Evolution and the Levels of Selection
Source: https://global.oup.com/academic/product/evolution-and-the-levels-of-selection-9780199267972 ↗
The definitive technical treatment of the multilevel selection problem — at what level does natural selection operate? Genes, organisms, groups, species? Okasha formalises what had been decades of often confused debate, using the mathematical framework of the Price equation to show when and how selection at different levels can be rigorously defined.
Where Sober and Wilson opened the group selection debate with philosophical argument and biological evidence, Okasha provides the formal architecture.
The book is demanding but extraordinarily clear for its level of technical ambition.
Essential for anyone who wants to move beyond slogans ("selfish gene" versus "group selection") to understand what the actual structure of the problem is.
Connects to every serious discussion of cooperation, altruism, and organisational evolution.
Central argument
Okasha argues that the longstanding debate over whether natural selection operates at the level of genes, organisms, groups, or species has been conceptually muddled, and that the Price equation provides the mathematical framework needed to give precise, non-arbitrary meaning to 'levels of selection.' His central finding is that selection at multiple levels can occur simultaneously and that the question is not which level is 'real' but how to partition the total selective effect across levels rigorously. This dissolves many apparent contradictions between gene-selectionist and group-selectionist accounts by showing them to be mathematically equivalent under certain transformations — different descriptions of the same underlying dynamics rather than competing empirical claims.
Critique
A substantive tension in Okasha's project is that the formal clarity he achieves through the Price equation may outrun its biological interpretability: the equation is mathematically valid under almost any partitioning, which means it can describe selection at any level without adjudicating which level is causally efficacious. Critics like Sterelny and Kitcher have pressed this point — formal decomposability does not settle the metaphysical question of where selection 'really' acts, it just shows the question can be posed precisely. The book is thus more successful as a conceptual clarification tool than as a resolution of the underlying debate about causal structure.
Why it matters for product
The multilevel selection framework maps directly onto one of the hardest problems in organizational design for product leaders: whether to optimize at the team level, the squad level, the tribe level, or the company level — and how to diagnose when local team fitness is coming at the cost of system-level outcomes. Okasha's insight that selection pressures at different levels can conflict, and that you need explicit formal accounting to see which level is actually being optimized, is a precise analogue to the situation where product metrics look healthy at the feature or team level while the overall product coherence degrades. A CPO who internalizes this can interrogate incentive structures and OKR hierarchies with the same rigour Okasha demands of biological models — asking not 'are we improving?' but 'at what level is improvement actually being selected for, and what does that cost elsewhere?'