Analytic Bullshit, Ben Kotzee

Can analytic philosophers talk bullshit?

The publication as a small book last year of Harry Frankfurt’s (1986) paper “On Bullshit” ignited great popular interest in “bullshit”: what is it “to bullshit someone” and why is there so much bullshit about these days? Prime examples of bullshit mentioned in the many popular contributions to the subject post-Frankfurt (just google for it) are found in “business-speak”, in advertising and in politics. What people enjoy about Frankfurt’s book, it seems, is that having a theory of bullshit available makes it possible now to do with a straight face what you previously had to hide in a cough: say that someone is talking bullshit. (We might say that, after Frankfurt, “bullshit” is a technical term.)

As far as I can tell, the theory of bullshit is in its infancy and I’m afraid that I don’t have much to add. Frankfurt distinguishes between honest assertion, lying and bullshitting as follows: In making an honest assertion (in telling the truth) one aims to say what is true and in lying, one aims to say what is not true… both in honest saying and lying one is guided by the aim of truth. In bullshitting, however, the speaker is unconcerned with the truth of what he says; the bullshitter pretends to make an honest assertion whereas he really is just mouthing off. As such, bullshitting is a faking of assertion: the bullshitter pretends to make an assertion, but actually asserts nothing. To people who know me, it will be quite clear why I’m interested in this: I work on the relation between truth, believing and assertion and think Frankfurt makes a very good point about the nature of assertion: honest saying and lying, as species of asserting, involve a concern with the truth of what one says. This comes into sharper focus when we consider the case of the bullshitter who pretends to assert, but, not caring about the truth of what he says, really ends up saying nothing (ends up not really asserting at all).

Be that as it may, Frankfurt thinks that bullshitting is a kind of dishonesty: whenever one speaks dishonestly in this way, what comes out one’s mouth is bullshit. Jerry Cohen disagrees with this characterisation of the relation between bullshitting intent and the shittiness of what one says. He holds that it is possible to talk bullshit without dishonest intent and mentions as an example the stuff emanating from French departments of philosophy and departments of literature in the English-speaking world overly occupied with French theory. (Holding up an example of bullshit, Cohen refers to Althusserian Marxism; he also mentions the writings on science of French theorists from Latour to Kristeva that Sokal criticises.) Cohen holds that there need not be any dishonesty on the part of these people: the problem is not that they are unconcerned with whether what they are saying is true, it is that there is a deficiency in the concepts and language that is deployed by people who “do theory”. What is wrong with much of French philosophy, Cohen thinks, is that it is unclarifiable nonsense and this he wants to distinghuish as a sort of bullshit in its own right. He provides a test for being unclarifiable nonsense that involves adding a negation-sign: if adding a negation-sign to what one is saying makes no difference to its intuitive plausibility, it is bullshit of the “unclarifiable nonsense” kind.

What Cohen calls bullshit, it is clear, is that kind of philosophy perpetrated by people whose names sound continental, that is obscure, a little bit avant garde and generally down on mathematics, science, logic and technology. Recently, I criticised Cohen’s “unclarifiable nonsense” account of bullshit, but offered the following (qualified) support for his views on French philosophy. I held that “postmodernism” is bullshit for the following reason: Assume that the central tenet of postmodernism is that there is no such thing as truth or that the word “true” is no more than a clever cover for whatever beliefs or attitudes are generally accepted in some culture (and that is accepted due to concealed coercion). The problem is this: If there is really no fact of the matter as to what is true and someone may become conscious of this, then there can be no honest speech and no lying; this is because, as Frankfurt holds, assertion and lying is characterised by aiming to say what is true and aiming to say what is not true, respectively. Being fully aware that there is no truth either way, no-one can honestly assert anything (or lie) at all; without truth, assertion looses its goal. All that can remain of speech, if there is no truth, is bullshit or pretending to assert (although just pretending to assert would require at least the idea of truth and truthful assertion to remain, itself a tension in the postmodernist’s position on truth).

My argument invited the accusation of tu quoque from an editor. Analytic philosophers, he suggested, shouldn’t cast stones. Bullshit is not confined to the continent and, in any case, I wouldn’t take it as alright if a “postmodern philosopher” made a blanket attack on analytic philosophy in the same manner as I did. He had me wrong – I am perfectly willing to consider any reasonable argument that there is something systematically wrong with the presuppositions and method of analytic philosophy, its just that I haven’t heard one, despite listening. (By the way, I’m not complaining about the editor, who liked the rest of the piece and accepted it.)

In the interest of fairness, though, my question to the blog is this: do analytic philosophers ever bullshit? Nothing suggests that an analytic philosohper can’t talk Frankfurt bullshit – that is, pretend to care about the truth of what they say when they do not: of course any analytic philosopher is just as capable of this form of dishonesty, psychologically speaking. What’s less clear is that an analytic philosopher can talk Cohen-bullshit. I would suggest that the true analytic philosopher cannot. This is because analytic philosophy is characterised by its reliance on the method provided by formal logic: by formalising the contentious parts of our work, we make absolutely clear what we mean. At least when we formalise our philosophy, we can of course be wrong, but not “unclarifiably unclear”. Precisely the ideal of analytic philosophy is to be clear.

Bearing in mind these two points, does anyone want to offer examples of analytic philosophers talking bullshit (of the Frankfurt or Cohen variety)? Specifically, I’m interested in the role that formalisation in the language of logic plays in making our work clear or unclear. To some – that is people who’ve never taken a first level course in logic – much analytic philosophy looks absurdly complicated and technical. I’m interested in this. Does the method provided by formal logic ever obscure rather than clarify, or is this just a matter of not being able to read the logic? Even assuming that everyone should be capable of following it (and why should they?), can someone think of an example where formalising a point or argument renders something that is clear obscure? Would people write to me with their nominations for the prize “most gratuitous formalisation in the language of logic of something that’s perfectly clear in English”?

Certainty and Knowledge: Jonny Blamey

Is there more to knowledge ascriptions than factivity and evidence?
Jack’s dog is dead. Jack got a phone call from the vet saying his dog died this morning. Jack’s evidence is undefeated, (the vet saw Jack’s dog die). Is this enough to determine whether: “Jack knows his dog is dead” is true or false?

Let’s start with a simple definition of knowledge:

S knows that p iff S has a rational degree of belief 1 that p and p is true.

To have a degree of belief 1 that p means that you would accept odds of 1:0 on p. To have a rational degree of belief 1 simply means that you would be rational to accept these odds. The odds are the ratio between the loss if p is false against the gain if p is true. Degree of belief 1 is unique because one can increase the loss indefinitely without affecting the gain. For example, if Jack were to a bet a penny for no gain that his dog was dead, his degree of belief that his dog was dead would be 1. If Jack were to bet £1 million for no gain, then his degree of belief would still be 1. It is perfectly rational to demand more evidence to risk £1 000 000 for no gain than to risk a penny for no gain. Therefore the evidential requirement for a rational degree of belief 1 varies with the size of the stake. Conversely, the evidence alone will not determine whether someone would be rational to have a degree of belief 1. So evidence and truth alone cannot determine knowledge ascriptions.

If certainty is the disposition to rely on beliefs on a loss/no gain basis, then we should expect truthful knowledge attributions to vary with the magnitude of the loss. Is Jack rational to be certain that his dog is dead on the basis of the vet’s phone call? Let us take two situations.

1 : The Dog bed.
Jack decides on the basis that his dog is dead to throw away the dog bed. He has nothing to gain by throwing the dog bed away today. But if his dog is alive, then he will have to get a new dog bed. At this stake size, he is certain that his dog is dead. He is prepared to risk the cost of a new dog bed for no gain. Given his evidence, Jack is rationally certain : Jack knows his dog is dead.

2: The Inheritance.
Jack stands to inherit £1 000 000 conditional on Jack showing that his dog is alive to a lawyer. If he phones the lawyers to tell them his dog is dead, he loses £1 000 000 for no gain. Instead he drives around to the vet to check for himself that the dog is dead on the off chance that there has been a mistake, or that the vet is lying. He is not prepared to risk £1 000 000 for no gain on the basis that his dog is dead. Therefore he is not certain that his dog is dead. Given his evidence he is rational not to be certain. Jack does not know his dog is dead.

Things in themselves, appearances, noumena, and phenomena, Jessica Leech

Consider two distinctions: the ontological distinction between things-in-themselves and appearances (what things are), and the epistemological distinction between noumena and phenomena (how an object may come to be known). One might take these distinctions to be co-extensional, i.e. all appearances are all phenomena, and all things-in-themselves are all noumena. I have reason to disagree with this (and the preferred interpretation leads onto an account of a distinction between self and world). The first conjunct is acceptable; all the appearances are the same as all the phenomena. Both ‘appearances’ and ‘phenomena’ are defined by experience, in the first case the general type of what is experienced, and in the second case how it comes to be experienced. In both cases the same things are going to be picked out; everything experienced in the empirical world is appearance, and everything here experienced must be experienced through sensible intuition and the understanding, and thus is phenomenal.
I do not accept the second conjunct; that things-in-themselves and noumena are co-extensional. Take a simple general case, the thing Y: Y1 is a thing-in-itself, underlying an appearance y1, y2 is the phenomenal object the experience of which corresponds to the appearance y1, and Y2 is the noumenal object postulated through y2. Roughly speaking y1 and y2, and Y1 and Y2 correspond, those numbered 1 being ontologically defined, those numbered two being epistemically defined, those in lower case being part of the empirical world of experience. Now contrast this with a special case, the thing Z, where Z is a thing-in-itself (we do not have experience or knowledge of Z in any way). In the same terms as above, we have the thing-in-itself Z1, but no appearance z1. As there is no appearance, and appearances and phenomenal objects go together, then there is no phenomenon z2 here either. Finally, as noumenal objects are postulated as a result of experience of a phenomenal object, with no z2 there is no Z2. Thus, one has the thing-in-itself, Z1, with no corresponding Z2. Therefore the ontological and epistemological distinctions in question are not co-extensive.
This conclusion is perhaps not so surprising, given the different foundations on which each distinction is based. The ontological distinction concerns the nature of things; what they are. Although what appearances are is the product of our intuitions and concepts, the distinction does not rely on a human subject. It is just that without human experience to provide appearances, there would only be things to satisfy one side of the distinction—the possibility of there being a thing-in-itself with no appearance is just the possibility of there being no experiencing subject to “create” the appearance. Conversely, the epistemological distinction is based deeply in human experience, as it depends on different ways of knowing objects, presupposing that there must be a ‘knower’ before anything can be distinguished.


Now take the self as a supposed object of investigation. Suppose there is a real self, a thing-in-itself, S1, then there may be a corresponding appearance s1. We might accept this, allowing that the appearance is the empirical self. The empirical self is phenomenal, thus we then also have s2, and further postulate S2, the noumenal self. Now, the phenomenal s2 is related by some knower to the postulated S2. We are purportedly classifying the self, or elements of the self according to the phenomena-noumena distinction. However, this distinction itself is driven by there being a (human) knowing subject. This knowing subject is the recipient of intuitions, and will apply the concepts of the understanding and so on, resulting in knowledge of the empirical self, and inference to the noumenal self, these being associated with that same knowing subject. However, it is unclear whether something that is a foundation of a distinction can have that distinction applied back to itself in this way, i.e. can a distinction founded in some x be applied to x?
This kind of self-referential problem is a serious and wide-ranging one. (Related issues arise in the study of logic, associated with Russell’s Paradox, the Liar Paradox and the use of diagonalisation in Gödel’s incompleteness theorems.) It is unclear if the distinction at issue here, based on how a subject has knowledge, can be applied to that very same subject. We would no longer be considering the case of how a separate object is known, but rather the special case of self-knowledge. Do we know the self through sensible intuition and the understanding, or do we postulate the self as underlying phenomenal experience? Above I have already rejected both accounts, at least where the self in question is that in a self-world distinction. This lends support to the claim that the phenomenal-noumenal distinction cannot indeed apply to the subject on which it relies, whereas the ontological distinction can.