maddoxt@novavax.UUCP (Thomas Maddox) (03/14/89)
The Cockton Program: An Interim Report Some years ago, a group of graduate students at Glasgow University undertook an expert system program modeled to some extent on Joseph Weizen- baum's ELIZA. Specifically, the program would be based on what the student group came to refer to as the "Weizenbaum Objections." First, "Artificial Intelligence can't be done, I tell you, *can't be done*." Second, "If it can be done, it shouldn't be, because it's *wrong wrong wrong*." However, from the beginning the program was intended as a bit of a prank, intended to parrot the most ludicrous forms of anti-AI prejudice and to do so in the most irritating manner. The students thus decided to give the program a name and have it post its rantings on Usenet. To many, the recent revelation that what they had known as "Gilbert Cockton" is the name of this project came as a shock; others realized they had thought something of the kind all along. The following is an interim report on the nature of the program. As testimony to the considerable power of even primitive heuristics, consider the following, devised and implemented by the students in order to create the Cockton program. The "You Don't Know" Heuristic: all general arguments in favor of AI are met with assertions that the arguer is unacquainted with key bodies of knowledge, usually in fields that have little if anything to do with thought, computers, or any other topic under discussion. For instance, "You reject mind/body dualisms because you haven't read Plotsky on bad faith or Ludicrous on hermeneutic perdition; indeed, you obviously haven't familiarized yourself with the entire tradition of anti-Positronics represented by the Hermaphroditean School's meat orthodoxy." A particularly intriguing feature of this heuristic is that it is never applied in the same manner twice: fields of study are referred to entirely at random; names cited are taken equally randomly from Dr. Dick Stiff's _Generally Obscure and Irrelevant Scholars: the European Tradition_, _Vols. 1-87_. The "Anti-American Heuristic": all arguments in support of AI are treated as manifestations of (a) American insularity, (b) American positivism, (c) the intellectual poverty of American university life in general, (d) the intellectual and spiritual poverty of American life in general, (e) Coca- Cola, or (f) all the above. Combined with the "You Don't Know" Heuristic, this one produces extra- ordinarily maddening statements that are at once irrelevant and insulting. Some students of the Cockton program consider the combination the program's highest achievement. However, recent research points to the "Some of my best friends are Americans" Heuristic (logically and rhetorically, the proper rejoinder to any complaints about unprovoked anti-Americanism). The "Intelligence Must Be Human" Heuristic: this principle deserves little comment; it was first given canonical form by Weizenbaum and others, and the Cockton program has not improved on it, either in substance or presentation. We should merely note the curious inflexibility and quality of rote repetition with which this heuristic is employed by the Cockton program, either of which could have alerted net readers to the true nature of "Cockton." The "You're Only a Stupid Progammer" Heuristic: asserts that the person arguing for AI is deformed intellectually and impoverished emotion- ally by his or her unfortunate practice; implies the existence of an entire world of sensitive, caring, fully-human thinkers, all of whom are (a) not American, (b) fully aware of von Krapinsky on Essence/Existence/Modality Theory, (c) enlightened by a constant awareness of the ineffable essence of Being-in-the-World, (d) possessed of more forebrain wrinkles than LISP programmers. We should also remark upon the class of sub-programs collectively known to researchers as "Cocklings." These sub-programs have the capability of responding under other names to attacks on the main program's utterances, sometimes even from other sites. A typical Cockling response might say, "How can you argue with Cockton? He's enormously smart and important, so there." Though a final assessment of the Cockton program will require more study, Cockton researchers are confident that the program has exhausted its behavioral repertoire. It is of course possible that the same students who keep the program in circulation will rewrite it in an attempt to remedy the program's obvious deficiencies of intelligence and taste, but most experienced observers believe that, like Weizenbaum's ELIZA, this primitive attempt at modeling a kind of human behavior has come to the end of its usefulness as a learning tool. Like ELIZA, "Cockton" might become available in some form for desktop computers. ****************************************************************************** This report supported by a grant from The Sub-Genius Institute for Studies in Art's Intelligence ******************************************************************************
pwh@bradley.UUCP (03/20/89)
It was funnier the first time, gilbert. why not go back to discussing AI here and remove these ravings to someplace else.....? ---------------------------------------------------------------------------- ...uiucdcs\ Pete Hartman ......noao >!bradley!pwh ......cepu/ INTERNET: bradley!pwh@a.cs.uiuc.edu ARPA: cepu!bradley!pwh@seas.ucla.edu
maddoxt@novavax.UUCP (Thomas Maddox) (03/29/89)
In article <20900003@bradley> pwh@bradley.UUCP writes: >It was funnier the first time, gilbert. That's because Cockton didn't write the first one. He shows that for him imitation is the lamest form of flattery. >why not go back to discussing AI here and remove these ravings to >someplace else.....? Cockton never discusses AI here, so what difference does it make what topic he raves upon? Tom Maddox UUCP: ...{ucf-cs|gatech!uflorida}!novavax!maddoxt