COMPARATIVE STANDARD IN
Which epistemic value is the standard according to which we ought to
compare, assess and design instuonal arrangements in terms of their
epistemic properes? Two main opons are agent development (in terms
of individual epistemic virtues or capabilies) and aainment of truth.
The opons are presented through two authoritave contemporary
accounts-agent development by Robert Talisse’s understanding in
Democracy and Moral Conict (2009) and aainment of truth by David
Estlund’s treatment, most prominently in Democratic Authority: A
Philosophical Framework (2008). Both opons are shown to be unsasfactory
because they are subject to problemac risk of subopmal epistemic
state lock-in. The ability of the social epistemic system to revise subopmal
epistemic states is argued to be the best opon for a comparave standard
in instuonal epistemology.
Institutional epistemology is the study of the epistemic performance of social
epistemic systems, institutional arrangements governing over large and complex
populations of epistemic agents
. One of the foundational concerns of insti-
tutional epistemology is according to which property of the social epistemic
system must we judge its epistemic performance – the question, then, of com-
parative standard. The present text argues the ability of the social epistemic
system to revise suboptimal epistemic states should be regarded as a compar-
ative standard in institutional epistemology.
The dierence between the comparative standard, a methodological device
for achieving the task of being more likely to attain knowledge, and the task
of attaining knowledge must be clearly delineated. I will argue that the system
which is designed to be able to revise suboptimal epistemic states, as opposed
1 The population is comprised of individual agents, and presumably, of communities
of agents which may be understood as a single agent (Page 2008. For authoritative work
on social ontology, see List and Pettit 2011; Tuomela 2013; Gilbert 2014). For the pur-
poses of this text, I will however use the terms “agent” for individuals and “community”
division of cognitive
Original Scientific Article
Received: 15.12.2018. Accepted: 23.02.2019.
PHILOSOPHY AND SOCIETY
VOL. 30, NO. 3, 321–462
Marko-Luka Zubčić: Faculty of Humanities and Social Sciences, University of Rijeka; firstname.lastname@example.org.
STUDIES AND ARTICLES │ 419
to the one designed to nurture agent development or attain the truth, is more
likely to succeed in search for knowledge. Both agent development and attain-
ment of truth as comparative standards decrease the likelihood that the social
epistemic system will reach a less suboptimal or the optimal epistemic state
– a justied social normative commitment (Brandom 2001) to a true belief.
The present discussion diers from the one on “procedure-independent
standard of correct decision”, a concept of frequent attention in literature on
epistemic democracy (Peter 2016), inasmuch as the question of concern is not
whether the decision made through the democratic procedure should be judged
according to some such standard or it is epistemically and politically justied
by the procedure itself. The interest here lies in a broader inquiry in social
epistemology – according to which standard should we design and assess the
epistemic output of any large and normatively complex population governed
by a any institutional arrangement? While epistemic democrats will feature
prominently in this area of social epistemology, the new methodological con-
cept was needed to distance us from the particular debates in epistemic de-
mocracy, and to allow us a viewpoint from which we will be able to judge the
total epistemic merit of any social epistemic system.
The plan of the text is the following. First, the two most relevant candidates
for the comparative standard of social epistemic systems, agent development
and attainment of truth, will be presented and it will be shown how they fail
to escape the threat of suboptimal epistemic lock-in. While agent development
will be supported by the work of Robert Talisse, namely his epistemic capa-
bility perfectionism, attainment of truth will be discussed in relation to the
work of David Estlund and his political and epistemological treatment of the
claim that those who are more likely to attain truth should exercise political
authority over others. Secondly, the ability of the system to revise suboptimal
epistemic states will be derived from the objections to both agent development
and attainment of truth as the third candidate for the comparative standard in
institutional epistemology. By focusing primarily on the threat the rst two are
unable to systematically stave o, the third candidate eectively tracks how
conducive the social epistemic system is to learning. It will also be shown that
the ability of the system to revise suboptimal epistemic states as a comparative
standard can be recognized as supported by work in pragmatism and political
economy, as well as in line with recent developments of “negative approach”
to institutional design as argued for by Miranda Fricker.
2. The Assessment of Comparave Standards: Agent Development
and Aainment of Truth
2 .1. Agent Development as the Comparave Standard
The design of a social epistemic system based on agent development as com-
parative standard would posit that the superior social epistemic system is the
COMPARATIVE STANDARD IN INSTITUTIONAL EPISTEMOLOGY420 │ Marko-Luka Zubčić
one which allows for the best epistemic development of its individual agents.
This may include, for instance, development of individual epistemic virtues
or capabilities. I will present the case for agent development as a comparative
standard through a specic argument for epistemic capabilities perfectionism
featured in Talisse’s Democracy and Moral Conict. Two objections from the
division of epistemic labour to agent development as the comparative standard
will be presented – rst, that there can be a combination of “good” and “bad”
individual epistemic traits and behaviours which combined give a collective
ly better epistemic output then exclusively a combination of “good” individual
epistemic traits and behaviours; and second, that there can be epistemic traits
which contribute to the development of the agent but are unknown at the point
of assessment, and which therefore cannot be accounted for by the assessment.
Both objections point to the central threat of a suboptimal epistemic state lock-in
which social epistemic system designed with agent development as a compara-
tive standard cannot avert. While I will focus on a specic account for illustra-
tion and clarity, the objections presented can be used to argue against any design
of the social epistemic system based on this particular comparative standard.
2. 1. 1. Epistemic Capabilies
I will rst explicate Talisse’s account of epistemic perfectionism. While his
primary argument in Democracy and Moral Conict is itself highly relevant,
sound and elegant, particularly with regards to the discussions in the funda-
mental discursive nature of epistemic agents, the focus here is on the argument
for epistemic perfectionism with which Talisse is concerned in the second part
of the book (Talisse 2009: 156–192).
Talisse’s primary argument in Democracy and Moral Conict is that in or-
der for individuals to develop any kind of epistemic life they must be able to
exercise their capacities for reason-exchange – epistemic agents are dened
by being able to engage in reason-exchange. This is a sound pragmatist claim.
Talisse, furthermore, argues that democracy is the basic institutional arrange-
ment which allows the agents to do so (Talisse 2009: 79–154).Once faced with
the Agent Ignorance Objection which challenges the thesis that inclusive de-
liberation in democracy is epistemically valuable by presenting evidence of
individuals in the contingent historical circumstances of particular democrat-
ic regime (namely, citizens of USA in the beginning of 21st Century) seemingly
ignorant of a multitude of political and scientic facts, Talisse endorses a form
of epistemic perfectionism aimed at fuller development of agents’ epistemic
capabilities (Talisse 2009: 156–185).
The rst thing to notice is Talisse’s quick concession to the argument based
on the contingent historical ignorance of agents. Despite his initial argument
not hinging on agents’ being knowledgeable (his argument posits that without
being able to engage in reason-exchange agents cannot be referred to as epis-
temic at all), Talisse does, in the second part of the book, grow concerned about
agents’ lack of knowledge on the subjects they are engaged in reason-exchange
STUDIES AND ARTICLES │ 421
on. Let me, before going into Talisse’s defense against Agent Ignorance Ob-
jection, rst point out that from the standpoint of institutional epistemology
the objection of agents’ ignorance need be conceded as relevant. Epistemic
agents should be taken to be “constitutionally” (Hayek 1978: 5) and “irreme-
diably” (Hayek 1982: 12) epistemically suboptimal. They would be more igno-
rant if they did not engaged in reason-exchange. They are less ignorant if they
have a chance to engage in the social epistemic system. They, however, remain
ignorant either way. As Talisse himself recognizes, this is why there is a need
for the social epistemic system in the rst place2. Individual epistemic agents
have severely limited epistemic capacities, including the ability to recognize
relevant evidence, make use of relevant data and concepts, and develop good
inferential practices. Social epistemic system, where populations of agents en-
gage in epistemic activity, as opposed to particular individual agents, is less so.
Talisse’s reply to Agent Ignorance Objection is as follows. Given Sunstein’s
valuable epistemological insights on inferior epistemic output of isolated nor-
mative communities (Sunstein 2009), Talisse rst diagnoses the epistemic life
of agents in question as lacking in trans-normative interaction. Their exposure
to reasons and evidence beside those their communities provide is too low.
Normative pluralism is the feature of a superior epistemic system because it
reduces agent ignorance. This is a sound design understanding and a sucient
answer to the objection. Talisse, however, proceeds to argue for epistemic per-
fectionism aimed at agent development, and posits a list of capabilities (Talisse
2009: 173–177) the “state” should foster in the agents.
The argument in the next section is whatever the capabilities, the list of in-
dividually virtuous epistemic features is a awed attempt at designing a supe-
rior social epistemic system. I will not argue against Talisse’s capabilities them-
selves. Their content is beside the point here – the focus on individual epistemic
development itself runs the problematic risk of suboptimal epistemic lock-in.
2. 1. 2. Two Objecons from the Division of Epistemic Labour
Two objections from the division of epistemic labour to any list of individual
epistemic virtues or capabilities which arises from understanding agent de-
velopment as the standard according to which we should design and assess a
social epistemic system are:
1. There can be a set of practices or traits which cannot be understood as
“good” epistemic practices or traits at the level of the individual, but
which contribute to the superior epistemic output of the social epistem-
2 Talisse himself claims: “Our epistemic dependence is unavoidable because each in-
dividual has limited cognitive resources. Individually, we simply cannot inquire into
every matter that is relevant to our beliefs; we must at some point rely on the epistemic
capabilities of others, we must defer.” (Talisse 2009: 141) He, furthermore, rightly points
out that “(...) each of us epistemically depends on an entire social epistemic system.” (Ta-
lisse 2009: 142) He proceeds to argue for the epistemic perfectionism nevertheless.
COMPARATIVE STANDARD IN INSTITUTIONAL EPISTEMOLOGY422 │ Marko-Luka Zubčić
2. There can be a set of practices or traits unknown to any agent (including
the assessor) at t1 which contribute to the superior output of the social
epistemic system at t2.
The rst objection may be illustrated by Zollman’s work in agent-based
simulations in social epistemology. In “The Epistemic Benet of Transient
Diversity”, Zollman (2010) shows how the population in which information is
shared among agents with extreme priors (and thus who conserve strategies
or theories despite having evidence to the contrary available) is signicantly
epistemically superior to the one in which the agents lack such (non-virtuous)
individual epistemic traits.
The second objection is more general, and requires only the concession
that at any time of assessment there is a possibility of unknown individual
traits which could contribute to better epistemic output. Therefore, a popu-
lation with the set of traits T and an additional trait n, unknown at the time
of assessment (which represents any time of assessment), may epistemically
outperform a population with the set of traits T. Yet, our conceptual apparatus
allows only for the assessment of the population with the set of traits T to be
recognized as a superior social epistemic system, while in the same time the
population with the set of traits T is suboptimal in relation to the population
with the T+n set of traits.
What both of these objections present are cases of suboptimal epistemic state
lock-in, and they could be understood as a twofold form of a single central ob-
jection, namely the objection that agent development as a comparative standard
runs the problematic risk of suboptimal epistemic state lock-in. As Mayo-Wilson,
Zollman and Danks (2011) observe, the divergence of prescriptions for superi-
or individual and group epistemic performance, seminally argued by Kitcher
(1990), is among the founding insights of social epistemology. There can be a
conguration of individual epistemic practices which cannot be described as
“good” from agent-level perspective which produces better epistemic output
then does the conguration of exclusively “good” epistemic practices. Further-
more, there can be agent-level traits benecial to the epistemic development of
the population which are unknown at the time of design or assessment – and,
as both Elinor Ostrom (Ostrom, Hess 2007: 68) and Friedrich Hayek (1960:
414) have noted, judging the presently best state of knowledge as the standard
risks suppressing the optimal development.
2. 2 Aainment of Truth as the Comparave Standard
The other comparative standard featured prominently in the social epistemo-
logical debate is the attainment of truth. Epistemic democrats in particular
have a tendency to describe features of their favoured social epistemic system
as “truth-conducive” or “truth-tracking” (Gaus 2011: 273). Two objections to
attainment of truth as comparative standard are: a) it is either conceptually
empty without additional specication of the comparative standard accord-
ing to which we ought to nd the truth or knowledge tracked, which in turn
STUDIES AND ARTICLES │ 423
is the controversy presently under investigation; b) or the requirement fol-
lowing from attainment of truth as the comparative standard is that we dele-
gate epistemic labour to those agents who are most likely to attain the truth.
I will focus on the second objection, and provide a discussion on David Est-
lund’s work related to the question of authority of those who are more likely
to attain the truth. I will argue that his arguments against epistocracy are not
satisfactory, and that a stronger epistemological argument against expert gov-
ernance, and thus against attainment of truth as the comparative standard is
required and possible.
The objection from division of labour to delegation of epistemic labour to
experts is that it, again, exposes the social epistemic system to risk of subop-
timal epistemic lock-in. The pluralism required for superior epistemic perfor-
mance of a large and complex population must be redundant, and thus there
are agents who are more likely to attain the truth. By delegating the totality
of epistemic tasks to experts, the social epistemic system is lacking means of
contesting the epistemic state the experts have attained. While they are more
likely to attain the truth, the experts will not necessarily attain the truth. They
are, moreover, still epistemically suboptimal. Therefore, the epistemic state
the experts attained may as well be suboptimal. The agents less likely to at-
tain truth are denied any possibility at contesting the epistemic state due to
being denied access to any epistemic labour, and thus their dierent and dis-
agreeing normative strategies cannot oer any contribution to breaking the
consensus. The system build on attainment of truth as comparative standard
has no means of contesting its optimality3.
2. 2. 1. Strong Polical and Weak Epistemological Objecon to Epistocracy
The authoritative argument in institutional epistemology against the delega-
tion of epistemic labour to those agents who are more likely to attain truth is
Estlund’s objection to epistocracy not being “generally acceptable in the way
3 Truth is a controversial concept (particularly, of course, within epistemology). So
controversial, in fact, that it appears ill-advised to use it for the robust design of a social
epistemic system. Another objection to attainment of truth as comparative standard,
therefore, could be that it would lead towards too much controversy as to the nature of
this particular concept, and therefore the assessment could not even begin. However,
it could be argued that the design of the social epistemic system need not proceed ac-
cording to any particular controversial theory of truth, but, following Estlund’s “mini-
mal” conception of truth (Estlund 2008: 25), merely posit that the best social epistemic
system is the one which produces the claim “X is f” when X is f. The systematic ap-
proach to institutional epistemology, thus, need not deal with truth in the manner the
rst objection implies. It may merely posit truth-conduciveness (very roughly, the abil-
ity to produce “X is f” when X is f) as a formal feature of a certain procedure. I will con-
cede this point. Attainment of truth as the comparative standard need not be dened
substantially as to settle the discussions with regards to theories of truth. The problem
I will focus on is that the attainment of truth as comparative standard leads to dening
the substantial agent trait of being more likely to attain truth as the denitive reason to
delegate the totality of epistemic tasks to those agents that feature this trait.
COMPARATIVE STANDARD IN INSTITUTIONAL EPISTEMOLOGY424 │ Marko-Luka Zubčić
that political legitimacy requires.” (Estlund 2008: 7; Estlund 2003: 58) I will not
presently engage with the majority of the particularities of his complex and so-
phisticated work, but will solely focus on the aspects relevant for this inquiry.
Estlund’s account is focused on endorsing what he calls the Truth Tenet,
claiming that “there are true (at least in the minimal sense) procedure-inde-
pendent normative standards by which political decisions ought to be judged”
(Estlund 2008: 30), and Knowledge Tenet, claiming that “some (relatively few)
people know those normative standards better than others” (Estlund 2008:
30), while rejecting what he calls Authority Tenet, a claim that “(t)he norma-
tive political knowledge of those who know better is a warrant for their hav-
ing political authority over others” (Estlund 2008: 30). Estlund rejects the
Authority Tenet on political grounds, and argues democracy is epistemolog-
ically justied because it is “better than random and is epistemically the best
among those that are generally acceptable in the way that political legitimacy
requires.” (Estlund 2008: 7) Democracy is the best social epistemic system be-
cause it is most likely to attain the truth among those social epistemic systems
which can have political legitimacy.
Estlund is concerned exclusively with political and moral epistemic materi-
als – whereas I am concerned with epistemic materials in general. More impor-
tantly, Estlund’s argument is concerned primarily with political authority, and
only secondarily with epistemological value. He proceeds to argue against the
Authority Tenet on political grounds – Authority Tenet cannot hold, because
the authority of “those who know better” cannot be held politically legitimate.
Estlund’s larger understanding of attainment of truth as the institutional
epistemic comparative standard allows for expertism to reign supreme in in-
stitutional epistemic labour for epistemological reasons albeit forbidding it for
political. In “Why Not Epistocracy?”, for instance, Estlund appreciates “Mil-
lian” scholocracy epistemologically but nds it politically problematic, and
eectively concedes general epistemological labour to experts and retains the
political-epistemological labour as a domain of the democratic. From this, it
follows there is an ought simpliciter (Case 2016) which is known to few and
should be followed in the design of a social epistemic system which lacks po-
Estlund does make a particular epistemological objection to Authority Ten-
et-Demographic Objection (Estlund 2008: 215 – 219) – which states that contin
gent groups of experts may “carry” epistemic vices or suboptimal traits which
override their relative epistemic superiority to other agents in the population.
This is the case – however, not for any contingent reason of suboptimal indi-
vidual epistemic traits, but for the necessity of a less likely revision of subop-
timal epistemic state in cases of normative centralization. Estlund’s objection
is too weak and, moreover, cannot withstand the philosophical denition of
experts as those who are more likely to attain the truth, and thus remain rel-
atively epistemically superior despite any suboptimalities they may “carry”.
From the standpoint of the division of epistemic labour both of these claims
can be accounted for. It is the nature of knowledge that it is conditioned on
STUDIES AND ARTICLES │ 425
redundant normative pluralism – experts themselves can attain the truth if and
only if operating under the conditions of a redundant normative pluralism.
The Authority Tenet may then be rejected on epistemological grounds.
There may be an ought simpliciter and it may be known to the few, but they
cannot know it without epistemic input from the redundant diversity of inquir-
ers. Thus comparing the social epistemic systems according to the likelihood
of knowing ought simpliciter is at best uninformative in design due to its triv-
ial claim of there being an ought simpliciter and some agents being more likely
to know it, and at worst epistemically distortive if improperly interpreted to
have no epistemic reasons why not to delegate epistemic labour exclusively to
experts. I will now present this objection to attainment of truth as a compar-
ative standard in some detail.
2. 2. 2. Epistemological Objecon to Aainment of Truth as Comparave Standard
Attainment of truth, or tracking of truth, is obviously an epistemic task of pri-
mary importance. Knowledge is at least a justied normative commitment to
a true belief. However, this claim does not necessarily translate into the attain-
ment of truth being the proper comparative standard of social epistemic systems
– it would leave us with an uninformative or confusing standard. The inquiry
on “Which of these two systems are closer to truth?” would merely return us
to the original question of “How do you compare which is closer to the truth?”
We are concerned with the epistemological comparative standard, and thus the
standard according to which access to the commitment to a justied true belief
may be more available to the population. As I will argue, it will be more available
to the population which is more likely to revise a suboptimal epistemic state.
The other available answers to the question “Which of these two systems
are closer to truth?” could be “Which system has more agents who are more
likely to know the truth?” or “Which system is run by those who are more like-
ly to know the truth?” Thus, expert-governed social epistemic system may re-
sult from the comparative standard of attainment of truth, if we were not to
understand it trivially.
There is a wealth of empirical evidence (Hastie, Dawes 2001; Kahneman
2011; Gaus 2008) that experts can and do tend towards suboptimal epistemic
state lock-ins. However, the case here involves experts understood much more
stringently as those agents who are more likely to attain the truth at the as-
sessment point t1. Therefore, the claim is not that experts tend to get stuck at
suboptimal epistemic states, but that the social epistemic system which dele-
gates the epistemic labour to experts has no institutional mechanism to prevent
them from arriving at and retaining indenitely the suboptimal epistemic state.
The epistemological problem with delegation of epistemic labour to those
more likely to attain the truth is as follows. Note that “being more likely to at-
tain truth at assessment point t1” does not translate into “necessarily attaining
the truth at t1”. Therefore, the experts will not necessarily attain the optimal
epistemic state – they might attain and indenitely conserve a suboptimal
COMPARATIVE STANDARD IN INSTITUTIONAL EPISTEMOLOGY426 │ Marko-Luka Zubčić
epistemic state. Having delegated the totality of epistemic labour to experts,
however, the social epistemic system has no means of contesting whatever
epistemic state the experts have attained.
Experts could be understood more or less as one really smart person, both
in its skillful excellence and in its cognitive limitations – namely, the lack of
conceptual and computational resources if working with the same normative
strategy (Page 2008). Adding agents to the expert community may improve the
speed of computation and introduce some cognitive diversity (as exists with-
in any given population of agents [Landemore 2012]), but will not prevent it
from getting stuck at the local optimum – save adding normatively dierent
and thus non-expert agents. Normative pluralism is a condition of the discov-
ery of the ought simpliciter.
If the social epistemic system delegates the epistemic task to those epistem-
ic agents who are more likely to attain the truth, it denies itself any systematic
ability to recognize (and revise) suboptimal epistemic state, and thus denies
itself the hedging mechanism against such a state – if “hedging” is understood
as minimization of risk of a bad epistemic “bet” on a particular strategy for
overcoming the suboptimal epistemic state.
The superior social epistemic system, of course, still needs and should wel-
come experts – just as it needs and should welcome really smart people. They
bring individually and relatively superior (but system-level suboptimal) epistemic
material into the epistemic pool. The division of epistemic labour in the supe-
rior social epistemic system does not deny expertize nor does it deny the possibil-
ity of hierarchical relations within which experts hold higher social and epistemic
“positions” – it merely does not fully reduce the epistemic labour necessary
for the superior social system to epistemic labour done exclusively by experts.
The central objection to the attainment of truth as comparative standard is
thus the same as the one to agent development – it runs the problematic risk
of suboptimal epistemic state lock-in.
3. Revision of a Subopmal Epistemic State as the Comparave
In the analysis so far the suboptimal epistemic lock-in, the inability of popula-
tions of epistemic agents to revise suboptimal epistemic states, has been shown
to be the primary threat the social epistemic system faces. Thus, the ability of
the system to revise suboptimal epistemic states appears to be the quintessen-
tial epistemically superior feature. The system which exhibits this feature is
more likely to succeed in the search for knowledge. This ability of the social
epistemic system to revise the suboptimal epistemic state should therefore be
regarded as the comparative standard of social epistemic systems. I will call it
Modest Epistemic Comparative Standard, MECS for short.
MECS is a regulative standard which tracks the ability of the social epis-
temic system to learn and allows for the development of minimal conditions
STUDIES AND ARTICLES │ 427
for satisfaction of the justication criteria of knowledge. It is regulative as op-
posed to positive comparative standards of agent development and attainment
of truth because it does not posit or depend on a substantial epistemic doctrine
of agent traits (either in order to develop them or to identify those agents most
likely to attain the truth) but designs the system able to withstand their worst
suboptimalities. It doesn’t ask how excellent are its agents – it asks wheth-
er the system can escape the deepest ignorance of its best agents. Doing so,
it advances primarily a social epistemic system which is capable, in the most
robust manner, of moving away from epistemic ills, of upgrading its epistemic
state, and thus a system which is capable of learning.
MECS cannot favour any group of agents – 1) experts and agents with rec-
ognized epistemic virtues are less likely to revise the attained suboptimal epis-
temic state because they can be expected to form a consensus on a particular
betting strategy, and subsequently a particular epistemic state, without means
of evaluating and contesting that particular strategy and state, and 2) other
agents are less likely by default. Therefore, they are both less likely to revise
a suboptimal epistemic state apart then they are together. Instead of favoring
agents with certain properties, MECS favours redundant normative pluralism
and disagreement as epistemically benecial (and instrumental) developments
within a population. Furthermore, redundant normative pluralism presents
minimal conditions for satisfaction of the justication criteria of knowledge –
social epistemic system featuring redundant normative pluralism opens the
Epistemic Contributions of its members to contest, and thus makes it possible
for them to be justied. As J. S. Mill observed with clarity, the ability to revise
suboptimal epistemic states, “to be set right when it is wrong” (Mill 2003: 103)
is the fundamental epistemic feature of epistemically suboptimal agents. For
this ability to develop, and the deepest desperate epistemic state to be over-
come, mere experience is insucient and discussion is required (Mill 2003: 102).
The history of institutional epistemological thought thoroughly supports
MECS. It is sound, as already shown, from the perspective of the division
of epistemic labour – redundant normative pluralism is fully justied by the
maintenance of the ability of the social epistemic system to revise suboptimal
epistemic states. Friedrich Hayek posits the task of competition to be to show
“which plans are false” (Hayek 1982: 117), to reveal bad epistemic “bets” agents
make in the search for knowledge under conditions of irremediable and con-
stitutional ignorance. The economists working with dynamic complex nor-
mative systems in polycentric governance studies and New Institutional Eco-
nomics have long considered the ability of the system to learn and adapt to
be of central importance for it epistemic performance (Ostrom 20015; North
1990; for the overview of the subject of institutional change, see Kingston and
Caballero 2009). Pragmatism’s key innovations in epistemology4 are positing
the testing and contesting of normative commitments as baseline epistemic
4 For a more comprehensive understanding of the connections between Hayek and
pragmatists, see Aligica 2014.
COMPARATIVE STANDARD IN INSTITUTIONAL EPISTEMOLOGY428 │ Marko-Luka Zubčić
practices, the experience of error as central to epistemic development (Brandom
2001) and learning as the key feature of the superior institutional order. Final-
ly, MECS is in line with Miranda Fricker’s negative approach in epistemic in-
stitutional design (Fricker 2015). Elaborating the ideal epistemic institutional
arrangement requires focusing on threats to the social epistemic systems, the
diagnosis of and solution to the points of failure the population is prone to.
“The ideal social organism will have a well-functioning immune system, and
you cannot design one of those without a proper understanding of its suscep-
tibility to disease.” (Fricker 2015: 74)
The objection to MECS surely cannot rest on any argument against revision
of the suboptimal epistemic states as a key feature of knowledge-acquisition.
MECS however may be accused of being too modest. Revision of a suboptimal
epistemic state does not imply reaching an optimal epistemic state – a subop-
timal epistemic state may be revised into second suboptimal epistemic state.
The response to this objection is two-fold: 1) given the agents are epistemically
suboptimal, the attainment of the optimal state is never guaranteed, and there
cannot be a social epistemic system which guarantees it, but 2) the revision of
the second suboptimal epistemic state is possible only in the social epistem-
ic system designed to be able to revise suboptimal epistemic states. If the ob-
jection would to lead towards the design of the social epistemic system such
that it would be compared according to its ability to reach an optimal state, it
would merely lead towards attainment of truth as the comparative standard,
and is therefore subject to the same objection.
We should judge the social epistemic systems according to their ability to
revise suboptimal epistemic states, the ability to get “unstuck” from worst ig-
norance. The epistemically best performing population needn’t have the smart-
est members nor is led by the wisest ones. It must, however, be most likely to
recognize when it is wrong.
Social epistemic system designed or assessed according to the comparative stan
dard of agent development or attainment of truth give rise to the problematic
risk suboptimal epistemic state lock-in. Lowering this risk should be regarded
as the comparative standard of the institutional arrangement governing over
a large and normatively complex population in its search for knowledge. The
superior social epistemic system is the one which learns best.
Aligica, Paul Dragos (2014), Institutional Diversity and Political Economy: The
Ostroms and Beyond. New York: Oxford University Press.
Brandom, Robert (2001), Making It Explicit: Reasoning, Representing and Discursive
Commitment. Cambridge and London: Harvard University Press.
STUDIES AND ARTICLES │ 429
Case, Spencer (2016), “Normative Pluralism Worthy of the Name is False”, Journal of
Ethics and Social Philosophy 11(1).
Estlund, David (2008), Democratic Authority: A Philosophical Framework. Princeton:
Princeton University Press.
–. (2003), “Why Not Epistocracy?”, in Neomi Reshotoko (ed.), Desire, Identity and
Existence: Essays in honor of T. M. Penner. Academic Printing and Publishing,
Fricker, Miranda (2015), “Epistemic Contribution as a Central Human Capability”, in
George Hull (ed.), The Equal Society. London: Lexington Books, pp. 73–91.
–. (2009), “Can There Be Institutional Virtues?”, in Tamar Szabo, John Hawthorne
Gendler (eds.), Oxford Studies in Epistemology vol. 3. Oxford: Oxford University
Press, pp. 235–253.
Gaus, Gerald (2011), “On Seeking the Truth (Whatever That Is) through Democracy:
Estlund’s Case for the Qualied Epistemic Claim”, Ethics 121(2): 270–300.
–. (2008), “Is the Public Incompetent? Compared to Whom? About What?”, Critical
Review 20(3): 291–311.
Gilbert, Margaret (2014), Joint Commitment: How We Make the Social World. New
York: Oxford University Press.
Hastie, Reid and Robyn Dawes (2001), Rational Choice in an Uncertain World:
The Psychology of Judgement and Decision Making. Thousand Oakes: Sage
Hayek, Friedrich August (1982), Law, Legislation and Liberty: A new statement of the
liberal principles of justice and political economy. London: Routledge.
–. (1978), “Coping With Ignorance”, Imprimis 7(7): 1–6.
–. (1960), The Constitution of Liberty: Denitive Edition. Chicago: The University of
Kahneman, Daniel (2011), Thinking, Fast and Slow. New York: Farrar, Straus and Giroux.
Kingston, Christopher and Gonzalo Caballero (2009), “Comparing theories of
institutional change”, Journal of Institutional Economics 5(2): 151–180.
Kitcher, Philip (1990), “Division of Cognitive Labor”, The Journal of Philosophy 87(1):
Landemore, Helene (2012), “Democratic Reason: The Mechanism of Collective
Intelligence in Politics”, in Hélène Landemore, Jon Elster (eds.), Collective
Wisdom: Principles and Mechanisms. New York: Cambridge University Press,
List, Christian and Philip Pettit (2011), Group Agency: the Possibility, Design, and
Status of Corporate Agents. Oxford: Oxford University Press.
Mayo-Wilson, Conor, Kevin Zollman and David Danks (2011), “The Independence
Thesis: When Individual and Social Epistemology Diverge”, Philosophy of
Science 78(4): 653–677.
Mill, John Stuart (2003), “On Liberty”, in John Stuart Mill (ed), Utilitarianism and On
Liberty. Oxford: Blackwell Publishing, pp. 88–181.
North, Douglass (1990), Institutions, Institutional Change, and Economic
Performance. Cambridge: Cambridge University Press.
Ostrom, Elinor (2005), Understanding Institutional Diversity. Princeton and Oxford:
Princeton University Press.
Ostrom, Elinor, Charlotte Hess (2007), “A Framework for Analyzing the Knowledge
Commons”, in Elinor Ostrom (ed.), Understanding Knowledge as a Commons.
Cambridge: The MIT Press, pp. 41–83.
Page, Scott (2008), The Dierence: How the Power of Diversity Creates Better Groups,
Schools, Firms, and Societies. Princeton: Princeton University Press.
COMPARATIVE STANDARD IN INSTITUTIONAL EPISTEMOLOGY430 │ Marko-Luka Zubčić
Peter, Fabienne (2016), “The Epistemic Circumstances of Democracy”, in Michael
Brady, Miranda Fricker (eds.), The Epistemic Life of Groups. Oxford: Oxford
University Press, pp. 133–149.
Sunstein, Cass (2009), Going to Extremes: How Like Minds Unite and Divide. New
York: Oxford University Press.
Talisse, Robert (2009), Democracy and Moral Conict. Cambridge: Cambridge
Tuomela, Raimo (2013), Social Ontology: Collective Intentionality and Group Agents.
Oxford: Oxford University Press.
Zollman, Kevin (2010), “The Epistemic Benet of Transient Diversity”, Erkenntnis
Komparavni standard u instucionalnoj epistemologiji
Koja epistemološka vrednost je standard prema kome se trebaju upoređiva, procenjiva i
dizajnira instucionalna uređenja s obzirom na njihova epistemološka svojstava? Dve kla-
sične opcije su razvoj agenata (u smislu individualnih epistemoloških vrlina ili sposobnos) i
doszanje isne. Opcije su predstavljene kroz dva autoritavna savremena iskaza – razvoj
agenata kroz rad Roberta Telisija u Democracy and Moral Conict (2009), te doszanje isne
kroz rad Dejvida Istlunda, najistaknuje u Democrac Authority: A Philosophical Framework
(2008). Ovaj članak pokazuje da su obe opcije nezadovoljavajuće jer su podložne problema-
čnom riziku “zaglavljivanja” u subopmalnom epistemološkom stanju. Članak argumentuje
da je sposobnost sistema da revidira subopmalna epistemološka stanja najbolja opcija za
komparavni standard u instucionalnoj epistemologiji.
Ključne reči: instucionalni dizajn, podela kognivnog rada, pragmazam, upravljanje znanjem,
epistemološke performance, socijalni epistemološki sistemi