From the ALWS archives: A selection of papers from the International Wittgenstein Symposia in Kirchberg am Wechsel

Introduction

The concept of essence is traditionally associated with questions of the form 'What is a?', where 'a' is a singular term. The traditional view is that there is a distinctively metaphysical construal of questions of that form, and that describing the essence or some essential features of a given object a is just giving a correct answer, be it complete or only partial, to the metaphysical question as to what a is. I will call such answersobjectual essentialist statements.

The concept of essence is also traditionally associated with questions of the form 'What is it to F?', where 'F' is a predicate expression like 'be a man' or 'be wise' or 'think'. The traditional view is that there is a distinctively metaphysical construal of these questions, and that a complete or partial correct answer to the metaphysical question as to what it is to F constitutes a description of the essence or of some essential features of F-ing. I will call such answersgeneric essentialist statements.

The concepts of objectual essence and of generic essence both occupy a central place in philosophy. It is then quite legitimate to wonder how they are to be understood. In the contemporary literature about essence, however, focus has been made on the objectual notion only, and the generic notion has been neglected. The source of such a negligence may be the thought that the generic notion can be quite easily accounted for, in terms of the objectual notion or in terms of other, more familiar concepts, e.g. in modal terms.

As I will show, this is not the case. The greatest part of the paper will be devoted to the rejection of a number of more or less natural proposals. I will end up with the suggestion that the concept of generic essence is primitive and that it can be used to define several other notions of central philosophical importance, in particular that of objectual essence, that of metaphysical necessity and that of analyticity.

Grammatical Points

Before all, some points about the formulation of essentialist claims are in order.

I will take it that all objectual statements can be put in the form:

    (O) It is true in virtue of what a is that p,

and that all generic statements can be put in the form:

    (G) It is true in virtue of what it is to F that p.

I will in particular assume that:

    (o) a essentially Fs

is equivalent to 'it is true in virtue of what a is that a Fs', and that:

    (g) An F essentially Gs,

understood as a form of generic statements, is equivalent to 'it is true in virtue of what it is to be an F that all Fs G'.

I shall take (O) to be the canonical form of objectual statements and (G) to be the canonical form of generic statements. But I will nevertheless feel free to use other modes of formulations for ease of expression.

Let me note here that beside the ''individual" essentialist statements we met so far stand "collective" statements. An objectual statement is collective if it is a statement to the effect that some fact holds by virtue of what several objects taken together are, like e.g. 'in virtue of what number 3 and number 4 are, it is true that 3<4'. Generic statements can also be collective, to wit 'in virtue of what it is to be a man and of what it is to be a number, it is true that nothing can be both a man and a number'. For the sake of simplicity, focus will be made on individual statements, and collective statements will be left aside until the last section. Nevertheless it will be obvious how the discussion could be adapted so as to include them.

Three Accounts Rejected

In this section I present and reject three accounts of generic statements, giving in each case the version of the account for statements of type (G) and the version for statements of type (g). Focusing on (G) should be sufficient, given that any statement exhibiting the predicational form (g) can be rephrased in sentential form. But since some may find some of the three accounts to be introduced much more plausible as applied to statements in predicational form than they are as applied to statements in sentential form, I shall explicitly mention and criticize the accounts of predicational statements.

I may appear quite natural to understand (g), construed in the generic way, as:

    (g₁) Necessarily, every F Gs.

(Here and below, 'necessarily' expresses metaphysical necessity.) Accordingly, one may think that the form (G) should be understood as:

    (G₁) Necessarily, everything which Fs is such that p.

But both proposals are incorrect. I take it that (g) entails (g₁) and that (G) entails 'necessarily, p' and so (G₁) as well. But I think that the converse entailments do not hold.

(Fine 1994) has forcefully argued against the standard modal accounts of objectual essence, namely the view that (o) should be understood as 'necessarily, a Fs' or as 'necessarily, a Fs if it exists', and the view that (O) should be understood as 'necessarily, if a exists, then p'. Similar reasons can be invoked against the proposed modal accounts of generic statements.

Necessarily, every cat is either loved by Socrates or not. But we do not want to say on that account that cats are essentially either loved by Socrates or not, nor that it is true in virtue of what it is to be a cat that cats are either loved by Socrates or not. Or again, it is necessary that every object is a member of some set (its singleton set, for instance) – or so we may suppose. So necessarily, every cat belongs to some set. But pace the friends of the modal account, it does not follow that cats essentially belong to sets or that it is true in virtue of what it is to be a cat that cats belong to sets. Finally, take any necessary proposition whatsoever. Then by the modal account, the proposition is true in virtue of what is it to whatever you like – be a man, be wise, love Jesus and so on. And of course, this is absurd.

Some may wish to analyze (g), understood in the generic way, as:

    (g₂) Necessarily, every F essentially Gs.

And accordingly, one may think that the form (G) should be understood as:

    (G₂) Necessarily, everything which Fs is essentially
such that p.

The new accounts arguably escapes all the difficulties encountered by the modal ones. For instance, one may perfectly agree that as a matter of necessity, every cat belongs to some sets, and still deny that my cat Nessie is by its very nature a member of some set.

Yet they must also be rejected: (g) does not entail (g₂), and (G) does not entail (G₂) either. For instance, it is true in virtue of what it is to be a bachelor that bachelors are unmarried; but many actual men are bachelors and fail to be essentially unmarried, and presumably also fail to be essentially such that all bachelors are unmarried.

Another objection, to all accounts presented so far indeed, is that if there can be no F and nothing which Gs, then they predict that an F is essentially however you like and that it is true in virtue of what it is to G that whatever you like. But intuitively, a round square is essentially round but not essentially green, and it is true in virtue of what it is to be a round square that round squares are square, but not that round squares are tasty.

In the light of the previous considerations, one might wish to invoke properties (or kinds) and claim that (g) should be understood as:

    (g₃) It is true in virtue of what the property of being an F is that every F Gs,

and (G) as:

    (G₃) It is true in virtue of what the property of F-ing is
that p.

The idea behind this move is quite natural. When we say that, say, men as such are essentially human, we do not really talk about individual men, be they actual or merely possible. Rather, we talk about the having of a certain feature, that of being a man, and we state what having that feature essentially involves. The proposed account of generic essence takes this fact seriously and goes one step further: it reifies features and construes generic essentialist statements as objectual essentialist statements about reified features (properties).

It is quite obvious that the new proposal beautifully escapes all the previous difficulties. But it still faces some difficulties. One problem, or so I think, is that generic statements do not as such commit one to properties, nor to any object whatsoever indeed. The statements 'a man is essentially human' and 'it is true in virtue of what it is to be a man that men are animals' are not ontologically committing, in particular they not commit one to properties – no more indeed than 'men are human' and 'men are animals'. An enemy of properties may very well take these statements to be meaningful, or even true. But even under the assumption that there is such a thing as the property of being a man, my view is that one should agree that the statements under consideration do not commit one to the property. For it is surely not absurd to claim that these statements are true and at the same time deny that the property exists.

Another, more radical problem is that some predicate expressions cannot possibly express properties, while there are corresponding true generic statements. Consider for instance the predicate 'is a non-self-exemplifying property'. There cannot be such a thing as the property of being a non-self-exemplifying property. For if the property in question existed, it would be the case that it exemplifies itself iff it does not. Now there are arguably plenty of propositions which are true in virtue of what it is to be a non-self-exemplifying property: the proposition that every non-self-exemplifying property is non-self-exemplifying, the proposition that every non-self-exemplifying property is a property, the proposition that every non-self-exemplifying property is an abstract object.

The previous counterexamples can be used against the view that (g) entails (g₃) and the view that (G) entails (G₃). It is plausible to say, though, that both entailments hold in case the corresponding property is available. But even when there is the property, the converse entailments do not hold. It may be held that it is true in virtue of what the property of being a quantity of water is that it is an abstract object (a property, a multiply located entity, ...). But it may be denied, at the same time, that it is true in virtue of what it is to be a quantity of water that being a quantity of water is an abstract object. Some essential features of the property of being a quantity of water may fail to pertain to what it is for something to be a quantity of water.

Generic Essence, Objectual Essence, Metaphysical Necessity and Analyticity

In reaction to the problems met by the modal accounts of objectual essence, Fine (Fine 1994) suggests that the notion should be taken as primitive, and makes two important claims. The first is that instead of "viewing essence as a special case of metaphysical necessity'', he says, "we should view metaphysical necessity as a special case of essence''. The idea is that for a proposition to be metaphysically necessary is for it to be true in virtue of what a given object is, or in virtue of what some given objects are. (Remember the distinction between individual and collective essentialist statements introduced in the footnote in the section on the forms of essentialist statements.) The second claim is that truth in virtue of the meaning of an expression is nothing but truth in virtue of the nature of the meaning of (or concept expressed by) that expression – and similarly for truth in virtue of the meanings of several expressions taken together. On the view that analyticity is truth in virtue of the meanings of some expressions, Fine then proposes a general picture where both metaphysical necessity and analyticity are to be understood in terms of objectual essence.

Given the difficulties we met in the search for an "analysis'' of generic essence, it is tempting to go the same way and take the notion to be primitive. An option one may then find attractive is to adopt the general Finean approach to essence and its relationships to metaphysical necessity and analyticity, by suitably modifying it so as to take generic essence into account. Let me here sketch one version of that option.

Let me use 'Ε_a, b, ...' for 'it is true in virtue of what a is, what b is, ... that' (collectively understood), and 'Γ_F, G, ...' for 'it is true in virtue of what it is to F, what it is to G, ... that' (collectively understood). The proposal is then to reduce objectual essence, metaphysical necessity and analyticity to generic essence according to the following equivalences:

• 1. Ε_a, b, ... p iff Γ_{be a, be b, ...} p – where 'to be a' is short for 'to be identical to a', and similarly for 'to be b' and the rest;
• 2. It is necessary that p iff ∃F ∃G ... Γ_F, G, ... p;
• 3. It is true in virtue of the meanings of α, β, ... that p iff Γ_{Mα, Mβ, ...} p – where 'Mα' is short for 'mean what α means', and similarly for 'Mβ' and the rest;
• 4. It is an analytic truth that p iff ∃α ∃β ... it is true in virtue of the meanings of α, β, ... that p.

The picture does not include Fine's own view about truth in virtue of meaning, because I think that the view is incorrect (it is presumably true in virtue of what the concept MAN is that it is a concept, an abstract object, and capable of being grasped by minds like ours; but one may well deny – and Fine would, I guess – that it is true in virtue of what the expression 'man' means that the concept Man is a concept or an abstract object).

The proposed view is, in my opinion, rather attractive. Of course, a perspicuous presentation and defense of that view is needed, but it would require much more space than is available to me here. This is something I hope to do elsewhere.