Full chapter text
Reason
Problem
At the centre of Horkheimer and Adorno’s joint project is an uncomfortable question: within the modern Enlightenment that
claimed to free human beings from fear, superstition, and subjection, is there a possibility that reason itself can become an
instrument of domination, calculation, and destruction? This question is not directed against science or reason; rather, it is a self-
critique of a form of reason that measures the efficiency of means while ceasing to ask whether the end is justifiable.
The early form of Dialectic of Enlightenment was prepared amid the experience of the Second World War, Nazism, exile, and
industrialised mass destruction. Because of this historical context, the book is deeply suspicious of optimistic narratives of
progress. If the expansion of knowledge, administration, production, and technology is not a guaranteed path to freedom, then it
is philosophically inadequate to equate ‘modernity’ simply with more science and more organisation.
The second problem concerns the broad use of the word ‘Enlightenment’. Horkheimer and Adorno do not treat it merely as an
eighteenth-century European movement; they extend the term to the much older tendency to use knowledge to classify the
unknown and make it predictable and controllable. This breadth strengthens their argument, but it also creates a risk: distinct
periods, societies, and forms of knowledge may be absorbed into a single narrative of domination.
The third problem concerns the relation between myth and reason. The ordinary narrative says that myth is superstition and
Enlightenment is its opposite. Horkheimer and Adorno argue that the relation is more dialectical: myth, too, attempts to
understand the world through order, naming, and repetition; and when Enlightenment turns every object into something
calculable, equivalent, and controllable, it can itself assume the form of rigid repetition and necessity.
The fourth problem is the domination of nature. The achievements of modern science cannot be denied; knowledge has
transformed human life in diagnosis, transport, communication, agriculture, energy, and countless other fields. But if nature is
regarded only as an object of use, human beings may begin to view their own bodies, desires, labour, and other people as means
as well. The philosophical examination of this possibility is a central part of their project.
The fifth problem concerns culture and autonomy. When industrially produced entertainment is criticised, a question arises:
does popular culture mould the individual into conformity, or does the individual reappropriate it with meanings of his or her
own? Horkheimer and Adorno’s analysis of the culture industry offers a sharp critique of standardisation, repetition, and
commodification; but if the audience’s activity is denied altogether, the critique itself can become elitist.
The sixth problem concerns the basis of critique. If nearly every form of modern reason is said to be contaminated by
domination, on what kind of reason does the critic rely in making this judgement? This question was later raised by many critics,
including Habermas. The chapter should therefore be read not merely as a collection of conclusions but as a philosophical
experiment struggling with its own contradictions.
Parallel question—‘How can reason criticise itself without collapsing into anti-scientism, worship of the past, or despair? While
seeing structures of domination, how can evidence, public reasoning, difference, human dignity, and the possibility of reform
still be preserved?’
Core Proposition
The core proposition of this chapter is that Horkheimer and Adorno do not reject modern reason; they criticise its narrowed
form. When reason becomes only the capacity for calculating means to ends, while the ends themselves escape social and ethical
scrutiny, even a highly rational organisation can serve inhuman purposes.
In Eclipse of Reason, Horkheimer clarifies this crisis through the distinction between ‘objective reason’ and ‘subjective reason’.
Older traditions of objective reason raised questions of justice, virtue, human ends, and the good life; subjective or instrumental
reason asks chiefly how efficiently a pre-given end can be achieved. Horkheimer’s concern is that as the latter form becomes
dominant, the question of the ‘right end’ may disappear from public reason.
Dialectic of Enlightenment gives this analysis a historical-philosophical form. The aim of Enlightenment was to reduce fear,
understand the unknown, and increase human autonomy; but when knowledge becomes a logic of total control, objects, nature,
labour, and ultimately human beings can be reduced to exchangeable units.
Their famous ‘myth–Enlightenment dialectic’ does not mean that science and myth have equal truth-value. The argument is that
both contain a tendency to bind the uncertainty of the world within an order; when the critical power of Enlightenment is not
applied to itself, it can harden into rigid systems, rule-governed repetition, and the dogma that ‘what cannot be measured is
meaningless’.
Adorno’s later negative dialectics is an attempt at a philosophical response to this problem. Concepts are necessary, but the
object is more than the concept. The term ‘non-identical’ points to the remainder that cannot be completely absorbed into
classification, name, number, or theory. The task of critical thought is not to abandon concepts but to remember their limits.
For this reason, Horkheimer and Adorno’s critique does not say ‘there should be no control over nature’. It asks: who determines
the purpose of control? Who bears its costs? How much qualitative difference remains in what is measured? Is technical success
being treated as equivalent to moral legitimacy?
The thesis of the culture industry is the cultural form of this broader critique. Entertainment produced within an industrial
structure may display apparent diversity while being governed by standardised forms, familiar formulas, advertising logic, and
the requirement of profit. In such a system, ‘choice’ may appear abundant while the actual cultural horizon remains narrow.
Parallel proposition—‘Reason is defended through its self-critique. Measure, but do not deny the existence of what cannot be
measured; seek efficiency, but ask whether the end is just; classify, but preserve the particularity of the person; adopt
technology, but examine its social value and distributive consequences.’
Principal Arguments
First argument—instrumental reason is not by itself complete reason. The capacity to answer ‘how?’ cannot replace the
questions ‘why?’ and ‘for whom?’ How efficient a railway system, data system, administration, or industry is constitutes one
question; whether its purpose is just constitutes another.
Second argument—the history of domination can extend from nature to human beings. The habit of treating external nature
merely as a resource can become connected with a discipline that represses internal nature—the body, desire, fear, and
sensibility. Horkheimer and Adorno’s language is often excessively broad, but the question of this relation remains relevant
today in environmental issues, labour, and the culture of health.
Third argument—exchange and equivalence can conceal difference. Numerical equivalence of value is necessary for markets,
classification is necessary for administration, and measurement is necessary for science. The problem begins when these useful
abstractions are treated as the whole truth of the object. Even where numbers are equal, experience, history, risk, and human
consequences can differ.
Fourth argument—the critique of myth is necessary, but so is the self-critique of Enlightenment. Evidence and reasoning are
indispensable against superstition; at the same time, a sentence such as ‘science says’ is not by itself a final argument, because
scientific conclusions are field-specific, method-dependent, revisable, and may be distinct from value judgements. Ethical and
political reasoning is needed in moving from scientific facts to public policy.
गजेन्द्र ठाकु र
Fifth argument—the culture industry sells familiarity and comfort through standardisation. When the structures of stories,
music, images, advertising, and news repeatedly create the same expectations, the audience begins to know in advance what it is
supposed to feel. This ‘pre-approved experience’ can reduce the space for surprise, difficulty, and self-reflection.
Sixth argument—pseudo-individualisation can substitute for real difference. Products, programmes, or styles may appear
different on the surface while their underlying structure remains the same. Today the question reappears in streaming
catalogues, social-media templates, recommendation feeds, and branded individuality; nevertheless, to call every popular form
false would be an unsupported excess.
Seventh argument—in the analysis of antisemitism, irrational hatred was connected with social structure, projection, the
authoritarian personality, and propaganda. The historical centre of this analysis was European antisemitism. Its broader lesson
is that a society can project its fears, failures, or contradictions onto a minority group; yet the specific history of every form of
hatred must be examined separately.
Eighth argument—negative dialectics does not say ‘abandon concepts’. Knowledge is impossible without recognition,
comparison, and conceptualisation. Adorno’s point is that the necessary violence of the concept—placing a particular object
within a general category—should remain self-aware. Thought should not exhaust its object completely; it should remain open to
difference and exceptions.
Ninth argument—‘determinate negation’ does not mean merely saying ‘no’. Critique reveals contradictions within a specific
claim, institution, or concept. If a society declares equality while access remains unequal, criticism does not impose an arbitrary
value from outside; it makes visible the distance between society’s own claim and its practice.
Tenth argument—Horkheimer and Adorno’s pessimism should be understood in relation to historical experience, not as an
eternal theory of human nature. Their severe language can be misread if detached from the background of fascism, war, exile,
and industrialised extermination. Yet historical situatedness does not abolish their questions; rather, it supplies criteria by
which they can be tested.
Eleventh argument—when critique calls every form of reasoning an expression of domination, it can cut away its own
foundation. If no standard of truth, evidence, dialogue, or better reasons survives, then the judgement ‘domination is wrong’
itself becomes merely another claim to power. Later Critical Theory works extensively on this normative deficit.
Twelfth argument—art is not mere decoration in the Horkheimer–Adorno tradition. For Adorno in particular, difficult and
autonomous art can preserve forms of experience that resist immediate use and market formulas by keeping a distance from
social standardisation. Yet autonomous art itself does not stand outside institutions, markets, and class-based accessibility.
Pūrvapakṣa
Pūrvapakṣa 1—‘This is a philosophy opposed to reason.’ If Enlightenment itself becomes myth, science is domination, and
concepts are violence, the result will be irrationalism. Without scientific achievement and public reasoning, modern rights,
medicine, justice, and democracy would be weakened.
Pūrvapakṣa 2—‘The history is excessively linear.’ To seek one logic of domination from ancient myth through modern science,
markets, fascism, and the culture industry may erase the specificity of different periods. Philosophical metaphor can become
heavier than historical causation.
Pūrvapakṣa 3—‘The critique of mass culture is elitist.’ To describe popular music, cinema, humour, serials, or digital content as
standardised illusion underestimates the audience’s enjoyment, creative reception, communal meaning, and capacity for
resistance.
Pūrvapakṣa 4—‘Economic analysis becomes weak.’ The broad civilisation-critique of Dialectic of Enlightenment pushes concrete
institutional analysis of class, production, labour organisation, and capital into the background. If domination is everywhere, it
can remain unclear what will change when a particular institution is changed.
Pūrvapakṣa 5—‘The positive standard of emancipation is unclear.’ If reconciliation, non-domination, or the desire to reduce
suffering remain only suggestive ideals, how are concrete decisions to be made in policy, institutions, and democratic conflict?
The critique is powerful, but the direction of reconstruction is less clear.
Pūrvapakṣa 6—‘Non-European history is absent.’ Colonialism, caste, slavery, empire, indigenous knowledge, multilingual
societies, and diverse forms of religion are not adequately centred in these foundational texts. Their universal claims therefore
cannot be applied to India or Mithila without revision.
Pūrvapakṣa 7—‘The language of domination over nature erases different uses of technology.’ Irrigation, vaccination, sanitation,
assistive devices for disability, and disaster forecasting are also knowledge-based interventions into nature, but they are not
necessarily oppressive. Use, ownership, consent, and consequences must be examined separately.
Pūrvapakṣa 8—‘Non-identity is excessively abstract.’ If the object always escapes the concept, this is useful as a general
epistemological caution; but what indicator, method, or decision does it change in social research? Its practical
operationalisation needs to be made explicit.
Uttarapakṣa
Reply to Pūrvapakṣa 1—the critique of reason is not opposition to reason but an expansion of its scope. Instrumental calculation
is necessary, but it is not sufficient. Science can provide reliable methods for facts; philosophy, ethics, and democratic reasoning
must add questions of ends, rights, and distribution.
Reply to Pūrvapakṣa 2—it is more fruitful to read Dialectic of Enlightenment as a ‘warning model’ than as a complete world
history. It shows the conditions under which an emancipatory institution can produce its opposite. Every historical case should
be examined on independent evidence.
Reply to Pūrvapakṣa 3—the culture industry’s severe generalisations should be corrected through empirical audience studies.
Examine ownership of production, algorithms, advertising, standardisation, and market concentration; but also include
evidence of audience use, remix, criticism, community, and resistance.
Reply to Pūrvapakṣa 4—bring economic structure back to the centre without making culture and psychology secondary.
Concrete variables such as ownership, labour, wages, debt, platform fees, the value of data, and advertising revenue can make
culture-industry analysis measurable.
Reply to Pūrvapakṣa 5—state the normative basis explicitly: reducing unnecessary suffering, human dignity, equal citizenship,
freedom of thought and expression, real options, non-arbitrary power, rights of appeal, and public reasons. Where these
standards conflict, open reasoning is required.
Reply to Pūrvapakṣa 6—the practical translation of non-identity is this: create categories, but examine residual cases; calculate
averages, but inspect distributions; build models, but test error subgroups; preserve diverse histories under one name; listen to
affected people’s own descriptions of themselves.
Reply to Pūrvapakṣa 7—in relations with nature, standards such as stewardship, reciprocity, precaution, and justice can be
added in place of domination. The technological question is not ‘intervention versus non-intervention’; it concerns benefit, risk,
irreversibility, consent, uncertainty of knowledge, and costs to future generations.
Reply to Pūrvapakṣa 8—critique must be applied to itself. If the critic regards the public as ‘stupid’, popular culture as ‘low’, or
his or her own theory as invincible, the critic can repeat the very logic of domination being criticised. Counter-position, revision,
and public testing should be indispensable conditions of Critical Theory.
गजेन्द्र ठाकु रक समानान्तर दर्शन — खण्ड २
Parallel reply-maxim—‘Do not narrow reason; make it multi-layered: empirical reason should test facts, practical reason should
test ends, public reason should test legitimacy, historical reason should test context, and self-critical reason should test the limits
of its own concepts.’
Indian Dialogue
In the Indian context, the first rewriting of Horkheimer and Adorno’s question comes through colonial modernity. Railways,
censuses, law, maps, education, industry, and bureaucracy were modern institutions; they increased some public capacities
while also becoming instruments of colonial control and classification. Therefore the simple formulas ‘modern institution =
emancipation’ and ‘modern institution = domination’ are both inadequate.
Gandhi’s Hind Swaraj raises sharp questions about modern civilisation, machinery, unrestrained desire, and the relation
between means and ends. The point of dialogue with Horkheimer and Adorno is the distinction between technological progress
and moral progress. But Gandhi and the Frankfurt tradition do not share the same philosophical sources, social programme, or
attitude to modern science; the comparison should therefore concern intersections of questions, not identity.
Ambedkar’s critique of caste provides a decisive correction for Critical Theory. Domination operates not only through capital,
administration, or mass culture, but also through hereditary hierarchy, social exclusion, religious legitimation, rules of marriage,
representation, and structures of educational opportunity. An Indian critical theory that treats caste as secondary will remain
socially incomplete.
In dialogue with Tagore’s critique of nationalism, one can see the tension between ‘organised collective power’ and living human
relationships. The administrative efficiency of a modern nation may be necessary, but when the nation itself becomes the final
moral end, individuals and diverse communities can be reduced to means.
The Constitution of India provides a positive public basis for reason: equality, liberty, dignity, religious freedom, the abolition of
untouchability, democratic representation, and legal remedies. Connecting Frankfurt’s negative critique with constitutional
standards in the Indian context clarifies not only ‘what is wrong?’ but also ‘against which public ideal is it wrong?’
In language policy, one form of instrumental reason may be administrative convenience alone—uniform forms, standardised
language, centralised examinations. But language is also a medium of life, identity, knowledge, and participation. Without a
balance between efficient administration and linguistic justice, smaller languages or dialects can become invisible.
In the Indian media market, culture-industry analysis is useful when ownership concentration, advertising, political access,
TRP/engagement, streaming recommendation, and the structure of language markets are examined. Yet because of vast
multilingual production, folk media, independent publishing, and diverse forms of audience agency, a uniform conclusion of
‘mass deception’ should be drawn cautiously.
Indian Parallel maxim—‘Do not become trapped in the opposition between accepting and rejecting modernity; examine each
institution separately—whose capacities it increases, whose voice it reduces, whom it classifies, what path of appeal exists, and
whether its declared purpose accords with constitutional dignity and equality.’
Mithila’s Parallel Perspective
In the context of Mithila, the self-critique of reason should first be applied to administrative and social classification. Categories
such as ‘village’, ‘caste’, ‘language’, ‘tradition’, ‘educated’, ‘backward’, and ‘modern’ are useful, but each contains internal
diversity. Produce data, but also write the history and limits of the categories.
In regional history, narratives of glory and narratives of decline can both become forms of standardisation. A ‘golden age’ story
can conceal inequality; a story that ‘everything is finished’ can render invisible living creativity, education, migration, women’s
participation, digital publication, and new economic forms. A non-identical perspective preserves exceptions and inconvenient
evidence.
Mithila’s manuscripts, Panji records, Nyāya tradition, folktales, songs, painting, theatre, e-journals, and digital archives do not all
belong to one cultural economy. Some are hereditary, some religious, some market-based, some voluntary, and some platform-
driven. Before applying the term ‘culture industry’, examine the actual structures of production, distribution, and reception.
Linguistic standardisation carries both benefits and risks. Standard spelling is useful for search, education, publication, and
digital processing; but if the standard is called ‘the only pure form’ and living regional forms are treated as inferior, the concept
dominates the object. The standard should remain a useful tool, not a weapon of social status.
The experience of migration reveals a new form of instrumental reason. Calculations of employment, wages, examinations,
credentials, and digital recruitment create opportunities for individuals; at the same time, qualitative dimensions of family time,
linguistic ties, care, insecurity, and identity may not appear in the table. The experiential costs of life should be studied alongside
economic gain.
Floods, rivers, embankments, irrigation, and water management make the question of domination over nature concrete. ‘Leave
nature alone’ is not a practical answer; ‘complete control over nature’ may also be an illusion. Local geography, old waterways,
scientific models, displacement, maintenance, distribution of benefits, and uncertainty should be studied together.
In education, rank, board results, and competitive examinations are measurable achievements, but they are not the whole of
education. Understanding through the mother tongue, critical capacity, creativity, knowledge of local history, library access,
digital facilities, and social confidence are also important. Do not turn an indicator into a complete synonym for the goal.
In digital Mithila, recommendation systems, short video, viral slogans, and engagement metrics determine the visibility of
cultural production. Algorithmic invisibility is possible for content in smaller languages. The solution is not complaint alone but
the creation of concrete structures such as metadata, open archives, searchable text, accessibility, multilingual tagging, and
independent editorial networks.
Mithila Parallel maxim—‘Measure, but write down what was left out; standardise, but preserve diversity; keep tradition, but
examine power; adopt technology, but also account for the qualitative costs to life.’
Contemporary Applications
In artificial intelligence, the question of instrumental reason is direct: what is model accuracy, latency, and cost? But go further:
on whom do errors fall more heavily? Is appeal possible? What permission exists for training material? How well are smaller
languages represented? Whose labour is replaced by automation? Efficiency does not take the place of moral legitimacy.
In algorithmic scoring, the practical meaning of non-identity is that a score is not a person. A credit score, risk score, ranking,
attendance index, or productivity score can be an indicator for decision-making; if treated as the person’s entire reality,
exceptions, context, and errors disappear. Human review and a right to reasons are necessary.
Social media is a new field for culture-industry analysis, but the old model should not be applied unchanged. Today users are
both consumers and producers; platforms earn through attention, advertising, and data extraction. Ownership, recommender
design, creator incentives, moderation, audience agency, and network effects must therefore all be examined together.
A streaming service may offer thousands of ‘choices’, but a few home-screen slots, the recommendation model, and licensing
agreements determine actual visibility. The number of choices and the structure of choice are different matters. To measure user
freedom, examine discoverability and switching costs as well as the number of available options.
In educational technology, an adaptive platform can improve a student’s practice, but when learning becomes only measurable
clicks, speed, and quiz scores, curiosity, discussion, writing, long reading, and creative failure can diminish. Technology should
remain a tool of pedagogy; pedagogy should not become an excuse for technology.
गजेन्द्र ठाकु र
In health systems, triage, imaging, analytics, and digital records increase efficiency. At the same time, the patient should not be
reduced merely to a ‘case’ or ‘risk category’. Qualitative dimensions such as pain, language, economic situation, responsibilities
of care, and informed consent are important alongside clinical data.
In environmental policy, cost-benefit analysis is useful, but things that are difficult to measure in money—biodiversity, cultural
sites, displacement, and risks to future generations—are not therefore zero. Instrumental calculation should be supplemented by
precaution, rights, distribution, and consideration of irreversible loss.
In workplace monitoring, a productivity dashboard can reveal bottlenecks, but continuous tracking can reduce autonomy,
privacy, and trust. Before indicators are created, purpose, data minimisation, retention, worker participation, and contestability
should be made explicit.
In publishing and academia, citation counts, rankings, and impact metrics can assist evaluation; but when they become ends in
themselves, subject choice, language, local archives, negative results, long-term editing, and public scholarship can be pushed to
the margins. A Goodhart-like problem is a contemporary empirical form of this critique.
In democracy, opinion polling, voter segmentation, and targeted messaging can increase the efficiency of political
communication; they also create a danger of reducing citizens to ‘predictable segments’. A healthy public sphere requires shared
facts, transparent finance, open debate, non-manipulative design, and access to opposing arguments.
Contemporary Parallel checklist—‘What goal is being optimised? Is the metric an adequate representative of that goal? What
value lies outside the metric? Who has been classified? How are exceptions treated? Who benefits? Who bears the cost? Is there a
right to challenge the decision?’
Chapter Conclusion
Horkheimer and Adorno’s self-critique of modern reason warns that knowledge, technology, administration, and markets do not
automatically produce emancipation. They can increase human capacities, but when ethical scrutiny of ends, qualitative
difference, and public accountability disappear, the same capacities can become instruments of domination.
It would be a mistake to read the ‘myth–Enlightenment’ formula of Dialectic of Enlightenment as equating science and myth. A
more appropriate reading is this: critical reason that does not criticise itself can harden into dogma. Enlightenment is defended
not by anti-Enlightenment but by self-enlightenment.
Eclipse of Reason clarifies the limit of instrumental reason: efficiency of means does not establish the justice of ends. Today the
practical value of this distinction is especially clear in AI, algorithms, platforms, policy measurement, education, health, and the
environment.
Adorno’s negative dialectics recognises the necessity of concepts while reminding us of their incompleteness. The contemporary
meaning of the ‘non-identical’ is that a person is more than a score, a language more than a standard, a society more than an
average, and history more than one narrative. Categories are necessary, but they are not final truth.
The critique of the culture industry is a powerful tool for understanding standardisation of production, market concentration,
and the attention economy; its limitation lies in underestimating audience agency and treating popular culture as uniform.
Today the theory needs to be reconstructed through platform economics and evidence about audiences.
The Indian and Mithila contexts add to this project the independent histories of caste, colonialism, language, multilingualism, the
Constitution, rural–urban relations, migration, and the environment. European experience offers useful questions, not ready-
made answers.
Parallel Philosophy takes three rules from this chapter: ‘Do not mistake efficiency for justice; do not mistake classification for the
object; do not mistake progress for automatic emancipation.’ It adds three positive rules: ‘Preserve evidence; preserve public
reasoning; keep the door to revision open.’
The final maxim—‘The enemy of reason is not reason but the narrowing of reason. Ask the purpose of a thought that knows only
control; show difference to a category that makes everything equivalent; demand an account from progress that conceals its
costs; and criticise a critique that itself becomes final truth.’