# Millianism and the Metaphysics of Propositions

The dominant view about the semantics of proper names is probably Millianism.  According to Millianism, the meaning (i.e., semantic value) of proper names is just their referent.  Most Millians cash out their view in terms of singular Russellian propositions.

(A Russellian proposition is an ordered tuple containing the objects and properties referred to by the names and predicates in the sentence expressing the proposition; for instance, the sentence ‘Aristotle is a philosopher’ expresses the proposition <Aristotle, the property of being a philosopher>.)

To solve several problems facing Millianism, such as the problem of handling names without a referent (’empty names’), some Millians appeal to pragmatic processes such as Gricean implicatures.  I will call this combination of Millianism and pragmatic processes ‘pragmatic Millianism’.  Versions of it have been proposed by Scott Soames and many others.

Here is how pragmatic Millianism works in the case of empty names (as proposed by Ken Taylor and Fred Adams and collaborators).  Since an empty name has no referent, it cannot contribute its referent to the singular proposition expressed by sentences containing it.  The result is a so called ‘gappy proposition’, i.e. a proposition with an empty slot corresponding to the subject of the sentence.  For instance, the sentence ‘Santa Claus is a philosopher’ expresses the gappy proposition <____, the property of being a philosopher>.  Gappy propositions have no truth values (according to most theorists), but listeners of sentences containing empty names have the intuitions that such sentences have truth values.  Where do such intuitions come from?  According to pragmatic Millians, they come from non-singular propositions implicated by the gappy proposition.  In our example, the implicated proposition might be expressed by a sentence along the lines of ‘The jolly fat man who brings presents to children at Christmas is a philosopher’.  According to pragmatic Millianism, listeners get a bit confused between what is actually expressed by the original sentence and what is implicated by it, so they attribute truth values to the original sentence.

Now, I find this theory somewhat implausible for a number of reasons having to do with the semantics of proper names and sentences in which they occur, but this is not what I want to ask about.  I want to ask about the interaction between the metaphysics of propositions and the pragmatic processes postulated by pragmatic Millians.

The problem is that I don’t understand how pragmatic processes such as inferring Gricean implicatures can possibly be defined over Russellian propositions.  On one hand, Russellian propositions are ordered tuples of things in the world.  On the other hand, inferences to Gricean implicatures are presumably defined over mental representations.  So it seems that in order to be metaphysically kosher, pragmatic Millianism should be formulated in terms of mental representations (of propositions?), rather than the propositions themselves.  But neither Taylor nor Adams et al., whose works I’ve been reading, ever qualify their talk of propositions in this direction.

Does anyone have any thoughts on what’s going on here?

1. Malcolm

When you mention pragmatic processes, what physical events are you referring to? As long as such discussions remain in some imaginary intellectual universe, the problems you point to are inevitable. A little neuroscience and Information Theory are in order.

2. gualtiero

I appreciate your instinct but for now, discussions of linguistic semantics are quite removed from neuroscience and information theory. There is still a question of what kind of entities should be posited by such theories. Even assuming that propositions are ok, I would think that what listeners process (what they infer implicatures from) are at best representations of propositions, not the propositions themselves. Or am I missing something?

3. Aidan

I guess I’m not really clear on what the question is here. If one thinks of conversational implicatures as the results of some inferential process (as philosophers tend to) and think of inference as some sort of transition between mental states (as seems natural) then it follows that a story about why some utterance in a discourse implicates this or that will be, at bottom, a story about a transition between certain types of mental states (communicative intentions, recognition of communicative intentions, shared background beliefs etc.). Philosophers of Language tend to speak of these things simply in terms of their content (that is in terms of propositions or sets of propositions). Do you imagine any particular complexity arising if they acknowledged that speaking in terms of propositions is just a shorthand for speaking in terms of mental states which have those propositions as their content? Is there some particular worry about vacuous names? As I read the post I thought there might be some unity-of-the-proposition-ish worry in the background but it never came up.

For what it’s worth: I’m also suspicious about the Millian account of the semantics of proper names and sentences in which they occur.

4. gualtiero piccinini

Thanks for your comment.  I agree that your reading of what is going on (talking about propositions to talk about the content of mental states) is the charitable reading of what philosophers of language are talking about.  But I am not very comfortable with this way of talking, especially when it comes to empty names.

Most Millians believe that since empty names have no referent, sentences containing empty names express gappy propositions.  This may sound plausible when we are thinking of meaning as something that pertains directly to language, without considering the role of the mind.  But if we think in terms of representations, what would a mental representation of a gappy proposition look like?  And why would you think that anyone would have such a representation in mind?  Some Millians even talk about “ways of grasping” gappy propositions.  What are those, and how do they work?  I’ve been doing a lot of readings on this topic, but I haven’t seen any answer to these questions.  The gappy proposition view may sound plausible when you derive it from Millianism.  When you think in terms of mental representations, it just seems bizarre and unmotivated to me.  This is one way to express my discomfort with gappy propositions vis a vis mental representations.

5. Aidan

Thanks, I think I have a better of idea of the issue now.

My knowledge of the literature you are engaging with is a couple of years old (and so probably degraded somewhat) but my impression is that “gappy propositions” are supposed to be something like functions from objects to propositions. I’m not a philosopher of mind – is there some particular worry about thinking that people can have mental representations of such functions? Without meaning to be glib: I’m pretty sure I have a representation of the function that takes an object as argument and returns the proposition that that object is jolly as a value. On the Pragmatic-Millian story isn’t some such function part of the content of my mental state when I hear a sentence like “Santa Claus is tall” uttered?

Again, I don’t think this is a good view, I’m just not quite seeing the force of your criticism.

Apologies if I’m way off base.

6. Matthew Arnatt

I want to say something about the last substantial paragraph in the original post. I think there is a problem of defining any sort of process over Russellian propositions (in the sense offered here). I’d been thinking of something like this anyway as I was reading early Christopher Peacocke again, who was talking about ‘content’ in terms of some adjudicated holism, which holism seemed to me to rely upon smuggling in some objects or other by way of settling something like the sustained reasonableness of a local holism (in the guise of a essential or revealingly essentialising process. I suppose what one might really want is something like objects that can be in a way parts of tautologies.)