BARBELITH underground
 

Subcultural engagement for the 21st Century...
Barbelith is a new kind of community (find out more)...
You can login or register.


Underground > Head Shop >

The Frege-Geach problem in ethics

  
 
Good Intentions
01:41 / 14.11.07
I'll start this off with a pre-amble, for those interested parties who don't study metaethics. I think this is necesary, because metaethics is a field that takes the most pressing philosophic problem - how should we live? - and makes it as finnicky and academic as any other branch of philosophy, meaning that questions of metaethics aren't much discussed except by experts (and their students, like myself).

Those of you who have sat down and wondered about what the nature of right and wrong is anyway might have come up with the idea that it's simply inappropriate to talk about a moral judgement being true or false. Everybody seems to agree, more or less, that wantonly setting children on fire is a pretty terrible thing to do, but judgements like 'abortion is wrong' is much, much more contested, to the point that it might seem that universal agreement on them is a pretty remote possibility. Subjectivism - the idea that a moral judgement is true or false depending on who you are - has some pretty enormous problems and I for one think it's simply an untenable position, because you are saying that the actual, honest-to-God, encoded-into-the-fabric-of-the-universe truth of a moral judgement changes willy-nilly from one person to the next. A position that is much more attractive to myself and most other people in the field is that moral judgements aren't truth-apt: they are less like factual claims and more like exclamations. The classical example of such an expressivist theory is Ayer's emotivism, also called the boo-hurrah theory, because to say 'it is wrong to burn children alive' is akin to saying 'boo, burning children alive!' and going 'helping your fellow man' is like 'hurrah, helping your fellow man!' When I say that everybody agrees that burning children alive is bad, what I am saying is that everybody disapproves of burning children alive, not that it is a truth of the world that burning children alive has the property of being evil. Moral codes then are expressions of some inner state, for Ayer's your emotional response to a given situation, and in Hume's terms 'children of the sentiments', hence the tern expressivism.

Note that this is not a normative, ethical claim. I'm not saying that every moral judgement actually says 'do what you feel is right'. People often judge the right thing to be soemthing they don't feel s right. This claim is metaethical - it is about the form and nature of moral judgements. It says that all moral judgements are seated on expressions of sentiment, even when the normative judgements make no reference to them: 'never lie' is not an expressivistic judgement, since it tells you to be honest even when you'd much rather not be, but expressivism claims that what drives you to never lie is a deep-seated attachment to honesty, not an intuition of the true moral code or something in that order. An excellent introduction and discussion of these types of questions is here: Article in the Stanford Encyclopedia of Philosophy

Now, the Frege-Geach problem is a knockout argument against emotivism: it runs as follows:

Moral judgements are supposed to be expressions of sentiment, and this seems plausible for statements like 'don't torment the cat'. But what about if such moral claims are embedded within valid (in the logical sense) arguments, like 'if it is wrong to torment the cat, then it is wrong to get little brother to torment the cat'. This argument only makes sense if the premise 'it is wrong to torment the cat' is truth-apt, either true or false, and remember than expressivism denies that this s the case. But that would render this perfectly ordinary use of a moral judgement nonsensical, thus, it can not be the correct account of that moral judgements actually are. You'll see that a theory as unsophisticated as Ayer's emotivism ('all moral judgements are merely expressions of emotional reactions') can not respond to this objection (raised by Peter Geach from a point he attributes to Gottlob Frege).

Because expressivism's opponents, like moral objectivism, have problems of its own, none of them less serious than the Frege-Geach problem, a lot of people (myself included) have stuck to expressivism, shoring it up against objections like these. The two strongest contemporary expressivist metaethics are Simon Blackburn's quasi-realism and Allan Gibbard's norm-expressivism. In quasi-realism it is noted that if enough people agree on certain judgements then these judgements can be handled as if they were cognitive and truth-apt, in the same way that it might not be a factual truth of the world that the Smiths is the greatest band in existence, but if you hang out at a Morrissey concert you could talk as if it was that way. In norm-expressivism, the current king of the expressivist roost, when we make day-to-day moral judgements we do so according to a certain moral code, one we understand and apply exactly as if it was handed down from the heavens, but the way we choose which system we use and what ultimately supports it is sentiment. Like, I might live my life by an exacting utilitarian code, but what brings me to do so is a deep-seated attachment to utilitarian values. In especially the case of norm-expressivism, the Frege-Geach problem is met because the argument 'if it is wrong to torment the cat, then it is wrong to get little brother to torment the cat' remains valid: if you use a norm where tormenting the cat is wrong, it is consistent with that norm not to get little brother to torment the cat.

*** Alert! Real question below! ***

Problem solved? Apparently not. An objection levelled at Blackburn's quasi-realism, which he in turn levels at norm-expressivism, is that an appeal to consistency is to give up the fight for expressivism, in Blackburn's words, consistency is 'on the other side of the Fregean divide'.

What on earth does he mean? I don't get it. It seems to me that he's saying that any system that has consistency in such a manner is already a cognitive (non-expressivistic) system, where judgements are at bottom truth-apt, not just approximately, and are about truths of the world, or what the speaker believes to be truths of the world (consider a rival metaethical theory, Mackie's error-theory, where all moral judgements are truth-apt, but all are systematically false, because there are no really existent moral properties which could make such judgements true). If the day-to-day ethical decision making is along the lines of a cognitive norm, in relation to which individual judgements are truth-apt, then I don't see how the Frege-Geach problem gets traction against norm-expressivism. I can see how appeals to consistency alone is not a sufficient response, because then it would be unclear whether a logical error is committed by someone who accepts that tormenting the cat is bad but sees no problem with getting little brother to do it, and whether such a person would merely be frivolous (this is how the argument was raised against Blackburn, btw, and is a big worry for quasi-realism). Earlier I had considered tackling this point at hand of Carroll's paradox which argues that any logical error simply is being frivolous itself, but now I see that that would not do the work I'd like (there's different standards for being frivolous in logic and being frivolous otherwise). So I'm left with no real grasp on what the Frege-Geach problem modified to norm-expressivism does, and correspondingly on how to tackle it.
 
 
Good Intentions
11:26 / 19.11.07
Perhaps I'll succeed in attracting a reply if I flesh out some more on the content of these various ideas.

Norm-expressivism is a theory I find very impressive, and I probably didn't go into enough detail about it. As I mentioned, according to norm-expressivism every single moral judgement is a product of a norm or standard that is itself an expression of a deeply-held attachment to some value: for instance, I try not to leer at women because this is part and parcel of my heartfelt feminism. Where an emotivist (naive expressivist) would understand feminism as a collection of different judgements like 'leering: boo!' and 'patronising women: boo!', norm-expressivism is more cogent by understand it as a simgle claim, like, 'upholding the dignity of women, please', and the individual judgements follow from that.

Something I didn't mention is that the content of individual judgements, according to norm-expressivism, are themselves expressivistic: what the norm is according to which moral judgements are made is a standard according to which single acts are judged to be worthy of praise or of shame. When I am committed to feminism (and it's an example of norm-expressivism's cogency that it makes such good sense of the term 'committed'), the story goes, I see leering at women as shameful, and treating women equally or contributing to their enfranchising as praise-worthy.

By the norm-expressivist account, the argument given as an example of the Frege-Geach problem above would run as such: if it is shameful to torment the cat, it is shameful to get little brother to torment the cat.

Now, to me this seems to solve the Frege-Geach problem. The argument remains valid : if you recognise that bringing about torment to cats is immoral (shameful), then bringing it about by way of convincing your little brother to do it would also have that quality.
But the response to this response still lingers in the literature: the idea that what we have here is no longer expressivism because it appeals to something that is propositional and cognitive, rather than being a 'child of the sentiments'. If this is the case, if norm-expressivism sneaks in some logical content not located in the appreciation of pride and shame, then the answer norm-expressivism offers to the Frege-Geach problem isn't an answer at all. I just struggle to see how that might be the case.
 
 
All Acting Regiment
11:46 / 19.11.07
This is heavy, and I'm trying to construct something to add to the Badiou thread, but I will get back to it. It's highly interesting - thanks for putting the effort in.
 
 
Lurid Archive
16:03 / 20.11.07
This is interesting, but I think it is being presented as too technical for people to really get a handle on it.

I'm not sure I can say anything helpful really...except...I think I've understood Carroll's paradox as saying something different than that "any logical error simply is being frivolous itself". I mean isn't it saying - in this context - that having moral intuition about particular things isn't enough, because you also need a basis for moral reasoning. So, you need some explanation of how you arrive at the conclusion that it is shameful to torment the cat (norm expressivist, if you want) but you also need an explanation for the validity of making the inference that it is therefore shameful to get little brother to torment the cat. This sounds like a silly logic thing on first glance...but I'm not sure that it is, since there are lots of moral situations where our conclusions, however we arrive at them, are dependent on the actors and the precise situation involved. It is somehow implausible - at least to me, and I've spent a good five minutes thinking about it - that our intuition directly applies to each of these potentially infinite situations individually. So, there must be some reasoning going on, which is probably neither clear or deterministic. I'm guessing that the Frege-Geach problem you mention starts to get more traction as you apply it to moral reasoning where conclusions do depend on the actors involved, and the precise circumstances - all those trains killing people in morally cryptic ways, for instance.
 
 
Good Intentions
10:19 / 21.11.07
No, the Frege-Geach problem is specifically a metaethical argument, against the metaethical position of expressivism and entirely unconnected to any particular situation. No matter how much detail you skecth in, it doesn't change the metaethical character of the situation a whit. Whereas doing (normative) ethics is studying, and being in search of, right action, and applied ethics is the application of normative theories to particular situations, metaethics is the study of moral judgements themselves.

I imagine that the person making the judgement that it is wrong to torment the cat has his reasons for doing so, and his explicitely stated reason would be normative ethical standard in effect (if he's a utilitarian, it's causing harm to the cat and those who like the cat, if he's a virtue ethicist it's because he's acting cruelly, etc). The Frege-Geach problem does not gain or lose traction from the amount of detail that goes into the judgement, though, because expressivism, which it attacks, is a theory about moral language, not about the content of moral judgements. One version of expressivism Frege-Geach blows out of the water is emotivism, which analysis 'it is wrong to torment the cat' as 'boo! tormenting the cat! boo!'. The Frege-Geach problem comes from embedding moral judgements into arguments (and inferences and other logically-sensitive structures): in the emotivism example, the cat-tormenting argument from my OP becomes:
If tormenting the cat, boo! then getting little brother to torment the cat, boo!
Which is nonsense. Because emotivism can't make any sense of a perfectly simple and commonplace task of moral judgements, moral inferences, it can't be the correct theory about what is happening when people are busy with moral talk. No matter how thorough or how cockeyed the reasoning the person making the judgement used was, emotivism can not be the correct theory for describing the form of that judgement.

My problem is that I think norm-expressivism's modification, where the cat-tormenting argument analyses to:
If tormenting the cat is shameful, then getting little brother to torment the cat is shameful.
Which seems perfectly fine to me.

I've tried substituting 'dangerous' for 'shameful' in an attempt to have a non-moral version of this inference to test it's validity, but it's not proven decivisive because for some senses of dangerous ('dangerous to me') the argument becomes invalid, and for others it stays valid but in a way that suggests that the moral version could not ('dangerous to the person doing it' stays valid, but the point of the moral inference is to place blame on Jimmy for getting little brother to torment the cat, while getting little brother to torment the cat would quite decisively move the danger onto little brother) while the sense of 'dangerous' that seems to most closely resemble 'bad' ('dangerous to someone', the way you could talk about an unstable building as being dangerous even if you don't have anything to do with anybody near the building) doesn't seem to clarify anything.
 
 
grant
13:16 / 21.11.07
Your problem there is that "shameful" has a public, social component in a way that "dangerous" doesn't. I'm not sure you can have that kind of social component without it being an ethical proposition, though.

"Shameful" is inherently about the behavior/attitudes of crowds and the way an individual relates to a crowd. I wonder if you could try using a term from epidemiology to divorce your argument structure from ethics - something about vectors of infection, maybe?
 
 
Lurid Archive
14:42 / 21.11.07
No, the Frege-Geach problem is specifically a metaethical argument, against the metaethical position of expressivism and entirely unconnected to any particular situation.

Maybe I'm not understanding something, then, or maybe I'm coming at this from too much of a formal logic background, but...my point with the details *was* about metaethics, in the sense that I was trying to justify my point about moral reasoning, as well as reaction, being necessary. That is, you can't just argue about the basis for moral judgements without considering also the structure of moral arguments. So I guess I'm not understanding you here, since it seems to me that examples of moral reasoning are quite significant to our understanding of the viability of a system of moral judgements. In your example, you are implicitly saying that a moral rule of inference - if it is wrong for me to do something, then it is wrong for me to get someone else to do it - is valid. Is this a general rule of inference that holds in norm-expressivism? How can we tell? Well, by looking at lots of similar sorts of inferences and seeing what the results would be. But if the conclusions change with the details, it isn't a sufficiently explained rule of inference - which is fine - but then we have to be more suspicious of using it implicitly.
 
 
Good Intentions
09:20 / 23.11.07
Ok, Lurid, I think I see where you're getting at. In case I don't, I'm going to go into some detail about what I'm getting at, and please forgive me if I seem to be writing out course notes What you are saying is of course entirely correct - you need a moral reasoning to make moral judgements, even if the theory you follow is something as thin and undeveloped as 'do what your intuitions tell you'. The thing is, the description you give of what must fill that gap is a perfect, perfect fit for the role of a normative theory of ethics, which is why I replied in the way I did. You are correct that it is a matter of metaethics that such a reasoning process is needed. The reasoning process is given by whatever normative theory is being used (*) . For the sake of the study of metaethics, the normative theory in effect is bracketed out, since it plays no role in the metaethical character of moral judgements.

I think the point of confusion is in this sentence: "you need some explanation of how you arrive at the conclusion that it is shameful to torment the cat". Not in order to make a moral judgement, no. For this to be have been the case, the normative theory in use must be 'do not do things that are shameful, but rather ones that are praiseworthy', with some more concrete description of what shame and praise attaches to. But this isn't really a plausible normative theory: it's not like substituting 'good' with 'praiseworthy' is informative in any way (+) . What norm-expressivism and its competitors instead do is to guess at the source of the good itself.

I can see more confusion ahead, since 'good' is a word that is hardly used tightly in natural language, and the varying ways in which the concept probably causes confusion here and elsewhere. I'm going to go more technical in the hope of clarifying what I mean. I apologise: this won't take long.

Normative theories describe the extension of 'good' in its sense as a predicative: right actions are those to which the predicate 'is good' can be used, and a normative theory is a description of the criteria needed for some action to go on the list of right actions (this list of right actions is what is meant by "the extension of 'good' in its sense as a predicate"). The metaethical theories discussed here (emotivism and norm-expressivism - quasi-realism is a little different: its a theory on the behaviour of expressivisms in general, and can be safely ignored here) are theories of under what 'the good' falls: in the case of emotivism 'the good' is projections of emotional responses, in the case of norm-expressivism 'the good' is expressions of a norm on which actions are shameful or praiseworthy. What these theories attempt to inform is to recognise which judgements are moral judgements, not what the correct moral judgement might be.

* At the risk of causing more confusion, I need to put in a disclaimer - there are theories which say that there is no plausible normative ethical theory, that is, there is no theory of what is the right action that fits for every scenario, and that each moral judgement should be made on a case-by-case basis (this view is called particularism). In case anybody was wondering. The main counter to this view is that almost everybody believe that we can do better than that, that even if there isn't a theory that isn't plausible for every scenario, then there can be one that fits for almost every one, and gives us some guide to what the right action in a scenario would be. Particularism seriously reduces the guidance ethics might provide for our lives. Particularism is a normative theory (in its simplest version it says 'judge each case on its merits') and not a metaethical one. If there is a correct metaethical theory, then it would also correctly describe particularist judgements.

+ The three most plausible normative theories are utilitarianism, filled out in some way, ('the greatest good for the greatest number' with some description of 'good'), virtue ethics ('live in accordance with the virtues', with some list of virtuous behaviours) and Kantian deontology ('act in such a way that you would want to be a universal law', which doesn't need some filling out but requires some serious forethought on the part of the practicioner).
 
 
Good Intentions
07:45 / 24.11.07
One thing I passed over: Lurid asks whether it's necessary for the moral inference I've been using as an example (I nicked it from Michael Smith), the cat-tormenting argument, to remain valid in any proposed metaethical analysis. I'm sure there are people who say it must, but most people agree that it must be a weakness of a theory if an apparently valid argument like the cat-tormenting argument can not be analysed as being valid by that theory. If a theory can not deliver on that, we might feel that we could do better.
 
  
Add Your Reply