Sorted by New

Wiki Contributions


The design looks great. Some questions

  1. What software did you use to make the book? (I'm guessing some variant of LaTeX.)
  2. What inspired this project? 
  3. Are there any plans to produce further volumes like these in future? And has anything like this been published previously (apart from Yudkowsky's Rationality book)?
  4. Do you plan to publish a "Best of the Best" collection - all-time bests, rather than best of that year? (I expect such a collection would exclude the Yudkowsky essays already published in his book.)

Real Social Engineering. That's already been happening! Foucault describes the development from sovereign power to disciplinary power to governmentality. I expect the trend to continue.

These are very helpful points. It's always interesting to see people's inner processes. Thanks for writing it!

Do you have a link to the post?


Thank you for your generous response. Upon rereading my comment, parts of it sounded overly strident and accusatory, and I retract the tone of my comment. I was overly critical and didn't mention the good parts of your post. I've edited it slightly to downplay those and have added portions to emphasise that your post is an overall good for LessWrong. There's not enough discussion of these things, and I hope that my earlier post makes it clearer that I think your contribution is valuable.

I'd welcome anything further on these that you write here, especially the possible posts on Xunzi and Mozi! I hope my comments haven't discouraged you. I'd like to see more of your work.

Our main disagreement

Our main disagreement perhaps came from the fact (only realised after reading your comment) that you were focusing on steelmanning the I Ching, while I focused on interpretations of Xunzi as a rationalist. Yes, I would have liked you to explain Xunzi better and more fully grapple with the philosophy in ways mentioned above. But now I realise that this wasn't your main aim and that it was uncharitable of me to ask you to do so.

Xunzi on ritual

I'll quote Berkson's paper in The Dao Companion to the Philosophy of Xunzi:

Li is not merely repetitive, instrumental action (as might happen on an assembly line, or when we brush our teeth every day), but has a symbolic element. It is also essentially social, and so would not encompass what we might call “private rituals” that we create for ourselves. We do not invent li, we inherit them from the sages. By participating in and investing ourselves in the rituals, we are engaged with tradition, and in turn preserve and transmit that tradition.
[. . .]
Xunzi argues that rituals accomplish the following things—
(1) create and sustain social order, fostering right relationships with others (ritual’s ordering function);
(2) harmonize human beings with the larger natural order (ritual’s ecological function);
(3) enable the cultivation of good character in individuals, in particular virtues, right attitudes, emotions, and ways of thinking (ritual’s developmental / cultivative function);
(4) connect people with and educate them about their history and tradition (ritual’s pedagogical function);
(5) provide the means of transition from one period of life or status to another, i.e., rites of passage (enabling ritual to create a new status, an example of its performative function);
(6) allow the appropriate expression of feelings and attitudes (ritual’s expressive function); and
(7) beautify our lives (ritual’s aesthetic function, e.g., turning the act of nourishing ourselves through eating into a beautiful family meal).
[. . .]
Chapter 27 of the Xunzi is filled with examples of particular rituals. They include how one responds when being summoned by one’s superiors, what clothes are appropriate to wear given the occasion and one’s position, how ambassadors should comport themselves when on missions, words to be said at a wedding, the appropriate gifts for different occasions, and even how to stand and where to look when in the presence of others. In chapter 20, there is an extended description of a village wine ceremony.

Is ritual exclusively social? All of Xunzi's examples of rituals are social, rituals must be socially handed down and taught, and one of their main purposes is to sustain social order. It's not conclusive, but if rituals could be individual, it seems strange that Xunzi didn't give a single example of one.

Our minor disagreements

The question of mental horsepower is a very subjective one. The reason I bring it up is the possibility that due to your comment readers would unfairly prefer Xunzi over other authors without knowing much about them. ("Oh, Xunzi? I've heard he's good. Very rationalist. Mencius, Mozi, Laozi? Probably not so good. Vaniver doesn't think they have as much mental horsepower.")

I do think you read Xunzi as a rationalist. My concern was that other readers would read your post and come away with the false implicature that there were no other rationalists in China, or that no-one else considered the possibility of the I Ching being used for perspective-taking. I apologise for the phrasing of "Nothing, except that it stunts true learning." It comes off as overly sarcastic and I've edited it in the original comment.

I was concerned about the mindset of readers more than their likelihood of reading Xunzi. Ignorance is not contagious, but framing is. The impression given by the post is that almost everything Xunzi says about the I Ching was said indirectly by Brian Eno. What reason, then, is there to read Xunzi? My concern is that readers would take the implicature that Xunzi adds essentially nothing new to Eno and so there's no real reason to read Xunzi at all.

Once again, I want to reiterate that I think your post is a net good. I misunderstood what you were aiming for, and your post succeeds very well in steelmanning divination. My criticisms only relate to the parts on interpreting Xunzi. I hope that my comments came across as useful rather than dispiriting, and I hope one day I'll be able to read your posts on Xunzi and Mozi.


This post makes the best case I’ve seen for a steelmanned version of divination. Unfortunately, many of its substantive points are also either wrong or misleading. I’ll start small, by giving the least serious misleading statement. Then I’ll point out two wrong statements along with one extremely misleading statement, and end with why this all matters.

Vaniver's post is good and LW is better for having it. This is intended as a discussion of how it could be made better, and the criticism here should not distract you from how useful the post is.

The smallest misleading statement

Let’s start with the smallest misleading statement. Vaniver gives the throwaway remark early on that he and another rationalist had the impression that “Xunzi simply had a lot more mental horsepower than many other core figures.” That’s because many of the core figures lived before Xunzi, who comes at the end of the classical (pre-imperial) period.

To the extent that Xunzi’s work is richer than theirs, it is richer because Xunzi can build on their work. Xunzi can argue against Mozi and Mencius; Mozi and Mencius can’t argue against Xunzi.

It’s like saying that Derek Parfit had more mental horsepower than Bertrand Russell, because he made finer ethical distinctions. Of course he did. Parfit can take advantage of Russell’s knowledge. Russell can’t take advantage of Parfit’s.

The historical mistakes

To see where the post could be better, let’s see what Vaniver was trying to do. He attempts to answer two questions in this post. The historical question: what did Xunzi mean by rituals, and why did he think they were a good idea? The second, conceptual question: how can we steelman his argument?

To answer these questions, the post makes three substantive claims. The first two are claims of historical interpretation. The third claim is more conceptual.

  • When Xunzi talks about divination, he probably means the I Ching.
  • It’s unclear what reasons Xunzi gives for divination. “Sadly, [figuring out the reasons] was left as an exercise for the reader; the surrounding paragraphs are only vaguely related.”
  • The I Ching can be steelmanned as a version of Oblique Strategies for life.

The first two historical statements are wrong. The third conceptual claim is correct, but is phrased in a way which doesn't allow full learning.

When Xunzi talks about divination, he does not mean the I Ching. Reading Xunzi makes it clear that he thinks of ritual and divination as social activities; Vaniver gets his historical interpretation wrong immediately by assuming an individualistic framework. If the I Ching is used at all, it's in a communal setting. All of Xunzi's other ritual examples are communal.

Xunzi’s reasons for supporting divination rituals are very clear. To quote the SEP article:

. . . rituals, in Xunzi’s conception, not only facilitate social cohesion, but also foster moral and psychological development.

Those are the questions of historical interpretation that Vaniver gets wrong.

The misleading conceptual answer: missing out on real learning

What of Vaniver’s core claim that the I Ching can be seen as steelmanned perspective-taking? That’s certainly true. As I said, this post makes an excellent argument for it.

This exact point was made seventeen centuries earlier by Wang Bi’s commentary on the I Ching, which Vaniver does not mention. Tze-Ki Hon, writing on “Wang Bi's sustained efforts to obliterate the fortune-telling aspects of the hexagrams” in his article “Human Agency and Change: A Reading of Wang Bi’s Yijing Commentary,” points out [pdf, p. 238]:

By stressing human agency and activism in his reading of hexagrams, it is clear that Wang Bi sees the Yijing as a series of metaphors. If indeed human beings have to constantly find the optimum balance between the demands of their surroundings and their own needs, then the purpose of studying the Yijing is to expand one’s horizons so as to be at ease with changes. Reading the Yijing becomes an occasion to develop a mental picture of one's surroundings, such that one finds out the opportunities and limitations in a given situation.

I doubt Vaniver would find anything to disagree with in the above paragraph.

The problem with Vaniver not mentioning Wang Bi is that it prevents full learning. The unfortunate outcome is that even though Vaniver is reading a text from a different time and culture, very little is learned. By reading Xunzi as Brian Eno, creative visionary and co-inventor of Oblique Strategies, he misses out on Xunzi as Xunzi, thinker on self-cultivation and architect of social order.

Reading the I Ching as Oblique Strategies gives him nothing new, as he already knew about Oblique Strategies beforehand; he ends up with what he started. Reading the I Ching as the I Ching gives him Wang Bi, philosophical perspectives on self-cultivation and living in an uncertain world, historical depth, awe at the line of commentators who found so much to interpret, and Oblique Strategies if he wanted it.

(Incidentally, there is also a missed chance to mention other applications of the Yijing, as seen in Kidder Smith Jr., Peter K. Bol, Joseph A. Adler and Don J. Wyatt’s book Sung Dynasty Uses of the I Ching. Just as Xunzi is richer for having come after Mencius, Zhu Xi is richer for having come after Wang Bi.)

The unfortunate consequences

There’s nothing wrong in Vaniver having come up with his interpretation on his own. The real problem comes about when people with no knowledge of Chinese philosophy read Vaniver’s post and come away with wrong views. They---being acute readers---will likely read the false implicatures facilitated by his post and come away with the impression that Xunzi’s rituals work on an individual level, that other thinkers are less worth reading than Xunzi, and that in ancient China the I Ching was mostly read as a divination manual rather than a tool for communal change or a series of metaphors.

This last one is worrying. There’s the mindset that these “old ways” made sense from a rational perspective which those in the past missed, and now here we are, the rationalists, elucidating why they worked. All in all, it’s an excellent way to miss out on reading the past honestly on its own terms and to not realise that there were rationalists in the past as well. (It is also quite a patronising way to read past texts, as though people now have a monopoly on rational thinking!)

This post would have been richer if he had emphasised the social aspect of divination and brought in some of Wang Bi’s specific interpretations of the hexagrams for concrete examples of interpreting them as fields of action. Instead, Wang Bi’s wheel is reinvented. With two hours of extra work, the post could have been so much richer. Both the SEP and the IEP are freely available, and I’m not saying anything groundbreaking.

And the shame is that there is a lot that is interesting in Xunzi for rationalists. Xunzi thinks that the hierarchical rituals help to create a more egalitarian society. What kind of hierarchical rituals are allowed? Why? How can we steelman that argument? What implications does it have for institutional design? He thinks that these rituals have certain standards and origins. Is there a concrete way of having metrics for these standards? To what extent can evolutionary explanations account for these rituals?

(For more on rituals, see Daniel A. Bell’s paper “Hierarchical Rituals for Egalitarian Societies” and chapter four, pp. 181--188 of Puett's To Become a God: Cosmology, Sacrifice, and Self-Divinization in Early China.)

There’s a lot in Xunzi that's useful and interesting. Misreading him is a sure way to miss it.

Why this matters

Why does this matter? You might think that the conceptual point is the real one that matters. Who cares about interpretation and scene-setting?

I've already partly answered this above when I talked about how Vaniver's perspective stunts full learning. But here's another way of looking at it. Imagine a post on artificial intelligence which began:

Many people talk about artificial intelligence (AI) nowadays. But what is AI? It's not clear at all. But we can take a stab at it by looking at the etymology of the word “artificial,” from the Latin artificialis, from artificium, meaning handicraft. We can thus conclude that “artificial intelligence” refers to the intelligence possessed by handicrafts---baskets, chairs, and the like.

That’s about how badly the post reads to me. Just as a knowledge of etymology does not automatically enable one to write well about AI, a knowledge of randomisation in rationality does not automatically enable one to write well about Xunzi. The post on artificial intelligence may make cogent points later, but the very way it’s phrased is so misguided as to make it harmful to readers who aren’t familiar with AI.

As I mentioned earlier, Vaniver's post is good and LW is better for having it. I hope this comment adds to making it better and more complete.