Alan Jacobs

What these errors add up to is that Watson really cannot process natural language in a very sophisticated way — if it did, it would not suffer from the category errors that marked so many if its wrong responses. Nor does it have much ability to perform the inference required to integrate several discrete pieces of knowledge, as required for understanding puns, jokes, wordplay, and allusions. On clues involving these skills and lacking search-engine-friendly keywords, Watson stumbled. And when it stumbled, it often seemed not just ignorant, but completely thoughtless.

I expect you could create an unbeatable Jeopardy! champion by allowing a human player to look at Watson’s weighted list of possible responses, even without the weights being nearly as accurate as Watson has them. While Watson assigns percentage-based confidence levels, any moderately educated human will be immediately be able to discriminate potential responses into the three relatively discrete categories “makes no sense,” “yes, that sounds right,” and “don’t know, but maybe.” Watson hasn’t come close to touching this.

In short, Watson is not anywhere close to possessing true understanding of its knowledge — neither conscious understanding of the sort humans experience, nor unconscious, rule-based syntactic and semantic understanding sufficient to imitate the conscious variety. (Stephen Wolfram’s post accessibly explains his effort to achieve the latter.) Watson does not bring us any closer, in other words, to building a Mr. Data, even if such a thing is possible. Nor does it put us much closer to an Enterprise ship’s computer, as many have suggested.

One might argue that literacy is unalloyedly a good thing – yes, I can think of counter-examples, but then again one always can – but it is pretty clear to me that reading, as in reading of literature, is not. What we read can affect us vitally, penetrate, stimulate and inform us, but not always in the right ways, or at the right times, or about the right things. If you think that reading the right things in the right ways is morally bracing, improves one’s discriminations and heightens sensitivity – basically, the Leavis line – then all you have to do is look at the behaviour of Dr Leavis himself to begin to doubt the thesis. Indeed, if it were true that wide and deep reading redounds wholly positively on the development of a wholesome self, consider a typical member of a university English department, and despair.

I feel like shrinking now, I am so embarrassed for what I have done and how many people I offended. I always meant for my work to offend the powerful and give comfort to the weak. Yesterday I did the opposite of that. I, of all people, should know the power of words and the power of social media. In a few minutes of insensitive and sadistic banter meant to childishly provoke a few acquaintances my thoughtless words were seen by many thousands, and hurt many thousands. In addition I have now associated people I know and causes I care about with my own words, which make it seem like I could ever support assault on a woman, or anybody, which I cannot and would not. I have failed and disappointed many people and hurt many people and I am deeply sorry for that and I wish I could convey my apology to them all.

Conventions of politeness aren’t based purely or even primarily on functional considerations—putting one’s elbow on the table doesn’t affect the taste of the food—and they are apt to become tools in the hands of pretentious fools. Yet the proposition that they should therefore be done away with is manifestly preposterous. Certain conventions will govern the ingesting of food in a civilized society. Humans abide by customs when they wish to please their companions or ingratiate themselves, and there’s no point in trying to convince them to stop.

So it is with the prescriptivist temperament in language. The employer will always want to know whether a job applicant can write and speak the way educated people write and speak, considerations of clarity notwithstanding. The ambitious student will always take secret pleasure in saying “It was she” instead of “It was her.” The conscientious parent will always encourage his children to mind their grammar, and partisans of various stripes will always cackle when disfavored politicians fall afoul of the rules (“The goals of this country is to enhance prosperity and peace,” “Give Al Gore and I a chance”). Descriptivists like to think of themselves as clear-eyed realists. But it’s they, and not the sentimental traditionalists, who wonder why we can’t dispense with all these arbitrary “rules” and just get along. Is there a dream more fanciful?

Perhaps the most important thing to be said about the Art Project is that despite all the talk of interactivity and the application of Google’s street view technology to the museumgoing experience, what we have here is just a new version of an old dream—the dream of bringing an ever expanding audience into contact with works of art. The greatest museums, some of which are featured in the Art Project, are dedicated to that democratic promise, which is a promise not that everything is for everybody but that anybody should have the chance to experience anything. I am passionately committed to that principle, but even when more people do indeed have access to more works of art, we are still left with the nature of the individual’s experience, which is finally the only experience that matters. Some will say that the Art Project allows users to customize their experience, at least up to a point, by deciding which galleries in which museums they will visit, and how they will circulate around the rooms. But advances in technology do not necessarily make art more democratic. And in any event Google places severe restrictions on what we are able to see. At the Museum of Modern Art we gain access to a gallery of late-nineteenth- and early-twentieth-century works, including Cézannes and van Goghs, but the adjacent gallery, full of Picassos, is just a blur, because his work is still under copyright, and thus not available for this particular tour. The Google Art Project may have the same relationship to museumgoing that Facebook has to friendship. It will take you only a small part of the way.

All right, dammit, I’m back. Posterous is seriously problematic.

Friends, enemies, and followers —

I’m giving up on Tumblr, which I have only been able to access intermittently for months now. Henceforth, More Than 95 Theses will be on Posterous. At least until I change my mind.

I am grateful that, thanks to the miracle of RSS, I’ll still be able to follow my favorite tumblelogs.

See you over at Posterous, I hope!

The New Atlantis (2)

Fabulous chalk lettering by Dana Tanamachi

Adiaphora: Our Refusal to be Spoken To