Difficult languages

« previous post | next post »

The December 17th Economist contains an article entitled "In Search of the World's Hardest Language". Such things usually make me groan, but this one is actually pretty good. At the level of detail one can reasonably expect in such a context, the facts seem to be correct, the range of languages considered is broader than usual, and it recognizes that there are multiple factors involved. There are, however, a few points worth making about this article, as well as inferior examples of the genre.

One is that it doesn't really define what is meant by "difficulty". This could mean how hard a fluent speaker has to work to speak or understand the language, how hard it is for a child to acquire the language as his first language, or how hard it is for an adult to acquire as a second language. The context suggests that the last is what is intended. Even there, do we mean how hard it is to acquire near-native fluency, or how hard it is to acquire the rudiments, or some intermediate level of functional ability? These may turn out to be different.

Another question is whether difficulty is a property independent of the learner. It is, I think, the received view, that how difficult a language is to learn depends in part on what languages the learner already knows. Navajo is by all accounts quite difficult for English speakers, but probably not all that difficult for speakers of Apache. Koreans seem to find Japanese grammar much easier to master than English speakers do, presumably because the syntax of the two languages is similar.

It is also quite possible that different aspects of languages are more or less difficult for different learners even if one controls for what languages they start with. Some people may be good at, say, learning complex morphology but not very good at learning an exotic sound system. Others may be good at the latter but not at the former.

A point that frequently arises is the idea that languages that pack a lot of information into words are difficult. Is it really self-evident that it is harder to pack deal with complex words than with multi-word phrases that convey the same information? If a language puts a lot into a word but does so in a transparent way, so that words are easy to parse, interpret, and construct, why should this be difficult? It may well be that the perception of difficulty here merely reflects unfamiliarity, which is likely true of quite a few of the features often cited as leading to difficulty.

Perhaps the largest problem with this entire genre is that factors that might lead to difficulty are enumerated but no empirical evidence as to which languages are actually difficult to learn is ever adduced. There is very little research on this question, so one can't criticize authors for failing to cite it, but in the absence of such research discussion of the factors that lead to difficulty is mostly mere speculation.

The article also seems a bit confused regarding about theoretical debates in linguistics, conflating the question of the extent to which universals particular to language exist and the related question of how and to what extent languages differ from one another, with the question of whether language influences thought. These are two quite different questions. For one thing, the examples cited by proponents of the Sapir-Whorf hypothesis are almost entirely lexical, while Chomsky's proposals regarding universals are almost entirely grammatical. It would be perfectly possible, in principle, for there to be a world in which all languages had exactly the same grammar, differing only in their sound systems and lexica, where the differences in lexicon influenced thought in accordance with the Sapir-Whorf hypothesis.

In any case, in my estimation support for the Sapir-Whorf hypothesis is even weaker than the article suggests. It mentions that critics consider the examples put forward "trivial", but not that even such "trivial" effects are disputed and, I think, usually wrong. Our colleague Lila Gleitman is a real expert at this: she can debunk most Sapir-Whorf claims within seconds.

The example given in the article epitomizes the problem nicely. It points up a correlation between language (exclusive reference to compass directions vs. use body-oriented directions) and behaviour (constant tracking of compass direction vs. the lack thereof), but doesn't even attempt to establish the direction of causation. Surely it is perfectly plausible that what is at work here is a cultural focus on external vs. individual orientation which has linguistic consequences. If you don't care about individual orientation you don't have words for it. The Sapir-Whorf claim would be that it is the lack of words for left and right that causes people to rely entirely on compass directions. Even if you consider this plausible, the facts cited in the article do nothing to favour this direction over the reverse.

Whereas it is hard to say much about which language is the most difficult to learn, the difficulty of learning writing systems is better understood. English does indeed have rather a bad writing system, but it pales in comparison to that of Japanese, particularly in its pre-WWII version, which feature more and more complex Chinese characters and historicizing kana spellings. And although we have little direct information about the difficulty of learning it, the vaguely similar writing system of Hittite must be another contender for the title of most difficult writing system.


  1. Matthew Stuckwisch said,

    January 2, 2010 @ 5:15 am

    Add his count of Spanish verb forms to the list of (minor) mistakes.

    Most (non-defective) verbs have 71 forms (57 unique)

    IMPERSONAL – 3 (3)
    | Infinitive – 1 (1)
    | Participle – 1 (1)
    | Gerund – 1 (1)
    INDICATIVE – 35 (30)
    | Present – 7 (7)
    | Preterite – 7 (6)
    | Imperfect – 7 (5)
    | Future – 7 (6)
    | Conditional – 7 (6)
    SUBJUNCTIVE – 28 (22)
    | Present – 7 (6)
    | Past – 14 (10)
    | Future – 7 (6)
    IMPERATIVE: 5 (2)
    |Informal: 3 (2)
    |Formal: 2 (0)
    The way to actually count this could be questioned, though, and result in different numbers based on different counting methods. If not too impressive when looking at some other languages, it is much higher than his count of 48.

  2. Peter Taylor said,

    January 2, 2010 @ 5:26 am

    @Matthew Stuckwisch, how do you get e.g. 7 distinct present indicative forms in Spanish? Are you counting both tú and vos? I suppose that if you're looking at Spanish as a whole they exist, but from the point of view of someone learning the language there are somewhere between 4 and 6 forms depending on the dialect. (4 for the dialects which have no distinct second person and use only usted/ustedes; 6 for Spanish Spanish and other (rare) dialects which still have a distinct second person plural).

  3. Alon said,

    January 2, 2010 @ 9:27 am

    @Peter Taylor: from the point of view of someone seeking to communicate with Spanish speakers, it makes sense to learn to recognise and understand both voseante and tuteante forms, even if they won't use more than one in production. Unless they are learning Uruguayan Spanish, that is, with its threefold T-V distinction between 'vos', 'tú' and 'usted'.

  4. Alex Case said,

    January 2, 2010 @ 9:46 am

    A lot of the same thoughts popped into my head (in a much less coherent way!) as I was reading the article on the bus, especially the point about long words not necessarily being more difficult than long sentences. The only criticism I would slightly disagree with is that the article does make clear that it is at least starting by looking at how languages are difficult for English speakers ("Assessing how languages are tricky for English-speakers…"), although it is indeed not clear if that switches at any point during the article

  5. Thomas Westgard said,

    January 2, 2010 @ 11:12 am

    The article is amusing but confusing when you start to think. The perspective keeps shifting. The English language is:

    1. Used as the medium of communication.

    2. Used for point of familiar comparison, as when pointing out that sounds spelled with "a" don't necessarily equal one sound.

    3. Used as the starting point for determining what is difficult. Click sounds are considerably easier if your mother tongue already uses them, but click languages are "difficult" because English has none.

    4. Depicted in a light equal to all other languages in the portion of the article dealing with brain function.

    Those last two are mutually exclusive.

  6. Matthew Stuckwisch said,

    January 2, 2010 @ 11:39 am

    Peter: This is why I said one could count them differently. As well one could ignore the future subjunctive since it's reserved to legal writing and other extremely formal texts. The 7 is indeed vos as well, since vos does have distinct forms from tú in the present and imperative forms. But that's also why I did a unique count. For someone learning the language, tú is absolutely necessary, vosotros and vos are more optional (though I suggest learning both). What dialect exclusively uses Vd/Vds? I've never run across a description of a dialect like that, thought given it happened with Portuguese in Brazil to a large extent it wouldn't surprise me in border countries.

  7. Katie said,

    January 2, 2010 @ 12:11 pm

    Aren't linguists often guilty of making the opposite claim–that all languages are equally difficult/complex? Or a slightly more nuanced version, that simplicity in one part of the grammar is made up for by complexity in another? Not only that, but when I've heard it made in intro classes, it's with far less explanation than this article gives. Really, I understand why this point gets pushed in intro classes–it's to counteract the "speakers of language X must be stupid because their language doesn't have Y" line of thinking–but if we don't actually have evidence for it, why do we insist on pushing it? Though surely we have enough evidence from first language acquisition to begin to answer these questions …

  8. Ed said,

    January 2, 2010 @ 12:37 pm

    "A point that frequently arises is the idea that languages that pack a lot of information into words are difficult. Is it really self-evident that it is harder to pack deal with complex words than with multi-word phrases that convey the same information? If a language puts a lot into a word but does so in a transparent way, so that words are easy to parse, interpret, and construct, why should this be difficult? It may well be that the perception of difficulty here merely reflects unfamiliarity, which is likely true of quite a few of the features often cited as leading to difficulty. "

    One thing I've noticed is that both English and Chinese are at one extreme of this scale, of using combinations of words to convey different meanings instead of one big word. And both have relatively complex writing/ spelling systems!

  9. Andrew Dowd said,

    January 2, 2010 @ 12:40 pm

    I remember reading a paper about L1 aquisition of Navajo, and being somehow relieved that the verbal morphology took such a long time to acquire.

  10. Clarissa said,

    January 2, 2010 @ 3:55 pm

    I think the writing system in Japanese looks more complex from the outside than the inside, at least at this point in my study–having studied some Japanese and some Chinese, at least I can always "spell" (write down in a syllabary) any Japanese phrase that I can hear, whereas that's absolutely not true at all in Chinese, even though Chinese is superficially more straightforward (since it has one system instead of three or four, depending on whether you include romanized letter). I may not be able to write it the same way a highly educated Japanese adult, but at least I can write down something that may be readable–whereas with Chinese I can't write it at all unless I already know how to write it. I have a terrible aural memory, and that's one reason why I switched to Japanese, actually…

    At any rate, the article itself fills me with a strong sense of déjà vu. It's not bad, but it would have been a lot better if it had incorporated Bill Poser's considerations. However, the intro, acknowledging the "fact-challenged" nature of most books about that "certain genre" of books about English, is pretty refreshing.

  11. Lane said,

    January 2, 2010 @ 3:59 pm

    Hi – I'm the author of the article and a frequent denizen around here, though I don't comment as much as I'd like…

    I'll cop to conflating the theoretical debates (innativity, universals, Whorfianism) a bit; I know they're distinct questions but I didn't disentangle them as well as I should.

    I'll also cop to leaving out the "voseo" in my Spanish counts – silly, since I speak Spanish heavily influenced by my time in Uruguay, a voseo country. The reason? My high-school textbooks taught six forms, including vosotros but no vos. I guess I got it into my head permanently that there are six "real" forms (my brain still conjugates like hablo, hablas, habla, hablamos, hablais, hablan), and have always considered the voseo optional/extra – even though I use it!

    But to one more of Bill's points, I'll defend myself. As someone else pointed out – I did say what I thought would be "hard", in a very brief and unscientific journalistic way: what would count as hard for an adult native English speaker to learn. Even then, Bill's right that the answers I came up with were my own musings, not based on evidence (which I never found). I hope it was entertaining to people nonetheless. The article was meant to be more fun, and to show/celebrate the diversity of languages not known to many Western readers, more than it was to come up with the definitive scientific answer to the question.

  12. peter said,

    January 2, 2010 @ 4:39 pm

    I'm surprised the article did not also mention so-called mother-in-law languages as another source of difficulty: Is it correct that in the most extreme cases (some Australian Aboriginal societies) children are expected to acquire multiple languages in order to converse correctly with everyone?

  13. Peter Taylor said,

    January 2, 2010 @ 5:50 pm

    @Alon, I would consider vos analogous to y'all: many native speakers will go a lifetime without once hearing it, so whether you need to recognise it or not depends on your specific circumstances and reasons for learning Spanish. Personally I read up about it in the Diccionario panhispánico de dudas about 15 years after starting to learn Spanish, because I was preparing a basic language briefing for a group of people who were heading to Argentina for a couple of months.

    @Matthew Stuckwisch, I understand that there are dialects in Colombia which use solely ud/uds, and there are certainly characters in Colombian novelas who address parents, children, friends who are close enough to call "hermana", etc. as usted.

  14. John Thayer Jensen said,

    January 2, 2010 @ 6:11 pm

    I confess I groaned. Difficulty of language seems to me almost entirely a matter of where you are coming from.

    The article made inclusive/exclusive and singular/dual/plural sound difficult. Having spent 8 years speaking (during the day) almost exclusively Yapese, which has both, I still find it annoying that English has not. I sometimes feel I am being unclear – well, I am being unclear – in English when, for instance, I say 'we' and there is an ambiguous situation. Likewise, the first person inclusive dual is very handy. And another discussion in this blog shows how lacking a singular/plural distinction in the the second person drives dialects everywhere to invent one.

    I don't think the Whorfian comments were very relevant. The writer quoted pointed out the different ways of thinking implied by different languages – but if I understand the pure quill Whorfian hypothesis, it was that there was no way for a person to change his way of thinking – no way, for example, for the Dyirbal ever to learn the concept left/right.

    Yapese, like the people she quoted, normally greet one another either by asking where one is going, or by telling where one is going. Yap being a smallish island, one is always either going inland or seaward. That's not the way we normally think in English, even when we live in islands (although when I lived for many years in Hawai'i, mauka (inland) and makai seemed normal.

  15. Bob Ladd said,

    January 2, 2010 @ 6:32 pm

    I notice that Arnold Zwicky has also commented (here and here) on the two articles on language in the Economist's Christmas special issue, and that he came up with a similar verdict to Bill's. I had the same reaction – compared to a lot of what's in the lay press, these pieces were really pretty good.

    The US Foreign Service institute (Language School) used to have a ranking of language difficulty, based on how long it typically took trainee diplomats to reach a certain level of facility. The levels were operationally defined along a scale of 1 (beginner) to 5 (near-native), and the definitions took all kinds of things into account including literacy and cultural awareness. If I remember correctly, the hardest languages the FSI dealt with in the early 1970s were Chinese, Japanese, Korean, and Arabic, and the easiest were the nearby European languages written in the roman alphabet. There were four bands of difficulty, but I don't remember which languages were in the middle two bands, or how many months of training were typically required for languages in each of the four bands. But at least at that very practical level, there is indeed some research on how difficult languages are (for L1 English speakers acquiring an L2).

  16. John Cowan said,

    January 2, 2010 @ 10:42 pm

    Difficulty in learning doesn't quite scale with remoteness from the native language. I'd have a very hard time acquiring RP British English, or Scots, for example — the L1-interference problem would be enormous.

    Jacques Guy wrote a short article on sci.lang years ago about the two Vanuatuan languages he did fieldwork on, one of which (Sakao) is much more complex than the other (Tolomako). Indeed, Sakao is tough by world standards, whereas Tolomako is remarkably simple.

  17. Sam said,

    January 3, 2010 @ 4:03 am

    peter asked whether children are expected to acquire multiple languages in order to converse correctly with everyone? Babies can pick up any number of languages they are exposed to effortlessly from first year on. Speech sounds stimulate the wiring of language synapses, which are sorted out and stored up as tunes, without specific meaning. This facilitates the imitation of sounds, without meaning, at around 13th month. Sounds are associated with meaning at around 15th month. They are used in context, without any reference to national boundaries. Babies join our playgroups in our centres in Hong Kong, conducted by native speakers from 5 different countries, at around 3 month. They can speak 5 languages by the time they are two, in context they are familiar with, without any difficulty or formal instruction.

  18. Andy said,

    January 3, 2010 @ 1:05 pm

    For English speakers, learning another language presents verb conjugations and noun cases. Yes, we do have verb conjugation, but it's not as complex as most other languages. Learning Russian, I found noun cases (Russian has 6) an alien concept, though it does get easier and more instinctive. Hungarian, however, has 35.

    In my opinion, no phonetic language can be ruled the most difficult. When you think about it, a non-phonetic language more or less requires years of immersion to reach fluency. A phonetic language can be learned much more readily through self-study.

    In case it isn't a clear, I take a phonetic language to mean one where each character translates directly to one sound. Spanish is a good example.

  19. Franz Bebop said,

    January 3, 2010 @ 2:08 pm

    @Lane: Thanks for writing this article, I read it in the print edition and enjoyed it.

    @Bill Poser: There is very little research on this question, so one can't criticize authors for failing to cite it, but in the absence of such research discussion of the factors that lead to difficulty is mostly mere speculation.

    Really? This question has not been researched much? I'm surprised.

    @Sam: They can speak 5 languages by the time they are two

    They "speak" five languages at a two-year-old level. Let's be fair.

    I don't agree that there is any support for the idea that language acquisition by babies is "effortless." That can only be true if "effort" is defined in such a way that it excludes all the things that babies normally do, such as endlessly repeating words that are said to them, verbalizing with parents and siblings, etc. Being a baby is hard work.

    Has any research been done that directly challenges the received idea that babies are better at language acquisition than adults? I often wonder how adults would fare if they were exposed to new languages in precisely the same sort of setting as babies. An adult with no need to worry about work, food preparation, washing clothes, or any adult hobbies, but able spend the whole day (morning till night, every day, no exceptions) playing with peers and encouraged by endlessly patient caregivers, and with absolutely no recourse to their native language, might learn a new language very quickly indeed. I would hypothesize that adults would do surprisingly well, perhaps a lot better than babies.

  20. Ellen K said,

    January 3, 2010 @ 3:50 pm

    @Andy: I think you are talking about writing systems that are phonetic, not languages.

  21. Coby Lubliner said,

    January 3, 2010 @ 6:13 pm

    Other examples can be given of Jacques Guy's finding that one of a pair of closely related languages is unequivocally easier or simpler than the other. In the ones I can think of the difference is mainly phonological. I have found Macedonian much easier than Bulgarian, Galician than Portuguese, and Norwegian (Bokmål) than Danish.

  22. sephia karta said,

    January 3, 2010 @ 7:17 pm

    @Franz Bebop: I had a small argument with someone about this the other day, and I feel the same as you do: I don't think it is a given that adults are worse at learning languages than babies if you consider the dedication babies get to apply to the task.

  23. Tyrone Slothrop said,

    January 3, 2010 @ 8:17 pm

    Does anyone here at Language Log even read either Sapir or Whorf anymore? I mean, on a non-trivial level, if you are going to critique either Sapir or Whorf a general familiarity with their actual writings might be helpful..I know since Pinker didn't need to read them to criticize Whorf, why should anyone else…but leaving that point aside…wouldn't it be nice to be familiar with their arguments and not the stalking horse "the Sapir-Whorf hypothesis?" (Yes, Hoijer made popular that unfortunate turn of phrase and he too should be read for what he wrote and how he reframed the writings of Sapir and Whorf.)

    There is, of course, much work being done in the Sapirian and Whorfian vein. The late Dell Hymes's work, for example, on ethnopoetics would certainly fit into the Sapirian tradition. Michael Silverstein's work on language ideologies certainly (and Silverstein is explicit about this) falls into the Whorfian tradition.

  24. John Thayer Jensen said,

    January 3, 2010 @ 9:21 pm

    I did indeed read both Sapir and Whorf very carefully – but, alas, a long time ago – when I was a linguistic student in the 1960's. I also read a fair bit of Hymes at the time, although less intensely – Hoijer, not. And I confess I haven't read any of them for a long time.

    Your point is well-taken.

  25. Rohan F said,

    January 3, 2010 @ 9:48 pm

    As a nine-year learner of Ubykh, I often have to stifle a laugh when the ease of learning languages is raised, and it was a welcome change to see some less-widely-spoken languages (including Ubykh!) being looked at in this article. (Though @Lane: the number of consonants usually considered native in Ubykh is 80, not 78.)

    With that being said, it's true that a tendency does exist to see highly agglutinative languages as difficult. But I remember David Rood, who works with Wichita, putting it very simply: highly agglutinative languages like Wichita merely do their complexity on the word level, where English does it on the sentence level. Ubykh morphology permits words of up to 20 morphemes (AFAIK, Ubykh and its sister-languages Abaza and Circassian are the only languages in the world that permit quadripersonal verbal agreement), but the syntax is for the most part pretty unremarkable and tends not to vary greatly from one sentence to the next.

  26. J. Goard said,

    January 4, 2010 @ 9:06 am


    Exactly right. Quite a few people seem to miss the obvious point that to whatever extent a certain thesis is poorly defined, its negation is equally poorly defined. It's absurd to claim both that the Sapir-Whorf hypothesis is terribly ill-defined, and also that it's false because "any thought can be expressed equally in any natural language" or some such LING 1 drivel.

    Once we move away from grand claims and consider measurable specifics, there is very strong evidence for effects of language on aspects of spatial cognition, for example. An example I know well, and teach, is the work of Soonja Choi on Korean versus English encoding of spatial relationships. Monolingual Korean and English children group spatial relationships very differently, in line with Korean verb usage or English preposition usage, and there is no (even slightly) plausible mechanism for "culture" to instill those differences first, thus subsequently shaping language. The only reasonable causal path is from linguistic classification to spatial cognition.

  27. vanya said,

    January 4, 2010 @ 9:32 am

    I agree with Franz and Sephia above. The idea that babies learn languages "effortlessly" seems like one of those commonplaces no one really challenges, but which fall apart in two seconds if you think about it. Many young children do acquire second languages from nannies or parents at a young age, and then promptly forget them completley by adolescence if they don't use them. (And kids who can say their colors and numbers in Spanish at age 3 do not "speak Spanish".) So how useful a skill is that? Anyone who has tried to raise a child bilingually in a language that's not spoken commonly around the child knows that it takes a great deal of effort and focus to help the child maintain the second language. And in the US those parents usually fail. I know many couples where one parent speaks something other than than English as their native tongue – some of these kids have a smattering of Bengali or Portuguese or whatever, mostly from traveling to see relatives. The only children who seem to me to be truly bilingual are the Spanish/English children.

  28. John Thayer Jensen said,

    January 4, 2010 @ 3:14 pm

    I had a friend for years – now lives in China – who effectively had no native language. He grew up in Malaysia, with a father who spoke Mandarin, a mother who spoke Hakka. The home language was a mix of the English spoken by the parents, occasional Hakka, Mandarin, and Cantonese-as-a-second-language by the mother.

    When I knew him he was struggling to do a Masters thesis in English. He, his thesis adviser, and I talked about the grave difficulties he was having and our conclusion was that he had not properly learned any language.

    Of course he had a human language but it was very much an idiosyncratic creole. He has since learned Mandarin, and, I think, learned it well.

    On the other hand, arguing in favour of a special children's ability to learn language, are all the "wolf children." But sorting out what is specifically linguistic from the other pathologies they have acquired is obviously not easy.


  29. Gary Lupyan said,

    January 4, 2010 @ 6:45 pm

    Rick Dale and I have a paper coming out sometime this month in PLoS ONE in which we conduct a statistical analysis of over 2,000 languages and show that morphological complexity correlates strongly with social/demographic factors. This paper offers exactly the kind of empirical evidence that the Economist article is lacking.

    The relationship between language structure and social structure is far from arbitrary. Indeed, languages spoken by more people (and those having more L2 learners) consistently have simpler morphology. Conversely, languages that are only learned and used by native speakers, have more complex morphology. We offer a comprehensive theoretical treatment of these results.

    Stay tuned: Lupyan, G. & Dale, R. (in press). Language Structure is Partly Determined by Social Structure. PLoS ONE.

  30. Greg Morrow said,

    January 5, 2010 @ 12:49 pm

    Given that the sound stream is continuous or nearly so, why do we chop it at two different levels, word and morpheme? Lupyan & Dale's work certainly implies that the word-chopping is easier than morpheme-chopping, which matches my prejudices, only I don't know why.

    Is it because morpheme interaction has a greater effect on sound production than word interaction? Vowel harmony in (agglutinative) Turkish doesn't work across word boundaries, for example, but changes how morphemes are expressed. So a language with simpler morphology has phonetics that are accessible to a more diverse population. More complex morphology is disrupted more easily by the phonetic divergence across a diverse population.

  31. Alan H. said,

    January 5, 2010 @ 6:57 pm

    Greg: There are other criteria you can use to determine whether something is a word or a morpheme. AFAIK, in morphology you usually don't have morpheme order changing the way you do at the syntax level. I am not aware of any language that allows free morpheme order, for instance, although I could be wrong about that and would be happy to be enlightened if I am.

  32. J. Goard said,

    January 6, 2010 @ 8:17 am


    Morpheme order's never free. What's the ontology?
    If it be free, none dare call it morphology.


  33. Elizabeth P said,

    January 6, 2010 @ 5:22 pm

    Just some comments
    1) When the Economist means "difficult" they apparently mean "difficult for an adult English monolingual speaker". I would have to agree that when languages are considered from an L2 perspective, some may be "harder" than others because differences in areal features, writing system, and lack of cognates from the original L1. This shouldn't be news to us.

    2) It looks like "word length" is another factor in determining "difficulty" (at least for many English speakers). Yet, I think we have to consider spelling conventions across languages. For instance, some languages like German are happy to spell compounds as one long word, while English tends to split them into separate words in terms of SPELLING but not necessarily pronunciation

    That is, English speakers have decided that "blackboard" is one word based on spelling, but actually so is "yellow jacket" (i.e. the bug with the stinger) because it has the same stress compound stress pattern as "blackboard". If English spelling were more like German, we might be having words like "yellowjacket" or "manillafolder" which now look "hard". In fact we do have place names like "Cooperstown","Harrisburg", "Selingsgrove", "Danville", "Reisterstown","Westminster".

    I bring this up because most "long foreign words" (e.g. "katakana", Japanese) are generally compounds in the original language and might have the same status as a word like "Cooperstown" or "Harrisburg" does in English. Although longer than a single root, they mostly have elements which speakers recognize; few native speakers consider them hard to pronounce (versus a Native American name like Okinawa, Aberystwyth, Conowingo, Wicomico or Mongahela which are "hard" for English speakers). Yet many of these names like "Aberystwyth" and "Okinawa" were compounds in the original language.

    3) I think it's easy to underestimate the inflectional morphology of English because of its relative poverty of Latin type inflections. Yet it can make up for it in other areas such as tense/agreement clitics which attach to a subject noun (I'm referring to 'm, 's, 're, 'll, 've and even ll've and d've). Is this hard? Not for a native speaker, but for an ESL speaker – who knows?

  34. peter said,

    January 6, 2010 @ 8:03 pm

    Franz Bebop said (January 3, 2010 @ 2:08 pm):

    "Has any research been done that directly challenges the received idea that babies are better at language acquisition than adults? I often wonder how adults would fare if they were exposed to new languages in precisely the same sort of setting as babies. An adult with no need to worry about work, food preparation, washing clothes, or any adult hobbies, but able spend the whole day (morning till night, every day, no exceptions) playing with peers and encouraged by endlessly patient caregivers, and with absolutely no recourse to their native language, might learn a new language very quickly indeed. I would hypothesize that adults would do surprisingly well, perhaps a lot better than babies."

    In addition, most adult learners have two characteristics every single baby lacks:

    1. Past experience of learning, including the learning of at least one language.

    2. An ability to reflect on their own learning experiences and to adjust their own learning strategies accordingly.

    These two characteristics should facilitate adult language learning, provided the language teaching methods make use of them. However, many adult language training programs use rote, repetitive learning by ear. This is very odd considering that adult students have at least 10 and maybe as many as 20 years of experience learning by eye (using reading and writing), most of which was not done by rote.

  35. Gary Lupyan said,

    January 7, 2010 @ 4:24 pm

    Regarding children being better language-learners than adults:

    One issue here is what counts as language-learning. The best evidence for children's better skills is in the domain of phonology. The age at which one starts learning a language predicts mastery. No controversy there, as far as I now.

    There is absolutely no evidence for kids being better at learning vocabulary than adults. The "2 years olds learn X words/day" kinds of claims are misleading. Adults learne fewer words because they run out of words to learn. See Paul Bloom's work and Bob McMurray's work on the vocabular explosion for more info.

    With grammatical issues, things get really interesting. Classic work by Johnson and Newport (1989) shows that age-of-onset (English L2 learners) predicts mastery of a variety of grammatical devices, and this agrees with common sense notions that people who start learning a language earlier are more fluent.

    What has *not* been shown, to my knowledge, is any lab-based task in which kids do better than adults at learning some new grammatical device, whether in a real language or in an artificial grammar. After all, if kids are just better at this type of learning, it shouldn't be so hard to demonstrate it. If someone knows of such a demonstration, please point me to it..

  36. Knut Holt said,

    January 13, 2010 @ 3:10 pm

    "I remember reading a paper about L1 aquisition of Navajo, and being somehow relieved that the verbal morphology took such a long time to acquire."

    This type of morphology is not unique. You find something very similar in spanish and Italian.

    A predicate in these languages often consist of 1- one or more enclitic pronouns and adverbs 2- an auxilaiary and 3- the main verb, clung together in a rigid fashion like one long word, allthough written as several words.. These are again composed of smaller units with a specific function, but with intricate connection between them. If you divide it all in the smallest unit s and compare with the Navaho template, you can see that there is a great structural smilarity.

    But is this type of structure difficult? Yes, it takes a long time to master in all its details because you must loook at many elements to grasp the right meaning of tense and aspect and what is indirect and direct object and a lot other things.

    Italian and Spanish also have predicate without auxiliaries. Those are somewhat simpler, but still you can have that intricate string of "prefixes", usually called enclitic pronouns and adverbs.

    An example:

    Gliel'ho rimandato – meaning I have sent it back to him.

    If you decopose it you get:

    glie-l-o-ri-mand-at-o, with the structure:

    indirect objest.diredct object- aspect/tense/subject – back-send -aspect/tense- object gender and number

  37. Gary Lupyan said,

    January 28, 2010 @ 2:05 pm

    The paper I mentioned earlier is available here:
    "Language Structure Is Partly Determined by Social Structure"

  38. Mauro Chiesa said,

    March 18, 2010 @ 9:24 am

    I am not a linguist, but I spent the first four years of my life speaking Friulan, then learned Italian for four years and then English, coming to North America, and then French. I subscribe to The Economist that prides itself foremost on economics, the (questionable) art of managing scarce resources, and in the context of the article, it may be simply "time" relative to the merits achieved by learning the language. The general readership of the magazine, businesspeople whose time is scarce and from whom globalisation requires ever-expanding linguistic demands, is faced with that dilemma on a daily basis.

  39. Oliver said,

    August 22, 2010 @ 12:04 pm

    Doesn't the amount of meta-information you have to learn with a word make a difference? A language like French or German which require you to learn the grammatical gender of a word should be more difficult than a language like Serbian where it is largely predictable.

RSS feed for comments on this post