{"id":"692b17de-5cc0-4cf2-a800-0e88e1b198ba","arxiv_id":"2411.17012","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"Cycles in process-enablement graphs define organisational closure, and closure-preserving graph maps let different theories of self-organisation be compared precisely.","lead":"This paper introduces process-enablement graphs, directed graphs whose vertices are biological processes and whose edges are direct enablement relations, and represents self-organisation as cycles in these graphs. The formalism then uses graph maps that preserve cycles to compare classical theories of life, including autopoiesis, (F,A)-systems, and autocatalytic sets.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Definition 2.5's counterfactual-necessity test erases redundant enablements, so the cycles said to capture self-organisation are not robust to common biological redundancy.","rationale":"The reader's weakest assumption concerns the reliable identification of direct-enablement arrows; I agree that this is the fragile point. My stress-test sharpens it: the problem is not merely that arrows are hard to observe, but that Definition 2.5's counterfactual necessity condition is logically too strong for redundant or disjunctive causation, which is common in biological systems. The paper's own RAF example acknowledges one degradation-dependent case but does not address redundancy generally. The mathematics in Sections 3 and A is internally consistent, and the conditional verdict is appropriate: the framework is promising but its empirical and modelling payload depends on a causal semantics that is neither formalised nor defended. Because my concern is a more specific version of the reader's, the verdict should remain CONDITIONAL rather than move to REJECT: the issue is addressable by refining Definition 2.5 or by restricting the framework's scope, and the illustrative comparisons could be rechecked under the refined semantics.","tokens_in":20033,"tokens_out":9472,"duration_ms":100767,"concrete_test":"Construct a minimal ODE or Boolean model with processes A and B that redundantly supply a resource required by C, with C feeding back to sustain A and B, so that A→C→A is the intended self-organising cycle. Delete A while B remains: C continues, so Definition 2.5 denies A→C; symmetrically, deleting B denies B→C; hence no cycle is recorded. Re-run the same model using 'p is a member of some minimal sufficient set for q' as the enablement condition; the A→C→A cycle reappears. If the two pe-graphs differ, the framework's cycle identification depends on an unstated choice of causal semantics and is not a faithful general account of self-organisation.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Definition 2.5 defines enablement via the counterfactual 'if p had not happened, then q would not be happening', and Definition 2.6 adds a spatiotemporal interaction condition. This makes every arrow a strict necessary-condition claim. In any system with substitutable causes — two isozymes catalysing the same reaction, duplicate transporters, alternative nutrient routes — each individual cause fails this test even though it is a genuine enabling condition and even though the system contains a real feedback cycle. The RAF discussion in Section 4.3 notices one such case (B3 stockpiling) and sets it aside by assumption, but the issue is general and untreated. The graph-theoretic core (Theorems 3.3, 3.4) is sound conditional on a fixed graph, but it cannot validate the arrows. The central application in Section 4.2 is assembled from hand-drawn arrows and an explicitly 'somewhat arbitrary' partition P; if even one arrow should be absent under a more careful causal analysis, the cycle structure of IP, FA and A3 changes, and the homorheisms φ5, φ6 in Theorems A.9–A.10 may fail. The framework's central promise — that cycles locate self-organising components and that homorheisms compare them faithfully — therefore rests on an unformalised and, in the presence of redundancy, arguably incorrect causal semantics for enablement.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"Brown and Vittadello introduce process-enablement graphs (pe-graphs), directed graphs in which vertices are contemporaneous processes and edges are direct enablements defined via a counterfactual-necessity condition plus a spatiotemporal-interaction condition. They define organisational closure as the condition that every vertex has both an incoming and an outgoing edge, prove that every closed pe-graph contains a cycle (Theorem 3.3) and that strictly closed pe-graphs are exactly cycles (Theorem 3.4), and develop weak graph homomorphisms that preserve and reflect cycles (homorheisms) to compare different models of the same system. The framework is applied to autopoiesis, (F,A)-systems, and RAF autocatalytic sets; in particular, an intermediate pe-graph IP is constructed to show that the (F,A)-system and autopoietic descriptions of the cell are linked by homorheisms. The paper stresses that the formalism is static and does not demarcate life from non-life.","tokens_in":20318,"tokens_out":8194,"duration_ms":75901,"significance":"The mathematical core is elementary, self-contained, and correct: Theorems 3.3, 3.4, and A.1–A.10 are proved in the text, and the homorheism concept gives a precise language for saying when two process-level descriptions capture the same organisational closure. The paper is also unusually explicit about its limitations, including the unproved status of the direct-enablement hypothesis (§2), the 'somewhat arbitrary' choice of the intermediate partition P (§4.2), and the high-degradation assumption in the RAF example (§4.3, footnote 1). If the enablement semantics can be made robust, the framework would be a useful tool for comparing theories of life. As it stands, however, the central claim that cycles of direct enablements capture self-organising components of real biological systems depends on a strict counterfactual semantics that erases redundant causes, a common situation in biology. The significance is therefore conditional on resolving that semantic issue.","major_comments":[{"comment":"The counterfactual-necessity test in Definition 2.5 is not robust to redundant enablements. If two processes p1 and p2 are substitutable causes of q (e.g., two isozymes catalysing the same reaction), then removing p1 alone does not stop q, so p1 does not enable q under the definition, even though the system may contain a real feedback cycle that includes p1. The RAF discussion in §4.3 notices one such case (B3 stockpiling) and sets it aside by assuming a high-degradation environment, but the issue is general and affects the central claim that cycles of direct enablements capture self-organising components. Please justify the strict-necessity semantics or modify Definition 2.5 to handle redundant/disjunctive causation, and revisit the applications with that modification.","section":"§2 (Definitions 2.5–2.6; see also §4.3, footnote 1)"},{"comment":"The homorheism comparison between FA and A3 is constructed, not discovered: the intermediate process set P is chosen 'somewhat arbitrarily' and the arrows in IP are qualitative direct-enablement judgements. Theorems A.9 and A.10 prove that φ5 and φ6 are homorheisms only for this particular IP. If any arrow in IP or in the target graphs were drawn differently—for instance, if a more careful causal analysis judged a given interaction not to be a direct enablement—the cycle structure could change and the homorheisms could fail. The conclusion that 'FA and A3 model the same self-organising processes' is therefore conditional on the hand-chosen P and on the arrow judgements. The authors should either supply a principled method for choosing P and validating the arrows, or explicitly present the FA/A3 correspondence as a proof-of-concept illustration rather than a derived equivalence.","section":"§4.2 (choice of P; Theorems A.9–A.10)"}],"minor_comments":[{"comment":"The status of loops is under-specified: Section 3.4 says a loop is shorthand for a finer-grained cycle, but Definition 3.1 allows loops as ordinary edges, and Theorem 3.4 calls loops cycles of length 1. Please state explicitly in Definition 3.1 or a following remark that edges of the form p→p are not direct self-enablements but shorthand for an unresolved cycle, so that the graph-theoretic and conceptual readings are consistent.","section":"§3.4 / Definition 3.1"},{"comment":"The vertex maps φ5 and φ6 are defined verbally in the text and illustrated by colourings in Figure A.1, but the main-text Figure 4 does not show the colouring. Add a sentence to the Figure 4 caption directing readers to Figure A.1, or include a small table of the maps.","section":"§4.2 / Figure 4"},{"comment":"Theorem A.8 gives a sufficient condition for reflection of closure, not a necessary one; the name 'Reflection test' is therefore slightly misleading. Consider renaming it 'A sufficient condition for reflection' or adding a comment that the test is sufficient only.","section":"§A.2.2 / Theorem A.8"},{"comment":"Consider rephrasing 'cycles within these graphs capture self-organising components' to 'cycles within these graphs represent self-organisation as defined here', since Theorems 3.3 and 3.4 are consequences of Definition 3.2 rather than independent empirical findings.","section":"Abstract / §5"},{"comment":"The role of the high-degradation assumption is clear in the footnote, but the main text says 'whether reactions directly enable each other depends on the nature of molecule degradation and the rate constants' — this makes the pe-graph structure depend on empirical rate parameters. It would be helpful to state explicitly that the comparison of R1 and R2 is therefore conditional on those parameters.","section":"§4.3"}],"recommendation":"major_revision","confidential_remarks":"The paper has a solid, self-contained graph-theoretic core and is transparent about its limitations. The main obstacle is the enablement semantics: the strict counterfactual test in Definition 2.5 erases redundant causes, which are common in biological systems, and this directly affects the central claim that cycles locate self-organising components. The FA/A3 comparison in §4.2 is also explicitly dependent on a hand-chosen partition and qualitative arrows. I do not see these as unfixable; a revision that addresses the redundancy issue and reframes the applications as conditional illustrations could make the paper acceptable. The novelty of homorheisms is modest but useful for the intended audience."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Worth a look if you work on formalizing organizational closure. The graph theory is elementary and correct, and the authors are unusually transparent about what the framework does and does not do. The genuinely new piece is the homorheism—a weak graph homomorphism that preserves and reflects cycles—and the explicit three-way comparison of autopoiesis, (F,A)-systems, and RAF sets through an intermediate graph. Theorems 3.3 and 3.4 are correct, and the appendix proofs check out. The authors also state plainly that the framework is static, does not demarcate life from non-life, and does not justify the direct-enablement hypothesis. That honesty is a real strength.\n\nThe soft spots are real but mostly in the application layer rather than the formal core. The stress-test note about redundant enablements has teeth. Definition 2.5 requires counterfactual necessity: if p had not happened, q would not be happening. In any system with substitutable causes—two isozymes, duplicate transporters, alternative nutrient routes—each individual cause fails that test even when it genuinely enables the downstream process. The paper notices this in the RAF discussion (the B3 stockpiling case) and sets it aside by assuming a high-degradation environment, but the issue is general and untreated. That means the cycles the framework identifies can miss real organizational closure in robust biological systems. The applications also depend on qualitative enablement judgements and, in Section 4.2, an explicitly 'somewhat arbitrary' partition P. If even one arrow is absent under a more careful causal analysis, the homorheisms in Theorems A.9–A.10 can fail. The authors acknowledge this, but it does limit the examples to illustrations rather than demonstrations.\n\nI also agree with the reader that the closure-as-cycles result is definitional: Definition 3.2 defines closure so that cycles are the unit, so Theorem 3.4 is a consequence of the setup, not an empirical discovery. That is not a flaw in a formal-language paper, but it means the framework's value is in providing a precise notation and comparison tool, not in discovering new biological facts.\n\nWho is this for? People working on theories of life, origins of life, and organizational closure who want a precise way to compare different models. It deserves a serious referee: the formal core is sound, the limitations are stated, and the redundancy problem is addressable by refining the enablement definition or explicitly restricting its scope. I would recommend conditional accept with major revision, asking for a treatment of redundant enablements and a clearer statement that the examples are illustrative.","headline":"A sound, honest formalism paper whose graph-theoretic core is correct; the real caveat is that applications depend on an enablement semantics that breaks under redundancy and on hand-drawn arrows.","tokens_in":20835,"tokens_out":1725,"would_cite":false,"duration_ms":17719,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Graph cycles give a common language for self-organisation across theories of life.","keywords":["self-organisation","organisational closure","process-enablement graphs","autopoiesis","(M,R)-systems","autocatalytic sets","graph homomorphisms","constraints"],"falsifier":"Choose a well-studied self-sustaining network, such as a prokaryotic cell, and test every candidate arrow by removing the upstream process and by checking physical interaction. If the system remains self-sustaining while some process that is necessary for the system's persistence has no incoming arrow from another process in the network, the claim that organisational closure shows up as cycles would fail for that case.","tokens_in":103,"feed_emoji":"🔄","tokens_out":8433,"duration_ms":134429,"temperature":0.7,"pith_summary":"This paper proposes that biological self-organisation can be studied uniformly by drawing directed graphs in which vertices are processes and an arrow $p \\to q$ means that process $p$ provides a necessary condition, with direct physical contact, for process $q$ to occur. The central claim is that the self-organising parts of any system are exactly the cycles of such a process-enablement graph: every closed system contains a cycle, and a strictly closed system is nothing but a cycle. If this is right, locating organisational closure in any model reduces to finding cycles, and comparing two biological theories reduces to comparing their cycle structures with maps that preserve and reflect cycles. The paper applies the formalism to autopoiesis, $(F,A)$-systems, and autocatalytic sets, showing that two seemingly different models of the cell describe the same self-organising processes under different partitions, and that whether an autocatalytic set is closed can depend on perspective.","feed_headline":"Cycles in enablement graphs reveal life’s self-organising core","feed_subtitle":"A new graph formalism lets biologists compare any theory of life by matching its cycles.","key_machinery":"The central object is the process-enablement graph, or pe-graph: a connected directed graph whose vertices are contemporaneous processes in a system and whose edges are direct enablements, defined by a counterfactual necessity check plus a spatiotemporal interaction check. The load-bearing identity is the theorem that a pe-graph is strictly closed if and only if it is a cycle, and the supporting notion of a homorheism, a weak graph homomorphism that both preserves cycles and reflects cycles, lets fine-grained and coarse-grained perspectives be matched cycle-for-cycle. Loops are used only as shorthand for an underlying cycle, so a loop at $S$ means a finer network of processes inside $S$ contains at least one cycle.","core_discovery":"On the paper's own terms, the discovery is that organisational closure has a precise graph-theoretic signature: a pe-graph is closed if every process has at least one incoming and one outgoing direct enablement, and strictly closed if and only if it is a cycle (Theorem 3.4). Consequently the fundamental unit of self-organisation is the directed cycle of mutually enabling processes, not the individual process or constraint. The authors build homomorphisms of pe-graphs that preserve closure, and homorheisms that also reflect closure, so that a fine-grained model and a coarse-grained model can be certified to contain the same self-organising features even when their process boundaries differ. In the worked comparison, an intermediate graph $IP$ is constructed with homorheisms to both the fabrication-assembly model and the autopoietic model of the cell, showing that their differently arranged cycles are re-articulations of one underlying set of enablements.","pith_inferences":["We infer that the framework yields an operational recipe: apply standard directed-graph cycle algorithms to verified direct-enablement networks and treat the resulting cycles as candidate self-organising modules for experimental perturbation.","We infer a testable extension: because the comparison in Section 4.2 rests on a partition the authors call somewhat arbitrary, one could search algorithmically for the coarsest partition that still yields homorheisms, turning a manual construction into a reproducible optimisation.","We infer that if direct enablements can be tracked over time, the same cycle-based language could record when organisational closure first appears in an evolving chemical system, connecting self-organisation to origin-of-life scenarios more directly than the static graphs in this paper."],"forward_implications":["If the central claim is correct, finding organisational closure in any model becomes a cycle-detection problem in a finite directed graph.","Two biological models that admit a homorheism are guaranteed to agree on which self-organising cycles are present, even when they partition the underlying processes differently.","The autopoiesis and fabrication-assembly accounts of the cell can be reconciled through an intermediate pe-graph, with every cycle in either model reflected in the other.","Whether an autocatalytic set is self-organising is partly perspective-dependent: one coarse-graining of the same reactions can erase a cycle while another reveals it."],"supporting_citations":[{"why":"Provides the organisational-closure definition and the constraint-closure framing that the paper re-expresses as closure of processes.","marker":"(Montévil & Mossio, 2015)"},{"why":"Supplies the (F,A)-system model of the cell that is compared with autopoiesis.","marker":"(Hofmeyr, 2021)"},{"why":"Supplies the autopoiesis criteria that the paper turns into the A1, A2, and A3 pe-graphs.","marker":"(Maturana & Varela, 1980)"},{"why":"Supplies the RAF-set formalism and the worked example re-drawn as pe-graphs R1 and R2.","marker":"(Hordijk, Steel, & Kauffman, 2012)"},{"why":"Supplies the weak graph homomorphism theory used to contract edges while preserving cycle structure.","marker":"(Hell & Nesetril, 2004)"},{"why":"Supplies the model-comparison approach via simplicial complexes that the paper translates into directed graphs.","marker":"Vittadello and Stumpf (2021, 2022)"},{"why":"Develops the enablement concept that the direct-enablement arrow formalises.","marker":"(Longo & Montévil, 2013)"}],"fun_headline_variants":["Cycles in enablement graphs capture self-organisation","Life's self-organising core is a graph cycle","One graph to compare every theory of life","Enablement graphs: cycles define life's organisation"],"cache_read_input_tokens":22912,"weakest_assumption_plain":"That a biologist can reliably decide, process by process, whether an arrow is a genuine direct enablement in the counterfactual sense and whether the two processes physically interact over a time interval; the paper states that focusing on direct enablements is a hypothesis it does not further justify.","fun_headline_variants_meta":{"raw":{"variants":["Cycles in enablement graphs capture self-organisation","Life's self-organising core is a graph cycle","One graph to compare every theory of life","Enablement graphs: cycles define life's organisation"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000638,"raw_usage":{"total_tokens":2932,"prompt_tokens":929,"completion_tokens":2003,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":545,"completion_tokens_details":{"reasoning_tokens":1943}},"tokens_in":545,"tokens_out":2003,"duration_ms":15697,"temperature":1.0,"reasoning_tokens":1943,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T12:38:10.075928+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Choose a well-studied self-sustaining network, such as a prokaryotic cell, and test every candidate arrow by removing the upstream process and by checking physical interaction. If the system remains self-sustaining while some process that is necessary for the system's persistence has no incoming arrow from another process in the network, the claim that organisational closure shows up as cycles would fail for that case.","supporting_citations":[{"cited_title":"APACrefauthors \\ 2021","cited_arxiv_id":null,"evidence_quote":"Supplies the (F,A)-system model of the cell that is compared with autopoiesis."},{"cited_title":", Steel, M","cited_arxiv_id":null,"evidence_quote":"Supplies the RAF-set formalism and the worked example re-drawn as pe-graphs R1 and R2."},{"cited_title":"\\ Stumpf, M P","cited_arxiv_id":null,"evidence_quote":"Supplies the model-comparison approach via simplicial complexes that the paper translates into directed graphs."},{"cited_title":"\\ Mont \\'e vil, M","cited_arxiv_id":null,"evidence_quote":"Develops the enablement concept that the direct-enablement arrow formalises."}],"review_version":1}