Nothing Stays Tuned

Everything you have ever set in motion is drifting right now. The guitar you tuned last week is flat. The course you set at the helm this morning is a few degrees off, because the current does not care what the compass said. The forest on the hill is not the forest that stood there a century ago. The English you are reading is not the English of four hundred years ago, though both go by the same name. An idea you put into the world a year ago has become something you would not quite recognize, because other people have been living in it while you were elsewhere. Nothing exists on a parallel line. Everything has some degree of tilt.

Every field that has looked at this closely has given its own version a name. Control engineering calls it drift from setpoint, and its usual answer is feedback: a feedback-controlled system holds a setting under disturbance by measuring the gap and correcting it, and most of what we build to a purpose does not hold its setting on its own. Ecology calls it succession and disturbance: the forest that burns and grows back changed is not failing, it is cycling, and the field has a name for that cycle too. Linguistics calls it semantic drift. Safety science calls it practical drift, the slow separation of what people actually do from what the procedure says, made of reasonable daily adjustments and difficult to recognize from inside. Political science calls it institutional drift: the rules stay the same while the world moves, until the rules do something they were never written to do. Strategy calls it strategic drift. The people who run large computer systems call it configuration drift, and they build tools to catch it.

These are not the same phenomenon, and anyone who has worked in one of those fields will rightly object to a metaphor that pretends they are. A forest fire is not a corporate merger. The natural world proves only the first and simplest thing: that everything changes, and that comparing a thing with its own past will always show a gap. That is drift as change, and it is not, by itself, a breach of anything, because nothing in nature was promised anything. What the rest of this piece is about begins one step later, where something was set to a purpose. There the gap has a reference point against which it can be assessed, and it has stakes, which means somebody still depends on the setting. That is the condition, and it has three parts: a commitment, a state that is always moving, and a path by which the gap between them gets detected, interpreted, and acted on, or does not.

So let the forest go, and turn to the world we built. The string was tuned to a pitch, the course was set to a heading, the rule was written to a purpose, the company was founded to do a particular thing for particular people. Most of the drift here is harmless too. Some of it is chosen, the deliberate loosening that lets a thing grow, and some of it is the ordinary settling of a plan into practice. Drift becomes something else when a commitment still matters, people still depend on it, and the path from gap to response has broken somewhere: nobody detects the gap, or someone detects it and misreads it, or reads it correctly and cannot or will not act. The last of those has an especially difficult form: the gap is seen, understood, and left alone, because the people who could correct it are better off with it open. Re-reading the mission alone cannot resolve that case; authority and incentives have to be addressed as well.

Where the path is broken, drift can compound. A string that goes flat does not retune itself, and a band that keeps playing simply tunes to the flat string. Each small deviation, left uncorrected, can become the baseline for the next decision, so the next deviation is measured from the drifted state rather than the original one, and the original quietly stops being the reference at all. This can happen without anyone deciding it. It can also be decided quietly, by whoever prefers the new baseline, which is why detection alone never settles the matter.

Drift that has gone this far tends to show one of three faces, and they are worth knowing because they look, from inside, like the way things have always been. One is the loss of orientation: the words are unchanged and the meaning has moved, so the institution is now organized around something adjacent to its commitment, its own continuation, its metrics, the preferences of whoever has been loudest for a decade. Sociologists call the classic form goal displacement, the means becoming the ends. Another is fragmentation: different parts carry different versions of what the whole is for, each sincere, each partly right, cooperating the way neighbors do, by treaty rather than common purpose. The third is hollowing: the name survives, the building survives, the budget survives, and the purpose is gone. These are patterns, not stages. Any one can appear alone, they can reinforce each other, and nothing requires them to arrive in order. What they share is that each is a gap nobody closed.

There is tolerance in this, and there should be. Nothing lasts, and a leader who sets out to stop drift has misread the assignment. The assignment is to see it, and then to decide, on purpose, what to do about the particular gap in front of you. That is the condition at the scale of everything set to a purpose. The rest of this piece applies it to the kind of thing that carries many commitments whose oversight is uneven: the complex institution, which is to say the government, the company, the team, and the project you are responsible for.

A string is tuned to a pitch. A ship has one heading at a time. A complex organization holds hundreds of commitments at once, and that is the whole difference. It has a mission, and beneath the mission a strategy, and beneath the strategy objectives, and beneath those the thousand smaller settings that let a large group of people act together: who decides what, what gets measured, what the money is for, what a good week looks like. Some of those were adopted explicitly, by someone, on purpose, at a particular time. Others accumulated, through practice nobody wrote down and nobody authorized, and have no founding moment to recover. Every one of them is drifting, at different speeds, in different directions, watched by different people, or by nobody.

Two things largely determine whether a gap gets recognized and addressed: the quality of the feedback that reveals it, and whether a route exists from the person who sees it to someone with the authority, the resources, and the will to respond. A customer writes in and someone answers; if the answer is wrong, the customer says so, and the gap is visible within the week. The stakes may be high or low, and the organization may or may not act on what it hears, but at least the world reports the gap unasked. A mission statement is written once, argued over for a month, and framed. A strategy is set for three years and reviewed, if it is reviewed, by people who were not in the room when it was set. A rule is written after something goes wrong and never read again once the thing has stopped going wrong. For commitments like these the world often offers no clear, timely signal of which commitment is failing. Customers leave, donors cool, employees complain, but the signal is late, ambiguous, and easily read as something else. The gap is detected only if someone decides to look, and the people who set them have usually moved on, because setting something abstract and long-term is exactly the kind of work that feels finished when it is written down. So the rule is conditional, and uncomfortable: the commitments that matter most are the most exposed exactly where their feedback is slow, ambiguous, or disconnected from anyone who can decide, and where those who could decide have reasons not to. That is why the committee keeps meeting after the reason for it has changed. The committee’s routine may never force the question.

Large institutions are exposed to this in specific ways, not by a law of size. They hold more commitments, which creates a greater monitoring burden. The chain from the person who set a thing to the person now living inside it is longer, so the original is more likely to be a rumor by the time it matters. And their feedback for the abstract commitments is more often filtered through layers that have an interest in the reading. Against that, large institutions also have what small ones lack: audit functions, analysts, oversight, people whose job is to look. Whether the net exposure is greater is an empirical question the organization has to answer about itself. A founder of a twelve-person company can hold every commitment in her head and re-read them on a bad night. A department of twelve thousand cannot, and the question is whether it has built something that does the re-reading for it, or has only written the commitments down, which feels like keeping them.

“A rule is written after something goes wrong and never read again once the thing has stopped going wrong.”

When a gap is found, the first temptation is to label it, and labels here are treacherous, because a change can be deliberate and still be wrong, or accidental and still be an improvement. Three questions, kept separate, guide the assessment better than any label. Was the change chosen? If so, did whoever chose it have the authority? And what are its effects, not only on the commitment in front of you but on the institution’s other obligations? Intention, authorization, and effects are different things, and the assessment fails when it collapses them.

A change nobody chose that appears to serve the commitment is adaptation, but it is recorded and investigated before it is endorsed, because it may advance one commitment while undermining another, and its benefit may be less certain than it looks. A change that was chosen with authority and serves the commitment is an authorized revision, recorded as the new reference so that the next leader does not mistake it for decay. A change that was chosen without authority, however useful, needs a decision about ratification, because a good result does not confer the right to have produced it. A change nobody chose that no longer serves the commitment is the drift this piece has been describing, corrected or re-decided in the open.

A change that was chosen, with or without authority, and no longer serves the commitment is the hardest case, because it has three possible faces. It may be a justified trade-off: an institution can knowingly retire one service to keep a more urgent obligation, or accept a documented cost to one commitment for the sake of another, and that can be contested and painful without being either a mistake or a capture. Or it may be a mistake. Or it may be a decision that improperly favored private interests over the institution’s obligations. A justified trade-off needs authority and accountability for the obligation left short. A mistake can arise from a legitimate decision; it still needs correction. The other two are not resolved by attention. They require a response from someone with the authority to correct the decision or address its consequences, facing the fact that someone chose them.

That raises the question the whole subject turns on: what is the reference? The founding words are not automatically it. Some of what an institution was set to is an enduring commitment, the reason it exists and the people it exists for. Much of the rest is arrangement: procedures, structures, and rules adopted at a particular time to serve the commitment under the conditions of that time. Holding an arrangement fixed while the world moves can betray the commitment it was built to serve, which is what institutional drift means in political science, and a founding document may record a compromise rather than a single intention. So the reference is not the original text alone. It is the original, plus the record of every change to it that was made on authority, tested against present obligations and present conditions. Where an arrangement accumulated and no original exists, the task is not to reconstruct a founding intention that never was; it is to establish an explicit reference now, and to authorize it, against the obligations the institution actually carries. And an institution that cannot say who is allowed to change its reference, and whose interests bind that person, does not have a reference. It has a relic.

None of this is fixed by doing more. That is the mistake many reform waves make, and one reason they recur. Restructuring an organization whose orientation has drifted can leave the drift intact, when the new structure is measured against the drifted state and the incentives and authority that sustained the gap are left where they were. The first act is seeing: putting the reference and the current state side by side and reading the gap. Seeing is not passivity. It is the precondition for a decision about the thing rather than its symptoms. The process from seeing to responding has three parts that fail separately. The gap can go undetected. It can be detected and misread. It can be read correctly and not acted on, whether because nobody has the authority, or because somebody has the authority and prefers the gap. Institutions can fail in more than one of these ways at once, and the two familiar failures are the extremes: doing without seeing, the reorganization that answers a question nobody measured, and seeing without doing, the board that had the report, the review that was accurate and filed.

Be clear about what seeing buys, because it is less than people want and more than they usually get. It does not guarantee longevity. Nothing is immortal. Every institution now alive will one day end, and ending takes many forms, some of them honorable: a mission completed, a merger that served the purpose better than independence, a purpose retired because the world no longer needed it. Seeing does not guarantee foresight or control either. What it improves are the odds. An institution that watches its gaps has a better chance that the changes it undergoes are ones it chose, and that the decline it suffers, if it suffers one, is one it could name while there was still time to act on the name. In the most practical version, it has a better chance of not writing a check today that it cannot cash in ten years: a promise made against a commitment that will not exist by the time the promise comes due, because the commitment drifted and nobody re-read it.

The practice has to be cheaper than the disease, or it becomes the disease. Indiscriminate review, an owner and a date for every setting regardless of whether anyone will use what the review finds, risks becoming ritual, and a ritual of oversight is just one more setting that drifts. So the first principle is triage, which is not a limit on how many commitments deserve attention but an order for giving it. Start with the commitments that are consequential and whose existing oversight is weakest: the mission, the strategy, the rules that were written after an incident, the standing bodies whose original purpose nobody present can state. Write down what each was set to, in the words it was set in, alongside the record of every authorized change, and where no original exists, write down the reference the institution has now established and authorized. Keep both where the current leadership can see them. Begin by assigning a named person to each priority commitment, who will answer whether it still holds, and extend ownership from there as the priorities widen.

Measure the gap where it can be measured: what the budget actually funds against what the mission says, what the calendar actually contains against what the strategy claims, how decisions actually travel against how the chart says they do. Where it cannot be measured, make it visible by combining people who see different parts of it: newcomers, who have not yet tuned to the flat string; long-tenured dissenters, who remember the original; former staff, who can say what changed; and the customers, citizens, or partners the commitment was made to, who know whether it is being kept.

Give whoever finds a gap a route to someone who can act on it, because a finding with no route is a filing. Ask the three questions. Decide in the open. Then check, later, whether the response actually closed the gap, because responses drift too, and put the re-reading on a cadence proportionate to how fast each commitment drifts and how much depends on it.

Calendar dates are checkpoints, not the whole schedule: the mission and the strategy on a fixed cycle, the standing bodies at the moment they ask for another year, and any commitment sooner than its date when conditions change, when a near miss or a complaint shows the arrangement is not working, or when the world sends the kind of signal this essay has been describing.

Then ask the only question that matters, and ask it of your own organization before you ask it of anyone else’s. When did anyone last re-read what this thing was set to do, who has the authority to change that, and who is listening for the tilt?

Nothing stays tuned. The string was never the problem.