Due in part to Eliezer's writing style (e.g. not many citations), and in part to Eliezer's scholarship preferences (e.g. his preference to figure out much of philosophy on his own), Eliezer's Sequences don't accurately reflect the close agreement between the content of The Sequences and work previously done in mainstream academia.
I predict several effects from this:
- Some readers will mistakenly think that common Less Wrong views are more parochial than they really are.
- Some readers will mistakenly think Eliezer's Sequences are more original than they really are.
- If readers want to know more about the topic of a given article, it will be more difficult for them to find the related works in academia than if those works had been cited in Eliezer's article.
I'd like to counteract these effects by connecting the Sequences to the professional literature. (Note: I sort of doubt it would have been a good idea for Eliezer to spend his time tracking down more references and so on, but I realized a few weeks ago that it wouldn't take me much effort to list some of those references.)
I don't mean to minimize the awesomeness of the Sequences. There is much original content in them (edit: probably most of their content is original), they are engagingly written, and they often have a more transformative effect on readers than the corresponding academic literature.
I'll break my list of references into sections based on how likely I think it is that a reader will have missed the agreement between Eliezer's articles and mainstream academic work.
(This is only a preliminary list of connections.)
Obviously connected to mainstream academic work
-
Eliezer's posts on evolution mostly cover material you can find in any good evolutionary biology textbook, e.g. Freeman & Herron (2007).
-
Likewise, much of the Quantum Physics sequence can be found in quantum physics textbooks, e.g. Sakurai & Napolitano (2010).
-
An Intuitive Explanation of Bayes' Theorem, How Much Evidence Does it Take, Probability is in the Mind, Absence of Evidence Is Evidence of Absence, Conservation of Expected Evidence, Trust in Bayes: see any textbook on Bayesian probability theory, e.g. Jaynes (2003) or Friedman & Koller (2009).
-
What's a Bias, again?, Hindsight Bias, Correspondence Bias; Positive Bias: Look into the Dark, Doublethink: Choosing to be Biased, Rationalization, Motivated Stopping and Motivated Continuation, We Change Our Minds Less Often Than We Think, Knowing About Biases Can Hurt People, Asch's Conformity Experiment, The Affect Heuristic, The Halo Effect, Anchoring and Adjustment, Priming and Contamination, Do We Believe Everything We're Told, Scope Insensitivity: see standard works in the heuristics & biases tradition, e.g. Kahneman et al. (1982), Gilovich et al. 2002, Kahneman 2011.
-
According to Eliezer, The Simple Truth is Tarskian and Making Beliefs Pay Rent is Peircian.
-
The notion of Belief in Belief comes from Dennett (2007).
-
Fake Causality and Timeless Causality report on work summarized in Pearl (2000).
-
Fake Selfishness argues that humans aren't purely selfish, a point argued more forcefully in Batson (2011).
Less obviously connected to mainstream academic work
-
Eliezer's metaethics sequences includes dozens of lemmas previously discussed by philosophers (see Miller 2003 for an overview), and the resulting metaethical theory shares much in common with the metaethical theories of Jackson (1998) and Railton (2003), and must face some of the same critiques as those theories do (e.g. Sobel 1994).
-
Eliezer's free will mini-sequence includes coverage of topics not usually mentioned when philosophers discuss free will (e.g. Judea Pearl's work on causality), but the conclusion is standard compatibilism.
-
How an Algorithm Feels From Inside and Dissolving the Question suggest that many philosophical problems can be dissolved into inquiries into the cognitive mechanisms that produce them, as also discussed in, for example, Shafir (1998) and Talbot (2009).
-
Thou Art Godshatter, Not for the Sake of Happiness Alone, and Fake Utility Functions make the point that value is complex, a topic explored in more detail in affective neuroscience (Kringelbach & Berridge 2009), neuroeconomics (Glimcher 2010; Dolan & Sharot 2011), and other fields.
-
Newcomb's Problem and the Regret of Rationality repeats a common debate among philosophers. Thinking that CDT must be right even though it "loses" to EDT on Newcomb's Problem, one group says "What can we do, if irrationality is rewarded?" The other group says "If you're so smart, why aren't you rich? What kind of rationality complains about the reward for irrationality?" For example, see Lewis (1981).
I don't think Eliezer had encountered this mainstream work when he wrote his articles
-
Eliezer's TDT decision algorithm (2009, 2010) had been previously discovered as a variant of CDT by Wolfgang Spohn (2003, 2005, 2012). Both TDT and Spohn-CDT (a) use Pearl's causal graphs to describe Newcomblike problems, then add nodes to those graphs to represent the deterministic decision process the agent goes through (Spohn calls them "intention nodes," Yudkowsky calls them "logical nodes"), (b) represent interventions at these nodes by severing (edit: or screening off) the causal connections upstream, and (c) propose to maximize expected utility by summing over possible values of the decision node (or "intention node" / "logical node"). (Beyond this, of course, there are major differences in the motivations behind and further development of Spohn-CDT and TDT.)
-
Many of Eliezer's points about intelligence explosion and machine ethics had been made in earlier writings Eliezer did cite, e.g. Williamson (1947), Good (1965), and Vinge (1993). Others of Eliezer's points appear in earlier writings he did not cite but probably had read: e.g. Minsky (1984), Schmidhuber (1987), Bostrom (1997), Moravec (1999). Others of Eliezer's points appear in earlier writings he probably hadn't read: e.g. Cade (1966), Good (1970), Versenyi (1974), Lukasiewicz (1974), Lampson (1979), Clarke (1993, 1994), Sobel (1999), Allen et al. (2000). (For a brief history of these ideas, see here and here.)
-
A Technical Explanation of Technical Explanation retreads much ground from the field of Bayesian epistemology, surveyed for example in Niiniluoto (2004) and Howson & Urbach (2005).
Do you have a Greasemonkey script that rips all the qualifying words out of my post, or something? I said things like:
Your comment above seems to be reacting to a different post that I didn't write, one that includes (false) claims like: "The motivations, the arguments by which things are pinned down, the exact form of the conclusions are mostly the same between The Sequences and previous work in mainstream academia."
Really? This is the default reaction I encounter. Notice that when the user 'Thomas' below tried to name just two things he thought were original with you, he got both of them wrong.
Here's a report of my experiences:
People have been talking about TDT for years but nobody seems to have noticed Spohn until HamletHenna and I independently stumbled on him this summer.
I do find it hard to interpret the metaethics sequence, so I'm not sure I grok everything you're trying to say there. Maybe you can explain it to me sometime. In any case, when it comes to the pieces of it that can be found elsewhere, I almost never encounter anyone who knows their earlier counterparts in (e.g.) Railton & Jackson — unless I'm speaking to someone who has studied metaethics before, like Carl.
A sizable minority of people I talk to about dissolving questions are familiar with the logical positivists, but almost none of them are familiar with the recent cogsci-informed stuff, like Shafir (1998) or Talbot (2009).
As I recall, Less Wrong had never mentioned the field of "Bayesian epistemology" until my first post, The Neglected Virtue of Scholarship.
Here's a specific story. I once told Anna that once I read about intelligence explosion I understood right away that it would be disastrous by default, because human values are incredibly complex. She seemed surprised and a bit suspicious and said "Why, had you read Joshua Greene?" I said "Sure, but he's just one tip of a very large iceberg of philosophical and scientific work demonstrating the complexity of value. I was convinced of the complexity of value long ago by metaethics and moral psychology in general."
Let's look at them more closely:
Lots of cited textbooks were written after the Sequences, because I wanted to point people to up-to-date sources, but of course they mostly summarize results that are a decade old or older. This includes books like Glimcher (2010) and Dolan & Sharot (2011).
Batson (2011) is a summary of Batson's life's work on altruism in humans, almost all of which was published prior to the Sequences.
Spohn (2012) is just an update to Spohn's pre-Sequences on work on his TDT-ish decision theory, included for completeness.
Talbot (2009) is the only one I see that is almost entirely composed of content that originates after the Sequences, and it too was included for completeness immediately after another work written before the Sequences: Sharif (1998).
That's too bad, since I answered this question at the top of the post. I am trying to counteract these three effects:
I find problem #1 to be very common, and a contributor to the harmful, false, and popular idea that Less Wrong is a phyg. I've been in many conversations in which (1) someone starts out talking as though Less Wrong views are parochial and weird, and then (2) I explain the mainstream work behind or similar to every point they raise as parochial and weird, and then (3) after this happens 5 times in a row they seem kind of embarrassed and try to pretend like they never said things suggesting that Less Wrong views are parochial and weird, and ask me to email them some non-LW works on these subjects.
Problem #2 is common (see the first part of this comment), and seems to lead to phygish hero worship, as has been pointed out before.
Problem #3, I should think, is uncontroversial. Many of your posts have citations to related work, most of them do not (as is standard practice in the blogosphere), and like I said I don't think it would have been a good idea for you to spend time digging up citations instead of writing the next blog post.
Predictable misunderstandings are the default outcome of almost anything 100+ people read. There's always a trade-off between maximal clarity, readability, and other factors. But, I'm happy to tweak my original post to try to counteract this specific misunderstanding. I've added the line: "(edit: probably most of their content is original)".
Remember that I came to LW with a philosophy and cogsci (especially rationality) background, and had been blogging about biases and metaethics and probability theory and so on at CommonSenseAtheism.com for years prior to encountering LW.
That is definitely not the spirit of my post. If you'll recall, I once told you that if all human writing were about to be destroyed except for one book of our choosing, I'd go with The Sequences. You can't get the kind of thing that CFAR is doing solely from Feynman, Kahneman, Stanovich, etc. And you can't get FAI solely from Good, Minsky, and Wallach — not even close. Again, I get the sense you're reacting to a post with different phrasing than the one I actually wrote.
Most people won't read the literature either you or I link to. But many people will, like Wei Dai.
Case in point: Remember Benja's recent post on UDT that you praised as "Original scientific research on saving the world"? Benja himself wrote that the idea for that post clicked for him as a result of reading one of the papers on logical uncertainty I linked to from So You Want to Save the World.
Most people won't read my references. But some of those who do will go on to make a sizable difference as a result. And that is one of the reasons I cite so many related works, even if they're not perfectly identical to the thing me or somebody else is doing.
I think a valid criticism can be made that while you were trying to counteract these three effects (which is clearly an important and useful effort), you didn't take enough care to avoid introducing a new effect, of making some people think the Sequences are less original than they actually are. (For example you didn't ask Eliezer to double check your descriptions of how the Sequences posts relate to the academic works, and you didn't give some examples of where the Sequences are original.)
This is bad becaus... (read more)