This appendix collects the mathematical definitions and results used throughout the book. It is self-contained for readers unfamiliar with differential geometry, simplicial complexes, or search algorithms; those with background in these areas may skip directly to the chapters. Full proofs and extended treatments can be found in Geometric Methods in Computational Modeling (Bond, 2026a) and Geometric Reasoning (Bond, 2026c).
Definition A.1 (Topological Manifold). A topological manifold of dimension $n$ is a Hausdorff, second-countable topological space $M$ such that every point $p \in M$ has a neighborhood homeomorphic to an open subset of $\mathbb{R}^n$.
Definition A.2 (Chart). A chart on $M$ is a pair $(U, \varphi)$ where $U \subseteq M$ is open and $\varphi: U \to \mathbb{R}^n$ is a homeomorphism onto its image. The coordinates $x^1, \ldots, x^n$ given by $\varphi$ are local coordinates on $U$.
Definition A.3 (Smooth Structure). A smooth atlas is a collection of charts $\{(U_\alpha, \varphi_\alpha)\}$ covering $M$ such that all transition maps $\varphi_\beta \circ \varphi_\alpha^{-1}$ are smooth ($C^\infty$). A smooth manifold is a topological manifold equipped with a maximal smooth atlas.
Clinical interpretation: A clinical assessment framework (e.g., a triage protocol, a QALY calculator, an ethics committee rubric) is a chart on the clinical manifold. Each framework provides local coordinates for the patient’s clinical-moral state. The transition map between charts is the conversion procedure between frameworks. The smoothness requirement says that small changes in a patient’s state should produce small changes in any well-constructed clinical assessment. (See Chapter 4.)
Definition A.4 (Riemannian Metric). A Riemannian metric on a smooth manifold $M$ is a smooth assignment of a positive-definite inner product $g_p: T_pM \times T_pM \to \mathbb{R}$ to each tangent space $T_pM$.
In local coordinates, $g = g_{ij} \, dx^i \otimes dx^j$, where $g_{ij}(p) = g_p(\partial_i, \partial_j)$.
Definition A.5 (Geodesic). A curve $\gamma: [0,1] \to M$ is a geodesic if it satisfies
$$\frac{d^2 \gamma^k}{dt^2} + \Gamma^k_{ij} \frac{d\gamma^i}{dt} \frac{d\gamma^j}{dt} = 0$$
where $\Gamma^k_{ij}$ are the Christoffel symbols of the Levi-Civita connection.
Definition A.6 (Geodesic Distance). The geodesic distance between $p, q \in M$ is
$$d(p, q) = \inf_\gamma \int_0^1 \sqrt{g_{\gamma(t)}(\dot{\gamma}(t), \dot{\gamma}(t))} \, dt$$
where the infimum is over all smooth curves from $p$ to $q$.
Clinical interpretation: The metric determines which clinical states are “close” — i.e., which transitions have low cost. The clinical geodesic is the minimum-cost path from the patient’s current state to an acceptable outcome. (See Chapters 4 and 6.)
Definition A.7 (Mahalanobis Distance). Given a positive-definite matrix $\Sigma \in \text{SPD}(n)$ and vectors $\mathbf{a}, \mathbf{b} \in \mathbb{R}^n$, the Mahalanobis distance between $\mathbf{a}$ and $\mathbf{b}$ with respect to $\Sigma$ is
$$d_\Sigma(\mathbf{a}, \mathbf{b}) = \sqrt{(\mathbf{a} - \mathbf{b})^\top \Sigma^{-1} (\mathbf{a} - \mathbf{b})}$$
When $\Sigma = I$ (the identity matrix), this reduces to Euclidean distance. When $\Sigma$ captures covariance between dimensions, the Mahalanobis distance weights the contribution of each dimension and its cross-dimensional interactions.
Proposition A.1. The squared Mahalanobis distance $d_\Sigma^2(\mathbf{a}, \mathbf{b})$ is a positive-definite quadratic form. It defines a Riemannian metric on $\mathbb{R}^n$ with constant metric tensor $g_{ij} = (\Sigma^{-1})_{ij}$.
Clinical interpretation: The clinical edge weight (Definition 4.2 in Chapter 4) uses the Mahalanobis distance with the $9 \times 9$ clinical-moral covariance matrix $\Sigma$. This captures the crucial cross-dimensional dependencies: a treatment with uncertain outcomes ($d_1$) is costlier when the trust relationship is fragile ($d_5$), because $\Sigma_{15}$ is non-zero. Euclidean distance would treat these dimensions as independent; the Mahalanobis distance captures their interaction. (See Chapters 4–5.)
Definition A.8 (Abstract Simplicial Complex). An abstract simplicial complex $K$ on a vertex set $V$ is a collection of non-empty subsets of $V$ (called simplices) such that: (a) every singleton $\{v\} \in K$ for $v \in V$; (b) if $\sigma \in K$ and $\tau \subseteq \sigma$ with $\tau \neq \emptyset$, then $\tau \in K$.
A simplex $\sigma$ with $|\sigma| = k + 1$ is a $k$-simplex. Vertices are 0-simplices, edges are 1-simplices, triangles are 2-simplices, and so on.
Definition A.9 (Weighted Simplicial Complex). A weighted simplicial complex is a simplicial complex $K$ together with a weight function $w: K \to \mathbb{R}_{\geq 0}$ that assigns a non-negative cost to each simplex.
Clinical interpretation: The clinical decision complex $\mathcal{C}$ (Definition 4.1) is a weighted simplicial complex. Vertices are clinical states (carrying 9-dimensional attribute vectors). Edges are clinical actions (carrying Mahalanobis + boundary penalty weights). Higher simplices are care bundles — jointly administered interventions that must be evaluated as a unit. (See Chapter 4.)
Definition A.10 ($A^*$ Search). Given a weighted graph $G = (V, E, w)$ with start vertex $v_0$, goal set $G \subseteq V$, and heuristic function $h: V \to \mathbb{R}_{\geq 0}$, $A^*$ search finds a minimum-cost path from $v_0$ to $G$ by expanding vertices in order of $f(n) = g(n) + h(n)$, where:
Definition A.11 (Admissibility). A heuristic $h$ is admissible if $h(n) \leq h^*(n)$ for all $n$, where $h^*(n)$ is the true minimum cost from $n$ to the goal.
Theorem A.1 (Hart, Nilsson, and Raphael 1968). If $h$ is admissible and consistent (i.e., $h(n) \leq w(n, n') + h(n')$ for every edge $(n, n')$), then $A^*$ returns an optimal path.
Definition A.12 ($\varepsilon$-Admissibility). A heuristic $h$ is $\varepsilon$-admissible if $h(n) \leq (1 + \varepsilon) \cdot h^*(n)$ for all $n$. An $\varepsilon$-admissible heuristic guarantees that $A^*$ returns a path costing at most $(1 + \varepsilon)$ times the optimum.
Clinical interpretation: Clinical reasoning is $A^*$ search on the clinical decision complex. The start node is the patient’s presenting condition. The goal region is the set of acceptable clinical-moral outcomes. $g(n)$ is evidence-based medicine (accumulated clinical cost). $h(n)$ is clinical wisdom (the moral-heuristic estimate). Admissibility requires that clinical training calibrates boundary penalties as lower bounds on true moral cost. The $\varepsilon$-admissibility spectrum classifies clinical heuristics into four categories: strictly admissible, $\varepsilon$-admissible, inadmissible, and gauge-variant. (See Chapters 6–7.)
Definition A.13 (Clinical Edge Weight). The weight of a clinical action $(v_i, v_j)$ on the clinical decision complex $\mathcal{C}$ is:
$$w(v_i, v_j) = \underbrace{\Delta \mathbf{a}^\top \Sigma^{-1} \Delta \mathbf{a}}_{\text{Mahalanobis distance}} + \underbrace{\sum_k \beta_k \cdot \mathbf{1}[\text{boundary } k \text{ crossed}]}_{\text{boundary penalties}}$$
where $\Delta \mathbf{a} = \mathbf{a}(v_j) - \mathbf{a}(v_i)$ is the change in the 9-dimensional attribute vector, $\Sigma$ is the $9 \times 9$ clinical-moral covariance matrix, and $\beta_k$ is the penalty for crossing clinical-moral boundary $k$.
Remark. The edge weight has two components. The Mahalanobis component measures the distance traveled on the manifold — how much the patient’s clinical-moral state changes. The boundary component measures the cost of crossing ethical thresholds — the discontinuous jumps in moral cost at professional norms (harm, consent, futility, abandonment, sacred values). The total weight is the sum of continuous and discontinuous costs. (See Chapters 4–5.)
Definition A.14 (Gauge Transformation). A gauge transformation $\tau$ is a meaning-preserving re-description of the clinical situation — a change in how information is framed, presented, or described that does not alter the underlying clinical-moral state.
Definition A.15 (Gauge Invariance). A clinical evaluation $E$ is gauge-invariant if $E(\tau(x)) = E(x)$ for all meaning-preserving transformations $\tau$.
Definition A.16 (Bond Invariance Principle). An ethical evaluation is valid only if it is gauge-invariant: the evaluation must depend on the clinical-moral state, not on its description.
Clinical interpretation: Informed consent is a gauge-invariance condition: the patient’s clinical decision must be invariant under meaning-preserving reframing. If “90% survival rate” and “10% mortality rate” produce different treatment decisions, the patient’s heuristic is gauge-variant, and consent is not genuinely informed regardless of what the form says. Algorithmic bias in medical AI is also a gauge-invariance violation: if a clinical decision support system produces different recommendations for the same clinical state described with different demographic labels, the system violates the medical BIP. (See Chapter 8.)
Theorem A.2 (Scalar Irrecoverability; Bond 2026c). Let $Q: \mathbb{R}^n \to \mathbb{R}$ be any function mapping an $n$-dimensional state to a scalar ($n > 1$). Then:
Proof sketch. By the rank-nullity theorem, the kernel of $dQ$ has dimension $\geq n - 1$ at every regular point. Information in the kernel is irrecoverable by the data processing inequality. Part (3) follows by constructing specific examples where the scalar-optimal and manifold-optimal paths diverge arbitrarily.
Clinical interpretation: This is the QALY Irrecoverability Theorem (Theorem 5.1). With $n = 9$, any QALY function destroys at least 8 dimensions of clinical-moral information. The destruction is mathematical, not a failure of implementation. (See Chapter 5.)