Selection Revisited — What Does Selection Actually Act On?

Where This Article Fits

This article is a clarification within the APS Evolutionary Dynamics corpus. It asks what natural selection acts on without assuming that APS can identify a universal unit, level, scale, or object of selection by framework stipulation.

Its task differs from the core Natural Selection article, which characterises selection as contributing to differential historical stabilisation. Here the narrower question is which biological entities and relations are relevant when that process is explained in a particular system. Genes, organisms, groups, lineages, and other biological organisations may be relevant to different selection problems, and their explanatory importance must be established relative to the explanandum.

Organised persistence provides APS with a way of investigating historical biological continuity, but this does not make organised persistence the universal unit of selection. Likewise, APS Scale concerns the extent required for an explanation and should not be identified with a level of selection. The article therefore clarifies the selection problem while leaving its particular biological resolution to evidence and target-specific analysis.

Why Revisit What Selection Acts On?

Natural selection is central to evolutionary biology, yet the apparently simple question of what selection acts on opens several different explanatory problems. Genes, organisms, groups, populations, and other biological entities can enter selection explanations in different ways, while debates over units and levels of selection have shown that terms such as object, unit, interactor, beneficiary, and level should not automatically be treated as equivalent (Okasha 2006; Godfrey-Smith 2009).

APS approaches these questions from the standpoint of living organisation, but it does not use that standpoint to nominate a new universal unit of selection. Its concern is instead with how selection contributes to evolutionary transformation where biologically relevant differences have differential historical consequences.

This distinction matters because two questions can otherwise become conflated. One asks which entities and causal relations participate in a particular instance or model of selection. The other asks how the evolutionary consequences of selection should be understood within the broader continuity and transformation of living organisation. APS has a distinctive perspective on the second question. That perspective does not, by itself, settle the first.

The resulting aim is therefore narrower than answering once and for all what selection “actually” acts on. APS seeks to clarify selection’s place within evolutionary explanation while leaving the relevant units, levels, entities, and causal dependencies open to specification for the biological case concerned.

What Selection Requires — and What It Does Not

For selection to have evolutionary consequences, there must be biologically relevant differences capable of producing differential historical outcomes. How those differences arise, persist, recur, or are transmitted depends upon the system under investigation. Inheritance, development, variation, ecological interaction, and other processes may therefore be important to particular selection explanations.

APS distinguishes this minimal requirement from the stronger claim that every concept appearing in its evolutionary architecture must be a universal prerequisite for selection. Agency, viability, organised persistence, inheritance, variation, adaptation, fitness, natural selection, and evolutionary transformation identify distinguishable explanatory concerns within APS. Their inclusion and ordering do not establish that every instance of selection depends upon them through one necessary causal or temporal sequence.

This also means that selection should not be assigned all explanatory responsibility for the alternatives among which differential outcomes occur. The processes generating, maintaining, reproducing, or modifying biologically relevant differences raise further questions. At the same time, APS need not assert the converse proposition that selection can make no contribution to the evolution of biological organisation. The relevant causal relations must be established for the particular evolutionary system.

Selection can therefore be given a bounded explanatory role without either making it the source of biological organisation or deriving its operation from the complete APS architecture.

Selection as Differential Historical Stabilisation: An APS Interpretation

Within APS, natural selection is interpreted as contributing to differential historical stabilisation among biologically relevant alternatives. The formulation directs attention to the fact that selection matters evolutionarily because differences have consequences extending through biological history.

This is an APS explanatory interpretation, not a proposed replacement definition of natural selection. Established selection theory provides formal and empirical ways of analysing Darwinian populations, fitness differences, inheritance, reproduction, and differential evolutionary outcomes (Godfrey-Smith 2009). APS asks a further organisational question: how should those outcomes be understood in relation to the continuity and transformation of living organisation?

“Historical stabilisation” therefore identifies the explanatory role APS assigns to selection within its evolutionary architecture. It does not specify in advance which entity is the unit or object of selection, nor does it imply that everything that persists through evolutionary history is itself selected.

The distinction is important. Historical persistence is an outcome requiring explanation; identifying the entities and causal relations through which selection occurs is part of providing that explanation. The two questions must not be collapsed into one another.

What Persists Is Not Necessarily What Selection Acts On

APS uses organised persistence to analyse the continuity of living organisation through change. This makes organised persistence relevant to evolutionary explanation, because evolution requires both historical continuity and transformation. It does not follow, however, that organised persistence is therefore a newly identified biological unit upon which selection universally acts.

A lineage may persist historically while the processes explaining that persistence involve genetic, organismal, ecological, population-level, or other relations. Conversely, an entity identified in a selection model need not correspond straightforwardly to the complete organisation that persists through evolutionary history. Questions about persistence and questions about units or levels of selection therefore intersect without being identical.

This distinction corrects an important possible overextension of APS terminology. To redescribe genes, organisms, groups, or lineages as manifestations of organised persistence would not by itself explain selection more successfully. APS gains explanatory significance only where its organisational analysis identifies relevant dependencies, resolves an explanatory problem, or otherwise improves explanation relative to a comparator addressing the same explanandum.

Organised persistence should therefore be understood here as an APS explanatory target and organising concept. It is not, merely by virtue of that role, the universal object, interactor, beneficiary, or unit of natural selection.

Units, Levels, and Organised Persistence

The units and levels of selection have been analysed in several ways within evolutionary theory. Different accounts distinguish among biological entities and the roles those entities occupy in selection processes, while multilevel approaches investigate circumstances in which selection can be represented at more than one level of biological organisation (Okasha 2006; Godfrey-Smith 2009).

APS does not replace these distinctions with a single organisational answer. Instead, it asks how the entities identified in a selection explanation participate in the persistence and transformation of living organisation. This may provide useful explanatory context, but context should not be confused with adjudication.

The framework therefore does not infer that genes, organisms, groups, populations, or lineages must always be the relevant unit of selection. Nor does it infer that “organised persistence” supersedes these alternatives. Which entities matter, in what causal role, and at what level of description depends upon the biological system, the selection model, and the explanandum.

This restraint is methodologically important. A framework can integrate several kinds of explanation without thereby showing that its own vocabulary supplies a superior account of the processes being integrated. APS must therefore remain answerable to the strongest comparator addressing the same selection problem.

Selection and Fitness

APS distinguishes fitness and natural selection because they answer different explanatory questions within its evolutionary architecture.

Fitness is interpreted in APS in terms of differential historical continuity: differences relevant to how biological organisation is represented or persists through historical time. Natural selection is interpreted in terms of differential historical stabilisation: the contribution selection makes to differential evolutionary outcomes among biologically relevant alternatives.

This is an analytical distinction within APS. It should not be read as a claim that fitness and selection are universally separate biological stages or that fitness must first arise as an independent process and then cause selection. Established evolutionary accounts connect fitness and selection in ways that depend upon how populations, reproduction, heredity, and differential success are modelled (Godfrey-Smith 2009).

The APS distinction is therefore useful insofar as it keeps two explanatory questions visible: what differences are relevant to differential historical continuity, and how does selection contribute to their evolutionary consequences? Their particular causal relationship must be specified for the biological system concerned.

Inheritance, variation, adaptation, fitness, and natural selection can accordingly occupy distinguishable positions within APS without forming a mandatory biological sequence.

Genes and Organisational Context

Gene-centred approaches have provided powerful resources for evolutionary explanation, and gene-level models can illuminate inheritance, differential representation, and evolutionary change. APS does not reject those achievements. Nor does the fact that genes function within larger biological systems establish that gene-level selection explanations are invalid (Dawkins 1982).

APS instead applies a narrower control. Explanatory success at the genetic level does not, by itself, establish universal causal or explanatory priority for genes. Genetic effects occur through biological contexts that include molecular, cellular, developmental, physiological, organismal, and environmental relations. Noble (2012) argues against assigning causal privilege a priori to a single biological level, while Dupré and Nicholson (2018) emphasise the processual and relational organisation through which biological persistence is maintained.

This organisational contextualisation should not be confused with a solution to the units-of-selection problem. Depending upon the explanandum and model, gene-level description may be highly informative. The relevant question for APS is what additional explanatory work, if any, is accomplished by situating such descriptions within the organisation that makes their biological effects possible.

APS therefore neither replaces genes with organised persistence as the “true” unit of selection nor grants genes automatic explanatory primacy. The appropriate explanatory extent must be determined by the problem being investigated.

Selection Across Biological Scales

Selection should not be described as inherently multiscale simply because living organisation can be investigated across multiple scales. APS Scale concerns the spatial, temporal, and organisational extent over which relations relevant to an explanation are investigated.

Some selection problems may require relations extending across several such extents. A biological explanation might, for example, need to connect processes occurring within organisms with population dynamics or ecological conditions. In other cases, a more restricted explanatory extent may be sufficient. Whether cross-scale relations are required must be demonstrated from the biological system and explanandum rather than inferred from APS architecture.

APS Scale must also be distinguished from a level of selection. Levels-of-selection analysis concerns the structure and attribution of selection within evolutionary models and biological systems (Okasha 2006). APS Scale asks how far an investigation must extend in space, time, and organisation to capture the relevant dependencies. The two can intersect, but they are not synonymous.

A multiscale description likewise does not establish a multiscale causal explanation. Where developmental, physiological, ecological, population, or lineage relations are claimed to interact causally in selection, those dependencies require independent biological support.

The value of Scale is therefore methodological rather than prescriptive: it prevents the explanatory extent of a selection problem from being fixed in advance.

Selection Within the APS Evolutionary Architecture

APS situates natural selection within a broader explanatory architecture of continuity and transformation. Selection remains a central evolutionary process, but its place in that architecture does not make it the source of all biological organisation, nor does the architecture determine the complete causal conditions under which selection occurs.

Agency, viability, organised persistence, inheritance, variation, adaptation, fitness, natural selection, and evolutionary transformation address distinguishable explanatory questions. Their arrangement allows APS to ask how viability-oriented organisation persists, how differences arise or are maintained, how differential historical outcomes occur, and how living organisation is transformed through evolutionary time.

Within this architecture, APS assigns selection the interpretive role of contributing to differential historical stabilisation among biologically relevant alternatives. Other processes may be involved in generating, maintaining, transmitting, reproducing, or modifying those alternatives. Which dependencies are relevant must be established for the particular biological case.

The architecture is therefore conceptual and explanatory. It is not, simply by virtue of its ordering, evidence for a universal causal chain, necessary temporal progression, or hierarchy of explanatory importance.

Explanatory Architecture

The explanatory structure of this article can be stated without turning it into a universal biological pathway.

Biologically relevant differences are required for differential evolutionary outcomes. How those differences are generated, maintained, reproduced, or transmitted depends upon the biological system.

Fitness, in APS, concerns differential historical continuity. It provides one way of analysing differences relevant to the representation or persistence of biological organisation through time.

Natural selection, in APS, is interpreted as contributing to differential historical stabilisation among biologically relevant alternatives. This identifies its role within APS evolutionary explanation without specifying a universal object or level of selection.

Evolutionary transformation concerns the resulting historical change in living organisation, but the causal relations connecting particular differences, fitness relations, selection processes, and evolutionary outcomes must be established independently.

This architecture therefore organises explanatory questions rather than prescribing a necessary sequence:

difference → differential outcome → historical stabilisation → evolutionary transformation

Even this compact representation should not be read as a universal causal chain. It identifies distinctions that an APS explanation may need to investigate; the biological dependencies among them remain empirical.

What APS Adds — and What It Does Not Establish

APS contributes an organisational perspective to selection by asking how differential evolutionary outcomes relate to the historical continuity and transformation of living systems. This perspective can help distinguish the persistence requiring explanation from the particular entities and causal relations through which selection occurs.

Its value cannot, however, be established merely by replacing the vocabulary of genes, organisms, groups, populations, or lineages with the vocabulary of organised persistence. Existing evolutionary approaches already distinguish units and levels of selection, analyse Darwinian populations, and examine the relation between selection and biological hierarchy (Okasha 2006; Godfrey-Smith 2009). Gene-centred explanations likewise cannot be dismissed merely because genetic effects depend upon wider biological organisation (Dawkins 1982).

APS can therefore claim explanatory gain only where its organisational analysis identifies a relevant dependency, resolves an explanatory problem, or otherwise improves explanation relative to the strongest comparator addressing the same explanandum. Integration, conceptual coherence, or broader vocabulary alone does not satisfy that criterion.

Accordingly, this article does not establish a universal unit of selection, does not claim that selection is necessarily multiscale, and does not derive selection from the complete APS evolutionary architecture. Its more limited purpose is to clarify how selection can be situated within an organisational account of evolutionary continuity while keeping the relevant biological dependencies open to investigation.

Conclusion

Asking what natural selection acts on remains a legitimate biological question, but APS does not answer it by nominating organised persistence as a new universal unit. The framework instead separates the question of what persists through evolutionary history from the question of which entities and causal relations participate in particular selection processes.

Within that distinction, APS interprets selection as contributing to differential historical stabilisation among biologically relevant alternatives. The formulation situates selection within evolutionary transformation while leaving the units, levels, causal dependencies, and explanatory extent of particular selection processes to biological investigation.

This bounded interpretation also preserves distinctions elsewhere in the APS evolutionary architecture. Fitness and selection can address different explanatory questions without becoming successive stages in a universal pathway. Genetic explanations can remain powerful without receiving automatic explanatory priority. Cross-scale relations can matter without making selection inherently multiscale. And organised persistence can provide a common explanatory target without becoming, by definition, the object upon which selection acts.

The resulting account is therefore not an alternative mechanism of natural selection. It is an attempt to clarify selection’s place within the broader explanation of biological continuity and transformation while remaining answerable to established evolutionary theory and to the biological evidence relevant to each case.

Continue Exploring

• Fitness — Differential Continuity and Organised Persistence
• Adaptation — How Living Systems Sustain Themselves Through Change
• Inheritance and Continuity in APS
• Variation in APS — Where Does Novelty Come From?
• What Is Evolution in APS?
• What Is a Species in APS?
• What Is a Taxon in APS?

Key Point

APS interprets natural selection as contributing to differential historical stabilisation among biologically relevant alternatives. This situates selection within evolutionary transformation without determining in advance the universal unit, level, scale, or complete causal architecture of selection.