The Propositional Nomic

an Agoran Thesis

by Sean "coppro" Hunt

Please note that this thesis requires UTF-8. I also apologize if my idea is not novel; a cursory search indicated it was, but such things are frequently misleading.

Nomic is, by its own title, the Game of Self-Amendment. And yet, it is so much more. Nomic is every game that ever was, every game that will be. And yet, Nomic is none of these. Every nomic has aspects in common, such as a dispute-resolution system (of some nature), and a victory system (of some nature). Nomics tend to fulfill different roles depending on the medium and intent of the game - a face-to-face game, for instance, will likely grow a ruleset that allows it to end with a victor, while an email game often finds itself ending only when it lies decaying and unloved (and consider what the Agoran populace would say if you attempted to declare someone the Grand Winner of Agora, ending it).

The one thing, however, that is truly common to every nomic, and that is pedanticism. Nomicites (as they are sometimes called) will all-too-often (or perhaps not often enough, if you enjoy it) argue over the most trivial of points - whether it is timing, phraseology, or the sort of coup that has now enterend the annals of nomic history as a "Lindrum". The devotion to the fine print does vary from game to game, as in some nomics it's acceptable to retroactively erase a misunderstanding (Agora's ratification process is an execellent is example), whereas in others, the rules must be followed to the letter (most famously in B Nomic, where more than a year's play was erased after the discovery of a misinterpretation of the rules).

Is it possible to have a nomic where pedanticism not an integral part of the gameplay? I don't think so. I do believe, however, that it is possible to create a nomic where a formal system of logics can be used to determine the effects of the rules. The most interesting aspect of codifying a nomic (as redundant as tha tseems) is the temporal aspect. Systems of temporal logic are typically very complex and hardly understandable to the casual viewer.

Suppose the following proposition were a rule, with P being the set of players, and A(x) meaning that a player is active (i.e. it is their turn). Standard mathematical symbols are used (this may be confusing to computer programmers; not in particular that the stroke | indicates NAND rather than OR).

(∃(p ∈ P)(A(p))) ∧ (∀(p ∈ P)(∀(q ∈ P)((p = q) ∨ (A(p) | A(q)))))

This proposition reads, in logically-minded English "For some player p, p is active, and for all players p, and for all players q, either p is q or it is not the case that both p and q are active" This can be simplified down significantly in English, until we get to "Exactly one player is active." A set-theoretic definition is also an option:

|{p|(p ∈ P) ∧ A(p)}| = 1

This assertion is that "The cardinality of the set of all p such that p is a member of the set of players and that p is active is one." Once again, this mathematical definition can be simplified down to "Exactly one player is active."

So let's take a little step back? Is this really helpful? All I've gone and done is used a bunch of complicated and funky symbols to say something that took five words. Did we really gain anything?

We haven't really gained anything here. Instead, we've forced ourselves to make a step we wouldn't otherwise have taken - we've forced ourselves to be explicit. The mathematical formulation had no benefit in this case, but we've stated explicitly a fact often taken for granted - and facts taken for granted are often the culprits for rules issues. In general, such is the case with nomic - formalizing the rules into mathematical symbols is not strictly necessary, but can improve the explicitness of the rules.

However, we've gained something else here. When you consider those two assertions I made earlier in the context of formal logic, you discover something else. You don't need to take them for granted. Rather, treat them as propositions which are not necessarily true. Call the earlier statements about player activity R1 (remember, they're equivalent). Suppose wanted to add a rule that said that any number of players could be active on Tuesdays. Add an additional definition - that DOW() is the day of the week as a number (0 = Monday, 6 = Sunday), and you could have the following rule:

R1 ↔ (DOW() ≠ 1)

And voila, you have now stated that R1 does not apply on Tuesdays. Using this as a framework, I present to you the following basic framework which could be used to construct what I call a "propositional nomic":

1. This game is defined by a set of rules. Each rule is either a metarule, a definition, or a proposition.
2. Metarules are in plain English, and provide the general framework of the game. Metarules each have a number.
3. Definitions are have a value of some sort (not necessarily a number or truth value), and may not reference propositions. Each definition shall have a name with which it can be referred, and may have a variable value (determined by a set of parameters or state outside the game).
4. A proposition is a statement of truth. A proposition is interpreted mathematically and may or may not be true. Each proposition has a natural number associated with it; the first proposition has the number 1, and each subsequent proposition has the following natural number. A proposition may be referred to as P1, where P is its number.
5. A definition or a proposition is referred to as an f-rule.
6. A metarule may define constructs to be referenced in f-rules.
7. No f-rule may be introduced nor amended such that a circular set of references within the f-rules would be created, and no f-rule may be introduced nor amended that is self-contradictory.
8. A terminal proposition is a proposition that is not referenced by any other proposition.
9. For the purposes of this metarule, let N be the number of propositions that are true and let M be the sum of their associated numbers. Together, they cannot have values that would lead to a logical paradox or uncertainty. N always has its maximal value, and M always has its maximal value for a given value of N. [ This is a precedence rule with a "majority rules" effect. ]
10. This game has a temporal aspect. The set of values referred to but not defined by the rules (including the text of the rules) or the outside environment are collectively referred to as the game state. The game state has whatever value it most recently had, except where an unique amendment or set of amendments is necessary to satisfy metarule #9. The game state is constantly reevaluated until no change is necessary. If the game state is stuck in an infinite loop, the last player to do something which affected the game state wins.

This small ruleset is designed to be very strongly specified, and should rarely, if ever, bear amendment. The majority of the game's meat would be contained within the definitions. Paradox is generally avoided with a strongly-defined precedence system and an explicit prohibition against circular reference. A temporal paradox wins a player the game, in traditional nomic fashion (though the method of selection could be better; it was written to provide a quick-and-dirty fix).

An example set of definitions and propositions is provided below.

Pl(p) is true if p is a player. P is the set of all persons. Post(t, p) is true iff t is a text and p is a person and, since the last evaluation of the game state, p has posted a message with text t.

P1: ∀(p ∈ P)(Post("I register.", p) → Pl(p)) P2: ∀(p ∈ P)(Post("I deregister.", p) → ¬Pl(p))

This very simple system provides a framework for players to register and deregister. If a person posts a message reading exactly "I register.", then P1 would be false if that person was not a player. The game state can be amended to make it true, so the game state is amended to make em a player. This is not a particularly detailed example, but exists to show how a propositional system can be used to introduce amendments to game state. Note that if a given message would cause an ambiguous change to the game state, metarule #9 will step in to prevent this from happening.

While I don't believe that a propositional nomic (at least given my sample ruleset) would be a very easy game to play, it would provide a very interesting experience. There are probably flaws with my metarules as formulated, but that is not in the design - the idea is that they are supposed to be airtight and provide a framework under which the definitions and propositions can have contradictions, but never such that they are fatal to the game.

Different rule frameworks can provide different insights into Nomic, and I hope that my suggestion of a propositional nomic here has been enlightening to you.

Thank you for taking the time to read this.