democratic reason: the mechanisms of collective intelligence in politics hélène landemore harvard...
TRANSCRIPT
DEMOCRATIC REASON:the Mechanisms of Collective
Intelligence in Politics
Hélène LandemoreHarvard University/Collège de France
Introduction
• How much does knowledge play a role in our justifications for democracy?
• Including all people in the decision-process means including a lot of “dumb” people
• Rule of experts seems more conducive to intelligent decisions
• Comparison rule of one, few, many
Introduction (cont.)
Three parts:
I Main concepts
II Mechanisms of democratic reason:
1. First mechanism: deliberation2. Second mechanism: majority rule
III Conclusion: the epistemic edge of democracy
I. Main concepts
Democratic reason: collective distributed intelligence of the people
Mechanisms: political cognitive artifacts
Cognitive diversity: plurality of cognitive “tools”
Epistemic competence
• Not virtue (or civic duty or impartiality)
• Not information (raw data)
• Individual vs. collective competence
CoEC= f(iEC, cognitive diversity of the group)
II. Mechanisms of democratic reason
1. Deliberation
Epistemic properties comes from:
1) Enlarging the pool of ideas and information2) Weeding out the good arguments from the bad3) Leading to consensus on better
solution“the forceless force of the better argument”
(Habermas)
Condition of optimal deliberation
Cognitive diversity matters MORE than individual epistemic competence
“Diversity Trumps Ability Theorem” (Page 2007)
Better to have a random group of relatively smart people than two or three Einsteins
E.g. 2: Guiding each other to theglobal optimum
Calvados: (Marseille (7), Caen (10))Corrèze: (Paris (8), Grenoble (9), Caen (10))Pas de Calais: (Grenoble (9))
Problem: feasibility of deliberation with large numbers.
Solution: representation
by election: recurrence and accountability
(by lot: recurrence and random selection)
Hypothesis: democratic representation is meant to preserve cognitive diversity on a smaller scale, rather than select the “best and brightest”
2. Second mechanism of democratic reason: majority rule
Supplements deliberation
Has its own epistemic properties
3 theoretical arguments:
#1 Condorcet Jury Theorem #2 ‘Miracle of Aggregation’
#3 ‘The Crowd Beats the Average Law’
#1 Condorcet Jury Theorem (CJT)
Among large electorates voting on some yes or no question, majoritarian outcomes are virtually certain of tracking the “truth,” as long as three
conditions are verified:
1) ‘Enlightenment’ Assumption
2) Independence 3) Sincere voting
#2 The ‘Miracle of Aggregation’
E.g.: Galton’s weight-contest experiment; information-markets’ predictive accuracy
1) Elitist version
2) Democratic version
3) Distributed version
Key: errors cancel each other out
The ‘Miracle of Aggregation’
Advantage compared to CJT: the average voter need not be epistemically competent at all
Problems:
1) The rationally irrational voter and systematic cognitive biases (Caplan 2007)
2) Empirical implausibility of an infinity of independent signals
#3 ‘The Crowd Beats the Average Law’ (Page 2007)
Given any collection of diverse predictive models,
Collective Prediction Error < Average Individual Error
Negative correlations, not independence
iEC matters AS MUCH as cognitive diversity
=>Democratic majority rule > rule of the random one, but not rule of the smart few
III Conclusion
1) Inclusive deliberation (direct or indirect) epistemically dominates deliberation among the smart few
2) Majority rule among the many epistemically matches majority rule among the smart few
1) + 2) = democracy epistemically dominates oligarchy
And economizes on virtue too!