THE ARCHITECTURE OF THE IMAGINARY: MAP, DIRECTRIX, ACTION, COMPLETION, AND REAL-GEOMETRIC ENCODINGS IN THE FULL STRUCTURAL ATLAS OF THE COMPLEX PLANE

This paper develops a formal distinction between \emph{map}, \emph{directrix}, \emph{action}, and \emph{completion} in the setting of the complex numbers. The motivating claim is that charting the place of the imaginary numbers within the complex plane is not the same act as sending a selected aspect of that structure through a projection, restriction, quotient, normalization, branch choice, or flow, and that neither of these is identical with extending a previously partial evaluative regime by a clause that forces a value at sites of nonconstruction. The former is representational; the second is operational; the third is semantic completion. To make this precise, the paper constructs a layered atlas of the complex plane \(\C\) and of the pure imaginary axis \(\ii\R\subset \C\), expanding the canonical ``subrooms'' of one-complex-dimensional space: algebraic, linear, Euclidean, polar, topological, branch-theoretic, holomorphic, harmonic, symplectic, oscillatory, dynamical, and projective. Central structural results include the classification of real-linear endomorphisms as maps of the form \(z\mapsto az+b\conj{z}\), the topological obstruction to a global argument, the principal logarithm on the slit plane, the identification \(T_1S^1=\ii\R\), the Lagrangian character of the real and imaginary axes, Hamiltonian realization of phase rotation, explicit dynamical classifications for power and affine maps, and stereographic/projective completion by the Riemann sphere. The paper then adds a second layer not present in standard expositions: a distinction between \emph{intrinsic zero} and \emph{completion zero}. The element \(0\in\C\) is intrinsically part of the field, but many appearances of the glyph ``\(0\)'' in mathematical practice arise from what may be called \emph{completion-by-clause}: extending a sparse coefficient family, an out-of-support recurrence, an empty aggregation, or an exceptional locus by an explicit rule assigning the value \(0\). Such completions are shown to be neutral only in zero-erasing contexts, especially additive aggregation. A depadding theorem for additive zero-extension is proved, and a complementary non-neutrality proposition shows why zero-completion fails to be harmless once reciprocals, logarithms, arguments, branch-sensitive operations, support-sensitive operators, or analytic continuation enter the scene. A third layer addresses representational pluralism. For the classes of constructions treated here, the mechanics of the complex plane need not be written with the literal symbol \(\ii\): they can be expressed exactly in real two-dimensional terms via a quarter-turn operator \(J\), via winding-number differential forms, via harmonic pairs, and more generally through alternate real-geometric atlases. This includes, at the level of principle, sweep-based encodings such as oriented boundary sweeps, sweeping nets, and saddle-map formalisms used to track phase or singularity structure without explicit imaginary notation. The paper argues that imaginary numbers are therefore not the only possible ontology of planar phase, but the most algebraically compressed one. An appendix provides a formal novelty-map table for the present manuscript and for several adjacent manuscripts supplied by the author, classifying their main components as \emph{likely new}, \emph{probably known}, or \emph{needs citation check}. The aim is not to settle priority definitively, but to separate classical mathematics from original synthesis, and theorem-level originality from interpretive or programmatic innovation.