Non Dubito Essays in the Self-as-an-End Tradition
← SAE第一批判← The SAE First Critique
SAE 基础文本 · 第一批判
SAE Foundational Text · The First Critique
第 01 篇,共 12 篇
Essay 01 of 12

先审自己

Critique Begins at Home

Han Qin (秦汉)

一把先转回来的刀

一门哲学通常在什么时候开始批判自己?最常见的答案,是等别人找出了问题以后。理论先把自己的体系搭起来,说明它能解释什么,再把异议安排在末尾:适用范围、可能反例、未来研究。自我批评因此像建筑竣工后的消防检查,重要,却不属于建筑本身。

《SAE第一批判》采取了相反的次序。它不是在 SAE 的理论旁边附上一份免责声明,而是把批判安排成框架继续成立的条件。SAE 一直要求每一项主张接受追问;那么第一个不能逃开的对象,就是这项要求本身。若“不得不被追问”只约束别人,不约束提出它的人,它就不是方法,而是特权。

所以这部批判的开场不是扩张,而是撤回豁免。那一条从“非”出发的公理要重新摆上桌;六条公设要各自说明自己是推出来的,还是承担下来的;八类幻相要先在自家文本里查;十六层发生序列也不能因为是框架最珍视的部分便免审。刀不是先对外,后对内。它必须在第一次落下时就可以转向握刀者。

批判不是宣布怀疑

这和泛泛地说“一切都可能错”不同。无限怀疑很容易摆出谦逊的样子,却不必做任何工作。只要说所有结论都暂时有效,理论便可以一面继续照常使用,一面把每项责任推给未来。

《第一批判》要求的是更具体的动作:一项主张说了什么,它在哪里成立,依赖了哪些前提,担保从哪里来,更换前提以后哪些条目要被送回重审。被审的不是作者是否真诚,也不是一套思想值不值得尊敬,而是主张实际出示的凭证是否超过它已经挣得的地位。

因此“批判”在这里接近公开记账。不是把所有账都判成坏账,而是先把可推、登记与假说分开;把公理、公设、分支公设与极限名分开;把已经走过的历史事实与能否从结构上推出的必然性分开。模糊的谦虚不能替代这些区分。

五章为什么按这个次序出现

全书五章不是五个并列主题。导论先问批判凭什么开庭;辖域论问每一项判决管到哪里;幻相论问一个“构”怎样冒领它没有取得的地位;公设论把形式工具、公设与分支逐项登记;发生论则把审计落到 SAE 的 0D—16DD 序列上。

这个次序很要紧。没有辖域,批判会把局部结论说成普遍真理;没有幻相,越界只会被当作普通错误;没有公设登记,形式化会把借来的规则伪装成中立背景;没有发生论,前四章就只审别人,不审 SAE 自己最重的道路。

四个附录也不是资料仓库。六栏分级表把发生论的判词摊平;五个构形检验六条公设能否彼此分离;七栏忠实性登记审查散文转成形式语言时保留和丢失了什么;跨文映射则处理同一术语在不同论文里的地址。正文作判,附录让判词可复核。

规则也必须有地址

这部批判给自己的第一条纪律,可以写成一句很朴素的话:没有哪位裁判天然坐在场外。提出辖域规则的人也得说明这条规则的辖域;诊断幻相的人也可能制造新的幻相;要求别人公开前提的人也必须公开自己使用的分支逻辑。

这种反身性不会制造一位更高的最终裁判。若每次追问都要再找一位绝对可靠的人来担保,链条只会无限上移。这里选择的不是终审,而是常设复核:地址可以被纠正,地位可以被降档,依赖可以重新传导,读法可以进入登记册。

于是“没有最终裁判”并不等于“谁说什么都一样”。开放的是复核,不是效力。一个可推的结论与一项尚待承担的公设仍然不同;一个把边界写明的局部判断,也比把局部偷换成整体的判断更有分量。平等的是进入登记的资格,不是每一项主张的成色。

读者不是观众

全书最后把自己“入册,立于被追问之下”。这不是漂亮的收尾,而是开头规则的兑现。作者不能宣布自己的自审已经完成,因为自问无法担保发现自己的全部盲区。文本写毕,只能说明作者这一轮作业结束,不能说明账已经关上。

因此读者进入的不是一座已经盖好的体系,而是一场仍在进行的审计。读者可以追问某条公设是否真的独立,可以指出某个形式写法删去了散文里关键的动作,也可以拒绝一条从结构事实到规范效力的移交。反对并不自动取胜,但它取得了一个正式地址。

《SAE第一批判》的第一项成果不是证明 SAE 正确,而是取消 SAE 不受审的资格。只有做到这一点,后面关于辖域、幻相、公设和发生的判词才不会成为另一套对外使用的器械。批判从家里开始,也必须把家门留给后来的人。

Turning the Instrument Inward

Philosophical systems often become self-critical late. They first establish their vocabulary, demonstrate their range, and build an account of the world. Qualifications arrive afterward under familiar headings: limitations, objections, future work. Self-criticism functions like an inspection performed after construction. It may be conscientious, but it is not part of the architecture.

The SAE First Critique reverses that order. It does not append a modest disclaimer to the Self-as-an-End framework. It treats self-audit as a condition of the framework's continued standing. SAE repeatedly insists that claims must remain open to interrogation. If that demand governs rival theories and ordinary agents but not the theory making the demand, it is not a method. It is a privilege.

The book therefore begins by withdrawing the exemption. The performative axiom associated with Negativa returns to the table. Six postulates must declare whether they are derived or undertaken. Eight forms of false claiming must be sought first in SAE's own corpus. Even the 0D–16DD sequence—the part of the framework with the greatest symbolic and systematic weight—receives no sanctuary. An instrument that can cut only outward has already failed its own test.

Critique Is More Than General Doubt

This project should not be confused with the soothing admission that “everything may be wrong.” Unlimited doubt can look humble while doing almost no work. A theory labels every claim provisional, continues to use it unchanged, and transfers responsibility to an unspecified future correction.

The First Critique requires a more exact practice. What does a claim say? Where does it have standing? Which premises does it require? What kind of warrant supports it? If a postulate changes, which dependent claims must be returned for review? The audit does not ask whether an author is sincere or whether a tradition deserves admiration. It compares the credential a claim uses with the status it has actually earned.

Critique here resembles public bookkeeping. Derivation, registration, and hypothesis must not share one unmarked column. Axioms, postulates, branch postulates, and limit-names must not borrow one another's authority. A fact about how a sequence was historically produced must not silently become proof that the sequence was necessary. Generic modesty cannot perform these separations.

Why the Five Chapters Have an Order

The book's five chapters are not a collection of neighboring topics. The introduction asks what gives critique standing at all. “On Jurisdiction” asks how far each judgment reaches. “On Illusion” explains how a construct claims a status it has not earned. “On Postulates” registers formal equipment, explicit undertakings, and branch choices. “On Genesis” finally subjects SAE's own graded sequence to the completed audit.

The order is functional. Without jurisdiction, a local result becomes a universal truth. Without an account of illusion, overreach appears as a casual mistake rather than a credential failure. Without a postulate registry, formalization presents borrowed logical rules as neutral background. Without the final audit of genesis, the first four chapters become instruments for examining everyone except the theory's most valued claims.

The appendices continue the work rather than storing supplementary material. A six-column table exposes the judgments made about the graded sequence. Five formal configurations test whether the six postulates are separable. A seven-column fidelity registry records what a formal rendering preserves, loses, or adds. Cross-text mappings give repeated terms explicit addresses. The main chapters judge; the appendices make those judgments inspectable.

Rules Need Addresses Too

The opening discipline can be stated simply: no judge occupies an innocent position outside the field. Whoever proposes a rule of jurisdiction must state that rule's jurisdiction. Whoever diagnoses illusion can generate a fresh illusion. Whoever demands explicit premises must disclose the branch logic used in the demand.

Reflexivity does not produce a higher and final court. If every audit required an absolutely reliable auditor, responsibility would merely move upward forever. The First Critique chooses standing review over final review. An address can be corrected, a status downgraded, a dependency propagated, and an interpretation entered into a public registry.

The absence of a final judge does not make all claims equivalent. Openness belongs to review, not to validity. A derivable result still differs from an adopted postulate. A local judgment whose boundary is explicit has more standing than one that quietly promotes itself into a total account. What is equal is access to the registry, not the quality of every entry.

The Reader Is Not an Audience

The book closes by placing itself in the register and under interrogation. This is not decorative humility. It fulfills the rule announced at the beginning. An author cannot certify that self-audit has found every blind spot, because self-questioning cannot guarantee its own independence. Completion of the manuscript marks the end of one round of work, not the closing of the account.

Readers therefore enter an audit in progress, not a finished monument. They may challenge the independence of a postulate, identify an action lost when prose became notation, or reject a transfer from structural description to normative authority. An objection does not win merely by appearing. It does, however, acquire a legitimate address at which it must be heard.

The First Critique's first achievement is not a proof that SAE is correct. It is the cancellation of SAE's immunity from examination. Only after that cancellation can its judgments about jurisdiction, illusion, postulates, and genesis avoid becoming a new set of tools reserved for use on others. Critique begins at home, and the door must remain open after the author leaves the room.

本系列依据《SAE第一批判》中文原论文重写;原论文为完整论证与权威底本。阅读原论文 ↗ · DOI ↗ This reader-facing essay is independently rewritten from SAE: The First Critique; the complete paper remains the authoritative argument. Read the paper ↗ · DOI ↗