4. So something than which nothing greater can be thought cannot be

4. So something than which nothing greater can be thought cannot be

Anselm and the ontological argument

We can think of something existing which cannot be thought not to exist.

Such a thing would be greater than something which can be thought not to exist.

So something than which nothing greater can be thought cannot be something which can be thought not to exist.

So something than which nothing greater can be thought cannot be thought not to exist.

Or, in Anselm's words:

For something can be thought to exist that cannot be thought not to exist, and this is greater than that which can be thought not to exist. Hence, if that than which a greater cannot be thought can be thought not to exist, then that than which a greater cannot be thought is not the same as that than which a greater cannot be thought, which is absurd. Something than which a greater cannot be thought exists so truly then, that it cannot be even thought not to exist. anselm and gaunilo Many have thought that the arguments of Proslogion 2 and 3 are bad ones. Gaunilo is a case in point. According to him, Anselm is wrong, because (a) we should not think of God as being in the mind or understanding, and (b) Anselm's case for God's existence entails unbelievable consequences.

Gaunilo challenges the claim that that than which nothing greater can be thought is in the understanding by insisting on the incomprehensibility of something than which nothing greater can be thought.

His basic point is: we do not understand what God (or that than which nothing greater can be thought) is, so God (or that than which nothing greater can be thought) is not in the understanding. He writes:

I can so little think of or entertain in my mind this being (that which is greater than all those others that are able to be thought of, and which it is said [i.e. by Anselm] can be none other than God Himself) in terms of an object known to me either by species or genus, as I can think of God

Himself . . . For neither do I know the reality itself, nor can I form an idea from some other things like it since, as you [i.e. Anselm] say yourself, it is such that nothing could be like it.14

Cambridge Companions Online © Cambridge University Press, 2006 Cambridge Companions Online © Cambridge University Press, 2006

brian davies According to Gaunilo, "Something than which nothing greater can be thought" is nothing but "a verbal formula," a string of words which fails to furnish the basis of a proof of God's existence.

Gaunilo's second main criticism of Anselm comes in the following (much quoted) passage:

They say that there is in the ocean somewhere an island which, because of the difficulty (or rather the impossibility) of finding that which does not exist, some have called the "Lost Island." And the story goes that it is blessed with all manner of priceless riches and delights in abundance, much more even than the Happy Isles, and having no owner or inhabitant, it is superior everywhere in abundance of riches to all those islands that men inhabit.

Now, if anyone tell me that it is like this, I shall easily understand what is said, since nothing is difficult about it. But if he should then go on to say, as though it were a logical consequence of this: You cannot any more doubt that this island that is more excellent than all other lands exists somewhere in reality than you can doubt that it is in your mind; and since it is more excellent to exist not only in the mind alone but also in reality, therefore that it must needs be that it exists. For if it did not exist, any other land existing in reality would be more excellent than it, and so this island, already thought by you to be more excellent than others, will not be more excellent. If, I say, someone wishes thus to persuade me that this island really exists beyond all doubt, I should either think that he was joking, or I should find it hard to decide which of us I ought to judge the bigger fool.15

Here Gaunilo seems to be saying (a) that thinking akin to Anselm's would successfully prove the existence of things we cannot seriously believe in, and (b) that something must therefore be wrong with Anselm's reasoning.

Is Gaunilo right in his critique of Anselm? A notable feature of it is its frequent failure to focus on Anselm's key formula: "that than which nothing greater can be thought." Sometimes Gaunilo more or less quotes this expression exactly. Mostly, however, he construes

Anselm as arguing for the existence of something which is, in fact, greater than all other things. Hence, for example, and early in his reply, he represents Anselm as holding that, if God exists only in the mind, "that which is greater than everything would be less than some thing and would not be greater than everything."16 This is also how Gaunilo seems to be understanding Anselm as he offers his lost island argument, for here he denies that there has to be a best island just because we can conceive of such a thing.

Cambridge Companions Online © Cambridge University Press, 2006 Cambridge Companions Online © Cambridge University Press, 2006

Anselm and the ontological argument

But Anselm is not seeking to prove the existence of a best anything. As he says in response to Gaunilo:

You often reiterate that I say that that which is greater than everything exists in the mind, and that if it is in the mind, it exists also in reality. However, nowhere in all that I have said will you find such an argument. For "that which is greater than everything" and "that than which a greater cannot be thought" are not equivalent for the purpose of proving the existence of the thing spoken of.17

That which is greater than everything, Anselm adds, can be thought of as possibly not existing, while that than which a greater cannot be thought cannot be thought of as possibly not existing. And Anselm is clearly right in at least one respect here: for "that which is greater than everything" is certainly not equivalent to "that than which a greater cannot be thought." Someone can believe that X is the greatest existing thing without needing to describe it as "that than which a greater cannot be thought." And that which is, in fact, greatest could be very imperfect indeed.

Yet, even supposing that Gaunilo's island is really not analogous to what Anselm means by "something than which nothing greater can be conceived," might he not still reasonably call on us to accept his first line of criticism of Anselm? Might he not ask us to wonder whether "that than which a greater cannot be thought" signifies anything intelligible? Might he not fairly suggest that we cannot really conceive of such a thing, that it cannot truly be said to be "in the understanding"?

Well, perhaps he can. For could there be something than which nothing greater could be thought? Do we really know that there could be such a thing? Maybe we do. But how? Anselm does not tell us. He assumes that as soon as we hear the phrase "that than which a greater cannot be thought," we should understand it as a label signifying a possibly existing being. But should we? Might there not, for example, be no limit to conceivable greatness? Let us suppose that we are thinking of X and that we cannot, as it happens, think of anything greater than X. Does it follow that X is something than which nothing greater can be thought? Obviously not. For, maybe, somewhere or sometime, someone might be able to think of something greater than X. Let us call this new something Y. Does it follow that Y is something than which nothing greater can be thought?

Cambridge Companions Online © Cambridge University Press, 2006 Cambridge Companions Online © Cambridge University Press, 2006

brian davies Obviously not. For, maybe, somewhere or sometime, someone might be able to think of something greater than Y. And how do we know that this process of being able to think of something greater cannot proceed ad infinitum? Anselm, at any rate, does not show that it cannot.

In that case, however, he has not proved that something than which nothing greater can be thought is in the mind. In reply to Gaunilo he insists that it is because (a) Gaunilo believes in God, so must therefore have in mind the notion of something than which nothing greater can be thought;

(b) We can understand the notion of there being something which lacks a beginning, which cannot not exist, which exists whole and entire at all times and places.18

Anselm's "something than which a greater cannot be thought" formula is not, however, forced on Gaunilo simply by his subscription to belief in God's existence. It is not even forced on him by his allegiance to Christianity. Many orthodox Christians have believed in God without claiming that God is "something than which a greater cannot be thought." Furthermore, even if " lacks a beginning, cannot not exist, and exists whole and entire at all times and places" is truly predicable of something, it does not follow that the thing in question is "something than which nothing greater can be thought."

We might wonder whether that formula signifies anything thinkable even though we might believe that there is something which lacks a beginning, cannot not exist, and exists whole and entire at all times and places. In this sense, we might agree with Gaunilo's claim that God is not "in the mind."

We might also agree with it for another reason. As we have seen, part of Gaunilo's case against Anselm rests on the suggestion that God cannot be thought since, even if he exists, he belongs to no genus or species. Without denying God's existence, Gaunilo here seems to be saying (a) that he cannot form a concept of God as he can of other things, (b) that his inability to do so means that God does not exist in the mind (his mind, anyway), and (c) that Anselm is therefore wrong to say that the existence of God can be proved just from an understanding of what God is. And Gaunilo is making a reasonable point here if, indeed, God, or that than which nothing greater can be thought, is taken to belong to no known genus or species, and if Cambridge Companions Online © Cambridge University Press, 2006

Cambridge Companions Online © Cambridge University Press, 2006

Anselm and the ontological argument

one cannot form a concept of what does not belong to any genus or species.

Yet "something than which nothing greater can be thought" is a formula we can work with negatively, so to speak, and with an eye on the notion of something that could be better. Though we might find our minds going blank when faced by the phrase "something than which nothing greater can be thought," we can surely make sense of the idea that X is not something than which nothing greater can be thought if we can think of something greater than X. We can surely say with some confidence that, for example, a rabid dog is not something than which nothing greater can be thought, for we can think of something greater than a rabid dog. And with this kind of example in mind Anselm might legitimately defend himself against the claim that "something than which nothing greater can be thought" cannot be legitimately employed in a case for God's existence in re. He might argue like this:

God is something than which nothing greater can be thought.

If we can think of something greater than X, then X is not God.

We can think of something greater than anything which exists only in intellectu.

So something existing only in intellectu cannot be God.

So God does not only exist in intellectu.

And he might add:

God is something than which nothing greater can be thought.

We can think of something which can fail to exist.

Something which can fail to exist is less great than something which cannot fail to exist.

So something which can fail to exist cannot be God.

10.

So God is not something which can fail to exist.

And this might be all that Anselm is arguing in Proslogion 2 and 3.

We need not take him to be claiming that "something than which nothing greater can be thought" can be proved to be a coherent formula or a description of something which could possibly exist. All we need to take him to be claiming is that something which is only in intellectu, and which might possibly fail to exist, is not something than which a greater cannot be thought. And that claim, perhaps, is hardly absurd. After all, is it not plausible to suggest that something

Cambridge Companions Online © Cambridge University Press, 2006 Cambridge Companions Online © Cambridge University Press, 2006

brian davies that exists in re by nature has the edge over something which is nothing but a figment of someone's imagination? the logic of anselm's reasoning Validity Anselm's Proslogion 2 argument seems to be formally valid. It begins with something like a definition: "God" is "something than which nothing greater can be thought," and it goes on to assert that something than which nothing greater can be thought is in the understanding. Then it introduces the suggestion that something than which nothing greater can be thought is not in re (the position of the "Fool"). Anselm's clear objective is to show that this suggestion cannot be true (his argument is what is known as a reductio ad absurdum: it aims to prove that, given certain premises, a particular assertion leads to contradiction and is, therefore, false). And (regardless of how we translate "si enim vel in solo intellectu est potest cogitari esse et in re quod maius est") Anselm moves to his conclusion by arguing: a.

If something is in intellectu but not in re, something greater than it can be thought. b.

If something than which nothing greater can be thought is in intellectu but not also in re, then something greater than it can be thought (from [a]). c.

Something than which nothing greater can be thought is in the understanding but not also in reality. d.

Something greater than something than which nothing greater can be thought can be thought (contradictory conclusion from [c]).

Anselm's reasoning here seems logically impeccable: given his premises, his conclusion appears inescapable. "Something greater than something than which nothing greater can be thought can be thought" is clearly self-contradictory. Given Anselm's reasoning, therefore, we ought to conclude that it is, indeed, absurd to say that something than which nothing greater cannot be thought does not exist in re. We may, of course, reject this reasoning by rejecting some of its premises. But those premises indeed seem to entail that it is

Cambridge Companions Online © Cambridge University Press, 2006 Cambridge Companions Online © Cambridge University Press, 2006

Anselm and the ontological argument

absurd to deny that something than which nothing greater can be thought exists in re.

Premises Are Anselm's premises true, however?

With respect to the first one, could it be that we have no reason to think of God as something than which nothing greater can be thought? Thomas Aquinas (1224/6–1274) rightly notes that not everyone has taken the word "god" (deus) to signify "something than which nothing greater can be thought."19 Yet Anselm is surely entitled to stipulate what he means by "God" for the purposes of an argument. Furthermore, insofar as he is seeking to engage with what we might call the Judeo-Christian concept of God, Anselm's

"something than which nothing greater can be thought" expression seems not inappropriate. As I have noted, it echoes the way in which Augustine speaks of God. In addition, it is hard to conceive of anyone in the Judeo-Christian tradition being prepared to say "There might, after all, be something greater than God." The idea that God is unsurpassably great seems to be part and parcel of Judeo-Christian theism, and to say that God is something than which nothing greater can be thought seems to be a succinct way of capturing what those in the Judeo-Christian tradition mean by "God." As Norman

Malcolm puts it: "God is usually conceived of as an unlimited being.

He is conceived of as a being who could not be limited, that is, as an absolutely unlimited being. This is no less than to conceive of Him as something a greater than which cannot be conceived."20

With respect to Anselm's second premise, could it be that "something than which nothing greater can be thought" does not describe anything possible? As I have said, one might think that it does not (as some have suggested, one might think that it may be compared with expressions like "greatest prime number"). As I have also said, however, one can entertain the thought of something than which nothing greater can be thought so as intelligibly to conclude that something

(e.g. a rabid dog) is not something than which nothing greater can be thought. The expression "something than which nothing greater can be thought" is hardly unintelligible. If that is so, however, we might, as Anselm does, ask whether there is something than which nothing

Cambridge Companions Online © Cambridge University Press, 2006 Cambridge Companions Online © Cambridge University Press, 2006

brian davies greater can be thought in re. We might also ask whether more can be said of something than which nothing greater can be thought other than that it exists in re.

Some would say that Anselm needs to demonstrate that there is no intrinsic absurdity in the notion of something than which nothing greater can be thought. But why should he feel the need to do so? And how could he succeed in doing so? People might claim that there are arguments to show that such and such an expression signifies nothing that could possibly exist, and politeness would then require us to examine their arguments. But what can one do to demonstrate in the abstract that a particular form of words signifies the concept of something possible (or does not express an impossibility)? By various arguments one might seek to demonstrate that there could be something rightly referred to by the form of words in question.21 Such arguments, however, will inevitably depend on what those who offer them take, without argument, to be possible.

Yet what of the premise "If something is in intellectu but not in re, something greater than it can be thought"? Many would reject it because, they would argue, it has to mean that existence in re is a perfection and, therefore, a property or characteristic of things, which it is not. And it certainly seems odd to speak of existence as a perfection, if only because " exists" does not specify a way in which that of which it is predicated differs from anything. When we say that something has a perfection, we generally seem to be noting some particular way in which it differs from or resembles a limited number of other things. But " exists" cannot serve to distinguish one thing from another since, so to speak, everything exists.

Anselm's argument, however, does not call on us to think of existing as a particular perfection. It asks us to accept (a) that thinking of something than which nothing greater can be thought is not the same as thinking of something than which a greater can be thought, and (b) that something only in intellectu cannot be thought of as something than which nothing greater can be thought. But (a) and (b) here are plausible claims. (a) is clearly self-evident. We would contradict ourselves by denying it. And (b) seems true since we surely can think of something greater than something which has nothing but the status of existing as an idea in someone's mind. As Anselm himself says, we can think of something which cannot fail to exist. And such a thing, so one might plausibly suggest, is greater than something

Cambridge Companions Online © Cambridge University Press, 2006 Cambridge Companions Online © Cambridge University Press, 2006

Anselm and the ontological argument

which exists only as an idea in someone's mind (and which, considered as such, can certainly fail to exist). Following Kant's discussion of what he called the "ontological proof," philosophers have often attacked the Proslogion while echoing Kant's assertion that "Being is obviously not a real predicate."22 And perhaps there is much to be said for this assertion, for it could be construed as suggesting that, as Gottlob Frege (1848–1925) put it, " exists" is not a first-level predicate (i.e. a predicate which ascribes a distinguishing property to an object or individual – by contrast, for example, with " is made of plastic" or " is bald"). And there are well-known arguments in favor of this suggestion.23 In the Proslogion, however, Anselm does not seem to be arguing that " exists" is or is not a first-level predicate.

Rather, he seems to be contrasting objects of thought and saying that one of them cannot be sensibly described as something than which nothing greater can be thought – the one in question being, of course, something which exists only in the mind (like, to use his example, the painting conceived by an artist thinking of what he might proceed to create). Anselm's main idea seems to be that of two thought objects, the first being something than which nothing greater can be thought, and the second being something which is only in intellectu, the first is greater than the second.

Thinking of God Yet from this idea it does not follow that the Fool is wrong to deny that God is in re just because he is prepared to accept someone's insistence that the word "God" means "something than which nothing greater can be thought." The Fool could always say: "I am happy to allow it to be stipulated that 'God' means 'something than which nothing greater can be thought.' But on that basis alone I do not have to agree that there really is something (God) than which nothing greater can be thought."

In other words, the Fool might rightly insist that one cannot establish the existence of something in re on the basis of a stipulative definition. Such definitions do not, by themselves, tell us anything when it comes to what really exists. For the most part, we produce definitions of what we take to be really existing things, and we do so on the basis of what we believe ourselves to know about them. If someone asks us what, for example, an elephant is, we would probably seek

Cambridge Companions Online © Cambridge University Press, 2006 Cambridge Companions Online © Cambridge University Press, 2006

brian davies to define the word "elephant" while relying on the reports of zoologists. Mere definitions of names, however, do not, by themselves, mean that there is anything corresponding to them.

So we might side with Anselm's fool if we start by supposing that "a being than which nothing greater can be conceived and which cannot be conceived not to exist" is, so to speak, a phrase to be read within quotation marks. In other words, the Fool could always say that "a being than which nothing greater can be conceived and which cannot be conceived not to exist" is just an expression which some people (though not he) use when talking (perhaps mistakenly) of what they take to be something.

Children refer to Santa Claus, and we can buy into their talk so as to agree, for example, that it is silly, for anyone who is seriously thinking about Santa, to deny that he delivers presents on December 25. But we would not thereby be committed to concluding that someone called Santa Claus delivers presents on December 25, since we do not seriously think of anything or anybody as being

Santa Claus. By the same token, so the Fool might reason, buying into the talk of believers, we might say that what believers think of as a being than which nothing greater can be thought, and which cannot be thought not to exist, certainly has to be asserted to exist by anybody seriously thinking of this object as such. But this does not mean that we are, on pain of self-contradiction, committed to thinking of anything being something than which nothing greater can be thought and which cannot be thought not to exist. Or, as Aquinas writes:

Even if the meaning of the word "God" were generally recognized to be "that than which nothing greater can be thought," nobody accepting this definition would thereby be forced to think of God as existing in the real world rather than existing as thought about (in apprehensione intellectus).

If one does not grant that there is something than which no greater can be thought, it cannot be proved that God [sc. as so defined] really exists. And those who hold that God does not exist do not grant that there is something than which no greater can be thought.24

Suppose, however, that we are prepared to think seriously of something than which nothing greater can be thought. In doing so, we would not have to suppose that there is any such thing in reality, for we can seriously think of what does not exist – as a painter can

Cambridge Companions Online © Cambridge University Press, 2006 Cambridge Companions Online © Cambridge University Press, 2006

Anselm and the ontological argument

think of a painting yet to be painted. All we would have to do is to entertain the thought of something than which nothing greater can be thought without presupposing either its nonexistence or its real existence. All we would have to do is to take it to be the thought of something possible, something the nature of which we might further reflect on. What then?

Well, if we are so prepared to think seriously of something than which nothing greater can be thought, it seems that we are committed to acknowledging an absurdity in the claim that it is only in the mind and able not to exist. At any rate, we are so committed if Anselm is right to say that we can think of something greater than what is only in the mind and able not to exist. Or, to put the point in another way, Anselm's argument is effective against us if we are prepared to refer to that than which nothing greater can be thought

"constitutively" as opposed to "parasitically."25

Our basic way of referring is constitutive. That is to say, when we refer to or think of things, we commonly do so without distancing ourselves from what other people think or believe. If I say "Let's think about London," we would normally go on to do so without worrying whether or not "London" is the name of a place in Britain, or a name used only in works of fiction or by deluded people who take it to be the name of a real city.

Yet suppose I say "Let's think about ghosts." We can do so since we can refer to ghosts parasitically. That is to say, we can latch on to what has been said about ghosts and we can think about them only on the basis of that. We can think about ghosts without being committed to anything other than a claim to understand what has been said or believed about them whether seriously (by people we may think of as deluded) or in works of fiction (by people whose writings we admire and find entertaining).

Now think about God considered as something than which nothing greater can be thought. You could think of God parasitically here.

You could take "God" to be a word which some people understand as meaning "something than which nothing greater can be thought."

And though you might come to agree that "God, considered as something than which nothing greater can be thought, is only in the mind and might fail to exist" is somehow contradictory, you would not be committed to supposing that there actually is a God who is (a) something than which nothing greater can be thought, (b) something

Cambridge Companions Online © Cambridge University Press, 2006 Cambridge Companions Online © Cambridge University Press, 2006

brian davies which is not only in the mind, and (c) something which cannot possibly not exist. But you would be so committed if you are prepared constitutively to think about something than which nothing greater can be thought, and if you are prepared to accept the premises

Anselm employs in Proslogion 2 and 3, and if you take his reasoning there to be formally valid.

If we constitutively think of something than which nothing greater can be thought, then we ought to concede that we cannot be (seriously) thinking of something which is nothing but an idea in the mind (like the thought of an unpainted painting). And if we constitutively think of something than which nothing greater can be thought, and if "not being able not to exist" signifies a possible perfection or great-making quality, we ought to concede that we cannot be (seriously) thinking of something which might be able not to exist. Someone might say that this claim only amounts to the suggestion that if God exists then He necessarily exists. Yet how are we to understand the "if " here? It seems to imply that it is possible that God does not exist. It also seems to imply that we might consistently assert that "If God exists, and it is possible that He does not, then He necessarily exists." But is this last assertion not self-contradictory?

Anselm and thinking of God Is Anselm arguing along the lines that I have just put forward? Does he view himself as starting from the possibility of constitutively referring to something than which nothing greater can be thought?

One might think that he does not since this would leave him assuming to begin with what he claims to be out to prove in Proslogion 2 and 3 (i.e. that God is in re and cannot be thought not to exist).

Anselm, however, is perfectly aware that one can think seriously of something without asserting or presupposing that it really exists.

One might reply to this point by suggesting that all Anselm is arguing is that a definition of God implies God's real existence. Yet Anselm does not say anything, either in the Proslogion, or in the reply to Gaunilo, to indicate that he is reasoning as simply as this. What he does say, however, clearly shows that he thought it absurd seriously to think of something than which nothing greater can be thought while also insisting that this is something which might not exist in re and which might possibly not exist.

Cambridge Companions Online © Cambridge University Press, 2006 Cambridge Companions Online © Cambridge University Press, 2006

Anselm and the ontological argument

So perhaps we might well read Anselm along the lines that I indicate above. An additional reason for doing so is that Anselm is unlikely to be taking his "Fool" to be referring parasitically to God when he supposes him to have in mind something than which nothing greater can be thought. It is implausible to suppose that Anselm

(or, for that matter, the author of Psalms 15 and 53) ever encountered an atheist of the sort who would say "Yes, I know what people mean by the word 'God,' but though I perfectly understand their way of talking, I cannot take it any more seriously than I take the talk of children when they refer to Santa Claus." Atheists like this just did not belong to Anselm's world (or to the world of the authors of the Psalms), so we should therefore not suppose him to be arguing with them in the Proslogion. That work (offered, note, as a sequel to the Monologion) clearly has in mind an audience composed of people whom Anselm would have expected to entertain the notion of something than which nothing greater can be thought constitutively rather than parasitically.

In this connection it is, perhaps, worth noting the tone which Anselm sets at the start of his reply to Gaunilo. He writes: "Since it is not the Fool, against whom I spoke in my text, who takes me up, but one who, though speaking on the Fool's behalf, is an orthodox

Christian and no fool, it will suffice if I reply to the Christian."26

This remark strongly suggests that Anselm saw his Proslogion 2 and 3 line of reasoning as directed to someone able and willing to think constitutively of God as something than which nothing greater can be thought. It suggests that Anselm is not writing for people who can only have, or are only willing to have, a concept of God in the way that we (most of us, anyway) have a concept of wizards or Santa

Claus. It suggests that Anselm, though without explicitly assuming that God exists, is chiefly concerned to argue that it is not foolish to believe that there is a God in re. It suggests that he is out to show that thinking seriously (as opposed to parasitically) of God cannot be intellectually reconciled with holding that God might not exist.

Notice, however, that Anselm's remark about Gaunilo being an orthodox Christian need not, taken in context, be read as suggesting that Anselm is merely preaching to the choir, that he is merely explaining what believers mean by "God" (as, for example, Barth said that he is). Rather, it suggests that anyone (believer or not) who can seriously think of something than which nothing greater can

Cambridge Companions Online © Cambridge University Press, 2006 Cambridge Companions Online © Cambridge University Press, 2006

brian davies be thought (anyone who can entertain this as a legitimate object of thought, and anyone who is prepared to accept what this object of thought implies) cannot reasonably conclude that God is but an idea in people's minds. People sometimes ask whether Anselm in Proslogion 2 and 3 (taken together with his reply to Gaunilo) is (a) trying to take those who believe in God into a deeper understanding of what

God is, or (b) trying to establish the existence of God without presupposing that God exists. The answer, perhaps, is that he is seeking to do both of these things. There is no reason why one cannot attempt to enrich the understanding of those who believe in God without also, and simultaneously, aiming to show that God exists (in re and not only in intellectu) without presupposing that there is any God at all. Given the "faith seeking understanding" bell that he rings in Proslogion 1, Anselm, in Proslogion 2 and 3 (and in his reply to Gaunilo) is clearly talking to believers. Taken as a whole, however, his line of thinking concerning that than which nothing greater can be thought does not seem bluntly to presuppose that there is a God.

Rather, it seems to be conceived of by Anselm as claiming that, without presupposing the existence of God, and given certain premises, it would be absurd for someone to say that there is no God. conclusion If that is how Anselm thinks of his Proslogion 2 and 3 arguments (and allowing for his reply to Gaunilo) then he defends himself well. His reductio argument seems valid, and his premises are hardly incredible. He has clearly not shown that everyone has to conclude that God is in re and cannot be thought not to exist (for some people may insist on thinking only parasitically of God as something than which nothing greater can be thought). Be that as it may, however,

Anselm has plausibly explained how we cannot seriously think of God as something than which nothing greater can be thought without also being committed to the conclusion that God is in re and that God cannot possibly not exist. And, since he does so without presupposing that there really is a God, one might well take him to have defended belief in God's existence in a significant way. People have offered many arguments for God's existence. Some, for instance, have claimed that God exists on causal grounds (as Anselm himself does in the Monologion). But there are more ways than one to skin a Cambridge Companions Online © Cambridge University Press, 2006

Cambridge Companions Online © Cambridge University Press, 2006

Anselm and the ontological argument

cat, and Anselm's Proslogion way is impressive. That, presumably, is why it has generated attention for several hundred years and is still being studied and discussed. notes 1. Immanuel Kant, Critique of Pure Reason a590/b618.

2. For an account and discussion of the ontological argument from Anselm to the present, see Graham Oppy, Ontological Arguments and Belief in God (Cambridge: Cambridge University Press, 1995). For some primary texts, see Alvin Plantinga (ed.), The Ontological Argument from St.

Anselm to Contemporary Philosophers (Garden City, NY: Doubleday, 1965). Note that what Kant called "the ontological argument" is something he found in the writings of Descartes. We have no reason to believe that Kant ever read Anselm.

3. All my quotations from the Proslogion come from Anselm of Canterbury: The Major Works (hereafter Major Works), edited with an introduction by Brian Davies and G. R. Evans (Oxford: Oxford University

Press, 1998).

4. Major Works, 87.

5. Major Works, 82.

6. Augustine, De Moribus Manichaeorum, 2, 11.

7. L. Annaei Senecae Naturalium Questionum libri viii (ed. Alfred Gercke, Stuttgart, 1907).

8. This point is stressed by G. E. M. Anscombe in "Why Anselm's Proof in the Proslogion Is Not an Ontological Argument," Thoreau Quarterly

17 (1985).

9. See De Casu Diab. 1 where Anselm (taking "being" to mean "being in re") says: "Just as from the highest good only good comes, so from the highest being only being comes, and all being comes from the highest being. Since the highest good is the highest being, it follows that every good is being and every being is good" (Major Works, 195–96). See also

Mon. 31 and Prosl. 3, where Anselm equates truly existing with existing greatly.

10. We know nothing of Gaunilo other than that he was a monk of Marmoutier (located close to the modern day French city of Strassbourg).

His claim to fame derives from the fact that Anselm directed that Gaunilo's reply to Anselm, and Anselm's reply to Gaunilo, should always be appended to published copies of the Proslogion. Readers of Anselm might usefully reflect on why he did this. A plausible theory is that he took his reply to Gaunilo to be a serious explanation of what he understood himself to be arguing in Proslogion 2 and 3.

Cambridge Companions Online © Cambridge University Press, 2006 Cambridge Companions Online © Cambridge University Press, 2006

brian davies 11. Major Works, 111–12 12. Major Works, 112–13; 116–17; 121.

13. See Norman Malcolm, "Anselm's Ontological Arguments," Philosophical Review 69 (1960). Here Malcolm claims to find two distinct arguments for God's existence as between Prosl. 2 and 3. Malcolm, however, does concede that there is no evidence that Anselm thought of himself as offering two different proofs in these chapters (p. 45).

14. Major Works, 161. For a similar critique of Anselm's line of reasoning, see Thomas Aquinas, Summa Contra Gentiles i, 11 and Summa Theologiae 1a, 2,1.

15. Major Works, 109.

16. Major Works, 105.

17. Major Works, 116.

18. Major Works, 111–13.

19. See Summa Contra Gentiles i, 10–11 and Summa Theologiae 1a, 2,1.

20. Malcolm, "Anselm's Ontological Arguments," 47.

21. Richard Swinburne attempts to do this with respect to the concept of God; see The Coherence of Theism (revised edition, Oxford: Oxford University Press, 1993).

22. Kant, Critique of Pure Reason a598/b626 f.

23. For a full-scale defence of it, see C. J. F. Williams, What Is Existence?

(Oxford: Oxford University Press, 1981).

24. Aquinas, Summa Theologiae 1a, 2,1.

25. For this distinction, see Gyula Klima, "Saint Anselm's Proof: A Problem of Reference, Intentional Identity and Mutual Understanding," in G. Holmstr¨om-Hintikka (ed.), Medieval Philosophy and Modern Times

(Dordrecht: Kluwer, 2000). I am greatly indebted to Professor Klima for personal discussions of the distinction and its bearing on the reasoning of the Proslogion.

26. Major Works, 111.

Cambridge Companions Online © Cambridge University Press, 2006 Cambridge Companions Online © Cambridge University Press, 2006

sandra visser and thomas williams

Anselm's account of freedom introduction1 According to Anselm's official definition, freedom of choice2 is "the power to preserve rectitude of will for the sake of that rectitude itself."3 From the point of view of contemporary metaphysics, this is one of the most unhelpful definitions imaginable. Does such freedom require alternative possibilities, for example? Is it compatible with causal determination? Is the exercise of such freedom a necessary and sufficient condition for moral responsibility? The definition sheds no light on these questions.

So we need to move on from Anselm's definition to Anselm's account of freedom. Here, though, we encounter the opposite problem. Where Anselm's definition seems not to answer these questions at all, Anselm's account seems to answer all these questions, sometimes with a yes and sometimes with a no. Consider the question about alternative possibilities. In De Libertate Arbitrii, Anselm seems clearly to deny that freedom involves alternative possibilities.

God, the good angels, and the blessed dead cannot do otherwise than preserve rectitude, but they are still free – freer, in fact, than those who are capable of abandoning rectitude.4 On the other hand, in De

Casu Diaboli Anselm seems to require alternative possibilities for freedom. For if an angel is to be just, Anselm says, he must have both the power to will rectitude and the power to will happiness. If only one power were given him, he would be able to will nothing but rectitude or nothing but happiness, as the case might be; being unable to will otherwise, his will would be neither just nor unjust.

Now justice, according to De Veritate 12, is rectitude of will preserved for its own sake. So an angel without alternative possibilities

Cambridge Companions Online © Cambridge University Press, 2006 Cambridge Companions Online © Cambridge University Press, 2006

sandra visser and thomas williams cannot have rectitude of will, and a fortiori cannot preserve rectitude of will; hence, an angel without alternative possibilities is not free.

In this paper we offer a reconstruction of Anselm's account of freedom in which this apparent inconsistency and others like it are resolved. As it turns out, the linchpin of this account is the definition of freedom. Anselm argues that the power to preserve rectitude for its own sake requires the power to initiate an action of which the agent is the ultimate cause, but it does not always require that alternative possibilities be available to the agent. So, while freedom is incompatible with external causal determination, an agent can, under certain circumstances, act freely even though he cannot act otherwise than he does. the definition of freedom and its roots in de veritate Freedom of choice is the power to preserve rectitude of will for its own sake. In order to understand what Anselm means to convey by this definition, we must first turn to his dialogue De Veritate, where the notion of rectitude is fleshed out in detail. Anselm's student asks for a definition of truth. Anselm replies that, so far as he remembers, he has never run across a definition of truth. Perhaps, he suggests, they can look for such a definition by examining the various things in which truth is said to exist.5 Thus they consider what truth is in statements, opinions, the will, actions, the senses, and finally the essences of things.

In each case, Anselm argues that truth is a matter of rectitude: that is, something's being or doing what it was meant to be or do.6 Thus the "rectitude of will" that figures in Anselm's definition of freedom is equivalent to truth in the will. The Devil, Anselm points out, is said to have abandoned the truth. He asks the student to explain what is meant by "truth" in that case. The student replies:

Nothing other than rectitude. For if, so long as he willed what he ought – that is, that for which he was given a will – he was in rectitude and in truth, and if when he willed what he ought not, he abandoned rectitude and truth, truth in that case cannot be anything other than rectitude, since both truth and rectitude in his will were precisely his willing what he ought.7

Cambridge Companions Online © Cambridge University Press, 2006 Cambridge Companions Online © Cambridge University Press, 2006

Anselm's account of freedom

Just as the truth or rectitude of a statement is the statement's doing what statements were made to do, the truth or rectitude of a will is the will's doing what wills were made to do.8

In Chapter 12 of De Veritate Anselm links rectitude of will with both justice and moral evaluation. Justice in its most general sense is equivalent to rectitude in its most general sense; whatever is as it ought to be has both rectitude and justice. The student objects,

"Shall we say that a stone is just when it seeks to go from higher to lower, since it is doing what it ought to, in the same way that we say human beings are just when they do what they ought to?"9

After some further discussion, Anselm notes, "I see you are looking for a definition of the justice that deserves praise, just as its opposite, injustice, deserves reproach."10 The justice that is the proper subject of moral evaluation is ultimately defined as "rectitude of will preserved for its own sake."11 Such rectitude requires that someone perceive the rectitude of his action and will it for the sake of its rectitude. Anselm takes the second requirement to exclude both compulsion and "being bribed by some extraneous reward."12

Since freedom of choice is by definition the power to preserve rectitude of will for its own sake, the arguments of De Veritate imply that freedom is also the capacity for justice and the capacity for moral praiseworthiness. So before turning to De Libertate Arbitrii it is useful to note how these equivalences must constrain Anselm's account of freedom, if he is to be consistent. It is both necessary and sufficient for justice, and thus for praiseworthiness, that an agent will what is right, knowing it to be right, because he knows it is right. That an agent wills what is right because he knows it is right requires that he is neither compelled nor bribed to perform the act. Freedom, then, must be neither more nor less than the power to perform acts of that sort. arriving at the definition of freedom in de libertate arbitrii Much that is initially puzzling about Anselm's account of freedom in De Libertate Arbitrii becomes clear when one reads it – as Anselm meant for us to read it – with De Veritate in mind.13 The first question is whether free choice is, or at any rate involves, the power to Cambridge Companions Online © Cambridge University Press, 2006

Cambridge Companions Online © Cambridge University Press, 2006

sandra visser and thomas williams sin. Anselm has two arguments to show that it does not. The first is as follows: God and the good angels have free choice; God and the good angels do not have the power to sin; therefore, free choice neither is nor entails the power to sin. But could not someone object

(the student asks) that the divine and angelic free choice differs from human free choice? Irrelevant, says Anselm: however much their free choice might differ from ours, the definition of free choice is the same and must apply equally to both.14

The second argument relies on the premise that a will is freer when it is incapable of sin than when it can be turned to sin. So if the power to sin is added to a will, its freedom is diminished; and if it is removed, the will's freedom is increased. Obviously, though, if something's absence increases freedom and its presence diminishes freedom, that thing cannot itself be identical with freedom, or even a part of freedom.

Both of these arguments are valid, but each relies on a controversial premise. In the first argument, Anselm assumes that God and the good angels have free choice, and the student raises no objection. But why should this assumption be so obvious? Since Anselm has yet to define free choice, we can only assume at this stage that free choice is something good, the lack of which would be a defect.

But when we come to Chapter 3 and the definition of free choice, it will turn out (as we have already seen) that no one can be just or praiseworthy without possessing free choice.15 It would be impious

(nefas) to deny that God and the good angels are just and praiseworthy, so it would also be impious to deny that they have free choice. So the controversial premise will turn out, in retrospect, to have been justified.

The disputable premise in the second argument is that a will is freer when it cannot sin. Here the student raises the obvious objection: "I don't see why a will isn't freer when it is capable of both [sinning and not sinning]." Anselm replies, "Do you not see that someone who has what is fitting and expedient in such a way that he cannot lose it is freer than someone who has it in such a way that he can lose it and be seduced (adduci) into what is unfitting and inexpedient?" The student, perhaps unlike the contemporary reader, replies, "I don't think anyone would doubt that."16

Anselm's interrogative argument for the questionable premise is philosophically revealing. Unlike most contemporary philosophers,

Cambridge Companions Online © Cambridge University Press, 2006 Cambridge Companions Online © Cambridge University Press, 2006

Anselm's account of freedom

he thinks of freedom as teleological. Freedom is a power for something, and that power is greater just insofar as it is less apt to fall short of its purpose. Specifically, beings have freedom for the purpose of having what is fitting and expedient; the more tenuous a being's grip on what is fitting and expedient, the less free that being is.

But if free choice is the power to hold on to what is fitting and expedient, and it is not the power to sin, does it make any sense to say that the first human beings and the rebel angels fell through free choice? The student formulates the problem acutely:

I cannot rebut your arguments at all, but it strikes me quite forcefully that in the beginning both the angelic nature and our own had the power to sin – if they had not had it, they would not have sinned. But if both human beings and angels sinned through this power, which is extraneous in this way to free choice, how can we say they sinned through free choice? And if they did not sin through free choice, it seems they sinned out of necessity. After all, they sinned either spontaneously17 or out of necessity. And if they sinned spontaneously, how was it not through free choice? So if it was not through free choice, they apparently sinned out of necessity.18

Anselm insists that human beings and angels did in fact fall through free choice:

It was through the power of sinning, and spontaneously, and through free choice, and not out of necessity that our nature, and that of the angels, first sinned . . . The fallen angel and the first human being sinned through free choice, since they sinned through their own choice, which was so free that it could not be compelled to sin by any other nature . . . They sinned through their choice, which was free; but they did not sin through that in virtue of which it was free, that is, through the power by which it was able not to sin and not to be a slave to sin. Instead, they sinned through that power they had for sinning.19

Though embedded in what looks like an unpromising bit of protoscholastic distinction-mongering, Anselm's point is both subtle and plausible. The argument clearly relies on taking arbitrium (choice) to be the power for self-initiated action. So when Anselm says that the arbitrium of angels and human beings before the fall was liberum

(free), he is saying that they had a power for self-initiated action that was not compelled by any external agency. To say that they sinned per liberum arbitrium (through free choice), as Anselm does twice,

Cambridge Companions Online © Cambridge University Press, 2006 Cambridge Companions Online © Cambridge University Press, 2006

sandra visser and thomas williams is simply to say that they sinned by an exercise of that power. But when he denies that they sinned "through that in virtue of which [their choice] was free," he is emphasizing the teleological nature of freedom; full-fledged freedom of choice is the power for self-initiated action for some good end, and the angels did not sin through that.

Finally, the potestas peccandi (power for sinning) through which the angels did fall is simply liberum arbitrium unsupplemented by freedom from sin.

Thus Anselm can consistently maintain that the primal sins were committed per liberum arbitrium and yet deny that the power to sin is a part of liberum arbitrium. If liberum arbitrium is simply the power for self-initiated action not compelled by any external agency, then liberum arbitrium neither entails nor includes a power to sin.

For liberum arbitrium can be perfected by something else, as yet unspecified, that renders it incapable of sinning. So the power for self-initiated action as such does not entail or include the power to sin, even though that same power, if unsupplemented by freedom from sin, is itself the power to sin. anselm's definition and its immediate implications Anselm's arguments in the first two chapters of De Libertate Arbitrii pull in two different directions. As we saw in the last section,

Chapter 1 hints at a normative definition: free choice is the power to hold on to what is fitting and expedient. Chapter 2, however, suggests a purely descriptive definition: free choice is a power for selfinitiated action not compelled by any external agency. In Chapter 3

Anselm opts unmistakably for a normative definition: "free choice is the power to preserve rectitude of will for its own sake." Anything that satisfies the normative definition will also satisfy the descriptive definition, since (as Anselm made clear in De Veritate) the power to preserve rectitude of will for its own sake requires that an agent be able to initiate his own action on the basis of what he believes to be right, and act for the sake of that rightness, without being either compelled or bribed.20 In keeping with this line of thought, Anselm goes so far as to say that freedom of choice consists in having the rational ability to know what is right in conjunction with the will by which one can choose it.21

Cambridge Companions Online © Cambridge University Press, 2006 Cambridge Companions Online © Cambridge University Press, 2006

Anselm's account of freedom

But the entailment does not work the other way around: a power could satisfy the descriptive definition without satisfying the normative definition. Suppose there were a capacity for self-initiated action that is at least sometimes free from external compulsion but was not bestowed upon its possessor for any particular purpose or designed with any particular end in mind. (The free will described by many contemporary libertarians is just such a capacity.) That capacity satisfies the descriptive definition but not the normative definition, and Anselm would not call that capacity liberum arbitrium. He would, in fact, find the very idea of such a capacity bizarre. For to suppose that such a capacity exists is to suppose that God created a power for which He had no particular purpose in mind – hardly the act of a rational creator. Accordingly, Anselm shows no interest in what we might call "garden-variety" freedom: freedom with respect to whether one has pasta or pizza for dinner, say. It is, indeed, difficult to imagine Anselm's God granting us a power for self-initiated menu choices, at least under that description. If we in fact have gardenvariety freedom, it will be only as a by-product of the morally significant freedom that interests Anselm. Accordingly, in De Concordia

1.6 Anselm explicitly acknowledges that his discussion in De Veritate and De Libertate Arbitrii concerns only the freedom necessary for salvation.22

Even so, we should not overestimate the importance of Anselm's opting for the normative definition. Although Anselm proceeds, in the remainder of the work, to derive a number of important conclusions using the normative definition, most of the arguments would work equally well if he used the descriptive definition. For example, he argues in Chapter 5 that no temptation forces anyone to sin unwillingly:

Student: But how is the choice of the human will free in virtue of this power [i.e. free choice], given that quite often a person whose will is right abandons that rectitude unwillingly because he is compelled by temptation?

Teacher: No one abandons rectitude otherwise than by willing to do so.

Therefore, if by "unwilling" you mean someone who does not will, no one abandons rectitude unwillingly. For a man can be tied up unwillingly, since he does not will to be tied up; he can be tortured unwillingly, since he does not will to be tortured; but he cannot will unwillingly, since he cannot will if he does not will to will. For everyone who wills, wills his own act of willing.23

Cambridge Companions Online © Cambridge University Press, 2006 Cambridge Companions Online © Cambridge University Press, 2006

sandra visser and thomas williams This argument assumes only that actions performed through free choice are uncompelled and self-initiated; Anselm need not appeal to the purpose for which human beings were given free choice. Even the argument that nothing is freer than an upright will (Chapter 9) depends explicitly only on the descriptive definition, although the influence of the normative definition is evident in Anselm's specifying the upright will. freedom and alternative possibilities in de casu diaboli The account of freedom that is in place by the end of De Libertate

Arbitrii seems to entail the falsity of what contemporary philosophers call the Principle of Alternative Possibilities (PAP), which states (roughly) that an agent performs an action freely only if it was causally possible for that agent to act otherwise than he did. Suppose, for example, that God commands the angel Gabriel to announce the Incarnation to Mary. Since this command is given after the good angels have been confirmed in goodness, it is not possible for Gabriel to do otherwise than obey God.24 And yet Gabriel announces the Incarnation freely, because in doing so he is preserving rectitude for its own sake: he knows that it is right for him to obey God, he wills that obedience for the sake of its rightness, and he initiates his own act of obedience. It follows that PAP is false; Gabriel acts freely even though he cannot do otherwise.

When we turn to De Casu Diaboli, however, hints of some version of PAP are everywhere. We hope to show that these new arguments extend the account of freedom in De Veritate and De Libertate

Arbitrii but are fundamentally consistent with it. For although free action does not always involve alternative possibilities, it often does; and the reasons why it does arise straightforwardly out of the account of freedom we have already sketched.

The first argument suggesting PAP comes in Chapter 5:

T: Do you think the good angels were likewise able to sin before the evil angels fell?

S: I think so, but I would like to understand this through reason.

T: You know for certain that if they were not able to sin, they preserved justice out of necessity and not in virtue of their power. Therefore, they

Cambridge Companions Online © Cambridge University Press, 2006 Cambridge Companions Online © Cambridge University Press, 2006

Anselm's account of freedom

did not merit grace from God for remaining steadfast when others fell any more than they did for preserving their rationality, which they could not lose. Nor, if you consider the matter rightly, could they properly be called just.25

It is tempting to see in this argument a straightforward endorsement of PAP. After all, Anselm seems to argue that if the angels who refrained from sinning had not been able to do otherwise – had not been able to sin – they would not have been free. They could no more have abandoned rectitude than they could have abandoned rationality, and it would be as incongruous to praise them for remaining upright as to praise them for remaining rational.

Despite the initial appearances, however, there is no appeal to PAP in this argument. Rather, the appeal is to the requirement that a free action have its origin in the agent rather than in some external cause. Consider the situation Anselm is envisioning. The good and bad angels were in exactly the same position before the fall: they were equal in nature, in knowledge, and in power. We know that the bad angels fell by exercising their power for self-initiated action: "It was through the power of sinning, and spontaneously, and through free choice, and not out of necessity that . . . the angelic nature first sinned."26 If it was not likewise possible for the good angels to fall, that could only have been because some external agency was preventing it; for there was, ex hypothesi, nothing internal to their own power of agency to account for that impossibility. And in that case, the good angels did not preserve justice through their own power, but out of necessity.

So Anselm is not assuming PAP. Alternative possibilities come into the picture as a kind of by-product. They are not constitutive of freedom; they just happen to be available, given the requirement that free action have its origin within the agent, in conjunction with the relevant circumstances of the particular case. No doubt alternative possibilities will often be available to agents exercising free choice, but nothing in Anselm's account requires that they always be.

A second passage that seems to involve reliance on PAP is the extended argument of Chapters 12 through 14, an argument we summarized at the beginning of this paper. Anselm argues that an angel must have both a will for justice and a will for happiness if he is to be morally responsible. If he had only one of these wills, he would

Cambridge Companions Online © Cambridge University Press, 2006 Cambridge Companions Online © Cambridge University Press, 2006

sandra visser and thomas williams be able to will nothing but rectitude or nothing but happiness, as the case might be. He would therefore will rectitude or happiness necessarily. Necessity, as we have already seen, is incompatible with freedom. Therefore, an angel without the alternative possibilities provided by the two wills would not be free.

What we shall now show, however, is that this quick summary misrepresents Anselm's argument. Once again, it is not PAP but the requirement of self-initiated action that generates the two-will theory – as we can show by offering a more careful recapitulation of Anselm's argument in Chapters 12 through 14. In Chapter 12 Anselm argues that in order for an angel to will anything at all, God must give the angel its initial will. For if some agent moves himself to will, he first wills to move himself. Hence, whatever does not yet will anything at all cannot move itself to will. "So it must be the case,"

Anselm concludes, "that the angel who has already been made apt to have a will but nonetheless does not will anything cannot have his first will from himself."27 His first will must therefore come from God.

Anselm's use of "will" (voluntas) in this argument and those that are to follow can cause confusion. He explains in De Concordia28 that voluntas can mean three different things. Voluntas can mean the "tool" that the soul uses in order to will (i.e. the faculty or power of will), the disposition of that tool to respond to certain features of what is proposed to it for willing (i.e. desire or motivation or dispositional volition), and the act in which that tool is employed (i.e. occurrent volition). Let us call these respectively "faculty," "disposition," and "volition." When God gives the angel its initial will, is He giving the faculty, the disposition, or the volition? Anselm does of course hold that God gives the faculty of willing, just as He gives every other creaturely power, but the faculty is clearly not what is at issue in the argument just stated. That argument seems to require that we interpret voluntas as volition: if there is no volition at all, there cannot be the volition by which the soul wills to employ its faculty of will in a particular way. But the argument would then be obviously mistaken. Surely if the angel has a disposition to will in a certain way, then so long as he has the faculty of willing (and there are no impediments to the use of that faculty), there is no reason why the angel cannot generate his first volition for himself. So what

Anselm must mean is that unless God gives the initial disposition, Cambridge Companions Online © Cambridge University Press, 2006

Cambridge Companions Online © Cambridge University Press, 2006

Anselm's account of freedom

the angel cannot have any volition. He thinks this because he understands volition as goal-directed: "we do not will anything at all unless there is a reason why we will it."29 The faculty of will does not engage in pseudo-Sartrean reasonless choice. So if the faculty of will is to be operative at all, God must give the angel at least one motivational disposition in response to which it can engage in actual volition.

In Chapter 13 Anselm asks us to suppose that God first gives this angel the will (i.e., the disposition) for happiness, and no other will. Can he move himself to will (i.e. have a volition for) something besides happiness? The teacher and student agree that he cannot:

I can't see how someone who wills nothing besides happiness would move himself to will anything other than happiness. After all, if he wills to move himself to will something else, he wills something else.

Therefore, just as he could not will anything at all on his own when no will had yet been given, so also he cannot have any other willing from himself if he has received only the will for happiness.30

For similar reasons, the angel will also be unable to refrain from willing happiness. So unless God gives him some other will, he will will happiness; and the higher his estimation of happiness, the more intensely he will will it. If he cannot have the best things, he will will lesser things – even "the base and impure things that please irrational animals."31 But no matter what he wills, his will is "the work and gift of God, just as his life and his power of sensation are, and there is neither justice nor injustice in it."32

In Chapter 14 we are assured that the same conclusions follow if the angel is given only the will for rectitude: he will not be able to help willing rectitude, and his will will be neither just nor unjust.

Now he cannot be happy unless he wills to be happy, and he cannot deserve happiness unless he also wills to be just. So if he is to be deservedly happy, he must have both the will for happiness and the will for justice.

At each stage of this argument, Anselm appeals not to anything like PAP, but to the requirement that the agent be able to initiate his own action. The angel can have no volition at all until God gives him a disposition to will in a certain way. If God gives him only the will for happiness, every volition of happiness will have its ultimate origin in God and not in the angel himself; his will is "the work and gift of Cambridge Companions Online © Cambridge University Press, 2006

Cambridge Companions Online © Cambridge University Press, 2006

sandra visser and thomas williams God."33 He will not have the power to originate any willing that he did not receive from God; by the descriptive definition of free choice, then, the angel is not free. Similarly, if God gives him only the will for rectitude, every willing of rectitude will have its ultimate origin in God; the angel will again lack the power to initiate any willing that is genuinely his own, so he will lack free choice. Only if God gives him both wills does he have that power.34 For then he has the power to will happiness as tempered by justice, and to will happiness without regard for justice. Neither of those volitions is received from God; both have their ultimate origin in the angel himself.

Even though, as we have argued, the arguments of De Casu Diaboli 5 and 12–14 are not driven by PAP, they do show an important connection between freedom as Anselm understands it and the possession of alternative possibilities. Freedom requires that an agent be able to initiate an action that is genuinely his own. Now creatures receive their wills – that is, both their faculty of will and their characteristic dispositions – from God. So if God makes a creature's will in such a way that alternative possibilities are never open to him, every volition of that creature will be "the work and gift of God."

He will not be able to initiate any action that is genuinely his own, so he will not be free. It is, therefore, not freedom as such, but creaturely freedom, that requires alternative possibilities.35 And even then, alternative possibilities are required only once, as the case of the good angels makes clear. The good angels had alternative possibilities with respect to their primal choice. Afterwards God made them unable to sin; in this way He closed off any alternative possibilities, but He did not destroy their freedom.

Thus, Anselm's arguments up to this point in De Casu Diaboli merely elaborate on the account of freedom that has been in place since the early chapters of De Libertate Arbitrii. But his discussion of the primeval angelic freedom takes an unexpected turn in Chapter 23 of De Casu Diaboli when he argues that the angels would not have been free if they had known for sure that they would be punished if they fell. He seems to say that their fear of punishment would have been so great that they would inevitably have willed to retain rectitude, not for the sake of rectitude, but for the sake of avoiding punishment. The just action – willing rectitude for its own sake – would not have been open to them, and by the normative definition of free choice, they would not have been free.

Cambridge Companions Online © Cambridge University Press, 2006 Cambridge Companions Online © Cambridge University Press, 2006

Anselm's account of freedom

The problem is that the good angels do have this knowledge now, thanks to the example of their fallen brethren. Anselm seems to have a dilemma on his hands. If the good angels are now just, they are preserving rectitude for its own sake, and not merely for the sake of evading punishment. But then there is no reason to think they could not have preserved rectitude for its own sake even if they had known then what they know now. On the other hand, if they are indeed merely trying to evade punishment, they are not just, not praiseworthy, and indeed not free, because their new knowledge is such as to preclude their preserving rectitude for its own sake.

The dilemma owes some of its force to a misreading. It seems natural to read Chapter 25 as arguing that the only reason why the good angels can no longer sin is that they are aware of the consequences of sin. But in fact the argument is more subtle. Anselm is interested in maintaining that even if their knowledge of the consequences of sin is the sole reason why the good angels can no longer sin, their not sinning is still to their credit.36 At the end of the chapter Anselm clearly denies that their inability to sin derives from this knowledge.

The teacher remarks, "But in fact you know – because it became evident earlier – that the reason why [the good angel] cannot sin is that by the merit of his perseverance he has attained such happiness that he no longer sees what more he could will."37

The back-reference is to Chapter 6. Chapters 4 and 5 had shown that the fallen angels sinned by willing some additional good that God had not yet given them, and that the good angels could have willed "that something extra" (illud plus) but chose instead to retain the just will that God had given them.38 Anselm then argued in Chapter 6 that as a reward for their perseverance in justice, God gave the good angels whatever it was they had passed up in the interests of justice. Thanks to this divine gift, there is nothing for them to will that they do not already enjoy.39 Now, at the end of Chapter 25,

Anselm makes sure we do not think he has abandoned this explanation of the sinlessness of the angels.

So Anselm does not after all argue that their knowledge of the consequences of sin renders the good angels unable to sin. But some version of the dilemma we posed above still threatens his account.

If the good angels after the fall can have this knowledge and yet retain free choice and choose rectitude for its own sake, why would this knowledge subvert the free choice of angels before the fall? In Cambridge Companions Online © Cambridge University Press, 2006

Cambridge Companions Online © Cambridge University Press, 2006

sandra visser and thomas williams particular, Anselm thinks that if an angel had this knowledge before the fall, it would necessitate his action.40 And yet after the fall, it does not necessitate his action. Why would the very same knowledge undermine freedom before the fall but be consistent with freedom afterwards?

Remember that the angels have only two wills: the will for justice and the will for happiness. Now imagine two angels, Gabriel and Michael, who are preserving the will for justice. Both know all the consequences of sin, but Gabriel knows this before the fall, whereas Michael first learns this after the fall, by noting the fate of the rebel angels. It follows straightforwardly from the account of freedom given thus far that neither angel can abandon his will for rectitude, but that Gabriel is necessitated, whereas Michael is free.

Consider Gabriel first. He can only will happiness and rectitude – and that is God's doing, not his. He cannot sin by willing happiness, because he knows full well he will not get happiness by sinning.

He cannot sin by willing rectitude, obviously. So his not sinning is entirely God's doing, not his own. He is necessitated by God's creative acts not to sin.

Michael's case is different. He too cannot sin by willing happiness, but that is because he already has all the happiness he can imagine as a reward for his decision to preserve rectitude. His retaining the will to rectitude, though, is his own doing, not God's. As we saw earlier, God gave him the will for rectitude and the will for happiness, but the decision to subject his will for happiness to the demands of rectitude was the angel's own doing. Moreover, he retained rectitude for its own sake. So, as long as he sustains that will, he is acting on his own, not out of any necessity. True, he has no temptation to abandon rectitude, but he retains rectitude under his own steam, so to speak, and not because of God's action.

Of course, according to our contemporary way of using modal terms, it seems obviously false to say that Michael is not acting out of necessity. Surely if it is not possible in those circumstances for Michael to sin, it is necessary in those circumstances that he not sin; Michael, it seems, is as much necessitated as Gabriel. Obviously Anselm has something different in mind when he speaks of "necessity," and since he has been regularly opposing necessity to freedom since the first chapter of De Libertate Arbitrii, we need to be clear about what exactly Anselm takes this freedom-threatening

Cambridge Companions Online © Cambridge University Press, 2006 Cambridge Companions Online © Cambridge University Press, 2006

Anselm's account of freedom

necessity to be.41 As one would expect from De Veritate, compulsion certainly imposes such necessity. More generally, what a contemporary philosopher would call causal necessity is also incompatible with freedom. Thus, when Anselm sets out in De Concordia to reconcile free choice with divine foreknowledge, he explains that the kind of necessity that attaches to what God foreknows is not the freedom-threatening kind that "brings it about that a thing exists"

(facit rem esse) or that "compels" (cogit) something to come about.42

Even what what God predestines does not happen "by that necessity which precedes a thing and brings it about" (ea necessitate quae praecedit rem et facit),43 for such causal necessity would destroy freedom.

But the examples of Gabriel and Michael show that Anselm's most fundamental notion is this: an action is necessary just in case its ultimate explanation is external to the agent. Causal determination makes actions necessary because it prevents an agent from initiating any action that is genuinely her own; Gabriel's knowledge of the consequences of sin makes his action necessary for exactly the same reason. No self-initiated action can ever properly be described as necessary, even if it is not possible for the agent to act otherwise in the relevant circumstances.

Anselm's most striking affirmation of this understanding of necessity in action comes in a discussion of God's action in Cur Deus

Homo. After Anselm has argued that in some sense God had to provide a remedy for sin, Boso objects: "If this is so, it seems that God is, as it were, compelled to secure human salvation by the necessity of avoiding impropriety (indecentia) . . . And how will we ascribe our salvation to God's grace if He saves us by necessity?" Anselm replies:

God does nothing by necessity, since He is in no way compelled to do or prevented from doing anything; and when we say that God does something as if from the necessity of avoiding dishonorableness – which He certainly does not fear – it is rather to be understood that He does this out of the necessity of preserving His honorableness. And this necessity is nothing other than the immutability of His honorableness, which He has from Himself and not from another and which is therefore improperly called necessity.44

Because God's immutable uprightness is "from Himself and not from another," every upright divine action will be self-initiated; and for Cambridge Companions Online © Cambridge University Press, 2006

Cambridge Companions Online © Cambridge University Press, 2006

sandra visser and thomas williams that very reason Anselm insists that no such action should be called

"necessary." Divine aseity in fact guarantees that every action God performs is self-initiated. So all of God's actions are free, even if He never has alternative possibilities available to Him. reconciling the two definitions It is instructive to see how Anselm's two definitions can be combined into a single general definition without doing violence to Anselm's theory. Recall that Anselm offers a normative and a purely descriptive definition of free choice. According to the descriptive definition, free choice is a power for self-initiated action. According to the normative definition, free choice is the power to preserve rectitude of the will for its own sake; the normative definition entails that the agent (1) is able to initiate his own action on the basis of what he believes to be right, and (2) is able to act for the sake of that rightness.

How can a normative and non-normative definition be reconciled?

The answer lies in Anselm's motivation for discussing free choice in the first place. Anselm's primary interest in free choice is how it bears on human responsibility for sin and the need for grace. Any other exercise of free choice is ancillary. Thus, his normative definition (his preferred one) explicitly builds in features central to his moral and theological concerns. Anselm believes that some goals are better than others. Specifically, he believes that while justice and happiness are our two most important goals, justice is incomparably more important than happiness. So, if God gave us free will for a purpose, and that purpose is to achieve the best goal through our own free action, then we are acting most freely when we seek to achieve that goal. Further, Anselm seems to think that following this goal is the most rational thing to do as well.

As we shall now show, however, one can abandon Anselm's own story about our ultimate goal without doing much damage to his account of free will. That is, one can accept a teleological account of free choice and reject the notion that the best goal is justice or that one acts most freely when one acts for the sake of justice. If there is no objective hierarchy of goals, an agent will not be more or less free depending on which goals he has chosen, but he will be more or less free depending on how well he satisfies Anselm's descriptive definition of free will.

Cambridge Companions Online © Cambridge University Press, 2006 Cambridge Companions Online © Cambridge University Press, 2006

Anselm's account of freedom

For if we look at exactly how his descriptive definition (the one Anselm uses when he is not concerned with ultimate goals) would function in actual examples, we will see where we build back in those teleological concerns – properly modified – without realizing it. The normative aspects of Anselm's second definition concern justice. As Anselm says in De Veritate 12, "Every will not only wills something but also wills for the sake of something. Just as we must examine what it wills, so also we need to understand why it wills."45 (Notice that this is presented as a general claim, without any reference to rectitude.) And again, "Every will has a what and a why. For we do not will anything at all unless there is a reason why we will it."46 So we do no violence to Anselm's descriptive definition if we reformulate it as follows: free choice is the power to attain one's goals for the sake of those goals. This definition, which we shall call the enriched descriptive definition,47 requires that an agent (1) be able to initiate his own action on the basis of what he believes will achieve his goal, and (2) be able to act for the sake of that goal.

The relationship between the enriched descriptive definition and Anselm's preferred normative definition becomes clear in one of Anselm's own illustrative examples in De Concordia:

Let us now offer an example involving an upright (that is, a just) will, freedom of choice, and choice itself; and let us consider how the upright will is tempted to abandon rectitude and how it maintains that rectitude by its free choice. Suppose someone is resolved to hold fast to the truth because he understands that it is right to love truth. This person surely already has an upright will and rectitude of will. Another person approaches and threatens to kill the first person unless he tells a lie. We see that it is his decision (in eius arbitrio) whether to abandon life in favor of rectitude of will, or rectitude of will in favor of life. This decision . . . is free, because the reason by which he understands rectitude teaches that this rectitude ought always to be preserved out of love for rectitude itself, and that whatever is offered to him as a pretext for abandoning rectitude is to be held in contempt, and that it is up to the will to reject or choose as the understanding of reason dictates . . . Hence, a decision of the will to abandon this same rectitude is also free and not forced by any necessity, even though it is assailed by the dreadfulness of death.

For although it is necessary that he give up either life or rectitude, nevertheless no necessity determines which he preserves or abandons. Surely in Cambridge Companions Online © Cambridge University Press, 2006

Cambridge Companions Online © Cambridge University Press, 2006

sandra visser and thomas williams this case the will alone determines which of the two he retains; nor does the force of necessity cause anything, where only the will's choice is operative.

And if there is no necessity for someone to abandon the rectitude of will that he has, it is clear that the power to preserve it – i.e. freedom – is not absent . . . In virtue of this freedom both the choice (arbitrium) and the will of a rational nature are said to be free.48

Anselm's arguments in connection with this example obliterate any distinction between the descriptive and the normative definitions of freedom. Anselm begins by appealing to the key elements of the enriched descriptive definition: the person in the example is free because he knows what goal he ought to aim at and has the power to choose accordingly, and no external force is operating so as to necessitate his choice. But since the goal that he ought to aim at is precisely the preservation of rectitude for its own sake, he satisfies the normative definition. What it is for him to satisfy the enriched descriptive definition is precisely the same as what it is for him to satisfy the normative definition; the two definitions, in other words, are equivalent. the usefulness of anselm's account of freedom

So in the end, the enriched descriptive definition of free choice turns out to be equivalent to the normative definition that Anselm prefers, given the assumption that reason shows us that rectitude of will is the paramount goal to be respected in all action. If we decline to join

Anselm in that assumption, the two definitions will not be equivalent; but for that very reason, the reformulated descriptive definition becomes a useful and interesting option for contemporary debates about freedom. We can accept it without committing ourselves to any substantive moral claims, and we disentangle the discussion of freedom from the specifically theological concerns that motivated

Anselm.49

The greatest advantage of the enriched descriptive definition of free choice is that it satisfies both incompatibilist and compatibilist intuitions about free will. Certainly Anselm takes very seriously the incompatibilist intuition that a free action cannot be causally determined. The reasons that a person has for performing a free action do not determine that he take that action. But while it is true that many

Cambridge Companions Online © Cambridge University Press, 2006 Cambridge Companions Online © Cambridge University Press, 2006

Anselm's account of freedom

free choices are entirely unpredictable, not all of them are. So while Anselm's account satisfies the intuitions of incompatibilists, there are other conditions in which it also satisfies some of the intuitions of compatibilists. What is central to Anselm's definition is that the action be self-initiated and consciously chosen, not that it be one of at least two possibilities. This aspect of Anselm's theory partially satisfies the intuition that as long as a person knows what he is doing and why he is doing it, his action is free, regardless of whether the agent had some other option available to him. Of course, while the compatibilist does not care whether the action is self-initiated, but only that the agent is doing what he wants, the Anselmian insists that the action be self-initiated. But the Anselmian can explain why there is a pull to say that a person who has chosen a particular course of action and is happy with it has sometimes chosen freely, despite a lack of alternatives.

Moreover, in light of arguments purporting to show that which goals and desires one finds oneself with are largely (or even fully) beyond one's control, the enriched descriptive definition helps to explain how it is fair (or just) to hold a person responsible for acting on whatever goals he finds himself with. According to Anselm, how one comes by one's goals is irrelevant. He in fact presupposes that the motivations of rational creatures derive entirely from outside themselves, although he of course thinks the external source is

God rather than heredity, upbringing, or the like. What is relevant to freedom is not the source of the motivations, but whether, when there is a decision to be made among competing goals, it is the agent himself who is doing the deciding. If the agent initiates the choice and is not determined by circumstances outside his control, then his choice is free and it is permissible to hold him responsible for his action. In the unfortunate, and indeed unlikely, instance in which a person has absolutely no good motives from which to choose, he is still responsible for the action that results from the motive he chose to follow.

What one might see as the greatest strength of Anselm's account – its ability to capture both incompatibilist and compatibilist intuitions – might also be its greatest weakness. We can imagine that a compatibilist would find it incredible that while one's decision to act on a desire might determine one's action, nothing determines which desire one opts to follow. It is true that the Anselmian can

Cambridge Companions Online © Cambridge University Press, 2006 Cambridge Companions Online © Cambridge University Press, 2006

sandra visser and thomas williams give an explanation of her free choice; the explanation will always be in terms of which desire she placed above the others, and she might have reasons for preferring that desire to another. But ultimately, when asked whether that preference determined her action, the Anselmian will say no. In fact, given the same situation, she might conceivably do something else – if there were more than one motive at work in her decision. And that, a compatibilist might well say, is hardly an appealing picture of the relation between free choice and reasons for action.

In reply, an Anselmian should note that the key point behind some brands of compatibilism (especially the freedom-entailsdeterminism varieties) is that unless my character determines or at least explains my actions, they are not really actions at all, but merely spasms. But the only plausible motivation for that view is the belief that free actions are those that the agent herself originates, those for which the agent is somehow responsible. And Anselm's theory secures that belief. The compatibilist simply refuses to face the problem that worries Anselm in Chapters 12 to 14 of De Casu Diaboli. If both the good and the bad angels are to have been free and responsible for their primal choice, it cannot be the case that anything about their desires, powers, or knowledge determined their choice either to preserve or to abandon rectitude. For their desires, powers, and knowledge were all owed to God. Therefore, if their desires, powers, and knowledge had determined their choice, that choice too would have been owed to God. God, not the angels, would have been responsible for it; the bad angels would not have been blameworthy, nor the good angels praiseworthy. Indeed, there would have been no distinction between good and bad, because they all had the same desires, powers, and knowledge, and would therefore have made the same

"choice." The angels would not have been agents at all, but inert conduits for divine agency.

In contemporary terms, Anselm's arguments amount to this claim: there is no responsible agency unless there is an element of radical voluntarism somewhere. If a certain set of cognitive and affective states, all of which have their origin outside the agent, guarantees a certain choice, the agent is not really an agent at all, but an inert conduit for external causes. An exercise of agency, therefore, is possible only where what the agent has "received" from outside does not guarantee one choice over another.

Cambridge Companions Online © Cambridge University Press, 2006 Cambridge Companions Online © Cambridge University Press, 2006

↑ Back to top