The Philosopher's Studio

The Normative (In Search of a Stalking Horse)

Normativity, as I ceaselessly tell my students, is a fundamental philosophical issue, perhaps the most fundamental.  It would appear central to our very characterization of ethics and epistemology, and would seem to  pervade virtually every other domain on the back of questions of meaning and truth.  However, as my students well know, normativity is a notion that is far more pervasive in its use rather than the subject of discussion itself.  This is particularly apparent in the domain of practical reasoning theory.  All of the authors we are consulting in our course on practical reasoning use normativity in their disputes, generally to claim that one's own theory, but not one's opponent's, can account for the normativity of practical reasoning.   But besides what amounts to mere gestures as to what normativity is or involves-- it 'guides', 'requires', provides a 'standard of correctness' or the virturally vacuous 'tells us what we ought to do'--nothing substantive is said. 

There is nothing peculiar about practical reasoning theory on this count.  To be told that epistemology is normative because it is in the business of justification is not particularly informative either.  Nevertheless, that epistemology is normative is supposed to matter and matter a great deal: it has been used to rule out certain construal or interpretations of the subject itself.  Dispositional theories of meaning, to use another example, whatever virtues (or other vices) they may have, are rejected my many as being unable to account for the inherent normativity of meaning.  But apart from the claim that a theory of meaning that does not account for its normativity is a theory that has nothing to say about what we ought to mean, hardly anything is said.

My students, and, I must say, myself, find this "disturbing."  Particularly distressing to them was their sad realization, while writing their essays on the question of whether instrumental reasoning is normative, that neither the Stanford nor Internet Encyclopedia of Philosophy has an entry on this most crucial notion (though the former has an entry on 'nothingness').  Nor, for that matter, does the rather compendious Cambridge Dictionary of Philosophy (Audi).  How could a notion not merely so ubiquitous in philosophical discourse, but apparently so significant, a notion that is believed to actually cut some heavy philosophical ice, go without more focused discussion?  I would like to know what some people think about the absence of such discussion (and why there are no entries in the aforementioned sources), but I would also like to here some proposals of just what people take the nature of normativity to be (apart from being 'related to reasons'!).  Without a fixed notion of the normative, even a loose, rough, and dirty one, disputes over satisfying normative constraints would seem interminable.  Moreover, without some such notion, being a normative nihilist is not terribly intelligible either.  Getting clear(er) on normativity would seem useful for realists and nihilists alike.   

Logic and Belief

The normative nihilist claims that there are no ahistorical, non-artificial, non-contingent (you get the idea) standards of correctness against which what actually happens can be measured.   One consequence of this view, I believe, is that logic cannot be understood as anything other than 'laws of thought' in the purely descriptive sense in which Frege hoped to reject.  This alone, no doubt, makes such nihilism unacceptable to many, no matter what other unattractive consequences it may have (such as the consequences for ethics and epistemology once we deny any substance to the notion of justification).  Be that as it may, I want to flesh this idea out a little in the hopes of seeing just how tenable (or untenable) such a view is.

To understand logic in terms of the laws of thought, I think, is to say something very close to the claim that valid inference patterns are exhibited by believers as a matter of psychological necessity.  Consider modus ponens.  Someone who believes modus ponens is a correct pattern of reasoning will claim that if a believer believes p, believes if p then q, then that believer ought to believe q.  Such an understanding of modus ponens is normative in that it implies that it is possible to fail to believe q (when one believes the premises).  The nihilist, in contrast, must deny that such failure is possible.  She has to claim that modus ponens is a description: a person who believes p, believes if p then q, believes q.  If they don't (believe q), then they simply cannot be said to believe one or the other (or both) of the 'premises'.

I have a worry (!) about this, but I first want to gesture at a reason for thinking that this idea is not entirely insane.  The idea that logic concerns the 'laws of truth' gets its bite from the intuitive idea that logic maps out the structure of reality.   A silly example might help to illustrate this point.  Consider the following argument:

1. All pigs fly.
2. Salmon are pigs.
__________________
3. Salmon fly.

This argument is valid: the truth of the conclusion is necessitated, or guaranteed, by the truth of the premises, assuming they are true.  In what sense is the (truth of) the conclusion necessitated? It appears that it is, or would be, in some sense 'metaphysical': that salmon fly does not 'follow from' the truth of pigs flying and salmon being pigs, rather salmon flying is (part of) what it is for pigs to be flying and salmon to be pigs.  That salmon fly is not some further fact that could conceivably fail to obtain if the world ceased to exist at an inopportune moment, the moment just before the truth of pigs flying and salmon being fished passed on to salmon flying.  These three facts are, in a certain sense, not three distinct facts at all; the obtainment of the first two just is (in part) the obtainment of the third. 

Much of this is impressionistic, and I'm not sure how to make it more precise at the moment.  However, I think it will suffice, if we are sufficiently charitable, for seeing why the nihilist thinks that logic might (also?) be descriptive of belief.  Given what we have just said about the world of flying salmon-pigs, its hard to imagine how a believer could be thought to believe that both pigs fly and that salmon are pigs and yet not believe that salmon fly.  If he doesn't believe this last, we are surely going to want to know what he thinks he is believing when he says he believes the premises.  I think that it is plausible to say that the presumption will be that he doesn't think he is believing what we think he is believing (or, perhaps, what we think we would be believing if we were to use those sentences to sincerely assert what we believed).

Now it will have been observed that this example is not one of modus ponens.  Does the introduction of the conditional form make a difference?  If a person believes that salmon are pigs, and that if salmon are pigs then salmon fly, is it possible for them not to believe that salmon fly?  I confess to finding it very hard to see that it is, but I have found that others are often more reticent.  Here, though, is a useful place to enter my worry.  Reasoning, whether in this case of modus ponens or the categorical syllogism above, is temporal.  Whereas it makes no sense to say that if the world were to go out of existence it might be the case that the conclusion of a valid argument would fail to obtain, it does not seem so senseless to say that in that circumstance it might be the case that a person might fail to draw the conclusion of a valid argument.   Moreover, whereas there is (so I think) an intuitive case to be made for not (completely) distinguishing 'the facts' (of the syllogism, for example), the intuitions are just as strong, if not stronger, that we need to distinguish the belief-content: thinking pigs fly is just not the same thing as thinking salmon fly, even if you think salmon are pigs.   Does this threaten the nihilist's view in any substantive way?  Are there deep issues between claiming that one, in some sense, already believes the conclusion of a valid inference pattern if one believes the premises and claiming that one cannot help but draw the conclusion of such a pattern if one believes the premises, provided one has the opportunity to do so?

This is just scratching the surface of this idea.  Much more needs to be said, particularly about other valid patterns, especially more complex ones (and not only complex ones: what about 'or-introduction'?).  I hope to say something, eventually, about these, but what we have at the moment is enough of the idea to take some shots at it, so I will leave it here for now.      

The Normative and the Natural

In the introduction to his beautiful yet tragically underappreciated Logic and the Basis of Ethics, Arthur Prior makes the following remark:

I am inclined to think that almost all that can be said, from a purely logical point of view, on the issue between naturalism and anti-naturalism, has already been said in two quite brief sections in Hume's Treatise of Human Nature (II. iii. 3, and III. i. 1) and in one quite brief chapter in Reid's Essays on the Active Powers (V. vii).  At all events, a thorough mastery of these three items would provide anyone with a complete set of tools for cutting away the thick growths of sophistry which seem in all periods to thrive on the soil of Moral Philosophy.

I am inclined to agree (though I find the derogatory use of 'sophistry' unfortunate!), and have been trying, in a variety of ways, to make a case for Prior's view ever since I first read LBE in graduate school.  At present, however, I want to follow Prior's suggestion and take a closer look at one of the more oft-discussed passages in the second Hume section referred to above: the famous remarks about 'is' and 'ought'.

It is, perhaps, thought by many that here we find expressed what has subsequently become known as 'Hume's Law', declaring the impossibility of deriving normative claims from descriptive ones, 'oughts' from 'iss'.  Presumably, those who think this have the following remark of Hume's in mind, namely

    ...a reason should be given, for what seems altogether inconceivable, how this new relation can
    be a deduction from others, which are entirely different from it.

This remark, however, is not a declaration of what is possible or not, but rather a challenge, or, better, an expression of incredulity.  In any event, full appreciation of Hume's stance on the matter is impossible, I think, unless we pay close attention to what he says before this remark.  The relevant passage in full:

In every system of morality, which I have hitherto met with, I have always remark'd, that the author proceeds for some time in the ordinary way of reasoning, and establishes the being of a God, or makes observations concerning human affairs; when of a sudden I am surpriz'd to find, that instead of the usual copulations of propositions, is, and is not, I meet with no proposition that is not connected with an ought, or ought not. This change is imperceptible; but is, however, of the last consequence. For as this ought, or ought not, expresses some new relation or affirmation, 'tis necessary that it shou'd be observ'd and explain'd; and at the same time that a reason should be given, for what seems altogether inconceivable, how this new relation can be a deduction from others, which are entirely different from it.

Hume, it is clear, has two challenges for those who traffic in oughts: First, assuming that ought is different from is, what does it mean?  What function does it perform in the statements in which it figures?  Second, assuming, again, that it is different from is, how does one deduce, or derive such statements from the more familiar is statements? It is this second claim that those who talk of Hume's law are focused on.  But what seems more fundamental is that Hume is expressing, whether feigned or sincere, ignorance of the significance of ought. 

If we take Hume seriously here, as I am inclined to do, his position is nihilistic about normativity: there simply is no other thing than the natural, descriptive world; nothing that ought picks out that is does not.  Nothing, at least, that he--or we--have any hope of understanding.   

One can, of course, disagree with Hume on these matters.  Reid, for one, did.  In the chapter referred to by Prior we find the following:

We may here observe, that it is acknowledged, that the words ought and ought not express some relation or affirmation; but a relation or affirmation which Mr. Hume thought inexplicable, or, at least, inconsistent with his system of morals.  He must, therefore, have thought, that they ought not to be used in treating of that subject.  He likewise makes two demands, and, taking it for granted that they cannot be satisfied, is persuaded, that an attention to this is sufficient to subvert all the vulgar systems of morals.  The first demand is, that ought and ought not be explained.   To a man that understands English, there are surely no words that require explanation less.   Are not all men taught, from their early years, that they ought not to lie, nor steal, nor swear falsely?  But Mr. Hume thinks, that men never understood what these precepts mean, or rather that they are unintelligible.  If this be so, I think indeed it will follow, that all the vulgar systems of morals are subverted.  Dr. Johnson, in his Dictionary, explains the word ought to signify, being obliged by duty; and I know no better explication that can be given of it. 

His second demand is, that a reason should be given why this relation should be a deduction from others which are entirely different from it. This is to demand a reason for what does not exist. The first principles of morals are not deductions. They are self-evident; and their truth, like that of other axioms, is perceived without reasoning or deduction. And moral truths that are not self-evident, are deduced, not from relations quite different from them, but from the first principles of morals

This is what we would refer to today as 'non-naturalism'.  Make of it what you will.  What I want to suggest here, though, is that is rather hard to see what other position on normativity one might take.  Specifically, I think there is here, in the combination of Hume's two challenges, an argument against the very cogency of 'naturalistic realism' about normativity.  If one is a naturalist, then, presumably, one thinks that ought is not different from is. But then it becomes hard to understand just what such a theorist does think normativity amounts to.  Furthermore, if ought statements are indeed translatable without remainder into is statements, why would it be 'altogether inconceivable' that they be deduced from other is statements?  True, not all such statement follow validly from all others, but some do.  'Hume's Law', however, isn't supposed to have exceptions.  What is open to the Humean, however, is to claim that the relevant class of is statement is actually about one's preferences, or motivations, and these are not deducible from descriptions of the various states of the world at which they might be directed.  But this is not something that a naturalist who also fancies himself a 'cognitivist' or a 'realist' will be comfortable with.  But what, then, is such a naturalist left with?  What does such a view actually involve?  I don't see how to make sense of it myself.  Hence I think that you either take normativity seriously and are therefore a non-naturalist, or you are a naturalist and you don't. 

 

What are naturalistic realists about morality thinking?

Many philosophers find non-naturalism a very odd doctrine, and indeed it is.  Whether its the version of Plato, Reid, Moore, Prichard, or Dancy, the idea that there are truths (knowable or not) about value or what ought to be that are in principle inaccessible by the senses strikes many as downright 'queer'.  Just what is the nature of these truths and, more importantly (assuming it is important), how might one try to know them?   The answers non-naturalists have given to such questions, moreover, tend not to showcase their philosophical talents to their best advantage.  The pre-existence of the soul, faculties of intuition, and evaluative sensibilities seem more like desperate hand-waiving than cogent epistemic theses.   

Nevertheless, I have always found the dogged tenacity of non-naturalists impressive: they see their path as the only conceivable way of saving morality (or normativity more generally).  For them it is either their way or no way.  Perhaps I respect them because I tend to agree: if there is nothing more than the natural world of the senses then there is no such thing as normativity (and, a fortiori, morality).  As a naturalist, therefore, I am a normative nihilist.  But this last statement of mine, one that both non-naturalists and I would take to be logically valid inference, is one many a 'naturalist' would reject. 

There are, apparently, a class of naturalists who want to claim that what is, say, good, or right, is something capable of expression in descriptive terms.  Perhaps utilitarians, or consequentialists more generally, belong to this class (I would think 'response-dependence' theorists, and most expressivists  belong here as well).  So, it might be claimed, goodness is pleasure, or utility, or some such perfectly natural thing (it really doesn't matter what).  In itself, this is unobjectionable, at least not on naturalist grounds per se (though naturalists can surely quibble amongst themselves on how to most effectively construe value, particularly whether it makes much sense to distance it from the individual psychologies of people, and thereby giving it an air of objectivity).  The problem is when such philosophers go on from here to claim that we ought to maximize pleasure, i.e. that it is right to do so, where that means something more than that maximizing pleasure would be good, i.e. be pleasurable.  For indeed such philosophers do often take themselves to be saying something more; they want to claim that it is in some sense wrong not to bring about the good (whatever it may be).  But what can this 'more' be? 

I think that this kind of naturalism faces a dilemma.  It either takes the naturalistic promise to define moral/normative notions descriptively and thereby deny that there is any significance to 'ought' (and right, etc) that cannot be cast in terms of 'is', (as nihilists claim) or they doggedly commit to the naturalistically irreducible significance of 'ought' (as non-naturalists maintain).  Sooner or later, the philosopher of value has to decide whether he is a nihilist or a non-naturalist.  There aren't any other options.   

Motivation for the evaluative deflationist

Assuming that 'is good' is a predicate used primarily for the purposes of communication and designation, rather than to pick out a particular property or set of properties, how would this affect the internalism/externalism/cognitivism/conativism debate about moral motivation?   I think that it would undermine its status as a debate. 

'Good' is among the most indiscriminately used terms in the language.  We use it, perhaps most frequently, to express our preferences, and in such cases a conativist internalism seems a natural description of what is going on. 

But, in part because preferences are so prone to conflict, both interpersonally and intrapersonally, the good-predicate is not limited in its use to occurrent personal preference.  Its utility for communication would be nil if it were.  Furthermore, the good-predicate is often used in the sense of 'good for', where that is naturally construed as relating, no doubt with varying degrees of remove, to preference.  Hence, eating spinach might be good for me.  So, anyway, I've been told.  How do I understand such a claim?  As that eating the stuff will further a healthy existence, something which will likely be presumed by those who say it that I prefer.  Whether or not I do, it will be natural for someone brought up in our linguistic environment to find the claim that spinach is good (for you) perfectly intelligible.  Moreover, someone brought up in our culture is likely to believe it.  That is, they will believe that spinach is good for them, where that means nothing more than that it is conducive to health.  But why think that such a belief need have any connection to the believer's preference for health, to say nothing of a preference for eating spinach?  Nothing, it would seem, and this would appear to be a conclusion amenable to the cognitivist externalist.

What I want to suggest, from these brief thoughts, is the following. 

1. Deflating goodness makes evaluative judgments not particularly interesting.  Often, they are, in effect, judgments of relations between certain descriptive properties or states, often including, though not necessarily, states of preference.  There is no reason to think such judgments lack truth-conditions, presuming you think there are such things (truth-conditions, that is).  So the question about whether we are to be cognitivists or conativists about evaluative judgment seems to lose a lot of its bite.  Some of those judgments, to repeat, are about the obtainment of certain states. But others, again, are more like expressions of preference.  There is no point to think that there can be only one answer.      

2. The connection between evaluative judgment and motivation is equally plural: internalism is what one would expect when the judgment is directly grounded in personal preference.  Externalism when it is not.

3. The moral seems to be that the idea that evaluative judgment has to be either conativist or cogntivist, and that the the link to motivation has to be either interanal or external, is bound up with the idea that there is a robust evaluative property of goodness.  If so, then I think a more useful response to the actual way our motivations relate to our evaluative judgments is not to debate when, why, and how they might be 'irrational' but that it shows, prima facie, at least, that there is no robust property of goodness.  Put another way, the moral motivation debate makes sense only on evaluatively realist assumptions and, furthermore, those assumptions seem to ignore just how we use evaluative notions themselves. 

Is there a 'normativity paradox'?

Here's a (very) half-baked thought I recently had. Normativity, if there is anything we can agree about it, is distinct from the descriptive; normative claims are not exhausted by their descriptive content (assuming they have any). But can creatures that communicate in natural languages communicate meaningfully in non-descriptive terms? To clarify slightly, by 'meaningful' I here mean articulable content (so I'm not including in such communication facial expressions, or the communication of feelings or emotions). I am also understanding 'descriptive' widely enough to include subjunctive and counterfactual claims. So, if I'm right about the 'essence' of normative claims, and about whats comprehensible in natural language, is it the case that normative claims are necessarily incomprehensible?

Back Again

I've been working and travelling a lot this summer, so its been a rediculously long time since I've posted anything. The summer was quite productive, however, and I went to a number of conferences and had many a stimulating conversation. Now that school is beginning up again (!), I hope to make use of this site more frequently, using it not merely as a repository of ideas but as a pedagogical device. In that vein, keep on the look out for some possible format changes. I will keep those interested updated.

What is the liar lying about?

I've been thinking about the liar paradox recently and can't shake the feeling that the problem is one of meaninglessness.  I realize that this is not a novel take on the matter, but nor is it particularly popular.  Consider the standard 'liar sentence'

(1) This sentence is false.

I don't see paradox here.  I see nothing: I can grasp no content here to which falsity is being ascribed.  The same applies to

(2) This sentence is true.

Being told that (2) is true tells me nothing informative about (2).  My initial suspicion is that the well-formdedness of (2), like (1), obscures the fact that neither of them are about anything; they are grammatical formulations lacking in content. 

I am not claiming that the problem here is one of self-reference.  For instance,

(3) This sentence has five words.

though self-referential, is perfectly meaningful; i.e. it has content.  This is why (4)

(4) The sentence 'This sentence has five words' is true.

is also meaningful (and, presumably, true).  But (5)

(5) The sentence 'This sentence is true' is true.

is, like (2), though grammatical, void of content; nothing is being said.  I don't see how a paradox can be generated out of such constructions.   

(BTW: nothing changes if we move to discussing the 'strengthened liar': (6)

(6) This sentence is not true.

is also missing content.) 

I have read somewhere that this is something along the lines of response that Strawson's 'performative theory of truth' gives to the paradox (though I have yet to track down where, exactly, he says this).  Does anyone know of any similar responses?  Anyone want to say what's wrong with it?  Most of what I read dismisses the meaningless response as unpromising, but I think they are conflating issues of well-formedness, self-reference, and content. 

Update: I have it on good authority that the 'lack of content' line is found in the work of Keith Simmons (and others). He (said authority) also thought that something like that line is surely right.  I agree, of course.  The question, for me anyhow, now becomes trying to explain why this is so.      

Truth for Hire

Frege presents two arguments that one might find question-begging. The first, explicit, argument is that the laws of truth are distinct from the laws of thought because the latter (somehow) enable us to distinguish between true and false assertions/beliefs. This assumes that that distinction cannot be drawn in doxastic (or at least a combination of doxastic and other non-alethic) terms alone. The second, implied, argument that only alethic laws (laws that, being themselves non-normative, are inviolable) will ensure that we believe as we ought simply presupposes that there is a normative standard applicable to belief, one that is not to be understood in terms of the believer’s peculiar motivations for fixing beliefs in the first place.

Obviously, someone who thinks the alethic is more or less reducible to the doxastic will not be moved by either argument. But, as Dave Shein is quick to remind me, such a response brings us dangerously close to succumbing to a vulgar game of burden tennis.   Since the last thing we want to be here is vulgar (the forum itself is vulgar enough), it seems that what we need to do is identify non-normative work for truth that cannot be duplicated by doxastic means. An inability to do so would, I think, wed completely the fortunes of truth to normativity; both notions would sink or swim together.   So, what, besides providing the 'standard for belief,'  is the truth good for?

Whats wrong with psychologism?

I've been rereading Frege's "The Thought," which he opens by claiming that truth is the 'aim' of logic, in the sense that "it falls to logic to discern the laws of truth."  Then, in an effort to clarify, he distinguishes between two senses of 'law'.  The first sense, which he associates with 'morals', picks out "regulations which ought to be obeyed but with which actual happenings are not always in conformity."   The second sense, which he associates with 'laws of nature', "are the generalization of natural occurrences with which the occurrences are always in accordance."  It is this latter sense that he intends when he speaks of laws of truth.   

It seems clear that the distinction that Frege is making here is between what we can call a 'normative' law and a 'natural', or 'existenially necessary' law (it is a fair question whether the idea of an inviolable 'law' makes any sense, but we can put that aside for the moment).  So the laws of truth, according to Frege, are not normative.  But things get confusing rather quickly when Frege tries to block the assimilation of laws of truth to another set of natural laws, namely the 'laws of thought,' "the generalization of thinking as a mental occurrence...a psychological law."  To attempt such an assimilation would, Frege tells us, "be a misunderstanding of the task of logic, for truth has not been given the place which it is due here.  Error and superstition have causes just as much as genuine knowledge.  The assertion both of what is false and of what is true takes place in accordance with psychological laws.  A derivation from these and an explanation of mental processes that terminates in an assertion can never take the place of a proof of what is asserted....In order to avoid this misunderstanding and to prevent the blurring of the boundary between psychology and logic, I assign to logic the task of discovering the laws of truth, not of assertion of thought."         

I fear I have misunderstood Frege's argument here.   It seems that psychologism is being rejected for its failure to make what seems like a normative distinction, between correct and incorrect, true and false, beliefs.  And what seems to make this failing significant is that true beliefs, rather than 'error and superstion,' are the beliefs we apparently ought to have.  But doesn't this suggest truth is a normative notion?  Has Frege mistakenly characterized the putative laws of truth as non-normative?  If not, however, what more can be said about such natural laws that enables us to distinguish them from the laws of thought?  And what is it about such laws that tells us why our beliefs should conform to their object?  To put these questions another way, just what is the natural, non-normative alternative to psycholigism in logic and why believe it?