Adsotrans

Bizarre entries from the Linguistic Data Consortium still surface every now-and-then like sea creatures from the deep. The latest I noticed was a surprising definition for 深入研究. The dictionary had it as “lucubrate”, which WordNet defines as “to elaborate on”.

True to form, both Dict.cn and Kingsoft appear to have copied the entry wholesale without lucubrating too much over accuracy or attribution. Still, a definite step-up for them from the depths of “nigger-brown“.

What do people feel about switching the output of the “popup annotation” (the default output from the Adso home page and the popups on newsinchinese.com to a more comprehensive definition (something that includes all of the glosses in the database)? Right now Adso chooses the most likely definition/pos given its understanding of Chinese grammar. I’m not sure that most people notice this selectivity until the system actually gets something wrong….

The reason the system is built this way is that Adso aims to offer gist translation and other sorts of generic semantic analysis functionality. Being able to hone down a word to a single definition is much more useful for search/semantic/translation applications than just showing a list of possible definitions. I strongly believe that structuring data in a way that enables this is important for the ability of the open source community to innovate in the long-term in this space.

But maybe this is the wrong approach for data-display, and especially for a project that aims to make it easy for people to collaborate around language. I’ve noticed that a lot of users (especially new contributors) like to pack multiple definitions into the popup windows rather than add new ones. I usually edit these during the review process, but don’t want people to feel disappointed if their edits “disappear” from one release to the next, especially if they aren’t really gone, but merely restructured. Also, if people really WANT more comprehensive definitions when they use an annotator, it is easy enough to change the system and get Adso to produce that sort of output. It’s actually much faster to do as well since this eliminates the hard work of grammatical disambiguation.

Curious if anyone has any thoughts on this, especially since the next review is coming up fast and it looks as if we’ll have more than a thousand new submissions.

Even those whose sole experience with China involve booze-fueled stopovers are probably familiar with the local custom of ignoring lines, regulations and courtesy except insofar as they provide opportunities for graft. This is cute for a while in the 入乡随俗 spirit of adventuring, but gets old fast once you realize there aren’t many other places where customs involve getting spit on while biking to work.

But rushing past that, let me devote this post to complaining about something I’ll call the Shanghai cab-shuffle. This behavior isn’t unique to Shanghai, but it’s particularly egregious here, perhaps because it’s impossible to get a cab in normal circumstances and desperate times call for desperate measures when you’re hustling through life and need to get to a club pronto.

Those with experience in northern China are probably nodding, and thinking “nothing special here”. And to be fair, the shuffle is not unknown in the northern hinterlands. You see it a lot during traffic jams or Mongol invasions when social order collapses and the Beihai economy picks up. Even in these circumstances though, I think the northern shuffle is a more furtive beast, and you’ve got a decent chance of shaming someone by calling them a dick in Chinese. In Shanghai, I don’t think anyone would bat an eyelid at this — screwing your neighbour is not only commonplace, it’s expected.

Put another way, Shanghai elevates hostile, feigned oblivion to an art. After all, getting a cab is not exactly a subtle diversion. At a minimum, it involves wading into the bike lane and pumping your arm using the sort of repetitive motion that centuries of evolution have trained human eyes to notice lest it involve wolves leaping out from undergrowth. In Shanghai, the atavistic response to seeing this is to walk two meters over and try to flag one down first. This is such pointless behavior it’s baffling to see it on a large scale: does no-one ever just snap in this country? But then astonishment turns to something else when you realize that people here really do value saving a bit of time over visibly and unapologetically screwing their neighbour. In a sign that irony might be divine but stupidity is definitely human, I noticed each of the people who shuffled me tonight getting progressively angry at being shuffled in turn. But when the fifth person showed up, something happened that I’d never seen before….

I saw hesitation on his face.

Had courtesy for the common man crept into someone’s heart? Was he wondering WHO he should stand behind? Curious which of our four lines was “moving fastest”, so to speak? Not at all! He was panicking because he couldn’t inch any CLOSER to oncoming traffic without wading right out into the road and getting flattened by a bus. Which was when I started walking home, contemplating that any city that requires you to wage pointless and silent battles for street positioning needs a heart.

In the summer of 2006, several of us in Beijing launched a project called “Adso Textbook”. The goal was producing an archive of manually-annotated selections from classic Chinese literature. I’m proud of the results, and happy that the readings are back online as part of a freshly-redesigned NewsinChinese.

That’s the other big news. After about a half-year hiatus, NIC is back online and hopefully improved as well. I’m still tweaking the site, but am happier with this version than the old one. The backend of articles isn’t updated as frequently as it was before, but the articles are all fully annotated in advance this time around, making it much more convenient to view full-text news.

As for the selections from stories, my two favourites are Dream of the Red Chamber and Lu Xun’s A Little Incident. My favourite poem on the site is So Gently I Go. The selection from Dream of the Red Chamber is particularly cool for making late Qing prose accessible and acting as a functional tutorial for non-modern texts. Once you have the basics down, it *is* possible to make headway in the book without smashing your head against a wall.