[Topics: Determinism, Foreknowledge, Free Will, Infallibility]
The Foreseeable Future:

The Implications of Demonstrable Infallible Foreknowledge with Respect to Free Will


David Hume Sketch by M.R.P. - infallible foreknowledge - free will - determinism

Sketch by M.R.P.


As promised at the end of my last post, this post explores a question which has plagued philosophers of metaphysics for millenia: can an action be demonstrably and infallibly foreknowable, and yet free? Although there are mountains of highly technical literature on this and related questions—with infinitely debatable minutiae (and this question’s own camps of more esoteric compatibilists and incompatibilists)—I will be providing a much less formal response to the question. For that reason, I will be trying to explain much of the typical philosophical jargon I use with common, everyday expressions. In the interest of this clarity, I would like to note that any use of the words ‘compatibilism’ and ‘incompatibilism’ below refer strictly to the sense in which they were used last week, concerning determinism and free will (rather than concerning infallible foreknowledge and free will).

By presenting my own account of an answer to this question in this way, I hope to accomplish two things. First, I have been told that my articles, which are intended for accessibility, have remained difficult for readers who are truly new to philosophy; I hope that the ideas seem more accessible in this format, and thus interested readers will be able to pursue the aforementioned mountains of literature at a later date. Second, I hope to help further clarify my perceptual model of free will sketched briefly in the prior post, and to come at my notion of the ‘inescapable practical illusion of free will’ from another angle.

A Hypothetical Case of Infallible Foreknowledge with Typical Responses:

Let us begin by establishing a case study: suppose that Hannah is looking at her bookshelf, and is about to decide whether to read Voltaire’s Candide or H.G. Wells’ The Island of Dr. Moreau. Now, what would it mean to infallibly know if Hannah is going to pick Voltaire or Wells? For someone to have infallible foreknowledge of some future event, it would have to be the case that their knowledge of that future event could literally not be wrong. To make this clearer, from now on I will occasionally refer to ‘infallibly knowing something’ as ‘knowing something with absolute certainty and correctness’ (even if that means side-stepping some very big epistemological problems).

So, for example, if someone knows with absolute certainty and correctness that Hannah is going to pick Wells, then Hannah will pick Wells and will not pick Voltaire. More generally, if an action can be foreknown infallibly, then it will happen in that way. Next, imagine that Hannah’s coworker Beth is watching a video of Hannah on Beth’s four-dimensional camera, and she acquires necessarily correct knowledge that Hannah will actually choose Voltaire. Now we can restate the question like this: If Beth is in demonstrable possession of absolutely certain and correct knowledge that Hannah will choose to read Candide, can Hannah’s choice still be considered free?

Just as in last week’s post, your response to this problem relies on your definition of free will. The folks in the incompatibilist group discussed last week are likely, but by no means guaranteed, to say that Hannah’s choice can not be considered free (or, under other circumstances, that if Hannah’s choice is free, it is only free because Beth’s knowledge is not infallible). The most likely reason for them to respond in this way is that Beth’s absolutely certain and correct knowledge removes the possibility of alternatives.

To put this another way, they might say that the decision is only free if there is a non-zero percentage chance that Hannah could choose to do otherwise than she will do. If Beth’s foreknowledge that Hannah picks Voltaire is necessarily correct, then there is a one hundred percent chance that Hannah picks Voltaire, and a zero percent chance that Hannah picks Wells (and a zero percent chance that she picks any other book, and a zero percent chance that she decides not to read anything, etc.). This means that Hannah can only choose to read Voltaire, which implies that the choice is determined, which means that the choice is not free on the incompatibilist picture.

Luis de Molina - infallible foreknowledge - free will - determinismThe folks in the compatibilist group are likely, but by no means guaranteed, to say that Hannah’s choice remains free. The most common methods of presenting their reasoning would be by either reconciling responsibility and lack of choice or by painting the knowledge itself as hypothetical in nature. In the former method, compatibilists (as well as some hard determinists) proceed by responding to the incompatibilist point about alternative futures with an account, likely following the work of Harry Frankfurt, of determined action which retains the responsibilities of free choice (Hannah is still responsible for choosing Voltaire, even if she could not have done otherwise). In the latter method, compatibilists proceed by offering a Molinist account of infallible foreknowledge as being more of a map of infinitely many precise causal relationships than a knowledge of exact future events (Beth’s camera is merely simulating the precise circumstances which would lead Hannah to choose Voltaire, which also happen to be the actual circumstances).

(For a more complete introduction to the different responses to a version of this topic framed as ‘theological fatalism,’ see this encyclopedia entry.)

A Phenomenological Response to the Hypothetical:

Now, my own response will require at least a small amount of explanation (and might stray a bit from the accessible goal outlined in the introduction), because, although I hold to the compatibilist definition of free will, and so contend that determinism and free will are compatible—and although I also contend that determinism and moral responsibility are compatible—I nevertheless find demonstrable infallible foreknowledge and free will incompatible. In short, I do not find most of the responses to this problem which reconcile demonstrable infallible foreknowledge and free will at all compelling.

In particular, I find the Molinist picture to be fallacious-by-omission, as the account of general foreknowledge it depicts nevertheless entails the sort of specific-event foreknowledge which it seeks to disavow. Besides, in this particular hypothetical, Beth’s camera, after all, is not a ‘Molinist possibility map generator’—it is a recording device, and it is providing an account of the event itself; Beth is not omniscient, and so lacks what Luis de Molina called middle knowledge (i.e. a priori knowledge of all counterfactuals).

Meanwhile, although the Frankfurtian picture (the compatibilist picture concerning responsibility) brilliantly demonstrates what compatibilists mean when they (or I) say something like ‘a free moral choice with no alternatives,’ I think the picture is stretched too far when it is used to say that the choice remains free in the face of infallible foreknowledge that is demonstrable. (If you are familiar with Frankfurt’s thought experiment, I would say that no one in the story, not even Black, has infallible foreknowledge; as far as Black knew, any number of contingencies could have prevented Jones from making the kill, even if they did not in actuality prevent her from doing so.) I do not think that a version of free will which is compatible with possessible, reportable, demonstrable, infallible foreknowledge is a version of free will which almost anyone would reasonably call free.


Harry Frankfurt - infallible foreknowledge - free will - determinismTo clarify why I have come to this conclusion, I would like to underscore two pragmatic points that I think are relevant and important. The first point is that humans have the experience of deciding, which may yet be determined (in fact, I find it likely that it is), but which gives a perception of autonomy. The second point is that no being to which humans have access, at present, has been able to provide a reliable demonstration of infallible foreknowledge. These two points are why I characterized my response last week as a phenomenological model of free will, as it deals explicitly with the pragmatic case of what humans can and can not perceive.

It is not merely the truth or falsehood of the notion that Hannah’s choice of reading material is theoretically foreknowable that makes the choice unfree or free; it is Beth’s possession of that infallible foreknowledge. Under the circumstances in the hypothetical, wherein Beth demonstrably infallibly foreknows (as in any universe where agents become aware of provable infallible foreknowledge), the inescapable illusion of human free will is shattered, and is not inescapable at all. In the demonstrable presence of infallible foreknowledge, it would no longer be the case, as I phrased it last week, that the compatibilists are describing lived experience.

In fact, if it were somehow to be a demonstrable reality for all agents, I would count the infallibility of Beth’s foreknowledge as evidence for hard determinism, which would defeat not just the compatibilist free will to which I hold, but also every existential instinct, outlook, and inclination which has informed literature and philosophy throughout history. Fortunately for the value and meaning I derive from my existential perceptions, however, a claim of infallibility can not be proven by any presently existing method; it can only be either disproven by the evidence, or merely consistent with the evidence as far as we can tell. So the inescapability of the perceptual experience of free will remains.


Many responders to this article on forums around the web have insisted that my adherence to compatibilism requires that my conclusion be otherwise. In effect, they are following closely one of the strains of argumentation outlined briefly at the end of the second section section above. But in contrast to my other articles, the responses have all shared in common a misunderstanding of my response to the hypothetical. So I have decided to add this small addendum to clarify my position further. My clarification shares its form with a sentence from my conclusion above, but with greater specificity:

It is not the mere existence of an objective, foreknowable sequence of events that harms the free will of Hannah in the hypothetical case under study; it is the simultaneous infallibility, possessibility, reportability, and demonstrability of the foreknowledge of that sequence by Beth that does so. Such epistemological attributes, taken together and only when taken together, would remove Hannah’s perception of her own freedom, which is wholly constitutive of freedom from my phenomenological and pragmatist perspective. They would move cognizance of the present and future into the territory of cognizance of the past, and the past is a necessarily determined realm. Anyone still unclear on my position here should first consider the difference between knowing something infallibly and knowing something with absolute certainty and correctness (my possibly misguided simplification above), and then consider the difference between knowing something infallibly and demonstrably knowing something infallibly.

Last, please be aware that I am not arguing that Hannah’s choice necessarily loses its status as a philosophically moral choice, only that it loses its status as a metaphysically free choice (moral responsibility is, after all, the only thing some modern compatibilists care about). Recall that I agree with Harry Frankfurt’s orientation toward moral responsibility, and even praised his case studies as “brilliant” elucidations above.

Some have replied, after thinking about it further, that my conclusion is a feature of the way my hypothetical is structured, but that slightly re-tooling the hypothetical (for instance, by considering infallible foreknowledge that is not necessarily demonstrable) leads to a more typical compatibilist perspective; to those replies I say, you’re correct. That is why the last few sentences of my original conclusion deal with the theoretical impossibility of such demonstrability, and the consequences of that for perception. Yes, if you change the hypothetical into a different hypothetical, then my response to it is different, and I am inclined to hold a compatibilist position.

[Topics: Determinism, Foreknowledge, Free Will, Infallibility]
The Foreseeable Future:

The Implications of Demonstrable Infallible Foreknowledge with Respect to Free Will

was last modified: December 9th, 2016 by Daniel Podgorski
Bookmark the permalink.


  1. I will link below a lengthy but interesting Reddit comment on how we may reconcile foreknowledge and free will. The key is to realize that we *are* a deterministic process as opposed to something else being “pushed around” by one. And if physics did not constrain our choices then they could not be constrained by us. By this light any process that predicts our choices would do so by *implementing* us, which is to say, it would do so by *being* us.


    • Thank you for referring me to an interesting read.

      If someone shows you a home video of their family in the distant past, did you have to be that person in order to see the decisions made in the video? Obviously not. Now imagine someone shows you a home video of their family in the distant future. Do you have to be that person in order to see the decisions made in the video? If you think the answer might be ‘no,’ then you are considering the sort of observational-future-knowledge hypothetical which the above article considers.

      This is a separate hypothetical from the sort of constraining-process-hypothetical considered in the comment you have linked. After all, it does not matter to the above article whether the infallible fore-simulation of Hannah has to ‘be Hannah’ or not, but merely what follows from the obtainability of the knowledge of that simulation by a third party (Beth).

      Unfortunately, the commenter in question also makes assertions that many thinkers would reject, and does so without justifying argumentation (e.g. that information processing should be understood as a part of a person which is independent from the minutiae of that person, and that causal determinism can be assumed true a priori).

      So, the problem which my article confronts, and which has troubled philosophers for millenia, is being ignored in that solution. On the separate issue which that person is discussing, however, their conclusions may be at least tenable for people already convinced of all of the same priors.

  2. I wouldn’t say that the problem is being ignored in the linked comment so much as it is being dissolved.

    The problem, stated more plainly, is that determinism implies fatalism (a troublesome conclusion that doesn’t jibe with our lived experience). But this is true only if there is a divide of some sort between “me” and “my determined choices”. Your hypothetical preserves this divide (and therefore the problem) by positing infallible foreknowledge that is derived independent of the subject.

    • From my prior response: “it does not matter to the above article whether the infallible fore-simulation of Hannah has to ‘be Hannah’ or not, but merely what follows from the obtainability of the knowledge of that simulation by a third party (Beth).”

      For that to be the set-up under study and remain inadequate, and for it to be true that the problem is being “dissolved,” you would have to deny all forms of observational knowledge that do not result from identical being. If this interpretation of the linked comment is correct, then its argument is in a far sorrier state than it seemed to my more charitable earlier interpretation.

      For instance, if true, your conclusions would not be limited to future decisions. I asked before, “If someone shows you a home video of their family in the distant past, did you have to be that person in order to see the decisions made in the video?” Well, if you are committed to the notion that choices can not be observed externally, then a video, whether of future events or past events, has given you no new knowledge whatsoever; as you still have not been (or, if you prefer, implemented) the person, you can know literally nothing about them.

      Even things you observe directly would not count as knowledge on the picture you have sketched. You could spend your entire life next to someone, watching their every move, but if you were asked what decision they made at a certain time (e.g. what college they decided to attend), you would have to say, “I do not know, as I do not believe in knowledge that is derived independent of the subject.”

      Perhaps you would respond, “No, in those cases the person was meaningfully implemented.” Then I would say, exactly; that is the same as the hypothetical case in the above article, and that is why the hypothetical remains a tenable topic for philosophical inquiry.

      I have never granted the divide you mention (between a person and their determined choices), and, frankly, I’m not aware of any professional philosophers that do.

  3. Interesting article, but I disagree with your conclusion.

    Take space as an analogy. I can move an object from location X to Y or from X to Z out of free will (if I’m able and free will exists). Whether you know where I move it, does not impact my free will. How could it? The same holds for the temporal domain. Assuming you don’t tell her about it (and create a Grandfather Paradox by it, which is a different problem), if someone else knows what book she will pick this will have no influence on her decision. How could it? We can also know the outcome of past decisions with certainty and that doesn’t indicate that they weren’t free. How would the mere introduction of a temporal divide change this?

    You can make the same point without any free will by considering stochastic processes. If you know in advance that the outcome of throwing a fair dice will be 3, this does not mean that the result was not random, just like knowing the result after the throw does not make it non-random either. It is random, because of certain features of the dice that make throwing it a stochastic process, not because of the uncertainty associated with throwing it. I believe that free will, if it exists, is similar to randomness in that respect, although I do not claim that it is otherwise related to it in any way.

    Here is the catch. If there is no free will, then someone may be able to correctly predict with hundred percent accuracy the outcome of someone’s decision without the need for precognition (=backwards causation), and since genuine precognition is a very problematic concept, it tempting to assume that prediction without precognition warrants the kind of knowledge in question, and so there is no free will. But that’s a fallacy (affirming the antecedent).

    • Great response! I agree with the vast majority of what you’ve written.

      When you say, “Assuming you don’t tell her about it (and create a Grandfather Paradox by it, which is a different problem), if someone else knows what book she will pick this will have no influence on her decision,” then, from a pragmatist perspective, you have granted the conclusion of my article: that free will remains if the perception of free will is not impinged upon by the foreknowledge.

      At base, my conclusion is really more about phenomena, perception, and experience than it is about metaphysics. But even so, as I said in the article, I do not find Molinism compelling. And part of the problem with the argument in your last paragraph, as you may be aware, is that a prediction with 100% accuracy is not the same as a metaphysically infallible prediction (I tried to make it clear in the article that any overlapping references to such cases were only for the sake of temporary ease of communication).

  4. I find free will to be such a human centric notion, such a sad proof of the shortsightedness evolution had to impose on us, another byproduct of the unfortunate reconciliation between cognition and abstract thinking on one side and the survival imperative on the other. Such arrogance on our part…

React to this Article