"It seems to me that this criticism depends on a confusion between two kinds of mistake. We may call them 'errors of functioning' and 'errors of conclusion'. Errors of functioning are due to some mechanical or electrical fault which causes the machine to behave otherwise that it was designed to do... Errors of conclusion can only arise when some meaning is attached to the output signals from the machine."
(449 Turing)
This passage brings me back to the scene in 2001 Space Odyssey when HAL remarks that "human error" is the only explanation for the discrepancy between the two HAL 9000's different results. At first, I thought this was just a defensive remark made on HAL's part - a prideful reluctance to admit he had made any sort of mistake. However, his statement about human error could also be read as him acknowledging an error of functioning within himself as a result of some mechanical/electrical fault caused by the humans who created him. Under this circumstance, HAL is not malfunctioning at all but is rather still operating without errors according to his system - perfectly performing the tasks he is capable of doing according to his programming. Rather, I don't know if it's possible for him to recognize it as an error in himself at all - HAL may only recognize that in the event of a discrepancy between two HAL units, the fault lies with the human element in the equation - HAL's programmers and technicians.For some reason, I find it easier to lay the fault at the feet of the programmers than the machine - perhaps that means I'm the type who would blame the parents for the outcome of the child.
The other explanation for HAL's behavior is that an error of conclusion occurred. Honestly, I'm still a little shaky on the exact meaning of an "error of conclusion, and Turings explanation moves a little too quickly for me to really grasp it. I'd love some help interpreting it so that I can further attempt to understand the potential reasons behind HAL's errors (if in fact they are errors at all).
-James G.
I like your specific reference to a portion of the Turing, here, and in a way, "errors of functioning" sets us up nicely for next week's class on the materiality of media (all those physical things that can go wrong with machines and their networks when we think about them less as abstract than as embedded in the real world). As for "errors of conclusion" vs. "errors of functioning," Turing almost seems to be making a distinction between understandable mistakes and unintelligible ones. A reasonable error in solving a math problem would be misplacing the decimal in the solution or being off by one in carrying a remainder. An unreasonable error would be answering a math problem with a picture of a duck or the blue screen of death. In a way, HAL's behavior is the product of an understandable error.
ReplyDelete