In every discussion I've had since, in any forum, nobody who thinks them to be incompatible can describe even vaguely what "free will" would be supposed to look like if it does not contain determinism inside it.
We haven't met, then. I'm natuarlistic libertarian and therefore an incompatibilist. I may not be right, but so long as I am not completely wrong, EY does not have "the" answer.
Anyway, on to the refutation of "FW requires determinism"..
II.1.i Introduction Compatibilism comes in a stronger form, which does not have a traditional name; We will call it supercompatibilism. According to supercompatibilism, free will is not only capable of existing alongside causal determinism, it cannot exist without determinism.
Here is case that free will requires causal indeterminism.
PRO: "Free will requires the ability to have done otherwise under the very same circumstances, which is only possible in a universe with some degree of indeterminism."
The following argument goes against the usefulness of indeterminism to free will:-
ANTI: "Freedom of a kind worth wanting requires rationality, and rationality mean following rules. Causal Determinism would ensure that the rules are followed. Indeterminism would disrupt the process of rational thought, and result in a capricious, irrational kind of freedom not worth having. An individual cannot call an action his or her own unless they can account rationally account for it, and unless it was caused by their own intentions. Reasons are causes, so to be rational is to be determined.
We will now argue against the ANTI argument, taking it a sentence at a time.
II.1.ii Objection 1: "Freedom of a kind worth wanting requires rationality"
Yes, but it is by no means limited to rationality. free will of a kind worth wanting must facilitate the whole gamut of human behaviour, including creativity, imagination, inventiveness etc.
II.1.iii Objection 2: "Causal Determinism would ensure that the rules are followed."
Casual Determinism only guarantees that everything follows the laws of nature, not that anything follows the laws of thought. If you believe the universe is deterministic, you have to admit that any lapse of rationally is just as determined as everything else. Indeed, we are more likely to look for a deterministic — in the sense of an external cause —explanation for uncharacteristically irrational behaviour than for rational behaviour — so-and-so was drunk, drugged, etc.
II.1.iv Objection 3: "Indeterminism would disrupt the process of rational thought, and result in a capricious, irrational kind of freedom not worth having."
Is that so ? Computer programmes can consult random-number generators where needed, including 'real' ones implemented in hardware. (like this) The rest of their operation is perfectly deterministic. Why should the brain not be able to call on indeterminism as and when required, and exclude it the rest of the time ? And if random numbers are useful for computers, why should indeterministic input be useless for brains ? Is human rationality that much more hidebound than a computer ? Even including such faculties as creativity and imagination ? Pseudo-random numbers (which are really deterministic) may be used in computers, and any indeterminism the brain calls on might be only pseudo-random. But it does not have to be, and if we assume it is not, we can explain realistically why we have the sense of being able to have done otherwise.
And is it so great to be compelled into rule-following rationality ? If you asked some what 5 and 7 make, you would expect the answer 12. But if you asked them the same question ten times in a row, you would expect them to object at some stage, and stop answering, unlike a pocket calculator which will spit out the same answer ad infinitum. Surely the choice of whether or not to follow a set of rules is part of rationality ?
II.1.iv Objection 4: "An individual cannot call an action his or her own unless they can rationally account for it and unless it was caused by their own intentions"
It is true that we would not consider an individual to 'own' a an action or decision if it had nothing to do with his beliefs and aims at the time he made it — that is, if we assume that indeterminism erupts in-between everything that happened to make him the individual he is, and the act itself.
But, we libertarians claim. an act is also not an individual's own if it entirely attributable to causes lying outside him, ultimately traceable to circumstances before he came into existence.
There is no need to be disheartened. If the causal origins of our actions cannot lie before our births or after our decisions have been made, they can still lie, just where they should: during our lifetimes.
Our actions can be determined by our preceding mental state, providing that our mental state is not itself entirely attributable to causes outside of ourselves. This means that, although we can pin actions to immediate purposes, we cannot trace back a chain of purposes-for-purposes ad infinitum.
I do not think that is any loss, since determinism fares no better. All purposes may be causes, but not all causes are purposes. The deterministic causal chain, if traced back, is bound to encounter factors which do not rationally explain, any more than a random occurrence does. Moreover, this process of looking for ultimate rational explanations is unusual to say the least. Our normal attitude is that John and Mary have their reasons, which are very much part of who they are, and that's that. II.1.v Reasons are causes, so to be rational is to be determined. Is it the case that our actions are determined by our reasons, so that indeterminism must entail irrationality?
The reasons we have for actions might be said to cause those actions, but there are a number of differences between reasons and causes, and in any case reasons are not determining causes, as will be argued.
Whereas any event can involve causation, only a small subset of entities in the universe, rational agents, can base their actions on reasons. Causes lie in the past, whereas reasons are generally directed towards some future state of affairs. A causal statement is not a direct substitute for a rational justification. If I kick someone and am asked why I did it, it is no justification — for all that it is true — to say that nerve signals from my brain caused my leg to flex. A Causal chain can continue back to the Big Bang, but a rational explanation for behavour cannot. An individual must explain their behaviour in tersm of their own aims an desires. Once they start claiming they were caused to behave in a certain way, by their environment, genes,etc, they are no longer offering reasoned explanaion. Causal explanation is 'classical' — everything is brought under a uniform, impersonal set of laws. Rational explanation is 'romantic' — people have their own unaccountably individual reasons for doing things. There is no single right answer to "what should I do", as it depends on what you as an individual want to do — nor is there any single right answer to "what should I want".
That reasons are not determining causes is established by the fact that one can have multiple reasons for multiple courses of action. If one accepts the slice of cake, one pleases the hostess; if one rejects it, one sticks to ones diet. If reasons and actions are chosen in pairs there will always be a reason for ones action irrespective of what is actually chosen. This is still true if a choice is made randomly. If one does choose a reason-action pair randomly, it is admittedly true that a reason-for-the reason cannot be given. But it is always the case that chains of reasoning cannot be pursued to infinity; they have to stop somewhere (or be circular). Since real-world reasoning is constrained in this way, limited amounts of indeterminism would not render people any less able to provide chains of reasoning than they are anyway, even though it remains the case that complete indeterminism entails complete irrationality.
If reasons are not determining causes, are they causes at all? It would certainly be a peculiar situation if reasons were causally completely detached from the actions they explain. At least the neural correlates of reasons need to be causes. Reasons themseleves arguably belong to a different language game so that it is a category-error to substitute talk of reasons directly for talk of causes an vicer-versa. This need not imply any ontological dualism, but rather a kind of anomolous monism.
It is sometimes said that we are free to do what we want, but not to choose what we want. The approach sketched so far is pretty much the opposite of this. If we have clear reasons for doing something and are in a rational frame of mind our actions follow almost inevitably. Freedom lies in the fact that our basic aims and goals, our basic nature is not inevitable, We might not have done as we did because we might not have been that kind of person. Freedom is not mere caprice, not does it lie in being the puppet of circumstances, it is self determination, a gradual evolution of selfhood.
To have an opinion about free will, you must first observe the existence of the issue.
Most people do this with introspection: The world outside you seem to conform to \ while inside you, it seems that indeed you control every movement and thought.
Lord Kelvin has voiced the above statement quite poetically.
Now, the keyword here is 'seem'. Your argument hitches on an anecdote from your own, non-optimal cognitive machinery.
What EY did was point at this 'seem' and explain it. He did not point at free will and explained it, he explained why the cognitive machin...
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:
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).