Darwinian Populations and Natural Selection
Source: https://global.oup.com/academic/product/darwinian-populations-and-natural-selection-9780199552047 ↗
What conditions must a population meet for natural selection to actually operate on it? Godfrey-Smith answers with a framework far more precise than anything in popular evolutionary writing — he identifies the parameters (fidelity of inheritance, variation, competitive interaction) that make a population more or less "Darwinian." The result is deeper than any Dawkins popularisation because it replaces metaphor with structure.
The book treats evolution not as a story but as a set of abstract conditions that can be present in varying degrees, which opens the door to thinking about cultural evolution, organisational selection, and other non-biological domains.
A rigorous, original contribution that connects naturally to the multilevel selection debate in Sober, Wilson, and Okasha.
Central argument
Godfrey-Smith argues that 'Darwinian population' is not a binary category but a graded one: populations can be more or less subject to natural selection depending on how strongly they exhibit three core parameters — variation among individuals, heritability of that variation, and fitness differences that translate into differential reproduction. By formalizing these conditions into a framework rather than a narrative, he shows that natural selection is an abstract logical structure that can be instantiated, to varying degrees, in non-biological substrates including cultural and social systems. The book's central contribution is replacing the metaphor of evolution with a precise set of conditions whose strength can be measured and compared across very different kinds of populations.
Critique
The very abstraction that makes Godfrey-Smith's framework powerful also limits its empirical traction: by treating the Darwinian parameters as continuous dimensions, the framework risks becoming unfalsifiable in edge cases, since almost any system can be described as 'weakly Darwinian' without a principled threshold for when selection is doing real explanatory work. Readers extending the model to cultural or organizational evolution — which the book explicitly invites — will find limited guidance on how to operationalize heritability or fitness outside biological contexts where these concepts have clear referents. The philosophical precision is genuine, but it purchases generality at the cost of predictive specificity.
Why it matters for product
For a CPO, the framework reframes a practical question — why do some team structures, discovery processes, or product bets persist and scale while others die quietly — as a question about whether the organizational environment actually satisfies the Darwinian conditions: are there enough distinct variants being tried, is there a reliable mechanism transmitting what works to the next iteration, and is competitive pressure strong enough to select against weak variants? This is a direct diagnostic for evaluating portfolio strategy or experimentation culture: a team running many A/B tests but with low 'inheritance fidelity' — where learnings are not systematically encoded into future decisions — is operating a population with high variation but broken heritability, and selection cannot accumulate. The model also sharpens thinking about multilevel selection, helping a product leader see when selection is operating on individual features versus teams versus the whole product line, and whether those levels are aligned or in tension.