The Puzzling Paradox of Sign Language(technologyreview.com)
technologyreview.com
The Puzzling Paradox of Sign Language
http://www.technologyreview.com/blog/arxiv/24521/
7 comments
Good comment. That said...
> '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.
> '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.
> 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.
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.
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 not sufficiently familiar with Japanese, but I can relate two specific stories from my own experience.
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.
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.
I think you meant to type the composite word "varsågod" (and not "Be so god" ^^), it's not a phrase. A similar problem exist with "lagom" which roughly means something like: "enough, not to much or to little.", with a strong positive meaning. It also goes the other way too as computer engineer doesn't have a good exact translation as the English "engineer" is more nuanced than the Swedish "Ingengör".
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
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
Thanks - corrected. My written Swedish was always lousy.
Another example is from French. The phrase "Je vous en prie"...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.
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.
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.
But largely "Je vous en prie" does not translate as "you're welcome and I mean that with the respect a subordinate accords his superiors". There's more, and I really, really can't explain it, because I simply don't have the words. It's not just that there's no short phrase - the fact that it's a short phrase does itself carry information.
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.)
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.)
I'm inherently suspicious when someone says something like "conveys information", but then says "logic and reason" don't work.
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."
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."
> I'm inherently suspicious when someone says something like "conveys information", but then says "logic and reason" don't work.
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.
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.
Well, I can't say much more really. Natural language semantics are tough, and I've clearly failed to communicate to you the idea that some phrases in other languages can't really be translated. They can sort-of be explained, sometimes, but even then I remain unconvinced that the true sense of the original can always be retained.
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.
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.
> 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.
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.
Right. I think I would file all of that under the "hard to translate" category.
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." ;)
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'm curious about the origins of language in East Asia then. I was always under the impression that Chinese dialects heavily influenced both Korean and Japanese (and visa versa) so I'm a little confused by considering both of those to be isolates.
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).
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).
(Disclaimer: I've not studied any of these languages in detail; the following is a mix of what I have studied, personal knowledge, and logical conclusions derived from both)
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.
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.
The story I hear (again from the Chinese) is that Korean is a dialect/divergence from Chinese that lost much of the tonality. I've not studied it personally, but from my knowledge it wouldn't be surprising for a Chinese speaker to believe Korean (and Japanese) lack tonal structure.
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.
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.
You're mincing words to tell us about Japanese. Your examples seem to indicate that Japanese more naturally conveys social situations and basic descriptions at the same time, making it more effective, not less. Russian has no future first person singular conjugation for 'to win'- this doesn't make Russian 'less effective' at communication, nor does English's use of "I will be able to" in place of "I will can".
If you want to make an argument on the effectiveness of communication, please use a quantitative or otherwise qualified argument, rather than exotic anecdotes.
If you want to make an argument on the effectiveness of communication, please use a quantitative or otherwise qualified argument, rather than exotic anecdotes.
It would be interesting to discover that some languages are less efficient than others, and what effect (if any) it has on the economy. I think Japanese probably is more complicated that English - so why does it not appear to slow them down? Or maybe it slows them down in some respect (maybe having to sacrifice a lot of childrens play time to rote learning?), and they compensate in other ways?
I don't know about Japanese, but in Chinese, all the numbers have words that can be said extremely quickly - si, qi, etc. Because our working memory is implemented as a 2 second loop of audio, this allows Chinese people to have a working memory that's closer to 10 items instead of the famous 'seven plus or minus two' that English speakers have. In 'Outliers' Malcolm Gladwell posits this as part of the reason why asians often seem to be better at math.
Spoken Japanese may be less compressed than English, but written Japanese is considerably more terse.
I did a ten-week course in BSL (passed with full marks!) and your assessment matches up with my experience, which was that it was like exaggerated enthusiastic telling of a story. It's like miming, too. Hearing people talk with their hands, bodies, faces, reactions, energy, rhythm, speed, etc. all the time already. They get along fine in clubs with loud music thumping away. Just watch someone chatting on the phone, they still make a tonne of physical gestures. And we've seen that textual interpersonal communication (i.e. pure English syntax and grammar) is hard, it's great for technical details or (with some effort) describing stories but sucks at that.
This is always what I understood, you have words for letters but you only really use them when you need to spell something out otherwise you use other words. I have never learned any sign language but I assumed it worked in the same way, or else there would be no advantage to signing over writing what you want to say down.
"This is only interesting if one doesn't realize that sign language is its own language, rather than simply signed English."
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.
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.
>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)
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)
"I file Technology Review with Wired and SciAm. That is, fluff. Hardly intellectual material imho."
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.
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.
>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.
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 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 just disagree. I have some familiarity with ASL and I feel they've gone off in the wrong direction from the start."
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.
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.
One obvious point that somehow isn't mentioned in the article:
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.
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.
"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."
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.
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.
> And processing language is a hard problem for people (and computers!)
Processing of some human languages (like English)...
Processing of some human languages (like English)...
Calling asl "signed english" is usually a good indication that the writer doesn't know much about the subject. Then they go on to say:
> "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.
> "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.
Just read the full article. It's no good. Analyzing ASL in terms of hand shapes only, on the grounds that they're the only feature that all linguists looking at ASL seem to agree on is really problematic. Computation of entropy is heavily dependent on the alphabet. If you say english has an alphabet at the syllable level, you'll get a very different amount of entropy than if you say the alphabet is phoneme level. If you go a level below phonemes and look at distinctive features (voiced/unvoiced, aspriated/unaspriated, front/back, high/low, ...) you'll get yet another entropy number altogether. These numbers will be wildly different from each other. Entropy calculations between languages must use comparably derived alphabets which cover the whole communication channel, or they will be meaningless. Argh. Junky article.
How would you calculate the theoretical maximum for a handsign? Or for spoken language for that matter.
I too find information theory a fascinating subject. Several commentors are disappointed by the trivial nature of the study (which I assume none of us have read). That is a problem with trying to apply theory to real world problems, You often have to make compromises in the quality of your interpretation of the problem in order to rigorously apply the tools of theory. I'm guessing the researches picked up some understanding of ASL during their research (if they didn't have some to begin with), but chose to frame their study so the parameters were easier to quantify.
There's a very nice sign in ASL to describe this article. It uses two hands, and one of them is a fist with the index and little fingers extended, like the horns of a bull.
(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!
(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!
Along similar lines as the article, I've always thought about how we can be more efficient when speaking. If you notice, there is a tendency to be more and more efficient on the web, with abbreviations/acronyms. Imagine a world where we say "lol" just as though we type it (not hard to imagine). Now take it a few steps further, by considering the vast amount of different sounds/tones our vocal chords can create, imagine if we keep simplifying the spoken language to a point where it becomes like one of those star-trek civilizations with their clicking sounds when they speak.
I believe the trend is inevitible.
I believe the trend is inevitible.
Based on how I, like, hear many people, like, talking, I'm, like, not so sure it's, like, inevitable.
Like, such as, like, Miss Teen USA South Carolina, such as.
http://www.youtube.com/watch?v=lj3iNxZ8Dww
http://www.youtube.com/watch?v=lj3iNxZ8Dww
What did they study? The accompanying photograph shows the alphabet, but I assume they aren't just considering using sign language to spell out words. Then they mention phonemes, but those letters aren't phonemes. Are there signs for phonemes? And ASL has signs for words.
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?'