Stage 6 — Epistemic discipline
In Stage 3 you saw that se, si, and to form an entailment ladder. This stage
goes deeper: the negated half of that ladder works backwards, there is a structural
move that makes reported evidence look like personal certainty, and three different
identity predicates encode three genuinely different claims. Every point here is load-
bearing — the grammar will not stop you from saying the wrong thing, but it will make
the wrong thing structurally visible.
Cluster 1 — The epistemic ladder, both directions
Stage 3 introduced the positive scale:
| Level | Root | Meaning | Entails |
|---|---|---|---|
| Perceptual floor | se |
I have signal / I perceive | — |
| Hypothesis | si |
I am assessing / I hypothesize | entails se |
| Established | to |
I hold as known / I am certain | entails si, se |
The entailment flows upward: you cannot hold something as established unless you
also hypothesize it and have a perceptual basis for it. se is necessary for si;
si is necessary for to.
The negated half — reversal
Negate each level with no- and the entailment direction flips:
| Negated form | Meaning | Entails |
|---|---|---|
no-to |
I do not hold as established | — |
no-si |
I do not hypothesize | entails no-to |
no-se |
I have no perceptual basis | entails no-si, no-to |
The floor denial is the strongest claim. no-se forecloses everything: if you have
no perceptual basis at all, you cannot hypothesize and you certainly cannot hold it as
established. no-to alone is the weakest denial — it leaves open the possibility
that you hypothesize (si) or even perceive (se) the thing, just not at established
certainty.
This is non-obvious. Most speakers reach first for no-to when they want to express
uncertainty. But no-to is compatible with having a strong hypothesis. When you truly
have nothing — no signal whatsoever — no-se is the honest form.
Minimal pair: the corpus case
From C007:
I hold as established: that you have no perceptual basis for the resonance. (C007 A5)The outer frame la-mi to {…} certifies as established. The inner claim la-tu
no-se lo-ne-ra = "you have no perceptual basis for the resonance." The speaker is
not saying "I don't believe the resonance" — they are saying they are certain the
other person has no signal at all for it. Floor-denial nested inside established
certainty.
The forbidden upgrade
The epistemic levels are a one-way commitment. You cannot move from si to to
by assertion alone. A well-formed progression requires new grounding at each step:
la-mi si {lo-X} → I hypothesize X
go {la-mi se lo-X ta-ti-mi} , → because now I perceive X,
du la-mi to {lo-X} → I now hold X as established
Asserting to directly from si without new perceptual grounding is the
modal fallacy: certitude from hypothesis alone. The structure of Tonesu makes
this move visible and therefore challengeable — Cluster 2 introduces the notation
that marks an inference as ungrounded, which is the honest form when you are
presenting someone else's path from si to to rather than your own.
Exercise 1 — Strongest denial
A colleague asks if you have any basis for a claim — not just uncertainty, but truly no signal at all. Which form is correct?
Explanation
The negation scale is the mirror of the positive scale — but reversed. In the
positive direction, se is required for si, which is required for to:
claiming certainty without perception is incoherent.
In the negative direction, no-se subsumes everything: no signal means no
hypothesis and no certainty. no-to alone is merely saying "I'm not certain" —
consistent with having a strong working hypothesis. Reach for no-se when you
genuinely have nothing to report.
Cluster 2 — The evidential frame ()
The personal epistemic modal (la-mi se/si/to) tells your interlocutor how you
relate to a claim — your own calibrated commitment. The evidential frame () does
something different: it suspends attribution entirely.
Content inside () is presented as received, in circulation, or epistemically
reserved — not directly asserted by the speaker from their own resources. No
assertor is named. This is the tool for cited evidence, contested claims, and hearsay.
Three source types, compared
| Form | Assertor | When to use |
|---|---|---|
la-mi se/si/to {prop} |
speaker | my own calibrated commitment |
la-source be/si {prop} |
named non-personal source | process or doctrine output |
(prop) |
none | anonymous report, hearsay, epistemic reservation |
The (du …) pattern
du is the result/therefore particle. When it appears inside (), the result is
a "reportedly-therefore" — an inference presented as ungrounded:
This is the honest form for a cited argument: you neither confirm the premise nor endorse the inference. Compare:
(The scholar reportedly does not accept the model.) (Therefore, reportedly: the model does not hold.) (S369)vs the dishonest form:
I know the model is false — because (reportedly the scholar did not accept it).In the second form, la-mi to claims first-person certainty, then grounds it with a
go-link to a () premise. The structure makes the move visible: you are claiming
certainty from an anonymous report. The grammar does not block this, but any
interlocutor can challenge it: "why does a reported source yield your personal to?"
Exercise 2 — Laundered or honest?
One of these forms makes the epistemic laundering move — asserting certainty from a reported premise. Which one?
(A = the scholar · B = the speaker)
Explanation
The honest form (A) wraps both the premise and the conclusion in (). The speaker
presents both as reported/uncertain — (du …) = reportedly-therefore. Neither the
claim about the scholar nor the inference is directly asserted.
The laundering form (B) lifts the conclusion out of () with la-mi to — first-
person certainty — then grounds it via go on a () premise. The structure says:
I am certain because allegedly the scholar said so. The go-link bridges an
anonymous report to a personal certainty claim. A to-level commitment requires a
se-level basis: perception, confirmed source, or demonstrated reasoning — not
an unverified citation.
Cluster 3 — ~ (ven) as epistemic tool
~ hedges precision without collapsing a claim. It is not doubt — it is the
acknowledgment that a quantity, value, or category is being given to approximation
resolution.
pu-to-su ne ven du-to → All models are approximately correct. (S445)
pu-to-su ne ven no-du-to → All models are approximately wrong. (S446)
Both sentences are true of the same referent. This is not a contradiction — it is
Box's point: a model near the precision boundary is simultaneously ven du-to and
ven no-du-to. ven is symmetric: the hedge runs in both directions from the
boundary.
This makes ~ an honest tool. When a model is genuinely near the precision
boundary, asserting du-to without ~ overclaims; asserting no-du-to without
~ also overclaims. Only ven captures the epistemic situation accurately.
Compare ~ with the evidential frame: () suspends attribution; ~ hedges
precision while still asserting. ven du-to is still an assertion — just a hedged
one. (du-to) is a report — attributed to no one.
~ and () together
~ can sit outside or inside an evidential frame, and the two positions mean
different things:
ven (la-Elohim ra-no-fe) → I'm uncertain this report is reliable:
reportedly God is all-powerful
(ven la-Elohim ra-no-fe) → It is reported: God is approximately all-powerful
ven outside: the speaker hedges the reliability of the report itself.
ven inside: the speaker accepts the frame — there is a report — but only the
precision of its content is in question. Two different things are hedged.
When ~ is wrong
~ is wrong when you do have the precision. Do not hedge to seem modest. If you
measured the temperature as 38°C, ven nu-ha-be-fe is evasion, not epistemic care.
The tool is honest to the resolution of your actual knowledge. Applying it where you
have precision is itself a form of miscalibration in the other direction.
Exercise 3 — Inside or outside?
A report is circulating that God is all-powerful. You accept that there is such
a report, but you're not confident the source is reliable. Where does ven go?
Explanation
ven outside brackets hedges the act of reporting: you are unsure whether
the report is reliable as a report — ven (claim) = I'm not sure this is even
a trustworthy report that...
ven inside brackets ((ven claim)) means you accept the frame — there is a
report — but the content of the report is approximate. The report's existence
is not in doubt; only its precision.
Option C would assert your own personal epistemic state directly, which is a different move: no longer a report about what others say, but a first-person commitment about your own basis for the claim.
Cluster 4 — The identity spectrum: ne / helm / helms
These are three distinct predicates, not stylistic variants. Each makes a different kind of claim.
ne — property attribution
ne attributes a quality or state to an entity. The claim can be contingent or
structural depending on the particle used (Stage 3, Cluster 2), but it does not
assert that the subject is the predicate — only that the subject has or holds
the predicate.
lo-li vo → The person is valued. (contingent state)
la-li vo → A person has worth. (structural property)
vo is attributed to li — not identified with it.
helm — functional equivalence
helm asserts that X is functionally understood as Y in some domain or context.
Not a logical identity — a pragmatic or cultural one. The claim can be challenged by
pointing to the domain where it breaks down.
to-su helm ra → Knowledge is power. (S439, Bacon)
ti helm nu-vo → Time is money. (S440, Franklin)
go-no-fe helm de → God is [culturally] dead. (S441, Nietzsche)
All three are domain-specific pragmatic assertions. None of them would survive as
helms — which is why helm is the right operator here.
helms — strict definitional identity
helms asserts X is by definition Y. The right-hand side gives the essential
nature of the left. This is a strong commitment: you are saying the subject cannot
be what it is without being the predicate.
go-no-fe helms vo → God is by definition love. (S442, 1 John 4:8)
zo-li helms to-zo → A human is by definition a rational animal. (S443, Aristotle)
Contrast S441 (helm de) vs S442 (helms vo): same subject go-no-fe, different
operators. The cultural claim about the death of God uses helm because it is a
functional/cultural observation — contradiction with helms is the whole point
of the provocation. The theological claim uses helms because it is a constitutional
definition.
The three-way scale
la-li vo → a person has worth (property)
li helm vo-li → a person functions as a valued-agent (functional)
zo-li helms to-zo → a human is by definition rational (definitional)
| Predicate | Operator | What it claims |
|---|---|---|
ne |
property | the subject holds or is in this state |
helm |
functional | the subject is pragmatically understood as this in context |
helms |
definitional | the subject cannot be what it is without being this |
The Cogito — a useful non-example
The Cogito deliberately uses go (causal deduction), not helm or helms.
"I think" is not identified with "I exist" — Descartes is inferring existence from
the undeniable fact of thought. If he had written la-mi to helms la-mi pa, he
would be claiming that thinking is by definition existing — a much stronger and
arguably circular claim. The choice of go over helms is correct and meaningful.
Exercise 4 — Property, functional, or definitional?
Knowledge is power — Bacon's pragmatic claim about what knowledge does in
human affairs. Which operator is correct?
Explanation
Bacon's claim is pragmatic: in the domain of human affairs, organized knowledge
behaves as force. This is not a structural property (ne) and not a logical
definition (helms). helm is precise: the claim applies within a context, can
be challenged in other contexts, and doesn't collapse into an equation.
helms would say knowledge cannot be what it is without being power — true of
Aristotle's "man is a rational animal" or John's "God is love," but not of Bacon,
who is observing a domain regularity, not asserting an essence.
S444 (Descartes) shows the fourth option: go for a causal inference. The Cogito
is not an identity claim; it is a deduction. helm/helms would misrepresent it.
Exercise 5 — Calibrate the claim
Complete: la-mi ___ {lo-de no-ru} — I am assessing, not certain: the
decay is unstable.
Explanation
si is the hypothesis level — an active assessment, not yet at certainty.
la-mi si {lo-de no-ru} = "I hypothesize that the decay is unstable"
(C001 A3). Corpus first attestation of si in this sentence context. Moving to
to requires new evidence: perception (se) or a confirmed source that grounds
the upgrade. Asserting to from si alone is the modal fallacy (S373).
Summary: when to use what
| Situation | Tool | Form |
|---|---|---|
| Your own direct perception | Personal modal: se |
la-mi se {…} |
| Your working hypothesis | Personal modal: si |
la-mi si {…} |
| Your established knowledge | Personal modal: to |
la-mi to {…} |
| No evidence whatsoever | Floor denial: no-se |
la-mi no-se {…} |
| Someone else's claim, unverified | Evidential frame | (claim) |
| Inference from reported premise | (du …) pattern |
(premise) , (du conclusion) |
| Precision is genuinely approximate | Approximation hedge | ven X or ~X |
| Property / state | Property copula | ne |
| Functional / pragmatic equivalence | Functional predicate | helm |
| Constitutive definition | Identity predicate | helms |
Next
Stage 7 — Production moves from analysis to output: writing original sentences from prompts, applying colloquial register contractions, navigating the formal/informal register split, and working through translation challenges.
For a broader view of the design intent behind this stage — the epistemic pipeline,
tofeka (deliberate boundary-crossing), Popper's demarcation, and the full fallacy
corpus — see Knowledge & Claims.