The Complete Roderick

RODERICK AT RANDOM

or

THE FURTHER EDUCATION OF A YOUNG MACHINE

I

Dead or dreaming? It seemed to Leo Bunsky that he had come out of retirement. Somehow he was back in his old office, working on Project Roderick again. And somehow the old heart condition had decided to stop tormenting him: gone was the breathlessness, the tiredness, the draining of fluid down into his feet until they doubled in size and burst his shoes. Without any medication or surgery, he was now cured. Everything was back to normal now, if that word could be used in these miraculous circumstances. Calloo, and also Calais! But what was the explanation?

He was dreaming. He was dead. Dreaming but dead. Neither. He had slipped through a ‘time-warp’ into a ‘parallel universe’ (Dr Bunsky was a reader of science fiction), probably through a ‘white hole’.

It didn’t matter; in any case there was plenty of work to do. He could live an unexamined life, until Project Roderick demanded less of his time, okay? Okay, and great to be part of this real-life science-fiction dream, a project to build a ‘viable’ robot. Roderick would be a learning machine. It would learn to think and behave as a human. All the team had to do was solve dozens of enormous problems in artificial intelligence that had defeated everyone else; from there on, it was science fiction.

Bunsky’s job at the moment was teaching simple computer programs to talk. So far he’d got a program to say Mama am a maam, but not with feeling. If Roderick the Robot was ever going to think as a human, it would of course need to learn and use language as a human: Mama am a maam was not exactly Miltonic, but it was a start.

How did people learn to talk? No one really knew. There were those who thought it might be a matter of training, like learning to ride a bike. Others seemed to imagine a kind of grammar-machine built into the human head. Still others tried teaching chimpanzees to talk while riding bikes. Chimps, so far, had articulated no theories of their own.

Bunsky found it easier to scrap general theories and consider the brain as a black box: language stuff went in and different language stuff came out. In between, some sort of processing took place. What Roderick the Robot would have to do, then, was to mimic the hidden processing. The robot would have to learn as human children learn, and that meant making the same kinds of childish mistakes. And only those kinds of mistakes. It was okay for Roderick to say Me finded two mouses on stair. It was not okay to say I found two invisible green guesses on the stair.

Leo Bunsky lifted his gaze to a file card tacked to the wall above his desk:

TO ERR (APPROPRIATELY) IS HUMAN

There was something he couldn’t remember, that made his head ache.

The door opened and one of the younger men in the project came slouching in. It was that interdisciplinary disciple with the unfortunate name, Ben Franklin. Bunsky didn’t know him well.

‘Leo, how’s tricks?’ He slumped into a chair and started flicking cigarette ash on the floor.

‘Fine, uh, Ben. Fine. Wish you’d use the ashtray, I know the place is untidy but—’

‘Yes, I found two invisible green guesses on the stair. Yours?’

‘Very amusing. Now if you’ll excuse me…’

Franklin stood up. ‘Busy, sure. Sure. I don’t suppose you need any help with anything?’

‘Sorry, no.’ No one ever wanted Franklin’s help. No one really trusted him, with his strange background: a hybrid degree in Computer Science and Humanities. A little too eclectic for serious research work. Dr Fong had hired him as project librarian and historian, but so far there weren’t many books and no history. Ben Franklin just sort of hung around dropping ash on the floor. ‘Sorry, Ben.’

‘Sure.’ After a pause, he sat down again. ‘Leo, you ever have any doubts about this project? About Roderick?’

‘Doubts?’

‘Kind of an ethical grey area, isn’t it?’

Bunsky felt the headache settling in, deepening its hold on him. ‘What grey area, for Christ’s sake? Building a robot, is that grey? Is that ethically suspect, to build a sophisticated machine? Is cybernetics morally in bad taste?’

‘Well, no, if you put it like—’

Bunsky was shouting now. ‘We’re not violating anybody’s rights. We’re not polluting any imaginable environment. We’re not cutting up animals and we’re not even screwing around with genetic materials!’

Franklin flicked ash. ‘Come on, Leo, what about long-term consequences? Don’t tell me it never crossed your mind that Roderick might be dangerous. First of a new species, of a very high order, has to be some danger in that.’

‘A mechanical species, Ben.’

‘But on a par with our own. And what if robots evolve faster and further? Where does that leave us? Extinct!’

Bunsky made his voice calm. ‘Let’s not be too simplistic there. Humans wouldn’t be in direct competition with robots, would they? Both species would use, let’s see, metal and energy. But robots wouldn’t need much of either resource. Should be enough to go around, eh?’

‘Eh yourself, what about intangible resources? What about things like meaning?’ Franklin put out his cigarette on the floor. ‘I mean look, it could be that humans feed on meaning. It could be that we only survive by making sense out of the world around us. It could be that this is all that keeps us going. So if we turn over that function to some other species, we’re finished.’

The headache began to throb, roaring waves of pain breaking over him, trying to drag him under. There were moments of dizziness and deafness, moments when Bunsky could hardly make out the empty smirking face before him. Franklin looked a little like a ventriloquist’s dummy sometimes.

‘You have a point, Ben. Too bad you’re such a goddamned jerk.’

‘What?’ Franklin paused, a fresh cigarette halfway to his smirk.

‘I said you have a point. Yes. I believe we do have a need to make sense out of the world. We see by making pictures, no? By sorting out the “blooming, buzzing confusion”. And of course we hear speech by making sentences out of it. I believe that the essence of human intelligence is that kind of hypothesis-building. Life is making intelligent guesses, you agree?’

‘Of course, Leo. Hypothesis-building.’

‘Then don’t you see? The robot is the ultimate human hypothesis. What better way of making sense out of the world, than through complete copies of ourselves? How better to model our inner and outer world? Ben, we need the robot — we need the idea of the robot.’

Franklin lit his cigarette and tapped it. ‘The idea, maybe, but I’m not worried about the idea. It’s the embodiment.’

‘Yes, we’re building a real robot. We’re doing what everyone has always wanted to do, down through the ages. If the Jews could have built a real golem, do you think they would have hesitated? Or the Cretans, wouldn’t they want a real Talos? If myths are just wishes, isn’t it obvious what they were all wishing for? Dolls and statues from the terra-cotta past… puppets and manequins today… The idea of the wish for the robot is so, so very powerful… What child doesn’t like Punch and Judy? What grownup won’t pause to watch a ventriloquist at work? My God, when I was a kid, do you know they even had a ventriloquist on the radio — on the radio!’

Franklin flicked ash. ‘I don’t get it.’

‘Edgar Bergen and Charlie McCarthy on the damned radio. You see, the novelty of a talking doll was so powerful, we didn’t even need to see the doll. There didn’t even have to be a doll! All we needed was the