tagalog?

    Yupik – Inuktitut (or Inupiat)?

    Greenland – Nunavut?

    Icelandic – Faroese?

    On the mutual intelligibility of the Scandinavian languages: aren't there classes in the standard school program in which youngsters in each of Denmark, Norway and Sweden are taught or at least exposed to the basics of the other two languages?

  95. Gassalasca said,

    May 29, 2014 @ 9:53 am

    A native speaker of Serbian here.

    Saying "Serbian and Croatian use different scripts" is oversimplifying things a bit. Whereas Croatian uses only the Latin script, Serbian uses *both*. What's more, taking everything into account, the Latin script is more common.
    Government documents and websites insist on Cyrillic, most shop signs are in Latin, and regarding books/magazines/newspapers both are used widely, but, again, overall, I think you'll see more things printed in Latin. Though when it comes to personal signatures, Cyrillic seems to still prevail.
    Finally, if you see a person using Cyrillic on Facebook, Twitter, blogs or forums, nine cases out of ten, they're right-leaning.

  96. Victor Mair said,

    May 29, 2014 @ 10:10 am

    The Japanese sure use a lot of loan words from English, and even make up plenty of their own Japanese-style English words, but I dare say that a monolingual speaker of English would understand far less than 1% of a Japanese conversation.

    Knowing a few isolated loanwords or cognates in a language other than one's own mother tongue does not make one in any way proficient in that other language or capable of understanding normal speech in it.


    Some references for Eigo shakuyō-go 英語借用語 ("English loanwords") in Japanese

    "Too many English loanwords in Japanese?"

    and vice versa:

    "Too many recent Japanese loanwords in English?"

  97. Gnoey said,

    May 29, 2014 @ 10:24 am

    I grew up speaking English and Mandarin, but hearing a lot of Cantonese and Hokkien (Fujian/ Amoy dialect) around me, and can understand the latter two quite well, probably 80-90% of the time, though I've never been formally taught. Interestingly, my sister, despite growing up in the same household and attending the same schools as me, cannot understand those two Chinese dialects.

    I have always found Thai and Vietnamese sounding quite similar to Chinese, probably because they are tonal languages, although I don't understand them at all. I remember when I first heard them, I thought they were some unfamiliar Chinese dialects! When I visited Laos, Laotian sounded like Thai, and I was told that most Laotians understood Thai because of the strong influence of Thai culture, but not the other way round.

  98. Peter said,

    May 29, 2014 @ 10:38 am

    @ Victor Mair

    "I don't quite know how to respond to your quiz as you outlined; instead I personally feel that Dutch, Flemish, Afrikaans and Yiddish are some form of German dialect, although Dutch and Yiddish I always have to "sound out" to comprehend due to their very strange spelling."

    This is very interesting, as a native dutch/flemish speaker i can understand some german, though very little but no Yiddish at all. Maybe my exposure to Yiddish has simply been to small and i'm therefor underestimating my understanding of it ?

    @Victoria Simmons


    There are several possibilities of what language may be called flemish. Either the intermediate language thought in schools in the flemish region of belgium. The collection of dialects spoken in the flemish region of belgium being, Limburgish, Brabantian, East Flemish and West Flemish. Or maybe just West Flemish.

    The intermediate is almost identical to dutch, with only few non shared words and mainly pronunciation differences sometimes smaller then the ones found between Flemish speakers themselves.

    Limburgish is closer to german and might be a bit more difficult to understand, grammar is unique and so are many words. Western flemish has a low intelligibility with dutch, this means that most of the flemish speakers don't understand (true) western flemish due to pronounciation and non shared words. Brabatian and Eastern flemish are both 90% intelligible with both Flemish and Dutch.

    Brabantian is so similar to standard dutch and intermediate flemish that it might even be impossible to distinguish is you aren't a fluent speaker.

    So i would say they are intelligible :)

  99. Mr Punch said,

    May 29, 2014 @ 10:52 am

    A couple of comments on "sound" and register:

    I'm an English speaker (American) with French as a second language. I can understand Italians from Turin much better than those from Naples – don't know if they're actually using the same words, but one sounds much more like French. I understand Catalan better than Spanish. I can't follow German at all. Dutch people speaking English sound to me like native speakers with unplaceable accents (maybe sort of Canadian), even when their vocabularies are quite limited.

    As to register – my father, a scientist of French origin who had studied German, reported that he could understand oral deliveries of papers in his field in any Romance or Germanic language.

  100. John Walden said,

    May 29, 2014 @ 11:22 am

    I find that as a fluentish Spanish listener I seem to understand spoken Portuguese better than the Spanish speakers I'm with at the time.

    They seem to have a narrower spectrum than I do of what constitutes a word close enough to their word for it to be recognisable. Which is odd because there are massive differences between pronunciations across the Spanish regions, but then a typical Spanish native is familiar with all of them and the differences are mostly with consonants: the vowels remain fairly unchanged between different Spanishes, as far as I can hear.

    Then the fact is that some people can understand cognate languages but won't. Portuguese speakers don't always care to understand Spanish.

    As an aside. do the scattered Mennonite communities across North America understand each other? And would they understand present-day Platt?

  101. julie lee said,

    May 29, 2014 @ 11:27 am

    My best language is English, but I can understand Mandarin and Cantonese, and all the Mandarin-like dialects like urban Sichuanese and Hunanese. Can understand a little Shanghainese, and can understand 0% Hokkienese, Taiwanese, Shangdongese, etc. etc.
    Taiwanese and Mandarin are very different in pronunciation even though there are many common words such as "eat (a meal)", "Jia bung吃飯" in Taiwanese and "chi fan吃飯" in Mandarin ("sig fahn吃飯“ in Cantonese, "je veh 吃飯" in Shanghainese, the "je" like a French "je").
    However, because of the very large shared vocabulary, I am confident I can pick up urban Taiwanese, Shangdongese, Shanghainese, Hokkienese, Suzhou-ese and many others very quickly if I learn from a native speaker, even though those topolects are very different from one another, and from the topolects I know, Mandarin and Cantonese.

  102. spz said,

    May 29, 2014 @ 11:29 am

    I'm a native speaker of southern German. I visited Amsterdam with ~7 years of English lessons behind me (which meant decent general understanding of English) but none of Dutch.

    I could understand the gist of people talking in the street (talking about musical street performances), but I was completely lost with written Dutch.

  103. Victor Mair said,

    May 29, 2014 @ 11:33 am

    From a Korean language specialist who is also proficient in many other East Asian and European languages:

    Kellen exaggerates by saying one can "get along all right" with Korean on the basis of Hakka proficiency. As you point out, the two languages are unrelated. Can an English speaker get along in Russia on the basis of some shared high-level vocabulary? Everyone knows the answer. Hell, I can't even get by in Paris.

    That said, the languages do sound similar to someone not native to either, because they both have CVC syllables and many monosyllabic morphemes from shared Sinitic vocabulary. I remember getting off an airplane in Taipei after 9 months in Seoul and being struck by how much the Taiwanese–not Mandarin–sounded "like" Korean. I mean phonology, not syntax, and certainly not understanding. My Korean, which I continued studying in Taiwan, never helped me comprehend the Taiwanese around me in any practical sense, or the Hakka I was exposed to in my Yongho enclave and through contacts with a colleague compiling a Hakka glossary.

    Because many of the modern non-standard Chinese languages split off (this is a lousy metaphor) from what became "standard" Chinese at about the same time J K V were absorbing Sinitic loans, and because they all preserved the Ancient Chinese phonology better than Mandarin, there are cases where you can guess a Sinitic word (not phrase, not discourse) spoken in an unfamiliar language on the basis of a familiar one if you know the context.

    My Vietnamese wife and I do this all the time. She'll ask me the meaning of an English word, usually some high-level term, and if I can't remember the Vietnamese but do know another Sinitic equivalent (e.g., Mandarin or Korean), I'll throw that at her or use the latter to manufacture a Vietnamese term from my knowledge of the SV sound transitions. When it works, which is most of the time, she'll offer a guess at the correct Vietnamese, which I can usually confirm from my passive recognition of spoken Vietnamese.

    If this seems pretty convoluted, it is. And it's certainly a far cry from "getting along all right." I should also point out that I'm not a typical example, having academic and practical experience in most of the languages. So my verdict is Kellen needs to temper his claim or risk being marginalized by serious students.

  104. julie lee said,

    May 29, 2014 @ 11:39 am

    A question:
    When I arrived in London some years ago and moved into an apartment, the movers, all white men, all spoke what I presumed to be Cockney among themselves. I couldn't understand the Cockney at all. It sounded like a foreign language, though English is my first language. Would Cockney English be a dialect or a language?
    (Perhaps if they spoke very slowly I'd have de-coded it, but at the speed they spoke I couldn't understand anything. What little English I spoke to them, they understood, such as "Put it here" or "This goes upstairs.") So if Cockney English is unintelligible to a Standard-English speaker, is Cockney a language?

  105. zoetrope said,

    May 29, 2014 @ 11:41 am

    A few factors that I imagine influence this:

    1. Personal desire to understand the other language

    It seems to me that many Spanish speakers (I live in Madrid) claim they can't understand any Catalán, Valencian, etc. but I can certainly understand a fair bit of these languages/dialects when I hear them spoken, so I have a feeling part of it is willful disinclination on their part to understand.

    2. Exposure to other languages in general (not necessarily the language in question)

    People who have never learned another language often seem to have more difficulty picking up on patterns that could lead to partial understanding of another language. And again, using Spanish as an example, because they don't distinguish between many phoneme pairs that other languages do, I think it can be harder for Spanish speakers to understand say, Italian (which differentiates between b and v, sh and s, z and s, etc.) than for Italian speakers to understand Spanish.

    Regarding the language pairs I have some experience with, I would say:

    – Egyptian vs. Levantine Arabic: 50% probably the closest two groups of Arabic, but with 0 exposure to the other, a lot of basic expressions would be unintelligible. As other commenters have noted though, most Arabs have some exposure to the Egyptian dialect through music, TV, etc.

    – French vs. other Romance languages: 5-10%
    Despite similar grammatical systems, French sounds so different from the other languages that I imagine there isn't much mutual intelligibility (until you get to Northern Italian dialects anyway)

    – Spanish vs. Portuguese & Italian: 50-80%
    Depending on which direction you're going from. As people have said, it seems easier for Portuguese and Italian speakers to understand Spanish than the other way around.

    – Madrid vs. other (non-Spain) dialects: 75-95%
    It certainly took me awhile to get used to the rhythm and melody of Mexican Spanish, but once I did it was mostly a question of different vocabulary items that made it a bit difficult to understand everything. I still have problems understanding everything my Argentinian friends say.

    – Spanish vs. Catalan: 20-50%

    – English vs. Romance languages: 5%

    – English vs. Dutch: 10%

    – English vs. German: 0-5%

    – Dutch vs. German: 10%??

  106. Rodger C said,

    May 29, 2014 @ 12:17 pm

    When I was stationed in Panama 1970-71 I'd occasionally, at night, catch the radio from Bonaire. Listening to Dutch was a fascinating and frustrating experience. Having had two years of college German, I felt that I was always just on the verge of understanding Dutch from one side or the other, but never quite. It was like trying to touch fingers around a treetrunk that was an inch too big.

    Sociolinguistic note: This news analysis program in Dutch was preceded by a gospel hour in Papiamentu, which at first I mistook for Portuguese.

  107. Darkwhite said,

    May 29, 2014 @ 12:29 pm

    Alon Lischinsky:
    Your figure for Swedish-Danish seems very low. In my experience, they have very little trouble conversing with each other or watching television in the other language.

    Victor Mair, regarding Korean:
    I would imagine the situation would be somewhat like someone going to France, speaking only English and German – you understand next to nothing at first, but you pick it up probably ten times faster than a truly foreign language.

  108. Vasha said,

    May 29, 2014 @ 12:40 pm


    In fact the increasing use of "Hindiisms" due to the popularity of Indian media in Pakistan seems to be a topic of reactions ranging from bemusement to annoyance to horror in the Pakistani media. See for example: ‘Vishvas’: A word that threatens Pakistan.

    Intriguingly, the reverse situation is not seen as a problem in India, with even some right wing politicians counting themselves as connoisseurs of Urdu poetry,

  109. Piyush said,

    May 29, 2014 @ 12:53 pm

    Sorry, I somehow managed to sign my last comment—posted in response to Vasha—as Vasha.

  110. GH said,

    May 29, 2014 @ 1:23 pm


    Well, it just goes to show how subjective this whole thing is. Swedish-Danish 20% seems maybe a little low to me, but I find your estimates of… well, everything much too high. If we're assuming someone with no prior exposure (which is almost inconceivable for most of these language pairs) I'd go with something more like:

    Norwegian-Swedish: 75%
    Certainly enough for conversation once you got used to some of the sound changes, but there'd be unfamiliar words in most sentences.

    Norwegian-Danish: 50%
    Sound system is different enough that even with a large shared vocabulary and grammar, someone with no familiarity would have great difficulty making out enough to be able to understand what was being said.

    Norwegian-German: 5%
    You might be able to catch the odd cognate here and there, but without any training to be able to spot the sound changes, many would be very hard to identify (plus you'd risk getting fooled by a whole lot of false friends), and without knowledge of the grammar you couldn't hope to understand any complex utterances. The idea that "casual exposure" would take the average person to 80% comprehension strikes me as a fantasy (at least of the spoken language; it would be somewhat more realistic for reading). Someone with two-three years of school German might be able to get there.

    Norwegian-Dutch: ?
    I don't know it well enough to say; presumably similar to German (maybe a little higher?).

    Norwegian-English: 2%
    As with German, except the large Romance influence dilutes the common features. This is very much a thought experiment, as it would be pretty much impossible to find any L1 Norwegian speakers who haven't had extensive exposure to English.

    Norwegian-French: <1%
    If you were lucky you might be able to recognize the odd loan-word, though they would most likely drown in all the stuff you didn't understand.

  111. prase said,

    May 29, 2014 @ 1:32 pm

    Whenever I write "dialects", I mean the two languages are dialects of the same language rather than that one is a dialect of another; there is no asymmetry in the relation, especially when we are interested in spoken forms only.

    Czech – Slovak: Nearly 100%, certainly dialects. Although I am a native Czech speaker, it is not easy to assess how difficult Slovak really is with no prior knowledge, as I have been exposed to it since childhood.

    Czech – Polish: A borderline case. Without prior exposure the level of understanding may be as low as some 20%, but after getting used to typical Czech – Polish phonological correspondences, one may get at 60% without actually learning any vocabulary. In my experience Poles have greater difficulty understanding Czech than vice versa, but how much this is a linguistic rather than sociological or cultural phenomenon ("we shouldn't try to understand a language of a four times smaller nation") I don't know.

    Czech – Russian: Distinct languages, 40%.

    Czech – Bulgarian: Distinct languages, 20%. Furthest one can get in the Slavic continuum.

    Russian – Ukrainian: 80%, dialects.

    Serbian – Croatian: 100%, the same language, not even dialects.

    In my experience it is nearly impossible to say how much I understand from a particular language, as with different individual speakers, contexts and situations, the understanding of the same language can vary anywhere between 0% and 100% (assuming there is a degree of intelligibility at all). Also, moving from a perfectly intelligible speech away to a completely foreign one, the amount of understood information doesn't decrease continuously; rather, there is a threshold where with a tiny increase of unintelligibility I lose the context and consequently my understanding rate falls drastically, say from 80% to nearly 0%. Therefore the percentages I have given are meant to represent the rate of understanding of a very slow and clear speech, with repetition available on demand. And of course, it's more often than not a wild guess, but that I understand is expected here.

  112. Martha said,

    May 29, 2014 @ 1:54 pm

    Bill W suggested Tagalog and Cebuano, and although I'm a proficient speaker of neither, I'd have to say there is very little mutual intelligibility. Although I've never interacted in Cebuano and have interacted VERY little in Tagalog, having a Visayan mother, I've been exposed to both languages my entire life. Once in Manila, my mother was speaking to a Visayan cousin and another cousin who had grown up on Luzon asked me if I could understand them, because she couldn't understand them at all. I could understand the gist, but I have to say she was exaggerating about not being able to understand them, because part of the reason I was able to understand them was because key words were in English.

    Which I think would be an issue regarding the mutual intelligibility of any Filipino language: the amount of English loanwords/codeswitching often prevents people from consistently speaking only one language at a time. (Although I must admit the the bulk of my observation has taken place in the U.S., where people perhaps do more codeswitching.)

  113. Pflaumbaum said,

    May 29, 2014 @ 2:01 pm

    @ Alexandru Pănoiu –

    Would you agree with my suggestion that Romanian is less easy to understand for Italians than Italian is for Romanians? Probably mostly because of the /ɨ/ and /ă/, and the Slavic influence in the lexicon?

    @ Julie Lee –

    Cockney is an accent, but the issue is complicated by the fact that many traditional Cockney speakers speak in non-standard dialect.

    So your builders may well have been saying things like 'He still ain't done nothing about it though', which night be realised as something like [ɪj stɪw ɐinʔ dɐn nɐfɪŋk əbaːʔ ɪʔ vɐw]. But you could quite easily have a Cockney speaker using Standard English 'He still hasn't done anything about it though', which might sound like [ɪj stɪw æznʔ dɐn ɛnɨfɪŋk əbaːʔ ɪʔ vɐw].

    Then there is the further complication of Cockney Rhyming Slang, which is essentially a kind of code and would have likely been opaque to you even if you'd understood the accent. (E.g. 'apples', short for 'apples and pairs', to mean 'stairs'). But it's not much spoken anymore. In fact Cockney in general is on the wane in London, where it's being replaced by accents like MLE. It's still alive and well though in the surrounding counties, in some of which it is itself replacing previously dominant accents.

  114. Victor Mair said,

    May 29, 2014 @ 2:20 pm

    @julie lee

    "… I am confident I can pick up urban Taiwanese, Shangdongese, Shanghainese, Hokkienese, Suzhou-ese and many others very quickly if I learn from a native speaker, even though those topolects are very different from one another, and from the topolects I know, Mandarin and Cantonese."

    You are right, the point being that you would have to learn these languages from a fluent speaker, just as a speaker of English would have to learn French, German, Spanish, Norwegian, Swedish, Dutch, Italian, etc., even though English shares a lot of cognates with them — in the case of French as much as 60%. No speaker of English is able to comprehend French, even a small amount of it, by virtue of being fluent in English. He / She must take the time to learn French. The same is true of Cantonese, Hakka, Taiwanese, Shanghainese, and all the other Sinitic languages for a Mandarin speaker.

    As for the Cockney issue, Pflaumbaum has explained it very clearly. It's an accent, not a language. It's like a very thick southern accent in America. If you listen to such accents for a few days, you get used to them, and — if you're not careful — you might even find yourself starting to talk a bit like them!

  115. David B Solnit said,

    May 29, 2014 @ 3:10 pm

    After all that, aren't there any empirical studies that actually take speakers of two linguistic varieties, sit them down, and find out how well they can understand each other? I'm not sure how you would extract a numerical measure out of that, but you would surely find out something that's distinct from all the (very interesting) impressionistic views and speculation that we've seen here so far.

  116. Darkwhite said,

    May 29, 2014 @ 3:29 pm

    I'm not as much objecting to the specific percentages, because we have no sensible metric here. As a trivial example, if you hear 'He punched her' and thinks it means the girl punched the guy, how many percent did you understand? It's almost inevitable that we come up with different scales. Personally, I try to reserve 0% for entirely unrelated languages such as Swahili, Russian and Korean, in which case it's tempting to put NOR-ENG in the twenties.

    What puzzles me is how his SWE-NOR is at 60% (it should likely be larger than SWE-DAN, but why such a large difference), and particularly the ESP-ITA at 80%. SWE-NOR-DAN will generally all just talk their respective languages with each other, even when they all speak good English. This is sort of possible with ITA-ESP too, but I have trouble imagining it's much more effortless than SWE-DAN.

    Not to start a huge discussion, I found his post very interesting and we're all just ballparking anyway, but I wondered about the rationale. For instance, I'd be very interested to learn that ESP-ITA, MEX-ARG, ESP-CATA are more similar than SWE-DAN, because I've always wondered about these language-pairs that I have no way of judging personally.

  117. Victor Mair said,

    May 29, 2014 @ 3:29 pm

    @julie lee

    Another reason why Cockney is a dialect or accent of English, not a separate language is that the movers were monolingual Cockney speakers, yet they understood what you were saying in standard English.

  118. Ben Hemmens said,

    May 29, 2014 @ 3:56 pm

    No formal study or contact and then? Does it count if you can start understanding what people are saying just by listening to them for a few hours, or a few days?

    If not, I'd have to record an experience of English-English, 50% or less: based on my very first trip to Scotland, for a job interview. It was a stormy day and I had flown over to Edinburgh in an unpressurized plane. Now its true that my ears weren't working properly afterwards, but when the train stopped in Cupar because of a fallen tree on the tracks, I really didn't understand what the people were saying. Everyone trooped off the train and a bus appeared and I followed them on to it. Now the language that most people speak in that part of the world, if you transcribed it, I think would turn out to be more Scots-colored English than pure Lowland Scots; the main problem is just the pronunciation and accent.

  119. Tim Friese said,

    May 29, 2014 @ 4:23 pm

    zoetrope said: "- Egyptian vs. Levantine Arabic: 50% probably the closest two groups of Arabic, but with 0 exposure to the other, a lot of basic expressions would be unintelligible. As other commenters have noted though, most Arabs have some exposure to the Egyptian dialect through music, TV, etc."

    I suppose it depends on how we count groups. As an urban Levantine speaker, I used to find bedouin-influenced gaaf dialects of Iraq, northern Saudi Arabia, and eastern Syria and Jordan to be easier to understand than Egyptian, although my exposure to Egyptian has reversed things for me. It may also help that eastern Syrians and Jordanians are used to making accommodations when speaking to urbanites, and my dialect is clearly 100% urban.

    Many people would claim that these dialects of eastern Syria and Jordan are Levantine by definition since they exist in states that form the Levant, but others would say they are properly Arabian peninsula dialects.

  120. GH said,

    May 29, 2014 @ 4:26 pm


    Yes, I agree that the numerical scale is highly arbitrary, and as @prase points out, in practice understanding of one language can easily range from 100% to just above 0% depending on a lot of different factors.

    I would also extend @prase's point that understanding does not decrease linearly with language distance, so that even though Norwegian-English are related and Norwegian-Korean are not (or only so remotely that we cannot detect it), that doesn't help you much until you learn enough to be able to draw those connections. It might be the difference between 0% and 5% comprehension, but hardly 20%, in my opinion (however we were to measure it).

    I don't know either language well, but Italian-Spanish 80% sounds surprisingly high to me, too. However, someone might argue that the Scandinavian languages are less mutually comprehensible in principle, and that a greater degree of understanding in practice is due to greater exposure.

    As for The Bridge and other cross-country TV shows, surely the foreign bits are subtitled in each country? I know that Norwegian TV always subtitles Swedish and Danish shows, and Norwegians are supposed to be the best at understanding both neighboring tongues.

  121. Victor Mair said,

    May 29, 2014 @ 4:40 pm

    @David B Solnit

    Linguists have attempted to devise mechanisms for arriving at the sort of "numerical measure" that you propose, but have come to the conclusion that there is no hard and fast way to determine the exact amount of comprehension that exists between any two speakers of two given languages. First of all, there are too many variables, and secondly, understanding by its very nature is subjective and impressionistic.

    We do know for certain when we don't understand a single word of what the other person is saying, and then we can say that intelligibility is 0%. That happens a lot when we travel to unaccustomed places. We feel totally lost, not being able to exchange so much as a coherent sentence with the other person. Other than that, though, we can never be sure exactly how much of what the other person is saying we understand. That's why all we can do is say things like "next to nothing", "a little bit", "about 5%", "about half", "around 75%", "almost everything".

    I have seen widely cited studies which attempt to predict the amount of intelligibility between languages on the basis of shared cognates, but I find them highly unsatisfactory (in fact, totally useless) because they don't take into account pronunciation, accent, grammar, syntax, idiomatic usage, and so forth.

    BTW, even for two native speakers of the same language, it's rare that we understand absolutely everything each other says. There are many reasons for this, including parapraxes, garbling, lapses, unintentional mispronunciation, differences in vocabulary and usage, and so forth. Even though I'm an educated native speaker of English, I find that I often miss about 5-10% of what's being said in movies because it is not clear enough, is mumbled, is covered up by other sounds, consists of jargon and slang with which I am unfamiliar (this is especially true of gangster talk and speech patterns of those under thirty or so), etc.

  122. Jongseong Park said,

    May 29, 2014 @ 4:43 pm

    Many people have already commented on Kellen Parker's extraordinary claim that knowing any two Sinitic languages will be enough preparation for understanding Korean in no time.

    As a native speaker of Korean, let me just say that it's like saying that someone who knows English and French will be all set to understand Modern Persian. Sure, Persian uses a lot of loanwords from English and French, and in certain registers and circumstances spoken Persian may be liberally peppered with such loanwords so that one may pick up on them, but there's a long way to go from that to getting a gist of what they're saying. And English and French at least share a very distant common ancestor with Persian, while there is no discernible genetic relationship between Korean and any of the Sinitic languages.

    Similarly, I can recognize quite a bit of English words or Sino-Japanese words in Japanese based on passing familiarity with how they are adapted to Japanese phonology. But these only help with content words. Despite the similarity in grammar between Korean and Japanese, if I didn't already know some Japanese particles, I'd be completely lost. This is a slightly different case because being able to speak English or knowing Sino-Korean words doesn't by itself guarantee that you would recognize their forms in Japanese. But I think what Kellen Parker is saying is equivalent to saying that this sort of obstacle to recognizing borrowed words in Korean is less if you know a couple of Sinitic languages. This may be correct, but this is well, well short of intelligibility.

    Getting the gist of what is being said in Korean would require at least getting some sense of how the real basic things like negation, declaration, question, etc. are expressed in Korean, and being able to recognize a few Sino-Korean content words will be no help here, any more than knowing Latin and Greek will help you understand a sample of English, however full of terms of Latin or Greek origin it may be.

  123. Darkwhite said,

    May 29, 2014 @ 5:26 pm

    GH: (Thanks for the stimulating conversation)

    I can't speak for Broen [The Bridge], but any number of Norwegian shows includes the odd Swedish or Danish character without any subtitling, for instance Rolf Lassgaard in Dag. Example here (the blonde is Swedish):

    As for the precise ENG-NOR percentages, I'm not sure if 20% is really too high, but it depends on the sort of situations you choose to imagine. When there's a lot of contextual clues and nothing too complicated is going on, a few cognates and similar grammar can really make a difference.

    And the reverse is of course also true – a random line from an epic poem will likely be nothing but nonsense, either for the purposes of NOR-ENG or NOR-KOR. That said, sitting in on a dinner table conversation in Vietnamese, I felt there was a very vivid difference between having a at least faint idea about the topic, when to nod and smile and when to pass the salt, as opposed to having literally no clue.

  124. Pflaumbaum said,

    May 29, 2014 @ 6:04 pm

    Sorry, a couple of errors in my post above – the curse of the LL comments section abides.

    For what it's worth, I meant "/ɨ/ and /ə/" in Romanian, not "/ɨ/ and /ă/". And I suspect most Cockney speakers wouldn't vocalise /l/ in 'still' before a following vowel.

  125. John Rohsenow said,

    May 29, 2014 @ 6:23 pm

    I guess we all have our stories:
    In about 1974 a distant in-law showed up in Ann Arbor from Canada
    with a friend from the Isle of Man; apparently he had lived in Canada for many years. Try as I might, and with several years of grad. works in linguistics at U of M, I STILL could only understand about 50% of what
    he said. His friends seemed to have no trouble and were used to "translating" for him.
    On the basis of having studied Hindi (sic) in India in the Peace Corps,
    as the resident linguist, I was delegated by our dean to make up a syllabus to introduce Hindi-Urdu (sic) to our faculty senate for approval. An Urdu-speaking engineering faculty member stood up in the Senate
    and swore that he could "not understand a word of Hindi", even after
    he replied when I spoke to him in my very rusty Hindi. (I assume he must have been talking about writing.) The motion passed anyway.
    My colleague the linguist Adam Makkai, a native speaker of Hungarian, said that he could not understand Finnish, but that here is
    a children's rhyme used in both languages which can be understood by
    speakers of both.
    My brother-in-law, whose first language was Quebec French (Quebequois?) said that they closer he got to Normandy, the easier it
    was for him to be understood.

  126. julie lee said,

    May 29, 2014 @ 8:24 pm

    Thanks, Pflaumbaum and Victor Mair, for responses to my comments.

  127. Kellen Parker said,

    May 29, 2014 @ 8:36 pm

    I feel that people aren't really responding to what I said at all, but rather to what they half-remember me saying by the time they got to the comment box.

    Yes, I said "you'd be set in a very short amount of time", but I hardly meant that as you'd be fluent.

    "in a week they'd be fine" is again being taken to mean that they'd be fluent in Korean. This is not my intention at all. My Cantonese/Mandarin speaking friend still studied, still spent a good amount of time trying to speak to people in Korean. It's not like she simply oozed fluency after a week of doing nothing. It wasn't a week to fluency, and it wasn't by doing nothing. But then I never said any of that.

    Victor said "I can pick up a few Sino-Korean words when I listen in on Koreans talking to each other… but do you really understand what they're discussing if they're talking at normal speed?"

    I think you're not accounting for context. If it's two people in line at the restaurant standing in front of you and looking at the menu, I don't see what's supposed to be so hard about picking out the foods they're discussing.

    I'm not sure how "get along alright" is an exaggeration, unless you think Dashan gets along alright in Chinese.

    Jongseong Park said "…Kellen Parker's extraordinary claim that knowing any two Sinitic languages will be enough preparation for understanding Korean in no time."

    It's only extraordinary if you think I meant fluency in three months, which I didn't and which there's not much in my original comment that can be taken that way.

    "I think what Kellen Parker is saying is equivalent to saying that this sort of obstacle to recognizing borrowed words in Korean is less if you know a couple of Sinitic languages. This may be correct, but this is well, well short of intelligibility."

    Nor do I claim it as such. To repeat myself from my original comment:

    "All that said, objectively quantifying mutual intelligibility is a mess, so I still feel —it's not really all that useful of a measure.—"

    I've used a bunch of phrases like "get along" and "do alright" without making any claims like you'd be fluent in a month. I stand by my original claim, which isn't really being reflected in responses to my comment.

  128. Son Ha said,

    May 29, 2014 @ 10:50 pm

    When I was working in Viet Nam and encountered minority languages like the Bana, J'arai, Khmer and E'de I seriously had no clue what they were trying to tell me. I then had to seek the help of an interpreter. When I was in Cambodia, same problem so Vietnamese-Ethnic Language or Khmer = 0

    Thai-Khmer: A Thai friend of mine said she could understand Cambodian workers in Thailand and many of her friends employed Cambodians for house-helpers since they understand more than half of the Khmer being spoken if not more. So I would say there is some mutual intelligibility.

    Hindi-Urdu: Yes mutual intelligibility. Our case workers fluent in Hindi don't need an Urdu interpreter when working short term in Pakistan.

    Vietnamese-Cantonese: When my siblings and I were smaller we watched many re-run of Hongkong wuxia series and were frustrated that we couldn't understand the Cantonese although we FELT that we should, it's almost like listening to a conversation outside of one's earshot and just ALMOST but never able to fully grasp what's going on. Though a friend of mine claimed she could understand a lot of Cantonese after several years of watching those shows.

    Nepali-Bengali: A Bangladeshi colleague working with me said he could listen in on a Nepali conversation and understand the majority of what's being said.

  129. Vanya said,

    May 30, 2014 @ 1:51 am

    From personal experience I would guess the intelligibility of spoken Highland Scots 0 for native American English speakers. I suspect many British people would have difficulty with certain Appalachian dialects in the US.

    Ukrainian-Russian. I was in Kiev once with Muscovites who told me that they could not understand a Ukrainan speaking tour guide. I found that odd because I, as a fluent but non-native Russian speaker, could follow about 70%-80% of what the guide was saying.

    German-Dutch: Oddly, since moving to Vienna and being exposed to significant amounts of Viennese dialect on a daily basis, I find I understand spoken Dutch better than I used to, although Dutch is the other end of the continuum. Maybe the "German" linguistic centre of my brain has expanded its definition of what is acceptable as a "German" sound.

  130. Jongseong Park said,

    May 30, 2014 @ 3:11 am

    @Kellen Parker:
    Thanks for responding. The problem with your claim is that you were making it in a discussion about mutual intelligibility, where other commenters cannot be faulted if they assume that you are claiming a degree of mutual intelligibility between Korean and Sinitic languages.

    Even you said at the end of your comment that "objectively quantifying mutual intelligibility is a mess", as if you had been talking about mutual intelligibility.

    But you were not. You admit that your "Cantonese/Mandarin-speaking friend still studied, still spent a good amount of time trying to speak to people in Korean", something you didn't specify in your first comment. When we are talking about intelligibility based on a prior language, we don't assume that we have studied the language in question. So you were making a comment about a language being easier to learn with other languages already in your arsenal, which is different from a question about mutual intelligibility.

    Japanese is the easiest language to learn for Koreans because of the similar grammar and a large amount of shared loanwords. I have no problems with the claim that Koreans who start studying Japanese and practice with Japanese speakers can already understand a fair amount in a few weeks. But that has no bearing on the question of mutual intelligibility between Korean and Japanese, which is near zero.

  131. joanne salton said,

    May 30, 2014 @ 3:22 am

    I remember Spanish speakers in Mexico had little difficulty with Brazilian TV. Koreans/Chinese/Japanese all understand virtually nothing of each others programmes.

    This is a much better measure than actual conversation because that depends so much on the attitudes as well as linguistic nous of those involved. I can't easily understand a Glaswegian if he/she treats me like a fellow Glaswegian.

  132. Alon Lischinsky said,

    May 30, 2014 @ 5:42 am

    @Darkwhite: well, I did make some caveats.

    First, I find non-Copenhageners much more understandable, almost to Norwegian levels, but Victor asked about the national standard, which I took to be the Copenhagen variety.

    Second, I specified that I was coming from Swedish. Danes find Swedish much easier to understand than the converse. And I learned my Swedish in Västerbotten; someone from Skåne would obviously give a very different answer.

    I've seen (and done) the each-speaks-their-own-language thing. Invariably, I've ended up asking Danes to translate (or speaking English), while I could get by with the Norwegians.

    As to ES-IT, again, caveats. My native dialect has a very strong Italian influence, with intonation patterns that are more closer to the Rome than the Madrid standard. Perhaps a Mexican or a Castilian will give you a different answer, but Argentines will usually get by in Italy without more than the occasional inconvenience.

    ES-CA is based in my own personal experience of moving to Catalonia with no prior exposure, and finding Catalan (well, the central and western varieties, but again we're talking about the standard here) trivially intelligible. (As zoetrope mentions above, though, many Spaniard will deny this intelligibility for sociolinguistic reasons.)

    This is obviously related to the ongoing Spanish influence on Catalan; the literary and elevated registers, which are less affected, are harder to follow, but standard everyday Catalan has an almost Spanish-like syntax. Historically speaking, Catalan and Spanish have more ‘Abstand’ than Danish and Swedish, but its ‘Ausbau’ character is more problematic, as virtually all L1 Catalan speakers are also fluent in Spanish.

    ES_MX-ES_AR is a dialectal difference, no larger than that between, say, Canadian and Australian English.

  133. JS said,

    May 30, 2014 @ 5:59 am

    It's of course incorrect to claim as Kellen Parker does that we could "put… a Hakka speaker in Seoul today and in a week they'd be fine," but it's understandable to me that recognition of the enormity of the SK component of Korean would inspire an observer to such exaggeration. We might more properly look at the situation (whether coming from [esp. southern] Sinitic to Korean or vice-versa) as providing a massive (massive!) aid in the acquisition of new vocabulary, the massiveness of which is to me not even scratched at by the sorts of analogies to Kellen's claim offered above (Uzbek to Russian; English to Russian; English plus French to Persian). The suggestion that 60% of the Korean lexicon is SK (Ho-Min Sohn, The Korean Language) seems typical; Korean leaners of Mandarin commonly report (with the same surely true in reverse) that their task gets "easier and easier" over time, reflecting the increasing ubiquity of cognate vocabulary as register rises above the colloquial.

    But again, this is decidedly not the same thing as mutual intelligibility. Which between Hakka or other Sinitic and Korean I'd have to assume is approximately zero. Equal, incidentally, to my assessment of my own (L1 American English) comprehension of spoken German, Dutch, etc., despite suggestions here that I ought to understand 1/20th or more of the information relayed (or is it 1/20th of the words said?) in these and some other languages…

  134. Brian Spooner said,

    May 30, 2014 @ 6:11 am

    I would say that there is a significant amount of mutual intelligibility among all Turkic languages, from Istanbul to Urumqi. There is a certain amount of mutual intelligibility among all the languages that are vernacular developments over the last few centuries in the areas that used Persian as the language of administration until recently (i.e. Ottoman, Safavid, Mughal, Uzbek empires). Similarly among Latinate languages. Interestingly there is little or no mutual intelligibility between any two Iranian
    languages I can think of. But after reading through the comments
    posted above, I would say that intelligibility always depends on a
    number of factors, eg: social class and education, accents, are the speakers interested in understanding each other? and especially whether we are talking about spoken or written language.

  135. aut said,

    May 30, 2014 @ 6:50 am

    ad 60. Vienna – Berlin:

    Both dialects of German, I'd say, although Viennese (as a Bavarian dialect with lots of influence of Slavic languages, Yiddish, etc.) would seem to diverge much more widely from the High German Standard (although it is getting increasingly closer to the Standard).
    Mutual intelligibility: in general very high (although some difficulties concerning pronunciation and lexis would probably arise, arguably more so for the Berliner).

  136. Ben Hemmens said,

    May 30, 2014 @ 6:58 am

    intelligibility of spoken Highland Scots 0 for native American English speakers

    Generally speaking, people from the Highlands should be more intelligible to foreigners than people from the Lowlands. Supposedly because the Highlanders only recently switched from Gaelic to modern English, whereas Scots in the Lowlands has had a separate existence from English English all the way back to early Middle English. It's the rather distinct varieties of the cities that an outsider is likely to have most trouble with.

  137. Jongseong Park said,

    May 30, 2014 @ 7:16 am

    @JS, I agree with your point about the enormity of the Sino-Korean component in Korean, especially in more elevated or specialized registers. This sort of thing is very difficult to quantify—one could reel off a couple of sentences where Sino-Korean is hardly used, or at the other end of the spectrum one could produce sentences where nearly all the content words are Sino-Korean. The closest parallel I can think of is the amount of French and Latinate vocabulary in English.

    But again, this is a separate discussion from mutual intelligibility. Having a large body of recognizable vocabulary is a necessary condition for mutual intelligibility, but it is nowhere near sufficient.

  138. Darkwhite said,

    May 30, 2014 @ 8:25 am

    Alon Lischinsky: Thanks for the follow up. I'm still somewhat baffled by the SWE-DAN number, but then again, my parents still refuse to believe me when I try to explain that I find SWE and DAN strenuous to decode. I'm now a bit wiser, and a bit more sorry to never have learned Spanish.

  139. Warren Maguire said,

    May 30, 2014 @ 8:47 am

    Vanya, Ben Hemmens: There's not really any such things as 'Highland Scots'. Scots varieties are spoken in the Scottish Lowlands and Northern Isles. Highland English is a recently developed variety, as Ben says, not very different from Standard English in Scotland, but with a noticeable Gaelic substrate accent, especially in the Hebrides.

    And although you will hear rather distinctive and difficult-to-understand forms of speech in the cities, much of this is to do with divergences in accent with some morphological differences. For really divergent Scots (dialect/language/whatever – neither of these terms can be objectively defined), rural places such as the Shetlands, Buckie or the Scottish Borders are the places to go.

  140. Doyle said,

    May 30, 2014 @ 9:25 am

    Rubrick said:
    "I think an interesting pair (apologies if it's already among the many comments) is American English / Caribbean English."

    How about Any Other English / Caribbean English? I (American) had a Jamaican neighbor where I used to live. I could understand him OK when he spoke to me, but he was still more difficult to understand than any of my other neighbors. When he spoke to his Jamaican friend, however, I could understand almost nothing. It was almost like when I was around my El Salvadorean friend and his mom. They must have been speaking patois, but I don't think I knew that at the time.

    What about:
    Other Types of English / Black American English (in particular, southern varieties of it)?
    Other Types of English / Gullah or Gullah-influenced English?

  141. Victor Mair said,

    May 30, 2014 @ 9:25 am

    I have many Chinese professor friends who have gone to stay in Korea to do research or teach for periods ranging from three months up to two years. In addition to Mandarin of various stripes, several of them also knew at least one other Sinitic language (Cantonese, Shanghainese, Hangzhouese, etc.). Not one of them came back to China speaking any Korean whatsoever, not even a sentence or two. At most they learned a couple of Korean words, usually badly pronounced.

  142. Alon Lischinsky said,

    May 30, 2014 @ 10:47 am

    @Darkwhite: according to the best-known study of such matters (Delsing & Lundin Åkesson 2005), Norwegians have the easiest time understanding other Continental Scandinavian languages, while Swedes have the hardest. That might account for the difference in perception (well, that and the fact that I acquired Swedish pretty late, and have nothing resembling native-like competence).

  143. Oscar Jenz said,

    May 30, 2014 @ 11:21 am

    Native speaker of American English, Finnish and a legacy speaker of Turkish here.

    Turkish — Persian
    Individual words here or there, mainly via Persian loans into Turkish and shared loans from Arabic. Beyond that nothing really, 5-10%.

    Finnish — Estonian
    Both are in the ‘Balto-Finnic’ sub-branch of Finnic, and the written languages seem very similar (some shared cases, a lot of easily recognisable shared roots etc.). However, intelligibility is only around 10%, as the pronunciation is fairly different and semantic drift has created a large amount of false friends that really impedes intelligibility for Finnish-speakers. A lot of older Estonians used to watch Finnish TV and listen to Finnish radio during the Soviet era, meaning that they tend to have a decent passive understanding of Finnish. Finns, on the other hand, have a minimal understanding of Estonian, ditto with younger Estonians vis-à-vis Finnish.

    Finnish — Hungarian
    Related via Uralic, but it takes a trained eye to recognise the few cognates that occur in everyday speech. Mutual intelligibility is pretty much 0% in an everyday scenario, except for shared English/French loans from 19th century onwards.

    Finnish — Swedish
    Different languages families (Uralic-IE), though there has been extensive contact between the dialect of Swedish spoken in Finland and Finnish since around 300AD. Finnish has a lot of loanwords from Swedish, meaning that some words can be identified. With no formal or social background, however, intelligibility is minimal, and Finns understand maybe 5% of Swedish lexis. This is because Finns are very aware of the influence of Swedish on Finnish, and even with no prior contact may look for shared words. Swedes from Sweden tend to have no idea that Swedish ever influenced Finnish, and have a roughly 0% understanding of Finnish in a real-life situation outside shared English and French loans. Metalinguistic awareness really accounts for the difference.

    Outside the ideal 'no-contact' scenario, Finns study Swedish for at least 3 years in school, and due to widespread bilingualism in the coastal south and in official documents most Finns have a decent passive understanding of Swedish. Swedes don't return the favour.

    Uyghur — Turkish
    A lot of everyday vocabulary is shared and recognizably related, as are shared Persian/Arabic loans. Written intelligibility is relatively high, 15%-25% depending on how educated the reader is: spoken it is somewhat lower.

    Mongolian — Turkish
    5-10% at best from shared loans/possible cognates, in an everyday situation 0%.

  144. marie-lucie said,

    May 30, 2014 @ 12:00 pm

    (Sorry, I did not write your names right away and now can't find them)

    I was homesick in Finland and hung out at a Chinese (I think Mandarin) restaurant because I found the staff conversations homelike and familiar, even though I speak no Mandarin–it sounded like Seattle!

    I had a similar experience when I when I went from Vancouver (BC, North of Seattle), where I had been several years, to Whitehorse (Yukon territory) some years ago. At first I felt quite out of place in Whitehorse, until I went to a restaurant where I felt very much at home. I wondered why I was feeling that way, and I realized that it was because the cooks were speaking Chinese (probably Cantonese), a language that I did not know but that I had heard many times in Vancouver's Chinese restaurants, although without paying aparticular attention to it.

    I'm an English speaker (American) with French as a second language. I can understand Italians from Turin much better than those from Naples – don't know if they're actually using the same words, but one sounds much more like French.

    What you heard in Turin was probably a form of the Piemontese dialect, which is indeed closer to French.

    My brother-in-law, whose first language was Quebec French (Quebequois?) said that they closer he got to Normandy, the easier it was for him to be understood.

    Indeed the speech of old rural Normandy sounds closest to old-fashioned Québécois speech. (I say "old" because in both places the schools are teaching varieties closer to "metropolitan" French, which is also spread by the mass media). The town of Tourouvre in Southern Normandy was a major centre of emigration to Québec in the 17th century, and now has a museum and research centre about it, with links to Canada. Many Québécois take advantage of organized tours to the region, where they encounter many people with the same last names as theirs.

  145. J. M. Unger said,

    May 30, 2014 @ 12:56 pm

    This is a bit off-topic, but harks back to an earlier, related thread.

    "Mutual intelligibility" is a fundamentally synchronic concept. Common origin (genetic relationship), on the other hand, is a diachronic concept. It seems perfectly all right to me to use the distinction between dialect and language as a code for indicating the epistemic status of the relationship between language X and speech variety Y. To say that Y is a dialect of X just means that the source of Y does not antedate proto-X and that no one demands rigorous proof of that fact. To say Y and X are both languages is thus not the same as saying that Y and X have become mutually unintelligible: it means that any who claims that Y and X have a common origin is obligated to prove it. Y can be a dialect of X in this diachronic sense and yet be mutually unintelligible to speakers of X.

    Apparently, this use of the word 'dialect' is too fastidious for some people. I was previously criticized for writing that Okinawan is a dialect of Japanese. Yet few flinch when someone notes the extent to which Danish and Swedish speakers can understand each other.

    Many Chinese deny that Cantonese is a language. I don't think the status of Cantonese as a language is just a matter of its being (almost) unintelligible to speakers of, say, Mandarin. The truth is that, apart from Chinese historiography, which is only circumstantial evidence from the standpoint of linguistics, we lacked valid grounds for saying that Cantonese and Mandarin had a common origin before comparative linguists provided them. This may seem curmudgeonly and contrary to common sense, but think of Vietnamese before Haudricourt's work: the whole scholarly world accepted that it was Sinitic. Now we know it's really Austro-Asiatic.

    If we let "obvious" cases dictate how we do typology, we set ourselves up to make false assumptions when we tackle difficult cases, which are, after all, the urgent ones in view of language death.

  146. Victor Mair said,

    May 30, 2014 @ 1:15 pm

    From Peter B. Golden:

    My wife, born in Shanghai, came to Hong Kong at about the age of 10. As she often remarked, during her first days in school in Hong Kong (in a Chinese school) she could not understand one word. The principal’s daughter, who knew some Mandarin, was assigned to sit next to her and translate. Although my wife had not formally studied Mandarin (in her Shanghai school Shanghainese was used), she knew Mandarin. Within a year her Cantonese was fluent and she spoke without a Shanghai accent. Her older brother and sister learned Cantonese, speak it fluently but with a pronounced Shanghai accent. Her younger sister and brother, like her, speak Cantonese without an accent.

    On Russian-Ukrainian-Belarusian: here my own personal history plays a role. I spoke “Russian” with my grandparents, who were from the region around the Belarusian part of the Dniepr River. Their “Russian” was actually a mix of Russian, Belarusian, and Ukrainian (in particular true of my grandfather’s speech). This is typical of what they term Surzhyk in Ukrainian (Russians also use the term) or Trasianka in Belarusian, a mixed language. My grandfather, from a family of rural artisans (rope-makers, sawmill workers, rafters – he did all of these things as well as working for the railroad painting and repairing RR stations), born in a place that is today described in the Russian Wikipedia as “an inhabited point” (not even rating the designation of “village,” in local parlance a vioska, much less a “small town, mestechko) was orphaned at 6, had no education and worked in various parts of Belarus’ and Ukraine (hence more Ukrainianisms in his speech). He also worked with groups of bosiaki (the “barefoot ones,” migrating bands of work crews) mainly from Kaluga Province in which southern dialects of Russian are spoken. That influenced his speech as well. He did not distinguish between Russian, Belarusian and Ukrainian. For him they were one language spoken in different ways in different places. He mixed vocabulary, grammatical forms and pronunciation from all three, even in one and the same sentence. At first (after I had become acquainted with literary Russian through reading), I thought it was just his manner of speaking based on the different areas in which he lived, lack of any formal education, etc. Only later, watching a Dutch documentary on Chernobyl’ in which local people were speaking, I was delighted to hear folks who sounded just like my grandfather. My guess, then, is that it is a local dialect extending from eastern Belarus’ to northern Ukraine. My grandmother (a farmer’s daughter) had three years of schooling at the local gorodskoe uchilishche, but was a serious reader who became passionately devoted to classical Russian literature. Her speech was “correct” Russian with only occasional Belarusianisms creeping in, usually in pronunciation. Her father, in addition to his farming, was also a “country musician” who was only too happy to play at weddings etc. His closest friend (whom they took in as he had no family), was a former serf whose master had recognized his natural musical talent and sent him to St. Petersburg for a musical education. I note this because Vasilii Vasilievich (my grandmother whenever she mentioned him always respectfully used his name and patronymic, some other, more distant, family members called him Diadia Vasia “Uncle Vasia”) brought, as a result of his “education” in St. Petersburg, a touch of the high Russian of the capital, which may have also impacted my grandmother’s speech.

    The comments in the blog are very interesting, but the best ones involve a personal note detailing contacts and the like. The “backstory” is usually most important in sorting out this kind of question of mutual intelligibility.

    With languages so much depends on personal experiences. In addition to my family linguistic background, I grew up in a neighborhood in NYC (Manhattan) in which many adults of parental age and virtually all adults of grandparental age spoke English with an accent of one sort or another. Irish brogues and Spanish predominated. Early on, I was sensitized to languages and – as you know – remain so. As a kid, I was suspicious of white-haired people who spoke English without an accent (I encountered a few – outside of characters in movies) – it seemed somehow “off” to me, not quite right.

  147. Ellen K. said,

    May 30, 2014 @ 1:18 pm

    After reading some of the comments, I'm thinking there's two issues, with regards to mutual intelligibility. There's if two speakers can understand each other in a conversation with each other. And whether a speaker of one language variety can understand an overheard conversation (under average listening conditions) in the other language variety. With related languages, the issue of different registers comes into play, though, particularly when talking about speakers of non-standard varieties.

  148. julie lee said,

    May 30, 2014 @ 1:58 pm

    @ Peter Golden

    I've enjoyed this post and conversation. Some of Peter Golden's experiences recall my own.
    Re. a Chinese person picking up other Chinese topolects, I've heard or seen adults and children pick up a local Chinese topolect quite different from theirs (e.g. someone speaking the Nanking topolect picking up the totally unintelligible Shanghai topolect) very quickly (six months to a year) when they were flung into the new environment. However, there are differences in aptitudes. My dad could never speak Cantonese (in South China) whereas my mom who cam from the same Central China county picked it up quickly as a adult.

    Also, we children in the family all learned Mandarin as a first language. But when my brother and I went to Chinese schools, we were ashamed of our speech, which was quite different from Standard Mandarin. We both worked hard to correct our speech. I always thought my mom and dad must have had poor ears for speech because their Mandarin was so different from what hey spoke in school. For teng "pain" they said "tin" (like in English). For leng "cold", they said "nin". For _he_ "drink" they said "ho", and so on.
    So for decades I spoke an artificial Standard Mandarin that I copied from schoolmates. I felt like a Cockney trying to speak King's English.
    Then a few years ago I was at Stanford Graduate Library's circulation desk and found that the librarian spoke just like my mom. She had just come from Hankow, China. Then it suddenly dawned on me, and I listened to her speech with delight. For I realized that there had been nothing wrong with my mom's ear or her speech. She was simply speaking standard Hankow Mandarin, not so-called Beijing Mandarin. Some Hunanese also speak like my mom, except their "tune" or "melody" is different, and they say Funan instead of Hunan. All their initial h- 's are f-'s, so that I often don't understand what they are saying, but can understand if they speak slowly and give me time to convert words like Funan to Hunan, f- 's to h-'s. Now I am going back to my original Hankow speech—with the shame gone. I'm retired so don't have to impress anyone anymore.

  149. Victor Mair said,

    May 30, 2014 @ 2:19 pm

    From a colleague who wishes to remain anonymous:

    I always find the “Swedish-Danish-Norwegian” discussion interesting, since it fits into all kinds of idealized tropes about the place. Indeed, as a Norwegian friend – a highly educated UN diplomat – told me and some other non-Scandinavian friends a few weeks ago: “we really don’t understand each other at all.”

  150. mira said,

    May 30, 2014 @ 2:37 pm

    I'm a fluent but non-native Czech speaker. I agree with Fil above who said that the supposed inability of younger Czechs to understand Slovak is overstated. I find that usually what people mean when they say they don't speak Slovak is that they're unfamiliar with certain funny-sounding Slovak words that aren't used in Czech. The fact that not knowing the Slovak word for "icicle" or "camel" is enough to qualify you as "not understanding Slovak" just confirms how similar the two languages are. I find it incredibly hard to believe that a native speaker of Czech would have a hard time understanding someone speaking standard Slovak. There are a number of American Idol-type television shows on right now that feature both Czech and Slovak contestants and judges speaking their native languages.

    I have the feeling that Slovaks have even less of a problem understanding Czech than vice versa, because they have more exposure to the other country's media, being smaller and all. In fact, Slovak prescriptivists are always very worried about "Czechisms" creeping into the language. I found a book called "How not to sin against the Slovak language" in a bookstore in Košice which was mostly about avoiding nasty Czechisms like "hranolky" (french fries) instead of "hranolčeky".

    I should admit that my Czechoslovakism might be colored by my having learnt Czech & spent most of my time in eastern Moravia, where family and other connections to Slovakia might be stronger than in other parts of the country (also where the purest and correctest Czech is spoken, according to the people who live there).

  151. Levantine said,

    May 30, 2014 @ 3:03 pm

    Michael Dunn, ordinary Turkish-speakers before Atatürk would not have been any more able than their present-day counterparts to understand Persian, even if they used a greater number of Persian-derived words. The differences between the two languages are too great to have been overcome by lexical overlaps.

  152. Catanea said,

    May 30, 2014 @ 4:26 pm

    I think no-one has addressed Catalan – Portuguese?
    When my Catalan was only a couple of years old, and I'd lost my French, I found myself grounded in Lisbon on a flight suspended because of weather.
    In my new-ish Catalan I had a fascinating conversation with my Portuguese taxi-driver all the way from the airport to the hotel. I thought it was very cool. I spoke Catalan, he spoke Portuguese (with wonderful zhe/dze sounds, and we talked about his family and job, and where I'd come from (Seattle) and I got a new idea about the possibilities of communication going on in separate languages without "translation".
    I'd call that 85%, maybe.
    And I would expect if he'd had conversations with Catalan-speakers in the past, they'd have had Spanish as well and see all relevant posts above about the mutual intelligibility of Spanish and Portuguese or vice versa. I had about three stock phrases of Castilian, none of which came up.

  153. Alexander Vovin said,

    May 30, 2014 @ 5:06 pm

    1. Russian — Ukrainian


    2. Hindi — Urdu

    It is almost the same language on thew spoken level, but, of course being more and more divergent for the last 70 yrs. Written conprehession is not possible.

    3. Hebrew — Arabic


    4. Japanese — Mandarin


    5. Osaka — Tokyo

    A very tricky question. Basically yes, but there are catches. It is much easier for an Osaka speaker to understand a Tokyo speaker, because the standard language is to some extent is based on Yamanote Tokyo speech with some heavy intrusions of Western Japanese. And understanding southern Osaka for Tokyo speakers might be sometimes difficult.

    6. Turkish — Persian


    7. Persian — Arabic


    8. Egyptian Arabic — Lebanese / Syrian / Jordanian Arabic

    9. Finnish — Estonian

    Possible, but sometimes with hilarious results.

    10. Finnish — Hungarian


    11. Finnish — Swedish


    12. Bengali — Hindi

    13. Nepali — Bengali

    14. Hindi — Nepali

    15. Marathi — Hindi

    16. Serbian — Croatian

    This is the same language exceopt for slang and some peculiar/idiomatic expressions.

    17. Czech — Russian

    Only on the written level for educated speakers.

    18. Czech — Slovakian


    19. Tamil — Telugu

    20. Malay — Indonesian

    Minor differences due to English lons in Malay and Dutch in Indonesian.

    21. Pashto — Persian

    No, different groups of Iranian.

    22. French — Italian

    Written for educated speakers, yes, spoken, I doubt

    23. French — Spanish

    Same as above.

    24. French — Portuguese

    Same as above

    25. Spanish — Portuguese

    Depending on the speakers' native dialects.

    26. Spanish — Italian

    Written for educated speakers, yes, spoken, I doubt

    27. Spanish — Catalan

    Written for educated speakers, yes, spoken, I doubt

    28. Madrid — Mexico


    29. Mexico — Argentina


    30. Mexico — Cuba


    31. Madrid — Puerto Rico


    32. Korean — Japanese

    No way.

    33. Korean — Mandarin

    No way.

    34. Vietnamese — Lao

    No way.

    35. Vietnamese — Cantonese

    No way.

    36. Toishan — Hong Kong

    37. Gujarati — Bengali

    38. Gujarati — Hindi

    39. Swahili — Arabic

    No way.

    40. Arabic — Berber

    No way.

    41. Somali — Amharic

    NO WAY! Different families of Afroasiatic!!!

    42. Tigrinya — Amharic

    43. Setswana — Swahili

    44. Kinyarwanda — Kirundi

    45. Malagasy — Seychellois Creole

    46. Seychellois Creole — Mauritian Creole

    47. Afrikaans — Dutch

    48. Afrikaans — English


    49. Sesotho — Xhosa

    50. Swedish — Danish


    51. Swedish — Norwegian


    52. Swedish — Lithuanian


    53. Swedish — French


    56. Swedish — Dutch


    57. Dutch — Flemish

    Mostly yes, but depending on a Flemish dialect, if an uneducated Flemish sopeaker is involved.

    57. Danish — Norwegian

    On the written level, yes, on the spoken — depends on which two of Norwegian variaeties are used.

    58. Dutch — Norwegian


    59. Dutch — German

    Ah, well, people in Aachen will understand dutch, but not those in Berlin.

    60. Vienna — Berlin

    If both speak Hochdeutsch Hannover, yes.

    61. English — Norwegian


    62. English — German


    63. English — Dutch


    64. English — Swedish

    65. English — French


    66. English — Italian


    67. English — Spanish


    68. English — Portuguese


    69. Mandarin — Sichuan (rural)

    70. Mandarin — Gansu (rural)

    71. Mandarin — Ürümchi (local Mandarin)

    72. Uyghur — Turkish

    This a one-way deal. Ughurs are likely to comprehend AnatolianTurkish, but not the other way around.

    73. Kazakh — Uyghur

    Limited understanding

    74. Kazakh — Kyrgyz

    Limited understanding

    75. Mandarin — Shanghai


    75. Mandarin — Cantonese


    76. Mandarin — Hokkien / Taiwanese


    77. Albanian — Armenian


    78. Latvian — Lithuanian


    79. Estonian — Hungarian


    80. Thai — Zhuang


    81. Thai — Khmer


    82. Tibetan — Mandarin


    83. Lhasa — Amdo

    Written yes, spoken most likely no.

    84. Sherpa — Bhutanese

    85. Rai — Limbu

    86. Gurkha — Newar

    87. Tibetan — Burmese


    88. Ubykh — Georgian


    89. Azerbaijani — Chechen


    90. Ossetian — Georgian


    91. Mongolian — Turkish


    92. Mongolian — Korean


    93. Quechua — Guarani


    94. Belarusian — Russian

    Another tricky question. Belorussian dialect (unlike the literary and standard Belorussian) represent a continuum gradually changing from Russian to Polish. So, the guys on the eastern extreme about 100%, on the western about 60%.

    95. Belarusian — Bulgarian

    Depends on what kind of Belorussian is meant, but iverall about 70-80%.

    96. Romanian — Italian

    I guess none, but I have no knowkedge of Romanian.

    97. Basque — Spanish


    98. Basque — Catalan


    99. Romanian — Romani

    No, Romani is a Punjabi dialect.

    100. Chippewa — Muscogee (Creek) / Miccosukee (mutually unintelligible Seminole languages)


    101. Wenzhouese — Mandarin (late addition!)


  154. Harbans Mukhia said,

    May 30, 2014 @ 11:31 pm

    Hindi-Urdu have the same grammar structure and the same gender identity of words. The scripts, however, belong to different families: Urdu is written in an adaptation of Arabic-Persian script, Hindi in Devnagari, the script of Sanskrit. Reader of one cannot read the other unless one learns the script. The two have a great deal of common vocabulary and are intelligible to speakers of each except when deliberate attempt is made to bring in heavy Sanskrit words in Hindi and Arabic-Persian words in Urdu.
    Hindi-Bengali both use Devanagari script, though modified in each. Both belong to the Sanskrit family, though Hindi has a much higher proportion of Persian-Urdu words. Accent and pronunciations are different, but easily accessible to both. It happens that Hindi is my mother tongue and Bengali my mother-in-law tongue and Urdu and Persian I learnt for academic pursuits.

  155. Alexandru Pănoiu said,

    May 31, 2014 @ 2:16 am

    @ Pflaumbaum:

    I agree that it is harder for Italians to understand Romanian than it is for Romanians to understand Italian. I think that there are three reasons. One reason is that some Romanian sounds do not exist in Italian (the vowels ă and â/î and, I think even more importantly, the softened consonants written -bi, -ci, -gi, -li, -mi, -ni, -pi etc.); another reason is that very many Romanian words end in consonants, making then sound "alien" to Italians (the first two rules when learning to fake Italian are "drop all diacritics and pronounce all letters as written" and "just add -o if the word still ends in a consonant"); and the third is indeed the Slavic influence (which appears much less important today than it was 200 years ago, before the massive influx of French and Italian words in the second half of the 19th century).

  156. John Walden said,

    May 31, 2014 @ 3:57 am

    It's difficult to imagine any way of testing some of these assertions. The test conditions would surely involve sitting down a speaker of each language, neither of them having had any previous contact with the other. No back stories (an excellent term!) of previous attempts to learn cognate languages and any number of other caveats. For example, I was able to follow groups of his Dutch friends talking to my brother in Dutch, but then I was an English speaker who had learnt German and who knows how much they were simplifying their Dutch for my brother's benefit? Too many variables to say that English and Dutch are mutually intelligible on the basis of one anecdote.

    By the same token on my first day of living in Spain I was able to follow some conversations from my knowledge of French and Latin. It proves nothing.

    One such assertion is

    "27. Spanish — Catalan

    Written for educated speakers, yes, spoken, I doubt"

    That might be true but to really test it you'd have to find a Catalan speaker who'd never heard Spanish before. There'd be Spanish speakers from outside Spain of course but I don't know if there's any remote village in Catalonia where Spanish TV has never been heard.The older inhabitants might well have done military service in Castilian-speaking Spain and so on.

    Then there's: talk about what? What would the test subject matter be? I would say that a group of scientists speaking their own Romance languages could easily talk about their shared speciality of, say, nuclear physics.

    As could soccer players: I never see the need for subtitling Italian and Catalan speakers on Spanish TV. Though, as I have said, I may have the advantage over a native speaker.

    Whether or not Dutch, German and Scandinavian-speaking farmers or fishermen could chat together about their specialities I don't know. Probably not.

    It's all too close to categorically call. More anecdote: I moved from a solid Spanish-speaking established neighbourhood to a more exotic transient cosmopolitan one and my Spanish went from being understood by next to nobody to overnight being understandable in shops and on the street, presumably because the Spanish speakers in my new barrio had opened their ears to mangled Spanish before. I'd bet that someone from the latter group would say that Portuguese was more intelligible, to them.

    The original condition of "Assuming no prior, formal study of or contact with the opposite language in a given pair (i.e., one is coming at these languages completely cold)" doesn't seem that easy to create in some of the examples. Which makes some of the answers rather speculative.

  157. Alon Lischinsky said,

    May 31, 2014 @ 4:33 am

    @John Walden:

    you'd have to find a Catalan speaker who'd never heard Spanish before

    In principle you could: there are heritage speakers in the Roussillon (France) or in Alghero (Italy), and even some revitalisation movements in the former. Of course, they would have their own French/Italian to draw on as well.

    I've met some rural folk in Catalonia who had minimal active competence in Spanish, but none who lacked passive competence.

  158. Victor Mair said,

    May 31, 2014 @ 6:13 am

    @John Walden

    "It's difficult to imagine any way of testing some of these assertions."

    But surely not ALL of them. And what "assertions" are you talking about anyway? The claims of commenters who have described their own personal experience? For example, when I say that I don't understand Korean, even though I know several Modern Standard Mandarin very well and bits of other Sinitic languages plus lots of Japanese (about one third spoken, two thirds or more written), am I being "speculative"? No, I'm stating a fact: I do not understand Korean.

    Similarly, when I state that I only understand about 5% of Hindi and less than that of Urdu, still less of Bengali, even though I was once fluent in Nepali, I'm not making a speculative assertion, although I am being impressionistic and approximative. I understand some Hindi, Urdu, and Bengali, but only a very little.

  159. John Walden said,

    May 31, 2014 @ 7:46 am

    @ Victor Mair, contritely.

    I'm truly sorry that my comment above does come across as overly belligerent.

    The speculations and assertions (obviously not the best-chosen of words and not helped by those scare quotes of yours) are being made by thoughtful and knowledgeable commentators giving their informed opinions about two languages they either speak or know enough about.

    I must add that I find the backstories very intriguing and I was only trying to make the point that no two are alike, not to deny their worth.

    And I only wanted to point out that the possibility of testing some of the pairs is remote. Not that the informed speculation (that word again) is pointless.

    I'm equally intrigued by the Ultimate Warrior questions like Ancient vs Modern Greek: you could find some very paper-fluent Ancient Greek scholars but for them to have had no contact with Modern and know how the Ancient version was really pronounced seems a stretch. But it's no less an interesting question.

    I hope the opinions and testimonies keep on coming. The original proposal

    'My main purpose in giving this quiz is to encourage people to think about the nature and reality of degrees of intelligibility between forms of speech.'

    seems to be working.

  160. Colin said,

    May 31, 2014 @ 10:21 am

    When I moved to Brussels, I was a native English speaker who also spoke French and German, but had no prior exposure to Dutch. I found I could pick up odd words from overheard conversations and quite a lot more from announcements and standard short phrases. Newspapers and the like could be more or less understood with some effort. To my ear/eye, Dutch is closer in vocab and word order to German than it is to English, so I'm surprised that some German speakers find it so incomprehensible. Flemings generally have accents that are quite recognisably different from a Hollandic accent, even when the speaker is speaking English. (By contrast, I never could figure out the distinctive features of a Belgian French accent that set them apart from other French speakers. Most of the differences noted on Wikipedia are where Belgians are more likely to observe 'official' French pronunciation rules than the French themselves, e.g. in preserving the distinctions brin-brun or j'irai-j'irais.)

    As for the dialects/languages thing:
    English, Dutch and German are all separate languages in the same group.
    'Flemish' could mean the Dutch spoken in Flanders proper (essentially confined to the provinces of West and East Flanders), which is a group of related dialects, or it could mean all forms of Dutch spoken in Belgium (associated with 'Flanders' as a political term for Dutch-speaking Belgium), in which case it doesn't have much linguistic significance. European Dutch is all one dialect continuum, and the dialect groups don't align with national borders. Differences in the written language are very minor.
    Belgian French is a dialect of French with some distinctive vocab, but close enough that mutual intelligibility is nearly 100%.

  161. Victor Mair said,

    May 31, 2014 @ 12:17 pm

    @John Walden

    Thank you very much for your kind reply.

    When I put a couple of words inside of quotation marks, I honestly did not mean for them to be scare quotes, but simply to indicate that I was quoting words that you had used.

    I wish to take advantage of this opportunity to state unequivocally that I am deeply grateful to each and every commenter on this thread who has shared his / her experience, views, and knowledge. This blog and all of the comments to it are an excellent example of why it is such a privilege to be a part of Language Log.

    If I ever need to tell anyone about a good place to go to learn about mutual intelligibility, this is the first place that I'll recommend.

  162. Piyush said,

    May 31, 2014 @ 1:50 pm

    @Harbans Mukhia,

    The scripts, however, belong to different families: Urdu is written in an adaptation of Arabic-Persian script, Hindi in Devnagari, the script of Sanskrit.

    I am not sure it is entirely accurate to describe Devanagari as the "script of Sanskrit". Sanskrit never had a canonical script of its own, and was written (and is still written) using different scripts in different parts of India (e.g., Grantham in modern day Tamil Nadu). Devanagari, on the other hand, is much younger than Sanskrit (having developed mostly in the last 100 years).

    Hindi-Bengali both use Devanagari script, though modified in each.

    The Bengali script is actually quite different from Devanagari, enough that someone familiar with Devanagari but not with Bengali would not be able to recognize more than a few letters.

  163. Pflaumbaum said,

    May 31, 2014 @ 4:51 pm

    @ Alexandru Pănoiu –

    I'm surprised to see the palatalised final consonants make that list – I would have thought that someone speaking a language lacking a palatalisation contrast would just fail to perceive them as any different from the unmarked consonants.

    Certainly in my experience as a native English speaker learning Romanian, I perceived these segments as a short voiceless vowel following an ordinary consonant. It was only after several years that I read more about it and started to 'hear' it as palatalisation.

    But then since the consonant inventory is so very similar in Italian, maybe you're right and variations would stand out more.

  164. JIm Manheim said,

    May 31, 2014 @ 5:53 pm

    I'm a native American English speaker who speaks decent but not fluent Indonesian. My impressions: Among adult speakers of Indonesian, Malay is pretty much fully intelligible. But there are occasional hilarious misunderstandings. The sounds of Malay are a bit different, though. There's a popular Malaysian kids' animation show called Upin & Ipin, and it's sometimes subtitled for children in Indonesia. If I try to talk to a Malaysian, we eventually communicate, but it's sort of as if someone with American English as a second language tried to talk to an urban Scot.

  165. S Frankel said,

    May 31, 2014 @ 8:54 pm

    If I can add a pair:
    Biblical Hebrew- Modern (Israeli) Hebrew:
    apparently minimal mutual comprehensibility.

    Biblical Hebrew is not one monolithic language, of course, but we'll ignore that complication. In Israel, all Jews have exposure to the language of the bible in school, but there have been tests done with non-Jewish speakers of modern Hebrew (mostly Israeli Arabs who have native ability in Hebrew), and they found spoken recitations of the text worse than incomprehensible: they would sometime latch onto a word that resembled a modern word (false friends of some kind) and construct a meaning around that.

    The modern Israeli pronunciation of Hebrew was used. If a historically accurate pronunciation had been used (which is approximated by Yemenite Torah cantillation), no doubt the texts would have been even more mysterious.

  166. Bill W said,

    May 31, 2014 @ 9:08 pm

    Marie-Lucie: "Many Québécois take advantage of organized tours to the region, where they encounter many people with the same last names as theirs."

    Both of them? Both Tremblay and Ouelette? Are there any other last names in Québec?

  167. marie-lucie said,

    May 31, 2014 @ 11:00 pm

    Tremblay is very common in Normandy, but Ouelle(te) is more likely to be Acadian.

  168. Pflaumbaum said,

    June 1, 2014 @ 5:41 am

    @ S Frankel –

    I'd expect that if a historically accurate pronunciation were used it would actually be easier for Arabic speakers to understand. (I realise this doesn't affect your main point about intelligibility of Hebrew with Ivrit.) The partial loss of aleph and ayin in particular has surely made Ivrit and Arabic less mutually intelligible.

  169. S Frankel said,

    June 1, 2014 @ 7:23 am

    @Pflaumbaum – Jews from Arabic-speaking backgrounds do, sometimes, pronounce aleph and ayin,and they also differentiate between Het (dot under the H) and khuf, so non-Jews from Arabic-speaking backgrounds may already be familiar with this pronunciation, although it is stigmatized and, as a result, recessive. And there are other differences between the classicizing and modern pronunciations that would probably impede intelligibility, such as the fricativazation of dalet(h), gimel, and taf (when they don't have a dagesh).

  170. Pflaumbaum said,

    June 1, 2014 @ 8:23 am

    That's what I mean. I'd expect Arabic speakers to get more from it than non-Arabic speakers, due to Arabic's greater retention of glottal/pharyngeal consonants (and notwithstanding the begadkefat alternations). So I'm not sure in principle they make a good test case for Ivrit-Biblical Hebrew intelligibility comparisons – even if as you report they didn't understand much when it was tried in practice.

    Wouldn't it be kind of like using bilingual speakers of Icelandic and English to test the mutual intelligibility of Modern and Old English?

  171. marie-lucie said,

    June 1, 2014 @ 9:12 am

    Colin: <i.Belgians are more likely to observe 'official' French pronunciation rules than the French themselves, e.g. in preserving the distinctions brin-brun or j'irai-j'irais.)

    The "official" French pronunciation is a conservative one, such as my own (older French native, Parisian parents but raised in Normandy). Current "Parisian" French has changed a lot during my lifetime, probably because of the "melting pot" that is the Paris area. So I don't think that the Belgians are making a conscious effort to speak "official" French, instead they are maintaining their own conservative pronunciation.

  172. Shira said,

    June 1, 2014 @ 1:59 pm

    Arabic-Hebrew: as a fluent Hebrew speaker I cannot understand a word of Arabic (save the few words/phrases I learned as Arabic words/phrases, but I wouldn't have been able to associate them with a Hebrew equivalent had I simply heard them in a conversation).

  173. Martin said,

    June 1, 2014 @ 4:44 pm

    Scandinavian languages have been covered quite a bit, but here are my thoughts, as a Norwegian: Norwegians generally do understand Swedes and Danes, at least if a little effort is made. Also vice versa if we assume the Norwegian uses an educated Oslo dialect, which is probably the closest you can get to a standard. ('Standard spoken Norwegian' is not a politically popular concept…)

    Of course we have all been exposed to the other languages quite a bit, but even without any prior knowledge a Norwegian and a Swede/Dane who needed to communicate would probably be able to, speaking slowly, clearly, and adding in the odd explanation of a phrase.

    I hear Swedes and Danes have a harder time of it, though.

    The more remote Norwegian dialects can be very hard to understand for outsiders. An anecdote: Two girls I know shared a room, they had just known each other for a day or so. One is from Oslo, the other from high up in Telemark, a valley in the mountains of southern Norway. The Telemark girl asked the other, 'læs att dynni' ('shut the door', which in Oslo would be 'lukk igjen døra'). The Oslo girl had no idea what this meant, but after a few repetitions decided it was probably 'les under dyna' (read under the duvet). Remote means, roughly, anything high up in the mountain valleys, the fjords, and possibly the far North. I certainly have no problem with other dialects, but it's hard to imagine what would happen with no prior exposure.

    Swedes tend not to understand me. My dialect is south west coast, but at least part of the problem is my horribly lazy diction. Maybe if I made an effort, but English is easier.

    Germans who come to live in Norway can learn the language very quickly, I know several who – as adults – have gone from zero to pretty much fluent in half a year. Don't know about initial intelligibility though.

  174. S Frankel said,

    June 1, 2014 @ 8:17 pm

    @Pflaumbaum – Hi, the point of testing Arabic speakers on Biblical Hebrew was not that they're Arabic speakers, but that they are Arab Israelis, who have native competence in modern spoken Hebrew and, unlike the Jewish speakers of Hebrew, have no exposure to the Biblical language.

  175. John Swindle said,

    June 1, 2014 @ 8:49 pm

    In a message above, Son Ha reports Thais hiring Khmer-speaking Cambodian workers because they can understand a lot of what they're saying. How does this come about? Aren't Thai and Khmer pretty different?

  176. Colin said,

    June 1, 2014 @ 9:25 pm

    marie-lucie: Indeed, they are features that have simply been preserved over the generations, not the result of present-day Belgians deliberately following the rules. I said 'official' because these pronunciations are still considered part of the standard language in France, as opposed to features of Belgian French that are considered archaic/obsolete in France (such as 'septante' and the use of 'déjeuner' to mean 'breakfast').

    I suspect Belgian French has relatively few specific innovations because most of its speakers have only a few generations of L1 French ancestors. In Wallonia, French displaced other langues d'oïl in a process similar to what happened in France. Brussels is a more abrupt and localised case of a language shift, which mostly cut the city off linguistically from its own hinterland in the course of three or four generations.

  177. Chaaak said,

    June 2, 2014 @ 12:47 am

    36. Toishan — Hong Kong

    I am a Cantonese speaker from Hong Kong. Cantonese was my only language until 4 y.o. when I had a maid who would speak to me only in English. I didn't speak any Mandarin until I entered the uni.

    I have studied Hokkien-Minnan-Taiwanese since 18, but have never heard Toishanese before. I was told that Cantonese and Toishanese are pretty close but a friend of mine said she couldn't understand a word. So I looked it up on youtube to see how it sounds like, and I found that I could make out most of what they say. I tried the same for Hakka and it worked for me too.

    I guess I must have figured out some of the sound correspondence rules and managed to do some 'calculation' in my head. I might understand Gan (贛) as well, but I will try not to study it so that I can do another mutual intelligibility test later on.

  178. Son Ha said,

    June 2, 2014 @ 3:55 am

    Thanks for following up. I crossed checked with my Thai friend and have to correct my erroneous belief that Thais and Cambodians can understand each other for case work or house work. So mutual intelligibility = nil.

  179. Pflaumbaum said,

    June 2, 2014 @ 4:47 am

    @ S Frankel –

    I know! I'm just saying they don't seem like a very good test group, since they generally know a language other than Ivrit that's related to Biblical Hebrew. This interference could affect the results in either direction.

    There are lots of non-Jewish fluent Ivrit speakers from Asia, for instance, living in Israel who'd be unlikely to have read any Biblical Hebrew before. They'd be a more logical choice IMO.

  180. Vanya said,

    June 2, 2014 @ 10:11 am

    No one has addressed Kazakh-Kyrgyz. When I lived in Kazakhstan I was told there was 70-80% intelligibility between the two. I was never a fluent speaker of either, but my limited exposure to both leads to me to believe that is probably true. Uzbek is a different branch, but I was told Uzbek and Ughur are very similar. Turks will often tell you that they can understand all Central Asian Turkic languages but I suspect they are grossly exaggerating. Kazakh, for example, has a lot of archaic vocabulary, and Persian, Arabic and Russian loanwords, and Kazakh grammar, particularly verb tense formation, differs from Anatolian Turkish to a far greater degree than say Italian grammar differs from Spanish or Croatian grammar from Russian.

  181. Greg Morrow said,

    June 2, 2014 @ 3:58 pm

    Native Standard American English speaker. I have a terrible ear, so the intelligibility of anything non-English is very poor, and even some accents of English can be hard to work out.

    But it's a LOT easier to read a language, because cognates are a big help. Four years of high school French and a monstrously large cognate vocabulary make written French pretty easy. And knowing both makes written Spanish somewhat intelligible. I would expect that there would be similarly early-rising learning curves for other Latin and German-branch sister languages… in writing.

    I'll mention two more things: Czech and Polish toponyms and isolated language fragments look a lot "friendlier" in terms of intelligibility than Russian, i.e., I'd expect West Slavic to be easier than East Slavic.

    And I've looked at a little Japanese. Romaji is surprisingly readable, aside from the vocabulary. The basic sentence grammar is pretty easy for me (an early breakthrough was that "no" and "ni" particles could be interpreted quite like French "de" and "à" prepositions — English spreads those grammatical functions over more prepositions). Because of restrictive syllable rules, the borrowed vocabulary is more translucent than transparent.

  182. Ray Dillinger said,

    June 6, 2014 @ 10:31 pm

    I find that an extensive vocabulary in English gives me sufficient cognates that, with time and effort and a few different contexts, I can puzzle together written documents of just about any western european language. If I'm looking at one sentence on a place card, I'm usually out of luck; but if I have a bunch of different sentences, on, eg, a menu or a kids' book or something like that, I can usually cross-reference the different bits and contexts and eventually figure out what most of it says.

    That said, I didn't understand any spoken French until I studied it, and I'm still picking out the occasional word and sentence of spoken Spanish, and despite having one set of grandparents who spoke Low German and another set who spoke a frozen sixteenth-century dialect of English, I'm at best getting about 15% of spoken German.

  183. thunk said,

    June 9, 2014 @ 10:01 am

    I just asked my father about the Belarusian-Russian pair (he grew up in Slutsk, 100 km south of Minsk, speaking Russian before heading off to university in St. Petersburg). According to him, about 80% of the words would be the same, with the remainder in some cases having wildly divergent meanings. Spoken intelligibility would be lower, but Belarusian speakers would understand more Russian than vice versa.

    Ukrainian would be another matter, but I guess it's more divergent than Belarusian or Russian, siding more closely with the former. As a (patchy) Russian speaker, I'd estimate that I'd understand 30 or 40 percent of spoken Ukrainian.

  184. Jelena said,

    June 11, 2014 @ 12:29 pm

    I am actually doing my PhD on mutual intelligibility of Slavic languages, so I can give you some preliminary numbers, for instance for written language:

    Czech-Slovak: around 97%
    Croatian-Slovene: around 68%, but Slovene-Croatian: over 80%
    Slovak-Polish: over 50%
    Bulgarian-Croatian: between 30 and 40%

    This is quite unrefined, based on a broad sample of participants and the level of intelligibility was measured by a cloze test. Still, hope it helps the discussion.

  185. Victor Mair said,

    June 14, 2014 @ 3:33 pm

    From a Swedish colleague:

    I wanted to follow up on the “Scandinavian language” business, since I’m currently reading Karl-Ove Knausgard’s literary phenom, which has much on how he hates Sweden and Swedes, and much of it revolves around mutual incomprehensibility.

    Thus when his wife is to give birth he’s terrified that the midwife in Stockholm will not understand him and thus things may go horribly wrong. Yet, here’s a passage that kind of sums it up (it’s my translation from the Swedish, vol 2 pg. 375):

    “The redhead came from behind the counter when the bell above the door rang, I said what I wanted, one of the big sourdough breads, six small baguettes, two cinnamon rolls, pointing at them simultaneously, since even the most simple Norwegian words are answered with “what?” when in Stockholm, she then put it all in a bag and rang it up on the register.”

  186. Alicia said,

    July 27, 2014 @ 10:26 pm

    I come at this from the perspective of a native (US) English speaker who learned German to near-native fluency age 18-20.
    After 2 years of college German, 2 months of immersion High/Standard German at the Goethe Institute, and 2 months immersion in a mixture of High German and Schwäbisch (a dialect in the south west of the country), my German was good enough for conversational purposes, although I was still missing chunks of university lectures.
    I spent a long weekend in the Netherlands, where I found I could understand nearly everything that was said, if the person enunciated with a modicum of patience and a bit of gesturing. It took more _concentration_ than understanding German, but communication was no problem. I spoke mostly in German, backed up by gesture. Once I asked directions from an old woman, and in the paragraph response the only word I didn't understand was "highway overpass," which, when I got to it and realized what her mime must have meant, I hardly blamed myself for not knowing.
    However. The next day t I met a cute German guy and we were walking somewhere and stopped to ask directions. I understood clearly what was said, but he was completely lost. I think the fact that German was somewhat "unnatural" to me meant I had more of my brain engaged. Also, I'd already had to adjust my ear for one "dialect" of German, so retuning it again apparently wasn't a problem. Or maybe he was just an idiot.
    I would be curious to hear whether non-native-but-fluent English speakers have the same experience with the various dialects of English (Aussie, Scot, etc), of finding them more easily understood than do native English speakers not familiar with that dialect. Or any other analogous situations.
    Note: I still understand Dutch 90% when spoken clearly, and it still sounds to me like perfectly orthodox German being spoken under water. Reading it, however, is about a 30% success. The orthography is completely different. Germans write the "eye" sound "ei," the Dutch write it "ij": WTF. With that kind of chaos, I'm amazed at the commenter above who claims she can do more with written than spoken.
    Another theory, about why the cute German couldn't understand the Dutch while the Dutch could evidently understand me, and the frequently cited claim that the Portuguese understand the Spanish more than the reverse, I think it's more likely a case of exposure than anything linguistic. There's enough German language Stuff produced in Europe that most Dutch speakers have already had occasion to develop some degree of aural diglossia, whereas a German is more likely to be able to remain blissfully ignorant. Likewise for the relative positions of Spanish and Portuguese, both in Iberia and in South America.
    Oh, later in the year, having achieved "near native" fluency in German, I went to Denmark. There I had that experience of feeling like I ought to be able to understand, but not getting more than 30% not matter how hard I listened. Rather than German underwater, it sounded like German through a garbage disposal. I do about 40% when reading. Swedish and Norwegian I understand even less, but can read about as well as Danish.
    Yiddish I can understand 90% if it is spoken or if I take the time to sound out an English speaker's attempt to write phonetic German. Which is somehow extreeeeemly difficult for me.
    The short bits of Afrikaans I've seen written looked like Dutch, except written by someone who wasn't on psychedelics. I understood 80% or so.

RSS feed for comments on this post