Literature

Synthesis: putting things together

To observe the textual codedness of digital systems of course implies the danger of generalizing and projecting one's observations of digital code onto literature as a whole. Computers operate on machine language, which is syntactically far less complex than human language. The alphabet of both machine and human language is interchangeable, so that ''text'' - if defined as a countable mass of alphabetical signifiers - remains a valid descriptor for both machine code sequences and human writing. In syntax and semantics however, machine code and human writing are not interchangeable. Computer algorithms are, like logical statements, a formal language and thus only a restrained subset of language as a whole.

However, I believe it is a common mistake to claim that machine language would be only readable to machines and hence irrelevant for human art and literature and, vice versa, literature and art would be unrelated to formal languages.

It is important to keep in mind that computer code,

{5: No computer can reprogram itself; self-programming is only possible within a limited framework of game rules written by a human programmer. A machine can behave differently than expected, because the rules didn't foresee all situations they could create, but no machine can overwrite its own rules by itself.}

and computer programs, are not machine creations and machines talking to themselves, but writings by humans.⁵ The programmer-artist Adrian Ward suggests that we put the assumption of the machine controlling the language upside down:

''I would rather suggest we should be thinking about embedding our own creative subjectivity into automated systems, rather than naively trying to get a robot to have its 'own' creative agenda. A lot of us do this day in, day out. We call it programming.''⁶

{6: quoted from an E-Mail message to the ``Rhizome'' mailing list, May 7, 2001}

Perhaps one also could call it composing scores, and it does not seem accidental to me that musical artists have picked up and grasped computers much more thoroughly than literary writers. Western music is an outstandig example of an art which relies upon written formal instruction code. Self-reflexive injokes such as ''B-A-C-H'' in Johann Sebastian Bach's music, the visual figurations in the score of Erik Satie's ''Sports et divertissements'' and finally the experimental score drawings of John Cage shows that, beyond a merely serving the artwork, formal instruction code has an aesthetic dimension and intellectual complexity of its own. In many works, musical composers have shifted instruction code from classical score notation to natural human language. A seminal piece, in my opinion, is La Monte Youngs ''Composition No.1 1961'' which simply consists of the instruction ''Draw a straight line and follow it.''⁷

{7: Galerie und Edition Hundertmark. George Maciunas und Fluxus-Editionen, 1990. no page numbering }

Most Fluxus performance pieces were written in the same notation style. Later in 1969, the American composer Alvin Lucier wrote his famous ''I am sitting in a room'' as a brief spoken instruction which very precisely tells to perform the piece by playing itself back and modulating the speech through the room echoes.

In literature, formal instructions is the necessary prerequisite of all permutational and combinatory poetry.⁸ Kabbalah and magical spells are important examples as well.

{8: Some historical examples have been adapted online on my website "http://userpage.fu-berlin.de/ ~cantsin/permutations"}

But even in a conventional narrative, there is an implict formal instruction of how - i.e. in which sequence - to read the text (which maybe or followed or not, as opposed to hypertext which offers alternative sequence on the one hand, but enforces its implicit instruction on the other). Grammar itself is an implicit, and very pervasive formal instruction code.

Although formal instruction code is only a subset of language, it is still at work in all speech and writing.

It is particularly remarkable about computing that the namespace of executable instruction code and nonexecutable code is flat. One cannot tell from a snippet of digital code whether it is executable or not. This property does not stand out in the alphabet of zeros and ones, but is solely dependent on how another piece of code - a compiler, a runtime interpreter or the embedded logic of a microprocessor - processes it. Computer code is highly recursive and highly architectural, building upon layers of layers of code.