Week 2 Study Guide: Classification, Data, and Making People

1. Bowker & Star, Sorting Things Out - "Race Classification and Reclassification under Apartheid"

Big picture: Classification systems aren't neutral bookkeeping — they're sites of political struggle that get buried under a technical/bureaucratic surface. Apartheid South Africa is the extreme case study of what happens when a state tries to force a messy human reality (race) into rigid legal boxes.

• Torque: The twisting/damage that happens to a person's life story (biography) when a classification system doesn't match their actual lived experience, appearance, relationships, or location. The system stays rigid; the person's life gets bent/torn to fit it (or doesn't fit, and they suffer for it).
• Aristotelian = categories with hard, logical, necessary-and-sufficient boundaries (the law's fantasy: "White," "Coloured," "Bantu," "Asiatic" as clean boxes).
• Prototypical = categories based on fuzzy resemblance, "best examples," and everyday judgment (what a policeman or tram conductor sees and feels is normal).
• Apartheid's contradiction: the law demanded Aristotelian precision but could only ever be enforced using prototypical, subjective judgment (appearance, "general acceptance," hair texture, etc.). This conflation is where all the horror and absurdity comes from.

Examples:

• 1950 Population Registration Act + Group Areas Act: everyone must be classified into 4 basic races (European, Asiatic, Coloured, Bantu/Native), which then determined where you could live, work, go to school, marry, etc.
• Pass books ("reference books"): required ID document for Black South Africans, 50+ pages, tracked birth, tax, employment, etc. — total surveillance infrastructure.
• Reclassification Boards: heard appeals from people who didn't fit their assigned category. Criteria were literally "appearance and general acceptance and repute" — not blood/descent (explicitly rejecting the American "one-drop rule," since almost all White South Africans had some Black ancestry).
• "Technologies of classifying": the pencil test (pencil stuck in hair — if it falls out, "white/coloured" enough; if it stays, "too woolly," you're Black), examining earlobes, cheekbones, skin-color scales (Fig 6.3), interrogating relatives, checking if you eat porridge or sleep on a bed vs. the floor.
• Vic Wilkinson: jazz musician reclassified across race lines five times in his life as he remarried and moved.
• Sandra Laing: born to white parents, expelled from her white school at age 10 for looking coloured; later reclassified white again, but never socially accepted; eventually lived with a Black man and asked to be reclassified again so she could live with him legally. Symbol of how legal classification and social/informal racism can contradict each other.
• David Wong: Chinese man whose White neighbors gave affidavits calling him "White" under the "general acceptance" clause — used almost as an act of protest — which triggered a government crackdown/amendment.
• "Dottie" and the Griqua people: cases where language and appearance clashed with assigned race, leaving people unable to attend any appropriate school.
• Ronnie van der Walt: boxer suddenly reclassified from white to coloured right before a big match — a targeted, humiliating "lesson" to others.

2. Ian Hacking - "Making Up People" (LRB, 2006)

Big picture: Sometimes the sciences (especially the "human sciences": psychology, psychiatry, sociology, clinical medicine) don't just discover pre-existing kinds of people — they actually create new ways of being a person that didn't exist before, meaning people start to experience themselves, and interact with others, through a new category (e.g., "the multiple personality," "the high-functioning autist").

The Looping Effect: When you classify and study people, they change in response to being classified (they learn the label, act it out, seek the diagnosis, community forms around it). Then the classification has to update to match the changed people. It's a feedback loop between the category and the people in it.

Statement A vs. Statement B:

A) "There were no [X] in 1955; there were many in 1985.": a claim about existence, often pointless to argue about. B) "In 1955 this wasn't a way to be a person; in 1985 it was a way to be a person, to experience oneself, to live in society.": a claim about social roles/experience

Examples:

• Multiple Personality Disorder (MPD, now Dissociative Identity Disorder)
  • A is (he argues) true, B is definitely true.
• High-functioning autism
  • A is false (autistic people surely existed in 1950), but B is still true (this specific way of being a person didn't exist yet).

The Ten "Engines" of making up people:

1. Count! – just counting a population as a group creates visibility (e.g., autism prevalence surveys).
2. Quantify! – turning something into a number (BMI, IQ).
3. Create Norms! – defining what's "normal" automatically defines deviance (Canguilhem).
4. Correlate! – statistical correlation (Galton) links traits to categories, sometimes spuriously.
5. Medicalise! – turn a behavior into a medical condition (child abuse, suicide, obesity).
6. Biologise! – find a bodily/neurological basis (removes moral blame — e.g., "it's not weak will, it's biology").
7. Geneticise! – push further to genetic causes (criminality, homosexuality "gay gene" search).
8. Normalise! – treatments/therapies designed to bring people "back to normal" (behavioral therapy, anti-craving drugs).
9. Bureaucratise! – institutions build systems (school special-ed screening) that in turn define who counts as belonging to the category.
10. Resist! ("Reclaim our identity") – the classified group pushes back and reclaims control of their own label (gay pride; autism self-advocacy; "Health at Every Size" for obesity).

Natural Kinds:

• A "real kind" (e.g., horse) has tons of properties bundled together beyond the defining one.
• A "finite kind" (e.g., white things) shares only the one defining property and nothing else.
• Hacking applies this to argue autism is closer to a "real kind" (bundle of many correlated traits/behaviors, part of someone's whole nature) while obesity is more of a "finite kind" (people share only being overweight, nothing else necessarily connects them). This is why "autism is who my son is" makes sense as a statement but "obesity is who I am" usually doesn't.

3. Lecture 4 - "Making Personal Data"

• Punch cards at Berkeley (1960s): early symbol of student data being reduced to standardized categories on a card — controversial (student protest imagery, "Do Not Fold, Spindle, or Mutilate").
• Classification and social order: classification isn't just descriptive — it actively organizes society (who goes where, who gets what).
• Circulation and durability: once data is recorded, it travels and persists beyond its original context — this is a source of vulnerability (data can be reused against the person later).
• Reflexivity: the act of measuring/representing a person can change their behavior (a preview of performativity).
• Vulnerability: personal data creates risk precisely because of its durability and circulation — once made, it's hard to control where it goes or how it's used.

4. Lecture 5 - "Making People Out of Data"

Core question of the lecture: What does it mean to say we "make people out of data"? How do representations/classifications bring about or shape the very people/phenomena they claim to just describe?

Performativity:

"The way that actions that describe the world (language, concepts, metaphors, models, classification systems, measurement systems, predictions, automated decision-making systems) can shape and even make the very phenomena they set out to describe."

What performativity is NOT:

• It does not mean "something worked well/functioned as expected" (a machine "performing" its task).
• It does not mean expressing an inner authentic state ("I performed my true feelings").
• It does not mean "pretending"/acting/fakeness (like "performative allyship" as an insult).
• Correct meaning: language/classification is a form of action that brings about new social realities, like saying "I do" at a wedding actually creates a marriage; it doesn't just describe one.

Personality as a historical product

• Before ~1900s: people were understood through "character" (moral virtue/vice), not measurable "personality."
• "Personality" became a bundle of measurable, describable traits starting around 1917, tied to: psychology as a discipline, new measurement/testing practices, and industrial organization (need to sort/manage large workforces — think early personality tests, IQ tests, punch cards).

"Know thyself" and self-quantification:

• Ancient idea (Delphic maxim) that self-knowledge requires external tools: writing, counting, mirrors — these externalize and objectify the self so it can be examined.
• Modern versions: personality quizzes (Buzzfeed/Myers-Briggs-style), the Quantified Self movement (Fitbit, Apple Watch, self-tracking apps), and even historical precedents like Benjamin Franklin's Book of Virtues (13-virtue self-tracking chart) and Santorio Santorio (1561–1636), who literally weighed himself and his food/waste daily — an early "quantified self."
• Contemporary/extreme example: "Looksmaxxing" — using facial measurement/self-tracking apps to rate and try to optimize one's physical appearance.

Key Takeaways:

1. The "self" is made, not simply found — it's produced together with technology and society, so it's contingent/variable, not fixed.
2. Quantified self-tracking is one specific technique by which the self gets "made" in a datafied world.
3. Ways of being a self/person are products of the performativity of classifications, sociotechnical systems, norms, and practices.
4. The "cyborg" is a useful figure for thinking of selfhood as porous and relational — built from personal, social, biological, and technical elements together.

Diagnostic questions:

• What does a representation or prediction do in the world? How does it bring about new states of affairs?
• What assumptions are being naturalized along the way?
• How does a classification system/representation/narrative/technology create and impose identities on individuals and populations?
• How does it affect people's agency and well-being? Whose agency or power does it serve?
• What makes this knowledge productive (i.e., generative of new realities, not just descriptive)?
• How does performativity help knowledge serve social control?

Classification systems — whether imposed by a violent state (apartheid), by expert institutions (psychiatry), or adopted voluntarily (self-tracking apps) — are never just neutral descriptions of a pre-existing reality. They are actions that create, enforce, and reshape the very people and identities they claim to describe (performativity), and because classifications are rigid while human lives are messy, this always produces some degree of strain, harm, or "torque" for people who don't fit cleanly.