Tasking Ariel in Graham Nelson's The Tempest

  ...but this swift business
I must uneasy make, lest too light winning
Make the prize light.
  --Prospero, The Tempest

In Graham Nelson's text-based "interactive performance" of The Tempest, the player/reader takes on the role of the fairy spirit Ariel, who must perform tricks in order to win his freedom. The game file contains the nearly complete play text, plus an additional 20% of iambic pentameter computer messages of the "You can't do that here" variety. In theory, text-based interaction sounds like a great way to experience Shakespeare's work in a new context. In practice, however, Nelson's program is likely to prove equally frustrating to fans of modern computer games (who are used to a much greater degree of interaction on a broader, shallower narrative field) and "serious" theater people (who will be put off by the interludes of puzzle-solving gameplay that interrupt the dramatic flow of events). Nelson's dramatic experiment is most valuable for the light it casts upon the nature of this particular computer-mediated genre.

The Tempest seems a natural candidate for interactive fiction -- not only because many IF games in the 1980s featured wizardry (and thus audiences might be attracted to the subject matter) but also because both Elizabethan drama and interactive fiction use language in order to stimulate the senses. Shakespeare and his contemporaries openly wrestled with the limits of Elizabethan stagecraft -- for example, in the prologue to Henry V, the Chorus apologizes for not being able to produce real kings and whole armies, imploring the audience: "Think, when we talk of horses, that you see them / Printing their proud hoofs i' th' receiving earth; / For 'tis your thoughts that now must deck our kings, / Carry them here and there, jumping o'er times...". Further, the speech with which Enobarbus describes Cleopatra's barge (Antony and Cleopatra II.i) is not only more economical, but far more elegant than any mechanical special effect the Elizabethan stage might have attempted.

"Mystery House," Williams 1980Shakespeare's Chorus would need to do a lot of explaining in order to draw audiences into the pictures from Roberta Williams's 1980 "Mystery House" (generally held as the first computer narrative that employed graphics).  Infocom's advertising campaigns played upon the notions that text games were more intellectual and did not require expensive computer systems.  After a boom in the early and mid 80s, the commercial market for text-based computer games fell before the arrival of inexpensive computer graphics. As steady improvements in 3-D graphics degraded the narrative elements of the next generation of commercial computer games (in which most characters exist only for target practice), a Usenet community of die-hard IF hobbyists (rec.arts.int-fiction and rec.games.int-fiction) created and shared new stories, reviewed each other's works, pondered the classics, and developed a unique critical vocabulary. Their online messages were just as likely to refer to the nature of Aristotelian mimesis as to the limitations of the Turing artificial intelligence test. 

Nelson -- who created the IF programming language Inform, and also some of the best IF of the 90s, is Marlowe (of the “mighty line”) and Shakespeare rolled into one. (His online persona also shares elements of Dr. Johnson, Lewis Carroll, and, of late, J.D. Salinger.) His epic works Curses (a delirious mythological and genealogical romp, 1993) and Jigsaw (a time-travel romance, 1995), did much to rekindle interest in "serious" IF. While Nelson's IF stories are in and of themselves notable, even more remarkable is Nelson's creation of the programming language Inform, an authoring system for IF, which he has generously donated to the worldwide gaming community. 

While Linda Hutcheon, discussing the postmodern tradition of emphasizing the receiver's role in constructing a text, offers interactive fiction as "the most extreme example I can think of" (77), the plots of most IF works are tightly constrained, such that the story does not advance until the player-reader has solved certain puzzles.  The puzzles can range from uttering a magic word, to finding the right key, to successfully mastering a complex simulation of a WWII "enigma machine" (from an extremely challenging chapter in Graham Nelson's "Jigsaw"); but owing to the technical difficulty of coding such puzzles, and the aesthetic difficulty of integrating such puzzles into the fabric of the story, the plots of most IF works are tightly constrained.  Aarseth's Cybertext is one of few critical examinations of electronic text that looks beyond canonical literary hypertext  (see Landow, Moulthrop and Shelly).  Montfort and van der Linde are among those who recognize the significance of Aarseth's efforts to expand the horizons of the attempt to theorize electronic literature. 

Playing The Tempest

Upon being greeted with the opening screen of text (see transcript Tempest-1) an experienced IF player would note the reference to "the sharp wind of the north" in the opening sequence (which actually comes from I.ii.225) and would try to "go north." The computer would print a description of this area, revealing the ship carrying the Duke of Milan and his train. A user who knows the play, or who can infer properly from the game's title and from Ariel's earlier presence "on the curl'd clouds" would gather that Ariel's job is to create a storm.  Simply typing "create storm" yields only the iambic pentameter error message, "That instruction, that verb, doth elude me." But an experienced IF player knows that "blow on" is one of the commands that game authors frequently employ; thus, typing "blow on ship" conjures up the storm. The storm is described via a passage lifted from Ariel's report to Prospero, and the player is awarded one point.

The ship (which now "founders on the wave") is still the only object in the vicinity, so the player sends Ariel in for a closer look (typing something like "enter ship" or simply "go in"). This action triggers a cut-scene, which represents the (slightly edited) opening dialogue. The individual speeches are visually separated by the symbol "<--->," which also signals that the computer is waiting for a keypress. When the scene plays out, and the narrative stops in order to give Ariel a chance to act, the screen displays the ">" prompt.   

When the interface works, it works well -- as in transcript Tempest 2, in which the user tells Ariel to open and then enter the hatches (which prompts the King's party to abandon ship, and thus brings them to Prospero's island). Opening and entering objects are actions commonly encountered in interactive fiction.

"The Tempest," submitted under the name "William Shakespeare" to the 1997 Interactive Fiction Contest, finished 27th out of 34 entrants. While it later won the 1997 "Best Use of the Medium" from XYZZYNews, upon its first release, reviewers were both surprised and frustrated by Nelson's respectful adherence to the Shakespearean plot. One of the competition judges, apparently as vexed by Nelson's The Tempest as the mariners he paraphrases, dismissed the entry:

What a clever idea! (Which, together with a ha'penny, will buy you a brick.)

I couldn't figure out what the hell to do. Even reading the beginning of the original play. I got as far as when the King's party jumped overboard, and then I was stuck. So I split. I split, I split, I split. (Plotkin)

The passive act of reading the dialogue far outweighs the interactive elements, as in this extended sequence. The user types several commands which the programmer must exclude from the narrative. In transcript Tempest-3, the first several commands ("look," "kiss miranda," "hit prospero," and "throw phial at prospero") do nothing to advance the plot, but they do establish the setting and lay down some of the rules of this fictional world.

Nelson is far too respectful of Shakespeare's text to allow any alteration to the plot. The resulting necessity to close off all other possible actions (except for the one action necessary to trigger the next scene) left many reviewers wondering where the "fun" in the game was supposed to lie. Others pulled out their copies of The Tempest in order to figure out what to tell Ariel to do, and then felt irked when the "reward" for solving a puzzle consisted in having to scroll past lines of online text that replayed the scene they had just studied. One poster, pondering the brave new world of interactive fiction, wondered:

Would it be possible to write a story-line exploring other possibilities of the 'Tempest' such as Miranda not falling in love (which I always thought was a litle [sic] too pat), the baddies killing each other... a sort of 'what if?' interactive fiction. (Olive)

Nelson himself replied as follows:

It might have been possible (for a better author, anyway), except that then it wouldn't have been a performance of "The Tempest"... it would have been more like those 19th-century outings for "King Lear" in which Cordelia lives at the end and marries Edgar while Lear goes off to an old people's home, on the grounds that everybody likes a happy ending. (Nelson, "Re: Tempest: still stuck!")

Despite the fact that the game will let the plot progress only towards the end that Shakespeare had already prescribed for it, some of the puzzles are fairly difficult to solve, simply because their solutions are not always clear. (Whereas "blow on ship" was the solution to the opening puzzle, one cannot "blow on flute" to wake Ferdinand, but must rather "sing". Such a restriction may be obvious to anyone who remembers Ariel's songs from this scene, but within the logic of the game, there is no reason why playing the flute, or simply shaking his shoulders, wouldn't do just as well.)

Nelson did not provide a hint file or a walkthrough (a full list of all the commands necessary to trigger the narrative's optimal conclusion). One USENET wag suggested that the full text of Shakespeare's The Tempest should suffice; yet even when the player knows what scene is supposed to happen next, it's sometimes hard to figure out how to tell Ariel to accomplish it. As Prospero notes (in the quote at the top of this page), to make a game too easy robs it of its pleasure; but the "prize" Miranda and Ferdinand stand to win is to live happily ever after. I for one felt silly re-reading a scene in order to figure out what to tell Ariel to do, and then having the game "reward" me by displaying that very same scene!

While most IF games permit the player a great deal of local freedom (wandering around a sprawling textual topography, interacting with complex props, and solving puzzles which unlock doors that lead to new exploratory and interactive possibilities), Nelson does not give the reader the opportunity to change the plot. This makes a certain amount of sense; after all, a player who takes on the role of Ariel should feel Ariel's constraints.  Whereas Shakespeare's textual world is rich and vivid, Nelson's version -- which used the very same words -- feels like a cardboard cutout. I don't see this as a weakness in Nelson's ability, but rather as evidence of the complex layering of textual meanings and interactive possibilities that native IF typically represents.

From Duncan Stevens's online review of The Tempest 

Certainly, the extensiveness of the Inform hacking is impressive, and the sheer concept of adapting a drama and making it interactive is novel -- but the game does not, in truth, meet all the challenges the task presented..... Though the gameplay limitations of Tempest are considerable, they are there for a valid reason, not simply inadequate coding -- and, as such, I decided they shouldn't count too heavily aganist [sic] the game. Though it doesn't "work" especially well, the concept as put into practice works about as well as it could, and the author should get some credit for a worthy effort. <http://www.sparkynet.com/spag/t-z.html#tempest>

From Paul O'Brian's online review of The Tempest

The Tempest is entertaining and innovative; it often feels quite magical to inhabit the Prospero/Ariel connection, and to take part in a groundbreaking interactive experience. I think that the game also has great potential as an educational tool, allowing readers to experience Shakespeare's language in a new and thrilling way.... The author's erudition is clear, from the simple choice of subject matter to the deft interweaving of other Shakespearean and Renaissance phrases into the play's text when necessary (for example, to the command "throw x at character" the game responds "I have no aim, no, no chance of a palpable hit.", a phrase echoing Hamlet). Such attention to scholarly detail recalls some of the finer moments of Nelson's epics, especially Jigsaw.   <http://ucsu.Colorado.EDU/~obrian/97rev4.html#tempes>

Works Cited

Category Tags