4. Too much conceiving: humans, machines, and meaning
¶ 1 Leave a comment on paragraph 1 0 [I]n this new theory the word information relates not so much to what you do say, as to what you could say. That is, information is a measure of your freedom of choice when you select a message. […] The two messages between which one must choose in such a selection can be anything one likes. One might be the King James version of the Bible, and the other might be “Yes.”Warren Weaver, “The Mathematics of Communication”
King James Bible (1631)
James Joyce, Ulysses (synoptic ed.)
¶ 4 Leave a comment on paragraph 4 0 Though the architects of information theory may not have wished their theory to become an ideology, that very situation developed over the last half-century. The ideological dimensions of information theory have been well documented by Nunberg, Hayles, and others in recent years, but less recognized is Shakespeare’s centrality to this theory as a touchstone example. In contrast to the humanities computing field of more recent decades, information theory’s most influential early figures were not literary scholars, let alone theatrical practitioners. Although formative events like the Macy conferences embodied the spirit of what we now call interdisciplinary, information theory’s origin and early takeup were decidedly on the scientific side of what C.P. Snow called the division of the two cultures, the arts and the sciences.
¶ 5 Leave a comment on paragraph 5 0 Even so, Shakespeare appears in the theoretical discourse about information and computing from the beginning, even in one of the founding works of modern computing, mathematician Alan Turing’s 1950 essay “Computing Machinery and Intelligence.” One of the first serious scientific discussions of artificial intelligence, this essay proposes criteria for what has become known as the Turing test, a scenario in which a human judge converses with a hidden interlocutor who may be another human or a machine; if the judge cannot tell which is which, the machine has passed the test. Notably, Turing’s proposal envisions this test partly in terms of a “sonnet-writing machine” that could simulate what Turing regards as an essentially human mode of communication. Taking an example from Shakespeare’s Sonnet 18 (“Shall I compare thee to a summer’s day?”), he equates Shakespeare with an aspect of human communication that his audience might regard as exceeding the machine’s capacity for modeling.
¶ 6 Leave a comment on paragraph 6 0 These very conditions materialized, with Shakespeare at their center, when an actual Turing test in 1991 at the Boston Computer Museum, in which ten human judges conversed via keyboard and screen with six artificial intelligence programs and four human confederates, with the confederates’ and computers’ identities concealed. Shakespeare again showed up among the topics of conversation selected for the test, along with “dry martinis”, “women’s clothing,” and other topics “of the sort appropriate for a cocktail party,” as one of the referees recounts. An episode of the PBS series Scientific American Frontiers captures some of the conversations between the judges and a human confederate, Cynthia Clay, whose knowledge of Shakespearean matters such as the authorship complexities of Pericles unexpectedly reversed the dynamics of the Turing test: judges and audience members misidentified her as a computer. As one audience member reasoned, “some of [Clay’s] answers seemed too studied, as if they were somehow canned opinions that came from a large database”; similarly, a judge thought no human could command so much information about Shakespeare. The subtext of the episode is revealing. Clay’s good-humored but exasperated response—“people, go to school!”—highlights the ostensible status of Shakespearean information as a kind of public commons, like baseball statistics, but hints also at a decline of Shakespeare as knowledge internalized by the public. By the time computing had advanced to the stage where a real Turing test was viable, a person like Clay had become an anomaly, and her personal command of information on Shakespeare most readily explained in computational terms. Between 1949 and 1991, Shakespeare had become an information domain where the boundary between human and machine was contested.
¶ 7 Leave a comment on paragraph 7 0 The Turing example is no exception, but part of a pattern in explanations of information as a measure of disambiguation rather than a quantity of knowledge. G.H. Hardy’s 1940 book, A Mathematician’s Apology, while predating information theory as such by a few years, nonetheless sets the tone by using Shakespeare’s texts to explain the difference between pure and applied mathematics. Hardy argues that it would be fallacious to suppose that our understanding of a theorem should be affected by the physical character of the means we use to represent it, like imperfectly drawn shapes on a lecture-hall blackboard: “It would be like supposing that a play of Shakespeare is changed when a reader spills his tea over a page. The play is independent of the pages on which it is printed, and ‘pure geometries’ are independent of lecture rooms, or of any other detail of the physical world.”
¶ 8 Leave a comment on paragraph 8 0 Similar idealism appears in a 1956 book by one of the founders of information theory, the physicist Léon Brillouin, who mentions Shakespeare at the point where he explains (following Shannon) that information theory does not deal with “the human value of information,” or what we might call meaning, but only with the quantifiable behavior of information in a communications system. Brillouin states that we can quantify the amount of information in any arbitrary string of 100 letters, regardless of whether they are (in his words) “a set of 100 letters selected at random […], a sentence of 100 letters from a newspaper, a piece of Shakespeare or a theorem of Einstein.” Information is to be distinguished from knowledge, with knowledge represented by Einstein, Shakespeare, and a newspaper. Compare this comment from the unpublished 1959 notebook of French biologist Jacques Monod: “From the point of view of the theory of information, the works of Shakespeare, with the same number of letters and signs aligned at random by a monkey, would have the same value. […] What could be considered ‘objective’ in the Shakespearean information that would distinguish it from the monkey’s information? Essentially the transmissibility.” Monod alludes to the commonplace idea of infinite monkeys at typewriters reproducing the works of Shakespeare by accident, though in popular culture Shakespeare’s works have somehow substituted themselves for the books in the British Museum, which was the original example popularized by British physicist Arthur Eddington.
¶ 9 Leave a comment on paragraph 9 0 These are just a few examples among many. The pattern that emerges reveals certain assumptions: that Shakespeare is ideal text, not material text or performance; that the plays are the products of a single authorial source (Cynthia Clay was unwilling to discuss The Two Noble Kinsmen as a co-authored play, and correctly guessed that human judges would not recognize the collaborative play Pericles); that these texts themselves possess no more internal complexity of transmission than any other book on the shelf; and that despite whatever cultural power we may locate in Shakespeare—what Brillouin calls “human value”—information theory holds a greater power to make information strategically mobile by asserting a fundamental equivalency between all texts.
¶ 10 Leave a comment on paragraph 10 0 Information theory’s claims for Shakespeare’s unexceptionality might seem as provocative as Benjamin Jowett’s controversial claim a century earlier that the Bible could be interpreted “like any other book,” but this approach also resonates with the contemporary thinking about literary texts that was coming from the New Bibliographers. F.P. Wilson’s 1945 book, Shakespeare and the New Bibliography, exemplifies this sympathy in a brash claim which nonetheless echoes the spirit of the Weaver epigraph above: “To a formal bibliographer a book is not the life-blood of a master spirit but a collection of pieces of paper with printing on them.” This hearkens back to W.W. Greg’s dictum from “Bibliography—An Apologia”: “what the bibliographer is concerned with is pieces of paper or parchment covered with certain written or printed signs. With these signs he is concerned merely as arbitrary marks; their meaning is no business of his.” Such bracketing of meaning for the purpose of determining a mechanical process mirrors Shannon’s similar gesture in most respects, including its contentiousness in the eyes of critics.