When a ludic intrudes into experience
In my work on ludics, I consider interactions as signs at a finer scale than game mechanics (Seraphine, 2025). I call a ludic a xenoludic when it intrudes into the experience designed by the game designer. The prefix “xeno” does not simply mean that something sits outside the program: the criterion is what enters the game experience.
An emergent mechanic or ludophrase is not, for that reason alone, a xenoludic: these phenomena are not at the more fundamental level of ludics. Being unforeseen does not define the xeno. I can also welcome a xenoludic deliberately, anticipate its arrival, and place boundaries around it without removing its xeno quality. In my formulation, an internal ludic is fully understood; a xenoludic can be tamed or directed.
Three game design situations
Imagine a bug affecting a category of processors that doubles the strength of one particular enemy’s attacks. This version of the game then has a different legisign: I call it a xenolegisign. It intrudes into experience, and its effects may produce a different game mechanic. That resulting mechanic sits at another level than the xenoludic. The xeno quality can also occur across the whole taxonomy of ludics; it is not limited to legisigns.
Multiplayer voice chat is less simple to isolate. It can bring encouraging or coercive dynamics into play that would be absent without communication among players. A threat of doxxing or swatting, for example, changes how someone plays. Coercion itself may be a ludic: the analysis need not be reduced to one spoken sentence.
Another possibility is to let each player use an LLM of their choice within a predetermined experience, trusting that the whole arrangement will work. That choice would introduce a myriad of xenoludics. Even a model trained for the game may sit at the frontier of the xeno while it remains a black box to the designer. This case returns us to what the designer understands about the ludics entering the experience.
Aesthetics and the interpretant
I do not define xenoludics as negative. When a designer does not account for them, they tend to become problematic; their value nevertheless depends on the interpretant. A speedrunner might exploit the xenolegisign found on a particular machine and find that it improves their experience. I can also imagine an MMO in which users bring their own agents: the variety of non-player characters might grow at lower cost to the designer, with tradeoffs and safeguards. The outcome could be positive for the intended experience.
The designer’s judgment should remain guided by the intended aesthetic. To borrow Marie Kondo’s image loosely, I ask what gives an experience the quality I seek and what draws it away from that quality. That quality may involve different emotions or ways of relating to a game, as I explore in the Ludic Framework (Seraphine, 2016). Welcoming a xenoludic calls for aesthetic judgment as well as a technical decision.
Testing possible experiences
In my approach, a xenoludic always brings some nondeterminism to the experience. Once I introduce xenoludics, probabilistic automated testing becomes inevitable. I need to test possible experiences in light of the aesthetic intention guiding the design.
During QA for a game with chat, I could have custom LLMs and human testers play out positive and negative interactions, then collect questionnaires. I could also deliberately introduce bugs across thousands of automated test runs to challenge a system that corrects problems at runtime. These protocols remain to be developed. I cannot think of every possibility; I can, however, train systems to prepare for them, without a guarantee that they will always respond as I intend.