Under What Conditions

An airspeed indicator reads under conditions. The pitot tube feeding it senses ram air pressure; the indication is derived against static pressure, and it is trustworthy only while the conditions of that measurement hold: probe clear or heated, ports open, airflow inside the angles the installation was designed for. The failures are conditions all the way down. Block the pitot inlet alone and the indication falls toward zero. Block the inlet and its drain together and the instrument behaves like an altimeter, reading confidently and wrongly as the airplane climbs. Performance charts make the same point without apology. They are not descriptions of the airplane; they are functions from conditions to capability, and the same aircraft that lifts off easily at sea level needs more runway than exists on a hot afternoon at altitude. Nothing in flying is true unconditionally. The discipline of airmanship is asking, before believing anything about the machine, the weather, or the pilot: under what conditions?

There is not always one answer. It depends on the conditions, and a pilot's competence is knowing which ones, measuring them, and respecting what they say. The pilot's word for the whole subject is performance, which is never quoted without its conditions.

A structure is designed against a bounded set of anticipated conditions: loads, movements, temperatures, and combinations specified by codes, hazard models, and the judgment of the engineers, with factors addressing the uncertainties they can name. The design basis is set ahead of time, and good practice then returns to the structure across its life: inspection, monitoring, strengthening when uses or hazards change. A roof that serves eight hundred years serves because someone climbs up and reads it. The aircraft has its envelope too, set ahead of time by design and testing. Flight compresses the rest of the cycle into the operation itself: conditions inside the aircraft, outside it, and with the pilot all have bearing on the outcome, none permanent, each able to act on the others, and establishing which ones presently obtain is the flying itself. Institutions are designed like structures and often inspected like them, at intervals, against the anticipated set; they operate like flights, in present conditions the whole time. Drift is this corpus's name for the gap that opens between the anticipated set and the present ones, and the conditions question is how the gap gets measured instead of discovered.

Across the laboratory, the same question returns. What Keeps a Record Standing asks under what structural conditions a verified record retains the capacity to correct the system it describes. The Correctability page's central section asks what keeps the path open from an accurate reading to a changed course. Memory Continuity asks what has to survive across time for an institution to locate its present against its past. Drift is defined against conditions that go on changing whether or not anyone tracks them.

The form is load-bearing. The corpus does not ask whether an institution is good; it asks under what conditions the institution can still be corrected. The second question can be tested. And it is the same question for any institution, any purpose, any administration. The conditions change from case to case. The question never has to.

Across the Neighboring Work sections of this site, each named tradition asks a conditions question about the object it studies: the conditions of sustained adaptation, of enforced safety constraints, of authentic records, of retained organizational memory, of verification's authority, of decisions judged on what was knowable. That shared form is part of why the rooms border each other so naturally, and it is why overlap has so far proved neighborly, not territorial: two fields can study the same object while conditioning different properties of it. This laboratory's object is correction: under what conditions reality retains the ability to correct a complex human system. Whether anyone has taken correctability itself as the object of a conditions science before is a question the site's visits leave open and continue testing; the components come from the neighbors' shops, and the assembly around this object is the work done here.

"It depends" ends a conversation when nobody can say on what. It begins a science when the dependencies can be named, measured, and tested. A performance chart is "it depends" with numbers on it. The corpus's instruments aim at the same standard for institutions: naming the conditions a corrective path depends on, so that "it depends" becomes a reading, and the honest answer to whether a system can still be corrected is a position against measured conditions.

English already has a perfectly good first-order question: what conditions obtain? That one is answered by a reading. Temperature, pressure, density altitude, the load on a roof; instruments exist because the first-order question exists. The question this page is named for takes a different preposition, and the difference is technical. "Under" places something beneath the conditions, and the something is always a claim. The airplane will clear the trees. The record can still correct the system. The institution can still change course. Under what conditions? The question never arrives alone. It takes a whole statement as its argument and asks for the domain over which the statement is true.

That operation happens one level up from observation. A reading is a value; the conditions question asks for the function the value came from, which is what a performance chart literally is: a claim about takeoff with its validity domain published beside it. The opening section made the move twice without naming it, because the airspeed indication is itself a claim, trustworthy under conditions of its own, and the conditions question governs the instruments as well as the world they measure. And because the question operates on claims, it can be asked of any sentence anyone says, including every sentence on this site. A corpus built on the question has to answer to it.

An option search inventories what exists. The conditions question asks what would have to exist, and "would have to" is future grammar: its answer describes a world, whether or not that world has arrived yet. This is what makes the question generative as well as diagnostic. Naming the conditions a solution depends on is the first draft of a specification for producing them.

Everything the question works from is past material, and that is no limitation. The performance chart was compiled from flights already flown; every number on it is a record. But a chart is a function from conditions to capability, and a function accepts combinations that have never once occurred. The pilot who consults it on a hot afternoon at a field no test pilot ever visited is reading a future out of the past, and reading it accurately. Novelty does not require new material. It requires a form that lets old measurements answer questions nobody has asked yet.

Every reader of this page already owns the forward version of the question, under the name a discipline gave it. To the engineer it is design: conditions specified before the structure exists, for loads that have not arrived. To the pilot it is planning: the chart consulted before the takeoff it predicts. To the physician it is prognosis, a function from present conditions to a future the patient has to decide about. To the archivist it is a retention schedule, written against futures a record has to survive. To the founder it is a specification for a world that will admit the thing being built. Each discipline counts its version as ordinary competence. The person out of options owns the question too, and usually does not know it, because no discipline handed it over with a name. The next section is for that reader.

Institutions can be asked the forward version as well. Under what conditions could this institution still be corrected ten years from now is a question about a world that has not happened, answerable with instruments that exist today. The question is as old as engineering, and it faces permanently forward.

"Out of options" is one of the most honest sentences a person under load ever says, and it is almost never a statement about the world. It is a report on the option space visible under present conditions, with the conditions themselves silently assumed fixed. The frame does not announce itself; from inside a problem, the conditions look like the world. That is why the question this page is named for so often goes unasked at exactly the moment it matters most: it is not in the space being searched. It lives one level up, where the search for a solution becomes a search for the conditions under which a solution would exist.

What the question offers has to be stated exactly, because nothing on this site trades in reassurance. Asking under what conditions a solution would exist promises no solution. It corrects a single inference, the one despair depends on: that exhausting every visible option under present conditions proves no solution exists under any. It proves no such thing. The condition-space has not yet been examined, and the examination can end three ways: the necessary conditions already exist and went unrecognized; they do not yet exist but can be produced; or they cannot be produced, or cannot permissibly be, and the honest answer is no.

The third verdict is what keeps the question honest. Elsewhere this site separates manufactured doubt from doubt that has a floor under it. The same floor belongs under hope: hope worth the name is built on a question whose answer is allowed to extinguish it, and that is the whole difference between hope and reassurance. Sometimes the conditions question opens a path. Sometimes it gives a person permission to stop searching. Both are orientation.