Critical Thinking with AI • Hist 300
After Steven Shapin, A Social History of Truth, Chapter 1. Trust isn’t a problem science solved — it’s the medium science is made of.
Act I
England, 1640s–1660s. The old anchors — religious, political, intellectual — have come loose. Who speaks for truth?
Act I • Slide 01 — Scene Setting: Europe in Crisis
Mid-1600s England: Civil War, regicide, Restoration. Religious fragmentation — no single church authority. Intellectual parallel: the rejection of Aristotle and scholasticism as reliable guides to knowledge.
Into this vacuum enters the Royal Society — a new institution dedicated to producing knowledge by observation rather than by appeal to tradition. But the question it inherits remains: whose word do we trust?
Act I • Slide 02 — What Should Replace Authority?
Thinkers like Francis Bacon distrust tradition and emphasize direct observation as the new foundation of knowledge. The cultural slogan: see for yourself.
If everyone must “see for themselves,” how can knowledge circulate? The ideal of independence immediately creates a new problem: most of what we know, we know through other people’s testimony.
Act I • Slide 03 — Shapin’s Opening Twist
“A social history of truth is not supposed to be possible.”
Truth is imagined as universal — independent of society, culture, persons. But Shapin asks us to look at how truth is actually made: in specific places, through specific relationships, by specific people deciding whom to believe.
Act II
Even the most committed empiricist cannot be everywhere at once. Science depends on strangers.
Act II • Slide 04 — The Practical Reality of Scientific Life
To gather data, natural philosophers rely on sailors, merchants, distant observers, and correspondents — people they have never met, in places they will never visit.
Tracking a comet requires observers across Europe to report its position night by night. No single person can do this alone. The resulting “fact” is assembled from dozens of separate acts of testimony.
Act II • Slide 05 — The Quiet Admission
“In securing our knowledge we rely upon others…”
“No practice has accomplished the rejection of testimony…”
Even the most “modern” science is built on other people’s words. The Baconian dream of pure first-hand observation is, in practice, impossible. Testimony is not a fallback — it is the normal condition of knowledge.
Act II • Slide 06 — The Information Explosion
The 17th century brings global exploration, an expanding print culture, and dense correspondence networks. The Royal Society gathers reports from across the world — natural phenomena, exotic specimens, indigenous knowledge — all arriving as testimony.
Expansion of knowledge = expansion of dependence on unverifiable sources. The wider the reach of science, the deeper the reliance on trust.
Act III
Trust isn’t just a scientific problem. Thinkers across Europe worry it is the foundation of civilization itself.
Act III • Slide 07 — Moral Fear of Social Collapse
Michel de Montaigne:
“We are men, and hold together, only by our word.
If it deceives us, it breaks up all our relations…”
For Montaigne, the liar doesn’t just violate a rule — they unravel the fabric of society. Language is the glue of human community. Betraying it is betraying everything.
Act III • Slide 08 — English Voices Echo the Fear
Henry Mason: lying “disturbeth humane society.”
George Mackenzie: it “striks at the root of all humane society.”
Proverb: loss of credibility = social death.
This is not a minor moral worry — it is existential. In a world already fractured by civil war and religious schism, the collapse of verbal trust feels like a second dissolution.
Act III • Slide 09 — Classical Inheritance
Cicero:
“To stand to one’s word… [is] the foundation of justice.”
The anxiety about trust has deep classical roots. What the 17th century adds is a moment of acute crisis — when the old structures that managed trust (church, monarchy, university) have been weakened or destroyed. The ancient question returns with new urgency.
Act IV
Everyone agrees society needs trust. Shapin extends the argument: knowledge itself has the same dependency.
Act IV • Slide 10 — From Social Order to Knowledge
“The relations in which we have and hold our knowledge have a moral character…”
Society needs trust — everyone agrees. Shapin’s extension: knowledge has the same dependency. To know something is to have received it from someone — and that someone’s credibility is part of the fact.
Act IV • Slide 11 — A Concrete Example: Comets
Observers across Europe report different sightings of a comet. To decide what is “true,” scientists ask: Who is skilled? Who is honest? Who is reliable?
What we know about comets depends on who we trust. The natural fact (the comet’s path) cannot be separated from the social judgment (this person’s word is credible).
Act IV • Slide 12 — Another Example: Experiments
Experiments are local: only a small number of witnesses are present. Knowledge spreads through reports and publications. Figure: Robert Boyle — whose air-pump experiments were known across Europe by readers who had never seen the machine.
Science travels as testimony, not as direct experience. The gap between the original experiment and the distant reader is bridged entirely by trust in persons and texts.
Act V
Shapin names what has been building across the chapter. Facts are not pure — they are hybrids.
Act V • Slide 13 — Shapin’s Formulation
“What we call ‘social knowledge’ and ‘natural knowledge’ are hybrid entities.”
Every scientific fact contains two things: (1) a claim about the world (the comet moved this way) and (2) a judgment about a person (I believe this because X reported it and X is credible). The two cannot be fully separated.
Act V • Slide 14 — What This Means
There is no purely “objective” knowledge that floats free of social relations. Knowing nature requires knowing who to trust. This is not a failure of science — it is the structure of all knowledge.
To study how knowledge is made, we must study how trust is organized — who counts as credible, under what conditions, and by whose judgment.
Act VI
If science runs on trust, why doesn’t it look that way? Because trust, once stabilized, becomes invisible.
Act VI • Slide 15 — The Strange Illusion of Objectivity
Scientists claim independence from authority — they defer to evidence, not persons. But in practice they rely on trusted witnesses at every stage of knowledge production.
Trust doesn’t vanish — it becomes taken for granted. When we stop questioning a source, we don’t stop trusting it; we trust it so deeply that the trust disappears from view.
Act VI • Slide 16 — Why Hide It?
To present knowledge as universal and impersonal, the social labor of building trust must be rendered invisible. The more trust is taken for granted, the more “objective” knowledge appears — not because the social work is gone, but because it is hidden.
Objectivity is not the absence of social foundations. It is what successful social work looks like from the outside.
Act VII — The Cliffhanger
Not all people are equally credible. Society must decide: whose word counts?
Act VII • Slide 17 — The Central Question
“Whom to trust?”
This is not just a philosophical puzzle — it has a social answer. In every historical period, certain kinds of people are presumed credible, others are not. The question is: what criteria govern that judgment?
Act VII • Slide 18 — Where the Story Is Going
The gentleman: socially independent, honorable, presumed truthful. He has nothing to gain from lying because he is not dependent on patronage or wages. His word is his bond — and his social position is the guarantee.
This becomes the foundation of early modern scientific credibility. The Scientific Revolution does not solve the problem of trust by eliminating it — it solves it by anchoring credibility in social status.
Summary
We receive nearly everything we know through the testimony of others.
The ideal of pure first-hand observation is, in practice, impossible.
Every scientific claim embeds a judgment about the credibility of its source.
Trust, once stabilized, becomes invisible — which is what “objectivity” looks like.
The question was not: How do we find truth? It was: How do we organize trust?
Bridge to Now
The problem of truth and credibility has never been solved. A new machine is not a new problem.
Bridge 01 • From “Is it true?” to “What networks?”
Knowledge always rests on trust in people. Testimony is unavoidable — especially in science.
AI outputs rest on trust in multilayered systems. Treat AI like a witness, not an oracle.
Instead of “Is this correct?” — “What view am I getting?”
Bridge 02 • Knowledge Is Still Collective
Knowledge is a collective good, built through networks of correspondence and testimony. No single person makes it.
AI compresses many voices into one output. Use it as a starting point — then compare with other sources and perspectives to re-insert the human plurality it flattens.
Triangulate! AI output + scholarly sources + primary evidence. Stir up and remix the “collective” that AI has collapsed into one voice.
Bridge 03 • All Knowledge Is Hybrid
Every "fact" contains a claim about the world and a judgment about a person. The two cannot be separated.
AI hides the human choices inside it — training data decisions, model design, cultural biases — behind the appearance of neutrality. It exaggerates the illustion of truth.
Ask AI to show uncertainty and provide alternative interpretations. Ask yourself: what invisible social networks and power dynamics shaped this answer? What perspectives might be missing?
Bridge 04 • From “Can AI Think?” — to “How Does AI Reorganize Trust?”
Science succeeds by stabilizing credibility — deciding who to trust, under what conditions, for what kinds of claims.
High trust: brainstorming, summarizing, drafting.
Medium trust: explanations, synthesis — verify!
Low trust: facts, citations, specialized claims — always check.
Match your level of verification to the stakes of the task and the consequences of being wrong.
One-Line Takeaway
Shapin’s lesson: Questions about knowledge was never “what is true?” It was “how do we organize trust?” (socially, institutionally, etc)