孔德：把缝隙灌满水泥的人
Comte: The Man Who Grouted the Cracks
一、那本日历
1849年，奥古斯特·孔德出版了一本日历。到了1852年，它已经出了第四版。
不是普通的日历。月份的名字不是一月、二月，而是摩西、荷马、亚里士多德。每天、每周、每月，都安排了值得纪念的人。诗人在那里。科学家在那里。圣保罗也在那里，只是被纪念的理由换了：他们不再替天上的真理作保，而是进入了孔德安排的人类历史。
十三个月。每月二十八天。三百六十四天排好，另有纪念日补足一年。连闰年多出的那一天，也有自己的位置。
他管这叫“实证主义日历”。
这不是行为艺术。孔德是认真的。他想让人类不再依赖超自然的权威，也仍然有可以共同敬重的人、可以共同度过的节日。上帝不再坐在那个位置，人类自己来。他给这套安排起了一个名字：人道教。有教义，有仪式，有等级。他自己站在精神领袖的位置。
一个拒绝用超越经验的本质解释世界的人，最后建了一个没有上帝的宗教。
最值得看的，不是笑他前后矛盾。是那些没有因为知识变得精确就消失的需要：哀悼一个人，爱一个人，想跟别人一起记住什么。
孔德知道需要还在。他要把它们也安排进去。
日历可以是一份邀请。孔德也把它看作过渡性的安排。问题是：当安排越来越完整，坐不进格子的人，还能不能让格子改一改？
这一篇从那本日历开始。纸很薄，背后却站着一个很大的愿望：不要再留下无处安放的东西。
二、他在堵什么
要理解孔德，先理解他在堵什么。
孔德1798年出生在蒙彼利埃，法国大革命开始才九年。他长大的世界里，旧秩序已经被掀翻，新秩序还没站稳。教会的权威碎了，王权的权威碎了，启蒙运动许诺的理性天国也没来。到处是缝隙。
他进了巴黎综合理工学院，跟圣西蒙共事几年，然后闹翻。后来做过辅导教师和入学考试考官，却没有得到期望的教授席位。工作失去了，朋友和支持者的资助使他还能继续写。一个想替整个人类安排位置的人，自己的位置并不稳。
但他在写。1830年至1842年，六卷《实证哲学教程》陆续出版。数学、天文学、物理学、化学、生物学、社会学。他把一门门知识摆到一起，想看见它们共同走过的路。
这条路叫“三阶段法则”。
神学阶段，用神的意志解释现象。形而上学阶段，把神换成抽象的本质或力量。实证阶段，不再追问不可把握的第一因和绝对本质，而研究现象之间的关系与规律。
雷为什么响？说是某位神发怒，和研究闪电、空气与雷声之间的关系，是两种不同的工作。孔德要求后一种。他不是不许人问所有的“为什么”，而是不再用一个摸不着的终极原因，代替实际可以展开的解释。
这一刀有力量。
但孔德还往前走了一步。他把实证阶段叫作人类理性的最终状态。
不是“我们现在从这里工作”。是“人类长大以后，就这样工作”。
他知道知识有限，也承认知识相对于人的处境。地图上的答案可以改，制图的最高方向却仿佛已经定下来了。未知可以留下，只要它继续按已经规定好的方式成为已知。
缝隙不一定被填平。它也可能被取消发问的资格。
三、知识只有一种合法形式
这套逻辑的力量极其巨大。
孔德不只是在说“科学好”。《论实证精神》给有意义的命题设了一道门槛：它应当能归结到某种特殊或一般的事实。拿不出这样的联系，便不能只凭一个漂亮词语要求别人相信。
问题在于，那道门槛能不能也被追问。
如果我问的不是“这一张表里哪个格子空着”，而是“为什么只能用这一张表”，这句话还有地方站吗？还是一开口，就被送回人类不成熟的阶段？
这就是本系列写孔德时，想画出的水泥。
柏拉图的墙，亚里士多德的地板，孔德的水泥。墙把答案放在远处；地板替眼前的东西分好位置；水泥则告诉你，两块板之间那个不服帖的地方，不必再看。
孔德当然不只是水泥。他承认观察需要理论，也知道不同学科不能简单压成一条公式。他给知识排次序，不是宣布自己已经知道一切。
可也正因此，那个“最终”才更值得问。既然答案要改，为什么规定答案如何成立的那一层，可以先把自己安排到终点？
把一扇门暂时关上，和抹平门框、说这里不该有门，不是一回事。
实证的方法可以清理许多空话。方法一旦变成所有问题的地主，也会把仍然有东西要说的人，变成非法住客。
水泥不是观察，不是分类。水泥是“不必再问”硬下来的那一刻。
四、克洛蒂尔德
问题是，孔德自己的生活里全是缝隙。
1826年，他经历严重的精神健康危机，接受过治疗。后来，他与妻子卡罗琳·马桑的关系破裂，1842年分居。生活没有像知识的次序一样，一格一格地安稳下来。
1844年，他遇见克洛蒂尔德·德沃。孔德爱上了她。1846年，她病逝。他此后长久地纪念她，把她放进自己的思想和仪式里。
但先把她从圣像里请出来一会儿。
她写作，发表过《露西》，想靠自己的工作获得独立。她愿意与孔德保持友谊，却不愿接受他的恋爱和共同生活安排。她不是没来得及说话。她说过不。
一个人被爱，并不因此欠下一份同样的爱。
孔德的哀悼是真实的。她的拒绝也是真实的。难处正在这里：当他把她放到体系最明亮的位置，她原来不同意的地方，还能不能一起留下？
本来是两个各有方向的人。后来，一个已经不能再写信，另一个还在替这段关系命名。
人道教并非从一次失恋或一次死亡中突然长出来。孔德很早就想把知识、道德与共同生活重新组织起来。克洛蒂尔德进入了这个计划，也使它有了一个再也不能当面询问的人。
他没有拆掉自己的构。他围着失去，建起了纪念的地方。日历，祈祷，仪式。房间比以前更完整，那个不在的人，却不能再亲自决定怎样进来。
穆勒和孔德逐渐疏远，分歧涉及女性地位、资助关系和思想方向，不是看见一个宗教名称便突然摔门。亲近过的人，也没有都留在孔德为他们安排的位置。
一个理论能安排纪念，能不能容纳被纪念者自己的话？这比“科学家为什么信起了宗教”更难回答。
你灌了水泥，水仍会从别的地方冒出来。这里的水，不是他的疾病给哲学下的判决。是一个活人不肯被安排说尽的地方。
五、模式匹配的天花板
把这件事放到今天，可以先看一种熟悉的工作：观察，分类，建模，预测。
从经验里找规律。给定条件，推算结果。哪里预测错了，就回来修模型。这很有用。天气预报，实验安排，许多日常判断，都离不开它。
但模式匹配不是认知的天花板，是认知的地板。
“我能预测”，和“我知道自己正在预测什么，为什么要预测这个”，中间还有一步。模型里的答案越来越好，模型没有问的事，却不一定跟着消失。
比如一套办法能准确判断谁会服从，却始终没有问：为什么我们只想得到服从？你可以把预测做得更准，问题仍然留在原处。
这不是说孔德本人只会匹配模式。他对思想的历史和方法有大量反思。这里要追问的是：当三阶段法则成了那道已经规定好的题目，人的反思还有没有可能走出题目？
一个方法可以让人认识世界，也可以反过来规定，人只许以什么方式遇见世界。
下一篇的波普尔，会把脚步朝另一边转。
不是先把支持我的事实越堆越高，而是先提出猜想，再主动寻找它可能失败的地方。让世界有机会说不。观察不只是给答案添砖，也可以把一块砖抽出来。
这里的“猜想在前”，不是说未经检验便有了康德意义上的先验知识。是提出问题的人先走了一步，却没有把世界必须赞同他的权利一起拿走。
方向盘回来了。检验还在。
但检验的标准，会不会又被拿去管理所有别的工作？波普尔区分科学与非科学，并不因此宣布后者无意义。可一条线离开书本，落到别人手里，也可能变厚。
下一篇写波普尔。他使问题转了一个方向。我们就沿着这个方向，再走一步。
六、他做对了一件事
但孔德做对了一件事：他把问题说得很尖锐。
给未知起一个名字，不等于已经解释了它。说某种本质在起作用，仍然要问：它怎样起作用？我们凭什么知道？
这一刀切出了可以工作的地面。
地面不是他一个人铺的。科学也没有等他出生才开始。但把分散的学科连起来，要求解释面对实际世界，是他真正做过的工作。不能因为要看缝隙，就把地板一并扔掉。
狄拉克会从这样的地面上走出去。不是没有经验，凭空听见一个方程。是已有的物理知识彼此挤压，数学的要求又把他带到尚未看见的地方。方程先到了，实验还要跟上。
屈原走的是另一条路。他不需要把《离骚》变成一份实验报告。诗让一个人的处境来到面前：不是告诉你在什么条件下必然产生一首诗，而是让你听见，人在这样的地方怎样说话。
麦克林托克会带着玉米来。“对有机体的感觉”，不是从没有观察的地方飘来的神秘提示。那里面有多年的种植、比较和等待。她看见一个差别，别人也许只看见背景。差别一旦出现，原来的问题也得改。
波兰尼会在后面把这一点说得更清楚：你能写下的步骤，不一定装得下你实际知道怎样做的事。
孔德要人交出根据。这份要求应当留下。但交出根据，不等于先把一个人的全部认识都压成表格，才许他开口。
地面可以站。边缘也还可以走。
七、三种封闭
孔德在这个乐章里的位置，从这里清楚起来。
柏拉图的墙：答案仿佛已经在上面完成。这是从上面封闭。
亚里士多德的地板：分类太完备，便容易把位置当成东西本身。这是从下面封闭。
孔德的水泥：不是声称所有事实已经知道，而是把某种求知方向规定成人类成熟的终点。这是从中间封闭。还没回答的问题里，有一部分先被宣布不必回答。
三种封闭。墙，地板，水泥。这是本系列的三幅图，不是三个人一生的全部。图的用处，是让我们看清下一步从哪里走。
本乐章十八个人，要做的事就是把那些封住的地方重新打开。波普尔让猜想接受可能推翻它的检验。狄拉克让我们听见，提出猜想之前和之后，数学结构怎样改变提问的人。屈原让美开口。费希特把认知者放回问题里。麦克林托克带来长期工作养成的感觉。薇依带来注意。龙树追问看似坚固的支点。
一直到波兰尼，把线索接起来：说不出来的东西，不一定是知识的残余。它也可能是你说得出来的那些话，一直站着的地基。
从“什么才有资格被算作知识”，到“知识靠什么才能发生”——这就是本乐章的弧线。
这些人不必同意彼此，路才可以走。他们带来的分歧，也是路的一部分。
孔德站在起点。他想给人类一张不会再迷路的地图。后面的人要问：地图没有画出来的地方，我们是不是就不能再去？
八、桥头
现在，让孔德走到这座想象的桥上。
他的脚步很重。他带着一本日历，上面写满人名。他在找一个合适的位置站，但他的站法跟别人不一样。别人还在看桥通向哪里，他先看这些人有没有各自的位置。
他把日历摊开在栏杆上，开始清点。伽利略，在。牛顿，在。莎士比亚也在——不是站在格子里，是桥下面的水。孔德听见水声，再低头看那个名字。
日历是他的地图。地图上，水声没有那么宽。
康德在远处站着。孔德看了他一眼。两个人都知道知识有边界，但康德站在那里，仿佛还在问边界怎样成为边界。孔德低头，把眼前能安排的先安排好。这一点距离，看起来很近，走起来是一整条路。
亚里士多德走过来看了一眼他的日历。笑了笑。分类这件事，他早就做过。但分类帮你找到一个人以后，那个人还会继续走。
孔德暂时没有看桥下面。
夜色渐深的时候，他从口袋里掏出一张小画像。克洛蒂尔德·德沃。他看了很久。她曾经写下的那句不，也在他们之间。画像没有替她收回。
然后他把画像收好，重新把日历铺平。日历的某一页被风吹起来，露出下面栏杆的缝隙。他把那一页按回去。
按得很用力。
注释
生平与克洛蒂尔德的书信摘引参孔德故居资料；日历据1849年文本及1852年第四版书目。1842年是与马桑分居，并非离婚；穆勒1841—1846年的通信与1865年的著作属于不同阶段。疾病、情感和婚姻不作为哲学真假的证据。
《教程》第一讲承认理论与观察相互需要；《论实证精神》§§12—18同时提出意义限制与知识的相对性。本文不把孔德等同于后来的逻辑实证主义。墙、地板、水泥是作者的比较性读法，不是对三位哲学家的完整概括。“先于某次检验”也不等于康德意义上的先验。
第八节为文学寓言，包括桥头会面与随身画像；莎士比亚延续本系列桥下的水声。科学、诗和默会认识的并置不是一条已证实的影响链；模式匹配的讨论只针对所描述的方法，不是对孔德个人或全部人工智能的能力诊断。
I. The Calendar
In 1849, Auguste Comte published a calendar. By 1852 it had reached its fourth edition.
Not an ordinary calendar. The months were named not January and February but Moses, Homer, Aristotle. Days, weeks and months commemorated people. Poets were there. Scientists were there. Saint Paul was there too, but the reason for remembering him had changed: these figures no longer guaranteed a supernatural truth. They entered Comte’s arrangement of human history.
Thirteen months. Twenty-eight days each. Three hundred and sixty-four days in place, with an additional day of commemoration to complete the year. Even the extra day of a leap year had somewhere to go.
He called it the Positivist Calendar.
This was not performance art. Comte meant it. Humanity should cease to depend on supernatural authority without losing the people it could honour together or the days it could share. God no longer occupied that position. Humanity would manage. His arrangement had a name: the Religion of Humanity. Doctrine, ritual, hierarchy. Comte himself in the position of spiritual leader.
A man who refused to explain the world through essences beyond experience ended by founding a religion without God.
The interesting response is not simply to laugh at the contradiction. Look instead at the needs that more precise knowledge had not made disappear: mourning someone, loving someone, wanting to remember something with other people.
Comte knew the needs remained. He wanted to arrange those too.
A calendar can be an invitation. He regarded this one as a transitional arrangement. But as the arrangement becomes more complete, can the people who do not fit still ask it to change?
This essay begins with that calendar. Thin paper, carrying an enormous wish: let nothing remain without a place.
II. What He Was Sealing
To understand Comte, understand what he was sealing.
Born in Montpellier in 1798, nine years after the Revolution began. He grew up where the old order had been torn down and the new one had not found its footing. The Church’s authority was shattered. The crown’s authority was shattered. The rational utopia promised by the Enlightenment had not arrived. Cracks everywhere.
He entered the École Polytechnique, worked with Saint-Simon, then broke with him. Later he held positions as a tutor and an admissions examiner, but never obtained the professorial chair he wanted. Jobs were lost. Friends and supporters helped him keep writing. A man trying to place the whole of humanity had no very secure place himself.
But he was writing. Between 1830 and 1842, six volumes of the Course in Positive Philosophy appeared. Mathematics, astronomy, physics, chemistry, biology, sociology. He brought the disciplines together to see the road they had travelled.
That road was the Law of Three Stages.
The theological stage explains phenomena through divine wills. The metaphysical stage replaces those wills with abstract essences or powers. The positive stage leaves inaccessible first causes and absolute essences behind, investigating relations and regularities among phenomena.
Why does thunder sound? Saying a god is angry and investigating lightning, air and sound are different kinds of work. Comte demands the latter. He is not banning every question beginning with why. He refuses to let an unreachable ultimate cause stand in for an explanation that can actually be developed.
The cut has force.
But he takes another step. He calls the positive stage the final state of human reason.
Not: this is where we work for now. This is how humanity works when it has grown up.
He knows knowledge is limited, and relative to our situation. Answers on the map can change; its governing direction seems settled. The unknown may remain, provided it continues to become known in the prescribed way.
A crack need not be filled. It can also be denied permission to become a question.
III. One Legitimate Form
The power of this framework was enormous.
Comte was not simply saying that science is good. His Discourse on the Positive Spirit sets a threshold for meaningful propositions: they must be reducible to some particular or general fact. Without that connection, an elegant word cannot demand belief.
But can the threshold itself be questioned?
Suppose I ask not which square in the table is vacant, but why we must use this table. Does the question have anywhere to stand? Or is it sent straight back to humanity’s immature years?
That is the grout this series wants to draw around Comte.
Plato’s wall. Aristotle’s floor. Comte’s grout. The wall puts the answer in the distance. The floor gives the things before us their positions. The grout tells us not to bother with the place where two boards fail to fit.
Comte is not merely grout. He acknowledges that observation needs theory, and that different sciences cannot simply be compressed into one formula. Arranging knowledge is not a claim to know everything.
That makes the word final more interesting, not less. If answers must change, why can the arrangement governing how answers count place itself at the destination already?
Closing a door for now is different from plastering over its frame and saying there should never have been a door.
A positive method can clear away empty talk. When it becomes the landlord of every possible question, it can also turn somebody with something left to say into a trespasser.
Grout is not observation. It is not classification. It is the moment when “no need to ask again” sets hard.
IV. Clotilde
The problem is that Comte’s own life was full of cracks.
In 1826 he experienced a serious mental-health crisis and received treatment. His marriage to Caroline Massin later broke down; they separated in 1842. Life did not settle into place as neatly as the order of the sciences.
In 1844 he met Clotilde de Vaux. He fell in love with her. In 1846 she died. He continued to commemorate her, giving her a place in his thought and rituals.
But first, let her step out of the sacred image for a moment.
She wrote, published Lucie, wanted independence through work of her own. She welcomed friendship with Comte but refused the romantic relationship and shared household he proposed. It was not that she had never had the chance to speak. She had said no.
Being loved does not create a debt to love in return.
His grief was real. Her refusal was real too. Here is the difficulty: when he places her in the brightest part of his system, can the things she disagreed with remain there as well?
There had been two people, each with a direction. Then one could no longer write a letter, while the other could still name the relationship.
The Religion of Humanity did not spring fully formed from a rejection or a death. Comte had long intended to reorganize knowledge, morality and common life. Clotilde entered an existing project. She also gave it a person whom he could never question face to face again.
He did not dismantle his construct. He built a place of commemoration around the loss. Calendar, prayer, ritual. The room became more complete. The absent person could no longer decide for herself how to enter it.
Mill’s estrangement from Comte grew through disagreements about women’s position, financial support and intellectual direction. It was not a door slammed at the sight of a religious label. People who had been close to Comte did not all remain where he placed them.
A theory can arrange a commemoration. Can it hold the remembered person’s own words? That is harder than asking why a champion of science founded a religion.
Pour grout and water finds another way out. The water here is not an illness pronouncing judgment on a philosophy. It is the part of a living person that an arrangement cannot finish describing.
V. The Ceiling of Pattern Matching
Bring this closer to the present. Begin with familiar work: observe, classify, model, predict.
Find regularities in experience. Given these conditions, calculate the outcome. If the prediction fails, repair the model. This is useful. Weather forecasts, experiments and countless ordinary judgments depend on it.
But pattern matching is not the ceiling of cognition. It is the floor.
Between “I can predict” and “I know what I am predicting, and why I am predicting this,” there is another step. The answers inside a model can improve while the things it never asks remain untouched.
A method might predict obedience accurately and never ask why obedience is all we want. Refine the prediction as much as you like. The question remains where it was.
This does not make Comte a man capable only of matching patterns. He reflected extensively on the history and methods of thought. Our question is this: once the Law of Three Stages becomes the assigned problem, can reflection still step outside the assignment?
A method can help us encounter the world. It can also begin dictating the only terms on which that encounter is allowed.
In the next essay, Popper turns in another direction.
Instead of piling up supporting facts, propose a conjecture and actively seek where it might fail. Give the world a chance to say no. Observation need not merely add a brick to an answer. It can pull one out.
Conjecture first does not turn an untested idea into a priori knowledge in Kant’s sense. A person has taken a step before this test, without acquiring the right to the world’s agreement.
The steering wheel returns. The test remains.
But might the test become a way of governing every other kind of work? Popper’s distinction between science and non-science does not make the latter meaningless. Yet a line can thicken when it leaves the book and passes into other hands.
The next essay is Popper’s. He gives the question another direction. We will take another step along it.
VI. What He Got Right
But Comte got one thing right: he made the question sharp.
Giving the unknown a name is not the same as explaining it. If some essence is said to be at work, we can still ask how it works, and how we know.
That cut opens ground on which work can be done.
He did not lay the ground alone. Science did not wait for his birth. But connecting dispersed disciplines and requiring explanations to answer to the actual world were real contributions. Wanting to see the cracks is no reason to throw away the floor.
Dirac will walk out from such ground. Not without experience, hearing an equation from nowhere. Existing physical knowledge presses against itself; mathematical requirements take him somewhere not yet seen. The equation arrives first. Experiment must still follow.
Qu Yuan takes another road. He need not turn the Li Sao into an experimental report. Poetry brings a person’s predicament before us. Not a law specifying when a poem must occur, but the sound of a person speaking from this place.
McClintock will bring maize. A feeling for the organism is not a mysterious message from somewhere untouched by observation. Years of planting, comparing and waiting live inside it. She sees a difference where somebody else might see background. Once it appears, the old question may have to change.
Polanyi will make the thread clearer later: the steps you can write down may not contain everything you actually know how to do.
Comte asks us to give grounds. Keep that demand. Giving grounds need not mean compressing everything a person knows into a table before letting them speak.
The ground can support us. Its edge can still be explored.
VII. Three Closures
Comte’s place in this movement becomes clear here.
Plato’s wall: the answer seems complete somewhere above. Closure from above.
Aristotle’s floor: classification becomes so comprehensive that the position can be mistaken for the thing. Closure from below.
Comte’s grout: not a claim that every fact is known, but the designation of one direction of inquiry as humanity’s mature destination. Closure from between. Some unanswered questions are declared unnecessary before they can be answered.
Three closures. Wall, floor, grout. Three pictures in this series, not the whole of three people’s lives. The pictures help us see where to take the next step.
The eighteen figures in this movement will reopen these sealed places. Popper exposes conjectures to tests that could defeat them. Dirac lets us hear how mathematical structures change the person asking, before and after a conjecture takes shape. Qu Yuan lets beauty speak. Fichte returns the knower to the question. McClintock brings feeling educated by long practice. Weil brings attention. Nāgārjuna questions seemingly solid supports.
Then Polanyi connects the thread: what cannot be articulated need not be knowledge’s residue. It may be the ground on which the things you can say have always stood.
From “what is entitled to count as knowledge?” to “what makes knowing possible?” That is the arc of this movement.
They do not have to agree for the road to continue. Their disagreements are part of it.
Comte stands at the beginning. He wants to give humanity a map on which it cannot lose its way again. Those who follow ask: if a place is not on the map, are we no longer allowed to go there?
VIII. The Bridgehead
Now let Comte walk onto this imaginary bridge.
His footsteps are heavy. He carries a calendar dense with names. He looks for somewhere to stand, but his way of standing differs from the others’. They are still looking at where the bridge leads. He first wants to know whether everyone has a place.
He spreads the calendar on the railing and starts counting. Galileo: present. Newton: present. Shakespeare too — not standing in his square, but the water beneath the bridge. Comte hears the water, then looks down at the name.
The calendar is his map. On the map, the water is not so wide.
Kant stands at a distance. Comte glances at him. Both know that knowledge has limits. But Kant seems still to ask how a boundary comes to be a boundary. Comte lowers his eyes and arranges what can be arranged before him. The distance looks small. Walking it is a whole road.
Aristotle comes over and glances at the calendar. Smiles faintly. Classification is something he has done before. But once classification has helped you find somebody, the person keeps walking.
For now, Comte does not look beneath the bridge.
As night deepens, he takes a small portrait from his pocket. Clotilde de Vaux. He studies it for a long time. Her no is still between them. The portrait has not withdrawn it for her.
Then he puts it away and smooths the calendar flat. A gust lifts one page, revealing a crack in the railing beneath it. He presses the page back down.
He presses hard.
Notes
Biography and excerpts from Clotilde’s letters: Maison d’Auguste Comte. The calendar dates refer to the 1849 text and the catalogue record for its fourth edition, 1852. Comte separated from Massin in 1842; this was not a divorce. Mill’s correspondence of 1841–1846 and his 1865 book belong to different phases. Health and private relationships are not evidence for or against a philosophy.
The Course’s opening lecture recognizes the mutual need for theory and observation. Discourse on the Positive Spirit, §§12–18, combines a restriction on meaning with relative knowledge. Comte is not simply identified with later logical positivism. Wall, floor and grout are the author’s comparative images, not complete accounts of the philosophers. Preceding a particular test does not establish the Kantian a priori.
Section VIII is a literary fable, including the meeting and the pocket portrait. Shakespeare continues as the sound of water beneath the bridge. The juxtaposition of science, poetry and tacit knowing does not assert an established chain of influence. Pattern matching names the specified procedure, not a diagnosis of Comte or all AI.