Language, Dolphins and Garage Cinema

May 11th, 2003
dolphin as projector?

What if dolphins communicate by sending and receiving images? What if humans can learn to do the same, on the fly, via computer mediation?

I know what you’re thinking: this guy’s been in California far too long. You’re probably right.

But bear with me on this.

As a kid growing up by the sea in Florida I was obsessed with bottlenose dolphins. I read everything I could find about them. When I was 13 I borrowed a fancy underwater microphone from an oceanographer and used it to record dolphin sounds at Ocean World, the local marine theme park. I played back the recordings into another dolphin tank. But I didn’t get much of a reaction at all. After gathering around this strange noise-making machine for a few minutes, the dolphins quickly grew bored and ignored the tape recorder. They were far more interested in my cheap watch. And dead fish.

Plenty of more serious research (and writing and movie-making) was devoted to the prospect that dolphins’ clicks and whistles might be a language. Scientists showed that dolphins convey instructions to one another, but still nobody has proven whether a high-level dolphin language exists.

This week I read a paper by Berkeley’s Professor Marc Davis that dramatically changed my thinking about this by pointing out that a dolphin language might not be based on words. Most linguistic dolphin research I’ve seen seeks dolphin sounds strung together as words, and I always unconsciously assumed that any high-level language must be based on words.

yukaghir love letterBut Davis points out in his paper that some written human languages don’t use words at all but instead directly represent meaning visually. The image to the right is a message written in such a language, by a member of the Yukaghir tribe in Siberia. (See Davis’ paper for more details and a translation of the letter.)

Like bats, dolphins use echolocation. They emit waves of sound and use the resulting echoes to pinpoint locations, sizes, shapes, densities, and even internal states and structures of animals and objects, with astounding precision and accuracy. If dolphins use their own sounds so skillfully to probe their environments and to “see” what’s around them, can they also use sound to create artificial imagery that’s “visible” to other dolphins?

Dolphins exhibit a superhuman ability to convey spatial instructions to one another. Nobody’s sure how exactly they work this out, but if you watch a group of dolphins carrying out tasks in which they have to quickly synchronize very complicated sets of movements — during a theme-park performance, for example, or during hunts in which they round up thousands of fish into dense schools — you’ll be amazed at their powers of spatial coordination.

Can you imagine dolphins sending each other visual cues mapped to real-world environments — or even sending entire artificial “video” scenes showing planned activities — on the fly?

This may be a stretch; it’s probably fiction and so far it’s not backed by much science. But it’s a very powerful idea that we can use. Even if dolphins cannot communicate this way, perhaps we will be able to, with the help of computers.

Davis and his Garage Cinema Research group at Berkeley are working on it. They’re designing systems that they hope will allow regular people to easily and quickly build video compositions, without putting forth the tremendous amount of time, expense and technical knowledge necessary for today’s film production. Thanks to smart systems that can recognize media assets and automate much of the video capture, editing and production process, Davis hopes to allow us all to “write” video as often and as easily as we “read” video today. The promise lies not just in replacing the current wasteful and corporate-dominated system of creating polished high-end feature films, but in providing humanity with a new, more powerful form of everyday communication.

A Pivotal Month

April 19th, 2003

Mein Gott, has it really been more than five weeks since my last entry?

It’s been a crazy, hectic, life-altering month. The top headlines:

The University of California at Berkeley - Cal BearsThe grad school decision: My quest for a Human-Computer Interaction (HCI) Master’s degree continues. I was accepted by four great Information Management & Systems schools: Berkeley, The University of Michigan, The University of North Carolina at Chapel Hill, and The University of Washington. For reasons I won’t go into, this was an unbelievably difficult decision, but after talking to and exchanging e-mails with more than 25 professors, students, and people in the industry, and after spending countless hours reading related Web sites, papers, etc., I chose Berkeley. I’m glad the agonizing decision’s over and I can’t wait to get started in the Fall.

CHI 2003CHI 2003: I gathered a lot of this knowledge during the CHI 2003 conference, which was held in my hometown of Ft. Lauderdale, Florida (!). How strange it was to see all these brilliant people discussing HCI less than five miles from where I was born and raised, in a region where you normally never meet anyone remotely interested in product development or software development, much less in HCI. I was just a bit overwhelmed while trying to learn as much as I could about the schools I was considering, working as a student volunteer, showing people around town, taking in presentations, tutorials and demos, seeking out my heroes in the field, and just finding my way around my first academic conference. It was an exhilarating experience and a lot of fun.

Stanford UniversityThe new gig: Last month I started my new full-time job at Stanford University’s Department of Dermatology. Among other duties, I’ll help to manage the input, cataloguing, storage, and retrieval the thousands of digital images that the department creates every month, as well as the associated medical records. It’s great fun so far; I’m the only computer person in the department, but the doctors and staff are super-smart and they seem much more open to change and to new technology than the users I worked with during two other medical gigs. I hope to pool efforts with people elsewhere on campus working on cool medical informatics projects like the Stanford MediaServer but there’s not much time until the Fall semester begins and I leave Stanford to start classes at Berkeley. I wish I had started this job a year ago…

Life should be a tad saner now, but now much. Among other things, I have to begin a Java class and possibly a data structures and algorithms class in preparation for Berkeley. These are busy days. But I’m lovin’ it all.

Next Entries »