Some Logico-Empirical Aspects of the Phrase “Could Have Acted Otherwise”
Free will (and thus moral responsibility) has often been defined by the criterion that “one could have acted/done otherwise.” This phrasing goes back at least to the time of Hume and remains in use today (The Information Philosopher collects various relevant quotations). I’ve always found this phrase perplexing. In what sense could I have done otherwise? I may regret an action after I have done it, but seemingly nothing can change the fact that I did what I did. I may wish I did otherwise, but could I really, at the time, have done otherwise?
There are several ways to understand this phrase. Here are three of them:
The Rewind/Rollback Experiment
A number of people (perhaps most famously Peter van Inwagen) have proposed variations of the following thought experiment: Imagine we’re being observed by some supernaturally-powerful entity. This “demon” (as such entities are conventionally called in philosophy) has the ability to rewind time to some earlier moment (such that every last atom is back to where it had been) and let history replay itself. Would the demon, having done this, observe us making the same decisions we made before, or would it observe us “acting otherwise”? What about after several rewinds, or a zillion rewinds? If some of the time we can be observed to “do otherwise,” what does this suggest? A straightforward implication is that the universe is nondeterministic: Given identical starting points — identical down to the last atom (I’ll ignore the quantum realm) — one can get different outcomes.
A rather more bold (less grounded) inference from the rewind experiment is that free will exists. In one and the same situation, sometimes the will makes one decision, sometimes another. But on what basis, we might ask, does the will make its differing decisions? Or must the will, to be truly free, be free of influence? And would such a will be “free” or just “capricious”?
The Apparent Similarity of Situations and Organisms
Another sense of the phrase “could have acted otherwise” derives from the observation that we often find ourselves in similar situations and sometimes make different decisions in those situations. Sometimes I get Thai food for lunch, sometimes I bring a sandwich. But, of course, the phrase “similar situation” is a generalization. No two existential situations are ever exactly the same, though they may be considered the same for certain purposes. Generically, the what-shall-I-have-for-lunch situation looks much the same from day to day. In brute existential terms, however, it differs. Both the organism and the environment have changed at least somewhat. Given different starting points, it’s not surprising to get different outcomes.
A similar interpretation of “could have done otherwise” arises from the fact that organisms (rather than situations) are sometimes similar to each other. And it can be observed that in an experimentally-controlled situation (such that all environmental variables are held as close to constant as possible), one organism emits one behavior, while another of the same species exhibits a different behavior. If two members of the same species behave differently in the same situation, it appears that each “could have done otherwise.” But existentially speaking, of course, the organisms are different, and it’s not terribly surprising if, in this case, different starting points also lead to different outcomes.
So we find that the “same” organism plus “similar” environments sometimes leads to different outcomes, and that “similar” organisms plus the “same” environment also sometimes leads to different outcomes. Hence, “could have done otherwise.” On the face of it, however, these universal propositions (inductive inferences) do not suggest or imply anything about the nature of free will or determinism. The same propositions could hold whether free will existed or whether the universe was entirely deterministic. Determinism is still possible because, existentially speaking, the organisms and environments involved in each case are existentially different.
“Could Have Done” or “Could Do”?
The observation that we sometimes do different things in similar situations suggests that perhaps the phrase “could have done otherwise” conflates the past and the future. The fact that we did one thing in a certain situation does not mean we need to do the same thing in future situations of like kind. Humans are among those creatures which are capable of “learning from experience.” If our action in one situation doesn’t quite work out as well as we would have liked (see illustration above), there is the possibility of our doing otherwise next time. In other words, while we could not have done otherwise last time, we could do otherwise next time.
Is It Relevant?
Thought experiments, empirical observations, and logical parsing are inconclusive regarding free will vs. determinism. In my view, the phrase “could have done otherwise” has no relevance to free will vs. determinism debate. What the phrase should do is simply point us to the fact that humans are among those natural systems which must sometimes choose between alternative means and ends. In some existential situations, we can do otherwise but we only do what we actually do. This doesn’t mean that we need to do the same thing next time. Whether this choice of what to do always happens deterministically or sometimes happens nondeterministically (as a result of a freely-operating will) is at present unknowable.* While science must in general assume determinism, I’d argue that behavior can be equally well or badly managed under various views of ultimate reality.
*Though, to be honest, I think determinists have stronger arguments. No doubt this is because I’ve been strongly influenced by the philosophy of science.