Week 6 Study Guide: Surveillance, Numbers, and Decisions

Big idea

1. Institutions need to make social order (control, discipline, security)
2. Surveillance watches people so deviations can be spotted and acted on (Lecture 13)
3. Numbers and "objectivity" make those institutions trusted by outsiders (Lecture 14)
4. Framing people as rational "decision-makers" lets institutions steer behavior with nudges (Lecture 15)
5. None of this is natural. Each is a historical product that is co-produced with politics and power
6. Watch for who is exposed, who is trusted, and who is left out

Foucault, "Panopticism" (Discipline and Punish)

Core idea: Modern power works less through violence and more by getting people to police themselves. Bentham's Panopticon is the model.

Plague town vs. leper

• Plague town (17th c.): town sealed, each street watched by a syndic, daily roll call at windows, everything written down and reported up, everyone fixed in place. This is the model of discipline (exceptional, violent, temporary).
• Leper = exclusion (cast out, binary in/out). Dream of a pure community.
• Plague = discipline (sort, track, individualize everyone). Dream of a disciplined society.
• In the 1800s the two merge. Asylums, prisons, schools, and hospitals both brand people (normal/abnormal, sane/mad) and individually track and train them.

The Panopticon

• Ring of cells around a central tower. The guard sees every cell, but inmates can't see the guard.
• Inmate is seen but doesn't see: "the object of information, never a subject in communication." Inmates are also cut off from each other (axial visibility, lateral invisibility).
• "Visibility is a trap."
• Power is visible but unverifiable: you always see the tower, but never know if you're being watched right now.
• So you act as if always watched and become "the principle of your own subjection."
• Power is depersonalized. It doesn't matter who is in the tower. Power lives in the arrangement (architecture and lines of sight).
• It's cheap and light (no chains), and preventative (acts before any offense).
• It is also a laboratory: you can observe, compare, and experiment on people. Knowledge follows power.
• It is generalizable: prisons, schools, hospitals, factories. It's a diagram of power, not just one building. The goal is to make society more productive.
• Bentham's "democratic" twist: the public can inspect the tower, so it supposedly can't become tyranny.

Key terms: discipline, panopticon, internalized surveillance, normalization, "docile" self-policing

NYT Opinion: Vasquez, "The N.Y.P.D. Is Teaching America How to Track Everyone Every Day Forever"

Core idea: The NYPD has built a huge, hidden, mostly unregulated surveillance system, and other cities and the federal government are copying it.

• NYPD has spent $3+ billion. It tracks via subway data, cameras, license plate readers, social media scraping, drones, DNA, iris scans, phone taps, and more.
• Data goes into the Domain Awareness System, which keeps it for years and predicts where you'll go and who you'll contact.
• For about 90% of its technologies, NYPD says no warrant needed.
• Gang database: 13,000+ people, 99% people of color. Example: a teen in the Marcy Houses can be added because of where he lives or who he posts with. Result: repeated stops, hours of questioning, phone seizures (24,000+ phones in 2024), undercover cops posing as teens online.
• Chilling effect: people avoid protests, change schools, "look over their shoulders." "Paranoia spreads like a contagion."
• Future-risk example: abortion is legal in NY but not elsewhere. The data could identify out-of-state visitors to Planned Parenthood, and no law stops the feds from asking for it.
• Federal level: DOGE breaking down walls between IRS, SSA, Medicare/Medicaid, and VA data. March executive order told agencies to remove data boundaries. Expanded Palantir contracts. USPS reportedly helping DHS track undocumented immigrants.
• State privacy laws all have carve-outs for law enforcement. We need to fix this loophole by enacting new laws that ban police from collecting and keeping this data in the first place.

Connection to Foucault: same logic (watching changes behavior), but now digital, everywhere, invisible, and unequal (differential exposure).

Porter, "Objectivity and Authority: How French Engineers Reduced Public Utility to Numbers"

Core idea: Numbers became powerful in public life because of politics, not just better science. Quantification is a public rhetoric for justifying decisions to outsiders who don't trust you.

• Two meanings of objectivity: impersonality (following rules) and fairness (impartiality). Numbers give the appearance of both. Science is like justice: "supposed to be blind."
• Mechanical processes seem fair because they stand outside private interests. Experts should look as "unimaginative" as possible, just applying rules.
• Math is "silent on many points" (it can't express tacit judgment), which is why it's clear and shareable to diverse readers.
• Who uses numbers? People under pressure from powerful outsiders (public, parliament, rivals). Secure elites don't need them because they trust their own judgment.
• Quantification needs bureaucracy, and is tied to democracy. Numbers are "a tribute to democracy" because personal authority isn't enough.
• Descriptive statistics (murders in Belgium) look like plain "facts." Complex calculations (is a canal worth building?) mainly defend decisions.

Case: French state engineers (Corps des Ponts et Chaussées)

• Elite engineers from École Polytechnique. Math was their special language: credible to the public but not comprehensible to it, so it fended off controversy.
• France required a "declaration of public utility" before taking land for roads, canals, and railroads. Engineers chose to define it with numbers.
• Public works were politically exposed (parliament, prefects, every deputy wanting a connection), so numbers gave protection.
• Takeaway: values were being argued inside the math. Standardizing methods meant standardizing values.
• Technocracy is not the same as quantification. French "technocrats" actually trusted instinct and general culture. They used numbers only where they couldn't control decisions (location and pricing of public works). Postwar U.S. was far more mechanical. Expertise "reflects the embeddedness of science in culture."

Lecture 13: Surveillance and Security

Core idea: Surveillance matters more than just privacy. It's a way of making social order, economic power and state-corporate partnerships, and repurposable infrastructure.

3 narratives

• 1984: control the population ("you can't escape")
• Panopticon: discipline ("monitor yourself")
• Global analytics: security. Find and preempt bad actors, separate signal from noise.

History: wartime internal enemies (4th Amendment issue) → 9/11, PATRIOT Act, warrantless NSA surveillance (identify and preempt) → Snowden 2013 (Verizon metadata on 120M subscribers, PRISM).

Normalized surveillance: ubiquitous, suspicionless, bulk, cheap, predictive, repurposable. What's new is precision.

Where it comes from: porous corporation-state boundaries (Flock, Palantir, Anduril, ICE vendors).

Surveillance culture: everyday but differential exposure (some are targeted, others aren't).

Alternatives: individual agency (not enough), contestation, regulation (GDPR Art. 5(c) data minimization).

Lecture 14: Quantification and Objectivity

Core idea: Trust in numbers is historical. Data is a rhetorical practice for persuasion.

History: Galileo (primary properties, control), Priestley vs. Lavoisier, census and political arithmetic, double-entry bookkeeping (technology of trust), Condorcet (numbers vs. secrecy and prejudice), rule of law ("judge as automaton," e.g. MLB ABS).

Porter:

• Objectivity adapts to the "suspicions of powerful outsiders."
• Quantification is a "technology of distance" that builds trust among strangers.
• It is a "weapon of the weak" where experts lack authority.

Varieties of objectivity: truth-to-nature, disciplinary (trust the expert), mechanical (rules/algorithms), intersubjective (agreement).

Why numbers persuade: transparent, condensed, commensurable, portable, rule-bound, appear fair.

Performativity: metrics change what they measure (U.S. News rankings).

Takeaway: objectivity is modern, co-produced with quantification, democracy, and law. It means public, uniform standards instead of personal trust.

Lecture 15: Making Decisions

Core idea: "Decision-making" is a recent concept (1940s-60s), a technology and product of history, not a natural fact. Agency is the ability to act, and context shapes who has it.

Before 1945: decisions = legal authority (Schmitt); choices = economic behavior (homo economicus).

Built for WWII logistics and Cold War strategy from:

1. Cybernetics (Wiener)
2. Behaviorism (Skinner)
3. Game theory (Prisoner's Dilemma)
4. Mind as machine/bureaucracy (Simon, bounded rationality)
5. Behavioral economics (Kahneman and Tversky)

Choice architecture:

• Nudges and libertarian paternalism: you're free to choose, but designers arrange the options.
• Power = second-order decision-making (shaping others' choices).
• Facebook 2010: voting messages raised turnout.
• Zuboff: goal is to change behavior at scale with "data pellets."
• Hypernudges (Yeung): decision → data → assessed against goals → architecture reconfigured → nudge again.

Co-production: tech and society make each other (avoids technological and social determinism). Models of rational decision ⇄ managerial social order.

Left out: relationships, creativity, obligation. AI inherits old assumptions about reasoning.