Keith Sedgwick DONNELLAN and Herbert Paul GRICE's distinction between identificatory and non-identificatory uses of descriptors, or the story of Jones's butler. Keywords: Grice as Russellian; Strawson -- Robbing Peter to pay Paul.

Speranza

Grice refers to Donnellan's essay, 'Reference & Definite Descriptions, Philosophical Review, vol.75 (submitted in 1964, but rejected, and finally published in 1966).  

 

Grice says he is not sure he is "wholly sympathetic towards the conclusions which he draws from the existence of the distinction" between what Donnellan had dubbed an "attributive" and a "referential" use of a definite description.

 

Grice's two corresponding usages are labelled, respectively, "non-identificatory" (and symbolised by italics) and "identificatory" (symbolised by capitals)

 

  1. Jones's butler will be seeking a new position.
     *The F* is G.
  2. Jones's butler got the hats and coats mixed up.
     THE F is G.

Grice refers to these different 'usages' of definite descriptions' pertain to pragmatics (utterer's intentions and conversational implicature), rather than semantics (logical form):
 

Grice writes: "It is important to bear in mind that I am NOT suggesting that the difference between these devices represents a difference in the MEANING or SENSE which a descriptive phrase may have on different occasions; on the contrary, I am suggesting that descriptive phrases have no relevant systematic duplicity of meaning: their meaning is given by a Russellian account."

 

Grice is again robbing Peter to pay Paul -- (Donnellan, interestingly, criticised BOTH Russell and Strawson, while Grice remained a faithful Russellian, vide his "Definite descriptions in Russell and in the vernacular").

 

Grice is here concerned with what we may regard as 'knowledge by acquaintance', or to use G Evans's phrase (in The Varieties of Reference), 'recognition-based identification'. For this Grice provides a linguistic device, namely, the addition, to any definite description, of the phrase, 'whoever he [it/she] may be". Grice notes that only the attributive (or non-identificatory) usage allows for the addition of the phrase:

   Jones's butler, whoever he may be, will be seeking a new position.
   *Jones's butler, whoever he may be, got the hats and coats mixed up.

Grice writes:

"We may say with respect to [the referential or identificatory use] that some particular individual has been 'described as' [or 'MISdescribed as', as the case may be. Speranza], 'REFERRED to as', or 'called' Jones's butler by the [utterer]. No such comments are in place with respect to [attributive or non-identificatory usages]."

 

It is in this context that Grice introduces the notion of a "dossier" associated to a definite description. In this he is, as Evans notes, more or less following Strawson in 'Identifying Reference & Truth Values' (in
Logico-Linguistic Papers), although not explicitly using the notion of 'information'.

 

Grice defines a dossier as follows:
 

"Let us say that utterer U has a dossier for a definite description D if there is a set of definite descriptions which includes D, all the members of which U supposes to be satisfied by one and the same item. In [the non-identificatory usages], unlike [the identificatory ones], the [utterer] intends the audience to think (via the recognition that he is so intended)
1. that U has a dossier for the definite description D which he has used, and
 2. that U has selected D from this dossier at least partly IN THE HOPE THAT THE ADDRESSEE HAS A DOSSIER FOR D WHICH 'OVERLAPS' the utterer's dossier for D.

 

This is labelled a 'tremendously important feature of referring expressions' by Evans, and given a kind of cognitive background:

"[An] organism must have more than the simple capacity to find previously encountered objects familiar; a recognitional capacity must be associated with (i.e. enable the subject to recall) the appropriate dossier of information. Under these circumstances, selection pressures would strongly favour organisms which possessed, or were capable to developing, this capacity. G Evans, The Varieties of Reference, Oxford University Press, p.276. (Grice also cited on p.50 and 306).

Grice proposes a method having to do with semantic interpretation (rather than the syntactical one involving the 'whoever he may be' idiom) when stressing the fact that the existence of BOTH these two non-identificatory and identificatory uses conform to truth-conditional semantics. The apparent difference is explained in terms of the distinction between what the utterer SAYS - rendered by means of the logical form - and what he has MEANS (rendered by the communicative intentions):
  

"the truth-conditions for a statement [of the identificatory type] no less than for a statement [in the non-identificatory use], can be thought of as being given by a Russellian account of definite  descriptions. Though what a [utterer] has SAID may be FALSE [in
the identificatory use] what he MEANT may be true (for example, that a certain particular individual (who is in fact Jones's gardener) mixed up the hats and coats.

 

Another device, of the linguistic type, that Grice introduces to distinguish between these two uses is the effects of the replacement of the definite description by a [proper] name, yielding:
 

 *Jones's butler* (let us call him Bill) will be seeking a new position.
  JONES'S BUTLER (let us call him Bill) got our coats and hats mixed up.

Here, there is an apparent asymmetry between which Grice analyses, precisely, in terms of the distinction between a truth-value associated to WHAT IS explicitly communicated (for the non-identificatory use), as opposed to WHAT IS MEANT (including conversationally implicate) (for the identificatory use):
 

Grice writes:

"A subsequent remark [to a non-identificatory use] containing [the newly introduced proper name] 'Bill' will have THE SAME TRUTH-VALUE as would have a corresponding remark in which *Jones's butler* replaces 'Bill'. If Jones has no butler, and if in consequence it is FALSE that *Jones's butler* will be seeking a new position, then it will be [also] false that Bill will be seeking a new position ... [However, in the case of an identificatory use], it will NOT be true that a a subsequent remark contaning [a newly introduced proper name such as] 'Bill' will have the same truth-value as would have a corresponding remark in which 'Bill' is replaced by 'Jones's butler'. For the person whom the [utterer] proposes to call 'Bill' will be the person whom he MEANT when he said, 'Let us call JONES'S BUTLER [sic in capitals. Speranza] Bill', viz. the person who looked after the hats and coats, [...] and if this person turns out to have been Jones's gardener and not Jones's butler, then it may be TRUE that Bill mixed up the hats and coats and FALSE that Jones's butler mixed up the hats and coats."

 

The distinction between the non-identificatory and the identificatory use is based thus on the way in which they react to the replacement of the definite description by a proper name.

 

Grice writes:

"[In the identificatory use] remarks of the form 'Bill is such-and-such' will be INFLEXIBLY TIED [cf. Kripke's rigid designation], as regards truth-value, not to possible remarks of the form 'Jones's butler is such-and-such", but to possible remarks of the form 'The person whom U meant when he said 'Let us call Jones's Butler 'Bill' is such-and-such'".

 

Grice further considers the issue of "vacuous" descriptions, in connection to Strawson's concerns. Grice writes of the different uses of 'vacuous definite descriptions', even when used in an identificatory way - including what he calls 'parasitic' uses.

 

Grice writes:

"It is important to note that, for a definite description used in explanation of a name to be employed in an IDENTIFICATORY way, it is NOT required that the item which the explainer MEANS (is referring to) when he uses the description should ACTUALLY EXIST. A person may establish or explain a use for a name A by saying 'Let us call THE F A' or 'THE F is called A' even though every definite description in his dossier for 'the F' is vacuous. He may mistakenly think, or merely deceitfully intend his hearer to think, that the elements in the dossier are non-vacuous and satisfied by a single item; and in secondary or 'parasitic' types of case, as in the narration of or commentary upon fiction, that this is so may be something which the utterer non-deceitfully pretends or 'feigns'.

 

Incidentally, when Donnellan replied to criticisms [by A. Mckay] to his 'Reference & Definite Descriptions' (which we have seen is cited by Grice in 'Vacuous Names'), as e.g. 'Putting Humpty Dumpty Together Again' (Philosophical Review, now repr. by OUP in the Donnellan volume) he relies on Grice's intentional stance. Donnellan is concerned with A. Mckay's allegation of 'linguistic anarchism' (as when an utterer can mean what he wishes, as Humpty Dumpty does in Alice in Wonderland). Donnellan writes:
  

Donnellan writes:

"In the analysis of meaning given by Grice [In Meaning [PR, 1957, but written in 1948], reprinted in Studies in the Way of Words], a speaker means something by an utterance when he has a certain complex kind of intention involving recognition on the part of his audience of his intention ... It does not follow from this analysis that speakers might, out of the blue, mean anything at all by any utterance."

 

Donnellan's example is using 'the book' to mean a rock.

 

Donnellan writes:

 

"Suppose there is a rock on my shelf that has been carefully carved to resemble a book and that I know the person I am speaking to cannot recognise it for what it is. I say to him  'Bring me the book with the blue binding'.  It seems at least plausible to say that I referred to a rock. When a speaker uses a definite description referentially [or as Grice would have it, in an identificatory way. Speranza] he intends his audience to take the descritption as characterising what it is he want s to talk about. p.214.

An interdisciplinary note: the linguistic-philosophical literature on definite descriptions is also dealt with, from a cognitive linguistic point of view, by P N Johnson-Laird and A Garnham, of the Centre for Research on Perception & Cognition at the University of Sussex, in their essay, 'Descriptions & Discourse Models' (Linguistics & Philosophy, 3, 371-393, and cfr. further work in Oxford by M. Sainsbury and D. E. Over).