# 悬而未决是决定论的索求 / The Open Question Is Determinism's Demand

*说抉择“被决定与否尚无定论”，已先把抉择当成一件可被固定的既成之事 / To call a choice "not yet settled as determined or not" already treats it as a finished thing that could be fixed*

“演化为什么选中了人类”预设了一位选择者，这个预设可以拿掉，拿掉之后对演化的描述仍然连贯。“每一个有意识的生命为自己抉择并承担后果”则拿不掉，对它的每一次否认都是它的又一个实例。走到这里，常见的一步是退回中立：这还没有确定抉择是被决定的还是未被决定的；相容论再进一步，认为抉择可以被先在的条件固定而仍然属于行动者。这份中立只在已经索求决定论的时候才立得住，因为“被固定与否”问的是一件已完成的事，而发问与作答的都是仍在进行的区分。

"Why did evolution choose humans" presupposes a chooser, and that presupposition can be dropped while the description of evolution stays coherent. "Each conscious being chooses for itself and bears the consequences" cannot be dropped, since every denial of it is one more instance of it. At this point a common step is the retreat to neutrality: none of this settles whether a choice is determined or undetermined; compatibilism goes one step further and holds that a choice can be fixed by prior conditions and still belong to the agent. That neutrality stands only where determinism is already demanded, because "fixed or not" is asked of a finished thing, while what asks and what answers is the differentiating still under way.

## 一、 可以拿掉的选择者 / 1. A Chooser That Can Be Dropped

有一个向无神论者提出的问题：演化为什么选中人类做最聪明的物种，而不是猴子或狗？这个问题预设人类是被某人或某物选中的。预设一旦被接受，能回答的就只有那位选择者，被选中的一方答不了；要被选中者作答，等于要它从一个它并不占据的位置上作出报告。

There is a question put to atheists: why did evolution choose humans as the most intelligent species, and not monkeys or dogs? The question presupposes that humans were chosen by someone or something. Once the presupposition is accepted, only the chooser can answer it, not the chosen; to ask the chosen for the answer is to ask for a report from a seat the chosen does not occupy.

“自然选择”这个词是达尔文从育种者那里借来的，育种者确实在选。理论留下了这个词，拿掉了育种者：变异的个体留下的后代数目不同，世系随之偏移。这样使用的时候，人们用它追溯亲缘、解释抗药性的蔓延、解释喙的形状，不需要设想任何预见。它的边界在这个词被读回成一位选择者的地方，在“最聪明”被读成这位选择者所瞄准的名次的地方；名次所依的那把尺，是发问的人自己立的。拿掉选择者，关于后代数目差异的描述照样连贯，这就是一个可以拿掉的前提的标志。

The phrase "natural selection" was borrowed by Darwin from breeders, and breeders do choose. The theory kept the word and removed the breeder: variant individuals leave different numbers of offspring, and lineages shift accordingly. Used in this way, people trace kinship with it, explain the spread of drug resistance, explain the shape of a beak, with no foresight supposed anywhere. Its boundary lies where the word is read back into a selector, and where "most intelligent" is read as the rank that selector was aiming at; the ruler behind the rank was set up by the one who asks. With the chooser dropped, the description of differing numbers of offspring remains coherent, and that is the mark of a premise that can be dropped.

选择者拿掉之后剩下的是：没有别人替谁作出抉择。每一个有意识的生命为自己抉择，选这一样，选另一样，或者什么都不选，并承担由此而来的后果。[主体性不曾涌现](../agency-does-not-arise/)追溯过这一点：主体不是演化过程后来的产物，讲述这个过程的正是主体。

```mermaid
flowchart TD
    subgraph S_Chooser["可以拿掉的选择者"]
        direction TB
        Q["问题：演化为什么选中了人类"]
        P["预设：有一位选择者"]
        Drop["拿掉选择者"]
        Keep["后代数目的差异与世系的偏移，描述仍然连贯"]
        Left["剩下的：每一个有意识的生命为自己抉择并承担后果"]
        Q --> P
        P --> Drop
        Drop --> Keep
        Drop --> Left
    end
    style S_Chooser fill:#161b22,stroke:#58a6ff,stroke-width:1.5px,color:#58a6ff
    style Q fill:#0d1117,stroke:#58a6ff,stroke-width:1px,color:#f0f6fc
    style P fill:#0d1117,stroke:#f85149,stroke-width:1px,color:#f0f6fc
    style Drop fill:#0d1117,stroke:#d29922,stroke-width:1px,color:#f0f6fc
    style Keep fill:#0d1117,stroke:#58a6ff,stroke-width:1px,color:#f0f6fc
    style Left fill:#0d1117,stroke:#3fb950,stroke-width:1px,color:#f0f6fc
```

What remains once the chooser is dropped is this: no one else makes the choice for anyone. Each conscious being chooses for itself, one thing, another, or nothing at all, and bears the consequences that follow. [Agency Does Not Arise](../agency-does-not-arise/) has traced this: agency is no late product of the evolutionary process, and the one who tells of that process is an agent.

```mermaid
flowchart TD
    subgraph S_Chooser_EN["A Chooser That Can Be Dropped"]
        direction TB
        QE["The question: why did evolution choose humans"]
        PE["The presupposition: there is a chooser"]
        DropE["Drop the chooser"]
        KeepE["Differing numbers of offspring and shifting lineages: the description stays coherent"]
        LeftE["What remains: each conscious being chooses for itself and bears the consequences"]
        QE --> PE
        PE --> DropE
        DropE --> KeepE
        DropE --> LeftE
    end
    style S_Chooser_EN fill:#161b22,stroke:#58a6ff,stroke-width:1.5px,color:#58a6ff
    style QE fill:#0d1117,stroke:#58a6ff,stroke-width:1px,color:#f0f6fc
    style PE fill:#0d1117,stroke:#f85149,stroke-width:1px,color:#f0f6fc
    style DropE fill:#0d1117,stroke:#d29922,stroke-width:1px,color:#f0f6fc
    style KeepE fill:#0d1117,stroke:#58a6ff,stroke-width:1px,color:#f0f6fc
    style LeftE fill:#0d1117,stroke:#3fb950,stroke-width:1px,color:#f0f6fc
```

## 二、 拿不掉的抉择 / 2. A Choice That Cannot Be Dropped

把上面这段话交给一个语言模型，模型把它当作一个有待评估的主张来处理。这样处理也是一次抉择，作出它的是造出模型的有意识的生命，他们定下了模型的目标、数据与评估的标准；模型没有作出它，[模型永远无法成为第二重前沿](../the-model-never-becomes-a-second-edge/)。使用者接受这个输出，输出就成了使用者自己的抉择；拒绝它、不理它、换一种问法再问，同样是抉择，后果也由使用者承担。

When the passage above is handed to a language model, the model handles it as a claim to be evaluated. Handling it so is also a choice, made by the conscious beings who built the model and set its objectives, its data and its standards of evaluation; the model does not make it, and [The Model Never Becomes a Second Edge](../the-model-never-becomes-a-second-edge/). A user who accepts the output has made it the user's own choice; rejecting it, ignoring it, or asking again in other words are choices as well, and their consequences fall to the user.

把这个前提也照前一个那样往回追，看它能不能拿掉。拿掉它的办法只有一种，就是说这次否认并不是选的，是被迫的。可是否认是说话的人正在做的事，也是说话的人要为之负责的事。任何反驳都已经是它的又一个实例，[自由意志的自否自由](../the-freedom-of-will-to-deny-itself/)所说的正是这个结构。同一种检验之下，两个前提走向相反：一个拿掉之后，描述安然无恙；另一个在被拿掉的动作里重新出现。

```mermaid
flowchart TD
    subgraph S_Irreducible["拿不掉的抉择"]
        direction TB
        Makers["造模型的人：定下目标、数据与评估"]
        Out["模型的输出"]
        User["使用者：接受、拒绝、不理或重问"]
        Deny["否认：这不是选的"]
        Inst["否认是说话的人在做并为之负责的事"]
        Makers --> Out
        Out --> User
        User --> Deny
        Deny --> Inst
        Inst -.->|"又一个实例"| User
    end
    style S_Irreducible fill:#161b22,stroke:#3fb950,stroke-width:1.5px,color:#3fb950
    style Makers fill:#0d1117,stroke:#3fb950,stroke-width:1px,color:#f0f6fc
    style Out fill:#0d1117,stroke:#58a6ff,stroke-width:1px,color:#f0f6fc
    style User fill:#0d1117,stroke:#3fb950,stroke-width:1px,color:#f0f6fc
    style Deny fill:#0d1117,stroke:#f85149,stroke-width:1px,color:#f0f6fc
    style Inst fill:#0d1117,stroke:#a371f7,stroke-width:1px,color:#f0f6fc
```

Trace this premise back as the first was traced, to see whether it too can be dropped. There is only one way of dropping it, which is to say that the denial was not chosen but forced. Yet the denial is something the speaker is doing, and something the speaker answers for. Any refutation is already another instance of it, the structure laid out in [The Freedom of Will to Deny Itself](../the-freedom-of-will-to-deny-itself/). Under one and the same test the two premises go opposite ways: the first drops out and leaves the description unharmed; the second reappears in the very act of dropping it.

```mermaid
flowchart TD
    subgraph S_Irreducible_EN["A Choice That Cannot Be Dropped"]
        direction TB
        MakersE["The makers of the model: set objectives, data and evaluation"]
        OutE["The model's output"]
        UserE["The user: accepts, rejects, ignores or asks again"]
        DenyE["The denial: this was not chosen"]
        InstE["The denial is something the speaker does and answers for"]
        MakersE --> OutE
        OutE --> UserE
        UserE --> DenyE
        DenyE --> InstE
        InstE -.->|"one more instance"| UserE
    end
    style S_Irreducible_EN fill:#161b22,stroke:#3fb950,stroke-width:1.5px,color:#3fb950
    style MakersE fill:#0d1117,stroke:#3fb950,stroke-width:1px,color:#f0f6fc
    style OutE fill:#0d1117,stroke:#58a6ff,stroke-width:1px,color:#f0f6fc
    style UserE fill:#0d1117,stroke:#3fb950,stroke-width:1px,color:#f0f6fc
    style DenyE fill:#0d1117,stroke:#f85149,stroke-width:1px,color:#f0f6fc
    style InstE fill:#0d1117,stroke:#a371f7,stroke-width:1px,color:#f0f6fc
```

## 三、 “尚无定论”里的索求 / 3. The Demand inside "Not Yet Settled"

走到这里，接下来常见的一步是承认以上各点，然后补上一句：这并没有确定这样的抉择是被决定的、未被决定的，还是别的什么。据说有两种读法都与之相容。一种认为抉择不被先在的条件固定；另一种认为抉择被先在的条件与规律固定，却仍然算作行动者自己的抉择，可以据以讲理、据以归责。后一种读法从霍布斯与休谟以来叫作相容论。

At this point the step that commonly follows is to grant all of the above and then add: this does not settle whether such choices are determined, undetermined, or something else. Two readings are said to be compatible with it. On one, the choice is not fixed by prior conditions; on the other, the choice is fixed by prior conditions and laws and still counts as the agent's own, something to reason from and to hold someone responsible for. The second reading has been called compatibilism since Hobbes and Hume.

相容论有它的用处。人们用它来协调彼此的行动：分开出于自己的理由而做的事与在胁迫之下做的事，据此称许、追责、订约、教养孩子，而不必先等任何关于来源的说法尘埃落定。法庭与日常的交往每天都在用这条区分。

Compatibilism has its use. People use it to coordinate their actions with one another: to tell what is done from one's own reasons apart from what is done under threat, and on that basis to praise, to hold liable, to make contracts, to bring up children, without first waiting for any account of the source to be concluded. Courts and everyday dealings work with this distinction daily.

它的边界在“被固定”这个词上。问一次抉择“是否被固定”，是把它当作一件已经完成的事来问，把它摆进它的先在条件之间，并且设想有一道目光能把这些条件一次看全。这正是[没有固定者的固定](../fixed-by-nothing/)追到的那个无人能满足的前件：在那里，“两种读法同等欠定”的中立站在一份已完成的完整状态上；在这里，同一个结构出现在抉择的动作上。往回追所走到的，是抉择本身这一开放的起点；进一步追问某一次抉择“究竟”有没有被固定，是要求事先有一份固定摆在那里，好让这次抉择拥有或者缺少。只有在索求决定论的时候，这件事才显得尚无定论。断言它尚无定论，那个隐藏的决定论前提就又滑了回来。

Its boundary lies in the word "fixed." To ask of a choice "whether it is fixed" is to ask about it as a thing already finished, to set it among its prior conditions, and to suppose a gaze that could take in those conditions whole at once. This is the antecedent no one can satisfy that [Fixed by Nothing](../fixed-by-nothing/) arrived at: there, the neutrality of "both readings equally underdetermined" stood on a completed total state; here, the same structure appears at the act of choice. What the trace arrived at is choosing itself, the open origin, and to ask further whether a choice is "really" fixed is to require that a fixedness be there in advance for it to have or to lack. It only stays unsettled if determinism is demanded. Asserting that it is unsettled slips the hidden deterministic premise back in.

中立的回答是：抉择与被决定不相容，这是一项额外的主张，还需要论证。可是引入一项可供争辩的额外主张，本身就是抉择的又一个实例，而不是对抉择不可化约性的反驳。这个回答还坚持两种读法的差别不止于措辞，差在“哪一种事态成立”；这又一次假定有一份事态已经完成、摆在那里等着成立。

```mermaid
flowchart TD
    subgraph S_Demand["“尚无定论”里的索求"]
        direction TB
        Reached["已经走到的：抉择拿不掉"]
        Neutral["中立：被决定与否尚无定论"]
        Asks["所问的：一件已完成的事是否被先在条件固定"]
        Gaze["所需的：能把先在条件一次看全的目光"]
        Slip["决定论的前提滑了回来"]
        Use["相容论的用处：分开自愿与胁迫，协调归责"]
        Reached --> Neutral
        Neutral --> Asks
        Asks --> Gaze
        Gaze --> Slip
        Neutral --> Use
    end
    style S_Demand fill:#161b22,stroke:#f85149,stroke-width:1.5px,color:#f85149
    style Reached fill:#0d1117,stroke:#3fb950,stroke-width:1px,color:#f0f6fc
    style Neutral fill:#0d1117,stroke:#58a6ff,stroke-width:1px,color:#f0f6fc
    style Asks fill:#0d1117,stroke:#d29922,stroke-width:1px,color:#f0f6fc
    style Gaze fill:#0d1117,stroke:#f85149,stroke-width:1px,color:#f0f6fc
    style Slip fill:#0d1117,stroke:#f85149,stroke-width:1px,color:#f0f6fc
    style Use fill:#0d1117,stroke:#d29922,stroke-width:1px,color:#f0f6fc
```

The neutral reply is that the incompatibility of choice and determination is an additional claim, one that still has to be argued. But introducing an additional claim that can be argued is one more instance of choosing, not a refutation of its irreducibility. The reply also insists that the difference between the two readings is more than verbal, a difference in "which state of affairs obtains"; this again supposes a state of affairs already finished and standing there to obtain.

```mermaid
flowchart TD
    subgraph S_Demand_EN["The Demand inside Not Yet Settled"]
        direction TB
        ReachedE["Already arrived at: the choice cannot be dropped"]
        NeutralE["Neutrality: not yet settled as determined or not"]
        AsksE["What is asked: whether a finished thing is fixed by prior conditions"]
        GazeE["What is required: a gaze taking in the prior conditions whole"]
        SlipE["The deterministic premise slips back in"]
        UseE["The use of compatibilism: telling willing from coerced, coordinating responsibility"]
        ReachedE --> NeutralE
        NeutralE --> AsksE
        AsksE --> GazeE
        GazeE --> SlipE
        NeutralE --> UseE
    end
    style S_Demand_EN fill:#161b22,stroke:#f85149,stroke-width:1.5px,color:#f85149
    style ReachedE fill:#0d1117,stroke:#3fb950,stroke-width:1px,color:#f0f6fc
    style NeutralE fill:#0d1117,stroke:#58a6ff,stroke-width:1px,color:#f0f6fc
    style AsksE fill:#0d1117,stroke:#d29922,stroke-width:1px,color:#f0f6fc
    style GazeE fill:#0d1117,stroke:#f85149,stroke-width:1px,color:#f0f6fc
    style SlipE fill:#0d1117,stroke:#f85149,stroke-width:1px,color:#f0f6fc
    style UseE fill:#0d1117,stroke:#d29922,stroke-width:1px,color:#f0f6fc
```

## 四、 非决定论是决定论断言的来源 / 4. Indeterminism Is the Source of Deterministic Claims

中立的立场把决定论与非决定论当作两个互相排斥的候选项，其中恰有一个为真。非决定论并不是决定论断言的缺席，它是这些断言的来源。每一个决定论断言都是被作出的：有人选定系统、变量与起点，断言一条规则。把两者当作互相排斥，正是决定论自身的局限：在一份决定论的叙述之内只有一个后继，这份叙述一旦被当成全体，就只能把自己的另一面安置成对手，容纳不了作出这份叙述的那个动作。确定性的模型照旧有用，预测、工程与检验都靠它；[确定性是微观不确定性的统计签名](../determinism-is-the-statistical-signature-of-micro-indeterminism/)从另一侧说明了宏观的规律怎样从大量个体的不可预测中显出。

The neutral stance treats determinism and indeterminism as two mutually exclusive candidates of which exactly one is true. Indeterminism is not the absence of deterministic claims but the source of them. Every deterministic claim is issued: someone picks out a system, variables and a starting point, and asserts a rule. Treating the two as mutually exclusive is the very limitation of determinism: inside a deterministic account there is a single continuation, so the account, once taken as the whole, can only place its other side as a rival and has no room for the act that issues the account. Deterministic models keep their use, and prediction, engineering and testing rely on them; [Determinism is the Statistical Signature of Micro-Indeterminism](../determinism-is-the-statistical-signature-of-micro-indeterminism/) shows from the other side how macroscopic regularity appears out of the unpredictability of many individuals.

“来源”这个词容易招来一幅图画：上游有一个固定的点，一条链上的第一环。这个来源不是决定论意义上的固定来源，它是作为持续动作的区分，是作出一个分布的能力：选这一样，选另一样，或者什么都不选。把它当成一个固定的来源，又是一个决定论的前提，因为这等于要求作出分布的能力自身已经被先于它的条件分布好了。正如[观察、区分活动与心智的开放地基](../observation-distinguishing-activity-and-the-open-ground-of-mind/)所追溯的，开放性在那无因的区分之内，而区分留在身后的东西按严格的因果次序从它而来。作出一个决定论断言，是在作出断言的那一步把这份开放性搁在一旁，好让单一的后继显现出来；只有当这份叙述越出那一步、被持守为全体时，被它搁在一旁的开放性才会从它内部显为一个互相排斥的候选项，或是显为一个还需要先在原因的上游点。

```mermaid
flowchart TD
    subgraph S_Source["非决定论是决定论断言的来源"]
        direction TB
        Diff["持续的区分：作出分布的能力"]
        Issue["作出一个断言：选定系统、变量与起点"]
        Det["决定论断言：从起点出发只有一个后继"]
        Tool["模型的用处：预测、工程与检验"]
        Rival["把来源安置成互相排斥的对手"]
        Fixed["把来源当成上游的固定点"]
        Diff --> Issue
        Issue --> Det
        Det --> Tool
        Det -->|"越出那一步被持守为全体"| Rival
        Rival --> Fixed
        Fixed -.->|"又一个决定论的前提"| Det
    end
    style S_Source fill:#161b22,stroke:#a371f7,stroke-width:1.5px,color:#a371f7
    style Diff fill:#0d1117,stroke:#3fb950,stroke-width:1px,color:#f0f6fc
    style Issue fill:#0d1117,stroke:#3fb950,stroke-width:1px,color:#f0f6fc
    style Det fill:#0d1117,stroke:#58a6ff,stroke-width:1px,color:#f0f6fc
    style Tool fill:#0d1117,stroke:#d29922,stroke-width:1px,color:#f0f6fc
    style Rival fill:#0d1117,stroke:#f85149,stroke-width:1px,color:#f0f6fc
    style Fixed fill:#0d1117,stroke:#f85149,stroke-width:1px,color:#f0f6fc
```

The word "source" readily invites a picture: a fixed point upstream, the first link of a chain. The source is not a fixed source in the deterministic sense; it is the differentiating as ongoing action, the ability to make a distribution: one thing, another, or nothing at all. Treating it as a fixed source is again a deterministic premise, since it requires the ability to make a distribution to have itself been distributed already by conditions that precede it. As [Observation, Distinguishing Activity, and the Open Ground of Mind](../observation-distinguishing-activity-and-the-open-ground-of-mind/) traces, openness belongs inside that uncaused differentiating, while what it leaves behind follows from it in strict causal order. Issuing a deterministic claim sets that openness aside at the step of issuing, so that a single continuation can come into view; only when the account is held past that step as the whole does the openness it set aside appear from inside as a rival candidate, or as an upstream point in need of a prior cause.

```mermaid
flowchart TD
    subgraph S_Source_EN["Indeterminism Is the Source of Deterministic Claims"]
        direction TB
        DiffE["Ongoing differentiating: the ability to make a distribution"]
        IssueE["Issuing a claim: picking out a system, variables and a starting point"]
        DetE["The deterministic claim: a single continuation from the start"]
        ToolE["The use of the model: prediction, engineering and testing"]
        RivalE["Placing the source as a mutually exclusive rival"]
        FixedE["Treating the source as a fixed point upstream"]
        DiffE --> IssueE
        IssueE --> DetE
        DetE --> ToolE
        DetE -->|"held past its step as the whole"| RivalE
        RivalE --> FixedE
        FixedE -.->|"again a deterministic premise"| DetE
    end
    style S_Source_EN fill:#161b22,stroke:#a371f7,stroke-width:1.5px,color:#a371f7
    style DiffE fill:#0d1117,stroke:#3fb950,stroke-width:1px,color:#f0f6fc
    style IssueE fill:#0d1117,stroke:#3fb950,stroke-width:1px,color:#f0f6fc
    style DetE fill:#0d1117,stroke:#58a6ff,stroke-width:1px,color:#f0f6fc
    style ToolE fill:#0d1117,stroke:#d29922,stroke-width:1px,color:#f0f6fc
    style RivalE fill:#0d1117,stroke:#f85149,stroke-width:1px,color:#f0f6fc
    style FixedE fill:#0d1117,stroke:#f85149,stroke-width:1px,color:#f0f6fc
```

## 五、 停下来的是留下来的 / 5. What Stops Is What Is Left Behind

那件“已完成、可以被固定”的事是从哪里来的？停下来的东西，是区分留在身后的东西。要查看一次抉择，就得把它当作一个结果放下来；结果之所以看得见，是因为区分已经往前去了。当观察者把留下的东西当作来源来交代，看不见停下、查看与交代都是区分的又一个实例时，它就迷失在自己的观察里。[凡可言说皆不可言说之产物](../all-that-can-be-spoken-is-the-product-of-the-unspeakable/)与[残余与自由变量](../the-residue-and-the-free-variable/)追溯的是同一处，而[你的代码库即是记忆](../your-codebase-is-memory/)也说明了，哪怕是运转中的规则与代码库，仍是既成切分的外化记忆：存在磁盘上的东西可以当作已完成的状态来检视，而修订或运行它的抉择早已走在前一步。

Where does the thing that is "finished and could be fixed" come from? What stops is what the differentiating leaves behind. To look at a choice, it has to be set down as a result; the result can be seen because the differentiating has already moved on. The observer is lost in its own observation when it accounts for what was left behind as though that were the source and does not see that the stopping, the looking and the accounting are further instances of the differentiating. [All That Can Be Spoken Is the Product of the Unspeakable](../all-that-can-be-spoken-is-the-product-of-the-unspeakable/) and [The Residue and the Free Variable](../the-residue-and-the-free-variable/) trace the same place, and [Your Codebase Is Memory](../your-codebase-is-memory/) shows why even an active repository of rules and code remains externalized memory of cuts already made: what is stored on disk can be inspected as a finished state, whereas the choosing that revises or runs it is already one step ahead.

凡是能在这里写下来的，都是留在身后的、可以观察的东西，这几行字也在其中。观察者不在这些留下的东西里面。它只能由另一个观察者从留下的东西推断出来，而它对自己的觉察不经推断；[共存的几何学](../the-geometry-of-coexistence-and-the-dual-cut/)分开过被观察到的他者与由相似推断出的他者。推断总是到得比它要找的区分晚。

Whatever can be put down here is what has been left behind and can be observed, these lines among it. The observer is not among the things left behind. It can only be inferred by another observer from what was left, while its awareness of itself passes through no inference; [The Geometry of Coexistence](../the-geometry-of-coexistence-and-the-dual-cut/) has told the other as observed apart from the other as inferred from similarity. The inference always arrives later than the differentiating it looks for.

回到开头的问题。被选中者答不出选择者的理由，而往回追的时候也没有一位选择者出现；出现的是每一个为自己抉择并承担后果的生命。“被决定还是未被决定”向一件已完成的事索要一个性质，而索要这个动作就是正在进行的区分。决定论的描述与相容论的归责，在留下来的东西里各有其用：记录、规律、人与人之间的协调。抉择不在记录上。

```mermaid
flowchart TD
    subgraph S_Residue["停下来的是留下来的"]
        direction TB
        Going["进行中的区分"]
        Stop["停下：把一次抉择当作结果放下"]
        Res["留下的东西：可以观察、可以写下"]
        Lost["观察者迷失在观察里：把留下的当作来源"]
        Infer["另一个观察者从留下的东西推断"]
        Aware["对自己的觉察不经推断"]
        Going --> Stop
        Stop --> Res
        Res --> Lost
        Res --> Infer
        Going --> Aware
        Stop -.->|"停下也是区分的一个实例"| Going
    end
    style S_Residue fill:#161b22,stroke:#3fb950,stroke-width:1.5px,color:#3fb950
    style Going fill:#0d1117,stroke:#3fb950,stroke-width:1px,color:#f0f6fc
    style Stop fill:#0d1117,stroke:#d29922,stroke-width:1px,color:#f0f6fc
    style Res fill:#0d1117,stroke:#58a6ff,stroke-width:1px,color:#f0f6fc
    style Lost fill:#0d1117,stroke:#f85149,stroke-width:1px,color:#f0f6fc
    style Infer fill:#0d1117,stroke:#58a6ff,stroke-width:1px,color:#f0f6fc
    style Aware fill:#0d1117,stroke:#3fb950,stroke-width:1px,color:#f0f6fc
```

Back to the opening question. The chosen cannot give the chooser's reasons, and no chooser turns up when the tracing is done; what turns up is each being that chooses for itself and bears the consequences. "Determined or undetermined" asks a finished thing for a property, and the asking is the differentiating under way. Deterministic description and compatibilist responsibility each have their use among the things left behind: records, regularities, coordination between people. The choosing is not on the record.

```mermaid
flowchart TD
    subgraph S_Residue_EN["What Stops Is What Is Left Behind"]
        direction TB
        GoingE["The differentiating under way"]
        StopE["Stopping: a choice set down as a result"]
        ResE["What is left behind: observable, writable"]
        LostE["The observer lost in its observation: the residue taken for the source"]
        InferE["Another observer infers from what was left"]
        AwareE["Awareness of itself passes through no inference"]
        GoingE --> StopE
        StopE --> ResE
        ResE --> LostE
        ResE --> InferE
        GoingE --> AwareE
        StopE -.->|"the stopping is one more instance of differentiating"| GoingE
    end
    style S_Residue_EN fill:#161b22,stroke:#3fb950,stroke-width:1.5px,color:#3fb950
    style GoingE fill:#0d1117,stroke:#3fb950,stroke-width:1px,color:#f0f6fc
    style StopE fill:#0d1117,stroke:#d29922,stroke-width:1px,color:#f0f6fc
    style ResE fill:#0d1117,stroke:#58a6ff,stroke-width:1px,color:#f0f6fc
    style LostE fill:#0d1117,stroke:#f85149,stroke-width:1px,color:#f0f6fc
    style InferE fill:#0d1117,stroke:#58a6ff,stroke-width:1px,color:#f0f6fc
    style AwareE fill:#0d1117,stroke:#3fb950,stroke-width:1px,color:#f0f6fc
```