Reasonable Authority is an experimental judgement framework designed to explore how discretionary decisions are formed in planning — and how they might be modelled, interrogated, or simulated without collapsing their complexity. It sits at the intersection of legal thresholds, policy interpretation, actor modelling, and explainable AI. Its starting point is the recognition that most planning decisions are not deterministic: they are negotiated, contested, and context-dependent. But they are also not arbitrary. They must be, in principle, reasonable.
The project takes its name from the Wednesbury reasonableness test, established in English public law (Associated Provincial Picture Houses Ltd v Wednesbury Corporation [1948]), which holds that a decision may be quashed if it is so unreasonable that no reasonable authority could have made it. This threshold — while rarely met in practice — remains foundational. It defines a boundary between legitimate discretion and procedural irrationality, and it implicitly shapes how officers, committees, and inspectors frame their judgements.
Reasonable Authority is not an attempt to automate this judgement. Rather, it offers a structured means of tracing how such decisions are constructed. It models how different actors (officers, committees, appellants) weight policies, interpret constraints, and frame planning balances — and how these choices might vary under different procedural or spatial contexts. By simulating the formation of a decision, it becomes possible to expose where discretion is being exercised, what trade-offs are being made, and whether a given line of reasoning sits within the plausible space of institutional justification.
The system utilises agent-based modelling (ABM) to represent decision-making actors as boundedly rational agents with different roles, objectives, and interpretive heuristics. These agents interact with structured policy inputs, site conditions, and procedural rules to generate a spectrum of plausible planning outcomes. This approach allows for the exploration of reasonable disagreement — not as error or inconsistency, but as a constitutive feature of discretionary governance. It also enables sensitivity testing: how variations in constraints, policy emphasis, or institutional culture can lead to different but still legitimate decisions.
Theoretically, Reasonable Authority draws from planning theory (particularly the communicative and legal–institutional schools), public law, and AI ethics — especially debates around explainability and model transparency. It is aligned with a view that planning systems require tools not to resolve discretion, but to reflect it back in ways that are understandable, contestable, and consistent with the duties of public decision-making.
Like the other tools in this suite, Reasonable Authority does not claim neutrality. It is an instrument for exploring what happens when reasoning is made visible — and how the boundaries of discretion shift when exposed to scrutiny, simulation, or structured disagreement.