I think the most revealing thing I can say about my relationship with Eliezer himself is that on the one occasion we’ve been face to face we were completing each others’ sentences within fifteen minutes of first meeting.
I wonder if Eliezer had the same impression.
I am not myself particularly concerned with the Friendly AI problem. I think it’s good and necessary that others are working on it, and I’m glad at least one of them is as bright as Eliezer. That makes one less thing for me to worry about.
That doesn't make much sense. It seems pretty obvious that, although he's smart, Eliezer isn't capable of solving Friendly AI single-handedly, and the amount of resources being devoted to it is much smaller than anyone who accepts that UFAI is a threat should want.
Edit: Ok, yes, there exist reasons for Eric Raymond not to devote attention to Friendly AI other than Friendly AI not needing more attention devoted to it.
I don't think the quote you've chosen indicates the author believes EY is capable of solving FAI single-handedly.
Unless you think UFAI is a problem so large that all others are irrelevant, there's room for people to work on other problems justifiably.
See this article. At any given time there's likely only one maximally efficient use of your time/energy/money resources on the margin. "Justifiable" sure (I don't know what that word means), but don't call it utility maximization.
He might share my opinion that it's better that smart people work on things that fascinate them and are important than things that might be more important but they find uninteresting.
Eric writes:
One major difference is that I learned techniques corresponding to much of the the Less Wrong analytical method from Alfred Korzybski’s discipline of General Semantics, a very long time ago.
Me too. One disappointment I had after coming here was finding that while Less Wrong might note that The Map Is Not the Territory, there was very little influence here from Korzybski himself, or even from the rest of General Semantics more generally. Which is too bad. It's good stuff, served up in conceptual chunks much like the sequences, and when I talk to people who have had real influence from GS, the level of conceptual sanity is always a cut above. The only example of someone with real knowledge of Korzybski I know of from LW is Richard Kennaway, who proves the rule again. I think even Dan Dennett had some oblique interest.
I remember Eric showing up on a small Stirnerite Egoist mailing list I was very involved in, peddling anarchism if I remember right. I don't think he stayed long. I seem to remember him being elsewhere too. Certainly Libernet. Paul Vixie's list? Maybe Extropians? He did get around. Very funny to have him pop up here too. It's a small world after all.
there was very little influence here from Korzybski himself, or even from the rest of General Semantics more generally. Which is too bad. It's good stuff, served up in conceptual chunks much like the sequences, and when I talk to people who have had real influence from GS, the level of conceptual sanity is always a cut above. The only example of someone with real knowledge of Korzybski I know of from LW is Richard Kenneway, who proves the rule again.
If I were to read a single Korzybski book, which one should I read? Would the answer be different if I wanted to read many, but wanted to know which one to start with?
Science and Sanity is his main work.
But be warned, reading it is likely to take more time than the average book.
I didn't recommend it, because you have to have a pretty high tolerance for ideological windbags to get through it. I do, but it was pushing my limits. EY stopped.
I've pinged Kennaway to see if he has an alternative recommendation.
Recommending a way into Korzybski is difficult. I read Korzybski the hard way, having been inspired to seek out "Science and Sanity" at the age of about 15, by mentions of K and General Semantics in some of Heinlein's stories, and found it was in the public library. (Ah, the Golden Age of Science Fiction -- being 15 years old.) I found it extraordinarily laborious, but about ten years later I read it again and found it as clear as day (albeit a rather long day). I've heard of others having a similar experience. During my earlier exposure I also read Hayakawa, and Stuart Chase's "The Tyranny of Words" and "The Power of Words" (but I think Chase goes overboard on dismissing abstractions as meaningless).
The thing is, though, all of these books are rather dated now. I can read them now and see what I saw in them then, but I question whether they will work for a modern reader. What more modern sources are there for these ideas? The General Semantics movement itself seems to have produced little, although the Institute of General Semantics and the International Society for General Semantics still trundle on. One of them produced a 5th edition of S&S, but it adds little but page count to the 4th.
I'll sound like an Eliezer fanboy, but the best answer I know to the question, "if not Korzybski, then what?" is the Sequences. Some themes of Korzybski:
The distinction between word and thing, the ladder of abstraction. See a tree falls in the forest, words being wrong, and more.
Implications of modern science for understanding reality. If what Eliezer writes on QM is not always quite right, it is not so very wrong. The same could be said of K's knowledge of the science of his time. I judge this not by my own knowledge of the subject (which is small) but the absence of anyone showing up to demolish swathes of it as nonsense. The most contentious thing seems to be whether MWI vs. Copenhagen is as much a slam dunk as EY believes, but I see that issue as only a sideshow anyway.
The inexhaustibility of the physical object, as opposed to our finite ideas of the object. ("More can be said about a single apple than about all the apples in the world." -- EY in "Twelve Virtues")
"The map is not the territory." (A point that I believe was first made in those words by K.) As someone else put it, if you draw a line on the map, two lanes of blacktop do not appear under your feet.
The principle of non-identity: nothing (quantum-mechanical arcana aside) is identical to anything else. When you describe two things with the same words, they are still different things.
And there are some topics in the sequences not covered by K, but which are essential to good thinking:
Causality. To borrow Dobzhansky's dictum on the role of evolution in biology, nothing in science makes sense except in the light of causality. I think it's something that goes under-appreciated on LessWrong, even by Eliezer, despite what he's written on the subject.
Neuroscience. Implications for "free will", "agency", "self-modification", "identity", etc. I think a lot of this will date as much as K's account of "colloidal" mechanisms, but however it turns out (if we knew, we'd be there already) it's what we have right now, and it matters.
Weaknesses of the Sequences:
S&S is copiously referenced; the Sequences sparsely.
Causality is not treated as well as it might be. Too much emphasis on statistics, not enough on mechanisms. (ETA: And it took someone else to point out to Eliezer the distinction between Bayesian graphs and causal graphs.) I would be interested to see Eliezer's response to Judea Pearl's paper "Why I am only a half-Bayesian." And systems of circular causality and the paradoxical phenomena they produce get almost no notice.
An excessive emphasis on Bayesian reasoning as sufficient for salvation, er, solving every problem. I'm quite willing to grant it a fundamental status in principle, and its heuristic usefulness when actual numbers are not and cannot be available. (Which Eliezer himself has said -- this is more a fault of LW than the Sequences.) But until someone comes up with an effective method of finding better models when the probability of the observations given any existing model has fallen to absurdly low levels, talk of "Bayesian superintelligences" is moonshine. I mean "effective" here in both senses: both practical and defined as an algorithm. Such a method would be equivalent to an AGI, so I'm not holding my breath. What does one do in the meantime? There are answers (e.g. Andrew Gelman on model checking), but are they good enough?
I would be interested to see Eliezer's response to Judea Pearl's paper "Why I am only a half-Bayesian."
I believe I saw that at some point, actually--I think I'm thinking of this comment and the associated thread, which unfortunately doesn't have much in a way of a response.
Hm, the context there wasn't originally about causality, which seems to come in as an aside. As far as I can judge from it, Eliezer is saying that causal models are things you can have probability distributions over, to be updated by evidence, just like every other sort of model of reality. They are not a separate magisterium from probability, and there is no need to be only half a Bayesian. That causality is not a statistical concept is of no more significance than that velocity is not a statistical concept.
It's not clear to me what he intended by mentioning Pearl's paper. As far as I know, Pearl nowhere considers probability distributions over causal models. He considers them only from the point of view of answering two-valued questions such as "Are these data consistent with this causal model?" or "Given this causal model, do these observations of some of its variables allow estimation of the magnitude of these causal effects?" He does not study questions such as "what causal model is most likely, given these data?" Some do, although as yet that seems to be limited to one group of researchers in neuroscience.
I understand your concern. Science and Sanity isn't an easy book to read.
On the homepage of the institute for General Semantics there a list of introductionary books: http://www.generalsemantics.org/store/53-general-semantics-intros?n=20&id_category=53
Maybe one of those would be better.
Yes. I realize that's saying a lot. But I bet you a nickel he's actually worse than you're thinking now.
Probably not for everyone on LW but for a good portion, yes. He wrote 80 years ago. He's no logical positivist.
Another quote (page 16):
In 1933, scientific opinion is divided as to whether we need more science or less science. Some prominent men even suggest that scientists should take a vacation and let the rest of mankind catch up with their acievements. There seems no doubt that the discrepancy betweeen human adjustments and the advances of science is becoming alarming. Is, then, such a suggestion justified?
The answer he finds is "no" but questions like that can irritate people on LW.
Any idea if this Kindle book is an acceptable substitute for the unabridged version? It appears to be a third of the length, but I wanted to read this on an airplane and historically 900 page books have been moderately troublesome for that.
The problem is not only that it's 900 pages but that different jargon get's introduced at different times in the book. One of the prefaces suggest to read the book once to get the general idea of the jargon and then read it another time to understand the arguments
It contains nice sentences like:
I, personally, have no doubt that this would mark the beginning of a new era, the scientific era, in which all human desirable characteristics would be released from the present animalistic, psychophysiological A-sematic blockages, and that sanity would prevail.
Yeah, I think this is a good example of the most annoying aspect - the endless claims of saving the universe awash in seas of jargon. An ideological infomercial that goes on and on and on. Just give it a rest.
Those sentences often do make points of how a specific point he makes relates to the overall whole. It's just often not obvious at first glance what he's trying to point to.
Another nice quote:
Many of the structural points are of genuine impotance and interest to professional scientists, teachers, and others, who seldom, if ever, deal with such structural, linguistic, and semantics problems as are here analysed. The layman who will read the book diligently and repeatedly, without skipping any part of it, will get at least a feeling or vague notion that such problems do exist, which will produce a very important psycho-logical effect or release from the old animalistic unconditionality of responses, whether or not he feels that he has 'understood' them fully.
My earnest suggestion, backed by experience, to the reader is to read the book through several times, but not to dwell on points which are not clear to him. At each reading the issues will become clearer, until they will become entirely his own. Superficial reading of the book is to be positevely discouraged, as it will prove to be so much time wasted.
A few paragraphs later:
The diffculty in passing from the old system to another of different structure is not in the non-aristotelian system as such, which is really simpler and more in accord with commmon sense: but the serious diffculty lies rather in the older habits of speech and nervous responses, and in the older semantic reactions which must be overcome. These difficulties are, perhaps, more serious than is generally realized.
Alfred Korzybyski
So, long ago I was rather put off Korzybyski and General Semantics by the brief discussion of them in Martin Gardner's "Fads and Fallacies in the Name of Science". A few sample quotations:
Neither movement [sc. general semantics and psychodrama, which Gardner happens to put in the same chapter -- gjm], it should be stated, approaches the absurdity of the two previously considered cults [sc. orgonomy and dianetics -- gjm]. For this reason, general semantics and psychodrama must be regarded as controversial, borderline examples, which may or may not have considerable scientific merit.
[Science and Sanity] is a poorly organized, verbose, philosophically naive, repetitious mish-mash of sound ideas borrowed from abler scientists and philosophers, mixed with neologisms, confused ideas, unconscious metaphysics, and highly dubious speculations and neurology and psychiatric therapy.
Korzybyski's explanation of why non-Aristotelian thinking has therapeutic body effects was bound up with a theory now discarded by his followers as neurologically unsound. It concerned the cortex and the thalamus. [...]
The simple reason is that Korzybyski made no contributions of significance to any of the fields about which he wrote with such seeming erudition. Most of the Count's followers admit this, but insist that the value of his work lies in the fact that it was the first great synthesis of modern scientific philosophy and psychiatry.
The impression I get from Gardner is that "the parts that are good are not original, and the parts that are original are not good". So, e.g., Korzybyski is of course right to say that black-and-white dichotomous thinking is harmful, and that it's useful to distinguish between a thing and a description of that thing, and that "is" can be a treacherous word. But we didn't need Korzybyski to tell us that. And much of what's distinctive in Korzybyski (again, I'm describing the impression I get from people like Gardner) is just plain wrong.
Is that all wrong? Is there substantially more to Korzybyski and "General Semantics"? Has present-day GS filtered out all the pseudoscience while preserving (or, better, newly finding) a lot of insight? Are there insights there that are unique to GS?
For what it's worth, I got a lot of good from Chase's The Tyranny of Words-- basically the idea that general statements are typically based on much less information than they imply-- and when I read Science and Sanity, there didn't seem to be much that wasn't covered by the Chase book.
On the charges of neologisms for old ideas and the repetitive claims, my recollection is that those were intentional moves on Korzybski's part.
The repetition was to develop semantic habits - to actually change people. The neologisms were for creating new semantic habits, without carrying the semantic baggage of the old terms for the "same" ideas into the new.
Korzybyski wrote Science and Sanity in 1933 and as such specific things he says about the brain turned out to be false. I don't think you will find a single person who wrote at the time who made no mistakes. Technology such as fMRI did allow us to discard a lot of ideas about the brain as wrong.
Is that all wrong? Is there substantially more to Korzybyski and "General Semantics"?
A bit above you mention that "is" can be a treacherous word. Yet you started both of those question with it. From the General Semantics perspective your insight that "is" is can be a treacherous word didn't help you.
It's a bit like teaching someone Bayes theorem. It's quite easy to teach it in a way that someone can use it in textbook problems but it's quite another thing to teach it in a way that someone actually uses it in daily life.
Someone well trained in General Semantics and who uses that framework as his default way of reasoning about the world wouldn't ask "Is that all wrong?". Discussion whether the rejection of that question is pseudoscience bring you again near the space of what Korzybyski calls flawed Aristotelian reasoning.
Of course when I speak about that way and ask whether or not your statement is Aristotelian in nature, I'm also using binary categories. That makes it easier to make my point, but at the same time I'm not moving inside the rules of the General Semantics framework. In Science and Sanity Korzybyski mostly tries to present the arguments for General Semanitcs in it's own logic. That makes the book very hard to read.
As far as the charge of being poorly organized and verbose goes, it comes from the book being hard to understand. Korxybyski didn't write in a way that's easily accessible. If he would have the world would likely look different.
As far as the charge of not having influenced anybody goes, Albert Ellis who developed Rational emotive behavior therapy (REBT) with is one of of Cognitive Behavior Therapy does say that he was influenced by several of Korzybyski's ideas when developen REBT.
He has also influenced a lot of other important people. There's a list at Wikipedia.
I don't think you will find a single person who wrote at the time who made no mistakes.
Nor am I looking for one. I'm extremely uninterested in questions like "was Korzybyski a very clever person?" and much more interested in ones like "does the stuff Korzybyski wrote contain enough insights I currently lack, and little enough distracting wrongness, to be worth my while reading?".
A bit above you mention that "is" can be a treacherous word. Yet you started both of those questions with it.
Yup. There is no valid inference from "X can be treacherous" to "one should completely avoid X". If Korzybyski claimed that one should simply never use "is" (I am doubtful that he did, as it happens[1]) then I disagree with him, which is not the same thing as failing to understand him or failing to have assimilated whatever useful points he may have made.
In the present case, note that "Is that all wrong?" is equivalent to "Does all that have the property of wrongness?" which doesn't (either implicitly or explicitly) contain an "is". (It arguably has the weakness of assuming that everything can be classified as "wrong" or "not wrong", but use of such language is a very useful shorthand; it would be intolerably cumbersome always to have to say things like "Does all that have the property of wrongness to such an extent that I should be regretting having written it?". And "Is there substantially more ..." is equivalent to something like "Does Korzybyski's work contain substantially more useful insight than what I wrote above would suggest?". Again, no "is" contained or implied.
Now, if you think I made an actual error as a result of writing in English rather than in E-Prime, that I'd have avoided if I had more of Korzybyski's ideas in my brain, go ahead and show me; but so far all you've done is to observe that my writing doesn't conform to some rules associated with him (I'm not sure whether they're his or his followers') which I currently see no sufficient reason to endorse.
As far as the charge of being poorly organized and verbose goes, it comes from the book being hard to understand.
It seems somewhat plausible that the causation might go the other way. Are you claiming, specifically, that in fact K's book is neither poorly organized nor verbose, but that because K didn't write it accessibly Gardner misunderstood it and this led him to mischaracterize it as poorly organized and verbose?
As far as the charge of not having influenced anybody goes [...]
I didn't see that charge. I saw a charge of not having contributed anything novel to human knowledge and understanding. Not being very influential (if indeed K wasn't) might be a consequence of that, though it's not hard to find very influential people whose influence didn't mostly come from such contributions.
I wonder what led you to see a "charge of not having influenced anybody" in what I wrote. It seems to me a rather curious error.
[1] My hazy recollection is that Korzybyski and his followers have no particular objection to "is" when used merely to apply an adjective to a noun: "the cat is blue", but that they object to using it in contexts that assert, or seem to assert, the identity of two different things ("that man is a liar", maybe).
but so far all you've done is to observe that my writing doesn't conform to some rules associated with him (I'm not sure whether they're his or his followers') which I currently see no sufficient reason to endorse.
Yes, I'm not trying to demonstrate that something is true but I'm trying to illustrate what Korzybyski's ideas are about.
In the present case, note that "Is that all wrong?" is equivalent to "Does all that have the property of wrongness?" which doesn't (either implicitly or explicitly) contain an "is".
The issue is more that the word is appears. Right/Wrong is black-and-white dichotomous thinking. Maps are not right or wrong but have degrees of accuracy in mapping certain teritory and usefulness for navigating the territory.
You will find somewhere a claim that humans have rougly 12 billion neurons. That happens to be clearly wrong and humans have more neurons but it doesn't matter much to the main thesis and it's not what makes reading the book hard.
Most sentence of the book are not clearly right or wrong. Mostly what's driven the kind of criticism that Martin Gardner gives is not that Korzybyski says thinks that are easily shown to be wrong but that Korzybyski says things, where it's not clear what point Korzybyski tries to make.
It's a bit like reading a Zen Koan. You don't ask yourself: "Is this claim wrong?" but "What is the author trying to tell me?"
I wonder what led you to see a "charge of not having influenced anybody" in what I wrote. It seems to me a rather curious error.
I refer to the quote "The simple reason is that Korzybyski made no contributions of significance to any of the fields about which he wrote with such seeming erudition".
degrees of accuracy
Yes, I already addressed this.
says things where it's not clear what point Korzybyski tries to make
This sort of writing is only worth the effort of engaging with if there's enough useful or interesting content there to warrant it. So we come back to my question: is there more there than if think from e.g. what Gardner wrote? This still isn't clear to me.
I refer to the quote [...]
... which doesn't say anything about whether he influenced anybody. At least, not when interpreted the way I meant it. Do you define "making a contribution of significance" to mean"influencing people"?
Has present-day GS filtered out all the pseudoscience while preserving (or, better, newly finding) a lot of insight? Are there insights there that are unique to GS?
I would be more suspicious of a work along these lines that was full of new insight. Science and Sanity is a work of synthesis, not of discovery, and sometimes we do need someone to put it all together and tell us what, once we hear it, we can easily dismiss as "but we knew all that anyway". "What is new is not good, and what is good is not new" is a misdirected complaint when made against a work of this sort. What is new in "The God Delusion" or "The Selfish Gene"? Only the presentation of those syntheses to the masses. Ask rather, what is old, but seen anew?
OK, I'm asking. (I have no objection to pop science, pop psychology, pop philosophy, etc., when done well.) What is "old, but seen anew"? Specifically: suppose me to be a longstanding LW participant, reasonably well read in science and philosophy; if I read (say) Science and Sanity or some later GS work, am I likely to come out the far end knowing more or thinking better, and if so what do you expect me to learn?
Specifically: suppose me to be a longstanding LW participant, reasonably well read in science and philosophy
Then reading S&S may well be supererogatory. If you've read all of the Sequences, you will recognise much in S&S, just as I, having read S&S, recognised much in the Sequences.
The impression I get from Gardner is that "the parts that are good are not original, and the parts that are original are not good".
So what does that make the LW sequences?
Obviously it doesn't make them anything. But I have heard similar criticisms levelled at them.
My own impression of the Sequences is that most of what they say is fairly standard-issue analytic philosophy / cognitive science / physics / whatever; that where they're novel they're right more often than (according to what I've read, which I repeat is rather little and I have no reason to think it very reliable) Korzybyski is when he is novel; and that there's very little in them that's just straightforwardly wrong as (again, according to what I've read) some key bits of Korzybyski are.
But what seems to me most useful in the Sequences is that they bring together a wide-ranging body of ideas that, despite involving quite a lot of philosophy, are generally not wrong and are selected with good taste. Works of philosophy usually tend (as it seems to me) to be either narrow or frequently wrong. And by and large they're clearly and engagingly written.
Perhaps fans of General Semantics would say the same about, say, Science and Sanity or later GS works. If they can correctly say so then that would be a reason to adjust my view of Korzybyski and/or his later followers. That's why one of the questions I asked was: "Has present-day GS filtered out all the pseudoscience while preserving [...] a lot of insight?" to which the answer could be yes even if there's very little original in GS.
From a discussion maybe a year ago, I recall him saying he gave up on Korzybski, but read Hayakawa, though I don't remember him characterizing how much he considered himself influenced by Hayakawa.
It's been over a decade now, but my recollection is that Hayakawa was predominantly focused on language, and left out much of the most interesting and illuminating stuff. Basically, he's tagged in my head as "popularizer, relative lightweight" in relation to Korzybski.
ESR is the one who got me into HPMOR, actually - the funny thing is, I'm not a hacker, I read him more for the politics and board games, but I find the computer stuff nifty(if opaque).
http://esr.ibiblio.org/?p=6549