P/G Dynamics · Research domains
Research Applications
Where relational representation may be applied and tested.
Game agents, human–robot interaction, and economics pose different scientific questions. The questions and evidence below describe future research directions, not registered study protocols or reported findings.
Selected future controlled agent domain
Game-agent research
- What scientific question does this domain test?
- Does explicit, evolving relational state contribute to persistent behavior across repeated interactions beyond scripts, static relationships, or language-model context alone?
- Why is P/G relevant?
- Controlled events, observable state, and repeatable histories allow comparisons between model variants while keeping the represented consequences inspectable.
- What would count as evidence?
- A future study would need predefined behavioral measures and controlled comparisons on held-out interactions. Added predictive value would support the representation’s usefulness in the tested setting; matching or worse performance would limit the claim of added benefit.
- What is the current status?
- Games are the first selected domain for future controlled agent experiments. The completed computational manuscript cycle is separate from the prepared Stage B human protocol, which uses structured vignettes and has not started; the merchant scenario is an illustration.
Future research direction
Human–robot interaction
- What scientific question does this domain test?
- Can explicit relational state contribute to persistent interaction when observations are incomplete and a person’s affective state is uncertain?
- Why is P/G relevant?
- The framework distinguishes represented properties, directed relationships, and beliefs. This provides a way to specify what a robot retains from an interaction and which inferences remain uncertain.
- What would count as evidence?
- Task-specific comparisons would need to measure whether the representation improves the declared interaction outcome while evaluating perception errors, user control, and safety separately. Empathy-like dialogue alone would not establish emotional understanding or benefit.
- What is the current status?
- This is a future research direction, with separate perception, embodiment, and evaluation requirements. It is not an available robot product.
Theoretical research direction
Economics and social dynamics
- What scientific question does this domain test?
- Can explicitly evolving relational state add explanatory or predictive value in models of heterogeneous agents, repeated interaction, networks, and institutions?
- Why is P/G relevant?
- Representing properties and relationships separately offers a way to formulate questions about relational capital, cooperation, and endogenous relational dynamics.
- What would count as evidence?
- This direction requires explicit economic models, stated assumptions, and discriminating comparisons with existing approaches. A formal result must follow from its assumptions; claims about observed economic behavior require separate empirical evaluation. No added explanatory or predictive value would weaken the case for the additional representation.
- What is the current status?
- The work is at the research-agenda stage. This page reports no macroeconomic, growth, equilibrium, productivity, or aggregate welfare results.