Apparently there is no recession in the machine translation market. Systran just posted no fewer than 8 openings for computational linguists!! See here for 7 here for 1-2 senior positions.
Too bad Systran (and every other MT company) gave up hiring real linguists after the dot com bubble burst. But alas, lots of NLPers should be happy. Work! Work! Work!
Thursday, September 30, 2010
bastardizing a snowclone!
Andrew Sullivan used this as the title of a post recently: Palinites, Latinos, Tea Partiers, Women, Oh My!
I love the X, and Y, and Z, oh my! snowclone as much as the next guy, but the construction has to be respected. You can't just add a fourth member of the list all willy nilly! There are rules!!
I love the X, and Y, and Z, oh my! snowclone as much as the next guy, but the construction has to be respected. You can't just add a fourth member of the list all willy nilly! There are rules!!
Wednesday, September 29, 2010
Rankings?
The newest National Research Council PhD program ranking are out. I'm not quite willing to dish out the $125 for my own personal copy. Anybody care to look to see if they ranked linguistics departments?
Tuesday, September 28, 2010
in praise of William Yardley
the death of writing!
This is pure wild speculation: I can imagine writing becoming obsolete within maybe 200 years. My logic is thus:
Writing as a technology has only been with us for a small time (6000 years or so, compared to the 100,000 years or so of homo sapien evolution (ohhh, let's not get into how to date homo sapiens)), and it has only been utilized by a large number of people for an even smaller time (maybe 200 years or so, before that most people were illiterate (probably still are)). Hence, writing is an unnecessary and cumbersome luxury we can happily live without should something better come along.
Now imagine that the computational linguists finally get off their lazy arses and give me a computer I can frikkin talk to and can talk back to me*. If I can talk to my computer like a human being, poof goes the keyboard, right? I give a 90% chance of having this as a viable alternative within 50 years**.
Once I'm unburdened of the clunky inefficiency of a keyboard, and once I can preserve and share ideas without a writing system (think bloggingheads), why oh why would I bother with the ridiculous tedium of representing my words in an altogether unnecessary form?
But, you say, writing provides the best way to preserve and share ideas*** because it lets us organize and review and get all meta. No. It does none of those things. We do that. We're just stuck with this third party representation in which to do those things.
But how would academics write papers and get tenure? Good question. First, tenure will die long before writing systems (I give it maybe 75 years). Second, imagine that instead of writing a paper, I can create a virtual me, encode it with a set of arguments about a topic then instead of reading my paper, you engage in a Socratic give-and-take with this virtual me on the topic. Call it iSocrates****
Thus endeth the prophesy.
*I want everything Kirk had. I got the cell phone. The Terrapins are working on a transporter, now give me a frikkin computer I can talk to!
**I just made those numbers up. People seem to like fake numbers, so okay, there you are.
***You didn't really say that. I made that up. This is what's known as a straw man argument. It makes this kind of bullshitting easier.
****Dear gawd that's a horrible name. Let's hope this whole iXXX trend goes the way of eXXX, eXtreme, XXXtech, XXXsoft, etc.
Writing as a technology has only been with us for a small time (6000 years or so, compared to the 100,000 years or so of homo sapien evolution (ohhh, let's not get into how to date homo sapiens)), and it has only been utilized by a large number of people for an even smaller time (maybe 200 years or so, before that most people were illiterate (probably still are)). Hence, writing is an unnecessary and cumbersome luxury we can happily live without should something better come along.
Now imagine that the computational linguists finally get off their lazy arses and give me a computer I can frikkin talk to and can talk back to me*. If I can talk to my computer like a human being, poof goes the keyboard, right? I give a 90% chance of having this as a viable alternative within 50 years**.
Once I'm unburdened of the clunky inefficiency of a keyboard, and once I can preserve and share ideas without a writing system (think bloggingheads), why oh why would I bother with the ridiculous tedium of representing my words in an altogether unnecessary form?
But, you say, writing provides the best way to preserve and share ideas*** because it lets us organize and review and get all meta. No. It does none of those things. We do that. We're just stuck with this third party representation in which to do those things.
But how would academics write papers and get tenure? Good question. First, tenure will die long before writing systems (I give it maybe 75 years). Second, imagine that instead of writing a paper, I can create a virtual me, encode it with a set of arguments about a topic then instead of reading my paper, you engage in a Socratic give-and-take with this virtual me on the topic. Call it iSocrates****
Thus endeth the prophesy.
*I want everything Kirk had. I got the cell phone. The Terrapins are working on a transporter, now give me a frikkin computer I can talk to!
**I just made those numbers up. People seem to like fake numbers, so okay, there you are.
***You didn't really say that. I made that up. This is what's known as a straw man argument. It makes this kind of bullshitting easier.
****Dear gawd that's a horrible name. Let's hope this whole iXXX trend goes the way of eXXX, eXtreme, XXXtech, XXXsoft, etc.
Monday, September 27, 2010
can language affect blood flow?
Do languages affect blood flow in the brain differently? Apparently, yes! In a recent fMRI study, researchers showed that Cantonese verbs and nouns are processed in (slightly) different parts of the brain than English nouns and verbs in bilinguals. The researchers used a lexical decision task to contrast the processing of English and Cantonese verbs and nouns in the brains of bilingual speakers.
Chinese nouns and verbs showed a largely overlapping pattern of cortical activity. In contrast, English verbs activated more brain regions compared to English nouns. Specifically, the processing of English verbs evoked stronger activities of left putamen, left fusiform gyrus, cerebellum, right cuneus, right middle occipital areas, and supplementary motor area. The cognition of English nouns did not evoke stronger activities in any cortical regions.
This is truly language affecting thought, no? The point of general interest to linguist is that bilingual speakers seem to process words in their two languages differently. Cantonese words are processed using diffuse brain regions and English words are processed using localized regions (this is a simplified explanation of course).
Now, I have to admit that this is not my specialty so I am not familiar with the background literature. However, as interesting as this is, I must say I have some serious questions about their methodology and underlying assumptions. I
First, they use orthography as their base for determining the "similarity" and "complexity" of languages. That is, if two languages use an alphabet, they are considered similar. While they give some passing references to other linguist measures, ultimately it is orthography that they use to compare "complexity" of stimuli (their word, not mine). So, they compared the mean number of strokes in a Chinese character with the number of letters in an English word to determine which was "more complex" than the other. I found this weird.
Then they made an assumption that Cantonese words are more ambiguous with respect to parts of speech. I do not klnow if this is true, but it certainly is true that English has plenty of POS ambiguity (just ask Eric Brill), so it's not obvious to me that this is a fair assumption. Furthermore, they provider no evidence for this. Unfortunately, they do not publish their actual sets of stimuli, so it's not possible (this morning while googling around) to look at which words they actually use, but I suspect there's plenty of ambiguity to be found in the English words.
Based on earlier work, they conjecture that morphological simplicity leads the brain to distribute where words are processed in the brain:
...a recent fMRI study examining monolingual Chinese adults in our own laboratory indicated that Chinese nouns and verbs activate a wide range of overlapping brain areas (without a significantly different network) than those reported in the English studies cited above (Li et al., 2004). Relatively fewer distinctive grammatical features of nouns and verbs at the lexical level are likely to be responsible for this finding, but the question may be addressed more directly by employing bilingual individuals.
And the corollary should be true: the fact that English has tense and number markings means English verbs and nouns are processed ion more isolated parts of the brain. This is my wording of their conjecture. I may be oversimplifying just a bit, but I'm trying to wrap my head around the underlying claim. It's not clear to me why this would be true.
Next (and this may be a bit nit-picky), they judged the level of bilingual proficiency using a self-assessment questionnaire. Call me a cynic, but I just don't trust people's perceptions of their own language skills.
Then, the researches used frequency data from really dated sources including Francis and Kuceras 1982. I love F&K as much as the next guy, but in the age of the BNC, Davies's freely available 400 million word COCA, and the redonkulous Web 1T corpus of 1 trillion words (yes, 1 Trillion!), I see no reason to use resources so old.
Their basic conclusions are a tad confusing too. They never clearly explained the connection between bilingualism and morphological complexity, imho. The interplay is complicated and requires thorough discussion, which they simply did not provide. When I used to teach writing to college freshmen, I always told them that their job when writing a paper was to make my job as a reader easy. Explain things clearly so I don't have to work too hard to figure out what you mean. These authors failed to make my job easy. I had to figure things out too much for myself.
Ultimately, they found something interesting, I'm just not sure what it means and without more thorough linguistic vetting of their underlying assumptions, their results remain a head scratcher.
Chan, A., Luke, K.K., Li, G., Li, P., Weekes, B., Yip, V., & Tan, L.H. (2008). Neural correlates of nouns and verbs in early bilinguals. Annals of the New York Academy of Sciences, 1145, 30–40. (pdf)
Chan, A., Luke, K., Li, P., Yip, V., Li, G., Weekes, B., & Tan, L. (2008). Neural Correlates of Nouns and Verbs in Early Bilinguals Annals of the New York Academy of Sciences, 1145 (1), 30-40 DOI: 10.1196/annals.1416.000
Chinese nouns and verbs showed a largely overlapping pattern of cortical activity. In contrast, English verbs activated more brain regions compared to English nouns. Specifically, the processing of English verbs evoked stronger activities of left putamen, left fusiform gyrus, cerebellum, right cuneus, right middle occipital areas, and supplementary motor area. The cognition of English nouns did not evoke stronger activities in any cortical regions.
This is truly language affecting thought, no? The point of general interest to linguist is that bilingual speakers seem to process words in their two languages differently. Cantonese words are processed using diffuse brain regions and English words are processed using localized regions (this is a simplified explanation of course).
Now, I have to admit that this is not my specialty so I am not familiar with the background literature. However, as interesting as this is, I must say I have some serious questions about their methodology and underlying assumptions. I
First, they use orthography as their base for determining the "similarity" and "complexity" of languages. That is, if two languages use an alphabet, they are considered similar. While they give some passing references to other linguist measures, ultimately it is orthography that they use to compare "complexity" of stimuli (their word, not mine). So, they compared the mean number of strokes in a Chinese character with the number of letters in an English word to determine which was "more complex" than the other. I found this weird.
Then they made an assumption that Cantonese words are more ambiguous with respect to parts of speech. I do not klnow if this is true, but it certainly is true that English has plenty of POS ambiguity (just ask Eric Brill), so it's not obvious to me that this is a fair assumption. Furthermore, they provider no evidence for this. Unfortunately, they do not publish their actual sets of stimuli, so it's not possible (this morning while googling around) to look at which words they actually use, but I suspect there's plenty of ambiguity to be found in the English words.
Based on earlier work, they conjecture that morphological simplicity leads the brain to distribute where words are processed in the brain:
...a recent fMRI study examining monolingual Chinese adults in our own laboratory indicated that Chinese nouns and verbs activate a wide range of overlapping brain areas (without a significantly different network) than those reported in the English studies cited above (Li et al., 2004). Relatively fewer distinctive grammatical features of nouns and verbs at the lexical level are likely to be responsible for this finding, but the question may be addressed more directly by employing bilingual individuals.
And the corollary should be true: the fact that English has tense and number markings means English verbs and nouns are processed ion more isolated parts of the brain. This is my wording of their conjecture. I may be oversimplifying just a bit, but I'm trying to wrap my head around the underlying claim. It's not clear to me why this would be true.
Next (and this may be a bit nit-picky), they judged the level of bilingual proficiency using a self-assessment questionnaire. Call me a cynic, but I just don't trust people's perceptions of their own language skills.
Then, the researches used frequency data from really dated sources including Francis and Kuceras 1982. I love F&K as much as the next guy, but in the age of the BNC, Davies's freely available 400 million word COCA, and the redonkulous Web 1T corpus of 1 trillion words (yes, 1 Trillion!), I see no reason to use resources so old.
Their basic conclusions are a tad confusing too. They never clearly explained the connection between bilingualism and morphological complexity, imho. The interplay is complicated and requires thorough discussion, which they simply did not provide. When I used to teach writing to college freshmen, I always told them that their job when writing a paper was to make my job as a reader easy. Explain things clearly so I don't have to work too hard to figure out what you mean. These authors failed to make my job easy. I had to figure things out too much for myself.
Ultimately, they found something interesting, I'm just not sure what it means and without more thorough linguistic vetting of their underlying assumptions, their results remain a head scratcher.
Chan, A., Luke, K.K., Li, G., Li, P., Weekes, B., Yip, V., & Tan, L.H. (2008). Neural correlates of nouns and verbs in early bilinguals. Annals of the New York Academy of Sciences, 1145, 30–40. (pdf)
Chan, A., Luke, K., Li, P., Yip, V., Li, G., Weekes, B., & Tan, L. (2008). Neural Correlates of Nouns and Verbs in Early Bilinguals Annals of the New York Academy of Sciences, 1145 (1), 30-40 DOI: 10.1196/annals.1416.000
Sunday, September 26, 2010
pullum bait
Jeremy Porter decided to adapt Strunk and White's infamous Elements of Style for Tweeting here. C'mon Geoffrey, you know you wanna respond..I dare ya...
Subscribe to:
Posts (Atom)