There are two things I’d like to do with your work that I think you may have a reasonable problem with.

The first is choosing how to present it. If I were to print TNC, I’d have to decide how to typeset it (especially those forum sections!) and what the cover would look like. The only part that would be yours is the text. I don’t know how I’d feel about letting someone do that to my Internet scribbles, but it was kind of implied in my ask, and it doesn’t seem like it fazes you—still, if you want to talk about what I have in mind, my DMs are open. Or I can just make it as vanilla as possible if you like.

Second, the reason I’d like to print TNC is because the college I’ve committed to has this thing where incoming freshmen add books to the library, and, well. I feel like TNC expresses an important part of my past, I guess, this sort of lonely modern intellectual craziness that I’ve never really seen in print, and I wanted to bring a bit of that with me because I’m not really over it yet. My copy of TNC would remain there indefinitely for students to find and borrow—with the AO3 URL printed inside, of course. This is the part I figured you’d probably say no to, but I thought it was worth asking.

And when it comes to selling, I feel I should note… I don’t know what they’ll end up doing with the collection decades down the line when the institution inevitably ceases to exist. I can talk to them about that if it matters.

Sorry for the very late reply – hope it isn’t too late!

This actually sounds really cool and I don’t have any problems with the basic concept.  A few things:

When you say “add books to the library,” do most people just donate an ordinary book they own?  If so, what do you know about the precedent (if any) for adding a book you’ve printed and binding?  If the book doesn’t exist anywhere else, will they still be able to enter it into their catalogue, assign it a call number, and all that?

I’d like to review some digital representation or mock-up of whatever you’re going to print before you print it.

I’d prefer that the book contain my real name (rather than “nostalgebraist”), and be shelved under my real name.  I can send it to you in a PM once you’re further along (I’d imagine it can be dropped in at the end of the design process as it would only appear once or twice).

I have a strong preference that the forum sections appear as close to the web versions as possible, with the exception of changes in the color scheme to make it readable in grayscale (or, if printed in color, readable on paper).  Since book pages are a lot narrower than the average monitor, this may mean the column with the post text will get really narrow and the posts will look really long, which is fine.  Wrapping individual posts across page breaks is also fine, even if it looks bad.  Basically, if it looks like someone has just hit “print” on their web browser, that’s actually appropriate and desirable.

Keep me posted!

I wish you’d published TNC, even self-publish on Amazon. I think I asked you about it a year or two ago in a reblog and you said the forum aesthetic wouldn’t transfer well to an epub or something like that. Maybe this is a good time to ask again if you’ve thought about it since.

Also, have you seriously tried and verified that the forum part of TNC can’t be presented well in the usual ebook formats, or is it just a hunch? I’m a big enough fan of it to possibly devote some time to mucking about with authoring tools I don’t know, to see if it can be done without much pain, if that’s the major thing preventing you from publishing it.

Going through an old notes file I’d forgotten about.   Found a part where I seem to be noting down things I didn’t like in Anna Karenina in preparation for some never-written screed:

Awkward writing seems much more likely to be a criticism of a bad translation, because that is not an especially common criticism of Lev Tolstoy.

I eventually wrote a more positive if kind of contentless review, in which I blamed the awkward writing on the translation.  Although years later I got an interesting comment which made me less sure of that.

Pevear&Volokhonsky translations of Russian classics (the AK you read is one of them) are hugely controversial among literary critics, especially those who can read the original Russian. I harbor an intense dislike towards them myself, ever since I tried to compare different translations of Bulgakov’s “The Master and Margarita” into English many years ago and was appalled by P&V’s choices. The final comment to your goodreads review links to one article critical of P&V translations, but in my opinion, the objections raised there (all of them can be categorized as “tin ear” problems) aren’t the most important ones. 

The chief problem with P&V translations is that they routinely take ordinary Russian idioms and translate them word-by-word into English, and they then become weird-looking examples of what the reader can only surmise must be the artistic expression of the author’s style. This is quite intolerable to look at if you know both languages well. In Russian we say “spit on smth.” to mean “nevermind smth., don’t worry about smth.”, this is a very common phrase that doesn’t really suggest an image of spitting (unless subverted or played with in humor or poetry of course). So in a P&V translation, instead of “nevermind the rug” [i.e. don’t take your shoes off, said to a guest] the host says “spit on the rug”, which fails to carry the intended meaning. 

There are numerous examples of this. Suppose you meet somebody you know and you notice something, like a bandage maybe, that makes you want to ask if they’d visited a doctor. If you want to emphasize your surprise, you might ask “Wait... have you been to a doctor?” In Russian, the colloquial device to produce the same feeling of surprise is to insert a [what] inside the phrase, with a particular intonation. So literally it becomes “you’ve, what, been to a doctor?” which is how P&V render it. A completely natural colloquial Russian sentence becomes, at best, stilted and awkward English that doesn’t fit the situation. More examples, and a more eloquent restatement of the principal objection, in Timothy D. Sergay’s long comment in this old NYTimes blog discussion.

I’d guess that most of your misgivings about the style of A&K were due to the unfortunate choice of a translation.

Testing a hypothesis: please reply (without checking Wikipedia) how many people you think are in the KKK.

I’d expect that KKK is ill-defined and there’re various “chapters” in different state of conflict and such, but the most coherent/populous/non-trivial cluster of humans calling itself KKK numbers around 30K people.

You wanted to say “nothing says that 0 is the only number that’s not a successor” (no number is without a successor). But that’s not right either. 0 is in fact the only number that’s not a successor, and this follows by the induction axiom. Simply consider the statement “if x is not 0, there’s y such that S(y)=x”, and apply induction.  Nonstandard models of PA are weird, but they’re weird in other ways. They all look like N followed by an infinite ordered chain of Zs. The members of the initial N are the “standard” numbers, and of all the Zs above it are “nonstandard”. Only one element, the 0 of the N, is not a successor; all the 0s of the Zs are successors to their -1s.

(you can of course define Peano Axioms to not include induction, only the properties of equality and S. But that gives a theory that’s spectacularly weak and not useful to anything - e.g. can’t define addition or prove it commutative - and that’s not what’s normally meant by PA or models of PA).

True, but it’s important also to realize that nonstandard models exist not only “because” PA is incomplete and there are statements that are neither provable nor disprovable in it. Even if PA was a complete theory, there’d still be nonstandard models of it, in the sense of not being isomorphic to N. First-order logic is just too weak to capture and fix the exact structure of a model, except in the trivial (well, in this context) case of a finite model.

I hesistate a bit too much to talk about this, since I’m not at all an expert in logic or model theory.

But I think the best way to understand model theory and the idea of “theorems which are independent of an axiomatization” is to look at much simpler axiom systems.

Keep reading

I don’t really follow the context of why you posted this, so maybe my remark is irrelevant, but treating this post as a thing unto its own –

This still isn’t elementary enough; to really explain it, you need to push harder, right through things that are blindingly obvious to you. Accepting that the symmetric group satisfies the axioms but not S, why do you conclude that S is not provable?

Because if a group satisfies axioms, it also satisfies any statement we can prove with those axioms. This is blindingly obvious to a mathematician, but it needs to be explicitly stated, considered and thought through in a logic course. It’s a soundness theorem. Why do we have soundness? Because we built our rules of proof, our thinking apparatus in such a way as to preserve truth, because we’re interested in things that are true. Even if we work purely in formal logic, syntactically, without considering models at all - or before models were invented as an explicit mathematical structure - we use logical axioms and rules of inference that preserve truth and not others. And this is far from a settled ground, cf. the excluded middle and intuitionism.

But to really convince ourselves that our methods of proof are sound, we need something like formal logic and model theory - a formalizable notion of truth and satisfiability. The need becomes acute when we get serious about expressing real mathematical proof as logical proofs, and not merely slinging syllogisms around. While we remain in propositional logic, we’re still doing “model theory” of a kind naively, when for example we explain to others that the correct conclusion from P->Q is not-Q -> not-P, rather than the all-too-frequently mistaken not-P -> not-Q. But that is easy to explain with a few helpful handwaves, or at most a Venn diagram. When quantifiers come in, we end up needing real model theory grounded in some sort of set theory, even for something as obvious as the basic soundness of our proof methods. In the 19th century they kept arguing whether “there exists that X” necessarily entailed that there exists at least something, that is that the empty model didn’t satisfy the existential quantifier, but did satisfy the universal one; some people thought it should be the other way around. Lewis Carroll has a few pages devoted to this in his mathematical book. We sorted this out, did the legwork, formalized math, wrote Principia Mathematica, so that now it’s easy and blindingly obvious to us that 

my colleague and I each have clients that know each other and are close friends, having met in foster care.

my colleague’s client is… how to put this… having a bad time. With the team that runs their case. This is the harlow’s monkey kid. The kid has been put in a group home with no case plan or treatment plan, and the team has said they’re just going to stay there “until.” While there, they keep fiddling with the kid’s meds, and the kid keeps having freak outs that they blame on the kid instead of the meds. While there, the kid is being forced to go through trauma therapy. And while there, the kid is not allowed to have contact with any person they know in the world outside of their team members. So they’re doped up, being observed at every moment for any possible “wrong” move, being forced to relive their worst memories, and then returning to a facility where they know nobody and can speak to nobody who knows them.

We have tried pleading with the team to let this client and my client visit, or have a phone call, or have a sleepover at a local foster home that has offered to host them both for a few days so they can be together. They have refused. They have told us this child does not have friends, not really, because they have RAD, and so their attachments are all superficial, and if they tell us this person is their best friend, they’re just being manipulative. They’ll have a new best friend next week, their caseworker told us.

So every time I visit my kiddo, I have them record a video of themselves for their friend on my smartphone. My coworker brings that to the kiddo, and they record a video back.

The videos are the sweetest, most heartbreaking things on the planet. These kids miss each other so much. But they both have RAD, so apparently it’s not possible that they could miss each other. It’s not possible they could derive some comfort from speaking to each other.

Never in my life did I think social work would involve having to perform a series of cloak and dagger moves to sneak a child a thirty second video of their best friend saying, “I love you! You’re going to be okay! You deserve a family! Please don’t give up!”

I just came upon this blog post from May 2016 and it’s the single most pro-Yudkowskian thing I’ve seen in ages.  I’m not sure why I’m posting about this except that it startled me; I’m used to thinking that these attitudes are rare these days even inside the “rationalist community.”

I don’t want this to come off as “lol this guy’s style is soooo 2009, how embarrassing.”  It’s just that … that post expresses a lot of the attitudes that I used to criticize a lot using the #big yud tag, and I’m realizing that I literally haven’t seen those attitudes in a long time

Very fresh, I love it. I couldn’t have written something like that and in a way that saddens me a little.

Thinking about it, over the years I’m pretty sure I wrote many more words contra Yudkowsky and stuff in the sequence than pro them. But if I knew someone who was interested in, you know, being less wrong, and never heard of LW or anti-LW, and if their choice was between reading the sequences and reading lots and lots of smart and sarcastic anti-LW writings (people like Hallquist, Kruel, Chapman, you, su3su2u1, etc.), I’d tell them to go to the sequence in a heartbeat. Now that’s certainly a false dilemma, one can do both, but I think one is much more likely to do both starting from the sequences. I suspect many, many people who come into the “rationality community” of today end up not reading any nontrivial amount of the sequences. There’s nothing terribly wrong with that, they’re just poorer for it, that’s all. They’re not getting to read the things and fiercely engage with them on their own, find their own disagreements. OP’s zeal in trying to nudge people from that path and towards reading the sequences is commendable, and is more important than the excesses of cool-aid that leak through the post.

Weird hypothesis: astrological sign correlated much more strongly with personality, back a few hundred years ago when type and availability of food depended heavily on the time of year.

I’ve thought of basically this same hypothesis before, actually. It seems fairly plausible, though I don’t know how one would go about researching it.

I’m not sure how to test our specific astrological system (since the time-of-year personality effects might depend on the specific region where you live), but one could presumably take a third-world culture, give everyone personality tests, and see if there were any correlates with birthday, right?

(Do WEIRD personality tests work on third-world people?  Is the Big 5 factor model a good description of people in other parts of the world too?)

This is almost certainly not true, because modern astrology - the kind where it all depends what zodiac sign you were born under - is only about eighty years old.

Ancient astrology was really complicated and mostly involved the position of the planets rather than the sun. It also dealt with cycles other than the annual - what hour of the day you were born, what day of the month, et cetera. The position of the sun in the zodiac signs was a tiny or minimal part of this.

In the 1930s, newspaper astrology columns caught on, and sun signs became popular as something really simple for everyone to calculate and which could apply to broad groups of people (eg you only needed twelve of them to cover everybody). This was only marginally compatible with any preceding astrological tradition.

So no, it’s unlikely that sun signs track some useful meaning from back in the past.

There are effects based on what season you’re born in - schizophrenia is more common in babies born in spring, probably because there are more infections during winter and that means more chance for a pathogen to screw up some crucial window in utero. And great hockey players are way more likely to have been born in the winter, for silly reasons. But none of these have big effects or let you make useful predictions.

I would not have guessed the sun sign forecasts were only 80 years old - thank you! This is amazing.

I found this link with a fascinating detailed history of how that came about - the story of the modern invention of both sun sign forecasts (”Look to your finances this week”) and sun sign delineations (”Scorpions are stubborn”):

Sun sign columns History, validity, and an armchair invitation

Look to your right. The first object you see is the object your clan has banned permanently.

my clan is a terrible quiltless dystopia

A copy of Godel, Escher, Bach.

Recursion is of the devil.

Sign me up for membership in your clan, then!

you don’t like geb?

I don’t. I kind of owe a big post on this–I meant to write one up when I finished, and then I stopped halfway through and I should probably admit to myself that the book annoys me too much for me to finish it.

Some of my objections are unfair–this list includes “Of course, anyone who’s taken a graduate level class in computability theory knows that,” which is rather missing the point and the target audience of the book.

But it glosses over a lot of subtleties that I think are very important to the argument Hofstadter is trying to make–or rather, that would undermine it somewhat. And the presentation seems overly glib, in a way that makes it more compelling as evangelism and as inspiration, but less accurate and tethered to reality.

I think I have sort of the same reaction to it that su3 had to Yudkowsky’s writings, honestly. “Oh cool, someone’s writing popular essays about that! That’s nifty. Well, I guess that bit is technically accurate. I suppose you could say this other thing, but it’s kind of misleading. Oh, but here’s a really nice bit of explanation. Wait, whoa, now he’s missing the point completely!”

See also my comment the other day that if you’re discussing formal logic and you use the word “true”, something has gone seriously wrong. Hofstadter and GEB are major offenders here.

aw ): It’s been a while since I read it, but I really love it. Most of the ideas discussed in the book were new to me at the time and very relevant to my interests and among other things, the book helped me grok the concept of proofs and like… mathematical rigor in general i guess. it’s also pretty fun to read.

tbh i dont really remember the bits where Hofstadter tried to actually make arguments, so I’m assuming they are as bad as you say they are.

w.r.t the “true” comment, my best guess is you mean that “true” is just “provable in some system” and that Hofstadter mistakenly talks as if there is an system-independent notion of mathematical truth? or do you mean something else?

I picked it up in part because @collapsethewavefunction was very enthusiastic about it. She said much the same thing you just did. (In her defense, she also told me I wouldn’t like it nearly as much as she had because I’d already been exposed to a bunch of those ideas before, and “it won’t blow your mind like it did for me when I was young.”)

Hofstadter is very much attempting to explain mathematical rigor and recursion in order to explain consciousness and how you should think. (again, Yudkowskian echoes, although I trust that Hofstadter actually understands most of the details far more than I trust Yudkowsky).

For the last bit, you have it basically right. A lot of popularizations of formal logic use the word “true” as a shortcut, and it never makes things clearer. Examples are rendering Godel’s theorem as “a sufficiently complex consistent system contains at least one statement that is true but unproveable”, or the rendering of Lob’s theorem as “If you can prove that a proof of p would imply that p is true, then p is proveable”.

Hofstadter makes this confusion essentially everywhere, and while some of that can be chalked up to the difficulties of popularization, the idea of “true” is actually pretty critical to his argument, and thus I get annoyed every time he does it.

It’s been a while since I thought about this, but… I’m not convinced. I think the following two claims do not contradict each other, and I support both:

1. You don’t really understand the full generality of Godel’s First Theorem until you learn and grok the purely syntactic version of the statement and proof, one that doesn’t use the notion of truth. You certainly need this to understand the Second Theorem in any but the most hand-wavy way.

2. To someone who didn’t study formal first-order logic, the “naive statement” of Godel’s First that uses goes “a sufficiently complex consistent system contains at least one statement that is true but unproveable” is a better and more clear way to explain the gist of the theorem than a purely syntactic formulation.

The benefit of the “naive statement” is that it glosses over things that such a reader has no grasp of, and focuses on the things they have an intuition for:

- it glosses over “recursively axiomatized system”, which is hard to explain without going into theory of computation

- readers who didn’t study logic do not appreciate the difference between “unsound” and “inconsistent”. “To prove false statements” and “to prove contradictions” sounds really really similar to them. As axioms go, why would you pick an axiom that is a false statement? This sounds very confusing, and they’re just not trained to think of it as a possibility. The sharp demarcation line between syntax and semantics is not there.

- on the other hand, everyone with >passing interest in math has an intuitive grasp of a “true statement about integers”, like 2+2=4, or one of those famous theorems. So the “naive statement” above does seem shockingly profound and clear: if you try to add more and more true axioms, you’re still never going to be able to prove all true theorems. There’s a limit to what we can prove etc. etc.

Compare this to the syntactic formulation like “a sufficiently complex recursively axiomatized consistent system is incomplete, that is, there’s a statement P it neither proves nor disproves”. To the naive reader discussed above, this just doesn’t connect to anything intuitive in their head. It doesn’t seem apriori important that for every P we must either prove or disprove it. They don’t have the training that lets them immediately grasp e.g. that this means (assuming we have a canonical model) “one of P, not-P is true but unprovable”. And if you don’t have a canonical model, like with set theory, completeness just isn’t that intuitively desirable. 

@slatestarscratchpad

Curious what you think of this claim that placebo effects are generally quite small

http://www.dcscience.net/2015/12/11/placebo-effects-are-weak-regression-to-the-mean-is-the-main-reason-ineffective-treatments-appear-to-work/

This seems like an important fact if true.

Find it plausible though more research/discussion needed. See this post, search “What if there were actually no such thing as”

What if the real placebo was the regressions we made along the way?

Oh, while I’m on this sort of thing – does anyone know about AI models of concept formation?  Like, have they been developing along with the rest of AI?  Are there current ones doing cool things the way Deep Dream and neural-style are doing cool visual perception things?

I ask partially because a while ago I claimed rashly that the current enthusiasm for deep learning would plateau eventually, because deep learning was just learning to mimic the sort of sequential abstraction that goes on in relatively early sensory perception (except without top-down input from more abstract/conceptual parts of the brain!).  So it would produce results that seemed very human-like in the sensory realms, without ever getting to the point of being able to abstract beyond “whorls like in Starry Night” or (in the case of that automated NSFW flagger) “things that look like butts” to the kinds of higher- and higher-tier concepts we deal in.

Concept formation in particular seems to me like it’s somehow “especially unsupervised” (I’m sorry if this is naive or doesn’t make sense, I am very tired rn).  Ask a neural network to look for dogs and it will gain an impressively high-level and robust sense of the visual appearance of dogs.  But no human comes into the world knowing that “dogs” are a thing they are looking for.  They just get the buzzing, blooming confusion, and are able to somehow infer that it is useful to lump certain roiling patterns of visual data together with certain auditory and tactile patterns (and higher-level induced patterns of animal behavior) together into this new concept, “dog,” without ever being pointed in that direction.  Meanwhile they do not invent, or do not maintain for long, “non-natural” classification systems like the ones in Borges’ Emporium.  I have never seen an AI that can do something like this.

In the bullshit amateur neuroscience district of my mind, it feels like there is a connection between: the “especially unsupervised” nature of concept formation, the lack of success (?) in designing artificial neural networks for concept formation, and the way in which (I think?) neural processes after early sensory processing get more and more interconnected and less separable into discrete stages (point 5 in this old post).  The amazing “style representations” that let you make anything look like van Gogh or Kandinsky (you may begin to notice how shallow my reading has been, how the reference points repeat) were obtained through a certain kind of averaging over the things “seen” by a visual classification network – and that network got supervised training on labelled data, and did as well as it could by being “very deep” (19 layers), i.e. (?) it got a lot of chances to form more and more abstract generalizations on its way to a preset goal.  Each representation feeding neatly into the next, nothing emerging that wasn’t told to emerge at the end.

Of course that would all break down if there are similarly “amazing” results in AI concept formation that I just haven’t heard about.  Which there may be.

(Apologies if the style of this post is too grandiose or otherwise off, blame sleep deprivation and long flights across the ocean and massive amounts of caffeine to compensate, I should go to bed very soon)

My answer to this depends on what you mean by “concept formation”. If you mean “given a bunch of data about dogs and cats, and instructed that there are such things as dogs and cats, can an AI figure out which data points are dogs and which are cats?”, the answer is “yes, AIs have done well on this task quite a while now”. We’re at the point with this stuff that computers often outperform humans, particularly with large datasets: it’s often worth it to feed a large dataset to a computer so it can find things no human would notice, and even if it doesn’t it will sure sort the data more quickly than a human would. If you mean “given a bunch of data about dogs and cats, and no instructions that there are such things as dogs and cats or how many groups to look for, can an AI figure out that there are two groups and separate them?”, the answer is still yes but it’s a more qualified yes. The algorithms for doing this are still very good but not as good as the supervised algorithms, and are usually used only when the human isn’t sure whether there are patterns in the data. They are already practically useful and capable of finding patterns that humans can’t, though, so they’re still quite good. If you mean “given a bunch of data about dogs and cats, and no instructions that there are such things as dogs and cats or how many groups to look for, can an AI give you a definition for ‘dog’ and a definition for ‘cat’?” the answer is “sort of, but it’s probably going to give what looks to a human like a very strange answer”. I don’t think it’s entirely fair to blame the AI for this though. If everything you knew about dogs was in terms of height, weight, and ear length, and all relative to cats, you would probably give a very strange definition of a dog too. (Also, sidenote: it’s not like human learning is entirely unsupervised either. You were taught, at some point, that dogs were a thing distinct from cats, and that blue was a color distinct from green. You didn’t figure that out purely from experience. The reason I bring up that second example is because my proof is that in some languages, blue is not a color distinct from green. It’s not like English babies go around wondering why adults are making these arbitrary distinctions between some grue things and other grue things, or like Welsh babies are confused about why people keep calling green things grue.)

(This was originally meant as a quick response to your post but turned into a longer, much more explicit statement of what I was trying to say in the OP)

By “concept formation” I don’t mean the same thing as “binary classification” (or even multiclass classification with a specified number of classes).  There are at least two related differences I have in mind:

(1) Human learning is, yes, not entirely unsupervised, but it is also able to take place in a continually changing perceptual world where one never gets some circumscribed “data set” known to contain some number of useful classes if analyzed properly.  If one had infinite storage space and computation time, one could just run a classifier algorithm on the data set consisting of all the sense data one’s ever seen, but clearly that’s impractical.  It’s possible that this can be done by just continually running classifier algorithms on recent sense data and maintaining a set of “cluster boundaries thought to be useful” which are gradually updated as new data comes in.  (I’m sure there is work like this out there, I just don’t know about it.)

(2) Humans learn hierarchies rather than flat categories, and are able (?) to generate new hierarchies on the fly.  Most work on hierarchical categorization, I think, gives the classifier a pre-built hierarchy which it uses as background knowledge to help with classification.  Some work on learning hierarchies starts with an initial hierarchy and evolves it, while other methods need no initial seed ( “adaptive” and “generative” approaches as defined in section 2.2 here, respectively).  The latter seems like what we want.  But even in this work, one usually (?) starts with a known set of flat classes – much more abstract than the data itself – and then the algorithm learns how to group them together (as in this paper or this one).

By contrast, in human learning, it seems that the set of flat categories we begin with is very primitive – not quite the raw sense data, but just the output of standard sensory processing applied to that data.  (E.g. not “retinal activation patterns,” but “shapes, patterns of motion, etc.”)  For instance, if we sometimes encounter dogs, our sensory apparatus might give us information like “visually furry,” “moves without an external force being applied,” and some sort of processed version of a barking sound (at the outset we do not have the concept of a “bark” yet).  In the high-dimensional space of these sensory features, there is a density peak around the conjunction of these three things (because dogs exist), but finding it by doing classification over the whole flat space of such sensory features is surely computationally intractable.

Instead, I imagine something like this is going on: we notice the immediate conjunction of some sensory features, generate a bespoke hypothetical category (“things that are furry, move on their own, make this sort of sound, etc.”), and then quickly try out that category on new sense data in the short term.  If we keep seeing the same conjunction (where there is one dog, there are often others!), we keep the category and put some work into refining it, so that perhaps we start to notice that dogs have various higher-level features we would never have defined at the outset.  (Even if standard auditory processing doesn’t give us “bark” as a primitive, once we have started to think that dogs are a thing, we may start to pay attention to higher-level features of their sounds and define “a dog’s bark” in terms of the auditory primitives.  More generally, once we have “grabbed onto” a “dog”-like concept, we can start noticing many higher-level features of dogs – that they are common pets, say – which would be far too abstract to be found in a brute-force examination of the primitives.)

@lambdaphagy linked to a very interesting paper in which a very large network (though still 10^6 times smaller than the human visual cortex) was able to learn the (visual) concept of “face” from a training set of unlabelled YouTube frames, less than 3% of which contained faces, just by optimizing for accurate compression of input images.  (“Compression” because some of the intermediate layers were much smaller than the input layer or the output layer [which was the same size as the input and meant to ideally match the input image].)  It also learned about cats and human bodies (which, as the authors drily note, “are quite common in YouTube”).

This is exciting, but it’s a result about how well existing architectures can do when you use a big enough network and run a single round of optimization on a huge computing cluster for 3 days.  There may be tricks that work specifically for unsupervised learning of hierarchical features from low-level primitives – along the lines of what I discussed two paragraphs ago – which could work much more efficiently than the standard approach used here.

Also, the kind of “propose concepts, retain/refine some and discard others” approach I described may be especially valuable for an organism or machine that must react continually to incoming stimuli – online learning by necessity – and doesn’t get to passively absorb a large training set and then, all at once, do a possibly costly optimization over it.  (Although perhaps one could do this while sleeping!)  For building robots or other kinds of artificial minds, it seems especially important to make an algorithm that can (say) pick up a rough version of a concept like “face” pretty quickly, perform information-seeking behavior to help it evaluate the usefulness of this candidate concept, and refine the concept while continuing to interact with the world.

Finally, my interests here are slanted toward the basic science approach “let us explore concept formation, perhaps with the result of efficient algorithms years down the line” and away from applied science approaches like “there is interest in knowing when there’s a dog in a YouTube video, what’s the best we can do in [short, market-driven timeframe].”  In other words, I’m thinking more in the spirit of the earlier AI researchers who liked to talk philosophically about whether their algorithms represented the “right approach” to the problem domain, as opposed to just the most efficient approach available – do perceptrons really get at “what concept formation is”? do Hopfield nets really get at “what memory is”?

As usual (?) in basic science, it may be better to try many ideas on very simple testbeds rather than to jump straight to the kind of real problem that humans solve.  “Discovering ‘dogs’ the way we do” is probably way too much to hope for at the outset.  But maybe we could experiment with similar discovery methods in simpler toy worlds.

Like, consider a computer program designed to play Super Mario Bros.  After some suitable (built-in?) pre-processing of the pure sense data (pixels, possibly sound), could it form a hierarchy of concepts including “enemy” and “block I can hit to get something,” where “enemy” has subtypes like “goomba” and “koopa”?  Could it sometimes sacrifice short-term performance (as predicted by its current, imperfect knowledge) to refine its concepts, say by testing whether a given object is indeed an enemy?  Could it do this all on the fly without needing offline optimization sessions?

I know there are people working on AI for video games, so maybe this thing I’m thinking of already exists?  Would love a pointer to it if so.

A relevant recent paper is 

Building Machines That Learn and Think Like People

I recently summarized it in a talk at a local rationality meeting, as a way of motivating myself to read it carefully. It’s very readable and particularly useful in that it takes into account plenty of very recent research (2015-2016). 

The authors’ main point is that “human-like learning” uses model-building crucially, where by model-building they mean much the same or exactly the same thing as you mean by concept formation. They look at specific examples of classifying handwritten characters and playing old video games, two areas where deep learning systems have been very successful recently, but their success still seems to be contained within a sharply limited functional area, and the authors believe that to get out of that area, machine learning needs to try and incorporate model-building. I particularly liked the way they explained and illustrated the amazing flexibility of models compared with trained neural nets, in that a model remains useful even if you change the goal function sharply. E.g. if you learn to play an Atari video game, you immediately get “for free” such skills as: explaining it to a friend; playing to lose; playing to reach a particular score rather than maximize a score; playing in parallel with a friend aiming to finish the level with identical scores; etc. etc. Any of this is a separate goal for a deep learning system, and all the training it did to learn to play the level really well - even better than you - is just useless in reaching these separate goals. There’s no “learning transfer”. This is just one example - they list several more important aspects in which human learning outperforms the best of deep learning systems in ways that seem to require fundamentally new ideas - but it’s the one which resonated with me the most.  

The authors have some specific ideas on how to try and augment current best-of-breed learning system with model-building-related features, but at least my impression was that these do not reach anything like the goal (if that is to be a goal) of formalizing a general concept of a “model” and explicitly managing such “models” in code. They suggest augmenting AI systems with features that resemble certain seemingly important building blocks of human learning. For example, human infants already early on have a good understanding of how the physical world works (the behavior of solids, liquid, gravity, objects being pushed, pulled, etc.), and we seem to use this understanding to make very efficient shortcuts in e.g. learning to play a video game. They say that an AI system which learns to play an Atari video game,  and is already very good at classifying various objects and classes of objects that appear in the game, might be able to make significant shortcuts by using a relatively simple physics model of the kind current physics simulators already use. [the rest of the paragraph is my opinion] While this may well be a promising approach (hard for me to judge), and while it may direct the attention in the “right” (?) direction of model-building, it doesn’t seem to be nearly general-purpose enough. If I play a video game in which my intuitive physical understanding helps me, and then suddenly the laws of physics in-game change (e.g. gravity reverses direction), I very quickly learn this by using the same general-purpose learning mechanism as I use for the game itself, and adapt. That is, “physics in this game” is for me a model as rich and as amenable to change as “how to play this game”, while the authors mostly discuss using “physics in this game” as a lower-level building block aiding the deep learning system, a building block that doesn’t undergo learning itself.  

But the reason that is so, as I understand it, is simply that nobody knows how to formalize explicitly, or reach implicitly with some sort of learning, models that are as rich and feature-full as human learning seems to use. 

@lovestwell​

There are a lot of little details here that we could go back and forth on forever.  I don’t really want to continue arguing over these details.

Your account of Dreger’s perspective, although internally coherent, feels like it’s reading a lot into the book I just read that wasn’t actually there.  The line you’re drawing is (I take it) between scientists “doing their thing” with potentially harmful results down the line, and scientists using actively unethical methods.  But I don’t remember Dreger ever drawing that distinction explicitly.  This is not me being coy or “perversely charitable” or something; I just don’t remember that being the thrust of the book I read.

(It is also not something I would naturally read in, because it doesn’t fit the facts as I see them.  Michael Bailey is in fact a scientific researcher, but the campaign to ruin his reputation was in response to a popular book he wrote which meant to illustrate a theory he didn’t himself develop – and the theory itself was developed by Ray Blanchard in a clinic [the Clarke Institute, or “Jurassic Clarke”] that has a reputation for clinical horror stories.  So what Bailey actually did is sort of analogous to some colleague of Maria New writing a popular book in which they interview some cherry-picked children who received prenatal dex talking about how great the results are.  Would the author of that book be “just a scientist doing their thing”?)

But in particular I want to reply to your concluding paragraph, because it seems to get at some core friction here:

I don’t understand this, so I apologize if I’m getting it wrong.  What I think you are saying is “you and Promethea believe that mere discussions among scientists of whether certain theories are wrong or right can be inherently harmful and deserve public shaming and nastiness, in which case you must be willing to give up the whole endeavor of scientifically adjudicating the truth or falsehood of those theories.”

I certainly don’t believe that.  I believe (like Dreger) that activism, and society in general, needs the free discussion of scientific ideas.  But I also think that not every statement by someone with a scientific professorship counts as a defense-worthy part of this free discussion.  At a certain point – as when someone writes a book for a general audience containing no new scientific content – they are acting simply as citizens, not as participants in the protected sphere of scientific discourse.  No idea should be inherently anathema in the academy, but no one spends all their time in the academy.

If a chemistry professor (after work) tells someone (not a colleague) that they should mix bleach and ammonia when they get home to make a super-great cleaning product – “trust me, I’m a chemistry professor” – they are not advancing an unorthodox scientific hypothesis in some way we ought to protect and celebrate.

You’re drawing a distinction between science and everything else that I don’t subscribe to, and which I did not intentionally ascribe to you either (whether Dreger subscribes to it or not I’m not sure). Your write “Activism, and the society in general, needs the free discussion of [scientific] ideas” and “No idea should be inherently anathema [in the academy]”, but from where I stand, both these statements improve when the bracketed parts are removed. I don’t know how to make a principled defense of the bracketed parts but not the whole; any such attempt falls victim to the volatility of the boundary.

Look; my position as far as I can see is very simple; Yudkowsky’s “Bad argument gets counterargument.  Does not get bullet.  Never.” pretty much covers it. Someone who’s making a genuine attempt to understand the world and/or explain their ideas to others is covered. It doesn’t matter if they’re writing a peer-reviewed article or a blog post on a personal blog: they do not deserve to be doxed, fired, subjected to an angry activist mob, etc. Now it so happens that scientists are much more likely to be engaged in trying to understand the world than people in general; and it so happens that “bullets” applied to explicitly scientific discourse have the greater potential to fuck with gaining more and better knowledge. And I think since Dreger is especially worried about that (as am I), she focuses on activists hindering scientists. But it doesn’t mean that “argument gets bullet” is virtuous w.r.t. a blogger or a popular book writer. I don’t know what Dreger thinks on that, but I sure don’t think so.

(of course, this also means that the onus is on me to distinguish between Bailey, whose book is protected by this principle, and New, whose actions aren’t. But to me, the difference between them is clear, as I tried to explain in my previous post)

Thus, to take an example, even though I happen to have a strong aversion to anti-Semites for many reasons, including personal ones, if you were to write a post trying to argue in good faith that Jews run the world, the thought of trying to dox you, get you fired, falsely accuse you of various kinds of misconduct etc.  would be extremely repugnant to me. I’m quite content with never having done anything like that, throughout a very long internet life of blogging that included many intense flame wars. 

I’m not sure what *your* position is, but based on the above - and I’m sorry if I’m misinterpreting you - there’s a genuine difference; following @socialjusticemunchkin, you believe that “nasty extremism” is in fact justified in cases where someone argues for a position you believe to be “indirectly harmful” in a major way. Is that a correct summary of your view? Do you, in fact, agree with and justify the actions of Andrea James et al against Bailey described in Dreger’s book (given that Bailey was merely writing a popular book with no new science, which removes him from the “protected sphere” in your words)? In case you do, how much farther would you be willing to go, and in case you don’t, what kind of nasty, directly harmful activism *do* you support against people who express “indirectly harmful” ideas?

Another post on Galileo’s Middle Finger, having finally finished the book.  (Previous posts: Maria New and prenatal dex, also various posts in the tag #michael bailey cw?)

Galileo’s Middle Finger (hereafter GMF) is a strange book.  On one level, the book’s content is pretty easy to make sense of: Alice Dreger has been involved in a number of dramatic academic controversies over the course of her career, and she figured (sensibly enough) that people might enjoy reading a book that retells these stories.  To some extent, she just presents the book as “a memoir of the controversies I’ve been involved in.”

However, she also claims that these stories are connected by an overarching theme, which is something like this:

“Scientists and activists often find themselves at odds, on opposite sides of angry battles.  But everyone should recognize that truth and justice are intimately connected: you can’t help the victims of injustice if you don’t care about the facts of the situation, and if you’re in a unique position to explore facts (such as an academic job), you ought to steer your investigations toward the social good – not by sacrificing the truth, but by looking for the truths that can help.  Activists need to be more concerned with truth, and scientists need to be more concerned with justice.  And if both sides followed this advice, they would be at odds far less often.”

All of this sounds very agreeable to me; I think I already more-or-less agreed with it before I read Dreger’s book.  But do Dreger’s accounts of various controversies actually serve as useful examples of this stuff?  Not always.  And Dreger’s attempts to link everything back to her theme produce some awkward results.

Besides a few minor subplots, there are three controversies narrated in GMF.  First, she narrates the controversy over Michael Bailey’s book The Man Who Would Be Queen.  Second, the controversy over Patrick Tierney’s book Darkness in El Dorado, which accused anthropologist Napoleon Chagnon of genocide as well as various other wrongdoings.  Third, Dreger’s investigations into Maria New and her struggles to get her criticisms recognized by government bodies and the public.

Of these three, it’s the Tierney/Chagnon case that most directly fits Dreger’s theme.  Tierney’s book was a work of shoddy hack journalism which made spectacular allegations that have been uniformly refuted by later investigators.  (N.B. Tierney made many allegations, and some of the more minor ones have been less clearly refuted, but those weren’t the ones that made headlines.)  Nonetheless, shortly after the book came out, the American Anthropological Association quickly endorsed Tierney’s book – the academic equivalent of reflexively believing a callout post without checking the sources.

Reading this in terms of Dreger’s theme seems straightforward: in its concern for justice, the AAA neglected the value of truth, and thus failed to even serve justice.

Even here, though, the theme strains a bit.  The Tierney debacle was not exactly a conflict between “activists” and “academics”; the people under-valuing truth in the service of justice were the academics of the AAA.  (Tierney could arguably be called an “activist,” but Dreger treats him – rightly, it seems to me – as a hack journalist from whom more concern for the truth cannot be expected.)

The Maria New story also lacks a clear instance of an activist failing to sufficiently value truth.  In that story, Dreger is the activist, raising ethical concerns from the outside about an established academic, and her activism is directly grounded in science that she believes that academic is ignoring.  She may intend this as an example of “activism done right” (about which more later), and/or as a case of an academic caring too little about justice.  But it’s not as though New is ignoring justice because truth is her only value; as Dreger notes, her prenatal dex work has produced little in the way of academic knowledge.  So again, it’s hard to see this as an illustration of the theme.

So far, it looks like Dreger has failed to exhibit an example of activists behaving badly, although this is crucial to her theme.  The third story (well, first as presented in the book), about Michael Bailey, is her main (and only) example of this.  But of the three stories, it’s that one that fits the theme least well.

Dreger’s account of the Bailey controversy shares a quality with her account of the Chagnon controversy: both are told as stories of lovable and humane, if out-of-touch, researchers being persecuted by ignorant people who don’t understand them.  Dreger spends a great deal of text talking about how much she personally likes Bailey and Chagnon – Bailey is a personal friend, Chagnon she met while investigating that controversy.  As “characters” in the book, they downright glow.  They’re funny, they’re good company, they both have cute and harmonious marriages.

It makes sense to write stuff like this in order to humanize people who have been demonized by others.  But one has to note here that none of this bears on the “truth” side of the things.  It’s certainly possible for someone to have committed genocide and still be a warm and sparkling conversationalist at the dinner table; it’s possible for Michael Bailey to be a great guy if you know him personally, and nonetheless to have been wrong about trans women.

With Chagnon, this tension never becomes relevant, because as a matter of simple fact, Chagnon was exonerated by multiple serious investigators.  With Bailey, the tension is glaringly relevant, because the issue of whether Bailey is actually right never gets fully addressed in Dreger’s treatment.  Indeed, she treats it almost as an irrelevant side issue.  Where is the value of truth here?

To be fair, Dreger does put her beliefs on the table about the issue.  But these beliefs seem to reveal little serious interest in the questions involved.  She seems to have uncritically bought the Blanchard-Bailey line – possibly because she only cares about these issues insofar as they affect her good friend Michael Bailey? – and to have done little investigation into academic work on transgender beyond this.

Astonishingly, for instance, the phrase “gender dysphoria” never appears in GMF at all.  (A word-search for “dysphoria” turns up only one result, in the title of a Blanchard paper cited in the endnotes.)  When Dreger presents her account of trans women, she talks about (for instance) transitioning as a choice made by feminine gay men in order to better fit into homophobic social environments, stressing that these people might not have transitioned if feminine gay identities were more accepted in their local environments.  I’m willing to believe this happens sometimes – but Dreger seems to actually not know that gender dysphoria is a thing.  This is in a book published in 2015.  One wonders if she’s ever even looked up the condition in the DSM (which changed the name from “Gender Identity Disorder” to “Gender Dysphoria” in the 2013 DSM-V, but even before that had included dysphoria as one of the two major diagnostic criteria).

Dreger has a page on her website, written after GMF was published, in which she responds to questions about “autogynephilia” and states her current positions.  Again, she never mentions gender dysphoria.  Of Blanchard’s androphilic/autogynephilic typology, she says that “I think what I’ve seen from the scientific clinical literature and socioculturally suggests this division makes sense.”  She does not provide any citations, and does not address critiques (see here) that the data show a continuum which does not separate well into two clusters.

I belabor all of this because Dreger’s indifference to the truth here simply makes GMF fundamentally incoherent.  I agree with Dreger’s theme; I have no clue how she thinks the Bailey story illustrates it.

But wait – Dreger’s claim is that activists value truth too little in their quests for justice.  Does this hold true for the activists who attacked Michael Bailey?

Again, Dreger seems to not much care.  She devotes a lot of space to the claims made by these activists, but mostly to express confusion over them.  Noting that some of them display what look to her like signs of autogynephilia, she scratches her head: why are they angry at a book for talking about autogynephilia?  One would think that someone in Dreger’s position – someone interested in getting to the bottom of situations where truth and justice appear to conflict – ought to answer a question like this.  Dreger doesn’t.  Her attitude is basically: “who are these weird people attacking my friend Michael?  What do they want?  They’re so confusing!  Michael is a scientist, so maybe they don’t like science?  Jeez, who knows!!!”

What she substitutes for consideration of these issues – and let me be clear, this is not nothing – is a detailed, blow-by-blow account of the nasty, dishonest ways in which the activists tried to ruin Michael Bailey’s reputation.  They were, in fact, really nasty.  But people don’t just do things like that for no reason.  What about the larger questions of truth and justice here?  Why do these activists believe Michael Bailey is so harmful?  Could it be the case that Bailey is harmful, to the point that defaming him is a net good?

Dreger never mentions this sort of idea, but it hangs uncomfortably over her whole book.  She bemoans the fact that her work on Maria New – which is generally polite and non-nasty, if very harsh on New – has failed to make appreciable waves in the world, beyond loading the first page of Google results with dex-critical pages.  On the other hand, Bailey’s book is now solely known as the subject of a stormy controversy, which received huge amounts of media discussion.  What if nasty activism is sometimes necessary to get the job done?  What if simply having both truth and justice on your side isn’t enough?  And, putting it the other way around, how can Dreger assume that the anti-Bailey crowd didn’t have truth and justice on its side, just because they were nasty and vicious to her friend?

In Dreger’s telling, Andrea James is a scary asshole who sends her possibly-physical threats via email, and Michael Bailey and Napoleon Chagnon are precious cinnamon rolls.  But fighting for truth-and-justice is not the same as identifying the Nice People and the Mean People.  These may in fact be (I hate to say it) largely unrelated endeavors.

A serious book about activism, science, truth and justice would begin with these disquieting possibilities, and then explore from there.  (One example that book might look at: Dreger’s earlier non-nasty activism for intersex people has gotten stuff done.)  Dreger’s book instead stays in an overly cozy universe, where “fighting for good” and “defending her lovable buddies against the bad meanies” can never be conflicting goals.

>how can Dreger assume that the anti-Bailey crowd didn’t have truth and justice on its side, just because they were nasty and vicious to her friend?

Because they sought to maliciously interfere with the scientific process, and censure a researcher for arguing a particular theory. Dreger is pro-science. Dreger is anti-muzzling. Andrea James et al unmistakably tried to muzzle Bailey and ruin his career because they didn’t like the theory he was proposing. Therefore they’re bad activists. Dregger, when she was an activist, didn’t try to stop Maria New because of articles New was publishing; she was trying to stop New’s unethical (in her view) actions towards patients. Therefore she was a good activist.

This view may be naive, but I struggle to understand how you see it as confused or why you continue to charge Dregger with ignoring the object level truth. Dregger believes that science is our best way of getting to the truth, and she’s not shy to repeat it again and again and again. She likes people who insist on doing science, including and especially science perceived as “problematic” by various activists. She dislikes activists who seek to destroy the careers of aforementioned people, and offers her own history of activism, which according to her did nothing of the kind, as a better alternative. This you dismiss derisively as “defending her lovable buddies against the bad meanies”, thereby missing the point. You make it seem as if she met Bailey, really liked him as a person, and continued to twist the facts on the ground into this lovable-Bailey theory, ignoring things that wouldn’t fit. The causality arrow goes in the other direction: those people became her lovable buddies *because* she perceived them as not compromising on science. 

OK, but what if the meanies are correct on object level, you keep asking. What is autogynephilia is bunk, doesn’t Dreger have a duty to address this possibility? Suppose a Fermat Last Theorem crackpot beat up several number theorists because they wouldn’t address his latest proof, and someone wrote about this, and you asked: but how come they don’t address this latest proof? Isn’t it conceivable that it might be correct? Yes, it’s conceivable, but also besides the point, the point being don’t beat up mathematicians. Did I load the dice too much in this analogy? I certainly did, so let’s make it a maverick algebraic geometer who beats up several established mathematicians because they give bad reviews to their articles and they can’t get published. Is it still maybe reasonable to say, don’t beat up other scientists, and whether or not your theory is correct is largely beside the point? Maybe it is, but how can we tell if you keep threatening you’ll thrash anyone who disagrees? Sorry, but in that climate references to published articles that entirely support your theory aren’t very convincing. It’s silly to try and talk about the object level while studiously ignoring the fact that the meta level is thoroughly poisoned.

P.S. You also complain at length that Dreger never tries to explain why the “meanies” seek to muzzle the “lovable buddies”, and what could their motivation be. She does, in fact, talk about it repeatedly; in the Bailey case, her view is that autogynephilia threatens the popular metaphor/theory of “a woman’s brain in a man’s body” etc. In her view, (some) trans activists see this theory as indispensable to promoting trans rights, and of course also as the truth; therefore autogynephilia is a vicious lie seeking to deny the reality of the trans-individual’s true gender. It’s a fair criticism that Dreger doesn’t try very hard to see things from trans activists’ perspective and that her explanations of their motives may be overly simplistic (in particular, that she prefers not to mention dysphoria at all is suspicious). But it isn’t a fair criticism that she doesn’t try at all.

I don’t see a clear-cut division between these two cases.  In both cases, the critics have an ethical case and a theoretical/academic case, and they would be vindicated if these cases add up to the conclusion that the researchers are not generating enough useful knowledge to justify the harm they are doing.

James et. al. didn’t just object to the Blanchard theory out of abstract disagreement; they believed that this theory was harming patients (insofar as it was believed by clinicians) and worsening the day-to-day lives of trans women (insofar as it was believed by the public).  Likewise, Dreger doesn’t just object to New’s methods; she also believes that New is not generating useful knowledge (her papers are few and methodologically poor).

In both cases, it is hard to separate “this theory is wrong” from “this theory is harmful,” since the two are deeply intertwined (as, per Dreger, the two tend to be).  Dreger is not just concerned for New’s own patients, but for the patients of other clinicians who have bought New’s take on the issue.  For instance, in GMF she writes

Note that in some respects it would be easier to make a case that Dreger is trying to “muzzle” New because of her theories than to make the analogous case for James and Bailey – because at least New is doing real, unique research.  (If only because everyone else has held back on doing the same research for ethical reasons!)  Bailey, by contrast, was merely writing a popular book providing narrative illustrations of a scientific theory he already believed was well-established.  Dreger in fact mentions this in the course of defending Bailey:

I don’t want to get too far down the rabbit-hole of these two cases and all of their “object-level” details, here.  My point is that once you look at such details, it becomes difficult to judge these cases on the sole basis of a simple “pro-truth, anti-muzzling” principle.

Moreover, such a principle cannot be applied until one looks closely at the merely “object-level” details.  Questions about the wrongness of muzzling a researcher depend on the details of what that researcher is doing and whether it is really contributing to the common store of knowledge.  Questions about research ethics are inextricable from questions about academic theories, because often the ethics will look different depending on which theory you believe.  (Dreger has to cite studies in order to convince people that prenatal dex is unsafe; I imagine Maria New would have some research-based argument for why it is safe.  This stuff is not orthogonal to the ethical questions!)

>I don’t see a clear-cut division between these two cases.  In both cases, the critics have an ethical case and a theoretical/academic case, and they would be vindicated if these cases add up to the conclusion that the researchers are not generating enough useful knowledge to justify the harm they are doing.

Huh? No, not at all. At least, that certainly isn’t Dreger’s position. In the Bailey case, she’s silent on *usefulness* of autogynephilia - she does believe it to be correct science, but “useful knowledge”? There isn’t any instance I remember where she contrasts the putative harm caused by Bailey’s theories (at least according to their critics - Dreger doesn’t believe they’re causing any real harm at all) with any usefulness they bring to the table. Indeed, she protests repeatedly that Bailey et al. support transgender people’s right to transition, even if they do not buy into the “X brain in the Y body” identity theory - but she doesn’t attempt to claim, as far as I recall, that therapists armed with the autogynephilia diagnosis do a better job of helping trans folks than therapists armed with the identity theory. It may or may not be true, but her opposition to activism in this case doesn’t hinge on this claim, and doesn’t attempt to balance harm vs useful knowledge. She simply opposes ideological attempts to hinder scientists doing their thing, as a matter of principle - no matter how useful their thing is in any particular case, or how true their conclusions are. 

In the Maria New case, her position is even more clearly far from what you take it to be. She does say that Maria New doesn’t produce much in the way of results, but to say that this forms a serious part of her objections to New would be misleading. It’s obvious from reading those chapters that in an alternate universe where Maria New scrupulously follows the principles of bioethics as Dreger understands them, does not try to misrepresent dex as standard of care, obtains fully informed consent from families etc. etc., but the drug itself is as dangerous and (un)helpful as in ours - it’s obvious that in that alternate universe Dreger does not choose to wage bitter war on Maria New. Her outrage is all about injustice done to the patients. When those are not New’s patients but other clinicians’, Dreger doesn’t merely object (pace your description) to them “buying New’s take”; she objects to them having been *misled* by New’s take into falsely believing that dex is the standard of care, and transmitting that false belief to their own patients. This is a crucial point. If New were honest in enlisting support for dex, if she didn’t try to mislead other clinicians about its true status, Dreger would not have objected. She might well have led an activist campaign to bring her own message to clinicians, but she wouldn’t have fought New personally. And the reason I know this is that this is what she tells us happened with her intersex activism at the beginning of her career: she’s very clear that she saw lots of doctors’ positions as being very harmful, but at the same time saw them as trying to behave ethically and do the best thing for their patients; so she lectured and gave talks and appeared at conferences, but didn’t try to “go after” those doctors the way she did with New.

So yes, I see a VERY clear-cut division. When a scientist tries to figure out how the world works (e.g. Bailey in her view) she’s never against that, however outrageous the theories thus formed may be to some. However, when the same scientist uses her position to influence patient care in ways she believes to be unethical (according to the established principles of medicine rather than her own private view), she’s against that. Bailey and New stand well clear of the dividing line, on its separate banks. 

Now you may well believe that Bailey, by virtue of trying to understand how the world works, causes “indirect harm” insofar therapists and the public “buy” autogynephilia. You may even believe, together with @socialjusticemunchkin in the post you just reblogged, that this may justify causing him “direct harm” in the way Andrea James et al had done. Under that view - a view which basically excuses and justifies not just Andrea James but Tierney and all the other “bad meanies” from Dreger’s book - for they all of course believed with fervor that they were working to prevent “indirect harms” of various sorts caused by scientists doing their thing - indeed Bailey can be seen to blend with New. But this isn’t Dreger’s view (or mine), and you shouldn’t criticize her for the confusion that’s not her own. 

(and if you do believe that - if you do think that someone writing an article in support of autogynephilia, for instance, is “causing indirect harm”, and thereby qualifies as “those few cases where nastiness and extremism might be okay”, to quote @socialjusticemunchkin - then your repeated insistence on discussing the object level, the actual truth of autogynephilia and other such theories - remains that much mind-boggling to me).

Another post on Galileo’s Middle Finger, having finally finished the book.  (Previous posts: Maria New and prenatal dex, also various posts in the tag #michael bailey cw?)

Galileo’s Middle Finger (hereafter GMF) is a strange book.  On one level, the book’s content is pretty easy to make sense of: Alice Dreger has been involved in a number of dramatic academic controversies over the course of her career, and she figured (sensibly enough) that people might enjoy reading a book that retells these stories.  To some extent, she just presents the book as “a memoir of the controversies I’ve been involved in.”

However, she also claims that these stories are connected by an overarching theme, which is something like this:

“Scientists and activists often find themselves at odds, on opposite sides of angry battles.  But everyone should recognize that truth and justice are intimately connected: you can’t help the victims of injustice if you don’t care about the facts of the situation, and if you’re in a unique position to explore facts (such as an academic job), you ought to steer your investigations toward the social good – not by sacrificing the truth, but by looking for the truths that can help.  Activists need to be more concerned with truth, and scientists need to be more concerned with justice.  And if both sides followed this advice, they would be at odds far less often.”

All of this sounds very agreeable to me; I think I already more-or-less agreed with it before I read Dreger’s book.  But do Dreger’s accounts of various controversies actually serve as useful examples of this stuff?  Not always.  And Dreger’s attempts to link everything back to her theme produce some awkward results.

Besides a few minor subplots, there are three controversies narrated in GMF.  First, she narrates the controversy over Michael Bailey’s book The Man Who Would Be Queen.  Second, the controversy over Patrick Tierney’s book Darkness in El Dorado, which accused anthropologist Napoleon Chagnon of genocide as well as various other wrongdoings.  Third, Dreger’s investigations into Maria New and her struggles to get her criticisms recognized by government bodies and the public.

Of these three, it’s the Tierney/Chagnon case that most directly fits Dreger’s theme.  Tierney’s book was a work of shoddy hack journalism which made spectacular allegations that have been uniformly refuted by later investigators.  (N.B. Tierney made many allegations, and some of the more minor ones have been less clearly refuted, but those weren’t the ones that made headlines.)  Nonetheless, shortly after the book came out, the American Anthropological Association quickly endorsed Tierney’s book – the academic equivalent of reflexively believing a callout post without checking the sources.

Reading this in terms of Dreger’s theme seems straightforward: in its concern for justice, the AAA neglected the value of truth, and thus failed to even serve justice.

Even here, though, the theme strains a bit.  The Tierney debacle was not exactly a conflict between “activists” and “academics”; the people under-valuing truth in the service of justice were the academics of the AAA.  (Tierney could arguably be called an “activist,” but Dreger treats him – rightly, it seems to me – as a hack journalist from whom more concern for the truth cannot be expected.)

The Maria New story also lacks a clear instance of an activist failing to sufficiently value truth.  In that story, Dreger is the activist, raising ethical concerns from the outside about an established academic, and her activism is directly grounded in science that she believes that academic is ignoring.  She may intend this as an example of “activism done right” (about which more later), and/or as a case of an academic caring too little about justice.  But it’s not as though New is ignoring justice because truth is her only value; as Dreger notes, her prenatal dex work has produced little in the way of academic knowledge.  So again, it’s hard to see this as an illustration of the theme.

So far, it looks like Dreger has failed to exhibit an example of activists behaving badly, although this is crucial to her theme.  The third story (well, first as presented in the book), about Michael Bailey, is her main (and only) example of this.  But of the three stories, it’s that one that fits the theme least well.

Dreger’s account of the Bailey controversy shares a quality with her account of the Chagnon controversy: both are told as stories of lovable and humane, if out-of-touch, researchers being persecuted by ignorant people who don’t understand them.  Dreger spends a great deal of text talking about how much she personally likes Bailey and Chagnon – Bailey is a personal friend, Chagnon she met while investigating that controversy.  As “characters” in the book, they downright glow.  They’re funny, they’re good company, they both have cute and harmonious marriages.

It makes sense to write stuff like this in order to humanize people who have been demonized by others.  But one has to note here that none of this bears on the “truth” side of the things.  It’s certainly possible for someone to have committed genocide and still be a warm and sparkling conversationalist at the dinner table; it’s possible for Michael Bailey to be a great guy if you know him personally, and nonetheless to have been wrong about trans women.

With Chagnon, this tension never becomes relevant, because as a matter of simple fact, Chagnon was exonerated by multiple serious investigators.  With Bailey, the tension is glaringly relevant, because the issue of whether Bailey is actually right never gets fully addressed in Dreger’s treatment.  Indeed, she treats it almost as an irrelevant side issue.  Where is the value of truth here?

To be fair, Dreger does put her beliefs on the table about the issue.  But these beliefs seem to reveal little serious interest in the questions involved.  She seems to have uncritically bought the Blanchard-Bailey line – possibly because she only cares about these issues insofar as they affect her good friend Michael Bailey? – and to have done little investigation into academic work on transgender beyond this.

Astonishingly, for instance, the phrase “gender dysphoria” never appears in GMF at all.  (A word-search for “dysphoria” turns up only one result, in the title of a Blanchard paper cited in the endnotes.)  When Dreger presents her account of trans women, she talks about (for instance) transitioning as a choice made by feminine gay men in order to better fit into homophobic social environments, stressing that these people might not have transitioned if feminine gay identities were more accepted in their local environments.  I’m willing to believe this happens sometimes – but Dreger seems to actually not know that gender dysphoria is a thing.  This is in a book published in 2015.  One wonders if she’s ever even looked up the condition in the DSM (which changed the name from “Gender Identity Disorder” to “Gender Dysphoria” in the 2013 DSM-V, but even before that had included dysphoria as one of the two major diagnostic criteria).

Dreger has a page on her website, written after GMF was published, in which she responds to questions about “autogynephilia” and states her current positions.  Again, she never mentions gender dysphoria.  Of Blanchard’s androphilic/autogynephilic typology, she says that “I think what I’ve seen from the scientific clinical literature and socioculturally suggests this division makes sense.”  She does not provide any citations, and does not address critiques (see here) that the data show a continuum which does not separate well into two clusters.

I belabor all of this because Dreger’s indifference to the truth here simply makes GMF fundamentally incoherent.  I agree with Dreger’s theme; I have no clue how she thinks the Bailey story illustrates it.

But wait – Dreger’s claim is that activists value truth too little in their quests for justice.  Does this hold true for the activists who attacked Michael Bailey?

Again, Dreger seems to not much care.  She devotes a lot of space to the claims made by these activists, but mostly to express confusion over them.  Noting that some of them display what look to her like signs of autogynephilia, she scratches her head: why are they angry at a book for talking about autogynephilia?  One would think that someone in Dreger’s position – someone interested in getting to the bottom of situations where truth and justice appear to conflict – ought to answer a question like this.  Dreger doesn’t.  Her attitude is basically: “who are these weird people attacking my friend Michael?  What do they want?  They’re so confusing!  Michael is a scientist, so maybe they don’t like science?  Jeez, who knows!!!”

What she substitutes for consideration of these issues – and let me be clear, this is not nothing – is a detailed, blow-by-blow account of the nasty, dishonest ways in which the activists tried to ruin Michael Bailey’s reputation.  They were, in fact, really nasty.  But people don’t just do things like that for no reason.  What about the larger questions of truth and justice here?  Why do these activists believe Michael Bailey is so harmful?  Could it be the case that Bailey is harmful, to the point that defaming him is a net good?

Dreger never mentions this sort of idea, but it hangs uncomfortably over her whole book.  She bemoans the fact that her work on Maria New – which is generally polite and non-nasty, if very harsh on New – has failed to make appreciable waves in the world, beyond loading the first page of Google results with dex-critical pages.  On the other hand, Bailey’s book is now solely known as the subject of a stormy controversy, which received huge amounts of media discussion.  What if nasty activism is sometimes necessary to get the job done?  What if simply having both truth and justice on your side isn’t enough?  And, putting it the other way around, how can Dreger assume that the anti-Bailey crowd didn’t have truth and justice on its side, just because they were nasty and vicious to her friend?

In Dreger’s telling, Andrea James is a scary asshole who sends her possibly-physical threats via email, and Michael Bailey and Napoleon Chagnon are precious cinnamon rolls.  But fighting for truth-and-justice is not the same as identifying the Nice People and the Mean People.  These may in fact be (I hate to say it) largely unrelated endeavors.

A serious book about activism, science, truth and justice would begin with these disquieting possibilities, and then explore from there.  (One example that book might look at: Dreger’s earlier non-nasty activism for intersex people has gotten stuff done.)  Dreger’s book instead stays in an overly cozy universe, where “fighting for good” and “defending her lovable buddies against the bad meanies” can never be conflicting goals.

>how can Dreger assume that the anti-Bailey crowd didn’t have truth and justice on its side, just because they were nasty and vicious to her friend?

Because they sought to maliciously interfere with the scientific process, and censure a researcher for arguing a particular theory. Dreger is pro-science. Dreger is anti-muzzling. Andrea James et al unmistakably tried to muzzle Bailey and ruin his career because they didn’t like the theory he was proposing. Therefore they’re bad activists. Dregger, when she was an activist, didn’t try to stop Maria New because of articles New was publishing; she was trying to stop New’s unethical (in her view) actions towards patients. Therefore she was a good activist.

This view may be naive, but I struggle to understand how you see it as confused or why you continue to charge Dregger with ignoring the object level truth. Dregger believes that science is our best way of getting to the truth, and she’s not shy to repeat it again and again and again. She likes people who insist on doing science, including and especially science perceived as “problematic” by various activists. She dislikes activists who seek to destroy the careers of aforementioned people, and offers her own history of activism, which according to her did nothing of the kind, as a better alternative. This you dismiss derisively as “defending her lovable buddies against the bad meanies”, thereby missing the point. You make it seem as if she met Bailey, really liked him as a person, and continued to twist the facts on the ground into this lovable-Bailey theory, ignoring things that wouldn’t fit. The causality arrow goes in the other direction: those people became her lovable buddies *because* she perceived them as not compromising on science. 

OK, but what if the meanies are correct on object level, you keep asking. What is autogynephilia is bunk, doesn’t Dreger have a duty to address this possibility? Suppose a Fermat Last Theorem crackpot beat up several number theorists because they wouldn’t address his latest proof, and someone wrote about this, and you asked: but how come they don’t address this latest proof? Isn’t it conceivable that it might be correct? Yes, it’s conceivable, but also besides the point, the point being don’t beat up mathematicians. Did I load the dice too much in this analogy? I certainly did, so let’s make it a maverick algebraic geometer who beats up several established mathematicians because they give bad reviews to their articles and they can’t get published. Is it still maybe reasonable to say, don’t beat up other scientists, and whether or not your theory is correct is largely beside the point? Maybe it is, but how can we tell if you keep threatening you’ll thrash anyone who disagrees? Sorry, but in that climate references to published articles that entirely support your theory aren’t very convincing. It’s silly to try and talk about the object level while studiously ignoring the fact that the meta level is thoroughly poisoned.

P.S. You also complain at length that Dreger never tries to explain why the “meanies” seek to muzzle the “lovable buddies”, and what could their motivation be. She does, in fact, talk about it repeatedly; in the Bailey case, her view is that autogynephilia threatens the popular metaphor/theory of “a woman’s brain in a man’s body” etc. In her view, (some) trans activists see this theory as indispensable to promoting trans rights, and of course also as the truth; therefore autogynephilia is a vicious lie seeking to deny the reality of the trans-individual’s true gender. It’s a fair criticism that Dreger doesn’t try very hard to see things from trans activists’ perspective and that her explanations of their motives may be overly simplistic (in particular, that she prefers not to mention dysphoria at all is suspicious). But it isn’t a fair criticism that she doesn’t try at all.

I learned this from somebody who added that nachos (the food) were named after their inventor, Ignacio Anaya, and so the words ‘Nazi’ and ‘nacho’ are cognate with each other.

This seems likely to be folk (and false) etymology. The long text in the OP is just a fanciful tale built around the core idea. It appears to be a well-attested story told confidently, but the author doesn’t know what he’s talking about. He thinks, for example, that “Nazi” was made from NAtionalsoZIalistische - he highlights these two syllables. That’s a silly, but understandable naive version to make up for someone with zero knowledge of German. The word “Nazi” is a *phonetic* shortening - “Nazi” is pronounced as the first two syllables of “Nationalsozialistische”. Wikipedia gets this right.

But the core idea is attested. If you follow the links from Wikipedia to etymonline to a German etymological dictionary etc., it boils down to three claims:

1. “Nazi” was a common shortening for Ignaz, a common Bavarian first name.

2. In that use and at the time, “Nazi” commonly stood for a stupid/inept/clumsy person, as common rural nicknames often do in various cultures.

3. Opponents of Nazis seized on that connotation deliberately and used the name “Nazi” to link the party with the idea of a stupid/inept/clumsy person, and this was understood by people at the time, including the Nazis themselves, who hated it, though they also tried to embrace it for some time.

1. is true beyond doubt, e.g. the nickname is well-attested as character names and sometimes titles in books and plays in late 19th/early 20th century.

2. seems to have little to no support from contemporary sources, and likewise 3. I couldn’t - so far, didn’t look very deep - find any remark made by anyone in 1920s or 1930s linking “Nazi” as the nickname for the party with “Nazi” as a stereotypically inept peasant named Ignaz. 

Seeing as “Nazi” was used occasionally in the beginning of the 20th century and the end of the 19th to refer to *other* ideologies with names starting “National...”, and seeing as “Sozi” as short for “Sozialistische” was already in use, it seems likely that Nazi was simply a phonetic shortening that became really popular with the opponents of the party. Think “dems” or “repubs”.