From Hold-Capacity to Crisis Triage: A Full-DQ Diagnostic Grammar for Financial Stress, Failure Channels, and Intervention Match

This paper develops a crisis-triage methodology for financial stress based on the Full-DQ grammar: drive pressure (D), holding-capacity (h), blockade (B), and boundary/contactability (σ). The starting point is the Hold-Capacity result — a reported and internally validated measure of h as the no-discharge viability radius across banking crises. Hold-Capacity is valuable because it asks whether an institution remains able to carry stress without discharging it. However, financial crises do not fail through capacity shortfalls alone. They also fail through accelerating pressure, blocked relief channels, boundary and confidence collapse, and mismatched interventions. The purpose of this paper is therefore not to replace Hold-Capacity with a new scalar score. It is to build a diagnostic grammar that can answer a different question: what kind of crisis is occurring, through which primitive channel, and which intervention class is structurally matched to it? The paper makes five main contributions. First, it formalizes Full-DQ crisis triage as a practical method rather than a general metaphor. D denotes pressure toward discharge; h denotes remaining capacity to hold difference; B denotes blocked conversion of available potential into effective relief; and σ denotes the boundary condition by which an institution, claim, market, or funding channel remains economically contactable and recognizable. Second, it defines an intervention taxonomy: I_D reduces pressure, I_h increases capacity, I_B reduces blockade, and I_σ repairs boundary/contactability. Third, it introduces Protocol v0.4, an archetype-based coding system for dominant crisis channels. Fourth, it applies the protocol to a public-evidence event atlas of thirteen financial stress cases: Black Monday 1987, LTCM, Bear Stearns, Lehman, AIG, Washington Mutual, Silvergate, Silicon Valley Bank, Signature Bank, First Republic, Credit Suisse, Western Alliance, and PacWest. Fifth, it states falsifiers and next tests, including inter-rater reliability, H-only difficult-case evaluation, intervention-match prediction, and eventual panelization using bank and market data. The event atlas is not presented as final validation but as a pilot corpus that clarifies the empirical object. The main finding is that financial stress events are better understood as coupled-channel archetypes than as scalar risk rankings. Run and confidence crises cluster as D+σ; confidence/contactability crises as σ+D; fire-sale and market-functioning events as B+σ+D; dealer/funding-market ruptures as D+B+σ; bank-run plus balance-sheet failures as D+h+σ; all-primitives ruptures as D+B+σ+h or h+B+σ+D depending on containment; and boundary-repair cases as A8a (σ_repair+D) or A8b (σ_repair+D+B). The A8 distinction is especially important: a useful crisis model must explain not only why systems break, but why they sometimes do not break despite visible stress. The paper concludes that Full-DQ is best treated as a crisis-triage grammar for timing, channel identification, and intervention matching, rather than as a universal replacement for existing risk scores.