3. How SAE might try to get normative prescriptions from its descriptive core

Let’s suppose that a breakthrough in SAE results in wide agreement that

(say) a certain kind of foundationalism captures perfectly well the considered

epistemic judgments of proponents of SAE. Does anything follow

about how we ought to reason or about what beliefs we ought to adopt? The

proponent of SAE might argue that in the given scenario, it follows that our

cognitive efforts should be aimed at adopting empirical beliefs that are

basic or that are appropriately related to basic beliefs—related in the way

described by the account that accords with our considered epistemic

judgments. But why? It doesn’t matter how deeply philosophers may have

considered and refined their epistemic judgments. We still need to know

what’s so great about philosophers’ considered epistemic judgments.

Proponents of SAE might respond to this challenge as follows: ‘‘We

can connect the descriptive results of SAE with normative prescriptions by

noting that normative, epistemic claims are a priori. It is natural, therefore,

to suppose that figuring out the truth about epistemology will involve the

close analysis of our epistemic concepts. To characterize SAE as a descriptive

endeavor (as you have done) might be correct, but it is misleading.

The theories of SAE aim to describe an essentially normative

concept (or set of concepts). And that’s why SAE is normative. To put it

crudely, discovering conceptual truths involves the accurate description of

concepts. So discovering conceptual truths about the epistemological involves

the accurate description of epistemological concepts. And this is

precisely what SAE does. And so even though this endeavor is descriptive

(it involves describing our concepts), it nonetheless yields normative, a

priori prescriptions. It tells us what it really is for a belief to be justified,

and so what we ought to believe.’’

Let’s grant for the sake of argument that epistemic claims are a priori

(BonJour 2002). It doesn’t follow that SAE is the proper way to discover

such a priori truths. Given that proponents of SAE disagree with each

other about the nature of justification and that not all of these views can

be true, we can distinguish between a priori beliefs (that are true or false)

and a priori knowledge. We are willing to grant for the sake of argument

that the theories of SAE give us a priori beliefs. But why suppose that they

give us a priori knowledge? We cannot always assume that a priori truths

can be easily read off of our deeply considered judgments. The history of

mathematics shows how difficult it can be to come to grips with a priori

truths. The long-held belief that the sum of the internal angles of any triangle equals 180 degrees may have been a priori, but it was never

knowledge. If proponents of SAE have not properly grasped the concepts

of epistemic evaluation, then no amount of careful armchair contemplation

and analysis is going to succeed in uncovering a priori epistemological

truths.

The diversity findings of Weinberg, Nichols, and Stich discussed in

section 2 bring home the possibility that proponents of SAE are busy

analyzing the wrong epistemic concepts. If there are significant intercultural

and intracultural differences in our epistemic concepts, then not all

of these groups will be able to read off a priori truths from an accurate

account of their use of epistemological expressions (unless, of course, one

defends a crude kind of epistemological relativism). It may be that philosophers

actually have the right concepts. But we need some reason to

think so.

Our naturalistic perspective suggests another way to think about the

possibility that proponents of SAE are focused on concepts that will not

yield a priori knowledge. We share a common, folk understanding of

physics (Carey 1985, Spelke 1994). Our primitive physical concepts are

swamped by the perceptual accessibility of superficial features of objects.

Our folk judgments treat pulleys as objects that use rope and eyelets, inclined

planes as tilted surfaces, and screws as a stem with twisted threads.

Our folk understanding of the world does not allow us to see the basis of

the screw in the inclined plane, as physics experts do (Chi, Glaser, and

Rees 1982). Possession of a good theory allows us to classify certain

problems as tokens of a type (e.g., problems soluble by energy equations

and problems soluble by Newton’s laws of motion). But our folk notions

of physics have a powerful hold on us. For example, when novices are

asked to explain what forces are acting on a flipped coin, they typically

identify a diminishing upward force as the coin ascends, no forces (or

balanced forces) when the coin reaches its peak, and then increasing downward

forces as the coin drops (Clement 1982). These subjects employ a

‘‘folk’’ theory that is quite reminiscent of Aristotlelian physics. The flipped

coin acquires ‘‘impetus,’’ which explains its upward motion; but the impetus

is soon sapped by gravity, at which point the coin falls. This way of

understanding the problem—as involving two forces, the impetus upward

and gravity downward—is so natural that even many who have taken

Newtonian mechanics will describe the flipped coin in these terms. But

from the perspective of Newtonian mechanics, ignoring air resistance, the

only force acting on the coin is gravity, even when the coin is moving

upwards. And that force, for all practical purposes, is constant.

Suppose the naıЁve physicist were to sit down and carefully analyze his

concept of impetus. He refines, codifies, and harmonizes his impetus

judgments with great care and Austinian attention to linguistic detail. Now

he constructs an account that captures, with total accuracy, his application

of his concept of impetus. As an attempt to tell us the truth about something

other than his own linguistic predilections, lovingly detailing his

naıЁve concept of impetus is a waste of resources. (Actually, we can grant

that understanding the notion of impetus as it was employed by ancient

and medieval physicists can be of significant value in understanding the

history of science.) The attempt to clarify the central concepts of highly

successful scientific theories—gene, function, superposition, force—can be

of great value. What’s interesting and important about these notions, and

what draws our attention to them, is that they do (or presume to do) real

explanatory, predictive, and practical work in a successful theory about how

some aspect of the world works. Providing a careful account of a concept

can yield worthwhile results—but only when the concept is embedded in a

high-quality theory. And it is here where the challenge to SAE can be put in

sharp relief by comparing it to Ameliorative Psychology. Unlike the normative

judgments of Ameliorative Psychology, philosophers’ considered

epistemic judgments have been incubated in happy isolation from what we

have learned about how best to reason about significant matters.

As far as we can see, there is only one line of argument for the idea

that the theories of SAE are capable of yielding correct epistemic judgments

(a priori or not). The proponent of SAE must make a case for a

special kind of expertise for himself in matters of reasoning and belief. This

claim to expertise would presumably depend on the claim that the

methods of SAE have allowed philosophers to home in on what knowledge

or justification really is. Just as medical doctors spend years studying what

disease really is and so end up with an expertise in matters of disease that

others lack, philosophers with the appropriate training have an insight

into knowledge and justification that others lack. But if proponents of SAE

are experts about epistemological matters, then it is reasonable to suppose

that they have some kind of documented success. As far as we know, however,

SAE does not have a documented track record of success in epistemological

matters. (In fact, in the past half-century or so, SAE has not

converged on an ever smaller set of plausible theories but has instead

spawned brand new theories about the nature of justification [e.g., reliabilism

and contextualism]. This suggests that SAE is failing even to approach

realizing the limited, descriptive goal of capturing the judgments

of philosophers.) Besides the lack of a documented record of success, philosophers are not the only ones who study what’s involved in good

reasoning. Ameliorative Psychologists can also reasonably claim to have

some expertise in what’s involved in good and bad reasoning. Further, as

we have already noted, these scientists have many documented successes in

helping people and social institutions reason better about matters of great

importance.

Proponents of SAE have many of the social trappings of expertise. But

as far as we can tell, proponents of SAE are experts about a purely descriptive

domain: their own epistemological views (and the views of others

who have been similarly trained). Unless proponents of SAE can offer

some evidence for thinking that they have some kind of expertise that

makes their judgments about epistemic matters more worthy of trust than

the judgments of East Asians, Ameliorative Psychologists, or plumbers, we

just don’t see how the proponent of SAE is going to overcome the naturalist

challenge.

Let’s suppose that a breakthrough in SAE results in wide agreement that

(say) a certain kind of foundationalism captures perfectly well the considered

epistemic judgments of proponents of SAE. Does anything follow

about how we ought to reason or about what beliefs we ought to adopt? The

proponent of SAE might argue that in the given scenario, it follows that our

cognitive efforts should be aimed at adopting empirical beliefs that are

basic or that are appropriately related to basic beliefs—related in the way

described by the account that accords with our considered epistemic

judgments. But why? It doesn’t matter how deeply philosophers may have

considered and refined their epistemic judgments. We still need to know

what’s so great about philosophers’ considered epistemic judgments.

Proponents of SAE might respond to this challenge as follows: ‘‘We

can connect the descriptive results of SAE with normative prescriptions by

noting that normative, epistemic claims are a priori. It is natural, therefore,

to suppose that figuring out the truth about epistemology will involve the

close analysis of our epistemic concepts. To characterize SAE as a descriptive

endeavor (as you have done) might be correct, but it is misleading.

The theories of SAE aim to describe an essentially normative

concept (or set of concepts). And that’s why SAE is normative. To put it

crudely, discovering conceptual truths involves the accurate description of

concepts. So discovering conceptual truths about the epistemological involves

the accurate description of epistemological concepts. And this is

precisely what SAE does. And so even though this endeavor is descriptive

(it involves describing our concepts), it nonetheless yields normative, a

priori prescriptions. It tells us what it really is for a belief to be justified,

and so what we ought to believe.’’

Let’s grant for the sake of argument that epistemic claims are a priori

(BonJour 2002). It doesn’t follow that SAE is the proper way to discover

such a priori truths. Given that proponents of SAE disagree with each

other about the nature of justification and that not all of these views can

be true, we can distinguish between a priori beliefs (that are true or false)

and a priori knowledge. We are willing to grant for the sake of argument

that the theories of SAE give us a priori beliefs. But why suppose that they

give us a priori knowledge? We cannot always assume that a priori truths

can be easily read off of our deeply considered judgments. The history of

mathematics shows how difficult it can be to come to grips with a priori

truths. The long-held belief that the sum of the internal angles of any triangle equals 180 degrees may have been a priori, but it was never

knowledge. If proponents of SAE have not properly grasped the concepts

of epistemic evaluation, then no amount of careful armchair contemplation

and analysis is going to succeed in uncovering a priori epistemological

truths.

The diversity findings of Weinberg, Nichols, and Stich discussed in

section 2 bring home the possibility that proponents of SAE are busy

analyzing the wrong epistemic concepts. If there are significant intercultural

and intracultural differences in our epistemic concepts, then not all

of these groups will be able to read off a priori truths from an accurate

account of their use of epistemological expressions (unless, of course, one

defends a crude kind of epistemological relativism). It may be that philosophers

actually have the right concepts. But we need some reason to

think so.

Our naturalistic perspective suggests another way to think about the

possibility that proponents of SAE are focused on concepts that will not

yield a priori knowledge. We share a common, folk understanding of

physics (Carey 1985, Spelke 1994). Our primitive physical concepts are

swamped by the perceptual accessibility of superficial features of objects.

Our folk judgments treat pulleys as objects that use rope and eyelets, inclined

planes as tilted surfaces, and screws as a stem with twisted threads.

Our folk understanding of the world does not allow us to see the basis of

the screw in the inclined plane, as physics experts do (Chi, Glaser, and

Rees 1982). Possession of a good theory allows us to classify certain

problems as tokens of a type (e.g., problems soluble by energy equations

and problems soluble by Newton’s laws of motion). But our folk notions

of physics have a powerful hold on us. For example, when novices are

asked to explain what forces are acting on a flipped coin, they typically

identify a diminishing upward force as the coin ascends, no forces (or

balanced forces) when the coin reaches its peak, and then increasing downward

forces as the coin drops (Clement 1982). These subjects employ a

‘‘folk’’ theory that is quite reminiscent of Aristotlelian physics. The flipped

coin acquires ‘‘impetus,’’ which explains its upward motion; but the impetus

is soon sapped by gravity, at which point the coin falls. This way of

understanding the problem—as involving two forces, the impetus upward

and gravity downward—is so natural that even many who have taken

Newtonian mechanics will describe the flipped coin in these terms. But

from the perspective of Newtonian mechanics, ignoring air resistance, the

only force acting on the coin is gravity, even when the coin is moving

upwards. And that force, for all practical purposes, is constant.

Suppose the naıЁve physicist were to sit down and carefully analyze his

concept of impetus. He refines, codifies, and harmonizes his impetus

judgments with great care and Austinian attention to linguistic detail. Now

he constructs an account that captures, with total accuracy, his application

of his concept of impetus. As an attempt to tell us the truth about something

other than his own linguistic predilections, lovingly detailing his

naıЁve concept of impetus is a waste of resources. (Actually, we can grant

that understanding the notion of impetus as it was employed by ancient

and medieval physicists can be of significant value in understanding the

history of science.) The attempt to clarify the central concepts of highly

successful scientific theories—gene, function, superposition, force—can be

of great value. What’s interesting and important about these notions, and

what draws our attention to them, is that they do (or presume to do) real

explanatory, predictive, and practical work in a successful theory about how

some aspect of the world works. Providing a careful account of a concept

can yield worthwhile results—but only when the concept is embedded in a

high-quality theory. And it is here where the challenge to SAE can be put in

sharp relief by comparing it to Ameliorative Psychology. Unlike the normative

judgments of Ameliorative Psychology, philosophers’ considered

epistemic judgments have been incubated in happy isolation from what we

have learned about how best to reason about significant matters.

As far as we can see, there is only one line of argument for the idea

that the theories of SAE are capable of yielding correct epistemic judgments

(a priori or not). The proponent of SAE must make a case for a

special kind of expertise for himself in matters of reasoning and belief. This

claim to expertise would presumably depend on the claim that the

methods of SAE have allowed philosophers to home in on what knowledge

or justification really is. Just as medical doctors spend years studying what

disease really is and so end up with an expertise in matters of disease that

others lack, philosophers with the appropriate training have an insight

into knowledge and justification that others lack. But if proponents of SAE

are experts about epistemological matters, then it is reasonable to suppose

that they have some kind of documented success. As far as we know, however,

SAE does not have a documented track record of success in epistemological

matters. (In fact, in the past half-century or so, SAE has not

converged on an ever smaller set of plausible theories but has instead

spawned brand new theories about the nature of justification [e.g., reliabilism

and contextualism]. This suggests that SAE is failing even to approach

realizing the limited, descriptive goal of capturing the judgments

of philosophers.) Besides the lack of a documented record of success, philosophers are not the only ones who study what’s involved in good

reasoning. Ameliorative Psychologists can also reasonably claim to have

some expertise in what’s involved in good and bad reasoning. Further, as

we have already noted, these scientists have many documented successes in

helping people and social institutions reason better about matters of great

importance.

Proponents of SAE have many of the social trappings of expertise. But

as far as we can tell, proponents of SAE are experts about a purely descriptive

domain: their own epistemological views (and the views of others

who have been similarly trained). Unless proponents of SAE can offer

some evidence for thinking that they have some kind of expertise that

makes their judgments about epistemic matters more worthy of trust than

the judgments of East Asians, Ameliorative Psychologists, or plumbers, we

just don’t see how the proponent of SAE is going to overcome the naturalist

challenge.