The Puzzling Paradox of Sign Language
technologyreview.com
technologyreview.com
These facts are patent to anybody with a passing acquaintance of American (or any other sign language). The 'paradox' they're discussing here is tantamount to saying, 'Japanese words tend to be longer than English words. So how do the Japanese communicate as effectively as we do?'
Define interesting. Personally, I find the entropic analysis of spoken Engligh vs. sign language pretty interesting, especially so when they relate it to channel capacity and other information-theoretic stuff. Your comment sounds anti-intellectual.
Not a thing wrong with that. Fwiw, I file Technology Review with Wired and SciAm. That is, fluff. Hardly intellectual material imho.
>I find the entropic analysis of spoken Engligh vs. sign language pretty interesting
That would be interesting if they weren't using so poor an approach. Constraining the analysis to handshapes goes too far. Like constraining English analysis to vowel sounds, or removing prepositions.
Maybe the paper gives the wrong impression but it sounds like the parameters of the study were determined by someone with very little understanding of how ASL actually works.
(edited: formatting)
Sure, TR and Wired are fluff. Except that the article links to an arXiv paper. This is not a standard TR article. Moreover, you can't present scientific research in half a page without making it fluff. That, too, is an interesting information-theoretic problem.
"That would be interesting if they weren't using so poor an approach."
You missed the forest for the trees. In case you didn't notice, the people who wrote the paper are electrical engineers. They used the information-theoretic approach that is used in communications theory to a problem outside the traditional scope of application of the theory. Sure, natural language is hard, but if you start your research by focusing on all the little details you will get nowhere. To me, the paper looks like a first shot at a difficult problem. If you can do better, I would love to hear about it.
"...it sounds like the parameters of the study were determined by someone with very little understanding of how ASL actually works."
That is not the point. The point is that someone who does indeed understand how ASL works can read the paper, find out what is missing and build on it. No one knows everything, and inter-disciplinary work is very hard. Your criticism is too hard, because no one ever built a theory in one single iteration.
I just disagree. I have some familiarity with ASL and I feel they've gone off in the wrong direction from the start. I suppose you could say that a future study could build on this one by completely debunking it. Perhaps that argument has some merit.
>Your criticism is too hard
Probably. But I got the feeling from the paper itself that this is someone's last-minute rush to finish a thesis or fulfill a grant obligation or something. I guess I'd like to see a little passion in my science. Sorry if that seems too harsh.
I know a bit of information theory, but I know zero of ASL. It's quite possible that they went in the wrong direction from the start, but someone can still write a paper to point that out and prevent other people from repeating the same mistake. There's value in going in the wrong direction: it serves as a warning to others.
"I guess I'd like to see a little passion in my science. Sorry if that seems too harsh."
Personally, I found the paper's presentation horrible. I would never submit something so visually unappealing under my name. I agree that it sounds like a last-minute rush to finish something. I also agree that there seems to be little passion in it.
However, let us look at the authors: 1st author in an EE undergrad, 2nd author is a post-doc, 3rd author is a professor. Of course, the undergrad did all the work, the post-doc guided him, and the professor secured the grants that paid for the effort. Despite all the paper's flaws, I still think it must be judged for what it is: an EE undergrad trying his luck outside his field... and failing, perhaps.
> 'Japanese words tend to be longer than English words. So how do the Japanese communicate as effectively as we do?'
I speak some Japanese. Japanese is actually a bit harder to communicate in than most other languages. The emphasis on social unity and modesty has evolved to make the language incredibly nuanced and complex. Nowadays, there's some factual statements that are almost literally impossible to say in Japanese because the words/phrases/grammar to say them don't match up.
Feynman touched on this in "Surely You're Joking" - there's different verbs for the exact same action based on the social situation and how well respected the person you're talking to is. So if you want someone to look at your garden, you say something like, "Would you please take-a-pasing-glance at my barely suitable garden?" But if you want to look at someone else's garden, you might say, "May I look-with-admiration upon you're beautiful garden?" If it's an official figure or a temple, it'd be something like, "May I quickly-glance-my-eyes-over-reverently on your most majestic of gardens?"
Different verbs. All mean "view" - but with different levels of politeness and appropriateness. One of the HNers who lives in Japan could probably comment more, I haven't spent all that much time there. Fascinating and beautiful language, but tricky to say things sometimes.
Care to elaborate? As a English/Mandarin/Japanese speaker, I am certainly acutely aware of the limitations of translating from one language in to another, but find such an assertion hard to accept without concrete examples.
I'm nowhere near fluent in Japanese, so I'm not the best person to get into it. I do have a couple good friends who are native speakers or expats with a very high level of understanding. A good mate runs a bilingual IT firm in Tokyo, and can handle even technical/legal Japanese to a pretty high level.
Sometimes I've asked, "How do I say XYZ?" and there's no real way to say it without describing it. To go back the other way, what would the English equivalent of "domo" or "daijobu" be? Daijobu is something like "don't sweat it", but that's not quite right. I'm not sure there's anything quite like domo.
But, that's common enough in other languages. German doesn't "cheers", which is a pretty useful and versatile British English word, but that's not what I mean here. From my general understanding, it's very hard to put together appropriate forms of disagreement in Japanese. Like, it might not possible to say something like, "You're wrong, but I respect the sentiment", or "Agree to disagree."
I'm not sure on those specific examples. I know one of the first things people learn about doing business in Japan is you'll rarely hear "no" from a subordinate, so you have to be very aware of something like, "That might be difficult..." - that might mean it's impossible.
Not entirely sure - I've spent a bit of time living and working in a lot of places, and picked up touches of a fair few languages. I've had people translate for me to some extent or other in German, Dutch, Spanish, Mandarin, and Japanese. By far, the most stumbling and "that's hard to say" I got was in Japanese. I'd actually be curious to have someone more skilled with the language elaborate more to see if I'm on the mark here.
That said, the character of the Japanese language is really beautiful - I love reading Japanese literature translated into English. You'll commonly see a very factual, plain description, followed by a very profound beautiful statement mixed in with minimal transition. Eiji Yoshikawa's "Musashi" is pretty much a must-read for anyone talented who feels some friction with more mainstream and normal society.
Now, let me clarify upfront that my minor (technically a "joint-degree", but effectively a minor) is in Linguistics, so this may get wordy and technical.
What you are mostly describing is a not a restriction of the language itself, but of social conventions. You can certainly say "いいえ、できません". While it is true that Japanese has a relatively substantial honorific structure when compared to English, there are other languages with even richer systems, such as Malay.
One thing that makes Japanese stand out amongst the ones you have listed is the fact that it is a language isolate, meaning there are no other languages that are "linguistically related". (While technically in the Japonic family, there is some controversy around the categorization of Japanese dialects, which muddies the family categorization as well). Korean is another isolate, and, as far as I know, the only language that is remotely like Japanese in terms of grammar and morphology. German, Dutch and Spanish all belong in the Indo-European family, and there are some obvious structural similarities. Mandarin Chinese is certainly not Indo-European, but it also lacks a non-trivial inflection system, and has a basic word order that is not unlike Germanic languages.
Japanese, on the other hand, does not share this trait. Consequently, there are certain grammar constructs in English that are essentially impossible to transform into Japanese; it takes a speaker sufficient in both to first completely comprehend an utterance and then re-express it. Add to this the many subtle semantic differences in words, and it's not surprising that you will get a lot of "that's hard to say"s. However, this doesn't mean that something is untranslatable. Speaking from experience, us polyglots do occasionally use it as an excuse that more precisely means "the exchange which must take place for me to extract more context out of you to effectively perform your request is something I don't want to spend energy on at the moment." ;)
I know the word order is vastly different and the honorific system obscures the language, but what about more fundamental tendencies of Mandarin such as the heavy dependency on chronology or the kind of general sounds of the phonemes. I know, for instance, that it can often take a second or two for me to differentiate Mandarin from Korean or Japanese when it being spoken quickly and loudly (such as when you first turn on a movie).
Structurally, Japanese and Korean is almost identical. The morphology system are similar, they have the same word order, are both tonal in a similar way, and there are even parallels that can be drawn in terms of their phonology; some words are even phonemically identical. However, their core inventory of phonemes is quite different: there are sounds in Korean that just do not occur in Japanese. Like Japanese, the categorization of Korean is debated; Wikipedia is helpful here if you want to know more.
Now, Mandarin Chinese is entirely different. The tonal system is different, it lacks an overt inflection system, word order is different, and the phoneme inventory is, again, different. While there is vocabulary sharing going on, it's hard to see any technical relationships between them.
I've heard a number of different stories on how the Korean peninsula and Japanese archipelago were settled by the Chinese, from the Chinese, but it just doesn't seem plausible considering the linguistics of the regions. What history I did study in Japan simply never mentioned it. And, to be honest, that is the extent of my historical knowledge of the languages.
Japanese on the other hand is considered more likely to be separate but high influenced due to word sharing from the many wars between China and Japan. this seems pretty plausible considering the sharing of the written language which is widely considered to be Chinese in origin.
It's all interesting because I wouldn't be surprised if a lot of the ancient history is obscured by the three proud nations.
While travelling and working in Sweden I picked up enough of the language (alas, now gone) to converse reasonably well over dinner with people I hadn't met before. They were fluent in English, my colleague was fluent in Swedish.
I asked about the word "varsågod." It seemed to have many translations, often different for different contexts, and I was wondering how they all perceived it. The consensus came only after about an hour of back and forth. There is no translation, even when the context is known.
The best I've come up with is "All is well," but that really, really doesn't cover it. Sometimes it means "You're welcome," sometimes it means "Here you are," and there are other contexts.
And the English translations don't carry the extra meanings, the baggage. It just feels untranslatable. No translation I've seen or heard carries all of the meanings and nuances.
Another example is from French. The phrase "Je vous en prie" is often translated as "You're welcome," but it's also very, very formal. You'd hear it from staff in hotels, perhaps, and perhaps in the very best restaurants.
But the point is that while it effectively means "You're welcome" it actually carries more information. It also says: and our relationship is a formal one, such as staff to employer, and I'm in the subordinate position.
There is no way to say that in English without spelling it out explicitly, and once you've done so you've lost the sense of the original anyway.
It's like explaining a joke. Once you've done so you've given the understanding needed, but lost the humor.
Similarly with so many things in translation. To carry all the meaning properly sometimes you have to explain or describe the meaning, and then it's no longer actually a translation.
The thing I find most interesting is just how many monoglots claim that this can't possibly be true and give many excellent reasons, while so many polyglots simply accept it as fact. My wife is fluent in French and works copy-editing translations from German (and other languages) into English. I've seen this problem "in action" as it were, and it's why good translators cost so much, while mediocre translators don't.
EDIT: corrected the Swedish word - thanks. My spoken and reading Swedish was always better than my written.
If you feel like procrastinating and want to have a laugh at how ambiguous and strange Swedish can be I recommend Mastering Swedish by slay radio: http://www.slayradio.org/mastering_swedish.php
It sounds as if you are saying the lack of a short code in a language makes a concept untranslatable.
But that doesn't seem right. In 1943, I couldn't translate a document about nuclear reactors into French or Spanish because the only languages with words for "nuclear reactor" were English and German. Does that make "nuclear reactor" untranslatable?
Of course, once the concept became relevant to French speakers, they borrowed or created a short coding.
From what you've said, you can translate "je vous en prie" to "you're welcome and I mean that with the respect a subordinate accords his superiors". That just doesn't have a short coding in English because the concept being encoded is rarely relevant to English speakers.
In programming terms, (map f lst) translates to c, it's just not as short.
Your point about "nuclear reactor" is well taken, but, ultimately, not really relevant. No amount of explanation makes a joke funny to someone who doesn't "get it." Further, many coordinate bilinguals never get this feeling, simply because they're coordinates. Compound bilinguals do get it, and find it impossible to explain.
And here I am, once again, trying to explain a joke to someone who doesn't get it.
(Please note - that last was an analogy. This isn't a joke, it is real that there are some terms that just don't translate, even when you expand your translation into an explanation. Some things need to be experienced, and cannot be appreciated from logic and reason alone.)
Would it be fair to say that the particular phrase, rather than conveying information, causes a particular emotional reaction in the hearer? I may not experience what you experience and some particular phrase might simply be a short code for that feeling.
But that doesn't mean the information cannot be conveyed. I can explain fairly easily what 'orange' means to a blind person: "Objects and parts of objects are associated to unique colors, which humans can distinguish via vision. [geometric optics skipped.] One particular color is orange."
They haven't experienced orange, in the sense that the "orange" set of neurons in their brain hasn't fired. But that doesn't mean they lack information when I say "object X is orange."
Actually, the study of semantics, which is highly logical, will often fail at describing the "true meaning" of an utterance. This is where pragmatics come in, and, frankly, a far more interesting area of study and occasionally even applicable in every day life.
The most useful thing I got out of my semantics course was that I never, ever, want to take that stream ever again.
Personally, I've experienced it, and perhaps the fact that I can't explain it to you, who haven't experienced it, is more convincing than the attempts at explanations.
If you want to make an argument on the effectiveness of communication, please use a quantitative or otherwise qualified argument, rather than exotic anecdotes.
The speed of human language isn't limited by the ability of "the language" to encode a stream of information. It is limited by the human ability to create and understand the bits of language information.
And that limit is much lower than the human ability to take in other streams of information. Vision lets you take in mega-bytes of information in seconds. Speech processing involves far less because language processing is such a hard problem for the brain.
Luria's the "The Working Brain" mentions that most brain damage degrades speech in some fashion and that's because such a large portion of your brain works on the speech recognition problem when you are speaking or listening.
And processing language is a hard problem for people (and computers!) because a language statement involves answering (at least implicitly) global questions about your store of information - "are all men are mortal", "do black swans exist" etc.
I've read that people with extraordinary memory or mental arithmetic ability often visualize shapes and perform operations on them to get at the answer. And I guess "normal" people use "subvocalization" instead. In light of your comment, that makes sense.
Processing of some human languages (like English)...
> "It turns out that the information content of handshapes is on average just 0.5 bits per handshape less than the theoretical maximum. By contrast, the information content per phoneme in spoken English is some 3 bits lower than the maximum."
Talking about the information content of speech symbols is likely to be entirely bunk, but I'm going to go read the full article and try to find out whether the summary is bad or if the research is really this confused.
(1) I'm pretty sure it was Klima and Bellungi's book The Signs of Language that pointed out that arm muscles are slower than vocal muscles, and therefore ASL does things that spoken languages can't do in order to maintain the same communication bandwidth. (That book was written ten years ago.)
(2) The handshape is only a very small part of what conveys meaning in sign language; one of the classic newbie mistakes in learning ASL is to look at your conversation partner's hands, rather than his or her eyes. A great deal of grammatical information is communicated purely by facial expression; for example, raised eyebrows can indicate a yes/no question, lowered eyebrows can indicate a wh-question, and just looking in one direction or shifting the body slightly can substitute for a pronoun. There are also movements of the mouth that act as adverbial modifiers for a sign, to indicate things like "almost", "carelessly", "with difficulty", "distant in time or space", and a whole bunch of other stuff.
(3) With regard to the hand and arm movements themselves, the location and movement of the signs are as significant as the handshape. The signs for "father" and "mother" differ only in location. The signs for "paper" and "cheese" differ only in movement. Skimming the article, it appears that the authors didn't bother taking location and movement because linguists disagree on how to categorize those other features. But that's no excuse for completely leaving them out of your analysis. That's methodological laziness.
(4) Modulation of movement also has grammatical significance which in English would be conveyed by modal verbs or adverbs. For example, a change in how you make the sign for "to be red" turns it into "to become red". The Klima and Bellungi book above has more of this kind of thing.
(5) There's also the ASL classifier system, which provides a concise way of using the relative position and motion of hands to indicate the relative position and motion of objects in physical or metaphorical space. I once saw a lecture at which a woman very eloquently used this to describe herself advancing through all four years of her college education while a friend of hers kept repeating her "prep" year. (Gallaudet has a pre-freshman year for students who, thanks to the ocean of suck that is the American deaf-ed system, don't arrive with adequate college preparation.)
There have been over thirty years of serious linguistic research into ASL, and judging from the references, these jokers didn't do more than strip-mine it for a list of handshapes. AAARRRGGGHH!
I believe the trend is inevitible.