Truth and Normativity

I am on the hunt for a deductive argument that is valid in point of logical form and that takes us from a premise set all of whose members are purely factual  to a categorically (as opposed to hypothetically or conditionally) normative conclusion.  Tully ( = Cicero?) the Commenter  offered an argument that I make explicit as follows:

1. It is snowing
2. For any proposition p, if p, then it is true that p.
Therefore
3. If it is snowing, then it is true that it is snowing. (2, UI)
Therefore
4. It is true that it is snowing. (1, 3 MP)
5. For any p, if p is true, then one ought to believe that p.
Therefore 
6. If it is true that it is snowing, then one ought to believe that it is snowing. (5, UI)
Therefore
7. One ought to believe that it is snowing. (4, 6 MP)

Does this argument do the trick?  Well, it is plainly valid.  I rigged it that way!  Is the conclusion categorically normative?  Yes indeed.  Are all of the premises purely factual?  Here is the rub.  (5) is a normative proposition.  And so the argument begs the question at line (5).  Indeed, if one antecedently accepts (5), one can spare oneself the rest of the pedantic rigmarole.

But I have a second objection.  Even if the move from 'is' to 'ought' internal to (5) is logically kosher, (5) is false.  (5) says that whatever is true is such that one ought to believe it.  But surely no finite agent stands under an obligation to believe every true proposition.  There are just too many of them. 

If one ought to do X, then (i) it is possible that one do X, and (ii) one is free both to do X and to refrain from doing X.  But it is not possible that I believe or accept every true proposition.  Therefore, it is not the case that I (or anyone) ought to believe every true proposition.  (One can of course question whether believings are voluntary doings under the control of the will, and (surprise!) one can question that questioning.  See my Against William Alston Against Doxastic Voluntarism.)

Still and all, truth does seem to be a normative notion.  (5) doesn't capture the notion.  What about:

5*. For any p, if p is true, then p ought to be believed by anyone who considers it.

The idea here is that, whether or not there are any finite minds on the scene, every true proposition qua true has the intrinsic deontic property of being such that it ought to be believed.   I say 'intrinsic' because true propositions have the deontic property in question whether or not they stand in relation to actual finite minds.

But of course plugging (5*) into the above argument does not diminish the argument's circularity.

Here is a possible view, and it may be what Tully is getting at.  Truth is indissolubly both factual and normative.  To say of a proposition that it is true is to describe how it stands in relation to reality: it represents a chunk of reality as it is.  But it is also to say that the proposition qua true functions as a norm relative to our belief states.  The truth is something we ought to pursue.  It is something we ought doxastically to align ourselves with.

This is murky, but if something like this is the case, then one can validly move from

p is true

to

p ought to be believed by anyone who considers it.

The move, however, would not be from a purely factual premise to a categorically normative conclusion. My demand for a valid instance of such a move might be rejected as an impossible demand.  I might be told  that there are no purely factual premises and that if, per impossible, there were some, then of course nothing normative could be extracted from them.

What say you, Tully?

Related articles

Truth and Truthfulness

From 'Is' to 'Ought'? Help for Hodges

Comments

[Tully Borland]

Great stuff, Maverick—both this post and the previous one. I must admit that I’m shooting from the hip here and appreciate your patient indulgence. I do find this subject very interesting even if I’m certainly no expert in this area.

My demand for a valid instance of such a move might be rejected as an impossible demand. I might be told that there are no purely factual premises and that if, per impossible, there were some, then of course nothing normative could be extracted from them.

This was actually my first inclination in reading your first couple posts on the issue. If we can’t allow any explicit normative words (or concepts) in the categorical premises, and we have to use all categorical statements (no “if fact, then ought” bridge principles), and we can’t have an assumed normative premise such that the argument is an enthymeme, and we’re working with (what you refer to as) syntactical validity, then the project seems obviously doomed from the start. At least it did to me when I first came across the challenge. That’s why some of my arguments haven’t been syntactically valid. So if there is a way to get from factual premises to an explicitly normative conclusion I think we’re going to have to adopt some other understanding of validity than syntactical validity. Consider something like (D3) or (D4) hereby adopted.

But if the content of the premises are purely factual (and we can’t have any assumed premises the content of which is normative), then what are we left with to get normativity into the conclusion? Well, there is (a) the truth of the premises, (b) the nature of the premises (assertions purporting to be evidence), and (c) the process of inferential reasoning itself. This latter notion was what I was invoking in my last comment that you didn’t understand. So please allow me start here again, and then I’ll return to (a) and your most recent post in a separate comment when I get the chance. Even if what I say about (c) fails, it will hopefully be somewhat more intelligible (if it is intelligible!).

In making an inference, we take for granted in the inferential process itself that we’re engaged in that the conclusion ought to be believed or at least ought to be given greater or lesser credence given the premises. When I assert “p, thus ~p” I take for granted the proposition that "~p should be given less credence than p" (in this case, zero credence). But what I’m suggesting is that, perhaps we should think that engaging in the inferential process itself provides the justification for the explicit normativity in the conclusion.
1. P
2. Thus, ~p.
3. Thus, ~p should be given less credence.
I’m suggesting that the inferential process itself justifies the further move to 3, and I’m also suggesting that there’s no normative premise being smuggled in. (Maybe this won’t work but it’s worth a try). Here is an analogy to motivate that latter claim.
When I reason inferentially using modus ponens (when I engage in the process), in order for that process to be the truth preserving process that it is, the following proposition must be true (if p & (if p then q), then necessarily q). But that proposition is not an implied premise in the argument. I’m justified in believing the conclusion because I’m justified in believing the premises and the inferential process is a truth preserving one. What I’m suggesting is that, similarly, perhaps it’s the case that I’m justified in believing (3) because I’m justified in believing (1) and (2), and by engaging in the process of inferential reasoning itself, I’m further justified in believing 3.

[Bill Vallicella]

>>When I reason inferentially using modus ponens (when I engage in the process), in order for that process to be the truth preserving process that it is, the following proposition must be true (if p & (if p then q), then necessarily q).<<

This formulation suggests that you are confusing the necessity of the consequence (necessitas consequentiae) with the necessity of the consequent (necessitas consequentiis). What you ought to say is that the following proposition is true:

1. Necessarily [(p & (p --> q))--> q]

as opposed to

2. (p & (p --> q)) --> necessarily q

The latter has plenty of counterexamples.

I take it that you are saying that an argument having the form of modus ponendo ponens is not an enthymeme with (1) as suppressed premise. That seems right.

Since I don't understand your move from p to ~p, let me give a different example:

Your wife likes flowers
Therefore
You ought to give her some.

Your point, I take it, is that this argument is not an enthymeme. It is valid as it stands. There is no need for some such auxiliary premise as:

If a person P likes X, then, ceteris paribus, one ought to give P X.

I'll have to think about this.

[Peter Lupu]

Bill,

What about this example.

Let A stand for any valid argument with P and Q as true premises and R as a conclusion. x is a rational person who is an expert in logical matters.

1. x carefully inspected A and knows that A is a valid argument.
2. x knows that P and Q are true.
3. x knows that R logically follows from P and Q.
4. x is a rational person who always accepts propositions that logically follow from true premises.

Therefore,

5. x ought to accept R as true.

Of course, you will balk at premise 4 on grounds that 'rational' is already a normative concept and therefore it imports a normative element into the premises. You may be right, but perhaps it is worth discussing it anyway.

[Bill Vallicella]

You might, however, reject my flower example, seeing it as an illict *is* to *ought* slide, while holdig that concepts like truth are relevantly different from the concept like liking.

[Tully]

The previous comment was a desperate attempt to try to make good on your challenge according to your rules. I now return to your most recent post and to one of my earlier suggestions that it may be the case that the idea of a purely factual (true) premise is problematic.

BV: Truth is indissolubly both factual and normative. To say of a proposition that it is true is to describe how it stands in relation to reality: it represents a chunk of reality as it is. But it is also to say that the proposition qua true functions as a norm relative to our belief states. The truth is something we ought to pursue. It is something we ought doxastically to align ourselves with.

I agree that truth is indissolubly normative, though I wouldn't know how to argue for it (any argument I would give would implicitly beg the question since I'm assuming the premises should be valued as true or likely to be true in serving as evidence). I'd further add that if a premise is true it is right, and rightness is clearly a normative concept. So regardless of whether I ought to believe what is true, I'm already employing a normative concept in any argument the premises of which I take to be true (even if the content of those premises contains no normative concepts).

And maybe premises themselves (even false ones) are indissolubly normative as well. A premise, I think, by its nature, is a statement intended to serve as evidence for a conclusion. As such, a given premise ought to be understood as purported evidence for the conclusion. If someone does NOT already believe that a premise ought to be understood as purported evidence for a conclusion, then any argument from a "purely" factual premise to a normative conclusion is bound to fail (since one would not understand how the premise is supposed to function in relation to the conclusion).
If you don't already believe in some sort of normativity, it's hard to see how you can rationally believe that you can rationally arrive at the belief in normativity by way of arguments. Rationality itself is a normative concept. Understanding is also a normative concept. If I don't think I'm understanding a premise correctly, then I'll have no reason to think that the premise is good evidence for the conclusion.
Hence anyone who doesn't believe in any normativity is irrational if he thinks he can arrive at the belief rationally by means of argumentation.

What I just said here might seem to conflict with my previous attempt in the comment above to try to arrive at an explicitly normative conclusion from premises the content of which is purely factual. In this comment I've assumed that one has to believe in some normative concepts if one rationally thinks one can rationally arrive at conclusions. Still, perhaps the above argument works as long as those implicit normative beliefs I think we need aren't serving as implied premises in the argument. (If all my background beliefs count as implied premises then that argument would seem to fail).

I hope some of this makes sense, but I've run out of time and will have to stop there.

[Tully]

Bill,

Yes, the necessity should go before the entire conditional. Slip of the brain.

Another oops: In that first argument I meant to say
1. p
2. It's false that ~p.
No wonder you didn't understand it!

Regarding enthymemes, I'm not clear myself on what counts as an implied premise. Like I said in that long post a second ago (before I saw your remarks), if my background beliefs count as implied premises then I'm going to have a WHOLE lot of arguments which are enthymemes.

Regarding rationality, it sure seems like a normative concept.

[Bill Vallicella]

Good example, Peter. Yes, I will balk at (4).

Here is the puzzle. Suppose I convince you that two of your beliefs are contradictory and so cannot both be true. Then I say: you ought to remove this contradiction somehow, either by rejecting one of the beliefs or showing that the inconsistency is merely apparent, or in some other way. How did the oughtness come into the picture? Why must you remove the contradiction?

[Peter Lupu]

Indeed!

The question which I would like to explore is this. Let us assume that no logically valid argument of the form is-ought is forthcoming. Thus, every argument of this form is invalid. What does this show about the fact/value distinction and about naturalism regarding morality? Can we learn anything substantive about morality and its place in the natural world in which we are bound to reside (at least for a while) from the invalidity of is-ought arguments? Can systematic invalidity teach us anything about the nature of the subject matter involved in such systematically invalid arguments?