Interactive fiction is any narrative which is at least partly determined by one or more choices on the part of the audience. It can be extremely rudimentary, as in Frank R. Stockton's famous "The Lady or The Tiger"--in which the author sets up two equally likely but diametrically opposed endings and terminates the story without concluding one or the other, leaving the outcome to the reader's imagination--or extremely sophisticated, as in the modern computerized interactive fiction of Emily Short--in which the reader is prompted at each step to direct the decisions of the protagonist and may do so using typed commands in natural language (descended from the classic Infocom adventure games like "Zork"). Other variations include Ayn Rand's play "Night of January 16th" (in which audience members vote democratically to determine the outcome), DVDs which include the ability to choose between alternate endings, and the much-beloved "Choose Your Own Adventure" (CYOA) books of my youth. Many permutations of techniques and media can be imagined.
Very often (albeit not necessarily), interactive fiction is written in the second person "you" voice, which is generally rejected by authors of traditional fiction due to its artificiality. Although a competent fiction author will never tell the reader how to feel, he or she is more or less obliged to describe some action on the part of the protagonist, and if that protagonist is the reader, as the second person implies, the author almost inevitably ends up telling the reader how he or she will or did respond to certain events. This is insulting to most adult readers, who, one would hope, know their own minds and hearts better than any author ever will. The 2nd person can work in certain situations, for instance "guided meditation" narratives, in which the reader has made a conscious decision to be suggestible. But generally speaking, requiring readers to suspend personal will as well as disbelief is asking too much.
Interactive fiction takes some of this burden off the second-person pronoun. By allowing readers to choose actions for themselves, the author can limit his or her prose, at least, to simply describing the setting, events, and actions of other characters. Authorial intrusion persists in the definition of the alternatives (which, as E. E. Schattschneider reminds us, "is the supreme instrument of power.") which the reader may choose, but at least on the surface the air of mind-control is gone. In truth, this superficial remedy is rarely satisfying to a mature mind, which knows a rigged game when it sees one, and thus second-person interactive fiction typically ends up being regarded more as a game or puzzle, or as children's fare, than as serious literature. Also, the extent to which the form facilitates ready connections between choices and consequences makes it ideal for instructional purposes, and this only serves to exacerbate the just-for-kids aura. An interesting counter-example here, from the adult world, is John Antal's "Armor Attacks: The Tank Platoon: An Interactive Exercise in Small-Unit Tactics and Leadership," which is basically a CYOA book to help prepare calvary officers for combat decision making. But even the original children's CYOA books, although clearly written mostly for entertainment purposes, contain a certain didactic overtone, best exemplified by a quote from the Page 1 "Warning" common to all of the books:
"From time to time as you read along, you will be asked to make a choice. Your choice may lead to success or disaster! The adventures you take are a result of your choice. You are responsible because you choose! After you make your choice, follow the instructions to see what happens to you next. Remember--you cannot go back!"
I think Edward Packard, who popularized the form, would agree that while the stories may be pure fluff, there is a certain moral inherent in the form itself, which, of course, is that the reader, a child, is responsible for the outcomes of his or her actions. Certainly an important life lesson, but here we have, accidentally, uncovered a potential stylistic flaw in the CYOA genre, which probably results from tension between the drive to entertain and the drive to instruct. If the lesson is one of responsibility, then clearly there should be strong causal connections between the reader's choices and the outcomes for the story, so that the reader is encouraged to think ahead about what the potential consequences of a certain course might be. But very often in the genre one finds a causal disconnect between choices and outcomes; an example, here, from Packard's "The Mystery of Chimney Rock" will serve to illustrate:
On p. 38, the reader, while searching for his or her mischievous cousin in a haunted house, and having been confronted by the witch who owns the place and who, in the guise of a kindly old woman, offers cheese and crackers, has chosen not to take "candy from strangers" and instead to run away. The witch then instructs her maid to block the reader's escape, but the maid rebels and denounces the witch and tries to lead the reader out of the house. At this critical junction the reader is asked to choose between immediate escape with the maid and returning to rescue his or her cousin Jane, who has been trapped in another part of the house. If the reader chooses to escape immediately, he or she finds a policeman waiting outside the house, that Jane has already escaped on her own, and that the witch has died of a heart attack and the curse on the house is thereby (somehow) lifted. On the other hand, if the reader chooses to go back and rescue Jane, he or she finds Jane waiting in the hallway and escapes together with her and the maid as before. This time, however, there is no policeman, the witch is not dead, and the cat which is the witch's familiar (or were-form) stalks them menacingly as they flee into the night, suggesting that they may not have seen the last of the curse.
Now, while one could argue about which is "the right" choice for a child in this situation, that debate misses the point altogether: Jane escapes regardless of the reader's choice to try to rescue her or not, and the only difference between the outcomes of the two choices is that in one the curse is lifted and in the other it is not. But there's not really any conceivable causal connection between the reader's choice to escape immediately or to try to rescue Jane and either of these two outcomes, which are thus basically random. The message of the medium is that one has a choice, but the message of the content is that those choices make no difference as to what actually happens, and that one might as well choose randomly. This, patently, is opposed to the spirit of the enterprise, and particularly if the point is to impress the importance of careful decision making.
But even if the goal is merely entertainment, such a disconnect between medium and content remains aesthetically offensive. To make the point, imagine the kind of extreme deterministic mockery of a CYOA book, complete with many possible choices of routes through the pages, all of which lead to exactly the same outcome by exactly the same story. The reader is offered choices and may make them, but none of them make any difference to the course of the story. And while there may be a certain amusing philosophical irony in the presentation of the illusion of choice while denying its actuality, the reader of such a book is nearly certain to feel put upon and insulted. (More amusing, perhaps, is the possibility of an anti-didactic CYOA book which, in the spirit of some of Shel Silverstein's "children's" poems, consistently rewards bad behavior and punishes good.)
These "gimmicky" approaches to the problem of a mature CYOA book may be clever and perhaps very amusing, but truly would not rise much above the level of elaborate jokes. To achieve my longstanding goal of writing a CYOA book that could succeed as serious fiction, for adults, a more fundamental strategy is required. And that, in spite of the title and direction of this essay, is to drop the second-person voice, and write in the third-, where the "rigged game" effect becomes no more of a problem than in normal non-interactive fiction. The first person may offer some interesting possibilities, as well, with the narrator acting as the reader's "agent" in the fictional world and reporting back the outcomes of his assignments. This agent would have a character all his own, and might choose to obey or ignore the reader's choices for his own reasons, or likewise to accurately or inaccurately report the outcomes of those choices.
Finally, my conclusions about the second person for serious interactive book-fiction purposes should not be taken to imply that I think second-person interactive fiction in general is a lost cause. In fact, I think quite the opposite. It is probably fair to speculate that second-person interactive fiction can succeed to the extent that it presents the reader with a realistic level of choice. Given the infinity of futures that spiral away from every instant of reality, however, I find it difficult to imagine that this goal can be achieved in the format of a book. But computer-based interactive fiction is another matter entirely. It is conceivable, indeed some would argue commonplace today, for a computerized universe to offer sufficient choice on the behalf of the "reader" at every moment to escape the impression that things cannot unfold otherwise than as they do. To do so, contemporary computerized interactive fictions turns increasingly to rule-based reality simulation and multi-user participation. It might even be argued that MMORPGs like Everquest, Second Life, and EVE constitute interactive fiction in its highest state of development to date. And there is no fundamental reason why these graphically-intensive universes could not be implemented textually, thus overcoming the reservations of those who would argue that fiction requires written language.