AI summary by TwinMind
The Future Of Reasoning
Vsauce
30m 2s
Introduction: The Extended Mind Concept
- Mind extends beyond brain through tools, communication, collaboration
- Brain described as "wet lump of fat and protein no firmer than tofu"
- Mind encompasses "tissue and wood and stone and steel and people"
- Communication allows accessing others' memories, perceptions, knowledge
- Humans as "techno Sapien powered by imaginations and passions"
Reason's Promise vs Reality
- Reason traditionally viewed as guide to "truer knowledge and better decisions"
- Increased life expectancy, reduced suffering, better cooperation
- Brain evolution takes millions of years, but fruits of reason grow rapidly
- Next 4 decades: expected to build equivalent of New York City every month
- More concrete installed outside US in last 2 decades than US installed in entire 20th century
- Quality of life rising globally: electricity, manufactured goods, food, comfort, transportation becoming more accessible
Climate Change as Hyper Object
- Nearly every life process releases molecules with lopsided electrical charges
- These absorb/re-emit thermal radiation, slowing escape to space
- Warmer air parcels create stronger, more extreme weather events
- Stakes: more lives lost, property damage, droughts, hunger, famine, refugees
- 51 billion tons of greenhouse gases released annually
- Need to reach zero emissions
- Timothy Morton's concept: hyper objects are massively distributed in time/space, viscous, sticky
- Potential "great filter" that civilizations struggle to overcome
Behavioral Inertia and Status Quo Bias
- Tendency to keep doing what already works
- Evolutionary advantage: organisms that survived long enough reproduced
- Helps prevent unintended consequences, preserve working ideas
- But slows innovation and adaptation
- Climate emissions invisible, consequences not immediate/local
- Impact future people, distant people, different/poorer populations
- Challenge: extending empathy beyond immediate tribe
The Nature of Reasoning and Inference
- Inference: extracting new information from existing information
- All living things make inferences (e.g., depth perception from 2D visual input)
- Reasoning: consciously considering why conclusions were reached
- Eratosthenes calculated Earth's circumference within percentage points using shadows and camel travel times
- Logic and mathematics emerged as abstract, general concepts
Puzzles vs Real World Problems
- Logic puzzle example: Paul (married) looking at Mary looking at Peter (unmarried)
- Answer: Yes, married person always looking at unmarried person regardless of Mary's status
- People typically get wrong initially but accept correction immediately
- Real world differs: subjective premises, competing assumptions, unclear definitions
Flaws in Human Reasoning
- Scientific study of reasoning shows humans are "bad at reasoning, lazy and biased"
- Flaws appear "almost programmed to be bad like intentional"
- Minefield experiment: magician switched rejected faces for chosen ones
- Majority didn't notice, easily provided reasons for choices they never made
- Even with political beliefs, nearly half fail to notice when answers reversed
- Built to give reasons for whatever we think, not actual reasoning process
Intuitions vs Reasoning
- Brains evolved to react brilliantly with little conscious input
- Example: recognizing someone upset without conscious analysis of facial expressions
- Intuition: knowing without knowing how you know
- Process: use conclusions to come to reasons, not vice versa
- Can reason deliberately when no strong intuition exists
- Love of puzzles suggests reasoning has survival value
Consumer Decision Biases
- People prefer products with more functions even when:
- Don't want those functions
- Never plan to use them
- Think they're pointless and overly complicated
- Reason: easier to justify decisions to others
- Avoid embarrassment when asked about missing functions
Social Theory of Reasoning (Mercier & Sperber)
- Reasoning didn't evolve for better individual decisions
- Evolved to help humans be social in cognitive niche
- Human advantage: cognition, reasoning, cooperation (not physical attributes)
- Enables planning hunts, building traps, coordinated strategies
- Hard to convince others of intuitions alone; reasons make persuasion easier
- Vital for coordinating group action
- Allows self-justification, expressing identity and values
Why People Give Weak Reasons Initially
- Waste of cognitive resources to construct "Grand Slam reasons" unnecessarily
- Can offload reasoning work to others through dialogue
- Example: restaurant choice conversation shows escalating reason quality as friend presses back
- Reasoning starts weak, improves if others push, tailored to intended audience
Confirmation Bias Explained
- Tendency to seek, prefer, interpret information confirming existing beliefs
- Frustrates ability to accept new inconvenient data
- Seems counterproductive for truth-seeking
- Social theory explanation: reasoning is group activity
- Division of cognitive labor: I argue for A, you argue for B
- Halves cognitive work compared to both researching both options
Wisdom of Crowds
- Collection of people can process information better than any individual
- Why big decisions require groups, not single person regardless of education/power
- Jelly bean jar example: average of all guesses closer to truth than any individual
- Mistakes balance out: some guess too high, others too low
- Disagreement evens out into spectacular accuracy
Modern Problems with Reasoning
- Reasoning evolved for social contexts with multiple perspectives deliberating
- Such contexts becoming less common
- Easier to be lone reasoner justifying particular viewpoint
- Internet gives more voices but enables disengagement from accountability
- Easy to find echo chambers where everyone agrees
- Technology creates more issues requiring opinions faster than ever
- Growing complexity/specialization makes well-informed reasoning difficult
- Look for people to defend our intuitions; reasons just need to be "good enough"
Democratization of Communication Problems
- Past: unconvincing reasons limited to sandwich boards
- Now: unpopular, nonsensical ideas can look professional and trustworthy
- People can challenge weak reasons, show contradictions
- But preparation for "debate that never comes"
- Individual role in societal decisions very small
- Without consequences, why not just enjoy reasoning as puzzle-solving?
Solutions: Return to Social Reasoning
- Real work: convincing others, using reason as evolved
- Future of reason may be past of reason
- National deliberation days: citizens join small groups, discuss opinions, compare reasons
- Tests show small targeted discussions lead to greater understanding
- Participants better understand beliefs, potential decisions, possible actions
Lottocracy Proposal
- Government where decisions made by people chosen at random (not elected leaders)
- Similar to jury system for individual fate
- Citizens representing thought differences (not just geography)
- Paid for time, learn from experts, deliberate until conclusion/recommendation
- Motivated by conscience only (not re-election, money, attention, power)
- Special interests couldn't target specific likely electees
- Would have to protect everyone since anyone could serve
- More neighbors would have direct experience over time
Historical Precedent
- Democracy originally used lotteries for many political positions in ancient Athens
- Aristotle: "appointment of magistrates by lot is thought to be democratic"
- Election of magistrates is "oligarchic" (government by small number)
Future Vision and Existential Security
- Building better arenas for deliberation
- Apply reason properly to hyper objects like emissions
- Lesson for people 100-1000 years in future
- Widening participation provides existential security
- Emissions won't be last hyper object civilization faces
- Good performance might lead future civilizations to recreate us
- "Could use the cooperative abilities they had"
- History as great teacher
- We will all someday be history and can choose what that means
