Almost all games build on some translation points with natural language, whether in general use or with a technical sub-language. Many RPGs borrow from the technical language of theater, film, and/or television to help inform and structure the acts of play. Beyond the system of play, these terms and the ideas that they package encourage play to encounter similar or related ideas and constraints. In this sense, these terms exert a sort of gravity on the events of play, attracting player behaviors towards certain patterns and away from others.

For example, many RPGs incorporate the notion of a scene, a particular theatrical idea about a set of events in a single time and place. Once you become used to breaking down a story into scenes it can be second nature, but it is far from the only way to structure a narrative. Scenes inhibit action which occurs across sizable distances or in fits and stops. They can easily encourage the notion that nothing interesting could be happening fictionally until it become part of a scene, and that if a scene happens it must be consequential to the story. These conceits are sensible from a theatrical perspective, but distinctly limit the scope of story to fit the constraints of theatrical expectations, which are by no means limits within an RPG.

It is impossible to avoid the gravity of terms, even outside of spoken language. Players typically expect to roll dice, and will do so even when uncalled for by the game. People respond to acted body-language (whether another’s or even their own) with tinges of real emotion. Expectations arise simply by calling a game an RPG. Design shouldn’t try to avoid this gravity, but understand and when necessary accept its warping effect on play. As long as this gravity serves most often to draw play towards your design goals it is a benefit, but care must be taken to make that happen.


Games which are built from multiple interacting sub-languages can present an experience of delving into the meanings of fictional events and player decisions or they can constrain the complexities of a variety of coupled situations. Before even creating the sub-languages and the translation points used to go between them, it is important to consider how players will recognize each sub-language. Each language within the game relies on a context. This context is the domain of play which it covers, and when that context arises, it serves as a learned signal to players to venture into that sub-language.

Deciding how to break down the experiences of play and in so doing how to setup the relationships between linked sub-languages. Whether those relationships represent layers of strategy or layers of meaning, the contexts for each sub-language should be clear. These contexts can be fictional experiences, such as gunfighting, seduction, or research. Or the contexts can arise in terms of the player experiences, such as performing or improving dialogue, collaboratively describing a new person, or drawing on a map. As always the root of these decisions are the design goals. One game may want to encourage players to consider the relationship between battle and romance, in another those two experiences can be broken apart to express something entirely different.

Even if the game your designing is meant to express a single unified language of play, you can rarely afford to ignore translation to and from that language. Translation is necessary for players to learn play from text or others and to discuss play. This translation of game elements into known modes of communication and experience is what makes the language of play accessible. It is possible to do without, and either trust in the natural human process of building new languages or even to intentionally wrap and set apart the experience of play.

At its simplest designing for translation means discovering or intentionally putting in points where the objects and actions of play closely relate to terms or dynamics in the other language. These points are the atomic elements of translation, the building blocks in which an experience in one language can begin to be expressed in the other. Both similarity and modeling (see my previous post) provide strong opportunities in building translation points – either in borrowing those used by your influences or in linking to the concepts you are simulating. This means cleaving close to your design goals can often give you all the translation points you need.

Even if translation points are not terribly difficult, it is important to keep them basic. Complex parts of your language of play will not easily translate in most cases. Instead the many moving parts need to be translated through the more basic parts of play to unravel the meaning. This produces short-hand concepts and terms, derived from our experiences comprehending larger parts of play, but these short-hand associations are not translation points.

What I call poetry in one language likely translates into another, but not in a simple and direct way. Calling them both poetry wraps an idea, but it doesn’t express the idea and certainly does not give me the tools to translate. For the same reason, “elevator pitches” – the short blurb describing a game experience, are heavily dependent on hard won understanding of myriad cultural and gaming fore bearers. They can be good targeted marketing, but they do instill understanding.

When designing languages, whether constructed “natural” languages (conlangs) or formal computation-style languages you should nail down some overarching goals. While in game design as in language design you can simply design ad hoc and see where it takes you, that approach is generally best for practicing design or exploring raw design space rather than utility. A conlang, a new programming language, or an RPG all present at least two very distinct approaches: similarity and modeling.

The similarity approach to designing a language explicitly starts with one or more direct influences from existing languages and (usually) a plan for a few places to innovate or synthesize upon that foundation. For these designs, which in RPGs could be called hacks or heartbreakers depending on context and inclination, this foundation isn’t really laziness. Knowingly or unknowingly familiarity is an important part of these languages. These RPGs and languages are clearly situated within the meta-dialogue of its influences, even if that position doesn’t turn out to be positive or prominent. 

On the other hand, modeling takes something like an imagined society without a sense of self, a logical calculus, or a distributed theatrical performance and tries to capture their dynamics and meaning. While familiar tools are useful for this approach, the overriding goal is to emulate or simulate the system. Fidelity is key.

There is a continuum between similarity and modeling, trading off between the goals of familiarity and fidelity.  Innovation has a place here too, and there is something thrilling in exploring a part of design space without overwhelming ulterior motives. But it is those motives which help turn the speculative exploration into a concrete discovery.