Predictive Policing
What Tonesu does to systems that claim to detect danger before guilt is visible.
Minority Report and Psycho-Pass look like two different futures.
- One claims to know a future act.
- The other claims to measure a present danger-state.
But structurally they are the same pressure test: a system takes a signal about danger and tries to treat that signal as enough to justify binding force. Tonesu is unusually good at breaking that move apart.
The shared collapse
English predictive-policing stories usually slide across several distinct steps without marking where the transitions happen:
| Step | Minority Report | Psycho-Pass | What Tonesu forces into view |
|---|---|---|---|
| signal | a forecast of a killing | a high crime coefficient / danger reading | si [PROP] = signal-level holding |
| time | future act | present threshold state | ta-ti-be keeps the forecast in future time |
| classification | future murderer | dangerous person | su-ka marks categorization as an act |
| coercion | detention | detention or execution | ka-ko and ka-de remain separate predicates |
| legitimacy | assumed because the system says so | assumed because the score exists | to-fe-li, to-fe-su, and to-fe-su-ki make authority procedural |
That last column is the point. Tonesu does not merely object to these regimes morally. It makes their hidden category jumps visible.
Minority Report: future act is not present fault
The Minority Report batch pressures the simplest pre-crime mistake:
a predicted future killing is treated as if it were already a present guilt-state.
Tonesu resists that in two moves.
First, the killing forecast stays in future time:
la-o-ka-su si [ta-ti-be la-na Anderton ka-de lo-na Crow]
The system signals that Anderton will kill Crow. It does not yet say that the killing has happened.
Second, the language can hold the forecast beside the target's current status without blur:
... / la-na Anderton no ne de-su
Anderton is not yet at fault.
That seems obvious in English once stated plainly, but predictive-policing stories depend on getting the audience not to insist on the distinction. Tonesu insists on it by grammar.
Psycho-Pass: classification is an action
Psycho-Pass pressures a colder version of the same problem.
The issue is not only whether the signal is strong. The issue is that the system classifies a person from the signal and then treats the classification as if it were neutral fact.
Tonesu has a verb for that step:
su-ka
su-ka means to impose structure or category. That matters because it prevents the regime from hiding behind passive language. The system is not merely discovering a dangerous essence. It is classifying a person into a structural category and then acting on that classification.
Once su-ka is explicit, the chain becomes visible:
- signal a thresholded danger-state
- classify a person
- confine or kill the person
The regime can still do all three. It just cannot pretend they are the same event.
The real fault line: unpublished standards
Psycho-Pass adds a deeper pressure that Minority Report only gestures toward.
What if the threshold that authorizes detention or lethal force has never been openly published in a form that others can challenge?
This is exactly where to-fe-su-ki matters.
to-fe-su= the standards frameworkto-fe-su-ki= the moment that framework enters published, challengeable status
Before to-fe-su-ki, the body has power but not legitimate epistemic force. It may have an internal rule, but that rule has not yet crossed into public, contestable knowledge.
That yields one of the strongest constitutional consequences in the language:
unpublished standards cannot bind.
If a regime enforces an unpublished threshold as if it were settled public knowledge, the problem is not only secrecy. The problem is epistemic inflation. Internal adoption is being treated as certified authority.
That is why the Psycho-Pass batch ends with a to-fe-ka charge at the meta-level.
The Tonesu repair
Across both stories, the repair is the same.
1. Keep signal distinct from fault
Danger-signals may justify attention, warning, or escalation. They do not automatically justify guilt.
2. Keep classification distinct from discovery
When a system categories a person, that is an institutional act. It should be named as such.
3. Keep standards distinct from secrecy
A threshold becomes legitimate only when it has crossed into published, challengeable status.
4. Route uncertainty upward before force is applied
to-fe-li exists precisely for disputed threshold crossings. If the signal is contested, the lawful step is adjudication, not silent conversion into custody or death.
What this reveals about Tonesu
Tonesu is not anti-prediction.
It is anti-collapse.
The language is comfortable with thresholds, warnings, danger-signals, and structured institutions. What it refuses is the compression of all of those into one opaque object called "the system knows."
That is why predictive policing is such a good stress test for the language. It forces Tonesu to show whether its epistemic and legal machinery can survive pressure from the exact kind of regime that benefits from blurred categories.
So far, the answer is yes.