Archive | Words RSS for this section

Gathering in the town square


The MIT Technology Review has a short article on one of my main projects, called Meadan, which I’ve not discussed at all on this blog. Mostly this is because we’re still in development, but recently we embarked on a closed alpha test phase so this seems like as good a time as any.

From the article:

The basic idea is simple: it’s a website that brings English and Arabic speakers together around daily postings of news articles, broadcasts, and events that are of common interest, and it gives users a platform to communicate through dialogues, blogs, and other exchanges. All the while, it allows users to pinpoint their location so that people can share views across continents. The hard part is creating a system that allows users to express their ideas in their native tongue.

What’s really interesting about Meadan — apart from its small part in removing barriers to rationale discourse between the West and the Muslim world — is how it uses social networking technologies both to create communities (the “traditional” use of social networking) but also to enlist users to rate, edit, and correct the English-Arabic machine translation. Social networking as language feedback loop. What’s missing from so much machine translation is a sizable corpus of informal conversation (not bizspeak or medical parlance, for instance). This is what helps MT learn grammatical and dialectical nuance and this is precisely the kind of conversation we envision on the Meadan site.

Much more on this in the coming months, of course.

Thamus (partially) vindicated

If you read yesterday’s post early in the day you may have missed that Richard Powers responded to it. I’m grateful for his very thoughtful reply. But what everyone wants to know is: how did he find the post? You mean, Richard Powers — MacArthur Genius, National Book Award Winner — reads John Tolva’s blog?

In a word, no. I e-mailed him and alerted him to the post. No buzz, no meme, no trackbacks, or digg swarms came to the attention of Mr. Powers. Just an old-fashioned note in his inbox that he graciously acted on.

Actually the story behind the reply is somewhat amusing. Powers initially commented but nothing showed up on the site. Usually that happens when the anti-spam script kicks in. But why would it block a regular comment? I was stymied and more than a little irked that this author had taken the time to respond and my site had black-holed his effort. And then it occurred to me that Powers was likely composing (or speaking) his response into an external application and then pasting it in-bulk into the blog comment form. This is a red flag for the spam script, since that’s exactly how bots dump garbage into blog posts, as a single pre-written chunk. I asked him to paste it in and then type a little at the end. You know, act like an old-fashioned human writer. Type a little. It worked. Congratulations, Mr. Powers. You appear to be human. Your comment shall be accepted. Turing test passed. (If you recall my in-person encounter with Powers you will find this as ironic as I do.)

Truth is, I’m more in awe of Powers’ talent than ever now that he’s erased any doubts I had about his composition-by-dictation (on a keyboardless machine — sheesh). I think my incredulity stems from the way I consider words spatially: objects on a page to be moved, sequenced, and arranged into thought. Almost like the visual arts. That’s a pretty narrow way to think about language, of course, but it has taken this little episode for me to realize just how much the tools of word manipulation I use form what I write. Perhaps even constrain what I write. I guess I need to fire up the speech reco on this MacBook and find out.

Speaking of, so to speak, it might be interesting to listen to The Echo Maker as an audiobook, if it exists. (Yikes, $120. What the hell?) Does it read out loud better because it was composed out loud?

If you are interested, here’s the tablet that Powers references in his reply. Litgeek!

One wonders what jacket-blurber and long-time Powers fan Sven Birkerts (of neo-Luddite Gutenberg Elegies fame) thinks of all this. The fate of writing in the electronic age!

By the way, if you liked The Echo Maker’s exploration of memory (and stories that begin with a mystery-shrouded car crash) you must read Michael Joyce’s seminal hypertext fiction afternoon, a story. A comparison of these two works would be interesting indeed.

Oh, and Prof. Turnbull you should engage Prof. Powers. If anyone can squeeze a publication from this, it’s you. 🙂

Thanks again, Richard!

Echoes and real voices


I recently finished the latest novel from Richard Powers called The Echo Maker. It was one of the finest books I’ve read in years. Powers is by leaps my favorite writer. His books are poems trapped in the novel form. The craft with words every bit as compelling as the stories they tell. I’m a big fan.

But there’s something different about this novel. It is still superbly crafted for sure, but the narrative engine revs louder. There was something about it. Something I couldn’t quite identify. As I was reading the book a fellow Powers fan friend of mine alerted me to an article in the NYT (login required), written by Powers, about how he composed the novel using only speech recognition software on a tablet computer. That was it, I thought. This must be the stylistic difference. A novel voice-crafted versus hand-crafted.

Except. The more I think about it, the more I can’t believe it. I work for IBM, a company deeply committed to speech recognition, text-to-speech, and machine translation. It is hairy, complex computing — bordering on AI. Personally I’ve been working in Arabic-English translation since 2000 and I know just how thorny the problems are in getting good recognition. I simply can’t believe an author as talented as Powers could create a book as linguistically complex as The Echo Maker using speech reco alone.

I’m not saying he’s lying. I’ve had some interaction with Powers, all positive. He kindly responds to e-mail, for one. And yet, there’s precedent for this tale of novel-by-dictation being fiction too. In 2002 at the Chicago Humanities Festival Powers delivered a talk called “Literary Devices” about an ELIZA-like machine that sucked him into an e-mail conversation that was as real as any human author’s output. I bought it. Most bought it. We bought wrong. The story itself was fiction — which only made it better. Humans falling for a story about a machine that tells stories indistinguishable from human stories. Amazing.

So, I guess I’m asking this. Mr. Powers, did you really dictate this whole novel? Or should we nestle comfortably in what is admittedly a damn good story even if you didn’t? All half-dozen readers of this blog are dying to know. And if we can’t tell your response from a computer impersonator we’ll obviously consider the dialogue valid. Do tell!

UPDATE: Powers replies. Wow. More on this in a bit …

See also the follow-up post Thamus (partially) vindicated.


You probably didn’t notice that I removed my blogroll recently. That’s not becuase I’m no longer reading other sites, but rather because I follow so many sites and my interest shifts so often that it seemed silly to call out a subset for special note.

But there’s one site that rises above the rest. Maciej Cegłowski’s Idle Words is this site. Ceglowski is a polymath polyglot and one hell of a writer. He seems partially powered by wanderlust, a tendency that gives his posts a rewarding freshness. Here are some excerpts.

On Chinese Mooncakes:

Mooncakes, of course, are the exact cultural analogue of the American Christmas fruitcake, that venerable Christmas pastry of astonishing density that brings people together by uniting the giver and receiver in a shared reluctance to eat it. The Chinese have not yet advanced as far as those intrepid Americans who store a received fruitcake for a year before re-gifting it to another victim, but there are promising signs that the failure to let mooncakes overwinter may just be a function of limited apartment storage space, solvable by applying economies of scale.

On flying from North America to Asia:

If you are a package of avionics software, the North Pole is a stressful place. Depending on how close by you pass, longitude and bearing can change extremely quickly (or converge into an unlucky singularity) and most autopilots throw up their hands and enforce a special wings-level lockout flight mode within a few miles of the pole, to keep from spiraling around it like a housefly circling a light bulb.

On NASA’s aimlessness:

NASA dismisses such helpful suggetions as unworthy of its mission of ‘exploration’, likening critics of manned space flight to those Europeans in the 1500’s who would have cancelled the great voyages of discovery rather than face the loss of one more ship.
Of course, the great explorers of the 1500’s did not sail endlessly back and forth a hundred miles off the coast of Portugal, nor did they construct a massive artificial island they could repair to if their boat sprang a leak. And we must remember that space is called space for a reason – there is nothing in it, at least not where the Shuttle goes, save for a few fast-moving pieces of junk from the last few times we went up there, forty years ago. The interesting bits in space are all much further away, and we have not paid them a visit since 1972.

On running the NY Marathon:

Mile 19
We’re at 115th street, and the crowd has thinned considerably. My legs are much more tired than I expected, and getting stiff – I stop at a water stand, and walk a block before running again. The next five miles will be walk-and-run, trying not to let my legs seize up like they crave to do. A man with a big synthesizer is playing some easy listening jazz number. I resist the urge to trample him (must conserve energy). Who the f*** plays elevator music to motivate tired runners?

Do yourself a favor and visit the site. Ceglowski is proof that blogging doesn’t have to be quick and daily to be satisfying.

TiVo the Tutor

My oldest son has taught himself to read. He takes every opportunity he can to sound out letters into words. Identifying road signs is a favorite pastime, though not without its hazards. Like yesterday when he sounded out the words “Road Closed” and let rip a bloodcurdling “No!” from his car seat that almost caused me to wreck.

But the best exercise he’s created for himself by far is to search for his favorite TV shows by spelling out their names in TiVo’s “Search by Title” feature. No one showed him how to do this; my wife and I rarely use Search by Title. TiVo is a perfect tutor, actually. He thinks of a show he likes — Hip Hop Harry, for example — then starts typing the letters he thinks make up the title. TiVo of course starts displaying what it thinks are matches which my son visually identifies. If he really screws up the spelling TiVo won’t show any matches and he’ll have to back up. And the reward for a correct spelling is that he gets to record his show. Positive reinforcement!

Gotta figure out how to get the microwave to teach him mathematics and we’ll be all set.


The Romantic concept of the poet inpsired from seclusion and singular inspiration seems to be alive and well, though once removed. My post on the isolation tank seems to have been the muse for a poem. Here’s an excerpt.

groupmind human searing feedback feedback human searing in human and groupmind and groupmind and and in to groupmind in feedback feedback and and

The weird thing (among many) is that it looks copy-and-pasted, but it isn’t. It’s non-random.

Here’s the full post/poem.

Update: some interesting stats from Chris.

504 is the dominant number in the sequence. If the (perhaps deliberately) malformed “groupmindand,” “nd” and “earing” were normalized then “prayer,” “and,” “feedback,” “groupmind,” “human,” in,” and “to” would all appear 504 times. 504 is also the HTTP status code for “Gateway Time-out.”

Just gets weirder, eh?

Story witch doctor

Recently I’ve been working with a really smart researcher in computational linguistics and, as is happening with increasing frequency with my colleagues, he happened upon my blog. The Icelandic connection with my last name (Tolva = “number witch doctor” = computer) was particularly interesting to him. He writes:

You’re right about the aversion to foreign words in Icelandic. I observed that there. The Icelandic “tala” for “number” appears to be related to the words “tal” in Danish and “Zahl” in German for “number”. “Tala” may also be related to “tell” and “tale” in English, because these English words go back to an Indoeuropean word “del” that means “count” or “recount”. There seems to be a semantic etymological connection between telling (a story) and counting. German “zählen” means “count”, and “erzählen” means “tell”. Danish “tælle” is “count”, and “tale” is “speak”. In English we can “recount” a story or give an “account” of some event(s). Maybe the semantic connection is that as you’re telling a story, you’re counting off the events?

So not only is my surname the made-up word for computer, but it has etymological connections to storytelling. Computers and narrative. Counting and recounting. It’s all so clear to me now. I suppose I am doing what I was destined to do.

(Of course, I’m not Icelandic at all.)

Corporate lingo watch

Ever since I first heard someone ask to “take this conversation offline” I’ve had a biochemical aversion to corporate parlance, especially when technical terms are used unironically to describe non-technical things. Today’s morsel went something like this:

“Thanks, Bob, I really like those ideas. One that I’d particularly like to double-click on was …”

Shouldn’t it be “click”? I mean, isn’t the implication here that this is an idea that should be followed, like one follows a link? Or is he double-clicking it to run it like an application? Start it up?

You know, if you’re going to lace discussions with technical metaphors that are already a minefield of business-specific terms you could at least strive not to sound like you’ve just discovered the mouse and GUI. Oooh, the pretty icons make my copy of WordPerfect come alive!

Two bitter posts in a row. Feels good.


The elevators in our building have a curious recorded floor announcement. The female voice has an unplaceable accent: nasal, snooty-almost-schoolmarmish, vaguely Canadian. Probably the result of a focus group on pleasing intonation gone wrong.

I rode the elevator down today with a construction worker. We didn’t speak. As we emerged from the car I heard him mutter to himself “damn foreign elevators.” And he clomped off.

I’m really not sure he was kidding.

Brown, I’d prefer you didn’t

I just remembered something that made me laugh a while back. I was talking to a friend of a friend who works for UPS. We were joking about the tagline “What can brown do for you?” when he told me about an internal effort to galvanize support for the new campaign. The execs were trying to think of something catchy when someone suggested “Operation Brownstorm.” This stuck. As a storm of brown might.

The teams, I was told, couldn’t believe that management would go with such a blatantly scatological reference. Not only that but employees were encouraged to “Get behind Operation Brownstorm!” Um, no thank you. I’ll stay right out here in front. A safe distance from the squall.

Lesson: when Googling for name inspiration be sure to deselect the mature content filter.

UPDATE: If you want to mount an awareness effort on the color brown, you might have a look at the Chicago Transit Authority’s Countdown to a New Brown. There’s built-in potty humor there too, of course, but no storming as far as I can tell.