Geometric Realization of Completed Source Relations Descent, Orbit Spaces, Invariant Relations, and Variational Response in Shadow Theory
Geometric realization of the minimal completion
Geometric realization of the minimal completion
The geometric evaluation map
Definition 5.1 (Declared readout and relation family).
Let be a declared Tier-1 readout: a map, defined on the reduced source, recording exactly the data retained by the Tier-1 description of the model. Let be an index set and let be invariant geometric relations in the sense of Definition 4.1. Write and
Definition 5.2 (Geometrically realized completion).
The geometrically realized completion of relative to is with corestriction and forgetful map so that by construction.
Note that is the image, not the ambient product. This restriction performs the essential mathematical work: it enforces all the compatibility relations among metric, bundle, connection, matter, boundary, and invariant data that Section 3 exhibited. The Cartesian product of independently written coordinate values is in general not realizable (Corollary 3.7, Remark 3.8).
The restricted realization theorem
Theorem 5.3 (Restricted geometric realization).
Fix an admissibility class , a readout , and a family of invariant geometric relations on . Then:
(Realizability.) For every there exists with .
(Factorization.) and for every .
(Terminality.) If is any surjection through which and every factor, say and , then there is a unique map with ; is surjective and satisfies and .
Thus is the canonical minimal completion of relative to , in the sense of Paper 3 [18], computed on the geometric source .
Proof. (1) is immediate: is by definition the image of . (2) holds by definition of and of the projections; well-definedness of and each as maps on is exactly the invariance hypothesis of Definition 4.1, which is where gauge and diffeomorphism invariance are consumed.
(3) Define for any ; the fibre is nonempty since is surjective. This is well defined: if then and likewise for every ; hence , i.e. . By construction , and since is surjective onto , so is . Uniqueness: any with agrees with on , hence everywhere. The compatibilities follow by evaluating on and using surjectivity of .
This is the terminal universal property established in Paper 3 [18], applied to the source set , the readout , and the relation family ; the additional content here is that is the reduced geometric source and that the are invariant geometric relations, so that all maps involved are functions of physical orbits rather than of gauge or coordinate representatives. ◻
Remark 29 (Relative, not absolute).
Theorem 5.3 is restricted in two independent senses, both essential. First, it realizes the completion only relative to the declared family : a different family gives a different , generally neither finer nor coarser. Second, it realizes only what geometry attains: is an image, and Corollary 3.7 shows that this image can be a proper subset of the abstract product. Neither the theorem nor anything else in this paper asserts that an arbitrarily nominated abstract completion of Paper 3 is geometrically realizable. It is realizable exactly when it is isomorphic, over , to for some admissible , readout, and invariant relation family, and establishing that in a given model is work.
When does the completion reconstruct the source?
Proposition 5.5 (Orbit reconstruction criterion).
The completion map is injective, equivalently, bijective onto , if and only if the family separates points of : for all , equality of the readouts and of every forces . In that case as sets over .
Proof. By construction if and only if the readouts agree and every agrees. Injectivity of is therefore literally the separation property. Surjectivity onto always holds, by Theorem 5.3(1). The identification is over because . ◻
Proposition 5.5 converts a question about completions into a question of invariant theory: does the declared family separate the orbits? No general answer exists, and none is asserted (Remark 4.6). Separation must be proved on the restricted domain relevant to the model. Section 6 carries out exactly such a proof in a case where the answer is affirmative and complete.