As Artificial Intelligence continues its quest to make computers more human, it seems strange that a 1984 book marketed as "the first book ever written by a computer" has mostly vanished from sight. Its esoteric title, "The Policeman's Beard is Half Constructed," probably equally attracts and repels, but it does adequately prepare readers for the quirky, often meandering and sometimes humorous text that awaits within. But given the book's vintage, one may wonder at the bold tagline "written by a computer." After all, 1984 stands as the computer history equivalent of the Punic Wars. Apple's Macintosh first appeared during that year's Superbowl, 3.5 inch floppies and CD-ROMs had just arrived on the market, IBM still ruled the PC market and Microsoft released MS DOS 3.0. Windows version 1.0 wouldn't debut until the following year. Programming languages available at that time included C, C++, Pascal, Smalltalk, LISP, BASIC, FoxPro, Prolog, Ada and others. LISP and Prolog even made accommodations for Artificial Intelligence. Yet the book's introduction says that the program created to write the book, called Racter, used "compiled BASIC on a Z80 Micro with 64K of RAM." When people think of Artificial Intelligence tools, even arguably in 1984, BASIC probably doesn't come immediately to mind, if it comes to mind at all. Designed initially as a language to make programming more accessible, the acronym BASIC stands for the overwhelming mouthful "Beginners' All-purpose Symbolic Instruction Code." Though sometimes derided as a "toy" language, BASIC can actually produce programs of substantial complexity. For many at the time, possibly including Racter's creator, William Chamberlain, it may have provided the easiest, or maybe the only, means of creating a computer program. Other languages, especially assembly, tended toward the esoteric and often required extensive training. As for the hardware, "64K of RAM" says it all. "Z80 micro" likely refers to the Zilog Z80 8-bit microprocessor introduced in 1976. By today's standards this all sounds quite technologically feeble, of course, but back then it represented a fair amount of computing power. Apparently enough to "write" a book.
The book's opening page introduces Racter as "The Author" and "the most highly developed artificial writer in the field of prose synthesis today." Computer file name limitations at the time didn't allow the use of Racter's full unabbreviated name, "raconteur." Then the introduction states "with the exception of this introduction, the writing in this book was all done by a computer." Further down, the text describes Racter in modern AI terms, such that "the programmer is removed to a very great extent from the specific form of the system's output" and "the computer forms output on its own." In other words, the programmer creates a generative machine, in this case for grammatical English, starts it up and then sits back and watches what happens. The resulting outputted text does not reside within the program, which itself handles verb conjugation, singular and plural forms, gender and combining multifarious elements together based on grammatical rules. Apparently, "word files" contained a pool of basic words that the program accessed for creation of its output. It grabbed the words, ran them through various programmatical structures and finally displayed the results that ended up in the final book. In this way Racter allegedly created prose "in no way contingent upon human experience." This aptly describes the book's main text, as it has a distant feel to it while nonetheless remaining strangely comprehensible, if rarely identifiable. Lines such as "this dissertation will show that the love of a man and a woman is not the love of steak and lettuce," "enthralling stories about animals are my dreams and I will sing them all if I am not exhausted and weary," "flounders and lobsters are munched by famished theoreticians who drink champagne and tepid seltzer" and "tomatoes from England and lettuce from Canada are eaten by cosmologists from Russia" demonstrate this effectively. Fans of Edward Lear, Monty Python and other nonsense-inspired authors will likely enjoy some of the mostly comprehensible passages. But other lines just seem random or exude a possibly contrived randomness. The book overall presents a mixed bag with high points, low points and numerous indifferent points. Some pieces take up a mere line or two, while others fill multiple pages. A few involve "conversations" with Racter, one between "Joan," presumably the illustrator, and another with "Bill." These evoke the less anarchic "ELIZA" program from the mid-1960s that simulated human conversation in a therapeutic manner that some users actually found convincing and comforting. Racter, based on the book's conversations, would probably not make a great therapist.
In 1993, a comp.ai newsgroup post by Jorn Barger, subsequently published in "The Journal of Computer Game Design," referred to the book as "largely prefab" and argued that the claims of the book don't necessarily hold up "under close scrutiny." It says that Racter's templates likely contained additional elaborations and "wacky" words that reflected Chamberlain's disposition and style, so in a sense Chamberlain may have unduly "influenced" Racter's prose. These claims arose from experimenting with a commercially available version of Racter that appeared in the same year as the book. Apparently this software, though interactive and generative, did not produce text as rich or as complex as that contained in the book. The post also stated that the "INRAC" compiler, used to create Racter, did not "include any sort of 'syntax directive' powerful enough to string words together into a form like the published stories." These criticisms don't dismiss the book outright, but they do insinuate that it may have overstated its claims and that perhaps Chamberlain, as an "invisible hand," unintentionally played a larger "co-author" role than the book's introduction suggests. Whatever the case, given the capabilities of the commercial version of Racter, the book's text probably derived from different techniques or from applying additional code to achieve its more florid results. Nonetheless, this arguably leaves a lingering asterisk over the book's tagline "the first book ever written by a computer." Not only that, the need for a human to write a computer program that produces computer generated text raises additional questions about the possibility of producing programmatic textual output devoid of human influence. How do programmers "remove themselves" from this process? The program itself will inevitably contain traces of human experience and interaction. The option of writing a metaprogram to create programs that produce generated texts quickly falls into an infinite regress. Also, how could the people who choose the words to input into such a program extricate their experiences, their personalities and their predilections from their selections? One could even question the point of producing text meant for the consumption of humans that attempts to eradicate signs of human experience. Such a pursuit might seem potentially and theoretically interesting for its own sake, but whether it would produce entertaining or intriguing prose worthy of mass publication remains questionable. If that summarizes the purpose of Racter's book, then it arguably provides an example leaning toward the negative side of this argument, because the text only rarely gives the reader something to hang onto from a literary perspective.
At one point in the recent past, obtaining a physical copy of Racter's book apparently presented an ominous Herculean challenge. Now the vast Internet marketplace has multiple copies readily available, though not all of them come cheaply. In any case, the entire book exists online in .pdf form for the uncommitted curious or for those who prefer intangibility. One downside of the .pdf copy concerns the rendering of the interesting illustrations scattered throughout the book: they look far better on paper than on the provided scanned .pdf. The book appears to have gone through only a single printing, which helps explain its pre-Internet scarcity, and it also suggests that it failed to obtain any mass appeal. It actually seems to have vaporized almost completely, despite some hints of cult status in experimental literature circles. As the first book to mass-market computer generated literature, it will likely retain some importance. Or perhaps its dangling questions and mysteries serve as a warning concerning the nebulousness of phrases such as "written by a computer." Concerning the present, does Racter in any way presage "Automated Journalism?" Over the past five years, this term has emerged as Artificial Intelligence programs begin to "write" and publish news. Based on large data sets, such programs seem to instantaneously churn out stories involving sports scores, financial returns, weather and just about anything reducible to numerical analysis. Looking back at Racter, one may hesitate to say that such programs "write" text without human influence. They draw on data from the human world, as does the program code created to produce such texts. They may replace physical human roles, but they have not eliminated "the human" from their workings. AI likely represents yet another extension of human sense capabilities, as McLuhan characterized technology such as telescopes, microscopes and television. As such, it remains fundamentally human regardless of the actual physical presence or absence of humans surrounding it. If nothing else, Racter remains a possible tool or litmus test to reference in such debates. Essentially, it raises sticky questions around claims that a computer has "done" something without human intervention. The possibility of a computer "doing" something "inhuman" feels almost metaphysical, at least at this point in time. Human intervention arguably seems present throughout Racter's book. Given that, claims that Racter, acting as some sort of independent agent, "wrote" the book, in a way that a human would "write" one, doesn't feel accurate. Once again, that prominent proclamation on the book's cover, "written by a computer," seems controversial at best. Still, "The Policeman's Beard is Half Constructed" will probably linger in the AI novelty section for a long time. Though the book itself doesn't contain answers to the myriad questions it raises, an analysis of the book and its claims can lead readers down some very interesting paths. For that, and possibly for that for alone, it might retain its curious ineffable appeal.
Loading your book clubs
There was a problem loading your book clubs. Please try again.
Not in a club? Learn more
Join or create book clubs
Choose books together
Track your books
Bring your club to Amazon Book Clubs, start a new book club and invite your friends to join, or find a club that’s right for you for free.
Download the free Kindle app and start reading Kindle books instantly on your smartphone, tablet, or computer - no Kindle device required. Learn more
Read instantly on your browser with Kindle for Web.
Using your mobile phone camera - scan the code below and download the Kindle app.
Flip to back Flip to front
Follow the Author
Something went wrong. Please try your request again later.
Customers who viewed this item also viewed
Page 1 of 1 Start overPage 1 of 1
About the author
Follow authors to get new release updates, plus improved recommendations.
