4.       Theory of Reference

Some hold that common-sense F-ism depends on a bad theory of reference. It is a theory of reference for M state names: they refer to whatever satisfies the common-sense clauses about pain. We think F-ism doesn’t inherit the faults of the relevant theory of reference.

The Description Theory of Reference

Abbreviations are useful

Many words can be thought of as abbreviations for descriptions. E.g. Gold has a lot of characteristic properties; so ‘gold’ means ‘the stuff that has the properties …’ with a list of properties of gold inserted.

The Definite Description Theory of names claims that all names are abbreviations of definite descriptions of the objects they name.

The referent of a name is whatever satisfies the description that the name abbreviates.

It’s plausible because we used ‘water’ before we knew it was H₂O. It was ‘the colourless, potable, liquid in rain.’ When scientists found the colourless, potable, liquid in rain = H₂O, they knew ipso facto that water = H₂O; and now we know why.

Descriptions and properties

Many terms are cluster terms.

For something to be referenced by ‘water,’ for example, it only has to have enough of the properties to fall under that concept (whether or not there are terms for all those properties.)

This sort of clustering avoids the problem of infinite regress in definitions.

It also allows us to appeal to properties for which we don’t have words.

DD theories allow for change in the meaning of terms by change in the associated descriptions

Objections to the description theory

The associated descriptions differ from person to person

Will your description of acid agree with that of your neighbour? Does the meaning differ among speakers?

                Sure. Why not? As long as we can have an ‘official meaning’ set by experts.

Ignorance of unique specifiers

Most of us can’t give uniquely identifying descriptions for the terms we use. Yet meaning is there.

Can you describe an elm tree well enough to distinguish it from a beech tree?

Can you sensibly say ‘I know nothing about polyethylene?’

All this is ok as long as you can appeal to the use of those who can provide full descriptions.

Twin Earth

Consider the case of Twin Earth. Twin Earth is a world elsewhere in the universe that looks in all ways just like Earth. The only difference is that their water (‘retaw’ we’ll call it) is not H₂O but X_YZ. (On Twin Earth, H₂O is black and tarry.) If an Earthling on Earth speaking Earthish asks ‘What is water on Twin Earth?’ the answer is ‘That black, tarry stuff.’ So water on Twin Earth doesn’t satisfy any of the descriptions of water, yet it is water. And X_YZ on Twin Earth satisfies all the descriptions of water and isn’t water. So ‘water’ on Earth can’t just be an abbreviation of descriptions – even under the cluster theory.

The Causal Theory

Causal theories are preferred these days. Here’s one version.

Water stereotypes

Water stereotypes are exemplars of water. For example, the sea, a puddle, a glass of it, rain.

We assume they are examples of a natural kind.

[A natural kind is a ‘true’ division of reality.]

We interact causally with the exemplars of the natural kind water with which we come in contact.

Note that our assumption that they are examples of a natural kind precedes our knowledge of that natural kind’s actual characterization.

[The natural kind water is actually characterized by its being H₂O.]

This is why we can say that the black, tarry stuff on Twin Earth is water: because it’s really H₂O – i.e. it’s an exemplar of the natural kind water – and it was H₂O that we were always causally connected to when experiencing water.

For similar reasons what the Twin Earthlings call water is not water.

The description theory cannot explain these two facts.

Exteme versus moderate causal theories

There’s a depth problem.

Any causal connection to an exemplar of n.k. water is only a section of a causal chain that extends between ourselves (and the term) and the exemplar; but this chain really continues beyond both.

What is special about the exemplar in this infinite chain?

Why does ‘water’ point there and not at the light from the water or the rainstorm that created the puddle?

There’s a spread problem too.

There are many n.k. to which each exemplar belongs. Any causal chain connects to all those n.k.

What makes the natural kind water special?

Causal considerations alone can’t answer these questions. The natural answer is to say that to grasp the meaning of ‘water’ is to know that the relevant cause is a n.k. whose exemplars have the right properties.

But this makes the causal theory partly a descriptive theory as well. We call it this causal descriptivism.

Rigid Designation

Definite descriptions are non-rigid

A non-rigid designator is a term that picks out different objects in different possible worlds. Most descriptions are nrds; e.g. ‘the tallest man alive.’ In particular, ‘the watery n.k.’ picks out H₂O on Earth and X_YZ on Twin Earth.

‘Water’ is rigid

A rigid designator is a term that picks out the same thing in all possible worlds. ‘Water’ does that. Therefore ‘water’ can’t refer to just the thing that is properly causally related to us and has the right properties. Now, we saw that the causal theory needed to be fixed so that water also satisfied a cluster of property descriptions; but this doesn’t mean that water satisfies that cluster in all possible worlds, only that ‘water’ refers to what has the right causal connection and properties in the actual world.

‘Actually’ as a rigidification device

Definite descriptions can be turned into rigid designators by rigidification.

‘The tallest man might not have been the tallest man if he didn’t eat his weeties’ = nonsense.

‘The actual tallest man might not have been the tallest man if he didn’t eat his weeties’ = not nonsense.

‘actual’ has the effect of pointing at the man in the actual world so described – and in fact pointing at him in all possible worlds, including worlds in which he doesn’t eat his weeties. It rigidifies the description.

Giving the meaning versus fixing the reference

‘Water’ doesn’t mean the kind satisfying the water descriptions and having the water related causal connections (i.e. being ‘the watery kind,’) it means the actual watery kind. It is rigidified.

We fix the reference of ‘water’ in any world as the kind that is watery in the actual world.

The significance of all this for Theory of Mind is

1.                    it helps understand what is necessary

2.                    it makes it possible that a term associated with a set of features may not denote whatever has those features but may have its reference fixed by them.

c.f. the clauses in common-sense functionalism.

The Necessary A Posteriori

The fact of rigidified definite descriptions means that some things can be both necessarily true and known a posteriori

Rigid designation and necessity

For example: from above we can say that ‘water is H₂O’ is true in all possible worlds, therefore it’s necessarily true. (Any sentence like that with rigid designators on both sides is necessarily true/false.)

OTOH ‘water is H₂O’ is not known a priori; it’s a scientific discovery.

Point is that claims of possibilities (e.g. of multiple realizability of Minds) guide us in our theorizing about Mind. Now we know that just because things may be known about our Minds a posteriori nothing follows about their necessity from that way of knowing about them.

Intuitions about possiblility

However, we don’t conclude from this that we can’t rely on our intuitions about possibility.

Note that we only know about reference fixing because of our intuitions about possibility.

Just be careful not to say X is only known a posteriori, so it might be true or it might be false.

This is all unsatisfactory – but what else can we appeal to but our intuitions when it comes to ‘possibility?’

A new kind of possibility and necessity?

Some distingish metaphysical modalities from logical ones. MPs and MNs are discovered a posteriori.

We think, however, that water is necessarily H₂O because our concept of water is the concept of the thing that has the watery properties – which includes being H₂O. This is just logical necessity.

The necessary a priori

So: ‘water is H₂O’ is necessary in the same way that ‘water = water’ is; although the latter can be known a priori, while the former can’t.

We haven’t found a new sort of modality, we’ve only found a new way that we use words.

The issue of essential properties

But even if ‘water is H₂O’ is not necessary, couldn’t it be that water was essentially H₂O?

If ‘water’ was just a definite description and not rigidified, then water could be non-H₂O in some PW. But it could not be the substance it is in this world and not be H₂O. So that is an essential property. We might say, though, that it wouldn’t be being correctly called ‘water’ in all PW.