Oxford Scholarship Online: Epistemic Evaluation: Purposeful Epistemology

Regress-stopping and Disagreement for Epistemic
Neopragmatists
University Press Scholarship Online

Oxford Scholarship Online
Epistemic Evaluation: Purposeful Epistemology

David K. Henderson and John Greco
Print publication date: 2015

Print ISBN-13: 9780199642632
Published to Oxford Scholarship Online: October 2015
DOI: 10.1093/acprof:oso/9780199642632.001.0001
Regress-stopping and Disagreement

for Epistemic Neopragmatists
Jonathan M. Weinberg
DOI:10.1093/acprof:oso/9780199642632.003.0008
Abstract and Keywords
Suppose that our norms of epistemic evaluation are meant to

promote the goals of diachronic reliability—getting and
keeping hold of truths—and dialectical robustness—facilitating
fruitful discussions, collaborations, exchanges, and more
importantly, disagreements. What implications should these
goals have for our norms of justification? This chapter argues
for two main upshots. First, we should have a somewhat
demanding internalist requirement on justification, but with a
few key exceptions—for example, for beliefs that are
practically infallible (such as very basic arithmetic beliefs,
which are immediately checkable). Second, our two epistemic
goals motivate a version of conciliationism about peer
disagreement—mutual loss of justification should motivate
conflicting parties to deploy resources to resolve their
disagreement. But here too there is an exemption; namely,
Page 1 of 27
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
University College London; date: 26 May 2016
Neopragmatists
cases where neither party has any resources they could deploy
in the near future to help resolve the disagreement.
Keywords: justification, reliability, peer disagreement, epistemic regress,

pragmatism
I. A Brief Metaepistemological Preliminary

I want to begin by making a distinction between two families
of epistemological projects. The first are concerned with the
fundamental questions about how, or even whether, the mind
can gain any rational hold of the world around it. This family
includes such debates as those concerning classical forms of
skepticism; the grounding of chains of reasons and the
existence of “the Given”; the possibility of a priori justification;
and the factivity, and factorability, of knowledge. Such
projects typically do not draw in any substantive way on our
larger theory of the world, but can seem rather to be of a sort
that, if they can be carried out at all, can utilize only the most
absolutely general or purely structural considerations of the
human condition. Let us call this family of projects the
epistemology of morning, in its efforts to discern what
happens at that very moment as light dawns upon the human
mind. The epistemology of morning strives to take as little as
possible for granted, and often its internal methodological
debates concern just how dimly glinting a set of resources can
be deployed while still yielding valuable insights.
We can contrast those projects with a different family of

projects: to learn what particular norms should guide our
inquiries in the world as we find ourselves in it; to glean what
lessons can be cultivated from the specific successes and
failures of the histories of our many cognitive endeavors; or to
wrestle with the problems and puzzles that arise in the latest
chapters of those histories. The pursuit of such goals requires
one to be deeply informed by our most accurate theories of
human cognition, including its biological and social
mechanisms, and the particular causal workings of its
interplay with the environment. These projects take as
background that skepticism is refuted (or perhaps was a
nonstarter in the first place), and that somehow-or-other our
senses and intellects achieve significant purchase on the
Page 2 of 27
Neopragmatists
(p.187) world. Let us call this family of projects the

epistemology of midday—the dimness of the dawn is taken, for
its purposes, to be past, and one strives to look at the
epistemic world with as much illumination as can be mustered.
Without putting too much weight on what is already a
somewhat strained metaphor, I hope that the image of the
brightness of early afternoon also suggests much of the day’s
work is yet to be done, and nowhere near yet that
metaphorical period of dialectical resolution we often refer to
as “at the end of the day.”
One could argue that one or the other of these families is more
important than the other, and I suppose one could read
Quine’s (1969) “Epistemology Naturalized” as urging us to
abandon the former for the latter, and folks like Fumerton
(1995) as contending that there’s not much in the latter that
really counts as philosophy. Although I will be pursuing in the
rest of this chapter a piece of the epistemology of midday,
nonetheless I would stress that I’m not at all trying to contend
that the more traditional projects should be displaced, but
merely complemented. Indeed, I offer this distinction in hopes
that it might suggest a way in which naturalistic and not-so-
naturalistic approaches to epistemology can dwell together in
a less combative relationship.
Now, even once one has decided to pursue an epistemology of

midday, one still faces the question of how one is to proceed in
doing so. One must figure out how to deploy the sorts of
scientific, historical, and anthropological information that one
is now planning on making use of. We cannot just say vaguely
that we are doing very wide reflective equilibrium and leave
our methodological considerations at that—to do so would be
to ignore all the meta-methodological content that would be a
part of any such reflections. We will have only pushed the
question back: upon reflection, however wide, how ought we
to move forward?
To the extent that one can identify natural epistemic kinds,

perhaps one could follow Kornblith (2003) in looking to
transform some epistemological questions into ones we could
answer by consulting relevant scientists, such as cognitive
ethologists. But it is not at all clear that there are many such
Page 3 of 27
Neopragmatists
natural kinds, maybe not even much more than knowledge,

and it’s not at all clear that knowledge itself counts (Kumar
(2014)). Philosophers could also perhaps hope to use the
techniques of scientists themselves to investigate questions of
epistemological interest, in the manner of experimental
philosophy. (For overviews, see Knobe and Nichols (2008a,
2008b); Alexander (2012); Knobe et al. (2012).) But it is not
clear that such methods can tell us much other than
descriptive psychological facts about ourselves and other
persons; such facts may be part of what we want to appeal to,
maybe even an important or essential part, but the question
immediately at hand is how to go about using them. Being told
that we’ll get more of them, and gather them in the manner of
the experimental sciences, does not address that question. For
normative epistemology, experimental philosophy may be of
greater service in a debunking mode, in picking up on possible
foibles and biases in our cognition about epistemic matters,
than as a full-fledged epistemological method in its own right.
(Just to be clear, I am not sure (p.188) that any practicing
experimental epistemologist would disagree with that
assessment. Experimental philosophy is more often mistaken
for a supposed replacement method in philosophy, than it is
actually put forward as one.)
I hope that there will turn out to be a number of different,

complementary methods that we may ultimately be able to
deploy in pursuing these sorts of projects, but I will just offer
one such method here, which takes Edward Craig’s (1990)
Knowledge and the State of Nature as a touchstone of
methodological inspiration. I suggest that we proceed by
imagining that we are going to re-engineer some set of our
epistemic norms, given what we now take ourselves to know
about the kinds of creatures we are and the kind of world we
operate in. (A sort of epistemic constitutional convention, if
you will.) We ask ourselves first: for the set of norms in
question, what epistemic purposes are they meant to serve?
To focus on specifically epistemic purposes is part of what
keeps this program from collapsing into a more amorphous
brand of pragmatism, such as that which Susan Haack has
attacked Steve Stich for advocating. With such purposes in
mind, we can then ask: what norms would best serve those
Page 4 of 27
Neopragmatists
purposes? And indeed what norms would best do so for

creatures like us, in accord with our best current
understanding of how creatures like us work? I have
elsewhere called this method “reconstructive neopragmatism,”
and the method earns the moniker “pragmatist” in its
attention to the real consequences of our norms, and in that it
consults closely with human practical realities, albeit a special
subset of them that pertains to our particularly epistemic
needs. The “neo” is there to make clear that this approach
should not be pinned on any of the classic pragmatists of a
century or so ago. And the “reconstructive” is meant to
emphasize that this project may be substantially revisionary,
though likely not radically so. (See my 2006 and 2007 for
further discussion and applications of this methodology.)
II. Getting the Regress Started

Here is a rough version of a kind of principle that many
philosophers have found intuitively attractive, but not
uncontroversially so:
(D) To be justified in believing p, an agent must be able

to cite an appropriate reason in defense of p, and that
reason must itself be justified.
This formulation intentionally leaves a lot of key

epistemological questions unaddressed, such as what more
constitutes an appropriate reason for p, and within what time
frame such a defense must be presented. Also, it may be that
principles like (D) only come into play for assertions, not
beliefs themselves. I hope to be able to elide such particulars
here.
Under the neopragmatist rubric, then, we ask: does a principle

like (D) well serve the epistemic goals that norms of
justification are meant to promote? I will conjecture here, with
only a brief defense, that the norms of justification are meant
to promote jointly the goals of diachronic reliability and
dialectical robustness. We wish to get a (p.189) hold of truths,
and to keep a hold of them once we’ve gotten them, and part
of the job of norms of justification is to steer us towards such
truths, and to keep us on track once we’re there. Moreover,
Page 5 of 27
Neopragmatists
we wish to have fruitful discussions, collaborations,

exchanges, and even more importantly, disagreements. Norms
of justification thus also serve to structure the interactions
between epistemic agents so as to promote the mutual success
of such interactions, ideally to maximize the extent to which
we can learn well from each other, and minimize table-
pounding and foot-stomping. I will not concern myself here
with the question of whether the former of these two “DR
desiderata” is in some sense prior to the other; creatures like
us, in a world like ours, can best pursue the truth in
promiscuously interactive ways, and it will be highly unlikely
outside of pathological cases that a set of norms can do as
good a job as it can of promoting either DR desideratum
without also promoting the other.
These two certainly do not exhaust our epistemic goals on the

whole. For just one example, one could argue that securing
truths of significant practical importance is itself an epistemic
goal (Bishop and Trout 2005). And there is of course the whole
literature on pragmatic encroachment in knowledge. But we
are concerned here with justification, which is primarily a
norm of permission. It is not the job of justificatory norms to
direct the targets of inquiry, but to ensure that the inquiry
taking place meets appropriate standards of epistemic
hygiene. So we will restrict our attention here to the two
desiderata already identified, while acknowledging that there
is obviously room for much more debate as to which goals we
do or should have, and which are or should be central to our
understanding of justificatory norms.
Now, with these DR desiderata in mind, we can see why it

makes sense generally to require that agents be able to muster
good reasons on behalf of their beliefs, reasons that they not
only can cite publicly but also which they can thereupon
defend from further challenge as needed. On the front end,
beliefs that are already able to pass such a requirement at the
time they are first formed will be more likely to be true than
ones that are not so able. But just as importantly, after a belief
has been formed, satisfying (D) will allow us to respond
actively should any concerns be raised about it—either
because of dissent, or the uncovering of new evidence, or
perhaps simply because we now wish to put it to more
Page 6 of 27
Neopragmatists
demanding epistemic or practical work than we had asked of it

until now. We can inspect its credentials, and if we find them
wanting, we can apportion further epistemic resources either
to shoring them up, or to rejecting the claim. Having citable,
defendable reasons for one’s beliefs enables a kind of
epistemic bookkeeping for those beliefs, which in turn enables
epistemic auditing of a form which is tremendously valuable
when one finds oneself needing to pinpoint sources of error
and dissent, and either eliminate them or at least quarantine
them.
Another way to see how we might benefit from adhering to a

principle like (D) is to consider what sorts of creatures would
not particularly benefit from such a norm. Creatures with very
limited inferential or reflective powers would gain little from
such a norm, because they would not have any resources to
exploit the extra (p.190) information that they were retaining.
Whatever it is that they are getting in through their
fundamental sources of evidence, that’s what they’ve got, and
there may be little that can be done to make them better or
worse. If they at least have some capacities for self-
monitoring, they may be able to distinguish good belief-
forming circumstances from bad, and thus guide their
cognition somewhat, but further cognitive bookkeeping will
not help them. (We have good reason to think that at least
some other primates have this sort of cognition, and indeed
rats and even pigeons; see references and discussion in Shea
et al. 2014, especially Box 2, p. 187.) At the opposite end of
the spectrum, creatures with spectacularly strong powers of
coherence-based inference and a very broad sampling of the
facts of the world might not be particularly aided by such
bookkeeping, because they can simply survey their
accumulated beliefs, which ex hypothesi are drawn from a
sufficiently comprehensive stock of beliefs about the world,
and can discern whatever outliers they may have among them.
Yet we lie in between these extremes. Our inferential

ambitions vastly outstrip our sources of evidence, yet we have
nontrivial, but nonetheless fairly limited, powers of holistic
inference. And the sources of evidence that we have present
us with only a very slender slice of the universe over the
Page 7 of 27
Neopragmatists
course of one or even many lifetimes. We are thus at risk of

our belief-forming practices being dangerously path-
dependent, subject to the luck of the draw concerning what
initial inputs we manage to get a hold of to start our inquiries
moving. What seems likely to us at one point in time will
influence what we choose to investigate in the near future,
and what look like successes early on will have a deep impact
on what shape the rest of our inquiries take. This is a good
strategy for maximizing the range of truths we might acquire
using our limited resources, but it is not without risk; in the
context of urging philosophy to “do better,” Timothy
Williamson has rightly warned that, “[a] few resultant errors
easily multiply to send inquiry in completely the wrong
direction” (2007, p. 288). This is the human epistemic
condition, and it makes it all the more important, therefore,
that we be able to recover from such errors where they may
occur, lest an initial wrong turn lead into a permanent blind
alley. In other words, it will not do for us to try to avoid the
risk path-dependence in our cognition in the first place—we
can only try to manage it.
One of the key jobs that norms like (D) can help accomplish for
us is to manage these risks of path-dependence, by requiring
us to carry with us whatever aspects of our cognitive past we
are still looking to put to work. If we find ourselves in a dead-
end, then we will know how to rewind the string and navigate
back, and attempt a different path from there. And that
possibility points to another advantage of being guided by
something like (D): we can deliberately change our minds as to
what we will count as our reasons for p, by consciously
considering p and what reasons we may have available for it.
When we do so, we are thus in at least a local and partial way
freeing ourselves from what our history with p may have been
to that point. But we can only do so if both our set of candidate
reasons and the specific set of them we (p.191) currently take
ourselves to be deploying are all generally available to us for
consideration in the first place.
Nonetheless, if we try to make explicit defendability into an

absolute requirement, we will end up with a norm that no one
can actually manage to satisfy. This is not an in-principle
Page 8 of 27
Neopragmatists
argument, and I do not offer here any argument that an

infinite chain of reasons would fail to justify, were an agent
(mirabile cogitatu!) to deploy such a chain. Our reasons (pace
Leite 2005; Klein 2007) are finite. Moreover, our reasons can
run aground fairly quickly, and there are claims we would
make for which most of us have no idea how to respond to a
request for further reasons for them. Many of our beliefs are
such that almost all agents end up “dumbfounded,” should
they be asked to defend them (see, e.g., Haidt (2001), see also
references and discussion in Sauer (2011)). Whether reasons
must come to an end somewhere as a matter of conceptual or
metaphysical necessity, as merely a matter of fact for
creatures like us, they always do come to an end, and not
simply because we reach a point at which no further reasons
happen to be demanded. So we will need to discern where
such a requirement can be relaxed, consistent with promoting
the DR desiderata.
(Reconstructive neopragmatism, as an epistemology of

midday, starts off by considering skepticism not as any sort of
threat; nonetheless, skeptical concerns are not irrelevant to
our discussion, as they serve as a kind of quality-control—an
unfulfillable norm will not help guide our actions, and should
thus be rejected.)
There are a number of places in which it looks like our

capacity for reason-giving screeches to a halt. Some of these
can be located in terms of the content of the claims; basic
moral claims, for example, or Wittgensteinian “hinge
propositions.” Another locus of dumbfounding is at our
appeals to our basic sources. As many have argued, we can’t
defend them, certainly not without circularity. But we can’t
have an exemption for any and all putative sources, lest we
allow in far, far too much, and lose the advantages in terms of
both reliability (by admitting in grossly unreliable sources) and
dialectical robustness (by allowing different people or groups
to effortlessly admit different, perhaps wildly divergent
sources). We need a set of exemptions from the demands of
(D), but they need to be tailored to the DR desiderata. And if
we can find ways of restricting (D) consistent with promoting
the DR desiderata, then it will be hard to see why we should
Page 9 of 27
Neopragmatists
adopt the absolute version over a pragmatically preferable

restricted version. And I think that such restrictions are
available, under very particular sorts of circumstances.
III. Regress-stopping for Neopragmatists

One such circumstance is when the appeal to the source in
question is one that is practically infallible, and is recognized
to be so. By “practically infallible” I mean that it is
extraordinarily unlikely, in any sort of circumstances that the
agent is at all likely to find herself, that the source could be
mistaken. In using such a source, the agent is putting herself
at very, very little risk of being wrong in the first place;
moreover, by (p.192) making an appeal that is recognized to
be practically infallible, there will be no need to settle cases of
dissent, since we can require as part of these norms that any
contending parties acknowledge that appeals to such a source
are beyond challenge, except under extraordinary
circumstances. (Should we have reason to think that such
extraordinary circumstances actually obtain, then we will want
a meta-exemption: in the presence of specific evidence that
the appeal to the source is, in this instance, not practically
infallible, (D) comes back online.)
This practical infallibility exemption yields at least two

important appeals to sources as legitimately exempted from
(D). First, it will apply to many simple cases of the a priori,
such as basic arithmetic, uncomplicated instances of modus
ponens, and the cogito inference (whether or not that really
counts as a priori). Second, it will apply to the most central
cases of introspection, such as a report on a current severe
headache. Are our appeals to such sources perhaps more than
practically infallible, but indeed infallible simpliciter? I don’t
know, but I’m inclined to doubt it. I suspect that no
philosopher’s liver is so strong that no amount of good scotch
can possibly produce in them a state in which they could
misdiagnose their own sensations and/or fumble their way into
affirming the consequent. But one-hundred-percent,
metaphysically or conceptually guaranteed infallibility will of
course count a fortiori as practically infallible, so if such
sources are completely infallible, that will certainly not pose a
Page 10 of 27
Neopragmatists
problem for this loophole. But I see no advantage to be gained

by making full infallibility the requirement for the loophole.
But the version of (D) we get even softened with this loophole
is still surely too restrictive. We need something that actually
gets us perception—but perception isn’t infallible, not even
practically. We might try to do so in the manner of truly old-
school foundationalists, and try to springboard into perceptual
claims about the external world based on introspective claims
about our inner happenings. But then we would have to
require agents to be able to defend some further reason that
could license their inference from a claim about the contents
of their minds to one about the world outside of them. And I
don’t see how we can expect typical agents would have
anything like the cognitive resources to defend such a reason;
indeed, it’s not even obvious to me that typical agents would
have the resources to devise and articulate such a reason in
the first place. This would be a locus of dumbfounding, and is
thus precisely the sort of place where we need relevant
epistemic work to be performed by the norms themselves,
rather than requiring it of individual agents.
An obvious candidate to consider here is simple reliability.

This candidate is obvious both from its prominence in the
literature, and also as a further variant of the conditions I have
already considered: since I have already suggested that we
don’t need full-blown infallibility in our sources to merit an
exemption, but merely practical infallibility, why not weaken
the condition further, to something less than even practical
infallibility but still surely something well above a 0.5 truth
ratio? Why, from a neopragmatist perspective, should we
bother requiring people to defend their appeals to reliable
sources? Or, if we want to include a nod towards dialectical
(p.193) robustness, why should we bother requiring people to
defend their appeals to sources that are reliable and

recognized as such?
A reliable-source exception would not, in fact, properly serve

the DR desiderata. I suggested earlier that much of the value
in having a principle like (D) is not just its promotion of
reliability on the front end of belief formation, but also on the
back end of belief maintenance and revision—the help it would
Page 11 of 27
Neopragmatists
give us in tracking down errors, localizing disagreements,

reducing path-dependence, and so on. But mere reliability by
itself does not help much with these other components to the
value of (D). Merely having used a reliable source to form a
belief that p will not be of much help should one later have
concerns about p, or for that matter, if one has concerns about
the reliability of the source (even if ultimately misleading
concerns). And a very common location in which dialectical
robustness is called for is when agents disagree over the
trustworthiness of a particular source. A key difference
between the practically infallible and merely reliable is that
we have a nontrivial chance of facing controversy in our use of
a merely reliable source; for example, one may have had an
unusually bad track record with the source, leading one (even
if ultimately mistakenly) to suspect that it is not really so
reliable. So there is possible work here for publicly available
reasons to do, in service of merely reliable sources of
evidence, while no such aid is called for with the practically
infallible.
Moreover, sometimes we are in a situation where we want to

acknowledge a source as being generally reliable on a topic,
but not regarding the specific context or the particular
question at hand. Consider ordinary visual capacities in
funhouse conditions, or using a standard magnetic compass
while in proximity to an MRI machine, or a speedometer when
one suspects that the car shop put on the wrong size of tires. A
reliable-source exception would render inert the demand for
citable reasons at points where it is all too important that such
a demand be legitimately imposed. As noted above, much of
the value of having citable reasons is to be able to backtrack
and correct for errors that may arise in the course of inquiry,
and to give us avenues for discussion should dissent arise.
When legitimate doubts are raised, we want our norms of
justification to enforce an obligation that they be addressed.
Now, if one possessed a theory of how the source works, and
in particular what errors it can fall prey to and under what
circumstances, and if one furthermore had good records of the
circumstances under which one received its deliverances, then
one could perhaps go back and try to discern where any error
might have crept in. But, obviously, if one has all that, then (D)
Page 12 of 27
Neopragmatists
will be more than adequately satisfied without the need of any

sort of special exemption. The mere reliability of a source is
not sufficient for us to exempt appeals to it from (D)—we still
need some reassurance that our need to locate and correct or
quarantine errors in the source will be addressed. If we frame
such conditions only in terms of reliability, then practical
infallibility seems to be the weakest such condition.
(Two points that lie beyond our current investigation. First, a

very similar line of reasoning would, I think, reveal DR-related
problems of trying to use epistemic (p.194) conservatism
principles to weaken (D) as well. Second, there is likely a
further anti-reliabilist corollary to all these considerations:
namely, if one does have reasons to deploy in defense of
appealing to a source, should those reasons only defend the
mere reliability of the source, then they will not necessarily be
sufficient to secure that source’s trustworthiness.)
What we need to do, it seems, is look beyond different degrees

of reliability, and consider other ways that an appeal to a
source of evidence can allow us to monitor and accommodate
errors successfully, other than the possession of appropriately
citable defenses. If the advantage of practical infallibility was
that little risk of error is incurred in the first place, we might
consider whether there might be circumstances in which such
a risk is present—but which is nonetheless a risk that can be
managed in a fairly immediate way, in which we can easily
investigate possible errors, and in which potential dissent
could instantly be settled. Practical infallibility is like riding a
train with very little chance of deviating from the desired path;
this other sort of circumstance would be more like driving
with a great GPS navigational system, able to respond flexibly
in real time to any difficulties that may arise. In either sort of
circumstance, (D) would have little work to do for us, and
could safely be voided.
But this is the situation we are in with paradigm instances of

perception, such as when we report on the layout of our
immediate surroundings. We have excellent resources for the
detection and correction of error with perception in such
cases. Our percepts themselves carry substantial information
as to whether they have been formed in more or less error-
Page 13 of 27
Neopragmatists
prone circumstances; when the lights are dim, we experience

the world as poorly illuminated, and not, or not just, as gray.
Should a visual report seem amiss, it is very easy for us to look
closer (or step back), to turn the lights up (or down), to look
from different angles, to check with each other’s takes on the
scene, even to apply various aids such as rulers or color
swatches. Perception’s virtues are thus partly a matter of our
psychological hardware, and partly our folk practices that
have been honed over millennia. Moreover, both the hardware
and the practices are very broadly shared, and recognized as
such. So when someone appeals to perception, their
interlocutors can generally acknowledge when such an appeal
has been made correctly. Often they can even just check for
themselves.
So we can allow regresses of justification to terminate at

appeals to the deliverances of perception, within boundaries
that are well and mutually understood, without doing harm to
the DR desiderata. We risk little harm to the goal of diachronic
reliability, because perception is basically reliable, and when it
makes errors we are generally very good at noting that we
have done so, and in taking steps to discern the matter more
clearly. And we risk little harm to the goal of dialectical
robustness, because those we are in dialogue with can easily
check, immediately and on the spot, whether we have made
any such appeals in accord with our best perceptual practices.
No further bookkeeping is needed, because the checking that
such bookkeeping is meant to enable is something that we can
already do without. We thus have, in addition to the practical
infallibility exemption, an immediate checkability exemption.
(p.195) Two quick clarifications are in order. First, we need a

meta-exemption here as well, for cases in which we have
reason to think that our practices cannot operate in their usual
way, for example when one thinks that one has ingested
hallucinogens. Second, although this exemption is of clearest
relevance for our perceptual practices, it is at least possible
that it may apply to other putative sources of evidence as well.
At least some uses of memory, for the very-recent-and-so-not-
yet-very-much-changed past, may be able to pass through this
particular loophole.
Page 14 of 27
Neopragmatists
IV. Taking Stock

We have thus seen that, although principles of epistemic
responsibility like (D) have great promise for our epistemic
goals of diachronic reliability and dialectical robustness,
nonetheless we ought to allow some key exemptions—at a
minimum, an exemption for the practically infallible (and
recognized as such), and another for the immediately
checkable (and recognized as such). It is surely not an
accident that these categories roughly correspond to some
important traditional categories of foundational sources of
justification: the a priori and perception. Yet this account does
not merely re-certify those categories as they have been
traditionally understood, though that would be an interesting
enough result if it did. But the correspondences in question
are only rough, and there are zones where these exemptions
do not map cleanly onto traditional areas of epistemic
foundations. I have just mentioned one place where the
immediate checkability exemption may be broader than the
traditional category of sensory perception, namely the very
recent past. Another candidate would be some places that
satisfy the practical infallibility conditions without being
anything like “truths of reason.” In particular, some
necessities of a merely nomological strength from rudimentary
physics or folk psychology may be ones that are highly unlikely
to change, and yet are widely enough built into our innate
cognitive machinery to meet with sufficiently universal assent,
and thus sufficiently unlikely to be challenged.
Yet a further important difference between arriving at these

categories via the neopragmatist method, and the way that
they have been traditionally understood, is that the latter have
historically tended to generate further mysteries that need not
arise on the neopragmatist view. For example, if the
foundational status of perception is meant to be grounded in
something like its phenomenological immediacy or its own
special form of introspective infallibility, then pressures come
into play that would push the foundational area into a retreat
from the perceptual judgments about our actual physical
surroundings, and down into “sense-data” or the like, which
then have their own infamous epistemological difficulties.
Similarly, a priori justification has had a difficult time
Page 15 of 27
Neopragmatists
shedding its reputation as a form of mystery-mongering, as

can be seen recently in accounts like BonJour’s (1998) attempt
to rehabilitate an Aristotelian conception of the relationship
between our thoughts and the properties we are thinking
about. These sorts of questions simply fail to arise when one is
(p.196) engaged in the reconstructive project. After all, our
immediate environments are more checkable than our interior
sensations, as the durable and objective tend to be when
compared to the fleeting and subjective. (See Schwitzgebel
(2008).) And the very characteristics of the objects of a priori
justification that on the one hand can often make them
troublesome by the lights of traditional epistemology—that
they timelessly lie beyond the causal realms—on the other
hand make them excellent candidates for things to have
practically infallible cognitions about, since their truth is thus
itself timeless and unchanging.
There is no architectonic that specifies that these two sorts of

exemptions from (D) are all and only the ones that the DR
desiderata would motivate. We may well ask: are these two
sufficient for our purposes? Clearly we need to consider
further how, and in what way, appeals to testimony, to
memory of the more distant past, and especially to moral
intuition can be admitted. Neither of these two exemptions
seem a particularly good fit for any of those sorts of appeals
(with perhaps the exception of any sufficiently universally
acknowledged moral truths), but maybe the grounds for
making an appeal to them can be constructed from
information about the world that is let in via those exemptions.
That is, perhaps we can start from these foundations and build
from them to yet further ones. But I somewhat doubt that this
would work for testimony at a minimum, on general anti-
reductionist grounds. Nonetheless, there is important room for
exploration here about the relation between “monitoring”
conditions and the epistemology of testimony (e.g., Fricker
(1994, 2006); Goldberg and Henderson (2006)). Further
investigation is warranted.
Another area where I suspect yet a further source of regress-

stopping might be needed is that of holistic inference. Simple
deductive inferences will be good candidates for practical
infallibility, and perhaps chains of them can be treated
Page 16 of 27
Neopragmatists
iteratively. But such deductive steps make up only a small

fraction of our inferential lives. Their non-deductive close
cousins, like statistical syllogism, will perhaps not be too hard
to handle using very similar machinery, plus some added
conditions to accommodate defeasibility (e.g., Pollock
(1995)), since basic principles like “If 99% of all F’s are G, and
a is an F, then (very likely) a is a G” will be good candidates
for practical infallibility. But we need also to accommodate
inferences that we express in such terms as “taking everything
into account, p,” or “weighing up all the evidence, p.” And
contra Fumerton (1995), even if we could make out such
ginormous “If (all my evidence) then p” conditionals to be
somehow psychological attainable—which, whatever their
possible status might be in principle, they surely aren’t so as a
matter of contingent fact—they clearly would often fail the
conditions on practical infallibility. There is too much very real
potential for disagreement in such cases.
So, let’s consider what our methodological framework can say,

at least on a preliminary basis about how to handle
disagreements about what follows from sets of evidence, when
that is not a matter of tractable deductive entailment.
V. The DR Desiderata and Peer

(p.197)
Disagreement
I now want to show how these sorts of considerations play out
with regard to a different epistemological question: peer
disagreement. We have considered two sorts of exceptions to
(D): practical infallibility and immediate checkability. A key
rationale for each exception involved a lack of problematic
disagreement in those areas—either none was to be expected,
in the former case, or any that arose would be settlable at
once and in situ, in the latter. But there is an active debate as
to what should happen when such conditions are not met, and
we find agents disagreeing with each other about the degree
of justification for a proposition even when they do not
disagree about any particular pieces of evidence that either
might cite to the other in defense of their evaluation.
There are a large range of views on offer about this question

and its many variants, but for the purposes of demonstrating
Page 17 of 27
Neopragmatists
the neopragmatist approach in action, let me just address a

fairly simple version of the issue. I will follow Christensen
(2009) and Elga (2010) in marking the rough distinction
between “Conciliatory” and “Steadfast” approaches to what
our epistemic norms should recommend in situations of
epistemic peer disagreement. The former view requires the
parties to the dissensus to engage in some substantial degree
of credence reduction, though the category is broad enough to
include adherents to the so-called “Equal Weight” view, as
well as those with different approaches to how the retreat
from one’s initial state of belief should be conducted. The
latter view allows for one to “stick to your guns,” perhaps in
general or perhaps at least in circumstances where one is
actually right about the evidential question at hand. Just
considering these two broad policy options, we can ask how
they each seem to fare in terms of our DR desiderata.
First, in terms of diachronic reliability, we can see that the

Conciliatory approach may seem to put us at risk of a
substantial loss of truth, whenever either party happens to be
correct. For if you are correct in the inference you want to
draw from our total evidence, and I am incorrect, then there is
a truth that you may have been justified in believing but which
you no longer would be. But this potential veritistic loss is
perhaps offset by a symmetric potential gain: since I will be
taking my own incorrect inference to no longer be justified, I
am now in a position to recognize that further investigation is
required. And upon that further investigation, I will have the
opportunity to recognize the soundness of your initial view,
and thereby get to partake in the goodness of its truth as well.
Indeed, since you may be restored to your correct view as
well, after such investigation, your loss need only be a
temporary one.
Since the Conciliatory view is one that involves a loss of

justification, it is appropriate to ask whether there is a risk of
skepticism here (a standard concern about such views). Since
we are operating at midday, we can help ourselves to the fact
that disagreement is nowhere near so widespread as to
threaten to bring our epistemic house down in a wholesale
fashion. In terms of more local skepticisms, though, an area of
inquiry with a truly unmanageable level of peer disagreement
Page 18 of 27
Neopragmatists
is one for which (p.198) we should take seriously the

possibility that it is simply not on a sound epistemic footing,
and for which some form of skepticism, hopefully a temporary
one, would be appropriate. And recognizing that it is an area
in a state of methodological disrepair would be a key first step
towards repairing it. The Conciliatory view thus hits the sweet
spot here: no actual risk of general skepticism, but
appropriately able to reach legitimate lower-case-“s”-skeptical
results where they might be called for.
If the Conciliatory view looks at first like it might put too much
truth at risk, then unsurprisingly, the Steadfast view looks at
first like it may be stronger in veritistic terms, since an
objectively correct party will typically, on such a view, get to
retain its initial degree of justification. In cases where both
parties are normatively correct, perhaps because the evidence
legitimately permits a range of responses (see Kelly (2010)),
then this is clearly a good result. And in the many cases in
which at most one party is correct, any incorrect party should
perhaps, on such views, already be evaluated as lacking in
justification, since they have formed a belief that is
normatively inappropriate given their evidence. So, if lack of
justification is supposed to be a motive to further
investigation, then the incorrect party should already have it.
It looks as though perhaps a Steadfast approach can do at
least as well as a Conciliatory one, but without that short-term
loss to the correct party. Moreover, there is a riskiness to the
Conciliatory view in this context, in that while the further
investigations may lead to both parties coming to the truth, it
may also lead nowhere, or even, in a bout of bad epistemic
luck, lead both parties to get misleading evidence that causes
them to settle in favor of an incorrect view. (Such a result
would be particularly disappointing in a case where both
parties started with normatively acceptable beliefs, though
perhaps we should expect that the resulting state after
Conciliatorily resolving such a disagreement would itself also
be normatively acceptable.) The Steadfast view requires no
such risk on the part of someone who is already doing things
right.
But we must consider how the incorrect party’s lack of

justification is supposed to manifest itself to them, to have the
Page 19 of 27
Neopragmatists
requisite rationally motivating force. They are, ex hypothesi,

unaware that they are incorrect in this matter—and thus, ex
hypothesi, incorrect about their own justificatory status. What
a Conciliatory rule here would do, which a Steadfast one
would not, would be to provide a novel signal that something
has gone epistemically awry. Part of the neopragmatist
approach I am advocating here is that we must consider the
consequences of our trying to live by the rules that we are
engineering. And a rule that says “the incorrect party must
respond to that signal, but the correct party need not do
anything” would be a rule that would never actually lead to
any actions in cases of peer disagreement, since everyone
would think the first clause applies to the other guys, while
taking the second clause to be the one that applies to
themselves. So at best, it’s not as clear a comparison as the
preceding paragraph tried to make it out: perhaps the correct
party can retain their truth, but it will be at the cost of the
incorrect party missing an opportunity to cease to be
incorrect. And of course in cases where both parties are
incorrect, then everyone loses!
(p.199) There is a further cost to the Steadfast view here, a

kind of higher-order riskiness. In addition to incorrect parties
losing a possible signal to correct the inferences they are
currently inclined to draw, we also collectively lose a signal as
to a possible weakness in our modes of inquiry more generally.
Points of inquiry where we find ourselves encountering what
we might call inferential noise are places where we should be
looking to add to our existing epistemic infrastructure. In
these cases, even if someone is getting it right, in a larger
sense we are not. One might compare to situations in which
different labs try to measure the same value using the same
apparatus, and get different results. It might be that one
group is just misusing the equipment—or it might be that the
different labs are doing everything fine in terms of the
parameters they already know to watch out for, but where
some unanticipated factor is influencing the equipment. Points
of peer disagreement can be places for us to turn unknown
unknowns into known unknowns, but only on a Conciliatory
approach.
Page 20 of 27
Neopragmatists
In all, Steadfastness may buy a limited amount of present

truth, by shielding the inferences of folks who are getting it
right, but at the possible cost of future truth, both for the
incorrect folks and for the investigatory community more
generally. If I am right that it is diachronic reliability that we
should engineer our justificatory norms to serve, then the
Steadfast view is making the wrong sort of trade-off, and the
Conciliatory view the right one.
When we consider Conciliatory and Steadfast views in terms

of the desideratum of dialectical robustness, it also tilts in
favor of being conciliatory. On Conciliatory views, the mutual
loss of justification will also serve as a motivation to pursue
further discourse, precisely as part of that process of trying to
restore lost justification to whichever party may be initially
correct, and to rectify any initially incorrect parties. Moreover,
the divergent takes on the disputed inference are themselves
removed as potential premises in the discussion, since both
parties are no longer justified in that inference. As a result,
there is less danger of fruitless question-begging or talking-
past. But with Steadfast views, such discourse is somewhat
foreclosed upon, since in retaining my current level of
justification in the inference, I also thereby retain—or, at least,
would seem to myself to retain—justification for downgrading
you as an interlocutor; and mutatis mutandis for your
evaluation of me. We will each seem to have grounds to
consider the other as a source of testimony not to be fully
trusted on the matter at hand, even if perhaps only one of us
will truly have such grounds. That is not a condition that
promotes full and fruitful discussion and debate.
But wouldn’t Conciliatory views provide a kind of motivation

not to engage in dialectic at all? That is, if I never encounter
anyone with a divergent judgment from mine, then I need not
risk the justificatory loss that would come with such
encountered divergence. Here is a point where it is important
to remember that this particular chapter of our justificatory
norms is meant to operate in the context of the entire set of
such norms. Obviously, other norms will exist that will serve to
foreclose on such strategies of epistemic evasion. It would
clearly serve both DR desiderata, for example, to have norms
of intellectual responsibility that will require a (p.200) degree
Page 21 of 27
Neopragmatists
of fair exposure of one’s views and inferences to the rest of

one’s community. Having said that, I should note that fans of
Steadfastness should consider similar strategies for dealing
with the arguments I have been presenting in this section—
perhaps the flaws I am identifying in that view are ones that
could be more than compensated for with other aspects of the
entire system of justificatory norms.
Note that what is doing a lot of work here is the way that the
dialectic and diachronic elements are entwined with each
other: we improve our methods of inquiry over time in part by
registering peer disagreements as signals of places in which
our current methods could be improved. Thus is the spirit of
an epistemology of midday. We are not worried about
preventing skepticism from taking hold, as that battle was (to
mix metaphors) already won at morning. But we are worried
about how to improve our cognitive lot, and managing inquiry
in a way that makes room for, or better yet, substantially
facilitates such epistemic amelioration, is one of the core
“points and purposes of epistemic evaluation.” And the
neopragmatist method can bring to view ways in which
positions in the peer disagreement debate can have very
different consequences with regards to those points and
purposes.
Let me close this part of my discussion by invoking one further

interesting complication that can arise from the diachronic
nature of our epistemic interests. For there will be some cases
of disagreement in which the disagreeing parties will have no
chance of any sort of near-term resolution of their
disagreement. Where two parties disagree about the outcome
of a complicated arithmetic problem, we would expect that it
could be resolved by some repeated and careful re-calculation,
and if need be, the deployment of a calculator. The
Conciliatory loss of justification can in such cases serve well to
require of the disagreeing parties that they take such
measures to resolve their dispute. But for other cases, where
such an increase in cognitive attention and activity are
unlikely to do much to help, then perhaps Steadfastness would
be called for instead. I am thinking in particular of different
research communities with different methodological
commitments: the only way for us to find out which
Page 22 of 27
Neopragmatists
methodology is in fact superior is by allowing and indeed

encouraging the disagreeing parties to work within their rival
paradigms, in hopes of our ultimately, at perhaps a much later
date, improving our epistemic resources. (See my 2007 for a
defense of a form of epistemic relativism along similar lines.)
Our DR desiderata favor Conciliationism for most ordinary
cases, but may well prescribe Steadfastness in cases where
that can better serve our long-term goals of finding out what
are better and worse ways for us to investigate the world.
VI. Concluding Questions

A further question would need to be addressed: in order to
support the desideratum of dialectical robustness, we required
that the practically infallible or immediately checkable also be
widely recognized as such; we will thus need (p.201)
ultimately to address what sort of “recognition” is to be

required. Do agents need to have justified beliefs that the
conditions for an exemption are met? Or is it enough that the
agents in a given community all simply believe, perhaps
without further justification, a given set of appeals to a source
to have the requisite property? Indeed, is it even required that
they believe at all, justified or not, that the appeals have that
property? Or is it enough that they do a decent enough job of
distinguishing apt from inapt appeals, without any particular
explicit cognizing of the conditions themselves? I suspect that
this very weak and externalist construal of “recognition” will
be all that the DR desiderata require, but it is not in any way
an obvious result.
One upshot of our considerations here is that it is the nature

of our practices with our various faculties, even more than the
faculties themselves, that license the regress-stopping. If
someone were to make undefended appeals to introspection
outside of the practically infallible subset of its range, that
person will be asking for epistemic trouble. If someone had a
different set of practices with perception that lacked
checkability—perhaps relying on special, subjective capacities
like putative “aura seeing”—then the DR desiderata require us
to require them to defend such appeals. When it comes to
stopping the regress of reasons, it’s not so much a matter of
what you’ve got, but what you do with it.
Page 23 of 27
Neopragmatists
Finally, I would like to offer a closing thought on the potentials

for naturalism in epistemology, particularly in the context of
pursuing what I called projects in the epistemology of midday.
We are used to the idea of naturalism as a metaphysical
constraint on the sorts of entities one can admit into one’s
ontology; or as a methodological constraint as to how certain
sorts of propositions can be known; or perhaps also as offering
the promise of contributing additional information as premises
in philosophical discussion. None of these has played much of
a role in my arguments here, at most with perhaps a bit of
history of science lurking in the background of my views about
how dialectical robustness and diachronic reliability are
mutually entwined. Rather, I think we have perhaps not yet
tried fully to capitalize on naturalism’s promise as a source of
theoretical resources for philosophizing as well. From purely
armchair considerations, the kind of exemptions I have been
arguing for here might seem strange and ad hoc. My hope is
that the framework I am advertising for here is one in which
we can see that they are well motivated and, well, natural.
When we consider epistemic norms as abstract truths
delivered to us by intuition, then perhaps a certain elegant
unity and simplicity will seem appropriate to them. But when
we consider them as tools that are ours to configure as best
we can to help promote our epistemic goals, then we must
allow our norms to warp a bit to match the contours of the
crooked timber of humanity, and in doing so, we must expect
them to display a certain amount of trickiness and complexity.
We know to expect such complexity in, say, university degree
requirements and state traffic codes—we should expect no less
in epistemology.
References
Bibliography references:
Alexander, J. (2012). Experimental Philosophy: An

Introduction. Cambridge: Polity.
Bishop, M., and J. Trout (2005). Epistemology and the

Psychology of Human Judgment. Oxford: Oxford University
Press.
Page 24 of 27
Neopragmatists
BonJour, L. (1998). In Defense of Pure Reason: A Rationalist

Account of A Priori Justification. Cambridge and New York:
Cambridge University Press.
Christensen, D. (2009). “Disagreement as Evidence: The

Epistemology of Controversy,” Philosophy Compass, 4/5: 756–
67, 10.1111/j.1747-9991.2009.00237.x.
Craig, E. (1990). Knowledge and the State of Nature. Oxford:

Clarendon Press.
Elga, A. (2010). “How to Disagree about How to Disagree,” in

R. Feldman and T. Warfield (eds) Disagreement. Oxford:
Oxford University Press, pp. 175–86.
Fricker, E. (1994). “Against Gullibility,” in B. K. Matilal and A.

Chakrabati (eds) Knowing from Words. Dordrecht: Springer,
pp. 125–61.
Fricker, E. (2006). “Varieties of Anti‐Reductionism About

Testimony: A Reply to Goldberg and Henderson,” Philosophy
and Phenomenological Research, 72(3): 618–28.
Fumerton, R. (1995). Metaepistemology and Skepticism.

Washington, DC: Rowman & Littlefield.
Goldberg, S., and D. Henderson (2006). “Monitoring and Anti‐

Reductionism in the Epistemology of Testimony,” Philosophy
and Phenomenological Research, 72(3): 600–17.
Haidt, J. (2001). “The Emotional Dog and its Rational Tail,”

Psychological Review, 108: 814–34.
Kelly, T. (2010). “Peer Disagreement and Higher-Order

Evidence,” in R. Feldman and T. Warfield (eds) Disagreement.
Oxford: Oxford University Press, pp. 111–74.
Klein, P. (2007). “Human Knowledge and the Infinite Progress

of Reasoning,” Philosophical Studies, 134(1): 1–17.
Knobe, J. and S. Nichols (2008a). “An Experimental Philosophy

Manifesto,” in J. Knobe and S. Nichols (eds) Experimental
Philosophy. Oxford: Oxford University Press, pp. 3–14.
Page 25 of 27
Neopragmatists
Knobe, J., and S. Nichols (eds.) (2008b). Experimental

Philosophy, i. Oxford: Oxford University Press.
Knobe, J., W. Buckwalter, S. Nichols, P. Robbins, H.

Sarkissian, and T. Sommers. (2012). “Experimental
Philosophy,” Annual Review of Psychology, 63: 81–99.
Kornblith, H. (2003). Knowledge and its Place in Nature.

Oxford: Oxford University Press.
Kumar, V. (2014). “‘Knowledge’ as a Natural Kind Term,”

Synthese, 191: 439–57.
Leite, A. (2005). “A Localist Solution to the Regress of

Epistemic Justification,” Australasian Journal of Philosophy,
83(3): 395–421.
Pollock, J. (1995). Cognitive Carpentry. Cambridge, MA: MIT

Press.
Quine, W. V. O. (1969). “Epistemology Naturalized,” in

Ontological Relativity and Other Essays. New York: Columbia
University Press.
Sauer, H. (2011). “Social Intuitionism and the Psychology of

Moral Reasoning,” Philosophy Compass, 6: 708–21, 10.1111/j.
1747-9991.2011.00437.x.
Schwitzgebel, E. (2008). “The Unreliability of Naive

Introspection,” Philosophical Review, 117(2): 245–73.
Shea, N., A. Boldt, D. Bang, N. Yeung, C. Heyes, and C. D.

Frith (2014). “Supra-Personal Cognitive Control and
Metacognition,” Trends in Cognitive Sciences, 18(4): 186–93.
Weinberg, J. (2006). “What’s Epistemology For?” in S.

Hetherington (ed.) Epistemology Futures. Oxford: Oxford
University Press, pp. 26–47.
Weinberg, J. (2007). “Moderate Epistemic Relativism and our

Epistemic Goals,” Episteme, 4(01): 66–92.
Page 26 of 27
Neopragmatists
Access brought to you by: University College London
Page 27 of 27

Oxford Scholarship Online: Epistemic Evaluation: Purposeful Epistemology

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Oxford Scholarship Online: Epistemic Evaluation: Purposeful Epistemology

Uploaded by

Copyright:

Available Formats

Regress-stopping and Disagreement for Epistemic

University Press Scholarship Online

Epistemic Evaluation: Purposeful Epistemology

Print publication date: 2015

Regress-stopping and Disagreement

Abstract and Keywords

Suppose that our norms of epistemic evaluation are meant to

Keywords: justification, reliability, peer disagreement, epistemic regress,

I. A Brief Metaepistemological Preliminary

We can contrast those projects with a different family of

(p.187) world. Let us call this family of projects the

Now, even once one has decided to pursue an epistemology of

To the extent that one can identify natural epistemic kinds,

natural kinds, maybe not even much more than knowledge,

I hope that there will turn out to be a number of different,

purposes? And indeed what norms would best do so for

II. Getting the Regress Started

(D) To be justified in believing p, an agent must be able

This formulation intentionally leaves a lot of key

Under the neopragmatist rubric, then, we ask: does a principle

we wish to have fruitful discussions, collaborations,

These two certainly do not exhaust our epistemic goals on the

Now, with these DR desiderata in mind, we can see why it

demanding epistemic or practical work than we had asked of it

Another way to see how we might benefit from adhering to a

Yet we lie in between these extremes. Our inferential

course of one or even many lifetimes. We are thus at risk of

Nonetheless, if we try to make explicit defendability into an

argument, and I do not offer here any argument that an

(Reconstructive neopragmatism, as an epistemology of

There are a number of places in which it looks like our

adopt the absolute version over a pragmatically preferable

III. Regress-stopping for Neopragmatists

This practical infallibility exemption yields at least two

problem for this loophole. But I see no advantage to be gained

An obvious candidate to consider here is simple reliability.

defend their appeals to sources that are reliable and

A reliable-source exception would not, in fact, properly serve

give us in tracking down errors, localizing disagreements,

Moreover, sometimes we are in a situation where we want to

will be more than adequately satisfied without the need of any

(Two points that lie beyond our current investigation. First, a

What we need to do, it seems, is look beyond different degrees

But this is the situation we are in with paradigm instances of

prone circumstances; when the lights are dim, we experience

So we can allow regresses of justification to terminate at

(p.195) Two quick clarifications are in order. First, we need a

IV. Taking Stock

Yet a further important difference between arriving at these

shedding its reputation as a form of mystery-mongering, as

There is no architectonic that specifies that these two sorts of

Another area where I suspect yet a further source of regress-

iteratively. But such deductive steps make up only a small

So, let’s consider what our methodological framework can say,

V. The DR Desiderata and Peer

There are a large range of views on offer about this question

the neopragmatist approach in action, let me just address a

First, in terms of diachronic reliability, we can see that the

Since the Conciliatory view is one that involves a loss of

is one for which (p.198) we should take seriously the

But we must consider how the incorrect party’s lack of

requisite rationally motivating force. They are, ex hypothesi,

(p.199) There is a further cost to the Steadfast view here, a