II-03 — Implementation Conditions: The Observed States of a System's Coupled Entities

This paper defines Implementation Conditions, the first of four foundational concepts in the Implementation Intelligence programme. An implementation condition is the observed state of an entity that a specified implementation object interacts with, depends on, or acts upon. The paper's purpose is to make possible a description of an implementation's setting that is complete before it is judged — one that records what is there without labelling it as helping or hindering, since whether a given state helps or hinders depends on the object and the outcome, neither of which belongs in the observation itself. Four properties of the concept are established. Conditions are object-scoped: the object determines, through its requirements, which entities are in scope, so the same setting yields different condition sets for different objects. Conditions are evaluatively neutral: a condition is a state, not a barrier or a facilitator, and the same state may obstruct one object while enabling another. Conditions are graded and time-indexed: they are recorded on a five-point gradient (absent, emerging, partial, established, saturated) and dated, so repeated observation produces a trajectory rather than a snapshot. And conditions record entity states but not the relations between entities, which belong to a separate concept, Implementation Architecture. The paper distinguishes conditions from the constructs most often confused with them — determinants, requirements, capacity, and readiness — specifies how conditions are observed and graded, including the requirement for a declared observer position, names the failure modes to which condition observation is prone (schema error, observer-bounded scope, and valence leakage), and states three open questions on which the concept's reliability depends, foremost among them whether independent observers derive comparable condition scopes from the same object.