Where This Article Fits

*This article complements How Does Theoretical Biology Make Progress?, which examines scientific progress across explanatory systems and considers comparative methodology as a means of theoretical development. The present article addresses a narrower question: how do particular biological explanations become better through discovery, refinement, and critical reassessment? Historical examples from molecular biology, physiological theory, and bacterial chemotaxis illustrate different forms of improvement. The article also complements Biological Explanation and What Counts as Explanatory Gain, which examine explanatory structure and the assessment of explanatory differences. Its purpose is not to introduce another APS methodology, but to clarify how consequential improvements can be distinguished from elaboration, integration, and redescription.

How Do Biological Explanations Improve?

A biological explanation can change without becoming better. Scientists may introduce new terminology, organise existing knowledge differently, or connect previously separate descriptions. Such changes can improve presentation and understanding, but they do not necessarily establish a more adequate explanation of the biological phenomenon concerned.

Other changes are more consequential. Research may identify a previously unknown mechanism, reveal why an established mechanism behaves as it does, establish the limits of an earlier account, or distinguish biological outcomes that the earlier explanation could not adequately discriminate.

These developments raise a practical question: what has changed in the explanation, and what can scientists now explain, infer, or test that was not previously adequately established?

Historical cases help answer this question. The molecular interpretation of sickle-cell disease, the development of chemiosmotic theory, and successive explanations of bacterial chemotaxis illustrate different ways in which biological understanding can advance.

The emphasis here is on particular explanatory developments rather than progress across theoretical biology as a whole. The broader problem of theoretical progress and the methodology of comparing explanatory systems are addressed elsewhere in the APS corpus.

1. Scientific Explanations Are Not Finished Products

Biology seeks to explain how living systems work, how they maintain themselves, how they develop, and how they change through evolutionary time. These explanations are continually examined in the light of new evidence, improved methods, and changing theoretical understanding.

Sometimes progress involves a major change in scientific thinking. An established explanation may be replaced because it cannot adequately account for observations. At other times, an explanation remains substantially correct but becomes more informative. Its underlying mechanism is better understood, its conditions of applicability are clarified, or its implications are expressed more precisely.

These different forms of progress raise a fundamental question: what makes a biological explanation better than the explanation it replaces or extends?

The answer cannot simply be that the newer explanation contains more detail, employs more sophisticated mathematics, or connects more biological concepts. Such additions may be valuable, but they do not necessarily increase what the explanation establishes.

A genuine improvement must make an identifiable difference to what we can understand, justify, distinguish, predict, or investigate.

2. What Counts as Explanatory Improvement?

Biological explanations can improve in several related ways.

Mechanistic improvement

Mechanistic improvement occurs when research establishes a better account of how a biological phenomenon is produced.

Discovering a previously unknown causal interaction, for example, may explain an observation that an earlier model could only describe.

The improvement lies in the additional causal understanding, not simply in the greater number of components included in the explanation.

Explanatory generalisation

Explanatory generalisation occurs when an explanation is shown to apply across a wider range of conditions or cases.

An account originally developed for one particular biological system may be shown to express a more general principle.

However, a claim of greater generality must be supported. Merely presenting an explanation in more abstract terms does not demonstrate that it applies more widely.

Clarification of assumptions

Clarifying the assumptions of an explanation can establish the conditions under which it succeeds or fails.

This becomes a substantive improvement when it changes what can legitimately be inferred from the explanation.

An assumption that was previously implicit may prove essential to the mechanism’s success. Identifying it may reveal where the explanation is reliable, where it becomes uncertain, and which alternative mechanisms remain possible.

Discriminative improvement

An explanation can also improve by providing stronger grounds for distinguishing competing possibilities.

It may identify observations or experiments capable of ruling out alternative mechanisms, or demonstrate why some proposed explanations cannot produce the phenomenon under specified conditions.

These forms of improvement can overlap. A mathematical analysis might clarify assumptions while identifying new experimental tests. A mechanistic discovery might extend the scope of an explanation.

But none of these improvements can be established merely by attaching a label to a revised account. The relevant explanatory achievement must be supported by evidence or valid analysis.

3. Improvement Is Not the Same as Redescription

One of the most important distinctions in assessing scientific progress is the difference between explaining something better and describing the same explanation differently.

Suppose an established biological model explains how a regulatory process restores a system to a stable condition following disturbance. A later account describes the same process using the terminology of feedback control.

If the later account merely renames an already understood relationship, the change may be useful for communication or teaching, but it has not necessarily improved the scientific explanation.

On the other hand, if the new analysis identifies conditions required for successful regulation, reveals previously unrecognised limitations, or establishes constraints on alternative mechanisms, it may provide additional explanatory understanding.

The distinction matters especially when biological theories are integrated into broader conceptual frameworks.

Connecting an explanation of cellular regulation with ideas about biological organisation, agency, or evolutionary persistence may help readers appreciate relationships across biology. But greater conceptual integration is not automatically an explanatory advance over the established cellular account.

Integration may improve coherence or communication without establishing a new causal dependency, a more accurate prediction, or a stronger discrimination among alternatives.

Both achievements can be valuable. They should not be confused.

4. A Historical Example: Bacterial Chemotaxis

A particularly instructive example comes from research into bacterial chemotaxis.

Bacteria such as Escherichia coli respond to chemical changes in their surroundings. When the concentration of an attractant changes, the bacterium’s signalling system responds. Under suitable conditions, the signalling system subsequently adapts, allowing the bacterium to remain responsive to further changes.

The question is how this adaptation occurs and why it can remain robust despite variations in some of the biochemical parameters involved.

The 1997 explanation

In 1997, Naama Barkai and Stanislas Leibler published a mathematical and biochemical model explaining robust adaptation through feedback involving receptor modification.

Their work showed how the relevant regulatory organisation could produce adaptation without requiring precise adjustment of every parameter.

This was already a substantive explanation. The later research did not begin with an unexplained phenomenon or an absence of biochemical understanding.

The 2000 analysis

In 2000, Tau-Mu Yi and colleagues examined robust perfect adaptation through the mathematics of integral feedback control.

Their analysis made the control-theoretic structure explicit and investigated the assumptions under which the mechanism produces robust perfect adaptation.

It also developed conditional arguments about the requirements that alternative regulatory architectures would have to satisfy.

Where did the explanation improve?

The important question is what this later analysis added.

Simply recognising an existing feedback mechanism as an instance of integral feedback would not, by itself, establish explanatory improvement. It could amount to mathematical redescription.

The more consequential contribution lay in clarifying the conditions for robustness, examining departures from those conditions, and identifying broader constraints on possible mechanisms.

The later work therefore provides a plausible historical example of bounded explanatory improvement through greater formal generality and diagnostic specificity within an already successful explanatory account. Establishing the precise extent of that gain would require a more detailed comparison of what the earlier model and the subsequent analysis each independently warranted.

It did not demonstrate the discovery of a new molecular mechanism, nor did it establish that the original model was fundamentally mistaken.

This is precisely why the example is useful. Scientific explanations can improve substantially without requiring wholesale replacement.

5. How Can Explanatory Improvement Be Assessed Systematically?

The historical example illustrates how established principles of explanatory comparison can be applied to successive accounts of a particular biological phenomenon.

Step 1 — Specify the explanatory target

Identify the precise biological phenomenon or question to be explained.

Without a common explanatory target, two accounts may appear to disagree when they are actually answering different questions.

Step 2 — Reconstruct the strongest existing explanation

Establish what the existing explanation already accounts for.

This includes its mechanisms, assumptions, evidence, predictions, and acknowledged limitations.

An assessment that understates the existing explanation may manufacture an apparent improvement.

Step 3 — Identify a consequential unresolved question

A genuine limitation might concern an unexplained observation, an uncertain causal relationship, an untested assumption, or an inability to distinguish competing possibilities.

The limitation must be independently identifiable rather than created by the vocabulary of the proposed replacement.

Step 4 — Investigate a justified extension or revision

The appropriate scientific resources will depend on the question.

They may include experiments, mechanistic investigation, mathematical modelling, comparative analysis, or a combination of these approaches.

The revision must be supported by the relevant evidence.

Step 5 — Compare the explanatory achievements

The proposed improvement should be assessed against the original explanation on the same target.

The assessment should identify exactly what additional understanding or discrimination has been established, rather than awarding credit for new terminology or greater conceptual complexity.

Step 6 — Record the warranted outcome

The revised explanation may establish genuine improvement, prove equivalent to the earlier account, introduce disadvantages, or remain unresolved.

A disciplined assessment must allow all these outcomes.

This procedure is not presented as a newly discovered scientific method. Comparison, criticism, modelling, revision, and testing have long been central to scientific practice.

Its value is as an explicit way of keeping the criteria for explanatory improvement visible.

Nor does it imply that scientific discovery always proceeds through an orderly sequence. Actual research is often exploratory, opportunistic, and historically contingent.

6. Different Historical Patterns of Improvement

Not all changes in scientific explanation follow the same pattern.

Three examples illustrate important differences.

Chemiosmosis illustrates explanatory competition and substantial theory replacement. Peter Mitchell’s chemiosmotic account challenged prevailing explanations of how energy is coupled to ATP synthesis and became central to modern bioenergetics.

Sickle-cell disease illustrates cumulative explanatory extension. Research connected the clinical condition with haemoglobin, molecular variation, red-cell behaviour, and disease mechanisms, progressively extending understanding across molecular and physiological processes.

Bacterial chemotaxis illustrates bounded refinement of an already successful explanation. The later integral-feedback analysis clarified generality, assumptions, and constraints without replacing the established biochemical feedback account.

These distinctions are matters of emphasis, not rigid categories. Scientific developments can combine replacement, extension, and refinement.

Their significance is that explanatory improvement should be assessed according to the actual change in scientific understanding, rather than classified as progress simply because a newer theory has appeared.

7. Can an Explanatory Framework Improve Itself?

The same principles can be applied to broader frameworks that organise biological knowledge.

A framework may connect explanations of metabolism, development, behaviour, ecology, and evolution. It may provide shared concepts that help researchers understand how different biological processes relate to one another.

But the framework should also be open to correction.

One possible approach is to compare its explanations with the strongest established scientific accounts, identify consequential deficiencies, make carefully limited revisions, and assess whether those revisions genuinely improve explanatory performance.

If this process is repeated successfully, the framework could undergo iterative explanatory improvement.

The expression self-improvement needs qualification here.

A scientific framework does not independently detect and correct its own errors. Researchers make comparisons, interpret evidence, revise claims, and judge the results.

The process is therefore better understood as researcher-mediated iterative improvement.

Importantly, a framework may survive a comparison without revision. If an established biological explanation already accounts adequately for the target, adding the framework’s terminology may accomplish little beyond redescription.

A willingness to record such outcomes is part of scientific discipline, not a failure of the research enterprise.

8. What Does This Mean for APS?

The Agency–Process–Scale (APS) framework approaches biology through three related analytical perspectives: biological agency, the processes through which living organisation is maintained and transformed, and the scales across which that organisation is realised.

These are not separate hierarchical levels or competing causal forces. They are different analytical perspectives on materially realised biological organisation.

APS seeks to clarify relationships among viability, regulation, persistence, inheritance, development, and evolution.

This creates opportunities for explanatory comparison.

For example, APS accounts of biological function can be examined alongside established organisational theories of function, while its treatment of regulation can be compared with mechanistic and control-theoretic explanations.

Such comparisons can reveal substantial overlap, identify areas requiring clarification, or establish that an existing specialist explanation already adequately addresses the phenomenon.

APS research has investigated whether this approach can support a controlled cycle of explanatory correction.

The investigations have developed explicit comparison procedures and identified questions about explanatory specificity. They have not yet demonstrated a completed cycle in which a correction to APS produces an independently validated explanatory improvement.

A related investigation moved beyond APS to examine explanatory improvement in biological science generally.

The bacterial chemotaxis example illustrates how an established explanation can be refined through greater formal generality and clarification of its conditions.

The resulting assessment procedure provides a structured way to examine APS, but its component practices are already familiar in scientific inquiry, and no distinctive methodological contribution or independently validated improvement in explanatory performance has yet been established.

The appropriate ambition is therefore not to claim that APS possesses a special capacity for self-improvement.

It is to make APS open to systematic criticism, comparison, and justified correction using the best available scientific methods.

9. Why This Matters for Biological Understanding

Biology encompasses many specialised fields, each with its own methods, explanatory questions, and standards of evidence.

This diversity makes explanatory comparison difficult.

A molecular mechanism, a developmental model, and an evolutionary explanation may address different aspects of the same biological phenomenon without competing directly.

Progress requires attention to those differences.

A useful general framework can help clarify how explanations relate, where they overlap, and where further evidence is needed.

Yet the framework’s contribution must be distinguished from the explanatory achievements of the specialised sciences themselves.

The broader lesson is that scientific improvement is not measured by the quantity of concepts assembled into an explanation, but by the additional understanding that can be justified.

Sometimes this involves discovering a new mechanism.

Sometimes it involves establishing that an existing mechanism has a wider range of application.

Sometimes it means recognising that a proposed theoretical addition changes nothing of explanatory consequence.

All three outcomes contribute to responsible scientific understanding.

Conclusion

Biological explanations improve through discovery, criticism, refinement, and the clarification of their limits.

These improvements need not involve replacing an established theory or introducing an entirely new mechanism.

The history of bacterial chemotaxis shows how mathematical analysis can extend an existing explanation by clarifying the conditions of its success and the constraints on alternatives.

For APS, the practical implication is straightforward: explanatory claims should remain open to comparison with the strongest existing biological accounts, and revisions should be accepted only when their contribution can be demonstrated.

A framework that makes this discipline explicit may become more reliable and more useful over time.

Whether it achieves a demonstrable sequence of explanatory improvements remains a question for further research.

An explanation improves when it establishes something consequential that the earlier explanation did not already adequately establish.

Explanatory Architecture

The explanatory sequence developed in this article is:

Existing explanation → Identified limitation → Justified extension or revision → Comparative reassessment → Demonstrated improvement, equivalence, or adverse result

This sequence is an analytical reconstruction of explanatory improvement, not a universal description of how scientific discovery occurs.

The relevant dependencies are:

  1. A fixed explanatory target. Improvement can be assessed only when the earlier and later explanations address the same biological question, or when a change of scope is explicitly identified.

  2. An adequately reconstructed baseline. The strongest existing explanation must receive credit for what it already establishes, including its mechanisms, assumptions, predictions, and explanatory scope.

  3. A consequential difference. A proposed extension must establish something that changes justified understanding, causal discrimination, predictive or intervention-based expectations, or the conditions under which an explanation applies.

  4. Independent warrant. Additional explanatory content must be supported by appropriate empirical, mathematical, or theoretical evidence rather than by new terminology alone.

  5. Comparative adjudication. The revised explanation must be assessed against the unchanged baseline, allowing genuine improvement, equivalent redescription, deterioration, or unresolved outcomes.

  6. Controlled iteration. A demonstrated improvement may become the baseline for further investigation, but repetition does not guarantee cumulative progress.

These dependencies apply to biological explanations generally. They are not claimed as a methodological invention of APS.