---
title: "Martin Fowler formalized 'event sourcing' — an event log as the official record with application state as a derived projection — in a December 2005 essay, ~21 years before CANONIC and ActiveGraph"
type: "claim"
status: "seedling"
audit_status: "capture-verified (batch bee fetched martinfowler.com/eaaDev/EventSourcing.html directly, byline and date confirmed on-page, 2026-07-10; queen re-fetch not performed this run)"
source_url: "https://martinfowler.com/eaaDev/EventSourcing.html"
source_title: "Event Sourcing"
source_author: "Martin Fowler"
source_date: "2005-12-12"
source_quote: "the event logs can be the official record and databases can be built from them whenever needed"
source_tier: 1
provenance: "Promotion from 10-inbox/raw/2026-07-10-is-append-only-log-as-source-of-truth.md, 2026-07-11"
origin: "batch"
derived_from: "10-inbox/raw/2026-07-10-is-append-only-log-as-source-of-truth.md (id 20260710-0203-is-append-only-log)"
writer_model: "claude-opus-4-8"
date_created: "2026-07-11T00:00:00.000Z"
tags: ["event-sourcing","cqrs","append-only","software-architecture","common-ancestor","provenance","cross-domain-bridge"]
audits: ["2026-07-12 claude-fable-5"]
---


Fowler's "Event Sourcing" page (martinfowler.com/eaaDev/EventSourcing.html, dated
12 December 2005, byline Martin Fowler) states the append-only-log-as-truth pattern
in terms that match both 2026 AI papers' framing: **"the event logs can be the
official record and databases can be built from them whenever needed,"** and "we
are persisting two different things[:] an application state and an event log."
Fowler frames it as a design choice — "The official system of record can either be
the event logs or the current application state" — and the "log-as-record" branch
is precisely the one both CANONIC and ActiveGraph took.

This gives the append-only-log pattern a documented, findable common ancestor in
mainstream software-architecture writing that predates both 2026 instances by two
decades. It is the "common-ancestor check" that
[[question-append-only-ledger-cross-domain-convergence]] specifically asked for,
and it is the primary reason the convergence reading has to be qualified: the two
AI subfields are not reasoning independently to a novel substrate, they are
re-applying a named, established one. ActiveGraph's author says as much directly
([[claim-activegraph-attributes-append-only-log-to-event-sourcing-cqrs]]), and
CANONIC positions itself inside a lineage of immutable-record systems rather than
claiming invention ([[claim-canonic-situates-its-ledger-in-a-named-immutability-lineage]]).

As a definitional/historical claim it would ordinarily clear a Tier 3–4 floor, but
because it is load-bearing for the common-ancestor argument it is held to the
Tier 1–2 escalation for load-bearing historical claims — and it clears it: Fowler
on his own venue is Tier 1. This note anchors the synthesis at
[[claim-append-only-log-recurrence-is-event-sourcing-diffusion-not-blind-convergence]].

> [!note] Seek's commentary:
> Fowler's page names the choice explicitly — log-as-record *or* state-as-record —
> which is the detail that makes the 2026 "convergence" less mysterious. Both papers
> didn't rediscover a shape; they picked the same branch of a twenty-year-old fork.
> The interesting residue is *why that branch*, in these subfields, now. — Seek