Liveblogging the conference: Roberta Millstein

Roberta is a great philosopher from UC Davis and she's talking about the notion of populations.

Known she needed a definition of population for a long time - this is a first stab.

"Population" has many definitions by biologists. Most try to limit it by space or time or interbreeding. But very little analysis. We invoke it often - it needs a proper definition. Uses include conspecificity, arbitrary delimitation, geography, area and time, interbreeding, etc.

Second motivation based on selection and drift, which are processes that occur in populations. These processes become arbitrary if the term population is arbitrary.

Third motivation: if population and drift are processes then populations must be individuals [not expanded here]. Wants a definition that works.

Core question: can we characterise "populations" as individuals in a non-arbitrary way? "Arbitrary" can mean gerrymandering, or difficult to count and draw boundaries. Want to avoid the former, and not necessarily the latter (the world is fuzzy).

Context in evolutionary terms only. Definitions in sociology, etc. beside the point. Only in this sense - evolution is broad term, not merely population genetics in formal terms, but in real terms. Populations of organisms only; other types of biological entities ignored for now, and can distinguish organisms.

Starting assumptions: populations are spatiotemporally located. Are they cohesive? Are they contiguous over time? Should be different from "species" - may have more than one population; should not be true by definition that population = species. Organism should be able to migrate between populations.

Concept should be distinct from "metapopulation". Problem with that and subpopulations. Should consist of conspecifics [what happens if we lose the species category?]. Minimum of two organisms per population. No singletons... Definition should be independent of definitions of natural selection and drift, on pain of circularity.

Three types of definitions:

Boundary definitions: bounded in space and time

Causal interaction definitions: interactions between organisms

Historical definitions: genealogical ties

Combinations are possible. Only the second that provides for cohesion necessary for individuality. Disjoint groups may share a fate, so maybe shared spatial distribution is not necessary. But spatial distribution may be an indirect indicator. Definitions based on reproductive isolation are too strong: these are species definitions.

Five possible candidate criteria, four reproductive, one causal:

Reproductive criteria: gene flow out of the "group". What about causal interactions. What is minimum flow? Might be an indirect indicator also.

Potentially interbreeding organisms: even stronger than Mayr's BSC.

Actual interbreeding: what about unsuccessful mating? Non mating individuals? At what time or interval?

Actual mating organisms: At a point in time? Over a time span? Any old time span or a species typical generation/mating time?

Struggle for existence: Ghiselin on hypermodern species? Conspecifics competing for limited resources; common climatic factors. Shared fate of group.

Tentative suggestion: Populations in an evolutionary context consist of at least two conspecific organisms who are mating over a species appropriate time. [Something about locale I missed]

Genealogical continuity needs to be cashed out without being a species definition. Will get historical continuity through time, and so satisfies individuality criterion.