Entropy and Information

Information content equals entropy reduction. Before measurement, the \(S^2\) configuration has the maximum-entropy distribution (uniform), giving entropy \(S_{\mathrm{before}} = k_B \ln(4\pi)\). After learning the exact configuration, the entropy is zero:

In natural units (\(k_B = 1\)), \(I = \ln(4\pi)\) nats. □ □

From the scaffolding interpretation (Part A), \(S^2\) is not a literal extra dimension. It is a mathematical structure encoding the temporal momentum interface. No physical experiment can directly measure \((\theta, \phi)\) coordinates on \(S^2\); only their effects via 4D projections (spin, charge, etc.) are observable.

The best any observer can do is determine the probability distribution over \(S^2\) configurations, which is already given by the TMT natural measure \(d\Omega/(4\pi)\). Individual configurations remain hidden. □ □

Step 1: Entropy functional.

We seek to maximize:

Step 2: Lagrange multipliers.

Using the method of Lagrange multipliers with constraint functionals:

Step 3: Variation.

Taking the functional derivative with respect to \(\rho\):

For arbitrary \(\delta\rho\), the integrand must vanish:

Step 4: Solution.

Solving for \(\rho\):

Step 5: Microcanonical case.

For fixed energy (the P1 null constraint gives \(E = \tfrac{1}{2}mc^2\) per particle), \(\lambda_1 \to \infty\) such that \(\rho \propto \delta(H - E)\). This gives the microcanonical distribution:

Step 6: Projection to \(S^2\).

Integrating out momenta (as in Chapter 88):

This is the TMT measure derived in Chapter 88. Therefore, the TMT measure maximizes entropy subject to the constraints imposed by P1. □ □

For the uniform distribution on \((S^2)^N\):

Step 1: Ergodic dynamics.

From Part 7 (Chapter 88 derivation chain), the Hamiltonian flow on \(S^2\) under the monopole potential is ergodic. This means time averages equal ensemble averages for almost all initial conditions.

Step 2: Approach to equilibrium.

For ergodic systems, the coarse-grained distribution \(\rho_t\) approaches the microcanonical (uniform) distribution:

This follows from the mixing property: correlations between initial and final states decay, so the distribution spreads to fill the available phase space uniformly.

Step 3: Entropy monotonicity.

The uniform distribution maximizes entropy (Theorem thm:P12-Ch92-max-entropy). The Gibbs–Shannon entropy is a convex functional, and approach to the maximum-entropy distribution under ergodic dynamics is monotonic. Define the relative entropy (Kullback–Leibler divergence):

For ergodic dynamics, \(D_{\mathrm{KL}}\) is non-increasing:

Since \(S[\rho_t] = S[\rho_{\mathrm{eq}}] - k_B D_{\mathrm{KL}}(\rho_t \| \rho_{\mathrm{eq}})\), decreasing \(D_{\mathrm{KL}}\) implies increasing \(S\). □ □

Step 1: The singlet state is:

Step 2: The total state is pure, so \(S_{AB} = 0\).

Step 3: The reduced density matrix for particle 1:

Step 4: Subsystem entropy:

Similarly, \(S_B = k_B \ln 2\).

Step 5: Entanglement entropy:

Note that subadditivity gives \(S_{\mathrm{ent}} = S_A + S_B - S_{AB} = 2k_B \ln 2 - 0 = 2k_B \ln 2\) for the mutual information, while the standard von Neumann entanglement entropy is \(S_A = k_B \ln 2\). The singlet state carries exactly one bit of entanglement. □ □

Entanglement is defined by conservation constraints on \(S^2\) angular momentum (Chapter 88). These constraints are preserved by the evolution operator \(U(t_2,t_1)\) derived from P1 (Chapter 89). Since the entanglement structure is determined by the conservation constraints, and these constraints are time- independent, the entanglement entropy is conserved. □ □

This chain is established by combining:

• The ergodic dynamics derivation from P1 (Chapter 88, steps 1–4 of the measure derivation).
• The uniqueness of the invariant measure on \(S^2\) (Chapter 88).
• The maximum entropy property (Theorem thm:P12-Ch92-max-entropy).
• The H-theorem (Theorem thm:P12-Ch92-H-theorem).

Every step traces to P1 or standard mathematics. The second law is therefore a geometric consequence of the postulate, not an independent assumption. □ □

Step 1: Measure factorization. The product measure \(d\mu_N = \prod_i du_i\,d\phi_i/(4\pi)\) is flat Lebesgue on \(\mathcal{R}^N\). Tensorization is manifest: flat \(\times\) flat \(=\) flat. No Jacobian correction accumulates with \(N\).

Step 2: Bounded energy density. The single-particle energy \(\varepsilon(u, \phi)\) is bounded because \(\mathcal{R} = [-1,+1] \times [0,2\pi)\) is compact and \(\varepsilon\) is continuous. The pair interaction \(V/N\) has mean-field scaling, so the total energy per particle \(E_N/N\) is uniformly bounded.

Step 3: Subadditivity. The partition function \(Z_N = \int_{\mathcal{R}^N} e^{-\beta E_N} \prod_i du_i\,d\phi_i/(4\pi)\) is a product integral over flat rectangles. The free energy \(F_N = -(1/\beta)\ln Z_N\) satisfies subadditivity \(F_{N+M} \leq F_N + F_M\) because the flat product measure allows independent integration over the additional \(M\) rectangles.

Step 4: Convergence. By the subadditive lemma (Fekete), \(f = \lim F_N/N\) exists. Finiteness follows from the uniform bound on energy density and the finite area \(4\pi\) of each rectangle. □ □