About Cats, Boxes and Enterprise Management Decisions

Three Ways a Fact Can Go Missing – Tegrity.AI

Paper 2 drew a line through the missing information an organization holds about itself: some facts are merely buried and can be recovered with effort; others were never decided, and can only be declared into being. That line is real — but it hides a third case between its two halves, and telling all three apart is what decides whether effort will ever pay.

A cat makes the distinction visible. Keep the physics at arm’s length while we borrow it — the caution near the end is not decoration — but as a picture, it is exact.

The box you can open

The first kind of missing fact is ordinary. The cat is alive or dead; you simply have not looked. The answer sits in a register, a table, a retired system, a departed colleague’s files — determinate, waiting, shut in a box you can still open. Who configured this. What it was for. Effort recovers it, because there is something there to recover. This is the factual case: the retrieval problem, and the one most tools are built to solve.

The box that stays shut

The second kind is stranger, and more common than people expect. The cat is still alive or dead — there is a fact of the matter — but the box will not open. The knowledge lives in people, in routine, in local practice that was never written down, and no inventory or interface returns it. You recover it only from those who have seen the cat: not from what they say outright, which is usually little, but from what they will not say — the hedge, the pause, the glance that answers a question the sentence steps around. This is the tacit case: inference, not retrieval, and slow, partial work. But the essential point holds — the cat is not in limbo. A single true state exists; it is merely hidden. Ask the right people the right way, long enough, and there is an answer to arrive at.

The box with no cat in it

The third kind is the one Paper 2 named, and it is unlike the first two in kind, not in degree. Here the cat is neither alive nor dead. There is no fact to find, because none has been made. «Who owns this application.» «Where does this object begin and end.» «Which source is the real one.» Nobody is withholding the answer; the answer does not exist, and will not, indefinitely, until someone with the authority to do so opens the box and, by the act of deciding, makes the cat one thing or the other. This is the declarative case.

Two things follow, and they are what set it apart. The decision does not discover the state — it creates it, the way «this department now owns X» becomes true because an authorized voice said so, and not one moment before. And the opening is itself an intervention: collapsing the question reassigns budget, boundary, responsibility and risk. A process-mining engine handed a case owner does not find one; the instant the owner is declared, the process it was measuring is already a different process. You cannot read this state without changing it.

Why the middle case is not the last

The costly error is to treat the second and third cases as one — to file both under «hard to get.» They could not differ more in what they demand. For the box that stays shut, effort eventually pays: there is a cat, and patience, skill and the right witnesses converge on it. For the box with no cat, effort returns empty forms forever, because you are asking people to retrieve a decision no one has taken. You can put an army of architects and twenty thousand trainees onto information that was never decided, and the pages come back blank — not from failure, but because there is nothing yet to collect.

Paper 2’s clinching test tells the two apart cleanly. Ask several capable people, independently, to reconstruct the fact. Where a cat exists, they converge. Where none does, ten able analysts return ten coherent, different accounts — not because they failed to find the order, but because each imposed one. Divergence is the signature of the empty box.

The tell — and why it is political

Here the three meet, and become useful together. You do not detect an undecided fact directly; you detect it through the second kind of knowing — through the tacit signals of the people around it. Not by finding the answer, but by finding its structured absence: the place where every witness hedges, where two maps of the same process disagree, where the question of ownership meets a silence no one will break. And the silence has a cause that is not ignorance. To declare is to accept an accountability most would rather not carry; the superposition persists because collapsing it costs the person who collapses it. The signature of the void is not that no one knows. It is that no one will decide.

A caution, kept from Paper 2

It is tempting to promote this cat from picture to physics, and the temptation should be refused. Nothing here is quantum; the analogy earns its place only as a way to hold three cases in view at once. The undecided fact is not indeterminate in nature — it is indeterminate because no authority has yet resolved it. The superposition is institutional, not physical: a matter of who has and has not spoken, not of the world’s own uncertainty. Paper 2 made the same refusal, and it stands.

What it means

Because the third kind cannot be found, the discipline is not to hunt it but to price it — before commitment, not after. How senior a decision must climb before the missing declarations become binding. How much downstream reassignment each one forces. How much value sits in the parts still undeclared. These are measurable, and measuring them is the subject of Paper 5. The cat in the third box cannot be found and cannot be inferred. It can only be priced — and then, by someone willing to carry it, declared.

Sources

Describing vs. doing (constative vs. performative) — J. L. Austin, How to Do Things with Words (1962).
Declarations and institutional facts — J. R. Searle, Speech Acts (1969) and The Construction of Social Reality (1995): the class of utterances that make a fact true by declaring it, under the right authority.
Tacit knowing — Michael Polanyi, The Tacit Dimension (1966): «we know more than we can tell.»
Ownership assigned, not discovered — RACI accountability practice; canonical source and lineage in data-governance practice (e.g., DAMA-DMBOK), where these are decided, not found.
Measuring changes what is measured — the Hawthorne studies (Western Electric, 1920s–1930s); Espeland, W. & Sauder, M., «Rankings and Reactivity» (American Journal of Sociology, 2007).

On terms: declarative is used here in the speech-act sense of Searle’s declarations — a fact brought into being by an authorized utterance — not the cognitive-science sense in which «declarative» means a statable, already-existing fact (which is the factual case above).