Gary Drescher’s Good and Real: Demistifying Paradoxes from Physics to Ethics

I first came across a throwaway reference to Gary Drescher’s Good and Real while going through the long series of Less Wrong posts by Eliezer Yudkowsky collectively called “the Sequence” a few years back, which I thought at the time was the most eye-opening thing I’d ever read, so the fact that Yudkowsky mentioned it as providing him with the impetus to finish writing the quantum physics sequence as quickly as possible so as to not be influenced by Drescher’s arguments (which apparently ran along very similar lines) was to me a positively glowing review, albeit something of a sidetrack. I told myself I’d look into Drescher’s book when I got the chance, but forgot about it.

Just now I saw it get a mention again deep within the comments section of Scott Alexander’s Slate Star Codex, in a post called Proving Too Much (which itself deserves a read). The idea behind the fallacy of an argument “proving too much” is that it doesn’t just prove your conclusion, it proves something else that we can agree is patently false (or absurd, or untenable, etc – which leaves open the possibility that our personal absurdity heuristics are just poorly calibrated, among other things, but I digress). The argument that we can’t reject the existence of a supernatural power on the grounds that we haven’t definitively established proof of nonexistence, for instance, shouldn’t be used to imply that said power might exist (and running away with it, that it has a high probability of existing, that we should act and live our lives as if it exists, etc): replace “supernatural power” with Bigfoot and notice that the exact same conclusion holds.

(This isn’t arguing for disbelievers of supernatural powers either: I’m just saying this particular argument is a poor one, and we should look for others if we’re honest about wanting to go down to the bottom of things instead of just raising or lowering the bar for evidence according to whether we lean towards or away from the conclusions.)

In any case I’ve digressed again, so let’s get back to Drescher’s Good and Real.

One thing that struck me when reading the reviews section for the book on Amazon.com was that Danny Hillis gave it a glowing review. Danny Hillis isn’t a philosopher, I’ll grant you that, but he’s one of the most original thinkers around. I first came across him in a profile for a Reader’s Digest article when I was probably twelve or whereabouts; the fact that I still remember him speaks a lot of how much the description of him made an impression on me, though the details are lost to time – Disney Imagineer; co-founder of Applied Minds, Inc; co-founder of the Long Now Foundation (yes, that Long Now Foundation); inventor, thinker, all-around awesome guy. Here’s his review:

I am proud to write a review for this book, because I am convinced that philosophers of the future will look back on it as being ahead of its time.

Drescher establishes a comprehensive framework for studying some of the most difficult problems in philosophy, starting with a mechanistic view of the mind. With these tools, he dissects some of the most perplexing philosophical problems, questions about mind and body, consciousness, cause and effect, and moral choice. Drescher demonstrates convincingly that many of our intuitions about free will and moral choice are not only not contradicted by a mechanistic view, but can be supported by it.

Last emphasis mine, because it’s the one thing I (as someone who intuitively espouses a reductionistic, mechanistic view of the world) have yet to be able to understand. Where, paraphrasing Death in a Discworld novel of Terry Pratchett’s, are the atoms of morality to be found? And hasn’t Sam Harris already shown in his book Free Will (convincingly enough to me) that, rehashing standard arguments to be found in the literature, free will and determinism are incompatible, that the compatibilist view is either unsound or baits-and-switches by redefining “free will” to mean something else? Free will and ethics in a deterministic universe has always been a topic of personal interest to me, so I found this review particularly attractive.

Here’s another review I found interesting because it basically dismisses Harris’ own attempt at grounding ethics within a naturalistic framework (where ethics isn’t just “built-in”), while supporting Drescher as saying it “goes much further”:

I found Drescher’s arguments sound and consistent, and his assumptions more than reasonable, and thus can agree with his general conclusions for the most part without much reservation. It goes much further than other recent attempts at grounding ethics within a naturalistic framework, such as Sam Harris’ failed attempts in his recent book ‘The Moral Landscape’, but I fear the necessarily more technical style of Drescher’s book will impede it from receiving the popular attention it deserves.

Same review also mentions the is-ought distinction and how Drescher resolves it, which I found really interesting because I had thought the is-ought divide impassable:

Drescher’s title ‘Good and Real’ alludes to the is/ought dichotomy of what there ‘is’ and how moral agents ‘ought’ to act. He presents solutions to both using reasonable assumptions based on modern scientific evidence and then extrapolating those into cleverly simplified toy models.

Underlying all of Drescher’s thinking is a foundational construction of the ‘real’ or what ‘is’ and can be summarized as a deterministic quantum-mechanical configuration space based on Everett’s many-worlds interpretation that sits statically and timelessly representing the possibility space of spacetime. Using this foundation, he offers a theory of the ‘good’ or what ‘ought’ to be done and can be summarized as following the rule of subjunctive reciprocity, which is the use of acausal counterfactual reasoning to justify following Kant’s categorical imperative. In reaching this conclusion, Drescher spends time reconciling notions of free-will with a deterministic universe and puts forth arguments for using acausal counterfactual reasoning as the preferred way of thinking about means-end relations that is more general than causal relationships but also more strict than mere evidential relationships.

And here’s the review that compelled me to write this blog post:

Explain to people that they are fully natural, caused creatures, that they don’t have contra-causal free will, and they often suppose you’re dallying with fatalism. Explain that there is no provable basis for morality outside the natural world, and they often assume you’re a moral relativist or nihilist. How can we construe human choices as anything but illusory if all we do is completely determined? How can we judge behavior right or wrong if there are no supernatural ethical foundations?

In Good and Real, computer scientist and independent scholar Gary Drescher mounts a mind-bending attack on these and other problems that arise when commonsense conflicts with the science-based view that we inhabit a purely physical, mechanistic, deterministic universe. (Please fasten your seatbelts.) Establishing that we are in such a universe is just one of his projects, set forth in a chapter called “Quantum Certainty.” Drescher explains and defends Hugh Everett’s relative-state interpretation of quantum mechanics in which there is no collapse of the waveform and in which the evolution of the (locally branching) universe in configuration space is fully deterministic. This unflinching fidelity to the mathematical quantum formalism is quite the opposite of pop-quantum physics, for instance as popularized by the film What the Bleep Do We Know, which gives the putatively undetermined conscious observer a special role in “creating” reality by collapsing the waveform. Here as elsewhere in the book Drescher draws a tough-minded, unpopular conclusion: sorry, we don’t create our own reality.

Nor is consciousness something that transcends mechanism. Rather, Drescher explains in “Dust to Lust,” it’s what happens when a representational system goes recursive and starts taking its own episodes of representing as objects of further representation. Consciousness isn’t something extra generated by recursion, it is recursion (of a particular type), and so not anything that can’t be instantiated by a sufficiently complex mechanism, for instance, ourselves. Many readers will object to such a characterization: after all, we’re not just machines, are we? Well yes, we’re organic machines, choice machines in fact, Drescher says, whose consciousness and rationality can best be explained as the complex deterministic functionality of achieving goal states that have many sub-goals. Sticking with science, there’s no reason to suppose we’re animated by something non-physical in our goal-seeking behavior, since that assumption does no explanatory work. It’s here that many will likely part company with Drescher, and hold out for extra-scientific claims about our cognitive capacities, for instance that consciousness transcends the brain. Such claims support a more “optimistic” view about human exceptionalism, in which our choices have contra-causal leverage over the world. But this refuses to let empirical findings drive our conclusions about reality – a no-no of the first order for scientific naturalists like Drescher.

The discussions of consciousness and quantum physics are joined by a consideration of time in the chapter “Going Without the Flow.” Drescher reminds us that, according to 100 year-old standard physics, all events are sitting statically in four dimensional space-time. The past, present and future just are – there is no cursor moving forward along the time dimension that temporarily endows each moment with reality. All moments are equally real, which means that the future is there, “waiting” to be discovered by consciousness, not created de novo by human action. Now we start to see the problem for our standard intuition about human efficacy: if the future is inalterable, aren’t choices futile?

Before tackling this problem, Drescher explains how the illusory impression of the flow of time arises, and further, given that basic physical laws don’t specify a temporal direction, why it is we only observe events evolving forward in time, not backwards. As is often the case in this book, readers will find the explanations challenging; not because the writing isn’t lucid (it is, and often entertaining) but simply due to the conceptual complexity and counterintuitiveness of the material, which sometimes translates, inevitably, into what are politely referred to as technicalities. Although the gist of his conclusions can be grasped without tangling with the tough parts, to decide if he’s right requires you grapple with them.

The last third of Good and Real is devoted to the twin problems of choice and ethics in a deterministic universe, and if your mind isn’t already stretched, this will definitely do the trick. If we are choice machines, whose every decision is etched inalterably in the space-time manifold, and whose consciousness isn’t privileged in creating reality, why bother to act for the sake of what already exists? Part of the answer is relatively straightforward: if we didn’t bother to engage in choice making behavior, which ordinarily includes considering alternative possibilities, then we wouldn’t be as likely to achieve our goals. And choices needn’t involve our being causal exceptions to nature:

“Thus choice…is a mechanical process compatible with determinism: choice is a process of examining assertions about what would be the case if this or that action were taken, and then selecting an action according to a preference about what would be the case. The objection *The agent didn’t really make a choice, because the outcome was already predetermined* is as much a non sequitur as the objection *The motor didn’t really exert force, because the outcome was already predetermined.* Both choice making and motor spinning are particular kinds of mechanical processes. In neither case does the predetermination of the outcome imply that the process didn’t really take place.” (p. 192, original emphasis)

But the rest of Drescher’s answer takes us way down the rabbit hole, first by means of the seemingly innocent example of safely crossing the street, followed by his solution to Newcomb’s Problem, a notorious thought experiment about choice and prediction that has divided philosophers for decades. It turns out, says Drescher, that it makes sense to act as if your choice had an effect on conditions preceding the choice, even though there’s no causal link between your choice and those conditions. There exists what he calls a subjunctive means-end relation, a non-causal link between action and desired states of affairs. Therefore, Drescher argues, it can be rational to act for the sake of states of affairs that you know already obtain. If this seems completely counter-intuitive, join the club. Making it intuitive or at least logically transparent is Drescher’s goal, which in my case was not achieved, at least at first pass (which says nothing about whether he’s correct, since it will likely take several passes to fully understand the argument).

The capstone of Drescher’s tour de force is to apply the rationality of appreciating subjunctive means-ends relations to the classic problem of the Prisoner’s Dilemma, and from that derive an ethics grounded in enlightened self-interest. Agents caught in the dilemma who are smart enough to grasp the reality of subjunctive means-ends links will see that it’s in their best interest to cooperate, not defect. This insight, generalized, becomes the rational basis for Kant’s categorical imperative and the golden rule. Unlike Kant, however, Drescher posits nothing beyond the physical space-time continuum and goal-seeking choice machines (us) to establish this most basic ethical maxim. So, perhaps, he has fully naturalized it.

The scope of Drescher’s ambition in this volume will not have escaped the reader. But he doesn’t come across as ambitious or overbearing, just curious and relentlessly logical, wanting to get to the bottom of the best puzzles that unvarnished reality offers. That he ventures into such diverse territory might make specialists suspicious, but Drescher seems to have done his homework. Deciding whether he’s right in any given instance will, however, require a close reading of his arguments and an evaluation of his evidentiary basis, for instance in consciousness studies, physics, game theory, and behavioral economics. Many of us non-specialists will likely have to reserve judgment, but can we suppose that standard intuitions about choice and reality, comforting though they be, are better than Drescher’s carefully thought out if counterintuitive conclusions? Here are the big questions, addressed by a gifted, independent-minded thinker, made real for us in all their perplexity, and it’s good that we should catch at least a glimpse of well-argued answers that form a satisfying whole. A deterministic, godless universe can, it seems, offer a sufficient basis for human efficacy and ethics.

Now I’m seriously reconsidering getting Good and Real.

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s