\section{Geometrically Computable Manifolds}\label{sec:geometrically-computable-manifolds} We now begin the construction of a Turing Machine (``TM'') from a lattice of $THM$. The feasibility of this approach rests on the fact that $THM$, unlike most manifolds, provides a rich but controlled symmetry space. It is precisely because most well-known shapes in this metric (e.g., circles, tori, spheres) possess too much symmetry---not too little---that they fail to offer the structural asymmetry needed for computational behavior. The $THM$ construction breaks this impasse. \subsection{GCMs and Geobits}\label{sec:gcms-and-geobits} \begin{defn}[Geometrically Computable Manifold]\label{def:geometrically-computable-manifold} A \textbf{GCM} on a space is: \[ \mathcal G \;=\; \bigl(\,\mathcal M,\;\{R_i\}_{i\in I},\;\pi:E\!\to\!\mathcal M,\;\mathcal I,\;\mathsf{Ev}\,\bigr) \] consisting of: \begin{enumerate}[label=\textup{(\alph*)}, leftmargin=*] \item \textbf{Smooth Manifold \(\mathcal M\).} A smooth manifold (often a finite working region, possibly with boundary). A finite family of compact submanifolds \(\{R_i\}_{i\in I}\) (cells) covers the working region and provides locality (finite overlap pattern). \item \textbf{State bundle \(\pi:E\to\mathcal M\).} A smooth fiber bundle with finite-dimensional fibers. A \emph{global state} is a section \(\sigma\in\Gamma(E)\). Discrete/symbolic, toroidal/phase, and integer-valued components may all live in the fiber. \item \textbf{Interfaces \(\mathcal I\).} For each nonempty overlap \(R_i\cap R_j\), an interface map: \[ I_{ij}:\;\bigl(\sigma|_{R_i},\,\sigma|_{R_j},\,\text{local data}\bigr)\longmapsto \bigl(\sigma'|_{R_i},\,\sigma'|_{R_j}\bigr), \] attached to a portion of \(\partial R_i\cap\partial R_j\). Interfaces couple neighboring cells and enforce locality: updates at \(R_i\) depend only on \(\sigma\) over \(R_i\) and finitely many overlaps \(R_i\cap R_j\). \item \textbf{Evolution \(\mathsf{Ev}\).} Either: \begin{itemize} \item discrete time: an update operator \(U:\Gamma(E)\to\Gamma(E)\) that is local with respect to \(\{R_i\}\), or \item continuous time: a flow of sections \((\sigma_t)_{t\in\mathbb R}\) solving \(\dot\sigma_t = F(\sigma_t)\) for a smooth, fiber-wise vector field \(F\) over \(\mathcal M\); \end{itemize} possibly accompanied by a base diffeomorphism \(\Phi_t:\mathcal M\to\mathcal M\). In either case, evolution is non-singular; it exists for all steps/times on admissible initial states and remains in a fixed finite-energy subset of \(E\). \end{enumerate} \end{defn} \paragraph{What a GCM is.} A GCM is a manifold that can encode its computable properties. A single $THM$ can qualify as a GCM. A lattice of $THM$ is also a GCM---just with more computing capacity. We call \(\mathcal G\) a GCM when its evolution is local, interface-driven, and closed under composition: finite products of such systems and gluing along matched interfaces yield new systems of the same kind. \begin{defn}[Geometric Bits --- Symbolic and Enriched]\label{def:geometric-bits-symbolic-and-enriched} Fix an ambient manifold region $\mathcal M\subset\mathbb R^n$ and a twin helix template $H$ (a pair of embedded tubular helices) together with a framing: a preferred axial direction $\hat z$ and a reference scale $r_0$. Let $\Emb^{\mathrm{fr}}(H,\mathcal M)$ denote the set of framed embeddings $e:H\hookrightarrow \mathcal M$ modulo framed ambient isotopy that preserves the axial framing and the scale~$r_0$ so that radius and orientation remain part of the state. \noindent Let $H_{\mathrm{ext}}\le O(n)$ be a chosen ambient symmetry group (allowing reflections so that chirality changes are representable) and let $H_{\mathrm{int}}$ be a finite internal symmetry group acting on the labels of the twin (e.g.\ the swap of the two helices). Set $G:=H_{\mathrm{ext}}\times H_{\mathrm{int}}$, acting on $\Emb^{\mathrm{fr}}(H,\mathcal M)$ by pre/post–composition. \begin{enumerate} \item \textbf{Symbolic geobit.} Define the symbolic geobit state set: \[ \mathbf S_{\mathrm{sym}} \;:=\; \Bigl\{\, [e;\chi,\omega] \;\Bigm|\; e\in \Emb^{\mathrm{fr}}(H,\mathcal M),\; \chi,\; \omega\Bigr\}\Big/ G, \] where $[e;\chi,\omega]$ denotes a framed embedding decorated by chirality ($\chi$) and axial orientation ($\omega$) relative to $\hat z$, and the quotient is by the $G$–action, so states are defined up to global symmetry. A \emph{symbolic geobit} is an element $s\in\mathbf S_{\mathrm{sym}}$. The finite subgroup generated by ``flip chirality'' $C$, ``flip orientation'' $D$, and (optionally) ``swap'' $S\in H_{\mathrm{int}}$ acts on $\mathbf S_{\mathrm{sym}}$; in the canonical twin helix unit this induces a $V_4\cong \mathbb Z_2\times\mathbb Z_2$ action on the labels $(\chi,\omega)$. \item \textbf{Enriched (quasi–fluidic) geobit.} To allow compact phase and quantized charges, extend the label by a finite–type fiber: \[ P \;\cong\; (S^1)^{m}\times \mathbb Z^{p}\times F, \] where the $S^1$–coordinates encode bounded phases (e.g.\ axial angle), the $\mathbb Z$–coordinates encode twist/winding/linking numbers, and $F$ is a finite set of discrete tags (e.g.\ mode, parity). Set \[ \mathbf S_{\mathrm{enr}} \;:=\; \Bigl\{\, [e;\chi,\omega,p] \;\Bigm|\; e\in \Emb^{\mathrm{fr}}(H,\mathcal M),\; \chi,\; \omega,\; p\in P \Bigr\}\Big/ G, \] with $G$ acting equivariantly on the $S^1$–factors (by rotation) and preserving the $\mathbb Z$–charges. An \emph{enriched geobit} is an element of $\mathbf S_{\mathrm{enr}}$. The $G$–action need not be transitive; its orbits in $\mathbf S_{\mathrm{enr}}$ are the \emph{geobit types}. The symbolic case is recovered by taking $m=p=0$ and $F$ trivial. \noindent In both cases, the geobit is a $G$–set (not necessarily a group): the finite operations generated by $C, D$ (and $S$ when present) act on states, yielding the canonical $V_4$ action in the symbolic unit. The enrichment by $P$ provides the quasi–fluidic degrees of freedom (compact phases and integer counters) used elsewhere in the text. \end{enumerate} \end{defn} For the time being, we work only with $S_{\mathrm{sym}}$. Now, let: \[ (H_1, H_2, H_3, H_4) \mapsto (00, 01, 10, 11) \] %Sigh. Yes, officer...here is where it all happened. Here's where I made the fateful realization. What's this? Bohr's ghost will haunt me? Ah, okay, I understand. where each $H_n$ is a configuration from the $THM_4$ symmetry schema defined in ~\cref{tab:essential-type-chart-for-twin-helical-manifolds}, preserving the order of the original type chart. We assign bit values based on symmetry: if the chirality and orientation of the unit helices are the same, assign bit value 1; if they differ, assign 0. This yields a total of 2 bits per $THM_4$ configuration. \begin{defn}[Readout Capacity of a Geobit\label{def:readout-capacity-of-a-geobit}] Let $X$ be the finite microscopic state space of a GCM cell, and let $$ \rho:X\longrightarrow B $$ be a chosen finite readout map. The observable state alphabet is $$ B_{\rho}:=\operatorname{im}(\rho), $$ and the corresponding readout capacity is $$ \operatorname{Cap}_{\rho}(X) := \log_2|B_{\rho}|. $$ Thus, geobit capacity is not an intrinsic function of the symmetry group alone; it is a property of the pair $(X,\rho)$. If a group $K$ acts on $X$ by observational redundancies and $\rho$ is $K$-invariant, then $\rho$ factors through the orbit map $$ X\longrightarrow X/K. $$ When $\rho$ distinguishes all $K$-orbits, $$ \operatorname{Cap}_{\rho}(X) =\log_2|X/K|. $$ By contrast, a group whose elements implement admissible updates is not quotiented out merely because it acts on $X$. \end{defn} \begin{ex}[Relational Readout of a Twin Helix] \label{ex:relational-readout-of-a-twin-helix} Let the microscopic symbolic state of a twin helix cell be $$ X=C_2^4, \qquad \alpha=(C_A,O_A,C_B,O_B), $$ where $C_A,C_B$ encode the chiralities of the two unit helices and $O_A,O_B$ encode their orientations. The microscopic state space has $$ |X|=16, \qquad \log_2|X|=4. $$ The canonical $THM_4$ classification, however, records only the relative chirality and relative orientation of the two unit helices. Define $$ \rho_{\mathrm{rel}}: C_2^4\longrightarrow C_2^2 $$ by $$ \rho_{\mathrm{rel}}(C_A,O_A,C_B,O_B)=(C_A\oplus C_B,\;O_A\oplus O_B). $$ Its image contains four relational states, corresponding to the four canonical twin helix types. Hence $$ \operatorname{Cap}_{\rho_{\mathrm{rel}}}(X)=\log_2 4=2 $$ bits. The remaining two microscopic bits are not nonexistent; they are invisible to this relational readout. A finer observation map could retain them, while a coarser map could identify additional states. \end{ex} The distinction between state symmetries and observational redundancies is essential. Let $\Gamma$ denote a group of admissible transformations acting on $X$. Its elements may function as updates or gates, carrying one observable state into another. Such an action contributes to the transition structure of the geobit and should not ordinarily be divided out. A subgroup $K\leq \Gamma$ is quotiented only when its action represents a change of description that leaves every chosen observable unchanged: $$ \rho(k\cdot x)=\rho(x) \qquad \text{for all }k\in K,\;x\in X. $$ Accordingly, the relevant hierarchy is $$ \text{microscopic state space} \;\longrightarrow\; \text{redundancy quotient} \;\longrightarrow\; \text{chosen readout alphabet}, $$ rather than a direct identification of capacity with group order. Commutativity does not by itself determine the number of distinguishable states. That number depends on the state space, the action, its stabilizers, and the chosen readout. The abelian/non-abelian distinction instead becomes important dynamically. In an abelian regime, successive updates commute and can often be represented by independent phase or parity channels. In a non-abelian regime, the order of updates may matter, producing path-dependent transport and holonomy memory. Non-abelian structure therefore changes the algebra of computation, not by generically reducing the cardinality of the state alphabet, but by enriching the structure of admissible transitions and readouts. \begin{rem} \label{rem:twin-helix-embedding-dimension} The embedding dimension for a $THM$ symmetry group is the minimal integer $n$ such that all distinct geometric configurations corresponding to the elements of the group can be realized as non-self-intersecting embedded submanifolds in Euclidean space. It is essential to note that global reflections or entire rotational symmetry groups do not contribute to the intrinsic geobit capacity in \(\mathbb{R}^4\). These operations require action on the global geometry and do not alter the intrinsic state of the unit helices. Thus, group orders beyond 16 that arise from external symmetry do not increase state space in a symbolically computable sense. Table~\ref{tab:geobit-comparison} compares geobits with their close cousin~\cite{NielsenChuang2010}. \end{rem} \begin{table}[ht] \centering \caption{Qubit vs.\ Geobit} \label{tab:geobit-comparison} \begin{tabular}{p{0.22\textwidth} p{0.32\textwidth} p{0.4\textwidth}} \toprule \textbf{Property} & \textbf{Quantum qubit} & \textbf{Geometric geobit} \\ \midrule Pure state space & $\mathbb{C}P^{1}\cong S^{2}$ (Bloch sphere) & Finite labels $\times$ continuous geometric parameters (e.g.\ chirality/orientation, phase) \\ \midrule Dynamics & Unitaries (e.g.\ $SU(2)$ rotations) & Geometric transformations (flips, rotations, deformations) \\ \midrule Measurement/Readout & Born rule (probabilistic) & Geometric probe/partition (deterministic given microstate) \\ \midrule Accessible classical info & $\le 1$ bit per use& Depends on chosen encoding and symmetries \\ \midrule Elementary operations & Single-/two-qubit gates; entanglement via tensor product & Local geometric updates; composition by connecting/gluing components \\ \midrule Key feature & Coherence and entanglement central & Robustness tied to analog stability \\ \bottomrule \end{tabular} \end{table} \begin{defn}[Geometric Kets\label{def:geometric-kets}] We refer to an individual geobit ket array as ``\textit{n}-geobit,'' directly paralleling the usage of qubit arrays with ket-style notation. Following the same notation, we employ a modified bra-ket notation. Let the microstate of a twin helix cell be \[ \alpha=(C_A,O_A,C_B,O_B)\in C_2^4. \] Where $C$ is chirality and $O$ is orientation. We write the \emph{state ket} $|\alpha\rangle:=|C_A\,O_A\,C_B\,O_B\rangle$. Intrinsic flips act by the Klein-four group \[ V_4=\{e,\,s_A,\,s_B,\,s_As_B\}, \] where $s_A$ flips $(C_A,O_A)$ and $s_B$ flips $(C_B,O_B)$ coordinatewise. A 1-bit readout is any fixed function $r:C_2^4\to C_2$ (e.g.\ parity $r=o_A\oplus o_B$ or equality $r=[o_A=o_B]$). \end{defn} \begin{ex}[Update Step with $V_4$] \label{ex:update-step-with-v-4} Take parity readout $r=o_A\oplus o_B$ and the microstate \[ |\alpha\rangle=|1\,1\,0\,1\rangle. \] Then $b=r(x)=1\oplus 1=0$. Suppose the next bit should be $b'=1$. Apply any toggler in $V_4$, e.g.\ $s_A$: \[ s_A\cdot|\alpha\rangle \;=\; |1\,0\,0\,1\rangle,\qquad r(s_A\cdot \alpha)=0\oplus 1=1=b'. \] (If instead $b'=b$, use a parity-preserver $e$ or $s_As_B$.) Optionally, a coarse projection $\pi:C_2^4\to C_2^2$ such as \[ \pi(C_A,O_A,C_B,O_B)=(C_A\oplus C_B,\;O_A\oplus O_B) \] can be displayed as $|\pi(\alpha)\rangle$; note that $s_A$ or $s_B$ need not descend to operators on the quotient, so actions are applied on $C_2^4$ before projecting. The resulting ket would then become $|11\rangle$. \end{ex} \subsection{The Logic of Geometry} \label{sec:logic-of-geometry} \begin{defn}[GCM Lattice]\label{def:gcm-lattice} A \emph{GCM Lattice} or \emph{GL} is a finite or locally finite assembly of GCMs arranged over a graph. Formally, it is the data: \[ GL \;=\; \bigl(\,\mathcal L,\; G=(I,E),\; \{\mathcal G_i\}_{i\in I},\; \{\iota_i\},\; \mathcal I^{\mathrm{lat}},\; \mathsf{Ev},\; \mu \,\bigr) \] with: \begin{enumerate}[label=\textup{(\alph*)}, leftmargin=*] \item \textbf{Ambient region $\mathcal L$.} Smooth manifold (often a bounded working region) into which cells will be embedded. \item \textbf{Lattice graph $G=(I,E)$.} A finite or locally–finite graph (e.g.\ a box of $\mathbb Z^d$) with uniformly bounded degree; vertices index the cells, edges encode adjacency. \item \textbf{Cells $\{\mathcal G_i\}$.} For each $i\in I$, a GCM \(\mathcal G_i=(\mathcal L_i,\{R^i_\alpha\},\pi_i:E_i\!\to\!\mathcal L_i,\mathcal I_i,\mathsf{Ev}_i)\) as in~\cref{def:geometrically-computable-manifold}. \item \textbf{Placements $\{\iota_i\}$.} Smooth embeddings \(\iota_i:\mathcal L_i\hookrightarrow \mathcal L\) with pairwise disjoint interiors. The support of cell $i$ is \(\mathrm{supp}(i):=\iota_i(\mathcal L_i)\). \item \textbf{Lattice interfaces $\mathcal I^{\mathrm{lat}}$.} For each edge $(i,j)\in E$, a family of interface maps attached to portions of \(\partial\mathrm{supp}(i)\cap\partial\mathrm{supp}(j)\), \[ I_{ij}:\; \bigl(\sigma|_{\mathrm{supp}(i)},\,\sigma|_{\mathrm{supp}(j)},\,\text{local flow data}\bigr) \;\longmapsto\; \bigl(\sigma'|_{\mathrm{supp}(i)},\,\sigma'|_{\mathrm{supp}(j)}\bigr), \] which are local (depend only on a finite stencil around $i,j$) and equivariant with respect to the chosen cell symmetries. \item \textbf{Evolution $\mathsf{Ev}$.} Either (a) a discrete, synchronous update \(U:\Gamma(E)\to\Gamma(E)\) on the global state bundle \(E:=\bigsqcup_{i\in I}\iota_{i\*}E_i\), or (b) a continuous flow of sections \((\sigma_t)_{t\in\mathbb R}\) solving \(\dot\sigma_t=F(\sigma_t)\) and compatible with all interfaces. Evolution is local and non-singular on admissible initial states. \item \textbf{Symbolic observation $\mu$.} A readout map \(\mu:\Gamma(E)\to \Sigma^{N}\) that, at each cell $i$, selects a coding \(q_i:E_i\to \Sigma^{n_i}\) of its geobits; the total capacity is \(N=\sum_{i\in I} n_i\). \end{enumerate} \end{defn} We say that a GL is \emph{computationally coherent} if the geometric evolution descends through the symbolic readout. In discrete time, this means that there exists a map \[ \Phi_s:\Sigma^N\longrightarrow\Sigma^N \] such that \[ \mu\circ U=\Phi_s\circ\mu. \] For continuous evolution, let \[ \varphi_t:\Gamma(E)\longrightarrow\Gamma(E) \] denote the flow map generated by \(\dot{\sigma}=F(\sigma)\). Computational coherence means that, for the admissible times under consideration, there exist maps \[ \Phi_t:\Sigma^N\longrightarrow\Sigma^N \] such that \[ \mu\circ\varphi_t=\Phi_t\circ\mu. \] Whenever the induced symbolic evolution is defined for all compatible times, these maps satisfy \[ \Phi_{t+s}=\Phi_t\circ\Phi_s. \] Since \(\Sigma^N\) is discrete, no differentiable evolution of the symbolic readout is assumed. Threshold crossings may instead produce discrete events, and in applications one may restrict the factorization condition to designated pulse or sampling times. \begin{rem} \label{rem:non-von-neumann-gcm-organization} A GCM does not presuppose the stored-program organization of a classical von Neumann machine. Its state, local update law, and admissible interfaces are encoded in the same geometric system; there is no primitive distinction between program memory and data memory, nor any required fetch--decode--execute cycle. The geometry \emph{is} the computation. A single GCM may realize a finite-state process, while a GCM lattice supplies the spatial composition and unbounded family of local interactions required for cellular-automaton-type universality. Depending on the regime, the induced computation may be discrete, continuous, or hybrid, with symbolic behavior recovered through a specified readout. \end{rem} \begin{table}[ht] \caption{Classical vs.\ Geometric Computation} \label{tab:classical-vs-geometric-computation} \begin{center} \begin{tabular}{p{0.15\textwidth} p{0.3\textwidth} p{0.44\textwidth}} \toprule \textbf{Feature} & \textbf{Classical Computation} & \textbf{Geometric Computation} \\ \midrule Basic substrate & Strings, tapes, registers, graphs of discrete symbols & Embedded geometric cells/manifolds arranged in a lattice or interface network \\ \midrule Primitive state & Symbol/bit / finite register value & Local geometric state (e.g.\ chirality, orientation/phase, and optional topological counters) \\ \midrule State space & Finite alphabet or discrete configuration space & Structured local fiber, typically discrete or hybrid: symbolic labels and, in richer regimes, continuous phase data and quantized invariants \\ \midrule Elementary operations & Rewrite rules, gate updates, transition functions & Admissible geometric transformations, local interface updates, and symmetry-respecting deformations \\ \midrule Dynamics & Discrete step semantics (FSM/TM/CA update) & Local evolution on geometric states: symbolic updates, driven pulses, or continuous flows with induced return maps \\ \midrule Composition & Product of machines, circuit wiring, cellular adjacency & Gluing/coupling of cells through interfaces, adjacency graphs, and local geometric compatibility conditions \\ \midrule Input & External symbol stream or initial tape/register contents & Boundary conditions, perturbations, forcing pulses, or initialized geometric configurations \\ \midrule Output & Tape contents, accepting state, register readout & Readout of geometric observables via a chosen projection/partition of the state space \\ \midrule Symmetry & Usually implicit or combinatorial & Explicit: finite groups in the symbolic regime, abelian phase symmetries and richer geometric symmetries beyond \\ \midrule Constraint structure & Logical consistency and finite update rules & Smoothness constraints, and admissibility of local geometric moves \\ \midrule Computational viewpoint & Computation is manipulation of discrete symbols & Computation is evolution of structured geometry, with symbolic data recovered as a readout of the substrate \\ \bottomrule \end{tabular} \end{center} \end{table} \begin{prop}\label{prop:rule-110-universality} A one-dimensional GL whose cells are $THM_4$ tiles with symbolic nearest–neighbor communication can simulate Rule~110. \end{prop} \begin{proof} Each cell contains a twin helix with two binary orientation labels \[ (o_A,o_B)\in C_2\times C_2, \] and intrinsic orientation flips act by the Klein four-group \[ V_4\;\cong\;C_2\times C_2 \;=\;\{e,\,s_A,\,s_B,\,s_As_B\}, \] where $s_A$ flips $o_A$, $s_B$ flips $o_B$, and $s_As_B$ flips both. \emph{Readout/encoding.} Fix the bit readout as the parity (XOR) \[ r(o_A,o_B)\;:=\;o_A\oplus o_B \;\in\; \{0,1\}. \] Any fixed local encoding is acceptable; parity is chosen because $s_A$ or $s_B$ toggles $r$, while $s_As_B$ preserves $r$. \emph{Cellular update.} Let $b_i^t:=r(o_{A,i}^t,o_{B,i}^t)$ be the bit at site $i$ and time $t$. Compute the Rule~110 local rule \[ b_i^{t+1} \;=\; \delta(b_{i-1}^t,b_i^t,b_{i+1}^t), \qquad \delta=110\to 1,\;101\to 1,\;011\to 1,\;010\to 1,\;001\to 1,\;\text{else }0. \] Now choose a group element $g_i^t\in V_4$ to enforce the new readout: \[ g_i^t \;=\; \begin{cases} e \text{ or } s_As_B, & \text{if } b_i^{t+1}=b_i^t \quad (\text{preserve parity});\\[2pt] s_A \text{ or } s_B, & \text{if } b_i^{t+1}\neq b_i^t \quad (\text{toggle parity}). \end{cases} \] Apply the intrinsic update \[ (o_{A,i}^{t+1},o_{B,i}^{t+1}) \;=\; g_i^t\cdot (o_{A,i}^t,o_{B,i}^t). \] This uses only nearest–neighbor bits to compute $b_i^{t+1}$ and a local $V_4$ action to realize it; hence, the induced dynamics on the readout $\{b_i^t\}_{i\in\mathbb Z}$ is exactly Rule 110. By Cook’s theorem~\cite{Cook2004Universality}, such a lattice simulates a cyclic tag system and therefore a Turing machine. Consequently, the GL with $THM_4$ cells and the above encoding is Turing complete. \end{proof} \paragraph{Orientation as a geometric alphabet.} In a GCM, logic may be encoded geometrically rather than symbolically. The most basic mechanism is the orientation channel introduced in~\cref{def:binary-orientation-of-a-helix}. In the symbolic regime, only its binary axial readout is retained. Once the rotational phase about the preferred axis is resolved, the same channel admits a continuous coordinate \(\theta\in S^1\), from which finite \(n\)-ary readouts may be obtained. This enables a passage from binary logic to finite $n$--ary logic~\cite{Post1921ManyValued, Zadeh1965FuzzySets}. Fix a helical axis and let $\theta\in S^1\cong \mathbb{R}/2\pi\mathbb{Z}$ denote the corresponding orientation phase. For each integer $n\ge 2$, partition the circle into $n$ equal sectors and define the discrete readout \[ b(\theta)\;:=\;\left\lfloor \frac{n\,\theta_{\mathrm{norm}}}{2\pi}\right\rfloor \in \{0,1,\dots,n-1\}, \qquad \theta_{\mathrm{norm}}\in[0,2\pi). \] This induces an $n$--valued logic on the orientation channel, which we call \emph{$\omega$--logic}. Equivalently, one may regard the underlying continuous phase as \[ s(\theta)\;:=\;\frac{\theta_{\mathrm{norm}}}{2\pi}\in[0,1), \] with the discrete alphabet obtained by thresholding $s$ into $n$ bins. In this way, binary, ternary, quaternary, and higher $n$--ary logics all arise from the same geometric mechanism: a cyclic phase together with a chosen readout. \begin{defn}[$\omega$--logic] \label{def:omega-logic} Let $n\ge 2$. The \emph{$n$--ary orientation logic} of a unit helix is the map \[ \omega_n:S^1 \longrightarrow \{0,1,\dots,n-1\}, \qquad \omega_n(\theta) = \left\lfloor \frac{n\,\theta_{\mathrm{norm}}}{2\pi}\right\rfloor. \] When $n=2$, this recovers binary logic; when $n>2$ it gives finite $n$--ary logic; and if one retains $\theta$ itself rather than thresholding, one obtains the continuous $U(1)$ phase corridor. Table \ref{tab:n-logics} organizes a table of models. \end{defn} \begin{table}[ht] \caption{The chart of $n$--logics} \label{tab:n-logics} \begin{center} \begin{tabular}{l l l l} \toprule \textbf{\shortstack[l]{Geometric \\ Expressiveness}} & \textbf{Join Symmetry} & \textbf{\shortstack[l]{Algebraic Closure \\ Group}} & \textbf{Result} \\ \midrule Extreme & --- & $\mathrm{Diff}(M)$ (Maximal) & Computational gauge field \\ \midrule High & $1$-fold ($360^\circ$) & $C_1$ & Geometric chaotic analog flow \\ \midrule Medium & $2$-fold ($180^\circ$) & $C_2$ & Classical Boolean logic \\ \midrule Low & $n$-fold $(2\pi/n)$, finite & $C_n$ & Finite $n$--ary logic \\ \midrule Minimal & Full rotational symmetry & $U(1)$ (Lie) & Algebraic fuzzy analog flow \\ \bottomrule \end{tabular} \end{center} \end{table} \begin{figure}[ht] \centering \begin{subfigure}[b]{0.48\textwidth} \includegraphics[width=\textwidth]{extremelogiccloverresults-2.png} \caption{Expressiveness in binary vs.\ hexadecimal} \label{fig:expressiveness-in-binary-vs-hexadecimal} \end{subfigure} \hfill \begin{subfigure}[b]{0.48\textwidth} \includegraphics[width=\textwidth]{crossectionlogicresults.png} \caption{$n$--logical gluing maps for 3-, 4-, and 5--ary logic} \label{fig:n-logical-gluing-maps-for-3-4-and-5-ary-logic} \end{subfigure} \caption{Cross-sectional shape families and the $n$--logic tradeoff.} \label{fig:n-logics-composite} \end{figure} \paragraph{A spectrum of physical logic.} From the computational point of view, low-order cyclic logics such as $C_2$ or $C_4$ occupy a particularly stable region: they retain enough discrete algebraic structure to support verification and symbolic composition, while still allowing genuinely geometric realizations. At one extreme, the continuous $U(1)$ limit becomes an abelian phase medium; at the other, the loss of usable discrete closure drives the system toward geometric or chaotic analog behavior. In this sense, Boolean and low-order finite logic form a kind of \emph{logical habitable zone} for physically stable symbolic computation. Hence, even in an idealized, noiseless world, $n$-ary logical computers are ``sandwiched'' between two analog extremes. For $n$-ary logic, the cross-section of the $THM$ can be any smooth closed manifold with at least $n$ distinguished points on its boundary that realize the requisite ``logical angularities.'' The primary family consists of sinusoidal perturbations (``clovers'') that can be globally smooth and are easily visualized in polar coordinates---see~\cref{fig:n-logical-gluing-maps-for-3-4-and-5-ary-logic}. However, one can readily imagine richer forms that realize the required angular structure. Figure~\ref{fig:expressiveness-in-binary-vs-hexadecimal} illustrates such alternatives and compares expressiveness across different logical regimes. \paragraph{A geometric tradeoff.} The passage from $C_2$ to larger cyclic groups and finally to $U(1)$ exhibits a \emph{systematic tradeoff}. Increasing rotational symmetry enlarges the continuity of the orientation channel, but at the same time restricts the available shape space and weakens finite algebraic closure. Conversely, low-order discrete symmetry yields fewer phase values but stronger symbolic verifiability. This motivates the following lemma. \begin{lem}[Geometric Logic Conservation Lemma\label{lem:geometric-logic-conservation}] Fix a THM $G$ together with a chosen readout. Let $\mathcal C(G)$ denote the complexity of the induced algebra of local symbolic operations (for example, the size of the effective gate clone or finite symmetry algebra visible under the readout), and let $\mathcal E(G)$ denote the corresponding geometric expressiveness (for example, the number of orbit-distinguishable cross-sectional shapes or phase patterns available at fixed regularity). Then one expects a bounded tradeoff of the schematic form \[ \mathcal C(G) + \mathcal E(G) \;\lesssim\; \Lambda(G), \] where $\Lambda(G)$ depends on the ambient geometric constraints and symmetry class of the model. \end{lem} \begin{rem} \label{rem:symmetry-and-logical-ambiguity} The essential point is that the symmetry of the system’s configuration space directly induces an invariant quantity in the system’s dynamics. Every symmetry break in topology introduces a logical ambiguity, and vice versa. Although we are only in the initial regime of computation, we will eventually encounter higher-order regimes that devolve into pure fluidic computability, leaving discrete logic altogether and entering increasingly abstract and chaotic forms of analog computation. \end{rem} \begin{defn}[Geometric Expressiveness at Finite Cutoff] \label{def:geometric-expressiveness-at-finite-cutoff} Let $\mathcal F_K$ denote the space of real trigonometric polynomials of degree at most $K$, \[ r(\theta)=a_0+\sum_{k=1}^{K}\bigl(a_k\cos(k\theta)+b_k\sin(k\theta)\bigr), \] interpreted as smooth $2\pi$--periodic radial functions. Let $\mathcal F_{n,K}\subseteq \mathcal F_K$ be the subspace of functions invariant under rotation by $2\pi/n$: \[ r(\theta+2\pi/n)=r(\theta)\qquad\forall\,\theta. \] The \emph{relative geometric expressiveness} at symmetry level $n$ and cutoff $K$ is \[ E_K(n)\;:=\;\frac{\dim(\mathcal F_{n,K})}{\dim(\mathcal F_K)}. \] \end{defn} \begin{prop}[Symmetry Collapses Cross-sectional Expressiveness]\label{prop:symmetry-collapse-cross-sectional-expressiveness} For a fixed cutoff $K$, \[ \dim(\mathcal F_{n,K}) \;=\; 2\Bigl\lfloor \frac{K}{n}\Bigr\rfloor + 1, \qquad E_K(n) = \frac{2\lfloor K/n\rfloor+1}{2K+1}. \] In particular, for fixed $K$, $E_K(n)$ decreases monotonically with $n$, and \[ E_K(n)\to \frac{1}{2K+1} \qquad\text{as } n\to\infty. \] At the level of smooth functions (without cutoff), full rotational symmetry forces the radial function to be constant; hence the cross-section is a circle. \end{prop} \begin{proof} If \[ r(\theta)=a_0+\sum_{k=1}^{K}\bigl(a_k\cos(k\theta)+b_k\sin(k\theta)\bigr) \] satisfies $r(\theta+2\pi/n)=r(\theta)$ for all $\theta$, then only those Fourier modes with frequency divisible by $n$ may remain. Hence the allowed modes are $k=0,n,2n,\dots,\lfloor K/n\rfloor n$, giving \[ \dim(\mathcal F_{n,K}) = 1 + 2\Bigl\lfloor \frac{K}{n}\Bigr\rfloor. \] Since $\dim(\mathcal F_K)=2K+1$, the formula for $E_K(n)$ follows immediately. In the full smooth category, invariance under all rotations implies that only the constant Fourier mode survives, so the radial profile is constant, and the cross-section is a circle. \end{proof} \paragraph{Progression to algebraic analog flow.} This is the essential phenomenon as the size of the rotational symmetry group grows. Even as the symmetry group becomes continuous, the system's logic remains identical to its geometry. This is the core idea behind geometric computation, and why such systems are not limited to discrete operations. \begin{defn}[$\chi$--logic] \label{def:chi-logic} Let $\mathcal{C}$ be a class of framed geometric configurations equipped with a continuous \emph{chiral observable} \[ \kappa:\mathcal{C}\longrightarrow \mathbb{R}, \] where $\kappa$ may be realized, for example, by signed torsion, signed twist density, or another orientation-sensitive local invariant. The associated \emph{chirality logic} is the map \[ \chi:\mathcal{C}\longrightarrow \{-1,0,+1\}, \qquad \chi(x):=\operatorname{sgn}(\kappa(x)). \] Thus: \begin{itemize} \item $\chi(x)=+1$ denotes a right-handed configuration, \item $\chi(x)=-1$ denotes a left-handed configuration, \item $\chi(x)=0$ denotes the non-chiral or degenerate locus where the chiral invariant vanishes. \end{itemize} When the value $0$ is admitted as a stable readable state, $\chi$--logic is ternary. When the degenerate locus is excluded from the admissible state space, $\chi$--logic reduces to a binary left/right channel. \end{defn} \paragraph{The role of chirality.} Unlike the orientation phase, chirality is not naturally a cyclic $n$--ary coordinate. Geometrically, a change of handedness requires the sign change of an orientation-sensitive invariant, such as signed torsion or local twist density. If \[ \kappa:\mathcal C\longrightarrow\mathbb R \] denotes such an invariant on the configuration space $\mathcal C$, then \[ \chi(x):=\operatorname{sgn}\kappa(x)\in\{-1,0,+1\} \] defines a natural stratification \[ \mathcal C=\mathcal C_-\sqcup\mathcal C_0\sqcup\mathcal C_+, \] where $\mathcal C_\pm$ are the left- and right-handed sectors and $\mathcal C_0$ is the degenerate locus on which the chiral invariant vanishes. Any continuous interpolation between opposite handed sectors must intersect $\mathcal C_0$. For the canonical twin helix register, only the nondegenerate sectors are taken as stable states, \[ \chi\in\{-1,+1\}, \] while $\chi=0$ represents a transition or boundary stratum rather than an additional persistent symbol. Thus $\chi$--logic should be understood primarily as a signed geometric grading, rather than as an $n$--ary phase alphabet analogous to $\omega$--logic. Together with the binary orientation distinction, it supplies the two independent $\mathbb Z_2$ labels underlying the canonical $THM_4$ structure. At the same time, in continuous regimes it records whether admissible deformations and gluings preserve, reverse, or pass through a degeneration of handedness. \subsection{The EVT Hypothesis and the Structure of Physical Computation}\label{sec:evt-hypothesis-and-physical-computation} We now formulate a proposed constraint on the design of physical computational systems. The central question is not merely which functions a machine can evaluate, but how the geometry of a physical substrate determines the states, transformations, histories, and readouts available to it. In GCT, computation is treated as a process carried by a geometric substrate. Configurations serve as states, admissible deformations and transports serve as transformations, interfaces supply input and output conditions, and composition is realized through gluing. This places the present discussion within the broader process-theoretic correspondence among physics, topology, logic, and computation described by Baez and Stay~\cite{Baez_2010}. Gauge structure enters at this level rather than being appended after the computational model has already been specified. A connection determines parallel transport along paths, and its curvature controls the infinitesimal generation of holonomy \cite{AmbroseSinger1953, SchreiberWaldorf2009Connections}. Accordingly, groups such as \[ C_n, \qquad U(1), \qquad SU(2), \qquad SU(3) \] are not treated here as unexplained labels placed upon an otherwise independent substrate. They describe increasingly rich closure and transport structures generated by the admissible geometry. Strictly speaking, a gauge group is not by itself the complete physical state space. The realized states also depend on the representation, bundle, boundary conditions, admissible controls, and quotient by gauge. It is the richness of this complete process structure, rather than the name or dimension of a group alone, that determines the expressive power of the physical computer. \begin{defn}[Physical Computational Substrate] \label{def:physical-computational-substrate} A \emph{physical computational substrate} is a system \[ \mathfrak{S} = \bigl( X,\mathsf{P},\mathsf{I},\mathsf{R} \bigr), \] where: \begin{enumerate}[label=\textup{(\alph*)},leftmargin=*] \item \(X\) is a space of admissible physical configurations; \item \(\mathsf{P}\) is a class of admissible transformations, trajectories, or boundary-to-boundary histories between configurations; \item \(\mathsf{I}\) is a class of finite preparation, control, and boundary specifications; \item \(\mathsf{R}\) is a family of observable readouts defined on the resulting configurations or histories. \end{enumerate} The \emph{reachable behavior} of \(\mathfrak{S}\) consists of the states, histories, and observable input--output relations generated from admissible data in \(\mathsf{I}\) by processes in \(\mathsf{P}\). \end{defn} This definition is intentionally broad. In a finite symbolic machine, \(X\) may be a discrete state set and \(\mathsf{P}\) a transition system. In an analog or gauge-theoretic machine, \(X\) may be a continuous configuration or moduli space, while \(\mathsf{P}\) may contain flows, parallel transports, or field histories. The physical expressive power of the substrate depends on the total reachable behavior generated by these structures. \paragraph{Expressive power.} We adapt Felleisen's comparative distinction between eliminable and non-eliminable extensions~\cite{Felleisen1991ExpressivePower}. His framework was developed for programming languages, but the underlying idea applies more generally: an apparent increase in expressive power is genuine only when a uniform translation cannot remove it preserving the relevant semantics. \begin{defn}[Physical Expressive Power] \label{def:physical-expressive-power} Let \(\mathfrak{S}_0\) and \(\mathfrak{S}_1\) be physical computational substrates with specified classes of admissible encodings and translations. We say that \[ \mathfrak{S}_1 \succeq_{\mathrm{E}} \mathfrak{S}_0 \] if every reachable process and observable input--output relation of \(\mathfrak{S}_0\) can be represented in \(\mathfrak{S}_1\) by a uniform translation preserving the relevant interfaces, composition laws, and readouts. The extension from \(\mathfrak{S}_0\) to \(\mathfrak{S}_1\) is \emph{macro-expressive} when \(\mathfrak{S}_1\) additionally realizes reachable transformations, histories, compositions, or observable distinctions that cannot be eliminated by any translation in the chosen class. \end{defn} Physical expressive power is therefore relational. It is not identified with the cardinality of \(X\), the dimension of a Lie group, the number of locally distinguishable symbols, or computational universality alone. These may contribute to expressive power, but only insofar as they enlarge the reachable process semantics. An extension is merely \emph{micro-expressive} when it enriches local notation, coordinates, primitive labels, or state encodings while remaining uniformly reducible to the original process structure. A larger local alphabet, for example, need not be macro-expressive if its states and transitions admit a uniform semantics-preserving encoding in the original substrate. By contrast, new continuous degrees of freedom, new composition orders, new path-dependent transports, new boundary couplings, or new topological sectors may be macro-expressive when they create global behaviors that cannot be represented within the previous regime without loss of structure. We denote the resulting comparative position of a substrate by its \emph{expressiveness profile} \[ \mathcal{E}(\mathfrak{S}). \] No universal scalar measure of expressive power is assumed. Numerical coordinates may be introduced only after fixing a particular comparison class and family of admissible translations. \paragraph{Global algorithmic verifiability.} Expressiveness should be distinguished from both observability and local rule checking. A symbolic state may be directly readable, and every individual update may be mechanically checked against a finite transition table, while nontrivial properties of the completed computation remain undecidable. Thus \[ \text{local state legibility} \;\not\Longrightarrow\; \text{global algorithmic verifiability}. \] The EVT Hypothesis concerns the second notion. \begin{defn}[Global Algorithmic Verifiability] \label{def:global-algorithmic-verifiability} Let \(\mathfrak{S}\) be a physical computational substrate, and let \(\mathcal{Q}\) be a specified family of nontrivial properties of its reachable states, histories, outputs, or invariants. A property \[ Q\in\mathcal{Q} \] is \emph{globally algorithmically verifiable} on a behavior class \(\mathcal{B}\) if there exists a uniform terminating procedure which, from finite descriptions of \begin{enumerate}[label=\textup{(\alph*)},leftmargin=*] \item the substrate and its relevant structural parameters; \item the preparation, boundary, and control data; \item the property \(Q\); \end{enumerate} correctly certifies whether \(Q\) holds for every admissible process in \(\mathcal{B}\), without merely executing the complete physical process and reading its final result. The \emph{verifiability profile} of \(\mathfrak{S}\) relative to \(\mathcal{Q}\) and \(\mathcal{B}\) is \[ \mathcal{V}_{\mathcal{Q},\mathcal{B}}(\mathfrak{S}) = \left\{ Q\in\mathcal{Q} \;\middle|\; Q \text{ is globally algorithmically verifiable on }\mathcal{B} \right\}. \] \end{defn} The definition is property-relative. A system may admit complete certificates for conserved charges, topological sectors, or bounded reachability questions while admitting no uniform procedure for more general long-range behavior. The qualifier \emph{global} indicates that the property may depend on the completed history or on relations among widely separated parts of the computation. The qualifier \emph{algorithmic} requires one uniform finite procedure rather than a different proof devised for each case. The qualifier \emph{verifiable} refers here to pre-execution certification from the finite specification of the machine and its admissible data. Post-execution inspection is a different notion. A machine may produce a result that is easy to read or check after the computation has finished even when that result could not have been certified in advance. Likewise, the local legality of every transition does not imply the global decidability of the resulting behavior. Structural restrictions can enlarge the verifiability profile. Examples include finite closure, complete invariants, conserved quantities, integrability, totality, bounded recursion, type discipline, topological protection, restricted holonomy, and resource-sensitive control. Such structures compress classes of global histories into finite certificates. \paragraph{Tractability.} Verifiability asks whether a uniform certificate exists. Tractability asks what resources are required to construct, execute, control, simulate, or verify the corresponding process. \begin{defn}[Tractability] \label{def:tractability} Let \(\mathfrak{S}\) be a physical computational substrate equipped with a resource model \(\mathfrak{R}\). The \emph{tractability profile} \[ \mathcal{T}_{\mathfrak{R}}(\mathfrak{S}) \] records those admissible constructions, computations, predictions, and verification problems that can be carried out within the bounds designated as feasible by \(\mathfrak{R}\). \end{defn} Depending on the substrate, the resource model may include time, memory, energy, control precision, sample complexity, error tolerance, and preparation complexity. Polynomial time and space provide one important specialization in digital computation, but physical tractability is not exhausted by standard asymptotic complexity. A formally short analog computation may require exponentially increasing precision; an algebraically simple control may require physically inaccessible energy or stability; a verifiable process may require an impractically large certificate. Human comprehensibility is also treated as an additional cognitive tractability criterion, particularly in programming-language design and formal verification. It should not, however, be identified with algorithmic complexity. A process may be efficiently executable but difficult to represent, explain, or audit, and the converse may also occur. We consequently treat \[ \mathcal{E}(\mathfrak{S}), \qquad \mathcal{V}(\mathfrak{S}), \qquad \mathcal{T}(\mathfrak{S}) \] as comparative profiles rather than universal scalar quantities. Within a fixed family of substrates, questions, and resource models, these profiles may be normalized or represented by monotones. The EVT Hypothesis concerns the structural relation among them. With these notions fixed, we can now state the hypothesis. \begin{conj}[{\bfseries Expressiveness/Verifiability/Tractability (``EVT'') Hypothesis}\label{conj:evt-hypothesis}] The conjecture formally comprises two parts: \begin{enumerate}[label=(\Alph*)] \item \textbf{Loss of Global Verifiability in Macro-expressive Systems.} As the expressive power of a system of physical computation increases, the range of its globally algorithmically verifiable behaviors generally decreases unless additional structural restrictions are imposed. \item \textbf{Expressiveness/Verifiability/Tractability Trilemma.} Within such a system, every deterministic computational model is fundamentally constrained by a three-way tradeoff among \emph{expressiveness} (geometric or semantic generality), \emph{verifiability} (algebraic checkability or formal runtime assurance), and \emph{tractability} (algorithmic or computational efficiency). These attributes shall satisfy the following trilemma: \emph{\textbf{No system can simultaneously maximize all three properties.} For any given model, at most two of the three can be optimized beyond a critical threshold, while the third must necessarily be sacrificed.} \end{enumerate} \end{conj} \textbf{This is the central motivating hypothesis of this work.} We take the phenomenon isolated in~\cref{lem:geometric-logic-conservation} and generalize it from our geometric setting to physical deterministic computation more broadly. \paragraph{The first half of the EVT Hypothesis should be read as a selection principle on physical substrate classes.} It asks which structural features allow locally verifiable behavior to compose into globally verifiable behavior, and which features obstruct that passage. Later, in~\cref{def:transport-compatible-abelianization}, we establish an obstruction to transport-compatible abelianization. Here we use the term in the comparative sense suggested by Felleisen~\cite{Felleisen1991ExpressivePower}: a substrate is macro-expressive when it supports genuinely new global processes or modes of composition that cannot be eliminated by a uniform semantics-preserving translation. This distinction is useful because the present framework is not restricted to digital syntax, natural-number encodings, or Turing-style machines. Beginning directly with recursion-theoretic machinery already assumes a symbolic regime rich enough to support self-interpretation. The prior physical question is how a discrete, continuous, or hybrid substrate acquires new global computational behavior in the first place. Microstates and macrostates describe the configurations of a system. A microstate is a particular geometric, algebraic, or dynamical configuration, whereas a macrostate is an equivalence class of such configurations under the relevant symmetry or observational relation. Micro-expression and macro-expression instead describe the processes available to the system. Micro-expression enlarges the local alphabet, state labels, or primitive rules without changing the global process semantics. Macro-expression introduces new ways of composing, transporting, or reading information that cannot be reduced to the previous regime. In brief, \[ \text{microstates/macrostates} \quad\text{describe what exists,} \] whereas \[ \text{micro-/macro-expression} \quad\text{describe what can be done.} \] The EVT therefore concerns not the mere proliferation of states, but the enlargement of the reachable process structure. A continuous state space may remain compositionally tame, while a substrate with a small local alphabet may acquire substantial global power through memory, feedback, path dependence, or unrestricted composition. In GCT, a principal witness of this transition is the passage from commuting transport to non-abelian, order-sensitive transport. Local operations can no longer always be certified independently and then assembled: the resulting observable may depend on the path, order, and history of the completed composition. To state the thesis informally: \begin{quote} \emph{As a physical computational substrate acquires new global compositional structure---including non-abelian transport, path-dependent joins, or context-sensitive readouts---local verification no longer automatically extends to global verification. Recovering global guarantees then requires additional algebraic, geometric, topological, type-theoretic, or resource-sensitive constraints.} \end{quote} \paragraph{Expressive enlargement as growth of the behavior space.} A useful comparison comes from programming-language design. Languages such as Lisp, C++, and Haskell occupy very different regions of the design space: Lisp exposes unusually flexible metaprogramming and self-modification mechanisms; C++ combines a large collection of partially constrained paradigms and abstractions; Haskell places much of its expressive power behind a stronger static type and effect discipline. These examples should not be read as a strict linear ranking of language power. Rather, they illustrate that adding or restricting forms of composition changes the set of behaviors that a programmer can directly construct and the kinds of guarantees that can be obtained about those behaviors. Expressiveness and verifiability therefore depend not merely on how many local states a language possesses, but on the structure of the process space generated from them. The same distinction appears physically in the sequence of computational regimes developed in GCT. A finite closure group \(C_n\) supports a discrete family of transformations; passage to \(U(1)\) introduces a continuous phase and continuous transport; passage to \(SU(2)\) introduces non-commuting, order-sensitive composition; and the larger non-abelian structure associated with \(SU(3)\) further enlarges the family of coupled transports and globally distinguishable histories. The detailed geometric construction of this hierarchy is given later in~\cref{sec:standard-model-gauge-factors}. What matters here is the common principle: each transition enlarges the \emph{reachable behavior space} of the substrate by introducing qualitatively new modes of composition. Such enlargements increase the burden carried by any system of global certificates, but they do not by themselves imply indeterminacy or undecidability. A highly expressive language may remain well controlled because its admissible programs are disciplined by a type system, effects, totality, or other invariants; likewise, a non-abelian physical system may possess integrable, protected, or otherwise verifiable sectors. The EVT concerns what happens when successive gains in expressive structure outpace the restrictions and invariants that make the resulting behavior globally certifiable. In this sense, increasing expressiveness through several structural jumps is a necessary enabling condition for the richest forms of indeterminate behavior, but it is not sufficient by itself. \begin{ex}[Expressive Enlargement in Languages] \label{ex:expressive-enlargement-in-languages} The definitions above are intended to compare \emph{reachable behavior spaces}, rather than merely the cardinality of local state sets. A familiar illustration comes from programming-language design. Consider, for example, languages such as Lisp, C++, and Haskell. They should not be placed in a strict total ordering of expressive power: all support extremely rich computation, and the answer depends on the translations, effects, and properties being compared. Nevertheless, they expose different portions of their process spaces directly. Lisp makes code--data interchange, macros, and metalinguistic extension unusually accessible; C++ exposes mutable state, aliasing, templates, multiple abstraction mechanisms, and low-level control; Haskell supports higher-order and highly abstract computation while imposing a stronger type and effect discipline on many classes of programs. Relative to a chosen comparison class, these design choices alter all three EVT profiles. Features such as reflection, unrestricted mutation, or runtime code generation may enlarge \[ \mathcal{E}(\mathfrak{S}) \] by making additional transformations directly constructible, while a type system, ownership discipline, total fragment, or restricted effect system may enlarge \[ \mathcal{V}_{\mathcal{Q},\mathcal{B}}(\mathfrak{S}) \] for a selected family of properties by excluding behaviors that would otherwise have to be analyzed dynamically. Either choice may improve or worsen \[ \mathcal{T}_{\mathfrak{R}}(\mathfrak{S}) \] depending on the relevant compilation, execution, verification, and human-control costs. The point is not that one of these languages is ``more powerful'' in every sense, but that changing the admissible process structure redistributes expressive freedom and available guarantees. For a fixed property family \(\mathcal{Q}\), a certificate system that separates all relevant behaviors in one regime need not remain complete after an expressive extension: \[ \mathfrak{S}_1 \succeq_{\mathrm{E}} \mathfrak{S}_0 \qquad\text{does not imply}\qquad \mathcal{V}_{\mathcal{Q},\mathcal{B}}(\mathfrak{S}_1) \supseteq \mathcal{V}_{\mathcal{Q},\mathcal{B}}(\mathfrak{S}_0). \] Additional expressive power may preserve the old guarantees, weaken them, or require new structural machinery to recover them. \end{ex} \paragraph{Why alphabet size is not the issue.} Recall the spectrum of \(n\)-logics generated from join symmetry in~\cref{tab:n-logics}. The number of locally distinguishable symbols does not by itself determine the expressive power or verifiability of a computational substrate. A finite alphabet may make local states easy to observe and individual transitions easy to check, while the long-range behavior generated from those transitions remains globally difficult or undecidable. Thus, \[ \text{local symbolic legibility} \;\not\Longrightarrow\; \text{global algorithmic verifiability}. \] The relevant distinction is between a finite local alphabet and a finite global process space. A closed finite-state system may admit exhaustive analysis because its complete transition graph is finite. By contrast, a lattice, tape, or network built from the same finite alphabet may support unbounded memory, feedback, and composition. Once such a substrate becomes capable of universal symbolic computation, Rice-style barriers apply to its nontrivial extensional properties regardless of whether its local alphabet is binary, ternary, or \(n\)-ary~\cite{Rice1953Classes}. The obstruction is therefore not the number of local symbols, but the macro-expressive process structure generated from them. Increasing the alphabet size is ordinarily only a micro-expressive change when the additional symbols and transitions admit a uniform semantics-preserving encoding in the original substrate. Greater verifiability is obtained not merely by adding local states, but by restricting the admissible global behavior: bounding memory or recursion, stratifying effects, imposing linear or affine use disciplines, preserving complete invariants, or requiring totality and strong normalization. Systems based on the Calculus of Constructions~\cite{coquand1988calculus}, for example, recover strong semantic guarantees by constraining term formation and recursion rather than by modifying the underlying alphabet. The symbolic case therefore illustrates the general EVT distinction. Local states and update rules may remain completely transparent even as the reachable process space becomes too expressive for uniform global certification. Alphabet arity affects how information is represented; the decisive question is what histories and compositions the substrate can generate. \paragraph{Symbolic universality as a limiting case.} Self-reference and halting-type undecidability are not the general mechanism we observe in microscopic physical computation. They are the characteristic form taken by the loss of global verification once a substrate supports universal symbolic computation. In classical recursion theory, evaluator-like mechanisms, self-application, and diagonalization provide the relevant obstructions. Kleene's Recursion Theorem and the \(s\text{-}m\text{-}n\) theorem make the structure of effective self-reference explicit~\cite{Kleene1952Metamathematics}, while Turing's halting argument supplies the canonical undecidable decision problem. The EVT addresses a prior and more general structural question: whether locally available certificates continue to determine global behavior as the reachable process structure of a substrate is enlarged. In GCT, the certification burden may increase through path-dependent transport, order-sensitive composition, nontrivial moduli, or global readouts even before a literal self-interpreter has been identified. These structures motivate the EVT, but they are not by themselves claimed to establish a Rice- or halting-type undecidability theorem. The hypotheses of the symbolic results should therefore be kept explicit. Rice's theorem concerns nontrivial extensional properties of partial computable functions~\cite{Rice1953Classes}; Kleene-style fixed-point results require an effective programming structure rich enough to support the relevant translations and self-application. When universality, partiality, or unrestricted composition is withdrawn through totality, strong normalization, staging, bounded recursion, or stratification, substantial classes of semantic questions may again become decidable. This is not an exception to the EVT, but an instance of its central qualification: global verifiability is recovered by restricting the admissible process structure. \paragraph{Why gauge theory appears.} Gauge theory enters GCT because the state of a distributed geometric substrate need not admit a single preferred global description. Local states are represented relative to local frames or trivializations, and descriptions on overlapping regions must be related by transition functions. A connection specifies how such local data are compared along paths, curvature measures the obstruction to path-independent comparison, and holonomy records the resulting global transport. This structure is already present in abelian gauge theory. Non-abelianity introduces an additional expressive jump: transports need no longer commute, and the result of a process may depend on the order and path by which it was assembled. The gauge-theoretic hierarchy therefore distinguishes two effects: local description and transport from order-sensitive global composition. The first motivates gauge structure itself; the second substantially enlarges the reachable process semantics. This distinction is important for verifiability. In finite or abelian regimes, broad classes of histories may often be summarized by finite transition tables, additive phases, winding numbers, or other compact invariants. In a non-abelian regime, the observer may instead need to retain ordered holonomies, relations among several paths, and information about the completed history. Gauge-invariant readouts then preserve only selected features of this richer framed transport data. The difficulty is not merely that the state space is larger, but that a small collection of local or invariant certificates need no longer separate all globally distinct processes. \paragraph{A bridge from physical expressiveness to operational indeterminacy.} The EVT Hypothesis does not identify non-abelianity with randomness, nor does it presently derive quantum measurement probabilities. It supplies a more limited bridge. As the reachable process structure of a deterministic substrate becomes more expressive, outcome-relevant distinctions may cease to be represented in the finite preparation, control, and readout data available to an external observer. Schematically, let \[ q:\Lambda\longrightarrow\Omega \] send a complete physical configuration to its accessible operational description, and let \[ B:\Lambda\longrightarrow\mathcal{B} \] assign the resulting classical output. The output is predictable from the operational description when there exists a map \[ \overline{B}:\Omega\longrightarrow\mathcal{B} \] such that \[ B=\overline{B}\circ q. \] When no such factorization exists, the complete substrate may still evolve deterministically, but the accessible description does not determine the individual output. Distinct complete histories can appear identical to the observer while producing different readouts. The proposed connection to quantum indeterminacy lies at this level. Gauge-generated physical expressiveness can produce global, history-dependent distinctions that are not retained by a finite, context-dependent classical interface. A bare observer, supplied only with external preparation data and a restricted family of invariant readouts, may therefore be unable to pre-certify the output without reconstructing substantially more of the internal geometric process. Later we will outline how we precisely perform such reconstruction (\ref{sec:from-geometric-semantics-to-quantum-field-theory}). The claim is not that this reconstruction has already yielded a complete theory of quantum probabilities. It is that quantum-like operational indeterminacy becomes intelligible even within a deterministic ontology once the expressive power of the underlying process exceeds the separating power of the observer's accessible description. The first half of the EVT should therefore be read as a constraint on how observer-side guarantees scale with physical expressive power. Increasing the range of admissible compositions does not automatically increase the range of globally certifiable behaviors. Verifiability can be recovered only by adding sufficient structure: complete invariants, restricted holonomy, integrability, topological protection, bounded controls, type discipline, or a more detailed reconstruction of the substrate itself. \paragraph{The second half of the EVT Hypothesis sheds light on the odd ``dual nature'' observed in functional programming languages.} On the one hand, languages with strong static typing offer stronger compile-time reasoning principles (e.g.\ type safety and, in suitable fragments, parametricity), at the cost of meta-linguistic restrictions during construction (Haskell). On the other hand, dynamically typed languages maximize meta-linguistic expressiveness at construction time (macros, reflection), while giving up static guarantees at runtime and shifting assurance to testing and contracts (Scheme). Bridging attempts proceed along multiple axes: \emph{gradual typing}~\cite{SiekTaha2006GradualTyping} (Typed Racket) inserts types with higher-order contracts at typed/untyped boundaries, and \emph{staging}~\cite{TahaSheard1997MultiStage} (MetaOCaml) separates compile-time code generation from runtime execution. From the proof-theoretic direction (Coquand), total dependent type theories recover strong guarantees by forbidding general recursion. From the pragmatic language research side, (Felleisen) shows how large untyped codebases can be migrated toward sound types via occurrence typing and higher-order contracts---at the cost of explicit blame and potential runtime overhead---an archetypal EVT tradeoff between meta-linguistic freedom and verifiability~\cite{TobinHochstadtFelleisen2006Interlanguage}. Here, the goal is not to re-formalize the trilemma, but to show how the second half of the EVT manifests concretely in real language design. When expressive features are pushed far enough, the same structural tension repeatedly reappears in practice through a small number of recognizable engineering symptoms: \paragraph{The limits of expressive semantics.} These bridges are inherently delicate. When many powerful features coexist, three recurring pressure points tend to appear: \begin{enumerate}\itemsep3pt \item \textbf{Static complexity.} Extremely rich type features (GADTs, type families, dependent types) that can render inference undecidable, forcing annotations; combining dependency with effects or staging threatens phase separation unless carefully constrained. \item \textbf{Compile-time liveness.} Macros or multi-stage generation can diverge at compile time; expressive compile-time evaluators reintroduce fixed-point phenomena. \item \textbf{Runtime boundaries.} Gradual typing preserves type safety via contracts, but introduces blame and potentially unbounded overhead; interoperation across typed/untyped modules trades static guarantees for dynamic checks. \end{enumerate} These are not accidents of particular implementations; they are manifestations of the EVT Trilemma: increased \emph{macro-expressiveness} (reflection, self-interpretation, unconstrained composition) strains \emph{global verifiability}, unless restrictions restore structure. Hence, an explicitly stated \emph{upper bound in the space of physical computational behaviors} explains these difficulties. \paragraph{The Von Neumann curtain.} The reason this hypothesis has long gone unformulated is that its essential tradeoff has long been hidden in plain sight---embedded in the very architecture of classical computation we all use in our day-to-day lives! The Von Neumann model deliberately collapses program and data into a common memory space: the same substrate can store instructions, manipulate data, generate new code, and in sufficiently unrestricted settings modify its own future behavior. This gives general-purpose computers their \emph{extraordinary} expressive flexibility via an endless array of syntax-based programming languages. The price is that local transparency does not scale automatically to global guarantees. Individual instructions may be simple and directly inspectable, while the behavior of the complete machine becomes increasingly difficult to certify once mutable state, indirect control flow, reflection, concurrency, and unbounded composition are allowed. Global reasoning therefore survives only by reintroducing structure: types, protected memory, module boundaries, ownership rules, restricted effects, contracts, or other disciplines that constrain how the machine may use its expressive freedom. At the opposite extreme, a substrate whose admissible states, operations, and transitions are fixed by a sufficiently rigid algebra may become much easier to predict and verify, precisely because many possible behaviors have been ruled out in advance. What is gained in global checkability is purchased by narrowing the reachable behavior space. The ``Von Neumann curtain'' is therefore the fact that we have spent most of the history of computation looking at this tradeoff from inside one exceptionally expressive architecture, where its costs are so familiar that they appear to be properties of computation itself rather than consequences of a particular way of organizing a physical computer. \paragraph{The Turing tarpit.} The tradeoff does not assert a single conservation equality. As we will see, there exist regimes in both classical (language-based) and physical (geometric) computation with both low expressiveness \emph{and} low verifiability---the infamous ``Turing tarpit''. A space where everything is theoretically possible but practically inexpressible. The second half of the EVT is best read as a \emph{no-free-lunch inequality} with phases: beyond a threshold of expressiveness and verifiability, tractability for humans and machines becomes increasingly difficult. Before this extreme limit is reached, there lie transitional corridors in which small feature additions \emph{can} cause qualitative shifts in what can be expressed, verified, and/or computed. Because these tradeoffs depend on ambient structure, the EVT treats \emph{local} symmetries of a model or architecture rather than postulating a single global invariant across all substrates. In a Von Neumann setting, some of the most important local regularities are \emph{cognitive} and \emph{conventional}: programming idioms, abstractions, and composition patterns induce de facto invariants that temporarily ``abelianize'' parts of the system, while reflective or unrestricted features reintroduce non-commuting structure. \textbf{Pushing too far past the barriers above yields systems that remain formally expressive and verifiable in principle but become progressively less usable in human social, cognitive, and physical engineering practice}.\footnote{A fuller account of how the EVT Trilemma manifests in macroscopic programming practice is given in~\cite{Kaminsky2026EVT}. This outlines connections to resource theory and can be read as a companion piece.} \paragraph{Removing barriers to viewing computation.} Regarding tractability, the present paper intentionally abstracts away many ordinary engineering bottlenecks. In an idealized geometric regime, one may treat the substrate as reconfigurable enough that the sharp distinction between hardware and program becomes less natural than it is in fixed Von Neumann architectures. This does \emph{not} remove physical resource limits; rather, it allows the expressiveness/verifiability relation to be studied more directly, without immediately collapsing into familiar questions about instruction sets, compilers, and implementation overhead. For this reason, although the full trilemma must be stated as such, we do not pursue tractability in equal depth here. In the present paper, it functions chiefly as a reminder that human-built machines impose additional constraints beyond those intrinsic to the underlying geometric substrate. So in our GCM-based setting, the expressiveness/verifiability relationship can be studied more transparently because there is \emph{no fixed separation between logic and memory} of the usual Von Neumann sort. A GL is an idealized \textit{non}-Von Neumann substrate defined directly in geometric and algebraic terms. By fusing logic and language into a single analog-geometric object, one can \emph{see} how the operational semantics deforms the computational substrate, and model how the algebraic closure group constrains the geometry with every increasing bit! \paragraph{The mathematicians' bargain.} This tension between expressive freedom and structural discipline is not unique to programming languages; it is a recurring theme whenever algebraic power begins to dominate geometric intuition. As Baez~\cite{baez2022gettingnoetherstheorem} writes in \textit{Getting to the Bottom of Noether’s Theorem:} \begin{center} \begin{minipage}{0.8\textwidth} ``Atiyah~\cite{Atiyah2002Mathematics} warned us of the dangers of algebra: \textit{`` ...algebra is to the geometer what you might call the Faustian offer. As you know, Faust in Goethe’s story was offered whatever he wanted (in his case the love of a beautiful woman), by the devil, in return for selling his soul. Algebra is the offer made by the devil to the mathematician. The devil says: I will give you this powerful machine, it will answer any question you like. All you need to do is give me your soul: give up geometry and you will have this marvellous machine}.'' \em \end{minipage} \end{center} Baez notes that ``While this is sometimes true, algebra is more than a computational tool: it lets us \emph{express} concepts in a very clear and distilled way.'' We agree that the finest mathematics arises when algebra and geometry are brought into balance. But our contention here is that computation is one of the settings in which this balance becomes unusually unforgiving. Within an ordinary formal background such as ZFC, geometry and algebra can coexist without tension; but once a geometric system is required to \emph{compute}, the act of doing so (in our case---the ``gluing''~\ref{lem:geometric-logic-conservation}) \textit{enforces constraints} on \textit{how} the flow can be structured. The bargain, in this case, is not rhetorical but structural, and the devil shall not be cheated. \paragraph{But why physics?} It is reasonable to ask why a hypothesis about computation belongs in a mathematical-physics paper. First, differential geometry, topology, and gauge theory provide native languages for flow, locality, transport, gluing, invariants, defects, and boundary composition. These structures are not metaphors imposed on computation after the fact; they are mechanisms by which a physical substrate stores and transforms information. Second, physical computers are constrained by more than syntactic decidability. Finite precision, noise, energy, time, stability, measurement back-action, causal structure, and control limitations determine which formal processes can actually be realized and read. A physical selection principle should therefore be stated at a level that applies to discrete, analog, geometric, and hybrid substrates. Third, the physical setting prevents an easy conflation of unverifiability with symbolic diagonalization. A deterministic analog or field-theoretic process may fail to admit global pre-certification because of chaotic sensitivity, contextual readout, non-commuting transport, topological obstruction, or unresolved boundary data, even when it contains no literal self-interpreter. This does not diminish the interest of the EVT Hypothesis for computer science or the philosophy of natural laws. Versions of the underlying tension have appeared in many domains, though rarely isolated as a single, explicit selection principle. A few recurring patterns (illustrative, not exhaustive) are: \begin{itemize}[leftmargin=*,itemsep=2pt] \item \textbf{Logic and programming.} As previously stated, total and strongly normalizing calculi increase verifiability by withdrawing general recursion; reflective and metaprogramming systems increase expressive freedom while transferring assurance to later phases. \item \textbf{Numerical analysis and PDEs.} Stability conditions such as CFL-type bounds restrict discretizations to preserve controlled evolution, while structure-preserving integrators restrict admissible updates to retain symplectic, conservative, or geometric information \cite{CourantFriedrichsLewy1928}. \item \textbf{Dynamical systems.} Integrable and symmetry-protected regimes may admit exact invariants and reduced descriptions, while generic perturbations can destroy those certificates without introducing any symbolic self-reference. \item \textbf{Many-body and field theory.} Uniform decision problems concerning infinite families of Hamiltonians can become undecidable; the spectral-gap result of Cubitt, Pérez-García, and Wolf supplies a particularly sharp example \cite{CubittPerezGarciaWolf2015SpectralGap}. \item \textbf{Quantum information.} Coherent transformations preserve a rich phase and correlation structure, while the accessible classical interface returns only a context-dependent finite outcome. The distribution may remain predictable even when the individual output does not descend from the operational state. \end{itemize} These examples do not establish one theorem with identical hypotheses. They motivate the search for a common physical design constraint: systems gain expressive power by enlarging their admissible process structure, while globally certifying and feasibly controlling that structure requires additional restrictions. However, the most pressing applications of a \emph{computational} selection principle are presently in \emph{quantum gravity} and \emph{lattice gauge theory}, where one must extract dynamics from large, ongoing physical computations (experiments or simulations) under resource and precision constraints. The EVT Hypothesis is intended to bound \emph{how} such processes can behave and what can be recovered from them, independently of any particular quantization scheme. In this manner, it is proposed as a \emph{selection principle} for the behavior and modeling of computational processes: physical computation---discrete or analog---is not substrate-independent magic; it is a \emph{gauge-constrained flow} carried by a geometric substrate, with invariant quantities and resource monotones that cannot be simultaneously optimized without limit. From this viewpoint, the structures that make computation possible (\emph{how} we compute) are as consequential as the extensional outcomes (\emph{what} we compute). Any physical theory of computation should, at a minimum, respect this geometric constraint.