Over at UD, Salvador Cordova criticizes a statement by Jerry Coyne that appeared in a USA Today column:
So if we don’t have free will, what can we do? One possibility is to give in to a despairing nihilism and just stop doing anything. But that’s impossible, for our feeling of personal agency is so overwhelming that we have no choice but to pretend that we do choose, and get on with our lives.
There are a lot of problems with Cordova’s post (which we can address in the comments), but I do agree that Coyne’s statement is problematic and warrants criticism. Ironically, Coyne’s misunderstanding is shared by many ID proponents — hence the title of this post.
Coyne’s error is in this statement:
…our feeling of personal agency is so overwhelming that we have no choice but to pretend that we do choose…
There’s no need to pretend, because we actually do choose — even if free will is illusory, and even if our choices are determined. Look at the definitions Google provides for the word ‘choose’:
1. Pick out or select (someone or something) as being the best or most appropriate of two or more alternatives.
2. Decide on a course of action, typically after rejecting alternatives: “he chose to go”.
These definitions work even if free will doesn’t exist, and they work even if every choice we make is ultimately determined by forces outside of our control. Determined choices are still choices — we gather information, we evaluate alternatives, and we make a decision based on our desires and our reasoning.
[Note: For the record, I am a compatibilist, meaning that I think we have free will, in a meaningful sense, even if determinism is true.]
As if on cue, the first comment in Cordova’s thread exemplifies the misunderstanding I lament in the OP:
Spot on – from one compatibilist to another. If choosing is something different from a determined/random activity then how do we know it is? What is there about the experiencing of choosing that tells us it is not determined/random?
Moreover, there is considerable evidence indicating that representation of one another as agents whose behavior arises out of what we see, know, believe and desire – e.g., the theory of mind literature – reflects evolutionary adaptations by means of which we conduct our profoundly social existence. We have “no choice” but to represent ourselves in this way in our day do day intercourse not because compelled by a philosophical conundrum, but because representing one another in these terms is a fundamental component of the suite of adaptations by means of which human beings survive.
This is a test post, Sunday May 12 10:54 EST.
Given the fact that the emergence of phenomena in complex assemblies of matter is itself so complex and unexpected, it should not be too surprising that the products of the mental activities of sentient beings would also have all the attributes of having been produced by the exercise of free will.
If mental activity itself is an emergent property of matter, then certainly the products of mental activities would simply be still more emergence on top of that.
Thus the products of mental activity would not be predictable from any deterministic view of the universe. Determinism is mostly a leftover illusion stemming from classical physics, which was, in retrospect, a very naive and incomplete understanding of matter and energy.
However, the emergence of new properties and new phenomena with increasing complexity has pretty much obliterated any notion that determinism is a workable notion in the real world of quantum mechanics and condensing matter.
Mike,
I’m not sure why you think that emergence isn’t compatible with determinism. They seem orthogonal to me. I can imagine deterministic forms of both strong and weak emergence.
Determinism to how many decimal places? Determinism applies to idealized classical systems.
Even Henri Poincare had already noticed the sensitivity to initial conditions of the subsequent trajectories of dynamic systems; and this is with idealized classical systems.
During the 1960s and 70s an entire field of mesoscopic physics emerged that explores the boundary between quantum indeterminacy and classical indeterminacy. These merge in any real systems; but this might not be apparent in their idealized representations.
One simply cannot snap two or more atoms or molecules together without some indeterminacy. Energy has to be released in at least the form of a photon or a number of photons that leave the system and interact with the rest of the universe. This leaves the bonding system in an indeterminate state; and the more complex the system, the more indeterminate its state.
The indeterminate level is not just at the quantum level, the system has macroscopic motion that is sensitive to its environment both classically and quantum mechanically. Systems at finite temperature are immersed in a heat bath of photons, phonons, or particles that spread the microscopic states of the system into some kind of probability distribution, either a Maxwell-Boltzmann, a Fermi-Dirac, or a Bose-Einstein distribution depending on the nature of the system.
The idealized representations of simple mechanical systems may show complete determinism out to as many decimal places as one cares to calculate; but real systems simply cannot be completely deterministic. The mere fact that matter interacts with matter and with electromagnetic and gravitational fields in just our baryonic universe alone makes strict determinism impossible.
This is not a nit-picking quibble on the meaning of words. The existence of condensed matter systems is a result of the second law; energy has to be released and spread around. The release of that energy – its magnitude and direction – leave a system in a necessarily indeterminate state. You can’t escape it.
I keep pointing out that little high school level physics chemistry calculation that scales up the charge-to-mass ratios of protons and electrons to kilogram-size masses separated by a meter. The energy of interaction is on the order of 10^26 joules or 10^10 megatons of TNT.
Now consider the fact that atoms and molecules in close proximity interact and radiate energy due to their accelerations and charge redistributions as they swing around each other and bind or scatter depending on their kinetic energies. The radiated energy is not an insignificant fraction of the energies of the interacting atoms or molecules. At the atomic/molecular level, electromagnetic interactions are by far the dominant interaction; being something like 10^36 times larger than gravitational interactions.
Consider the fact that everything around you is at some temperature. Every system, living or non-living, is exchanging photons, phonons, or particles with its environment. You simply cannot have such systems in completely determined states.
Soft-matter systems associated with biological organisms are among the most complex of systems that exchange energy and matter with a larger environment while existing within a very narrow energy window. There is nothing determinate about soft-matter systems; not just in the quantum mechanical sense, but in the classical sense as well.
Mike,
Your comment answers a different question, which is “Do you think mental activity is deterministic?”
I’m asking why you think that emergence implies indeterminism, as you seem to be claiming here:
That seems wrong to me. Emergence and determinism are fully compatible, as far as I can tell.
Seems to me Coyne is simply saying that IF determinism is the case, (1) we can never know that, and (2) what difference does it make?
I think Mike is saying that determinism is a sort of hare-brained notion arising out of a simplistic understanding of obsolete classical physics. It doesn’t actually MEAN anything, in the sense that it has no operational definition.
keiths,
Flint puts it a bit more concisely than I did; determinism has no operational meaning in real condensed matter systems. It is an extrapolation of classical physics in the limit where there are no other mechanisms of interaction other than classical gravity and Newton’s laws of motion.
In that world, matter consists of hard spheres or point particles with no other properties than mass. Further classical extrapolations include things like idealized springs; but all the questions of how material objects came to be assembled in the first place are not even part of the picture.
In real matter, all sorts of interactions are taking place. In soft-matter systems, such as us, there is no way to escape the contingencies of these interactions. The operations of our neural networks hang on the verge of chaos, in that a slight increase in temperature would make these networks go chaotic – and eventually come all apart – and a slight decrease in temperature would shut them down completely.
So our thoughts and our thinking processes exist in a world of contingency at the most fundamental level. Strict determinism has no real application to systems such as our neural networks.
Introductory physics and chemistry courses – as well as those science courses for non-majors in college – give little hint of what experimental physics is really like. We experimentalists spend much of our time trying to isolate the effects of the phenomenon we are investigating. That isolation is artificial in the real world; it is nearly impossible to totally isolate a system down to one type of interaction and still collect data from it.
The emergence of phenomena in complex systems is the result of multiple interactions among the constituents of the system and between the constituents and the surrounding environment. The complex systems of living organisms are not just complex, they are heterogeneous systems; and that results in the emergence of phenomena that are even more unpredictable and contingent.
Flint,
No, in that USA Today op-ed he actually says that we do know that our decisions are determined, and that this means we do not have free will.
He defines free will as the ability to make undetermined choices:
Since he denies that we have free will, he is therefore affirming that our choices are determined.
Determinism is actually an extremely useful notion both in philosophy and in other fields such as computer science. It is possible to build deterministic systems on an indeterministic substrate, and computers are a prime example of this.
Mike,
Keep in mind that chaos, in the technical sense of the word, is actually deterministic — so to claim that neural networks are chaotic is really an argument for determinism, not against it.
The computation is deterministic. The computer probably isn’t.
It depends on the sense in which you are using the word ‘deterministic’. If a computer has interrupts enabled and is running an operating system that uses virtual memory, its state will not be deterministic with respect to time or physical addresses, but it may still be deterministic with respect to the virtual address space at each instruction boundary.
To carry out a deterministic computation, the computer itself has to behave deterministically in some critical respects. If ‘MOV EBX, EAX’ doesn’t deterministically copy the contents of EAX into EBX, your program is in big trouble. The great strength of digital information processing is that it uses nonlinearity to immunize the system against a certain amount of noise, thus preserving this determinism.
Anyway, determinism ceases to be very important in the free will debate once you understand why libertarian free will is incoherent and why indeterminism can’t rescue it.
Deterministic chaos takes place in classical, closed, isolated systems. In reality, very few systems are sufficiently classical, closed, and isolated for deterministic chaos; and such systems cannot fulfill the requirements of a living organism. They are modeled on computer by removing all interactions with an external environment. Real systems interact with a wider environment; and the interactions among constituents involve the radiation of energy in some form.
All biological systems are open; they could not exist if they weren’t. They exchange matter and energy with a larger environment. They have to be in some sort of heat bath.
When I used the word chaotic in the case of the nervous system, that example would not be deterministic chaos; it would be chaos in the more colloquial sense. Energy is being added to the system so that the average kinetic energy per degree of freedom (i.e., the temperature) is increasing. The system is proceeding in the direction of coming apart.
There may be some differences in the way that philosophers use the idea of determinism; but in classical physics it means that, if you know the initial positions and momentums of all the particles that make up a system, you can in principle, if not in practice, know the positions and momentums of everything into the past and future. Deterministic chaos brings in the fact that a non-linear shstem or a system of sufficient complexity does not permit sufficiently precise knowledge of intitial conditions to be able to predict outcomes; but patterns do emerge.
But that is not what real systems are like; not even in principle. In order for complex systems – such as the soft-matter systems that make up living organisms – to even exist, there must be exchanges of matter and energy with a wider environment; and there must be a temperature range in which the system can maintain the billions of degrees of freedom that allow the emergence of organization and coordination among the constituents. Temperature can only be maintained by being in “contact” with a larger heat bath.
I am not sure I understand what is meant by “libertarian free will;” but free will is not unrestricted. We don’t defy gravity and jump naked off cliffs without consequences. We still operate within constraints; but there are enough degrees of freedom and contingency that we have broad ranges of choice. The better choices generally tend to preserve the integrity of the rather fragile systems that make those choices. Learning those choices, or having the tendency to make those choices being “wired into” the nervous system, leads to a higher probability of survival.
So I think that the bottom line with living systems in particular is that determinism is irrelevant to any philosophical arguments about free will.
I should add that there are driven chaotic systems that do not conserve energy. The energy supplied to them keeps the energy of the system constant by replacing energy lost. These are often referred to as being in dynamic equilibrium; energy or matter flow is constant. The paths these systems take in phase space cluster around “attractors” and produce patterns that remain relatively static for a given rate of energy or matter flow.
But the soft matter systems of biology do not fit into that category; they are too heterogeneous and they fluctuate and do not trace out constant patterns in a “phase space.” If they did, they couldn’t adapt to changing environments. There is a lot more “slop” in biological systems.
The mere fact that they mutate keeps them in a state of constant change; and they mutate because they are soft-matter systems.
You’re an ex-smoker (like Lizzie), I think. My problem as a non-smoker is trying to understand why theists think determinism and free will make any sense to start with.
The theological approach to understanding seems not to work because of the reluctance to use reality as a test for philosophical or religious concepts. Testing models (Mike’s comment is apposite) against reality has outstripped the philosophical approach.
I don’t think Cordova understands many of the issues here, but I’m fairly sure Coyne doesn’t, either.
One very tricky point worth making is that, in general, there’s only a methodological distinction between our theoretical vocabulary and our observation vocabulary. (There’s an argument for this, but I’ll treat it here as a point of departure.) So it can be difficult to know if “free will” and “determinism” are observation-terms (part of how we “immediately” classify and describes our experiences) or theoretical-terms (part of the conceptual framework by means of which we explain, predict, and manipulate the world). And which theoretical terms we’re committed to using depends on what theories are currently accepted in light of the available evidence.
Above, we’ve seen a really good point nicely made: that “determinism” is a theoretical concept. It was originally at home in the theological metaphysics of the early Christian philosophers who worried about human responsibility and divine omniscience, and was then transported over into 17th- and 18th-century metaphysical interpretations of classical physics. But it just doesn’t jive with the metaphysics of contemporary physics, let alone chemistry and biology.
I would say, however, that the same basic point holds for the concept of “free will”: it’s also a theoretical term that arises along with the notion of determinism and has persisted along with it through the centuries. It would be much better, I think, to begin with describing our subjective experience of volition — e.g. of the difference between raising one’s arm and having someone else move one’s arm, or the difference between raising one’s arm and trying to raise one’s arm.
And here a bit of ethology can help, too. It might seem odd to assert that my cat has free will — and equally odd to deny it! But it’s not odd to point out that my cat can perform intentional actions, do things because he wants to do them. The relevant difference here between me and my cat is that I take adopt the standpoint of rationality and ask if my motives and inclinations are good (enough) reasons for my actions.
So I think that very question, “is free will compatible with determinism?” is a bad question. A much better question would be, “is our subjective experience of agency compatible with our objective knowledge of emergent processes?”. And to that question, I think that the answer has got to be a tentative yes — tentative because I am not at all confident that subjective experience and objective knowledge can be unified within a single, comprehensive metaphysical system.
I think one of the first stumbling blocks in any “free will” discussion is that the term itself is a tautology. “Unfree will” doesn’t make any sense. So the issue it seems to me is not whether will is free, but whether will exists. In other words, is there such a thing as volition? And it seems to me there patently is – to call it “illusory” is as silly as saying that “love” or “justice” are “illusory”. It’s a perfectly good word to describe our ability to make informed choices. The more tricky question is: whose are those choices? In other words, the question is not “is will free” but “who is free?” And I think an sentient organism is a perfectly valid unit to regard as a volitional agent.
Mike, you are conflating determinism with predictability. Determinism (barring inevitable philosophical nitpicking) is a fairly simple concept: an event is determined in some theory if there is a set of prior events that, combined with the laws of the theory, make this event a necessary outcome. This can be expanded to the entire world at any given time slice.
Whether the world is deterministic in the above sense, and whether this is even knowable, is an open question. It is not even clear whether quantum physics is indeterministic. In one sense QM may be seen as deterministic: after all, it describes the state of a system as a deterministic evolution of its wave function (that is, if we give up wave function collapse of the Copenhagen interpretation as a genuine feature of the theory). Carl Hoefer, in the SEP article on causal determinism, having reviewed other fundamental physical theories (including even Newtonian physics), concludes that “quantum mechanics is one of the best prospects for a genuinely deterministic theory in modern times!” In practice though, we are more interested in observables than in the wave function. John Earman, in a similarly scoped review, is more circumspect, concluding that “There is no easy way to summarize the status of determinism in QM.” I should note that Earman is dismissive of the Everett (many-worlds) interpretation, which is often presented as deterministic, but it is held by a substantial number of physicists and cosmologists.
Mike Elzinga,
Mike, you are conflating determinism with predictability. Determinism (barring inevitable philosophical nitpicking) is a fairly simple concept: an event is determined in some theory if there is a set of prior events that, combined with the laws of the theory, make this event a necessary outcome. This can be expanded to the entire world at any given time slice.
Whether the world is deterministic in the above sense, and whether this is even knowable, is an open question. It is not even clear whether quantum physics is indeterministic. In one sense QM may be seen as deterministic: after all, it describes the state of a system as a deterministic evolution of its wave function (that is, if we give up wave function collapse of the Copenhagen interpretation as a genuine feature of the theory). Carl Hoefer, in the SEP article on causal determinism, having reviewed other fundamental physical theories (including even Newtonian physics), concludes that “quantum mechanics is one of the best prospects for a genuinely deterministic theory in modern times!” In practice though, we are more interested in observables than in the wave function. John Earman, in a similarly scoped review, is more circumspect, concluding that “There is no easy way to summarize the status of determinism in QM.” I should note that Earman is dismissive of the Everett (many-worlds) interpretation, which is often presented as deterministic, but it is held by a substantial number of physicists and cosmologists
Lizzie: I don’t think “free will” is a tautology. For one, there is a trivial sense of “free will” as unconstrained will. That is, you both want and can do something (raise a hand, leave a prison cell, fly like a bird). But that’s not an interesting definition, so let’s set it aside. Still, I think there is a cogent concept of free will, distinct from will simpliciter, that is relevant to the usual debates. You can raise a hand voluntarily, or you can raise a hand because you are given an order that you cannot resist. In both cases it’s a willed action (even in the latter case there is, strictly speaking, a choice to be made), but only one can be said to be free. You might also raise your hand involuntarily while trying to regain balance, or unconsciously in your sleep. It’s unclear whether such reflexive and unconscious actions are never willed, but it is safe to say that they are not freely willed.
The distinction between free and not-free will that I was trying to pump with the above examples may seem fuzzy. And so it is, in my view. I have come to regard free will – as the concept is popularly applied – not as a metaphysical category, but as a psycho-social one. And as things in that domain are rather fluid and fuzzy, so is free will. A useful concept that tends to closely track free will is responsibility, or ownership. Freely willed acts are those that we own, for which we assume responsibility and hold others responsible. (Here I mean all actions, not just consequential actions that affect other people, as the word “responsibility” might suggest.) With simple physical actions like raising one’s hand, we intuitively understand when we do that “voluntarily,” and when not. Things get more ambiguous when more complex interactions come into play: does one run out of a burning house voluntarily, or is this action forced by circumstances? Does one voluntarily give up the wallet to an armed robber? Does one voluntarily cooperate with an oppressive regime? Are we fully responsible for things we do in altered states of mind (drunk, drugged, sleep-deprived)? We often say that we “weren’t ourselves” when we were in such states, so can our actions be considered quite free? What about being influenced and manipulated by another person? Or by false impressions, or lack of information? There aren’t always straightforward answers to these questions, but that’s to be expected if one doesn’t take free will (or its lack) to be a simple matter of fact.
This view of free will is not dependent on any physical or metaphysical considerations, and so the question of determinism is strictly orthogonal to the question of whether or not we have free will (clearly we do, and nothing that we learn from physics can change that).
That makes sense to me. Certainly I tend to think in terms of “degrees of freedom”. So I guess will with only two degrees of freedom (“salute or die!”) is a lot less free than most of us would want.
So, OK 🙂 And that makes even more sense of the notion that we havefree will – that we have, mostly, far more than two degrees of freedom.
As I said above, the word “determinism” may be used by philosophers differently from what it is used by physicists.
In physics, determinism means one can track the positions and momentums of a system of particles into the past and future by knowing their initial values and the classical laws of physics (Newton’s laws).
It says, in effect, that no matter what time or position one cares to look into the system, its state at that instant can be traced to its states past or present.
No matter how one chooses to use determinism, the point is that it is irrelevant in any discussion about some “conflict” between “free will” and “determinism.”
The substrates upon which our thought processes are built are not deterministic in any sense of that word. One cannot link any current state of the system unambiguously to any specific state in the past and use a prior state to predict precisely the current state.
The mere assembly of such a substrate involves uncertainties and contingencies that make every substrate different from the substrates of every other individual. The processes that take place on these substrates are subject to temperature and environmental perturbations. Thus their emergences, while they may have similarities among the various substrates from which they emerge, have many major differences.
The “deterministic” evolution of the wave function in Schrödinger’s equation is not an observable. Its absolute value squared is a probability density; it doesn’t determine a specified outcome.
But, as I pointed out above, such indeterminacy in soft-matter systems doesn’t even have to reach down to the quantum mechanical level; it is already part of the assembly and behavior of the system due to the fact that the system has many degrees of freedom and is immersed in a heat bath at finite temperature. The fact that it is highly sensitive to temperature also tells us that the phenomenon of thought emerges from a highly contingent system that is sensitive to the environment in which it is immersed.
I don’t know how deeply to go into the mathematics on a site like this; but in a deterministic system, the world lines of all constituents of a system are “frozen” in a 4-D space-time diagram. Any change in any part of any path in such a diagram means that the entire path is changed.
There are some implicit assumptions underlying such a diagram; one being continuity, another being the assumption of a continuous “time” coordinate, which in turn assumes the existence of “disembodied” clocks.
In reality, clocks are themselves systems that are singled out as reference systems. Thus, “time” is a correspondence between states of a given system and states of the system designated as the clock.
If one projects all world lines of a system backward in “time” in such a diagram, one encounters intersections with world lines coming from outside the system that are within the light cones of the individual constituents of the system. One doesn’t have to go backward very far before the sensitivity to initial conditions becomes so extreme that it drops into the quantum mechanical realm.
But even before that, one has to take into account all the world lines of the billions upon billions of photons, phonons, and particles whose world lines are entangled with the world lines of the system of interest.
So, it doesn’t make any difference if one drops into the quantum mechanical realm or simply runs into enough complexity to make calculations impossible, one quickly arrives at a point where “determinism” has no operational meaning in any real universe.
That is a matter of debate, even if we are only concerned with observables.
While what you say about practical indeterminism is certainly true, I think that many people are disturbed by the prospect of determinism even in principle, making it possible for at least a Laplace’s demon, if not anyone we know, to have knowledge of all that we are going to do in the future. That picture of the world frozen in the 4D spacetime is what gives many people the willies.
Perhaps that concern is misguided, regardless of whether determinism is the case*. There is only one way that the world will eventually evolve, so from a timeless perspective it is a frozen 4D block, any way it goes. Whose perspective that is, you might ask? Well, you got me there. But then, the Lapace’s demon is not a real perspective either.
* or whether it even makes sense to say that one or the other is the case. In the context of such discussions, a certain reductionist fundamentalism is tacitly assumed. But that position itself should not be taken for granted.
Real perspectives involve making predictions from past observations using world models. Determinism/indeterminism is a property of models, and that property serves as the lowest theoretical bound on predictability. But as you say (and no one that I know disagrees), that theoretical bound is unreachable in any case.
SophistiCat,
It is not possible, even in principle, for a condensed soft-matter entity in a 4-D space-time to measure the positions and momentums of anything within itself and use that information to predict its future or its past.
For one thing, it has no way of knowing when and how many photons, phonons, or particles will intersect the world lines of any of its own constituents. There is no observational technique that allows, even in principle, any observation of an incoming photon, phonon, or particle without disturbing their world lines as well as the world lines of any of its own constituents, thereby changing the world lines of the incoming as well as its own constituents.
All decisions it makes within its neural networks involve the emission of photons, or of phonons interacting with other matter and eventually dissipating their energies in the form of photons that leave the system along the light cones of all of the constituents of the system. Those decisions involve energy input in the form of phonons and photons and particles coming in on their own world lines.
The idea of a Laplacian demon has no operational method of confirmation – it is little more than a gussied-up version of a supernatural being for which we cannot conceive of any way it can interact with the natural world and communicate what it knows to the world lines of our constituents.
If it did somehow do that, we would perceive it as a violation of conservation laws that changed the world lines of the constituents that make up us. What would be the nature of interactions with a Laplacian demon that could change the world lines of our constituents? They would have to be baryonic in some way; and if that were so, then what is the nature of the Laplacian demon?
The notion of determinism may be useful in simple or idealized models; but there is no operational procedure that exists, even in principle, that permits a sentient, soft-matter entity to learn anything about it. Thus, just as with the notion of a supernatural being, determinism is an idea without any operational handle on its existence.