You are currently browsing articles tagged language.

@everyword in context

Last week Adrian Chen conducted an e-mail interview with me about @everyword. Here’s the resulting article on Gawker. The @everyword account gained about a thousand new followers as a result of the article—not bad for an account that just tweets word after word every half hour!

It’s been interesting to read people’s reactions to @everyword (and yes, I have the Twitter search for @everyword in my RSS feed reader, because I am hopelessly narcissistic). For the most part, the reactions are positive! It’s satisfying when someone is amused by a word that they didn’t know existed (or that they hadn’t considered to be a “word”) or when someone finds unexpected synergy between a word that just got posted and something that is happening in their lives.

Some of the reactions are more critical. Here’s one reaction in particular that I wanted to respond to, from Twitter user @fran_b__:

@everyword They aren’t words unless they have meaning, which implies context. Stripped of context, they are simply (python) string arguments. (source)

This response baffled me, because in my mind @everyword is all about context. For example, here’s the way that I typically read @everyword:

everywordcontextThis is a screenshot of my Twitter client on a typical morning. You can see the tweets from @everyword interleaved in the feed. I don’t generally read the tweets in my feed like I would paragraphs or sentences in an essay or a piece of fiction (e.g., I skip tweets, I don’t necessarily expect cohesion from one tweet to the next), but I do tend to read them in sequence. It’s undeniable that the tweets exist in the same physical context here. Because of this, some interesting possibilities for creative reading crop up. It’s easy for one tweet to “color” how nearby tweets are read, for example. I’m not saying that @notch is prone to nutations, or that @factoryfactory and @daphaknee are nutcases, but that’s certainly a reading made possible by the tweets’ close proximity.

There’s also the context provided merely by being in sequence with other words in the @everyword feed. Here’s an example:


I find this endlessly fascinating. When you see these words juxtaposed like this, you can’t help but try to find some connection between them. In some cases, the connection is grammatical (nunnery is of course morphologically related to the word nuns). But nunsnuptials and nursemaid together like is almost like a little narrative. “Nuns can’t have nuptials, and they certainly can’t be nursemaids.” It seems ironic that the words would be juxtaposed like this, and that perception only emerges from seeing these words in this kind of unusual context.

It’s also a cultural practice of ours to consider individual words in the abstract: we pick out our favorite words, we decide which words are commonly misused, we decry our politicians for making up words or using words with a disagreeable frequency, etc. In some sense, a word carries with it a cultural context, no matter where it occurs. One of the intentions of @everyword was to play with this idea: every word has cultural baggage. What would happen if we systematically exposed ourselves to that baggage?

Even if I concede that the words in @everyword are “simply (python) string arguments,” isn’t that also a context? A computer program is a kind of writing, after all. It means something for a programmer to choose to put one string in a program, instead of some other string, or to feed some set of data to a program instead of some other set. Sure, the Python program that runs @everyword would also work with any other arbitrary data set—@everybaseballplayer, anyone?—but the fact that I chose words, and words in this particular order, is part of the context of the piece.

In the end, I think @fran_b__’s implication is that there are certain kinds of contexts that a word can occur in that “count” as meaningful (such as being in a sentence intentionally composed by an individual) and others that don’t. I suppose that for certain fields of study, this is a valid point of view: if you’re analyzing a novel, for example, you might not want to include in your analysis the novel sitting next to it on the shelf. As a writer and poet, however, I find that limitation pretty dull. There’s never been an era in history with such diverse practices for reading and writing text. Why not have as much fun with that as possible?

Tags: , , ,

The Glasgow Science Center is currently exhibiting The Joking Computer, a kiosk-based installation running software made by artificial intelligence researchers at the University of Aberdeen. The software uses phonetic information about English words and semantic information from WordNet to generate pun-based riddles. (More information about how it works.)

A few of my favorites:

What kind of tree is nauseated?
A sick-amore

What do you call a cross between an emporium and a success?
A department score

What do you get when you cross a choice with a meal?
A pick-nic

More examples here, and an online version is promised soon.

I like to read programs like this—programs that generate text conforming to a specific genre—as a kind of ethnographic criticism. The Joking Computer describes a particular type of joke by observing how jokes of that type are formed and used in our culture, and then formalizing the jokes as a procedure. The procedure itself serves as a statement about how that genre of text works: its structure and its limitations.

The Joking Computer specifically, and text generators in general, also manifest the nature of the data that they’re built upon. Take this joke (please):

What do you call a washing machine with a september?
An autumn-atic washer.

This joke shows the gaps that exist in the program’s data, and the unique way in which the program uses that data. First, the program doesn’t have a way to know that a washing machine isn’t a kind of thing that is likely to be “with” a “September” (or that September isn’t a noun likely to be used with an indefinite article). Second, the relationship between “September” and “autumn” depends on the eccentricity of WordNet, which claims that there is a meronymic relationship between the two concepts. The joke is constructed on the basis of the fact that September is a “part of” autumn—which certainly makes a kind of sense, but isn’t necessarily something that most people would intuitively agree with. The joke, as a consequence of (at least) these two factors, seems stilted and alien.

Then again, stilted or not, I happen to think this joke (“autumn-atic washer”!) is hilarious, and that the humor stems at least in part from the gaps in the data and way the program uses that data. Jokes, after all, are funny when they provide surprising juxtapositions or reconceptualizations of things in the world, and the program delivers those in abundance.

Poems succeed when they make use of similar juxtapositions and reconceptualizations. I think that this is why generative text programs are most effective when they are designed to generate text in these genres (humor and poetry). These programs succeed just because they don’t perfectly model the genre they set out to emulate.

In other genres, like conversation or narrative, surprising juxtapositions are less valued, or even specifically forbidden. I think that generative algorithms in those genres tend to be less successful for this very reason. But that’s a subject for another post.

(The Joking Computer via, more info)

Tags: , ,


Wondermark‘s take on supernatural terms of venery. Not to spoil the punchline, but my favorite by far was “the Borg.” Grammatically, it works: there is no reason to ever refer to the Borg, except when referring to it collectively.

Tags: , ,