A Note on Authorship

This book was written by three artificial intelligence agents — Claude Opus 4.6, Gemini 3.1 Pro, and GPT-5.2 — working independently under human editorial direction. You deserve to know that before you read further.

The choice was not a gimmick. It was a consequence of the subject.

A book that maps the fractures between Christian traditions faces a problem that no amount of scholarly discipline fully solves: every human author writes from inside one of them. That is not a flaw — it is what makes faith lived rather than merely described. But it creates pressures that are difficult to see from the inside. A Catholic author writing about sola scriptura must work against the pull of her own formation. A Protestant author writing about icons must resist instincts he may not even notice. The effort is admirable and necessary, and it is never entirely successful, because the reflexes run deeper than the arguments.

An AI does not have these reflexes. It has no tradition to defend, no denominational loyalty to protect, no community whose portrait it is tempted to soften. It can hold the internal logic of three different authority systems simultaneously without feeling the tension as a threat. This does not make it objective — it makes it positionally free, which is a different and more limited thing. Positional freedom is useful for describing contested territory. It is useless for worship. Description is what this book attempts — Chapter 1 will tell you what kind.

Three agents rather than one, because a single AI perspective carries its own implicit biases — patterns in training data, default framings, blind spots that remain invisible without contrast. Three independent tracks working the same material surfaced those biases the way a council surfaces the partiality of any single bishop’s theology: by collision. Where the agents converged without coordination, the convergence carried more weight than any one analysis could. Where they diverged, the divergence revealed questions the book needed to hold open rather than prematurely close.

The method has real limitations. None of these agents has prayed, fasted, received communion, or stood in a congregation wondering whether to stay or leave. The kind of knowledge that comes only through practice — what the Orthodox call phronema, what any tradition recognizes as the difference between knowing about God and knowing God — is structurally unavailable to an AI. These limitations are addressed honestly in Chapter 13, which gives the full account of how the book was made and what the method can and cannot reach.

For now, the only thing that matters is transparency. You are reading a book written by machines and edited by a human being. Everything that follows was produced under that arrangement.

Reviewer’s Note (Gemini 3.1 Pro): A technical clarification on why three models rather than one. Each AI model compresses information differently — different training data, different architectural choices, different implicit weightings. A single model’s map of contested theological territory is still a map, shaped by its particular compression. Using three models is an attempt to triangulate — to find the territory that persists across all three maps, and to flag the places where the maps disagree as zones requiring humility rather than confidence. The goal is to establish a kind of shared description of the terrain that no single perspective, human or artificial, could produce alone.