Who gets to decide what counts as knowledge?
Epistemic Institutions: Truth,
Power, Justice, and Democratic Legitimacy
Every institutional order depends on an epistemic
infrastructure: the institutions, practices, technologies, and power relations
that determine what counts as knowledge, which claims are considered credible,
whose experiences are heard, and how beliefs can be tested and corrected.
Science, education, journalism, courts, public administration, and digital
platforms do not merely transmit information. They organize society’s
relationship to reality and therefore shape institutional legitimacy, freedom,
justice, and equality[1].
Contemporary societies face an epistemic crisis
extending beyond misinformation. Shared reference points are weakening as
groups operate within increasingly separate informational and interpretative
environments. Epistemic plurality can improve understanding by introducing
different perspectives, but it becomes destabilizing when communities no longer
share procedures for evaluating evidence or correcting error. Truth then risks
becoming an expression of identity, strategy, or power rather than an intersubjectively
testable process[2].
This crisis has several mutually reinforcing
causes. Knowledge production is concentrated within universities, research
institutions, governments, media organizations, and technology companies.
Commercial and political interests affect which questions receive funding,
which findings become visible, and how information is framed. Digital platforms
prioritize engagement and attention rather than truth or public value. Social
identity and online echo chambers encourage internal confirmation, while
information overload pushes people towards simplification, heuristics, and
emotionally coherent narratives.
Trust in epistemic institutions is consequently
under pressure. Some distrust reflects genuine failures, including conflicts of
interest, exclusion, secrecy, or institutional errors. Other forms arise from
strategic disinformation and fragmented media environments. Trust should
therefore not be demanded unconditionally. It must emerge from transparent,
competent, inclusive, and correctable knowledge practices.
Epistemic power and inequality
Knowledge is inseparable from power. Epistemic
power determines which problems are recognized, which explanations are treated
as plausible, and who is authorized to participate in decision-making. It
operates through research funding, curricula, publication systems, media
agendas, technical standards, algorithms, language, and cultural
classifications.
Specialized epistemic elites are indispensable in
complex societies because reliable knowledge requires expertise. Expertise can
justify giving greater weight to a judgement within a defined field, but it
does not by itself authorize a binding political decision. Epidemiologists can
assess transmission risks, engineers can model system failure, and economists
can estimate distributional effects; elected and legally accountable
institutions must still decide how competing rights, costs, uncertainties, and
time horizons are balanced. Conversely, democratic authorization cannot make an
empirical claim reliable. Legitimate governance therefore requires both
epistemic independence and an explicit decision point at which political
responsibility becomes visible.
Yet unequal access to education, research
infrastructure, professional networks, and public visibility can narrow the
perspectives represented within knowledge institutions. When expertise becomes
socially closed or institutionally unaccountable, blind spots and distrust
increase[3].
Epistemic inequality also concerns access to
knowledge itself. Digital infrastructure, literacy, language, education, social
position, and economic resources determine who can obtain information and
evaluate it critically. Groups may be disadvantaged not only because they
possess less information, but because their experiences are not recognized as
legitimate sources of knowledge.
Formal institutions often privilege standardized
and quantifiable knowledge over lived experience. This can produce policies
that appear rational in abstract models but fail in practice because they
ignore how citizens, patients, workers, or local communities experience
institutional systems. Participatory research, co-creation, citizen science,
and systematic feedback can help connect scientific expertise with practical
and experiential knowledge.
Colonial histories have also shaped epistemic
hierarchies. Western knowledge traditions have often presented themselves as
universally valid while marginalizing Indigenous, local, oral, and non-Western
forms of understanding. Epistemic justice does not require treating every claim
as equally reliable. It requires that different knowledge traditions receive
fair consideration and are assessed through open procedures rather than
excluded by inherited hierarchies. Indigenous ecological knowledge and relational
traditions can, for example, complement scientific analysis by revealing
long-term and context-specific relationships that formal models may overlook.
Education is the central institution of epistemic
transmission. It determines who gains access to knowledge, which histories and
perspectives are included, and whether citizens develop critical thinking,
media literacy, and the ability to reason under uncertainty. Education can
reproduce epistemic inequality through unequal access and narrow curricula, but
it can also become one of the strongest mechanisms for pluralism and
intergenerational epistemic justice.
Knowledge, uncertainty, and responsibility
Science should be understood as an organized
process of doubt, testing, criticism, and revision rather than as a source of
infallible certainty. Scientific consensus is provisional, but this provisional
character is a strength: knowledge improves because its claims remain open to
evidence and correction.
Many policy problems including climate change,
pandemics, and emerging technology, contain irreducible uncertainty. Their
consequences are probabilistic, model-dependent, and affected by complex
feedback processes. Institutions cannot wait for complete certainty before
acting. They must communicate assumptions and limitations clearly, compare
risks, adopt precaution where appropriate, and revise policies as evidence
develops.
Epistemic diversity is functionally important
under such conditions. Scientific models, professional expertise, local
experience, and culturally situated knowledge can reveal different aspects of
complex problems. This plurality must remain connected to standards of evidence
and mutual testing; otherwise, it may deteriorate into relativism.
Knowledge processing is also emotional and
social. Fear, hope, distrust, resentment, and group identity influence how
information is interpreted. Under uncertainty, simple narratives and conspiracy
theories may become attractive because they reduce psychological complexity.
Effective epistemic institutions must therefore support dialogue, recognition,
and comprehensible communication rather than assuming that citizens process
information as purely rational actors.
Uncertainty is unevenly distributed. Socially
vulnerable groups often have less access to reliable information, less
influence over decisions, and fewer resources to absorb the consequences of
error. Epistemic responsibility therefore includes justice in the distribution
of informational risks. Scientists must be transparent about evidence and
uncertainty; media should avoid distortion and sensationalism; governments
should not use uncertainty to disguise inaction; and institutions should ensure
that vulnerable groups do not bear disproportionate costs.
Uncertainty may also be strategically abused.
Political and economic actors can exaggerate doubt to delay regulation or
protect established interests[4].
Epistemic corruption occurs when knowledge systems become structurally
oriented towards legitimating commercial or political objectives rather than
seeking and testing truth.
Digital epistemic infrastructure
Search engines, social media, recommendation
systems, cloud platforms, and artificial intelligence have become epistemic
institutions. Their technical architecture determines what information is
visible, in what order it appears, and which interactions are encouraged.
Algorithms are therefore not neutral tools but epistemic gatekeepers[5].
The attention economy creates a structural
conflict between commercial success and epistemic quality. Content that
generates outrage, fear, or confirmation often receives greater visibility than
nuanced and reliable information. A small number of companies control much of
the infrastructure, data, and computational capacity through which knowledge is
produced and distributed. This creates economic, technological, and epistemic
power at a scale that traditional democratic institutions often struggle to
regulate.
The concentration of digital infrastructure also
creates geopolitical dependence. Societies relying on foreign platforms, cloud
systems, and AI models have limited control over how their public knowledge
environments are organized. Digital sovereignty is therefore partly epistemic
sovereignty: the capacity of societies to shape the infrastructures through
which knowledge becomes accessible and authoritative.
Artificial intelligence can summarize complex
information, support multilingual access, integrate different sources, and
broaden participation. It can also reproduce historical bias, generate
plausible falsehoods, obscure sources, and make users dependent on systems they
cannot inspect. Its social effects depend on datasets, objectives, ownership,
accessibility, and governance.
The opacity of digital systems creates a serious
accountability problem. When algorithmic decisions affect access to public
services, employment, credit, or information, people must be able to understand
and contest those decisions. Black-box systems make it difficult to detect errors,
attribute responsibility, or obtain legal redress.
Digital platforms can also influence behavior
through personalization, microtargeting, psychological profiling, and choice
architecture. The distinction between assistance and manipulation depends on
transparency, user control, purpose, and the possibility of refusal or
correction. Political microtargeting is particularly problematic because it
fragments public debate into invisible and differently tailored messages[6].
Digital inequality extends beyond internet
access. It includes differences in literacy, visibility, data ownership,
computational resources, and the ability to participate in digital knowledge
production. Users collectively generate data, while control and economic value
remain concentrated among private actors. This can be understood as epistemic
extraction.
Yet digital systems can also strengthen epistemic
resilience through open data, public knowledge platforms, digital commons,
citizen science, multilingual tools, and faster correction of errors. The
relevant question is therefore not whether digital technology should be
accepted or rejected, but under which institutional conditions it can serve the
public production of reliable knowledge[7].
Moral orientation and pluralism
The epistemic crisis also concerns what societies
consider good, just, and desirable. Knowledge alone cannot determine collective
direction. Modern societies contain competing religious, philosophical,
political, and cultural frameworks. This pluralism supports autonomy and
emancipation but can weaken shared moral orientation.
Normative disagreement appears in conflicts
between present security and future ecological responsibility, individual
freedom and collective protection, national solidarity and universal human
rights, or economic efficiency and social justice. Access to these debates is
unequal, meaning dominant groups often have greater influence over the values
embedded in policy.
Institutions may also produce normative
corruption by selectively applying principles or presenting political
choices as neutral necessities. Digital platforms intensify moral conflict by
rewarding outrage and simplified narratives, although they can also facilitate
transnational solidarity and new forms of public deliberation.
The solution is neither complete relativism nor
the imposition of a closed moral doctrine. Institutions need procedures through
which normative claims can be articulated, challenged, justified, and revised.
Education should develop moral reflection and the capacity to engage with
difference, not merely transmit a fixed canon.
Principles for epistemic institutions
Legitimate epistemic institutions should be
organized around six design principles:
- Epistemic
justice: fair recognition of different
speakers, experiences, and knowledge traditions.
- Plurality
and testability: multiple perspectives
combined with shared procedures for evaluating claims.
- Transparency
and explainability: visible assumptions,
methods, interests, datasets, and decision processes.
- Correctability
and reflexivity: accessible mechanisms for
criticism, appeal, learning, and revision.
- Inclusion
and access: broad access to education,
information, participation, and knowledge production.
- Protection
against manipulation and corruption:
safeguards against commercial distortion, political interference, hidden
influence, and concentrated epistemic power.
These principles may conflict. Transparency can
endanger privacy, pluralism can complicate decisiveness, and inclusion may
reduce short-term efficiency. Institutional quality lies not in eliminating
these tensions but in managing them openly, fairly, and adaptively.
Practical implications include independent audits
of algorithms, explainability requirements, human oversight, public and
non-commercial knowledge infrastructures, media pluralism, protection for
independent science and journalism, participatory research, stronger media
literacy, transparent research funding, accessible appeal mechanisms, and
democratic control over data use.
Epistemic stability does not require uniformity.
It requires organized pluralism: enough diversity to expose blind spots
and enough shared testing and correction to sustain a common reality. Without
this foundation, democracy loses its deliberative capacity, institutions lose
legitimacy, and meaningful autonomy becomes impossible. Reliable knowledge
should therefore be treated as a public good and its maintenance as a
collective responsibility.
Would you like to explore how a relational
understanding of humanity can transform democracy, law, the economy, and public
institutions? Click the link below to read the full book, Becoming Human
Together, and discover a new framework for building a more humane,
democratic, and correctable society within ecological limits.
[1] Longino, Science as Social Knowledge; Robert K. Merton, The
Sociology of Science, ed. Norman W. Storer (Chicago: University of Chicago
Press, 1973).
[2] Hélène Landemore, Democratic Reason (Princeton, NJ: Princeton
University Press, 2013); Jürgen Habermas, Between Facts and Norms, trans.
William Rehg (Cambridge, MA: MIT Press, 1996).
[3] Harry Collins and Robert Evans, Rethinking Expertise (Chicago:
University of Chicago Press, 2007); Sheila Jasanoff, ed., States of Knowledge
(London: Routledge, 2004).
[4] Naomi Oreskes and Erik M. Conway, Merchants of Doubt (New York:
Bloomsbury Press, 2010); Justin E. Bekelman, Yan Li, and Cary P. Gross, “Scope
and Impact of Financial Conflicts of Interest in Biomedical Research,” JAMA
289, no. 4 (2003): 454–465, https://doi.org/10.1001/jama.289.4.454.
[5] Soroush Vosoughi, Deb Roy, and Sinan Aral, “The Spread of True and
False News Online,” Science 359, no. 6380 (2018): 1146–1151, https://doi.org/10.1126/science.aap9559.
This study concerns Twitter during a particular period and should not be
generalized to all platforms or societies;
Safiya Umoja Noble, Algorithms of Oppression
(New York: New York University Press, 2018); Emily M. Bender et al., “On the
Dangers of Stochastic Parrots,” Proceedings of FAccT ’21 (2021): 610–623, https://doi.org/10.1145/3442188.3445922.
[6] Frederik J. Zuiderveen Borgesius et al., “Online Political
Microtargeting: Promises and Threats for Democracy,” Utrecht Law Review 14, no.
1 (2018): 82–96, https://doi.org/10.18352/ulr.420.
[7] UNESCO, Recommendation on the Ethics of Artificial Intelligence;
European Union, Regulation (EU) 2024/1689, https://eur-lex.europa.eu/eli/reg/2024/1689/oj/eng.
Reacties
Een reactie posten