Institutional memory: uses and misuses

Institutions frequently announce that they are preserving their memory. They commission digital archives, record exit interviews and require teams to document decisions. The metaphor suggests a warehouse in which experience enters, is labelled and can later be retrieved intact, yet organisational memory behaves less like storage than like repeated reconstruction.

A record does not carry its former significance automatically. It becomes useful only when somebody in a new situation recognises what kind of question it might answer—and what assumptions accompanied the original answer. This is why organisations can possess extensive documentation and still repeat mistakes: a project report may state that a pilot failed because users resisted it.

Years later, that sentence appears to settle the matter. What has disappeared is the meeting in which 'resistance' meant that the software conflicted with an older process nobody had authority to change. The label survives because it is portable, while the institutional constraint fades because it was awkward, distributed and never owned by a single document; memory has not simply been lost but compressed in a way that favours conclusions over the conditions that produced them.

A record can therefore be accurate sentence by sentence and still mislead as an institutional memory if it removes the sequence of hesitation, conflict and revision through which the recorded conclusion acquired its force. Later readers inherit an answer that appears self-sufficient, although the question to which it responded has quietly changed. The usual response is to demand fuller records, which can help, although completeness is not a neutral cure.

Every additional field imposes work, and workers quickly learn which details the template rewards. A database designed around deliverables encourages people to narrate uncertainty as a sequence of completed steps. It may preserve more words while narrowing the range of experiences that can be described.

Documentation is therefore not merely a container for practice; it is one of the forces that shapes practice, because people begin to act in ways that will later be legible to the system. Tacit knowledge is often invoked as the missing element, and the phrase is useful if it reminds us that judgement is embodied in habits, relationships and timing. It becomes misleading when treated as a mysterious substance that experts possess and novices lack.

Much apparently tacit knowledge can be made discussable, but not by asking an expert to list everything they know. It emerges when an actual case interrupts routine: why was this exception tolerated, why did that warning matter, why was a technically correct solution rejected? The productive handover is less a lecture than a series of jointly examined surprises.

There is also a danger in assuming that memory should always be maximised, because organisations need forms of forgetting. A procedure created after one crisis may become an unquestioned ritual long after the conditions have changed. Archived decisions acquire authority partly because they are archived; their survival can be mistaken for continuing relevance.

Deliberate forgetting does not mean deleting evidence to avoid accountability. It means marking some rules as provisional, setting review dates and allowing obsolete categories to lose their grip. Without such mechanisms, institutional memory becomes a method by which the past votes repeatedly in decisions it no longer understands.

This argument can sound hostile to archives, but the opposite is true: good records make reinterpretation possible. They preserve alternatives considered, dissent expressed and uncertainty remaining, rather than only the official rationale. They also distinguish evidence from the decision that was made with it.

That distinction matters because later readers may face a different risk tolerance, legal framework or public expectation. A useful archive does not command obedience; it enables an informed disagreement with predecessors. The most resilient organisations therefore treat memory as a social practice supported by records, not as a technical asset completed once files have been uploaded; they create occasions for old decisions to be retold by people who interpret them differently.

They invite newcomers to ask naïve questions that expose settled language. And they design documentation with an exit route: a way to challenge, revise or retire what has been preserved. The aim is not to remember everything.

It is to remain capable of learning from what was remembered without becoming governed by its original frame.