By Garrett Pendergraft
Editors note: Because of a major snowfall in Savannah, GA during the recent Eastern APA, the Blog of the APA will be hosting papers by people who missed the opportunity to present or who would like a larger audience for their work. If you were on the schedule for the Eastern and would like to take advantage of this opportunity, please fill out the submission form.
Abstract: The concept of freedom is the mental entity that partially constitutes our thoughts about freedom. A conception of freedom, on the other hand, is a view of what freedom is and what it requires. When we examine important philosophical concepts, such as the concept of freedom, we often do so by asking questions about the corresponding conception. My goal here is to explore a particular question of this type—namely, the question of whether freedom requires that we be able to do otherwise, holding fixed the past and the laws of nature. I will argue that the correct conception of freedom does not include this incompatibilist requirement. My argument will depend on an important distinction between relatively “austere” and relatively “opulent” conceptions of a given concept, the relative austerity of the compatibilist conception of freedom, and the claim that we should adopt a presumption of austerity.
In this paper I will argue that the compatibilist conception of freedom is more likely to be correct than the incompatibilist conception. My argument (§5) will depend on an important distinction between relatively “austere” and relatively “opulent” conceptions of a given concept (§2), the relative austerity of the compatibilist conception of freedom (§3), and the claim that we should adopt a presumption of austerity (§4).
2. Setting up the terms: austerity vs. opulence
Let’s begin by distinguishing the concept of freedom from a conception of freedom. Setting aside difficult issues about the nature of concepts themselves, we can characterize the concept of freedom in a minimal way: the concept of freedom is the mental entity that partially constitutes our thoughts about freedom. A conception of freedom, on the other hand, is a view about what freedom is and what it requires. It is a set of conditions that tell us when the concept—the mental entity—is correctly deployed. These conditions specify when it is appropriate to label an action or a choice as ‘free.’
The distinction between concepts and conceptions is important, because when we examine a philosophical concept we often do so by asking questions about the corresponding conception. When we ask, for example, what it takes to know some proposition, we are asking—figuratively speaking—what knowledge requires in order to be instantiated. Sometimes these questions lead to a full-blown conceptual analysis, but it’s possible to explore such questions without attempting such an analysis. My goal here is to explore the concept of freedom in this more limited way, without attempting a full-blown analysis. In particular, I will be asking whether freedom requires that we be able to do otherwise, holding fixed the past and the laws of nature. I will argue that the correct conception of freedom does not include this incompatibilist requirement.
This type of question about what a concept requires can be used to characterize important developments in the recent history of philosophy. For example, for many years the prevailing conception of knowledge was that it required only justified, true belief. One thing we learned from Gettier’s (1963) famous counterexamples to this prevailing conception was that knowledge (and thus the correct conception of knowledge) is more demanding than we thought it was.
Sometimes we are urged to revise our conceptions in the other direction, toward fewer requirements. Consider Hilary Putnam’s (1962) example of robot cats. “All cats are animals” appears to be an analytic, necessary truth. But if we were to discover that the things we have been calling cats are actually robots controlled by Martians, then (arguably) we should admit that we were wrong about cats being animals; as it turns out, cats are not animals. In other words, we thought that proper deployment of ‘cat’ required applying the term only to animals—but consideration of the possibility of robot cats suggests that cathood (and thus the correct conception of cathood) is less demanding than we thought it was.
We need a way of describing these competing conceptions relative to each other. Thus, following Graham and Horgan (1994; cf. Horgan and Graham 1991), I’ll use the terms austerity and opulence to distinguish between less demanding and more demanding conceptions. So our pre-Gettier conception of knowledge was more austere—less demanding—than our post-Gettier conception is; and our pre-Putnam conception of cathood was more opulent—more demanding—than our post-Putnam conception is.
Our guiding assumption, then, will be that when we are examining different conceptions of some concept, we can situate those different conceptions on a continuum of demandingness and then make comparative judgments between them. A conception that is less demanding, when compared to some other conception on that same continuum, is relatively austere; whereas a conception that is comparatively more demanding is relatively opulent.
3. The relative opulence of incompatibilism
With this framework in hand, we can now turn to the concept of freedom. The philosophical dispute over ‘freedom’ (or ‘free action,’ or ‘free will’) is in part a dispute over whether the correct conception of freedom is less demanding (i.e., relatively austere) or more demanding (relatively opulent). Let the relevant requirements be represented as necessary conditions on the truth of statements like “S acted freely in performing A.” We can then distinguish between different conceptions of freedom by identifying the requirements that are included in one conception but not the other. For example, the following requirement is widely accepted and thus included in most conceptions of freedom:
The ability to do otherwise: An agent S acts freely in performing an action A only if there is some point at which S has both the ability to perform A and the ability to refrain from performing A.
This, in other words, is the requirement that S be able to do otherwise. If there’s any such thing as a commonsense conception of free will, then it probably includes this requirement. But despite the commonsense agreement on this particular requirement, we don’t have to move very much farther on the continuum toward opulence before we encounter requirements that mark a sharp distinction between opposing views on free will.
According to the incompatibilist about freedom and determinism, freedom requires something else in addition to the mere ability to do otherwise. This additional ability is sometimes referred to as the ability to extend the given past:
The ability to extend the past: An agent S acts freely in performing an action A only if S is able to do otherwise, consistent with the actual laws of nature and the past history of the actual world.
This requirement (let’s call it “the incompatibilist requirement”) entails that a statement such as “S acted freely in performing A” cannot be true unless the world is indeterministic. In short, the incompatibilist requires—the incompatibilist claims that freedom requires—that (the thesis of) determinism be false.
The compatibilist, on the other hand, denies that freedom makes this type of demand. Clearly, then, there is an important sense in which the incompatibilist conception of freedom is more demanding—more opulent—than the compatibilist conception. Thus the dispute between compatibilists and incompatibilists can be construed as a debate over whether the correct conception of freedom is relatively austere or relatively opulent.
4. The presumption of austerity
As I hope to have shown above, the distinction between austerity and opulence is a fruitful starting point for characterizing competing conceptions of important philosophical concepts, and in particular for characterizing two rival conceptions of freedom. At this point I would like to move beyond mere description and start asking whether we can use a conception’s relative degree of demandingness to draw conclusions about its ultimate viability. Is a conception’s relative austerity or opulence any sort of guide to its correctness? I argue that the answer is Yes: that relatively austere conceptions are more likely to be correct than their relatively opulent counterparts, and thus that there should be a presumption of austerity. This presumption, as we will see, can serve as a premise in a new argument for compatibilism.
Several prominent voices in the literature have advocated for something like a presumption of austerity (e.g., Graham and Horgan 1994), but they haven’t said much in support of this presumption. So I’ll need to say a little bit in its defense before I use it to develop my argument for compatibilism.
Because the conceptual requirements we’re interested in are (or entail) empirical requirements, let’s narrow our focus a bit and flesh out the above characterization of austerity as follows: If there are competing conceptions of a concept, and the first conception requires the truth of an empirical claim whereas the second conception does not, then the second conception is relatively austere. I will argue that relatively austere conceptions are more likely to be correct—that we should adopt a presumption of austerity—because this presumption provides the best explanation of our intuitive reaction to various hypothetical cases.
Consider the relatively straightforward concept of flying ability (Horgan & Graham 1991). This is a concept that we have a pretty good grasp of, which is to say that we are pretty good at categorizing creatures according to whether they are able to fly. Now suppose we are told that in order for something to be able to fly, the ratio of its wing surface area to its body mass must be above a certain threshold; i.e., we are told that the correct conception of flying ability requires this ratio. If we then discover that, say, a bumblebee doesn’t satisfy this requirement, we are faced with a choice: we can maintain our commitment to the ratio requirement and revise our judgment that bumblebees can fly, or we can retain our judgment that bumblebees can fly and abandon the ratio requirement. It seems pretty clear that in this case the right thing to do is retain our judgment and abandon the requirement. Here we have two competing conceptions of flying ability, one of which is relatively opulent in virtue of endorsing the ratio requirement. Given the empirical discovery that bumblebees don’t satisfy this requirement, it seems pretty clear that we should reject the requirement and thus endorse the relatively austere conception that doesn’t include it.
Or consider the concept of causation. Suppose that some ancient philosopher—let’s call him “Lucretius”—believed that causation demands a lack of empty space: that for all x and y, a causal relationship between x and y requires a lack of empty space between x and y. This is a plausible demand, but notice that it entails the falsity of various empirical claims, and in particular the falsity of various claims about the atomic structure of matter. Now suppose further that a physicist travels back in time and presents Lucretius with a conception of causation that is consistent with matter having an atomic structure. Lucretius would then be in the following situation: he would be considering two competing conceptions of causation that capture the same phenomena but differ with respect to their austerity. And in that case it seems clear that he should drop the relatively opulent conception that demands a lack of empty space and instead endorse the relatively austere conception that does not include that demand.
I am not trying to suggest here that our fictional Lucretius was bad at science. We can’t always anticipate empirical discoveries, and sometimes the best we can do is revise our conceptions after the fact. But when we are aware of a potential empirical discovery that would falsify one of the requirements of some conception, and we are also aware of a competing conception that does the same explanatory work without including that requirement, we should prefer the relatively austere conception. We should prefer the conception that’s resilient with respect to the potential empirical discovery.
When we encountered Putnam’s robot cats case for the first time, we were in a situation analogous to Lucretius’ situation described above. In light of the possibility that Putnam raised, we could give up on the existence of cats or we could give up on the requirement that cats be animals. The correct option is arguably the second one: rather than claiming that there aren’t any cats, we should instead admit that the correct conception of cathood is more austere than we thought it was.
In each of the cases above, an empirical discovery (or a potential empirical discovery) prompted a comparison between a relatively austere conception and a relatively opulent conception. And in each case, the relatively austere conception is the conception more likely to be the correct one. If this is the appropriate reaction to such cases, then we should ask if there’s a general principle that explains the appropriateness of these reactions. I think there is such a principle, and that we can formulate it as follows. First, if a conception of some concept entails the falsity of an epistemically open empirical claim, then it’s possible that a competing conception will capture the relevant phenomena equally well—will explain the relevant data—without entailing the falsity of the empirical claim. If we then discover a competing conception that realizes this possibility—a conception that can do the same theoretical work but with fewer empirical commitments—then we should prefer the conception that does not dictate the relevant empirical results. We should prefer the conception that is resilient with respect to the empirical discovery in question.
In short, we should adopt a presumption of austerity:
The presumption of austerity: When we are evaluating competing conceptions of some concept (both of which can explain the relevant data), and one of the conceptions is relatively austere, the relatively austere conception is more likely to be correct.
In the cases considered above, the conditions specified in the presumption of austerity are satisfied: the relatively austere conception captures the phenomena, or explains the data, as well as the relatively opulent conception. The presumption of austerity tells us that in these cases we should prefer the relatively austere conception, and this guidance matches up with with our intuitive judgments about the relevant cases. I see no other unifying feature of these cases that could explain these intuitive judgments, so it seems to me that the presumption of austerity provides the best explanation of these judgments and is thus a principle we should adopt. I will now employ this principle in a simple, novel argument for compatibilism.
5. A new argument for compatibilism
The truth of determinism is an epistemic possibility for us, which means that whether some conception of freedom rules out the truth of determinism is relevant to the austerity of that conception. This should prompt us to compare competing conceptions of freedom in terms of their relative austerity. As we saw above, the compatibilist conception of freedom is more austere than the incompatibilist conception. Thus we have the materials for a simple argument for compatibilism:
- When we are evaluating competing conceptions of some concept (both of which can explain the relevant data), and one of the conceptions is relatively austere, the relatively austere conception is more likely to be correct.
- The compatibilist conception of freedom can explain the relevant data.
- The compatibilist conception of freedom is relatively austere (i.e., less demanding than the incompatibilist conception).
- Therefore, the compatibilist conception of freedom is more likely to be the correct one.
I argued for (1) in §4, and for (3) in §3. Thus my remaining task consists of defending (2)—call it the explanatory premise.
First, I should note that although the explanatory premise describes the conception itself as capturing the phenomena, or explaining the data, properly speaking it would be a theory or account built on the conception that would do the capturing or explaining. This complicates the question of whether a compatibilist conception (or any conception) of freedom can explain the data, but I hope to show that we can still make progress on such questions.
Second, I need to say a little more about which data count as relevant when it comes to evaluating competing conceptions of freedom. At a minimum, these data include our belief in our freedom, our experience of making a causal difference in the world, our experience of deliberating between genuine options, and our moral responsibility practices (broadly construed).
Let’s begin with our belief in our freedom. How a theory explains this data set will depend first of all on whether the theory includes the claim that we do in fact have free will. If the theory does include the claim that we have free will, then the explanation is simple: We believe that we have free will because we do in fact have free will (and because we have evidence of this fact). If the theory includes the claim that we do not have free will, then the explanation of our belief in free will is going to be more complicated. Either way, though, the compatibilist and the incompatibilist are going to offer explanations of our belief in freedom that are on a par. So this particular data set presents no problem for the explanatory premise.
For similar reasons, both compatibilist and incompatibilist explanations of our experience of making a causal difference in the world are going to be roughly on a par. (The fatalist explanation of this data set will perhaps be more complicated.) Comparative judgments about explanations of our causal experiences will turn on the underlying account of causation, and not on whether the theories in question accept or reject the incompatibilist requirement.
Consideration of the third data set (our experience of ourselves as deliberating between genuine options) complicates the situation because the notion of a “genuine option” is fraught with controversy. Nevertheless, there is a straightforward explanation of this data set, namely that we experience ourselves as deliberating between genuine options because we do in fact have genuine options. As long as there is at least one viable compatibilist theory that can give this answer, then we can safely claim that the compatibilist conception is able to explain this deliberative data set in a way that is roughly on a par with the best incompatibilist theories.
The last data set, involving our responsibility practices, is a large and diverse set including our practices of holding each other responsible and the reactive attitudes and behaviors associated with those practices. Consideration of this data set complicates the situation even more, but I do think the following observation is true independently of the specific details of the relevant theories: Given that much of the work on moral responsibility in recent decades has been done by compatibilists, it seems fair to say that there are at least some viable compatibilist theories that are able to explain the data consisting of our moral responsibility practices.
I conclude that we have good reason to believe that theories built on a compatibilist conception of freedom are able to explain the data. Thus, given the presumption of austerity (and the fact that the compatibilist conception of freedom is relatively austere), we have good reason to believe that the correct conception of freedom is one that rejects the incompatibilist requirement.
Garrett Pendergraft is an Associate Professor of Philosophy at Pepperdine University. Most of his research explores compatibility questions involving free will and moral responsibility.