Is a Global Sociology Possible?

Part 6 has covered six traditions and two bodies of theory. This lesson asks the question they were all leading to, and it does not resolve it — because it is not resolved.

If the discipline's categories were built from a narrow range of cases and presented as general, then adding non-Western material to an unchanged framework is not a correction. It is the same operation performed more inclusively.

So what would a genuinely general sociology be, and can it be built?

A conference, and four positions in a room.

The panel is on global sociology. Four people speak.

The first says: the discipline's problem is coverage. We have too much research on rich countries and not enough on the rest. The remedy is more data, better sampling, more comparative studies, and more scholars from more places. This is a resource problem and it is being solved, slowly.

The second says: the problem is not coverage but concepts. Adding cases to a framework built elsewhere produces studies of how far other societies approximate a norm they did not set. The remedy is to generate concepts from other societies' material — chaxugeju, asabiyyah, coloniality, closure — and test them everywhere.

The third says: the problem is deeper. The criteria for what counts as knowledge were themselves set by one party. Peer review, the journal system, the English language, the citation apparatus, the conference: these are not neutral containers. The remedy is to build separate institutions with separate criteria.

And the fourth says: all three of you are speaking English, in a European city, at a conference funded by a Western foundation, and you will publish in journals indexed in Web of Science. What exactly is being proposed?

Nobody in the room has a good answer to the fourth speaker, and the discussion moves on.

This lesson is about that unanswered question.

The problem has three levels, and they require different remedies.

Level one: coverage. Who and what gets studied. This is measurable, it is genuinely bad, and it is improving.

The evidence: studies of the discipline's leading journals consistently find the great majority of published research concerns a small number of wealthy countries; the "WEIRD" critique in psychology — Henrich, Heine and Norenzayan's finding that the overwhelming majority of published behavioural research uses subjects from Western, Educated, Industrialised, Rich, Democratic societies, who are frequently outliers on the measures concerned — applies with modification to sociology.

Level two: concepts. Whose categories organise the analysis. This is the level Part 6 has principally addressed.

Level three: the criteria of knowledge itself. What counts as evidence, argument, theory, rigour and contribution. This is where the arguments become genuinely hard, because a critique of the criteria has to be conducted in some criteria.

And each level has a plausible remedy that fails to address the level below it. More data does not fix the concepts; better concepts do not fix the criteria; and abandoning the criteria removes the ground on which the first two arguments were made.

The positions, stated properly

Four positions, and each has serious people holding it.

One: universalism. Social processes are general; concepts developed anywhere can apply everywhere; and the remedy is more and better comparative research. The strongest version is not naive: it holds that concepts must be tested rather than assumed to travel, and that failure to travel is itself a finding.

Its defenders point out that the alternative — a separate sociology for each society — makes comparison impossible and thereby removes the ability to identify what is distinctive about any case.

Two: indigenisation. Each society should develop concepts from its own material and its own intellectual traditions. The strong version holds that cross-cultural generalisation is not possible; the weak version, that concepts must be generated locally before being tested comparatively.

Its problem is the strong version's: if concepts cannot travel, the claim that they cannot travel is itself a claim about all societies, and therefore travels.

Three: alternative discourses. Syed Farid Alatas's position (6.6.1), and this course's preferred one: the problem is not that sociology is too universal but that it is insufficiently universal — built from a narrow range of cases and presented as general. The remedy is more sources of concepts, tested comparatively.

Four: decolonial rupture. Mignolo's delinking (6.7.1): the criteria themselves are products of coloniality, and reform within them reproduces them. What is required is epistemic disobedience and the building of alternative institutions.

Its problem is the one the fourth speaker raised, and Mignolo has been pressed on it repeatedly.

Two proposals that have institutional form, and are worth assessing.

Raewyn Connell's Southern Theory (2007) argues that mainstream social theory is not merely incomplete but structurally misleading: classical sociology was built on the imperial experience — its data came from the colonies, its questions were shaped by empire — while systematically erasing the colonies from view. Her positive proposal: read theory produced in the global South as theory rather than as data, and treat the "metropole" as one region rather than as the centre.

The criticisms: "Southern theory" flattens enormous diversity — Africa, Latin America, South Asia and the Pacific have little in common beyond a position; geographical origin is a poor proxy for the actual variable, which is institutional position (a scholar at a Delhi institute and one at Harvard are differently placed regardless of birth); and the framing risks reproducing a North/South binary as essentialised as the one it opposes.

Michael Burawoy's Global Sociology takes a different route: not a new theory but a reorganisation of the discipline's practice — provincialising the American centre, building genuinely reciprocal international collaborations, and taking the International Sociological Association seriously as a body rather than as a conference series. He has been candid that his own attempt to do this as ISA president was harder than he expected, and his account of why is more useful than most position papers.

And Gurminder Bhambra's Connected Sociologies offers the most careful analytical proposal: not two sociologies to be brought together, but the rewriting of the discipline's foundational narratives — of modernity, of the nation state, of industrialisation — as connected histories from the start, in which the colonies are not a later addition but a constitutive part of the account.

Bhambra's version is the one most compatible with what this course has argued, because it is a claim about the content of the standard account rather than about who produces it.

The abstract argument becomes usable when applied to a specific concept, so take four from this course.

Case one: secularisation. Built from the European experience of declining religious practice, and for decades treated as a general feature of modernisation. Applied globally it failed: the United States, most of Latin America, Africa, South and Southeast Asia and the Middle East did not follow. The concept was not wrong; it was a description of one region's trajectory presented as a general law, and the field has since disaggregated it into three separable claims (3.2.6), which is what a proper generalisation looks like.

Case two: class. Marx's and Weber's schemes were built from European industrial societies with formally free labour. Applied to societies with caste, with large informal sectors, with unfree labour persisting alongside wage labour, or with strong kin-based economic units, they capture some of the variance and miss a great deal. Adding those societies as cases does not fix this; the categories have to be rebuilt, and the Indian, Latin American and African literatures have been doing exactly that.

Case three: the family. The nuclear household as the unit of analysis, embedded in survey instruments, census categories and welfare systems worldwide — and a poor fit for extended households, polygynous arrangements, matrilineal descent, transnational split households and fostering practices. The measurement error is not random; it systematically undercounts arrangements that do not resemble the template.

Case four, running the other way: closure. Ambedkar's account of endogamous closure (6.2.2) was built from Indian material. It travels: it describes Burakumin exclusion, Roma exclusion, aristocratic reproduction, professional licensing and racial endogamy in the American South. A concept generated at the periphery turned out to be general.

Which is the whole argument of "insufficiently universal" in one comparison. Secularisation was a regional finding presented as universal; closure is a regional finding that actually is universal. The difference was established by testing, not by the concept's origin — and the discipline tested the first and not the second, for institutional reasons.

What has actually changed

An honest accounting, because the argument is frequently conducted as though nothing has moved.

What has changed:

Composition. The International Sociological Association has members in over a hundred countries; the largest national associations outside the US and Europe are substantial; and Chinese, Indian and Brazilian sociology are large fields with their own journals and traditions.

Curricula. Du Bois, Ambedkar, Fanon and postcolonial theory appear on syllabi where they did not thirty years ago. This course could not have been written from the standard textbooks of 1990.

Publishing. Open access, regional journals, and — the most consequential — the availability of texts. Ibn Khaldun, Ambedkar, Fei, Alatas and Quijano are all readable now by anyone with a connection, which was not true in 1990.

And the empirical base. Comparative datasets covering most of the world's population now exist for health, education, employment, values and demography, which changes what can be claimed.

What has not changed:

The journal hierarchy. The highest-ranked general journals remain overwhelmingly US and UK based, with editorial boards to match, and career advancement everywhere is measured against them.

Language. English is the condition of visibility. A first-rate article in Portuguese, Hindi or Arabic will be read by a fraction of the audience an inferior one in English reaches, and the labour of writing in a second language falls entirely on one side.

Funding. Research agendas in poorer countries are substantially shaped by external donors, which is Alatas's captive mind operating institutionally (6.6.1).

Citation. Theory is cited from the metropole; cases are cited from elsewhere. The pattern is measurable and has been measured: work from the global South is cited disproportionately as evidence and disproportionately little as theory.

And the division of labour. The oldest pattern: the periphery supplies the data and the centre supplies the theory, which is the arrangement Part 6 has been documenting from 6.1.1 onwards.

Go deeper

The hard question

Can the criteria of knowledge themselves be criticised without using criteria?

The relativist answer: no, and therefore all criteria are equally arbitrary. This position is not held by any serious participant in the debate, because it destroys the ground for the criticism being made — if all criteria are arbitrary, the claim that Western criteria are parochial has no force.

The Habermasian answer (5.4.2): the criteria for assessing claims are implicit in the act of making them. Anyone offering an argument has committed to the possibility of its being assessed, and that commitment supplies a standard that is not the property of any tradition.

Its problem: the specification of what counts as a good argument is itself culturally variable, and the Habermasian answer has been accused of universalising a European conception of rational discourse — which is 5.4.2's standing criticism.

The pragmatist answer: judge by fruitfulness. Does a concept let you see something you could not see otherwise? Does it generate questions that survive contact with evidence? This is Weber's ideal type criterion (4.3.8), and it has the merit of being usable.

Its problem: fruitfulness is judged by someone, from somewhere, for some purpose.

And the position this course takes, offered as a working rule rather than as a solution:

The criteria that survive are the ones that hold up when applied to the party applying them. A standard of evidence, argument or rigour that its holders will not accept when turned on their own work is not a standard; it is a gate.

That is a minimal criterion and it is not nothing. It is what "state what would show you wrong" (4.3.9) amounts to, and it is applicable across traditions because it does not specify what counts as evidence — only that whatever does must count symmetrically.

And the reflexive turn this course owes

This course is an instance of what it has been describing.

It was written in English, by a system trained overwhelmingly on English-language material, drawing on sources that are disproportionately Anglophone and disproportionately available online — which correlates with wealth, language and institutional position.

Its Part 4 has ten lessons on Weber and six on Du Bois, and its Part 6 gives one lesson each to Chinese, African, Latin American and Islamic traditions. That allocation is a judgement, and it reproduces the structure this Part criticises even while criticising it.

A defensible reason exists: a reader needs to be able to read the discipline as it is, and the discipline as it is has ten lessons' worth of Weber in it. Teaching what is there is not the same as endorsing how it got there.

And that reason is also exactly what every syllabus says. It is how the arrangement reproduces itself — each individual choice defensible, the aggregate a structure.

The honest position is to state it rather than to resolve it, and to note what would be different if the constraint were removed: Ambedkar in Part 4 rather than Part 6; Fei beside Tönnies; Ibn Khaldun with the founders rather than the precursors; and chaxugeju, closure and asabiyyah in the general theory vocabulary rather than as regional contributions.

This course has said six times that a canon is an institutional product. It would be dishonest to present its own selection as anything else.

Where this leaves a reader

Four practical rules, which are what Part 6 is actually for.

One: ask where a concept was built. Not to disqualify it, but because a concept built from one range of cases will fit those cases best. This applies to concepts from everywhere, including the ones in this Part.

Two: test rather than assume travel. Whether a concept generalises is an empirical question. Secularisation did not; closure did. Neither outcome was predictable from the concept's origin.

Three: read work from other traditions as theory. The single most consequential change available to an individual reader. If Fei's chaxugeju is read as a fact about China rather than as a formal alternative to the bounded group, the division of labour reproduces itself in your own reading.

Four: apply the symmetry rule. Whatever standard you apply to a claim from elsewhere, apply to one from the centre. The claim that Confucianism explains East Asian growth and the claim that Protestantism explains European capitalism should meet the same evidentiary bar — and neither does.

"Global sociology means studying more countries." Coverage is one level of three. Concepts and criteria are the other two, and more data does not fix them.

"Southern theory means theory from the South." Connell's argument is about reading it as theory rather than as data — and the criticism is that geography is a poor proxy for the actual variable, which is institutional position.

"Indigenisation means a separate sociology for each society." The strong version says so and is self-undermining. Alatas's alternative discourses is the workable version: concepts generated anywhere, tested everywhere.

"Nothing has changed." Composition, curricula, publishing and the empirical base have changed substantially. The journal hierarchy, language, funding, citation patterns and the theory/data division of labour have not.

"The criteria of knowledge are arbitrary." No serious participant holds this, because it destroys the ground for the criticism being made.

"This course has solved the problem." It reproduces the structure it describes, in its allocation of lessons and in its sources, for reasons that are defensible and that every syllabus also gives.

Why it matters

It is the question the discipline has to answer and has not. The evidence of the previous seven lessons is that categories were built narrowly and presented generally, and that traditions producing first-rate work were not read. What follows from that is genuinely unsettled, and a course that pretended otherwise would be teaching a conclusion the field has not reached.

It gives four working rules that an individual reader can actually apply: ask where a concept was built; test rather than assume travel; read other traditions as theory; and apply standards symmetrically. None of these requires institutional change to begin.

It supplies the symmetry criterion, which is the most defensible minimal standard available: a rule that its holders will not accept applied to themselves is not a standard.

And it closes Part 6 with the reflexive move the Part has been demanding of everyone else. This course is an instance of the arrangement it describes, its selection is a judgement, and stating that is the minimum that the argument of the previous seven lessons requires. A course that made this criticism of Weber, Dumont and the Chicago School and exempted itself would have failed its own test.

Three levels, requiring different remedies: coverage (measurable, bad, improving — the WEIRD critique applies with modification); concepts (whose categories organise the analysis — Part 6's principal subject); and criteria (what counts as knowledge — where the arguments become genuinely hard, because the critique must be conducted in some criteria).

Four positions: universalism (concepts must be tested, not assumed to travel); indigenisation (self-undermining in its strong version); alternative discourses — Alatas's claim that sociology is insufficiently universal, built narrowly and presented generally, this course's preferred position; and decolonial rupture (delinking, which faces the fourth speaker's question).

Institutional proposals: Connell's Southern Theory, criticised for flattening diversity and using geography as a proxy for institutional position; Burawoy's reorganisation of practice; and Bhambra's connected sociologies — rewriting the foundational narratives as connected histories from the start, which is a claim about content rather than about producers.

"Insufficiently universal" in practice: secularisation was a regional finding presented as universal and failed; closure, built from Indian material, travels to Burakumin, Roma, aristocracies and professions. The difference was established by testing, and the discipline tested one and not the other.

What has changed: composition, curricula, availability of texts, and the comparative empirical base. What has not: the journal hierarchy, the English-language condition, donor-shaped agendas, citation patterns, and the theory/data division of labour.

The working criterion: a standard its holders will not accept applied to their own work is not a standard but a gate — which is symmetrical without specifying what counts as evidence.

And the reflexive admission: this course is written in English from disproportionately Anglophone sources, gives ten lessons to Weber and one each to four traditions, and reproduces the structure it criticises for reasons that are defensible and that every syllabus also gives.

Three levels of the problem — coverage, concepts, criteria.

WEIRD critique — the finding that published behavioural research draws overwhelmingly on Western, Educated, Industrialised, Rich, Democratic subjects who are frequently outliers.

Indigenisation — the development of concepts from a society's own material and traditions.

Alternative discourses — Alatas's programme: concepts generated anywhere, tested for general applicability.

Southern theory — Connell's argument for reading work from the global South as theory rather than as data.

Connected sociologies — Bhambra's rewriting of foundational narratives as connected histories.

Insufficiently universal — the diagnosis that the problem is a narrow base presented as general, not universalism as such.

The symmetry criterion — a standard must be acceptable to its holders when applied to their own work.

The theory/data division of labour — the pattern in which the centre supplies theory and the periphery supplies cases.

One — check a concept's origin. Take any concept you use and find out from what material it was built. Then ask which cases it fits best, and why.

Two — run the travel test. Take a finding from a Western society and one from elsewhere, and ask what evidence exists that either generalises. The asymmetry in how the question has been asked is usually the finding.

Three — read one text as theory. Take Fei, Ambedkar, Quijano or Ibn Khaldun and read it for its general claims rather than for what it tells you about a country. Note what changes.

Four — apply the symmetry rule. Find a claim about a non-Western society and one about a Western society, and hold both to the same evidentiary standard. Frequently one will fail.

Five — audit this course. Identify what Part 6's allocation of lessons leaves out, and what a differently constructed version would look like. If you can do this, you have read the Part properly — and you have also done to it what it did to Weber.