The Hofstadter Prize for Machine-Written Narrative was formally established by the Society for Analytical Engines at its annual meeting in February 1994, at which time the rules of the competition were ratified. Submissions were accepted through April 1996, after which began the long process of validating and judging. Winners were announced in late 1997 and were originally scheduled to be published in a single bound volume for distribution to members of the Society at its 1998 annual meeting. For reasons complex and obscure there then followed a period of more than four years during which the project was in a limbo of sorts. With the publication of Cheap Complex Devices this unfortunate situation is at least partially remedied, and the world can see at last that the era of human storytelling supremacy has ended.
Although the Hofstadter Competition had its birth in February 1994, its conception, at which I was present, occurred ten months earlier, in April of 1993. It was during that month that, while a member of the Human Factors Engineering group of Sun Microsystems, I attended the annual convention of the Special Interest Group for Computer-Human Interaction, which was held that year in Amsterdam. One night, dining alone in a charming subterranean restaurant (whose name I have since forgotten), I chanced to meet some other SIGCHI conventioneers, who invited me to join them at their table. There were about eight people in their group; I have an imprecise recollection. Nor can I remember how many of that number were men and how many were women, or where any of them called home.
Among them were hardware engineers and software engineers, linguists and cognitive scientists. I soon learned that the members of that dinner party were from two groupsߞcompetitors in the realm of Artificial Intelligence. Their particular speciality was Human-Language Storytellersߞcommonly called “Hals”ߞand they had come together at this restaurant to agree upon the rules for a Hals storytelling contest to be sponsored by the Society.
Wine flowed, voices overlapped each other, there was much good cheer and telling of jokes. But as the evening progressed the jokes became crisper, more biting, bordering on cruel. I began to feel awkward and uneasy, as if I were a clueless guest at a wedding where family tensions were palpable if inchoate.
After the initial exchange of pleasantries nobody said much to me or seemed to care that I was there. Which was a relief.
The dinner plates were cleared; dessert came and went. There was no interruption in the flow of wine. I did my best to keep up but I was out of my league: I couldn’t match their drinking or their repartee. I needed air. I attempted to leave money but was rebuffed with one collective voice. I woozily stood, made my farewellsߞand spent the next several hours walking along the canals.
Years went by. After losing my job in a downsizing at Sun I became disenchanted with the Silicon Valley rat-race, and in the fall of 1997 I moved to a small island off the coast of Maine. My last official act as a Sun employee was to convert ten years’ worth of stock options. I had never married, and have no children: I was awash in money. What I needed was something to do with my life.
It was at that time that I was approached, via e-mail, by the Hals Contest Subcommittee of the Society for Analytical Engines. It seemed that I had made some impression on the Amsterdam dinner party after all. I must have mentioned that earlier in my career I had been a technical writer. In fact, I had been recognized as a master of the craft by the Society for Technical Communication, from whom I received, in 1988, the coveted Award of Distinguished Technical Communication. Now the committee needed a disinterested technical communicator to edit and publish the results of their inaugural artificial storyteller’s competition.
As originally conceived, the Hofstadter commemorative was to contain two computer-written works of fiction: A novella called Bees, and a novel called The Bonehead Computer Museum. Along with them was an introduction written by the contest committee that explained the rules by which the winning entries had been judged. The Society had decided to publish privately and needed someone to manage the production. Did I want the job?
Remembering the odd tension of that subterranean dinner, my first inclination was to say no. But I was intrigued and my vanity bested my timidity. Two days after getting the offer I sent my acceptance.
Over the next several weeks and months the bytes arrived. I did nothing but collect them. Sometimes a chapter came entire, sometimes only a paragraph, or a sentenceߞa word! And then it was done. The writing being complete, I undertook to edit. It was child’s play. I corrected some obvious grammatical errorsߞfewer than a dozen, all toldߞeliminated a few instances of redundancy when they appeared to be the result of transmission glitches, and smoothed some ruffled Postscript. I chose the typeface, page size and layout; I made arrangements with a printer and arranged for a small private printing. All told, hardly more than a week’s work, for which I was paid quite well. This work was completed in early 1997. That could have been the end of my involvement with this project. It should have been.
But having read the work in question, I wasn’t happy with the plans for its publication. I had just edited an extraordinary and historic document. The Technical Report of the Artificial Fiction Subcommittee of the Society for Analytical Engines, 1993 contained two software-written novels which, while certainly imperfect, were the most compelling evidence ever of a truly human sensibility in a computer program. Moreover the Report also contained a scholarly introduction to these artificial fictions that explained in very accessible terms just how these programs achieved their magic. It just didn’t seem right to me that such a work should be privately published. So I decided, without consulting the Committee, to seek an established publisher.
A long coincidental chain led me to the New York City offices of literary agent Joe Regal, who, despite reservations about the authenticity of the work I wanted him to represent, took me on as a client and set about finding a publisher for the manuscript. Joe’s first assignment to me was to come up with a catchier title for the book. I proposed Cheap Complex Devices.
In deciding to take me on as a client, Joe was betting that the Technical Report I brought him could be shaped into a book that would make money for his agency. Paradoxically, money was not a consideration for me; my only concern was to find a publisher with sufficient stature to secure for this book the audience it warranted.
Alas, the publishing world failed to appreciate the significance of the manuscript that Joe brought them and not a single offer was made. So I was back to square one, and I prepared to publish the Technical Report myself as I had been hired to do in the first place. It was then that I made a most horrific and embarrassing discovery: I no longer had the source to one of the two novels, namely, The Bonehead Computer Museum. With this discovery I began my surreal stroll down a nightmarish path to which I still see no end.
It began like this: one day while I was in the offices of Joe Regal, literary agent, discussing strategies for publishing Cheap Complex Devices, my illegally parked car was towed from 29th Street to the New York City impound lot by the East River. That was embarrassing, but the remedy was easy enough: I paid the ransom and retrieved the car. Oddly enough, the impound lot clerk looked enough like me that he could have been my twin brother, and we joshed about that as I paid the fine. Only later did I discover that my laptop and a paper manuscript of The Bonehead Computer Museum had been stolen from the trunk of my brand new BMW coupe.
It is horrible to admit, but it is the truth and cannot be escaped: I had no backup. I had lost the only copies I had (paper and electronic) to a full-length novel written entirely by a computer program. By good fortune I had made copies of the remaining parts of the book. The Committee that had hired me of course would have a copy of The Bonehead Computer Museum, but for weeks I was too embarrassed to request one. Alas, my pride was to cost me dearly.
I finally mustered the courage to write to the Committee to tell them what had happened. But I never heard back from them. In fact, my e-mail bounced, and my subsequent efforts to find them by web surfing proved fruitless. I was baffled and disturbed by their vanishing act, but mostly I was upset about the lost manuscript. I felt an obligation to the book itself, and I dreamed obsessively about how to restore its integrity so that I could publish it.
And then things got really weird.
Twenty months or so after the manuscript to The Bonehead Computer Museumhad been stolen from my car in the New York City Police impound lot, the book itself, slightly revised, appeared for sale under a different title.
I first learned of this after reading a review on the geeky website called Lashout. The purported author of this book was passing himself off to the credulous masses as some kind of Silicon Valley archetype, but I recognized him as the now-retired New York City Police Department detective who bore a strong physical resemblance to me.
The editorial changes made to The Bonehead Computer Museum by this con artist (who had been demoted, for some infraction unknown, to the position of impound lot clerk) in every instance detracted from its overall quality. The chief “improvement” that he made was to make all the female characters gorgeous and irresistibly attracted to the protagonist. He also introduced a lot of hackneyed cliches and typos into what had been a clean manuscript.
Nevertheless the underlying novel was so good that not even his amateurish ministrations could ruin it. The reviews were raves and the sales were strong. He attained the stature of cult hero among the savvy set. So this human burglar had successfully ripped off and debased a superior work of computer-written fiction and ridden his crime to minor fame. But that’s not all he had done.
The Bonehead Computer Museum contained such a wealth of information about the workings of computers, the computer industry, biochemistry and so forth that it was hard to imagine a sidelined cop on a scutwork detail having written it. To account for the discrepancy, this liar invented a fictional persona that bore a striking resemblance to me. He gave this fictional writer a name obscenely close to my own. This car-lot clerk had been a detective, after all. He knew what he was doing, and ripped off my soul just as convincingly as he had ripped off the purloined novel.
Legal considerations unfortunately prevent me from naming the perpetrator of this literary fraud or explicitly stating the title he gave to his stolen and mangled thriller.
In the intervening years I have tried every way I can imagine to reclaim The Bonehead Computer Museum, to prove that its true author is a software construct. But the time has come to admit defeat and publish that to which I hold clear title. The gravely wounded book that you hold in your hands, Cheap Complex Devices, here published without its integral and deeply missed The Bonehead Computer Museum, includes the manuscript precisely in the state it was in on the day my car was towed away from 29th Street, Manhattan, while I sat upstairs discussing with my friend and agent Joe Regal just how to obtain respectability for this epochal literary construction.
I cannot take credit for writing the words that follow; they speak for themselves. But it was I who chose the title for the collection.
And I chose the epigrams.
John Compton Sundman
Stanhope Island, Maine
July, 2002