Orientalism and the Production of the Other

This course has said four times that a canon is an institutional product and that exclusion carries an epistemic cost — with Ibn Khaldun (3.1.1), with Martineau (3.3.3), with Du Bois (4.5.5), and with the women reclassified out of the discipline (4.6.1).

Part 6 is where that stops being a caveat and becomes the subject.

And it opens with the book that made the argument unavoidable: Edward Said's Orientalism (1978), which showed that a vast, learned, respectable body of Western scholarship about the East had produced the object it claimed to be studying.

A description written by someone who had been there, and one written by someone who had not.

In 1834, Rifa'a al-Tahtawi, an Egyptian scholar sent to Paris as imam to a student mission, published an account of French society (4.6.3). It describes French political institutions, the press, the theatre, marriage customs, dress, the position of women, the education system and Parisian manners. It is observant, comparative, frequently critical, sometimes admiring, and constantly aware that he is describing one society from the standpoint of another.

Nobody in Europe read it. It was not translated into a European language for well over a century.

In 1835, Edward Lane published An Account of the Manners and Customs of the Modern Egyptians, which became the standard European reference on Egypt for generations, went through many editions, and shaped how educated Britons understood a country most of them would never see.

And here is the asymmetry that Said's book is about.

Al-Tahtawi's account is understood, when it is discussed at all, as an Egyptian's impressions of France — a document about its author.

Lane's account was understood as knowledge about Egypt — a document about its object.

One is testimony. The other is science. And the difference is not in the quality of the observation. It is in who was writing, from what position, backed by what institutions, in what language, and — decisively — with what power to make a description stick.

Said's argument is that this asymmetry was not incidental to European knowledge of the East. It was constitutive of it.

The problem Said poses is more radical than "European scholarship was biased," and it is important to state it precisely because the weaker version is what critics usually attack.

The weak claim: Western scholars held prejudices, made errors, and should have been more careful. This is true, uncontroversial, and not Said's argument.

Said's claim: Orientalism was a discourse in Foucault's sense (5.7.2) — a regulated system of statements that produced its object.

"Orientalism is not a mere political subject matter or field that is reflected passively by culture, scholarship, or institutions... it is rather a distribution of geopolitical awareness into aesthetic, scholarly, economic, sociological, historical, and philological texts... it is, above all, a discourse that is by no means in direct, corresponding relationship with political power in the raw, but rather is produced and exists in an uneven exchange with various kinds of power."

Three components:

An academic tradition — chairs, journals, societies, dictionaries, grammars, translations, museums. Vast, learned, and genuinely erudite.

A style of thought based on an ontological and epistemological distinction between "the Orient" and "the Occident."

And a corporate institution for dealing with the Orient — "by making statements about it, authorising views of it, describing it, by teaching it, settling it, ruling over it."

The last is the one that makes it a sociological argument. Knowledge and rule were not two separate things that happened to coincide. Governing required knowing, and knowing produced the object it governed.

What the discourse produced

Said's analysis identifies a stable set of oppositions running through European writing about the East from the eighteenth century onwards, across authors who disagreed about everything else.

The West is	The East is
rational	irrational, mystical
developed, dynamic	backward, stagnant, unchanging
masculine, active	feminine, passive
adult	childlike
self-governing	requiring government
historical	timeless
individual	a mass, a type

Note four features of this list.

It is a set of relations, not observations. Each term on the right is defined by its opposite on the left. The East is what the West is not — which is Saussure's differential meaning (5.7.1) operating on a civilisation.

The Occident is constituted by the contrast. Said's sharpest point: Orientalism was never principally about the East. It was how Europe produced its own identity — "the Orient has helped to define Europe (or the West) as its contrasting image, idea, personality, experience."

"Timeless" is the crucial one. If a society has no internal history, then it has no development of its own, and change can only come from outside. The denial of history is the justification for intervention.

And it operates as a type. The individual disappears; there is "the Oriental," about whom generalisations hold.

Latent and manifest Orientalism — Said's distinction, and it is what makes the argument survive counter-examples.

Manifest Orientalism — the explicit content: stated views about Islam, Arabs, India, China. This varies enormously. Some Orientalists admired what they studied; some devoted their lives to it; some were sympathetic critics of empire.

Latent Orientalism — the underlying structure of assumptions: that there is an Orient; that it is essentially different; that it can be known as a totality; that its own inhabitants cannot know it properly; and that the Western scholar can speak for it.

Said's argument is that the latent structure persisted across radically different manifest positions — that the admiring Orientalist and the hostile one shared a framework, and that admiration is not the opposite of the structure but a variant within it.

This is why "but some of them loved the East" is not a refutation.

The abstract argument becomes concrete in administration, and India supplies the best-documented case.

Before British rule, caste in India was — as this course has said repeatedly (3.1.2, 4.3.6, 4.6.3) — regionally varied, locally negotiated, occupationally fluid at the margins, and contested. Jati identities were numerous, local and shifting.

The colonial state needed to know its population — for taxation, for recruitment, for land settlement, for law. And from 1871 it conducted decennial censuses of enormous scope.

The census required each person to have one caste. Enumerators needed categories that were mutually exclusive, exhaustive and stable across a subcontinent. Officials consulted Brahminical texts — the varna scheme of four orders — because these were written, systematic, and available, and because the scholarly apparatus of Orientalism had established them as the authoritative account of Indian society.

And then the effects.

Groups petitioned to be reclassified. Once caste was an official category attached to entitlements, disabilities and recruitment eligibility, it became worth contesting. Caste associations formed specifically to lobby the census commissioner.

The categories hardened. What had been fluid became fixed by enumeration, repeated decennially, and increasingly treated as the description of a real thing.

And the ranking became official. The 1901 census under H. H. Risley attempted to rank castes by "social precedence" and included anthropometric measurement — nasal index and cranial measurements — as an indicator of racial origin, on the theory that caste reflected the mixture of Aryan and pre-Aryan stock. This was serious science by the standards of its day, conducted by a competent official, published by a government, and comprehensively wrong.

The scholarship on this is substantial and its central finding is agreed: Bernard Cohn's work on the census as an "investigative modality," Nicholas Dirks's Castes of Mind (2001), and Arjun Appadurai on "number in the colonial imagination". Caste was not invented by the British — this is the overstatement to avoid, and Dirks has been criticised for approaching it. But the form it took by 1947 — enumerated, ranked, all-India, textually authorised and politically mobilised — is a colonial product, and it is now the form in which it operates.

This is Foucault's power/knowledge (5.7.2) with an archive. The knowledge did not describe a pre-existing object. It produced one, and people then lived in it.

Said's argument, and its critics

Orientalism is one of the most influential books of the twentieth century and it has substantial defects. Both facts belong here.

The criticisms that landed:

It is empirically selective. Said's archive is principally British and French, principally literary and philological, and principally about the Arab Middle East. German Orientalism — the largest and most philologically serious tradition in Europe, and one belonging to a country with almost no Middle Eastern empire — is barely discussed, which is awkward for a thesis linking knowledge to colonial rule. Robert Irwin's For Lust of Knowing (2006) is the most sustained attack and documents errors of fact, mistranslation and misattribution.

It homogenises. Said treats a diverse tradition across two centuries as a single discourse, and in doing so does to Orientalism something close to what he accuses it of doing to the East. Sadik Jalal al-Azm — an Arab Marxist philosopher, not a defender of empire — made this point in 1981 and called the result "Orientalism in reverse": an essentialised West opposed to an essentialised East.

It is theoretically inconsistent. Said uses Foucault's discourse theory and Gramsci's hegemony, which do not sit together comfortably, and he simultaneously insists on the humanist value of individual authorship and the truth of representation — which Foucault's framework denies. Critics on both sides noted this; Said acknowledged the tension.

And it leaves no room for agency. If the discourse is total, how did anyone think otherwise, and how is Said writing? This is the same objection made to Foucault (5.7.2), and Said's later work — Culture and Imperialism (1993), with its concept of contrapuntal reading and its attention to resistance — is partly an answer to it.

What survives all of it, and this is the substantial point: that European knowledge about colonised societies was produced within a relation of power, that it constructed its object in ways that served governance, and that the resulting categories outlived the empires and are still in use. No serious scholar disputes this now, and before 1978 it was not the standard view.

The general mechanism

Strip away the specific archive and Said's argument has a form that applies well beyond his cases.

One: the observer's position is erased. Knowledge produced from a particular standpoint is presented as knowledge from nowhere. This is Haraway's "god trick" (5.6.1) and Harding's target (5.6.1), arrived at independently.

Two: the object is essentialised. Diversity and internal conflict are flattened into a type with a nature.

Three: history is removed. The society is described in a timeless present tense, which makes its condition an expression of its essence rather than an outcome of processes — including processes the observer's own society caused.

Four: the object is denied the capacity to know itself. Local accounts become data; the external account becomes analysis. Al-Tahtawi's testimony against Lane's science.

Five: the knowledge becomes administratively useful, is embedded in institutions, and produces the categories in which people are then governed and eventually understand themselves.

And six: the categories outlive the arrangement that produced them. This is the most consequential feature and the one most relevant to a sociology course. Empire ended. The categories did not.

And now the test this course owes: does the mechanism apply to sociology itself?

Run the six points against the discipline as Parts 4 and 5 presented it.

Position erased: Weber's comparative religion studies (4.3.10) frame the question as why the West developed rational capitalism and other civilisations did not — taking one outcome as the norm and others as absences, from a standpoint not examined.

Essentialised: "Confucian society," "the Hindu social order," "traditional society" as an analytical category.

History removed: modernisation theory's "traditional" societies awaiting development (5.1.3); Marx's Asiatic mode of production (4.1.1); the ethnographic present tense.

Self-knowledge denied: the Atlanta Studies unread (4.5.5); Ambedkar's caste theory absent from the canon while Weber's textual account was standard (4.6.3); Alatas's captive mind describing scholars trained to answer other people's questions.

Administratively useful: the census, the tribal classification, the criminal tribes legislation, the development index.

Categories outliving the arrangement: "developing country," "traditional," "the West," and the entire apparatus of comparative measurement that takes one trajectory as the scale.

Every point lands. Which is why Part 6 exists, and why it is a Part of this course rather than an appendix to it.

Go deeper

The colonial production of knowledge

Beyond Said, a substantial body of research on how colonial states knew their populations.

Bernard Cohn's "investigative modalities" — the historiographic, observational, survey, enumerative, museological and surveillance modalities through which the British came to know India. His central point: these were not neutral instruments but forms that produced particular kinds of knowledge and particular kinds of object.

The Criminal Tribes Act (1871) is the sharpest case, and this course flagged it in 3.2.7. Whole communities were designated as criminal by heredity, registered, restricted in movement, and subjected to surveillance — on a theory that combined European criminology, caste as understood through Brahminical texts, and administrative convenience. Around thirteen million people were covered by the time it was repealed in 1952. The communities are still called "denotified tribes" and are still disadvantaged, which is the category outliving the arrangement.

The Indian Ethnographic Survey, the tribal classifications of Africa, and the mapping of "martial races" for recruitment all operated similarly: a classification imposed for administrative purposes, which then organised access to resources, and around which people organised.

Mahmood Mamdani's Citizen and Subject (1996) makes the most consequential argument for Africa: colonial rule created a bifurcated state — civil law for citizens in urban areas, "customary" law administered by chiefs for subjects in rural areas — and in doing so invented and hardened "tribe" as a political category. His analysis of Rwanda (When Victims Become Killers, 2001) traces the Hutu/Tutsi distinction's transformation from a fluid social and occupational category into a racialised legal identity, recorded on identity cards by the Belgian administration and used in 1994 to determine who was killed.

That is the mechanism at its most consequential, and it is not a metaphor.

The complications, which honesty requires

Colonial knowledge was not monolithic or simply imposed. Recent scholarship — C. A. Bayly's Empire and Information (1996) is the standard reference — has documented the extent to which colonial states depended on indigenous informants, existing information networks and local knowledge, and frequently misunderstood what they were told. The archive is a record of a collaboration and a series of misreadings, not a unilateral imposition.

Colonised people used the categories. Caste associations, tribal identities and religious classifications were taken up, contested and mobilised by those they were applied to — sometimes to enormous political effect. Ambedkar used the census categories to establish the scale of untouchability and to demand representation. Agency runs in both directions.

And some colonial scholarship was genuinely valuable. Texts were preserved, languages recorded, monuments documented, and works of real learning produced. Ibn Khaldun's return to circulation was mediated by European Orientalist translation (4.6.3), which is an uncomfortable fact that the strong version of Said's thesis handles badly.

The defensible position: colonial knowledge was produced in a relation of power, constructed its objects, and served governance — and it was collaborative, contested, sometimes valuable, and taken up by those it described for their own purposes. The strong thesis is more useful as a starting hypothesis than as a conclusion.

And the reason this is the opening lesson of Part 6

Because the rest of the Part is not a tour of exotic alternatives.

If the mechanism above is right, then the categories with which the discipline describes non-Western societies are not neutral instruments that happen to have been developed in Europe. They are products of a specific relation, carrying assumptions about what is normal, what requires explanation, and what counts as a proper society.

Which means Part 6 has two jobs, and they are different.

The first is recovery: the traditions of social analysis that developed elsewhere — Indian, Chinese, African, Latin American, Islamic — read as theory rather than as data (4.6.3).

And the second is correction: examining what the discipline's own categories carry, and what would have to change for a genuinely general sociology to be possible.

The second is harder and more important, and the last lesson of this Part (6.8) asks whether it can be done at all.

"Said said all Western scholarship about the East is worthless." He argued that it constituted a discourse producing its object. He also spent his career in Western universities using Western critical theory, and said so.

"Orientalism means being prejudiced about Eastern people." It means a system of knowledge production with institutional backing, which constructed "the Orient" and thereby "the Occident."

"But some Orientalists admired the East." Latent Orientalism is the structure of assumptions shared across manifest positions. Admiration is a variant within it, not its opposite.

"The British invented caste." They did not. The all-India, enumerated, ranked, textually authorised form in which it operated by 1947 is substantially a colonial product, which is a different and better-supported claim.

"Colonial knowledge was simply imposed." Bayly and others have shown extensive dependence on indigenous informants and networks, and colonised people used the categories for their own purposes.

"Said's book is unreliable, so the argument fails." Irwin and al-Azm identified real defects. The core finding — that knowledge and rule were mutually constitutive and that the categories outlived the empires — is not seriously disputed.

Why it matters

It explains why Part 6 is necessary. If the categories were produced in a relation of power, then adding non-Western material to an unchanged framework is not a correction. It is the same operation, performed more inclusively.

It supplies the six-point mechanism, which is transferable: position erased, object essentialised, history removed, self-knowledge denied, knowledge made administratively useful, and categories outliving the arrangement. Apply it to any body of expert knowledge about a subordinated population — clinical, criminological, developmental, educational — and it will usually find something.

It documents categories that outlived their empires and are still killing people. The denotified tribes; the Rwandan identity cards; the tribal classifications of the African bifurcated state. This is not intellectual history.

And it turns the discipline's own tools on the discipline. Said used Foucault and Gramsci — European theory — to expose what European theory had done, which is either a contradiction or the only available method. Part 6 will keep meeting that question, and 6.8 is where it is faced directly.

Said's Orientalism (1978): not a claim about prejudice but about discourse — an academic tradition, a style of thought based on an ontological distinction between Orient and Occident, and "a corporate institution for dealing with the Orient... by describing it, by teaching it, settling it, ruling over it."

The discourse produced a stable set of oppositions in which the East is defined as what the West is not — and thereby produced the Occident: "the Orient has helped to define Europe as its contrasting image." Its most consequential term is timeless: a society without internal history has no development of its own, and change must come from outside.

Latent and manifest Orientalism: the underlying structure of assumptions persisted across radically different explicit positions, which is why admiring Orientalists are not counter-examples.

The colonial census is the mechanism with an archive: caste required to be singular, categories drawn from Brahminical texts, Risley's 1901 ranking with anthropometric measurement, groups petitioning for reclassification, and the fluid becoming fixed. Caste was not invented by the British; the enumerated, ranked, all-India form was a colonial product — Cohn, Dirks, Appadurai.

The criticisms that landed: empirical selectivity (German Orientalism barely discussed); homogenisation — al-Azm's "Orientalism in reverse"; theoretical inconsistency between Foucault and humanism; and no room for agency.

The general mechanism, in six points: position erased; object essentialised; history removed; self-knowledge denied; knowledge made administratively useful; categories outliving the arrangement. And every one of them applies to sociology's own treatment of non-Western societies.

Beyond Said: Cohn's investigative modalities; the Criminal Tribes Act, covering some thirteen million people, whose "denotified" communities remain disadvantaged; and Mamdani's bifurcated state and the Belgian identity cards that determined who was killed in Rwanda in 1994.

And the complications: Bayly on colonial dependence on indigenous informants; colonised people using the categories, including Ambedkar; and the fact that Ibn Khaldun's return was mediated by Orientalist translation.

Orientalism — the discourse constructing "the Orient" as an object of knowledge, and thereby constructing "the West."

Latent Orientalism — the persistent structure of assumptions underlying varying explicit positions.

Manifest Orientalism — the explicit stated content, which varies widely.

Discourse (Said's usage) — after Foucault: a regulated system of statements producing its object, in exchange with power.

Essentialisation — the attribution of a fixed nature to a population, flattening diversity and conflict.

Denial of coevalness — the representation of another society as belonging to an earlier time (Fabian's term for the timeless present tense).

Investigative modalities — Cohn's term for the forms through which colonial states produced knowledge of their populations.

Bifurcated state — Mamdani's term for colonial rule dividing civil law for citizens from customary law for subjects.

Denotified tribes — communities designated criminal by the 1871 Act, still so labelled after its 1952 repeal.

Orientalism in reverse — al-Azm's term for an essentialised West opposed to an essentialised East.

Contrapuntal reading — Said's later method: reading imperial texts together with what they exclude.

One — run the six-point mechanism. Apply it to any expert literature about a subordinated population — clinical, criminological, developmental. Check each point. Most will land somewhere.

Two — find a category that outlived its arrangement. Take an official classification in your own country and trace when and why it was created. Ask whether the reason still exists and whether the category does.

Three — check the tense. Read any general account of a "traditional" society and note whether it is written in a timeless present. Then ask what history that tense removes.

Four — compare two accounts. Find a description of a society by an outsider and one by an insider from the same period. Ask which is treated as evidence and which as testimony, and why.

Five — read the introduction. Orientalism's first thirty pages state the whole argument. Read them, then read a page of Irwin's critique, and decide what survives.