Sir Peter Strawson’s ironically titled paper of 1950 ‘On Referring’ argues that expressions do not refer in themselves but rather people refer when using expressions.(Lycan, 2008: 20) Influenced by Wittgenstein’s ‘Use Theories,’ Strawson did not consider a sentence to hold a truth value in its abstraction but rather it is in virtue of the way in which Human Beings utter and react to the sentence that give it the attribute of truthfulness and as such he considered his theory to be more adaptable than Russell’s ‘Theory of Definite Descriptions’ in regard to ordinary language use. (Austin, 1978: 531) In this essay I will critically analyse Peter Strawson’s argument in relation to Bertrand Russell’s theory with the intention of demonstrating that Strawson’s position on ordinary language is unsubstantiated and that the consequences of his language model has deeper philosophical implications which marks his theory, not Russell’s, as untenable. I will initially outline Russell’s argument in the light of Strawson’s critique of the ‘Theory of Descriptions’ to determine the flaws it uncovers in Russell’s theory. I will then attempt to offer Strawson’s alternative theory through his objections to the ‘Theory of Descriptions,’ to prove that his attempt to justify his theory as less problematic to the use of ordinary language is misguided. I will then finally attempt to engage in Keith Donnellan’s solution to consider whether it proposes a more successful and balanced approach to Russell’s and Strawson’s theoretical problems of meaning and reference.
The most efficient model for demonstrating the dilemma concerning Russell and Strawson is through Russell’s famous example;
‘The present king of France is bald’ (Russell, 1905: 485)
This sentence appears to be of a subject – predicate form with the grammatical subject ‘present king of France’ forming the ‘denoting phrase’ that Russell profoundly analysed in his paper of 1905 ‘On Denoting.’ Russell proposed that there are three types of denoting phrases, however for the context of this essay I will deal exclusively with those ‘denoting phrases’ that denote a single definite person or object also known as definite description akin to ‘present king of France’ above. According to Russell declarative sentences of the structure of (1)do not possess the object represented by the grammatical subject in the sentence but rather conceal the more complex existential expression in virtue of their logical structure; there is one and only one x that is y, with x representing the grammatical subject and y representing the predicate. It is when the expression is altered into logical form that Russell makes the distinction between the logical subject and the grammatical subject of the sentence which he proposes as the solution to the puzzles of non-referring definite descriptions. (Morris, 2007: 51)
So if we convert (1) into formal logic;
2) ‘∃x(F(x) & ∀y(F(y) → x= y) & G(x))’ (Ludlow 2007)
Russell proposes that in instances at least of ‘denoting phrases’ being the grammatical subject of the sentence, the misunderstanding that leads to the puzzles occurs in virtue of the denoting phrase being an incomplete symbol on logical analysis. (Perkins, 1972: 201)This is deceptive in that it is not obvious in ordinary language and as such it is in fact x that is the logical subject of the sentence. Consequently Russell contends that ‘denoting phrases never have meaning in themselves, but that every proposition whose verbal expression they occur has meaning.'(Russell, 1905: 480) In other words whilst the sentence (1) may appear to be of a subject -predicate form, it is in fact not in virtue of the logical proposition concealing the quantifier of the expression and subsequently Russell argues that the ‘denoting phrase’ contributes meaning above its referent in that ‘it is not actually a singular term but a quantificational structure.’ (Lycan, 2008: 14) With this in mind, if we analyse the complex existential claim asserted by the definite article ‘the’ in (1)
a)(∃x)Fx
b)(€¢x)( Fx → (Gx) (Fy → y= x))
c)(€¢x)( Fx → Gx) (Lycan, 2008: 14)
It is necessarily the case that each condition needs to be satisfied in order for the proposition to be true, however in this instance condition (a) is false in virtue of nothing in the world satisfying the condition of being the present king of France and consequently conditions (b) and (c) cannot be satisfied . As a result of uncovering this logical structure it is evident that the grammatical subject in these classes of sentences are not singular terms in so much as ‘the king’ does not feature in the logical form of (1) and as such ‘denoting phrases’ cannot hold meaning independent of the proposition . Russell argues that this demonstrates how it is possible for sentences that express a non-referring denoting phrase to be meaningful and hold a truth value despite not referring. (Russell, 1905: 484)
Conversely Strawson contended that the statement was not false but rather devoid of truth value in that it misfired in its attempt to refer, and further argues that this mistake occurs in virtue of Russell’s disregard to the distinction between uniquely referring expressions, that which ‘mentions or refers to some individual person or single object,’ (Strawson, 1950: 324) and a sentence beginning with such an expression. He formulates the distinction between the expression itself, a use and utterance and similarly a sentence, its use and utterance. (Strawson, 1950: 325) If we reconsider example (1), and suppose that it is uttered once during the reign of Louis XIV and once during the reign of Louis XV, in this instance the same sentence has been uttered on two difference ‘occasions of use’ (Strawson, 1950: 327) and it would be logical to admit that it was uttered about different men and as such the statement asserted could either be true or false on either occasion. That is the man who uttered it during the reign of Louis XIV may have been true in virtue of the King being bald whilst the other man may have been incorrect in asserting such a statement. Subsequently it is clear that the same sentence can be used differently on two different occasions of the utterance. Furthermore, if the same sentence (1) had been uttered concurrently during the reign of Louis XIV, then it could be said that the same sentence had been used to express the same statement to the extent that as a result of successful reference to a present king of France the statement must necessarily hold a truth value, however the utterances of sentence (1) would have been different. (Strawson, 1950: 328) Therefore it must be the case, on the evidence presented, that it is possible to use the same sentences to make different statements with altering truth values, and it is in virtue of not recognising the distinction between sentences and expressions and their uses that Russell does not appreciate that it is only through the ‘token use’ (Cohen, 2008: 1) of a sentence that one can make a statement and similarly through the use of an expression or sentence that one can refer to a particular person through context laid out by the statement made and as such an (Lemmon, 1966: 90)
‘Expressions cannot be used to mention, or to refer to, anything, any more than a sentence can be said to be true or false. The same expression can have different mentioning uses, as the same sentence can be used to make statements with different truth values…. mentioning is not something an expression does, it is something that someone can use an expression to do.’ (Strawson 1950: 327)
Accordingly, Strawson concludes that just as the use of a sentence generates the truth value of a statement, referring or mentioning is the function of the use of an expression or sentence in that it is the speaker who uses the expression or sentence that refers and therefore it is illogical to speak of the ‘same things’ about sentences their uses and utterances and about unique referring expressions their uses and utterances for they belong to different categories. (Lemmon, 1966: 91)
So if we again reconsider;
‘The present king of France is bald’ (Russell, 1905: 485)
From the distinctions demonstrated in ‘On Referring’ Strawson challenges Russell’s assertion that this statement is false. He argues that as a result of Russell ‘confusing expressions with their use in a particular context’ (Strawson, 1950: 326) it appears that the sentences appear meaningful and false. However as for mentioned, Strawson considers that whilst the sentence itself holds meaning it is the token use of the sentence on a particular occasion that alters the sentence into a statement which holds truth value.(Cohen, 2008: 3) Consequently Strawson argues that in virtue of (1) being a sentence in itself that it is in fact meaningful in so much as it provides general direction for use, however since any present token use of it does not refer then it is not false but rather lacks truth value in so much as it misfires in its attempt to refer to the (nonexistent) present King of France. (Strawson, 1950: 327) Unlike other false statements such as, ‘the present Prime Minster of England is female, which is ‘defective’ as a consequence of the speaker incorrectly referring to Gordon Brown as female, (1) is defective since the speaker has failed in his attempt to identify a referring subject and as such it should be disregarded as ‘spurious use’ (Strawson, 1950: 327) rather than considered false. (Lycan, 2008: 20)
Strawson further develops his objection by demonstrating that it is the grammatical subject of the sentence that refers to the referent whilst it is the predicate that attributes a characteristic to the referent. However when reference fails it is inconceivable to attribute something to which we cannot refer and as such no attribution take place and therefore the statement cannot hold a truth value. (Cohen, 2008: 3) Unlike sentences akin ‘the present Prime Minister of England is female’ (1) does not attribute anything to the king of France since he does not exist. Consequently Strawson takes up a position which is strikingly similar to Frege who distinguishes between the referent and the sense of a sentence and as such it is possible for a sentence to express more than one sense, or furthermore a sense can be expressed by different sentences. As a result sentences akin to (1) which have a sense but no referent are meaningful in so much as you can understand them, however they lack the quality of truthfulness. (Lemmon, 1966: 94) Strawson perceives Russell as ignoring this distinction and as a consequence he misapprehends the identifying role of the grammatical subject between entailment and presupposition in that anyone who use the sentences (1) to make a statement ‘implies’ that there is a present king of France. (Strawson, 1950: 332) The notion of implies is equivalent to presuppose but it is in no way comparable to the formal logical notion of entails that Russell uses to state the existential conditions asserted by the definite article ‘the’ that necessarily needs to be fulfilled in order for the statement to be true. On the contrary it is to imply in the special sense that the existential conditions are satisfied.(Cohen, 2008: 3) If we use example (1) Russell would argue that ‘The present king of France is bald’ is true if and only if ‘There is a present king of France’ is true and vice versa, however Strawson contends that it is the case that ‘The present King of France is bald’ only holds a truth value if and only if ‘there exists a present king of France,’ if this condition is not met then the statement holds no truth value. ‘So once more I draw the conclusion that referring to or mentioning a particular thing cannot be dissolved into any kind of assertion. To refer is not to assert, though you refer in order to go on to assert.’ (Strawson, 1950: 335)
It is in this that Caton (1959) realised that Strawson did not recognise the distinction between ‘failing to refer to or to mention something and referring to or mention something’ (Caton, 1959: 540) in ordinary language usage and as such it is apparent that Strawson has not actually accounted for the way in which we use ‘refer’ in ordinary language. In other words, when using language we use it the same regardless of whether we are referring to something which exists or does not exist. (Caton, 1959: 540) Consequently, it is only possible to use a sentence or expression thinking it refers even though it does not, and surely even if one did use a non referring expression without intention surely it would appear as if one has referred? Consequently it appears as if it is not that the speaker fails to refer but rather he uses a sentence or expression with the intention of referring to something he believes to exist but actually does not. (Caton, 1954: 542) The implication of this position is that when a speaker utters a statement which intends to refer to an existing thing but fails, the speaker will be corrected by someone as there is no existing king of France. It is just not substantive of our approach to ordinary use of language to suggest that a speaker could use a referring expression successfully or conversely unsuccessfully and furthermore is it not the case that when we utter a sentence, unless we are consciously aware that it is factious, that we would believe that what we are referring to is successful given the linguistic and social conventions that govern our language and thought? Strawson argues that this failure of reference could occur, however given the account above how often would a speaker refer to something believing it exists when it does not, surely even if this was the case, given our social pedigree for awareness, it would be so rare that surely it would be absurd to accept it as a foundation to a critique?(Caton, 1954: 543) Furthermore if Strawson argues that his proposal is less problematic to ordinary language then surely the counterexamples used need be to be more inclined to towards our natural intentions of speech, and as such it seems that a Gricean distinction between intended and expressed propositions could almost accommodate for Russell’s Theory of Descriptions in this instance in that the meaning of the sentences that appear object-dependant studied by Russell are incomplete without the context of use that Grice studied through speaker-meaning. (Ramachandran, 1993: 1)
If we consider the example;
(3)The book case is filled with DVD’s
Strawson uses awkward example such as (3), which expresses that there is one and only one bookcase and it is filled with DVD’s when applied to Russell’s analysis, to devalue Russell’s theories but Neale (1990) developed ideas from Grice and argued that if (3) is applied to the notion of restricted quantification the absurd illusion disappears. That is the domain over which we take the quantifier is adequately altered and consequently the quantifier silently recognises the context of the speech act. Whilst this is by no means a solution it is evident that when a sentence such as (3) is used a contemporary, Russellian could argue that it is in its use that the statement is completed and furthermore this is through the restriction of quantification.(Lycan, 2008: 22)I admit that there is still an ambiguity and awkwardness about the non-existent reference however it has become apparent that in ordinary language it would highly unlikely that a speaker would use (1) with the intention of referring but could it be as a condition of the non-intention of the speaker that the uncomfortable nature of truth arises? (Ramachandran, 1993: 4)
Keith Donnellan (1966) attempts to offer a pragmatic approach to the problem by suggesting that both Strawson and Russell are incorrect in their analysis of definite descriptions by claiming that they both commit to the error that when nothing satisfies the condition of being the referent that the truth value of the sentence or use or the sentence is affected. (King, 1984: 14)Donnellan argues that this is not the case but rather a speaker could use a non-referring expression to express a true sentence despite the grammatical subject being a non-referring expression and this misunderstanding occurs in virtue of neither Strawson nor Russell appreciating both the referential and the attributive use of a definite description as a consequence of not analysing the interpretation of a sentence on specific occasions of use in the specific context.(Lycan, 2008: 24) If we consider the example;
(4) The person who bought that Ferrari is rich
Donnellan argues that that statement could be uttered twofold;
It could refer to the fact that the Ferrari was £300, 000 and as such it must be the case that the person who purchased the vehicle was wealthy, this is known as the attributive use of the description.
It could also refer to Henry the man who purchased the vehicle after haggling the price down to £250 000, this is known as the referential use of the description.
Russell’s theory is an application of the attributive use of definite description in speech acts, whilst Strawson’s argument against Russell is on the basis of him not recognising the referential role of definite descriptions, however Strawson neglects the attributive use and as such his theory is similarly problematic. (Cohen, 2008: 2) Furthermore, the distinction between the referential use and the attributive use runs deeper, in that the truth values are different given there being nothing that satisfies the condition of being the referent in a use of a sentence as a direct consequence of the presuppositions in each case being different. In the case of the attributive use the assertion misfires as a result of their being nothing to satisfy the condition of the referent however in the referential sense it may still have a truth value given that the object that is being referred to may satisfy the description given and if it does not then the given speech act is misleading. As a result Donnellan demonstrates that out of context it is ludicrous to try and determine what a definite description is, because it is context dependant in that within a particular speech act it could be used for either referential or attributive use and as a consequence both Strawson and Russell’s arguments are deficient in some capacity. (Cohen, 2008: 3)
However, given that the question at hand is whether Strawson’s critique has deeper philosophical implications for Russell’s theory, and that we have accepted that both theories are not sufficient if we accept Donnellan’s theory then could it be the case that given that Strawson’s theory is not practical for ordinary language, is it worth accepting in light of the detrimental effect it has on the analysis of language given the arise of truth gaps? Subsequently by accepting Strawson’s thesis on the non truth value of non-referring expression then we permit for these ‘truth value gaps,’ whereby certain sentences cannot be subject to logical analysis because they lack a truth value. (Quine, 1953: 439) By analysing sentences under the scrutiny of formal logic there is the possibility of refraining from ambiguity because all sentences are ascribed a truth value through the proposition expressed by their existential logic properties. (Lemmon, 1966: 105) According to E. J Lemmon (1966) this is only sufficient if logic is not liable to imitate ordinary language to the effect that it manifests itself in the same manner. As Russell seemed to acknowledge this is not the case and in fact the role of logic is to control language under conditions where ordinary language cannot be regulated and therefore surely it is a given that it does not necessarily accurately represent ordinary language in every respect? Besides is it harmful to accept a situation whereby the logician is aware of the problem of non-referring descriptions, would it not be adequate to merely realise this and maintain the logical structure which has been demonstrated as efficient and practical? As far as the difficulty demonstrated by Strawson in this instance it only holds if and only if Russell’s intention was for the logical proposition to mimic ordinary language. (Lemmon, 1966: 107)
In conclusion I would argue that whilst neither Strawson nor Russell’s argument provides a wholly adequate account of definite descriptions, if we evaluate the purpose of each theory they are of differing successes. Russell provides a scientific approach to analysing definite descriptions which generally provides a sufficient measure for reference whilst still encompassing difficulties. Strawson on the other hand attempts to offer an improved theory in regard to ordinary language use however, the counterexamples he provides are not intuitive uses of language. Furthermore the implication is that it is unlikely that such occurrences would occur intentionally and as a consequence the position Strawson takes up is not philosophically practical given the arise of truth gaps when accepting his theory.
Word count (3032 exclusive of reference and quotations)
