(or, How the way I talk about the free will problem hasn’t much changed since 2007)
Time for another “blast from the past” post, the past once again referring to my days of writing essays in university. Back in my college days, I was very passionate about philosophy (in many respects, more so than today!), but I was also very lazy when it came to studying what actual philosophers through the ages had to say. I guess my attitude has always been inclined towards reinventing the wheel. Accordingly, I was the only heavily committed member of the undergraduate philosophy club (eventually I became its vice president) who wasn’t a philosophy major or even a minor. I don’t think I even tried to register for a philosophy course until my third year.
But eventually, with some elective credits left to fill, I decided to sign up for a sort of easy Philosophy 101 course which was called something like “Intro to Philosophy”. This was what was known as a BS course (and the acronym BS didn’t stand for Bachelor of Science). We were given a much-too-perfunctory rundown of most of the ethical and metaphysical puzzles that have plagued humankind since the time of the ancient Greeks in a series of lectures by the main instructor which I don’t recall much about.
The aspect of the course I found slightly more engaging was the discussion section meetings, which were run by a graduate student who led my section (a small subset of the class members) in discussions and debates over the topic of that week’s lecture. My main recollections of the discussion leader involve a young guy sitting back in his chair pausing between delivering philosophical speculations, clucking his tongue, cocking an eyebrow, and starting out again, “Or maybe…”.
I guess my experience was about what one could reasonably expect from a General Education course, and being someone who found serious essay-writing stressful (at least when it was for a grade!), I was sometimes grateful not to be taking one of the heavier philosophy classes that some of my fellow philosophy club members were doing. We did have to do a little writing for the discussion section leader, though, in the form of a couple of fun mini-essays called squibs (a term I’ve otherwise never encountered outside of Harry Potter). The first course topic was free will, and this was the inspiration for our first squib assignment.
Recently I found the Microsoft Word document of what I submitted for this assignment, and I thought I’d put it here and comment on it as I did for the Bertrand Russell project I put on this blog a few weeks back. It’s dated September 17th, 2007, so I think it’s reasonable to assume I was taking this course for that fall semester and that this was right near the beginning. Unfortunately, I can’t recollect much about the prompt for this mini-essay, apart from the fact that we were given a couple of short texts representing two opposing opinions on the free will problem, which apparently were works of philosophers Peter Van Ingwagen and A. J. Ayer. I haven’t looked into what paper of Ayer we were given, as his name is only briefly mentioned in my squib, but I’ve deduced that the Van Ingwagen article I was firmly opposing must have been “The Incompatibility of Free Will and Determinism”, which can be found here. I recall getting a B or B+ on this assignment and feeling kind of disappointed. Admittedly, my present-day self finds my own argument against Van Inwagen’s 7-step demonstration of incompatibilism mildly difficult to follow, and perhaps the discussion section leader felt similarly.
Anyway, here’s the squib.
Like all frequently discussed philosophical ideas, both free will and determinism are concepts that are relevant to the way we process the world around us. In particular, the notion of free will is important to us in analyzing how we humans are able to control our actions and how moral responsibility for them should be assigned. For us humans, it is natural to covet free will, in hopes that one is truly in control of one’s own choices. However, determinism is viewed by many to be a threat to the free will that we so desire. Peter van Inwagen, who attempts to demonstrate in his paper that determinism implies a lack of free will, evidently holds such a view. However, I feel that the free will that we naturally crave can still be defined in a way such that van Inwagen’s arguments against it do not apply. Furthermore, this form of free will would actually imply determinism rather than render determinism impossible. In this paper, I will first argue that van Inwagen’s argument relies on an amorphous definition of free will, and then I will argue that, contrary to what van Inwagen asserts in part IV of his paper, the type of free will that is relevant to us does rely on determinism.
Van Inwagen begins his argument against compatibilism by defining free will as concerning “the power or ability of agents to act otherwise than they in fact do”. He does not specify exactly what he means by “power” or “ability”, but does acknowledge that the verb “can” (which is clearly related to “power” and “ability”) is difficult to define. He does not go on to attempt a definition, but rather, he proceeds to revamp the language we use in “talking about ability and inability in complicated cases”. However, his translation of the typical “can” statement still uses the word “can” and therefore does not answer any of our questions concerning what “can” actually means. Hence, van Inwagen’s idea of free will is not well defined.
It seems to be very difficult to come up with a coherent definition of “can” that does not use words like “power” or “ability”, but one can still come up with several possible candidates. For instance, “can” may be defined in terms of what “might be”. Thus, the statement “I can drive to [nearby city] tomorrow” would mean, “I might drive to [nearby city] tomorrow.” But if one believes in an objective reality, it is clear that either I will drive to [nearby city] tomorrow, or I will not; there is no uncertainty except in the limited minds of humans. Hence, such a “might” statement is meaningless unless we are referring to a lack of knowledge as to whether or not I will drive to [nearby city] tomorrow. This definition seems to make the truth of “can” statements, as well as the existence of free will, impossible regardless of the validity of determinism. On the other hand, one can define, “I can do X” as something along the lines of, “No physical barriers outside of what goes on in my Decider will (by themselves) cause me not to do X.” This is an example of a possible meaning of “can” which van Inwagen ignores and for which van Inwagen’s argument does not succeed. (Note that this is only one example, and probably not a sufficient definition for the “can” required for free will.)
Some may question my apparent assumption of the existence of an objective reality in the above paragraph. To the minds of some thinkers, it is most logical to analyze the world subjectively (since we cannot prove an objective reality), and therefore, the “might” definition has plenty of meaning because of how little we know for sure. I would respond by saying that if we are assuming subjectivity, then the entire topic of discussion makes no sense, because the notion of determinism is meaningless in a subjective form of reality. Determinism relies on an objective view of the universe, of the existence of entities outside of our own knowledge. We do not know the state of every particle in the universe or the absolute certainty of all laws of physics, and hence, if reality is analyzed purely subjectively, then determinism has no meaning. If determinism has no meaning, then the question of whether or not it is compatible with free will has no meaning.
Van Inwagen’s argument, when one adopts the definition of “can” as given above (“I can do X” means that no barriers outside of my Decider cause me not to do X), fails at step 4. His step 4 states that “if J could have rendered P false, and if the conjunction of P0 and L entails P, then J could have rendered the conjunction of P0 and L false.” Translated, this reads, “If no physical barriers outside of J’s Decider caused P to be true, and if the conjunction of P0 and L entails P, then no physical barriers outside of J’s Decider caused the conjunction of P0 and L to be true.” But this has no validity: it could be that the conjunction of P0 and L was caused to be true by physical barriers outside of J’s Decider, and that P0 and L entail P by means of a chain of causes and effects that operate only inside of J’s Decider. If this is the case, then it is still true that no physical barriers outside of J’s Decider caused P to be true. Therefore, step 4 of van Inwagen’s argument does not hold for that particular definition of “can”. In this way, we can see that van Inwagen’s argument against compatibilism is not watertight for particular definitions of “can”. Furthermore, it might be argued that the types of “can” which lead to the most relevant forms of free will for us humans (such as the type of free will described by Ayer in his paper) are examples of such definitions which render van Inwagen’s argument inconclusive.
I would go further than that, however. I would argue that the meaning of “can” which leads to the form of free will that we, as humans, desire to have is not only compatible with determinism but dependent on some degree of determinism. Van Inwagen is obviously aware of this position and touches on it in part IV, although he chooses not to tackle the details. He does make two central claims: first, he seems to agree that the type of free choice that makes sense to us relies on events having causes, and second, he claims that every event having a cause in this way does not imply determinism. I argue that the latter claim is not true. Suppose that some sort of free will exists so that at this instant I have a choice: I could finish this paper to turn in by midnight (choice A), or I could give up and throw my laptop through the window (choice B). I choose A, and even van Inwagen would agree that something is causing me to do so. If it is the sort of free will that we find meaningful in regard to moral responsibility, then the causes of my choice must somehow involve my personal character; let’s call this part of the cause C. Van Inwagen would agree that C caused me to choose A, but he would claim that C (in conjunction with all other prior circumstances as well as the laws of nature) did not determine that I would choose A. In other words, with C remaining the same, I might have chosen B instead. But then, my personal character, C, is bizarrely inconsistent. There is absolutely no reason to give credit to my personal character if the exact same character under the exact same circumstances might also lead me to throw my laptop through the window. If we cannot give credit to my character for my actions, then there is effectively no point in having free will, and it is not the type of free will that we so crave. (Of course, I am simplifying the situation; in reality, more of my internal attributes such as personal experience and immediate desire would come into play to give me moral responsibility.)
For these reasons, it seems to me that van Inwagen’s assertions of the incompatibility of determinism and free will are not true. His central argument relies on an undefined verb, “can”, and ignores possible definitions of “can” which would render free will compatible with determinism. In fact, it is impossible for determinism not to operate at some level while the intuitive form of free will exists. It seems to be human nature to desire a certain type of free will, and to my thinking, this type faces no threat from the possible truth of determinism. On the contrary, determinism is, in a sense, free will’s best friend.
 I’m not sure this is the most appropriate use of “covet”, which usually implies a situation of jealousy. However, in a weird way it’s refreshing to see my 11-years-younger self using a word I find jarringly unexpected, in the midst of the exact same phrases and expressions I use when talking about free will stuff on this blog!
 I probably wouldn’t attack this interpretation of “can” in terms of belief or disbelief in an objective reality today. I’ll attempt a rephrasing and expansion of this point below.
 One of my favorite anecdotes for years about that class was the fact that, in order to avoid committing to answers to conundrums like the mind-body problem, we were required to use slightly pedantic yet quite nontechnical-sounding words like “thinker” in place of “brain” or “mind”. What I wrote in this essay is evidence that in fact we referred to one’s mind as one’s “Decider” with a capital D, at least in the context of choice-making.
 This phrasing made me take a double-take at first, and I realized that I would now convey its meaning by inserting the quantifier “some” before “particular”. I think my evolution in how I use these grammatical constructions has come with my vastly increased experience in mathematical writing: there is a crucial difference between the sentences “We cannot apply this result for particular values of x” and “We cannot apply this result for some particular values of x“.
It seems fair to say that I was a pretty staunch compatibilist in 2007, and I’d say I haven’t changed much since then. Here are two of my college-student self’s main points which I’d like to reemphasize perhaps a little more lucidly.
First of all, the whole “can” thing. I still believe that one of the main issues incompatibilism has to deal with but seems unable to is whether a definition for “can” (and all the related words like “ability”, “power”, etc.) can be found or “can” is meaningless in the most purely metaphysical sense. It it’s meaningless, then the only choice remaining to incompatibilists is to assert that free will is impossible because it’s meaningless so a fortiori it is incompatible with determinism. As far as I’m concerned, when some universally intuited concept is rendered meaningless under the current definition, it’s time to look into some alternate definitions, which sets us on the road to compatibilism.
It seems to me that in our everyday language, “can” is used in three senses that I can think of. The first sense is the one I discussed in the squib, although I’m not sure I explained myself using the clearest language there. I say “I can go to the supermarket in an hour” when nothing seems to be preventing me, which is almost another way of saying that I might go to the supermarket in an hour. But this is essentially nothing more than an expression of my limited subjective knowledge of what I might or might not do; in objective reality, either I will go to the supermarket in an hour or I won’t. The second sense is in the context of a physical law; for instance, “Nothing can travel faster than the speed of light.” Without getting into the full discussion of exactly what physical laws are, I’ll just say that I regard them as rules we have come up with to model our universe in accordance with perceived empirical data. Thus, the statement of that law boils down to the assertion “Nothing ever does travel faster than the speed of light” — that is, it’s an absolute statement extrapolated from what we’ve observed so far, which we choose to believe based on the philosophy of science. Finally, there is the sense of “can” that we use when we say things like “Bachelors can’t be married” and “An odd integer can’t end in 4” — that is, in the context of expressing impossibilities in the abstract world of logic. And I still hold that none of these uses of “can” are applicable to the notion of someone making a free, indeterministic decision.
The other point in the squib I want to reiterate is the idea that the ability to make meaningful choices directly depends on some degree of determinism, even though this is probably the oldest point in the compatibilist book. I’ve said it before, and I’ll say it again: if someone exercises what we think of as their “free will” to do the Right Thing, then the freeness of their action is only practically relevant in terms of how much credit they deserve for it, and that deserved credit is proportional to how much the decision resulted from some aspect of their personal character. And I suppose my conclusion that personal character operates via deterministic mechanisms depends on materialism (one needs to view character as material and subject to the mechanics of physical causation), which is itself a controversial position that I don’t want to get into defending just now. I suppose my point is that 11 years later, I still strongly hold to the slightly whimsical-sounding sentiment that determinism is free will’s best friend.