The first note, preservation of type, asks whether the developed form remains recognizably the same kind of thing as the original. A software system whose architecture has been steadily altered by AI-generated additions may still execute its functions, but if the pattern of dependencies, the structure of abstractions, and the conceptual backbone have been quietly replaced, the type has not been preserved.
The second note, continuity of principles, asks whether the fundamental logic that governs the idea remains operative. Each AI addition may respect local conventions while collectively undermining the architectural principles that gave the system coherence. The principles are still nominally endorsed; they are no longer operative.
Power of assimilation asks whether the idea can absorb new material from its environment without losing its character — the mark of a living tradition as opposed to either a dead one or a dissolving one. Logical sequence asks whether the development follows from the original by a recognizable chain of reasoning, even if the chain is not formally deductive. Anticipation of the future asks whether earlier stages contain hints or foreshadowings of later developments — a test that is retrospective but not therefore arbitrary.
Conservative action upon the past asks whether development preserves and deepens earlier expressions of the idea rather than repudiating them. Chronic vigour asks whether the developed form is more alive, more active, more capable of engaging with its environment than a merely stagnant version would be.
Application of the notes requires judgment, not calculation. They are not a checklist. They are the kinds of questions that a formed steward of a tradition asks — and that a tradition without formed stewards cannot answer for itself. In the AI age, the formation of such stewards becomes the decisive institutional question.
The notes were developed across the last chapters of the Essay on the Development of Christian Doctrine. Newman's motivation was polemical as well as philosophical: he needed to show, to himself and his readers, that specific Catholic doctrines not explicit in the first centuries could nonetheless be legitimate developments of the original faith.
The notes have been adopted, adapted, and critiqued across disciplines. Theologians have refined them; historians of ideas have applied them to non-religious traditions; in 2024–2025, a small but growing literature in software architecture has taken them up as a framework for evaluating the integrity of AI-assisted codebases.
The notes are seven, not arbitrarily many. Each names a different dimension along which a living tradition can remain whole or come apart.
They operate together, not individually. A development that satisfies some notes and fails others may be partly legitimate and partly corrupting; the judgment of the whole is the aim.
They require trained stewards. Their application is not algorithmic; it depends on the illative sense of someone formed in the tradition.
They are portable across domains. Theology, biology, software, and curriculum have all proven responsive to the framework, with appropriate adjustments.
They expose corruption that surface metrics miss. Individual additions can pass every local test while the cumulative result violates multiple notes — the characteristic AI-era failure mode.
Whether the notes are exhaustive, or whether additional criteria are needed for domains Newman did not anticipate, is a live question. Scholars working on AI applications have proposed additional notes — for example, transparency of provenance — that speak to concerns specific to machine-generated contributions.