Niche construction

Power and colleagues (2015) describe the evidence for feedback between ecological and evolutionary dynamics as intuitive. They say it is already recognised wherever populations shape their own future selective pressures, with niche construction as the example, but that its full implications have not been realised. Their illustration is invasional meltdown, where species long in contact help one another invade a new environment. They describe a simple positive feedback. Species that occur together at high density coevolve to compete less with each other, and this makes high-density coexistence more likely in future¹. They present this as a formal result from population genetics.

Watson and colleagues (2016) ask whether niche construction is systematically good for the constructor. They argue it is only when the benefit falls differentially on the individual bearing the trait, which lets the trait be treated as an extended phenotype. A trait that reduces competition with another species gives its bearer no advantage if every member of the species gains equally². They conclude that evolved ecological relationships may change ecological dynamics without creating benefits for the species or the community. Their own models of social niche construction cover small pairwise collectives and larger homogeneous ones. They note that an integrated model is still needed³.

Constant, Ramstead, Veissière, Campbell and Friston (2018) reframe niche construction theory in variational terms. They list predictions about non-random, organism-dependent biases on selection pressures, and about stabilisation of niches across scales and generations⁴. Ramstead, Constant, Badcock and Friston (2019) call this the variational approach to niche construction. They argue that organisms offload part of the work of computing expected free energy onto a constructed environment. On this account, lasting changes to the niche act as unambiguous cues to the actions available⁵.

Reid (2023) reviews Physarum, which senses its own trail of extracellular slime and so builds a map of ground it has explored. Reid reports that Sims and Kiverstein (2022) read this as extended cognition and as cognitive niche construction⁶. Tung and colleagues (2024) note that work on extended phenotype considers factors outside individual bodies. They add that the single embryo is still usually taken as the self-contained unit of study⁷. Levin (2025) uses the term for the brain. He suggests that repetitive thoughts shape the neural hardware so that such thoughts persist and grow⁸.