Monday, December 17, 2007


I know it's been a while since I last posted. It's finals week, which means I'm basically studying all the time. I have a bunch of posts in the works,

I just took my first final, physics (8.02, intro electricity & magnetism). I felt really good about that one; everything seems to have come together really well and comprehensibly in my head in the past several weeks. I give the credit to the professor's very intuitive style of teaching and understanding everything, and to his fondness for conceptually difficult questions with easy math. That is how to teach physics, IMHO. (I have a lot more to say about the course format, TEAL ("technologically enabled active learning"), but that's a whole nother post.)

I'm not really so worried about tomorrow, when I have my midterm in calculus (18.02A). I'm in a kind of funnily-put-together course, where you spend the first six weeks reviewing single-variable calculus at high speed, and then follow the regular curriculum for multivariable calculus, finishing over the January independent activities period. It's primarily intended for people who took AP Calculus AB. I took AP Calculus BC and passed the AP exam with a high score, which would qualify me to jump straight into 18.02 (straight multivariable calculus), but at the beginning of the year I felt like I needed the review. I could have done without it, but it was nice to have. The only problem was that it put me farther behind in terms of math lagging physics (a perennial problem), but the math in my physics course is not actually that hard (e.g. all the line or surface integrals we ever have to do are simple cases that reduce to multiplication problems).

The one I'm worried about is chemistry (5.112), on Wednesday. Most of the content on the test will be drawn from the last several weeks of the course, and that's the material I understand the least. The course is taught by two different professors -- one does the first half and the other does the second half -- and the second guy is very difficult to understand, for various reasons which I will elaborate in the future. It's a blessing that it comes last, really; I have about twelve hours this afternoon/evening for studying, and I don't intend to devote all of them to calculus.

MIT really treats its students very well during finals week. There are free breakfasts in all the dorms, as well as one in the lobby of the building where the biggest finals are held. (And not cheapo lousy free breakfasts, either -- good ones. This morning, for the first time all semester, I had eggs and pancakes! It was awesome!) There are a few free lunches and dinners scattered around, and all the food vendors in the Student Center are giving discounts. Alpha Phi Omega holds "Finals Coffeehouse", with snacks in a room of the Student Center that used to be a 24-hour coffeehouse but is now just a sort of miscellany room with tables. MIT Medical holds relaxation seminars, even. This is characteristic of MIT, really: they pound you into the ground academically, and then they reach down and lift you back they can pound you again.

More content to come over winter break. I've got a lot of saved-up commentary on various aspects of the teaching here.

Sunday, December 9, 2007

What's This `Conlangery' Business? [Linguists and Conlangers, part 1]

[This will be part of a series of posts about conlanging and its interaction with academic linguistics.]

So, what is conlangery? It's not a popularly known art, so here's a primer for y'all who haven't ever heard of it.

The word conlang is short for "constructed language". Well-known conlangers include Ludwig Zamenhof (Esperanto), Marc Okrand (Klingon), and the revered JRR Tolkien (who once wrote that he invented Middle-Earth largely in order to give his beloved Elvish languages a place to live). We invent languages: spoken languages, signed languages, artificial siblings and alternative scripts for natural languages, artistic languages, logical languages, international auxiliary languages, mind-extending languages, you name it.

(I should warn you that conlangers, being naturally fond of playing with words, habitually `overgeneralize' and apply all kinds of grammatical forms to the word `conlang'. Just for starters, it's both a noun and a verb. A conlanger is someone who conlangs, what they create is a conlang, and the art in general is conlanging or conlangery. On top of that, the con- prefix has become productive, meaning it can apply to all kinds of invented/imagined things: conlangs, conworlds, conscripts (meaning writing systems, not military draftees), conreligions, consolarsystems, conbiology...the list goes on.)

The earliest known conlanger is Hildegard von Bingen, a 12th century German abbess. She's best known for her gorgeous music, but she also invented a language called Lingua Ignota ("unknown language"), which supposedly came to her by divine inspiration.

The best known conlanger...well, it depends who you talk to, but most people will name either Zamenhof or Tolkien, and here we come to a major split in the world of conlangery: that between auxlangers and everybody else. Auxlang is short for "international auxiliary language", and not to make any sweeping generalizations here, but the sort of people who make auxlangs are also the sort of people who have fiery and/or highly unconventional political agendas, and a lot of them are also shamelessly self-promoting. Unfortunately, if you're auxlanging in earnest, this isn't really fixable. You have to want the entire world unified, in some sense, by your language. And you have to think your language is not only `good' enough for everyone in the world to speak, it's `better' than the dozens and dozens of old failed auxlangs. A certain amount of crankery is inherent to the practice of serious auxlanging.

(Quite apart from the above, the vast majority of linguists and conlangers have a horrible visceral reaction to the thought of losing most -- if not all -- of the world's linguistic diversity. Who wants to replace all that culture, knowledge, and sheer beauty with something necessarily bland?)

(Yeah, and, in case you couldn't tell, I'm totally biased, and I freely admit it. Auxlangs are not all bad; any conlangery effort bears some useful fruit.)

So what about the `rest' of the conlangers? People conlang for all sorts of reasons, but the one that really unites us is, trite as it sounds, is love: love of language, its beauties, its intricacies, its elegances; and love of playing around with that in systems of our own creation.

This is not to say that there aren't divisions within the body of non-auxlangers. Probably the biggest group is the artlangers, like Tolkien, who conlang for aesthetics and elegance. (This is not to say that all artlangs end up looking like Tolkien's Elvish languages, all full of L and R and vowels everywhere with nary a `guttural' consonant. Conlangers have as much variation in taste as the general population. Plus, for example, there's a lot to be said for a language in which you can swear effectively. Imagine trying to shit-talk someone in Quenya or Sindarin.) At the very least, many hundreds of artlangs have been made, or at least started. My own language Tlharithad is a young artlang, though I sadly haven't had time to work on it since the beginning of the semester. You can find quite a few well-developed artlangs, associated with the fantastic conworld of Verduria, at Virtual Verduria, made by Zompist. I have to apologize to the many, many, very worthy artlangers that I haven't linked to, but for non-conlangers, Zompist's work makes a good place to start.

On the flip side of the same coin, you have the engelangers -- engelang is short for "engineered language" -- who design their languages to achieve a particular goal. There are logical languages, like Lojban, which are designed to eliminate ambiguity. And there are other engelangs, whose design goals don't really fall naturally into groups. A seminal example is Ithkuil, which has about five times the information content per syllable of natural languages. In the words of its creator, John Quijada, Ithkuil is "systematically designed to blend a high degree of communication of cognitive intent and meaning with a high degree of efficiency, i.e., to allow speakers to say a lot in as few syllables as possible." While I'm not intimately familiar with the language, I know the making of Ithkuil involved a lot of mindbending reorganization of cognitive concepts. For example, if you're indoors, the spatial axes around which you organize your speech are placed with respect to the long axis of the room! (Of course, JohnQ freely admits that Ithkuil is extraordinarily difficult to learn, and has in fact created a somewhat simplified version, called Ilaksh, for those of us without superhuman vocal tracts.)

It can be argued that artlangers are just engelangers whose design goal is that of beauty, rather that something more conventionally associated with the words "design goal". It's a little like the distinction between architects and sculptors: you have people who are clearly one or the other, people who are mostly one with a little of the other, and people like Michelangelo or Maya Lin who straddle the boundary so well that no one ever finishes arguing about which category they fall into. (And, as good linguists, we have no problem with this, knowing that strict definitions are artificial constructions, and everything is better described in terms of generalizations from prototypes.)

Next up: the enmity, such as it is, between conlangs/conlangers and professional linguists / linguistics research.

Thursday, December 6, 2007

Another essay on killing lab mice

Ran across this essay, which aired on NPR's All Things Considered five years ago. I especially liked this bit:
When I started, sometimes I had to go walk around the lab building afterwards to take a breath and gather myself. Now it still upsets me whenever I have to kill mice, but I'm used to it. I think that probably some scientists do get over it completely; the mice become tools. Other scientists hire technicians to handle and kill the mice. The researcher works with the cells of the mouse once it's killed, but never encounters the living mouse.

But that's not how we do things where I work. In fact, the people I get along with best in my lab sometimes even talk to their mice. While we're herding a mouse towards one side of a cage or another we might say, "Come on, sweetie." Maybe if we're injecting a mouse with something and it squirms, we say, "OK, OK. It's going to be just a second." When we put it back in its cage, we might say, "There you go." And while the mouse is sniffing and inspecting its cage mates, we say, "There's your buddies." I was in the mouse room with a colleague of mine and I noticed her doing this. And I said, "You talk to your mice, too." She said, "Doesn't everybody?"

It really captures the tension between mouse-as-tool and mouse-as-animal. And the -- well, I don't want to call it a happy medium, but the medium that most good scientists find. I dare say that working exclusively with cells, and never encountering the organisms they come from, has got to be dissociative. Seems like it'd be a good idea to keep that perspective in the back of your head, that these little spots in a dish came from a living animal, and serve multifarious purposes in that animal, and do things besides sit in culture and express your marker protein or what have you. They grow, they thrive, and above all, they are more complicated than they seem, even when you take that last into account.

A scientist who gets too debilitatingly upset over the death of a mouse will never get anything done, and a scientist who doesn't care about the mice as living creatures will have less perspective and, in all likelihood, get worse results for not being personally involved with their care.

I don't recall ever talking to the mice, but I didn't handle them all that much. Pretty much all I did was scruff and snip, and then they couldn't listen anymore. If I have to do more involved mouse procedures in the future, I probably will get in the habit of talking to the mice, especially given that I already even talk to inanimate objects when I'm working on them.

Stir-fried wikipedia, anyone?

The highly estimable Language Log points out yet another instance of Chinese menu translator ingenuity. Wikipedia: now with even more uses!

For more on this phenomenon, see (marginally NSFW), which is devoted to collecting this sort of unintentionally hilarious bad English, mostly from Japan. Aside from just giggling at the bad translations, it's interesting to occasionally catch a nugget of linguistic insight. Well, I mean, the fine folks at Language Log can catch them all the time, but I'm an amateur, so I have to take what I can get.

A while ago, featured a shirt (which I unfortunately can't seem to find now), bearing the sentence "We are dumb and haven't intelligence apes." And I jumped out of my seat, because it made perfect sense to me. In English, that sentence means "we are dumb and do not possess apes of intelligence." Presumably the shirt-writers meant "we are apes who are dumb and do not possess intelligence." (Perhaps dumb here means mute rather than unintelligent -- a relatively subtle distinction, but I've seen Japanese<==>English dictionaries that do very well at this. Or, y'know, maybe it's just redundant.)

What English handles as a relative clause, "apes who do not possess intelligence", Japanese handles by effectively turning the verb "not possess intelligence" into an adjective. To say "we are apes who are dumb and do not possess intelligence" in Japanese, you say something along the lines of "we are dumb and non-intelligence-possessing apes", which is clearly the origin of the T-shirt.

Of course, there's also the complete nonsense, e.g. "I smell the smelly smell of something that smells smell", and the grammatically-correct-but-thematically-inappropriate text, a la this classic example.

Monday, December 3, 2007

So, where do animal protocols come from?

My dad asked, in a comment to "My Experience with Lab Mice", why the accepted protocol is to gas them with CO2 instead of nitrogen. He writes:
Anoxia through O2 starvation is demonstrably painless. Many research pilots go through it, to the point of unconsciousness, and the general comment on recovery is "Did something happen?"
CO2 overdose triggers the breath reflex, which O2 starvation does not, at least in humans.

I did a fair bit of searching, but all I could find was protocols describing various euthanasia methods. I couldn't find anything that motivated them. Sure, I found some discussion of why toe clipping is discouraged, but that's because toe clipping isn't The Method for identifying mice (anymore). CO2 gassing appears to be one of The Methods, if not The Method, for euthanizing a bunch of mice -- hence, not much debate or discussion.
I don't know how many readers I've got, but I'm throwing the question open: who decides what the `official' animal protocols are? (Does it differ if you're a university vs. a company?) How do new methods get invented, approved, and adopted? What happens when an alternative method gets officially discouraged/banned? Where can one go to find out all this information about a specific protocol? And if it's the case that O2 starvation by N2 surfeit is painless, why isn't it The Method?

Gmail, how I love thee

Forgive me for taking a moment to extol the virtues of Gmail. I appreciate it a lot more since coming to MIT, because I get about ten times as much email as I did in high school. And no, I'm not getting paid for this. But people here complain a lot about the volume of email, and if more people would use Gmail then the amount of sniping would drop five-fold.

Gmail has tags instead of folders. Tags are so much better than folders! I have two major ways I categorize my email: by where it was sent to (me directly, my MIT email account, certain mailing lists) and by why I should save it (it points to a resource, it contains someone's contact info, it's one of those "please retain this email for your records" messages). So, to a first approximation, pretty much all the messages I save falls somewhere on a two-dimensional grid. A lot of messages fall in more than one place, and there are miscellaneous tags, and all kinds of stuff you just can't do with folders.

(I'm always a little annoyed by blog platforms that say things like "This entry filed in [list of tags]". They're just trying to pretend tags are the same as folders, no idea why. Inertia perhaps? Bah! Embrace the tags for what they are!)

Since Gmail has tags instead of folders, to get things out of your inbox you just hit "Archive". This puts the conversation in a big all-purpose bin, and you can find it later by searching for text, tags, or sender.

There's also a really nice filtering setup. For instance, I have all my email from my MIT address forward to my Gmail, and I have Gmail tag it `MITmail' so I know what was sent to where. I subscribe to a lot of mailing lists, and I have many of those automatically tagged as well. One list in particular, the Reuse list (for recycling / handing off old computers, furniture, unneeded coupons, etc etc etc...), gets a lot of traffic, and I don't generally want to see Reuse messages in my inbox unless I've got free time to go pick something up. So I have Gmail do two things: tag the messages `reuse', and archive them directly, so they don't appear in my inbox and get in the way of more important things.

On top of tags, you can also mark converstions (or individual messages) with a little yellow star that shows up by the subject line. This is really helpful because I keep several types of messages in my inbox (instead of archiving them): reminders for events, reminders for things I need to do, and messages I'll need to refer to within the next two weeks or so. Reminders get starred, so they stick out visually and I actually get reminded of them. References don't get starred because I only need them when I'm looking for them specifically.

Probably the best feature, though, is this: Back-and-forth emails with the same subject line are organized into conversations. This is just fantastic. It keeps everything related to one topic in one place, instead of having individual messages in several threads scattered randomly throughout your inbox. You can read an entire thread on one page. If a new message appears while you're backreading the thread, Gmail pops up a little "Update Conversation" box so you don't reply redundantly. And since the whole thread is on one page, Gmail hides each message's quoted text, all the lines that begin with piles of >>>s. And you can delete an entire flamewar with one click. Can't beat that.

I know a lot of people are concerned about the direction Google is headed, or that they will end up "owning all the information in the world". Yes, there are legitimate concerns, but I think the danger is way overblown, and Gmail seriously saves me a lot of time and aggravation. (Are you listening, stupid Yahoo email account that I keep for signing up for potentially spammy things? Grr.)


Many, many thanks to Coturnix, of A Blog Around The Clock, for linking here!

I realize this happened several days ago. I plead guilty of falling out of the habit of checking my Technorati page because nothing ever happened on it.

And, I may as well take this opportunity to apologize for the recent lack of activity. Homework is no excuse; there are lots of people out there who are busier than MIT undergrads (though you'd have a hard time getting some of my friends to agree to that)