(i) The speaker means to refer to Jones, who was caught as he was killing Smith with a power saw.
(ii) The speaker means to refer to whoever murdered Smith with a power saw.
(iii) The speaker means (ii), but he does not know that in reality Smith is not dead (he was saved in extremis by Doctor Jekyll). In fact, there should not be a referent for the expression Smith’s murderer, but the principle of charity prompts one to imagine that the speaker means to refer to the unsuccessful murderer (who would still be mad, and incompetent into the bargain).
(iv) The speaker means (ii), but the speaker is probably mad, because no one has made an attempt on Smith’s life. The listeners understand that the speaker is making a hallucinatory reference to an individual or a situation from the possible world of his beliefs.
(v) The speaker believes (mistakenly) that Smith has been murdered, that the murderer is Jones, and that everybody knows this. If the listeners do not know that the speaker harbors these strange beliefs, we are in situation (iv). If the speaker goes on to make his beliefs explicit, the listeners will understand that he is referring to Jones. Now it will be necessary to decide whether the speaker thought Jones was mad because Jones was Smith’s murderer or for other reasons (with the result that the speaker will still think Jones mad, even though Jones did not murder Smith).
(vi) Smith really has been murdered, and the speaker believes that the murderer is Jones (while everybody knows it was Donnellan). The interlocutors do not know the speaker’s beliefs and think he means to say that Donnellan is mad (which is clearly false, because Donnellan has murdered Smith for scientific reasons, so as to be able to work on the difference between attributive and referential use). I imagine that, if the conversation went on for a little, it might be possible to clear up the misunderstanding but—as in (v)—extra information will be required to establish whether the speaker means to refer to Jones, though Jones is innocent, as a madman.
(vii) Smith really has been murdered, and the speaker believes that the murderer is Jones (while everybody knows it is Donnellan). But the listeners know that the speaker is biased against Jones and has repeatedly stated that he believes him to be Smith’s murderer, and therefore they understand that the speaker means to refer to Jones.
(viii) The Smith murder trial is coming to an end, and in the dock Donnellan is listening to the sentence that officially defines his guilt. The speaker (a psychiatrist) has just entered the courtroom and thinks that Donnellan is a certain Jones he had known in the mental hospital. He is therefore referring to Jones and not Donnellan. Naturally the listeners believe he is referring to Donnellan. But I imagine that they will ask him to explain his judgment, and in the course of the conversation the referential misunderstanding might be cleared up.
This is a set of cases in which the reference is negotiated, and in which we cannot speak of an act of reference that is independent of the intentions and knowledge of the speaker and that points to a haecceitas of which the speaker knows nothing.
What does Nancy Want?
But the same distinction between referential and attributive use leaves many borderline cases uncovered. Let’s take a look at another famous example, reworked for the occasion.16
Let’s suppose that I say, Nancy wants to marry an analytic philosopher. We can give two semantic interpretations, (i) and (ii), of this sentence that are possible even when the sentence is uttered out of context, and at least three pragmatic interpretations, (iii) through (v), that depend on some inferences regarding the speaker’s intentions. Interpretations (iii) through (v) can be attempted only after a decision has been made between (i) and (ii):
(i) Nancy wants to marry a determined individual X, who is an analytic philosopher.
(ii) Nancy wants to marry anybody, as long as he is an analytic philosopher.
(iii) Nancy wants to marry a determined individual, an analytic philosopher: she knows who he is, but the speaker doesn’t, because Nancy has not told him the name.
(iv) Nancy wants to marry a determined individual X, an analytic philosopher: she has also given the speaker the name and introduced them to each other, but out of discretion the speaker has thought it more fitting to avoid going into details.
(v) Nancy has taken a fancy to a fellow and wants to marry him; she has the told the speaker who he is; the speaker happens to know that the person in question is an analytic philosopher. At this point it is irrelevant whether or not Nancy knows that the fellow is an analytic philosopher, or whether or not the speaker has told her. The fact is that, as Nancy is doing her dissertation on Derrida, the speaker thinks that the two of them will never understand each other and that their marriage is doomed to fail. He tells his interlocutors (who know Nancy’s ideas very well) of his perplexity.
Interpretations (iii) through (v) depend on interpretation (i), that is to say, on the decision, which has been made, to consider the sentence referential. The listeners will presumably ask for more information about this X, and in that case the speaker must either confess that he does not know him (case (iii), both he and the listeners must accept the reference on trust), or justify reticence (case (iv), only the listeners must accept the reference on trust), or provide instructions for X’s identification or retrieval (by opening our «white box»). Or the listeners are uninterested in the identity of the speaker (this piece of gossip is juicy only because X is an analytic philosopher), and there the matter stops.
This leaves interpretation (ii), which at first sight would seem to point to an attributive use of the sentence. But, above all, we should note that attributive use (a la Donnellan) is also a case of referring. As a matter of fact, although the speaker did indeed define whoever killed Smith as a madman, in reality his supposition was that Smith was killed by a specific individual (albeit an unknown one as yet), and that was the individual he was referring to, even though on trust. Talking of Smith’s murderer was like talking of the first victim of the Second World War. The unknown X who killed Smith was mad; that precise X who was killed before anyone else was unlucky. But madness and bad luck are predicated of an X, who, while remaining socially or historically or juridically indefinable, is ontologically defined.
But here we are talking not about whom Nancy may marry (in which case that man, even unknown, would still be one person and one only). Nor are we talking about whomever Nancy will marry, that is, her possible husband, in which case it would be as if a pregnant woman talked of the child to be born to her in a few months: whatever it may be, it will certainly be the son/daughter born of her womb at a fairly specific moment and equipped with a given genetic inheritance (or it might not be born, and this is precisely why it is possible). We are talking about whom Nancy would like to marry. The entity in question is not only possible but also optative.
The individual whom Nancy is said to want to marry is not only undefined as yet but also might never even come on the scene (and Nancy would remain single). To the extent that she is prepared to marry whoever has the property of being an analytic philosopher, she is in love with a property, as if she wishes to marry whoever has a mustache. Perhaps during her wildest erotic fantasies, Nancy has assigned a face to this imprecise X, imagining that he looks like Robert De Niro. She is prepared to compromise on looks, height, and age, as long as her X is an analytic philosopher, and therefore Kripke or Putnam would be equally good for her, but certainly not Robert De Niro.
Nancy (or whoever is talking about her intentions) is referring not to an individual but to a class of possible individuals, and therefore she is not performing an act of reference. Nancy’s X is a general object, such as cats in general. And since I feel it is inopportune to talk of referring in the case of general objects, the sentence ought to be translated as Nancy has the property of appreciating analytic philosophers (in general) and of desiring them as possible husbands, or Among their many properties, analytic philosophers also possess that of being desirable to Nancy. Even though this would still be a reference to Nancy, it would not be a reference to any specific analytic philosopher.
We should also consider that it is by no means certain that Nancy wants to marry whoever is an analytic philosopher. It might mean that she intends to get married, has not yet decided with whom, but definitely wants the chosen one to be an analytic philosopher. Also, she does not intend to throw in her lot with just any analytic philosopher, only with an analytic philosopher she likes. If a marriage broker suggests she try Marco Santambrogio (who has the dual property of being an analytic philosopher and remarkably good-looking) Nancy might grumble, for example, because she does not appreciate his vis polemica.
Before saying
