Immanuel Kant · 伊曼努尔·康德
Critique of Pure Reason · 纯粹理性批判
Undated 未注日期
纯粹理性批判——读书笔记
Reflections · Pages
康德那里的两个无限(infinite/indefinite),一个是全体在经验中作为限制被数学性地在“直观”中一次给出,但这只是说明某个直观作为结果包含了所有条件,而所有条件如果要被完全展开,需要被不断回溯,被给出的因此不是全体本身,而是被包含在结果当中等待展开的全体;另一个则是单纯的一阶一阶往前追溯
感觉美国人真的读不懂康德啊……二律背反(antinomy)最重要的就在这个“律(nomos)”上啊……这不是一个逻辑矛盾,如果是逻辑矛盾那整个问题就崩了。这也不是一个科学的验证问题,因为任何科学的证实或者证伪都是依赖于直观的;只有发现一个与理论矛盾的肯定性的直观才可以证伪理论。无条件之物因此是不可证伪也不可证实的,因为既没有确证它的肯定性直观,也没有能证伪它的肯定性直观,任何能证伪无条件之物的直观自己也得是同一层次的、无条件的,而这肯定是不可能的。所以这不是contradiction,而是conflict,是严格意义上的法律问题,并进一步是如何运用法律的问题,是一个政治问题、边界问题……
其实康德B436左右讲得蛮清楚的,任何统一性都是对所有统一性官能的统一。所以才说,动力性观念可以介入数学性范畴,并将直接被给予的数学性现象也给扭成一个序列,把数学性综合放进动力性综合底下,扭成一串连续的测量运动……也呼应了之前的“质的时间变化”了……
对康德先验辩证论的几点评论:
1. 康德给出了一种幻象的分类学。经验幻象、先验幻象和超验幻象各自有着不同的意味,经验幻象来源于想象官能错误的判断力,先验幻象来源于对脱离感性的纯粹知性的肯定性运用,超验幻象则来源于理性整体必然的奠基行动。
2. 但是对于先验幻象和超验幻象的区分我们不能这么轻易放过。两者之间的关系要更暧昧复杂。事实上,先验幻象作为对经验对象的否定性限制,一旦作为肯定性的奠基被给出,就成为了超验幻象。相应的,超验幻象一旦被把握为独立于其奠基作用的肯定性的理念,就成为了先验幻象。
3. 这让我们对康德处对理性整体和知性整体的处理更加困惑了。理性整体和知性整体之间究竟是一个什么关系?简单来说,知性整体规定了“每一个”,而理性整体规定了“全体”。但仅仅这样说是不够充实的。我们需要一个结构性的差异。如果说感性是“多中之一”的平面结构,想象力是“一—多”的褶子结构,知性是“多上之一”的阶次结构,那理性就是“大全”的内涵(comprehending/B367)结构。
4. “理性是主体的、知性是客体的”这一划分本身是很空洞的。在B383,康德明确指出了理性的客观用法。至于B362,康德不过是说,理性如果只是为了整理知性已经得到的知识的话,那理性就只会是主观的;所以他接下来才会发问,理性到底有没有先验综合的客观性?理性是不能独立于知性的,因为知性是理性的材料。但问题是,知性整体能不能独立于理性整体而获得其客观性?如果可以,那理性整体的“主观性”就可以获得充分决定:理性整体的主观性是指,知性整体无需理性也可以获得客观性;而理性整体只有客观用法,本身却没有客观性,因为它只作用于概念。在B364,康德认为知性整体是不需要理性整体的,这是因为知性整体所拥有的概念都不是从理性推断出的,而是反映在经验中使经验成为可能的。可这让我们怎么理解理性的奠基作用呢?客观用法和客观性该怎么分开呢?
5. 对于B364,我们需要更认真地阅读。知性是自发地生成了使经验成为可能的概念并对经验根据概念进行了综合。知性的概念因此是一个活动性,是对感觉材料的有限综合活动。知性的视角是主体无法改变的视角。人类主体总是已经被放置在知性当中。理性则是依赖于知性的,为知性提供了源头和奠基,且指引着知性活动的方向。知性整体因此总是一个活动整体,并总是在活的当下获得明见性;理性整体则是一个在这样的活的当下被建立起来的外部,一个“无根”的奠基。理性整体彻底地依赖于无根的、绝对的理念,因此其知识是绝对不可能成为对象的,因为任何普遍的对象,只要不是绝对的,都只能是相对的、依赖于其它条件的;也是因此理性整体被认为不可以指涉可能经验。所以我们必须要放弃一种对康德所述“整体性”的简单理解,仿佛整体无非是部分的集合,或者整体本身无非是一个排外的部分。不是这样的。整体意味着「视角」,任何整体对康德来说都涵盖了所有官能的共同作用,只不过这些官能被赋予了不同的组织存在的结构。简而言之,只有在知性整体中,在知性的“多上之一”的阶次结构中,理性才被视作“奠基”而获得客观的作用;但在理性整体所设定的“大全”结构中,理念只能给予知性知识一个统一性,但本身无法进入经验,也无法被理解。所以德勒兹的洞见是很深刻的,康德那里,任何整体都并非是静态的,而是动态的、随视角变动而变动的协调。
6. 这一点是bonus:虽然康德的主旨很好理解,但其实在他的批判哲学里面,他对知性整体和理性整体的演绎并没有那么清楚。知性整体的问题是,为什么“本体”“因果”和“交互”都是对于可能经验来说必要的。“本体”很好理解;然后因为主体总是经历着事件,所以为了让事件之间获得必然性,成为真正的事件,“因果”也是必要的。可是“交互”这个法则不是恰恰违反了“因果”吗?毕竟“交互”指的是,A可以在B前面,B也可以在A前面。这不是摧毁了因果的必然性吗?但这其实是误解了康德的论证。康德要指出的是,如果时间的顺次连接预先没有必然性,那任何“事件”的时间位置都是无法决定的。同样,如果没有共时性,那任何“集合”的客观性也是无法决定的,任何有内容的时间位置也是不可能的。简单来说,任何经验的整体,如果被认识为“事件”的整体,都需要因果决定作为主导,而交互决定则作为最小单位内容(任何瞬间都是一个空间的展开);如果被认识为“事物”的整体,则需要交互决定作为主导,而因果决定作为最小单位形式(任何空间都是被压缩到最小瞬间的彼此交替的因果决定)。所以我们又一次能看到,康德任何系统的每一个部分都是对一整个整体的视角演绎。这就是系统的意义。
7. 从定然判断(categorical judgement)得到无限主体不难,只要不断从species上升到genus就行了。在这个过程中,无论是哪一项都可以不断上升不断变得更普遍,直到成为绝对主体。从假设判断(hypothetical judgement)得到无限开端就要脑筋急转弯一下了:假如A,那么B;假如C,那么A;假如D,那么C……光这样一直下去,似乎永远达不到无限开端,永远只是一个新的条件。因此,真正能够达到无限开端的判断形式应该是:假如A,那么B,有A,所以B;假如C,那么A,那么B,有C,所以A,所以B;假如D,那么C,那么A,那么B,有D,所以C,所以A,所以B……无限开端只有在这个扩张的形式下才可以变得更加普遍;如果这里列出来的最后一个判断是绝对的判断,那D就是「绝对开端」。因此,因相对于果的普遍性和genus相对于species的普遍性是不同的普遍性。而交互律的普遍性也不一样了,这个普遍性不是「上升-下降」的阶次的普遍性,也不是「展开-收缩」的序列的普遍性,而是「包含-占据」的平面的普遍性。其形式是:在B中,是A,或是非A;在C中,是B,或是非B;在D中,是C,或是非C……最后才能得到「绝对实在」;如果这里最后的一个判断是绝对判断,那D就是「绝对实在」。
8. 关于实践和理论的分野也挺有意思的,尤其是这个分野其实是一个量的关系……如果知识不能完全决定一个对象,那这个知识就是理论上没用的。但是只要能部分实践出一个理念,那这个理念就是实践上有用的。
对5.进行一个补充,知性的判断是直接的,理性的判断联系是中介的。但知性整体依旧依赖于理性的设定作为奠基,尽管知性本身的官能作用是直接的。理性是知性潜藏的地基。
对康德经验知识系统阐释的评论:
1. 如之前说过的,数学性范畴直接开显出单个表征,而动力性范畴开显出表征所依赖的基底。因此对于纯粹知性的数学层面和动力学层面来说,对象的涵义是不同的。知性在数学层面开显出的仅仅是「意识对象(objects of consciousness)」,也就是对象作为在场的表征;但只有在动力学范畴内,对象才真正被知性活动构建为「实存对象(objects of existence)」,这种对象是发生和事件,其被给予性并不是“在场”,而是“缺场”,因此勾勒出了经验本身的被给予性。在这种被给予性中,时间的模态,过去、现在和未来,本身都已经被设定了。而恰恰是这种“缺场的被给予”使在「恒在(permanence)」中被给予的「本体(substance)」及其诸模态变化成为可能。也是因此动力学的知性给出的总是遵照超越经验的大全而来的「规定性(regulativeness)」。
2. 康德在这一部分改变了之前的描述。之前,图示是作为中介连接感性和知性,但在这一部分(B224),图示承担了将感性和知性分开的功能。这一种分离是必要的,否则要么感性经验将直接决定概念辨析,或者相反;也就是,要么像休谟那样,概念盲从于经验,要么像莱布尼茨那样,经验直接地是概念的自我生成。只有通过图示,才可以让概念仅仅决定直观被给予的可能性,又让直观仅仅丰富概念的内容。假如取消图示,直接使用感觉材料进行概念思辨的话,要么,本来不相干的材料被强行塞入一个概念框架(比如巫术,巫术中,虽然A被认为必定导致B,但A和B本身在经验中可以毫无干系),因为感觉材料本身不涵盖任何判断因果的标准,任何材料彼此之间都可以被强加因果上去;要么,其概念没有严格的普遍性。只有通过图示,“什么是因果关系”才可以获得一个在经验中可以确证的必然回答,即“那在经验中必然跟随A的就是A的后果”。简单来说,图示就是相当于给了一个科学的判断标准,只要/只有在经验中辨认出了“A只能被B跟随”的事态,A和B就/才可以被认定为因果对子,也不会出现“不确定A是不是导致B”的怀疑论,或者“A应该导致C”的独断论。
3. 康德的经验构成论证其实是奠定在两个前提上的。一个是,任何对杂多的「摄入(apprehension)」都是连续的,任何表征也是连续的。第二个则是所有「摄入」都遵循一个使整体经验成为可能的原则。第二个前提的意思是,任何表征要被决定,就需要和其它表征关联,而且其关联需要有必然性。这两个前提结合起来,就可以得到,因为摄入的连续性,要么A表征跟随B表征,要么B表征跟随A表征;而这又带来三种情况,要么A表征跟随B表征是必然的,要么B表征跟随A表征是必然的,要么A表征和B表征,无论哪个跟随哪个,其伴随关系是必然的。前面两种情况是因果律,后面一种情况则是交互律。前面两种情况都强制要求连续性;而最后一种情况,事件的连续性则是可以被调转的,因此比起连续性,其更体现共时性,可以从A得到B,又从B回到A,循环往复。所以经验可能对象有三个,根据永恒概念组织出的本体(substance),根据因果概念组织出的事件(event),还有根据交互概念组织出的集合(assemblage)。
4. 因此对B275「驳斥笛卡尔唯心主义」的正确理解离不开对经验构成论与感性论的正确理解。康德实际上说的是,首先,任何主观的心理现象都是连续的表征综合。但是这一表征综合有其客观性和必然性。这些综合都依赖于「恒在」。这个「恒在」本身不能只是一个恒在的表征,因为任何表征都需要被再次综合,因此不可能存在任何恒在的表征。「恒在」因此是诸表征的基底表征,而这个表征需要直接指涉对象;换句话说,「恒在」必须是一种直观。可是「恒在」不能是对主体自己的智性直观,因为虽然时间是一种「自我触发(auto-affection)」,但这一触发体现为向外的投射,「恒在」则是通过时间这一内感知被投射出去的,并因此,在这一部分,康德颠倒了之前的判断,时间至此不是简单的「内感知」,而是「自我抛掷」,“时间是我被抛出自身之外”。在感性论部分,空间依赖于时间获得综合,但在这里,时间依赖于空间才获得超越性。时间自己无法被感知,只有指涉外部对象时间才获得客观性。就像是时间只有作为线段才能被表征一样。
自我意识设定了一个经验的自我。这个经验的自我是在时间中获得了决定的。这个决定要求一个决定的基底。这个基底需要被直观充满。但这个直观的源头,那个“智性的我”,必然是外在于我的,因为时间,作为感知,将它抛出了。我只能通过将它认识为外物来认识这个“智性的我”。这就是主体的分裂,主体将自己内在的纯粹活动建构为外在的世界这一视域;所谓的“内在性即超越性”。正是在这样的思考中,康德才能说,内在时间的决定依赖于外在时间的决定,或者说,「表征摄入」的时间必然呼应于「表象(appearance)」的时间。因此像Allison或Dummet那样,仅仅将康德这一部分解读为“如果要记住我心理状态的顺序我必须得写日记”的话,可能是相对有些肤浅了。
所以后面康德才会说任何心理显象,哪怕是疯人的幻想或者梦境,都已经设定且使用了外部了
所以其实感性论里面所谓的“时间涵盖比空间更广”指的不是时间可以关涉内在的心理状态而空间不可以,因为这些心理状态首先本身必须是被空间感知充实的,没有“浑身发热”就没有“愤怒”,没有“脑内声音”就没有“思考”,这些都是具身的、区域化的感知。真正重要的区分在于,时间同时直接关联着思维活动,因此尽管活动都需要被表征为外在,但其依然保留有意识的内在活性。所以虽然时间的客观性依赖于空间的充实,但时间同时让空间获得了主观意识的活性。缺失了空间的时间只是无内容的经过,缺失了时间的空间只是僵死的杂多。
所以说康德这里的颠倒是非常惊人的,时间作为内感知其实是比空间还要外在的,空间直接给予了直观,时间的活动却是在空间当中被中介解读出来的。这一下颠倒其实更强调了到底什么叫“内在”,其实内心是在“空间”之内被空间包裹着的那个内部,而不仅仅是“与外部世界相对的内心”。也是从这一步颠倒,康德直接进入了phenomena/noumena的范畴。
对康德纯批先验分析论的几点评论:
1. Wolff之前说他认为康德的图示(schematism)是无效论证,因为说时间和知性一样纯粹,并且时间是感性的形式,就意味着时间是感性和知性的中介,这就像是在说,因为水在杯子里,水从水龙头流出来,所以水是杯子和水龙头的中介一样,这是非常trivial的。但他这是误解了康德的意思。康德这里有两个要点是需要牢记的,一个是,时间作为中介是知性联系感性的必然条件,而不是充分条件;另一个则是康德已经预先划出了纯粹知识中所有的元素,在当中只有时间可以作为中介。因此,Wolff的批评是无效的。他想批评的无非是,一,水作为中介并不具有统一杯子和水龙头的效力;二,水作为中介可以统一两者,但水杯和水龙头之间的空气也可以统一两者。康德对第一点的回应是,哪怕时间并没有直接沟通感性和知性,时间至少给知性的主动性(也就是说,知性,而非时间,才是这一统一活动的源头)提供了必要的桥梁。对第二点,康德的意思是,在两者间可以充当中介的只有时间,没有其它选项。
2. 所以康德在这里真正的设计是,已知感性和知性的统一,求使两者统一成为可能的中介。这里Wolff的批评大致合情理,说康德提出的「容纳模型(subsumption model)」只是描述了中介过程,但没有明确解答为什么需要中介。如果知性就是活动性,难道不能让感性简单地屈服于知性吗?
3. 但这一批评虽然表面上有道理,其实已经走偏了,没有意识到康德更深层的问题域。为什么感性和知性之间需要一个中介?因为感性对象作为“多中之一”,其特殊性是绝对特殊,任何对绝对特殊的判断都是没有办法获得任何普遍性或者必然性的。而知性的概念作为“多上之一”,要被运用到感性对象上,其普遍性更是无法在感性对象的特殊性内获得觉察的,除非将感性单纯化为下级阶次,知性化为上级阶次,然后讨论「分有(participation)」——简而言之,放弃先验哲学,转而回归传统形而上学。可是先验哲学的信条就是,普遍性没有办法在直观中被给出,直观不是一个“更特殊”的普遍:这是一早就被决定了的。因此,要同时让概念和直观之间的一-多对子和概念和直观各自内部的一-多阶次共同完成一个体系,康德必须证明,一个普遍的概念可以对应于一个特殊的直观。这一点,他通过「想象力」的作用做到了。直观当然不是一个相比其它普遍来说“更特殊”的普遍,但直观作为特殊是处于一个普遍之特殊(纯粹直观)中的;概念当然也不是一个相比其它特殊来说“更普遍”的特殊,而是作为普遍的活动得以刻画一个特殊的普遍(纯粹想象)。因此,康德得以保留先验哲学的原初性:概念的普遍活动在想象力中被刻画为一个被给予的纯粹直观图像,而直观的特殊呈现图像则被包含在这一图示图像中。
4. 因此,想象力呈现出和知性和感性都完全不同的结构。如果说知性是多上之一的阶次结构,感性是多中之一的平面结构,那想象力就是一-多互渗的透视结构。一个图示图像“笼罩”了特殊图像呈现自身的视野。尽管这一笼罩是自上而下的,但这个笼罩本身也投射出一个给予性的超越视野。这也是海哥康德论最大的洞见所在,就是对想象力结构的澄清。只有通过想象力的中介,知识才成为可能;因为通过图示,知性不需要再“一个一个一个”地去判断特殊,而可以给出对“每一个特殊”的普遍判断,因为其判断直接地决定了整个视野。这也是为什么先验判断力比经验判断力更为优越,因为经验判断力只能达成对特殊的容纳,而先验判断力则可以达成对整个特殊之视野的容纳。这个视野就是可能经验,而在这个视野中被给予的特殊总体上来说就是可能经验对象。“对象在使经验成为可能的诸条件对直观杂多的限制中获得被判断的基础并因此被决定为可能的”。
5. 在这样的视野中,有些对象开显出自身,这一开显是直接的、无根的,这种就是“数学的”,比如感知,是没有理由的。有些对象开显出自身,却因此同样指示出使这一开显成为可能的决定性的黑暗土壤,这中就是“动力的”,开显的同时开显出自己所依赖的必然性。
其实时间性这个,在康德那里,更像是为了奠基知识来的。对任何对象的知识要成为可能,首先对知识的知识要成为可能,任何的判断(综合)都必须是能被决定(直观)的。任何综合因此必须是一个被给出的杂多,除非能够直接地在思内部去直观综合的活动。否则这个综合的活动自己必须是作为一个杂多的表征被「给予」出来的。这个被给予的对综合本身的直观其形式就是时间,无时的活动因此被表征为时间内的「运动」。而这个运动也需要一个超越性,因此在这上面时间必须依赖于空间,只能通过「画线」被表征为外部。时间因此依赖于空间被外化为表征,又依赖于知性去表征综合本身。这个构思听起来好像很危险,仿佛必然会导致无限递归,比如“既然知识依赖于对知识的知识,那对知识的知识是否依赖于对知识的知识的知识?”,但这问题就完全错误理解了康德的用意。在康德这里,对知识的讨论可以被划分出两个方面,一个是“阶次”,一个是“结构”,这两个方面彼此之间毫无影响。无论你想要第几级的知识,你的知识终究是得来自于直观和概念的综合;无论你选定哪个阶次,只要你有这个概念和相应的直观表征,那就是知识。一个杂多被综合要求「杂多」本身与「综合」本身的综合,这一纵向的综合(veritative synthesis)就是知识结构的问题。但一个杂多被综合后,其「综合过程」要成为杂多与其它「综合过程」综合,这一平面的综合(predicative or apophantic synthesis)是知识阶次的问题。在康德讨论“对知识的知识”的时候,他的整个问题域不在「阶次」,而在「结构」。范畴是知识的形式,是最高的统一性,而不是自己还需要被再次综合的表征。换句话说,尽管范畴是综合,但本身已经是最高的综合,而对最高综合的表征进行再综合本身也是属于最高的综合活动的。因此,无关于任何阶次并超出任何确定的阶次,只要有综合过程被直观被表征,对结构的知识都已经达成了;这也就解释了「时间」的统一性,因为「时间」是知识结构中的对子,是被动性,而「统觉」则是知识结构中的主动性。因此,正确的说法并不是“知识依赖于对知识的知识,对知识的知识依赖于……”,而是“知识就是知识的对象”。同样的,并不是“综合本身要被综合”,而是“综合的表征总是被综合着”。拥有阶次性且无限递归的并不是综合的活动,仿佛在综合之上还有综合,而是对综合的表征;表征总是不断被和其它表征综合,当中包括对综合的表征。因此,“对知识的知识”并不是指“在一个综合之上还有另一个综合”,因为「知识」与「综合」并不是同义词,而是指“综合这一活动(知识的概念部分)与综合作为表征(知识的直观部分)的结合”,所以“对知识的知识”其实就是“对概念的直观”或“对直观的概念”。
简单来说,“知识依赖于对知识的知识”其实应该被理解成,“对直观的理解依赖于对理解的直观”,所以这不是一个上升的阶次的关系,而是一个互相依赖的关系。
因为被给予了直观,所以“知道”直观要求理解;因为会理解,所以“知道”理解需要被给予一个直观。
所以其实最后康德的用词很有意思,understanding determines appearance,concepts are determined by intuitions
海哥所谓的形而上学基本问题其实就是沿着这样的思路走的:首先,基本问题就是“一-多”(存在作为存在-存在作为存在者全体)问题;然后,这一问题被重述为有限性问题;接着,这个有限性问题成为了有限存在的问题,也就是此-在的问题;再然后,此-在的有限性该怎么理解?,海哥很鸡贼地说,此-在的有限性就在于“问”,因为此-在不完全理解有限性或者自身的存在(「理解」永远只是「前抛」),所以有限性就恰恰在这一“问”中被开显出来;那到这一步,哲学的基本问题就被转换成了哲学自身的可能性问题,这也是海哥在康德那里看到的最重要最本质性的转折,也就是形而上学哥白尼革命的真正含义。「存在与时间」里面,这些步骤就构成了基础本体论的大部分内容,一直从「存在问题」到「被抛」,再到「烦」,等等等等。但最后一步则是,这一种「烦」或者有限性其实就是来源于此在的时间性与人的向死而生。此在的时间性因此使哲学成为可能,也因此与形而上学的基本问题属于一个整体;正因为此在在对时间的理解中才能向存在「发问」,任何对存在的理解都依赖于对时间的理解。
海哥跳大神解读康德时间论
1. 想象力综合概念和直观且让它们在自身中获得本质,这也就意味着在想象力中时间才成为时间。正是因为通过想象力的综合时间才显现出彻底的时间的样态(同一性、连续性等等),那么时间更根本的本质(先验的结构)同样要通过想象力的综合解决,也就是过去、现在和未来。
2. 过去、现在和未来是先验的时间形式,其必须被奠基在纯粹想象力中。「现在」来自于“直观综合”,即材料在表征空间中被把握为“多中之一”;「过去」来自于“想象力综合”,这里的“想象力”是纯粹想象力的本质,也就是“一-多综合”,也就是诸表征被放置在一个更高的统一表征之下;「未来」来自于“概念综合”,也就是“多上之一”,就是说那预先已经被投射到表征空间的整体。
3. 海哥这里很明显他感兴趣的图示(schematism)并不是概念或者理性给感性的立法,因为他没有直接提到任何范畴的综合,也没有对范畴与这些综合与时间根本结构本身的关系进行讨论。因为对海哥来说,理性的这些范畴根本上是从属于纯粹想象力的,是纯粹想象力的自我认识和自我奠基。因此,当纯粹想象力同样通过“部分-整体”这个角度为自己奠基后,对这具体带来了什么范畴的讨论就不重要了,因为从这个角度自我奠基的纯粹想象力已经将诸表征时间化了。既然正是通过纯粹想象力的这些个「综合」知识或者纯粹想象力本身才是可能的,那被综合下来的直观本身自然也要带有这些不同「综合」的痕迹,也就是「现在」「过去」和「未来」。换句话说,「现在」「过去」和「未来」不仅仅是时间的形式,同时也必然是纯粹想象力的综合所依赖的纯粹概念,尽管它们作为纯粹概念应该叫什么名字有待商榷。总之,时间虽然是经验中的连续的现在,但根本却是奠基在纯粹想象力上的。而纯粹想象力所进行的那些「综合」实际上就是原初时间,那使经验时间成为可能的时间。
4. 但海哥还不满足于此。仅仅在纯粹想象力里面发现原初时间只能让人给出对原初时间的一种“上手”的解读。原初时间并不是躺在纯粹想象力里面的一个东西;相反,它是那使得纯粹想象力本身成为可能的纯粹想象力的本质。纯粹想象力之所以成为可能,就是因为「一-多」的对子关系,而这个对子关系,如果人不是有限的,不可能存在。纯粹想象力奠基在人的有限性上。但人的有限性奠基在时间性上,因为有限性根本上来说是直观的有限性,也就是时间的超越性。时间总是对立于人构建出人的视角。正是因为直观(时间)的这一超越结构,或者说表征的结构,人才完全实现其有限性。所以纯粹想象力在进行这些「综合」的时候,已经预设了时间的超越性了,因为任何想象力的「综合」都是针对「表征」的综合,过去只是被统一在一个表征下的其它表征,现在只是某一个表征,未来只是统一了所有表征的表征;真正重要的不是表征如何被综合起来,而是在进行「想象力综合」并构造出「过去」的同时,「过去」本身已经是对立于想象力的构造活动的超越性视野了。因此更本质地开显出人之有限性并让纯粹想象力综合表征的活动成为可能的便是原初的时间性,或者说「过去」「现在」「未来」相对于人的超越性。因为人的存在就是有限的此在,原初的时间性也就是人的存在,或者说人的此在。
海哥魔改康德论想象力:
1. 想象力是知识的本源,同时为感性和知性奠基。这并不是因为想象力处在中间的位置可以连接上下级官能,因为在本源的场所中,上下级的区别并不重要,土壤可以同时滋养地上与地下的生物。重要的是知性和感性都要在想象力的场所中开显出其本质。
2. 感性相对好理解,因为想象力就是对表征的纯粹综合,感性的被动性在想象力中被主动综合为一个图示,让诸表征预先被综合起来,并让时空作为直观形式获得确定的先验内容,也因此让时空成为可能。因为时空并非是空形式,而是非-对象,有着自身的结构(连续性、同一性等等),而正是因为这些结构直观才成为可能,而不是直观先凭自己给出内容,再由想象力综合;先有想象力综合过,直观形式才是可能的。所以海哥说先验感性论在schematism之前是无法理解的。
3. 知性的理解相对困难一些。想象力作为原初综合,它不仅仅将表征综合起来,而且这些表征的综合本身作为“整体”自己也要被综合。这就是自我意识或者统觉,“我思我思”。在将诸表征联系起来时,同时也将联系的法则与诸表征一样联系起来。这给予了知性“活动-被动”的二重性,知性一方面是法则,一方面又是被表征的法则,且只有在被表征的情况下其法则才能有效。否则就会产生“A和B作为表征被综合在一起,但这个「一起」却并不是「整体」,无法被表征也无法被思考”的悖论。此外,任何“我思”其实都预先被综合过了,我之所以能获得“A导致B”的知识是因为「因果」这个范畴已经被我预先把握住了,我“知道”这个范畴是什么以及这个范畴是普遍的,这种把握必然是对我自身活动的表征(因为我是有限的,无法理智直观也无法自我创造),如果缺少这个表征,我将根本无法作出判断获得知识。所以统觉建立在想象力上,自我认知的知性也就建立在想象力上,而且知性和感性一样,是自我给予的活动,也因此有着自己的直观。
4. 不仅思辨理性建立在想象力上,实践理性也建立在想象力上。因为思辨理性是想象力的自我给予,这过程中包含了想象力的自我设限,也就是自由。在实践理性的领域内,自我在“当下”的自我意识中被开显为「道德人」,而既然任何开显都是对存在本身的决定,自我的存在(自我意识之奠基),也就是先验想象力,也需要被如此决定,以至于能够从想象力获得给道德奠基的可能。这里需要的就是对先验感觉的演绎,也就是「实践理性批判」的内容,如何从想象力演绎出对道德律令的尊敬。
感觉康德需要想象力的中介是因为他认为感性的多中之一(多的原则)和知性的多上之一(一的原则)不能直接联系起来,因为感性的杂多必须是结构上无限的(“多”必须是群,哪怕是“一”群“一”),概念必须是结构上有限的(“一”必须是法,必须将多个表征完全统一在一个表征下面),而像莱布尼茨那样一步步从有限加到无限是没办法解释感性生成的,因为两者是质或者结构上的区别,而不是单纯数量上的区别。“一”和“多”在这里是“质”,而不是“量”。因此需要一个中介,也就是想象力,将“一”「表征」成“多”,然后将“多”「表征」成“一”,这样尽管一和多依然有质的差别,但至少在现象的领域能够达成和谐。
就是说虽然想象力中有直观也有概念,但因为想象力是纯粹综合,它把所有概念都单纯把握为同一个视野中的表征,把它们都投射到同一个平面上并进行联系,这样,它把一表征为多,让概念获得直观,并得以和其它直观综合;同样,它的综合也建立起统一的形式,并将这形式和内容区别开。简单来说,知识是因为想象力才成为可能,因为本来概念作为多上之一,哪怕底下有无限数目的其它概念,它还是究其本身而言只是一个一;只有当这个一通过想象力在直观中被表征为多中之一的时候,它和其它表征的联系才能被建立,因此任何综合都依赖于直观。换成直观也是一样,作为多中之一,无论数量再怎么被缩小也还是多,除非被想象力综合成一个整体,否则无论如何都不会产生知识。
5. 想象力并不是瞎想的能力,因为首先想象力不仅想象不在场的东西,也使对在场的东西的认识成为可能,甚至在场本身的可能性也是想象力奠基的;其次想象力并不只想象“东西”,而是想象那使任何“东西”成为可能的视域。对想象力的人类学理解因此忽视了想象力在本体论知识上至关重要的中介地位。想象力并不是与有限的现实相对的无边际空想的能力;恰恰相反,想象力是那让无限的杂多得以被限制且出现在人类认知当中的能力。当然相比于感知经验,想象力更强调可能性而非实在性,并因此看起来似乎更“不受限”一点,但其实恰恰相反,正因为想象力强调可能性,它结构上不能是纯粹的无限的多(这里的多是结构与质的多),而只能是相对更有限的,“一个三角形”“几个x”,而不会是感知给予的纯粹杂多。这个相对有限的可能概念需要感觉的杂多去充满才能拥有现实性。这一点上面,也就是可能性与有限性相关,现实性与无限性相关,康德和莱布尼茨是相似的。
康德的模型大体上就是,感性接收被表征的多中之一(one among many),想象力将这些表征出的多中之一连接成一个表征整体,而这个整体性则来自于作为多上之一的纯粹知性(one above many);感性提供未被确定的、和其它诸多对象作为表象并置的对象(内容/objective reality),而知性则将这一对象给「确定」下来(形式/objective validity),纯粹知性则给了一个达成这种「确定」的视域和奠基,也就是,只要想象力开始对表征进行确定或综合并生成诸概念的阶级让诸表征被统一在某个表征之下的话,那这些表征在之前就必须已经被预先综合进纯粹知性底下成为「那可被思者」了,并且只有纯粹知性达到自我知识,认识到纯粹综合的必然规则之后,这样的总体上的知识才是可能的。这些都是很清晰的。除此之外,先验逻辑像还有比如知性统一性的问题,其实康德无非就是(和亚里士多德对实在进行抽象不同)将目光转向先验的知性本身并获得一个直接的自我知识,而在对「同一、真理、完备」的讨论里也示意了在自我认知时知性是如何将量与质与关系进行综合的(同一是只有“一个”的前提;真理是从这个前提能推出“多少个”不同质的结果;完备是推出结果的整体)。另一个问题我觉得康德解决得不够好,就是交互性和析取判断(要么A要么B)之间是怎么联系的。
而如果直观感性带来的表征要能够被思,本身就已经得有被这样思的可能了,也就是不能只是单纯的质料,而哪怕是质料也得是被预先「形成」过了的质料;也就是时间作为图示(schematism)。
因为直观是对概念的充实,因此概念获得确定的同时,其直观也要有如此的“确定”,那就是直观的形式要呼应其概念,这就是schematism的来源。
康德反对亚里士多德是因为亚里士多德把范畴放入实在去了;康德反对传统形式逻辑(非先验逻辑)则是因为只有在知性的行动(即判断)中我们才能获得对对象的先验规定,形式逻辑只在乎思考的秩序,而不在乎思考的秩序如何让知识成为可能并为自己提供对象的客观性(即知性在「判断」的行动前预先规定了什么是可能的经验对象)。
因为缺失先验的这一角度,传统逻辑没有办法解释它所包含的那些规定是哪里来的,是来自事物本身还是来自心灵;只有在澄清了逻辑自身的来源是知性的统一性而对象只有在逻辑当中才获得确定的可能和奠基的时候,逻辑才真正充实起来。
无限判断就是当A不是B且是非B的时候;如果仅仅说A不是B,A只是从B中排除了,但A并没有被确定是不是存在或者存在在哪儿;通过“A是非B、A是非C……”无限的叠加才能够确定A到底是什么
Longuenesse对析取判断的解释是:传统逻辑只是对已经预先被综合过了的诸概念进行分析和澄清;作为判断的官能,知性需要在判断的同时将概念预先综合成可判断的,所以事物先被综合成了交互的/互相决定的,只有在有了这样一个“多个独立部分共同组成的整体”的情况下,析取判断才是可能的。因此康德不是认为对概念的判断和被判断的事实总是完全重合,而是形式逻辑的判断和被判断的内容有相同的源头;先在内容上获得了综合,判断才是可能的。因此问题不是单纯的“什么对象对应这样的判断”而是“什么(被综合后的)对象允许了这样的判断”。这其实也挺好理解的,因为范畴的作用就是让表象成为可以被思被判断的对象,这是一个奠基关系,而不是一个简单的一一对应或类比的关系。
并且为每个类别的判断奠基的同类别综合只被运用于特定种类的表征上,如量用在直观上,质用在感知上,等等。
但是这些特定种类的表征被综合在一个总的“对象”的概念之下。或者说这些不同种类的表征无非是同一个表征的不同面向,和其它表征放置在一起时应该被如何决定。假设一个表征A只能被思考量,另一个表征B只能被思考质,那A和B之间就没有关系了。A和B没有关系就意味着A不能在B中被思考,或者说当A和B并置时,思考变得不可能。但这是荒谬的,因为任何表征都是被统一在“我思考”这一统觉下的;且对象就是那“可被思的”。A和B没有办法一起思考就意味着A无法被B确定,在B之下A不再是对象,也就不可能出现在经验中。
康德感性论的可能性
1. 时间与空间的纠缠,时间作为「内感知」,几何的绘图过程,运动作为纯粹科学同时关涉于时空(或许正因如此康德后面又不认同运动纯粹科学的地位);以及「内」「外」感知翻转的可能性,比如「在此」的心灵的时间变更。
2. 概念可以直接关涉直观也可以间接关涉直观。空间与时间的概念都直接关涉着直观,也因此仅仅通过辨析其概念就可以达到其直观的给予性。时空的概念在最本质的决定中并不是对不同的种属的时空的综合,而是对时空直观的直接对应。
3. 直观指涉并决定对象,但任一感知仅仅部分指涉对象。任一感知都无法直接指涉整个对象(任一感知都是不完整且相对的),因为「整个」对象需要被「整个」感知给予,还需要「整个」时空形式对其决定。直观通过整个感知获得了对象的材料,并通过这些材料充实了直观,在这种情况下感知才可以说带来了知识。也就是说,只有通过诸材料彼此间对对象的整体决定,质料侧的知识才是可能的。而形式侧的时空感知则因为本身就是决定了对象可能性的纯粹形式,其知识本来就囊括了对象整体了。质料侧的感性知识就是所谓经验直观,而形式侧的感性知识就是所谓纯粹直观。
4. 纯粹直观总是关涉于一个「图像」,或者说,一个「半对象」。通过想象力给予出的这样一个对象,如三角形或运动的球体,其意义仅仅在于,这样一个对象处在纯粹直观的时空内,仅仅支撑起了对任何一个先验的可能对象的象征。换句话说,想象力给予了一个直观的内容,但真正被捕捉到的是对作为纯粹形式的时空本身的知识。因为时空是那使一切对象成为可能的,通过对想象力的掌握(那进行变更的力量),我们从质料转向关注形式,并从此确证纯粹直观的先验性。
5. 与莱布尼茨相反,康德拒绝概念与感知仅仅只有程度上的划分。莱布尼茨认为,本体无非就是无穷的概念,感知则是对这一概念的抽象,但是康德会认为每个概念的表征都是无穷的,但不是概念每个层次的表征抽象都是感知,比如对「权利」这个概念的抽象并不会带来感知;同样,对感知的完全澄清也不会带来概念,而只会是更多的感知。毕竟,感知是一个没有深度的包含结构,而概念是一个有深度的等级结构。当然,这个批评对莱布尼茨是无伤大雅的。因为莱布尼茨那里,无限的概念和最高概念并不一样,而是分立首尾两端。向最高概念那端进行的回溯性抽象就是概念的抽象,而向着靠近无限概念那端进行的前抛性抽象则是感知的抽象。
6. 思是有限的。但思给予了一个将无限放入有限性的方式,也因此人可以在思中有限地放置物自体,让它成为一个有限的、不可知的无限性,也就成为了限制本身。有限的无限被转化为无限的有限。
第五点非常重要,这直接揭示了康德那里的直观和概念是按照完全不同的两个原则、官能、结构组织的;直观是多中之一(one among many),概念则是多上之一(one above many)。直观规定了概念的确定范围,概念拓展了直观之间的联系。想象力在与概念相关涉时生成形式直观,而在与直观相关涉时生成质料直观。
这样也可以解释为什么康德最终选择使用“线”来图示时间,因为时间的图像最终被把握为那个「半对象」,表征了知性统一的活动。
第四点,经验直观-纯粹直观-想象力-概念,正因为想象力是普遍表征的官能,它才能使纯粹直观与经验脱钩;同样,因为想象力根据纯粹概念(图示)进行表征,它让依赖它成为表征形式的时间成为被图示化的视域。
海哥论康德:
1. 此在是一个本体论结构,此结构规定了存在者被解读为“那此在所朝向的”。在康德处,这样的本体论结构被决定为先验主体对对象的先验决定。
2. 在有限的此在的这个本体论结构中,我们会遭遇一个问题,那就是,对这一结构的本体论知识是如何可能的。这一问题可以被简化成三个步骤或三个面向:首先,我们知道此在朝向的存在者总是超越的,或者物自体是不可知的;但我们依然确认“存在本身就要求了物自体的设定”,也就是,主体-对象这一结构适用于一切对象;这就意味着,我们对我们不应有知识之存在有知识,我们清晰地认识到此在的结构是本体论意义上的结构,而不是只适用于部分存在,那这说明“物自体”必然有着“接受物自体这一决定”的特质;也就是说,“物自体”在其超越性中包含了内在性,而这一结构,包括物自体在内,的多个组成部分要求一个更高的统一性奠基。也就是说,任何被视作这一结构外部的同样构成了这一结构,现在需要的就是解释这样的外部是如何在结构内部被生成的。
目前看下来我觉得海哥的大体方向是正确的,但在细节演绎上很乱暴。比如他将整个结构的统一性的问题简化成感性和理性的统一性的问题,又比如知性的生成被简化成它那“允许直接的表征被再次表征”的功能。这可能与他只关注纯批而不太关注三大批判的整体有关。
AI translation from the Chinese original. The original is the authoritative text.
Kant’s two infinities, infinite and indefinite: in one, the totality is mathematically given all at once in intuition as a limit within experience. But this only means that a certain intuition, as a result, contains all its conditions. If all those conditions are to be fully unfolded, they must be continually traced backward. What is given, therefore, is not the totality itself, but the totality contained in the result and awaiting unfolding. The other is simply a tracing forward, step by step.
I really feel that Americans cannot read Kant… The most important thing in “antinomy” is precisely the nomos, the law… This is not a logical contradiction; if it were, the entire problem would collapse. Nor is it a question of scientific verification, since any scientific confirmation or falsification depends on intuition. Only the discovery of a positive intuition contradicting a theory can falsify it. The unconditioned can therefore be neither falsified nor confirmed: there is neither a positive intuition confirming it nor a positive intuition falsifying it. Any intuition capable of falsifying the unconditioned would itself have to be unconditioned and on the same level, which is certainly impossible. So this is not contradiction but conflict, a legal question in the strict sense, and further a question of how to apply the law: a political question, a question of boundaries…
Kant actually explains this quite clearly around B436: every unity is a unification of all the faculties of unification. That is why dynamical ideas can intervene in mathematical categories, twisting even immediately given mathematical appearances into a series, placing mathematical synthesis under dynamical synthesis and turning it into a sequence of continuous movements of measurement… This also echoes the earlier “qualitative change of time”…
Some comments on Kant’s Transcendental Dialectic:
1. Kant provides a taxonomy of illusion. Empirical, transcendental, and transcendent illusion have different meanings. Empirical illusion arises from the imaginative faculty’s erroneous power of judgment; transcendental illusion arises from the positive employment of pure understanding detached from sensibility; transcendent illusion arises from the necessary founding activity of reason as a whole.
2. But we cannot let the distinction between transcendental and transcendent illusion pass so easily. Their relation is more ambiguous and complex. In fact, transcendental illusion, as a negative limitation of empirical objects, becomes transcendent illusion once it is given as a positive foundation. Correspondingly, transcendent illusion becomes transcendental illusion once grasped as a positive idea independent of its founding function.
3. This makes Kant’s treatment of the whole of reason and the whole of understanding even more puzzling. What exactly is their relation? Put simply, the whole of understanding determines “each,” while the whole of reason determines “all.” But this formulation is not sufficiently full. We need a structural difference. If sensibility has the planar structure of “one among many,” imagination the folded structure of “one–many,” and understanding the hierarchical structure of “one above many,” reason has the comprehending structure of the totality, comprehending/B367.
4. The division “reason is subjective, understanding objective” is itself empty. At B383, Kant explicitly identifies an objective use of reason. As for B362, Kant merely says that if reason only organizes knowledge already obtained by the understanding, it will be merely subjective. That is why he next asks whether reason has an objectivity of transcendental synthesis. Reason cannot be independent of understanding, because understanding is its material. But can the whole of understanding acquire objectivity independently of the whole of reason? If it can, the subjectivity of the whole of reason can be fully determined: it means that the whole of understanding can acquire objectivity without reason, while the whole of reason has only an objective use and no objectivity in itself, because it acts only on concepts. At B364, Kant thinks the whole of understanding does not require the whole of reason, because its concepts are not inferred from reason but reflected in experience as what makes experience possible. But how, then, are we to understand reason’s founding role? How are objective use and objectivity to be separated?
5. We need to read B364 more carefully. Understanding spontaneously generates the concepts that make experience possible and synthesizes experience according to them. Its concepts are therefore an activity, a finite synthetic activity directed toward sensory material. The perspective of understanding is one the subject cannot alter. The human subject is always already situated within understanding. Reason, by contrast, depends on understanding while providing its origin and foundation and guiding its activity. The whole of understanding is thus always a whole of activity, acquiring evidence in the living present; the whole of reason is an outside established in that living present, a groundless foundation. The whole of reason depends entirely on groundless, absolute ideas, so its knowledge is absolutely incapable of becoming an object: any universal object, unless absolute, can only be relative and dependent on other conditions. This is also why the whole of reason is considered incapable of referring to possible experience. We must therefore abandon a simple understanding of Kantian wholeness, as though a whole were merely a collection of parts or itself merely an exclusive part. It is not. Wholeness means perspective. For Kant, every whole encompasses the joint operation of all the faculties, except that these faculties are assigned different structures for organizing being. Put simply, only within the whole of understanding, within its hierarchical one-above-many structure, is reason considered a foundation and acquires an objective role. Within the totality-structure posited by the whole of reason, however, ideas can only give knowledge of the understanding a unity; they cannot themselves enter experience or be understood. Deleuze’s insight is therefore profound: in Kant, no whole is static; each is a dynamic coordination changing with perspective.
6. This point is a bonus. Although Kant’s main intention is easy to understand, his critical philosophy does not deduce the whole of understanding and the whole of reason all that clearly. The problem concerning the whole of understanding is why substance, causality, and reciprocity are all necessary for possible experience. Substance is easy to understand. Since the subject always experiences events, causality is necessary to give relations among events necessity and make them genuine events. But does the law of reciprocity not violate causality? After all, reciprocity means that A can precede B and B can precede A. Does this not destroy causal necessity? Yet this misunderstands Kant’s argument. His point is that, without a prior necessity in the successive connection of time, no event’s temporal position can be determined. Likewise, without simultaneity, no assemblage’s objectivity can be determined, and no temporal position with content is possible. Put simply, any whole of experience, if known as a whole of events, requires causal determination to govern, with reciprocal determination supplying the content of its minimal unit: every instant is an unfolding of space. If known as a whole of things, it requires reciprocal determination to govern, with causal determination supplying the form of its minimal unit: every space is a set of alternating causal determinations compressed into the smallest instant. Again we see that each part of any Kantian system is a perspectival deduction of an entire whole. This is what system means.
7. Deriving an infinite subject from categorical judgment is not difficult: one simply keeps ascending from species to genus. In this process, any term can continually ascend and become more universal until it becomes the absolute subject. Deriving an infinite beginning from hypothetical judgment requires a twist: if A, then B; if C, then A; if D, then C… Continuing in this way seems never to reach an infinite beginning, only another condition. The form of judgment that can genuinely reach an infinite beginning should therefore be: if A, then B; A holds, therefore B. If C, then A, then B; C holds, therefore A, therefore B. If D, then C, then A, then B; D holds, therefore C, therefore A, therefore B… Only in this expanding form can the infinite beginning become more universal. If the last judgment listed here is absolute, D is the absolute beginning. Thus the universality of cause relative to effect differs from that of genus relative to species. The universality of reciprocity is different again: not the hierarchical universality of ascent and descent, nor the serial universality of expansion and contraction, but the planar universality of inclusion and occupation. Its form is: in B, either A or not-A; in C, either B or not-B; in D, either C or not-C… Only finally do we obtain absolute reality. If the last judgment here is absolute, D is absolute reality.
8. The division between practice and theory is also interesting, especially because it is actually a quantitative relation… If knowledge cannot determine an object completely, it is theoretically useless. But as long as an idea can be partially put into practice, it is practically useful.
A supplement to 5: judgments of understanding are immediate, while reason’s connections among judgments are mediated. Yet the whole of understanding still depends on reason’s positing as its foundation, even though the faculty of understanding itself operates immediately. Reason is understanding’s hidden ground.
Comments on Kant’s systematic exposition of empirical knowledge:
1. As noted before, mathematical categories directly disclose individual representations, while dynamical categories disclose the substrate on which representations depend. Objects therefore have different meanings on the mathematical and dynamical levels of pure understanding. On the mathematical level, understanding discloses only objects of consciousness, objects as present representations. Only within dynamical categories does the activity of understanding genuinely construct objects as objects of existence. These objects are occurrences and events; their givenness is not presence but absence, thereby outlining the givenness of experience itself. Within this givenness, the temporal modalities of past, present, and future have already been posited. It is precisely this absent givenness that makes possible substance, given in permanence, and its modal changes. This is also why dynamical understanding always gives a regulativeness following from a totality that transcends experience.
2. Kant changes his earlier description in this section. Previously, the schema mediated between sensibility and understanding; here, at B224, it assumes the function of separating them. This separation is necessary. Otherwise, sensible experience would directly determine conceptual discrimination, or the reverse: concepts would blindly follow experience, as in Hume, or experience would directly be the self-generation of concepts, as in Leibniz. Only through the schema can concepts determine merely the possibility of intuition’s givenness, while intuition merely enriches conceptual content. If the schema were canceled and sensory material directly used for conceptual speculation, either unrelated materials would be forcibly stuffed into a conceptual framework—as in magic, where A is taken necessarily to cause B although A and B may have no experiential connection whatsoever—because sensory material itself contains no criterion for judging causality and any materials can have causality imposed upon them; or the concept would lack strict universality. Only through the schema can “what is a causal relation?” receive a necessary answer confirmable in experience: “What necessarily follows A in experience is A’s consequence.” Put simply, the schema provides something like a scientific criterion of judgment. If and only if the state of affairs “A can only be followed by B” is identified in experience can A and B be recognized as a causal pair. There will then be neither the skepticism of “it is uncertain whether A causes B” nor the dogmatism of “A ought to cause C.”
3. Kant’s argument for the constitution of experience actually rests on two premises. First, every apprehension of a manifold is continuous, as is every representation. Second, all apprehension follows a principle making experience as a whole possible. The second premise means that any representation, to be determined, must be related to other representations, and that relation must be necessary. Together, these premises yield the following: because apprehension is continuous, either representation A follows representation B or B follows A. This gives three possibilities: A necessarily follows B; B necessarily follows A; or, whichever follows the other, their accompaniment is necessary. The first two are the law of causality; the last is the law of reciprocity. The first two strictly require continuity. In the last, the continuity of events is reversible, so it displays simultaneity more than continuity: from A we can obtain B, and from B return to A, cycling repeatedly. Thus there are three possible objects of experience: substance organized according to the concept of eternity, event organized according to causality, and assemblage organized according to reciprocity.
4. A correct understanding of B275, the refutation of Cartesian idealism, therefore requires a correct understanding of the constitution of experience and the Aesthetic. Kant is actually saying, first, that any subjective psychological phenomenon is a continuous synthesis of representations. Yet this synthesis has objectivity and necessity. All these syntheses depend on permanence. Permanence itself cannot merely be a permanent representation, because any representation requires further synthesis, so no permanent representation can exist. Permanence is therefore the substrate-representation of representations, and this representation must refer directly to an object. In other words, permanence must be an intuition. But it cannot be an intellectual intuition of the subject itself: although time is an auto-affection, this affection appears as an outward projection. Permanence is projected outward through time as inner sense. Thus in this section Kant reverses his earlier judgment: time is no longer simply inner sense but self-projection—“time is my being thrown outside myself.” In the Aesthetic, space depends on time for synthesis; here, time depends on space for transcendence. Time itself cannot be perceived and acquires objectivity only by referring to external objects, just as it can be represented only as a line segment.
Self-consciousness posits an empirical self. This empirical self is determined in time. Its determination requires a determining substrate, which must be filled by intuition. Yet the source of this intuition, the “intellectual I,” must be external to me, because time, as perception, has thrown it out. I can know this intellectual I only by knowing it as an external thing. This is the subject’s splitting: the subject constructs its own immanent pure activity as the horizon of the external world—“immanence is transcendence.” It is through such thinking that Kant can say determination of inner time depends on determination of outer time, or that the time of representational apprehension necessarily corresponds to the time of appearance. To read this passage, as Allison or Dummet do, merely as “I must keep a diary if I am to remember the order of my psychological states” may therefore be somewhat superficial.
This is why Kant subsequently says that any psychological appearance, even a mad person’s fantasy or a dream, has already posited and employed the outside.
Thus the Aesthetic’s claim that time has a wider scope than space does not mean that time can relate to inner psychological states while space cannot. These states themselves must first be fulfilled by spatial perception: without bodily heat there is no anger, without voices in the head there is no thinking. All these are embodied, regionalized perceptions. The genuinely important distinction is that time also directly relates to thinking activity. Although all activity must be represented as external, it retains the immanent vivacity of consciousness. So although time’s objectivity depends on spatial fulfillment, time also gives space the vivacity of subjective consciousness. Time without space is merely a contentless passage; space without time is merely a dead manifold.
Kant’s reversal here is therefore astonishing: time, as inner sense, is actually more external than space. Space directly gives intuition, while temporal activity is interpreted through mediation within space. This reversal emphasizes what “inner” really means: inwardness is the interior within space, enclosed by space, not merely “the mind opposed to the external world.” From this reversal, Kant proceeds directly to the categories of phenomena and noumena.
Some comments on the Transcendental Analytic in Kant’s first Critique:
1. Wolff previously said that he considered Kant’s schematism an invalid argument: to say that time is pure like understanding, and that time is the form of sensibility, and therefore that time mediates between sensibility and understanding, is like saying that because water is in a cup and flows from a tap, water mediates between cup and tap. This is utterly trivial. But he misunderstands Kant. Two points must be remembered. First, time as mediator is a necessary rather than sufficient condition for understanding’s connection to sensibility. Second, Kant has already delineated all the elements of pure knowledge, among which only time can mediate. Wolff’s criticism is therefore invalid. What he wants to criticize is simply, first, that water as mediator has no efficacy in unifying cup and tap; second, that even if water can unify them, the air between cup and tap can do so too. Kant’s answer to the first is that, even if time does not directly connect sensibility and understanding, it at least supplies the necessary bridge for understanding’s spontaneity: understanding, not time, is the source of this unifying activity. His answer to the second is that only time can mediate between them; there is no other option.
2. Kant’s actual design here is thus: given the unity of sensibility and understanding, find the mediator that makes this unity possible. Wolff’s criticism here is broadly reasonable: Kant’s subsumption model only describes mediation without explicitly answering why mediation is needed. If understanding is activity, could it not simply make sensibility submit to it?
3. Although plausible on its surface, this criticism has gone astray, missing Kant’s deeper field of questions. Why is mediation needed between sensibility and understanding? Because a sensible object, as one among many, has an absolute particularity, and no judgment of an absolute particular can acquire any universality or necessity. A concept of understanding, as one above many, is to be applied to sensible objects, yet its universality cannot be apprehended within their particularity—unless sensibility is reduced to a lower rank and understanding to a higher rank, followed by a discussion of participation. In short, one would abandon transcendental philosophy and return to traditional metaphysics. But transcendental philosophy’s tenet is that universality cannot be given in intuition; intuition is not a “more particular” universal. This was decided from the outset. Thus, to let both the one–many pair between concept and intuition and the one–many hierarchies internal to each jointly complete a system, Kant must prove that a universal concept can correspond to a particular intuition. He achieves this through imagination. Intuition is certainly not a universal that is more particular than other universals, but as a particular it lies within a universal particularity, pure intuition. A concept is certainly not a particular more universal than other particulars; rather, as universal activity it can delineate a particular universal, pure imagination. Kant thereby preserves the originary character of transcendental philosophy: the universal activity of concepts is delineated in imagination as a given image of pure intuition, while intuition’s particular image of presentation is contained within this schematic image.
4. Imagination therefore displays a structure completely different from both understanding and sensibility. If understanding has a hierarchical one-above-many structure and sensibility a planar one-among-many structure, imagination has a perspectival structure in which One and Many interpenetrate. A schematic image encompasses the field in which particular images present themselves. Although this encompassing proceeds from above, it itself also projects a transcendent field of givenness. This is the greatest insight of Heidegger’s Kant book: the clarification of imagination’s structure. Knowledge becomes possible only through the mediation of imagination. Through schematism, understanding no longer needs to judge particulars one by one by one; it can give a universal judgment of each particular, because its judgment directly determines the entire field. This is also why transcendental judgment is superior to empirical judgment: empirical judgment can only subsume particulars, whereas transcendental judgment can subsume the entire field of particulars. This field is possible experience, and the particulars given within it, taken together, are possible objects of experience. “Objects acquire a basis for being judged, and are thereby determined as possible, through the limitation of the intuitive manifold by the conditions making experience possible.”
5. Within such a field, some objects disclose themselves immediately and groundlessly. These are mathematical: perception, for example, has no reason. Other objects disclose themselves while also indicating the determining dark soil that makes their disclosure possible. These are dynamical: in disclosing themselves, they disclose the necessity on which they depend.
Temporality in Kant is really more a matter of founding knowledge. For knowledge of any object to be possible, knowledge of knowledge must first be possible: any judgment, or synthesis, must be determinable, or intuitable. Every synthesis must therefore be a given manifold, unless the activity of synthesis can be intuited directly within thought. Otherwise, that synthetic activity must itself be given as a representation of a manifold. The form of this given intuition of synthesis itself is time; timeless activity is thus represented as motion within time. This motion also requires transcendence, so here time must depend on space and can be represented as external only by drawing a line. Time therefore depends on space for its externalization as representation and on understanding for representing synthesis itself. This conception sounds dangerous, as though it necessarily led to infinite recursion: “If knowledge depends on knowledge of knowledge, does knowledge of knowledge depend on knowledge of knowledge of knowledge?” But this entirely misunderstands Kant’s intention. His discussion of knowledge can be divided into two aspects, level and structure, which do not affect one another. Whatever level of knowledge you want, your knowledge must ultimately come from a synthesis of intuition and concept. Whichever level you choose, if you have the concept and its corresponding intuitive representation, you have knowledge. Synthesizing a manifold requires a synthesis of the manifold itself and synthesis itself; this vertical, veritative synthesis concerns the structure of knowledge. But after a manifold is synthesized, if its synthetic process becomes a manifold to be synthesized with other synthetic processes, this planar, predicative or apophantic synthesis concerns levels of knowledge. When Kant discusses knowledge of knowledge, his whole field of inquiry concerns structure, not level. Categories are the form of knowledge, the highest unity, rather than representations themselves requiring further synthesis. In other words, although categories are synthesis, they are already the highest synthesis, and resynthesizing a representation of the highest synthesis itself belongs to the highest synthetic activity. Thus, independently of any level and beyond any determinate level, whenever a synthetic process is intuited and represented, knowledge of structure has already been achieved. This also explains time’s unity: time is one side of the pair in knowledge’s structure, its passivity, while apperception is its activity. The correct formulation is therefore not “knowledge depends on knowledge of knowledge, which depends on…” but “knowledge is the object of knowledge.” Likewise, it is not that “synthesis itself must be synthesized” but that “representations of synthesis are always being synthesized.” What has levels and infinite recursion is not synthetic activity, as though another synthesis stood above synthesis, but the representation of synthesis. Representations are continually synthesized with other representations, including representations of synthesis. Thus knowledge of knowledge does not mean another synthesis above one synthesis, because knowledge and synthesis are not synonymous. It means the union of synthesis as activity, the conceptual part of knowledge, with synthesis as representation, the intuitive part of knowledge. Knowledge of knowledge is therefore really intuition of concepts or concepts of intuition.
Put simply, “knowledge depends on knowledge of knowledge” should be understood as “understanding intuition depends on intuiting understanding.” This is a relation of mutual dependence, not of ascending levels.
Because intuition is given, knowing intuition requires understanding; because we can understand, knowing understanding requires being given an intuition.
So Kant’s final wording is actually interesting: understanding determines appearance, concepts are determined by intuitions.
Heidegger’s so-called fundamental problem of metaphysics follows this line of thought. First, the fundamental problem is the One and the Many: being as being and being as the totality of beings. Then this is reformulated as the problem of finitude. Next, finitude becomes the problem of finite being, of Da-sein. Then, how should Dasein’s finitude be understood? Heidegger slyly says it lies in questioning. Because Dasein does not fully understand finitude or its own being—understanding is always only projection—finitude is disclosed precisely in this questioning. At this point, philosophy’s fundamental question is transformed into the question of philosophy’s own possibility. This is also the most important and essential turn Heidegger sees in Kant, the true meaning of the Copernican revolution in metaphysics. In Being and Time, these steps constitute much of fundamental ontology, from the question of being through thrownness to care, and so on. The final step, however, is that this care or finitude actually arises from Dasein’s temporality and human being-toward-death. Dasein’s temporality therefore makes philosophy possible and belongs to one whole with the fundamental problem of metaphysics. Precisely because Dasein can question being only through an understanding of time, every understanding of being depends on an understanding of time.
Heidegger’s shamanistic reading of Kant’s theory of time
1. Imagination synthesizes concept and intuition and lets them acquire their essence within it. This also means that only in imagination does time become time. Since time manifests its full temporal character—identity, continuity, and so on—through imaginative synthesis, time’s more fundamental essence, its transcendental structure, must likewise be resolved through imaginative synthesis: past, present, and future.
2. Past, present, and future are transcendental forms of time and must be founded in pure imagination. The present comes from intuitive synthesis, the apprehension of material as one among many within representational space. The past comes from imaginative synthesis, where imagination is the essence of pure imagination, the one–many synthesis that places representations under a higher unifying representation. The future comes from conceptual synthesis, one above many: the whole already projected in advance into representational space.
3. Clearly, the schematism that interests Heidegger here is not the legislation of concepts or reason for sensibility. He does not directly mention the synthesis of any category or discuss the relations among categories, these syntheses, and time’s fundamental structure itself. For him, reason’s categories are fundamentally subordinate to pure imagination, its self-knowledge and self-founding. Thus, once pure imagination has also founded itself from the perspective of part and whole, it no longer matters which specific categories this brings about, because pure imagination founding itself from this perspective has already temporalized representations. Since knowledge, or pure imagination itself, becomes possible precisely through these syntheses of pure imagination, synthesized intuition must naturally bear their different traces: present, past, and future. In other words, present, past, and future are not merely forms of time; they must also be pure concepts on which the syntheses of pure imagination depend, although what they should be called as pure concepts remains debatable. In sum, although time is the continuous present within experience, it is fundamentally founded in pure imagination. The syntheses performed by pure imagination are actually originary time, the time making empirical time possible.
4. But Heidegger is not satisfied with this. Merely discovering originary time in pure imagination permits only a ready-to-hand interpretation of originary time. Originary time is not something lying inside pure imagination; rather, it is the essence of pure imagination that makes pure imagination itself possible. Pure imagination is possible because of the One–Many pair, and this pairing could not exist if humans were not finite. Pure imagination is founded on human finitude. Yet human finitude is founded on temporality, because finitude is fundamentally the finitude of intuition, the transcendence of time. Time always stands opposed to humans and constructs their perspective. It is through this transcendent structure of intuition, or time—the structure of representation—that humans fully realize their finitude. Thus, when pure imagination performs these syntheses, it has already presupposed time’s transcendence. Every imaginative synthesis is a synthesis of representations: the past is merely other representations unified under one representation, the present a particular representation, and the future the representation unifying all representations. What truly matters is not how representations are synthesized but that, while imaginative synthesis constitutes the past, the past itself is already a transcendent field opposed to imagination’s constitutive activity. What more essentially discloses human finitude and makes pure imagination’s synthetic activity possible is therefore originary temporality: the transcendence of past, present, and future relative to humans. Because human being is finite Dasein, originary temporality is also human being, or human Dasein.
Heidegger’s drastic rewriting of Kant on imagination:
1. Imagination is the origin of knowledge, founding both sensibility and understanding. This is not because it occupies a middle position connecting higher and lower faculties: at the place of origin, the distinction between higher and lower does not matter. Soil can nourish organisms both above and below ground. What matters is that understanding and sensibility must both disclose their essence in the place of imagination.
2. Sensibility is relatively easy to understand. Imagination is the pure synthesis of representations; in imagination, sensibility’s passivity is actively synthesized into a schema, synthesizing representations in advance and giving space and time, as forms of intuition, a determinate transcendental content, thereby making space and time possible. Space and time are not empty forms but nonobjects with their own structures—continuity, identity, and so on. Intuition becomes possible precisely because of these structures. It is not that intuition first supplies content by itself and imagination subsequently synthesizes it; only after imaginative synthesis are the forms of intuition possible. This is why Heidegger says the Transcendental Aesthetic cannot be understood before schematism.
3. Understanding is somewhat more difficult. As originary synthesis, imagination not only synthesizes representations; their synthesis itself, as a whole, must also be synthesized. This is self-consciousness or apperception: “I think that I think.” In connecting representations, it also connects the laws of connection as it connects representations. This gives understanding an active–passive duality: it is both law and represented law, and its law can be effective only when represented. Otherwise, the paradox would arise that “A and B as representations are synthesized together, but this together is not a whole and can neither be represented nor thought.” Moreover, every “I think” has actually already been synthesized. I can know that A causes B because I have already grasped the category of causality. I know what this category is and that it is universal. This apprehension must represent my own activity, because I am finite and can neither intuit intellectually nor create myself. Without this representation, I could make no judgment or acquire knowledge. Apperception is therefore founded on imagination, as is self-knowing understanding. Like sensibility, understanding is an activity of self-giving and therefore has its own intuition.
4. Not only speculative reason but practical reason too is founded on imagination. Speculative reason is imagination’s self-giving, which includes imagination’s self-limitation, or freedom. Within practical reason, the self is disclosed as a moral person in present self-consciousness. Since every disclosure determines being itself, the self’s being—the foundation of self-consciousness, transcendental imagination—must also be determined in this way, so that imagination can provide the possibility of founding morality. What is needed is a deduction of transcendental feeling: the content of the Critique of Practical Reason, how respect for the moral law can be deduced from imagination.
I feel Kant needs imagination’s mediation because he thinks sensibility’s one among many, the principle of the Many, cannot be directly connected to understanding’s one above many, the principle of the One. The sensible manifold must be structurally infinite: the Many must be a group, even one group of one. Concepts must be structurally finite: the One must be a law, completely unifying multiple representations under one representation. Adding step by step from finite to infinite, as Leibniz does, cannot explain the genesis of sensibility, because the difference is qualitative or structural, not merely quantitative. The One and the Many here are qualities, not quantities. A mediator is therefore needed: imagination represents the One as Many and then represents the Many as One. Although One and Many remain qualitatively different, harmony can at least be achieved in the phenomenal domain.
That is, although imagination contains both intuition and concept, as pure synthesis it grasps all concepts simply as representations within the same field, projecting them onto one plane and connecting them. It thus represents the One as Many, gives concepts intuition, and allows them to be synthesized with other intuitions. Likewise, its synthesis establishes a unified form and distinguishes it from content. Put simply, knowledge becomes possible because of imagination. A concept as one above many, even with infinitely many other concepts beneath it, is in itself merely a One. Only when imagination represents this One in intuition as one among many can its connections with other representations be established. Every synthesis therefore depends on intuition. The same holds for intuition: as one among many, it remains Many however far its quantity is reduced. Unless imagination synthesizes it into a whole, knowledge cannot arise.
5. Imagination is not the ability to fantasize aimlessly. First, it not only imagines absent things but also makes knowledge of present things possible; even the possibility of presence itself is founded by imagination. Second, it does not merely imagine things but the horizon making any thing possible. The anthropological understanding of imagination therefore overlooks its crucial mediating role in ontological knowledge. Imagination is not a capacity for boundless fantasy opposed to finite reality. On the contrary, it is the capacity through which the infinite manifold can be limited and appear in human cognition. Of course, compared with perceptual experience, imagination emphasizes possibility rather than reality and therefore seems somewhat less restricted. In fact, the reverse is true: precisely because it emphasizes possibility, it cannot structurally be the purely infinite Many—Many here in the structural and qualitative sense—but can only be relatively more finite: a triangle, several x’s, rather than the pure manifold given by perception. This relatively finite possible concept requires the sensory manifold to fulfill it before it can have reality. On this point, the association of possibility with finitude and reality with infinity, Kant resembles Leibniz.
Kant’s model is roughly this: sensibility receives the represented one among many; imagination connects these represented ones among many into a whole of representation; and this wholeness comes from pure understanding as one above many. Sensibility provides an undetermined object juxtaposed as an appearance with many other objects—content, objective reality—while understanding determines it—form, objective validity. Pure understanding provides the horizon and foundation for this determination. That is, whenever imagination begins determining or synthesizing representations and generates hierarchies of concepts to unify representations under a representation, those representations must already have been synthesized in advance under pure understanding as what can be thought. Only when pure understanding attains self-knowledge and recognizes the necessary rules of pure synthesis does such knowledge as a whole become possible. All this is clear. Beyond this, transcendental logic includes questions such as understanding’s unity. Unlike Aristotle’s abstraction from reality, Kant simply turns toward transcendental understanding itself and obtains direct self-knowledge. In discussing identity, truth, and completeness, he also indicates how understanding synthesizes quantity, quality, and relation in self-knowledge: identity is the premise that there is only one; truth concerns how many qualitatively different results can be deduced from that premise; completeness is the whole of the deduced results. Another question I think Kant handles inadequately is how reciprocity connects to disjunctive judgment, either A or B.
If the representations brought by sensible intuition are to be thinkable, they must already have the possibility of being thought this way. They cannot be mere material; even material must already have been formed in advance. This is time as schematism.
Because intuition fulfills a concept, when the concept acquires determination, its intuition must have a corresponding determination: intuition’s form must correspond to its concept. This is the source of schematism.
Kant opposes Aristotle because Aristotle puts categories into reality. He opposes traditional formal, nontranscendental logic because only in understanding’s activity, judgment, can we obtain transcendental determinations of objects. Formal logic cares only about the order of thought, not how that order makes knowledge possible and provides itself with the objectivity of objects—what understanding determines in advance, before judging, as a possible object of experience.
Lacking this transcendental perspective, traditional logic cannot explain where its determinations come from, whether from things themselves or from the mind. Only when its own source is clarified as understanding’s unity, and objects acquire the possibility and foundation of determination only within logic, does logic become genuinely fulfilled.
Infinite judgment is when A is not B and is non-B. If we merely say A is not B, A is only excluded from B, without being determined as to whether it exists or where it exists. Only through the infinite accumulation of “A is non-B, A is non-C…” can we determine what A actually is.
Longuenesse explains disjunctive judgment as follows: traditional logic merely analyzes and clarifies concepts already synthesized in advance. As the faculty of judgment, understanding must, while judging, synthesize concepts in advance into what can be judged. Things are therefore first synthesized as reciprocal or mutually determining. Only when there is such a whole jointly composed of multiple independent parts is disjunctive judgment possible. Thus Kant does not think judgments about concepts and judged facts always coincide completely; rather, formal logical judgment and its content have the same source. Judgment becomes possible only after content has been synthesized. The question is therefore not simply “what object corresponds to such a judgment?” but “what synthesized object permits such a judgment?” This is actually easy to understand, because categories function to make appearances into objects that can be thought and judged. This is a founding relation, not a simple one-to-one correspondence or analogy.
Furthermore, the synthesis of the same class that founds each class of judgments applies only to particular kinds of representations: quantity to intuition, quality to perception, and so on.
But these particular kinds of representations are synthesized under a general concept of object. Or rather, these different kinds are simply different aspects of the same representation, concerning how it should be determined when placed with other representations. Suppose representation A could be thought only in terms of quantity and representation B only in terms of quality. There would then be no relation between A and B. This would mean A could not be thought in B, or that thinking became impossible when A and B were juxtaposed. But this is absurd, because every representation is unified under the apperception “I think,” and an object is what can be thought. Being unable to think A and B together would mean A could not be determined by B: under B, A would cease to be an object and could not appear in experience.
The possibilities of Kant’s Aesthetic
1. The entanglement of time and space, time as inner sense, the process of geometrical drawing, motion as a pure science concerning both space and time—perhaps this is precisely why Kant later rejects its status as pure science—and the possibility of reversing inner and outer sense, for example, temporal variation in the mind situated here.
2. Concepts can relate to intuition directly or indirectly. The concepts of space and time both relate directly to intuition, so merely analyzing their concepts can reach the givenness of their intuition. In their most essential determination, the concepts of space and time are not syntheses of different species and genera of space and time but direct correspondences to spatial and temporal intuition.
3. Intuition refers to and determines objects, but any perception refers to an object only partially. No perception can directly refer to the whole object—every perception is incomplete and relative—because the whole object needs to be given by the whole of perception and determined by the whole spatiotemporal form. Intuition acquires an object’s material through perception as a whole and fulfills itself through this material. Only then can perception be said to bring knowledge. In other words, knowledge on the material side is possible only through the materials’ reciprocal determination of the object as a whole. On the formal side, spatial and temporal perception is itself the pure form determining the possibility of objects, so its knowledge already encompasses the whole object. Sensible knowledge on the material side is empirical intuition; sensible knowledge on the formal side is pure intuition.
4. Pure intuition always concerns an image, or a half-object. An object given through imagination, such as a triangle or a moving sphere, has meaning only insofar as it lies within the space and time of pure intuition, supporting merely a symbol for any transcendental possible object. In other words, imagination gives intuition a content, but what is actually grasped is knowledge of space and time themselves as pure forms. Because space and time make all objects possible, through our mastery of imagination—the power of variation—we turn our attention from material to form and thereby confirm the transcendental character of pure intuition.
5. Against Leibniz, Kant rejects a merely gradual distinction between concept and perception. Leibniz thinks substance is simply an infinite concept and perception an abstraction of this concept. Kant, however, would say that every concept’s representations are infinite, but not every abstraction of representation at every conceptual level is perception. Abstracting the concept of right, for example, does not yield perception. Likewise, complete clarification of perception does not yield concepts, only more perception. After all, perception has an inclusive structure without depth, whereas concepts have a hierarchical structure with depth. Of course, this criticism does little harm to Leibniz. For him, the infinite concept and the highest concept are not the same but stand at opposite ends. Retrospective abstraction toward the highest concept is conceptual abstraction; projective abstraction toward the infinite concept is perceptual abstraction.
6. Thought is finite. But thought provides a way of placing infinity within finitude. Humans can therefore finitely place the thing in itself within thought, making it a finite, unknowable infinity and thus limitation itself. Finite infinity is transformed into infinite finitude.
The fifth point is extremely important. It directly reveals that Kantian intuition and concept are organized according to two completely different principles, faculties, and structures: intuition is one among many, concept one above many. Intuition determines the definite scope of concepts; concepts extend the connections among intuitions. Imagination generates formal intuition when related to concepts and material intuition when related to intuition.
This also explains why Kant ultimately chooses a line to schematize time: time’s image is ultimately grasped as that half-object representing understanding’s unifying activity.
On the fourth point: empirical intuition–pure intuition–imagination–concept. Precisely because imagination is the faculty of universal representation, it can detach pure intuition from experience. Likewise, because imagination represents according to pure concepts, or schemata, it turns time, which depends on it to become a form of representation, into a schematized horizon.
Heidegger on Kant:
1. Dasein is an ontological structure determining beings as interpreted as what Dasein is directed toward. In Kant, this ontological structure is determined as the transcendental subject’s transcendental determination of objects.
2. Within this ontological structure of finite Dasein, we encounter a problem: how is ontological knowledge of this structure possible? This can be simplified into three steps or aspects. First, we know that the beings toward which Dasein is directed are always transcendent, or that the thing in itself is unknowable. Yet we still affirm that being itself requires positing the thing in itself: the subject–object structure applies to every object. This means that we have knowledge of being of which we should have no knowledge. We clearly recognize Dasein’s structure as ontological rather than applicable to only some beings. This means the thing in itself must have the characteristic of accepting the determination “thing in itself.” In other words, it contains immanence within its transcendence, and the multiple components of this structure, including the thing in itself, require a higher unity to found them. That is, whatever is considered external to this structure also constitutes it. What we now need to explain is how such an outside is generated within the structure.
From what I have read so far, I think Heidegger’s general direction is correct, but his detailed deductions are rough and disorderly. For example, he reduces the question of the whole structure’s unity to the unity of sensibility and reason, and reduces the genesis of understanding to its function of allowing immediate representations to be represented again. This may have to do with his attention to the first Critique rather than to the three Critiques as a whole.