六条公设，不是六块地基
Six Postulates, Not Six Foundations
公设意味着承担
哲学体系喜欢把重要命题叫作“基础”。这个词既能表示工作从这里开始，又容易暗示它们不能被替换、已经由更深处证明。《第一批判》刻意改用公设：独立设定、可拒绝、可替换的承担。
公设并不因为可拒就随意。承担一条公设，意味着接受它带来的推论、成本与依赖；拒绝它，也要在另一分支上重建相应工作。公开承担比把选择藏进“常识”“逻辑本身”或“认识的本性”更严格。
六条公设不是“非不能非”的六个定理。若能由唯一公理推出，它们便不该叫公设。它们连接零点与 SAE 的具体方法，却不把这条连接冒写成唯一道路。
P1：把认识理解为划界
第一条是操作公设：在这套方法里，把认识理解为划界或否定。认识某物，就是作出区分，把“这”从“非这”中标出。它是一座桥，使施行性公理进入具体作业。
这条桥可以拒绝。有人可以把认识首先理解为参与、体验、显现或实践，而不接受划界是基本动作。拒绝 P1 不构成自相矛盾，只意味着他不进入这套操作语言。
P1 与后五条不同，它是整套形式化的入口。可以不走，却无法一面拒绝入口，一面要求后面的账目仍按原样成立。
P2 与 P3：余项不可灭，账目可传导
P2 说每次收纳都有亏，任何处理不能把余项彻底消掉。其直觉是：清除一项亏欠的动作本身又形成新的划界。P3 说余项可以转化、换相与迁移，但不能无后继地从总账消失。
两条看起来很近，却承担不同工作。P2 保证每一刀有非空欠账；P3 规定条目离账必须有转化后的去处。若把它们混成一句“总有余项”，依赖链便无法说明某个结果来自不可灭，还是来自守恒。
它们仍是公设。形式构形可以保留 P2 而拒绝 P3，也可以用不同的余项概念重写二者。分离测试的意义不是证明它们真实，而是证明一条没有被另一条暗中包含。
P4 与 P5：构会沉淀，步骤彼此离散
P4 是沉淀公设：一次划界可以留下相对稳定的构，使后续作业不必每次从零开始。没有这条，认识只剩瞬时事件，记忆、概念与依赖都难以登记。
P5 是离散公设：作业可被区分成步骤，步骤之间能标出先后与地址。连续经验当然可以存在；这条公设只承担我们能够把其中的作业切成可审计单位。
稳定不等于永恒，离散也不等于宇宙本体由格子组成。P4 与 P5 都是方法承担：为了记账，我们承认构可以保存、步骤可以辨认。把方法条件升级为存在论结论，又会落入冒领。
P6：刀可以回到握刀者
第六条是反身公设：否定能够以自身的施行者与产物为对象。主体可以审查自己的判断，理论可以把自己的文本、规则和形式化送回审计。
反身不等于完整自知。恰恰相反，它只说明刀能转回来，不保证转回来以后看见全部。外部追问仍有自问无法自证的独立价值。
六条合在一起，给出 SAE 认识作业的一套可用骨架：以划界开始，余项不被消灭而可传导，区分沉淀为构，作业可分步，刀能回身。它们不是六块压在所有哲学之下的地基，而是这门哲学署名采用的六项条件。知道它们可以替换，才真正知道自己正站在哪里。
A Postulate Is an Undertaking
Philosophical systems often call their central commitments foundations. The word can mean only that work starts here, but it also suggests that the commitments cannot be replaced or have already been demonstrated from something deeper. The First Critique deliberately calls its six commitments postulates: independent, rejectable, replaceable undertakings.
Rejectability does not make a postulate arbitrary. To assume one is to accept its consequences, costs, and dependencies. To reject it is to rebuild the relevant work on another branch. An explicit undertaking is stricter than hiding a choice inside “common sense,” “logic itself,” or “the nature of cognition.”
The six postulates are not six theorems of Negativa. If the sole axiom derived them, they should not be called postulates. They connect the zero point to SAE's particular method without falsely claiming that the connection is the only possible road.
P1: Cognition as Demarcation
The first is operational: in this method, cognition is understood as demarcation or negation. To cognize something is to make a distinction, marking “this” against “not-this.” P1 is the bridge from the performative axiom into concrete work.
The bridge can be declined. Another philosophy may understand cognition primarily as participation, experience, disclosure, or practice without treating demarcation as the basic operation. Rejecting P1 is not performatively contradictory. It means declining this operational language.
P1 differs from the next five because it is the entry into the formal package. One may refuse the entrance, but not refuse it while demanding that all later bookkeeping remain available unchanged.
P2 and P3: Indestructibility and Conservation
P2 states that every intake has a deficit and that no treatment eliminates remainder altogether. The motivating thought is that an operation clearing one deficit makes another demarcation. P3 says that remainder may transform, change type, and move, but may not leave the ledger without a successor.
The postulates look similar while performing different work. P2 ensures a nonempty debt at each cut. P3 governs how an entry may leave the account. If both collapse into “there is always remainder,” dependencies can no longer show whether a later result relies on indestructibility or conservation.
They remain postulates. A formal configuration can preserve P2 while rejecting P3, or reconceive remainder and rewrite both. A separation test does not prove either true. It shows that one commitment has not been silently packed inside the other.
P4 and P5: Sedimentation and Discreteness
P4 is the sedimentation postulate. A demarcation may leave a relatively stable construct, allowing later work to avoid beginning from nothing. Without it, cognition would consist only of vanishing events and could scarcely register memory, concepts, or dependencies.
P5 is the discreteness postulate. Work can be distinguished into steps with order and address. Continuous experience may exist; the undertaking says only that operations within it can be marked as auditable units.
Stability does not mean eternity, and discreteness does not mean that reality is fundamentally made of boxes. Both are methodological undertakings. For bookkeeping, constructs may persist and operations may be distinguished. Promoting those conditions into an ontology would be another credential failure.
P6: The Instrument Can Turn on Its User
The sixth postulate is reflexivity. Negation can take its own performer and products as objects. A subject may examine a judgment; a theory may return its texts, rules, and formalization to the audit.
Reflexivity is not complete self-knowledge. It says that the instrument can turn inward, not that it will see everything when it does. External interrogation retains an independence self-review cannot certify.
Together the six postulates provide a working anatomy of SAE cognition. Demarcation begins the operation. Remainder cannot be eliminated but can be propagated. Distinctions sediment as constructs. Work can be separated into steps. The instrument can return to its user. These are not six foundations beneath every philosophy. They are six conditions this philosophy assumes under its own name. Knowing that they can be replaced is part of knowing where one stands.