how Is influences Ought

Posted in Uncategorized by sarkology on June 19, 2012

The is-ought gap states that one cannot derive an ought from an is. Which is to say factual propositions have no bearing on normative claims, i.e. what should be, as opposed to what is the case. We can exclude the trivial case where ‘ought’ is construed instrumentally, in the sense that if we have agreed that we want to achieve X, reality forces us to do Y, i.e. it is the case that only Y can bring about X. This part is uncontroversial.

There are problems with this. Assuming naturalism, what ought to be would in fact be the case, since anything that there is, whether normative or descriptive, resides within reality, i.e. they are natural phenomena. Perhaps it’s something like the status of mathematical truth. They have an objective character (false vs. true mathematical statements) analogous to the normativity of ought-statements. But they do not reside ‘concretely’ within reality. Instead they are ‘implicit’ within the behavior of natural phenomena. For example, to say that 641 is prime is to say that if you try to factor it, you will get no non-trivial prime factors. This is the case for all ‘concrete’ natural phenomena embodying this ‘abstract’ arithmetical phenomena, whether they be Window or Mac (ha).

But there is a useful distinction to be drawn. This is not so much motivated by the ontology of factual vs. moral (that will be how I refer to oughts) claims. Ontology is overrated anyway, at least in the metaphysical sense. The distinction is this. When doing moral philosophy we want to be influenced by moral claims and we do not want to be influenced by factual claims. Of course, it isn’t straightforward to distinguish moral and factual claims at the margin, since ultimately, everything is factual. But no matter how we eventually arbitrate between a claim being moral vs. factual what we seem to be trying to do is to have a system of reasoning within which certain claims are admitted and from which certain other claims can be derived. Presumably, this system is not exhaustive of all possible claims for reality.

So I suggest we look at it this way: Moral theorizing is a ‘game’ we play by certain rules. The rules state roughly that only moral claims can be employed to derive other moral claims. All this is in the view of figuring out what is ‘good’, what we ‘should’ do, etc.

That is all well and good. But theory does not translate straightforwardly into practice. For example, the Bayesian epistemology prescribes precise rules for updating on evidence. But ‘evidence’ in practice does not conform straightforwardly to evidence in theory. For example, a scientific journal article makes a certain claim. That presumably provides quantitative rational evidence in favor of that claim. But what if there was publication bias? Similarly, an argument may be presented to you by another person which supports a specific claim. But what if that person cherry picked that argument and ignores certain other contrary arguments? Updating on Bayesian evidence presumes that the evidence was sampled randomly, if this condition fails, the rationality of the update is compromised.

The situation is similar for a class of ways in which is can influence ought. Say you preach Travel as a moral Good. Then someone else comes up with this ad hominem saying that you preach Travel not because you think it enlightens and uplifts but because Travel is expensive and is the domain of the privileged, and that you are trying to distinguish yourself from poor and under-privileged folks who cannot so readily afford travel. What’s up with the $5000 DSLR, he adds (ha). The case of ad hominem also happens to be analogous (in the sense that it carries rational evidential weight) but let’s focus on the attempted is-ought derivation here. The critic states that you are motivated in preaching Travel because you are trying to signal privilege. Notice first of all that it is certainly possible for both to be true. It could be simultaneously true that Travel is a moral Good and also that you are only preaching it to signal privilege. Assuming that you are in fact so ulteriorly motivated, should this fact be admitted within the moral philosophy arena? Should we subtract from our belief that Travel is Good based on the evidence that you were ulteriorly motivated? I argue yes. This is an is-ought gap we should cross.

The moral philosophy game is supposed to do moral philosophy work. Publication bias in science is a problem because it compromises the work that science does. Similarly coming up with moral arguments in a biased fashion compromises moral philosophy. Knowing that you are ulteriorly motivated suggests that you have not impartially explored the search space of moral arguments. So what if you possess with you these arguments for the Good of Travel, if you have in fact ignored the arguments against it? For example, that there are cheaper and less discriminating ways for one to engage with other cultures and to broaden one’s moral horizon. There is no unit of evidence. “Evidence” is an abstraction which we cast upon reasoning to analyze its functioning. A “piece” of evidence can be broken up into two conflicting evidences, or it could be lumped together with further evidences. The point is lost if we exclude anything which does evidential work, since then what we are doing would not be described as Bayesian updating, and hence would not lead to the rational results which we were after.

There are probably other ways in which is should influence ought. This is only one of them. Also, this is not to say that such ad hominem attacks should be permitted in moral discourse. Social institution design is another matter.

Tagged with:

2 Responses

Subscribe to comments with RSS.

  1. gwern said, on June 24, 2012 at 2:56 am

    > Updating on Bayesian evidence presumes that the evidence was sampled randomly, if this condition fails, the rationality of the update is compromised.

    Question: are you claiming that no Bayesian techniques can deal with any kind of filtering of data? Or expressing – in an unfortunately categorical way – the observation that the simplest naivest possible models involving Bayesian updating do not work completely in different settings?

    • sarkology said, on June 24, 2012 at 11:36 am

      Well at the most abstract level Bayesian updating is very coarse. So for example, a biased piece of evidence is simply modeled as weak evidence. If you want to include more structure in a more concrete model you certainly can. My point is not to show the inadequacy of Bayesian reasoning, but to show that evidential strength is not intrinsic to a “piece” of evidence, whatever that vague folk notion might mean.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: