The Real Distinction between Essence and Existence

John Robert Toby Lamont


The idea of a real distinction between essence and existence was famously argued for by Avicenna. There are various ways in which this distinction has been characterised, but we will understand it as holding that in contingent beings, existing is different from any essential property, and indeed from any accidental property; existing, in fact, is different from the attribute of having any sort of property at all, where a property is considered to be a feature of a thing that gives it some kind of determinate character. This distinction makes a fundamental difference to accounts of the divine nature, the nature of created being, and the character of God’s act of creation. I do not know enough about Islamic philosophy to comment on the history of views on this distinction in Muslim thought, but it became the subject of extensive debate among Catholic scholastics. It was adopted by St. Thomas Aquinas, argued for in his De Ente et Essentia and other works, and made into a central feature of his metaphysics. It was however rejected by other scholastics, and found its most eminent opponent in the Jesuit philosopher and theologian Francis Suarez. The debate over the existence of the real distinction grumbled on in Catholic circles for centuries. As late as the 1950s the important manual of philosophy and theology produced by the Spanish Jesuits was arguing for the Suarezian view,[1] while the learned Canadian Thomist Joseph Owens was defending his version of the real distinction in the 1980s.[2]

It is rather presumptuous to attempt a new contribution that has the purpose of overcoming such a lengthy intellectual impasse. I do not aspire to defend the precise views of Avicenna or St. Thomas in this question, or even to give an accurate account of their thought on the nature of the distinction between essence and existence. I only seek to offer arguments in favour of their being some kind of real distinction between the two, that will if successful hopefully cast some light as well on the insights of these thinkers. I think this task may be made more feasible by the possibility that the conceptual tools of analytic philosophy may help in looking at the problem from a new angle. There are debates about the historical origins of analytic philosophy, which is currently the dominant school of philosophical thought in the English-speaking world. In my view the main element in its genesis was the development of set theory and modern predicate logic towards the end of the nineteenth century. These developments are among the great achievements of the human mind, and are both directly and indirectly relevant to central philosophical issues. They have been used by analytic philosophers in giving accounts of existence, but these accounts have scarcely ever considered the question of the existence of a real distinction between essence and existence, a question that emerges from a scholastic philosophical tradition that is foreign to analytic philosophers. I nevertheless believe that such accounts can be useful in thinking about the question.

A starting point for considering the question of whether there is a real distinction between essence and existence is to ask what would be involved in there being no such distinction. On this view, existence would simply be one among the many properties held by things. It would have the highest degree of generality, since it would be possessed by everything, and it would be one of the least informative of properties; in these respects it would resemble a property like being self-identical. We would not have to rule out the position, held by many philosophers, that there are different kinds or degrees of existence, in order to postulate this property. We would only have to specify that the kind of existence we are talking about is that single simple feature that makes the difference between things being there and not being there. On this hypothesis, existence is not the same as other properties, such as for example the property of being a man. But then existing and (e.g.) being a man are different properties. But this is nonsense; since to be a man is to be an existing thing. So the idea of existence being a property of things is wrong, since it leads to nonsense, and we should accept that there is a real distinction between essence and existence.

One way of getting out of this conclusion is Meinong’s. He held that there are two sorts of things, existing ones and non-existing ones; so on his view, existence would be the property that is possessed by existing things but not by non-existent ones. He was excoriated by Russell for this,[3] in terms with which I am happy to associate myself, and I will just assume that he was wrong.

A more subtle way is the view that may have been Aristotle’s (Aristotle scholars disagree about the view that Aristotle actually held, and I am not competent to judge which of them is right). On this view, a property like being a man is a certain specification of the more general property of existing; it is a particular way of existing. This would explain why existing is different from the property of being a man, without postulating the real distinction. But here we can use the point made by Avicenna, that the content of the property of being a man does not tell us whether or not there are any men, and it is possible that there might not have been any men. Thus existence is not included in the content of the property of being a man, as it ought to be if being a man is a specification of the more general property of existence. It is of course true that something cannot be a man without existing. But this results from the fact that ‘something’ here must be understood as ‘some existing thing’ (rather than ‘some fictional thing’ or ‘some imaginary thing’), if it is to be true that something cannot be a man without existing. It is thus the ‘something’ that implies existence here, not the content of the property of being a man.

It seems that Aristotle himself would not have accepted this line of reasoning. He states: ‘He who knows what human - or any other – nature is, must know also that man exists; for no-one knows the nature of what does not exist – one can know the meaning of the phrase or name ‘goat-stag’, but not what the essential nature of a goat-stag is.’[4] He held in consequence that it was impossible for essential kinds to begin or cease to exist; he thought that they always were and always will be. But as Avicenna held, this is wrong. The nature of (e.g.) humans does not tell us whether or not humans exist, and thus it is possible for an essential kind to cease, begin, or never exist at all. Experience confirms this; once there were no men, and now there are no dodos.

In Avicenna’s day and for long afterwards it was taken for granted that existence is something that belongs to particular things. On this assumption, the reasoning given above for a real distinction between essence and existence looks persuasive. However, this assumption is now no longer generally made, and so this reasoning is likely to be seen instead as providing support for the prevalent view of existence among analytic philosophers, which is that existence is not a property of individuals. The view was first clearly stated by Frege. He held that the word ‘is’ is ambiguous, and has four different possible meanings:

i)                    the ‘is’ of identity, as in ‘Hesperus is Phosphorus’

ii)                   the ‘is’ of predication, as in ‘Socrates is a man’

iii)                 the ‘is’ of existence, as in ‘Whales exist’

iv)                 the ‘is’ of class inclusion, as in ‘all oranges are fruits’.

This ‘is’ of existence he interpreted as expressible by the existential quantifier in the predicate logic he devised, and as assigning a property not to individual objects, but to properties. To say that whales exist, in his view, is to say something about the concept ‘whale’, viz. that it has instances. Frege’s view of the ‘is’ of existence was adopted and strongly argued for by Russell. Ever since, it has tended to be taken for granted by analytic philosophers, who have assumed that it has laid to rest all the traditional philosophical disputes about the nature of existence.

Russell sets forth his version of the position in this way:

It is perfectly clear that when you say ‘Unicorns exist’, you are not saying anything that would apply to any unicorns there might happen to be, because as a matter of fact there are not any, and therefore if what you say had any application to the actual individuals, it could not possibly be significant unless it were true. You can consider the proposition ‘Unicorns exist’ and can see that it is false. It is not nonsense. Of course, if the proposition went through the general conception of the unicorn to the individual, it could not be even significant unless there were unicorns. Therefore when you say ‘Unicorns exist’, you are not saying anything about individual things, and the same applies when you say ‘Men exist’. If you say that ‘Men exist, and Socrates is a man, therefore Socrates exists’, that is exactly the same sort of fallacy as it would be if you said ‘men are numerous, Socrates is a man, therefore Socrates is numerous’, because existence is a predicate of a propositional function, or derivatively of a class…you sometimes know the truth of an existence-proposition without knowing any instance of it. You know that there are people in Timbuctoo, but I doubt if any of you could give me an instance of one. Therefore you clearly can know existence-propositions without knowing any individual that makes them true.[5]

Russell prefers to speak of existence as a property of propositional functions rather than of properties or concepts, but this distinction is not important for our purposes, and we will consider the thesis that existence is a property of properties. The above passage is better as an exposition of his view than as an argument for it. We can accept that ‘Unicorns exist’ is not making a statement about particular unicorns, and that it is telling us something about the concept or property of being a unicorn, without being committed to saying that existence is not a property of particular things. The statement ‘birds have wings’ need not be understood as saying anything about particular birds (as it can be understood by someone who does not know of the existence of any particular bird), and it can also be understood as telling us something about the property of being a bird – viz. that it includes having wings – but that does not imply that having wings is not a property of particular things.

Although ‘Unicorns exist’ is not used that effectively by Russell in the above passage, it does belong to the principal category of evidence that is used to support the Frege/Russell view. This category is that of statements that assert the existence of nonexistent things, or that deny existence to things. The basic structure of the argument from such statements is that if you hold existence to be a property of individuals, these statements get you into insoluble problems, such as asserting that there are things that do not exist, or assigning the property of non-existence to things. I will not try to overthrow these arguments, something that in my opinion has been done in a satisfactory way by Colin McGinn and Barry Miller.[6] I will only make a general remark on the underlying problem with them, which is that they try to use semantic considerations, i.e. considerations that have to do with the way language relates to and describes the world, to derive substantive metaphysical conclusions. The trouble with this method of argument is that there usually turn out to be acceptable ways of changing the semantics in order to avoid the metaphysical conclusions, which is what McGinn and Miller manage to do.

There is a further argument against existence’s being a property of individuals which is also rejected by McGinn, but which I think has more to be said for it than he allows. It is alluded to by Russell, in the remark that ‘There is no sort of point in a predicate which could not conceivably be false. I mean, it is perfectly clear that, if there were such a thing as this existence of individuals that we talk of, it would be absolutely impossible for it not to apply, and that is the characteristic of a mistake.’[7] I think we need to change talk of predicates to talk of properties here, and to distinguish between a property’s necessarily being true of everything, and its not saying anything about what it applies to. McGinn is right in pointing out that there are perfectly good properties, like self-identity, that necessarily apply to everything. However, the difficulty with existence lies not in its necessarily applying to everything, but in its not ascribing any feature to things. The source of this difficulty lies in the fact that part of what it is to be a property is to make a difference to the things that have it, to determine them in some way. This is connected to what the identity conditions for properties must be. An account of the identity conditions for properties can only be given in terms of how properties determine the things that have them, with different properties determining things in different ways, and the same properties determining things in the same way. Identity conditions for existence cannot be given in this way, however, since existence, if it belongs to individuals, is prior to all determination; a thing can be determined in one way or another only if it already exists. The case is not parallel with a necessarily universal property like self-identity. This property, as McGinn points out,[8] tells you a lot about the thing that has it, because it means that it is not identical with every other thing. So this feature of existence makes possible the following argument for the Frege/Russell view of existence. Existence cannot be a property of individuals, because if we understand it as belonging to individuals, it fails to satisfy the individuation conditions for properties. It is a perfectly good second-level property of properties, however, because it tells us something perfectly clear and determinate about properties, viz. that they have instances.

Of course this argument can be used as a reason for the real distinction between essence and existence, as well as for the Frege/Russell view of existence. However, the Frege/Russell can be presented as a superior account on the grounds of its being quite clear, unlike the bizarre and almost incomprehensible thesis that existence is a sort of attribute of things without being a property of things. What is needed to make the case for the real distinction is further argument in favour of it and against the Frege/Russell view. I think the following line of argument succeeds in doing this.

Begin with the assumption that every possible entity is either universal or particular. Then, suppose:

1) Only particulars exist.

In this case, existence will not be a universal, because there will be no universals. But it will not be a particular either. It cannot be, since it is precisely the common feature we attribute to particulars in the course of stating 1. So, according to number 1, existence does not exist.

One might reject this reasoning by appealing to a nominalist account of existence, according to which we only need to posit particular existences of particular things rather than any general characteristic of existence that is the same for different things. However, such an account is clearly more problematic than a nominalist account of e.g. the property of being red. In the very course of stating a nominalist view of redness, by claiming something like ‘there is no such thing as a property of redness in itself, but only particular red things’, we seem to be ascribing a general attribute, that of existence, to the particulars we are talking about. We can adapt an argument of Russell’s to make this point.[9] We cannot express the statement ‘Only particulars exist’ by listing all the particulars P1….Pn and saying of each of them that it exists, because such a statement does not rule out the possibility of more things existing besides P1….Pn, which things might not themselves be particulars. We have to add to this the claim ‘and nothing else exists’. But the existence referred to in this further claim cannot be any of the existences attributed to P1….Pn, which have all already been mentioned and said to exist; it has to be some other attribute, which cannot attach to any of the particulars that are supposed to be the only things that exist. Without postulating this other attribute we cannot say that only particulars exist; but the postulation of this attribute means that the statement that only particulars exist is false. Another way of arguing that the statement ‘only particulars exist’ requires us to understand the existence it ascribes to particulars as a general rather than a particular attribute, is to pick up on remarks by Russell quoted above, and point out that the truth of this statement is independent of which actual particulars are held to exist. Any particulars could exist, as far as it goes. But this means that the existence that is being talked about in the assertion ‘only particulars exist’ is not an existence that belongs to any given particular. But if there is an attribute of existence that does not belong to any given particular, a nominalist account of existence must be false. (Arguments of these sorts can also be given to oppose the claim that existence is an abstract rather than a concrete particular, a trope.)

This sort of reasoning might be thought to push us towards the Frege/Russell view of existence. But this view does not get us out of the woods. For if existence is a property of properties, that means that properties themselves must exist. So, we can consider the hypothesis:

2) Only particulars and first-level universals exist.

For the purposes of this argument we do not need to specify whether universals are supposed to exist prior to their instances or only in their instances. Nor do we need to consider what differences there may be between universals, properties or Russell’s propositional functions, as the argument works out in the same way for all of them; we will give the arguments as it runs for universals. First-level universals will be universals that have particulars rather than other universals as their instances (second-level universals will have first-level universals as their instances, and so on). If 2 is correct, existence will not be a particular, for the reasons given in discussing 1. But neither will it be a first-level universal, since 2 describes first-level universals as falling under it. Since according to 2 all that exists are particulars and first-level universals, and existence can be neither of these, it follows from 2 that existence does not exist, just as it does from 1. To again try and avoid this conclusion, rather than 1 or 2 we can suppose:

3) Particulars, first-level universals and universals higher than first-level ones exist.

There are different ways of specifying 3. One – it can be called 3a – is as the claim that particulars and a finite number of higher-level universals exist. In this case, we just take the top level n, whatever that is, and apply the same reasoning to it as we did to first-level universals with 2. 3a requires that existence be a universal on level n + 1, since it attributes the general quality of existence to universals on level n as well as to everything below. But it also states that universals on level n + 1 do not exist. Thus it implies that no universals on level n can exist, due to the fact that existence, according to the hypothesis, does not exist on level n + 1. Hence, existence on level n cannot exist; and so on down.

The second way to specify 3 – it can be called 3b – is as the claim that particulars and an endless series of higher-level universals exist. But this requires us to postulate existence as a general property of particulars and all the endless higher-level universals. We then have existence as a further general, existing property on a level of universals corresponding to the first transfinite ordinal. Even if we accept this as a possibility, there is then nowhere further to go, because there is no number that is one up from the first transfinite ordinal; but we are required to postulate a property on a level one up from the first transfinite ordinal, if we are to say that some property on the level of the first transfinite ordinal exists.

One might try to avoid this conclusion by generalising the Frege/Russell account of existence, and saying that instead of existence being a second-order property, it is the property of a universal on any level having something that instantiates it. But this position makes an illegitimate appeal to the notion of instantiation (a kind of point that has been made by Colin McGinn[10]). We can distinguish between falling under a universal and instantiating a universal. Round squares can be truly said to fall under the higher-level category of types of impossible objects. But making this true statement cannot just be the same as saying that round squares exist; if it were, ontology would be a lot simpler than it is. What must make the difference between falling under a universal and instantiating it is that when a universal is instantiated, there is something that exists that has it. But this just brings back in the notion of existence that was supposed to be explained away by the definition. And it brings it back as a single property that is alleged to occur on every level of particulars and universals. As such, it then becomes open to the objections given above to 3a and 3b.

There is thus no way of avoiding the conclusion that according to 3, existence does not exist. But 1, 2, and 3 exhaust all the possibilities, if we accept the initial assumption that the only possible entities are either universals or particulars. (It is perhaps worth noting that the argument does not require one to suppose that second-level universals apply only to first-level ones, and so on. Clearly, higher-level universals will apply to all universals and particulars that fall underneath them, but this does not vitiate the argument.) So, existence does not exist. And, since there is no such universal as existence that anything other than existence could be an instance of, nothing exists; and, in fact, nothing could exist.

Now obviously I could not propose this argument if its conclusion were true. Things do in fact exist. But since the argument works if we assume that existence is a universal, it follows that existence cannot be a universal, or a property. It thus establishes the conclusion that there is a real distinction between essence and existence.



[1]. See L. Salcedo and J. Iturrioz, Philosophiae Scholasticae Summa, I: Introductio in philosophiam, logica, critica, metaphysica generalis, Madrid: Biblioteca de Autores Cristianos, 1953, esp. 567 ff.

[2]. See e.g. Joseph Owens, Aquinas on Being and Thing, Niagara Falls, NY: Niagara University Press, 1981.

[3]. See Bertrand Russell, Essays in Analysis, London: George Allen & Unwin Ltd., 1973, and Introduction to Mathematical Philosophy, London: George Allen & Unwin Ltd., 1919.

[4]. Aristotle, Posterior Analytics, bk. II, ch. 7, 42b4-10, in The Basic Works of Aristotle, ed. Richard McKeon, New York: Random House, 1941, p. 166.

[5]. Bertrand Russell, The Philosophy of Logical Atomism, ‘V: General Propositions and Existence’, in The Collected Papers of Bertrand Russell, vol. 8, The Philosophy of Logical Atomism and Other Essays, 1914-19, ed. John G. Slater, London: George Allen & Unwin, 1986, pp. 204-5.

[6]. See Colin McGinn, Logical Properties, Oxford: Clarendon Press, 2000, and Barry Miller, A Most Unlikely God, Notre Dame: University of Notre Dame Press, 1996; The Fullness of Being, Notre Dame: University of Notre Dame Press, 2002. Miller is often arguing against C. J. F. Williams, whose What is Existence?, (Oxford: O.U.P., 1981) is the most comprehensive defence of the Frege/Russell view, relying heavily on argument from negative existential statements.

[7]. Russell, 1986, p. 211.

[8]. McGinn, 2000, p. 12.

[9]. See Russell, 1986, p. 206.

[10]. See McGinn, 2000, ch. 2.


© Copyright 2006 SIPRIn. All Rights Reserved.

 Print This Document

Save This Document on Your System