A Role for Could, Would, and Should in Game Input

Games controlled by sentences rather than buttons raise the design question of what subset of English to implement. Frequently this becomes a question of what the "verbs of gameplay" are, to take game designer Chris Crawford's metaphor literally. But his otherwise strong analogy misleads the designer of a sentential-input game.  Because a button press translates so naturally into an imperative sentence executed at that moment, it is easy to assume language-based game input should be imperative, and the verbs-of-gameplay metaphor only reinforces that line of thinking.  As a counterpoint, let us examine the utility of the three modal auxiliaries "could", "would", and "should" in multi-character games of (preferably dramatic) one-upmanship.

In any strategic board game, each player evaluates his options, evaluates his opponents' options, and attempts to predict his opponents' plans in order to account for them in his own plans.  Each player asks himself "what could" he do, "what could" each opponent do, "what should" each opponent do in order to win, and with that knowledge, "what would" each opponent do, given each his temperament and natural inclinations. It is only after considering these questions he finally settles on what he will do.  Our trinity of modals encompasses brainstorming, evaluating, predicting, and planning.  In computer game terms, they encompass AI.   

Allowing agents to query an AI about its internal state is a useful feature for interactive story.  It allows a human player to discuss the intentions of his NPC rivals with his NPC allies and get non-scripted answers sensitive to the gamestate and the state of the particular NPC AI in question. In the prose medium of interactive fiction, the same information can be presented outside the double quotes as a kind of interactive interiority.  When dialogue comprises a good deal of sussing out rivals' plans, it helps move the primary currencies of gameplay from money, magic, and munitions to plans, permissions, and predictions.  And when asked modal questions, characters can choose to lie, forming second-order currencies such as trust and perceptiveness. 

Drama is more than culpability and second-guessing of course, but those soft currencies are better grist for drama than spell points.  Characters especially powerful or especially mysterious draw questions about "what could" or "what would" they do, which can act as a reward for or proof of an especial player's agency.  "Should" is a particularly strong word, even a cantankerous word among many artists I know.  It implies judgement, and judging another's work can lead to conflict.  But it is precisely this aspect of the word that is so useful inside a fiction.  Many man versus man, man versus self, and man versus society conflicts can be expressed as a contradictory pair of "should" statements.  "Should" is also a friend to characterization:  how someone answers such a question alludes to what he believes most important.  Interestingly, a goal-directed AI can have its goal stated in plain English as a "should" statement.  So can the goal of a game.  

Contemporary author Tom Clancy once said, "The difference between real life and fiction is that fiction must make sense."  If real human beings continue to baffle researchers for the foreseeable future, then perhaps traditional narrative should continue to clarify and articulate characters' inner lives. Readers, game players, and current AI technology all like clarity.  If traditional narrative hinges on "will he or won't he", then perhaps interactive narrative should hinge on "would he or wouldn't he": the player clarifies his options as per his rivals' possible reactions.  And if intention, obligation, and possibility are so baked into human language they form their own lexical category, then perhaps we game designers should clarify, "why not modals?"