Moral epistemology is an area of metaethics that I feel most
uneasy about. I’ve done the least research in this area, and I’m never quite
happy with what I read. However, I have tried to get at least started. So, this
post is a Groundwork to my moral epistemology (not that the presented view will
be original). I want to start from Sidgwick’s intuitionist position as
described by Roger Crisp in the section 3.4 of his book. Call this the ‘Crispwick’
view which I used to find very appealing. I will then explain why I have come to
doubt this view, and why these doubts give me a reason to be drawn into a view
which I call ‘Social Reliabilism’.
The Crispwick view gives four necessary conditions for being
justified in believing a normative proposition (I’ll use ‘that p’ for this
proposition). On their own, each of these conditions is insufficient for
justification, but jointly they are sufficient. So, when a believer who
believes that p is in circumstances in which all the four conditions are
satisfied, then and only then she justifiedly believes that p. My worry is that
I don’t see how the person could have justification to believe that p when all
the four conditions are satisfied if she is not already justified to so believe
when the first three conditions are satisfied.
The first condition is clarity. So, the believer must have a
clear and precise understanding of the concepts in terms of which the believed
proposition that p is phrased. This also means that the believer should know
the inferential connections of the proposition to other normative propositions
and what kind practical consequences accepting it could have. But, given that
the intuitionist does not think that moral propositions are analytically true,
conceptual competences cannot be enough for knowing them to be true. So, we
need more for justification.
The second condition is reflection. This means that the
believer should find the truth of the proposition convincing after hard, unprejudiced
reflection. This is required for both ruling out distorting emotional
gut-reactions, acceptance based on social pressures, and so on, and for ‘seeing’
the synthetic truth of the proposition. The reflection needn’t offer arguments
for the proposition as some of these propositions are basic truths.
The third condition is consistency. Even if my belief that p
satisfies the first two conditions, I am not justified in believing that p
unless I first check that I do not also believe that not-p (when this belief
also satisfies clarity and reflection). It’s not clear how much this condition
adds. It is questionable whether one can satisfy the first two but not
consistency. Inconsistencies are often signs of unclarity.
It is crucial, according to Crispwick, that satisfying the
first three conditions is not sufficient for being justified in believing that
p. If there are others who satisfy the first three conditions in believing that
not-p, then even if I satisfy those conditions I am not justified to believe
[EDIT] So, the fourth condition is then consensus. One is only justified in believing that p when one satisfies the first three conditions if everyone else in the same community also believes that p – or, at least all the epistemic peers who satisfy all the previous three conditions believe that p rather than not-p. Consensus in this sense adds justification, whereas disagreement undermines it.
My problem is that I’m not sure I see why the fourth
condition should matter for justification on the intuitionist view. I keep
thinking about desert island cases. Imagine that Ann and Ben are on a desert
island. Ann believes that p (and satisfies the first three conditions) and Ben
believes that not-p (and also satisfies those conditions). Neither one of them
has a justified belief in this case because there is a blatant disagreement
between two halfs of the population. Now, let us assume that a series of individuals
start arriving to the island one by one. All these individuals believe that p
(and satisfy the three conditions). Because these individuals previously
belonged to no epistemic community, they do not justifiedly believe that p.
But, now, when Chris (the first newcomer) arrives on the island, he and Ann
begin to have some justification for believing that p given that we now have a
majority view. When you add another newcomer David, who didn’t previously have
justification for believing that p, they become even more justified in
believing that p because the level of consensus is increased. So, on this view,
you get justification out of adding more people who didn’t have justified
beliefs before. Likewise, you can remove justification by adding people who
didn’t have justification for believing the opposing view. This I find
But, I do share the intuition that consensus matters. I just
think that there is a better view than intuitionism to explain why this is so.
Consensus can come about as a result of various different mechanisms. You can
get it by brain-washing (North Korea, Fox News), using the sword (crusades), or
open, reflective discussion in which all individuals from different standpoints
can take part in the game of giving and assessing reasons (Sweden in the 80s).
I have the intuition that the consensus arrived in the first two ways lacks any
epistemic value (given that the results can be wildly different), whereas the
consensus arrived at the third way matters.
So, I want to think that we form our normative beliefs in
social settings. Different social settings have different social
normative-belief-formation-systems that tend to gravitate towards consensus. If
one thinks that there are normative facts, then some of these social
belief-formation methods are more reliable than others. Which ones are more
reliable depends on what the normative facts are and which of the social ways
of belief-formation get them more often than not right. And, as a result, an
individual has a justified normative belief when her belief is formed through
education and reflection within a reliable social system (and it satisfies the
first three conditions from Sidgwick).
I haven’t given any arguments for this proposal, but here’s
one reason why I like it. The Crispwick view is by and large an internalist
view. Just by doing a priori reflection we can know that the first three
conditions are satisfied, and we can do little polling to see if there’s
consensus. So, on that view, it is easy to know whether we have justified beliefs
or not. In contrast, the reliabilist view I sketched is an externalist view –
the reliability of different social methods of belief-formation depends on the
external facts. Given that we have no independent methods of seeing what those
facts are, we do not know whether our own social method is reliable or not. So,
we do not know whether we have justified normative beliefs. I quite like this
modesty. One can still accept that we have justified beliefs that our normative
beliefs are justified, if we think that the social consensus formation system
we use is reliable more generally.