Is Clarity Getting More Expensive?

A conjecture, now testable: that rising system entanglement is driving up both the cost and the risk of knowing how your own work runs — wave after technological wave

The series so far has built a present-tense problem and shown that it can be measured: information about how the work runs is missing and partly un-retrievable (claims one and two); the declarations needed to close the gap are costly and arrive in a paradoxical order (claim three); rebuilding to escape it is a blind bet (claim four); and the cost and the risk of all this can be measured, directly and indirectly (claim five).

Having the measures lets us ask the question the series has carefully avoided until now: is this getting worse over time?

We must be exact about what we are and are not claiming. We do not claim it is proven that organizations’ knowledge of their own processes decays across the years. The honest state of the public evidence is that it shows a durable present deficit and a striking failure to converge despite decades of tooling — not a measured year-on-year decline. What follows is therefore a conjecture — but one we now hold to be testable, and a strong one: that the cost and risk of clarity are rising over successive technological waves, and are borne less and less, so that process knowledge plausibly erodes over time.

The driver is one thing: entanglement. Each technological wave tends to add systems and wire them into what was already there, and — as the earlier articles argued — adding is cheap while removing is costly, so the connections accumulate and rarely dissolve. The dependency graph grows denser wave after wave. This premise is itself measurable, and it is the load-bearing one; the rest follows from it.

The cost, component by component

Recall the three costs from the third article, and ask honestly which way each is moving.

The cost of acquiring information — unclear, and we say so. Here the trend genuinely cuts both ways. There are more and better tools for discovery than ever, which pushes the cost down; there are also more systems to discover and monitor, which pushes it up. We do not pretend to know the net direction. This is not where the conjecture draws its strength.

The cost of the decision — rising, and changing in kind. This is where entanglement bites. The more coupled the systems, the higher a declaration’s effects reach, so the further up the organization the decision must travel to be made at all — the governance bottleneck, and it tightens as coupling grows. But a second, sharper movement layers on top: the decisions that rise are also becoming more technical. There was once a clean division of labour — a technical function settled technical questions, such as which source is canonical and which is not, with clear authorship and lineage. Under heavy entanglement, «which source is canonical» is still a technical question, but its consequences now spread so widely that it can no longer be settled technically; it has to be settled at the governance level. So the governance level is required to make technical decisions, and to understand their technical consequences well enough to make them. The result is a comprehension bottleneck stacked on the authority one: fewer and fewer people must make informed decisions across an ever-wider and ever-more-technical range. That is the comprehension bottleneck in a single image — and the point at which this series begins to touch a second one.

The cost of absorbing the decision — rising with entanglement almost by definition. Once a declaration is made, the organization must absorb its consequences, and the more coupled the systems, the more of the chain a single change forces to adjust. More entanglement, more to absorb.

Two of the three components rise with entanglement; the third is ambiguous. So the net cost of clarity tends upward — and for a specific reason, that entanglement is climbing, not as a vague pessimism about «complexity.»

The risk

The risk moves the same way, and here we are most careful, because the argument is so intuitive it risks becoming a tautology. The thought is simple: a thing that carries value is more likely to be disturbed the more dependencies run through it, because more of the surrounding decisions and chain-effects can reach it; and the more complex the whole, the less anyone controls its effects, so the higher the chance that something escapes control — and that the something is load-bearing. Because this is almost true by definition, we deliberately hold it as a conjecture rather than assert it as a law. That keeps it honest, and keeps it testable.

What we can say, and how it gets settled

So the claim is bounded. We cannot say the information about processes is provably getting worse. What we can say is that there is a strong possibility it is — across technological waves, driven above all by rising entanglement, which makes estates progressively harder to know — because the cost and risk of clarity rise with near-certainty on two of three counts and plausibly on the third, and because that rising price is, on our conjecture, paid less and less.

The reason to state this now, rather than earlier, is that it is no longer only an intuition. The measures from the fifth claim make it reviewable: measure entanglement not by raw graph density — which can fall as an estate grows even while coupling worsens — but by how far a change propagates, how many feedback cycles the dependencies form, and how tightly domains are coupled; measure the cost of clarity as discovery effort and the declarative-to-informative ratio; and compare estates that have lived through many technological waves against those that have lived through few. If the more-layered estates are not more entangled, or their cost and risk of clarity are not higher, or clarity is being held steady against rising entanglement, the conjecture is wrong. That test does not require waiting a decade; it can be run across a cross-section today.

This is the one claim in the series we put forward as a conjecture rather than a finding — and, like the rest, not as an accusation. It is a structural tendency to be measured, and the fifth claim is what finally makes the measurement possible.

Status: conjecture, not finding. The series establishes a present deficit; this claim proposes — and shows how to test — a trend, without asserting one.