Michael Shermer has a new book out called Truth: What it is, How to Find it, and Why it Still Matters, and I’ve mentioned it before. I’m reading it now, and there’s a lot of good stuff in it. But one of the twelve chapters—the one on free will—is, I think, misguided and confusing. In the preceding link you’ll find a video he made about free will, as well as my critique of it. You may not want to read this post if you’ve read the previous one, but the video differs slightly from the article I discuss below.
So here’s my take 2 on Shermer’s views, recently expressed in a longish article in Quillette. (Michael was kind enough to send me a pdf, so I presume he wants my take.) Read it by clicking on the screenshot below, or find it archived for free here.
In short, Shermer is somewhat of a compatibilist—or so I think, for though doesn’t seem to fully on board with libertarian “you-could-have-done-otherwise” free will, but neither does he accept physical determinism. Further, he doesn’t seem to think that “you could have done otherwise” is even testable, as we’re never in the same situation twice.
He’s right about the untestability criterion. But that doesn’t matter, for even if we were in the identical situation, with every molecule in the universe exactly as it was the first time, there are fundamentally unpredictable events of the quantum kind that might lead to slightly different outcomes. And the more distant in the future we look, the more divergent the outcomes will be. I’ve already noted that the future is probably not completely determined because quantum events could be cumulative. In evolution, for instance, natural selection depends on the existence of different forms of genes that arise by mutation. If quantum effects on DNA molecules can lead to different mutations, then the raw material of evolution could differ if the tape of life is rewound, and different things could evolve.
Further, if quantum phenomena affect neurons and behavior, it’s possible—barely, possible, I’d say—that in two identical situations you could behave differently. I don’t believe that, and neither does neuroscientist Robert Sapolsky, but quantum phenomena that affect molecular movement or positions do not give us free will, as our “will”, whatever that is, doesn’t affect the physical behavior of matter. And so, if we use Anthony Cashmore’s definition of “free will” as given in his 2010 paper in PNAS (the paper that made me a determinist), fundamentally unpredictable quantum effects do not efface free will. Cashmore:
I believe that free will is better defined as a belief that there is a component to biological behavior that is something more than the unavoidable consequences of the genetic and environmental history of the individual and the possible stochastic laws of nature.
Cashmore takes care of quantum effects by lumping them as the “possible stochastic laws of nature.” (Some physicists think that quantum mechanics is really deterministic though it seems otherwise.)
But Shermer doesn’t talk much about quantum physics—in fact, he doesn’t mention it at all. He simply argues by assertion, saying that yes, we could have done otherwise, and we could have done so on the rather nebulous bases of “self-organization” and “emergence”. Let’s take the assertions first. I’ll have to quote at greater length than usual:
Since philosophers love to employ thought experiments to test ideas, here’s one for you to consider (feel free to plug yourself and your spouse or significant other into the situation): John Doe is an exceptionally moral person who is happily married to Jane. The chances of John ever cheating on Jane is close to zero. But the odds are not zero because John is human, so let’s say—for the sake of argument—that John has a one-night stand while on the road and Jane finds out. How does John account for his actions? Does he, pace the standard deterministic explanation for human behaviour (as in Harris’s and Sapolsky’s definitions above), say something like this to Jane?
Honey, my will is simply not of my own making. My thoughts and intentions emerge from background causes of which I am unaware and over which I exert no conscious control. I do not have the freedom you think I have. I could not have done otherwise because I am nothing more or less than the cumulative biological and environmental luck, over which I had no control, that brought me to the moment of infidelity…
Could John even finish the thought before the stinging slap of Jane’s hand across his face terminated the rationalisation? If free will is the power to do otherwise, as it is typically defined by philosophers, both John and Jane know that, of course, he could have done otherwise, and she reminds him that should such similar circumstances arise again he damn well better make the right choice… or else.
This is argument against free will by assertion alone. What his wife is evincing here is her illusion of free will. Nobody denies the fact that we feel that we could make real choices. But that doesn’t mean that we do.
But where’s the evidence that John Doe could have refrained from his one-night stand? He is correct in thinking that he could have not done otherwise (how could he unless some undefinable, nonphysical “will” affected his libido?), but his wife, subject to the universal illusion that our behavior is more than “the unavoidable consequences of the genetic and environmental history of the individual and the possible stochastic laws of nature”, believes in some undefinable property called “will” that could change the outcomes of a given situation. She thinks that John could have chosen not to fall prey to the allure of that other woman.
So Jane gives John a slap (that slap, too, was determined). And the slap could change John’s future behavior so that he refrains from other affairs, for, like all vertebrates, we learn from experience. That’s the result of evolution. (Keep kicking a friendly dog and see how long it remains friendly!). He concludes what’s below (bolding is mine): But nobody with any neurons to rub together argues that changing behavior via learning somehow violates determinism.
More from Shermer:
But this is not the universe we live in. In our universe (unlike the one in which thought experiments are run), the Second Law of Thermodynamics and entropy means that time flows forward and no future scenario can ever perfectly match one from the past. As Heraclitus’ idiom informs us, “you can’t step into the same river twice,” because you are different and the river is different. What you did in the past influences what you choose to do next in future circumstances (the technical name for this is “learning”), which are always different from the past. So, while the world is determined, we are active agents in determining our decisions going forward in a self-determined way, in the context of what already happened and what might happen. Our universe is not pre-determined but rather post-determined, and we are part of the causal net of the myriad determining factors to create that post-determined world. Far from self-determinism being a downer, it’s the ultimate upper because it means we can do something about the future, namely, we can change it!
I don’t really understand this paragraph, nor the part in bold. In what sense are we active agents in determining our decisions in the future? Shermer doesn’t tell us, but he seems to be thinking of some nonphysical power of “will” to change the physics that governs our brains and behaviors. In fact, there is redundancy here: we determine our decision because our behavior is self-determined!
Apparently Shermer rejects physical determinism because, given the present, more than one future is possible. The laws of physics are likely to be, at bottom, unpredictable, though their effects on “macro” phenomena are probably minimal, and their effects on the behavior of human and other creatures is unknown. Shermer is even somewhat rude to determinists like Sam Harris and Robert Sapolsky (and, implicitly, me, as I’m with them): we are hidebound reductionists plagued by “physics envy” (bolding is mine):
Do determinists really fall into the trap of pure reductionism? They do. Here is the determinist Robert Sapolsky defending his belief that free will does not exist because single neurons don’t have it: “Individual neurons don’t become causeless causes that defy gravity and help generate free will just because they’re interacting with lots of other neurons.” In fact, billions of interacting neurons is exactly where self-determinism arises. But Sapolsky is having none of that: “A lot of people have linked emergence and free will; I will not consider most of them because, to be frank, I can’t understand what they’re suggesting, and to be franker, I don’t think the lack of comprehension is entirely my fault.”
Determinists like Harris and Sapolsky have physics envy. The history of science is littered with the failed pipe dreams of ever-alluring reductionist schemes to explain the inner workings of the mind—schemes increasingly set forth in the ambitious wake of Descartes’ own famous attempt, some four centuries ago, to reduce all mental functioning to the actions of swirling vortices of atoms, supposedly dancing their way to consciousness. Such Cartesian dreams provide a sense of certainty, but they quickly fade in the face of the complexities of biology. We should be exploring consciousness and choice at the neural level and higher, where the arrow of causal analysis points up toward such principles as emergence and self-organisation.
So what is there to behavior beyond atoms moving around according to physical principles? Shermer doesn’t tell us, but he seems determined (excuse the pun) to convince us that we do have free will, and it seems to be of the libertarian sort! He even evokes the mysteries of consciousness, which many people, including Francis Crick, think is best studied not from a “top down” approach, but from a reductionist “bottom up” approach. And we know from various experiments and observations that we can affect our notion of “will”, making us seem like we have it when we don’t (people who suddenly confabulate a purpose when they behave according to stimulation of the brain), or making us seem like we lack it when we are actually acting deterministically (e.g., ouija boards). We can take away consciousness with anesthesia, restore it again, or alter it with psychedelic drugs. All this implies that yes, consciousness and “will” are both phenomena stemming from physics.
Shermer rejects bottom-up approaches, raising the spectres of “self-organization and emergence” as arguments against Cashmore’s form of free will:
This we have through the sciences of complexity, in which we recognise the properties of self-organisation and emergence that arise out of complex adaptive systems, which grow and learn as they change, and they are autocatalytic—containing self-driving feedback loops. For example:
Water is a self-organised emergent property of a particular arrangement of hydrogen and oxygen molecules.
Complex life is a self-organised emergent property of simple life, where simple prokaryote cells self-organised to become more complex eukaryote cells (the little organelles inside cells were once self-contained independent cells).
Consciousness is a self-organised emergent property of billions of neurons firing in patterns in the brain.
Language is a self-organised emergent property of thousands of words spoken in communication between language users.
That list goes on, but it’s muddled. First, what do we mean by “self-organized” properties? Is water “self organized” beyond behaving in a glass in ways that are consistent with, but not necessarily predictable from, the behavior of a single water molecule? Ditto for complex life. In what sense are life and water “self-organized” rather than “organized by physics”? Yes, there are emergent properties, like the Eroica emerging from the pen of Beethoven, himself an admirable collection of organic molecules with the emergent property of writing great music.
Let’s dismiss “self-organization,” which seems like a buzzword that doesn’t advance Shermer’s argument, and concentrate instead on “emergence.” Yes, water is wet. “Wetness” is a quale evinced in our consciousness, yet the properties of water that make it feel wet are surely consistent with, and result form, the laws of physics, just as the “pressure” of gas in a container is an emergent property of a bunch of gas molecules acting as a group. But nobody says that gas molecules have free will, even though some of their properties are “emergent.”
The issue here is not whether emergence is something predictable from a reductionist analysis, but whether it is something physically consistent with its reductionist constituents. If the laws of physics be true, then that consistency does nothing to efface determinism. Shermer’s failure is that he neglects to tell us the nature of something called a “will” that interposes itself between molecule and behavior. And often, with greater knowledge of physics we can predict emergent properties from a reductionist analysis. (The gas laws are one such thing.)
I’ll draw this to a close now, adding one more note. Shermer’s failure is twofold. He fails to suggest how an undefined “will” can affect the behavior of matter, and he mistakes determinism for predictability, a rookie error. If quantum mechanics is a good explanation of physics, then the future is not 100% predictable, even if we had perfect knowledge of everything, which of course we don’t. And physicists tell me that quantum effects were important at the Big Bang, so at that moment the future of the entire universe was unpredictable. That says nothing about free will.
Shermer closes with another paragraph that I don’t understand; it sounds in some ways (this may anger him, but I apologize) like Deepak Chopra:
It may seem odd to think of yourself as a past-self, present-self, and future self, but as suggested in this language, your “self” is not fixed from birth, destined to a future over which you have no control. We live not only in space, but in time, and as such no matter the pre-conditioning factors nudging you along a given pathway—your genes, upbringing, culture, luck and contingent history—there is always wiggle room to alter future conditions. The river of time flows ever onward and you are part of its future.
Act accordingly.
This is more argument by assertion alone. I’m not sure what he means by “act accordingly”, much less “wiggle room.” Of course we can be influenced by what we read, but we don’t have a “will” that could alter what we do at any given moment. As Cashmore said in his article:
Here I argue that the way we use the concept of free will is nonsensical. The beauty of the mind of man has nothing to do with free will or any unique hold that biology has on select laws of physics or chemistry. This beauty lies in the complexity of the chemistry and cell biology of the brain, which enables a select few of us to compose like Mozart and Verdi, and the rest of us to appreciate listening to these compositions. The reality is, not only do we have no more free will than a fly or a bacterium, in actuality we have no more free will than a bowl of sugar.
I don’t mind being like a bowl of sugar, or, rather, a complex piece of animated meat. I admire Shermer for all he’s done to further skepticism and attack quackery, but I think that on the issue of free will he’s gone awry.
From AI:
