Concept study / Information & adaptationConceptual illustration. Shapes, colors, and motion are not measured data.
A population can change across generations because heritable variants leave different numbers of descendants in a particular environment. An individual can also change behavior during its lifetime, and a cell can regulate its internal conditions without a genetic change. Hill’s phrase adaptive system spans these very different processes. To use the phrase scientifically, the project must name the level, timescale, and mechanism involved. A shared outcome such as persistence does not erase those differences. The useful research question is whether a carefully defined mathematical relationship transfers between mechanisms, not whether every kind of change can be relabeled as the same process.
History can open or close possibilities
Blount, Borland, and Lenski investigated the emergence of citrate utilization in an experimental Escherichia coli population. Their work showed that the population’s earlier evolutionary history affected access to this innovation. This supports a concrete form of historical contingency: earlier changes can alter what later changes are accessible. It does not demonstrate a foresighted population or a universal blueprint. In the Hill atlas, this example should be linked to an explicit chain from inherited state to later evolutionary possibility, with the experimental scope preserved.
Choose the right unit of analysis
We propose a comparison sheet that distinguishes the individual, lineage, and population. For each, record what is inherited, what is sensed, what changes, and what counts as success. A population-level increase in a trait can occur even when no individual learns that trait. Conversely, a learned behavior can disappear when the learner dies. A valid cross-scale analogy should preserve the relevant relationships rather than replacing each mechanism with a vague appeal to improvement. This makes disagreements about adaptation much easier to resolve.
A useful nonlaboratory contribution
Researchers can contribute reproducible analyses of public evolutionary datasets or clearly labeled simulations before proposing new experiments. Define the fitness proxy and compare competing explanations such as selection, drift, and historical constraints. Where a simulation uses an explicit objective, disclose it: the programmer’s objective is not evidence that nature contains the same one. A result should identify whether predictability improves when historical state is included, and whether that improvement persists on data not used to select the model.
The mathematical lens
A frequency change is an observation; its cause requires a model of selection, drift, mutation, and transmission.
Turn it into a question.
- Question
- Does a measured history improve predictions of later evolutionary outcomes?
- Method
- Compare history-aware and history-free models on held-out lineages or simulation runs.
- What would count against it?
- No predictive improvement weakens the particular historical-state model, not evolution itself.
Name the inheritance mechanism and the timescale before comparing adaptation.
Follow the original work.
These sources support the mechanisms and limits discussed here. They do not endorse Hill’s framework. Notes distinguish research proposals, reviews, and observations.
Continue the investigation
Edition 02 · 18 October 2026 · AI-assisted synthesis. See editorial method.