Source
Lloyd, “Units and Levels of Selection”, Stanford Encyclopedia of Philosophy (2005; substantive revision 2024), LIT-160, read in full in NOTE-184.
What was actually shown
Four questions (§2, §5). What is the interactor, an entity that “interacts, as a cohesive whole, directly with its environment in such a way that replication is differential” (Hull 1980: 318)? What is the reproducer, Griesemer’s generalisation of the replicator? What manifests accumulated engineering adaptations at a level? Who benefits in the long run? The roles are “distinct and separable”. One entity may fill several.
The interactor is adaptation-neutral (§2.3). It needs only a higher-level character whose relation to fitness cannot be reduced to lower-level characters, detected by contextual analysis. It does not need engineering adaptations: accumulated, design-like change beyond the ancestral range of variation, as against any trait that is a direct product of selection.
The diagnosis traced through the debates (§4). G. C. Williams’s 1966 attack was on group benefit, and his critics “never considered groups as interactors, only as interactors-cum-manifestors”; Williams later accepted the interactor criterion (1990: 504). The paleontologists’ early demand for “emergent properties” in species selection fused the same two roles. The current definition of species selection is an interactor definition, with geographic range as the standard trait. Dawkins’s question is the beneficiary question, and he accepts vehicles as the “proximal subjects of selection”.
What could have come out otherwise. A dispute could have survived the separation of roles intact. Lloyd’s reading is that the ones she examines do not. She chose the disputes and reads them from one side.
What this does not say
- That the separation is accepted. Godfrey-Smith’s Darwinian populations deny that the unit question is ambiguous (2009: 111). Lloyd’s reply, that his analysis still uses all four roles, is an assertion with selective quotation (NOTE-184, C9).
- That the conflation is “the most common error”. That is asserted, not counted (C2).
- That kin and group selection models are not equivalent. Her case rests on faulting Dugatkin & Reeve (1994). The opposing technical literature is quoted, not refuted (C4).
- That the cheater problem is a fallacy. The support is a result reported secondhand and one survey of mutualisms (C5).
- That the multilevel Price equation should be abandoned. That is credited to Griesemer in correspondence and left as “an open question” (C7).
- What a biological individual is. The entry keeps units of selection apart from individuality on purpose (§4.3, n. 99). Its view is functional-role pluralism about the relata of selection, not a theory of identity.
Connections
- Bourrat (LIT-157; THEORY-465) builds on the reproducer this entry champions and defines a “Darwinian individual”, the notion Lloyd objects to as a fusion of roles (n. 80, n. 115). Her entry predates his paper and criticises his earlier work as mistaking “methods for metaphysics” (n. 109).
- The record’s reading advises citing the entry firmly for “units of selection is four questions” and attributing its other rulings to Lloyd and the Evolutionary Change school (NOTE-184, Bearing).