Patterns of Drama

Teodric Abstract: This document explores some aspects of the anatomy of drama. The underlying purpose of this text is finding a workable model for use in interactive software, specifically computer games and computer generated interactive fiction. To this end, the document starts with an analysis of Georges Polti's 36 dramatic situations and from those are extracted the fundamental properties and components necessary to build drama. Finally, the document outlines how to apply this information to a real application.

The document is part of the Haven project .

Availability: Public.

Contents

Background - Purpose - Format - Polti's Cases - Elements Recognized in Polti - Reorganization - Making Use of the Elements - Links - Bibliography and Sources - Copyright

Background

This text is based on a list published in many places on the web. That list in turn is based on a book by Georges Polti called " The Thirty-six Dramatic Situations" (original French title " Les Trente-Six Situations Dramatiques ", 1868?), published in 1916 (the book unfortunately seems a bit hard to find, Amazon for instance only had a used copy last I looked - there is however an entry for it ). The ISBN for the English translation is apparently 0-87116-109-5. One edition was published by Writer Paperback Library.

The list is reprinted/published in many places on the net:

• The WordPlay Archives ( http://www.wordplayer.com/archives/poltisitu.01-12.html )
• Polti.co.uk (http://www.sfarr.demon.co.uk/polti/ ) (Also has a link to a UK "Polti Appreciation Society", if you feel a need for that...)
• A copy of the list is at http://harris-donahue.tripod.com/harrisdonahue/id15.html .
• Info at IBM Research: http://www.research.ibm.com/knowsoc/connections_biblio_arts.html .

The list of Polti has obvious overlap (Enmity of Kinsmen (#13) and Rivalry of Kinsmen (#14) for instance are obviously quite similar) and may lack some categories - one post (archived here ) on the usenet newsgroup alt.movies.silent by David P. Hayes for instance argues that

There is a lot of overlap of categories in Polti's list. "Enmity of Kinsmen" and "Rivalry of Kinsmen" are two separate categories. "Self-Sacrifice for an Ideal" and "Self-Sacrifice" for Kindred" are another two. "Adultery," "Murderous Adultery," "Crimes of Love," and "Involuntary Crimes of Love" count as four situations, although they scarcely seem to qualify as plots apart from one another. Likewise, there are four separate numbers given to "Erroneous Judgment" and three variations on this: "Slaying of Kinsman Unrecognized," "Fatal Impudence," and "Mistaken Jealousy."

Nowhere in Polti's list is there a place for "Invention," "Intellectual Development," "Character Transformation," or "Character Reading Leading to a Change in Circumstances." Therefore, there is no slot on the list where one could categorize the main plot thrusts of "Pygmalion," "Woman on the Moon," "The Country Girl," or even such modern stuff as "The Lawnmower Man"--actually, sci-fi is cut off before it's even looked at.

Regardless of the flaws of the original list or its value as a tool in conventional literature and drama, this text is part of an attempt to make use of the basic structure of Polti's work.

Purpose

Using Polti's list as a starting point, we set up a framework for defining situations in which events can take place. The goal is not to adhere to Polti's list but to get material useful for creating drama or modelling creation of dramatic situations.

Other attempts at using Polti's list have been made:

• Loren J. Miller wrote "The 36 Plots", published at rpglibrary.org ( http://www.rpglibrary.org/articles/storytelling/36plots.htm )
• A page at robotwisdom.com makes use of Polti ( http://www.robotwisdom.com/ai/stories.html )
• The Computer as Storyteller is the subject of a short text at glasswings.com.au ( http://www.glasswings.com.au/modern/pixil/comtel4.htm )

The underlying purpose of this text is finding a workable model for use in interactive software, specifically computer games and computer generated interactive fiction.

Format

The format selected for this approach to Polti's system is (loosely) that of design patterns - if you are unfamiliar with the concept a quick web search should get you started. The idea comes from Alexander's design patterns for architecture, its use here stems from the recent interest in design patterns for software engineering.

Polti's Cases

(This list contains all 36 situations from Polti. It is a word-for-word copy of several identical listings found on the web with only capitalization and format changed. The contents of Polti's text is in the Actors and the Specifics sections for each situation)

Elements Recognized in Polti

[This section so far is a notepad...]
Using Polti as input and adding other items worthy of interest.

I believe that the "system" of Polti is useless. It is based in part on values, in part on effect, instead of being based on structure. The use one can derive from Polti is therefore mostly as a starting point for gathering cases to be covered. Polti does not appear to understand or use the concept of abstraction at all.

Cf situations #7 (Falling prey to misfortune) and #23 (Necessity of sacrificing loved ones); There is obvious overlap between the two, such that in some cases they only differ in focus - one is the story of the victim, one of the betrayer.

Cf situation #24 (Rivalry of superior and inferior); it is in its essence a case with two well defined actors who only need to be equipped with a context within which the specified relationship exists - saying mortality/magic ability/victory/power/wealth/honor is that context satisfies nearly all the cases Polti mentions. Add the case of rivalry with a memory, of double rivalry (valid or not), and possibility of circumstances such as imprisonment (one actor wielding additional powers over the other) and they're all covered. Instead, Polti divides the case into masculine and feminine, without covering the equal circumstances of the two (and with no plausible reason for it).

The solution is to use structure of the situation as the means of describing it instead of Polti's arbitrary categorization of events which is based on genre and values.

Re-examining Polti's List

We divide a situation into three main parts: the primary cause, the attempted resolution by the actors and the actual consequences. The primary cause also needs to consider modifications, motivators and context.

Causes of Conflict or Tension

This analysis recognizes 10 distinct causes of conflict or tension.

Circumstances that create conflict out of nothing. In themselves, these do not create a dramatic situation, but they are very useful initiators for events that do so.

Outer factors

1. Disaster. Typically a natural force or similar.
2. Discovery. Discovery of material assets, capability or knowledge.
   

Circumstances that build up tension, triggering events and actions or bringing existing tension to increased urgency

1. Love, Lust. The primary emotional positive mover.
2. Need, Want, Greed. The primary material positive mover.
3. Ambition, Hunger. The primary social positive mover.
4. Madness
5. Danger, Threat, Opportunity. Very broad. Applicable to a wide range of causes and contexts, including chess-move maneovering and plotting.
6. Vengeance. Requires pre-existing cause.
7. Enmity, Rivalry. Does not always require but may be well served by pre-existing cause.
8. Duty, Taboo, Obstacle to love. Requires set of circumstances that make a hindrance.

Modification, Context and Motivators

The situation must be further quantified to be useful. Context and motivators are mandatory parameters, while modifications are optionally complicating circumstances. 6 distinct motivators and 12 distinct contexts are recognized here - each can have multiple applications in any situation. 5 different modifications are recognized.

Modifications

1. Plea for help. Proximity modifier.
2. Delay. Formal modifier only.
   
3. In error. Formal modifier only. May make information discovery a likely course of action?
   
4. Sacrifice, Duty. Compositional modifier. May make circumstances that taken each on its own are harmless turn into critical conflicts.
5. Deception, Betrayal, Plotting or schemeing by a third person. Complicator that removes the true cause of events one step. Any of the above causes or situations may be the true motivator of the initiator.

Context

1. Marriage, Spouse, Intimate, Lover
2. Family, Ascendant, Descendant
   
3. Clan, Kin
4. Friendship
5. Status Controller, Tyrant, Monarch, Ruler, Protector, Benefactor, Owner. Jailor, ruler etc.Various forms of material or social superiority/inferiority.
6. Organizational. Military organizations for instance create a particular type of contexts.
   
7. Religious, Deity
8. Geographical. Village, neighbor etc.
9. Teacher, Mentor, Tutor
   
10. Professional
11. Business
    
12. Ownership (of item or asset), Debtor.

Note that contexts are the proper place to tie in positional description and provides the relations of actors. The context can be said to be the first thing to establish in many narratives, usually followed by motivators and the tension builders.

Motivators

1. Faith, Honor, Ideal
2. Love, Desire
3. Ambition
   
4. Want or Lack for Item
5. Knowledge, Capability
6. Debt, Personal or Material

Responses, Means and Complications

Responses to circumstances

1. Enigma, Mystery. Information discovery.
2. Obtaining. Resolution for material causes or for proxies of other causes.
3. Revolt. Response to physical or social threat. Changing a situation by conflict.
4. Pursuit, Rescue, Abduction. Response to physical or social threat. Changing, avoiding or postponing a situation by evasion or stealth.
5. Enterprise. Response to any cause. Catch-all for otherwise unclassifiable resolutions.

Means

1. By force: Theft, Robbery, Abduction, Conquest
2. Presenting information: Truth, Falsehood
3. Transfer of resources: Bargain, Purchase, Gift
4. Demonstrating capability or power: Show of force, solving a problem
5. Debt cancellation: Calling in a debt, Fulfilling repayment

Complications

1. Degree of urgency. Is the mover prepared to break laws, taboos, kill, injure etc, how much is the mover prepared to pay a proxy?
2. Degree of potency. How powerful are the various movers in the respective Contexts?

Resolution and Consequence

Resolution

• Successful
• Partially successful
• Failure
  

Consequences

1. Removal of an actor: Death, Injury, Exile, Ejection, Denounciation, Demotion, Humiliation, Imprisonment, Punishment
2. Reevaluation: Someone's view of someone changes, changes in status, Hierarchical changes
3. Change of possession: Ownership of item, Reward
4. Ability: Knowledge gained or lost, capability gained or lost
5. Need fulfilled
6. Repayment: Cancellation of debt

Approaching the Problem

Introduction of Tension

A key aspect of building up dramatic situations is introduction of tension. Tension must be injected into the situation somehow. Building dramatic situations is about how this tension is introduced and how it is resolved. Part of the analysis of to be undertaken here consists of identifying what is part of the dramatic situation and what is part of the form of introduction of tension.

Challenges in Modelling

There may be problems with some of the cases in Polti and with dramatic situations in general in that they may be hard to model and there may be some problems related to modelling as a whole. Some of these are summarized here.

• Deception or misinformation, lack of knowledge

Modelling lack of knowledge seems harder than modelling knowledge itself. The problem becomes apparent when trying to abstract plots based on situations where one actors actions would have been different if that actor had known the real truth of affairs, such as when involuntary or unknowingly effecting the death or suffering of unrecognized relatives (as in several of Polti's cases). On the other hand, it may be the case that those dramas represent situations that are the most contrived, and that they may be less desirable for automatically generating. A world of unrecognized siblings is not generally something to strive for.

The modifiers In error and Plotting or schemeing by a third party both fit here.

• The Double Bind

There is a need to be able to create situations in which the flow of events is non-trivial. Interlocked events may be one way of achieving this.

• Situation Complexes

With this we mean intertwined situations, where actors are simultaneously involved in several situations. This is extremely common in literature and cinema. Arranging complexes of the type seen in advanced drama presents a new level of challenge to the whole problem.

The modifier Sacrifice, Duty is a good example of a situation requiring the ability to model both double binds and situation complexes.

• Planning
  

The crucial part of the modelling may be the planning of the actions to be taken by actors. At this time I see this mostly as an AI problem. In some cases the plan is trivial - if the desired change state is acquisition of an item the alternatives may be purchase or theft, and the decision of which one to use is simply influenced by means available (money or barter to trade with) and disposition (whether theft is permissible or acceptable to the actor). Once the plan is made, execution can be delegated to another actor (i.e. the player).

• Extension
  

New settings or conditions may represent challenges. What if something not previously included in the model of the world needed inclusion? Say for instance, that dreams or prophesies were suddenly a requisite (as motivators, means, information or whatever). Unless such additions can be represented using the existing building blocks, this may be a problem.

Reorganization

This section introduces some things needed to model drama that were beyond the scope of Polti.

Proxies

Proxies are actors working for or representing another actor. They may be active parts of the events taking place or they may be just messengers or carriers of means. In a model situation, the proxy role is the most likely for a player to take on.

Further Exploration

Further exploration of the topic can be found in A Deeper Look at Conflicts .

Making Use of the Elements

Some notes on the suitability of certain mechanisms, and the usefulness of some of the situations in generated stories.

Need for External Prerequisites

To make reasonable use of certain patterns, external and pre-existing state must be established. A murder mystery cannot happen unless someone has been murdered (well, unless it's really a mystery of deception, not murder), a person to be found must first be missing. The murder must have had a reason or story, the missing person did not just poof. These events most likely require background creation and introduce a second problem - it is no good to waste a NPC's life in a murder mystery unless a player finds out about it.
Large scale events are required for many situations - war and natural catastrophes being examples, both with far-reaching consequences like destruction of property, loss of life, refugees and general hardship. These large events must either be introduced manually or be driven by a large-scale event engine. When present, large events can be used to trigger smaller events as cause or storyline for dynamic content. Large scale events are very lucrative settings for drama. A changing world is an effective setting making less engaging, smaller stories deliver a more appealing drama.

Note on Expert Systems

There is a noteworthy connection to problems in expert systems. Expert systems frequently need to quantify or justify their reasoning to the user in order to maintain credibility. For instance, a medical advisor system will need to show why a certain action should be taken.
The same may be true for a system attempting to draw in a user into the plots it designs. While a user may be satisfied with just reaping the tangible benefits of performing an action within the system, to further draw in the user it is necessary to make the events seem credible or sensible within the context at hand. This information could take different shapes - it could be the story told by an actor to the player in order to explain what needs to be done and why, or it could be things the player needs to find out in order to complete a mission. Regardless of how it is presented, it serves to justify the actions and draw the user into the flow of events.

Links

Beyond the links mentioned above, here are some for further reference and investigation:

• http://nickm.com/if/ (with a large list of references)
• Richard Bartle ("writer of the first MUD") has some resources on his web pages, including a text by Phil Goetz called "Interactive Fiction and Computers" ( http://www.mud.co.uk/richard/ifan194.htm )
  

It would be very interesting to find any attempt at catalogueing plot types of non-european drama. Polti's list was probably focused on classic European works. The topic of Polti's list as related to Shakespeare's dramas is sometimes brought up but while it is tempting to try to map the 36 situations to Shakespeare's 36 plays, doing so is not very helpful. The more interesting point is instead how very easy it is to envision shakespearean drama based on each of the items in Polti's list. Indeed, Shakespeare had a fine perception for the possibilities in dramatic situations.

Bibliography and Sources

• Georges Polti, The Thirty-six Dramatic Situations , 1916 (original French title Les Trente-Six Situations Dramatiques , 1868), ISBN 0-87116-109-5
  

Copyright

This text was written in its entirety by Olof Ekström but is based on a book, see the background section. For more information about the author of this page, see Olof Ekström's personal information in the Project Profiles document.

Copyright © 2002 Olof Ekström/Extro System. All rights reserved.

Bälinge/Uppsala, Sweden, April-July 2002