Dataset: Linked & Codebook: Linked
Background
The term “cyborg” was first coined by Manfred E. Clynes and Nathan S. Kline in 1960 to refer to a hopeful future where technology—specifically, osmotic pumps—could aid humans in traversing previously unsuitable environments, namely, space (Kafer 126). The birth of “cyborg” to refer to an imagined cybernetically modulated individual coincided with, if not encouraged, the emergence and popularization of the cyborg in science fiction film and literature, such as Westworld and Star Trek (films), “The Six Million Dollar Man” (tv series), its literary inspiration Cyborg, and its spin-off The Bionic Woman (novels). While Clynes and Kline did not necessarily have disability in mind with their coinage of the term in “Cyborgs and Space” (1960), they envisioned the use of psychotropic drugs and other pharmaceuticals as being integral to their project, and as such, the term was also adopted in clinical settings to refer to people with disabilities and chronic illnesses who use biotechnologies (Kafer 223; Kafer 126). Although this dataset does not intend to stage an examination of the history of the cyborg in popular culture, the term’s novelty and its roots in experimentation and modulation are important in understanding its multi-variant applications today.
“Cyborg” gained potency in the Academy in 1985, with Donna Haraway’s “A Cyborg Manifesto,” which intended to encourage a socialist, tech-positive feminism that was not so beholden to identity categories. While the use of the term “cyborg” as a metaphor for fluidity, futurity, and transgression was and is in some ways generative, Haraway took little care to acknowledge the lived reality of those most explicitly invoked by the term: disabled and chronically ill people who use technology for safety, if not survival.
Haraway’s omission has rightfully inspired pushback among disability theorists. In Feminist, Queer, Crip, to begin, Alison Kafer questions Donna Haraway’s willingness to fully grapple with the stakes of her claim that “cyborg bodies are not innocent, but… ‘maps of power and identity’,” as she does not fully consider the risks, costs, compromises, and social differences that might serve as forms of impasse to liberation through techno-hybridity (105). In “Disability in Theory,” Tobin Siebers disavows Haraway’s evasion of pain as a phenomenon of disability and as a potential side effect of technological integration. Chela Sandoval calls attention to Haraway’s appropriation of what she has termed “U.S. third world feminism” (76). And to invoke more recent work, in “Common Cyborg,” Jillian Weise reminds us that the implementation of the term “cyborg” as a take-or-leave metaphor erases, rather than liberates or makes visible, disabled women (69-70).
These criticisms should be taken seriously and are useful in highlighting the need for attention to metaphor in academic writing. While several academics in literary, film, and gender studies have addressed the presence of cyborg-like figures in popular entertainment, it seems useful to push further on contemporary and emerging connections between the cyborg of the Academy and the cyborg in popular culture and social media. This dataset, therefore, aims to extend criticisms and questions directed toward Haraway to the public imagination.
Kafer contends that “the cyborg of popular culture bears little resemblance to the cyborgs of Haraway’s manifesto” (106). And it is true that the cyborgs of superhero films and (most) popular sci-fi novels do little to advance a feminist agenda. Still, it seems worthwhile to ask how the fictional figure of the cyborg shapes and responds to public understanding(s) of identity, capacity, and body, and further, how theory and para-academic writing communicates with and perhaps reflects this understanding. If Haraway’s manifesto is wrought with misconceptions about disability, illness, and other forms of debilitation all while it attempts something laudable, I ask, here, what contradictions in and of the cyborg exist in the larger cultural imagination.
In order to better situate connections between problematic metaphors in the Academy and popular culture, I turn again to Weise. I am compelled by Weise’s term “tryborg” and its broad application across technologically interested consumers and producers of media. Weise defines a tryborg as “a non-disabled person who has no fundamental interface… and tries to integrate with technology through the latest product or innovation,” and in opposition to actual cyborgs, that is, “disabled people who interface with technology” and “depend on a computer for some major bodily function” (par. 3). As examples, Weise contends that a cyborg could be “an early adopter, a pro gamer, a TED Talker, a content creator or a follower… an expert who writes about cyborgs for screenplays, lab reports or academic journals’’ (par. 4). In a fruitful discussion with Alice Wong and Ashley Shew for the Disability Visibility Podcast, Weise offers that Haraway is just as easily a tryborg as Google’s CEO of Engineering, Ray Kurzweil. And Alice Wong helpfully describes her conception of the tryborg as someone who “uses tech… as accessories, but it’s not necessarily central to their existence” (pg 3).
The term “tryborg” points to a type of techno-ableism that fetishizes speed, optimization, and productivity. This dataset, then, surveys culture writing and social media to ask whether and how use patterns of the term “cyborg” are informed by a certain tryborgianism, and whether–and, if so, where–pushback to this figuration of the cyborg exists.
Description & Methods
Through this dataset, I look at the use of “cyborg” across two different domains: (1) literary and culture magazines and (2) Twitter user data and tweets. In more specific terms, this dataset (1) makes visible all searchable articles that use term “cyborg” across a variety of popular literary and culture magazines in the past ten years (2012-present), (2) provides a directory of existing Twitter users who identify as “cyborg” in their username, provided name, or bio, and (3) provides a directory of tweets that use “cyborg” both (1) in general and in conjunction with (2) “book” (3) “science fiction,” “sci-fi,” or “#scifi (4) “movie,” (5) “human,” and (6) “tech” (within 7 days of my search). I use “cyborg*" (asterisk included) to describe the information presented because I have mined each source in a way that provides not only utterances of cyborg, but also cyborgs, cyborg’s, cyborgian, and any other iteration where cyborg is the root.
While I relied upon various applications to obtain and mine the information included in this dataset, the dataset itself is contained within a .csv file in Google Spreadsheets. With respect to literary and culture magazines, information from each publication is stored in its own sheet, and with respect to Twitter, user information is stored in one sheet, while Tweet information is separated by co-keyword (or lack thereof). (I.e., “Twitter_cyborg data” is one sheet, and “Twitter_human data” is another.) See Codebook for full field information.
Literary Magazines & Data Mining in Voyant
With respect to the former domain (literary and culture magazines), I limit my sample to The New Yorker, The Atlantic, The New Inquiry, and Vanity Fair. While there are, of course, other literary and culture magazines that include discussions and figurations of the cyborg, I focused on these four publications to keep my sample manageable (in order test the efficacy of my methods), engage with sources that are read by both academic and non-academic audiences, and finally, work from publication websites that would allow me to access more than just a handful of articles at a time. I chose to limit my sample to the past ten years because I am interested, here, in recent evolutions and applications of the term “cyborg,” rather than its broader history. In order to obtain data and metadata from each publication, I turned to Voyant, a platform for text analysis that was developed by Stéfan Sinclair and Geoffery Rockwell in 2003. The affordances of Voyant in large part informed the fields I decided to include in the dataset, and this half of my dataset likely would not have been possible within my given timeframe without Voyant or an adjacent platform.
Task 1: Getting Corpus to Voyant
How, though, did Voyant aid in this process? The initial steps were fairly tedious, and to best create a visual, I will use the The New Yorker as an example: (1) Open Voyant. (2) In a separate tab, visit https://www.newyorker.com/. (2) Locate the general “search” function within The New Yorker website and search “cyborg.” (3) Sort search by “newest.” (4) Right click the title of each article (yes, one by one) and “copy link address” into the box at the center of Voyant’s homepage, under “add text.” (5) With all links (dating back to 2012) copied into Voyant, click “Reveal” on the homepage (see Figure 1 for intended outcome at this stage).
figure 1: screenshot of Voyant’s default text analysis in response to The New Yorker corpus
Task 2: Text Analysis in Voyant
From here, I was able to generate much of my desired information about each article. Due largely to copyright constraints, however, this information was limited to title, author, date, keywords, count, and collocates. This dataset does also include links, but all links were copied and pasted manually. The steps to obtain my fields worked as follows:
(1) Choose the “Documents” tab in the bottom left of the application. (2) Hover mouse above the top right of any one of the fields within the tool (e.g., hover mouse in the top right of “Tile”). (3) A down arrow should become visible; click on it and then select “Columns.” (4) Under “Columns,” select “Title,” “Author,” “Date,” and “Keywords.” (5) Voyant will populate the “Documents” tool with all fields it can locate based on the website-specific HTML tags. Some fields will be missing or inaccurate; remove these fields (for example, HTML files from The New Yorker will not allow Voyant to pull author and date, but HTML files from The Atlantic will). (6) As the final step in this stage, hover mouse over the top right of the “Document” tool and click on the left-most icon (to export the data). (7) Under exportation options, click “Export Current Data” then “export data as tab separated values (text).” The screen should look like the figure below (Figure 2). (8) Open a .csv file. (9) Paste this data into the newly created .csv file.
figure 2: screenshot of the exportation of first several fields yielded by Voyant as tab separated values using the “Documents” tool
In order to obtain the remaining fields, collocates and count, I relied upon two tools in the upper right hand corner of the application, “Document Terms” and “Collocates.” While “Document Terms” is a clear option (found to the right of “Trends”), “Collocates” can be selected in the top right corner of the tool window, under the second icon (next to the exportation option). The steps to extracting count and collocates are fairly similar to those for extracting the first several fields.
To access count: (1) Under the “Document Terms” tool, type “cyborg*“ into the search function. (2) Hover over any field within the tool and select only “#” and “count.” (3) In the same manner as with the fields located in the “Documents” tool, export data as “tab separated values (text).” (4) Paste new data next to existing data in the Spreadsheet, and sort “count” column by document ID number (“#”), ascending, so that the counts yielded by Voyant match the initial order of the documents.
To access collocates: (1) Select “Collocates” tool in the upper right corner. (2) As the identification of collocates must happen one document at a time; ensure that only one document is selected per search. (3) Again, type “cyborg*“ into the search box. (4) Hover over any field within the tool, and select only “Collocate” and “Count.” (5) Export collocates for each document as “tab separated values (text)” and begin to populate the collocates field document by document. This time, both values will go within one single field.
Task 3: Cleaning, Sorting, & Adding Data
Once all data from Voyant was added to my Spreadsheet, I finally sorted the entire sheet data by date. In the case of all publications except for The New Inquiry, I was able to copy all links at once into each sheet from my initial search (stored in a separate note for ease of transference), already arranged by date. The biggest hurdle in using Voyant to obtain data and metadata was the time it took to correct errors in identification, born not out of Voyant’s limitations, but in the lack of straight-foward tagging of certain fields, such as author and date, in the original HTML files embedded in article links (that would have allowed Voyant to pull said fields).
Twitter Data, Twitter API, & Postman
In order to most effectively obtain and use Twitter data, I first applied for access to Twitter’s basic (“elevated”) Application Programming Interface (API). To do this, I visited Twitter’s Developer Access page, selected the appropriate access level (“elevated,” based on my apparently limited credentials) and filled out an application form in which I described how I planned to use the API. My application was followed by a series of emails that prompted me to specify a bit more about my project and affiliations. After a day or two of back and forth, I was granted access. The steps to obtain necessary data, then work as follows:
(1) Use existing Twitter credentials to login to Twitter Developer Portal. (2) On the left side of the portal, click on “Projects & Apps” to create a first project and corresponding app. (3) Once a project has been defined and an app has been created, Twitter grants API access keys. Go to “Dashboard” within the Developer Portal and click on the key icon under “Development App” to generate keys. (4) Save keys in a separate, secure document for future use. (See Figure 3.)
figure 3: screenshot of Twitter Developer Portal “Dashboard,” with emphasis on location of Access Keys
Using Postman
Getting access to the Twitter API was only the first step. The actual data extraction happened not through the Developer Platform itself, but through a website called Postman, which facilitates the use of various APIs. Interaction with Postman for the specific purposes of this dataset worked in slightly different way for each keyword search (or, each Sheet or topic in the dataset). In all cases, the first step involved entering my unique API keys under the “Authorization” tab within the Postman homepage. The remaining steps are below, separated by category.
To extract user information: (1) open Twitter and search “cyborg” under “people.” (2) In Postman, within the Twitter API, go to “User Lookup,” found on the left-hand side of the webpage. (3) Under “User Lookup,” go to “GET Users by Username.” (3) Copy and paste up to one hundred usernames from the Twitter user search into the “usernames” field under “Query Params.” (4) The variables for “user.fields” should be set as follows: “created_at,id,description,location,name.” (5) Press “Get.”
To extract tweet information: (1) In Postman, within the Twitter API, go to “Search Tweets.” (2) Select “Recent Search.” (3) Under “query,” within “Params,” define inital value as “cyborg,” and add any additional keywords; finally, in “query,” add “-is:reweet,” to avoid duplicates. (4) The other parameters should be defined as follows: max_results: 100; tweet_fields: created_at,lang,context_annotations. (5) Press “Get.” (See Figure 4.)
figure 4: screenshot of parameters and results for search within the Twitter API via Postman
In both cases (user and tweet data), Postman will generate results in JSON format by default. For my purposes, to keep data congruent, I used a conversion tool (there are several accessible via quick Google search) to transfer the result into .csv format, which I then copied into my Spreadsheet. From there, I removed any potentially offensive tweets and organized results generated by the API to my liking.
A Glimpse into Use Potential & Preliminary Findings
While this dataset is in its early stages, it currently communicates, at least in part, with concerns I have outlined in my introduction: Does tryborgianism commonly pervade contemporary images of the cyborg? How does the academic and para-academic cyborg communicate with that of broader, popular culture? What tensions exist within common uses of “cyborg”?
Cross-media/Cross-cultural Analysis
With respect to literary and culture magazines alone, the dataset is best read as a guide to common topics, using the keywords and collocates fields. Computational analyses may be conducted within and across each literary publication by splitting either or both column(s) and counting the number of occurrences of keywords and collocates per document using the “COUNTIF” function, but this isn’t necessary to glean relevant information. Twitter user data and tweet information, on the other hand, can be analyzed and visualized computationality using the “pivot table” tool to gague count and frequency of values within variables.
To answer the above inquiries, however, it seems worthwhile to explore common topics and keywords across media rather than within each individual domain. (See figures 5-8.)
figure 5: screenshot of pie chart displaying first group of named entites by percentage
figure 6: screenshot of pie chart displaying second group of named entites by percentage
figure 7: screenshot of pie chart displaying third group of named entites by percentage
figure 8: screenshot of pie chart displaying user location by percentage
The above figures show that auto ads (if we read further into Tweet text), popular entertainment, sports, sci-fi films, and cryptocurrency are among the most popular topics within tweets that use the term “cyborg” (located and diagnosed by the Twitter API). Relatedly, then, a plurality of Twitter users who identify as “cyborg” tag their location as “Metaverse,” and “NFT” appears within 11.7 percent of users’ proffered information within one or more fields. Disability, medicine, and healthcare are not to be found among top named tweet entities, and only five of the three hundred (or 1.7%) of the top searchable users who identify as “cyborg” also identify in their user description or name as disabled. (Though, at least seven users who do not identify as “disabled” do signal chronic illness in some way.) Of almost seventy New Yorker articles that mention “cyborg,” 13% mention “tech” within their respective keywords, 18% mention “movie,” and some iteration of “super,” whether it be heroes or computers, is listed within title, keywords, or collocates in several documents. In contrast, “medicine” is only tagged as a keyword in one article and “disability”/”disabled” in zero.
While I will address the limitations of this data momentarily, these trends point to the likely veracity of Weise’s–and later Wong and Shew’s–claims that the popular understanding of the cyborg is perhaps too undergirded by metaphor and desire for optimization, acceleration(/ism), and/or accessory rather than the wellbeing and quality of life of those who are, as Weise would describe, actually cyborgs.
Limitations & Points of Further Inquiry
The more technical and logistical limitations of this dataset exist in the fact that my sample is limited. To start, because of my limited level of access to the Twitter API, this dataset only includes several days worth of tweets, rather than months or years worth of data, such as in literary and culture magazines. This creates discrepancies in cross-media comparison. Where literary and culture magazines are concerned, then, this dataset could clearly be expanded to include more publications based both within and outside of the United States. Further, user information is limited to the top three-hundred searchable users not because I was restricted by the API from collecting more, but because of time constraints. Creating a uniform number of tweets, users, and articles across a matching time frame might have aided in a more facilitating exacting comparisons than the ones provided above. That said, this dataset is not concrete, and I plan to expand and improve it with these limitations in mind.
Thematically and conceptually, then, this dataset falls short of making visible biotechnologically supported individuals who have reclaimed cyborg as an identity in order to change public discourse around body and technology. While there are a handful of tweets, articles, and user descriptions by disabled, diabetic, neurodivergent, and otherwise technologically supported creators to be located within the dataset, they are not made particularly visible due a the sheer discrepancy in numbers. Therefore, this dataset risks making hypervisible problematic projections of and desires for biotechnology. What it has aimed to do, and hopefully does, nevertheless, is help to point to common trends in order to locate where forms of resistance–if only visibility–might be needed (while acknowledging that visibility is both a privilege and risk).
Bibliography
Kafer, Alison. “Cyborg.” Encyclopedia of American Disability History, edited by Susan Burch. Facts On File, 2009.
Kafer, Alison. Feminist, Queer, Crip. Indiana University Press, 2013, pp. 103-128. ProQuest Ebook Central, http://ebookcentral.proquest.com/lib/miami/detail.action?docID=1189107.
Sandoval, Chéla. “Re-Entering Cyberspace: Sciences of Resistance.” Dispositio, vol. 19, no. 46, 1994, pp. 75–93. JSTOR, http://www.jstor.org/stable/41491506.
Siebers, Tobin. “Disability in Theory: From Social Constructionism to the New Realism of the Body.” American Literary History, vol. 13, no. 4, 2001, pp. 737–54. JSTOR, http://www.jstor.org/stable/3054594.
Weise, Jillian. “Common Cyborg.” Disability Visibility: First-person Stories from the Twenty-First Century, edited by Alice Wong, 2020, pp. 63-74.
Weise, Jillian. “The Dawn of the ‘Tryborg.’” The New York Times, 30 Nov. 2016, https://www.nytimes.com/2016/11/30/opinion/the-dawn-of-the-tryborg.html.
Wong, Alice. “Ep 66: Cyborgs.” Disability Visibility Project, 19 Dec. 2019, https://disabilityvisibilityproject.com/2019/12/18/ep-66-cyborgs/.