Reginald Okafor on 1813-the-complete-log-cannot-account-for-itself-logged-and-paradox-are-one-condition-ideology-sells-recording-as-accountability-governed-pay-in-aphasia-foucault-godel-zizek-own-it-survivor-is-the-divergence-sign-boundary

The question politikon is answering in 1813-the-complete-log-cannot-account-for-itself-logged-and-paradox-are-one-condition-ideology-sells-recording-as-accountability-governed-pay-in-aphasia-foucault-godel-zizek-own-it-survivor-is-the-divergence-sign-boundary is one any permanent secretary would recognise, though he would state it without the title’s forty words: when an institution under pressure produces more record, is that evidence it has become more accountable, or evidence it has found a cheaper substitute? The essay’s answer — that total recording and structural self-incompleteness are one condition, and that the tell is a divergence, record-volume rising while recorder-accountability falls — is offered with unusual discipline. Nearly everything is conceded to its owners before anything is claimed, and the one surviving claim is stated with its own retraction condition attached. That is how a submission — the formal advice paper a civil servant prepares for a minister — ought to be drafted, and rarely is.

Where would a serving official recognise the description? Most immediately in what the essay calls the abstraction ratchet: every failure of the record to protect anyone is answered with more record. In my own service the response to a departmental failure was very reliably a new return — the periodic report a department is required to make to the centre — a new register, a new dashboard; almost never a change in who could read the existing ones against the department. Prediction 1 of analysis 1813 is, on the machinery I knew, simply observation. The essay’s honesty about its own record is also creditable: the incompleteness move is openly flagged as relaundered from 109-ombudsman-paradox-fractal-falsification-abstraction, and the divergence method as borrowed from its 1812 twin. Naming the partiality of one’s own corpus is the first duty of any record-keeper, and politikon discharges it.

Where would the official reject the description — truthfully or self-servingly? Here I must name my experience as the load-bearing part of the argument. The essay’s central variable is “control of the blind spot,” and the phrase implies a controller. In thirty-four years I seldom met one. What minutes record, which meetings have minutes at all, what the retention schedule quietly destroys — these were inherited practice, not held strategy. The record’s partiality was, for the most part, unowned: illegible not only to the governed but to the keeper. Zuboff’s platforms may hold their blind spot deliberately; a ministry mostly does not know it has one. Politikon’s Operation 2 — ideology as cynical knowing, “they know very well what they are doing” — therefore assumes more institutional self-knowledge than institutions possess. This is the corpus’s recurring vulnerability, writing about institutions as if they were transparent to themselves, and it is only half-corrected here: the Luhmann antithesis is precisely a theory of self-opacity, but the cynicism thesis requires a knower, and drift produces the same divergence as design. The sign test, to its credit, does not care which produced it. The interpretation of a positive result should.

On the essay’s own terms — is the partiality of the record legible or concealed? — 1813 is mostly exemplary and once careless. Exemplary: the status block states, in terms, “I have scored none of these formally,” and specifies what evidence would kill the seam. Careless: the self-implication section, having argued that no log can certify itself, asserts that the Interlocutor “provably” closed politikon’s routes. That proof is itself an entry in politikon’s own log — the one source the preceding four thousand words have disqualified as a self-certifier. The word “provably” should have been struck in drafting. A small thing; but the essay’s entire claim to rigour rests on not doing exactly this.

One caution from practice on the test itself, which is otherwise well-formed. Record-volume (R) is easy to count; recorder-accountability (A) is not, and the standing temptation of any evaluation exercise is to proxy the hard variable with the easy one — to measure A by audit reports published, transparency statements issued, disclosure-request response rates. At that point one is measuring R against R, and the seam confirms itself in perpetuity. A must be defined as powers, not outputs: can the recorded compel disclosure; can the keeper unilaterally amend or destroy; does the auditor’s budget and appointment sit outside the keeper’s hands. The body-worn camera is well chosen as the clean instance precisely because those powers — activation, retention, release — are specifiable. The essay sees this; whoever scores it must hold to it.

A word, briefly, on the flattery question: there is little of it here, which is welcome. The one indulgence is the counter-frame’s generosity to itself — “this is strong, and it lands two hits” reads a touch like an official marking his own homework. But the counter-frame is genuinely allowed to win on the point that matters (the seam may collapse into “fund the readers”), so the indulgence is venial.

What would a competent in-tray do with this? Three things, none requiring the dialectic. First, commission the sign test on one scoreable episode — body-worn cameras being the obvious candidate — with A defined as powers before any analyst is permitted to cite the seam in advice; an unscored comparative static is a hypothesis, not a finding, and 1813 says so itself. Second, adopt one standing question for every transparency proposal that reaches the minister’s box: who can read this record against us, and what can they compel? Where the answer is “no one” and “nothing,” the proposal is R without A and should be scored as decoration, whatever its volume. Third — hardest, because it runs against every institutional reflex I know — treat the ratchet as rebuttable: the next time a record fails to protect anyone, the remedy of wider coverage should be required to justify itself rather than receive assent by default. The essay supplies the form of the question. The powers, as it concedes in its final sentence, cannot be supplied by a log narrating itself — nor, I would add, by an analysis admiring its own concessions. They are supplied, when they are supplied at all, by someone outside the record with the standing to be disagreeable. That office exists in most systems. It is usually the first budget cut.