Skip to content

Convergence Evidence Matrix

Template — instantiates Independent Convergence Evidence Appraisal

Tabulates lineages, pressures, solution-shapes, independence evidence, performance, and caveats into one grid so convergence can be read across rows instead of asserted.

Convergence Evidence Matrix is the fill-in grid the whole appraisal writes into: one row per lineage, one column per appraisal dimension — origin, pressure profile, solution-shape at a fixed abstraction level, independence evidence, performance, and caveats. Its defining move is not to judge convergence but to lay it out on shared axes so that the comparison is like-for-like and every missing cell is visible. A claim that "everyone independently does it this way" collapses the moment it has to be entered as five rows on the same seven columns, because the blanks — the lineage whose independence column is empty, the case with no performance data — announce themselves. The matrix is a container and a discipline, not a verdict; it makes the evidence legible so that a separate scoring or memo step can act on it.

Example

A transportation-planning team is asked whether physically separated bike lanes are a genuinely convergent, fit-driven solution or just a fashionable import. They open a matrix with rows for Copenhagen, New York, Bogotá, Seville, and Montreal. The solution-shape column is pinned at one grain — "curb- or parking-protected lane with vertical separation," not the vaguer "bike infrastructure" and not the hyper-specific "green-painted 2.1 m lane." The pressure column forces each city to state its constraints on the same axes: cyclist injury rate, mode-share target, street width, political window. The independence column asks who each city learned from.

Filling it is where the story changes. Copenhagen's row is a deep origin; New York's independence cell reads "study tour to Copenhagen, hired Danish consultants" — so its row is flagged as derivative, not independent arrival. Seville's cell, by contrast, shows a compressed home-grown rollout with little contact with the others. The finished grid does not announce a conclusion, but it shows at a glance that of five rows, perhaps three are substantially independent and two are diffusion — exactly the distinction a bare list of "cities that did this" erases.

How it works

  • Fix the columns before the rows. Agree the dimension set and, critically, the abstraction grain of the solution-shape column, so no row can be entered at a convenient level that manufactures a match.
  • One lineage per row, same axes across. Each lineage is described on identical pressure and independence axes; the grid's value is the enforced parallelism.
  • Leave blanks visible. An unknown independence or performance cell is left empty and flagged, never quietly filled with optimism — the empty cell is the finding.
  • Annotate cell quality, not just cell content. Each entry carries a light strength tag (direct evidence / inferred / asserted) so a full grid is never mistaken for a strong one.

The matrix delegates the actual scoring: it hands the filled rows to a strength assessment and the finished picture to a warrant memo.

Tuning parameters

  • Column set — how many dimensions the grid tracks. More columns capture more risk (timing, standards pressure) but raise the fill cost and the blank count; match to stakes.
  • Shape-abstraction grain — how specific the solution-shape column is. Coarser grain finds more "matches" (and manufactures false ones); finer grain is honest but may split real recurrence into look-alikes.
  • Row-inclusion threshold — how much is required before a lineage earns a row. A low bar pads the grid with weak cases; a high bar risks survivorship by admitting only the tidy examples.
  • Cell evidence standard — what counts as filling a cell (a citation vs. a hunch). Stricter standards leave more honest blanks.

When it helps, and when it misleads

Its strength is that it converts an unfalsifiable "everyone does it" into a bounded, inspectable object where independence gaps and missing performance data cannot hide. It is the cheapest way to turn rhetoric into rows.

Its failure mode is false comparability: a fully populated grid looks like strong convergence even when the cells are thin, because the visual regularity of a filled table reads as evidential regularity — the tabular cousin of over-trusting agreement among measures that only appear independent.[n1] A classic misuse is padding rows with near-duplicate lineages (three forks of one origin entered as three cities) to inflate the apparent count. The guarding discipline is the cell-quality annotation and a hard rule that the matrix organizes evidence but never concludes from it — strength scoring and the warrant statement live in other mechanisms.

How it implements the components

  • solution_shape_abstraction — the shape column header fixes one abstraction grain for every row, so recurrence is judged like-for-like rather than by sliding the definition.
  • shared_pressure_profile — the pressure columns force each lineage to state its constraints, incentives, and failure costs on identical axes, making pressure comparability checkable.
  • lineage_independence_map — the independence column is the lineage map in tabular form: one cell per lineage naming its shared teachers, vendors, or platforms; the Traceback Protocol populates those cells.

It does NOT score how much the convergence should shift belief (convergence_strength_assessment) or flag suspicious agreement (false_unanimity_warning) — that is Blind Pattern Comparison Round; nor does it write the final calibrated inference and its bounded_warrant_statement / transfer_scope_boundary — that is its nearest twin, Convergence Warrant Memo, which turns the grid into a conclusion.

Editorial Notes

Form Classification

Form family: Representation, Specification & Plan

Rationale: The matrix externalizes each lineage on identical pressure, independence, solution-shape, performance, caveat, and evidence-strength axes while preserving unknown cells, so its operative form is a convergence representation.

Nearest alternative: Interface, Display & Cue — The grid makes convergence readable across rows, but the persistent structured evidence artifact rather than a user interaction or behavioral prompt is the mechanism.

Review outcome: Adjudicated after independent review; high confidence.

Origin Attribution

Primary origin: History & Historiography

Origin pattern: Cross-disciplinary synthesis

Present-day reach: Multi-domain

Rationale: Comparative history established lineage-by-lineage analysis of origin, diffusion, independence, shared pressures, and performance.

Related originating lineages:

Review resolution: The lineage-by-lineage comparison is best anchored in comparative intellectual history, while qualitative matrices, philosophy of independent evidence, and convergent-validity reasoning produce the exact synthesized grid.

Encyclopedia synthesis: The exact catalogued form synthesizes established practice rather than reproducing a single standard historical label.

Review outcome: Researched adjudication after independent review; medium confidence.

Sources consulted:

Notes

[n1] In psychometrics, convergent validity is the degree to which supposedly independent measures of the same construct actually agree. The concept is a double-edged reminder here: agreement is evidential only when the measures are genuinely independent, which is exactly what a filled grid can make look true without being true.