It strikes me that the solution proposed by Maimonides to the freewill problem has been wrongly equated with the solution put forward by Boethius. In what follows, I try to explain how the two proposals differ. I’d love to hear people’s feedback on this, as I’m no real Maimonides scholar. Have other people noted this difference? Do let me know.
The freewill problem, in its medieval formulation, arises from the fact that our free will seems to be incompatible with God’s foreknowledge. The problem can be stated as follows:
- A being with a free will often has the choice between two mutually exclusive actions, action x and action y, at some time, t.
- God is omniscient, and therefore knows that I will chose to perform, say, action x at time t.
- Knowledge is factive, which is to say that it’s either a relation between a mind and a fact, or at least, a relation between a mind and a true proposition; a proposition made true by a fact. Knowledge is, therefore, always accompanied by the existence of a fact.
- Given 3, if God knows that I will perform action x at time t, then it is a fact that I will perform action x at time t.
- If it is a fact that I will perform action x at time t, then I cannot choose to perform action y at the same time.
- Therefore, I am not a free being with control over the course of my own life. Given any choice, I am always bound to act in the way that God already knows that I will act.
A popular ‘solution’ to the medieval free will problem among contemporary people of faith is to suggest that God’s foreknowledge doesn’t interfere with our actions: he knows what we’re going to do, from afar, but he doesn’t compel us to act in any given way. A couple of randomly selected Christian missionary websites will help me to illustrate how such a solution is supposed to get off the ground. A man called Justin wrote into the Come Reason Ministries (www.comereason.org) troubled by the medieval problem of free will. This was the response that he received:
Those who argue in this manner [i.e., those who take the medieval problem seriously,] make the mistake of thinking that because God possesses knowledge about a specific matter, then he has influenced it. That does not follow at all. Just because God can foresee which choice you will make, it does not mean you couldn’t still freely choose the other option … Let me give you an example. I have a five year old son. If I were to leave a chocolate chip cookie on the table about a [sic] hour before dinner time and my son was to walk by and see it, I know that he would pick up the cookie and eat it. I did not force him to make that decision. In fact, I don’t even have to be in the room at all. I think I know my son well enough, though, to tell you that if I come back into the kitchen the cookie will be gone. His act was made completely free of my influence, but I knew what his actions would be …
One might worry that this example about the child falls short of real knowledge. The father knows the child enough to know what’s most likely, but, if the child really has free will, then surely the father doesn’t know with any real degree of certainty what the child will actually do. Perhaps, unbeknown to the father, the child has suddenly come down with a stomach bug that’s devastated his appetite, even for chocolate. The father can only claim to know what’s most likely. Surely, God’s foreknowledge is supposed to be more certain than a knowledge of what is most likely! The analogy isn’t tight enough to prove its point, unless it’s willing to give up the notion that God knows anything more than the statistical likelihood of our future courses of action.
The second website improves upon the first with the popular move of claiming that God’s living outside of time is what makes it the case that his knowledge doesn’t interfere with our freedom. The authors of www.biblebell.org imagine a terrorist called Fred who plants a bomb on flight 999:
The airport where Flight 999 originated has security cameras. These cameras make TV tapes of the surroundings of each aircraft. After Flight 999 explodes, some FBI agents view these TV tapes. The tapes reveal Fred planting the bomb on the airplane. At the moment that the FBI agents are viewing the tapes, it is no longer possible for Fred to change his mind about planting the bomb. This fact does NOT mean that the FBI agents are forcing Fred to plant the bomb. In a manner of speaking, God has ALWAYS had the video tapes that the FBI agents viewed. Thus, God’s omniscience means that God knows our free will decisions BEFORE those decisions are made. It does NOT mean that God forces us to make those decisions.
I take it that their talk of God having had access to the video tapes before they were even taken is utilising a metaphorical notion of ‘before’ that really indicates some sort of atemporal perspective. The idea is that, sitting from his perspective, outside of our time line, it’s as if God is watching the past, the present, and the future, concurrently, on a number of television screens.
That God’s watching the future doesn’t mean that he’s influencing it in anyway. When I re-watch a recording of a football match, the goal keeper has already succeeded or failed to save a certain penalty. I know before I watch it whether the goalkeeper succeeded or failed, yet I’m still watching a free human agent. And, despite the fact that I know which way he’s going to dive, I’m still watching someone with the freedom to dive the other way. This response to the medieval problem, which seems to be so popular with many religious people today, is, in fact, an echo of the solution proffered by Boethius.
Boethius claimed that God is not in time and that he therefore has no temporal properties. God may have beliefs, but he doesn’t have beliefs at any given time. It is a mistake to say that God knew yesterday what I will do today, because God doesn’t live in time. Linda Zagzebski explains: ‘The way Boethius describes God’s cognitive grasp of temporal reality, all temporal events are before the mind of God at once. To say “at once” or “simultaneously” is to use a temporal metaphor, but Boethius is clear that it does not make sense to think of the whole of temporal reality as being before God’s mind in a single temporal present. It is an atemporal present, a single complete grasp of all events in the entire span of time.’
Aquinas adopted the same solution and used a circle as an analogy: just as a timeless God is present to each and every moment of time, the centre of a circle is present to each and every point on its circumference. Boethius thought that God’s knowledge of future-p, given that he knows it from his atemporal location, doesn’t necessitate p in any strict sense of the term ‘necessitate.’ The truth of a proposition about the future might be logically necessitated by God’s foreknowledge of it. But the fact in question might still be a contingent one. This was the distinction that Boethius sort to describe between ‘conditional necessity’ – that a proposition about the future will have if God knows it to be true – and ‘simple necessity’ which a fact will have only if it was coerced by outside forces into being. Just because p is conditionally necessary, given God’s knowledge, it doesn’t make it necessary in the simple sense; p can still be a contingent fact that God happens to know.
This Boethian approach to the freewill problem – that God’s knowledge doesn’t interfere with our actions – is popular in contemporary Jewish literature. One example is Rabbi Touger’s outreach book, What We Believe, in which a Boethian solution to the problem is proposed, and attributed, erroneously, in my view, to Maimonides. A distinctly Boethian solution to the problem can even be found in the erudite studies of Lord Sacks, the Chief Rabbi of the UK, Covenant & Conversation – where he utilised the analogy of a person watching a pre-recorded football match.
The Boethian approach also exists in classical Jewish texts. Tosfot Yom-Tov, in his commentary to Pirkei Avot (3:15), quotes R. Moshe Almoshino as having suggested a distinctly Boethian solution to the problem, once again in the name of Maimonides (but the Tosfot Yom-Tov is, at best, non-committal as to the accuracy of such an attribution). He also quotes a Boethian solution proposed by R. Abraham ben David.
Nevertheless, R. Almoshino and R. Abraham notwithstanding, I want to argue that the great Jewish philosophers of the Medieval period were united in their rejection of a Boethian solution to the freewill problem.
One might think, given the following expert, that I’m overlooking one of the fathers of Jewish philosophy, Saadya Gaon: ‘[We have no reason to believe that] the knowledge of the creator concerning things is the cause of their existence.’ Indeed, it seems as if Saadya Gaon is arguing that God has a foreknowledge that doesn’t cause us to act, and because it doesn’t cause us to act, it doesn’t interfere with our freedom. But this Boethian interpretation is undermined by Saadya’s later claim (Ibid) that God knows what man will do because, ‘God knows man’s nature.’ This could imply that God’s knowledge of future contingents isn’t factive: it’s merely statistical. Because he knows our nature, he can predict with stunning accuracy what we will do. That doesn’t mean that there is already a fact of the matter as to what we will do, but that there is a mind that is uncannily good at predicting how we will write our own futures. If this was Saadya Gaon’s position, it would certainly have been bold, giving God less than certain knowledge of the future (a view that would later be echoed by Gersonides).
A full interpretation of Saadya Gaon’s position on free will would require extensive research. His comments in Emunot Va’Deot are, at best, inconclusive. Either way, we can say with some surety that the three major Jewish philosophers to have written extensively and systematically about the free will problem Maimonides, Gersonides and Crescas were united in their rejection of a Boethian solution (I won’t explore Crescas’ position in this blog entry, but he adopted a compatiblism strikingly reminiscent of the much later compatibilism of Harry Frankfurt)).
As shall become clear, its absence from all of the major works of Jewish philosophy illustrates, not that the Jewish philosophers hadn’t entertained the Boethiun solution, but that it left them unsatisfied. The Jewish philosophers were acutely aware that the problem of God’s foreknowledge has little or nothing to do with the fear that God’s knowledge influences our actions; the problem is that if God has any sort of epistemic relationship to the future, then the future must exist; and, if there is some tenseless perspective from which the future can already be said to exist, then our freedom to choose between competing life alternatives seems to collapse, quite independently of God’s knowledge.
God might be watching the future ‘concurrently’ with the present; and he might not be interfering with what he sees; but the fact that there is a future for him to see is enough to undermine the sense in which human beings can be said to have control over their own lives. It seems to me that this was the realization that forced all of the classical Jewish philosophers to search for an alternative solution to the medieval problem of free will.
Maimonides and Gersonides
Maimonides wasn’t willing to give up on free will, nor was he willing to give up on God’s foreknowledge. What was more, Maimonides wasn’t willing to take the Boethiun route out of the problem. In his Guide for the Perplexed, Maimonides arrives at the conclusion that (1) we are free, that (2) God has foreknowledge and that (3) it is impossible for human beings to understand how (1) and (2) are compatible. As Seymour Feldman points out, Maimonides’ conclusion is simply an echo of Rabbi Aqiba’s statement that ‘all is foreseen, yet permission is given’ (Mishna, Avot 3:15). On the surface, it seems as if he’s completely side-stepping the problem. But, Maimonides tries to justify his startling conclusion upon his theory of religious language.
Maimonides was famously dismissive of the power of language to describe God discursively. As Feldman explains: ‘Since the gap between God and man is infinite, no property true of man can be predicated by God unequivocally. Everything we say about God is, for Maimonides, radically equivocal.’ In order to steer ourselves away from vulgar falsehoods, Maimonides famously suggests that the best way to talk about God, if we have to talk about him at all (Maimonidies was fond of the psalm that promotes silence as the most appropriate form of praise of God), is to negate predicates of him. Instead of saying that God is wise, we would be better off saying that God is not foolish.
There are many reasons for adopting this negative theology, and this isn’t the place for an extensive exploration of the Maimonidean philosophy of language, but, as Feldman points out, the most basic reason for preferring the negative formulation, in this instance, over the positive formulation, is that God’s wisdom is ‘so unique and incommunicable that we should explicitly recognize its absolute difference’ from the normal connotation of the term ‘wisdom’. The use of the negative formulation is thus a tool for communicating the complete transcendence of the Divine.
Just as, for Maimonides, God’s wisdom is radically different from any other instance of wisdom; so too is God’s knowledge radically different to human knowledge. Maimonidies lists five ways in which Divine knowledge is evidently different to human knowledge. And, because of the fifth item on the list, Maimonides’ solution to the problem has been wrongly equated to that of Boethius. The fifth item of the list is that even though God knows future contingent events, His knowledge does not annul the contingency of these events. That obviously looks like his solution to the freewill problem, and it obviously reminds us of Boethius. But we shouldn’t overlook the other four items on the list! Here is the list in full:
(1) Despite knowing many things, God knows them in such a way that his knowledge is indivisibly one.
(2) God knows non-existent things
(3) God’s knowledge ranges over infinite classes
(4) Even though he knows temporal facts, his knowledge is eternal and unchanging
(5) Even though God knows future contingent events, His knowledge does not annul the contingency of these events.
These five qualities illustrate how radically different God’s knowledge must be from any sort of knowledge with which we’re familiar. How does Maimonides arrive at this list? The first item is there because we know God to be One, in some radical sense of the word ‘one’, and we also want to attribute to him knowledge of many things. I’m not exactly sure what (3) is doing on the list other than attributing omniscience to God, as opposed to the finite set of propositions that any human being can know at any given time. (4) is there because we know God to be unchanging but want to attribute to him knowledge of temporal facts. (5) is there to solve the freewill problem, seemingly a la Boethius. The question is, what is (2) doing on the list?
Unlike humans, God is said to be able to know things that don’t exist. If we look back to my initial formulation of the medieval free will problem, we can see the role that is played by the factivity of knowledge (premise 3). Knowledge is thought to be a relation (whether direct or mediated via a true proposition) to a fact. If God knows the future, then there must be some future fact to which he stands related. The existence of these future facts is what seems to limit my freedom. Boethius’ attempt at a solution is unhelpful because it doesn’t get rid of these future-facts. Maimonides does. Thus Maimonides adds significantly to Boethius. Not only must God’s knowledge not influence our actions, it mustn’t stand him related to any existent facts about the future.
Maimonides takes it that we can’t understand what most of the items on this list mean? What does it mean to know many things in an indivisible way? What does it mean to have infinite knowledge? What does it mean to stand related to a fact that doesn’t exist? Does that mean that there are non-existent facts in some sort of Meinongian world of shadow being? I don’t think so. The idea is that we’re not supposed to be able to understand the items on the list, even though our theological commitments have led us to utter them! To be able to understand how God can know our future acts and yet leave us completely free is clearly beyond our ken; it would entail an intimate knowledge of Divine knowledge, which is unique, incommunicable, and, therefore, nothing like human knowledge. To be able to understand this would be to be God himself. And thus, Maimonides concludes, along with Rabbi Aqiba, that ‘all is foreseen, yet permission is given’; it’s just that we can’t claim to understand how this is possible.
To sum up: God can know things that don’t exist. God’s knowledge is different to ours. It doesn’t require any existent entity such as a fact as its object. As far as Maimonides is concerned, we can say that the future doesn’t yet exist. That is what allows for our freedom. But God can still know the future because God’s knowledge is nothing like human knowledge. To ignore the second item in Maimonides’ list, as many people do, is to miss the point completely. Maimonides is saying, contra Boethius, that if the future exists, we cannot be free. He concludes that the future doesn’t exist, even though God knows it.
Gersonidies famously attacks Maimondies’ solution to our problem. His attack stems from dissatisfaction with Maimonides’ theory of language. His argument moves in the following three stages. Firstly, there must be some continuity of meaning between the way we talk about the world and the way that we talk about God. If this had not been the case, then we wouldn’t even be able to formulate relevant negative expressions about God. How can the phrase that God is not foolish be meaningful to us if the word ‘foolish’ doesn’t bear some relation to the word as it is normally used? And, if the word ‘foolish’ does have something in common in both religious and mundane discourse, then so would the word ‘wise’, which leaves us wondering why we couldn’t just have started by saying that ‘God is wise’, without the unnecessary detour through his not being foolish.
Secondly, without the assumption that our everyday language has some continuity of meaning even when applied to God, it becomes impossible to know which negations are appropriate and which are not. Throughout the early chapters of the Guide, Maimonides borders on the obsessive as he tries to demonstrate the non-corporeality of God and the appropriateness of negating corporeal predicates of God. This is because corporeality entails certain imperfections. But this whole Maimonidean obsession seems to depend upon the very continuity of language that he wants to deny. Feldman explains:
[W]e only know that these attributes imply such imperfections on the basis of information drawn from the world around us, which are expressed in and through human language. In other words, we assume some continuity in our negations of certain attributes. Why not in our affirmations? If a certain property implies an imperfection when ascribed to man, and hence we refrain from applying it to God, then we assume that the attribute, or predicate, has at least something in common when it is applied in both cases. Otherwise, we could say that when the attribute, which in its ordinary uses implies an imperfection, is applied to God it does not imply an imperfection. Thus, we would not be prevented from saying “God has a nose.” The limitations that we place on certain attributes imply that these attributes have a certain core meaning in all the contexts of their use. And it is only because of this core meaning that we are able to exclude this attribute from being a divine attribute.
The third stage, Feldman explains, is to suggest a replacement for the Maimonidean theory of language. Gersonidies accepts that predicates are not true of God in the same way that they are true of human beings. God, for instance, is the primary exemplar of mercy. Anything that is merciful is merciful in respect of their resembling to some degree this quality of God. For that reason, nothing can be merciful in exactly the same way that God is merciful. ‘On this picture, God is the model for human action. But he can only be a model if there is some measure of similarity between divine and human mercy. Maimonides’ theory of negative attributes places God so far beyond the human ken that we no longer understand what it is to worship and imitate Him.’
Having dismissed Maimonides’ theory of language, it is no longer an option for Gersonides to have his cake and eat it. For Gersonides, the medieval problem of free will is a real problem. We will either have to deny human freedom or deny that God knows the future. For Gersonides, it is much more important to salvage the notion of free will. Without free will, the commandments of the Torah make no sense. Why issue a set of commandments for human beings if they have no freedom to obey or disobey? Gersonides therefore chooses to limit God’s knowledge. In fact, he follows Aristotle in limiting God’s knowledge to exclude knowledge of particulars. But this isn’t directly relevant to the Gersonidean theory at hand. Even if God were able to have direct knowledge of particulars, Gersonides maintains that he still wouldn’t be able to have any knowledge of a contingent future particular. The future doesn’t exist. If we want to maintain that we are free to regulate our own choices, we have to accept that the future doesn’t exist. If the future doesn’t exist, and if knowledge is factive, then God cannot have knowledge of the future because there is no future for him to know.
An omniscient intellect would know future necessities. If, at t1, I pull the trigger of a gun which, given the surrounding environmental factors, causally necessitates a determinate future trajectory of the bullet, then, God knows the future trajectory of that bullet, because he knows future necessities. But, if we are free, then much of our future is yet to be written. Future contingencies do not yet exist, and, therefore, even an omniscient intellect would have no knowledge of them. To be omniscient is to know everything that there is. The future isn’t, so it’s even closed off to omniscient eyes.
Gersonides and Maimonides are worlds apart, theologically, on this issue. For Maimonidies, God knows everything that we’re going do (although Maimonides can’t claim to understand what this thesis actually means). For Gersonides, God is ignorant of future contingencies; he is ignorant of the choices that we are going to make freely. But, there is an important thesis that these two thinkers share. The point of commonality is this: both thinkers accept that if we are truly free, then there can be no atemporal perspective from which it could be said that our futures ‘already’ exist. For Maimondes and Gersonides, unlike Boethius, the medieval freewill problem can’t be solved until we accept the following thesis: even though the past and the present exist, the future is yet to be written; the future doesn’t exist from any perspective whatsoever.
It is only the failure to recognise the significance of the second respect in which Mainomides thinks that Divine and Human knowledge differ – i.e., a failure to recognise that, according to Maimonides, God’s knowledge extends to things that don’t exist – that could lead you to misinterpret his position, placing two much weight upon the fifth respect of difference, such that you could wrongly identify the positions of Maimonides and Boethius.
One might counter my claim. One might argue that Maimonides is committed to the existence of the future. After all, Maimonides insists that the miracles reported in the Bible were written into the laws of nature already during the six days of creation (cf. Maimonides’ commentary to the Mishna in Avot 5:6). This entails that, at the beginning of time, God saw that the Jewish people would, at a precise moment in history, require the sea to split, causing God to write into the nature of the tides that, at that exact moment, the sea would split. This in turn might lead you to think that, according to Maimonides, the future need of the Jewish people already existed at the moment of the creation of the world; that, from some atemporal perspective, a la Boethius, the future already exists.
But, for two reasons, this would be an over-interpretation. Firstly: Maimodies’ position is precisely that God can know the future even though the future doesn’t exist. Secondly: the eventualities that God foresaw, such as the need for the sea to split at a very specific moment in time, may not have been future contingencies; perhaps, for one reason or another, such as the laws that govern societal development and international-relations, the historical eventualities in question were actually future necessities – which God could have knowledge of even according to Gersonidies.
It seems that careful attention to the words of Maimonides reveal that he had more in common with Gersonides on this issue than ordinarily meets the eye: both of them shirk a Boethian solution; both of them deny the existence of the future in order to secure human freedom. Boethius, Maimonides and Gersonides all believe that we are free and that God is omniscient. But what Maimonides and Gersonides have against Boethius is their claim that human freedom is incompatible with the existence of the future; that human freedom entails the rejection of the metaphysics of eternalism.
 Boethius, The Consolation of Philosophy, V.6, 25–32. Translated by Victor Watts. London: Penguin Classics, in 1999.
 Zagzebski, L. ‘Foreknowledge and Free Will’, The Stanford Encyclopedia of Philosophy (Fall 2008 Edition), Edward N. Zalta (ed.), http://plato.stanford.edu/archives/fall2008/entries/free-will-foreknowledge/
 Acquinas, T. Summa Contra Gentiles, I.66. Translated by Anton Charles Pegis. Indiana: University of Notre Dame Press, in 1991.
 Boethius, The Consolation of Philosophy, V.6, 27–30. Translated by Victor Watts. London: Penguin Classics, in 1999.
 Touger, E., What We Believe. Merkos Linyonei Chinuch, 2000.
 Sacks, J., Covenant & Conversation; Genesis: The Book of Beginnings; Maggid, in 2009. Cf. pp. 281-3.
 R. Saadya, Emunot Va’Deot. I quote from Alexander Altmann’s scholarly, though abridged, translation. The Book of Doctrines and Beliefs, East and West Library, 1946, pp. 122-3.
 Feldman, S., Feldman, S. (1987) The Wars of the Lord. Volume 2. New York: The Jewish Publication Society, pg. 77.
 Psalm 65:2, cited in Maimonides Guide, I.59.
 Feldman, S., Feldman, S. (1987) The Wars of the Lord. Volume 2. New York: The Jewish Publication Society, pg. 77.
 Guide, III.20.
 cf. Milchemet Hashem III.3
 Feldman, S., Feldman, S. (1987) The Wars of the Lord. Volume 2. New York: The Jewish Publication Society, pg. 79.
 Ibid., pg. 80.
 Milchemet Hashem III.4