• 1  Daniel Fischlin (ed.), OuterSpeares: Shakespeare, Intermedia, and the Limits of Adaptation, Toront (...)

“Shakespeare…is no longer just textual or theatrical in any conventional sense of the terms.”
(Daniel Fischlin)1

• 2  Marjorie Perloff, Unoriginal Genius. Poetry by Other Means in the New Century, Chicago, University (...)
• 3  Jan Kott and Boleslaw Taborski, Shakespeare Our Contemporary, Garden City (N.Y.), Doubleday, 1966; (...)
• 4  “dat het dramatische model niet langer de evidente uitdrukking is van het postmoderne wereldbeeld. (...)
• 5  John Drakakis, “Shakespeare in Quotations,” in Susan Bassnett (ed.), Studying British Cultures. An (...)
• 6  “Alles voor Heiner Müller hoort in het museum.”
  
• 7  “wie of wat is er nu precies dood als Shakespeare dood is [...]?” (Erwin Jans, “Drama na het drama (...)

1For many, Shakespeare is still an unchanging“contemporary” of ours, as Jan Kott and Boleslaw Taborski phrased it back in the 1960s.3 From this perspective Shakespeare’s work tends to be treated as a bible—hence his status of bard, a term kin to prophet. For Belgian playwright Paul Pourveur Shakespeare is dead, get over it! (2003) and no longer has something to tell us in our 21^st century world. With this title for a play Pourveur declares Shakespeare a dead playwright, but also declares dramatic theatre dead as a “model [that] [...] no longer express[es] [...] the postmodern worldview.”4 With more nuance, John Drakakis suggests that not the work of Shakespeare itself, but its use can signal the postmodern: “With the endless circulation of Shakespeare quotations now sharing the same status as other forms of representation, we have entered the realm of the postmodern.”5 Pourveur’s provocative anti-tradition statement “Everything preceding Heiner Müller belongs in the museum” nevertheless confirms an established idea that the postdramatic canon starts with Heiner Müller, author of the paradigm shifting theatre text Die Hamletmaschine (1977).6 As dramaturge Erwin Jans pertinently put it: “who or what is actually dead if Shakespeare is dead [...]?”7

• 8  Hans-Thies Lehmann, “From Logos to Landscape: Text in Contemporary Dramaturgy,” Performance Resear (...)
• 9  Jane Gallop distinguishes “literal,” “theoretical” and “other authorial deaths” of authors. Jane G (...)
• 10  From personal interviews by the author with Annie Dorsen: a Skype talk on June 22nd and a live tal (...)

2Long after Peter Szondi’s announcement of the “crisis of drama” in Theorie des modernen Drama’s (1956), and in the trail of Barthes’s declaration of “The Death of the Author” (1967) followed, in the theatre, several deaths: The Death of the Playwright? (Page 1992), The Death of Character (Fuchs 1996) and even, implicitly, “The Death of the Text”—at least according to a not unpopular interpretation of Hans-Thies Lehmann’s concept of Postdramatic Theatre (2006) as visual, post-text theatre. This reduction makes abstraction of the specificity of the dramatic model: the dominance of its logos over the other theatrical ‘languages’ of performance (sound, movement, embodiment...), imposing a teleological narrative of a moral conflict to be overcome.8 Die Hamletmaschine shatters this logos attributed to human reason, along with its language. In the postdramatic paradigm, the “literal” death of Shakespeare seems far less interesting to stage than his metaphoric death.9 By the latter, avant-garde director-auteurs position themselves vis-à-vis theatre tradition, its canon and/or the auctorial model of the genius. “Maybe I wanted to kill Hamlet,” confessed Annie Dorsen, educated in directing at what she experienced as the fairly conventional Yale School of Drama, and director of the data-processing Hamlet adaptation A Piece of Work (2013), that I will discuss.10

• 11  Fischlin qtd. in Daniel Fischlin (ed.), OuterSpeares: Shakespeare, Intermedia, and the Limits of A (...)

3However, with Mark Fortier I tend to argue that programmatically killing Shakespeare on stage also means putting him in the spotlight and creating what Daniel Fischlin calls “the Shakespeare effect,” a “function of Shakespeare’s cultural pervasiveness, in which echoes, resonances, and direct integration of that effect are in evidence in a given play.”11 So, if Shakespeare is revived by killing him, in what form, then, may the killed author rise from the dead?

• 12  Andrew Bennett, The Author, London, Routledge, 2005, p. 95, 100-102; Sujata Iyengar, “Upcycling Sh (...)
• 13  Andy Lavender, Hamlet in Pieces: Shakespeare Reworked: Peter Brook, Robert Lepage, Robert Wilson, (...)
• 14  Thomas Leitch, “Twelve Fallacies in Contemporary Adaptation Theory,” Criticism, vol. 45, no. 2, Sp (...)
• 15  Andrew Bennett, The Author, London, Routledge, 2005, p. 29.

4With the “author as data processor” I label a specific contemporary auctorial model by referring to the 20^th century technology of digital information processing, not to state that this intermedial approach of writing as the adaptation and appropriation of text data is new, but to historicize it. We seem to witness—(since about half a century?) the reemergence of a tendency towards less author-centered conceptions of writing reminding of, but of course not identical to, conceptions of the past. In Shakespeare’s time, the concept of individual authorship was only on the verge of being established: our “bard” was, in practice, an actor furnishing the theatre with text, in collaboration with other actors, appropriating circulating texts, plots and themes.12 Even what is deemed the play of all plays “is certainly not ‘original,’” as Shakespeare drew on an extant Ur-Hamlet.13 Nevertheless, Shakespeare’s adaptation of the Ur-Hamlet is considered an “original” text. Such a view illustrates two of the “Twelve Fallacies in Contemporary Adaptation Theory” pointed out by Thomas Leitch: “Source texts are more original than adaptations” and “Adaptations are intertexts, their precursor texts simply texts.”14 Ironically, by the distributive power of print culture and through the retrospective lens of Romantic and individual authorship, Shakespeare the adapter and in a way data processor avant-la-lettre “became, after his death, the very model of the universal genius.”15

• 16  Andrew Epstein, “Found Poetry, ‘Uncreative Writing,’ and the Art of Appropriation,” in Joe Bray, A (...)
• 17  Goldsmith qtd. in Marjorie Perloff, “Moving Information: On Kenneth Goldsmith’s ‘The Weather,’” Op (...)
• 18  For the idea of “strong” and “weak” auctorial models, see Marysa Demoor, Ingo Berensmeyer, and Ger (...)

5In Unoriginal Genius (2010) Marjorie Perloff states that writing in the 21st century has become less a question of inventing than of repurposing text. Some go even as far as to claim that in the 2010s “the mash-up rule[s]” and “becomes a trope for intermediality in the digital age.”16 Or in the highly citable words of Perloff’s follower Kenneth Goldsmith: “I used to be an artist, then I became a poet; then a writer. Now when asked, I simply refer to myself as a word processor.”17 The next model we may expect him to proclaim is that of the coder (from the artist to the nerd...). This does not keep Goldsmith from predicting, like a bard, that authorship today tends to shift towards what I would more generally call a “data processor” on a continuum ranging from “strong” auctorial models like the romantic conception of the genius to “weak” models like those of oral traditions, collaborative (digital) writing, appropriation and so-called Uncreative Writing.18

• 19  Bertrand Gervais, “Is There a Text on This Screen? Reading in an Era of Hypertextuality,” in Ray S (...)
• 20  Marjorie Perloff, “Moving Information: On Kenneth Goldsmith’s ‘The Weather.’” Open Letter. A Canad (...)

6The author as data processor foregrounds the act and the material of writing as the processing of text data, facilitated in western “linked computer culture” by data’s hyper-availability, potentially cut-and-paste-able, and ‘mashable’ in no time.19 Perloff calls this literary practice “moving information,” paraphrased by Goldsmith as “the act of pushing language around as well as the act of being emotionally moved by that process.”20

• 21  For the concept of “transcoding,” see: Lev Manovich, The Language of New Media, Cambridge (Mass.), (...)

12Technically, “moving information” is what makes new (= digital) media “new”. Digital computing has in certain ways increased the mobility of information; it can be “transcoded” from one digital format, “platform” or medium to another and vice versa, seemingly as easy as a magic trick:21

• 23  Bernard Stiegler, “Memory,” in W.J.T. Mitchell and Mark B.N. Hansen (eds.), Critical Terms for Med (...)
• 24  Lev Manovich, The Language of New Media, Cambridge (Mass.), MIT Press, 2000, p. 28; Bernard Stiegl (...)
• 25  Lev Manovich, The Language of New Media, Cambridge (Mass.), MIT Press, 2000, p. 20, 25-26.
• 26  Marjorie Perloff, Unoriginal Genius. Poetry by Other Means in the New Century, Chicago, University (...)
• 27  “un espace à dimensions multiples, où se marient et se contestent des écritures variées, dont aucu (...)
• 28  Siobhan O’Flynn qtd. in Sujata Iyengar, “Upcycling Shakespeare: Crafting Cultural Capital,” in Dan (...)

13Transcoding is only possible because “data” are the result of a process of breaking information up into small and uniform units, an ongoing cultural and historical process which Bernard Stiegler calls “discretization.”23 Discretizing or sampling analog information means “turn[ing] continuous data into discrete data [...],” “the process whereby the currents and continuities shaping our lives become discrete elements,” for instance “[w]riting, as the breaking into discrete elements of the flux of speech [...].”24 Digital information is per definition discretized, it consists of uniform units in binary code, or simply, “data.” For the first time, it seems, a writing tool simultaneously functions as a provider of information that invites to manipulate that same information: the computer converges the functions of data storage, processing and production.25 The Internet as a huge cultural digital database of information is technologically integrated into the writing process.26 Perhaps more than “the text,” as Barthes has it, the computer, or better, the author as data processor is, “a multi-dimensional space in which a variety of writings, none of them original, blend and clash.”27 It is the digital nature of data, in the sense of discrete standardized units in binary code, that enables authors and auteurs as data processors to easily move information. Linda Hutcheon and Siobhan O’Flynn define the practice of adaptation in terms of “a transcoding process.”28 Transcoding, here, can be a metaphorical act as well as a technological reality.

• 29  “FB Theater,” Kaaitheater, [Online] https://www.kaaitheater.be/en/agenda/fb-theater [accessed 7 Ma (...)
• 30  Eric Paulos, “The Human Body as Multimedia,” [Online] http://besser.tsoa.nyu.edu/impact/f96/Review (...)

14In Facebook Theatre (2015) Ivo Dimchev has an offstage dramaturge rewrite the audience’s real-time Facebook comments about the performance into dialogues, which then serve as the input for the performers via earpieces—whose output is again commented on, and so on, in a text-to-text transcoding loop.29 In cyber artist Stelarc’s groundbreaking 1990s telematic performance Ping Body: An Internet Actuated and Uploaded Performance (1996) abstract “movement” or activity on the Internet (e.g., textual input by users) is transcoded by means of a muscle stimulator into Stelarc’s own physical movement onstage.30 What these kinds of performances explore is how far technology’s capacities can push the limits of human capacities of data processing (of language data and ideas for the performance; of physical stimuli).

• 31  Mark Fortier, “Beyond Adaptation,” in Daniel Fischlin (ed.), OuterSpeares: Shakespeare, Intermedia (...)
• 32  Marjorie Perloff, Unoriginal Genius. Poetry by Other Means in the New Century, Chicago, University (...)
• 33  John Drakakis, “Shakespeare in Quotations,” in Susan Bassnett (eds.), Studying British Cultures. A (...)
• 34  Sujata Iyengar, “Upcycling Shakespeare: Crafting Cultural Capital,” in Daniel Fischlin (ed.), Oute (...)
• 35   John Drakakis, “Shakespeare in Quotations,” in Susan Bassnett (ed.), Studying British Cultures. A (...)

15Having databases at hand is not a condition for adaptation and appropriation practices. Historically, writing seems always to have been a form of (intertextual) rewriting—“there is nothing beyond adaptation” and in this sense even the most provocative adaptation is “not innovation but renovation.”31 Yet, a poetics of “citationality,” as Perloff recycles Antoine Compagnon’s idea of “récriture,” indeed seems “the logical form of “writing” in the age of literally mobile or transferable text.”32 What digital culture radically facilitates is the recombination of data due to their connecting potential. Discretized, data are open to potential new connections (with other data and/or other contexts). This echoes dramaturgical practices that connect material from various sources into new semantic and aesthetic contexts. But whether “reproducing, assembling, and hijacking pre-existing” discourses and, I would add, canons, is used for contestation or confirmation, citing’s “erasure of context from individual quotations, or from those works as a whole that we designate “literary,” is an irreducibly ideological operation,” the more so in the case of canonical œuvres like Shakespeare’s.33 Legitimizing citation practices tend to fetishize canonical works, either by citing them literally, or by referring to them in crafted “Shakesthings,” as Sujata Iyengar calls goods of relatively low material value that have been “upcycled” by means of Shakespeare’s “cultural capital.”34 Conversely, citing practices that resist the temptation of coherence and essence stand for “the idea of Shakespearean texts as sites of contestation, as opposed to being repositories of cultural wisdom.”35

• 38  Here, Shepherd and Wallis paraphrase W.B. Worthen. Shepherd, Simon, and Mick Wallis. Drama, Theatr (...)
• 39  Jerome Fletcher, “Introduction,” Performance Research 18, no. 5 (October 1, 2013), p. 1. [Online] (...)

20This is even more the case for stagings of digital Shakespeare texts, as the relation between Shakespearean text and performance is mostly asymmetrically conceived, considering “the written text as the repository of authorial truth.”38 Fletcher, then, proposes that “[r]ather than seeing [writing] as the end-point, the outcome of the digital device or apparatus, we can consider the question of how writing performs throughout the entire apparatus/device,” by which he seems to mean dispositif.39 Would it, however, not be as relevant to inverse Fletcher’s question: how does the digital apparatus perform through writing for and in the performance space?

• 40  Pascale Aebischer and Nigel Wheale qtd. in Daniel Fischlin, OuterSpeares, Toronto, University of T (...)

21Apart from the observations by Cornish, Fletcher, Lehmann and others, that the text has been symbolically displaced in postdramatic theatre, and that the Shakespeare text is being “increasingly decentered” in adaptations across genres and media, what does the more literal kinesis of such texts look like?40 Text data are literally moved around in the form of cards or printed playtexts, text appears and disappears, is scrolled, is projected on all kinds of ‘screens’, including bodies, deforming words. The kinesis of text data applies to analog and digital texts alike, in low-tech as well as high-tech performances, staging the real-time generation of texts in natural languages or in code. But also in a dramaturgical sense language is being moved around in the theatrical time-space. Either machine language (code) or natural language is written word for word and edited in real time. Alternatively, text is hypertextually composed by combining already available, text blocks or “lexia.”

• 41  Aravind Enrique Adyanthaya, “Towards a Poor Techno-Theatre,” Performance Research 18, no. 5, Octob (...)
• 42  I discuss Edit Kaldor’s Web of Trust in more detail, as well as Annie Dorsen’s Hello Hi There in: (...)
• 43  Emma Cocker, “Live Notation: Reflections on a Kairotic Practice,” Performance Research 18, no. 5, (...)
• 44  Emma Cocker, “Live Notation: Reflections on a Kairotic Practice,” Performance Research 18, no. 5, (...)
• 45   Fletcher refers to N. Katherine Hayles’s idea of the digital text as “a performance.” Yet, such l (...)

23In his self-declared “Poor Techno-theatre” Porto Rican Aravind Enrique Adyanthaya stages what he calls “act-writing,” acting in the form of writing.41 Edit Kaldor, in her social media performance Web of Trust (2016), invites logged-in participants present in the theatre or online to write and edit the performance’s and social network site’s working principles.42 Live Coding practices in general approach “programming as performance art, performance art notation as code” and are interested in performance as the “revelation of digital code.”43 In all of these and similar cases, text is projected on a screen while being written in the theatrical space, so the text’s and technology’s performativities converge. The “cursor” becomes “the location of action and retraction, of cut and paste, of deletion and erasure, insertion and manipulation.”44 The text’s live composition, here, is a performance in itself.45

• 46  Annie Dorsen, A Piece of Work, Brooklyn (N.Y.), Ugly Duckling Presse, “Emergency Playscripts 5,” 2 (...)
• 47  David Berlinski, The Advent of the Algorithm: The 300-Year Journey from an Idea to the Computer, S (...)
• 48  I discuss What You Will. Something by Shakespeare (2007) [Elk wat wils. Iets van Shakespeare] as a (...)

24Distinguishing her work from what she considers “multimedia performance,” with her digital “algorithmic theater” Annie Dorsen aims to “investigat[e] [...] the dramaturgy of algorithms themselves, which is to say their way of ordering the world, the particular kinds of meanings they make, the kinds of narrative structures they imply.”46 Algorithmic theatre, however, need not be “digital” in the common meaning of “related to computers.” In general terms, “[a]n algorithm is an effective procedure, a way of getting something done in a finite number of discrete steps. [...] an algorithm, speaking loosely, is a set of rules, a recipe, a prescription for action, a guide [...]”47 As a dramaturge in collaboration with director Lucas Vandervost I worked on what was, in retrospect, an analogue algorithmic performance using an artisanal Shakespeare-database.48 What You Will. Something by Shakespeare (2007) [Elk wat wils. Iets van Shakespeare] of the Belgian theatre company De Tijd was a far from “techy” looking, dedramatized collage of words and phrases from the whole dramatic œuvre of the Bard, metaphorically “passed through the paper shredder” (though we started by using scissors and ended up using a computer).

• 49  Orion Maxted, “Cybernetics vs. Decartes.” Etcetera, [Online] https://e-tcetera.be/cybernetics-vs-d (...)
• 50  I briefly discuss the idea of the “delegation of text production” in a review of Jisun Kim’s Deep (...)
• 51  Annie Dorsen, A Piece of Work, Brooklyn (N.Y.), Ugly Duckling Presse, “Emergency Playscripts 5,” 2 (...)
• 52   I have seen an unpublished video recording of the performance. Annie Dorsen, A Piece of Work - wi (...)
• 53  Annie Dorsen, A Piece of Work, Brooklyn (N.Y.), Ugly Duckling Presse, “Emergency Playscripts 5,” 2 (...)

25Building on his previous lo-fi algorithmic performances, Orion Maxted’s recent work [THE BRAIN] approaches the audience collectively as “a giant brain that is capable to thinking, solving problems and making a performance.”49 Algorithms enable human authors to delegate authorship to human or software-run performers by providing instructions.50 In Dorsen’s earlier Hello Hi There (2010), the algorithms designed for the two chatbot-performers made them produce dialogue. Dorsen has a point, then, when she considers the “writing of algorithms [...] creative work, as there are always multiple ways of solving any particular problem,” but also the writing by the algorithms.51 The ones developed for her Hamlet-adaptation A Piece of Work (2013) “were designed to act as artistic partners: co-writers, co-directors, co-designers, and performers” on stage to “autonomously create and perform an adaptation of Shakespeare’s play according to algorithmic principles.”52 The algorithms “are simple—they skip, sort, replace, and sequence” words of Hamlet.53

• 54  Constant, “Public Domain Day: The Death of the Authors,” Constant vzw, [Online] http://publicdomai (...)
• 55  Annie Dorsen, A Piece of Work, Brooklyn (N.Y.), Ugly Duckling Presse, “Emergency Playscripts 5,” 2 (...)
• 56  N. Katherine Hayles, How We Think: Digital Media and Contemporary Technogenesis, Chicago/London, T (...)
• 57  Thomas Leitch, “Twelve Fallacies in Contemporary Adaptation Theory,” Criticism, vol. 45, no. 2, Sp (...)

26Dorsen could have chosen less mythical canonical data to work with, like in the yearly “Death of the Authors” project of media art association Constant, based in Brussels, the same city where the Kaaitheater hosted several of Annie Dorsen’s shows. Treating literary texts of which the rights have expired as databases, Constant’s algorithms produced “generative novels” and a “bot opera,” featuring chatbots behaving differently with data and interacting with printers, projected images and the audience.54 But treating Hamlet as a text database is a choice implying more than just a desire for textual experimentation and technological insight. Considering A Piece of Work “a literal ‘hamletmachine’ [...] [Dorsen] chose Hamlet because of its status as the canonical humanist theater text. Harold Bloom, an emblematic spokesperson for this view, names Hamlet the play that invents what we humans understand ourselves to be.” For Dorsen, the claim about Hamlet’s “consciousness of his own consciousness,” as Bloom puts it, “made the play the obvious choice to explore how algorithms might change the way theater constructs representations of the human.”55 Accordingly, she reduced the play of all plays to data “that the computer can understand,” to say it with Katherine Hayles.56 The emphasis Dorsen puts on Hamlet would almost make one forget another fallacy from Leitch’s list: “Adaptations are adapting exactly one text apiece.”57 For sure, Dorsen is adapting Shakespeare’s Hamlet, but also Heiner Müller’s Hamletmachine and, in quantity perhaps as important as these literary precursors are the computer code and software models that Dorsen and her team adapt in order to have a machine generate versions of Hamlet.

• 58  Ioana B. Jucan, “Sys.Begin to Sys.Exit Software Performs A Piece of Work,” TDR-the Drama Review-th (...)

28As Annie Dorsen has already explained and reflected on mainly the software and technical aspects of A Piece of Work in her introduction to the book publication of the text, and Iona Jucan’s article looks deeper into the logic of algorithms, I will focus more on what is relevant for the idea of the author or performer as a data processor and on the related aspects of discretization, citationality and kinesis.58 The way the text is produced, the way it appears in the performance and the way it functions will be central to the discussion.

• 59  Dorsen, Annie, “A Piece of Work,” [Online] http://www.anniedorsen.com/showproject.php?id=14 [acces (...)
• 60  Sarah Bay-Cheng, “UNSEEN Performance Criticism and Digital Recordings,” Theater 46, no. 2 (2016), (...)
• 61  Sarah Bay-Cheng, “UNSEEN,” Theater 46, no. 2 (2016): p. 80, [Online] https://doi.org/10.1215/01610 (...)
• 62  Sarah Bay-Cheng, “UNSEEN,” Theater 46, no. 2 (2016): p. 83, [Online] https://doi.org/10.1215/01610 (...)
• 63  Steven Shaviro qtd. in Sarah Bay-Cheng, “UNSEEN,” Theater 46, no. 2 (2016): p. 81, [Online] https: (...)
• 64  Sarah Bay-Cheng, “UNSEEN,” Theater 46, no. 2 (2016): p. 78, [Online] https://doi.org/10.1215/01610 (...)

29Published interviews, private talks with Annie Dorsen as well as her Introduction to A Piece of Work in Emergency Playscript’s publication of the play text provided me with information concerning the production of the text. For my analysis of the text’s effect and function in the performance, I rely on a video recording of the performance at the Brooklyn Academy of Music in December 2013, where it ended its tour, after having travelled to Seattle, Oslo, Bergen, Vienna, Rotterdam and Paris that same year.59 Even though I have not seen the performance live, I have seen it, partially—after all, “what [does] it [mean] to ‘see’ work in the context of telematic performance and digital media”?60 Sarah Bay-Cheng raises the question in her article “UNSEEN Performance Criticism and Digital Recordings,” in which she convincingly argues that not only the experience of the performance’s document (e.g., a video recording) is incomplete, but also “the experience of copresence is not itself a complete version” of a performance.61 What Bay-Cheng proposes, then, is not to take video recordings of performances at face value, but “to look with a perspective learned from reality television, where the viewer questions both the accuracy of the media representation and the reality on which the documentation is based, while simultaneously attending to the affective relations developed by the screen versions.”62 In that respect, I am lucky that A Piece of Work was recorded with a fixed camera, mainly from a central point of view, showing about five rows of audience members sitting below the camera. When other camera angles are adopted (e.g., from aside, when the “ghost” is speaking), it is still able to locate the place of the camera, even in the case of the few close ups on the human performer (on the evening of the recording this was Scott Shepherd, not Joan MacIntosh). Technically, there is quasi no difference whether the synthetic voices, the text and the algorithms running the lighting and sound effects are generated by a computer in the theatre or by my laptop—both can be considered “machines for generating affect,” in the words of Steven Shaviro.63 This is especially the case here, because the textual and sensory data generated by Dorsen’s algorithms could be exactly reproduced (if Dorsen wanted to), contrary to a performance that mainly relies on biological or ecological agents (humans, animals, the weather). However, I am well aware that the presence-effect even of non-human agents like the algorithm-driven lighting, sound-effects and projected text cannot be denied.64 I therefore also count on my twelve years of experience as a dramaturge, observing the differences between rehearsals, try-outs, premieres and ‘days after’ as well as between performances and their recording(‘s camera angles and close-ups), photographs, reviews and audience accounts. All these factors help me in trying to imagine what the “live” experience of the performance in the theatre, being copresent with the audience, of performing algorithms in combination with one human actor, could have been like, without claiming that I know what they were like as if I had been there.

• 65  “WordHoard - The Texts: Provenance, Copyrights and Licenses,” [Online]  http://wordhoard.northwest (...)

30Dorsen’s endeavor would probably have been impossible as a solitary author, due to a lack of time and financial means if she had not been able to rely on an already discretized digital version of Hamlet. This was made available by the academic WordHoard project, which turns canonical texts into databases of tagged linguistic “data.”65 Since tagging is distinguishing items from other items (e.g., words, grammatical constructions) it also means isolating and hence discretizing them.

• 66  Annie Dorsen, A Piece of Work, Brooklyn (N.Y.), Ugly Duckling Presse, “Emergency Playscripts 5,” 2 (...)
• 67  Annie Dorsen, A Piece of Work, Brooklyn (N.Y.), Ugly Duckling Presse, “Emergency Playscripts 5,” 2 (...)

31Dorsen and programmer (and media theorist) Mark Hansen used WordHoard’s digital and tagged Hamlet-data but developed their own algorithms to generate text from this database. They relied on the algorithmic principle “Markov chain,” a model to process data that generates combinations and variants whose occurrence is probable according to the characteristics of the data (e.g., recurrent grammatical patterns). Typical of Markov chains is that the process of generating text is “memoryless”: when the algorithm produces sentences by recombining data, this is disconnected from previously produced text data.66 Hence, the recombined text data’s eventual coherence is not due to intelligence as understanding but as the mimicking of “the statistical order of the words in the original text” (and the audience’s expectations).67

• 68  Annie Dorsen, A Piece of Work, Brooklyn (N.Y.), Ugly Duckling Presse, “Emergency Playscripts 5,” 2 (...)
• 69   Annie Dorsen, A Piece of Work, Brooklyn (N.Y.), Ugly Duckling Presse, “Emergency Playscripts 5,” (...)
• 70   Annie Dorsen, A Piece of Work, Brooklyn (N.Y.), Ugly Duckling Presse, “Emergency Playscripts 5,” (...)

32To increase coherence, Dorsen and her team added “a small set of simple semantic and theatrical rules to impose on the generated text,” for instance “a character may only speak after entering, or should not say his/her own name”.68 In doing so, she reintroduced—though in a minimalist way—the kind of realist theatrical conventions that she intended to question in the first place. An additional constraint imposed by Dorsen was to limit a given task to only specific, quantifiable parts of the play (e.g., a famous scene, a specific number of lines etc.) Finally, Dorsen and her team built in what can be considered as yet another interpretative filter by tagging each word of the play with what they called “emotion scores”: “We broke our data down in several different ways to allow for a variety of uses.”69 These tagged data served as transcodable input for the sound, music, video and lighting systems of Dorsen’s hamletmachine or “autonomous theatrical intelligence”.70

• 71  Espen Aarseth, Cybertext: Perspectives on Ergodic Literature, Baltimore (Md.), Johns Hopkins unive (...)
• 72   Espen Aarseth, Cybertext: Perspectives on Ergodic Literature, Baltimore (Md.), Johns Hopkins univ (...)
• 73   Annie Dorsen, A Piece of Work, Brooklyn (N.Y.), Ugly Duckling Presse, “Emergency Playscripts 5,” (...)

33Espen Aarseth, in Cybertext (1997), considers three basic relations in “human-machine collaboration” when it comes to the production of (literary) text: “(1) preprocessing, in which the machine is programmed, configured, and loaded by the human; (2) coprocessing, in which the machine and the human produce text in tandem; and (3) postprocessing, in which the human selects some of the machine’s effusions and excludes others.”71 Hence, “The key question in cyborg aesthetics is therefore, Who or what controls the text?”72 Dorsen acknowledges the coauthorship of programmers Dylan Fried and Mark Hansen (and performer Scott Shepherd, who did additional programming in a later phase) as well as that of the algorithms—without erasing herself from the picture: “[t]o the extent that this adaptation of Hamlet was ‘written’ by me, it was these so-called configurations files that I wrote” (i.e., part of the preprocessing).73

• 74  Annie Dorsen, A Piece of Work, Brooklyn (N.Y.), Ugly Duckling Presse, “Emergency Playscripts 5,” 2 (...)
• 75  Johan Callens, “The Wooster Group’s ‘Hamlet’, According to the True, Original Copies,” Theatre Jou (...)
• 76  Annie Dorsen, “Talk about ‘A Piece of Work’. Group Self-Interview,” TDR-the Drama Review 59, no. 4 (...)

34In the end, Dorsen and her programming team developed five different algorithms to generate ‘new’ text from the Shakespeare-data in real time, mimicking the five-act structure of the play. The subtitles Dorsen gave to the acts loosely indicate what each specific algorithm does.74 (I will only discuss the algorithm’s main tasks here.) In “Act 1: SCENE” the algorithm mainly skips: it reproduces the first 4% of every scene, scene by scene, with the effect, more or less, of an abridged version. “Act 2: LINE” relies on keyword detection and sorting: the algorithm selects lines from the entire play that start with a specific keyword (“what,” “I” or “O”), and reproduces the lines according to their length, from the longest to the shortest. This has a poetic effect of rhyme, repetition and, as the lines get shorter, disintegration. In “Act 3: GRAMMAR” the available data is limited to the soliloquies, in which the algorithm first identifies words according to their grammatical categories (e.g., nouns) and grammatical structures. Then, it generates new soliloquies by replacing original words of the play by other words of the play pertaining to the same grammatical category, and according to the common grammatical structures in Hamlet. This act is played by a human performer, either Scott Shepherd or Joan MacIntosh, whose relations to the Wooster Group inevitably evoke Liz LeCompte’s contrary citational version of the same classic text, or better, of the “theatrofilm” filmic record of the classic’s staging.75 (Shepherd and MacIntosh’s use of earpieces in A Piece of Work also recalls LeCompte’s systematic transmission of data to the performers through earpieces during performances.) For those who know the soliloquies, the replacement of words by the algorithm draws attention to the potential and factual variety in (the Folio and Quarto versions of) Shakespeare’s text. Acts 4 and 5, then, use Markov chaining: the algorithm produces variants on the original by mimicking it, i.e., reproducing statistically probable combinations of letters, words, grammatical structures etc. Act 4: WORD” “generate[s] new scenes by re-sequencing words” and “Act 5: LETTER” “generate[s] a new final scene [...] by resequencing letters” from the original Act 5, Scene 2 of Hamlet.76

• 77  N. Katherine Hayles, How We Think: Digital Media and Contemporary Technogenesis, Chicago/London, T (...)
• 78  Lev Manovich, The Language of New Media, Cambridge (Mass.), MIT Press, 2000, p. 43, 230-231.
• 79  Lev Manovich, The Language of New Media, Cambridge (Mass.), MIT Press, 2000, p. 233.

35The five subtitles of the acts indicate an increasing “atomization of language,” dramaturgically corresponding to the character’s falling into pieces and coinciding with the audience’s experience of speech either falling apart or becoming hardly recognizable.77 Furthermore, the gradual movement from Act 2’s keyword repetition to the replacement of words by words of the same category in Act 4 and the recombination of letters in Act 5, corresponds to a shift identified by Lev Manovich as a manifestation of “computer culture”: the relative retreating of syntagmatic (storytelling) modes to the background of culture, and of paradigmatic modes of composing or meaning making to the foreground.78 Paradigms underpin the logic of “database forms” (or older cultural forms like the encyclopedia) and facilitate citationality.79 Especially A Piece of Work’s Act 3: GRAMMAR shows paradigmatic writing at work, orally mediated by a human performer:

• 81   [Online] https://doi.org/10.1162/DRAM_a_00502 [accessed 29 March 2019]. (...)
• 82   Ioana B. Jucan, “Sys.Begin to Sys.Exit Software Performs A Piece of Work,” TDR-the Drama Review-t (...)

37Why does the citationality of Hamlet matter in this case of human-machine coauthorship? As Iona Jucan remarks, algorithms are normally designed to make big data “intelligible” to humans, whereas the algorithms of Dorsen and her team make a small database of understandable Shakespeare relatively “unintelligible.”81 A Piece of Work “disarticulate[s],” the text of a highly articulate and canonical character, for a long time considered the essence of humanism.82 In the meantime, there is more or less of a consensus in Shakespeare and literary scholarship about the impossibility of attributing any final meanings to the text, revealed as the product of multiple versions and borrowings by Shakespeare. Yet, outside this circle of specialists, the idea still seems to dominate that Hamlet is a human character and condition rather than a text.

• 83  Thomas Cartelli, “Essentializing Shakespeare in the Shakespeare Aftermath: Dmitry Krymov’s Midsumm (...)
• 84  N. Katherine Hayles, Writing Machines, Cambridge (Mass.), MIT press, 2002, p. 25.
• 85  Hans-Thies Lehmann, Postdramatic Theatre [1999], transl. Karen Jürs-Munby, London, Routledge, 2006 (...)

38Stagings like Annie Dorsen’s refocus attention towards canonical works as texts in the first place, to such an extent that Thomas Cartelli speaks of A Piece of Work in terms of “the production’s textual monopoly.”83 Dorsen’s performance turns Hamlet into what Katherine N. Hayles calls a “technotext,” not because it looks digital, but because it puts the materiality of language in the spotlights, “interrogates the inscription technology that produces it” and “mobilizes reflexive loops between its imaginative world and the material apparatus embodying that creation as a physical presence.”84 This echoes the self-reflexive tendency of postdramatic theatre to foreground its medium specificity and processual character.85

• 86  Ronald Geerts, “Tekst als object. Over de herwonnen autonomie van de dramatekst,” in Claire Swyzen (...)
• 87  Annie Dorsen, “Talk about ‘A Piece of Work’. Group Self-Interview,” TDR-the Drama Review 59, no. 4 (...)

39Projecting text onto a screen is the most common way in which postdramatic theatre and performance turn text into an object.86 The projected text data from Shakespeare’s playtext in A Piece of Work, including stage directions, are generated in real time and transcoded into other media: synthetic computer voices, video, lighting, music, a fog machine. Metaphorically, what Dorsen herself calls “text-as-data” are also transcoded into the human voice of the performer (either Scott Shepherd or Joan MacIntosh, depending on the evening) by means of an earpiece, a human brain and a speech organ.87 For example, text data like the word “ghost” set in motion a fog machine, spotlights that illuminate the fog and lend the ghost a specific voice and oversized font that semantically emphasizes absence all over. Dorsen exposes the theatrical apparatus as a machine that produces an effect of presence by means of human, but mainly non-human, mechanical, but mainly digital technologies as performers. With the projected code that begins and ends the performance and structures it in between Dorsen indicates the meta-level of the theatre as a machine.

• 88  Claire Swyzen and Kurt Vanhoutte (eds.), Het statuut van de tekst in het postdramatische theater, (...)
• 89  Sue-Ellen Case qtd. in Philip Auslander, “Liveness, Mediatization, and Intermedial Performance,” D (...)
• 90  Aarseth, Espen, Cybertext: Perspectives on Ergodic Literature, Baltimore (Md.), Johns Hopkins univ (...)

40But Dorsen also deploys: an isolated theatre curtain (now and then moved by a wind machine); spotlights; a stage with a rectangular hole in it, suggesting a prompter’s box (the algorithms as prompters?) or a trap for disappearance tricks (exit human consciousness?)—all stock elements of 19^th and 20^th century dramatic theatre. Dramatic and postdramatic elements, even in digital performance, coexist in the “continuum” between drama and postdrama.88 To create this automatically generated theatre and this kinetic effect, Dorsen’s hamletmachine seems to have internalized (Dorsen and co. have programmed it to apply) some conventions of dramatic theatre. These increase the probability that most of its output will be relatively intelligible for humans acquainted with those conventions. As certain words are emotionally coded (i.e., culturally) and transcoded into voices, lighting etc. in a manner that tends to illustrate the text, at bottom, in a realist tradition (e.g., Polonius gets the voice of an old man, etc.), a minor degree of psychologizing characteristic of 19^th and 20^th century dramatic theatre reenters the stage where it was meant, dramaturgically, to exit by means of citationality. Unavoidably, Shakespeare’s lines not only perform as language, but, similar to Gerardjan Rijnders’s and Guy Cassiers’s performance with projected text, Rotjoch (1998), the lines attributed by the author to characters on the page again perform as characters on stage: due to the quality of their voice, the soundscape (e.g., trumpets) accompanying them and the typography of their projected words. Of course, also the anthropomorphizing tendency of the audience plays a role here. The words even seem to acquire an auratic presence despite their ghostly, “screenic” being.89 Yet, compared to her algorithmic performance Hello Hi There (2010), heavily constrained by the dramatic convention of dialogue, in A Piece of Work Dorsen seems to have much further explored “the computer’s potential for combination and world simulation in order to develop new genres that can be valued and used on their own terms [... and to] focus on the computer as literary instrument: a machine for cybertext.”90 How much further could she have gone?

• 91  John Drakakis, “Shakespeare in Quotations,” in Susan Bassnett (ed.), Studying British Cultures. An (...)
• 92  Thomas Cartelli, “Essentializing Shakespeare in the Shakespeare Aftermath: Dmitry Krymov’s Midsumm (...)

41Contrary to Drakakis’s insight that “Shakespeare now is primarily a collage of familiar quotations” the dramaturgical approach of Dorsen more often than not defamiliarizes citations by reconnecting them with other data.91 Does citationality then still have the effect of citationality when citations are difficult to identify? Perhaps it suffices to know that we hear quotes from X without being able to identify them, so Hamlet still stays “an always already known commodity”—the commodification of Shakespeare being the object also of Pourveur’s previously mentioned play.92 Perhaps, in the era of big data, performances like A Piece of Work ask for another attitude of the audience towards text in the theatre: being prepared to experience language patterns (visually and aurally) instead of wanting to rationally understand word by word. Human or non-human, managing the risk of citational and recombinant aesthetics apparently means limiting the random aspect of algorithmic performances and settling for some sort of randomness that has to take into account not only technological, but also human limits.

• 93  Lavender, Andy, Hamlet in Pieces: Shakespeare Reworked: Peter Brook, Robert Lepage, Robert Wilson, (...)

42Crash testing a classic like Hamlet, then, is not simply an act of crashing it against the wall in contestation, but rather a difficult exercise in precisely estimating how hard the crash may be in order to safeguard what is crashed. Safeguarding here means: keeping it sufficiently recognizable in its brokenness, its state of being data processed. The more the cultural object goes “in pieces,” to cite Andy Lavender, the more it risks no longer to be recognized.93 And when it is no longer identifiable, then neither is the performance of the crash test as a form of cultural-technological provocation. This is perhaps the new craft of authors and auteurs as data processors: calculating the probability that the chosen object will come out of the crash sufficiently broken, but still identifiable enough.