As soon as something has been found to have a recognisable and quantifiable structure, it becomes readily available for computer modelling and simulation. The work of various critics proved suggestive of a paint-by-numbers approach to plot construction. Therefore, computer science researchers and programmers such as Kevin M. Brooks of the MIT Media Lab [Bro96a]; Marie-Laure Ryan who is the author of Possible Worlds, Artificial Intelligence and Narrative Theory [Rya91]; James Meehan who developed Tale-spin [Mee76]; N. Dehn who developed Author [Deh81]; Michael Lebowitz who developed Universe [Leb85]; Masoud Yazdani developer of Roald [Yaz96]; and Melanie Anne Phillips and Chris Huntley developers of Dramatica [PH94] all began creating computational Scheherazades. Many of their systems automated story plotting with enough variables to produce thousands of artificially generated plots. However, the results usually remain threadbare and only nominally interesting, as in this example output from the program Tale-spin:
WONDERFUL SMART LADY BUXLEY WAS RICH. UGLY OVERSEXED LADY BUXLEY WAS SINGLE. JOHN WAS LADY BUXLEY'S NEPHEW. IMPOVERISHED IRRITABLE JOHN WAS EVIL. HANDSOME OVERSEXED JOHN BUXLEY WAS SINGLE. JOHN HATED EDWARD. JOHN BUXLEY HATED DR. BARTHOLOMEW HUME. BRILLIANT HUME WAS EVIL. HUME WAS OVERSEXED. HANDSOME DR. BARTHOLOMEW WAS SINGLE. KIND EASYGOING EDWARD WAS RICH. OVERSEXED LORD EDWARD WAS UGLY. LORD EDWARD WAS MARRIED TO LADY JANE. EDWARD LIKED MARY JANE. EDWARD WAS NOT JEALOUS. LORD EDWARD DISLIKED JOHN. PRETTY JEALOUS JANE LIKED LORD EDWARD...
Sample as presented in "Computational Story Writing"
[Yaz96]
The production of as many coherent plots as possible is the sort of task for which a computer is eminently suited. The output may then be viewed on screen or printed out, the production is uniquely computer based but the experience of the results is not. Nevertheless, it must be emphasised that what is produced is merely a plot, not a fully realised story.
Good writer[s know] that there is no single, ideal "form" for the short story--a fixed and sanctified way of handling things--or even a set of forms, already determined, into which [they] can fit [their] own story. [They know] that, when [they set] out to write a story, [they are] really engaged in a process of exploration and experiment: [they are] exploring the nature of [their] characters and the meaning of their acts, and, too, [they are] exploring [their] own feelings about them.
Many of the creators of these programs recognise this and offer their programs as a tool to assist the skilled writer, though not without wistfully dreaming of the ultimate storytelling machine as in Stanislaw Lem's "Tale of the Three Storytelling Machines of King Genius." [LtMK74]
Most of the plot programs I have seen to date are based exclusively on the work of literary critics. A creative writing instructor of mine used to say that critics do not criticise in the same way that writers write. Though the "how to write" book has become an established genre of its own, some authors take the time to provide genuine insights into their creative processes. Such insights could very well be invaluable in the creation of plot generation tools. Of course even writers using such tools must remember that storytelling requires a deeply personal commitment. Good storytelling engages the audience intellectually, emotionally and aesthetically. Without storytellers who can explore these dimensions afresh each time on the behalf of the audience, the results are bound to be sterile and may ultimately even be doomed to failure.
Other fields of research within computational storytelling may prove more fruitful where the aim is to provide the tools for creating fully immersive, interactive worlds under the direction of a team of people including writers, artists and programmers who will then develop plot, character and visual detail. The Oz Project [Bat88] explores the development of either graphical or textual worlds into which are introduced intelligent agents, characters with a given set of behaviours that they can independently execute. Without pre-determined goals and events, the results are an entertaining fictional world, but not a story. However, the addition of intelligent agents opens up some fascinating possibilities for the storyteller.
Copyright © 1997 Katherine Phelps