Intro to Language: The Referential Theory of Meaning

by Luke Muehlhauser on May 2, 2010 in Intro to Language

Part 2 of my Intro to Language series.


Few people know that before he came to the United States and became a movie star, Charlie Chaplin performed in a comedy troupe with his parents in Shanghai, who raised him in the British Settlement in the Huangpu district of China’s largest city.

There is a reason few people know this. None of it is true. But remarkably, as you read through that opening sentence – let’s call it sentence (1) – you understood it perfectly, and quite easily, and not by having seen it before, and not by having ever read any sentences very similar to it.1 How is this possible?

One clue is that (1) is a string of English words, each of which you understand individually, and you also understand something about how words work when they are arranged together using the rules of English.

But how is it that a sequence of marks or noises is meaningful?

Referential Theory

One common sense theory is that expressions have meanings because they stand for things: they mean what they stand for. Words are like labels. “Charlie Chaplin” denotes the person Charlie Chaplin. The word “cat” stands for a member of Felis catus. And the sentence “Charlie Chaplin kicked a cat” stands for Charlie Chaplin having kicked a cat, presumably because “Charlie Chaplin” denotes the person Charlie Chaplin, “kicked” stands for the act of kicking in the past, and “a cat” refers to a member of Felis catus.

This Referential Theory of Meaning is attractively simple, but it has some problems.

Problem 1

Not every word refers to an actual thing.

First, some words don’t refer to anything that exists. “Pegasus” does not denote anything real, because there is no winged horse after all. Also, consider the sentence “I saw nobody.” But to what does “nobody” refer?

Second, consider:

(2) Ralph is fat.

What does “fat” denote? Perhaps it denotes something abstract, like the property of being fat. Or as Plato would have called it, The Fat Itself. So perhaps when we express (2) we mean that Ralph exemplifies the property of being fat. On this view, “is fat” means something like “has fatness.”

So maybe what we’re doing is joining together the name of a person (Ralph) with the name of a property (fat) by using “is.” But then what does “is” stand for? Perhaps it stands for the relation of “having.” So (2) really means something like “Ralph bears the having relation to fatness.”

Uh-oh. Now it looks like we need to explain what “bears” refers to. And this could go on into infinity, and we’d never be able to work out the referential meaning of the sentence.2

Third, there seem to be nouns that do not name individual things or kinds of things or even abstract things like the property of being fat. Think of words like “sake” and “behalf.”3 I might do something on another’s behalf, but this “behalf” doesn’t seem to be a thing or even an abstract object. These nouns are meaningful, but they do not seem to get their meaning by referring to anything at all.

Fourth, if we consider words other than nouns, they often fail to refer to anything at all: “very,” “of,” “a,” “yes,” and “alas.” These words don’t refer to anything, and yet they are meaningful.

If it’s correct that many words do not refer to things, this is a problem for the Referential Theory of Meaning.

Problem 2

Referential Theory treats a sentence as a list of names for things to which the words refer. But a list of names says nothing:

(3) Bob Jill Washington Phyllis

So how could we get meaning from a list of words that refer to things? There must be something else going on, too.

Problem 3

There is more to meaning than reference. Some words can refer to the same thing but not share the same meaning, for example “Joseph Ratzinger” and “the Pope.”

If not Referential Theory, then what?

There are other theories of meaning that surpass the Referential Theory, though they have their own problems. But before we discuss them, we will have more to say about how reference works. Even if reference cannot account for how we use language by itself, it is still an important part of how language works.4

  1. Here it will be obvious to some that I am imitating the opening of William Lycan’s superb Philosophy of Language: A Contemporary Introduction. The first many parts of this course will be following Lycan’s book closely. []
  2. This infinite regress was discovered by Bradley in 1930, Appearance and Reality, p. 17-18. For a further discussion of Bradley, see Loux, Metaphysics: A Contemporary Introduction, chapter 1. []
  3. These are two of Quine’s examples in Word and Object. []
  4. A modern, systematic attack on Referential Theory is found in Waismann, The Principles of Linguistic Philosophy, chapter 8. []

Previous post:

Next post:

{ 9 comments… read them below or add one }

MarkD May 2, 2010 at 10:34 am

Having worked in machine translation, we loved Quine’s Gavagai and “words shall be known by the company they keep” kinds of semantic holism over any kind of referential theory. Throw in some neuro-cognitive results concerning priming and semantic distance and you get “words shall be known by the company they keep and by their relative distance from one another in a hyperdimensional coreferential space.” But I’m racing ahead…

A special case of Loftus’ Problem of Miscommunication: translations are so subject to pragmatic and semantic interference that there is so sense that we can claim that they are ever accurate. I think I tried that with an analysis of “God is Light” years back (“so you think He is the electromagnetic spectrum, or just the visible part of it? Oh, so you mean by light some kind of moral goodness. Well, why not just say that?”) and had some traction in the argument.


Jeff H May 2, 2010 at 11:43 am

So (2) really means something like “Ralph bears the having relation to fatness.”

Uh-oh. Now it looks like we need to explain what “bears” refers to. And this could go on into infinity, and we’d never be able to work out the referential meaning of the sentence.

If I could nitpick a little bit here… I don’t like this line of reasoning, because you are trying to replace “is” with “bears the relation to”, and then pointing out that we don’t know what “bears” means. But really, it becomes a red herring – to settle on the meaning of words without being circular, you cannot keep substituting it with other words. If you knew nothing about language, a dictionary would not help much, because you’d be looking up definition after definition. Sooner or later (according to this referential theory anyway), you’d have to point to an object in order to equate a word with some other non-linguistic object. So you can point to a cat and say “cat” to equate the two.

I agree that the referential theory is incoherent once you get into abstract objects and other parts of speech, but not because there is an infinite regress, but simply because they don’t have a non-linguistic referent that one can equate with the word. It doesn’t matter that there are other words that could replace the word that also don’t have a referent. If I can’t point to something that is “grue”, it doesn’t matter at all if I say, “Well, ‘grue’ is also a synonym for ‘bleen’, and that also doesn’t have a referent!” That’s irrelevant in this case, because working out a theory of language by equating words with other words is not going to get you very far.

Anyway, I definitely look forward to when you get to Wittgenstein. His idea of language games is superb :)


Scott May 2, 2010 at 11:50 am

I just finished a class on Phil. of Language. We focused on Locke’s initial attempts, then Frege, Russell, and Kripke. My prof called Wittgenstein the “linguistic elephant in the room”, but never discussed it (he taught on LW the previous semester). Fascinating stuff.


John W. Loftus May 2, 2010 at 7:22 pm

Thanks Luke. It’s been awhile since I studied this but it’ll be interesting to hear from others. I could create a “private language” that only one other person understands. I can use different symbols for words and define them as I like using a syntax that is completely different than any other language. Words and sentences mean what I intend them to mean and that’s it. Although, to go deeper, I don’t even understand completely what I say because I don’t understand all of the implications of what I say, and sometimes I don’t even say what I thought I’ve said. The Problem of Miscommunication does indeed concern the problem of language, as Mark D pointed out.


alex December 6, 2010 at 11:08 am



Luke Muehlhauser December 6, 2010 at 12:45 pm


Read the first sentence here. You clearly don’t know what the word ‘plagiarism’ means.


kabir October 18, 2011 at 2:14 am

how we are identified as human being ? what do you mean by animal and rational?


KR January 14, 2012 at 6:27 pm

You’ve copied en bloc from Philosophy of Language by William Lycan, and only changed Hitler to Chaplin in an example. This is decidedly plaigarism. It’s not even a summary. The only thing you’ve put some effort into is changing the odd word. Pathetic!


Luke Muehlhauser January 14, 2012 at 7:47 pm

It’s not plagiarism when I *say* in the original post it’s from Lycan’s book.


Leave a Comment