Blameless Language Guide
Word choice.
Overview
Blameless language is word choice at the sentence level. The same incident can be described in ways that point at people ("the engineer caused X") or at systems ("the system allowed X"); the discipline picks the latter every time. Five patterns carry most of the work: system framing, voice choice, component naming, "could have" instead of "should have," catching counterfactual hindsight bias.
- System-framed word choice. "The system allowed" rather than "the engineer caused." Same fact, neutral actor.
- Active vs passive voice deliberately. Passive voice describes systems; active voice often names individuals. Pick per sentence.
- Component naming. Roles or services, not people. Depersonalises the analysis without losing precision.
- "Could have" beats "should have" plus hindsight check. "Should have" implies fault; "could have" describes options. Counterfactual "if only X" is hindsight bias to catch and edit out.
The approach
Template with examples, peer review every postmortem, retroactive edits when blame language slips through, train new on-callers, document the language guide. The discipline scales because the rules live in writing rather than tribal review knowledge.
- Template with examples. Good and bad examples in the team's PM template. Supports new postmortem authors without re-explaining each time.
- Peer review. Second reader catches blame language the writer normalised. Consistency by review.
- Retroactive edits. Old PMs can be edited when blame language is found later. Ongoing improvement of the archive.
- Train new on-callers plus documented language guide. Blameless framing has to be taught explicitly; the team's language guide lives in writing for reference.
Why this compounds
Each blameless PM compounds the next one. Engineers share what they actually saw; system patterns surface across incidents; the team's incident culture matures past the blame-default mode of human storytelling. By year two, blameless framing is the team's natural style rather than a deliberate choice per sentence.
- Better PM quality. Blameless framing produces real root causes. Learning is the output.
- Better psychological safety. Engineers share what really happened. Honest PMs follow.
- Cultural reinforcement. Language shapes thinking. Reinforces the engineering culture the team is trying to build.
- Year-one investment, year-two habit. First PMs establish the language; subsequent ones run on the template.