Deposited preprint

Begleiten können

Ownness and the Warrant Structure of the Critical Architecture

Cite this as

Montaruli, Alessio. “Begleiten können: Ownness and the Warrant Structure of the Critical Architecture.” Preprint, Version 1.0, 2026. DOI: 10.5281/zenodo.21745017.

The instrument this study uses is the four-part substitutive specification of focal reference established in Across the Categorial Difference, §3; §2.6 depends on it, and the dependency is declared at §10.4. The category whose differentia is at issue is stated in The Who as a Modal Category and developed in Volume I.

---

Status of this study

All parts in prose. Part Two is the argument; Parts Four through Six support it.

One claim in this study is offered at less than the strength of the rest, and it is the largest: that the relation which marks a discourse as ontological is title-holding rather than grounding or dependence. It rests on a check of the relevant literatures and not on a survey of them, it is flagged where it is made, and falsifier 2 turns on it. No citation is outstanding. The apparatus is complete, and the note at the end records what was checked against what.

One absence-claim has been withdrawn. An earlier statement of this argument treated the reading of the Ich denke as a focal term as unstated in the literature. It is Schürmann’s, and §2.0 states the debt before the argument rather than acknowledging it afterwards.

§4.4 restates seven differences whose seats are elsewhere in the series. Each is named at the point of use, and the wording at the seat governs.

How this is written

Technical vocabulary is taught rather than assumed. Each term the argument needs is introduced with the case that gives it content, and used only afterwards. Where a term belongs to the literature it is marked as such and its source given; where it is inherited from the author’s own work, the seat is named.

German is shown rather than reported, because the argument turns on a modal that no English rendering carries without loss. Muß … begleiten können appears in the original wherever it does work, with a gloss in the same breath. The same holds for the two French formulations at §2.5, checked against Des hégémonies brisées.

Two registers are kept apart throughout, and the discipline is stated so that a reader can hold the study to it. Where this study says what an architecture lacks, that is a diagnosis made from the present account’s position and never an exposition of anybody’s doctrine. Expound in his terms, diagnose in ours, and never print the second as the first. And where it says an architecture could not house something, that is never a charge that its author suppressed it: the strain is in the architecture and not in the thinker.

Where a page is cited it has been verified. The Kant is quoted from the Akademie text. All Schürmann citations were checked page by page against the Lilly translation and, for the passages on which the disagreement turns, against Des hégémonies brisées; the two editions run at a constant offset of 105. Four citations remain to be confirmed and are listed at the end rather than left unremarked. None bears on the argument.

Provenance

The doctrine of modal coincidence, the category of the who, and the programme of a-metaphysical ontology are the author’s own, developed in Positive No-Ground: The Common and the Locus of Answerability (2026a) and Ethical Ontology, Volume I: The Modal Coincidence (2026b), both published in full, and stated in The Calculus of Modal Coincidence (2026c).

This study’s instrument is the four-part substitutive specification of focal reference, established in Across the Categorial Difference: Focality, Univocity, Ownness (2026d), DOI 10.5281/zenodo.21684041, §3. §2.6 depends on it, and if that account of focality fails, the demarcation claim goes with it. The dependency is declared at §10.4 so that a reader who wishes to attack the whole rather than the part knows where to begin.

The reading rule under which Kant is treated throughout Part Three is Schürmann’s, and §2.0 states that debt, together with the disagreement, before the argument opens.

What is constructed here is the argument from the double modal at §2.3, the three-level reply to the spontaneity objection at §2.4, the two-horn structure at §2.5, the bivalence finding at §3.2, and the third mode of archic occupation at §2.4-bis. Where the study inherits it says so and names the seat; where it constructs, it says that too.

Abstract

Two readings press on any statement of the differentia of the who-category, and they close each other’s exits. Heard as a condition of the possibility of someone appearing, ownness becomes a transcendental condition. Heard as what cannot be said, it becomes a depth beyond predication. Defending against either supplies the other, so an adequate formulation must refuse both from a single positive clause. This study answers both from one place, and the place is a modal: Kant’s muß … begleiten können at B131.

This paper argues that both readings fail for one structural reason, and locates the reason inside the architecture from which the first borrows its vocabulary. The critical apparatus has two positions for the content of cognition, the what-kind delivered by concepts and the this delivered by intuition, and a third term which is not a position of that order but the one on whose warrant both appear.

From the double modal at B131-132, muß … begleiten können, the paper argues that the relation between the Ich denke and the phenomenon is title-holding rather than causal or constitutive, on the ground that a right which need not be exercised cannot be a production; and it checks a four-part specification of substitutive reference against the passage. That reading is set beside the only other answer anyone has given to the same question, Schürmann’s, on which the können establishes an active relation of the one to the many. The paper argues that both readings are required, since the active reading alone leaves a transcendental reading of ownness a quieter slot to retreat into, and only the title reading occupies it.

Two consequences follow. A warrant structure produces not a plurality of ways of being but a bivalence, titled or nothing for me; and the thing in itself is the residue such a structure requires and cannot describe. The residue is read as accommodation rather than failure: what Kant refused to let the architecture erase, marked unsayable rather than denied.

The paper’s demarcation claim is narrowed against the contemporary grounding literature, where a relation is already proposed as the mark of the metaphysical. What is claimed is not that a relation demarcates but which one does, and the difference is that grounding determines while title-holding does not.

Seven falsifiers are stated, of which the first would defeat the structure outright, and the vanishing test the paper applies to others is applied to itself.

Part One · The encounter, and the statement

1.1 · The datum

A woman comes down the stairs in the morning and is greeted by name.

The greeting is not preceded by an examination. Nobody in the kitchen pauses to establish that the person on the stairs is she, weighs what is available, and then, the identification secured, releases the name. The name is already on its way before she reaches the last step.

Two features of that ordinary scene carry everything below, and both should be stated before any vocabulary is introduced.

There is no mediation. Nothing stands between the greeting and the one greeted, representing her to it. Nobody stops to establish anything, and no step has been skipped.

And nothing is hidden. No inner sanctum stands behind the face, holding in reserve what the greeting failed to reach.

1.2 · The statement, printed whole

The differentia this article defends is stated as follows, and it is printed entire because everything in Part Two follows from one clause of it.

Ownness is not a condition of anything and not a depth beneath anything. It is the completeness of this obtaining within a manner that stays entire. And that is a how, not a what: there is nothing further it consists in, and nothing withheld that a better vocabulary would reach.

Two readings attach to that sentence, and they attach at the approach rather than at the sentence itself.

The transcendental reading hears where the non-possibility falls as a condition, and what holds the possibility open as a horizon. It slips past every guard the account carries, and the reason is worth stating early: those guards were built against terms promoted from above or imported from outside, and a transcendental condition sits on neither side. It sits on the side of the conditions of access.

The mystical reading hears there is nothing further it consists in as apophasis: a ground exceeding predication, a depth that a better vocabulary would approach. It is caught by a guard the account already carries, but only if a reader has not already formed the picture.

1.3 · Why one sentence has to do both refusals

The two readings close each other’s exits, and this is the constraint under which the statement above was written.

Defend against the transcendental. Nothing is constituted, no faculty reaches it, it grounds nothing, it makes nothing possible. Every clause of that defence is a step toward the mystical, since what is left after all the conditioning has been denied begins to look like something that exceeds conditioning.

Defend against the mystical. This is exact, it has criteria, it has published defeaters, it can be shown to fail in a particular case. And the exactness begins to look like a condition-apparatus, since a structure precise enough to be tested is a structure precise enough to be doing work.

A formulation is adequate only where both refusals fall out of the same positive clause.

Here they fall out of a how, not a what. Not a condition, because nothing is conditioned. Not a depth, because nothing is withheld.

1.4 · The refusals are read off, not stipulated

This matters for what an opponent has to do, and it is easy to miss.

Neither refusal is a rule protecting a doctrine. Both are descriptions of the encounter at §1.1.

No mediation is the transparency of the greeting: carriage without thematic arrest, the name reaching her, nobody establishing anything. Nothing hidden is what the predecessor volume performs on the other: releasing the other from the office of ground while keeping everything the other actually does, with no reserve behind the face authorising the face.

So an opponent must show the encounter to be otherwise. He may not merely show an argument to have a gap, because at this stage no argument has been run.

1.5 · Three refinements, or the statement over-claims

Nothing hidden is not everything given. Properly fulfilled who-directed reaching discloses the whole modal coincidence without exhausting the life. What is unhidden is the someone. What remains inexhaustible is the life: its manners, its history, what it will yet do. Collapse the two and the result is either a hidden core, which is the mystical reading arriving by the back door, or the claim that we know people completely, which is false.

No mediation is not no route. Reaching runs through routes: perceptual, bodily, practical, addressive, mnemonic, testimonial, archival. A route is not a medium. It is how this one rather than another comes to be at issue, and it never stands between as a representative. The reaching does not constitute what it reaches, which is precisely what a medium would do.

No mediation is not infallibility. Misreaching is ordinary. A nurse can mis-track an infant and be corrected by the bands. Immediacy is not a guarantee, and the corrigibility is the evidence that something was being reached for, since what answers to something beyond itself does not constitute what it answers to.

1.6 · The control

Everything above describes how ownness is given. It does not establish that anything has been given, and the case that shows the difference is the one the rest of this article keeps returning to.

In encountering the burned canvas there is also no mediation and nothing hidden. And no one is there.

An original painted canvas burns. The most exact copy is then made, pigment for pigment, and the original painted object has not been restored. Nobody thinks it has. The canvas is deictically singular, non-restorable, and indelible in having-been. And it is no one. Nobody stands before the ash and asks who.

So the two refusals describe how ownness is given when it is given. They do not do the disclosing.

That work belongs to the minimal pair and the trial constellation and is not re-run here. Without this control, the statement at §1.2 would look like a who derived from the phenomenology of encounter, which is the first charge a critic will bring and the one this article must not deserve.

1.7 · The two additions, and where they are tested

The two readings can now be stated as what they are.

The transcendental reading adds a condition nobody met. The mystical reading adds a depth nobody found.

Stated after the encounter rather than before it, both are additions to what is shown, and the burden lies with whoever is adding. That is a fair position and it is not yet a strong one, because a burden can be discharged, and a reader who finds the additions natural will discharge it without difficulty.

What follows is stronger.

Part Two argues that there is nowhere to add either, and it argues this inside the architecture from which the transcendental reading borrows its vocabulary. That is not an incidental choice of venue. A reading that hears condition in the statement at §1.2 is hearing a word with a history, and the history has an address. If the apparatus that gave the word its sense has no position for what the word is being applied to, the reading has been answered where it lives rather than where it is convenient to answer it.

Part Three then shows what that same apparatus does with a residue it requires and cannot name, which is the second half of the same finding. Parts Four and Five return to the readings themselves, having by then something better to answer them with than a denial.

Part Two · The warrant

§2.0 · The debt, stated before the argument

Nothing in this part is reached without Schürmann, and the debt should be printed before the argument rather than acknowledged after it.

The reading that the modern hegemonic fantasm is instituted in the synthetic unity of apperception is his. It arrives by way of a chiasm, Kant with Luther, in which an inherited obedience becomes an autonomy; and the finding is stated flatly. Perhaps God, man and world reign supreme in Kant, but self-consciousness, reflected and determined through that relation as a transcendental cause, reigns as the fantasm of last resort. The distinction he draws in the same place matters for everything below: ultimate referents differ from supreme ones through their relational nature. What makes a term ultimate is not its elevation but the relation everything else stands in to it.

The reading of the Deduction as an instituting document is his. He calls it a legislative text, an instituting document for referential consciousness, and identifies what it enthrones: a principle for these principles, of which all that is required in the justification is that it be irreducible and thus ultimate. That last phrase is the archic function stated as a property of the term rather than diagnosed from outside it, and this article does not improve on it.

And the focal vocabulary is his, which should be said plainly because this part uses it throughout. It is his in the French as well as in the translation: pros hen is his own transliteration, used throughout and set beside apperception in the same sentence, and sens focal de l’être is his phrase. He credits Owen’s rendering explicitly. And he attaches to it something the present argument depends on: that in speaking of a focal meaning of being he means to preserve not only the relationality Aristotle established but the fantasmic character Kant detected in it, citing the focus imaginarius. None of that is imported here from Aristotle scholarship. It is inherited.

He asked the question this part answers

The decisive thing is not any of the above. It is that he put the question in the place where this article puts it, on the same sentence, and marked the same word.

Faithful to his habit of saying the essential in the first paragraphs, Kant summarizes this relation in these perplexing words, which open the deduction: “The ‘I think’ must be capable of accompanying all my representations.”

“To be capable of accompanying” , how are we to understand this? It is Kant who emphasizes it. A curious beginning to a document of enthronement. Could not the I-think also be capable of not accompanying my representations?

Every step this part takes at §2.3 is taken after that. The sentence is identified as the load-bearing one. The modal rather than the accompanying is identified as what carries it. Kant’s own emphasis is noticed and reported. And the question is asked in the form that makes it a question rather than a curiosity: what would it mean that the term might also not accompany?

He also sees at once what the question threatens. If the Ich denke may fail to accompany, is one to envisage a focal point fitted with an on-off switch, and would anything in it still be formal? On that reading, he writes, the nomothesis seems stillborn. Two pages later he closes it: with an intermittent apperception there is no more law, and therefore nothing left to legislate over.

So the modal is not overlooked and its danger is not underrated. What §2.5 contests is the answer, and the answer was given in full knowledge of what was at stake in it.

The article’s contribution begins where his answer begins, and not before.

The rule under which he is contested

There is one more debt and it is the one that governs the register of §2.7 and of Part Three.

The precept inherited here holds that metaphysics is phenomenological response given archic form, and that the destitution of a principle may release the phenomenon a thinker preserved in excess of his own system rather than defeat him along with it. Schürmann’s civil servants of humanity are champions of the singular even while they plead for hegemony. On that rule, a thinker who builds an architecture around a principle is not thereby a man who saw nothing. He is a man who found something and gave it the only form his epoch made available.

That rule is what licenses the treatment of Kant at §2.7, where the character is found genuinely reached and genuinely restricted. It is also what licenses the treatment of the residue in Part Three, where the Ding an sich is read as what Kant refused to let his architecture erase rather than as what it failed to say.

So the disagreement below is possible because of Schürmann in a strong sense. The rule that makes it charitable is his, and the position it contests is his.

What is contested, and how much

One clause.

He asks what begleiten können means and answers, at §2.5’s exhibit, that it is an active relation of the one to the many, a spontaneous doing that is absolutely first. This article answers that a right which need not be exercised cannot be a doing.

Everything else stands. The office, the fantasm, the enthronement, the relational character of ultimacy, the focal vocabulary, the identification of the sentence, the isolation of the modal, and the reading rule under which the whole is conducted. §7 sets out what the disagreement costs him and states it as tribute rather than as correction.

§2.1 · Where the homonym holds, and where it does not

One thing has to be settled early or everything below will be read as a quarrel with the first Critique, which it is not.

Kant’s categories are categories of the understanding. They are derived from the forms of judgement, they are concepts by which an object is thought, and Kant nowhere offers them as divisions of being. The point is not new and should not be presented as though it were: it has been stated in print that Aristotle conceives his fundamental principles as expressing modes of being, whereas Kant relates them as modes of thinking. Nothing here improves on that formulation and nothing here needs to.

But the homonym holds only where the word appears.

A pros hen is not a device for ordering concepts. What it orders are categories of being: substance and the rest are said to be with reference to a primary case, and the ordering is of ways in which something is, not of ways in which something is thought. And the two doctrines come apart. Politis and Steinkrüger record that there are philosophers who hold that being exhibits a pros hen structure while declining Aristotle’s doctrine of categories altogether. The structure and the table are separable, and the separation is already in the literature.

Which locates the confrontation somewhere other than where a reader expects it.

The projects share a subject matter after all. Not at the table, but at the term that titles it.

The twelve are modes of thinking and the point about them is conceded entire. The ontological work in the architecture is being done elsewhere, by the term on whose warrant everything in the inventory appears, and that is where the rest of this part looks.

§2.2 · Two positions, and a term that authorises both

The critical architecture divides the work of cognition in two, and the division is exhaustive of cognition’s content.

Concepts give the what-kind. Intuition gives the this. Neither works alone, and the sentence that says so is among the most quoted in the book: thoughts without content are empty, intuitions without concepts are blind. Between them the content of any cognition is accounted for. There is no third kind of thing to be cognised alongside a kind and a this.

The Ich denke is not a third position of that order. It delivers no further sort of content. It does not stand beside the kind and the this, adding a third ingredient to what is cognised, and nothing in the architecture asks it to. What it does is make both possible: there is no objectivity outside it, and nothing counts as an object except through it.

Which is where the question arises that the rest of this part exists to answer. What kind of relation is that? Makes possible covers several relations that are not the same relation, and the difference between them is the whole of the argument. §2.3 takes the sentence in which Kant states it.

§2.3 · Muß … können, and the four parts

The sentence

The claim at issue is not that the Ich denke accompanies all my representations. That would be false, and Kant does not say it: most representations are never thematically ascribed to anyone, and nothing in the architecture requires them to be.

The sentence is:

Das: Ich denke, muß alle meine Vorstellungen begleiten können.

Must be able to. A necessity governing a capacity, and the modal is doubled deliberately. Kant emphasised it himself, which is not a reading imposed on the text but a typographical fact about it.

Read what the doubling does. What is required of a representation is not that the Ich denke attend it, but that there be room for it. No event is demanded. Nothing has to happen with respect to any representation on any occasion. What is demanded is that a certain accompaniment be possible, and a representation admitting no such room would be, in Kant’s own phrase, für mich nichts.

So the requirement falls in a particular place, and the placing is the finding:

The phenomenon’s standing as phenomenon is not a fact about the phenomenon. It is a fact about what is possible with respect to it, and that possibility is held elsewhere.

Which is a warrant. A right that need not be exercised is still held, and it is held by whoever holds it.

(Translation note, and it belongs in a footnote rather than the text. The double modal survives in Guyer and Wood and in Pluhar. Kemp Smith’s construction (not to be confused with Houston Smit, at §4.6) relocates the necessity and obscures whose capacity is at issue, which is why the argument below cannot be run from that rendering.)

The four parts, checked

The instrument is not devised here and should be quoted rather than paraphrased, since everything in Part Two is a use of it.

The companion study distinguishes four readings of what a focal term does, three of which are in the literature already: the derivative case may be understood by way of the primary, accounted for through it, or caused by it. The fourth is the one it establishes:

Focality becomes ontological substitution when the primary case holds the derivative’s title to count as a being on the derivative’s behalf, so that the derivative appears in the ontological inventory on a warrant it does not itself hold.

And it is specified positively, in four parts, rather than as a list of denials:

The structure has four parts, which the tenant already gave us: a title to count as a being; a case that holds it; a case that appears by it; and an appearing-on-another’s-warrant which is neither the derivative’s own claim nor a causal consequence of the primary. Those four are what make this a shape rather than a denial.

Two things about that specification bear on what follows.

The word title is a coinage and is declared as one. The study introduces it deliberately, on the ground that the existing vocabulary of dependence, priority and responsibility has no term for the relation. So when §2.6 differentiates title-holding from grounding, it is not disputing anybody’s use of a term of art. It is proposing a relation for which no term was available.

And the specification was made on a different text, for a different purpose. The four parts were established by reading Aristotle’s inventory at Metaphysics Γ.2, in a study whose argument is about focality and univocity and which says nothing about the first Critique. They were not shaped to fit the passage they are about to be applied to. A reader who suspects the fit has been arranged can check the order in which the two were written.

All four hold here, and the fourth is the one on which everything turns.

A title to count. Standing as phenomenon, which is to say standing as something for me rather than as nothing for me. Not standing as this phenomenon rather than that one, which is a different question and is answered by the forms of intuition. Standing as one at all.

A case that appears by it. Every representation. Every object of possible experience. The scope is unrestricted within the architecture and Kant says so.

And an appearing-on-another’s-warrant which is neither the derivative’s own claim nor a causal consequence of the primary. Both negatives hold, and checking them is the decisive step, because it is where this relation separates from the merely explanatory and the merely causal.

Not the representation’s own claim. Nothing in a representation titles it. There is no feature a representation has, or could have, in virtue of which it stands as one for me. The condition is stated over it, not found in it.

Not causal. The Ich denke does not produce the representation, does not bring it about, does not make it occur. And because of können, it need not even accompany it. What it does is hold a standing possibility, and holding a standing possibility is not a way of causing anything.

The holder, corrected

The remaining part is the case that holds the title, and here an earlier statement of this argument was wrong in a way worth printing rather than repairing quietly.

It is natural to say that the Ich denke is what cannot itself be accompanied by any further representation, and so is where the forwarding stops. On the German that clause attaches elsewhere.

…weil sie dasjenige Selbstbewußtsein ist, was, in dem es die Vorstellung: Ich denke, hervorbringt, die alle andere muß begleiten können und in allem Bewußtsein ein und dasselbe ist, von keiner weiter begleitet werden kann.

The subject of von keiner weiter begleitet werden kann is die reine Apperception, the self-consciousness, and not the representation it produces. The standard translations render it so, parenthesising the middle clauses as modifying the representation. §2.4 takes up what the hervorbringt establishes about where the acting is located; here only the correction matters.

And the correction strengthens the finding rather than weakening it. The terminus of the architecture, the point at which nothing further stands behind, is not even the item that does the titling. The Ich denke titles; pure apperception is what nothing accompanies. Two positions, and neither of them is occupied by anybody.

The tenant is really there. The right is held elsewhere. And the landlord need not visit.

That is the reading. §2.4 defends it against the objection Kant’s own text supplies, and §2.5 sets it beside the only other answer anyone has given to the same question.

§2.4 · Spontaneity, and where the act is located

The strongest objection to §2.3 is not brought from outside the critical philosophy. It is brought from two pages of it, and a reader who knows the Deduction will have it ready before the section ends.

At B130 all combination is an act. The manifold’s being brought together is not given with the manifold and cannot be given with it; combination is an operation of the understanding, and Kant marks the whole family of such operations as spontaneity. Then at B132, of the Ich denke itself:

Diese Vorstellung aber ist ein Actus der Spontaneität.

So the very term this article describes as holding a right and doing nothing is called an act by Kant, in the paragraph that introduces it. If the Ich denke is an act of spontaneity, then a relation of possible accompaniment looks like the wrong description of what it does, and the landlord need not visit looks like a claim the text contradicts.

The reading is held in print and recently. Busse, pressing Kitcher in the Kantian Review symposium, argues that Kant’s main point in this stretch is the spontaneity or self-made-ness of combination rather than any requirement of reflexive awareness. Liang reads the Ich denke as an act of self-ascription of mental states, citing B132 for it. The objection is not a construction for the purpose of answering it.

What Kant means by the phrase, on the same line

Take the clause whole rather than the noun alone. Kant glosses it himself, immediately, and the gloss is the beginning of the reply:

Diese Vorstellung aber ist ein Actus der Spontaneität, d. i. sie kann nicht als zur Sinnlichkeit gehörig angesehen werden.

That is: it cannot be regarded as belonging to sensibility. Spontaneity here is stated as the contrast with receptivity, and the contrast is where Kant places the whole weight of the term. It marks which side of the architecture the representation falls on. It is not, in that sentence, a claim about anything the Ich denke performs upon the representations it may accompany.

This does not dispose of the objection. Combination at B130 is genuinely an operation, and calling it spontaneity there does mean an activity of the understanding. What the gloss establishes is narrower and it is enough to open the question: from the fact that the Ich denke is an act of spontaneity it does not follow that its relation to what it accompanies is a relation of acting. Where the act is located has to be read, not assumed.

Where the act is located

It is stated in the sentence that follows, and the German settles it:

…weil sie dasjenige Selbstbewußtsein ist, was, in dem es die Vorstellung: Ich denke, hervorbringt, die alle andere muß begleiten können und in allem Bewußtsein ein und dasselbe ist, von keiner weiter begleitet werden kann.

Three things are in play, not two.

The production relation runs from apperception to the Ich denke. The accompanying relation runs from the Ich denke to the representations. They are different relations, between different terms, and only the first is a producing.

Which is why the objection, granted entirely, touches nothing. There is an act here, and Kant names it: apperception brings forth the representation. That act falls wholly on the subject side, between the self-consciousness and the representation it generates. Nothing in it makes the Ich denke produce, constitute, or bring about the representations it may accompany. The relation to those is given by the modal, and the modal is begleiten können.

And the objection’s strongest form is answered in the same place. A term required to accompany every representation might still be construed as doing something on each occasion, a synthesis performed, an act done and repeated. A term that must merely be able to accompany them stands in no such relation to them, since on any given occasion nothing need occur, and the standing of what is titled is unaffected by whether anything does.

What is granted, and what is denied

The distinction the reply turns on is worth stating flatly, because a reader can accept the first line and miss the second.

Synthesis is granted. Combination makes a manifold into one representation, it is an operation of the understanding, and nothing here contests it. That is what B130 is about and it is not what §2.3 is about.

Titling is denied to be an act. What makes any representation count as one for me at all, rather than as für mich nichts, is not something done to it. It is a possibility held with respect to it, and held elsewhere.

The two are different questions and Kant’s text keeps them apart. Synthesis concerns how a manifold becomes one representation. Titling concerns how a representation stands as a representation for me. The first is answered by an act; the second by a modal.

(Guard, and it limits what the section claims. The construal above is the received one. The standard translations render von keiner weiter begleitet werden kann as attaching to the self-consciousness, with the middle clauses parenthesised as modifying the representation, and nothing in the reading of the German is offered here as a discovery. What is the article’s own is the inference from it: that a relation of possible accompaniment cannot be a relation of production. The philology is inherited; the consequence drawn from it is not, and it is the consequence that should be attacked.)

(Second guard. Draft v1.0 of this article attributed the terminal position to the representation Ich denke, writing that it cannot itself be accompanied by any further representation. On the German that clause attaches to die reine Apperception. The correction strengthens rather than weakens the finding, since it means the architecture’s terminus is not even the item that does the titling, and it should be printed as a correction rather than absorbed silently.)

§2.4-bis · Heidegger temporalises the office

An earlier statement of this argument held that the Ich denke and the forms of intuition are two deictic positions, and that Heidegger collapsed them into one. That reading is withdrawn. The Ich denke points at nothing, there is no this in it, and it is not thisness at a subject-pole. And what Heidegger does to it is not a collapse.

He temporalises the principle. He takes the term that authorises representation and grounds it in time as pure self-affection, which he names the transcendental Urstruktur of the finite self as such. And he records on the same page that Kant did not see it: the Critique brought time and the Ich denke together in their original selfsameness, without, to be sure, expressly seeing this himself. The identity is offered as a finding about Kant, and marked as one by the man who made it.

Which is a third way of occupying the office, and it is the one this article’s guards were least prepared for.

The account’s guards were built against terms promoted from above, as a first principle is, and terms imported from outside, as an external ground is. This is neither. The principle is not deposed and not replaced. It is grounded in something more original, and the result is renamed a ground-structure. That is an archic occupation performed in the open, on somebody else’s architecture, across three consecutive paragraphs, and a guard-set aimed at promotion and importation does not see it coming.

It is the strongest exhibit this account has for its own archic function test, and the test should acquire it as a case.

(Three guards travel with this. First, it is Heidegger reading Kant and must never be printed as Kant: the identity thesis is his finding, offered as such, and disputed since Cassirer’s review. Second, this article need not adjudicate the dispute and does not. On his reading the position collapsed into is time; on the standard reading the Ich denke is a formal condition of which there is no intuition and from which only that I am follows. Neither is someone, and the fourth row is empty either way, which is why the whole exegetical question can be conceded at no cost. Third, the recoil is not used. That Kant drew back from a root he had himself established is the most contested claim in the Kantbuch and nothing here needs it.)

§2.5 · The können taken two ways

The question left standing at §2.2 was what kind of relation makes possible names, and §2.3 answered it in one direction: the four parts of the substitutive structure hold of B131-132, and the relation is title-holding. That answer is not the only one available, and the one that competes with it has been in print for over twenty years.

Schürmann puts the question in the same place and answers it differently.

What is invariable is the “being able to accompany” itself, which is precisely an active relation of the one to the many. From the very beginning the können establishes the occurrent referent.

And, on the same page:

To be able to accompany is to translate, without mediation, originary acting into derived actions. Unity means to make a unity: to unify, and legislation is the immediate action of the I-think upon the categories.

And, three lines further:

“To be able to accompany” describes a spontaneous doing that is absolutely first.

He has also seen what the modal threatens and refused the obvious escape. Two pages earlier he considers whether apperception might be produced on some occasions and not on others, and rejects the reading: an apperception that came and went would lose its legislative status, and there would then be nothing for the deduction to enthrone. The können is therefore not intermittency. It is, on his reading, the mode in which an originary acting is transmitted.

(The French has been checked, because Lilly records in a note on the same page that Schürmann plays on the ambiguity of faire, and a disagreement resting on a translator’s choice would not be worth having. It does not rest on one. All three formulations are built on agir and actif, which carry no such ambiguity and which are Schürmann’s own technical vocabulary: rapport actif, précisément, de l’un au multiple; un agir originaire en actes; un agir spontané absolument premier. Where faire does appear in these pages it governs the making of experience and the heading of the lettered section, not the three formulations above.)

And the reading is not left to be inferred from those formulations, because he glosses the modal himself. Twice.

The first gloss is a paraphrase: the deduction posits self-consciousness as a process such that it always relates, toujours, and this, he writes, is what pouvoir means, effectively, effectivement, to the contents of the understanding.

The second is an analogy, and it is the mechanism. To heat the earth, the sun must be able to emit a heat of at least five thousand absolute degrees. That says that if it heats our planet it must be made in a certain way, and to say so is to describe its way of being, sa façon d’être. Likewise, he writes, Kant describes the way of being of consciousness: to serve as ultimate instance it must be able to accompany, which is to say, must be such that it accompanies, doit être telle qu’elle accompagne.

Two things follow, and the first is agreement. Schürmann takes the können to give a manner of being, and this article takes it to give a manner of being. Neither of us reads it as a claim about occurrences. The disagreement is about which manner, and it is now locatable in a single paraphrase.

The second is where the analogy fails. The sun in that example is heating the earth. Its capacity is inferred backwards from an exercise that is actually occurring, which is why must be able to can be converted without loss into must be such that it does. Kant’s case has no such exercise to infer from. Most representations are never accompanied by anything, and nothing in the architecture requires that any be. So the conversion of muß begleiten können into always relates effectively is not a reading of the modal but a replacement of it, and what is lost in the replacement is precisely the können.

This article answers the same question with the contradictory. A right that need not be exercised cannot be a doing. Where nothing is required to happen on any occasion, and the standing of what is titled is unaffected by whether anything happens, what remains is a right held and not exercised. The relation is title-holding, and title-holding is not action.

Both answers cannot be right. The disagreement is one clause wide, it concerns the same sentence, and each of us takes the emphasis to be Kant’s own. What follows is why the article needs both readings on the table rather than only its own.

Why the second horn is necessary

Take the transcendental reading of ownness at its strongest and ask what defeats it.

Suppose only the active reading were available. Then the Ich denke is an originary acting, legislation is its immediate action upon the categories, and the phenomenon stands where it stands because something is being done. To show that ownness is not that, one need only run the doctrine’s existing guards. Ownness clears nothing. It secures nothing. It holds nothing open. Every verb that would require somebody to be doing something is barred at the level of grammar, and the tell is printed with each guard so that a drafter catches it in the sentence rather than in the doctrine.

That refutation succeeds, and it is worth almost nothing.

Because the critic does not need the acting reading. He retreats to the quiet slot. Very well, he says: not an acting condition. A formal one. An enabling one. A condition that does nothing on any occasion and conditions all the same, in the way that the pure forms of intuition condition without labouring, and in the way a transcendental condition has always been understood to do its work without being an event. Nothing in the anti-action guards touches that, because those guards were built against a term that acts, and this one does not act.

And the quiet slot is where the transcendental reading of ownness actually lives. Nobody proposes that ownness performs. What is proposed is that it conditions: that it is what must hold for someone to appear at all, that the appearing of a who is possible because ownness obtains. That reading slips past the anti-bearer grammar, because a condition is not a substrate. It slips past the archic function test, because a condition does not command. It slips past grounds nothing, because conditioning is not grounding. Three refusals, and a condition neither stands under, nor commands, nor grounds. It conditions, and none of the three denials names that relation.

So the second horn is not a refinement of the first. It is what closes the retreat.

Read the können as title-holding and the quiet slot is occupied. A title is exactly a relation that does nothing on any occasion and is not thereby idle. It licenses appearance in an inventory. It requires no event, no act, no exercise, and it is fully operative while nothing whatever is happening. It is, in short, the best available description of a relation that is not an action and is not therefore nothing.

And ownness is not that either. A title licenses appearance in an inventory. Ownness licenses nothing, appears in no inventory, and there is no case whose standing it holds. It is not a title because a title is held by a member of an inventory on another member’s behalf, and ownness is held by nobody and licenses nothing.

Two horns, and the apparatus exhausted

The two readings are not rivals for the article’s purposes. They are the two ways the critical apparatus can construe a relation between a term and what appears by it, and between them the apparatus is used up.

Read the können as acting, with Schürmann, and the Ich denke is an originary doing that legislates. Ownness is not that, and the doctrine’s grammatical guards establish it.

Read the können as title-holding, and the Ich denke holds the standing of what appears without doing anything. Ownness is not that either, and §2.8 says why.

Between legislation and title there is no third relation in the critical architecture, and ownness is neither.

Which is the whole of the structural argument. §§2.2 and 2.3 established that the apparatus has two positions for content, the what-kind and the this, and a term that titles both. §2.7 will show that ownness is not either position. This section shows it is not the term either, on either available reading of what the term does.

The fourth row is empty, and it is empty on both readings. That is a stronger result than a comparison could produce, because a comparison invites the reply that the differences are matters of degree, and there is no degree of a position that does not exist.

(Guard, and it governs everything above. The disagreement with Schürmann is one clause wide and must not be printed as a general correction of him. He established the office, named apperception the fantasm of last resort, isolated the double modal, and asked what it means, and every one of those is inherited here. The article contests his answer to the last question and nothing else. §7 states what that costs him, and states it as tribute.)

(Second guard. The article does not need Schürmann’s reading to be wrong in order to run. If the active reading is correct and the title reading is not, the first horn stands, ownness is still not an originary acting, and only the retreat to the quiet slot survives unclosed. The article would then be weaker and not defeated. This should be said, because a reader who takes §2.5 as a refutation will overestimate what the section is doing.)

§2.6 · What makes a discourse ontological

The four parts checked at §2.3 belong to a distinction drawn elsewhere, and it is worth restating why the distinction was drawn at all. Four relations are commonly run together under the single heading of one case being prior to another.

A case may be understood by way of another. A case may be accounted for through another. A case may be such that it would not exist without another. And a case may appear in the inventory on a warrant it does not itself hold, which is to say that another case holds its title to count as a being on its behalf.

The first three are relations to a word, an account, and a cause. Only the fourth is a relation to standing. And standing in the inventory is what being a being consists in, if it consists in anything at all.

A thinker in whose architecture the fourth relation operates is doing ontology, whatever he calls it, and whatever the other three also hold of him.

Why this is not circular

Everything turns on the level at which the claim is made, and the level is operating, not naming.

A relation operates in an architecture whether or not its author identifies it, describes it, or has a word for it. That is not a special allowance made for this argument. It is the ordinary situation of any structural claim about a text: the syllogistic form of an argument operates in it whether or not its author had read the Prior Analytics.

So the claim is not that Kant secretly held the present account, and nothing here would be improved if he had. The claim is that the relation obtains in the architecture, and the demonstration is the check at §2.3, which runs the four parts against B131-132 and finds them.

There are two such demonstrations, and they are the reason the claim is not built on one case. The first is Aristotle’s inventory at Metaphysics Γ.2, where the derivative categories enter not on their own warrant but as determinations of what primarily is. The second is §2.3. Two constellations, two demonstrations, one relation.

The objection, which is the most cited position in the field

A reader from analytic metaphysics will have stopped several paragraphs ago, and the objection is a good one.

The claim that a relation rather than an inventory marks the ontological is Schaffer’s. It is the governing thesis of the most cited paper in metaontology of the last twenty years:

On the now dominant Quinean view, metaphysics is about what there is… I will argue for the revival of a more traditional Aristotelian view, on which metaphysics is about what grounds what.

And the structure is the one described above. A hierarchical reality ordered by priority in nature. Primary entities forming the sparse structure of being. Grounding relations generating an abundant superstructure of posterior entities. The primary being, as Schaffer puts it, all God would need to create.

Fine holds a version of the same. His realist metaphysics concerns not merely the nature of things but whether they are real, and if considerations of ground were abolished, very little of the subject would remain.

So the demarcation move is not this article’s, and any suggestion that it is should be withdrawn before the section goes further. What remains is a question about which relation demarcates, and that question has an answer.

What Schaffer is reviving

The word he uses is exact and worth taking at face value. He calls it a revival.

Heidegger’s account of what is being revived is available and it is not obscure. Across the history of metaphysics, ground appears as logos, as aitia, as ratio, formulated as a principle only in the seventeenth century by Leibniz; and the thesis stated in Der Satz vom Grund is that for metaphysics ground and being are the same, Sein und Grund: das Selbe.

So metaphysics is about what grounds what is not a novel metaontological proposal set against an older tradition. It is that tradition’s own self-description, offered affirmatively by someone who does not read it as a diagnosis.

Thomson’s reconstruction makes the fit closer still. Onto-theology performs two foundings at once: a top-down justification from a highest being, and a bottom-up ground-giving on a most basic being; the epoch-grounding systems combine them. Schaffer performs the bottom-up half explicitly, calls it neo-Aristotelian, and is right about the lineage.

(This is a diagnosis and is not printed as exposition. Schaffer does not claim to be doing anything other than metaphysics, and on Heidegger’s account he is doing exactly that, thoroughly and in the open.)

And the analysis being applied here is not new either. Schürmann worked it out across two books, tracing the archē as what simultaneously begins and commands through its epochal occupants, and setting out the hegemonic relation as a relation of the one to the many, with the concealment of its own positing as its signature. He could not aim it at Schaffer, who published sixteen years after his death. What this section does is apply a completed analysis to a body of work written after it, by philosophers who describe their project in the terms the analysis diagnoses. That is an application. It is not a discovery.

The differentiation is one word

Grounding, in the standard characterisation, is constitutive determination, as opposed to causal or probabilistic determination or explanation. The grounded is what it is in virtue of its ground. That is the relation’s content and its proponents are explicit about it.

Title-holding does not determine. That is what §2.3 establishes and it is established from the modal: a right that need not be exercised on any occasion cannot be making anything be as it is. The tenant is really there and the landlord did not build the house.

The distinction can be tested without any of the present vocabulary. My parents determine my existence completely. Remove them and I am not here. That is determination of the fullest kind available, and it settles nothing whatever about whether I appear in the inventory on my own account or on theirs. A cause brings about. A proxy stands in place of. The two come apart, and the whole of §2.3 is an argument that they come apart at B131-132 as well.

So the claim narrows and becomes decidable. Not that a relation demarcates, which is Schaffer’s. But that the relation which demarcates is title-holding rather than grounding, constitution, or dependence, because determining what something is leaves untouched the question of whether it stands in the inventory on its own account.

A second differentiation runs alongside the first and is independent of it. On the dominant assumption, grounding is factive: it relates facts or propositions rather than entities. An apple explains nothing; the fact that the apple is red may. Which is why it has been objected that substances grounding non-substances is an unattractive reading of Aristotle, and why the nearest formulation to this one relocates the relation to the admission of a status rather than placing it between the entities themselves.

Title-holding is not factive. The Ich denke does not hold the title of a fact about a representation. It holds the representation’s title, and what is at issue is the standing of the item rather than the obtaining of a fact about it. So the relation proposed here differs from grounding twice over: it does not determine, and it does not relate facts.

One further observation, and it is the rival literature’s own. Grounding relations appear to require grounding in their turn, generating either an infinite regress or an arbitrary termination. That difficulty is conceded from inside, and it has the shape the closure theorem describes.

Three further candidates

Three positions remain, and a survey that omitted them would leave the claim with unexamined doors.

Structure. On Sider’s account the mark of the metaphysical is joint-carving: a question is substantive insofar as it is posed in terms that match the world’s own distinctions, and a discourse is metaphysically serious insofar as its ideology is natural rather than merely convenient. That is a third candidate beside grounding and dependence, and it is not this one. Structure is a mark of naturalness in the vocabulary. Title-holding is a mark of standing within an inventory. The first asks whether the terms a discourse uses carve; the second asks, of an item already admitted, on whose warrant it counts. A discourse could carve perfectly and contain no relation of the fourth kind, and a discourse could contain one while carving badly. The two marks are independent, and neither entails the other.

Commitment. One clarification prevents a systematic misreading, and it is worth a paragraph because half this article’s readers will make it. The vocabulary of appearing in an inventory is Quine’s, and on his criterion what a theory says there is is what must lie in the range of its bound variables for its assertions to be true.

That is not what is claimed here. Commitment concerns what a theory admits. Title-holding concerns, given that a case is admitted, on whose warrant it counts. The criterion of commitment is silent about structure among the committed, and deliberately so: two theories with identical commitments may differ in whether one admitted case holds another’s title, and nothing in the criterion detects the difference. The fourth relation is not a rival criterion of commitment. It is a claim about what a criterion of commitment does not reach.

Deflation. A deflationary answer is available and should be named rather than left implicit. On the easy approach to ontology, existence questions are settled by the application conditions of the terms used to ask them together with whatever empirical facts those conditions call for, and disputes about what really exists are misconceived rather than difficult. If that is right, there is no substantive question of admission for any relation to answer, and this section has no subject matter.

The disagreement is prior to anything argued here and is not adjudicated. What can be said is narrower: the fourth relation is not a candidate answer to an existence question of that kind. It does not say whether a case exists. It says on whose warrant an admitted case counts as a being, and the deflationary machinery is not aimed at that.

Corkum, and the gap this fills

The closest formulation anywhere comes not from metaontology but from Aristotle scholarship, and it uses the article’s own vocabulary independently.

Corkum’s account of priority in the Categories concerns, in his words, the provision of a rationale for inclusion in the ontology: redness has its ontological status, which is to say it is an item of a certain category in the ontology, in virtue of there being red substances.

And the relation is placed with unusual precision. His proposal is not that primary substances ground other entities. It is that a primary substance’s admission of its ontological status grounds the admission of the ontological statuses of other entities, and that this suffices for the substance to be prior to and separate from them. What stands in the relation is the admission rather than the entity, which is as close to the structure described here as anything in print.

That is the inventory, the title, and the holding, arrived at on a different text and without any of the machinery used here.

And his critics have named the gap. Peramatzis and Koslicki both record that it is insufficiently specified in Corkum what constitutes the admission of an ontological status.

Two scholars have said in print that the explanandum is under-specified. The four-part structure is a specification of it.

That is the section’s actual contribution, and it is smaller and more defensible than the claim it replaces. Not that the question is unasked. That it has been asked, and answered incompletely by the person who asked it best, and that the four parts complete the answer.

And the consensus this contests

One position remains, and it is the oldest.

Owen, Patzig and Shields hold that the pros hen is what makes a science of being possible and unified: without it the senses of being are merely homonymous and there is no single subject matter. Politis and Steinkrüger narrow this, arguing that the pros hen structure is necessary only, and not sufficient.

The present claim separates two things that reading holds together. The pros hen secures unity. Whether it is also what makes a discourse ontological is a further question, and the answer given here is no: what makes a discourse ontological is the title relation, which the pros hen happens to instantiate in its focal form but does not exhaust.

The tradition conflated being ontological with being unified, which is what the companion study’s argument predicts, since both available devices exist to secure unity. Separating them is what allows an ontology to be ontological without being unified, and that is the third structure the who is the first worked case of.

(This is the largest claim in the study, and it rests on a check of the grounding, dependence and metaontology literatures rather than on a survey of them. It is stated at the strength a check supports and no higher, and falsifier 2 is the form an attack on it would take.)

(Guard, and it is the one a reader will feel forming before he can name it. A single mark by which all prior ontology is measured, supplied by the present account, is itself a focal arrangement, and this is the last body of work that can afford to build one by accident. The reply is that the fourth relation is not scalar. Either it operates in an architecture or it does not; there is no more and less of it; so nothing partitions, no order forms, and no primary case emerges against which the rest could be ranked. The test that convicts an archic term is run here on this one, and it returns nothing to convict.)

(Second guard. The relation found operating licenses the finding that a discourse was ontological. The relation not found operating licenses openness and not denial. Nothing here is a machine for declaring what was not ontology.)

§2.7 · Kant reaches the character, on another route, and pays for it

Everything above is structural, and a structural finding can look like a charge of blindness unless the record is set out. So it should be set out. The character this article is concerned with does appear in Kant, stated with complete precision, and stated in its native form.

In the kingdom of ends everything has either a price or a dignity. What has a price admits something else in its place as an equivalent. What is raised above all price, and therefore admits of no equivalent, has a dignity.

Read what that sentence does. Substitutability is the criterion, and the character is stated as a refusal of it. Not as a property present in the case and absent from the other, but as what fails when a substitute is put in place. A refusal is the character’s native grammatical form, and Kant reaches it in that form on the first attempt.

And then it is restricted twice, in the space of two pages.

First by a capacity clause: the dignity belongs to a nature insofar as it is capable of the morality in question. Second by a ground: autonomy is the ground of the dignity of human and of every rational nature.

Which is the reception principle running exactly as it predicts. A character constitutive of the who cannot fail to operate, whether or not anyone notices. It appears to any thinker who genuinely encounters the phenomenon, and Kant encountered it. But it appears as the differentia of a category of being only where it is reached ontologically. The register in which it shows up is fixed by the route, and not by the acuity of whoever takes the route. A moral route delivers it as Würde. A deictic route delivers it as mineness. Only an ontological route delivers it as a differentia.

So Kant is not a case of failing to see. He is the best available case of seeing on the wrong route, and the wrongness is not a fault.

Because §§2.2 to 2.6 explain why the route had to be the other one. A constitutive character of a category of being has nowhere to arrive in an architecture whose categories belong to the understanding, whose two content-positions are the kind and the this, and whose third term titles both without being a further kind of content. There is no position in that apparatus for such a character to occupy. It lands in the practical philosophy because the practical philosophy is the only register left unoccupied, and it lands there in the only form that register can give it.

And the cost of the second restriction is visible on the page rather than inferred.

Ground a categorial character in a capacity and the character tracks the capacity’s distribution.

Kant’s does. Where autonomy is absent, or not yet present, or no longer present, Würde in the strict sense cannot follow. The newborn in her first hour, the person whose capacities have receded, the animal: each of them either receives the character by an extension the ground does not licence, or does not receive it. That is not a moral complaint against Kant. It is what happens to any categorial character grounded in a capacity, and it is why the present account declares its differentia and grounds it in nothing at all.

§2.8 · The finding

The apparatus has now been walked. What follows from the walk is a set of four refusals, and each has its own reason.

Ownness is not a kind, not a this, not a principle, and not a title.

Not a kind. A kind is given by a definitional content, and a definitional content is a such: articulable, and therefore repeatable, and therefore something a substitute could be given along with everything else. The case that shows this is the same case throughout: match every listable point and the divergence remains, so the divergence is not among the listable points.

Not a this. Thisness answers which occurrence is this, and it answers it of bands, storms, coastlines, canvases and persons without distinction. It is region-neutral, which means nothing whatever is marked off by it. Two indiscernible wristbands are two, and nobody has produced a person.

Not a principle. This refusal is not invented for the occasion; it is the doctrine’s own, already printed: a mode is not a principle. Non-vicariousness is the mode by which ownness is the differentia it is, as ownness is the character by which the category is the category it is. A mode grounds nothing, produces nothing, and stands beneath nothing.

And not a title, which is the refusal this part adds. A title is held by a member of an inventory on another member’s behalf; it licenses appearance; it can be lodged, and what holds it can be named. Ownness is held by nobody, licenses nothing, and there is no case whose standing it holds. Nothing here appears on its warrant, because it issues none.

That fourth refusal is the third applied one level down, and it is worth seeing that it is. A mode is not a principle bars ownness from the office of what grounds. Ownness is not a title bars it from the office of what licenses. The two are the same refusal at two heights of the same architecture, which is why one guard, correctly stated, catches both.

Set them against what the apparatus contains and the result is a table with an empty row.

The critical architecture has two positions for the content of cognition, and one term that titles both. Ownness is neither position, on the reasons above. And it is not the term, on either available reading of what the term does: not on the reading that makes the können an originary acting, and not on the reading that makes it a right held and not exercised. §2.5 closed both, which is why it needed both.

The article’s answer to Kant turns out to be a guard the doctrine already carries, applied to an architecture instead of to a sentence.

Which is a better place to end than a comparison would have been. A comparative argument, of the form this how resembles a transcendental condition, and here is how it differs, invites the reply that the differences are matters of degree, and degrees are exactly what the account cannot afford to concede anywhere. A structural argument says there is no position for it to be a degree of.

What remains is the objection that the empty position could be filled after all, by something the walk did not look for. §3 answers that, and it answers it with the residue.

Part Three · The residue, and how to read it

§3.1 · Für mich nichts, read exactly

An earlier statement of this argument glossed für mich nichts as meaning that a representation admitting no room for the Ich denke would be no phenomenon at all. That over-reads the sentence, and the correction is worth making because it improves what follows rather than qualifying it.

The clause is a disjunction.

…welches eben so viel heißt, als die Vorstellung würde entweder unmöglich, oder wenigstens für mich nichts sein.

Either impossible, or at least nothing for me. Two disjuncts, and wenigstens marks the second as the weaker.

Read what that does. Kant reaches first for an absolute modal claim, that such a representation would be impossible, and then withdraws from it. The fallback he settles on is not a claim about what can exist. It is a claim about what stands for me, and für mich is his own relativiser, supplied unprompted in the sentence that introduces the whole apparatus.

So the alternative is not titled or nonexistent. It is titled or untitled.

Which is exactly what the reading at §2.3 predicts. A warrant does not settle whether something is. It settles whether something appears in the inventory on a warrant it holds or does not hold. A representation that cannot accommodate the Ich denke is not thereby annihilated. It is untitled, and nothing for me is how the untitled has to be reported when every reporting term is itself titled.

And this is why the residue can be at once required and unsayable, which is the difficulty every reader of the first Critique meets and few resolve. It is nothing for me without being nothing. Kant’s own retreat, from unmöglich to wenigstens für mich nichts, is the architecture registering that difference in the first paragraph in which it could.

§3.2 · Bivalence, not plurality

Running the modal

Take any candidate whatever and put the double modal to it.

Either it can accommodate the Ich denke with respect to its representation, and then it is a phenomenon, titled, an object for me.

Or it cannot, and then it is nothing for me, in the exact sense §3.1 established: untitled, and unreportable in any titled vocabulary.

There is no third position, and the thing in itself is an attempt to occupy one.

And it cannot be stated. Every word available for stating it is on the warrant. Something is on the warrant. Object is on the warrant. Cause and exists are on the warrant. To use any of them of the untitled is to use the vocabulary outside the conditions of its use, which is why the limiting concept is a name for a position the architecture forbids naming.

The sharpest objection to that claim is Hegel’s, and it should be met here. He finds it surprising that anyone keeps repeating that we do not know the thing in itself: on the contrary there is nothing we can know so easily.

The objection fails, and it fails into agreement. What Hegel knows so easily is that the concept is empty. Abstract from everything consciousness makes of an object, from every deliverance of feeling and every specific thought about it, and what is left is utter abstraction, total emptiness, describable only as a beyond. That is not a characterisation of the residue. It is the finding that there is nothing to characterise, which is what §3.1 and this section have been saying. The two accounts differ over what to call the result and agree that no positive content is available at the end of it.

And Hegel registers the connection this study argues for, by a different mechanism and two centuries earlier. The caput mortuum, he writes, is still only a product of thought, such as accrues when thought ends in abstraction unalloyed, and it is the work of the empty Ego, which finds an object in this empty self-identity of its own.

The residue traced to the I. His mechanism is abstraction rather than titling, and he draws from it a conclusion this study does not, that the emptiness is therefore known. What he does not ask is on whose warrant the phenomena appear, which is why the diagnosis stops at the emptiness rather than at what produces it.

The Grenzbegriff is evidence for the reading

The weak form of this charge, that Kant populated a second world with unknowable objects, is a caricature nobody holds and this article does not raise.

He says the opposite, and says it plainly. The noumenon is a Grenzbegriff, a limiting concept, whose employment is merely negative and whose office is to mark where sensibility’s claim stops. He saw that the position was untenable as an object and tried to hold it open as a boundary.

That is not a difficulty for the present reading. It is its best exhibit.

The restriction to the merely negative is what a thinker does when he has registered that a position is required and has seen that the vocabulary cannot occupy it. Kant is not being convicted of an oversight. He is being credited with seeing the split, and the architecture is being charged with requiring one.

The finding

A warrant structure does not produce a plurality of ways of being. It produces a bivalence. Titled, or nothing for me. And a bivalence is the opposite of a plurality.

Which is why the twelve are no counter-example, and why §2.1’s homonym point matters more than it looked. The twelve are a plurality of concepts, and all of them are on the warrant. There is no plurality of ways of being anywhere in the architecture. There is one way, being-for-me, and its shadow.

So the critical philosophy is the first answer to the plurality problem carried to its limit. Not many senses ordered toward a primary one, with derivative cases still in the inventory and still sayable. One sense, and a residue that cannot be said. The focal arrangement at full extension: the distance from the primary case is either zero or infinite.

That is where this article and the companion study meet. The companion argues that a scalar differentia makes its category a quantity, whose portions stand in an order that a focal term names rather than imposes. Kant is that argument’s limit case, and the limit is reached when the order has collapsed to two values and the second is unsayable.

§3.3 · Accommodation, not suppression

This section states the register for the whole of Part Three, and everything above should be re-read under it.

The residue is not what the architecture failed to say. It is what Kant refused to let the architecture erase.

Consider what suppression would have looked like, since it was available and costs nothing. It would have said: appearances all the way down. Nothing further is required. There is no remainder, the question of what affects sensibility is malformed, and the vocabulary is complete because there is nothing outside it for the vocabulary to fail at.

He does not say that. He admits the remainder and marks it unsayable, and the Grenzbegriff, restricted to the merely negative, is the most careful accommodation the architecture permits. He builds a room he cannot furnish rather than deny there is anyone to house.

Which changes how the well-known instability should be read. The transcendental object, the thing considered as it is in itself, double affection: these are not three symptoms of a theory buckling. They are three successive attempts at accommodation, each a fresh effort to house something the architecture required and could not name. That the room cannot be furnished is shown as rigorously as anything in the literature. It is not shown that the room should not have been built.

And one word needs its subject fixed, or the whole section is heard as fault-finding. The marks are all present: a negative formulation where a positive one is unavailable, a stopping point the author cannot cash further, a restriction that narrows what has just been admitted, a declaration that the matter can only be held open as a boundary.

The strain is in the architecture. It is not in the thinker.

Say whose, or a reader hears an indictment. And every pejorative in the vicinity goes with it: no evasion, no papering over, no failure to. What the marks show is not a theory collapsing but a thinker working, and the effort of accommodation visible on the page.

§3.4 · The reception principle’s second exhibit

Part Two found the character reached genuinely and confined by its route. Part Three has now found something else reached genuinely and confined by the same route, and the two are the same finding twice.

The Ding an sich is to categorial plurality what Würde is to ownness.

Set them side by side. In the Groundwork, non-substitutability is reached in its native negative form, on the moral route, and then restricted by a capacity and grounded in autonomy. In the first Critique, a remainder is reached that the one titled way cannot cover, through the warrant structure, and then restricted to the merely negative and held open as a boundary. Level two reached at full strength in both cases. Level three blocked in both. The register fixed by the route in both.

And the calibration matters, because the temptation here is to give Kant more than the evidence allows and thereby spend §3.2’s own result.

What he reaches, with complete fidelity, is that something must be admitted which the one titled way cannot cover. That is not a small thing and nothing here diminishes it.

What he does not reach is ways, in the plural. The architecture gives him one sense and a shadow, and a bivalence is not a plurality. So the residue is categorial plurality registered in the only register available: as a single unsayable remainder rather than as a second way in which being obtains.

Which is the whole of what this part claims about him. Not that he missed the question. That he registered it, and that the form in which he could register it is the form his architecture allowed.

§3.5 · The rival dissolution, and where its own literature has arrived

A reader trained in Anglophone Kant scholarship has an objection ready and it is not a small one. He does not think there is a residue at all.

On the two-aspect reading, thing in itself names a way of considering an object rather than a second object. Appearances and things in themselves are not two sets of entities but one set considered under, or in abstraction from, the epistemic conditions of our knowing. If that is right, nothing untitled is being posited, no unsayable remainder is generated, and §3.2’s whole finding is aimed at a doctrine nobody holds. This has to be met here rather than at §4, because a reader holding it will stop at §3 and never reach the differences.

The meeting is short, and it does not require the two-aspect reading to be wrong.

Ask what affects sensibility.

Something must, or the forms have nothing to form and receptivity receives nothing. And the same object considered differently does not answer it, because the considering is done under conditions and the affecting is what supplies those conditions with material. The difficulty is old, it is precise, and it is not this article’s invention.

It is what the two-aspect reading’s own literature has spent a century on. Vaihinger’s trilemma set out the options and found each unattractive. Adickes developed double affection to hold them together. Guyer pressed the charge that the methodological version is either trivial or an abandonment of the position it means to defend. Van Cleve pushed the two-object reading precisely here. Stang has reopened double affection on its merits rather than as an embarrassment. And Anderson has recently argued that the metaphysical versions of the two-aspect view risk collapsing into a Leibnizian position on which spatiotemporal cognition ultimately refers to a non-spatiotemporal reality.

The two-aspect route relocates the residue to affection. It does not remove it.

Which means the affection literature is not background to be acknowledged before the argument begins. It is affirmative evidence for the argument, and it comes from scholars with no interest whatever in the present thesis. A structure that generates an untitled remainder will generate it wherever the account is pressed hardest, and the place it has been pressed hardest for a hundred years is affection.

§3.6 · Meillassoux, who is an opponent and not a precursor

The nearest statement anywhere of the unsayability is the correlationist circle: we cannot think the in-itself without thereby making it a correlate of thought, so that what is reached is never the in-itself but always the in-itself-for-us.

Structurally that is close to §3.2’s claim that every characterising term is on the warrant. Substantively it goes the other way, and the divergence should be printed rather than left for a reader to find.

Meillassoux classes Kant a weak correlationist, and the classification turns on exactly the point at issue. For the weak correlationist the in-itself cannot be known but can be thought: it may be posited as existing, and as non-contradictory, and those are claims about it. That is why Kant is not, on his account, at the stronger position where even thinking it is forbidden.

§3.2 holds that it cannot be stated at all. Not that it is unknowable while remaining thinkable, but that the vocabulary in which any claim about it would be made is titled throughout, so that exists and non-contradictory are as unavailable as object and cause.

The article’s claim is therefore stronger than Meillassoux’s reading of Kant and inconsistent with it.

Which is a better position than agreement would have been, since a shared conclusion reached from an opposed premise is worth less than a disagreement that can be adjudicated on the text. Christias has already drawn the comparison between Sellars and Meillassoux on the passage to the in-itself, and the present divergence is located inside that comparison rather than beside it.

§3.7 · Flattening does not dissolve the residue

One further position has to be met, because it is the one that looks most like the present dissolution and is not it.

Object-oriented ontology removes the hierarchy. Bryant’s flat ontology rejects any ontology of transcendence or presence that privileges one sort of entity as the origin of all others, and to that extent it does what §2.6 says a non-focal ontology must do.

And it keeps the residue. Harman universalises withdrawal: every object withdraws from all its relations, so that the thing in itself is not abolished but extended to everything. Bryant’s objects are likewise withdrawn from all relation. The noumenon survives the flattening intact and is multiplied by it.

Which is the exhibit this part needs, and it is more useful than an absence would have been.

Flattening the origin relation is not removing the title relation.

The privileged origin is what these ontologies remove: no entity is the source from which others derive. The title relation is untouched, because nothing in the removal addresses whether an entity appears in the inventory on its own account. And so the residue persists, and is generalised. Brassier’s objection to the flat ontologies is the critical literature to consult here, and it is aimed at the levelling rather than at the residue.

The dissolution §2.6 describes therefore requires more than the rejection of hierarchy. It requires the substitutive diagnosis, which is what identifies the relation whose removal removes the remainder.

§3.8 · Existence is not a real predicate

One objection is short, entirely internal to Kant, and will be raised in the first five minutes by anybody who knows the book. It should be answered in the text rather than a note.

If existence is not a real predicate, how can there be a title to count as a being? Kant’s thesis is that is adds no determination to a concept: a hundred real thalers contain no more than a hundred possible ones. If being is not a property, nothing can hold a title to it, and the whole vocabulary of §2.3 is a category mistake.

The reply is a distinction Kant himself draws, and draws in the sentence §3.1 examined.

The title in question is not a title to exist. It is a title to count as an object of possible experience: to stand in the inventory, to be something for me rather than nothing for me. That is not the existence thesis’s territory. Kant’s claim is that existence adds no determination to a concept. It says nothing whatever against there being conditions on what counts as an object for me, and the Critique is very largely an account of those conditions.

And für mich nichts is Kant marking exactly this difference where the difference first becomes necessary. He does not write would not exist. He writes would be nothing for me, having just declined the stronger option. The relativiser is his, and it is doing precisely the work the objection says cannot be done.

Schürmann puts the same point at the centre of the architecture rather than at its edge: it is impossible to conceive of being as a real predicate, and consequently being adds no thought content to what we conceive. That is the objection’s own premise, granted in full. What it leaves untouched is the question of what counts as an object of possible experience, which is the question the title answers.

§3.9 · What this part claims, at the strength it can carry

This is the article’s most exposed stretch, and the exposure should be printed here rather than discovered by a referee.

The residue has two sources, not one.

The forms of intuition and the Ich denke jointly title appearing. Something is a phenomenon for me only if it can be given under the forms and can accommodate the Ich denke. So a defender may say that the thing in itself is required by the ideality of space and time by itself, that Kant’s own route to it runs through the Aesthetic rather than the Deduction, and that the warrant structure is idle in the derivation.

The reply is a narrowing and not a denial.

What the ideality of space and time yields, by itself, is that appearances are not things as they are independently of our forms of receiving them. What it does not by itself yield is the unsayability: that the residue admits no characterisation whatever, because every characterising term is titled. That second result comes from the warrant, and it is the result §3.2 rests on.

So the claim is that the warrant structure produces the residue’s unsayability. It is not that the warrant structure alone produces the residue.

Stated at that strength the claim holds. Stated higher it does not, and an opponent who defeats the higher version has not yet touched this one.

Two further items belong here rather than in a note.

Jacobi’s dilemma is the classical form of the difficulty and should be cited from the 1787 supplement rather than from the textbook paraphrase: without the presupposition one cannot enter the system, and with it one cannot remain within it. The present reading agrees with the diagnosis and disagrees with the inference. That both horns hold is exactly what a warrant structure predicts; that this shows the system incoherent does not follow, since a structure may require what it cannot say without contradicting itself.

And Miles has recently argued that the transcendental distinction is best read as a meta-metaphysical thought experiment designed precisely to forestall objections of Jacobi’s kind. That reading is compatible with this one and sharpens it. A thought experiment devised to forestall the objection is another instance of accommodation: the architecture registering a difficulty it cannot resolve, and constructing a device to hold the position open. Which is §3.3’s register, arriving from a source with no stake in it.

Part Four · The transcendental case, answered

(Six of the seven differences below have a published seat, named at the point of use, and this article’s wording is not the seat’s. Where the two differ the seat governs. The seventh is marked.)

4.1 · The case against, built from the account’s own sentences

Part Two argued structurally, which is the stronger way, and this part now runs the transcendental reading at full strength and answers it. The reading’s evidence is not supplied by its critics. It is supplied by the account.

The condition-form. Exhaust the manner and non-vicariousness is not false but unstatable. That is a condition on the statability of the differentia, stated by the account, in the account’s own words, in the place where the differentia is being defended.

The horizon-form. The manner remains wholly sayable of another, and that is what holds the possibility open. Read holds open as a verb with an agent and a horizon has been installed in the sentence that was meant to prevent one.

The faculty-form. The categorial eye is a historically situated and corrigible capacity of a practice of inquiry, by which the category is fulfilled where sensible indication would stop one category too early. A capacity whose exercise discloses a categorial difference is, to a transcendental ear, a faculty.

And the argument-form, which is conceded outright. The bounded closure theorem reasons that every proposed constitutive explanation of being-this or being-own must either offer an explanans characterisable without presupposing the target datum, or concede that it articulates the obtaining rather than grounding it. That is a transcendental argument. It reasons from what must hold if a certain kind of claim is to be possible, and nothing is gained by denying it.

What is denied is not that the account runs a transcendental argument. It is that a transcendental argument yields a transcendental condition.

4.2 · Why the existing guards do not catch it

§2.5 needed this in outline and it should now be stated in full, because the vulnerability is real and was found by looking rather than reported by a critic.

The account carries three refusals and each is aimed somewhere else.

The anti-bearer grammar refuses a substrate: something standing under and bearing what is said of it. The archic function test convicts a term that commands: one that authorises applications, organises a field toward itself, and receives what is forwarded until nothing else answers. And the doctrine’s own formula says the differentia grounds nothing.

A condition neither stands under, nor commands, nor grounds. It conditions.

None of the three denials names that relation, which is why a transcendental reading passes all three untouched, and why Part Two had to proceed structurally rather than by adding a fourth denial. A fourth denial would have been a fourth thing for the reading to pass. What defeats it is not a refusal but a demonstration that the position it needs does not exist.

4.3 · What a transcendental condition is, reconstructed

Offered as a reconstruction from Kant rather than an exposition of him, and with one item flagged as contested rather than leaned on.

On the standard reconstruction, a transcendental condition (i) is a condition of the possibility of experience or of objects of experience; (ii) is a priori, not learned from what it conditions; (iii) is necessary, in that its failure removes not a truth but the object; (iv) is subject-side, belonging to the constitution of the cognising subject rather than to what is cognised; (v) is established by a deduction, a proof of entitlement, rather than disclosed; and (vi) constitutes its object as an object of experience.

The sixth is contested and nothing here rests on it. Allison’s two-aspect reading takes the conditions to be epistemic conditions rather than conditions imposing form on a mind-independent material. Strawson grants the necessity of the conditions while rejecting the idealism that would make them constitutive in the strong sense. Each is granted what he grants, and the differences below are chosen to hold on all three readings.

4.4 · Seven differences

First, and sharpest: statability, not possibility.

A transcendental condition failing removes the object. This one failing removes the sentence.

Exhaust the manner and no who is annihilated. Nothing is taken from any obtaining. What becomes unavailable is the statement of the differentia: the refusal has no term, so it refuses nothing, and the character cannot be said.

The companion study puts it in one sentence, and the sentence is what the transcendental reading quotes against the account: exhaust the manner and non-vicariousness is not false but unsayable, since a refusal of substitution where no substitute is possible refuses nothing. (Seat: the companion study, §7.3.)

Second: nothing subject-side.

The non-exhaustion of the manner is not a feature of any cogniser, and it holds where no cogniser is in the vicinity. Nothing here belongs to a faculty, an apparatus, or the constitution of whoever articulates it.

The hard case is not Kant. It is the reformed subject of post-Husserlian phenomenology: embodied, temporalising, world-involved, not a bundle of faculties, and therefore able to claim the non-exhaustion as its own structure without any of the objectionable machinery.

The answer is the last cat. Felinity with no further occurrences anywhere: her ownness is not diminished, because the manner remains sayable of another whether or not another obtains. The non-exhaustion is modal and not a count, and a reading on which the refusal weakens as occurrences grow scarce has let the actual decide the possible.

A reading that made the non-exhaustion the structure of any subject would have to make it fail where the subject is absent, and it does not fail. (Seat: the companion study, §7.3, where the case is put against the same collapse under a different name. The application to the reformed phenomenological subject is this article’s.)

Third: access never constitutes, and there are two opponents here rather than one.

Against the transcendental idealist, the corrigibility of reaching, and the shape of the correction rather than its mere occurrence. A misreaching is corrected by resistance from within the field of the occurrence it was reaching for. What answers to something beyond itself does not constitute what it answers to.

Against Allison’s two-aspect reader, who is harder because he grants that: epistemic conditions are precisely conditions under which something can be an object for us, and corrigibility is at home inside them.

The answer is the third term, and it is inherited entire. Deictic determination is the ontological determination of an occurrence as this one, settled independently of anybody’s reaching. Deictic reaching is the regional route by which this occurrence rather than another comes to be at issue, defeasible and sustained within a relational and historical field. Linguistic deixis is one vehicle of that reaching among several and nowhere the paradigm.

Remove the first and the other two collapse into each other, and the corrigibility becomes unintelligible, since a misreaching would be answerable to nothing. The two-aspect reader must then either make the reaching constitutive and surrender the correction, or keep the correction and concede a determination his conditions do not constitute. There is no third option, because the demonstrative he might reach for is a vehicle of the access and not something prior to it. (Seat: the companion study, §3, where the argument is made against a reflexive objection and where the shape of the correction, rather than its mere occurrence, carries the evidence.)

Fourth: defeasible.

A purported disclosure of ownness is defeated where it tracks role, status, possession or convention rather than the case; where it disappears under reframing without resistance from the region; or where structurally symmetrical cases receive unequal treatment without relevant reason.

And the first defeater is run against the account’s own central example. Suppose the owner of the burned canvas reads his own grief as the loss of someone, the presence in the frame, the studio’s companion. The first condition convicts the reading: his refusal of the copy is sufficiently explained by possession, by fidelity to the original making, and by the conventions of originality, with no remainder directed at anyone. What is left standing is a true grief over the historical singularity of a making.

A transcendental condition is not defeated by cases. This is defeated by three, and one of them has been used on the account’s own exhibit. (Seat: the companion study, §7.1, where the defeaters are stated and where the conviction is performed.)

Fifth: regional, and provisional.

The finite-living articulation, this life cannot be somebody else’s living, is carried to no untried region. The restriction is binding and is not a hedge: a definition through life would settle the machine, the civic, the sacred and the divine negatively in advance, by wording rather than by trial. What travels is the unrestricted statement, and the life-sentence stays where it was earned.

And the boundary so drawn is provisional in a way worth stating. It is drawn at living, on an account of living the argument has not renewed. It marks where the trials have gone, not where the character stops. A transcendental condition does not have a research frontier. (Seat: the companion study, §7.1.)

Sixth: disclosed, not deduced.

No definition is available and none is possible, and the impossibility is not a limitation of the account. A definitional content would be a such: articulable, and what is articulable is repeatable in principle, which is what makes it articulable. So a substitute could be given it along with everything else, and the demand for a definition is not a demand the account fails to meet but a demand whose satisfaction would defeat the account it was meant to complete.

What stands in a definition’s place is what a definition could not have been: a case beside a case, matched at every listable point, diverging in one respect legible in the intentional form of the loss itself. A deduction proves an entitlement. A pair exhibits a divergence, and can be walked to and refused. (Seat: the companion study, §7.1.)

Seventh: discovery, not validity.

The path by which the distinction was found explains its discovery. Convergent regional controls warrant it, and may defeat it. A deduction’s conclusion is not exposed to the region in which it was reached.

(Seat: the companion study, §7.1 for the warranting half, where the pair is said to be one member of a convergent constellation, substitution, loss, death, address and attribution, and not a test to be applied; and §§8.2 and 8.4 for the discovery half. There the route is reported as having been walked for other reasons and before the diagnosis was formulated, which forecloses the suspicion that the access was chosen because it would deliver the result, and the conclusion is stated as a fact about a history, not a restriction on a field.)

4.5 · Two neighbours, used and not claimed

Stroud is used and no novelty is claimed against him. He grants that transcendental arguments can establish what we must believe or how things must appear, necessity claims within thought, and objects to the bridge from thought to world. The substitution test he presses is retained here. The first difference above is in his spirit rather than against it, and he is cited for what he grants.

D’Oro is differentiated at one fork, and the fork is truth-valuedness. Her absolute presuppositions, following Collingwood, are truth-valueless commitments that make inquiry possible: the yardsticks of knowledge rather than items of it.

The closure theorem’s condition has a truth value. It is an adequacy condition on a candidate explanans, and a proposed explanans can be shown to repeat the datum or to install a constituent the phenomena have refused. That showing is true or false. So the condition is a condition on explanations rather than a presupposition of explaining.

(Had it turned out to be a presupposition of the practice of explaining, it would collapse into hers and the position would be owed to her. It does not, and the reason is stated in one sentence so that a reader can test it rather than take it.)

4.6 · Rivals on the office of the Ich denke

Part Two’s reading has rivals that are not transcendental readings of ownness at all, and they should be met on their own ground.

Kitcher and Longuenesse are rivals rather than neighbours, and the Kantian Review symposium on Kant’s Thinker is the reference debate. Busse’s contribution has already appeared at §2.4 as the printed form of the spontaneity objection, and Liang’s recent reading of the Ich denke as an act of self-ascription belongs beside it.

Smit requires a differentiation rather than a rebuttal, and it is clean. He argues that B131 does not concern the possibility of self-ascribing beliefs, and that Kant advances it to identify an a priori warrant we have as thinkers: a warrant to prescribe to given representations their conformity to principles of thinking. The vocabulary collides and the direction does not. His warrant is an entitlement the thinker holds over representations. The warrant at §2.3 is a title the representation lacks and the Ich denke holds on its behalf. His is normative and subject-side; this one is inventorial. They are compatible and arguably complementary, and saying so costs nothing.

Ensslin is cited once, as the nearest published formulation of a capacity that need not be exercised, read through Agamben’s potentiality.

And the distance is legible in his subtitle, which reads The “I can” that Accompanies none of my Representations or Actions. Kant’s claim is that the Ich denke must be able to accompany them all. The inversion is not incidental to a passing sentence; it organises the chapter. Ensslin’s I can is a potentiality that persists precisely by not passing into act, whereas the können at §2.3 is a right whose holder need not exercise it and whose non-exercise leaves every representation titled all the same. He also declines to develop the relation to Kant, and treats apperception as the ground of all experience, which is the position §2.4-bis diagnoses rather than an alternative to it.

4.7 · Longuenesse’s own, and a dispute declined

Longuenesse reads the use of I in I think as premised on taking a mental activity to be one’s own in virtue of taking oneself accountable for the correctness of its contents, with the dependence running both ways.

That is a vocabulary collision and it should be disarmed with an inherited instrument rather than a new one. The account distinguishes the who, which concerns which modal coincidence this is, from the locus, which concerns where attribution becomes non-forwardable. Longuenesse’s own is locus-side. The ownness of this article is who-side. The two cross without coinciding, and nothing in either requires the other.

And one dispute is named and declined, as §2.4-bis declines the identity thesis. Whether the I in I think refers, as Longuenesse holds against Wittgenstein and Anscombe, or does not, is orthogonal to what office the term occupies. Newton takes the second side and takes it further than most: against Evans, for whom the empirical I is encountered as an object of reference, she argues on Sartrean lines that it is nothing at all, not even a perspective. Her paper and Smit’s appear in the same issue, which is a fair measure of how unsettled the question is. This article’s question is the office. Conceding the whole referential question costs it nothing, which is also why Brook and Forgione appear here as neighbouring literature on first-person semantics, cited once, and not as rivals: their question is whether I refers, and it is not this one.

Part Five · The separation, and three answers to it

5.1 · Held possibility is not a horizon

The horizon-form of the transcendental reading fastens on one phrase, and the phrase can be repaired without losing anything.

Nothing holds the possibility open. There is no keeper of the place and no act of holding. The non-identification is not a second operation performed alongside the completeness. It is what completeness-without-exhaustion looks like when it is described by what fails to happen.

The place for another is not provided, reserved, or maintained. It is what is left standing by the manner’s never having been consumed.

Obtaining-as-this is not a use of the manner. Saying is not spending. So nothing was spent here and nothing is left over there, and the availability of another place is not an achievement but a non-event.

And every verb that would need somebody to be doing it is barred: clears, secures, preserves, guarantees, provides, reserves, keeps. A drafter who reaches for one of them has installed the horizon the reading was looking for, and the guard is written at the level of grammar for exactly that reason.

5.2 · Schematism, common root, no separation

Here the account’s positive difference from the critical architecture can be stated once, and it is better than a denial.

The schematism exists because suchness and thisness are separate in Kant’s architecture. The categories are heterogeneous from intuition; subsumption requires homogeneity between what subsumes and what is subsumed; so a third thing is needed to mediate, and the transcendental schema is that thing, a transcendental determination of time homogeneous with both sides.

Heidegger’s answer to the same problem is a common root. He asks whether the two stems of knowledge grow from a gemeinsame Wurzel, and answers that the transcendental imagination is not an external bond fastening two extremities together but the originally unifying centre from which both arise.

The account’s answer is that there was no separation. The manner and this occurrence were never independently complete, and there is therefore no seam. A coincidence is neither transcendent to its occurrence nor immanent in it, because both require a separation it never had.

A mediating term and a common root are two ways of repairing a separation. The account denies the separation.

That is a stronger contrast than a two-term one would be, because the third position is not merely the absence of Kant’s problem. It is the absence of the problem that generated both prior answers.

And the guard travels with the sentence and is never relegated to a note. This is the absence of a separation, not the presence of a fusion. Nothing was one before it was two, and nothing is being recovered.

(This is the guard to watch hardest, and it is unlike the others. It catches a slip in register rather than in grammar: a sentence that is doctrinally correct and reads as metaphysics. It is also the only guard here that a sympathetic reader would breach on the author’s behalf, since prior to both is exactly the phrase that wants to become an original unity, and a reader in sympathy will supply the unity as a kindness.)

5.3 · An objection to pre-empt

Is the transcendental imagination not itself a third position, and does §2.2’s two-position claim not fail?

No. The imagination is a faculty of synthesis operating between the two positions, and it delivers no third kind of content. Ask what it contributes to cognition and the answer is the mediation of a kind and a this, not a further sort of thing to be cognised.

The two positions are positions in the content of cognition. The imagination is an operation upon them. §2.2’s claim concerns the first and is untouched by the second.

And the same answer disposes of the parallel move on the Ich denke. That term is not a third position either, which is exactly §2.3’s finding, arrived at from the other direction: it delivers no content, and what it does instead is hold a title.

5.4 · One sentence on the ontological difference

A reader will make this identification unprompted, so it should be blocked, and blocked in one sentence rather than argued.

The relation described in Part Two is not the ontological difference, and the reason is the holder. Substance is a being, the primary one. The Ich denke is a representation, or the self-consciousness that produces one. The office is filled by a member of the inventory, and Being is not a being.

Everything further belongs elsewhere. Whether the ontological difference can itself be restated from the plural ways in which being obtains is a question this work has prepared and not answered, and it is reserved. It is deposited terrain, to be argued where it belongs and not presupposed here.

Part Six · The mystical reading, named and refused

This part is short by design. The reading is caught by a guard the account already carries, and dignifying it with a long refutation would concede that it had a foothold.

6.1 · The precedent, credited

There is a tradition of statability-limits that are not transcendental conditions, Eckhart’s most sharply, and the present account shares something real with it: the refusal to cash a stopping point into a content. Where the tradition stops, it does not then describe what it has stopped before.

The debt for the without why is acknowledged, with one correction. The rose that blooms without why is Angelus Silesius rather than Eckhart, from the Cherubinischer Wandersmann; the ohne Warum theme, and the just man who acts without a why, are Eckhart’s. The two are regularly conflated and should not be.

6.2 · The divergence, which is exact

Apophasis withholds before a surplus. What is not said exceeds saying. The silence is a mark of excess, and the withholding is a form of praise.

This account withholds before a type-mistake. What is not said would be a what where a how is at issue. The silence is a refusal to change categories, and there is no excess anywhere in it.

One withholds because there is more. The other because there is nothing of that kind to have.

6.3 · The prior unity, refused in the same breath

The sentence at §5.2 invites the mystical reading more than any other in the article, and the invitation is declined where it is issued rather than afterwards.

Prior to transcendent and immanent names the absence of a separation. It does not name an original oneness. Nothing was one before it was two. There is no unity behind the distinction, waiting to be recovered by a thinking that has learned to stop distinguishing.

6.4 · Janicaud, and why the answer is not a protestation

Janicaud’s charge is the contemporary form of the objection: that a phenomenology has smuggled a theological surplus in under a phenomenological description, and that the description’s rigour is what disguises the smuggling.

The answer is not a denial of intent, which is worth nothing, and not a declaration of method, which is worth little more.

The answer is the defeaters. A purported disclosure of ownness fails where it tracks role, status, possession or convention; where it disappears under reframing without resistance from the region; where structurally symmetrical cases receive unequal treatment without relevant reason. And one of those has been run against the account’s own central example, convicting the canvas-owner’s reading of his own grief.

A depth cannot be defeated by a case. This is defeated by three, and one of them has been used.

6.5 · Why one answer serves both readings

The two readings this article was written to block now converge, and the convergence is the article’s shortest result.

Part Two found no position in the critical apparatus for ownness to occupy. Not a kind, not a this, not a principle, not a title.

The mystical reading proposes to fill that position with a depth. The transcendental reading proposes to fill it with a condition. One approaches from the side of what conditions; the other from the side of what conditions cannot reach. Both are attempts to occupy a place the architecture does not have.

One answer serves both, and it is the same answer: there is no position there to fill.

Part Seven · No office to refill

7.1 · The problem, and why a denial cannot answer it

A position that diagnoses an archic office has to prove that it has not quietly refilled it, and the proof cannot be a denial, because a denial is what every occupant offers.

Nobody in the history of first philosophy has claimed to be installing a sovereign term. Each has claimed to be describing what was already there, or recovering what had been forgotten, or clearing away the constructions of predecessors so that something more original could show. The claim of innocence is not evidence of it, and an account that answers the charge by repeating the claim has said nothing.

So the answer has to be structural, and this part gives it. It also sets out the mode of occupation this article discovered, which the existing guards do not catch.

7.2 · The vanishing test

The test the companion study proposes is a single question, and it is worth quoting entire, because the article runs it on others and then on itself.

Think of someone who spends every day arguing against God. Their whole vocabulary, the arguments, the objections, the case against, requires God in order to be about anything at all. Take God away entirely and they are not victorious; they are silent. Everything they had to say was about the thing they were against.

That is the test. Ask of any position: what would it still be able to say if the thing it opposes simply vanished?

If the answer is “nothing,” then the thing has not been left behind. It is exactly where it was, still setting what can be said, still supplying the terms in which it is opposed, still the condition on which the opposing position is intelligible at all. The opposition does not merely fail to escape it. It depends on it, and the dependence is invisible from inside, because everything the position has to say is already said in its terms.

(Montaruli, “Across the Categorial Difference: Focality, Univocity, Ownness,” §3, DOI 10.5281/zenodo.21684041, https://ethicalontology.org/articles/across-the-categorial-difference.)

The last clause is why the test cannot be passed by declaration. A position that has not escaped what it opposes will not notice, because noticing would require a vocabulary it does not have.

The test is uncomfortable to run and it is meant to be. It is run on this article at §9.

7.3 · Three operations, and what each requires

Three well-known positions perform an operation on the metaphysical principle, and each needs the principle present in order to perform it.

Undoing requires something constructed, since there is nothing to undo where nothing was made. Weakening requires something at full strength, since a weakened principle is a principle described by comparison with itself. Deposing requires an occupant, since a vacant office cannot be vacated.

None of the three is thereby refuted. Each is a real operation and each does real work. But run the test on any of them and the answer is the same: remove the principle and the operation loses not merely its object but its vocabulary.

And the test has been run in print on the source this article inherits from. It has been argued that ontological anarchy remains parasitic on metaphysics, being oriented throughout around the archē it opposes, and that the opposition of archē and anarché should be undermined rather than occupied. Nancy’s co-originality in being-with is proposed as the way past it.

That objection is not answered here and is not conceded. What should be said is where this article stands with respect to it. A position that names itself by a negation is exposed to the charge, and anarchy is such a name. The account inherited here is not so named: it does not oppose principles, and its differentia was not reached by removing one. Whether that difference survives the objection is a question for the seat of the doctrine and not for this article, which uses the vanishing test rather than defending the position from which it was borrowed.

7.4 · The fourth mode, and why it was invisible

§2.4-bis found a mode of occupation that the account’s guards were not built for, and it should be entered here as the part’s main structural finding.

The guards were built against terms promoted from above, as a first principle is, and terms imported from outside, as an external ground is. Both are visible in the grammar of the sentence that installs them.

The mode found at §2.4-bis is neither. In the formulation used there, the principle is not deposed and not replaced but grounded in something more original, and the result is renamed a ground-structure. The office survives, its occupant survives, and what has changed is the account of where the occupant stands.

A guard-set aimed at promotion and importation does not see it coming, because nothing has been promoted and nothing has been imported.

That is why the archic function test should acquire it as a case rather than merely noting it, and why this article carries the finding forward as a deposit rather than spending it here.

7.5 · The account’s own answer, which is structural

Now the account’s own position, and the answer is not that it has declined to install a term. It is that there is nowhere for one to go.

No assembly, so there is nothing to be a part of. A term that governs the parts requires parts, and the removal test at the coincidence found none: strip the manners into their manuals and no residue sits on the other side holding them.

No proprietor, so nothing holds anything. Ownness is not ownership, and the manner is ownerless while the obtaining is own.

And no aspect of the obtaining to which the character attaches, so there is nothing detachable to be held as a title.

That third one is the load-bearing refusal, and it is worth seeing why. A title held elsewhere must be lodged somewhere. The landlord holds a deed; a deed is a thing; it sits in a drawer and can be transferred. Here there is nothing to lodge and no holder-position to lodge it in. Not because the position is occupied by someone modest, and not because the holder is distant or hidden, but because the same constraint that keeps the bearer out keeps the deed out with it.

7.6 · The evidence that the policing is at the right level

One further piece of evidence is available and it is not an assurance.

The account’s guards are grammatical. Each is stated with a tell: a syntactic signature catchable while drafting rather than a doctrine to be affirmed afterwards. Any verb that would need somebody to be doing it. Two verbs sharing ownness as a subject, joined by and. Any sentence beginning the openness is and continuing with a predicate.

Guards written against grammar are guards written by someone who expects the failure to arrive as a sentence rather than as a doctrine.

Which is how this failure does arrive. Nobody decides to install a sovereign term. The term arrives in a clause that seemed harmless, doing a small piece of work that needed doing, and by the time it is visible as a term it has a history of service behind it.

7.7 · Schürmann, and what the disagreement costs him

The part closes with the exhibit it has been building toward, and it should be read as tribute rather than as correction.

Schürmann diagnosed the archic office more precisely than anyone. He identified the modern hegemonic fantasm in the synthetic unity of apperception. He read the Deduction as an instituting document. He isolated the double modal and asked what begleiten können means. He distinguished ultimate referents from supreme ones by their relational character, which is the office described as a relation rather than as an elevation.

And he described the relation between that term and what appears by it as legislation, as action, as the making of a unity. An active relation of the one to the many. A translation of originary acting into derived actions. A spontaneous doing that is absolutely first.

That is not carelessness and this part does not report it as such. It is accuracy of a particular kind. An archè commands, and he was reading an archè, and the vocabulary of commanding is the correct vocabulary for what an archè does. A diagnosis of the office written in any other vocabulary would have been a worse diagnosis.

The man who saw the office most clearly still named the relation in the vocabulary of the thing he was diagnosing.

Which is the strongest available confirmation of the companion study’s claim that the substitutive reading is unstated, and it comes from the most sympathetic source that claim could have. Not that nobody looked. That someone looked with more care than anyone before or since, at the same sentence, and found there an acting rather than a title.

And the cost of the disagreement is therefore small and should be stated as small. If this article is right about the können, one clause of his reading is wrong and the architecture he built around it stands. If it is wrong, the article’s second horn fails and the first, which is his, survives to do most of the work anyway.

(Reflexively. If the apparatus of this article begins legislating rather than describing, if the four parts start organising claims toward themselves, or the demarcation becomes a machine for declaring what was not ontology, then it has become the thing its own test convicts, and the test runs here as anywhere. §9 states this as a falsifier and means it.)

Part Eight · What is shared

8.1 · Four inheritances

An article that diagnoses an architecture should say what it takes from it, because a diagnosis that took nothing would be suspect and would probably be wrong.

Finitude taken seriously, and not as a deficiency to be repaired by a faculty we might have had if we had been made differently. The limits are not failures of a cognition that should have been otherwise.

The refusal of intellectual intuition, and with it the refusal of any standpoint from which the whole could be surveyed. There is no place to stand outside the inventory and count it.

The corrigibility of claims, and the insistence that a claim’s exposure to correction is a feature of its standing rather than a weakness in it. This account’s defeaters are the same commitment in another register.

And the refusal of a completed inventory. There is no list which, once closed, would have said everything.

On all four this account is Kant’s inheritor and says so.

8.2 · One neighbour, named and declined

A reader will supply the comparison unprompted, so it should be made and settled rather than left to form on its own.

The Tractatus limit is a limit on saying, set by logical form, and what lies beyond it is shown rather than said. The present limit is not a limit on saying. It is a type-refusal: what is withheld is not a sayable-but-unsayable content but a content where a manner is at issue. The two vocabularies converge on cannot be said and diverge entirely on what the cannot is doing.

Longuenesse has already run this comparison and found a disanalogy at a neighbouring point, since Wittgenstein takes the I used as subject to have no referent while on her reading Kant’s I refers. She is credited for having run it; the result here is a different disanalogy and is not borrowed from hers.

8.3 · What the findings add up to

Four results have been established and they can now be set beside one another. Nothing further is argued here.

A term in the architecture that is, in the words of its most careful reader, irreducible and thus ultimate.

Phenomena that appear on its warrant and hold none of their own.

A bivalence rather than a plurality of ways of being: titled, or nothing for me.

And a residue that the structure requires and cannot say.

Even real metaphysicians are ontologists.

That sentence has to be printed, because without it everything above reads as an accusation, and it is not one. A thinker who builds an architecture around a term of this kind is not thereby a man who has substituted machinery for the phenomena. He is a man who found something and gave it the only form his epoch made available, which is the reading rule inherited at §2.0 and applied to Kant throughout Part Three.

And the structure is named rather than the man. No sentence of the form Kant is a real metaphysician appears in this article, and none is needed. The four findings describe an architecture of a certain kind, and a reader who wishes to dispute the description has four results to attack rather than one label to dismiss.

Schürmann’s own account of what the critical philosophy accomplishes is the exhibit here, and it is stated affirmatively by him rather than extracted by us. The Deduction achieves the ennoblement of the modern fantasm: degrees of being are transmuted into functions, essences into judgements and so into statements, universals into operations of the understanding, the magnified hypostasis into a maximised subjective event, and subordinated particulars into subsumed intuitions. Five transmutations, one architecture, and the older structure visible under each.

8.4 · Why the misreading this article answers was intelligible

Which returns the article to where Part One left it, and supplies something Part One could not.

If the critical architecture is an inventory ordered by a term that titles everything in it, then a reader who hears where the non-possibility falls and reaches for condition is doing something intelligible. He is hearing a how as the kind of thing the architecture he was trained in is full of. The word arrives with a history and the history has an address, and at that address the word means something that ownness is not.

So the refusal at §1.2 stops being a stipulation and becomes the diagnosis of a specific misprision with a specific cause. The transcendental reading of ownness is not a careless reading. It is what happens when a categorial claim about a manner is received in a register built for conditions, by a reader who has no reason to suspect that the register is doing the work.

(This part is written for a reader of the series, with the Kantian overhearing. Nothing in §8.3 is offered as a contribution to Kant scholarship, and a Kantian who declines the whole of it loses none of Parts Two through Seven.)

Part Nine · What would show this wrong

An argument that cannot fail has not said anything. Seven defeaters are stated in the form that would actually work, in descending order of what their success would cost.

1. A position in the critical apparatus for ownness. Produce a third thing beside concept and intuition, or show that ownness is a kind, a this, a principle, or a title after all. Part Two fails and the article’s structure with it. This is the one to attack first and everything else is downstream of it.

2. Grounding or dependence specifies admission to ontological status without residue. Show that metaphysical grounding, or existential or essential ontological dependence, gives a complete account of what constitutes admission to the inventory. §2.6 collapses into Schaffer or into Fine and the four-part structure is unnecessary. Corkum’s critics have said that the explanandum is under-specified; if one of the existing relations specifies it, this article has no work to do.

3. The können sustains only the active reading. Show that begleiten können cannot bear a title-holding construction and that Schürmann’s active relation is the only available reading of it. §2.5’s second horn fails. The first horn survives, which is his, and the transcendental reading regains the quiet slot that the second horn was written to close.

4. The residue is not the warrant’s doing. Show that the unsayability of the thing in itself follows from the ideality of space and time alone, without the titling structure. §3.2 fails at the strength stated at §3.9, and Part Three loses its principal result. (Note the strength before attacking it. The claim is that the warrant produces the residue’s unsayability, not that it alone produces the residue. An argument that defeats the second has not touched the first.)

5. Statability conditions collapse into possibility conditions. Show that the differentia becomes unstatable cannot be held apart from the object is removed. §4.4’s first difference fails and Part Four with it.

6. Corrigibility is accountable within a constituted field. Show that a misreaching corrected by resistance from within the region is fully intelligible on a two-aspect reading, without the third term. §4.4’s third difference fails.

7. How and what do not divide as claimed. Show that the refusal of a definitional content is a concealed content-claim rather than a type-refusal. §2.8 and Part Six fail together, since both rest on the division.

9.1 · And a condition on the article’s own conduct

One further defeater is of a different kind and it is not a formality.

If the apparatus of this article begins legislating rather than describing, if the four parts start organising claims toward themselves, or the demarcation at §2.6 becomes a machine for declaring what was not ontology, or the two-position finding becomes a tribunal from which architectures are judged, then the article has become the thing its own test convicts.

The vanishing test at §7.2 runs on this article as on any other. Ask what would remain of it if the transcendental reading of ownness simply vanished, and if the answer is nothing, then the reading has not been answered here but depended upon.

(The answer this article would give: Parts Two, Three and Seven would remain, since the demarcation, the residue, and the fourth mode of archic occupation are findings about an architecture rather than replies to a reading. Part One and Parts Four through Six would go, being replies. That is a real answer and it can be checked, which is the only kind worth giving.)

Part Ten · Relation to the series, and three deposits

10.1 · Received and not re-argued

Each of the following is inherited and cited to its seat rather than defended here.

The statement of the differentia. The three-term structure of differentia, mode, and differentiating character. The four-part substitutive specification of focal reference. The minimal pair and its three defeaters. The bounded closure theorem. The threefold distinction of determination, reaching, and linguistic deixis. The reception principle in its three levels. The who and locus distinction, which does new work at §4.7. And from the predecessor volume, the reading rule at §2.0 and the account of the common that stands behind §4.4’s second difference.

10.2 · Contributed here

The argument from the double modal at §2.3. The three-level reply to the spontaneity objection at §2.4, which locates the producing between apperception and the Ich denke rather than between the Ich denke and the representations. The two-horn structure at §2.5 and the exhaustion claim that follows from it. The bivalence finding at §3.2 with the residue result. The third mode of archic occupation at §2.4-bis. The three-answer contrast at §5.2, in the form that has the account denying the separation rather than solving it.

10.3 · Three deposits owed to the series

First, the archic function test acquires a third mode of occupation. A term is refilled not only by promotion from above and importation from outside but by being grounded in something more original and renamed a ground-structure. The guards as written are aimed at the first two and do not catch the third, since nothing has been promoted and nothing imported. This should be recorded where the guards are kept, with §2.4-bis as its worked case.

Second, the who and locus distinction acquires a new application. It answers accountability-based readings of ownness in the analytic literature, where taking a mental activity to be one’s own is analysed through taking oneself answerable for its contents. That own is locus-side. This account’s ownness is who-side. The instrument is inherited; the application is new and should be recorded at the terminology’s seat.

Third, the companion study’s historical claim acquires its Kant clause. That study predicts that any ontology recognising a plurality of ways of being, and lacking a non-scalar differentia, will either install a focal term or fail to say why its plurality is a plurality of being.

Kant is the case a reader will raise against it, and the raising is fair. The warrant structure is conspicuously non-scalar. Titled or nothing for me, with no degrees anywhere and no more-and-less to partition. Here, it seems, is an architecture that escaped the scale two centuries before a non-scalar differentia was proposed.

§3.2 supplies the answer, and it is not that the case is unimportant. The architecture escaped the scale by abandoning the plurality. A bivalence is not a plurality of ways of being, so the prediction’s antecedent is not satisfied and the case is not a counterexample to it. But the case is worth recording positively rather than merely excluded: the tradition’s one non-scalar architecture purchased its non-scalarity at the cost of categorial plurality itself. Which is exactly why it is no third answer. It does not answer the plurality question. It dissolves it.

That belongs at the companion study’s seat, as the Kant case of its own completeness claim.

Fourth, a methodological rule. Rule 8 governs unread primary sources. It does not govern secondary characterisations of an unread primary, which failed twice in this article’s preparation, once while being used to withdraw a claim.

A secondary characterisation of a primary text may not become load-bearing while the primary is unread. Not even to withdraw a claim.

Enter at the doctrine’s methodological section as its own rule rather than as a note on Rule 8.

10.4 · One dependency, declared

§2.6’s contribution depends on the companion study’s. The specification of admission to ontological status is stated in the vocabulary of substitutive focal reference, and that vocabulary is the companion’s. If the companion’s account of focality fails, §2.6 has no instrument and the demarcation claim goes with it.

That is not a weakness to be managed. It is the ordinary situation of an application, and it is declared here so that a reader knows where to attack if he wishes to attack the whole rather than the part.

Notes and references

Cited by part. Where a claim is inherited the seat is named; where it is constructed, that is said at the point of use.

The author’s own work

Montaruli, Positive No-Ground: The Common and the Locus of Answerability (2026a), published in full with DOI. The reading rule at §2.0, the account of the common, and the who and locus distinction at §4.7.

Montaruli, Ethical Ontology, Volume I: The Modal Coincidence (2026b), published in full with DOI. The statement of the differentia, the three-term structure, the minimal pair and the trial constellation.

Montaruli, The Calculus of Modal Coincidence: The Grammar of A-Metaphysical Ontology (2026c), DOI 10.5281/zenodo.21497200. The bounded closure theorem at §4.1.

Montaruli, The Individuating Factor: Ownness, Jemeinigkeit, Haecceitas (2026e), DOI 10.5281/zenodo.21608145. The relation to Jemeinigkeit and to haecceity, defended against a different objection.

Montaruli, Across the Categorial Difference: Focality, Univocity, Ownness, Preprint, Version 1.1, 29 July 2026. DOI 10.5281/zenodo.21684041; https://ethicalontology.org/articles/across-the-categorial-difference. Quoted directly at §2.3, for the four-part specification and the definition of the fourth reading, and at §7.2, for the vanishing test. The four-part substitutive specification at §3, quoted directly at §2.3 and this study’s instrument throughout Part Two; the two-devices argument at §5; the third answer at §7; the seats named at §4.4.

Part Two

Kant. Kritik der reinen Vernunft, B130-132, A51/B75, A137-147/B176-187, B672; Grundlegung zur Metaphysik der Sitten, AA 4:434-436. The German is quoted from the Akademie text. Translations consulted: Guyer and Wood; Pluhar; Kemp Smith, whose construction of B131 relocates the necessity and is discussed at §2.3.

Schürmann. Broken Hegemonies, trans. Reginald Lilly (Bloomington: Indiana University Press, 2003), Part Three, section I, On the Consciousness That Determines (Kant with Luther). Cited at pp. 460, 462, 463, 469, 486, 489, and note 17 at p. 634. Verified page by page. French original: Des hégémonies brisées (Mauvezin: Trans-Europ-Repress, 1996), checked: the section is De la conscience qui détermine; the three formulations and the two glosses of the modal are at pp. 568-569; sens focal de l’être and the focus imaginarius are in the corresponding note. The offset between the two editions is a constant 105. See also Heidegger on Being and Acting: From Principles to Anarchy, trans. Christine-Marie Gros with the author (Bloomington: Indiana University Press, 1987), for the standing analysis of metaphysics as ground-thinking.

On the spontaneity objection. Ralf Busse, in the symposium on Patricia Kitcher’s Kant’s Thinker, Kantian Review 19 (2014); Kitcher, Kant’s Thinker (Oxford, 2011); Yibin Liang, “Kant on Pure Apperception and Indeterminate Empirical Inner Intuition,” Ergo 11.41 (2024): 1119-1150.

On the demarcation. Jonathan Schaffer, “On What Grounds What,” in Chalmers, Manley and Wasserman, eds., Metametaphysics (Oxford, 2009), 347-383. Kit Fine (2012a), 41. Ted Sider, Writing the Book of the World (Oxford, 2011). W. V. O. Quine, “On What There Is” (1948). Amie Thomasson, Ontology Made Easy (Oxford, 2015). Phil Corkum, “Aristotle on Ontological Dependence,” Phronesis 53, no. 1 (2008): 65-92, where the notion is weakened from a capacity for independent existence to the independent possession of a certain ontological status; “Substance and Independence in Aristotle,” in Hoeltje, Schnieder and Steinberg, eds., Varieties of Dependence: Ontological Dependence, Grounding, Supervenience, Response-Dependence (Munich: Philosophia Verlag, 2013), 36-67, for the question whether things have claim to their ontological status independently of standing in a relation; and “Ontological Dependence and Grounding in Aristotle,” Oxford Handbooks Online in Philosophy (2016), DOI 10.1093/oxfordhb/9780199935314.013.31, which carries the formulation this study takes up and the note recording the gap. Michail Peramatzis, Priority in Aristotle’s Metaphysics (Oxford, 2011), 243 n. 11. Kathrin Koslicki, “Varieties of Ontological Dependence,” in Correia and Schnieder, eds., Metaphysical Grounding (Cambridge, 2012), 186-213, and “Ontological Dependence: An Opinionated Survey,” in Hoeltje, Schnieder and Steinberg, eds., Varieties of Dependence (Munich: Philosophia Verlag, 2013), 31-64, at 8 n. 6 of the preprint pagination. Christopher Shields, Order in Multiplicity (Oxford, 1999), ch. 4. Vasilis Politis and Philipp Steinkrüger, “Aristotle’s second problem about a science of being qua being,” Ancient Philosophy 37, no. 1 (2017): 59-89, arguing that the pros hen structure is necessary only and not sufficient for a science of being qua being. (The preprint circulates under the longer title, “…about the possibility of a science of being qua being: a reconsideration of Metaphysics Γ 2.” The journal version is the shorter.) Michel Maya Aranalde, “Reflexões sobre os sistemas categoriais de Aristóteles, Kant e Ranganathan,” Ciência da Informação 38.1 (2009): 86-108.

On ground. Heidegger, Der Satz vom Grund, GA 10; Vom Wesen des Grundes, in GA 9; Identität und Differenz, GA 11. Iain Thomson, Heidegger on Ontotheology (Cambridge, 2005).

§2.4-bis. Heidegger, Kant und das Problem der Metaphysik, GA 3, §34, Die Zeit als reine Selbstaffektion und der Zeitcharakter des Selbst, pp. 188-195. Translations: James S. Churchill (Bloomington: Indiana University Press, 1962), from the second German edition; Richard Taft, 5th enlarged edn (Bloomington: Indiana University Press, 1990), where §34 opens at p. 134. Taft gives the Urstruktur sentence at p. 137, as pure self-affection provides the transcendental, primordial structure of the finite self as such, and the identity thesis at p. 138: time and the “I think” no longer stand incompatibly and incomparably at odds; they are the same. Taft is the translation of record here. On the identity thesis: Ernst Cassirer, review, Kant-Studien 36 (1931): 1-26.

Part Three

On the two-aspect reading and affection. Henry Allison, Kant’s Transcendental Idealism, revised edn (Yale, 2004), 42-45. Paul Guyer, Kant and the Claims of Knowledge (Cambridge, 1987), 336-342. James Van Cleve, Problems from Kant (Oxford, 1999). Erich Adickes, Kant und das Ding an sich (Berlin: Pan, 1924). Nicholas Stang, “Who’s Afraid of Double Affection?,” Philosophers’ Imprint 15.18 (2015), and “Kant’s Transcendental Idealism,” Stanford Encyclopedia of Philosophy. R. Lanier Anderson, “Transcendental Idealism as Formal Idealism,” European Journal of Philosophy (2022), DOI 10.1111/ejop.12753. Rae Langton, Kantian Humility (Oxford, 1998); Lucy Allais, Manifest Reality (Oxford, 2015).

On correlationism. Quentin Meillassoux, After Finitude, trans. Ray Brassier (Continuum, 2008). Dionysis Christias, “Toward the Thing-in-Itself,” in Gironi, ed., The Legacy of Kant in Sellars and Meillassoux (Routledge, 2018), 21-40.

On flat ontologies. Levi Bryant, The Democracy of Objects (Open Humanities Press, 2011). Graham Harman, The Quadruple Object (Zero, 2011). Ray Brassier, “Deleveling: Against ‘Flat Ontologies’.”

Other residue diagnoses. F. H. Jacobi, “Ueber den transzendentalen Idealismus,” supplement to David Hume über den Glauben (1787), JWA 2,1: 109; English in di Giovanni, ed. and trans., The Main Philosophical Writings and the Novel Allwill (McGill-Queen’s, 1994), 336, and in Sassen, Kant’s Early Critics (Cambridge, 2000), 173. The two translations differ and must not be conflated. Hegel, Encyclopaedia of the Philosophical Sciences, Part One, §44, quoted from William Wallace, trans., The Logic of Hegel (Oxford: Clarendon Press, 1874), 77, the translation of record here; the caput mortuum recurs in the Doctrine of Essence. Salomon Maimon, Essay on Transcendental Philosophy, trans. Midgley et al. (Continuum, 2010). P. F. Strawson, The Bounds of Sense (Methuen, 1966). Murray Miles, “Much Ado about ‘Something (Etwas)’,” Kantian Review 30.2 (2025): 235-258, DOI 10.1017/S1369415425000147.

Parts Four to Six

On transcendental arguments. Barry Stroud, “Transcendental Arguments,” The Journal of Philosophy 65.9 (1968): 241-256, and Understanding Human Knowledge (Oxford, 2000). Giuseppina D’Oro, “Between Ontological Hubris and Epistemic Humility,” British Journal for the History of Philosophy 27.2 (2019): 336-357, DOI 10.1080/09608788.2018.1471660. Robert Stern, ed., Transcendental Arguments (Oxford, 1999).

On the office of the Ich denke. Houston Smit, “Kant’s ‘I think’ and the agential approach to self-knowledge,” Canadian Journal of Philosophy 49.7 (2019): 980-1011, DOI 10.1080/00455091.2019.1615682. Béatrice Longuenesse, I, Me, Mine: Back to Kant, and Back Again (Oxford, 2017). Alexandra M. Newton, “Kant and the transparency of the mind,” Canadian Journal of Philosophy 49, no. 7 (2019): 890-915, DOI 10.1080/00455091.2019.1565613, in the same issue as Smit above. Her earlier “Kant on the Logical Origin of Concepts,” European Journal of Philosophy 23, no. 3 (2015): 456-484, DOI 10.1111/j.1468-0378.2012.00545.x, argues that the form of a concept is the self-consciousness of a capacity to judge, so that the Ich denke is the Concept, and expressly brackets the question of reference at its n. 39. It is noted here so that it is not mistaken for the paper bearing on §4.7. Andrew Brook and Julian Wuerth, “Kant’s View of the Mind and Consciousness of Self,” Stanford Encyclopedia of Philosophy. Luca Forgione, Kant and the Problem of Self-Knowledge (Routledge, 2019). Felix Ensslin, “Potentiality in Agamben: The ‘I can’ that Accompanies none of my Representations or Actions,” in Mark Potocnik, Frank Ruda and Jan Völker, eds., Beyond Potentialities? Politics between the Possible and the Impossible, Subjektile series (Zürich: Diaphanes, 2011), 121-136. The subtitle is the inversion discussed at §4.6.

§6. Angelus Silesius, Cherubinischer Wandersmann (1657), I.289. The rose is his and not Eckhart’s; the ohne Warum theme is Eckhart’s. Dominique Janicaud, Le tournant théologique de la phénoménologie française (Combas: Éditions de l’Éclat, 1991).

§8. Ludwig Wittgenstein, Tractatus Logico-Philosophicus (1921), on the limit of saying.

Citations owed

No citation in this study is outstanding. Every page reference has been checked against an edition, and the two Schürmann passages on which the disagreement at §2.5 turns have been checked against the French original as well as the translation.

• Articles The other deposited preprints and their identifiers.
• Doctrine The standing statement of the category whose differentia is at issue here.