Direct and Indirect Reference

London Ed asks:

Exactly what does ‘refer’ mean?  And when we talk about ‘direct reference’ and ‘indirect reference’, are we really talking about exactly the same relation, or only the same in name?

The second question got me thinking. 

The paradigms of direct reference are the indexicals and the demonstratives.  The letter 'I' is not the word 'I,' and the word 'I' -- the first-person singular pronoun -- has non-indexical uses.  But let's consider a standard indexical use of this pronoun.   Tom says to Tina, "I'm hungry."  Tom refers to himself directly using 'I.'  That means: Tom refers to himself, but not via a description that he uniquely satisfies.  The reference is not routed through a reference-mediating sense.  If you think it is so routed, tell me what the reference-mediating sense of your  indexical uses of the FPS pronoun is. 

As I understand it, to say of a singular term that it is directly referential is not to say that it lacks sense, but that it lacks a reference-mediating sense.  The indexical 'now'  does have a sense in that whatever it picks out must be a time, indeed, a time that is present.  But this very general sense does not make a use of 'now' refer to the precise time to which it refers.  So 'now' is directly referential despite its having a sense.  

Consider the demonstrative 'this.'  Pointing to a poker, I say 'This is hot.'  You agree and say 'This is hot!'  We say the same thing.  The same thing we say is the proposition.  The proposition is true.  Neither the poker nor its degree of heat are true.  The reference of 'this' is direct.  It seems to follow that the poker itself is a constituent of the proposition that is before both of our minds and that we agree is true.  But then propositions are Russellian as opposed to Fregean.  The poker itself, not an abstract surrogate such as a Fregean sense, is a constituent of the proposition. 

To say that a singular term t indirectly refers to object o is to say two things.  (i) It is to say that there is a description D(t) that gives the sense of t, a description which is such that anything that satisfies it uniquely satisfies it. (ii) And it is to say that O uniquely satisfies D(t).

Note that for the indirect reference relation to hold there needn't be any real-world connection such as a causal connection between one's use of t and o.  It is just a matter of whether or not o uniquely satisfies the description encapsulated by t.  Satisfaction is a 'logical' relation.  It is like the 'falling-under' relation.  Ed falls under the concept Londoner.  The relation of falling-under is not 'real': it is not causal or spatial or temporal or a physical part-whole relation.

Indirect reference is just unique satisfaction by an item of a description encapsulated in a term.  If 'Socrates' refers indirectly, then it refers to whatever satisfies some such  definite description as 'the teacher of Plato.'  Direct reference, on the other hand, has nothing to do with satisfaction of a description.

So I think London Ed is on to something.  When we talk about ‘direct reference’ and ‘indirect reference’, we are not  talking about exactly the same relation. The two phrases have only a word in common, 'reference.'  If all reference is indirect, then direct reference is not reference. And if all reference is direct, then indirect reference is not reference.  There are not two kinds of reference. 

The reason, again, is that indirect reference is just unique satisfaction of a description whereas direct reference has nothing to do with satisfaction of a description.

Comments

[David Brightly]

I think I see what you are getting at here, Bill, that 'direct reference' and 'indirect reference' seem rather different kettles of fish. But why say that there are not two kinds of reference? It's like saying that 'direct descendance' (being a child of) and 'indirect descendance' (having at least one intermediate ancestor) don't 'fit together' to make 'descendance in general'.

[BV]

Hi David,

I think your case is quite different. I am a descendant of both my parents and my grandparents.

But there aren't two kinds of ducks, decoy and real, or two kinds of leather, artificial and real. I take Ed's suggestion to be that indirect and direct reference are not two different kinds of reference.

[London Ed]

Correct. (Although David has a point too). We are in one of those rare moments of mutual agreement.

The background to this is my puzzle about essential (or 'internal') relations versus accidental (or ‘external’) relations. I think the direct referentialists want to characterise Russellian thoughts as somehow internally related to their object. A thought like ‘I am hungry’ is somehow essentially related to the person who thinks it: we can’t have the thought without the object it is about.

But ‘whoever discovered the elliptical motion of the planets was German’ isn’t essentially related to the discoverer in this way. The relation of satisfaction is accidental or ‘external’.

A minor quibble: Frege also held that the reference relation was essential. He says that the word ‘common name’ is confusing, because it suggests that the common name stands in the same relation to the objects that satisfy it as the proper name does to a single object. But if so, he says, a common name that nothing falls under (e.g. ‘unicorn’) would be as illegitimate as a proper name that designates nothing, which is false. The word ‘planet’ has no direct relation to the Earth, but only in indirect one. “If I utter a sentence with the grammatical subject ‘all men’, I do not wish to say something about some Central African chief wholly unknown to me”.

This suggests that he believed the common name/falling under relation was accidental, whereas proper name / designation was essential.

Note also that Frege’s word translated as ‘reference’ is bedeuten, from ‘deuten’ which I think means point out or ‘signify’.

[BV]

There are many difficult issues here.

The definite description 'the man who discovered the elliptical orbits of the planets' is satisfied by Kepler. But it might have been satisfied by some other person. So I agree that there is a clear sense in which the relation of satisfaction in this case is accidental or external.

But let 'A' be a rigid designator of our world, the possible world that happens to be actual. 'The man who discovered the elliptical orbits of the planets in A' is satisfied by Kepler, but could not have been satisfied by any other person. So in this case the relation of satisfaction is not accidental.

[London Ed]

>>But let 'A' be a rigid designator of our world, the possible world that happens to be actual. 'The man who discovered the elliptical orbits of the planets in A' is satisfied by Kepler, but could not have been satisfied by any other person. So in this case the relation of satisfaction is not accidental.
<<
Ooh that's tricky. Someone else must have already made that point, surely?

[BV]

Why, because I'm not smart enough? Truth be told, I got the idea from Plantinga who said something along these lines years and years ago. May have been in his article "The Boethian Compromise."

What I am suggesting above is that the issue is not whether satisfaction is accidental or essential, but whether, when one thinks about x, x itself is a constituent of one's thought or a representative of x.

Right now I am thinking about Ed. Surely the man himself with all his parts and properties is not before my mind. I am thinking of him under various descriptions which do not come close to exhausting all his properties, monadic and relational. This suggests that what is before my mind is something like a Fregean sense (accompanied by some memory images perhaps, but not necessarily). Of course, my thinking reference is not to the sense or to the images but to the man himself, a concrete chunk of the external world. So I say that in a case like this my reference to Ed is mediated, not direct: it is routed through sense.

Unfortunately, when I think about myself --as when I think: I am happy! -- the above model breaks down. I don't think of myself via a description but directly.

[london ed]

>>Why, because I'm not smart enough?
No no no. It just sounded like the kind of thing that someone must have addressed at some point. You are of course very smart.

[london ed]

I wonder if there are three types of relation to consider.

(1) The relation between purely descriptive thoughts and what falls under them. The odd Boethian case aside, I think we all agree on this. Purely accidental.

(2) Between first person thoughts and their object where the object (the thinker) is as it were contained in the thought. This is the relation of whole to part, and if anything is candidate for an essential relation, this is it.

(3) Between a 'demonstrative' thought and its object. Suppose I point to the Morning Star (Venus) and say 'that planet is very hot'. This surely cannot be the relation of whole to part. How can Venus itself, a massive celestial body weighing trillions of tons, with an incandescent a
nd sulphurous toxic atmosphere that is poisonous to all life, be part of a thought? But if it isn't, how can the relation be essential?

[Peter Lupu]

Bill ,Ed, David, etc.,

I am confused by this thread.

Reference is a relation between a referential item (linguistic or mental) and a putative object. So in both cases of direct and indirect reference the goal is to fix the reference between a referential item (e.g., name) and an object. The question is what exactly mediates the relationship and what is the nature of the mediators. All reference is mediated. The mediators (whatever they are) are the means that determine reference. The difference between direct reference vs. indirect reference theories is about the nature of the mediators and whether the mediators constitute the meaning of the referential item in addition to fixing its reference, if any; the difference is not about what sort of a relation is determined by the mediator (i.e., reference). In direct reference theories the mediators are initial baptism (or something analogous) and causal transmission. Expressions which qualify as rigid designators do have mediators, but lack meaning (not even the mediators, if any, are the meaning of such expressions). In the case of indirect reference, meaning fixes reference and the mediators are descriptions.

Viewed in this light, what is Ed's problem again?

[BV]

Peter,

I think you are using 'mediate' in a broader way than we are.

Suppose I say to you 'This poker is hot.' I am using the demonstrative 'this' to denote or designate a red hot poker in the fireplace.

Now one question is this: Is the reference in a case like this achieved or constituted by the referent's satisfaction of a definite description? Presumably not. Then the reference is not mediated by that description or routed through the sense of that description. If so, then contrary to what you say, not all reference is mediated.

Now suppose there is mediated reference and immediate reference. Are they two different kinds of reference or is only one of them genuine reference?

Ed and I appear to agree they they are not two different kinds of reference any more than artificial and real are two different kinds of leather.

[David Brightly]

Ed,

(2) Between first person thoughts and their object where the object (the thinker) is as it were contained in the thought.

Isn't the thought contained in the thinker? Ergo, ...

[Peter Lupu]

Bill,

I use 'mediator' in a broad sense: i.e., any manner involved in fixing the reference of a term. Now, in your example involving 'this' there is a mediator; e.g., Kaplanian character could be counted as one.

My suggestion is that, regardless of the nature of the mediator, the mediator's purpose is to fix a referential relation between a referential item and an object. Therefore, the relation of reference is the same, even when the manner in which the relation is fixed changes from case to case. So I do not think that there are two (or more) kinds of reference relations or that so-called "indirect reference" is really not reference. I think that we have one relation of reference with different ways of fixing this relation.

[Quinn Truckenbrod]

I have always been confused by discussions about reference. It seems to me that reference is a kind of speech act. Words do not refer to things. Human beings use words to refer to things. It is a kind of action, not a relation. Every act of reference occurs at a certain time and place, just as every promise and every command happens at a certain time and place. The words 'Barack Obama' on their own do not refer to anything unless and until some person uses them in a speech act to refer to something. And a person may use those words to refer to the current president of the United States or to a house cat he may have given the name to.

[London Ed]

Peter: The difference between direct reference vs. indirect reference theories is about the nature of the mediators
<<

As Bill has pointed out, direct reference is not mediated, by definition. Direct reference = unmediated reference.

David >>Isn't the thought contained in the thinker? Ergo, ...

Happy Christmas David.

[BV]

Quinn,

You make excellent points that people need to bear in mind. I agree with everything you say except >>It is a kind of action, not a relation.<<

Referring is a speech act, hence a kind of action. We do it using words. Not just mouthing words, of course, but using words thoughtfully. (Thinking reference underpins and is logically prior to linguistic reference.) But when we refer to things (successfully) we set up a relation of reference.

So while referring is an action, we can still speak of the reference relation, the relation between a word used thoughtfully by a person and something else to which the person refers.

And then one can ask a question like this: Is the reference relation assimilable to the the satisfaction relation or not?

[BV]

Peter,

You are not engaging the precise question that is on the table. No good purpose is served by taking a word that is being used narrowly, and has been defined, and then using it broadly.

[Peter Lupu]

Ed,

Amazing!
Every direct reference theorist I know off talks about reference being mediated (Kripke, Kaplan, etc.) First, according to Kripke, the reference of names is fixed by a baptism that may include demonstratives, descriptions, etc. Second, then reference is transmitted by way of a causal chain, which may involve descriptions, but does not have to. Kripke often speaks about the "intention" to refer to whatever object the name is originally used to refer. In fact, direct theorists do not even object to having descriptions used to pick out the reference on a given occasion. They do, however, object, to (1) meaning fixes reference; and (2) descriptions associated with a name fix the reference across possible worlds. So I am not sure what you guys are talking about. Which direct reference theorists you have in mind?

[BV]

Ed,

I agree that the three cases you mention are different.

As for (3), Venus itself, that massive chunk of physical reality, cannot be in my head. Nor can it itself be 'in' or 'before' my mind. Venus is infinitely propertied; nothing before our minds is infinitely propertied. So there is a clear sense in which Venus itself cannot be a part of a thought. But that is not to say that Venus cannot be a constituent of a singular Russellian proposition.
It is just that these Russellian propositions cannot thoughts that we can think. (God could think them, perhaps.)

Pointing at Venus, you say 'That is hot!' You thereby express a demonstrative thought. We agree that Venus itself cannot be a constituent of that thought. So what is the constituent of the thought? It has to be some surrogate item, something like a Fregean sense. Perhaps: the bright object in my visual field. But that is an incomplete object and you, of all people, don't want to say that.

You've got yourself a nice little puzzle.

Suppose I think: Ed is sleeping. You yourself with all your parts and properties are not in my head or in my mind or in my thought. So what is the subject constituent of my thought? Perhaps an incomplete object encoding such properties as being male, being in his 50s, being a philosopher of language, being a Londoner and a few others which, taken in conjunction, apply to you and you alone.

But then how does this incomplete object connect with you?

[London Ed]

>>Peter:"Every direct reference theorist I know off talks about reference being mediated (Kripke, Kaplan, etc.) "

This paper by Schiffer mentions Salmon, Soames and Braun. He characterises their view as: propositions are Russellian or structured propositions, whose components are objects and properties, and singular terms are typically directly-referential, i.e. such that their only contribution to the proposition expressed is their
referent. I.e. unmediated by a ‘sense’.

[London Ed]

Peter >>I use 'mediator' in a broad sense: i.e., any manner involved in fixing the reference of a term. Now, in your example involving 'this' there is a mediator; e.g., Kaplanian character could be counted as one. <<

Ah right, I see how you are using the term 'mediated'. In which case I agree with Bill: "No good purpose is served by taking a word that is being used narrowly, and has been defined, and then using it broadly."

[London Ed]

As for Kaplan, he says that that demonstratives are “terms which “refer directly without the mediation of a Fregean Sinn [or individual concept, set of properties, etc.] as meaning.” ( “Demonstratives”. Draft #2, Department of Philosophy, UCLA, 1977, p. 1.) My emphasis.

[BV]

Ed,

Peter may just be playing with us, trying to get our goat, or even trying to undermine the precious sliver of agreement that we seem to have achieved! [grin]

You need to tell me whether you like this formulation: direct and indirect reference are no more species of a common genus than are artificial and real leather.

I take it that that is the idea you had, put in my own way. It is worth exploring. There might even be the makings of a jointly-authored article here, although I fear that it working out the idea we will diverge irreconcilably . . . .

[London Ed]

>>You need to tell me whether you like this formulation: direct and indirect reference are no more species of a common genus than are artificial and real leather.

I agree with you (if we take 'direct reference' and 'indirect reference' as they are commonly understood).

On working out the ideas further, we haven't yet discussed the connection with essential and accidental relations.

I believe the direct referentialist holds that with indirect reference (such as with definite descriptions) there is an essential relation between the term (either a linguistic term or a mental term) and a Fregean sense, or meaning, or content, but an accidental relation to the object that satisfies the reference. But with direct reference, the essential relation is between the term and the object itself.

In this case, the relation cannot be the same, for in the descriptive case it is a combination of an essential and an accidental relation, in the Millian case it is essential only.

Some direct referentialists (like Kaplan, probably) hold that the object is part of a structured proposition or thought. Others balk at this (how could Venus be part of a thought?) and hold that the existence and identity of the thought depend on the existence of the object.

Bill, surely you are a direct referentialist of the latter type, for do you not hold that the object of a thought is an intentional object of some kind? (I may have mistaken you).

[David Brightly]

Ed,

In this case, the relation cannot be the same, for in the descriptive case it is a combination of an essential and an accidental relation, in the Millian case it is essential only.

Are you sure there is a fact of the matter here? After all, the relation less-than, defined over the natural numbers appears essential, but defined over, say, populations of cities, it seems accidental. More generally, how do we define 'sameness' for relations?