A common trope in discussions of fatalism is that arguments for fatalism are invariably guilty of a “modal fallacy”, specifically the fallacy of conflating “necessarily, if p then q” with “if p, then necessarily q“. In fancy academic jargon this is known as conflating the necessity of the consequence (i.e., of the whole conditional, if p then q) or necessitas consequentiae with the necessity of the consequent (i.e., the consequent of the “if p then q” conditional, namely, q) or necessitas consequentis.
The “modal fallacy” charge against fatalism is leveled by, among others, William Lane Craig (e.g., The Only Wise God, ch. 4) and by Norman Swartz in his Internet Encyclopedia of Philosophy article on “Foreknowledge and Free Will”.
I maintain, however, that the “modal fallacy” charge against fatalism misses the mark. The charge is itself a fallacious straw man. It misrepresents how fatalistic arguments actually work in order to make them easier to “refute”.
In what follows I will first explain the “modal fallacy” charge in more detail and then explain how the charge misrepresents fatalistic arguments by looking at the actual structure of fatalistic arguments. I will show that the latter (when intelligently formulated) are demonstrably valid. Fatalistic arguments may, and I think do, have a false premise, but they are not guilty of a logical error, as the “modal fallacy” charge alleges.
1. How fatalistic arguments work according to the “modal fallacy” charge
According to those who promote the “modal fallacy” charge, fatalistic arguments run roughly as follows: First, it supposed that there is something that singles out or specifies a unique and complete extension of the actual past and present as the actual future. Most commonly, this “something”, which I’ll call a future specifier, is thought to consist in one of three things:
- A “complete, true story” of the future—a tenselessly or omnitemporally true description of how the actual future is going to unfold.
- God’s infallible knowledge of just such a story, i.e., God’s having infallible foreknowledge of exactly how the actual future is going to unfold.
- God’s creative decree that the entire course of creation history shall unfold in precisely such-and-such a manner.*
*Note: The first type of future specifier (a) is sometimes said to be the basis for “logical fatalism”, whereas the latter, (b) and/or (c), are sometimes said to be the basis for “theological fatalism”. But this distinction between two different kinds of fatalism is bogus. These don’t represent different kinds of fatalism, but rather different ways of arguing for fatalism, differing only in what they take the relevant future specifier to be. In each case the reasoning process runs exactly the same.
For our purposes it doesn’t matter what type of future specifier the fatalist focuses on. So let’s just call it S. The fatalist’s first move, then, is to posit the existence of a future specifier, S.
(1) Future specifier, S, obtains. (assumption)
After positing a future specifier of some sort, the fatalist’s second step is the innocuous and obvious observation that the assumed specifier (S) entails a specific future (F). That is, the fatalist supposes
(2) Necessarily, if S then F. (definition of a future specifier)
This affirms the necessity of a consequence, that is, a complete if–then statement. And this step is indeed necessary because it is true by definition. It follows from the definition of a future specifier that it specify a particular future as the actual one. For example, if God infallibly knows (b) a specific “complete, true story” of the future, then, necessarily, the story that God infallibly knows to be the case will be the case. To suppose otherwise would require that the story be either partly false or incomplete or that God be fallible, all of which are contrary to the supposition that God infallibly knows a complete, true story of the future.
Next, the fatalist’s third step—according to those who promote the “modal fallacy” charge—is to confuse the necessity of the consequence (if S then F) with the idea that the consequent, the specified future F, is itself necessary, given S.
(3) If S then, necessarily, F. (illicit derivation from (2) or from both (2) and (1)—the “modal fallacy”)
To see that this derivation is illicit, consider that all (2) says is that F is conditionally necessary given S. All it says is that, necessarily, F will happen given S. (3), however, makes the much stronger claim that F is categorically necessary given S. It says that F must happen given S. But that doesn’t follow from (2) with or without (1). That something will happen (given certain conditions) does not entail—at least, not without additional, substantive assumptions—that it must happen (given those conditions).
Finally, from (1) and (3), the fatalist concludes that the specified future F is in fact categorically necessary.
(4) Necessarily, F. (from (1) and (3))
As an instance of modus ponens this step is logically valid, but the argument from (1)–(4) on the whole is invalid because it is predicated on the logically fallacious inference from (2) (or (2) and (1)) to (3).
The above, in essence, is the sort of story that “modal fallacy” proponents tell about fatalism. Unfortunately for them, the story is wrong.
2. How fatalistic arguments actually work
Steps (1) and (2) of the above reconstruction need no modifications, but step (3) where the dreaded “modal fallacy” occurs gets fatalism wrong. We need to replace that step in the argument with a better representation of how fatalists actually reason. To see what that step needs to be, it is helpful to work backwards from the fatalistic conclusion.
Crucially, (4) should not be construed as claiming that the specified future F is necessary in the same way that <If S then F> is necessary, as (2) says. In other words, the word “necessarily” in (2) does not mean the same thing that it does in (4). In (2) it means logical necessity because <If S then F> is true by definition and so cannot be denied without inconsistency. The kind of necessity that the fatalist attributes to the future in (4), however, is not logical necessity, but rather now-unpreventability (to use A. N. Prior’s apt phrase). The fatalist is not trying to show that only one future is logically possible, but rather that only one future is concretely realizable or causally possible since all other (logically) possible futures are in some way precluded or blocked by reality (not mere logic!).
So that we don’t take “necessarily” in (4) to mean logical necessity and thereby take the fatalist to be arguing for a much stronger conclusion than he needs, we need to modify (4) to specify the kind of necessity in view.
(4*) Now-unpreventably, F.
Now, how can a fatalist hope to derive a conclusion like this? To get an idea, take a quick perusal of the Stanford Encyclopedia of Philosophy entry on “Foreknowledge and Free Will”. Several fatalistic arguments are detailed in the course of the article, and all of them make use of some kind of “fixity” principle to the effect that some facts, such as certain truths about the future, or God’s knowledge of such, are “now-necessary”. This is the same idea as now-unpreventability. Further, all of the fatalistic arguments discussed in that article involve a transfer of necessity principle. The gist of fatalistic arguments is to transfer now-unpreventability from a future specifier to the future that it specifies. (Incidentally, none of the fatalistic arguments discussed in that article commit the so-called “modal fallacy”. This should confirm that it is indeed a superficial charge. Indeed, the SEP article on “Fatalism” doesn’t even mention “modal fallacy” or “necessity of the consequent”.)
The fatalist’s maneuver is not hard to motivate. In the first place, everyone believes that some facts are fixed or now-unpreventable. Plausible candidates include: mathematical theorems, basic principles of morality, causal laws, facts about the actual past, God’s creative decree, and basically anything that is beyond anyone’s power to bring about. Arguably, neither you nor I nor any other creatures have any say-so about such matters. Perhaps God has say-so about some such things, but we don’t. And even God is in some sense bound by His own nature and by His own decisions and actions (e.g., God can’t undo His own creative decree nor can He, say, undo the Incarnation). God may originally have been free to create or not create, but once the decision to create is made God can’t return to the not-having-decided-yet state.
Now, to the idea that some facts are fixed all the fatalist needs to add is that the fixed facts (whatever they are) collectively specify a unique and complete future. Given that (and an appropriate transfer of necessity principle) and fatalism logically follows. Instead of (3), then, we should have
(3*) Now-unpreventably, S.
The core logic of fatalist arguments, then, looks like this:
(1) Future specifier, S, obtains. (assumption)
(2) Necessarily, if S then F. (definition of a future specifier)
(3*) Now-unpreventably, S. (assumption)
(4*) Now-unpreventably, F. (from (2) and (3*) by transfer of necessity)
Notice that there is not even a hint of the dreaded “modal fallacy” here.
Since (2) is true by definition, there are only three places where the argument can be challenged: (1), (3*), and the inference from (2) and (3*) to (4*). Denying (1) is what I call the open futurist response to fatalism—it basically says, there are no future specifiers. Denying (3*) is what I call the preventable futurist response—it denies that any future specifiers are wholly fixed. Both of these responses have serious advocates.
As for denying transfer of necessity, this is a non-starter, for the relevant transfer principle
(TN) If now-unpreventably S and necessarily <If S then F>, then now-unpreventably, F.
is demonstrably valid. We can easily see this if we express “now-unpreventability” in terms of causal necessity, or what holds in all causally possible futures. If some alternative future F* (where F* ≠ F) were causally possible then F would not be now-unpreventable because it would still be possible to concretely realize F* instead, thereby “preventing” F. It may also help to observe that logical necessities are now-unpreventable and therefore causally necessary. So clarified, TN entails
(TN*) If S obtains in all causally possible futures and if <If S then F> obtains in all causally possible futures, then F obtains in all logically possible futures.
This is a logical truism for the exact same reason that the parallel transfer principle involving logical necessity is an axiom in every standard system of modal logic: If ALL entities in some domain (e.g., causally possible futures) are such that S and ALL of those entities are also such that <If S then F> then modus ponens requires that ALL of those entities be such that F.
3. Summing up
When intelligently formulated, fatalistic arguments are demonstrably valid. Hence, they do not commit a “modal fallacy”. The charge that fatalists do commit such a fallacy itself commits a fallacy, namely, it attacks a straw man version of fatalism.
Furthermore, there are only two promising ways to be an anti-fatalist: Either (1) open futurism or (2) preventable futurism. For my part, I think open futurism is the way to go (hence the name of this blog). Preventable futurism invariably leads to some very implausible metaphysics (e.g., backward causation) and is in the end, I submit, ontologically incoherent. But that’s an argument for another day.