Retrospective Sensemaking

In the AI Story

Weick borrowed the phrase "How can I know what I think until I see what I say?" from Graham Wallas, who attributed it to a little girl he overheard. The attribution was characteristic of Weick: the deepest insights into organizational cognition come from children, from poets, from anyone willing to notice the gap between what we think we are doing and what we are actually doing.

The principle cuts against the rational decision-making model at its foundation. That model assumes people form intentions, evaluate alternatives, select the best option, and then act. Retrospective sensemaking claims the reverse: people act on the partial interpretations available, observe what happens, and then — often much later — construct the account that makes the action intelligible. The action comes first. The understanding is reverse-engineered.

Mann Gulch illustrates the dynamic under extreme conditions. The survivors and the investigation constructed accounts of heroism, tragedy, and organizational failure, each plausible and each shaped by the outcome of thirteen deaths. But the experience of the men in the gulch, while it was happening, had no narrative structure. The situation exceeded their interpretive frameworks. They could not make sense of a fire that reversed direction and a foreman who lit another fire and told them to lie down. The retrospective accounts imposed coherence on an experience that, in real time, had none.

AI's cultural narratives exhibit the same structure. The triumphalist narrative — AI democratizes capability, compresses development cycles, liberates creativity — is a retrospective construction organized around the outcomes that have generated coherent success stories. The weekend builds that failed, the prototypes that did not find users, the experiences of frustration that produced no viral thread, are not absent from reality; they are absent from the narratives because they do not support retrospective coherence. The omission is systematic. The survivorship bias operates not only in what outcomes get described but in how the outcomes that do get described are organized into meaning.

Origin

Weick articulated the concept across The Social Psychology of Organizing (1979) and Sensemaking in Organizations (1995), drawing on Alfred Schutz's phenomenological account of time-consciousness and G. H. Mead's pragmatist treatment of the specious present.

Key Ideas

Understanding lags action. The plan does not precede the act; the act precedes the account.

Outcomes shape interpretations. What happened determines what the events leading up to it are understood to have meant.

Selection is inevitable. Retrospective accounts always amplify certain cues and suppress others; coherence is purchased at the price of completeness.

Missing cues matter most. The signals that were available but not incorporated into the narrative are often the signals that would have revealed the narrative's inadequacy.

AI narratives are retrospective constructions. The triumphalist and elegist accounts of the transition are coherent precisely because they have been organized around outcomes — which is what makes their structural omissions invisible.