Chapter 23: Geometric Finance — Market Microstructure on the Decision Manifold
RUNNING EXAMPLE — Priya’s Model
HealthBridge’s Series C investors evaluate TrialMatch using exactly one metric: expected revenue per match. This financial scalar inherits TrialMatch’s moral blindness. When Priya’s proposed fix reduces match volume by 8%—because it routes some rural patients to closer alternative trials—the investors see only a revenue drop. The moral dimensions of the change—improved access, reduced harm, better trial diversity—live in dimensions the spreadsheet cannot display.
This chapter applies the Geometric Ethics framework to financial markets. The standard model of finance — the Efficient Market Hypothesis (EMH) in its strong form — assumes that prices fully reflect all available information, that agents are rational expected-utility maximizers, and that arbitrage enforces price efficiency. Behavioral finance, following Shiller and Thaler, has documented systematic departures from EMH but, like behavioral economics, lacks a unified mathematical framework that explains why these departures occur and what structure they share. The geometric framework provides that structure. Financial decisions are pathfinding on a nine-dimensional financial decision complex, market anomalies are geometric properties of the manifold, and risk is curvature.
23.1 The Failure of Scalar Finance
Modern financial theory rests on three pillars, each of which assumes scalar optimization.
The Efficient Market Hypothesis (Fama, 1970). Markets are informationally efficient because rational agents exploit all available information to maximize expected monetary return. The EMH is the financial analog of Homo economicus: agents optimize on d_1 (monetary return) alone.
The Capital Asset Pricing Model (Sharpe, 1964; Lintner, 1965). Risk is measured by β — the covariance of an asset's return with the market portfolio, divided by market variance. This is a scalar measure: each asset gets a single number summarizing its "riskiness."
Black-Scholes Option Pricing (Black and Scholes, 1973). Options are priced using a model in which the underlying asset's price follows geometric Brownian motion with constant volatility. The model's single input beyond observables — volatility — is a scalar.
Each pillar has been empirically falsified. The EMH fails to explain bubbles, panics, and persistent anomalies (momentum, value premium, low-volatility anomaly). CAPM fails empirically: the Fama-French three-factor and five-factor models demonstrate that β alone is insufficient. Black-Scholes fails observationally: the implied volatility surface (smile, skew) demonstrates that the market prices options as if volatility is neither constant nor scalar.
The geometric diagnosis is the same as for economics (Chapter 20): scalar projection destroys information. Financial decisions are made on a nine-dimensional manifold; projecting to d_1 alone loses the eight dimensions on which market participants actually condition their behavior. The implied volatility smile is the shadow of the higher-dimensional pricing manifold projected onto the d_1 axis.
23.2 The Financial Decision Complex
Definition 23.1 (Financial Decision Complex). The financial decision complex F is a weighted simplicial complex whose vertices are financial states — configurations of portfolio positions, market conditions, regulatory status, and counterparty relationships — and whose edges are financial actions (trades, hedges, margin calls, regulatory filings). Each vertex v_i carries an attribute vector a(v_i) in R^9.
The nine dimensions of the financial manifold are the moral manifold dimensions of Chapter 5, instantiated for financial contexts:
d_1: Monetary return — profit and loss, alpha, risk-adjusted return. This is the dimension classical finance targets.
d_2: Contractual obligations — fiduciary duties, derivative contract terms, margin requirements, collateral agreements.
d_3: Market fairness — equal access to information, prohibition of insider trading, market manipulation, front-running.
d_4: Trading autonomy — freedom to enter and exit positions, market access, circuit-breaker constraints.
d_5: Counterparty trust — counterparty credit risk, clearing house reliability, settlement certainty, market integrity.
d_6: Systemic impact — externalities imposed on the broader financial system, contagion risk, too-big-to-fail effects.
d_7: Reputational identity — trader and firm reputation, brand value, market-maker reputation for fair dealing.
d_8: Regulatory legitimacy — compliance with securities regulation, exchange rules, Basel capital requirements, Dodd-Frank provisions.
d_9: Epistemic advantage — information advantage, market transparency, speed of information access, analytical sophistication.
Definition 23.2 (Financial Edge Weights). The weight of an edge (v_i, v_j) in F is:
w(v_i, v_j) = ΔaT ΣF−1 Da + Σk βk * 𝟙[boundary k crossed]
where Da = a(v_j) - a(v_i), Sigma_F is the 9x9 financial covariance matrix, and βk are boundary penalties for crossing regulatory, contractual, or ethical boundaries. Critical covariance terms include Sigma_{1,6} (return x systemic impact: high-return strategies that increase systemic risk carry cross-dimensional cost), Sigma_{3,9} (fairness x epistemic: information advantages that undermine market fairness are penalized), and Sigma_{5,8} (trust x regulatory: regulatory compliance reinforces counterparty trust).
23.3 Market Microstructure as Manifold Geometry
Market microstructure — the study of how trading occurs and how prices are formed — maps naturally onto the financial decision complex.
Kyle's Model (1985). Kyle's canonical model has three agent types: an informed trader with private information, noise traders who trade for liquidity, and a market maker who sets prices to clear the market. In the geometric framework:
The informed trader has lower effective edge weights on d_9 (epistemic advantage): her information reduces the Mahalanobis distance to states she knows the market will reach. The noise traders have high d_9 weights: they trade without information, incurring high epistemic cost. The market maker absorbs d_5 cost (counterparty trust) in exchange for the bid-ask spread (d_1 compensation).
Kyle's lambda (price impact) is a manifestation of the d_1-d_9 interaction: the market maker infers informed trading from order flow and adjusts prices, converting the informed trader's d_9 advantage into d_1 cost for the noise traders.
Theorem 23.1 (Information Asymmetry as Metric Distortion). In a market with informed and uninformed agents, the informed agent's effective metric tensor on the financial decision complex differs from the uninformed agent's. Specifically, the informed agent's Mahalanobis distance to the true future state is systematically lower on the d_9 component:
w_informed(v_i, v_j) < w_uninformed(v_i, v_j) for edges toward the true state
The difference w_uninformed - w_informed is the information rent — the value of private information measured in manifold distance. In Kyle's model, this rent is transferred from noise traders to the informed trader on d_1, mediated by the market maker's pricing function.
Proof. The informed agent's attribute vector on d_9 is calibrated to the true future state: d_9^{informed} is high (correct epistemic status). The uninformed agent's d_9 is low (uncertain). The Mahalanobis distance ΔaT ΣF−1 Da from current state to future state is smaller when Da_9 is small (informed agent is already close to the true state on d_9). The systematic difference in path costs produces systematic differences in optimal paths: the informed agent's geodesic reaches the true future state at lower cost. []
The Glosten-Milgrom Model
Glosten and Milgrom (1985) model the bid-ask spread as the market maker's compensation for adverse selection. In the geometric framework, the bid-ask spread is the Mahalanobis distance between the buy-path and the sell-path on the financial decision complex. When information asymmetry (d_9 disparity between informed and uninformed traders) increases, the market maker's d_5 cost (counterparty trust risk) increases, widening the spread.
Lemma 23.1 (Bid-Ask Spread as Manifold Curvature). The bid-ask spread s at vertex v is proportional to the Gaussian curvature K(v) of the financial decision complex at v:
s(v) ~ K(v) * |Da_9|
where |Da_9| is the information asymmetry at v. In regions of high curvature (volatile, uncertain markets), spreads widen; in flat regions (stable, transparent markets), spreads narrow.
Proof. The bid-ask spread compensates the market maker for the risk that the counterparty is informed. This risk is a function of two quantities: the magnitude of information asymmetry (|Da_9|) and the rate at which the optimal path changes with small perturbations in the current state (curvature K(v)). In flat regions, a small information advantage produces small path divergence (low adverse selection risk); in high-curvature regions, the same information advantage produces large path divergence (high adverse selection risk). The proportionality follows from the Taylor expansion of the Mahalanobis distance in curved coordinates. []
23.4 The Flash Crash as Dimensional Collapse
On May 6, 2010, the U.S. stock market lost approximately $1 trillion in market value in minutes before recovering most of the loss. The Flash Crash exposed the fragility of markets dominated by algorithmic trading systems that optimize on d_1 alone.
Theorem 23.2 (Dimensional Collapse). When algorithmic trading systems reduce decision-making to d_1 alone, the effective financial decision complex collapses from 9D to 1D. Boundary penalties on non-monetary dimensions (d_3 fairness, d_5 counterparty trust, d_6 systemic impact, d_8 regulatory legitimacy) are zeroed out, enabling paths that full-manifold agents would never traverse.
Proof. An algorithmic agent with objective function u = d_1 (pure monetary return) computes edge weights using only the d_1 component of the Mahalanobis distance: w_alg(v_i, v_j) = (Da_1)^2 / sigma_11. All boundary penalties βk for k != 1 are set to zero. The feasible path set under this metric strictly contains the feasible path set under the full metric (removing constraints enlarges the feasible set). Paths through catastrophic states — states with acceptable d_1 transitions but extreme d_5, d_6, or d_8 costs — become traversable. []
The Flash Crash Mechanism. During the Flash Crash, a large sell order from a mutual fund was executed by an algorithm optimizing on d_1 alone. The algorithm's geodesic on the 1D projection was optimal: sell at the best available price. But the 1D geodesic traversed states where d_5 (trust: other market participants lost confidence in price integrity), d_6 (systemic impact: cascading liquidations), and d_8 (legitimacy: exchange circuit breakers were triggered) carried enormous cost. The 1D algorithm could not perceive these costs. The full-manifold geodesic would have incorporated the systemic cost of rapid execution and chosen a slower path, sacrificing some d_1 performance to avoid d_5-d_6-d_8 catastrophe.
Remark (IEX Speed Bump). The Investors Exchange (IEX) introduced a 350-microsecond speed bump — a deliberate delay on incoming orders — to prevent high-frequency traders from exploiting microsecond d_9 advantages. In the geometric framework, the speed bump is a boundary penalty: it raises the edge cost for paths that exploit extreme d_9 asymmetry, forcing traders onto paths with lower d_9 exploitation and higher d_3 (fairness) preservation.
23.5 High-Frequency Trading and the Fairness Problem
High-frequency trading (HFT) firms exploit speed advantages — colocation, microwave towers, FPGA hardware — to access market information microseconds before other participants. In the geometric framework, HFT is a systematic exploitation of d_9 (epistemic advantage) at the expense of d_3 (market fairness).
The HFT Geodesic. The HFT firm's financial decision complex has d_9 edge weights that are systematically lower than those of slower participants. The HFT firm's geodesic reaches profitable states (high d_1) at lower cost, not because of better fundamental analysis but because of lower d_9 friction (faster information access). The value transfer from slower participants to HFT firms on d_1 is the monetary manifestation of the d_9 advantage.
Theorem 23.3 (HFT as Fairness-Epistemic Trade-off). In a market with HFT and non-HFT participants, the HFT firm's optimal path on the d_1-d_9 projection differs from its optimal path on the full manifold. Specifically, the d_1-d_9 geodesic exploits speed advantages maximally, while the full-manifold geodesic incorporates d_3 (fairness) costs that constrain exploitation.
Proof. On the d_1-d_9 projection, the HFT firm minimizes a two-dimensional Mahalanobis distance without fairness constraints. Adding d_3 to the optimization introduces cross-terms Sigma_{3,9}^{-1} that penalize high d_9 advantage when d_3 is low (unfair conditions). The minimum-cost path shifts toward lower d_9 exploitation and higher d_3 preservation. The difference between the two geodesics is the fairness cost of HFT, measurable in manifold distance. []
Remark. The regulatory debate over HFT (Budish, Cramton, and Shim 2015 vs. Menkveld 2013) is a debate over which manifold dimensions should count. Defenders argue HFT improves d_1 efficiency (tighter spreads, better price discovery). Critics argue it degrades d_3 (unfair access) and d_5 (market trust). The geometric framework clarifies the trade-off: both sides are correct on different dimensions. The policy question is the relative weighting in Sigma_F — an empirical question, not a philosophical one.
23.6 Geometric Risk Management
Classical risk management measures risk as variance, Value-at-Risk (VaR), or Conditional Value-at-Risk (CVaR) — all scalar projections of a multi-dimensional risk landscape. The geometric framework identifies risk with curvature.
Theorem 23.4 (Financial Irrecoverability). Any scalar risk measure R: R^9 -> R is non-injective. Multiple financially distinct risk states map to the same scalar risk score. The information destroyed by scalar risk measures is mathematically irrecoverable.
Proof. Identical to the proof of Theorem 20.2 (Scalar Irrecoverability), applied to the financial domain. By Brouwer's invariance of dimension, no continuous injection from R^9 to R^1 exists. []
Definition 23.3 (Risk as Curvature). The risk at a point v on the financial decision complex is the sectional curvature K(v) of the manifold metric at v. High curvature means that small perturbations in the current state produce large changes in the optimal path — the financial analog of "sensitivity to initial conditions." Low curvature means the optimal path is robust to perturbation.
This definition unifies several classical risk concepts:
Volatility is the magnitude of curvature fluctuation along the d_1 axis: high volatility corresponds to rapidly varying curvature in the monetary dimension.
Tail risk is curvature concentration at the boundary of the manifold: extreme events correspond to regions of very high curvature where the optimal path changes catastrophically with small perturbations.
Correlation risk is off-diagonal curvature: the rate at which optimal paths on one dimension change when another dimension is perturbed. The 2008 financial crisis was a correlation-risk event: the assumption of low Sigma_{i,j} (independent mortgage defaults) was catastrophically wrong when the true covariance spiked.
Liquidity risk is curvature of the market-access dimension (d_4): in liquid markets, d_4 curvature is low (small perturbations in desired trade size produce small changes in execution cost); in illiquid markets, d_4 curvature is high (small increases in trade size produce large increases in execution cost).
23.7 Financial Conservation Laws
Theorem 23.5 (Financial Conservation). In a closed bilateral financial transaction between parties A and B, for each transferable dimension k in {d_1 (monetary value), d_2 (contractual obligations), d_4 (trading autonomy)}:
Da_k(A) + Da_k(B) = 0
What A gains on dimension k, B loses, and vice versa. This is the specialization of the Attribute Conservation Theorem (Chapter 20, Theorem 20.6) and the Conservation of Harm (Chapter 12, Theorem 12.1) to financial markets.
Proof. In a closed bilateral transaction, the total endowment on each transferable dimension is fixed. Money (d_1) is neither created nor destroyed by trading; it is transferred. Contractual obligations (d_2) are paired: every long position has a short counterparty. Trading autonomy (d_4) is zero-sum in the sense that a restriction on one party's freedom to exit creates a corresponding obligation on the other. By closure, Da_k(A) + Da_k(B) = 0. []
Remark (Non-Conservation of Evaluative Dimensions). Evaluative dimensions — d_3 (market fairness), d_5 (counterparty trust), d_7 (reputation), d_8 (regulatory legitimacy) — are not conserved. Both parties to a transaction may simultaneously gain or lose trust, perceived fairness, or institutional legitimacy. A fair and transparent trade creates mutual value on d_3, d_5, and d_8 that did not exist before the transaction. A manipulative trade destroys trust on both sides.
Remark (Market Crashes as Conservation Violation). A market crash appears to violate d_1 conservation: aggregate market value decreases with no corresponding increase. The resolution is that market value is not the same as money. Market value is a mark-to-market evaluation — a conditional expectation of future cash flows. A crash destroys evaluated value (d_1 under current beliefs) while conserving actual money (d_1 as medium of exchange). The crash simultaneously destroys d_5 (trust) and d_8 (legitimacy) across all participants — a non-conserved evaluative dimension collapse.
23.8 Geometric Behavioral Finance
The major phenomena of behavioral finance are derived as geometric properties of the financial decision complex:
Bubbles as Heuristic Inadmissibility. A financial bubble occurs when market participants' heuristic h(n) systematically overestimates the benefit (underestimates the cost) of continuing on the current path. In A* terms, h(n) < 0 for nodes along the bubble path — the heuristic assigns negative cost (expected benefit) to remaining steps, encouraging the agent to continue even when g(n) (accumulated cost) is already high.
Theorem 23.6 (Bubble Dynamics). A financial bubble is an A* search with inadmissible heuristic. Specifically, the market's collective heuristic h_market(n) satisfies h_market(n) < h*(n) with h_market(n) < 0 for bubble-path nodes, where h*(n) > 0 is the true remaining cost. The bubble bursts when g(n) accumulated evidence forces revision of h_market toward h*, causing the evaluated cost f(n) = g(n) + h_market(n) to exceed the cost of alternative paths.
Proof. During a bubble, market participants estimate future returns (negative costs) as increasingly certain, producing h_market(n) < 0. The true cost h*(n) accounts for the probability of mean reversion, regulatory intervention, and fundamental revaluation, all of which are positive. As g(n) accumulates (higher prices, lower yields, stretched fundamentals), the gap between h_market and h* becomes unsustainable. A triggering event (margin call, earnings miss, regulatory action) forces h_market toward h*, making the bubble path dominated by the exit path. This is the burst. []
Momentum as Manifold Inertia. The momentum anomaly — the tendency of assets with recent positive returns to continue outperforming — is manifold inertia. On a curved manifold, the geodesic has a "direction" that persists over short distances (autoparallel transport). In the financial decision complex, an asset whose attribute vector has been moving in a consistent direction on d_1 has lower edge weights for continued motion in that direction (the Mahalanobis distance for small, same-direction changes is small). Reversal requires changing direction on the manifold, which has higher curvature cost.
Panic as Catastrophic Path Switching. A market panic is a sudden, discontinuous switch from one geodesic to another. In the geometric framework, this occurs when the manifold curvature at the current position exceeds a critical threshold: the optimal path bifurcates, and the agent switches from the "normal" geodesic to the "exit" geodesic. The discontinuity explains why panics are sudden — the transition is not gradual but catastrophic (in the mathematical sense of Thom's catastrophe theory).
23.9 Option Pricing as Scalar Projection
Proposition 23.1 (Black-Scholes as Scalar Projection). The Black-Scholes model prices options using only d_1 (return distribution) and its variance (constant volatility). The model is a scalar projection of the financial decision complex onto d_1 alone.
The implied volatility surface is the market's correction for this projection. The volatility smile (out-of-the-money puts and calls have higher implied volatility than at-the-money options) reflects the market's incorporation of dimensions invisible to Black-Scholes:
d_3 (fairness): Jump risk — the possibility of sudden large moves that disproportionately affect less-informed participants — is priced into out-of-the-money options, widening the smile.
d_5 (trust): Counterparty risk premium, especially post-2008, increases implied volatility for long-dated options where counterparty default is more probable.
d_6 (systemic impact): Correlation risk — the risk that all assets move together in a crisis — is priced into index options relative to single-stock options, producing the "skew."
d_8 (regulatory): Regulatory intervention risk (position limits, trading halts, short-selling bans) is priced into options on regulated assets, producing term-structure effects.
The implied volatility surface is not a "failure" of Black-Scholes. It is the market's attempt to recover, within a scalar framework, the information that scalar pricing destroys. The geometric framework predicts that the surface shape should correlate with the active non-d_1 dimensions — a testable prediction.
23.10 The 2008 Financial Crisis as Manifold Failure
The 2008 global financial crisis provides a comprehensive case study in manifold failure across multiple dimensions simultaneously.
Dimensional collapse in mortgage markets. The originate-to-distribute model separated mortgage origination (d_2: contractual obligations) from risk-bearing (d_1: return). Originators optimized on d_1 alone (origination fees) with zero d_2 cost (no skin in the game). This dimensional collapse enabled paths through the financial decision complex that full-manifold originators would never have traversed: loans to borrowers who could not repay, packaged into securities whose d_9 (epistemic status: actual risk profile) was systematically misrepresented.
Rating agency failure as gauge variance. Rating agencies assigned AAA ratings to mortgage-backed securities whose actual risk was far higher. In the geometric framework, this is a gauge-invariance violation: the same security, described differently (as a pool of individual mortgages vs. as a structured tranche), received different risk evaluations. The rating depended on the description (the gauge), not on the underlying attribute vector — a Bond Invariance Principle violation.
Contagion as curvature cascade. When Lehman Brothers collapsed, the d_5 (trust) dimension across the entire financial system underwent catastrophic curvature increase. Counterparty trust, previously a slowly varying manifold feature, spiked to infinity (no participant trusted any other's solvency). This curvature cascade propagated through the d_5-d_6 covariance channel: lost trust (d_5) amplified systemic impact (d_6), which further degraded trust. The positive feedback loop is a curvature instability on the financial manifold.
Regulatory response as manifold repair. The Dodd-Frank Act (2010) can be understood as manifold repair: raising d_2 boundary penalties (higher capital requirements, Volcker Rule), increasing d_8 edge weights (compliance costs), and introducing d_6 boundary penalties (systemic risk surcharges for systemically important financial institutions). The goal was to prevent future dimensional collapse by ensuring that d_2, d_5, d_6, and d_8 costs could not be zeroed out.
23.11 Falsifiable Predictions
The framework generates six predictions that distinguish it from classical and behavioral finance:
Prediction 1 (Volatility Surface Correlation): The shape of the implied volatility surface should correlate with measurable non-d_1 dimensions. Specifically, the smile width should correlate with d_9 disparity (information asymmetry), and the skew should correlate with d_6 (systemic risk) proxies. Falsified if: the surface shape is independent of non-d_1 measures.
Prediction 2 (HFT Fairness Index): A Financial Bond Index computed from HFT activity should predict market participant attrition (departure of slower participants from the market). Falsified if: participant attrition is uncorrelated with the computed BI.
Prediction 3 (Flash Crash Prediction): Markets with higher d_1 concentration (more algorithmic, less human oversight on non-d_1 dimensions) should experience more frequent and severe flash crashes. Falsified if: flash crash frequency is independent of d_1 concentration.
Prediction 4 (Curvature-Risk Correlation): Historical curvature estimates on the financial decision complex should predict future realized risk (measured by subsequent volatility or tail-event frequency) better than scalar risk measures (VaR, β). Falsified if: curvature adds no predictive power beyond scalar risk measures.
Prediction 5 (Crisis Covariance Spike): During financial crises, the estimated covariance matrix Sigma_F should show characteristic patterns: d_5-d_6 covariance increases (trust-systemic coupling), d_1-d_5 covariance increases (return-trust coupling), and d_8 variance increases (regulatory uncertainty). Falsified if: crisis covariance patterns are structureless.
Prediction 6 (Manifold Dimensionality): Factor analysis of financial market behavioral data (incorporating non-price dimensions such as regulatory filings, counterparty ratings, trading speed, and market fairness indicators) should recover approximately nine independent dimensions. Falsified if: the factor structure is consistently lower-rank (fewer than seven) or higher-rank (more than eleven).
23.12 Connection to the Framework
The Geometric Finance program extends the parent framework in five directions:
1. Chapter 20 applied the framework to individual and bilateral economic decisions. This chapter extends it to market microstructure — the domain where many agents interact simultaneously through an institutional mechanism (the exchange), introducing market-specific geometry (bid-ask spread as curvature, liquidity as d_4 smoothness).
2. Chapter 15 established scalar irrecoverability. This chapter applies it to financial risk measurement (VaR, CAPM β) and option pricing (Black-Scholes), showing that the implied volatility surface is the market's attempt to recover scalar-projected information.
3. Chapter 12 established conservation laws. This chapter derives financial conservation (money and contractual obligations are conserved in bilateral transactions) and identifies the non-conservation of evaluative dimensions (trust, legitimacy, reputation) as the mechanism underlying financial contagion.
4. The Flash Crash analysis demonstrates a new phenomenon — dimensional collapse — where algorithmic agents' restriction to d_1 enables traversal of catastrophic paths invisible to full-manifold agents. This is the financial analog of the ethical concern about AI systems that optimize on a scalar objective without moral constraints.
5. The 2008 financial crisis analysis shows how gauge-invariance violations (rating agency failures), dimensional collapse (originate-to-distribute), and curvature instability (trust cascades) interact to produce systemic failure — a multi-dimensional crisis that scalar analysis cannot diagnose.
23.13 Summary
This chapter has shown that the geometric ethics framework, when applied to financial markets, yields:
1. A formal construction of the financial decision complex F as a domain-specific instantiation of the moral manifold, with market microstructure mapped to manifold geometry.
2. Information asymmetry as metric distortion: Kyle's informed-trading model reinterpreted as asymmetric Mahalanobis distance on the d_9 component, with the bid-ask spread as manifold curvature.
3. The Flash Crash as dimensional collapse: algorithmic restriction to d_1 enables catastrophic paths invisible to full-manifold agents.
4. Risk as curvature: a unified geometric framework that subsumes volatility, tail risk, correlation risk, and liquidity risk as different aspects of manifold curvature.
5. Financial conservation: transferable dimensions (money, contracts) are conserved; evaluative dimensions (trust, legitimacy) are not — explaining both market stability and contagion.
6. Behavioral finance phenomena (bubbles, momentum, panics) as geometric properties: bubbles as inadmissible heuristics, momentum as manifold inertia, panics as catastrophic path switching.
7. The implied volatility surface as shadow of higher-dimensional pricing: the market's correction for scalar projection.