Auditing Dependency Claims in Biological Explanation: A Comparative Method and Three-Case Test

1. Introduction: The Comparison Problem

Biological explanations identify many different ways in which one phenomenon depends on another. A change in a signalling molecule may causally alter a developmental outcome. A cellular activity may constitute part of a larger mechanism. A structure may enable a process without producing it. A constraint may delimit the transformations available to a system. Spatial organisation may determine which interactions are possible, while developmental or evolutionary history may explain why a present organisation exists. These claims all concern dependency, but they do not identify the same kind of relation, support the same counterfactuals, or require the same evidence.

Causation remains central to biological explanation, but it does not exhaust it. Scientists often use causal language broadly to describe any condition whose alteration changes an outcome. That usage is frequently harmless, yet it can obscure whether a claim concerns productive influence, constitution, enablement, constraint, spatial organisation or historical origin. The present analysis does not oppose causal explanation or attempt to settle the metaphysics of causation. It asks only that causal claims be distinguished from adjacent dependency claims when the distinction changes what evidence is relevant or what explanatory conclusion is warranted (Woodward 2003; Craver 2007; Reutlinger 2016).

This heterogeneity presents a comparative problem. Biological accounts commonly integrate molecular interventions, cellular structures, regulatory networks, mathematical models, historical reconstructions, and higher-order descriptions of organisation. Such integration is often scientifically productive. It can nevertheless conceal differences between the explanatory roles assigned to individual findings. Evidence that a factor changes an outcome does not by itself establish that the factor constitutes the phenomenon, enables its occurrence, constrains its possible forms, or explains its historical origin. Conversely, a condition that is indispensable to a system may not be the salient cause of the outcome selected by a particular contrast.

Established philosophy of science already supplies powerful resources for analysing these distinctions. Dependence-based approaches connect explanation with information about the relations supporting a phenomenon (Taylor 2023). Interventionism analyses causal relevance through controlled variation and invariance (Woodward 2003), while counterfactual approaches extend dependency analysis beyond narrowly causal cases (Reutlinger 2016). Mechanistic accounts require materially organised entities and activities (Machamer, Darden, and Craver 2000), and constitutive explanation distinguishes relations between a phenomenon and its component organisation from ordinary causal relations (Craver 2007). Explanatory pluralism, meanwhile, cautions against expecting every scientifically legitimate account to conform to one explanatory form (Mitchell 2002).

The problem, therefore, is not that biology lacks accounts of dependency. It is that the relevant analytical operations are distributed across explanatory traditions and are not always applied through a common comparison. An inclusive biological narrative may contain several individually warranted claims while leaving unclear which relation each claim establishes, what material organisation realises it, what evidence supports that relation rather than an adjacent one, and whether the resulting synthesis improves on the strongest established explanation.

The governing question of this article is consequently limited: can a common, relation-typed and materially explicit audit of dependency claims improve the comparative assessment of biological explanations without introducing a new theory of explanation or presuming that every case requires improvement?

The article addresses this question by developing and testing a comparative dependency-hypothesis audit. A dependency hypothesis is treated simply as a testable explanatory claim that a specified relation obtains between identified relata under stated conditions. The proposed contribution does not consist in identifying dependency as important, extending counterfactual analysis beyond causation, introducing mechanistic or constitutive explanation, or discovering organisational constraints. Each of those achievements has substantial precedents. The residual proposal is a common procedure for bringing their relevant demands into one comparative assessment while preserving differences among relation types.

The test is deliberately asymmetric. Planarian regeneration serves as the principal positive case: the procedure must change at least one defensible assessment of explanatory sufficiency or evidential strength rather than merely reorganising familiar findings. Pristionchus pacificus mouth-form polyphenism serves as the transfer case: the same procedure must operate in an environmentally responsive, irreversible developmental switch without importing assumptions specific to regeneration. Mammalian glucose homeostasis serves as the calibrated null: if the strongest endocrine, mechanistic and control-theoretic account already distinguishes and evidences the relevant dependencies, the procedure must return no material explanatory gain. These cases are selected for diagnostic contrast rather than as a representative sample of biology.

These cases are diagnostic rather than representative of biology. Their sequence tests capacity, transfer and restraint. A positive result alone could reflect favourable case selection; a second positive result alone could repeat the same explanatory architecture. The null is therefore constitutive of the design. A method that always reports improvement would not discriminate explanatory gain from redescriptive complexity.

The thesis defended is correspondingly restrained. Biological explanations advance heterogeneous dependency claims whose adequacy cannot be assessed by a single evidential test. A common comparative procedure may improve assessment where relations are conflated or claims exceed their evidence, while properly declining to claim improvement where an established explanation already performs this work. The intended contribution is an integrative methodological instrument, not a new biological mechanism, dependency ontology, or universal theory of explanation.

2. What Is Already Established

The proposal begins from extensive anticipation rather than a methodological vacuum. Its most direct conceptual precedent is the view that explanation provides information about dependence. Taylor’s account of explanatory distance holds that explanations are supported by backing relations and allows those relations to include causal, grounding, mereological, conceptual, and other forms of dependence (Taylor 2023). This directly defeats any claim that explanation as dependence, or pluralism about dependence relations, originates here. What remains open is whether heterogeneous biological dependency claims can be subjected to a common, materially disciplined comparison.

Interventionism supplies the strongest established operational account of causal dependency. Explanatory claims answer what-if-things-had-been-different questions by identifying how interventions on one variable would change another, while invariance specifies the range across which the relevant generalisation remains stable (Woodward 2003). These resources already establish the importance of controlled variation, background conditions, scope and causal difference-making. The present proposal cannot claim novelty for any of them.

Nor can it define every dependency through intervention. Constitutive relations may not permit their relata to be independently manipulated without altering or destroying the phenomenon being investigated. Mechanistic philosophy makes this problem especially clear. Mechanisms explain through organised entities and activities whose productive and spatial organisation gives rise to regular changes (Machamer, Darden, and Craver 2000). Constitutive mechanistic explanation further asks how components and their organisation are relevant to the phenomenon they compose, imposing evidential demands different from those governing causal relations between separable variables (Craver 2007). Mechanistic analysis therefore already provides a powerful model of material explicitness, decomposition, localisation and relation-sensitive evidence.

Counterfactual dependence is also not confined to productive causation. Reutlinger argues that causal and non-causal explanations can be understood through the counterfactual dependencies they reveal (Reutlinger 2016). This blocks any suggestion that the present method first generalises dependency analysis beyond causal explanation. It also reinforces a central limitation: counterfactual information must be interpreted in relation to the kind of dependency claimed. A counterfactual pattern alone does not determine whether the relation is causal, constitutive, enabling, structural or constraint-based.

Explanatory pluralism provides a further anticipation. Integrative pluralism holds that complex phenomena can require multiple models and explanatory perspectives whose contributions need not be reducible to a single privileged account (Mitchell 2002). The coexistence of heterogeneous explanatory forms is therefore established. The remaining issue is comparative discipline: pluralism permits multiple accounts, but it does not by itself determine whether an integrated explanation has kept their relations, evidence and sufficiency conditions distinct.

The closest biological-organisational precedents narrow the residual contribution still further. Biological autonomy describes organisms as self-maintaining organisations whose component constraints and processes contribute to maintaining the conditions under which the organisation persists (Moreno and Mossio 2015). Closure-of-constraints accounts formalise mutually dependent constraints that act on underlying processes and are maintained through their collective organisation (Montévil and Mossio 2015). Organisational accounts of function ground functional attribution in a component’s present contribution to maintaining an organisation that, in turn, maintains the component (Mossio, Saborido, and Moreno 2009). Cross-generation cases demonstrate that even this account requires careful decisions about organisational continuity, temporal scope and the bearer of functional contribution (Saborido, Mossio, and Moreno 2011).

These approaches already make materially organised dependence central to biological explanation. More directly still, mechanism and autonomy have been integrated within a single explanatory architecture connecting mechanisms, control, self-maintenance and closure of constraints (Bich and Bechtel 2021). Integrated biological control has likewise been analysed as an organisation of materially realised constraints rather than as an aggregate of isolated regulatory causes (Bich and Bechtel 2022). Critical examination of organisational normativity and teleology further shows that establishing closure or contribution does not automatically establish every stronger normative or purposive conclusion associated with it (Corti 2023).

The consequences for novelty are decisive. The present article must not claim to originate dependence-based explanation, counterfactual testing, intervention, invariance, mechanism, constitution, constraint, explanatory pluralism, biological autonomy, organisational closure, organisational function, integrated control, or materially explicit biological explanation. Nor does placing these resources together establish novelty by itself. An integration is methodologically consequential only if it changes a defensible assessment, identifies an evidential mismatch or relation conflation, or demonstrates that an apparent omission is already adequately handled.

The residual contribution is therefore procedural and test-dependent. Established approaches provide nearly every component operation, but they ordinarily develop those operations within particular explanatory families. The proposal retained here asks whether placing relation type, material basis, evidential warrant, alternatives, scope, failure and strongest-rival comparison under one common discipline can improve assessment across heterogeneous biological explanations—and whether that same discipline can explicitly refuse improvement when no gain is warranted. The next section specifies that procedure; the three subsequent cases determine whether the residual is substantive enough to survive its acknowledged precedents.

3. The Dependency-Hypothesis Audit

The dependency-hypothesis audit is a repeatable procedure for comparing explanatory claims, not a new theory of explanation or an ontology of biological organisation. It does not assume that every biological phenomenon has one privileged kind of dependency or that all dependencies can be tested in the same way. Its purpose is to make explicit what an explanation claims, how that claim is materially realised and evidenced, and whether clarifying those matters changes its assessment relative to the strongest established account.

A dependency hypothesis states that specified relata stand in an explanatorily relevant relation under stated conditions. The relata may be variables, entities, activities, components, constraints, structures, spatial arrangements, temporal sequences or historically produced organisations. Calling the statement a hypothesis does not imply that its relation is necessarily causal or experimentally manipulable. It means that the relation, its explanatory relevance and its evidential warrant remain assessable.

The audit begins by fixing the explanandum and foil. The explanandum identifies the phenomenon to be explained; the foil identifies the relevant alternative, absence, failure or contrast. This prevents apparent gain from being manufactured by changing the explanatory question during comparison. A factor may explain why one outcome occurred rather than another while remaining insufficient to explain how the complete phenomenon is constituted or maintained.

The dependency claim must then identify its relata and the relation asserted between them. Relation typing is provisional and claim-specific, not an exhaustive taxonomy. Causal dependencies concern changes that produce or modify an outcome and are appropriately investigated through intervention, controlled variation, invariance and supported counterfactuals (Woodward 2003). Constitutive dependencies concern how components and their organisation compose a phenomenon; their assessment may require localisation, decomposition, inclusion and carefully constrained mutual manipulation rather than intervention between independently existing variables (Machamer, Darden, and Craver 2000; Craver 2007).

Enabling dependencies identify conditions or capacities required for a phenomenon without treating them as its salient productive cause. Constraint dependencies identify restrictions on possible processes or transformations. Organisational constraint accounts may additionally specify how constraints are materially maintained through mutual dependence (Montévil and Mossio 2015; Bich and Bechtel 2021). Structural dependencies concern topology, arrangement or formal organisation and may be supported by model–target correspondence, derivation, invariant structure and comparative perturbation. Spatial and temporal dependencies concern where and when relevant interactions occur, including order, duration, delay and critical windows. Historical dependencies concern how an existing organisation arose; evidence for historical origin must not be substituted for evidence of present causal or constitutive operation.

These types may overlap within an explanation, but they should not be collapsed. A molecular intervention can establish causal relevance without establishing that the manipulated factor constitutes the larger phenomenon. Indispensability can identify an enabling condition without demonstrating explanatory priority. Persistence after perturbation may establish stability without uniquely identifying a stored target. Evolutionary origin may explain why a regulatory architecture exists without constituting its present operation. Counterfactual dependence is therefore an important explanatory resource but does not determine relation type by itself (Taylor 2023; Reutlinger 2016).

Material specification records the entities, processes, structures or conditions through which the proposed relation is realised. This requirement is resolution-sensitive. Mechanistic claims ordinarily require detailed material organisation, whereas a legitimate structural or abstract explanation need not be converted into a lower-resolution mechanism merely to pass the audit. The question is whether the degree of material specification is appropriate to the explanatory work claimed.

Evidence must be matched to the relation. Interventions, knockouts, pharmacological perturbations and controlled environmental changes can support bounded causal claims. Imaging, localisation, decomposition and component manipulation can support constitutive or spatial claims. Time-course evidence can support temporal order and critical-window claims. Comparative, phylogenetic and genomic evidence can support historical claims. Mathematical and computational models may reveal structural relations, dynamics or invariance, but fitted parameters and simulated dependencies are not automatically material mechanisms. Every evidence entry must therefore record what it establishes and what remains inferential.

Alternatives, scope and failure conditions complete the assessment of the individual claim. Alternatives identify competing relations, mechanisms or interpretations compatible with the evidence. Scope records the organisms, regimes, spatial and temporal ranges, background conditions and explanatory resolution within which the claim is warranted. Failure conditions state what would defeat or materially weaken the claim: failure of the relation to obtain, incorrect typing, absent material realisation, insufficient or mismatched evidence, survival of a stronger alternative, loss of the result outside an undisclosed scope, or absence of comparative gain.

The audit then reconstructs the strongest established explanation. The comparator must be the best integrated scientific account available in the frozen corpus, not a textbook simplification, isolated factor or obsolete position. This requirement is essential to the redundancy test. A richer description counts as explanatory gain only if it improves assessment relative to what the strongest account already distinguishes and supports.

Five decision rules govern the final verdict. A changed assessment occurs when the audit defensibly alters the relation assigned to a claim, its evidential strength, scope, sufficiency or comparative standing. Genuine explanatory gain requires such a change to improve understanding of the explanandum relative to the strongest account—for example, by exposing relation conflation, evidential mismatch, a materially consequential omission, an untested alternative or an overstated sufficiency claim.

Redescription occurs when the audit reorganises established findings without changing any defensible explanatory assessment. Redundancy occurs when the strongest comparator already performs the relevant relation typing, material specification, evidence matching, qualification and failure analysis without meaningful loss. A principled no-gain verdict records this sufficiency as a result rather than treating it as failure. It demonstrates that the procedure can refuse improvement where improvement is unwarranted.

The method succeeds only if these verdicts are evidence-governed. Positive findings cannot be inferred merely from biological complexity, and null findings cannot be predetermined by case assignment. The same fields, comparator standard and failure rules must be applied in every case. The audit’s residual methodological value consequently depends on whether it can distinguish assessment-changing clarification from disciplined redescription while preserving the legitimate plurality of biological explanation (Mitchell 2002; Bich and Bechtel 2022).

4. Positive Test: Planarian Regeneration

Planarian regeneration provides the principal positive test because its established explanation integrates cellular supply, wound responses, positional information, signalling polarity, tissue remodelling and bioelectric regulation. The explanandum is restoration of the context-appropriate body plan after amputation: missing structures must be regenerated with correct anterior–posterior polarity, proportion and termination. The principal foil is not failure to produce new cells, but production of viable tissue with incorrect identity, location, polarity or scale.

The strongest established account begins with an injury response followed by neoblast proliferation, migration and differentiation. These adult stem cells generate the cell types required to replace missing tissue. Regeneration also requires positional information that distinguishes anterior from posterior and specifies which structures belong at particular locations. Muscle cells express position-control genes across the body and reset aspects of this coordinate information after injury. Wnt signalling favours posterior identity, whereas injury-induced notum expression at anterior-facing wounds inhibits Wnt activity and supports head regeneration. Differentiation, patterning, tissue turnover and remodelling then reconstruct proportional anatomy rather than merely filling the wound (Reddien 2018; Witchley et al. 2013).

This account is already mechanistically rich. The audit does not replace it. It asks whether distinct achievements within it support the same explanatory conclusion.

Neoblast availability is an enabling and causal dependency. Depletion and transplantation evidence establishes that neoblasts are required for producing regenerated cell types and that suitable neoblast populations can restore regenerative capacity (Reddien 2018). The material basis includes proliferating stem cells, their progeny and differentiation pathways. This evidence strongly supports cellular supply. It does not, by itself, explain why the supplied cells acquire the correct identities or assemble into the appropriate form. The assessment therefore changes from “neoblasts explain regeneration” when that phrase implies whole-form sufficiency to “neoblasts explain the indispensable production of replacement cells within a larger patterning organisation.”

Muscle-expressed position-control genes contribute constitutive, spatial and enabling dependencies. Muscle is not merely contractile tissue in this context: its distributed gene expression materially realises positional information used during tissue maintenance and regeneration. Perturbations and expression mapping support the claim that muscle provides instructions associated with regional identity (Witchley et al. 2013). The evidential limit is equally important. Position-control-gene expression identifies a spatially distributed specification system, but no single expression map constitutes the regenerated anatomy or independently explains how all tissues interpret, coordinate and terminate their responses.

Wnt/notum regulation supplies a more specifically causal, constraint and temporal dependency. The orientation of a wound and the early regulation of Wnt signalling help constrain whether anterior or posterior structures develop. Manipulating pathway components changes polarity outcomes, while the timing and location of notum expression connect wound orientation to anterior specification (Reddien 2018). Here the causal inference is strong but bounded: the pathway explains a decisive polarity choice, not every subsequent feature of head or tail construction. Calling Wnt/notum a switch is therefore appropriate for the selected contrast between alternative polar outcomes, but insufficient as a complete account of form restoration.

Tissue remodelling adds constitutive and temporal dependencies. Planarian form is restored through coordinated addition, loss, reassignment and proportional adjustment of tissues. Existing tissue is not a passive background surrounding a newly produced blastema. It participates in rescaling and integrating old and new structures over time (Reddien 2018). Evidence for these processes supports an account of regeneration as transformation of an existing organisation, but it does not establish that remodelling is directed by a single system-wide controller.

Bioelectric perturbations introduce a distinct causal and constraint claim. Transient alteration of physiological voltage states can modify early gene expression and regenerated anterior–posterior outcomes, including persistent changes expressed in later rounds of regeneration (Durant et al. 2019). The material basis includes ion-channel-dependent voltage patterns and their interactions with molecular signalling. These experiments establish that early bioelectric conditions can make a durable difference to polarity. They do not uniquely determine whether the persistence is stored in a dedicated anatomical memory, distributed across altered tissues, maintained through gene-regulatory states, or reproduced by another stable physiological organisation.

This distinction changes the assessment of target-memory language. Before the audit, persistent polarity after a transient intervention may be described as evidence that planarians store a target morphology guiding regeneration. After the audit, the evidence supports a narrower claim: an experimentally altered physiological state can persist or be reinstantiated strongly enough to bias later anatomical outcomes. A stored anatomical target remains one interpretation, but it is not uniquely established by the observed counterfactual dependence. Levin, Pietak, and Bischof’s anatomical-homeostasis framing appropriately identifies the explanatory importance of stable large-scale outcomes, but it does not eliminate alternatives concerning their material storage and implementation (Levin, Pietak, and Bischof 2019).

The comparison therefore produces genuine but limited explanatory gain. It does not add a regenerative mechanism or overturn the integrated account. It changes three defensible assessments. First, cellular sufficiency is separated from morphological sufficiency: generating required cell types is necessary but does not explain their correct spatial organisation. Second, pathway importance is separated from whole-form sufficiency: Wnt/notum and bioelectric perturbations establish bounded polarity dependencies rather than complete control of regeneration. Third, persistent anatomical effects are separated from proof of a stored target: persistence strengthens the demand for a materially explicit account but underdetermines its proposed realisation.

These conclusions are scope-bound. Much evidence concerns Schmidtea mediterranea, with cross-species extension requiring independent support. Experimental interventions may produce states outside normal regeneration, and severe perturbations can alter several coupled processes simultaneously. Alternative explanations include distributed positional information, long-lived gene-regulatory change, tissue composition, altered ion-channel expression and recurrent network dynamics. The positive verdict would fail if the strongest established account already made all three distinctions explicitly, if the cited perturbations did not support the attributed dependency, or if the revised assessments depended only on replacing moderate claims with exaggerated targets.

The result is accordingly diagnostic rather than revolutionary. Established stem-cell, positional, signalling and bioelectric explanations remain primary. The audit improves their comparison by preventing evidence for cell production, regional specification, polarity choice and persistent state from being treated as interchangeable evidence for complete anatomical explanation.

5. Transfer Test: Pristionchus Polyphenism

The transfer test asks whether the audit works in a prospective developmental system without importing assumptions from regeneration. Pristionchus pacificus develops one of two adult mouth forms: a narrow stenostomatous form associated primarily with microbial feeding or a wide eurystomatous form equipped for predation. The explanandum is how environmental conditions experienced during juvenile development produce a discrete, effectively irreversible adult morphology. The foil is development of the alternative mouth form under an otherwise relevantly comparable genotype and developmental regime.

The strongest established explanation integrates environmental cues, neuronal and endocrine signalling, chromatin regulation, a dosage-sensitive genetic switch and morph-specific development. Population density, nutrition and pheromonal conditions influence mouth-form frequency. Stage-specific pheromone production also allows adults to affect juvenile phenotypes, demonstrating that the relevant environment is partly produced by conspecifics rather than consisting only of an external physical condition (Werner et al. 2018). Dafachronic-acid signalling provides an endocrine route through which environmental conditions can influence developmental outcome (Bento, Ogawa, and Sommer 2010).

These inputs converge on a switch architecture involving EUD-1, SEUD-1 and NHR-40. EUD-1, a sulfatase expressed in neurons, is a dosage-sensitive regulator whose alteration can strongly bias or fix mouth-form outcome (Ragsdale et al. 2013). This establishes EUD-1 as a powerful causal switch node. It does not establish that EUD-1 alone senses every environmental cue, constitutes the threshold, specifies every morphological feature or executes the complete developmental programme.

The audit changes the sufficiency assessment accordingly. Before auditing, EUD-1 may be described as the developmental switch and therefore appear to explain the mouth-form decision generally. After auditing, it is more precisely a causally decisive component within a distributed switch architecture. NHR-40 acts downstream of EUD-1 and contributes to the developmental decision, while SEUD-1 has an opposing dosage-dependent influence and operates with EUD-1 and NHR-40 in controlling switch outcome (Kieninger et al. 2016; Bui, Ivers, and Ragsdale 2018). The dependency is therefore not simply phenotype-on-EUD-1. It is a threshold relation realised through the relative activities and dosages of interacting regulatory factors.

This also changes the ontology of the threshold. A threshold need not be a separate material structure or a single molecular concentration at which development changes direction. Here it is a system property of interacting signals, gene dosage, enzymatic regulation and developmental state. Genetic manipulation demonstrates causal control over morph frequency and switch output, but the threshold is constituted by relations among those factors. Its exact quantitative form may vary across genotypes, environments and species. Treating it as relational avoids both reifying it as an independent controller and reducing it to one switch gene.

Chromatin and antisense regulation provide enabling, causal and developmental dependencies. Altering chromatin regulators changes eud-1 expression and mouth-form outcome, while antisense-eud-1 RNA contributes to regulation of the switch gene (Serobyan et al. 2016). These findings materially connect chromatin state with switch sensitivity. They do not, however, demonstrate that chromatin directly detects every environmental cue or stores a complete representation of the experienced environment. Before the audit, chromatin regulation may appear to provide the missing route from environment to phenotype. Afterwards, the evidence supports a narrower conclusion: chromatin and antisense regulation condition expression and developmental responsiveness within the switch, while the complete causal chain from particular cues to those states remains only partly resolved.

Developmental execution must also be distinguished from switch selection. Once the regulatory system biases one outcome, downstream tissue differentiation constructs the corresponding mouth morphology. The conserved multi-gene locus containing eud-1 and functionally opposed genes helps organise the switch, but locus architecture is neither identical to the environmental decision nor sufficient for morphogenesis (Sieriebriennikov et al. 2018). The irreversible adult outcome depends on the temporal closure of developmental possibilities: cues and regulatory states have different effects before and after the sensitive juvenile window.

Evolutionary history performs further explanatory work without substituting for present operation. Co-option of endocrine signalling, lineage-specific emergence of eud-1, conservation of the multi-gene locus and evolutionary changes in relative dosage help explain how the switch architecture arose and diversified (Bento, Ogawa, and Sommer 2010; Ragsdale et al. 2013; Bui, Ivers, and Ragsdale 2018). They do not constitute the proximate developmental process producing a particular individual’s mouth form. The audit therefore separates historical origin from current cue integration, threshold determination and morphological execution.

Alternative interpretations remain. Morph frequency may reflect combined effects of nutrition, crowding, pheromones, developmental rate and genotype rather than one environmental input. Endocrine, neuronal and chromatin pathways may operate in parallel or sequentially, and switch genes may integrate signals whose immediate substrates remain incompletely identified. Findings from laboratory strains and manipulations do not automatically determine the relative contribution of these dependencies in natural populations.

The transfer verdict is positive but moderate. The audit changes four assessments: EUD-1 is a decisive node rather than the complete switch explanation; the threshold is a relational property rather than an independent controller; chromatin regulation supports developmental responsiveness without completing the environment-to-phenotype chain; and evolutionary origin is distinct from present developmental operation. These changes preserve switch genetics and reaction norms as the primary explanations. They show that the audit transfers to irreversible prospective development without invoking anatomical targets, regenerative memory or restoration of prior form.

6. Null Test: Mammalian Glucose Homeostasis

The null test concerns adult mammalian regulation of circulating glucose across feeding, postabsorptive fasting and ordinary activity. The explanandum is how glucose appearance, uptake, use and storage are coordinated within a viable operating range while tissues receive fuel. The principal foils are persistent hyperglycaemia after nutrient entry and hypoglycaemia during fasting, exertion or excess insulin. The comparator is the strongest integrated endocrine, mechanistic and control-theoretic account, not a simplified insulin–glucagon diagram.

Rising glucose is linked to insulin secretion through a well-characterised causal and temporal chain: β-cell metabolism raises the ATP/ADP ratio, closes ATP-sensitive potassium channels, depolarises the membrane, opens voltage-dependent calcium channels and triggers insulin exocytosis (Rorsman and Ashcroft 2018). Electrophysiology, secretion dynamics, genetic models and pharmacological interventions establish both the material basis and the inferential limits of this account, including differences between human and rodent β-cells.

Islet output is constitutively, spatially and causally organised through β-, α- and δ-cell interactions, vasculature, innervation and paracrine insulin, glucagon and somatostatin signalling. Incretins—especially GLP-1—potentiate glucose-dependent insulin secretion, modify gastric emptying and contribute neural and temporal modulation of nutrient appearance (Holst 2007). These dependencies are already qualified by route, glycaemic state, receptor integrity and species-specific islet architecture (Röder et al. 2016).

Insulin lowers circulating glucose through distributed organ-specific causal and feedback relations. It suppresses hepatic glucose production, promotes storage and supports skeletal-muscle and adipose glucose uptake. During fasting, reduced insulin and increased glucagon promote hepatic glycogenolysis and gluconeogenesis. Mechanistic and tracer studies distinguish hepatic appearance, peripheral disposal, storage and substrate competition rather than treating “lowering glucose” as one undifferentiated action (Petersen and Shulman 2018).

Defence against hypoglycaemia is temporally ordered. Reduced endogenous insulin precedes glucagon and epinephrine responses, with cortisol, growth hormone and behavioural responses contributing over longer intervals. Recurrent hypoglycaemia can lower autonomic and symptomatic response thresholds, demonstrating that the regulatory range and its defences are historically adaptable rather than fixed around an immutable set point (Cryer 2013).

Quantitative models supply a complementary structural and control description. Minimal-model analysis separates insulin sensitivity from glucose-mediated disappearance, but its parameters remain dependent on experimental regime and model assumptions (Bergman et al. 1979). Models of reciprocal α–β-cell signalling likewise test how feedback architecture affects overshoot and coordinated hormone secretion without converting a simulated circuit into the complete material mechanism (Garzilli and Itzkovitz 2018).

Applying the audit changes no defensible assessment. Causal dependencies are already linked to intervention, infusion, clamp, tracer, genetic, pharmacological and electrophysiological evidence. Constitutive and spatial dependencies are materially specified through islet organisation and distributed organ action. Feedback and temporal dependencies are distinguished by secretion dynamics, nutrient state and counter-regulatory order. Enabling and constraint conditions—including receptor integrity, tissue sensitivity, glucose effectiveness and available substrate—are explicitly qualified. Evolutionary history enables the inherited regulatory architecture but is not substituted for its present operation.

Alternatives, scope and failures are also established. Type 1 diabetes removes endogenous insulin and changes intra-islet and counter-regulatory relations. Type 2 diabetes involves interacting insulin resistance, β-cell compensation and failure, inappropriate glucagon activity and altered incretin effects. Exogenous insulin does not reproduce endogenous timing or portal delivery, while recurrent hypoglycaemia alters future defence. Species, diet, activity, route and measurement regime constrain generalisation.

The verdict is therefore principled no gain. The audit produces an organised synopsis but does not change relation typing, evidential strength, scope, sufficiency or failure analysis relative to the strongest account. This is not methodological failure. It shows that the procedure can recognise comparator sufficiency and refuse improvement where its work has already been performed. The null thereby distinguishes diagnostic assessment from a method designed merely to redescribe every complex system as requiring additional organisation.

7. Comparative Result

The three applications used the same audit fields, comparator standard and decision rules but produced differentiated outcomes. This pattern—not the number of positive verdicts—provides the principal evidence for the method’s value.

In planarian regeneration, the audit produced bounded positive gain. It separated replacement-cell production from restoration of correct form, distinguished polarity difference-making from whole-form sufficiency, and showed that persistent effects of transient bioelectric perturbation do not uniquely establish a stored anatomical target (Reddien 2018; Durant et al. 2019; Levin, Pietak, and Bischof 2019). The established molecular, cellular and physiological findings remained unchanged; their explanatory relations and warrant were assessed more narrowly. The result demonstrates diagnostic capacity because at least one defensible judgment of sufficiency and evidential strength changed.

The Pristionchus pacificus case tested transfer to prospective, cue-sensitive and irreversible development. Here the audit distinguished EUD-1 as a causally decisive node from the complete switch architecture, treated the developmental threshold as relationally realised across interacting factors, restricted the evidential implications of chromatin regulation, and separated evolutionary origin from present developmental operation (Ragsdale et al. 2013; Serobyan et al. 2016; Kieninger et al. 2016; Bui, Ivers, and Ragsdale 2018). Because these revisions did not depend on restoration, target memory or prior anatomical form, the result supports procedural transfer across distinct explanatory architectures.

Mammalian glucose homeostasis returned no material explanatory gain. The strongest integrated account already distinguishes glucose sensing, islet organisation, distributed organ action, feedback, temporal counter-regulation, pathological breakdown and model-dependent quantification (Bergman et al. 1979; Röder et al. 2016; Rorsman and Ashcroft 2018; Petersen and Shulman 2018). Applying the audit reorganised these findings but changed no defensible judgment of relation, evidence, scope, sufficiency or failure. The null is therefore affirmative evidence of calibration: the procedure did not manufacture improvement merely because the system was complex and highly organised.

The strict novelty test remains restrictive. Dependence-based explanation, intervention, counterfactual assessment, invariance, mechanism, constitution, explanatory pluralism, constraint closure and integrated control are inherited resources rather than contributions of this article (Woodward 2003; Mitchell 2002; Taylor 2023; Bich and Bechtel 2021). Relation typing, material specification, evidential matching and scope qualification also have strong precedents within their respective explanatory traditions. The three cases do not reverse those findings.

The strict redundancy test nevertheless leaves a procedural residual. The complete audit requires the same record to specify the explanandum and foil, type the claimed relation, identify its material basis, match evidence and inferential limits, register alternatives and failure conditions, reconstruct the strongest comparator, and issue an explicit gain or null verdict. Its contribution is not the aggregation of familiar questions alone. It lies in applying them under a common comparative discipline and demonstrating assessment change, transfer and refusal across deliberately differentiated cases.

The paper may therefore defend three claims. First, the audit can detect relation conflation and overextended sufficiency claims. Second, it can transfer across at least two distinct biological architectures. Third, it can recognise when an established account is already sufficient. These claims require qualification: the cases are diagnostic rather than representative, the observed gains are moderate and judgment-dependent, and further applications may overturn particular classifications or verdicts.

The paper must not claim a new theory of explanation, universal dependency taxonomy, biological mechanism, organisational ontology or general superiority over mechanistic, interventionist, organisational or control-theoretic accounts. Nor can three cases establish universal applicability. The surviving contribution is narrower: a repeatable comparative instrument whose value is case-sensitive and whose credibility depends as much on its calibrated null as on its positive results.

8. Objections, Limits and Failure

The strongest objection is that the audit merely bundles familiar tools. Dependence, intervention, counterfactual comparison, mechanism, constitution, pluralism and organisational explanation already provide its component operations (Woodward 2003; Mitchell 2002; Taylor 2023; Bich and Bechtel 2021). This objection is substantially correct. The contribution survives only if the common workflow changes defensible assessments and distinguishes gain from no gain. If the three verdicts could be reached as clearly without it, the audit is redundant.

A second objection concerns its relation categories. Causal, constitutive, enabling, constraint, structural, spatial, temporal and historical dependencies do not form an exhaustive taxonomy and may mix ontic relations with explanatory roles. The categories are therefore provisional and claim-specific. Their function is diagnostic: to prevent evidence supporting one relation from being extended without argument to another. The method fails if classification is arbitrary, makes no evidential difference or suppresses legitimate alternative analyses.

APS provenance raises a further concern. The procedure could covertly privilege organisational explanation or require acceptance of an APS account of life. It does not. Organisational accounts are among its strongest precedents and comparators, not its premises (Montévil and Mossio 2015; Bich and Bechtel 2022). If the audit cannot be applied by researchers who reject APS, its asserted neutrality fails.

The cases may also appear cherry-picked. They were selected diagnostically: one contested positive case, one structurally different transfer case and one mature null. This supports proof of procedural capacity, not prevalence or generality. Three cases cannot establish representative utility across biology. Broader claims require independent applications selected without expectation of a positive result.

The biological analyses face more specific objections. Planarian target-memory claims are heterogeneous and often qualified. The positive result survives only against their strongest moderate form; it fails if it depends on an exaggerated target or overlooks distinctions already explicit in the comparator (Reddien 2018; Levin, Pietak, and Bischof 2019). In Pristionchus, reaction norms and switch genetics remain the primary explanations. Transfer gain exists only if separating switch node, relational threshold, developmental execution and history clarifies their warrant rather than renaming established findings (Ragsdale et al. 2013; Bui, Ivers, and Ragsdale 2018).

The glucose null might seem predetermined or ceremonial. It is informative only because the same fields and strongest-comparator rule were applied and no assessment changed. If the comparison relied on a textbook feedback loop, the null would be trivial; if a material change were suppressed to preserve calibration, it would be invalid.

These points separate present limitations from method failure. The demonstration is limited by three non-representative cases, literature-based reconstruction, expert-dependent judgments and absence of prospective experimental testing. The audit itself is defeated if relation typing does not affect admissible evidence; different standards are used across cases; weak comparators manufacture gain; redescription is counted as improvement; null verdicts are structurally unavailable; or independent analysts cannot reproduce the decisive assessments.

The lean search remains closed because further acquisition was unlikely to change the advancement decision. It should reopen only for a named precedent, evidential omission, authentication failure, reviewer objection or journal requirement capable of altering a frozen verdict. Case reassessment is warranted if stronger evidence changes relation, scope or sufficiency. Claims must then be narrowed or withdrawn. If the residual procedure becomes redundant, the paper should be redirected from journal submission to APS Research as a methodological synthesis rather than defended through expanded governance or stronger rhetoric.

9. Conclusion

A common, relation-typed and materially explicit audit can improve the comparative assessment of biological explanations, but it does not do so universally. Its value depends on whether clarification changes a defensible judgment relative to the strongest established account.

The three applications differentiate the conditions under which such change occurs. In planarian regeneration, the audit separated cellular supply from restoration of correct form, bounded the whole-form implications of polarity mechanisms and distinguished persistent anatomical effects from unique evidence of a stored target. The result demonstrates diagnostic capacity. In Pristionchus pacificus polyphenism, the same procedure separated a decisive switch node from the complete developmental architecture, treated the threshold relationally and distinguished present operation from evolutionary origin. Because these conclusions required no assumptions about regeneration or restoration, they demonstrate transfer. In mammalian glucose homeostasis, the strongest integrated account already typed, materialised, evidenced and qualified the relevant dependencies. The resulting no-gain verdict demonstrates discrimination: comparator sufficiency was recognised rather than recast as an omission.

None of the audit’s component operations is claimed as new. Dependence-based explanation, counterfactual testing, intervention, invariance, mechanism, constitution, explanatory pluralism, constraint analysis and organisational explanation already supply them in substantial form. The residual contribution lies only in their common comparative discipline: fixing the explanandum and foil, typing the claimed relation, specifying its material basis, matching evidence and inferential limits, recording alternatives and failure conditions, reconstructing the strongest comparator, and requiring an explicit gain-or-null verdict.

The demonstration establishes capacity, not universality. Its cases are diagnostic rather than representative; its classifications are provisional; and its verdicts remain open to stronger evidence, alternative relation assignments and independent reassessment. The method fails if it manufactures gain through weak comparators, treats redescription as improvement, applies unequal standards or makes null outcomes unavailable. Further evidential support would require independent applications capable of reproducing, changing, or defeating its judgments.

The practical implication is straightforward. An arrow in an explanatory diagram should not be treated as evidence merely because the surrounding account is coherent. It represents a dependency claim that must be identified, typed, materially specified and tested against relevant evidence and alternatives. Some arrows will be strongly supported; others will require qualification, and some may prove to add no explanation at all. A framework earns scientific credibility not by protecting every connection it proposes, but by making those connections open to assessment, revision and rejection.

The journal-level contribution is therefore limited but assessable: a repeatable comparative instrument for determining when heterogeneous dependency claims alter explanatory sufficiency and when the strongest biological explanation should instead be judged adequate. Its credibility rests jointly on bounded positive gain, transfer across a distinct architecture and principled refusal of gain.