«Creativity Support for Computational Literature By Daniel C. Howe A dissertation submitted in partial fulfillment of the requirements for the degree ...»
Funkhouser writes  that “in I-VI, Cage employs elaborate processes and contributes significant input in generating his non-intentional work. He composes or identifies a source text that he uses as an ‘oracle’, and asks it what words to use for each letter of the (vertical) poem, a process that, he writes, “frees me from memory, taste, likes and dislikes” [Cage 1990]. Mesolist lists all words in the source that satisfy the mesostic rule, then IC selects words from the lists. The forty-five characters to the right and left of the chosen words in the original text (“wing words”) are included, and Cage removes those he does not like [Cage 1990]. To prepare these lectures, Cage writes, “four hundred and eightyseven disparate quotations have been put into fifteen files corresponding to the fifteen parts of [his text] Composition in Retrospect: method, structure, intention, discipline, notation, indeterminacy, interpenetration, imitation, devotion, circumstances, variable structure, nonunderstanding, contingency, inconsistency, and performance” [Cage 1990]. The source texts for the lectures included Composition in Retrospect, and a range of other sources, including writings by Henry David Thoreau, Ralph Waldo Emerson, L.C. Beckett, Fred Hoyle, Marshall McLuhan, Buckminster Fuller, Gene Youngblood, and from articles that had appeared in daily newspapers.
After giving the initial lectures, he realized the need to establish a simple notational system that would instruct him to take a breath when reading the work aloud (i.e., “ ’” [space apostrophe]), and where to stress syllables that, “would not normally be stressed but should be” (i.e., bold typeface).” [Cage 1990] As a short example of one of Cage’s mesostics, this passage from Finnegan’s Wake:
Just a whisk brisk sly spry spink spank sprint of a thing theresomere, saultering, Saltarella come to her own. I pity your oldself I was used to. Now a younger’s there. Try not to be part! Be happy, dear ones! May I be wrong!
For she’ll be sweet for you as I was sweet when I came down out of me mother. My great blue bedroom, the air so quiet, scarce a cloud. In peace in silence.
becomes the following mesostic:
As Cage writes in the preface to an essay titled “Anarchy”, these works “… do not make ordinary sense. They make nonsense…. If nonsense is found intolerable, think of my work as music, which is…a question of repetition and variation, variation itself being a form of repetition in which some things are changed and others are not” [Funkhouser 2007]. A similar sense of the language as musical is found in works by Mac Low and some by Hartman (with and without Kenner). Kenner himself describes the look of Cage’s Sentences on the page as “Chant, therefore Voice” [Kenner 1995].
In her discussion of this work in Radical Artifice, Perloff observes that Cage prefers “to let us participate in the process whereby unfinished news items and bits of information…can be absorbed into the rhythms of individual consciousness; they remain discrete entities that we restructure according to our own predilections” . This focus on process is a trait that procedural and computer literature often manifests. The procedural (or machinic) process is naturally highlighted (as discussed in the lessons section) and is cast against both our ‘traditional’ processes of writing and inevitably, as noted by Perloff in the fragment above, our thought processes, and even, the processes of consciousness itself.
Cage’s manipulation of chance elements mirrors the work of Jackson Mac Low (discussed below).
4.4.8 Jackson Mac Low Mac Low, already an established print poet when he began experimenting with computer processes, was one of the first American poets to use chance methods in his precomputer work90. These methods were not, however, the only means that Mac Low deployed in his larger attempt to solve the problem of producing so-called "egoless" poetry and music, a project particularly suited to the use of computational processes. Like Cage, Mac Low made use of the concept of nonintentionality, in which the artist no longer makes all decisions in her compositions, but instead lets chance control certain elements.
Mac Low created his first computer poems at the Los Angeles County Museum in the summer of 1969, using a PFR-3 programmable film reader designed for graphics applications connected to a DEC PDP-9 computer. Mac Low’s program, he explains in Representative Works: 1938-1985, selected and combined words from a list of short messages he had composed. The program’s database (the “message lists”) and selection process allowed Mac Low to create “an indeterminate poem, of which each run of the printout is one of an In the late 1950s Mac Low started experimenting with chance methods using a copy of One Million Random Digits, and 100,000 Normal Deviates, created by the RAND Corporation for use in Monte Carlo algorithms.. He continued using this book in his work throughout the sixties and beyond [Mac Low(b) 1997].
indeterminable number of possible realizations” [Funkhouser 2007]. The following excerpt
demonstrates the program’s style:
THE TREES TURN TOWARD THE LIGHT. (p214-15) Mac Low also used deterministic methods to generate poetry whose form and content were not known in advance, but could be reproduced given identical initial conditions [Mac Low(b) 1997]. His approach to writing poetry was literally “experimental”. The primary question he would ask was, “how can I achieve a certain effect?” as opposed to “what will happen if I implement this particular algorithm?” As his son, Mordecai-Mark notes [Mac Low(b) 1997], his approach was similar in many respects to that of an applied mathematician or computer scientist who studies both the general properties of algorithms and their adaptation to specific applications. He was not, of course, attempting to prove theorems or support particular theories with his experiments, but rather to empirically invent techniques of artistic production meeting certain criteria.
Mac Low began working with computers sooner even than many scientists. While he was an instructor at New York University in the late sixties, he took advantage of a free course in FORTRAN (one result was a short poem exploring output of a flawed program that showed gravity increasing as the height of the simulated fall increased). This was during the period when programs were written by hand on coding forms, which were then run through a computer in batch mode. In 1969 he was offered a fellowship by the Los Angeles County Museum of Art, to make so-called 'verbal' artworks at an IBM facility in Los Angeles, and he subsequently ended up working with Information International, Inc., "Triple-I", a company that went on to become one of the dominant computer graphics and printing companies.
There he worked with John Hanson, the VP of programming, and his assistant Dean Anschultz, to write assembler language code for displaying poetry on a Tektronix vector graphics screen, and eventually for printing it out as well, using a programmable film reader driven by an early minicomputer [Mac Low(b) 1997].
Throughout the seventies, before the widespread availability of microcomputers, his work made more use of audio electronics than computers. He finally got his own machine in 1987 which “let loose the usual flood of manuals over his workspace” [Mac Low(b) 1997], but gradually gave him additional tools as well. He made at least one further attempt to learn to program (this time in the C language), but eventually came to rely on already written software.91 In fact, much of Mac Low's computer-based work was realized in collaboration with Charles O. Hartman (see following section), author of Virtual Muse: Experiments in Computer Poetry. In the 1980s, Hartman automated many of Mac Low’s procedures for His use of software often involved creative ‘misuse’, as discussed in Chapter 3: Pedagogy, as when he completely filled the available space in his word processor's glossary with phrases drawn from work of and about Kurt Schwitters [Mac Low(b) 1997].
computer, including Mac Low's ‘diastic’ procedure, originally developed in 196392 and automated by Hartman in the late 1980's as part of the DIASTEXT program, which Mac Low began using in earnest in 1989.93 These programs were used extensively in his 42 Merzgedichte In Memoriam Kurt Schwitters (1994), a series of poems constructed by a variety of procedural methods, all employing found texts relating to the artist Kurt Schwitters.
Mordecai-Mark describes the project:
The first poem in 42 Merzgedichte in Memoriam Kurt Schwitter] [Mac Low, 1994], was written by impulse-chance-selection from sources about and by Schwitters, the 2nd through the 30th by computer-aided chance operations, and the 31st through the 42nd by computer-automated diastic methods, which in some of the last Merzgedichte were supplemented by use of Hugh Kenner and Joseph O'Rourke's program TRAVESTY, which produces what Kenner calls "pseudo-texts," determined by letter-group frequencies in English [Mac Low 1998].
As Mac Low himself notes, the writing of several sections in this piece involved Kenner and O’Rourke’s TRAVESTY program94, based on Shannon’s n-gram procedure.
"Diastic" is a word coined by Jackson from the Greek words "dia" (through) and "stichos"(a line of writing, a verse) and is contrasted to "acrostic." (from "akros" (an extreme, such as the letter at the beginning or end of a verse line). "Acrostic" reading-through procedures draw words and other linguistic units from source texts by "spelling out" their titles with linguistic units that have the letters of the words in the titles as their initial letters.
One reads through a source text and finds successively linguistic units spelling out the title as follows: the units spelling out individual words comprise single lines (often long ones) and the series of lines spelling out the whole title comprises a stanza. (The "asymmetries" are nonstanzaic but still partially acrostic.) He also experimented extensively with DIASTEX4, an improved version of the program which allows the user to choose and employ a separate index instead of using the whole source text as the index.
TRAVESTY was written by literary critic and James Joyce expert Hugh Kenner wrote, in collaboration with the programmer Joseph O’Rourke, and its, a text recombination program Kenner, a well-known literary theorist, was co-author (with Hartman), of Sentences, a volume of poems generated also with the help of TRAVESTY (For a full-description of TRAVESTY and its many incarnations, see Noah Wardrip-Fruin’s dissertation ”Expressive Processing” ). Just as John Cage used the computer to facilitate work that he had previously performed manually, Hartman’s program mechanically accomplished—with some variation and advancement—the procedural work that Mac Low had practiced for many years.
According to Funkhouser , once his attention became focused on Schwitters, Mac Low devised a computer program that would randomly select linguistic units that his initial poem for Schwitters ("Pieces O' Six: XXXII") stored in a “glossary” in Microsoft Word and process these fragments into what Mac Low describes as, “entirely new constellations”. Over the course of two years, Mac Low implemented modifications to the program, its glossary, and made other adjustments to create a substantial body of poems.
With 31st Merzgedichte in Memoriam Kurt Schwitters he began to incorporate DIASTEXT
and TRAVESTY into his process; his uses of the programs were quite precise:
I utilized these programs in different ways, employing earlier Merzgedichte as source texts: (1) For the 31st Merzgedichte, I ran the 25th Merzgedichte through DIASTEXT alone. (2) For the 32nd, I ran the 4th through DIASTEXT alone. (3) For the 33rd, I ran the 2nd through DIASTEX4 alone.
(4) For the 34th, I ran the 8th through DIASTEX4 alone. (5) For the 35th, I ran the 9th through DIASTEX4 alone. (6) And for the 36th through the 42nd, I ran the 29th first through TRAVESTY, asking for “low-order” output--i.e., scanning for sequences of very few characters, to insure the outputting predominantly of letter strings that aren’t real words (pseudo-words), along based on the Markov model. Dubbed Travesty, its source code was published in a 1984 issue of the popular computer magazine BYTE [Kenner & O’Rourke 1984]. For the algorithm, Kenner credited the “long-ago idea from the Father of Information Theory, Claude Shannon.” The code was adapted in 1990 by Larry Wall, creator of the Perl programming language, and published as a programming example in the 1st edition of the book Programming Perl. The second edition of the same book featured examples of “Perl poetry”[Wall and Schwartz 1988]”. 25 (see p. 94).
with a few real words, most of them embedded in pseudo-words--and then through DIASTEX4.
An excerpt from 29th Merzgedichte in Memoriam Kurt Schwitters:
Alsend brub HAgmes mencess kInces AumeIng As has been noted repeatedly, Mac Low and Cage share at least some goals as artists, and their work is often discussed in similar ways, though Mac Low himself suggests that too
strong a parallel has been drawn between the two:
The thing is that there is too much pairing of John’s and my work, despite our strong mutual regard. We're both concerned with intentionality and nonintentionality and started doing this sort of work from understandings of Buddhism, especially Zen & Kegon as taught by Daisetz Suzuki at Columbia University in the 40s and 50s. But I seldom used “pure” chance operations after 1960. My algorithmic work is often mistakenly thought to be chancegenerated, as they say, and I too used to think it was “chance-generated” work, but I realized sometime in the 80s that the only chance involved is in the making of mistakes (and after a certain point--especially book publication--the mistakes must be accepted as integral to the works).