Friends of Semantic Compositions

July 2008

Sun Mon Tue Wed Thu Fri Sat
    1 2 3 4 5
6 7 8 9 10 11 12
13 14 15 16 17 18 19
20 21 22 23 24 25 26
27 28 29 30 31    

Site Statistics

Blog powered by TypePad

April 20, 2007

No hockey for you, s'il vous plais

Via ESPN's Page 2, not normally a source of stories described as poignant or controversial, comes the incredible news that an 8-year-old ice hockey player in Quebec has been banned from his league because he doesn't speak French. SC has written about Quebec's outrageous Bill 101 in the past, but this is a new low in language politics. Or is it?

About the only fact of the case that is not in dispute is that Jared Murray, the player in question, is in fact 8 years old. This matters because, as Gare Joyce summarizes the decision to ban him:

1. He's too big to play with other 8-year-olds. At 4-foot-9 and 110 pounds, he's at least a head taller than the other kids on the ice.

2. He's too good to keep games fair and competitive.

It's far from unprecedented for kids to have their ages questioned if they seem a little too big or too talented; witness the cases of Danny Almonte or the early-90s scandals that cost the Dominican and Philippine Little League World Series teams. In this case, however, nobody is disputing his age. As for the second issue, whether or not he's simply too talented and ought to be playing in a higher league, young Mr. Murray scored 51 goals in 17 games, which is admittedly a very impressive number. However, as this hockey blog points out, Wayne Gretzky scored 6x as many goals at the same level of competition without getting bounced out of the league. If you consider Gretzky too extreme an example to be relevant, consider that the allegedly too-dominant Jared Murray only managed to bring his team to a third-place record in a 16-team league.

Let's stipulate for the sake of argument, though, that these issues at least ought to have been debated by the adults in charge of the hockey league. Canadian novice hockey has two categories of play, with A and B-level teams so that boys of roughly equal talent level can face off against each other. Mario Lemery, the president of Hockey Outaouais (the governing body for the league in question), claims that the team was asked to move to the A level:

"We asked if the complete team would go up to the A," Mr. Lemery said. "They didn't want to. It's no fun if they cream everyone else, 7-1, 8-1. They should play in A. No one wants to play against those teams."

He said Fort Coulonge agreed to move up. After the Shawville team wouldn't move up, he offered to let the rest of the team play in the B division as long as their star player didn't play. That's when the whole team decided to boycott.

"At the beginning of the year, they said we should play in the B because we're not strong. They were not honest at the beginning of the season," Mr. Lemery said. "I asked them to go A and they said no.

"It's no fun squashing every team. No one wants to play the team."

In other forums, though, the team's coach claims that this request was never made (see this CTV.ca story, and also this comment at the HockeyAnalysis.com blog by an individual claiming to be the coach).

Since nobody from either side is making an admission against their interest (i.e., the hockey team acknowledging getting a request, or the league acknowledging it wasn't communicated in English), it's hard to be sure of precisely if or when Jared Murray's eligibility was first questioned. Is it possible that the request was made through a letter sent only in French? It's not an unreasonable hypothesis. A look at the official Hockey Outaouais website indicates not the slightest concession to English-speaking members of the league; not one page is available in anything other than French. This makes it easy to credit the claim of Gare Joyce's original ESPN.com article that "[N]ot one of Shawville's volunteer officials went to league meetings where only French was spoken (emphasis added), which basically shut out the town where few residents are bilingual."

Ordinarily, when a language majority appears to be running roughshod over language minorities, the debate revolves around whether there should be an outreach effort by the majority (i.e., creating bilingual documents, providing translators) or the minority (i.e., working to become bilingual in the majority language). This scenario breaks the mold because while the majority/minority roles are obvious in context, the French speakers are also in the minority role in the larger context of Canada. Revisiting the HockeyAnalysis.com discussion, a number of comments from disgruntled English-speaking Shawville residents suggest that they feel Hockey Outaouais has a pattern of timing decisions like this to make it difficult for Shawville players to compete. It's possible that some, maybe even all, of the decisions made by the league's officials, are correct on the merits. But it's also possible to imagine sitting down over drinks with Hockey Outaouais officials, and hearing them say (in French, of course), "Payback's a bitch". SC certainly won't claim to be able to solve the historical grievances that underlie Canadian language politics. Would it be too much to ask for them not to be inflicted on 8-year-olds, though?

January 24, 2007

Theodore Dalrymple does complexity

Last LSA-related post, and one that had to sit for a week due to work intruding on stuff that matters again. (Note to employers: That's a joke. Obviously, I put work first, or this post wouldn't be so late.)

A recent note from the Linguist List drew SC's attention to this article by Theodore Dalrymple titled "The Gift of Language". Dr. Dalrymple (a retired M.D. writing under a pen name; his real name is Anthony Daniels) is one of the gloomiest human beings alive, as anyone who has ever read his ruminations on his work in the English prison system knows. That said, he's also a sharp observer of human nature. The article he has written in this case ties in quite nicely with the last LSA symposium your host wishes to write about, which was titled "Approaches to Language Complexity".

Dr. Dalrymple opens by noting that many of the prisoners he dealt with were, shall we say, communicatively challenged:

For the most part, though, I was struck not by the verbal felicity and invention of my patients and those around them but by their inability to express themselves with anything like facility: and this after 11 years of compulsory education, or (more accurately) attendance at school.

With a very limited vocabulary, it is impossible to make, or at least to express, important distinctions and to examine any question with conceptual care. My patients often had no words to describe what they were feeling, except in the crudest possible way, with expostulations, exclamations, and physical displays of emotion. Often, by guesswork and my experience of other patients, I could put things into words for them, words that they grasped at eagerly. Everything was on the tip of their tongue, rarely or never reaching the stage of expression out loud. They struggled even to describe in a consecutive and logical fashion what had happened to them, at least without a great deal of prompting. Complex narrative and most abstractions were closed to them.

He then goes on to observe that in many cases, conflicts between the inmates and their jailers that dragged on for months could be quickly resolved when he stepped in, which he attributes not merely to superior command of language, but to the sociological facts implied by it.

Dr. Dalrymple then proceeds from these observations to a theoretically muddled, yet intuitively reasonable, series of ruminations about the ideology of language complexity. It might have been quite entertaining to have him in the audience asking questions at the aforementioned LSA panel, as he asks many of the right sorts of questions (well, "right" from SC's perspective). Although he grounds much of his discussion in an engagement with Steven Pinker's book, The Language Instinct, we'll step back from that to lay out what he alleges are the "orthdoxies" of linguistics that don't accord well with his experience:

  • [E]very child, save the severely brain-damaged and those with very rare genetic defects, learns his or her native language with perfect facility, adequate to his needs
  • [T]he faculty of language is part of human nature, inscribed in man’s physical being, as it were, and almost independent of environment
  • [L]anguage itself is always rule-governed; and the rules that govern it are universally the same, when stripped of certain minor incidentals and contingencies that superficially appear important but in reality are not
  • [N]o language or dialect is superior to any other and...modes of verbal communication cannot be ranked according to complexity, expressiveness, or any other virtue
  • [T]he refusal to teach formal grammar is both in accord with a correct understanding of the nature of language and is politically generous, inasmuch as it confers equal status on all forms of speech and therefore upon all speakers

There is much here that might occasion protest from linguists about nuance: few people would seriously defend the position that all non-brain-damaged individuals are equally skilled users of their languages, which is a different point from claiming that they attain implicit awareness of the phonology, syntax and so forth. Nobody is claiming that someone who scores a 300 on the verbal SAT is as gifted a language-user as someone else who scores an 800 (to take a crude proxy for facility with English) -- but they are claiming, rightly so, that Mr. 300 is as capable of recognizing what sentences produced by other people are grammatical within their language, what words are phonologically possible (even if they don't recognize the words themselves). Ms. 800 may know more of those words, but knowledge of language from the perspective of a linguist is quite independent of the size of your lexicon, or your ability to write well.

As for the language faculty being tied to one's environment, surely Dr. Dalrymple does not mean that the chair he sat on while he wrote his essay played an integral role in its composition ([Oh yeah? YOU try writing 3,870 words standing up -- ed.]). More seriously, his comment about environment suggests a lack of awareness of the extensive debate over whether or not the "poverty of the stimulus" is a valid hypothesis (hint: it's not); SC guesses that most linguists would vehemently deny holding any idea along the lines of "environment doesn't matter".

However, he's hit on a basic truth about the squeamishness of linguists in discussing language complexity, and at long last, that's where the tie to the LSA panel comes in (for abstracts, see page 63 of the PDF here). All of the tenets sketched above reflect a worldview which was genuinely present at the discussion, which might be summed up as "Linguists won't willingly be tools of opression". Perceptions that one language is somehow less complex than another might readily be used to sign off on notions of racial supremacy, grounded in the idea that complexity is a proxy for intelligence. Of course, this would require some further assumption that one's native language is inexorably tied to race or location of birth, which is patently absurd, but that's an assumption that nonspecialists have repeatedly demonstrated they're all too willing to make.

What the panel discussion showed is that even highly trained professional linguists have no idea what it means to say that "one language is more complex than another". This is not to say that they didn't do interesting work -- far from it. However, no two talks even covered the question of complexity at the same level of linguistic representation, which underscores the danger of trying to discuss these questions with someone who doesn't have an understanding of basic linguistics -- who even knows that there are multiple levels of language representation? How can you tell someone who just wants a binary yes/no answer to the question of relative complexity that Chinese is more phonologically complex than Spanish (judged by number of tones), that Spanish is more morphologically complex than Chinese (judged by verb derivations), and that nobody has any way of saying systematically that one of these languages is more complex than the other overall? We'll briefly discuss each presentation from the panel just to give a sense of how complicated it all was; SC is working only from his own notes, having misplaced his handouts:

Ian Maddieson kicked things off with a discussion of the complexity of syllable shapes (i.e., consonant-vowel patterns), and how that compared to sound inventories (mostly consonants) in the same languages. Working from a hypothesis which he attributed to Matisoff 1973 (but he didn't say which 1973!), he expected that more complicated syllable shapes should correlate inversely with consonant inventories -- Matisoff's hypothesis apparently being that complexity in one area is compensated for by simplicity in another area. Using a 583-language sample from the UPSID database, he looked at 3 levels of syllable complexity, 4 levels of tone, and the aforementioned consonant inventories. Surprisingly, he found that bigger consonant inventories correlated positively with both syllable and tone complexity, and no relationship between either consonant and vowel inventories or syllable structure and vowel inventories. Syllable complexity and tones showed a slight negative correlation (i.e., less tonal complexity in languages with more complex syllables), but overall, his conclusion was that there is no evidence of "balancing" across the various inventories of phonological material. Aware that this might be seen as evidence for absolute statements about overall complexity, he hastened to propose the possibility that these measures are too simple, and that perhaps some measurement of the inherent complexity of the specific sound segments is in order (but how would one judge whether or not a language with 5 "complex consonants" is more complex than one with 20 "simple consonants", all other factors being equal?).

Johanna Nichols did another database-driven study, starting off with an attempt to define what was even meant by complexity. Specifically, she proposed a heuristic equation: complexity = # of discrete elements + # of possible intersections of elements + # of constraints on those intersections. This is a problematic definition -- aside from assuming that all 3 of these numbers are really of the same dimension, and can be appropriately summed, it backs away from the central insight of Optimality Theory, namely that some constraints are more important than others. Granting that all constraints are (in principle) equally available to all languages, it's the ranking of constraints that produces the apparent complexity of surface forms. Either that third term in her equation is null, or the number of constraints isn't as important as finding out which ones are ranked the highest.

Leaving aside that latter argument, she looked at 7 attributes of morphosyntax in the AUTOTYP database, in a sample containing 124 languages. Having misplaced his handout for the talk, SC can't say what they were, but she had produced a classification of languages into broad categories of low, medium, and high complexity based on the number of attributes each language possessed, and found very solid clustering in the middle category. This doesn't quite demonstrate that all languages are equally complex, but suggests that there's a certain range which is typical (which still might make those speakers of the languages with the magic 7 our linguistic Ubermenschen). She also examined -- and your host's notes say this wasn't in the handout, so he's off the hook -- how her complexity categories matched up with cultural factors, such as the presence of a pre-20th century writing tradition, the sorts of documents produced, and the number of speakers. That last variable probably can't be reasonably untangled from historical contingencies, but controlling for a variety of continental factors (not in SC's notes), she didn't find evidence that complexity correlated in any manner with population size.

Cristophe Coupe spoke on behalf of a research team consisting of himself, Francois Pellegrino, and Egidio Marsico, on an interesting attempt to look at issues of phonetic complexity cross-linguistically. By using a text translated into several different languages, and having native speakers read it aloud, they could use information theory to estimate whether or not the rate of speech production was in any way influenced by the complexity of the language being spoken. In order to do so, they needed to pick a level of language representation that is more or less independent from unit to unit, which rules out anything dependent on word order. So they went with rates of syllable production, a less complex language presumably being able to be spoken faster. A syllable's information value was computed by its probability of occurrence, but SC doesn't recall what corpora the probabilities were calculated from. They found that the production rate of syllables varied by language, but not so much by speaker -- perhaps this shouldn't be a surprise, as languages invariably come from cultural contexts where faster or slower speech may be the norm (try listening to Spanish news broadcasts in Madrid and Los Angeles if you doubt that culture matters here; SC is much more comfortable trying to follow the news in Madrid). More to the point, they found that languages with more informative syllables (more informative roughly equating with less probable) were produced more slowly, suggesting that at least one measure of complexity appears to demonstrate the compensation effect that Ian Maddieson's talk undermined.

Sheri Wells-Jensen gave a fascinating talk based on her doctoral dissertation, featuring the world's greatest study protocol. Understated in her abstract as "study participants narrat[ing] a fast-paced silent film", she had speakers of English, Hindi, Japanese, Spanish and Turkish narrate the last thirteen minutes of The Cat in the Hat. Not the insipid Mike Myers gross-out flick, but the inspired cartoon version voiced by Allan Sherman, Daws Butler, and Thurl Ravenscroft. After collecting the data, with the help of native-speaker consultants, she compiled a database of 1,300 speech errors from the narrations. SC weeps at the certain knowledge he'll never run a study this clever.

Again quoting the abstract:

The data were used to examine two interrelated hypotheses about the relationship between language structure and the speech production system: Hypothesis A "As measured in this way, languages are equally complex;" and Hypothesis B "The patterns of distribution of different types of errors will be distinct from one language to another." Both of the hypotheses were
supported.

It turned out that the overall error rates were similar across languages, but the sorts of errors made were very different; while English speakers made more pronunciation gaffes than Japanese speakers, Japanese speakers made more lexical errors (to give an example, and not necessarily a correct one sans handout). This tends to support the "all languages are equally complex" idea, but one might conceivably spin it to say, "ah, anyone could mispronounce a word when speaking quickly, but look how perfect my language is otherwise".

SC hates to give short shrift to Douglas Whalen's brain-imaging talk, but truthfully, the experimental protocol bothered him (SC, not Dr. Whalen). Dr. Whalen and his team attempted to correlate syllable complexity with greater brain activation, particularly in language-related areas of the brain (as opposed to merely hearing-related areas). This involved sitting people inside an MRI machine with headphones, and playing them nonsense syllables (versus control examples of music notes), while recording their brain activity. By Dr. Whalen's own account, the big challenge was keeping his participants awake. Some significant correlations were found between speech sounds and levels of brain activity, but SC won't pretend to apply the results to this discussion, because he just isn't convinced that a linguistically relevant phenomenon was being examined.

As even this cursory review makes clear, it's very hard to figure out what is even meant by complexity, or how to measure it. Any sort of professional pronunciation on "what the most complex language is" thus is ripe for abuse, and this places Theodore Dalrymple's "orthodoxies of linguistics" in a somewhat more reasonable light. Assuming that language learners as human beings are roughly equal, it would do them a grave disservice to take the equivocal scientific results of today and start labeling their languages with terms that will be used to make primarily political and economic judgments.

But does that make Theodore Dalrymple's conclusions wrong? Should Bill Cosby just concede that you can be a doctor with "all that...coming out of your mouth"? In SC's view, this is where the linguistics community would benefit from finding common ground with its critics. It is no imperialist who writes:

I am not of the ungenerous and empirically mistaken party that writes off [people unable to speak standard English] as inherently incapable of anything better or as already having achieved so much that it is unnecessary to demand anything else of them, on the grounds that they naturally have more in common with Shakespeare than with speechless animal creation. Nor, of course, would I want everyone to speak all the time in Johnsonian or Gibbonian periods. Not only would it be intolerably tedious, but much linguistic wealth would vanish. But everyone ought to have the opportunity to transcend the limitations of his linguistic environment, if it is a restricted one—which means that he ought to meet a few schoolmarms in his childhood. Everyone, save the handicapped, learns to run without being taught; but no child runs 100 yards in nine seconds, or even 15 seconds, without training. It is fatuous to expect that the most complex of human faculties, language, requires no special training to develop it to its highest possible power.

That's the end of Theodore Dalrymple's essay, and it strikes SC as being as sound a middle ground as one could ever hope to find. In the rush to make sure that nobody is unfairly stigmatized for speaking a "simple" language, whether a nonstandard dialect of English or the tongue of 1,000 villagers in a Borneo jungle, linguists are ending up tarred with the accusation of a hopelessly patronizing cultural relativism (think the SIL debate, with economics sitting in for religion). It's a view that the profession has reached with the best of intentions, but for the sake of being able to break people out of poverty, maybe it wouldn't hurt to advocate a little more forcefully that everyone ought to meet a few schoolmarms.

January 11, 2007

SIL: We're on a mission from Dallas

A couple of linguistics bloggers have mentioned the panel on missionaries and linguistics which took place at the LSA 2007 meeting. Your host attended this discussion (except for the talk by Courtney Handman, which he skipped in order to go hear Claire Bowern's talk on Nyikina), and took some notes. While The Tensor already made the most comment one most hopes to hear in contexts like this -- no chairs were thrown in the making of this production -- SC wishes to discuss his view of what went on at the panel, as he found it mildly dispiriting for many of the usual reasons that underlie his general refusal to discuss religion or politics at work.

As many of the participants found it a matter of good faith in argumentation to briefly disclose their own views on the ontological question undergirding the discussion, your host will do likewise. Longtime readers know SC makes no secret of being a practicing Conservative Jew, although he has generally restricted my posts in that regard to issues of language, cultural practices, or the desperate longing for leavened bread that generally strikes SC shortly after he stops eating in preparation for the first night of Passover. Your host is under no illusions about the views of most visitors to this site, but it's not the purpose of what goes on here, and as long as nobody wants to throw punches over things unsaid, we're all happy. Readers who wish to inquire further along these lines are welcome to do so by e-mail, but proselytizing in the direction of either Jerry Falwell or Richard Dawkins will not be well-received.

Having said that, the questions raised by the panel are threefold:

  • Does the missionary goal of organizations like SIL negatively impact the quality of the work they do as linguists?
  • Does the missionary goal of organizations like SIL negatively impact the communities that they operate among?
  • Given the divergence of goals between academic and missionary linguistics, should any sort of formal recognition be extended between the two communities?

The panelists responded to these in a variety of ways; what follows is based on my notes, which should not be treated as quotes in any case, and handouts where possible.

Lise Dobrin and Jeff Good, who organized the discussion, framed their talk as a series of questions not necessarily to be answered by them, but in need of discussion. They introduced several premises meant to be treated as the givens from which the discussion would start, but which SC found to be debatable notions in themselves:

  • Missionary linguists impose Christian values on their hosts even while conducting only their linguistic business
  • Academic linguists have recently discovered endangered language preservation as a moral cause (treating it as a human right)
  • The goals of missionary and academic linguists diverge because academics treat languages as ends in themselves, while missionaries ultimately subordinate their language work to the goal of evangelization
  • Different goals include different working priorities, such as SIL's shifting of spending on software from Shoebox development to font development, which isn't helpful to academic linguists

Starting from these premises, they ask some further questions:

  • Is it desirable for academic linguists to be dependent on missionary-developed tools when their goals diverge?
  • Is it desirable for SIL to have de facto control over things like the upcoming ISO standardization of language codes based on Ethnologue merely because nobody else has done the work, especially when there is disagreement over the empirical adequacy of SIL's categorizations?
  • Should academic linguistics be reconfigured in the 21st century to eliminate these dependencies?

While Dobrin and Good would argue they were just raising the questions, SC is concerned that it is hard to suggest anathematizing people or organizations without provoking the suspicion that an affirmative answer is the intended result. They emphasized repeatedly that they were not questioning individual motives or work, but rather the existence of formal relations with an organization that ultimately has missionary goals.  The attempt not to cast individual aspersions is admirable, but as a number of commenters pointed out afterward, it was very hard to determine what practical effects were meant to be had by dissociating from SIL without treating the collection of SIL linguists any differently.

They were followed by William Svelmoe, a historian from Saint Mary's College, who did not address the questions of the panel from the standpoint of professional relevance, but rather came to provide some helpful historical context. William Townsend, the founder of SIL, originally trained as a preacher, but eventually wound up as a missionary to Guatemala. His work there convinced him of the need to learn about local languages in order to communicate effectively with the populations, and that it was important this be done well, because if the Bible was to be presented as the word of G-d, it would not be particularly credible if G-d was seen to have a hard time speaking the local language. This would be particularly unfair in light of the fact that the responsibility would actually lie with human translators; therefore, Townsend undertook to establish "Camp Wycliffe" to train missionaries.

Fortunately for Townsend, Ken Pike shared his religious convictions and had the drive to be a great linguist, and thus set out to help him by establishing a school for translation, based in Texas. Townsend was thus free to concentrate on training the budding linguistic missionaries in their missionary skills specifically, which he did through a school based in Mexico. They clashed frequently over the need to emphasize getting good linguistics done versus good missionary work, but ultimately settled on a formula whereby the linguists would worry about linguistics, and would go into the field as members of SIL, whereas the job of translating the Bible would be carried out under the rubric of Wycliffe Bible Translators. Since all SIL members are Wycliffe members, this can be seen as something of an accounting fiction, but the take-home point is that the linguists are in no way supposed to proselytize as part of their work. SIL has since signed contracts with a number of governments to do language preservation work, contracts which specifically preclude their proselytizing (although not their contribution to Bible translation); however, the degree to which any individual is capable of keeping these goals separate cannot be predicted merely from their assent to such contracts. After all, it takes a committed person to go live in a foreign, and generally quite poor and remote, country for 10-20 years at a time.

Prof. Svelmoe was followed by Courtney Handman, whose talk SC missed in favor of Claire Bowern's, and so the next speaker heard by your host was Patience Epps, a specialist in Amazonian languages. Prof. Epps was there to argue for a separation between the LSA and SIL, and spent a number of slides establishing that SIL members view their work as a means to the end of "making disciples of all the nations" (quoted from Matthew 28:19). One example of the alleged dishonesty of SIL is provided from this page of publications by the late Nathan Waltz. While it lists his scholarly works, it makes no mention of Old Testament summaries or hymnals that he has also worked on. We'll address the relevance of this point later.

Prof. Epps claims that because some SIL members have been demonstrated to engage in active proselytizing, that this demonstrates the inherently proselytizing nature of the organization -- specific examples being adduced in support of this claim. She then made a point SC wishes to highlight for later discussion as well (and this is a quote, taken from her handout):

Translation of a text from a different language and culture into the target language is not documentation of the target language and culture itself.

We then get to the heart of her argument, that SIL's work is allegedly incompatible with group rights to self-determination. In a citation of Nettle and Romaine somewhat reminiscent of Jack Hitt's writing on language death ([be nice -- ed.]), she quoted their statement that "Every language is a living museum, a monument to every culture it has been vehicle to". In case the Hittite nature of that remark is questioned, she then cites the same authors as saying "a way of life disappears with the death of a language". A number of slides are then spent on the argument that while cultures may change organically, the introduction of Christian beliefs is specifically damaging because individuals with money, power, and prestige come into these societies, and no matter how scrupulously they may behave, it is impossible for the locals not to make connections between the outsiders' culture and their wealth. This then leads to an abandonment of the prior culture for reasons other than internal change, which SC thinks it is fair to say Prof. Epps considers an illegitimate prospect. Prof. Epps noted that the wealth brought by the outsiders includes food and medicine, and that this is a welcome charitable contribution, but questioned why the charity couldn't come "with no spiritual strings attached" (again a quote from the handout).

Prof. Epps was followed by Ken Olson, arguing the case for SIL (he's an active member, as well as a professor at the University of North Dakota, but spoke in his personal capacity only, not as a spokesman for the organization). Prof. Olson briefly acknowledged the dual nature of the organization from its founding, but gave a "top ten reasons to consider SIL a scholarly organization" which really had just one that mattered -- the existence of over 13,000 scholarly publications in the SIL bibliography (which itself states that it has grown to 20,000 entries; it's not clear when Prof. Olson's figure dates from). That's a fair amount of language work to do in 70 years.

Prof. Olson addressed the same self-determination issue that Prof. Epps brought up, both reiterating that active proselytization is not sanctioned by the organization (although he did not address what SIL does to actually prevent such efforts by non-conforming fieldworkers), and bringing in a quote from Matthew Dryer, who commented (in a personal communication to Prof. Olson):

The question of whether SIL's religious activities have a negative impact on indigenous communities misses a fundamental point. What right do we as academics...have to decide what is best for indigenous communities? Isn't that for them to decide?

Missionaries do not force people to become Christians. They simply give them the choice...I have made the choice not to be a Christian. Why shouldn't people in indigenous communities also be allowed to decide whether or not to be Christians?

Prof. Olson also raised the fact that charges of "ethnocide" have been previously brought against SIL by professional anthropologists, and found lacking by the American Anthropological Association. Finally, he produced evidence from a number of case studies that SIL involvement has actually helped grow the number of speakers of some endangered languages, weakening the charge that their influence leads people to "go Western" and abandon their historic roots.

Finally, Daniel Everett, a former SIL member, gave a speech opposing the organization (his work has been discussed by Geoff Pullum, especially with regard to anti-SIL bias, here). Prof. Everett is of the opinion that SIL exists to fulfill prophecies in the book of Revelation, and that this warrants a cautious stance toward them, but lauded many of the activities the organization has carried out to date regarding language documentation and the development of useful software tools. Since SC's laptop battery had run out by Prof. Everett's talk, and there was no handout, he must be especially cautious in what he says here, so we'll restrict ourselves to one comment Prof. Everett made that did not have the effect he intended.

Discussing a tribe he had worked with, and describing them as "hyper-empirical", he recounted a conversation that he had with some of their members one day where they asked him to stop talking about Jesus. It seems that because he had not personally seen Jesus, the locals concluded that he had no evidence for his existence. Therefore, they would like him to cease discussion of the topic. From this, Prof. Everett concluded that the people SIL workers visit would prefer not to hear about Christianity, and that they should not be spreading things which said people do not wish to hear about. In SC's view, this illustrates Prof. Dryer's claim rather clearly -- Prof. Everett's community had heard the claims, weighed them, and found them wanting. It hardly was the case that they could not handle the introduction of other cultures' views into their community.

SC found this debate wearying, as noted above, because so much of it seemed to be post hoc justification of already-held positions. Nobody in the room was going to shift their views on religion on the basis of what was presented -- it seemed even less clear to your host that anyone would shift their views on SIL, either. If you came in with an innate dislike/distrust of Christianity, you were likely to side with Prof. Epps, and take her anecdotes as sufficiently damning of the organization. If you came in with positive or neutral views, you were likely to side with Prof. Olson, and take his smoothing over of the existence of instances of coercion by resorting to their high-level policies as sufficiently redemptive.

Having said that, and not speaking as a person with any history of fieldwork abroad, it was SC's own opinion that the anti-SIL view is much ado about very little. Returning to the questions that formed the basis of the discussion, the answers seemed to be:

  • SIL's missionary goals do not appear to have systematically undermined the quality of the scholarship produced. Ethnologue may contain arguably wrong classifications, but it was not at all demonstrated that these stemmed in any predictable way from the specifically missionary goals that had sent the workers who produced it into the field.
  • SIL's missionary goals do not appear to have negatively impacted the communities they serve in a systematic way. While evidence of specific cases of coercive behavior was produced, and merits intervention in those cases, it is not at all clear that this behavior is widespread, nor that any kind of large-scale "ethnocide" is ongoing.
  • While the goals of missionary linguists extend beyond treating languages as ends in themselves and are thus not identical with those of academic linguists, it is neither clear from the above two answers that any kind of organizational malfeasance exists, nor that a specific formal tie between SIL and LSA exists to be severed.

The problem here is not only that the answers to the questions which inspired the panel are either favorable to SIL, highly ambiguous, or that the questions are too ill-formed to answer (i.e., nobody could agree on what relationship presently existed between LSA and SIL). It's that the posing of the questions itself seems to follow from a prejudgment against Christianity. SC will readily grant the veracity of every charge Prof. Epps leveled in its specifics -- this still leaves him believing that the organization as a whole is no threat to indigenous communities. An effort to show good faith on SIL's side might involve the public articulation of clearer policies for dealing with coercive proselytizing by actively rooting out offending individuals. This hardly requires renunciation of their beliefs, though.

More seriously, the insistence that  language documentation not result in the translation of any documents from English/Hebrew/Latin into endangered languages makes no sense unless it is grounded in specific animus against the documents in question. In machine translation, the gold standard for training statistical algorithms is to use parallel corpora, where the semantic content is identical and the structure of the documents at the level of both individual sentences and higher-level discourse can be cleanly mapped from one language to the other. People spend fantastic sums of money just to get parallel corpora of single pairs of languages, like the Canadian Hansards (the minutes of Parliament proceedings, in both English and French), or the 5 official languages used by the United Nations. SIL is an organization producing a parallel corpus spanning as many of the world's languages as possible. This is potentially an enormous gift to the field of machine translation (in fairness, SC is not aware of anyone using the data for this purpose at present, although it is not obvious why this should be the case). Irritation with the fact that the Christian Bible is the chosen document instead of a myriad of culturally-specific tales (which would be extraordinarily expensive to render in numerous other languages) is very hard to comprehend unless prior bias against Christianity is involved.

Returning to Prof. Epps' concern that Nathan Waltz had done religious work not mentioned in his c.v., it was not at all clear that his academic work qua academic work was in question. So the man wrote hymns. What if he had written erotic fan fiction centered around Mario, Luigi, and Princess Toadstool? Would that have impugned his academic work? It might not have been a credit to him personally, but surely the answer is "no". And if he had written such works and then proceeded to introduce them to the populations he worked with? Would that have been a lesser disservice to them than the sharing of religious ideas? More than a few concepts which might be foreign to the indigenous cultures would be passed along. If the transmission of any specifically Western knowledge or beliefs is an unacceptable incursion into the right of self-determination, all linguists, anthropologists and sociologists need to get out of the field right now and not return until they are sure that the abstract documentation of the local language and culture for the sake of disinterested knowledge is a value of the communities being studied as well as their own. Obviously this is not a serious proposal -- the point here is rather that secular scholars possess their own values, and it has not been demonstrated that those values are shared by the communities they visit any more than the missionaries' values are. So long as systematic coercion is not demonstrated, the relevance of non-scholarly activities by SIL workers is exactly as relevant as those of secular academics -- which is to say, not at all.

Last in the list of issues, the question of whether or not the academic community should cease its reliance on SIL-developed grammars and software struck SC as entirely ill-posed. If the work is lacking in scholarly utility, then yes, of course it should. But if that's not in question, then we return to the problem above -- namely, that the specifics of extra-academic work are irrelevant. Clearly, there are cases where this would not be true; the Nazis conducted gruesome experiments on their prisoners, and the question of relying on their medical work has been addressed by practitioners in that field . The reductio ad Hitlerum is grossly inappropriate here, though -- SIL linguists aren't murdering anyone. They're bringing food and medicine. It's hard for SC to understand why this taints their software while the unknown values of presumably secular developers should be assumed to be pure and beyond reproach.

One might read all this and conclude that your host merely had his own pro-SIL biases reaffirmed. While he nominally has no brief for specifically Christian belief, perhaps he merely sides with the SIL folks for fear that this was all just rehearsal for religious tests in reviewing linguistic work. Actually, SC is now curious to read some of Prof. Epps' citations -- it certainly sounds like specific wrongs have been committed, and this was something SC had not been aware of prior to attending the panel discussion. However, it certainly is the case that he came away from the discussion convinced that everyone involved might have benefited from reflection on a parable dear to the faith of the SIL founders. Jesus told those who would stone an adulterous woman not to do so unless they were free from sin; none could, and she lived. But this did not prevent him from then admonishing her: "Now go, and sin no more".

January 09, 2007

Have lexicon, will travel

At the recently-concluded LSA 2007 conference, Mark Liberman gave a plenary speech titled "The Future of Linguistics". SC hopes that Prof. Liberman will make his slides available online for the perusal of those who didn't attend -- aside from being a bracing challenge to the field to do a better job of publicizing itself and staying relevant, it featured a hysterical Photoshopped cover of the third issue of the best popular linguistics magazine never published.

Humor value aside, Prof. L. raised a number of points that should be of concern to anyone with a linguistics degree and sans tenured position. To sum up the key points: basic linguistic skills like the technical analysis of sentence structure, formerly key components of education, are taught late (if at all) to most people, and not by linguists. Despite the centrality of linguistics to much of 20th century science (particularly analytic philosophy and anthropology), linguistics departments are not present at many (most?) universities, and training in the key findings of the field is far from systematically available. In particular, he made an interesting comparison between the 20th century histories of linguistics and psychology, and questioned why it was that linguists are outnumbered by almost 40:1 by psychologists (as measured by your choice of professional society affiliations or introductory course enrollments).  Along the way, he suggested that perhaps part of the problem facing the field was the relatively late development of its modern incarnation, with much of the work taking place in the 1940s and '50s, as opposed to the decades earlier establishment of modern psychology.

This immediately struck SC as the one false note in an otherwise outstanding analysis, and during the Q&A session that followed, your host raised an alternative comparison, to the development of computer science. In the 1940s and '50s, computer science was an activity largely carried out in the head of John von Neumann, and the ENIAC researchers at Penn. Yet today, you can walk into any Barnes & Noble ([wait, you walk into bookstores? what about Amazon? -- ed.]) and pick from a dozen shelves' worth of books on how to do various things with computers, no small number of which carry the economically beneficial effect of helping you earn a living. Indeed, just in the past year, your host has acquired books on: PostgreSQL (his database of choice), Java Server Faces (the web interface tool of other people's choice on his project), Java Server Pages (related to Faces), JDBC (Java connectivity to databases), multithreaded programming, and statistical analysis using databases. All of these are directly related to your host's ability to earn a paycheck, a statement lamentably untrue of most of his recent linguistics-related purchases. So the question put to Prof. Liberman was: what is the linguistic equivalent to the professionalization of computer science that will make it attractive to more people? Prof. L.'s reasonable response was that this was a complicated question which could spark a lot of further dialogue, and so later in the day, SC promised him that said dialogue would be coming. Here goes:

Let's start out by defining our terms a little more carefully. Following numbered instructions from a book is not quite the same as doing basic research on difficult theoretical questions, and so we should distinguish the jobs we are talking about. If a researcher is a computer scientist, and the guy who picks up a how-to book is a programmer, then maybe a similar distinction is in order here. Since someone seeking to do professional linguistic work probably will have some programming tools in their resume, we'll need a different word. SC likes "mercenary linguist", or maybe "linguist of fortune", but since the PR value of those names is low, how about "linguistic technician"? Your host is unsure where this places master's degree holders like himself -- more knowledgeable than someone doing paint-by-numbers, but not a full Ph.D. -- but for the moment, we're trying to imagine the career path of someone whose knowledge is limited to an as-yet hypothetical LSA/ACL certification guide.

In order to make the analogy work, a linguistic technician's tasks ought to mirror those of a generically qualified programmer's in some abstract respects. Without pretending to an exhaustive consideration of the desiderata, it seems to SC that a linguistic technician's tasks need the following properties:

  • They must be repeatable -- a dozen equally trained technicians should all come up with the same answer, broadly construed, to the same problem
  • They must need to be done frequently -- the problem of connecting a database to a web server is long-since solved, but individual programmers write code to implement specific solutions for it every day
  • They must be subject to statistical quality control -- while language defects might be more subjective than software defects, money is only going to be made available if defects can be corrected efficiently with finite resources

This is not merely the declaration of linguistics to be a specialization of computer science. Practically every field that isn't only academic comes in both researcher and technician flavors; one can be a full medical doctor, or break into the field much more quickly (albeit with a much lower ceiling) as a nursing assistant or licensed vocational nurse. One can be a Ph.D. chemist or biologist and run a lab -- or one can be a lab technician and know how to carry out the various procedures without necessarily being qualified to direct original research. One can be a lawyer -- or a paralegal.

Objections might be raised that various classes of programmer are more skilled than this; web designers need to be graphic artists of at least better talent than the average randomly-selected person, but insofar as this is the case, that only demonstrates that technicians come in varying degrees of certification and competence. Another objection might be that almost any task that passes the above test is likely to be something involving computer programming, and thus that all I'm about to demonstrate is that a little extra linguistics coursework should be part of programmer certifications. Perhaps so, but if that's really the case, then we might have to face up to the possibility that Prof. Liberman's speech is grounded in a mistaken premise, and that the best we should hope for is getting more would-be lawyers and journalists to take courses in grammar so as to employ a few more Ph.D.s than is possible at present. Your host would not be writing this if he believed such to be the case.

We return again to the question -- what does a linguistic technician do? If genuine standards for representing grammars or ontologies existed, perhaps full-time grammar analysts would spend their time reviewing the results of low-scoring parses in a company's internal document management system, and determining company-specific improvements to be made. Maybe future theory checkers could be employed to manually translate legal arguments into predicate calculus, then run them against automatic theorem provers to make sure that cases are constructed soundly (yes, SC realizes this is precisely the sort of thing the research community would like to automate, but constructing a suitably broad-coverage application might be more expensive than having a few individuals responsible for this within a firm). Perhaps some future senior translation engineer, not actually capable of speaking the languages they are overseeing, will be responsible for overseeing metrics of bulk translation efforts, and farming out specific high error-rate translation pairs for correction to specialists (perhaps using services like Amazon's Mechanical Turk to do it). These are all ideas that meet the criteria above: there are demonstrably better and worse answers, if not always uniquely correct answers, they're all tasks that would need to be done daily in any enterprise, and they're all subject to quality controls using metrics that either already exist or can easily be conceived of with no great leap from the present state of the art.

Of course, in order for jobs like these to exist, there need to be applications. Of all the software applications that SC has seen to date which give end-users or administrators access to any kind of linguistic data tuning (names will not be forthcoming, to protect the guilty), the manuals usually come with the suggestion that all you need is some generic software engineer/system administrator to do the necessary fiddling. Aside from the fact that many of these systems are not yet at the point where much linguistic sophistication would do a lot of good, this represents a failure of marketing on the part of linguistics as a profession. Why shouldn't an application specializing in language technologies do better when managed by a specialist in language?

As noted above, though, conceiving of applications purely in terms of software reduces the notion of the language technician to a software engineer with language skills. This is a mistake. As Prof. Liberman noted in his speech, English departments have by and large abrogated the responsibility of teaching grammar and composition. Why shouldn't technical writers be certified as competent drafters of language by linguists? It might produce better results than Phil. Why not train a new breed of paralegals to be the "theory checkers" described above? Speaking of legal applications, why not train someone to specialize in analyzing the sort of terrible ambiguities that come up in courts over grammatical issues as trivial as comma placements? Couldn't the cost savings in litigation fees alone support paying to have a few such people on staff in every legislative chamber in the country? We already pay for "legislative analysts" to outline the expected legal and economic effects of bills before they become laws; there's no reason that language analysis shouldn't be part of the package.

SC does not pretend to have come up with anything like an exhaustive list of potential jobs for language technicians, but the thrust should be clear. In an increasingly professionalized society, obsessed with credentialing, one of the great failures of the linguistics field is that it has not managed to lay claim to accrediting those purporting to be expert in it. As Prof. Liberman astutely observed, other fields have been allowed to take over the intellectual domain of linguistics, and the quality of the training that results is disturbingly subpar. Taking some responsibility for fixing the problem would be enormously beneficial -- pure research, fieldwork, endangered language preservation, and other non-commercializable work are luxuries that need to be paid for, and the perception that linguistics needs to be funded will only come when administrators and industry see money in it. In order to do this, we need to figure out what linguistic technicians can do, and how to make the non-academic world want to hire them.

January 06, 2007

SC's New Reading List

So your host mentioned yesterday that he had gone on quite a tear through the publishers' displays yesterday. Some of the titles he purchased are things that were overdue for his shelf; some just caught his eye and begged to be read. After collecting the three that were being held for me this afternoon, it's time to tally up the list since they'll probably be getting plenty of mentions in the near future. The order below reflects only their order in the bags I unpacked this evening, and the listings are by title and authors.

Some of these are of obvious interest to SC, being computationally relevant, but your host tries not to be too intellectually narrow. Truthfully, the book by Paul Postal sold itself about 10 seconds after SC opened it up to a random page and encountered the sentence "If It Doesn't Follow Automatically, Then It's Pretty Much Got to at Least Virtually Follow, and If Not, Don't Worry; It Is Still Unquestionably Natural" (italics and caps in original). There's no question it will be full of dead-on reasoning, but it also looks a bit cranky, or perhaps grumpy, both of which are meant very much as compliments in this case. SC likes grumpy.

Full disclosure is in order on one of these purchases; your host worked with Sergei Nirenburg on a project sponsored by ARDA while he (Prof. Nirenburg) was preparing the manuscript. However, he (SC) has not read any of the content of the book as it was organized for the book -- flipping through it suggests that some of the material is from documents seen before, but your host did not serve in any sort of reviewing capacity.

Anyway, there's plenty more to say about the actual events of the conference -- Mark Liberman's talk on the future of linguistics was particularly stimulating, and your host already promised a reply to it afterward; the symposium on missionary linguistics and academia was contentious (albeit discussed in a genuinely civil manner). More on all that tomorrow.

SC does the LSA

Where to start? To quote The Sound of Music once again, we'll start with the beginning, because it's a very good place to start.

While 2006 was an annus horribilis for this blog, your host had never had any plans of retiring ([It just looked that way for months at a time. Define "plans". -- ed.]).  The lack of productivity aside, it had been your host's intention to go to the annual Linguistic Society of America conference from the day he found out it would be in Anaheim. If nothing else, he hoped it would be an opportunity to meet some of the people who made it so worthwhile when it was a more regular activity. Beyond that? Well, as he told a former professor who he was pleased to run into on the first day of the conference, it might be a chance to pick up some inspiration and find something to write about again. That alone would be wonderful.

It had been my hope to write something each day of the conference, and thus, properly speaking, the first LSA-related post ought to have occurred on Thursday night. It's not like I didn't hear things worth posting about -- The Tensor gave an interesting talk which I even asked questions about in the Q&A session (without him knowing who I was). Would've introduced myself at the end of the session, too, but I lost track of him shortly before it wrapped up. Whoops. In any event, following the session your host ended up going to visit with an old friend from college, who happened to be in Orange County as well, just for a couple of days. Being up until midnight, and having to be up at 7 am to make the drive back to Anaheim from LA the next morning, getting a post done just wasn't in the cards.

What SC could not have predicted was what a wonderful day the second day of the conference would turn out to be. The presentations were interesting -- the entire panel on approaches to complexity was worth hearing (and worth its own post), Heidi Harley's discussion of the amazing theory-busting data that she's pulled out of Yaqui has me excited to read more about it, and as much fun as the endangered languages symposium was, it was worth sneaking out in the middle to the session titled "That" to hear a talk titled "What does that mean?". Not a whole lot as it turns out, which was a point made in an interesting and experimentally novel way. The official SC wallet is meanwhile screaming in agony from damage inflicted on it by a trip to the publishers' exhibit room -- your host came home with four books, and has to pick up three tomorrow in cases where he bought the only copy they'd brought. After the third such incident, your host remarked to a publisher's representative that it looked like he wouldn't leave the room until every publisher there owed him a book on Saturday afternoon. Lots of good stuff, much of which I hope to write about as I read, and not all of it computational in nature.

But the highlight of the day came at 5 p.m., when I went to the hotel bar to meet up with other linguistics bloggers. The Tensor was there first, and when I went up to introduce myself, one of the first things he said was "you owe us a couple of posts about how that jury turned out". He's right, of course, making a point which Dad SC and the aforementioned friend visiting for a few days have been pushing themselves, but more importantly, as soon as he said that, I knew it was going to be a good evening among friends.

I proceeded to have a wonderful time getting to finally meet many, not to say all, of my favorite linguistics bloggers -- the aforementioned Tensor and Heidi Harley,  Mark Liberman, Claire from Anggarrgoon, the Literal-Minded Neal Whitman, Russell Lee-Goldman -- and a few more who I should be reading regularly. Had other longtime favorites like Polyglot Conspiracy, Languagehat, and EFL Geek been there, the hotel staff might have had to drag me out of the bar. Having missed the last two LSA conferences after planning to go to both, it was my first time meeting any of them, and I'm pleased to report that every last one is as friendly and interesting an individual as you would guess from reading their blogs. Despite my prolonged absences from the linguistic blogosphere in the last year, I felt as welcome as though I had never been gone, and I thank them all for it.

I get to break one interesting piece of news that many people at the gathering don't know about because we didn't all leave at the same time. While Ben Zimmer has already noted the selection of "to be plutoed" as this year's Word of the Year on Language Log, those of us who gathered in the bar missed the announcement altogether when it happened. So when your host and a couple of new acquaintances ran into a camera crew from the Discovery Channel, at the conference to cover this earthshaking story ([don't you mean "Pluto-shaking"? -- ed.]) none of us had any idea what to say when they asked us to offer our reactions. Once they realized we weren't kidding about not knowing the word that won, the crew cut us a deal -- they'd tell us the word, but only if we'd let them film our reactions. Was SC ready for his close-up? Who knows? But you might be able to find out, if that clip makes it to TV. Your host had to sign a waiver form giving them permission to use his image without compensation, despite his rates being far more reasonable than Steven Seagal's already, but the most important part of the form to him was the part where he gave them contact information so they could call to tell him if he'll be on. If so, rest assured that the information will appear here immediately ([and be bumped to the top every day until it happens -- ed.]).

All told, it was a wonderful day at the conference, and I'm very much looking forward to another day of more good sessions than I can possibly make it to. I'm currently planning on going to the morning sessions on corpus-based investigations and computational approaches to linguistic analysis, followed by Mark Liberman's must-hear speech on the future of linguistics, and concluding with the certain-to-be-uncontroversial symposium on the tension between missionary and academic linguistics. The LSA convention has been enormously interesting for me, and I regret only that I waited so long to go!

March 24, 2006

Play Lexicon, lose your mind

Anybody who comes here on purpose, as opposed to via Google, probably already subscribes to the Linguist List. But whether or not you subscribe, you may not have visited their website since their annual fundraising drive began on March 17th. SC hadn't until today, and thus wasn't aware of a fiendishly difficult puzzle game they've put up on the site.

It's called Lexicon: Game Without Rules, and it's hurting your host's brain just thinking about how hard it is. As of this writing, SC is stuck on the third screen of the game, which isn't particularly impressive, because the first screen is a gimme, and the second screen only requires a bit of trivia that most linguists probably already know. All there is to say about the third screen is the first dozen theories you will think of are probably all quite reasonable, and they're certainly all wrong. SC shudders to imagine what he'll find on the fourth screen, if he gets that far (and if you do, please don't post any answers, here or elsewhere).

While you're there, consider making a donation. The Linguist List is a terrific professional resource, and deserves support from the community.

[UPDATE @ 1:48 a.m. 3/25/06: I'm past the third screen. It's possible to outthink yourself with this game.]

October 15, 2005

Breaking up is hard to do

Perhaps if Neil Sedaka had heard about text messaging, he might have changed his mind.

While this is more of a Polyglot Conspiracy-type story than an SC one, your host could not help but be intrigued by an Agence France Presse article suggesting that perhaps as many as 15% of people between 18 and 35 are using text messaging to end relationships. The research, done by Macquarie University graudate student Natalie Robinson, is covered in more detail in a press release from the school.

As is often the case with science reporting, it's wise to be careful in taking the news at face value. SC did a double-take on reading the AFP report's line that:

Macquarie University researcher Natalie Robinson studied the texting habits of 100 young people aged 18-35 and found SMS messaging increased when relationships were beginning or going through a rocky period.

Could it really be that 35-year-olds -- people even less likely to be on the cutting edge of technology than the 28-year-old SC -- were as likely to use text messaging to dump people as 18-year-olds? Probably not. The AFP report is rather misleading about the sample, which skews heavily toward the young end of the range as the Macquarie press release makes clear:

Which led Robinson to undertake a study of the text message habits of around 100 young adults, three-quarters of whom were first-year university students. All of the subjects were in a dating relationship, or were engaged or married.

In the absence of access to the actual paper (Dr. Robinson doesn't appear to have a home page, and no journal is cited in either account), it's possible that there are more 30+ people in that group than one might expect from a typical population of "first-year university students", but SC would be greatly surprised if this was so.

There might be something to her comment that it's hard to interpret whether text messages are a sign of affection or a sign of suspicion, given the lack of "facial expression, tone and other pertinent cues". As this purely anecdotal BBC story demonstrates, it might be that the proper interpretation of frequent messages is "my boyfriend is not merely a control freak, but an addict who frankly needs to have his phone confiscated -- and his computer, too". (Found via a rather amusing blog on all things SMS-related.) If you're sending 700 messages a week, as the individual in this story did, it strikes SC as rather likely that any cues missing from text alone are more than adequately replaced by other indicators. But with some more sane level of text messaging -- say, 600 per week -- one might need to have a conversation with one's significant other on the exact meaning of text messages. Preferably not over the phone.

The least surprising news from the Macquarie press release might be the effect that this work has had on Dr. Robinson's personal...um...text-life:

But while text messaging seems set to continue its meteoric rise in popularity worldwide, Robinson's own phone has suffered a steep decline in incoming messages.

"It was interesting to read other people's text messages," she says "but I have found that since doing this research very few of my friends send me text messages any more!"

Is there any surer way to make people cautious about their language use around you than to tell them you're doing research on it?

September 28, 2005

¿Cómo se dice "Buenos dias" en coreano?

The politics of bilingualism have been a frequent topic of discussion in these pages. Last year, we had a look at a controversy in Maryland, over fast-food workers not speaking English in primarily English-speaking neighborhoods. Earlier this year, we took a look at a North Dakota state legislature bill that would have mandated demonstrated English competency in state schools. What these stories have in common (aside from the chest-thumping of various ideologues) has been an undercurrent of concern that immigrants aren't learning English sufficiently to integrate with the communities in which they live. Now courtesy of the Los Angeles Daily News, an interesting look at the development of economically interdependent communities in Los Angeles where bilingualism is critical to success -- but neither of the languages is English. As the story begins:

Peruvian immigrant Miguel Aliaga always knew that coming to Los Angeles would mean a long struggle mastering a new language. He just never figured that language would be Korean.

The story goes on to provide some anecdotal examples of Korean immigrants who have come to feel the same way about Spanish:

A Korean immigrant - by way of Argentina - Martin Paik writes a column, "Hola Amigo," in the Korea Times that provides conversational Spanish lessons in Korean. He doesn't speak English and finds little reason to, living in Los Angeles.

"In California, Spanish is more important than English," said Paik, a Seoul native. "I haven't found any inconvenience because I don't speak English. ... I don't need to speak English. If you can speak Spanish, you can drive, employers can have clients, you can order in restaurants, you can do anything."

Paik receives his credit card bills in Spanish and orders his office supplies in Korean. He teaches Spanish in Korean at a school he runs in a largely Latino neighborhood near Koreatown.

Most of the 200 students at Martin Spanish School speak little or no English. The only hint of English in the instruction books - which he wrote himself - are on the cover page.
...

Yoon Seong, a 60-year-old Korean - by way of Spain - said he feels fortunate to know Spanish. He lives in West Hills and said, unlike many of his Korean friends, feels no need to move to Koreatown.

"For me being here, the Hispanic community is the only world for me. I don't need English here. All that you need in California is Spanish."

One wonders what language these quotes were extracted in. But these are admittedly only anecdotal examples; does any data exist to support the alleged trend? As the article itself notes, no government agency is currently following this phenomenon. And Census data doesn't give us a whole lot to go on. The 2000 Census' profile of Los Angeles county tells us that 54.1% of Angelenos primarily speak a language "other than English", and slightly under half of that group rate themselves as speaking English "very well". So at most, about a quarter of the population could be going non-English-bilingual, but we're not done with the numbers just yet.

The Census provides us with a number of cross-tabulations of English proficiency against various demographic variables. This data doesn't allow us any way to directly estimate the number of people who might be learning a second non-English language, but our goal here is to put an upper bound on it. In order to do that, we have two types of tools: 1) counts of the members of the population who speak English "not at all", and 2) counts of members of "linguistically isolated" households, defined by the Census Bureau as "one in which no member 14 years old and over (1) speaks only English or (2) speaks a non-English language and speaks English "very well." In other words, all members 14 years old and over have at least some difficulty with English." Note that these numbers are based off of the "long form" version of the Census, which was only sent to 1/6 of the population; further details about methodology can be found here.

To start off with, we'll use table P19 of those cross-tabs to calculate the percentage of people within each linguistic subgroup who might be candidates for non-English-bilingualism. Note that the data set is for Los Angeles county only, since that's the only place the article alleges this trend is taking place (which isn't to say it might not be interesting to repeat in other places, or nationally). In each column, we're computing the percentage of members of each subgroup who report that they speak no English at all -- not as a percentage of the total population, just as part of the subgroup.

Age group Spanish Other Indo-European Asian/Pacific Other
5-17 2.9% 0.5% 0.7% 1.0%
18-64 14.7% 2.4% 3.6% 2.9%
65+ 23.4% 15.5% 20.7% 7.7%

Table 1: Percentage of each subpopulation with no English ability

It might be instructive to compare the rate of zero English proficiency with the rate of native/nonnative birth -- are the people who don't speak English at all doing so because they're isolated while here, or because they didn't start out that way? We can do this using table PCT12 (same link as above). Alas, while they break it down by nativity for the whole 5+ population, they don't subdivide it by the age groups above. Still, it might provide something of a baseline against which to evaluate the numbers above:

  Spanish Other Indo-European Asian/Pacific Other
Native 1.2%
0.3%
0.5%
0.6%
Non-native 19.3%
5.7%
6.2%
3.9%

Table 2: Native vs. foreign-born percentage of subpopulation with no English ability

Now we're getting somewhere -- native-born members of every subpopulation almost universally speak at least some level of English. The non-native percentages of those with no English skills also are notably larger than the percentages of 18-64 year-olds in all groups, which would tend to support the idea that those who aren't acquiring any familiarity with English are the first-generation immigrants, and not their descendants. Of course, we have to be reserved in adopting that conclusion since we have no way of positively determining a correlation between the two tables -- nothing in the available Census tables (at least not in this summary file) tells us how many members of the 18-64 group are native-born, and how many are not. We also haven't done anything to see if there's a trend, say by comparing the above numbers with 1990 Census data, but some of the numbers above are pretty close to statistical noise as it is, and not worth checking for trends.

Where there is some cause for concern is in the number of "linguistically isolated" individuals. Table PCT14 tells us that some 16% of the total Los Angeles-area population live in households where everybody speaks a language other than English, and nobody over 14 speaks English "very well". Fully 51% of those people are between 18 and 44, while 24% are 17 or under. This is harder to write off as merely being an immigration effect, as the distribution skews too young to just be old people who aren't going to learn. We have no way of knowing what percentage of these individuals are immigrants themselves and how many native-born, nor can we tell how many of them actually have zero English fluency (the standard we've been using above). It's possible that the younger generations are learning, and will integrate eventually. Still, if the acquisition of English counts as any kind of proxy for integration, it's no benefit to the younger population to have minimal English exposure at precisely the times when it would have the maximum impact on their ability to integrate and succeed.

Addressing SC's review of Samuel Huntington's Who Are We? last year, Geoff Nunberg wrote:

In fact, English is too useful and important to imagine that any immigrant group would be willing to turn its back on it in order to maintain a marginal, ghettoized existence. Whether the acquisition of English will continue to bring with it a sense of belonging to a national culture depends entirely on the economic and social opportunities that assimilation offers to immigrants, and on our ability to refashion the idea of American citizenship to meet new challenges.

The numbers presented above suggest that he's not wrong -- certainly, as far as native-born individuals go, the pull of English is too strong. But when Korean immigrants go on record as saying "all that you need in California is Spanish", it's worth at least asking the question of whether or not turning one's back on English still actually is a choice to maintain a "marginal, ghettoized existence".

September 17, 2005

Losing touch with the past

Over at Language Log, Arnold Zwicky and Sally Thomason have written about the recent passing of Murray Emeneau, a president of the Linguistic Society of America in 1949. Prof. Zwicky's post is a particularly poignant reflection on his long and illustrious history in the field (though he's too modest to say so himself), and it's quite something to see how the roster of LSA presidents tracks his professional evolution. Alas, in reading their obituaries for Prof. Emeneau, as well as Languagehat's comment that "he was a giant in the field", your host had to acknowledge that despite 4 years of graduate schooling at two institutions, he had no idea who Murray Emeneau was.

This is not to say that SC could rattle off a list of publications and ideas put forth by everyone on the aforelinked roster, but between reading The Linguistics Wars and Pieter Seuren's excellent Western Linguistics: An Historical Introduction, your host had never come across his name in the history books, either. Consulting the Berkeley obituary, one finds that "Prof. Emeneau is also generally seen as having initiated the modern field of areal linguistics". So off to my bookshelf to consult William Croft's Typology and Universals; alas, no mention there, either, despite Sally Thomason's averral that his original publication on the subject is "a ground-breaking article, and...not outdated even now".

This all leaves SC rather disturbed. The aforementioned history books are disposed towards syntax (in the case of The Linguistics Wars, necessarily so, although MIT phonology gets substantial treatment in there as well),  and their coverage of early linguists focuses -- at least so far as the twentieth century is concerned -- on those who bore most directly on Chomsky and his program (Boas, Sapir, Bloomfield). Even though they both eschew the English uber alles approach to language research, it's surprising to see someone who covered so many languages within the large Dravidian family not even rating a mention. There are all sorts of interesting things about Dravidian languages that make trouble for English-derived theories -- for example, Kannada has nine prefixes, and they mean "fore, back, down, two, big, intense, clear, red, sweet", which wreaks havoc on syntactic and morphological theories that try to treat prefixes as functional projections.* SC didn't learn that from Prof. Emeneau, but he has to wonder what else he might have gotten out of him instead. The history of linguistics is not only the history of syntax, and especially not only the history of research on English, but too often, that's what it gets reduced to, and we're all poorer for it.

*Aronoff, Mark, and S.N. Sridhar.  1988.  "Prefixation in Kannada".  In Michael Hammond and Michael Noonan (eds.), Theoretical Morphology: Approaches in Modern Linguistics, 179-192.  San Diego:  Academic Press.

June 20, 2005

Incompletely understood

Wonderful interview in Edge of a philosophy professor named Rebecca Goldstein, on her new book about Kurt Godel, Incompleteness: The Proof and Paradox of Kurt Gödel (courtesy AL Daily). Your host hasn't previously encountered Prof. Goldstein's work, but is very much looking forward to reading the book on the strength of her observations about Godel the man, as well as his oft-quoted, oft-misunderstood incompleteness theorem.

Godel was a firm believer that both natural and formal languages aren't quite up to the task of describing the realities they correspond to, but he wasn't quite prepared for the cracker barrel philosophizing that his theorem would give rise to. Prof. Goldstein has a keen eye for the mischief in question:

Gödel mistrusted our ability to communicate. Natural language, he thought, was imprecise, and we usually don't understand each other. Gödel wanted to prove a mathematical theorem that would have all the precision of mathematics—the only language with any claims to precision—but with the sweep of philosophy. He wanted a mathematical theorem that would speak to the issues of meta-mathematics. And two extraordinary things happened. One is that he actually did produce such a theorem. The other is that it was interpreted by the jazzier parts of the intellectual culture as saying, philosophically exactly the opposite of what he had been intending to say with it. Gödel had intended to show that our knowledge of mathematics exceeds our formal proofs. He hadn't meant to subvert the notion that we have objective mathematical knowledge or claim that there is no mathematical proof—quite the contrary. He believed that we do have access to an independent mathematical reality. Our formal systems are incomplete because there's more to mathematical reality than can be contained in any of our formal systems. More precisely, what he showed is that all of our formal systems strong enough for arithmetic are either inconsistent or incomplete. Now an inconsistent system is completely worthless since inconsistent systems allow you to derive contradictions. And once you have a contradiction then you can prove anything at all.

More than her understanding of Godel as a theorist, though, SC is fascinated by her take on Godel the man:

Here's what I think. Gödel was irked by Wittgenstein. He not only held meta-mathematical views that were deeply at odds with Wittgenstein's—and though Wittgenstein wasn't a positivist, his views on the foundations of mathematics, especially in the Tractatus, were in the positivist vein—but he was irked, too, I think, by the fuss that those around him, the positivists of the Vienna Circle, made about Wittgenstein. Maybe he was even irked by the fuss that Wittgenstein made about Wittgenstein. We only let people get away with that sort of stuff if we think they're worthy. And by Gödel's lights, Wittgenstein wasn't.

This is followed by the acknowledgement that her work is necessarily more speculative than it would be if the man had just displayed more emotion in his life:

Gödel was a reticent man, an opaque man. He doesn't give one a lot with which to try to imagine the inner man. A novelist is trained in the art of inhabiting characters, both real and imagined. A lot of the novelist's skill resides in trying to insinuate oneself into others' inner lives. Gödel is a hard one to penetrate. I'm fairly confident that there was some strong emotion connected with Wittgenstein; I can construct a fairly convincing story to this effect. But in the end it might be a made-up story. It's compelling to me, for what that's worth, and it makes sense, psychological sense. And there's even some written evidence.

It might all be speculation, but it's very entertaining speculation, and by the time you finish reading the interview, it's hard to think that anybody is better qualified to be doing it than Rebecca Goldstein. SC has previously read one other book purporting to deal with Godel the man, as well as the math, but it was among the most disjointed, randomly organized books he's ever read on anything. This should be a treat.

April 29, 2005

Los Angeles, Mexico

This bit of vindictive schadenfreude by the L.A. Times, kicking the city of San Diego while it's admittedly down, cannot go unpunished. Bringing up J. David Dominelli? (If you don't know who that is, don't worry -- someday your host will write about how SD never really gets scandals appropriate to the 7th-largest city in the U.S.). That really is scraping the barrel, Times. Therefore, today's composition will focus on an irritating little incident going on in Los Angeles at the moment, as part of our ongoing coverage of language politics in California.

The setting? A Spanish-language TV station has recently put up billboards which very unsubtly argue that Los Angeles is part of Mexico, not California. The Times describes it thusly:

New billboards advertising a Spanish-language newscast on KRCA-TV Channel 62 were intended as an attention-grabber for its core audience, but instead have struck a nerve with activists seeking to curb illegal immigration.

The billboards show two cable newscasters sitting in front of the downtown skyline, with "Los Angeles, CA" printed above. The "CA" is crossed out, and "Mexico" is stamped alongside in bright red letters. Underneath are the Spanish words, "Tu ciudad. Tu equipo." — Your city. Your team.

([You can see a picture here. -- ed.])

SC only learned of this controversy when Lenard Liberman, the station executive quoted in the story, appeared on Larry Elder's radio show yesterday afternoon to defend the billboard. He largely repeated the same comments he had made previously to the Times:

"We tell the story behind L.A.., and we tell the story behind Mexico," he said. "If they find that offensive, I'm sorry. But you just have to drive around L.A. to know that this is a Hispanic city."

It's hard to parse this as meaning something other than that the ad is intended to convey a sense of disregard for living in California. Other executives for the broadcaster try to put a more positive spin on it:

"All we are saying is, 'It's your city, your town, your team,' " said Andrew Mars, vice president for sales for Liberman Broadcasting. "We are a team that's educating and informing the Spanish-language marketplace."

Of course, it's impossible to get inside people's heads, and speculating about motives would be pointless. However, even taking both executives at their word, the lack of sensitivity to the perceptions of the English-speaking community is surprising. It's at least arguable that perceptions of previously having toyed with the idea of a "reconquista" cost Cruz Bustamante a chance to become governor of California in 2003, and Antonio Villaraigosa a chance to become mayor of Los Angeles in 2001. In suggesting that Spanish-speakers view themselves as resident aliens, regardless of their actual citizenship status, KRCA's executives have chosen to lend credibility to these views, a grave disservice to precisely the audience they are trying to reach.

April 28, 2005

Coming soon: the language czar

It's no great secret that since September 11th, 2001, foreign language skills have been a concern at the Pentagon in a way that they were not previously. But now, courtesy of Government Executive magazine, we learn that they've come up with a solution to the problem: create a bureaucracy for it.

In 2004, the DoD held a conference at the University of Maryland to discuss the shortage of qualified translators in strategically important languages, which resulted in the commissioning of that most useful of products, a white paper. Quoth the article:

"This task requires guidance and incentives from the federal level," the report said. "It is urgently recommended that a national language authority be appointed by the president to serve as the principal adviser, advocate and coordinator in the federal government, and to collaborate with state and local governments, academia and the private sector for improving our national foreign language and cultural understanding capabilities."

OK, so the paper doesn't actually say the word "czar" anywhere (and czars being the equals of kaisers and kings, how come we never talk about having a "drug kaiser" or "terror king"? [or queen -- ed.]). But the idea is the same; a new federal official, albeit probably not Cabinet-level, will end up overseeing a centrally coordinated effort to standardize langauge teaching. Assuming it doesn't end up being just one more piece of hot air from a Washington committee, the paper proposes to establish a "National Foreign Language Coordination Council" with responsibilities to include:

  • Develop and coordinate a process for identifying, assessing, and distributing a comprehensive list of needs from all sectors for foreign language ability and cultural understanding
  • Engage in regular consultation with language professionals from all sectors and with leaders in the Nation’s heritage communities regarding needed action
  • Design and oversee – with appropriate government and private sector support – a system that ensures coordinated foreign language and regional studies programs in K-12 and postsecondary training in order to raise the level of understanding of all students and produce highly skilled language professionals
  • Develop and oversee the implementation of foreign language and English language skill-level certification standards, teacher certifications, and graduation criteria for foreign language majors and for the language component of non-language majors (e.g., international business, national security studies, public administration, and health care)
  • Advocate funding for applied foreign language research into issues of national concern and provide channels for dissemination of research results

There are more, but your host has selected these to demonstrate the broad scope of the proposed mandate for this bureaucracy. If enacted -- and with all sorts of Pearl Harbor-style comparisons to Sputnik, this is as excited as government bureaucrats get -- this will completely change how languages are taught and how funding priorities are ordered in language-related work in the United States.

So are linguists in on the conversation? Judging by the roll of speakers, apparently not. Right now, it's all university and high school administrators, people named "Ambassador", "Undersecretary" or "General", and business executives. A couple of participants were involved with applied linguistics departments, but other than that, linguists are not involved in this emerging community.

The question that remains is: should they be? The answer isn't so much "yes" as "where to start?". Linguists have been lamen