A market can clear after its culture stops happening.

Liquidity records motion. It cannot tell us whether judgment, responsibility, or learning remains inside the motion.

Stock markets are mechanisms, but they are also disciplines. People build theses, risk reputations, argue about value, learn from error, and inherit practices from other people.

An automated market may keep producing prices while these human activities become ceremonial. Functional continuity can conceal cultural loss.

TEST: Raise AI penetration in the field. Throughput improves while the practice's human indices contract.

HX02 PLURALITY FIELD3.2 EVENTS/S

Human theses / model convergence

Each moving mark is an invented market participant or model agent. Click the field to introduce a dissenting thesis.

SYNTHETIC THOUGHT EXPERIMENT

HUMAN JUDGMENT ↔ AI PENETRATION 0%

HX03 CULTURAL BOOKINDEX / 100

SCENARIO H0

The human market

Orders carry attributable choices. Rival interpretations persist. Errors can become shared lessons.

The indices above are authored visual indices. They are not measurements, forecasts, or claims. They must be authored, because the capacities they point at do not appear in any measurement (see below).

MEASURED REAL MARKET RECORD · 2011–2025

Absorption ratio: share of return variance in the first principal component of a ~700-stock universe (Kritzman–Li; Sharadar prices). It is the measurable analog of convergence. It spikes in crises (2011 Euro, 2020 COVID) and does not slide; cross-sectional idiosyncrasy has risen since 2017 (3.2 → 4.7%/day).

MEASURED READING. The measurable market did not converge culturally. It had crises, and grew more idiosyncratic. And when the market's internal structure was measured directly, the concentration of cross-characteristic dispersion, its return content collapsed entirely into the standard characteristics (a slope of t=+2.12 falls to t=−0.06 once those characteristics are controlled). The human residual, judgment, apprenticeship, accountability, is precisely the part no characteristic set captures. A dashboard cannot see the loss you describe. That is the empirical case for your normative one: the boundary must be a rule, not a metric.

HX04 MARKET BOUNDARYPROPOSED

Keep the market-facing chain human. Keep its infrastructure deterministic and inspectable.

PERMITTED

Human-authored research and orders
Calculators and fixed statistical tools
Deterministic matching and limit rules
Named stewardship and oral defense

OUTSIDE THE BOUNDARY

Model-originated investment theses
AI ranking or recommendation
Autonomous order choice or routing
Generated explanations after the fact

HX05 PRESERVATION LEDGERWHAT THE SPACE KEEPS ALIVE

01 / DISSENT

Prices need rival readings.

Efficiency can compress difference. A human-only venue protects the social production of competing interpretations.

02 / MEMORY

Skill needs transmission.

Apprentices learn judgment from situated stories, defended choices, and witnessed mistakes, not outputs alone.

03 / ADDRESS

Responsibility needs a person.

Consequential allocation should end at someone who chose, can explain, and remains answerable for the choice.

HX06 SOURCE NOTEEDITORIAL

Conceptual source: Anastasia Berg and Jon Baskin, On Radical Preservation, The Point, August 2, 2026.

This market proposal is an independent application of the essay's framework. The essay does not discuss finance, supply market evidence, or imply that its authors endorse this argument.

Countercase: AI can widen access, lower research costs, reduce clerical work, and challenge some human biases. The proposed boundary must answer those benefits.

READ THE EDITORIAL SOURCE ↗