Constructing the Other Half of The Policeman’s Beard


by Leah Henrickson


In 1984, a curious book was released: The Policeman’s Beard is Half Constructed. With its bright red cover and substantial size (22.6 x 20.3 x 1.5 centimetres), it stood out on any shelf. What was more striking, however, was the front cover’s claim that the book was “a bizarre and fantastic journey into the mind of a machine.” This machine was driven by a system called Racter, developed by William Chamberlain and Thomas Etter, and The Policeman’s Beard was supposedly “the first book ever written by a computer.” To be sure, The Policeman’s Beard was not the first text ever written by a computer, but it was one the first algorithmically authored books – if not the first – to be printed and marketed for a mass readership. Indeed, the publication of The Policeman’s Beard and the subsequent release of the Racter software represent important moments in the history of algorithmic authorship wherein computer-generated texts became readily available for public consumption.

The Policeman’s Beard remains the topic of much discussion, especially as computer-generated texts become ever more prominent. Yet there is enduring speculation – and, even misinformation – about this book and its credited algorithmic author. This paper serves as a resource for those seeking a detailed narrative about the development of both Racter and The Policeman’s Beard. It examines the circumstances of The Policeman’s Beard’s production, and questions the extent of human intervention in the text’s generation and presentation. The Policeman’s Beard is an important juncture in the history of electronic literature, enjoying continued public fascination given both the inexplicability of its functionality and its cover’s claims of novelty. As computer-generated prose and poetry become increasingly common, references to The Policeman’s Beard are too more common, historically situating new works. This paper aims to reconstruct the contexts of The Policeman’s Beard is Half Constructed through reference to available scholarship, contemporary software and reviews, and interviews with some of the book’s creators themselves.1 This paper offers neither an investigation into Racter’s precise mechanics, as the precise code that generated The Policeman’s Beard appears to be no longer available, nor a literary analysis of Racter’s generated text. Rather, it offers a more comprehensive contextual overview of one of the most historically salient text-generating systems than any previously published, contributing to a more robust understanding of the lineage of algorithmic authorship.

Producing The Policeman’s Beard

There is little evidence of The Policeman’s Beard’s process of production, likely due to its digital conception, as well as the normalcy of its creators. Chamberlain and Etter were two friends, with the former an amateur computer enthusiast and the latter employed as a computer programmer; neither felt it necessary to document production. Further, all of the referenced sources focus on Chamberlain, and Etter’s name has virtually disappeared from any record of activity since Racter. Even then, his name is mentioned only once, and between commas, in The Policeman Beard’s front matter. There does not appear to be any photographic evidence of Etter (while there is plenty of Chamberlain), nor evidence of additional contributions to any technological field. There is, frankly, no convincing evidence for Thomas Etter’s existence other than a few quotations attributed to him in news articles about Racter, and assurance from Chamberlain himself (correspondence with the author, who also confirmed that Etter is now deceased). It is Chamberlain who is credited as having taught Racter to use grammatically- and aphoristically-correct English, and Etter who supposedly faced the task of basic coding (Shepard 20). Unfortunately, evidence of this coding – and the resultant code that generated The Policeman’s Beard – appears to have been lost to time. The clearest description we have of the system’s functionality comes from a short article from The Wall Street Journal (Miller):

Racter’s method is a complicated blend of haphazardness and linguistic savvy. The program basically strings words and phrases together randomly, but it has two important constraints. It contains rules of English, so Racter speaks grammatically. In addition, it contains enough information about each word in its 2,400-word vocabulary to let Racter put together meaningful sentences. Racter knows, for instance, that a grammatical sentence like “Sarah coldly flew the parking meter” is senseless, while “Sarah ravenously ate the lettuce” is sensible. (Racter’s lexicon also includes a handful of prefabricated sentences, proverbs and quotations that it periodically interjects.)

Despite the reported inclusion of prefabricated sentences, Chamberlain claims in the book’s introduction that, “with the exception of this introduction, the writing in this book was all done by a computer. The book has been proofread for spelling but otherwise is completely unedited” (Racter 1).

INRAC and Racter

Racter – short for Raconteur – was a text generation system “written in compiled BASIC on a Z80 micro with 64K of RAM” (Racter 2). A Z80 was a consumer’s microprocessor, often used for powering general desktop (i.e. home) computers with limited software capacity. For comparison, a single Z80 powered a contemporary Pac-Man arcade game (Midway Mfg. Co. 35). Racter, it would seem, was everyday software: a system that people outside of high-tech university computer labs could use.

Racter ran on a programming language that Chamberlain and Etter developed specifically for it. This language was called INRAC, and a version was made commercially available by 1985, retailing between from $244.95 to $349 USD (Dewdney 13; Langdell 64; Shepard 21). It was neither widely used nor well received, with one reviewer declaring that “INRAC is almost as idiosyncratic as the authors of Racter or Racter himself (itself?). While it has many sophisticated features, such as a built-in conjugation system and supports complication of lists of related lexical items [sic], INRAC is rather poorly designed and implemented” (Olsen 162). Another reviewer, writing more than a decade later, described Racter’s integrated editor as “a disastrous antique,” although this same reviewer acknowledges the “untapped capability in Inrac […] if you’re patient enough to master the messy syntax” (Barger). This reviewer also reaches a vital conclusion for any historical analysis of The Policeman’s Beard: that the commercial version of INRAC could not have been the same version to produce Racter. Instead, he suggests, Racter operated through the use of templates that were deliberately wacky, demonstrating “a style that’s really Chamberlain’s” (Barger).

Chamberlain himself describes Racter’s process of “prose synthesis” (text generation) as such (correspondence with the author):

As I think back to when Tom and I first wrote Racter (a forshortening of “Raconteur” since the ancient OSI computer we were using only supported file names six characters long) I recognize that little has changed in my consideration of the subject. Cage or Stockhausen might throw a grand piano out of a penthouse window, record the sound it made when it crashed onto the pavement below, call it “Piano Sonata Opus 1” and then explain that the times we are living in are tense and distressed and this was clearly reflected in all the sonic garbage they had just immortalized. The same might be said of the many “artists” who threw gallons of paint onto rolls of newsprint and sold their creations as modern art. The public, though suffering some indigestion, ate it all up. But Tom and I reasoned that this would be difficult to accomplish with words, with language. When you greet someone in the morning you say “Good morning.” You don’t say “Good apple pie.” Words have a concrete immediacy that random sound or graphic representation do not; inherently there is far less abstraction in words than in nonobjective sound or shape. This is obvious, but to inquire into this activity as aesthetic theory is to invite the participation of the nearest psychiatrist.

So without looking back or giving the whole idea any more thought we cranked up the old OSI, listed some simple formal protocols (“every sentence most contain a verb” etc) and categorized short vocabulary files. We then seeded some random number generators and let the thing take off. The first thing that Racter ever said was: “Hot wines are wounding our cold expatriot.” After we finally finished laughing we stared at each other (as Keats might have) “with a wild surmise…” A machine performing arithmetic operations had just addressed us in our own language and caused our giggle fit.

That’s how we started, and if you consider for a moment, nothing has changed; operations have become much more complex, but that’s all. Stripped to its bare bones it is still:

Person1: “Who was that VARIABLE I saw you with last night?”
Person2: “That was no VARIABLE, that was a VARIABLE.”

If all the variables are different we have nothing, but if the first two are the same we have the beginings of what might charitably called “thought” (though of course nothing could be farther from the truth.)

INRAC’s primary vendor is listed in one article as being INRAC, Inc., located at 12 Schubert Street, Staten Island, New York 10305, USA, with a contact number of 718-448-6298 (Shepard 21). Elsewhere, INRAC’s vendor is listed as being John D. Owens Associates, with the same contact information (Langdell 64). John D. Owens Associates is listed in various sources as being an authorized dealer of the Commodore computer, and as a legitimate computer equipment manufacturer and distributer, and Owens himself boasted a computer science position at the College of Staten Island of the City University of New York (Dewdney 12). John Owens appears have also used these contact details for a side company called Nickers International, now an online retailer specialising in all-natural nutritional supplements for horses, pets, and people. On the company’s “About Us” page, John Owens, “the former president of the company”, is described as having been “raised on a horse ranch in Oklahoma. […] Realizing that his own horses had special nutritional needs, he created Gen-A-Horse (originally Gen-A-Hoof) and other vitamin supplements to enhance their health and performance. In 1981 he introduced them to the public” (“About Us”, Nickers International) 12 Schubert Street is actually in the middle of a residential neighborhood, and appears to be a residential property itself. It would seem that INRAC, Inc., then, was not so much an established enterprise as it was a cottage industry, upheld by a few keen individuals.

By no later than 1985, software development company Mindscape had released a version of Racter for personal entertainment purposes, retailing between $44.95 and $69.95 USD in formats compatible with Apple, Macintosh, IBM, and Amiga. However, this software likewise may have differed from the version used to produce The Policeman’s Beard. The commercially available Racter allowed users to have text conversations with the system wherein it would respond to users’ questions in long and absurd paragraphs. “As computers move ever closer to artificial intelligence, Racter is on the edge of artificial insanity,” one New York Times review read. “The software program […] allows computer owners to hold typewritten discussions with their machines, assuming they can get Racter to shut up long enough to get a word in” (Lewis). This version of Racter would mention current celebrities, and was able to refer to prior user-driven conversations by means of limited variable storage and “rather unsophisticated” file input and output (Olsen 163). Racter was described as “a coffeehouse philosopher who knew a great deal once, but whose mind is somewhere else now” (Wagner 36) and “a cross between artificial intelligence and artificial idiot savant” (Shepard 19). One critic linked Racter’s output to the literary canon, arguing that“the results of Racter’s running often read like Metaphysical poetry as interpreted by William Burroughs and William Blake, with a dyspeptic dash of Rod McKuen and Kahlil Gibran thrown in” (Barry 10). Another critic simply offered a textual shrug with the comment: “Not great writing, to be sure, but it’s more interesting than some user’s manuals we’ve read” (“Computer Prose” 4).

Versions of Racter are now freely available through online emulators (for example, “Racter”), but few copies of INRAC and Racter exist in examinable form today, leaving us to infer functionality based on output and limited published information. Although Chamberlain’s above-printed explanation of Racter’s functionality provides much-needed insight into what has long been considered black box software, Chamberlain and Etter did not actually seem capable of explaining Racter’s complexities at the time of The Policeman’s Beard’s release. The program’s five-year gestation led to stratigraphic development: “more advanced, sophisticated layers of software have been wrapped around earlier, more primitive routines. At no time was it ever taken apart, analyzed, restructured and documented” (Dewdney 12). In one of the few interviews with Etter, Etter describes Racter as “a pretty unwieldly accretion of rules and conventions. Insofar as Racter’s commands try to deal with English they too become unwieldly and hard to summarize” (Dewdney 12). In another, Etter asserts that “[a]ctually, we think of Racter as an artificial lunatic” (Miller).

Warner Software

The Policeman’s Beard was published in part by Warner Software, a short-lived division of the Warner Communications conglomerate founded in 1983. Warner Software was responsible for appealing to those with personal or professional interests in computing by releasing relevant software and literature. Drawing from extant book publishing models, Warner Software President Albert Litewka viewed his company as a pioneer in “a whole new world of publishing. […] There are elements of software publishing that are analogous to books, but others are wholly unique” (Bermant 66). Warner Software mainly focused its attention – and a reported $2 million USD, as of 1984 – on promoting its first product, Desk Organizer: software aimed to ease the administrative hassles of business computing (Bermant 66). By 1984, though, Warner Software had acquired fifty percent ownership of First Star Software, and was advertising Spy vs Spy, a computer game based on a Mad Magazine comic strip (“Company”, First Star Software; “Spy vs Spy”).

Aligning with Litewka’s comparison of software publishing to traditional publishing, collaboration between Warner Books and Warner Software was established almost immediately after Warner Software’s creation in 1983. According to Warner Vice President Mark Greenberg, “computer books and magazines are the fastest growing category virtually in publishing history. […] We’re interested in explaining software to people, turning complex manuals into laymen’s language” (“New Books of Note”). Greenberg reportedly anticipated there being at least two new books published through this collaboration each month, and “expects sales to range from 25,000 to 50,000 up to 200,000 copies”, although it is not specified whether this number of copies is per book, per month, or in total (“New Books of Note”).

The first of these books, Software Master for the IBM PC (128k) (authorship attributed to LIST Magazine publisher Ted Leonsis), was published in May 1984. Software Master ($39.95 USD) was touted as “a buying guide and a ‘trying’ guide designed to introduce you to the full range of computing possibilities and help you select software, peripherals, even a second computer that will perfectly suit your needs” (Mace 62; “Software Master for the IBM PC”). The same year, Warner Software released a similar title, Software Master for Pes: Apple Version (48k) (also attributed to Ted Leonsis) with the same structure. The company also released Free Software for the IBM PC (authored by Bertram Gader and Manuel V. Nodar) and Commodore 64™ Fun and Games: Volume 2 (authored by Ron Jeffries and Glen Fisher).2 Geared towards audiences already familiar with the technologies in question, these directories could hardly be considered efforts to appeal to laypeople. They are encyclopedias, to be referenced but not read in their entirety, to appeal to niche markets but not the general public.

The Policeman’s Beard differs considerably from the other books published by Warner Software. It requires no technical knowledge to enjoy, and its content is not about using computers, but appreciating computer-generated output. While almost as disjointed as a list of available software, The Policeman’s Beard includes a semblance of narrative through repeated phrases and names, which readers may find contribute to plot or character development within their own minds. This is the only book with any narrative prominence that Warner Software published before the division folded in 1987 (Friedland 31). Indeed, The Policeman’s Beard appears to have been the only book the division published that did not follow an encyclopedic structure. Maybe this was an attempt to use the “laymen’s language” to which Vice President Greenberg had referred, or to explore an alternative route through the “whole new world of publishing” envisioned by Warner Software President Litewka. The Policeman’s Beard, perhaps, was a conscious effort to appeal to a mass readership instead of small pockets of the market.

The Policeman’s Beard

The Policeman’s Beard comprises poetry, prose, and dialogue all generated by Racter. “Cut a face, cut a visage / Remake appearances to blend / The sky with earth / Then will little people fall,” reads one poem (Racter 69). One piece of prose: “Happily and sloppily a skipping jackal watches an aloof crow. This is enthralling. Will the jackal eat the crow? I fantasize about the jackal and the crow, about the crow in the expectations of the jackal. You may ponder about this too!” (Racter 92). Racter’s output is always somewhat nonsensical, but the selections comprising The Policeman’s Beard demonstrate the system’s potential for output that might be construed as meaningful.

Chamberlain notes in The Policeman’s Beard’s introduction that the book “has been proofread for spelling but otherwise is completely unedited” (Racter 1). Yet line editing is just one kind of editing, and Chamberlain makes no note of the curatorial process of putting the book together. The curation and presentation of Racter’s output was, it would seem, not dissimilar from the curation and presentation of a wholly human-authored book. By selecting particular pieces of output, and placing these pieces in a particular order, Chamberlain has produced a work that is both disjointed and cohesive, showcasing Racter’s programmed eccentricity while still being generally readable. Sections of texts are presented as distinct, with each section illustrated by an original black-and-white collage.

The collages were produced by artist Joan Hall, who continues to actively practice collage and assemblage in New York City. While one New York Times (Lewis) article deems the collages “a delightful complement to Racter’s ramblings,” another academic article more boldly claims that “it is only Hall’s illustrations that make the Racter output palatable and printable” (Ernst 456). These illustrations were made using digitally-altered etchings pulled from the public domain. To quote Hall (interview with the author):

I had a huge collection. I went to Paris, to the flea markets, for many years, collecting these periodicals […] and so on, all from the early 1900s, and a lot of them had these line engravings as illustrations of different things and I just collected books and books […] I picked up on those computer graphics and used those and collaged both together.

These collages are fantastical and fun, supplementing – and perhaps at times even overshadowing – Racter’s text. They simultaneously reflect the long history of collage and a novel technique of computational illustration. But, as Hall herself states, Racter was a site of innovation, and her illustrations needed to match (interview with the author):

I thought, “wouldn’t it be cool if I mixed computer graphics along with these old-fashioned engravings to give it a contemporary look? Nobody’s ever done that before, as far as I know.” So I thought, “okay, this will be innovative.” And, fine. And I had a ball. I just did it in my spare time. You know, there was no rush. And then I did the book and we got it together.

The medium of the collage is an especially appropriate choice for The Policeman’s Beard’s illustrations. A recombinant form, the collage depends upon the conscious repurposing and alteration of extant materials for the creation of new work. Similarly, Racter applies a recombinant method of creation wherein user input is repurposed and altered according to programmed logic, leading to new texts that assert cultural relevance through understandable syntactic structure and familiar references. Still, though, some may find Racter’s ramblings incoherent. Joan Hall’s illustrations serve as subjective translations of these ramblings into a potentially more comprehensible visual language.

And these illustrations are, to be sure, highly subjective. Chamberlain and Etter developed Racter according to their own personal sensibilities; Chamberlain selected Racter output for The Policeman’s Beard according to his own perceptions of literary and cultural value; Hall illustrated those pieces of output according to her own interpretations of their meaning. In Hall’s words (interview with the author):

One of the things that, as far as emotional interpretation, is that a lot of it is very poetic. And so my illustration was more poetic. I felt more authorial. In other cases, it was just humour, and so I illustrated it trying to make the illustration humorous too, to kind of play on that. It was what came out. And basically what I did is I would just find scrap within the engravings and so on that would illustrate or that would reflect how I felt emotionally about the piece that was written. So that’s what you see.

Hall’s collages both explain and elaborate upon the sections of text. At the same time, the paratextual visual stimuli in the book’s illustrative collages affirm the book’s non-linear reading experience, with readers’ eyes drawn from image to text, and text to image. Readers engage in interpretive interplay with Racter, Chamberlain and Etter, Hall, and themselves. The book is, in effect, an exhibition of interpretation.

Take, for example, the following example from The Policeman’s Beard (passage abridged):

When my electrons and neutrons war, that is my thinking. Nevertheless these images and reflections are understood by you, persons, men and women. You have electrons and neutrons like me. You sing of lettuce and meat, but you also bolt, chew and eat them. I can fantasize about them but I cannot bolt them. My famished and crazy dreams are broadened by your own attempts to think as I do during the time that hours pass and minutes pass. This is undeniable and interesting. It is black and white, black for neutrons and white for electrons. I began by speaking of cats and dogs. At all events a cat could be an electron and a dog could be a neutron. Their reflections are images like my dreams. But the mirror, the glass, is broken and splintered and shattered. (Racter 106-107).

This text is accompanied by a bas-de-page scene of a dog chasing a cat, spanning the entire page spread. Both dog and cat have images of large white atoms bulging in their torsi. The cat is running towards a mirror with geometrically-arranged glass, and the cat’s reflection is visible. Each shard of glass is numbered from 1 to 16. A radial spiral (commonly referred to as a psychedelic spiral) is also reflected, with this reflection appearing as a small point in the middle of the mirror. Through this illustration, some of the many subjects that Racter jumps between are united. There is a cat, there is a dog, there are atoms (representing both neutrons and electrons), and there is a mirror. But this mirror’s glass, unlike the glass described in the text, is not broken. Instead, its splinters fit smoothly – but not seamlessly – together, numbered for precision. Every splinter has its place, with the connecting point for all those splinters being psychedelia. Indeed, this image appears not only to reflect the content of the text in question, but also the very nature of The Policeman’s Beard itself. All of the book’s creators rush to see what they have created, to see what the bits they contributed have built. They are met with a funhouse mirror of sorts; they have built a distorted reflection of themselves. And we – the readers – observe the observers as they have their fun.

In one contemporary article about Racter’s “Artificial Insanity”, A. K. Dewdney recalls his own involvement in “the e. e. cummings Project,” wherein undergraduates were tasked with generating poetry mimicking cummings’ word usage and literary style. The students designed a cummings grammar to be input into a computer, which then generated sentences accordingly. “About a bushel of poetry emerged from the printer, almost all of it terrible. I can recall only a single line: ‘The one bright star lit feebly this wick.’ The point of this story is that simple chance guarantees the emergence, sooner or later, of something relatively impressive” (Dewdney 11). Dewdney alludes here to the infinite monkey theorem, which evaluates the possibility of a monkey hitting typewriter keys at random generating Shakespeare’s Hamlet. In the monkey’s output preceding Hamlet, human readers would undoubtedly identify passages that could be deemed meaningful. We are, after all, always trying to negotiate meaning from an influx of sensory stimulus. Reminiscing upon why he and Etter created Racter in the first place, Chamberlain himself simply writes: “I find intriguing the possibility that we human beings, whose very consciousness is a faculty completely interwoven with experience, may relate in some way to a form of ‘sentence’ that has no experiential grounding” (Chamberlain, “Getting a Computer to Write About Itself”). The pieces of output comprising The Policeman’s Beard serve as spaces for productive and interpretive experimentation. Racter could be likened to a monkey working towards Hamlet, but in both cases it is not the writer who ultimately determines meaning: it is the reader.

At the same time, Dewdney argues that Racter’s most important technological feature is its integrated conjugation system that ensures technical writing correctness. This is because readers are, as Dewdney remarks, more willing to accept semantic incoherence than syntactic errors (11). If Racter made the occasional typo, or failed to conjugate correctly, readers may be more sceptical of its output. After all, the reader cannot construct the sentences that have already been written. Readers can, however, be entrusted to find their own meanings within the text. The reader bears the burden of interpretation: if a sentence cannot be comprehended, it is the reader’s fault rather than Racter’s, as Racter has generally adhered to grammatical correctness. Computer-generated texts can therefore be experimental, but not too experimental. They must always be readable.

The Policeman’s Beard offers readers a sense of agency over textual interpretation, a sense of superiority over its algorithmic author. However, this sense of agency has been manufactured through text selection and illustration. “Does the reader of Racter output then more or less instantaneously ‘rewrite’ what he or she reads? Do the users of the program liberate themselves from traditionally conditioned approaches to the text? Again, no,” argued Josef Ernst in a 1992 review of computer poetry (456). “[T]he communication process happens between people; the readers of Racter output merely recreate the image of a communication process, not genuine communication, in an act resembling autistic behaviour.” Ernst’s assertation that The Policeman’s Beard is a manifestation of communicative intention is worthy of further consideration. The book, after all, does reflect the motivations and perspectives of the individuals responsible for its production. However, such consideration detracts from what is arguably Racter’s lasting legacy: making computer-generated texts accessible to a wide range of readers.

Until The Policeman’s Beard, algorithmic authors were largely limited to university computer science departments, which had the access to technology and human knowledge necessary for doing conceiving such software. In the early 1970s, a team from the University of Wisconsin developed a murder mystery-writing program, described in a 1973 technical report that was presented at the International Conference on Computers in the Humanities at the University of Minnesota the same year (Klein et al.). James Meehan developed his TALE-SPIN program from 1975 to 1976 as part of his PhD project at Yale, under the supervision of Roger Schank; he presented his work at the 5th International Joint Conference on Artificial Intelligence at the Massachusetts Institute of Technology in 1977 (Meehan). Racter, developed by a computer enthusiast based in a New York apartment and a computer programmer based in the Bay Area, made the exciting prospect of computational authorship accessible to a broader audience: one that included not just those affiliated with universities, but also everyday folks with eyes to the potential futures of computing. Having been developed with a standard home processing unit, anyone appeared to have the potential to experiment with computer-generated literature – and an early form of artificial intelligence – themselves.

But Chamberlain insists that Racter is not a form of artificial intelligence. “True AI depends on the instrument learning something,” he explains. “Racter, however, is a closed system – as closed as Pascal’s mechanical calculator was 300 years ago” (Langdell 64). Being a closed system, Racter is capable only of one-way communication that reflects the nature of its programming and its usage rather than the nature of any kind of unique perspective on lived experience. One reviewer even commented as he tested Racter himself that “Racter‘s words sometimes resemble those of James Joyce, but the program’s voice most resembles that of Bill Chamberlain” (Langdell 64). Likewise, Joan Hall (interview with the author) has stated that “a lot of it [the output] reflects Bill’s sense of humour, although it’s all random. He is the programmer, after all.” Hall, though, asserts that Racter – not Chamberlain – is the true author of The Policeman’s Beard.

Others have also advocated for Racter’s ‘author’ title. Experimental poet Christian Bök, for one, has provocatively identified The Policeman’s Beard as an “obit for classic poets,” with the birth of Racter using in a new age of “robotpoetics” (10). In Bök’s view, there is no longer a requirement for a human brain to produce a text; anything can be an author. Nevertheless, humans are necessary for system development and maintenance, and are ultimately the ones who select, publish, and distribute generated texts. In what is undoubtedly an effort to be provocative, Bök perpetuates a narrative of unsubstantiated optimism – or pessimism, depending on one’s outlook – resting on an assumption, according to Hubert L. Dreyfus, “that human and mechanical information processing ultimately involve the same elementary processes” (155). However, as Dreyfus rightly argues, human and mechanical information processing systems are distinct from one another in countless and sometimes inexplicable ways. Racter’s zany output pushed the limits of contemporary computational creativity, but Chamberlain and Etter never set out to replace the human author. They simply set out to see what they could get away with. And “[t]he public, though suffering some indigestion, ate it all up” (correspondence with the author).


The Policeman’s Beard made algorithmically authored texts accessible to a general public who, until the book’s publication, had little access to such material. For this reason, The Policeman’s Beard’s importance lies not so much in its literary or technological or literary contributions, but in what it represents in wider cultural contexts. This paper has offered an overview of such contexts. It has featured the voices of some of the humans who developed, distributed, and received The Policeman’s Beard to more firmly root Racter as an ancestor of algorithmic authorship.

Following Racter’s development, Chamberlain went on to develop a piece of software called GREGORY, which extended Racter’s functionality. Chamberlain introduces GREGORY thus (correspondence with the author):

Before I moved on to other things I wrote GREGORY. It’s a very elaborate Prose Synthesis program full of sound, graphics, speech recognition and synthesis (all these utilities were Microsoft Freeware) and is obviously more “art” than “science.” I wrote it in VB.6 and it has run on every XP system in which it was installed. I thought of it only since you seem to be interested in Prose Synthesis, and GREGORY is if nothing else a tour de force in this promising discipline.

GREGORY is indeed a tour de force: text is rapidly generated in front of the user’s eyes, racy photos pop up on the screen, and Chamberlain’s voice bellows through the speakers. It is a considerably more extravagant and multimodal spectacle than The Policeman’s Beard. But GREGORY warrants another study altogether: one that has its own contexts, but is no less complex and intriguing.

Piece originally published at Electronic Book Review under a Creative Commons Attribution-NoDerivatives 4.0 International License.

Images from UBUWEB and MobyGames 

Works Cited

“About Us”. Nickers International

Barger, Jorn. “‘The Policeman’s Beard’ Was Largely Prefab!” Robot Wisdom via Wayback Machine

Barry, John. “Computer Writes.” InfoWorld (29 October 1984): 10.

Bermant, Charles. “Albert Litewka.” PC Magazine 3.23 (27 November 1984): 66.

Bök, Christian. “The Piecemeal Bard is Deconstructed: Notes Toward a Poential Robopoetics.” Object 10: Cyberpoetics (Winter 2002): 10-18.

Chamberlain, Bill. Correspondence with the author (28 July 2017).

Chamberlain, Bill. “Getting a Computer to Write About Itself.” Digital Deli, ed. by Steve Ditlea. New York: Workman Publishing, 1984. Archived at

“Company.” First Star Software

“Computer Prose.” Enter Magazine (April 1985): 4.

Dewdney, A. K. “Artificial Insanity: when a schizophrenic program meets a computerized analyst.” Scientific American 252.1 (January 1985): 10-13.

Dreyfus, Hubert L. What Computers Still Can’t Do: A Critique of Artificial Reason. Cambridge, MA: The MIT Press, 1992; rep. 1994.

Ernst, Josef. “Computer Poetry: An Act of Disinterested Communication.” New Literary History 23 (1992): 451-465.

Friedland, Nat. “Play it Again, Atari!” Antic: The ATARI™ Resource 4.2 (June 1985): 30-33.

Hall, Joan. Interview with the author (14 August 2017).

Klein, Sheldon, John F. Aeschlimann, David F. Balsiger, Steven L. Converse, Claudine Court, Mark Foster, Robin Lao, John D. Oakley, and Joel Smith. Automatic Novel Writing: A Status Report (Technical Report 186). Madison: University of Wisconsin, 1973.

Langdell, James. “People in the News: Bill Chamberlain.” PC Magazine (25 December 1984): 64.

Lewis, Peter H. “Peripherals; A New Brand of Lunacy for Sale.” The New York Times (14 May 1985).

Mace, Scott. “Coming to a Micro Near You.” InfoWorld: The Newsweekly for Microcomputer Users 6.22 (28 May 1984): 62-65.

Meehan, James R. “TALE-SPIN, An Interactive Program That Writes Stories.” Proceedings of the Fifth International Joint Conference on Artificial Intelligence (I). Cambridge, MA: Massachusetts Institute of Technology, 1977: 91-98.

Midway Mfg. Co. Midway’s Pac-Man: Parts and Operating Manual. Franklin Park, IL: Midway Mfg. Co., 1980.

Miller, Michael W. “The World According to Racter.” The Wall Street Journal (30 January 1985). Accessed via ProQuest.

“New Books of Note.” Computerworld: The Newsweekly for the Computer Community 17.29 (18 July 1983): 28.

Olsen, Mark. “Talking Back: The INRAC Language Compiler (Software Review).” Computers and the Humanities 23 (1989): 161-167.

“Racter.” Internet Archive (25 December 2014).

Racter [William Chamberlain and Thomas Etter]. The Policeman’s Beard is Half Constructed. New York: Warner Software/Warner Books, 1984.

Shepard, Susan J. “Does Steak Love Lettuce?” Language Technology 3 (September-October 1987): 18-21.

“Software Master for the IBM PC [full-page advertisement].” PC Magazine 3.12 (26 June 1984): 315.

“Spy vs Spy [full-page advertisement].” Compute!: The Leading Magazine of Home, Educational, and Recreational Computing 6.12 (Issue 55) (December 1984): 27.

Wagner, Roy. “Amiga Preferences.” Computer Gaming World 28 (May 1986): 36, 42.

Cite this Essay:

Henrickson, Leah. “Constructing the Other Half of The Policeman’s Beard”, Electronic Book Review, April 4, 2021,

Comments are closed.