Interlude 1
EXE TXT
If texts are rendered executable in encounters with readers, how might readers hack into their code? Lisa Samuels and Jerome McGann designate the type of expanded scholarly poetics presented in these interludes as a mode of “deformance,” a practice-based approach that moves “beyond conceptual analysis into the kinds of knowledge involved in performative operations.”1 In their foundational essay “Deformance and Interpretation,” they begin with a consideration of Emily Dickinson’s proposal to read poems backwards, line by line, as a tactic for releasing hidden vectors of meaning within a work. By reworking the material of the poem, they contend, new knowledge might emerge from its tactical rearrangement. If this simple operation contains the potential for “rethinking our resources of interpretation,” then Textz and the little databases I examine all offer substantial inventories prepared for modes of interpretive poetics.2 Other practitioners have used more expansive terms for this kind of activity: “creative scholarship,” “critical making,” “applied poetry,” “qualitative digital humanities,” and, most simply, “remix,” (among others).3 Trained in the creative forms of critical knowledge building in the expanded lineage of L=A=N=G=U=A=G=E, discussed in the following chapter, I prefer to group my creative scholarly interventions in this book under the heading of “media poetics.”
This “continuation of poetry by other means,” in Charles Bernstein’s formulation, is emphatically prepared for questions of meaning-making in experimental forms and formats that appear online.4 The collection subordinates the “content” of text and theory to the poetics of transmission, encryption, and format. In alignment with Bernstein, media archeologist Lori Emerson emphasizes the illuminative capacities of media poetics within the play of reading and writing alongside networks and algorithms that also read and write their users.5 Given the contingent specificities of format and context in the play of “readingwriting” the criticism of digital objects, this project invokes strands in the digital humanities that turn not only to data visualization and information design for interpretative inroads, but also to experimental art and poetry. Further, the project contends that these modes can work in cooperation, given the right conditions. Textz navigates a path between code and concept, collection and dispersion, text and software—all within a coherent set of aesthetic and political aims. In the following interlude, I explore a scholarly mode that might deploy one vector of Textz’s poetics of the variable format.
When I began conceptualizing the preceding chapter, one outrageous idea occurred to me. What if I were to exclusively deploy texts within the Textz collection to theorize the site? With its transhistorical assemblage of media theory, internet criticism, and related readings, this seemed a productive gesture. How might a collection theorize itself? Or, how could I develop an approach that would allow Textz to “speak for itself,” as it were? This idea was abandoned for a number of reasons: institutional, representational, and critical alike. The gains afforded by studying the site’s operations from the present seemed to outweigh the benefit of inhabiting the discourse of the collection itself. At least this seemed to be the case for the chapter proper, which by necessity focuses more on the contexts and techniques of the site than on the works that it contained. This decision also closed off opportunities for close and sustained reading of specific texts within the collection (tactics to which I’ll return in chapter 4). Given these limitations, I began to consider a scholarly poetics project that might accomplish the same task beyond the book you’re reading.
This work, entitled EXE TXT, endeavors to perform the collection’s self-theorization in both format and process. It includes a final document (as both a plain text file and print-on-demand book), a python script, and an archival record of the complete set of Textz textwarez. My intervention includes a text that was constructed with a relatively simple python script that extracts sentences containing variants of the strings “text” and “software” from the complete set of textwarez (including multilingual strings like texte, testo, and texto). In this way, it queries how text might meet software within the works that Textz hosts. The complete python script is as follows:
import re
import sys
import random
def write_lines(filename, data_lines):
with open(filename, ‘w’) as f:
f.write(‘ ‘.join(data_lines))
f.write(‘\n’)
def match_word(text, word):
return re.search(r’\b{0}\b’.format(word), text)
def split_sentences(filename):
with open(filename, ‘r’) as f:
data = f.read()
data = data.replace(‘\n’, ‘ ‘)
parsed = re.split(r’([.!?])’, data)
text = parsed[0::2]
ending = parsed[1::2]
return [t+e for t, e in zip(text, ending)]
if __name__ == ‘__main__’:
if len(sys.argv) < 3:
print(‘usage: {0} input_file word’.
format(sys.argv[0]))
sys.exit()
filename = sys.argv[1]
word = sys.argv[2]
basename, ext = filename.rsplit(‘.’, 1)
outfile = “{0}_{1}.{2}”.format(basename, word, ext)
lines = split_sentences(filename)
target_lines = [line.strip() for line in lines if match_word(line, word)]
random.shuffle(target_lines)
write_lines(outfile, target_lines)
Figure 1i.1. EXE TXT perfect bound paperback book (6 in. × 9 in.; 242 pp.); image captured by “Library of Artistic Print on Demand” (apod.li/exe-txt).
Figure 1i.2. EXE TXT perfect bound paperback book, open spread; image captured by “Library of Artistic Print on Demand” (apod.li/exe-txt).
The output of this script was then edited for coherence and consistency within the generic conventions of the academic monograph. The line breaks are mine, as well as certain minor typographical adjustments. Nevertheless, the aim is to have the collection speak for itself as clearly as possible—to offer a poetic intervention via code that surfaces otherwise unseen connections in the little database. It is presented in plaintext ASCII format as another work of textwarez, which I first published with the internet publisher Gauss PDF in 2015, adding its critical form into conversation with the experimental poetry projects published there. Like all linear textual formats, it presents one narrative trajectory among the limitless potential routes a user might follow through the database.
However, it is also a little database, or inventory of effects, in its own right. The reader/user is invited to repeat this operation on the Textz corpus using the enclosed python script, both of which are included in the publication EXE TXT. Simple terminal instructions for executing the python script on the corpus are contained in a “readme.exe.txt” file. More than a statement of authorial invitation, it is a prepared command that initiates a program called textwarez for any user of the book. The final two lines of the python script read “random.shuffle(target_lines) / write_lines(outfile, target_lines).” With these two lines of code, the extracted sentences are randomized and written in succession to a new text file. You may select your own terms for generating your own random-access Textz study. Or you may rewrite the program or throw it out altogether to generate something new from the corpus. Offered as both a media poetics of Textz and a tool for future play, EXE TXT hopes to render itself variable, encompassing any subsequent texts derived from its mechanism.
See:EXE TXT (2015).