The Document’s First Sentence

The mill has been running for more than twenty years. This spring it received a new effluent permit, stricter than the old one. Management issued a document whose first sentence read: “This mill ought to meet the effluent standard and not let anyone downstream be harmed.” A few items followed: the monitoring post samples daily and reports any exceedance within two hours; for whichever shift an exceedance occurs on, the cause is established and entered in that shift’s appraisal.

At the monitoring post is a technician who has done the job for more than ten years. On the shop floor there is a shift supervisor. The household downstream has no employment relation with the mill and has never signed anything with it. They simply share a river.

To whom is “This mill ought” addressed? To the director? He signed it, but when he is transferred the sentence does not go with him. To the monitoring technician? He takes the samples, but “not let anyone downstream be harmed” is nothing he can achieve alone, and nothing he should carry alone. To the mill? The mill is a permit, a set of rules, some positions, a few buildings. It cannot read a document.

It looks like a question of wording. It is the first question institution theory has to answer. To write what an institution ought to do, one must first say to whom the ought is said, where it comes from, and how it reaches persons without legislating for any of them.

What It Binds Is the Post

A foundational paper of this system, One’s Own Law, had already drawn the division needed here. It separates law from ethics. Law is an external construct: “It does not claim to be morality; it simply says: these are the rules, and violating them has consequences.” Ethics is different: it tries to extend a subject’s inner law into a demand on others, and that is a boundary-crossing.

How Is Institution Possible draws the line more finely on the side of institutions. It sets accountability apart from duty, in two columns: source, and what is bound. A duty’s source lies in the bearer’s own law, and it binds the subject himself; accountability is a contentful requirement that falls on a member’s position, and “It binds the position and not the subject itself.” And: “only a subject can bear a duty, whereas both constructs and subjects can bear accountability.”

So what an institution can say to a person becomes clear. It can make demands on his position: sample daily, report exceedances. That is accountability, and it holds the post. What it cannot make is the other kind, the kind that binds him himself, because that kind can come only from his own law. If the mill, instead of saying what his post should do, says how he as a person should be — if it requires him to take a demand from outside as one from within — it has stepped into the place of his own law. That is exactly what One’s Own Law called ethics: “law disguised as morality.” It is legislating for him from outside.

Saying ought to a construct is a different matter. “A construct has no first person,” and it has no law of its own. An ought said to it replaces no one’s law; there is none there to replace. Institution theory has to write oughts and may not legislate for persons. So the oughts of this series take the construct as their normative object.

This is not a convenient exit. The document says “This mill ought,” not “You ought,” but the sentence still lands on persons in the end: the technician has to go out and sample, the supervisor has to run his shift. Calling it a rule of the mill does not clear the account of what it does to people. When someone says “This is the mill’s rule, not me asking you,” it still has to be checked on whom the rule lands, how, and whether rightly: whether an accountability really stands at that post, and whether it crosses that person’s line. Moving the addressee to the construct only means the ought no longer legislates for persons from outside. It is not a certificate of exemption.

The converse holds as well. “Your post must keep honest records” has a “you” in it, yet it speaks of a position; if it meets the conditions of an accountability, it stands. The line is not whether the sentence contains a “you,” but what a demand binds and on what grounds.

On What Grounds

Addressed to the construct is not the same as coming from it. The mill has no first person; it produces no oughts of its own. These oughts come from the people who share this entanglement, and from the entanglement itself. The technician, the supervisor, the workers, the household farming fish downstream are all tangled up with the mill, and the river is where they are tangled.

And on what grounds? The system rests them on a premise it undertakes openly: “A human being is an end in itself.” How Is Subjecthood Possible states what kind of premise it is: “This postulate is a normative premise this paper openly undertakes. It is not derived from the operation of negation, and it does not pose as the result of any physical test.” It can be refused; whoever refuses it founds a branch of his own, and the branch is not struck from the record. But being refusable does not mean that, within this stance, a construct escapes review because it does not accept it.

From “a human being is an end” to “the mill ought…” there is still a step: from a person’s standing to demands on a construct. The postulate does not take that step, and How Is Institution Possible stopped short of it on purpose. There are two such steps, one for how a person treats people and one for how a construct is arranged. The step at the construct this series takes openly, as its own undertaking, and does not pass it off as a derivation. The step at the person it leaves alone, for the Critiques. Neither stands in for the other.

The oughts are addressed to every construct with institutional identity, and that is wide enough to include the bad ones: “slavery, caste systems and colonial administrations all have institutional identity.” The same paper adds at once: “classification issues no legitimacy.” A bad institution is still an institution, so it is still among those the ought is said to; it fails the ought precisely because the ought is said to it too. And the ought does not wait for a construct to start running on the premise that people are ends. Whether or not it runs that way, the ought is there.

Firm Here, Careless There

The mill is not all bad, and not all good. The monitoring technician is firm about one thing: the number is what it is, and no one talks him out of it. In another thing he uses the new apprentice as an errand boy. The shift supervisor often sees people as numbers behind an output figure; yet last year, when a worker’s child was ill, it was he who went and won the man three months’ leave. The household downstream is no different. The same person treats people as ends in one matter and as means in another.

The system calls this mixed reality, and says that “mixed reality is the normal state”; a community where everyone acts by his own law is only an ideal type. So an institution does not presuppose that the people inside it act by their own law. Its presupposition is mixed reality. The mill and the household downstream are tangled together without having first recognized one another, and institution theory stands exactly on that stretch where recognition has not happened yet and living side by side has to go on anyway.

What is mixed must be said precisely. “End in itself” has two senses, standing and realization, and “Standing intact, realization poor, is a combination with content, not one defined away in advance.” The people in the mill and downstream are unevenly realized; their standing is all intact. What is mixed is realization, not standing.

Three things follow. Guarding lines cannot wait until everyone holds one another in non dubito; when it starts is set by events, not by prejudging people’s levels. Cultivation between people cannot presuppose that everyone acts by his own law; that is the business of the fifth essay. And an institution’s recognition of a person stops at the event and does not become an identity. If the supervisor treats people as numbers today, the mill can check him and limit him on this matter. It cannot brand him “a man who only sees numbers.” “This identification is event-level, not permanent classification.” Or, as the upstream proviso has it: “There is no finally condemned evil person, but there are diagnosable evil operations — this paper’s proviso therefore judges operations, not persons.” However accurate such a reading is, it is a judgment of what he did. It never reaches his standing.

One thing is not said here: that the burden of guarding rises and falls with some “degree of mixing.” There is no such quantity.

Three Layers, and Those Without a Post

The document landed in three places. On the mill: this mill ought to meet the standard and not let anyone downstream be harmed — an ought, said to the construct. On the monitoring post: sample daily by the standard, record honestly, report exceedances within two hours — not an ought but an accountability, binding the position; and a position, On Accountability says, is “a single registration event.” When the technician takes up the post, that is a registration, and the accountability lands on it, not on “him as a person.” On the technician himself: he does not change the numbers. Whether that is fear of blame, fear of an accident, or simply his conviction that numbers are not to be changed, the mill does not avow it for him; that he followed the rules does not prove which of these is his own law.

Three layers, three grammars. The construct takes the ought; the position takes accountability; the person takes his own law. To the person, no ought is written.

The household downstream has no box in these three layers. They are not the mill’s people and hold no registered position there. But they cannot get away from that river, and they are already tangled up with the mill. So the interior has to be drawn by entanglement, not by identity. A construct’s interior has two parts: the people in the entanglement it handles, and the people its acts reach who have no way to avoid it by leaving. The technician is in the first part and could resign; the household is in the second. Drawing them inside only means they cannot fall out of the mill’s oughts. It gives the mill no authority to manage them, teach them or redirect them: reach of influence is not reach of authority.

How, then, does the ought land for them? Not through a position of theirs; they do not have to carry an accountability for the mill because the mill affects them. It lands through the mill’s constraint on itself, and through the mill’s other positions that have a source. The five oughts apply to them on their own terms, undiminished for lack of a post; the channel by which the mill is questioned must be within their reach as well. If a higher institution exists with the matching authority, it does its own guarding within its range; but “someone above will handle it” is no reason for this layer to be absent.

Two kinds of “leaving” have appeared here, and they are not the same. The technician can resign his post; that is exiting a registration, and “Exit terminates the answering load the position currently carries” — without erasing what he did or failed to do before. The household cannot leave the river; that is not being able to get out of the entanglement. The first is a matter of accountability; the second is a matter of where the interior is drawn.

Landing on a Post Takes Three Things

How does “not let anyone downstream be harmed” become what the technician does every morning? First, how it cannot land: not as an ought on a person. The document cannot write “the monitoring technician ought to care about the people downstream.” Whether he cares is his own law; if the mill writes it, it is legislating for him.

It lands on positions, and to land it needs three things together. A source: the requirement comes from the permit and the rules made under it, not from some manager’s passing wish. A content: “sample daily by the standard, record honestly, report exceedances within two hours” can be identified; “be responsible for the environment” cannot. And a registration: only when the technician takes up the post does the requirement actually stand at his position — though the registration supplies no content of its own and does not by itself prove the requirement right.

Once it has landed, there is more to check. If what it binds is a person, his position must still be preserved — “as a subject position able to answer, to refuse, and to revoke the registration”; On Accountability calls this the T5 gate. If what it binds is a construct — say, the mill’s demands on an outside testing lab — it is checked against the conditions for a construct’s accountability instead. And passing the gate settles nothing more: “That T5 has not given way is not a sufficient condition for a demand’s being warranted or for its being free of violation.” A demand can meet every condition of an accountability and still cross a person’s line; that is judged separately.

The whole of it comes down to one sentence of How Is Institution Possible: “A construct may require a position to do something; it may not require a person to think something.”

This rule needs special care where cultivation is concerned, because that is where it slips most easily. The mill starts a training program: reading drawings, electrical work. It can register “hold classes on time, have materials ready, adjust shifts for whoever wants to come” as the accountability of some post. It cannot register “get the workers cultivated,” nor make “the workers identify with the mill’s direction of development” a result it can compel. Task output and operating skill are another matter: whether someone can read a drawing, whether a process step is up to standard, can be tested and written into a post’s requirements, and checked as accountability and against the line. The one who teaches may also do it by his own law: he teaches because it is his job and because he wants people to learn. Accountability and duty can sit on the same act; they do not stand in for each other. The mill can neither demand his share of his own law nor certify it for him.

The One Who Does the Work Is an End

The mill is a means. The people who work in it are not. The technician, the supervisor, the man who teaches in the training program do the mill’s work, and what they do is accountability; they themselves are ends. This has two sides.

One side is protection. The mill can establish the content of the technician’s role: what the post requires and whether he did it. From that alone it cannot certify his own law, let alone avow it for him. “An external arrangement can establish the content of a role, but cannot by itself certify what stands within the bearer’s own law.” If his not changing the numbers comes from his own law, it is his. The appraisal can say “records accurately.” It cannot write “out of conscience,” and it cannot require conscience of him.

The other side is not dissolving. That autumn there was an exceedance. Two phrases came easily in the mill. One: “It was the mill’s decision.” The other: “The monitoring post reported late.” The first hides people inside the construct; the second pushes the construct’s account onto one man. How Is Institution Possible blocks both: “A construct’s name may not be used to shield an individual responsibility that has independent grounds,” “nor may a construct’s failure be transferred automatically onto some member.” The mill has its account; the one who signed, the one who opened the valve, the one who should have reported and did not, each has his own. None of them settles another’s.

So “I was only following the rules” fails on both sides: it cannot make the one who carried them out vanish into the construct, and it cannot make the construct vanish behind him. The fine grain of that sentence belongs to the next essay.

The Mill Does Not Open the Valve

All along this essay has spoken of “the mill’s acts”: the mill discharges, sets rules, runs a training program. But the mill does not open the valve. Power Theory has a hard sentence here: “Apparatuses do not wield power; individuals invoke apparatuses to wield power.” The mill discharges because a dispatcher opened a valve and a director signed a production plan. The construct is the address of attribution, not the subject of the act; the address and the people who acted are both there. That is where the two sentences of the previous section come from.

Not everything the mill does is a constraint. The textbooks handed out in the training program, the new instrument at the monitoring post, need not act on anyone’s rights. Only the part of the mill’s acts that works as constraint — who must do what and when, what happens to whoever does not, what goes into whose river — is judged by the line that Rights Theory draws between limit and violation. That line is read off structure, not motive. Applying it to a construct’s acts, once they have been traced back to the positions that invoke them, is a step this series takes as its own undertaking.

Is the training program the mill cultivating its workers? Rights Theory’s line between cultivation and colonization was drawn for another person’s questioning, and the mill is not someone questioning someone; it is a construct providing conditions. Carrying that line over takes three things with it. First, refusal and change of direction are preserved: a worker may stay away, may come and then stop, may learn electrical work and go off to do something else entirely, and the program can take none of that away. Second, it is named by the elevation actually achieved, and the elevation is located. After a year, one worker found that what he wanted was electrical work, not a lifetime at the paper machine; he went and got his license on his own and later changed jobs. His own law grew in that matter. Merely picking up a skill need not be that, and where nothing grew it is not called cultivation — though neither the program nor the worker is thereby judged a failure; whether the mill did its part is checked separately. Third, good intentions, design aims and labels do not vouch for it. However kindly meant, however much the sign reads “Growth Plan,” that proves nothing.

Now the main line can be stated. Toward its internal subjects, an institution holds to anti-colonization and strives for cultivation; it itself remains open to questioning. Holding to is the first two oughts: the mill crosses no one’s line, and it guards the lines between the people inside it. Striving for is the third and fourth: conditions for cultivation, room for people to cultivate one another. Open to questioning is the fifth, and also what the first four look like in time. Holding to does not mean the mill can always manage it; whether it can, and what happens when it cannot, is for the theories that carry judgment, and an ought that cannot be met for now is not withdrawn. The “colonization” of the main line is the narrow one, a boundary-crossing disguised as cultivation; the lines that holding to defends are wider. A discharge that presses on the bodies and livelihoods downstream, a configuration that seals off a person’s way of refusing, need not be dressed up as cultivation to fall under the first two oughts.

That mill, that river, will be needed in the essays that follow.