The Branch Method
The branch method is the analytical strategy that muddling through actually practices. It contrasts with the root method, which begins at fundamental values and derives optimal policy deductively. The branch method begins at the branch: the current situation, with its existing institutions, commitments, and constraints. It compares a small number of alternatives — typically two to five — that represent incremental modifications of the status quo. It evaluates these alternatives against the specific dimensions on which they differ, rather than against a comprehensive accounting of all possible values. And it selects the alternative whose marginal consequences are most acceptable to the relevant parties.
In the AI Story
The branch method is not a degraded form of comprehensive analysis. It is a different analytical strategy adapted to different conditions. Comprehensive analysis works when the system is well understood, the variables are known, the relationships are modeled, and the consequences are predictable. The branch method works when these conditions do not obtain — when the system is too complex, the variables too numerous, the relationships too poorly characterized, and the consequences too uncertain for comprehensive modeling to succeed.
The AI transition is a paradigm case of conditions in which the branch method is indicated. The technology is evolving faster than any model can track. The consequences of specific interventions are systematically unpredictable. The values at stake — productivity versus depth, capability versus agency, efficiency versus meaning — are in irreconcilable tension. No comprehensive framework can resolve these conditions. What can proceed is incremental comparison: specific interventions in specific contexts, evaluated against their practical consequences, revised based on observed outcomes.
The Orange Pill itself is a case study in the branch method operating under the rhetorical cover of comprehensive analysis. Edo Segal's framework of the Swimmer, Believer, and Beaver is not a comprehensive taxonomy of all possible responses to the AI transition. It is a comparison of three concrete alternatives, each defined in relation to the others, each evaluated against its practical consequences. This is branch method analysis, and its power derives from its constraints.
The branch method has a distinctive advantage: it accommodates disagreement about values without requiring resolution. Two analysts who disagree about whether depth or productivity is the more important value can still agree on the practical consequences of specific interventions. The disagreement is relocated from the theoretical domain (where it cannot be resolved) to the practical domain (where democratic deliberation can operate).
Origin
Lindblom introduced the branch/root distinction in his 1959 article as a deliberate alternative to the vocabulary of rationality. The choice of metaphor was pointed: roots are deep, singular, foundational; branches are many, distributed, and proliferating. The metaphor itself encodes the epistemological claim.
Key Ideas
Start from the current situation. Analysis begins not with abstract values but with existing commitments, because existing commitments constrain what is politically and institutionally possible.
Limited alternatives. Consider two to five options, not all possible options. The finitude is a feature: it permits serious comparison rather than shallow enumeration.
Marginal analysis. Evaluate alternatives on the dimensions where they differ. Ignore the vast dimensions on which they are identical. This is how cognitive resources are conserved for the comparisons that matter.
Incremental commitment. Choose the alternative whose consequences are most acceptable, and revise if the consequences are worse than expected. No step is so large that its failure cannot be absorbed.
Appears in the Orange Pill Cycle
Further reading
- Charles Lindblom, The Science of 'Muddling Through' (1959)
- Charles Lindblom and David Braybrooke, A Strategy of Decision (1963)