When it first began in the 1950s the task of programing a computer to translate text from one natural language into another was simply called machine translation, a term still in use today. It took the prospect of losing funding in the early 1960s to prompt the field to generalize its activity and rebrand itself as computational linguistics. The International Journal of Man-Machine Studies was founded in 1969. The title suggests that it might publish articles on the ergonomic design of kitchen appliances; lathes, power drills, circular saws and other woodworking tools; or perhaps more plausibly because of the greater complexity, the design of airplane cockpits. But no, such articles never appeared – at least not that I remember. It was devoted to articles about the interaction between humans and computers and, correspondingly, changed its name to The International Journal of Human-Computer Studies in 1994.
In what sense are computers kinds of machines? Computers are not machines in the ordinary sense of the word. The dictionary on my computer defines machines as “an apparatus using or applying mechanical power and having several parts, each with a definite function and together performing a particular task.” That’s not what computers do. Though many computers have such machines as component parts – I’m thinking mostly of the motors driving hard drives – they do not characterize the primary function of computers and one can get very powerful computers that use flash memory and so contain no motors at all.
I can hazard, without much investigation, an at least superficially plausible historical explanation. Before the electronic digital computer came into being out of the chaos of the Second World War, we used a wide variety of electro-mechanical devices, that is machines, for calculation and tabulation. IBM, after all, is an acronym for International Business Machines, which was founded in 1911 as Computing-Tabulating-Recording Company – the name change came in 1924. According the Wikipedia “they manufactured machinery for sale and lease, ranging from commercial scales and industrial time recorders, meat and cheese slicers, to tabulators and punched cards.” That is to say, business machines of various kinds. In my youth, back in the days when IBM dominated the mainframe business, the Selectric Typewriter was the premier electric typewriter, but the company is no longer in that business.
My argument, of course, is not specifically about IBM, as important as that company has been and remains in the computing industry. It is rather but an example from a commercial environment awash in mechanical and electromechanical machines serving various business purposes. Moreover computer has had a long history of use as job category: people who perform calculations. In this context, then, it would have been natural to think of electronic digital computers as machines. On the one hand they descend from an electro-mechanical heritage and, on the other hand, computers were people.
The usage has a metaphorical aspect as well. Machines, first mechanical and then later electro-mechanical, are artificial things, some of them perhaps even miracles of rare device. They are designed and built to human specifications to suit human purposes. We understand how they work, part by part, assembly by assembly. Their functions and operations are thus fully explicit. And perhaps chief among these machines, historically and philosophically, is the clock, which gave rise to the notion of a clockwork universe. If the universe is a giant clockworks, then how can we be any different? And so we arrive at La Mettrie’s 1747 Man a Machine (L’homme Machine).
It is the property of fully explicit design that transfers to electronic computers. And it does so on two levels: 1) that of the device itself, and 2) that of the computational procedure to be run on the device. Alan Turing gave us an abstract account of the first level while John von Neumann figured out how to implement that abstract device in hardware. The result is a physical machine, a computer, that can enact any computational procedure (the second) that is explicitly described in appropriate detail. It is this two-level specification and design that distinguishes electronic digital computers from all prior machines.
That is the essential difference between them. It is a difference in kind, of ontology as the term has come to be used in the cognitive sciences.
No comments:
Post a Comment