L1.5 Lottie’s fifth lesson on concepts

The next day is Friday. Lottie thinks of the weekend stretching ahead of her. It helps that there is never teaching in the afternoon on any day and that, as the Potting Shed’s only non-adult, no other duties – aside from being educated – fall on her. Sometimes this does mean that she thinks she ought to try just a bit harder.

Her tutor, Hipparchus, asks her about the last thing they’d discussed and, today, Lottie has given it some thought. “It depends…” she hazards. “It would seem odd to use my knowledge that the person ‘over there’ didn’t want to be like the losers who drink al-co-hol (she spreads out the enunciation) and just drank Svalbarði while my colleague, being a loser himself, didn’t know this, as a reason to think – just because of that – that he wasn’t ogling her!”

Hipparchus decides to ignore the extra details that have now arrived in the thought experiment. “So even though the definite description doesn’t apply to that person, your colleague might still have been talking about them.”

“He meant her! Them!”

“So, can we tweak the example so it goes the other way? Is there a way to make it plausible he might be talking about someone who happens to be drinking a martini from a mug, though no one knows this?”

Hipparchus begins to dust the leaves of his plants seemingly in no hurry. Eventually Lottie hits on a solution. “Suppose I spat in the martini, then whoever in the party drank it – if anyone did – would be drinking my spit!”

These encounters do seem bracing. Most unlike his Cambridge supervisions. 

“Yes. Fortunately for that person, a martini has almost no martini in it. The light merely has to pass through the bottle and shine on the glass of iced gin. But you’re right. Keith Donnellan, whose example it is, suggests in a paper ‘Reference and definite descriptions’ that we can refine Russell’s account by thinking that there might be two different uses of a descriptive phrase. In one, one might have no particular person in mind. It is a purely descriptive thought. He calls that ‘attributive’. It selects someone, just as Russell thought, who satisfies the description if someone does. In the other case, one already has someone in mind and merely uses a description to communicate the object of one’s thought to others. He calls this ‘referential’. The former behaves as Russell says. The latter does not. In response to the question of whether someone speaks truth if what they go on to say about the object of thought is true but the description they use to highlight it is not: he says ambiguously that they say something true.

Now there is a response to this by Saul Kripke who thinks that Donnellan’s points are not so critical of Russell after all. But the point I wish to draw from it is that we need to look at the use of a descriptive phrase in context to see whether it is expressing a purely descriptive thought (attributive) or whether it is an accidental means of communicating a singular thought (referential).”

They both pause and blink. But Lottie has nothing she wants to say now.

Are any proper names really proper names?

“I want to say something about two other things today. First, I want to address your question yesterday of whether there are any proper proper names and, second, I want to mention a link to natural kind terms.

One challenge for Russell’s account of what look like, but are not in fact, proper names is that if apparent names are really short-hands for descriptions, he will need to set out what the specific descriptions are. And now, given his worries about fallibility, is there any chance these could be wrong? I mean: what would it mean for them to be a wrong?”

There is another long pause. Well, it is Friday. Lottie says:

“It might be that no one fits the description. Yesterday, there might have been no 20th visitor to the Potting Shed. No one might be drinking the martini I spat in. No one person might have written Hamlet except lots of monkeys. Stupid monkeys!”

“So what does that do to the theory? After all, it is always possible to discover that some historical person, whose name we know, didn’t do exactly what we thought they did? It would be odd if we had to be exactly right or they never existed.”

“Maybe they just need to have done most of what we think they did?”

“That’s a good and sensible suggestion and one taken up by John Searle (in his paper ‘Proper names’). He thinks that names stand for a cluster of descriptions only some of which need to be true of someone to connect them to that name. Hold that thought. Thinking back to sense and reference and the Clark Kent vs Superman example, does it help?”

“Obviously! The ‘Clark Kent’ sense has lots of nerdy journo boring things and the ‘Superman’ sense is all the good heroic stuff. Different descriptions even if, almost unbelievably, it’s the same man.”

“Yes that’s why it is often assumed that if you believe in Frege’s distinction of sense and reference you have to believe in a descriptive theory of apparent proper names. Again, let’s hold that thought.”

“No I won’t. Well, anyway, there’s something else you must tell me. If names aren’t short for descriptions, what are they? Or, I suppose, what do people who think Frege is stupid think about names?”

A causal non-descriptive picture of names

“Well, one option was popularised by Saul Kripke in his book Naming and Necessity. He has lots of criticisms of the idea of names as descriptions. For example, if you use a description to connect the name ‘Gödel’ to a person, you had better not use the description ‘discoverer of the Incompleteness Theorem’ unless you can also do more to link that name to a theorem than say ‘the theorem discovered by Gödel’. In general, we do not know enough for our descriptions to specify uniquely the right individual. Also, even if Gödel did invent that theorem, it is possible to consider possible worlds where he did not. So being its inventor is not a necessary aspect of our understanding of who he is.

And so instead, Kripke suggests that the right account is that there is the right sort of causal connection between someone’s baptism and the use of their name ever after by other people, passed on in conversations over time. This looks like a reference-only theory of names. As long as you latch onto a name being used in a conversation, you can carry on using it about that person even if you know nothing about them. It is object-dependent or singular but, by contrast with my thought that that football, at several removes. 

In fact, Hilary Putnam (male, just so you know) has a similar account of the identity of natural kinds (in ‘Meaning and Reference’ (1973) and ‘The Meaning of “Meaning”’ (1975)). If we baptise some substance as ‘gold’ then ‘gold’ names whatever natural kind has the same sort of essence as that sample. ‘Gold’ is the name of some essential structure. The cluster of descriptions we might hold true of it – heavy, metallic, malleable, yellow – need not be true of ‘gold’ because that is not how the name latches onto that kind of substance. Further the properties need only be contingently true of gold. But gold must have its essential structure in all possible worlds. Interestingly, ‘gold’ cannot name its instances as any instance need not have existed.” (Hipparchus returns to this later, Ed.)

“So is singular or object-dependence the opposite of having a sense?”

“No, though philosophers often used to think so. Russell did. In fact, one central aim of Gareth Evans’s influential book The Varieties of Reference is to set out and defend a neo-Fregean account of singular and object-dependent thoughts combining insights from both Frege and Russell. An example might be keeping track of a day as it recedes into the past, thinking of it as today, yesterday, last week. Or tracking a cat as ‘that cat’, seen in the distance, to ‘this cat’ when it approaches. In both cases, there is a changing way that the day or the cat is thought of but which is not captured in a description. There are senses though the thoughts are also singular.

Let me leave you with a paradoxical idea. We have considered the idea that apparent proper names are really disguised descriptions. That would explain the link between a name and what it names and it is one way they might have senses. But it is subject to criticisms. Such names do not seem to behave like descriptions. An alternative is to explain the link between a name and an object via a causal connection. We’ve not had time to discuss this but one problem is that the links in the causal chain have to be the right ones, preserving the right link back to the named object and there are famous examples where this is not so. The problem is that we cannot appeal to our knowledge of what a name names to explain what constitutes making the right connection to past use. The right connection – one which intuitively gets the name right, linking it to the right initial baptism – should fall out of some independent specification of causal connections. But, like many attempts at reduction in philosophy, that looks difficult.

So think about this paradoxical idea as a third option. Proper names are names if they are used in the characteristic way expressive of singular – though in this case not demonstrative – thoughts. One feature of that, in the complete opposite spirit to a descriptive theory, is that it is always a standing possibility that one asks whether some particular description really holds good of the person named. And yet, at the same time, how we pick out the referent of the name is always via some description or other. The finesse of this account is that it accepts that descriptions have a role in picking something out but exactly not to characterise something as a name in the first place.”

“No, sorry. No. You’ve lost me.” And Lottie walks away happy not to have to think about any of this for the next three days.

Lesson 6