Thursday, April 19, 2012

Robot Weekend: Reflections


Read Jeffrey on April Madness first.

UPDATE: And let me introduce you to a blog all about medieval robots.

Last weekend, I had my own April madness (in my case, more like April silliness) of flying from Paris to State College, PA, to give a version of my worms paper at the Penn State Institute of the Arts and Humanities "Robot Weekend: Being Human Gizmos." How could I say no to Michael Bérubé? Could you? (and, incidentally, if you haven't read his recent MLA column, do so as soon as you can, and PARTICIPATE in the crowd-sourced project on non-tenure-track faculty).

The Institute's to be praised for putting--as you'd get from its title--the humanities in touch with the arts, something we tend not to do, preferring not to (or not take the trouble to) contaminate our commentary with the messiness of creation. Not our readers, of course! We're a messy bunch, I hope.

The arts, in this case, were represented by our papers, no doubt, and especially by Anthony Clarvoe's new play "Gizmo," a hilariously rhizomatic take on Čapek's R.U.R., and by Laurencio Carlos Ruiz's puppets (including one of Bob Flanagan!). Penn State treated me right, I met some great people, and I learned about some horrifying things and some less horrifying things. And, seriously, BABELers: if we're not reading Susan Squier, we need to get on it, and soon.

I'm sharing with you an expanded version of some of my notes. These aren't comprehensive, not even close, and they don't presume to represent the papers and presentations I heard. Instead, what you're getting are my ideas. I'm representing my reactions (like this automaton). If you're an old hand at robot studies, much of what follows will probably strike you as old hat. If you're new to it--like I am--then maybe you'll find something worth playing with.
  • I proposed that we should not look for the real breakthrough with artificial intelligence in machines that think as well or better than we do, nor in machines that love or suffer tragedy like we presume ourselves to do (and here, I direct you to the conclusion of R.U.R.) I don't think we need to keep swinging between, or trying to undo, the old divisions of thought and feeling (notoriously, the conclusion of Metropolis). Instead, I think the breakthrough will come when we build machines that play, and not in the sense of playing chess or in following any set of rules. I mean play in the sense of taking pleasure in things and each other, and in suspending the rules temporarily in order to bring a new world into being for its own sake. R.U.R.'s brilliant, but imagine if Čapek had ended it not with love and sacrifice, but with the robots jumping rope? Or singing karaoke?
  • All of this, of course, eludes the fundamental question of the distinction between artificial and "natural" or "innate" intelligence.
  • Note that the same problem of distinction applies to autonomy (eg, humans), semi-autonomy (eg, robots), or, well, whatever falls outside or below this threshold, like automata. Derrida's analyses of sovereignty and of reaction vs. response require more attention to these distinctions. More straightforwardly, studies of the cognition of elite athletes demonstrate that the athletes work best when they work automatically: pitchers who think about their motion, who becomes less 'robotic,' lose their effectiveness. Consider, finally, the entry on automata from Diderot's encyclopedia: "engin qui se meut de lui-même, ou machine qui porte en elle le principe de son mouvement." [Instrument which moves by itself, or machine which contains within itself the source of its motion]. If we follow deconstruction mechanically, we know what to do with this. On the one hand: the automaton moves itself; on the other, it is the object of an internal intention, so to speak, the object to a subject within itself, a motion-imparting subject that we will never be able to get at.
  • Notably, actors in Gizmo had wanted to know their "motivation." What's a motive but that which moves us but which isn't accounted for in the action itself? A motive's a kind of extra power. The actors want to know, in other words, what makes them more than robot. Alternately, a "motive" is precisely the search for the mechanism: actors who want to know their motives are robots who want to know what makes them tick. They're not looking for free action so much as looking for a different way to wind themselves up.
  • We had several papers on robot caregivers for disabled people. I won't presume to summarize (or mangle) their arguments. Instead, I'll just (again) offer my ideas, in this case, on the chronological difference of robots. Robots operate at a different speed than (most of) us do. Either they operate far more quickly (in terms of information retrieval and processing) or they operate far more slowly, in the sense that they might long outlast us, or they operate in a kind of stasis or quick temporal looping, in the sense that an action or set of actions will be repeated identically, whether done 10, 100, or 1000 times. See Disney's Sorcerer's Apprentice.
  • Proposed: what distinguishes robots from artificial intelligence is that robots are bodied. Counterproposal: HAL or Alpha 60 [in Alphaville] can be hurt, and seem to finally be localizable, and Skynet (in the Terminator films) does its ill through materializations (the terminators, nuclear warheads, etc.). Development: it's impossible to imagine a disembodied mind: it's all (necessarily) material, in the sense that everything has a place.


Jeffrey Cohen said...

Nice account, Karl: sounds like a richly provocative weekend. And it is great to see a medievalist among the robot theorists!

Another tack that theorizing the automomatic might take is to consider how much that isn't human actually follows your tripartite scheme at the end. Is history a robot? Temporality? The elemental?

Apropos of I'm not sure what, I also keep thinking of Benjamin and his fascination with the chess playing Turkish automaton. Though I guess that's a puppet rather than a true robot. Or maybe there isn't a huge difference between puppets and robots?

Jonathan Jarrett said...

Loads to think with here, Karl, and it's late enough and I've drunk enough that I shouldn't try, but, I can't resist throwing Asimov in there at your final paragraph: a robot that doesn't have to materialise to act, well, isn't that indistinguishable from a god? Sufficiently advanced technology...

(Now Blogger says, "Please prove you're not a robot" and of course, hasn't fixed its OpenID provision yet so I can't except by not being `authenticated', which is probably significant unbeknownst to them but, if I were a robot that knew enough to care about such a distinction, I'm pretty sure Turing's ghost would pass me...

E. R. Truitt said...

Karl, this sounds wonderful. Where can I learn more about it?

Karl Steel said...

First thing to say: for some (stupid) reason I didn't review the first issue of postmedieval, which of course has a lot of robotics and cyborgs in it, not least of all from ER Truitt herself.

So, please, reread her essay, and also have a look at Julie Singer, certainly relevant for the conference, and also Jeffrey's colleague JG Harris, on automata. Ruth Evans also has things to say about cyborgs.

So: ERT, the schedule's here. Most relevant for your work, I think, are the papers by Jenny Rhee, Chloe Silverman, and perhaps Joan Landes.

JJ: nice point. I immediately tried to think of a faith that hadn't materialized its god in some form. Maybe Islam? Even neoplatonism thinks in terms of emenations. (oddly, someone at the conference cited Plotinus's Enneads as a kind of spiritual/materialist monadic cosmology, which strikes me as a very odd way to engage with Neoplatonism....).

Someone at the conference, also, observed that one problem with the Turing Test is that it doesn't account for stupidity or deception. Thoughts on that?

Jeffrey: it was good to be the medievalist! The early modernist (Kellie Robertson's homegirl Debra Hawhee) acquitted herself very well (too), with a paper on animal energy in theories of metaphor in ancient rhetoric texts: as I recall, she observed that language, that supposedly rational nonanimal place, needs animal/material support to do its work. There's a lot more to her paper, of course.

The distinction between puppets and robots might be that between having a human present and not having one present?

In re: your SECOND paragraph: continue to mull it over.

E. R. Truitt said...

Karl, thanks for the link to the program, and also for the above-the-fold shout out to Medieval Robots. I was part of a conference on automata with JL back in 2009.

Re: the difference between puppets and robots. I've been thinking a lot about this. Automata is an early modern word, and medieval "automata" are most often called things like statues or images. But they could often work by hidden human agency, often in conjunction with some other technology. Maybe the whole distinction between "puppet" and "robot" needs to be historicized.

Jeb said...

Copy of RuR and the insect play was the first play I bought along with Sartre's the Flies. Still somewhat attached to both.

"What's a motive but that which moves us but which isn't accounted for in the action itself?"

I think their is a difference of approach here. Motive is a thing method actors seem to have a bit of a thing about; classical or declamatory style it's not something you would raise.

The approach here would be simple to speak the text let the rhythm take care of the rest (the approach developed at the Old Vic specifically for Shakespeare and restoration comedy)

The act of doing is key in classical style, which suggests may be different ways of thinking here.

People who come to acting from English lit. often have a difficulty here at first as they are trained to sit and discusses not work on their feet. Acting is observation and doing rather than discussing.