P.F.Strawson  INDIVIDUALS 　

Chapter 8  Logical Subjects and Existence                       2008/01/15 澤邊興平

[1]

Paragraph1 [extending beyond the paradigm cases]

 Recall the ways in which non-relationally tied items may collect each other:

 Predicates of particulars collect particulars in a way which contrasts with the ways in which particulars collect their predicates. When one non-particular principle of collection collects another rather in the way in which a predicate of particulars collects a particular, then we may say that it operates, in respect of the other, as a principle of collection of like things.(226)

 Whenever you have something which can be identifyingly introduced into a proposition, and can be brought under some principle of collection of like things, then you have the possibility of that thing's appearing as an individual, as a logical subject. . . . nothing can be identifyingly introduced into a proposition without satisfying that it can be brought under some general principle of collection of like things. (226-27)

Paragraphs 2, 3 & 4 [use of singular definite substantival expressions as the sign of logical subject/individual]

This feature is, however, not infallible. Two types of problematic cases:

 1) "The man-in-the-moon does not exist."

 2) "The man-in-the-moon lives on cheese."

Here, 'the man-in-the-moon' is an expression that seems to refer to a particular. The first type of sentences either explicitly assert or deny the existence of what seems to be referred to. The second kind conjoins an ordinary sort of predicate to that seemingly referring expression, which in fact does not refer to anything.

 The subjectival expression in the first kind of cases cannot be construed to be referential, because to do so would be to treat it as carrying a certain presupposition which is denied by the predicate part.

 Some strategies to make sense of the type 1 cases:

 a) there is no definite, specific thing referred to in the expression (except possibly the moon), it is merely about the existence/inexistence of a man-in-the-moon, i.e. indefinite existential claim, or its denial.

 b) it refers to a concept, and claims/denies that it is instantiated.

 c) a Russellian alternative: it refers to a propositional function and it claims that it is 'sometimes true' or 'never true'.

 In contrast, the second type of cases is less problematic. It can be analyzed very much similarly to the sentences where the subject part is an expression that does in fact refer to some existent particular:

The apparent referring expression really is such. Its role is to introduce a particular, and its failure to do so is a failure of fact, the factual failure of the presupposition it carries. . . . because of the falsity ... we in some cases deny a truth-value to the proposition as a whole. . . . Or ... can see it simply as operating in a different realm of discourse, ... of myth, fiction or fancy than that of fact. ... within limits ... we can presuppose existence and allocate truth-values as we choose. (228)

Paragraph 5 [the principle of collection of like things as a characteristic feature of non-particulars does not cohere well with the 'conceptual' or 'propositional function'-construal of the type1 cases.]

 Of the three strategies above-named, the first raises no problem, but for the second and the third where concepts and propositional functions are cast for the role of subject-terms, our intuition seems to falter somewhat:

Are we to say that having instances or being 'sometimes true' are principles of collection of like concepts, or of like propositional functions? (228-29)

 Two responses:

 First. Having a specific (finite) number of instances is a valid priniciple of collection of like concepts. If this is plausible, then extend it to include the case where the number is zero, i.e. not being instantiated at all too is a valid principle of collection of like concepts.

 Second. Instead:

we might simply plead that there are principles of collection of like concepts; that there are subject-predicate sentences bringing concepts under such principles (229)

 Those responses, especially the latter are not entirely convincing, far from it perhaps:

while the expression 'is instantiated', as applied to concepts, does not have just that kind of completeness which would, in the paradigm cases, disqualify it from ranking as a predicate-expression, neither does it have just that kind of incompleteness which we found characteristic of predicate-expressions in the paradigm cases. ... We should very shortly find ourselves in a familiar region of paradox if we tried to claim for it this kind of incompleteness. (229)

[2]

Paragrpaph1 [nominalistic reduction of non-particulars]

 More empirically or nominalistically minded philosophers are reluctant to embrace non-particulars as individuals, as logical subjects. They feel that such use of non-particular-referring expressions can be paraphrased by using appropriate quantifications strictly over particualrs:

Characteristically, this reductionist programme aims at replacing sentences involving references to non-particulars by sentences involving quantification over particulars. (231)

 This programme turns out to be successful as regards certain types of non-particulars; not very successful, or outright unconvincing in some other cases. Those non-particulars that resist this reduction will be called more 'entrenched' than those which do not.

 Examples of less well-entrenched, i.e. where reduction seems natural and satisfyingly explanatory: qualities, relations, states, activities, species, etc. An example of reduction:

'Anger impaires judgement' ⇒'People are generally less capable of arriving at sound judgements when they are angry than when they are not'

 Examples of better entrenched non-particulars: sentence-types, word-types, numbers, propositions, and other 'types'.

Paragrpah 2 [why the varying degree of entrenched-ness?]

 Two ways in which a non-particular may be well-entrenched:

1) because the reductionist programme is hard to carry out vis-a-vis the non-particular,

2) because there is little desire to carry out the programme for this non-particular.

Call the first type, logical entrenchment, the other, psychological entrenchment.

An example:

the interesting case of noun-clauses headed by the conjunction, 'that'. Sometimes we regard, and perhaps speak of, what  they introduce, as facts; at other times we do not commit ourselves in this way. Philosophers, seeking a general word for items that may be so introduced, a word that does not commit us in the way in which the word 'fact' does commit us, have used the expression 'proposition'. Of course, facts and propositions alike may be introduced, not only by being specified in a 'that'-clause, but in other ways as well. There is no reason to suppose that facts are better entrenched, logically, than propositions. But there is every reason to suppose that facts are better entrenched psychologically than propositions. (232)

Paragpraph 3 [types as model-particulars]

 An appropriate model for non-particulars of these kinds is that of a model-particular―a kind of prototype, or ideal example, itself particular, which serves as a rule or standard for the production of others. (233)

 Examples given are all artefacts, man-made things. This should not, however, mislead us to think that the issue here is strictly functional, it also involves fairly exact specification:

to produce an instance, one must conform more or less closely to more or less exact specifications. Fully to describe a non-particular of this kind is to specify a particular, with a high degree of precision and internal elaboration. (233)

Paragraph 4 [Another way of being more particular-like]

 Other well-entrenched non-particulars such as numbers and propositions do not exhibit the kind of feature described above, the feature that non-particulars such as 'types' possess. But there are other ways in which numbers and propositions are analogous to particulars:

Particulars have their place in the spatio-temporal system, or, if they have no place of their own there, are identified by reference to other particulars which do have such a place. But non-particulars, too, may be related and ordered among themselves; they may form systems; and the structure of such a system may acquire a kind of autonomy, so that further members are essentially identified by their position in the system. (233)

Words and sentences (as types) are well-entrenched in both of those senses:

the type-word can be thought of, on the one hand, as an exemplar for its own physical tokens (particulars), and on the other, as a unit of meaning, a rule-governed member of a language-system. (233-34) 

[3]

Paragraph 1 [Platonistic zeal and nominalistic zeal are both out of place]

 Particulars, esp. basic particulars hold the pre-eminent place among logical subjects. Still, to conclude from that that non-particulars must be banished from the scene entirely is to go too far, and not being sufficiently reflective over the sense of particulars' preeminence among logical subjects.

Paragraph 2 [the idea of existence involved in the above dialectic between nominalist and Platonist]

 Whenever Fc can be asserted, it entails/(is inferred from it) that ∃xFx .

So:

Subject-expressions can, and predicate-expressions cannot, be replaced by variables of existential quantification. And since '∃xFx ' is to be read 'There exists something which F ', it follows that a thing which can be referred to by a logical subject-expression is the sort of thing which we can say exists. (235)

Paragraph 3 [two questions]

 1)Why should it always be subject-expressions which give place to the apparatus of  

   quantification, and never predicate-expressions?

 2)Granted that (1) can be satisfactorily answered, why should we construe the 

   resulting quantified sentence in the recommended way?

 To elaborate further they may be filled out as follows:1)

We are invited to think of an existential statement as one which is entailed by any member of a range of propositions with varying subjects and a constant predicate and which itself contains that same predicate. But we can easily form the idea of a range of propositions with varying predicates and a constant subject. Can we not equally well form the idea of a proposition entailed by any member of this range and itself containing that same subject? What is the reason for the onesidedness of the logician's picture? (235)

and:2)

We read '∃xFx ' as 'There exists something which F '. What compels this reading? Why should we not, making such amendments as may be grammatically necessary, read it as 'F* exists', where 'F* ' is a singular substantival expression designating the property predicated in the original proposition? (235)

Paragraph 4 [approaching the first question from the Russellian treatment of existentially quantified statements]

 The first question is tantamount to ask, on the basis of 'Socrates is wise', why do we have

(1)  '( . . . is wise) is sometimes true'

but not

(2)  '(Socrates . . . ) is sometimes true'  ?

 As we saw in chapter 6, the expression 'is wise', in introducing its term, carries no empirical presupposition; it identifies its term for us whether or not we know or think that anyone is wise. In other words, all we need to know is the language and not some concrete empirical data. Therefore, (1) succeeds in giving a genuine empirical information as conveyed by a statement 'There is someone who is wise'.

For the expression of the form (2), the situation is different:

A condition of the referring expression, viz. 'Socrates', performing its role is that a presupposed empirical fact, or facts, should be known to its user or hearer. Consequently there is no way of construing (2) which allows to 'Socrates' the role of a referring expression. We might construe (2) as stating, in effect, that Socrates exists, that the presuppositions of a certain referring use of 'Socrates' are satisfied; but we cannot in that case also take 'Socrates' as having that referring use in (2). If on the other hand, we try to construe 'Socrates' as already having this use in (2), then there is no statement which we can construe (2) as making; all that it tries to say is already presupposed by the referring use of 'Socartes'. (236)

Paragraph 5 [analysis of the proposed answer to the first question continued]

 If , following Russell, we are to explain existence-statements in terms of the idea of the truth of a subject-predicate statemennt; and if we are to take, as the model of a subject-predicate statement, that in which a universal is predicated of a particular; then it is clear that it must be referring expressions and not predicate-expressions which give place to the apparatus of existential claim. This apparatus cannot be intelligibly joined to referring expressions which can occur in the model type of subject-predicate statement. (237)

That an already identified item, of whatever type, has some (unspecified) property or other, i.e. falls under some (unspecified) principle or other of collection of like things, is never news; that something or other unspecified has an already identified property, i.e. falls under an already identified principle of collection of things, is always news. The former can never be regarded as part of what is asserted by a proposition in which an identifeid thing and an identified principle of collection of suchlike things are assertively tied; but the latter is always part of what is asserted by such a proposition. (237)

Paragraphs 6&7 [trying to meet the second question/objection]

 Answering the first question in the manner described in the last two paragraphs, makes it harder to answer the second question:

why should we think the force of (1) above, i.e. of '(. . . is wise) is sometimes true' is better rendered by 'There exists someone who is wise' than by 'Wisdom exists' ? Or, why, when I say that Socrates is wise, am I to be regarded as committed to the view that there is such a thing as a wise man, but not to the view that there is such a thing as wisdom? (237)

But perhaps the folowing line of argument might help:

So long as we confine our attention to what may reasonably be said to follow from such a proposition, in the way of existence-claims, there seems no reason for preferring the empirical claim that there exists some instance of the predicated term to the empirical claim that the predicated term exists. . . . But if we turn from the question of what is entailed by the statement as a whole to the question of what is presupposed by the use of its term-introducing parts, the situation is altered. The subject-expression, introducing a particular, carries a presupposition of definite empirical fact; the predicate-expression, introducing a universal, does not. Here is an asymmetry regarding presupposed existence-claims which may be the ground of the preference for one mode of statement of the entailed existence-claim over the other. (238)