Contribution
It argues that intrinsic teleology can be naturalised through organisation, not through selection, and says which kind of circularity does the work. Earlier organisational accounts (Kant, Bernard, Piaget, Maturana and Varela) tied teleology to self-determination but left closure too loose to say why the water cycle is not teleological. Building on Rosen and on Kauffman’s work–constraint cycle, it makes closure a closure of constraints, acting on a thermodynamic flow at another level of causation. That lets it separate closure from cycles of transformation, and place dissipative structures, at most, on a minimal rung that is teleological but not organised.
Key insight
Circularity is “necessary but not sufficient for intrinsic teleology” (MS p. 2). What counts is whether the system’s activity “contributes to specifying the conditions under which the circular relation as such can occur” (MS p. 16). A river only obeys or modulates its bed. An organism makes the constraints that make it. So teleology sits at the level of constraints, and closure is “by construction an inter-level causal regime” (MS p. 15).
Assumptions
- Telos is identified with the conditions of existence. Biological systems “are what they do” (MS p. 2). The goal of an organism and its own existence “are one and the same thing”, which is what makes the teleology intrinsic, against the extrinsic teleology of artefacts (MS p. 2, after Jonas).
- Constraints are defined by conservation at a time scale. Constraints are “local and contingent causes, which reduce the degrees of freedom of the dynamics on which they act (Pattee, 1973), while remaining conserved (at the time scale which is relevant to describe their causal action) with respect to those dynamics” (MS p. 14). Catalysts are constraints, substrates are material causes (fn 17).
- A relational, not dynamical, description. Following Rosen, closure is described relationally, “abstracting closure from time”, so final causation does not invert the arrow of time (MS p. 13, and again p. 20).
- Rosen’s efficient causes are read as constraints. “Organisational closure should be understood as a closure of constraints” (MS p. 14, citing Moreno & Mossio, in press). The formal definitions are not given here; the paper refers to Mossio et al. 2013 (fn 17).
Key results
The paper is philosophical argument; it has no formal results.
- Closure versus cycles (MS p. 15). “Cycles refer to circular chains of transformations that, under the effects of externally determined constraints, can be described at one level of causation”. Closure is “a circular causal regime in which a set of constraints described at one level of causation maintains itself by canalising processes and reactions … at another level of causation”.
- The water cycle is not teleological (MS pp. 16–17). Its chain of transformations “does not generate its own constraints”, which “are external to the circular dynamics, and independent from them”: the slope of the ground. The river digs its bed, but “the water cycle would exist even though the river did not modulate its own bed”. Cycles “simply realise” their causal regime, so “cycles do not self-determine. Therefore, they are not teleological regimes.”
- Closure grounds teleology, normativity and function (MS p. 15). “The set of constraints subject to closure are the set of biological functions.”
- Non-biological closure (MS p. 18). Closure “requires a relatively high degree of complexity”, and “there are no clear examples of non-biological organisationally closed chemical networks, apart from abstract theoretical and computational models as, for instance, the well-known autocatalytic sets”. A material realisation would be “‘at the edge’ of the biological domain”. Footnote 21 adds that proto-biological closure cannot be excluded a priori.
- Dissipative structures (MS pp. 18–19). Whether they self-constrain “has not been settled”: Prigogine and Stengers, Bickhard and Ruiz-Mirazo say yes; Anderson and Stein, Crutchfield, and Minati and Pessa say no. The authors have elsewhere taken the first view (“as we did elsewhere”, MS p. 19). If they do self-constrain, the account must “concede that dissipative structures realise intrinsic teleology” — but of a “radically different kind”. They “do not realise closure; rather, they would be minimally teleological, insofar as they would generate a single macroscopic constraint (the structure itself) that, supposedly, maintains itself by acting on its own boundary conditions”. They “are not organised” and “their parts have no functions” (MS p. 19).
Claims
| id | claim | strength | support |
|---|---|---|---|
| C1 | Intrinsic teleology is grounded in self-determination understood as self-constraint, not in circular causation as such. | moderate | Conceptual argument, §§2–3 (MS pp. 10–17) |
| C2 | A cycle of transformations whose constraints are external (the water cycle) does not self-determine and is not teleological, even if it modulates its constraints. | moderate | The river-bed argument, MS pp. 16–17 |
| C3 | Closure is an inter-level regime: constraints at one level canalise far-from-equilibrium processes at another. | moderate | Definition-level claim, MS p. 15, fn 19 |
| C4 | Closure among constraints realises the division of labour: “each component makes a specific and complementary contribution”, so closure itself captures organisation. | weak | Asserted with citation to Moreno & Mossio, in press, MS p. 11; not argued here |
| C5 | Dissipative structures do not realise closure; if they self-constrain at all, they generate a single macroscopic constraint and are minimally teleological, unorganised, and without functions. | weak | Stated, not demonstrated; the self-constraint half is explicitly left open (MS pp. 18–19) |
| C6 | No non-biological material system is known to realise closure; only abstract models such as autocatalytic sets do. | weak | ”To our knowledge”, MS p. 18, with two citations |
Concepts
- Self-determination — the effects of a system’s activity “contribute to establish and maintain its own conditions of existence” (MS p. 2).
- Self-constraint — self-determination in which the system produces (some of) the constraints acting on its own activity (MS p. 19).
- Constraint — a local, contingent cause that reduces the degrees of freedom of a dynamics while staying conserved, at the relevant time scale, with respect to it (MS p. 14).
- Closure — “an organisation of mutually dependent constraints” (MS p. 16); “a network of mutually dependent constitutive constraints” (abstract).
- Cycle — a circular chain of transformations, under external constraints, describable at a single level of causation (MS p. 15).
- Minimal teleology — the status a dissipative structure would have if it self-constrains through its single macroscopic constraint (MS p. 19).
Connections
It is a step in Mossio’s programme. Mossio, Saborido and Moreno (2009, LIT-570) is cited for functions as constraints under closure (MS p. 3, p. 15), and Moreno and Mossio’s Biological Autonomy (LIT-536), then in press, for closure of constraints and for the division of labour (MS pp. 10–11, 14). It does not cite Montévil and Mossio (2015, LIT-582), which came later and gives closure formal definitions. Its treatment of cycles (the water cycle, MS pp. 11, 16–17) runs parallel to LIT-582’s footnote 25 on the hydrologic cycle, which NOTE-468 records; whether that footnote cites this paper was not checked. On teleology it sets itself against selected-effect accounts (Ruse, Millikan, Neander), the genetic-program view (Jacob, Monod, Mayr), cybernetics (Wiener, Ashby) and Maturana and Varela’s rejection of teleology, and against Toepfer’s inclusion of the water cycle (MS pp. 3–10, 16).
Bearing on the record
-
THEORY-072 — partly; it supports the line but not the account’s wording of it. What the paper says about differentiation as against closure, exactly:
- It never names differentiation as a condition. The word does not occur in the text; the only occurrence is the title of Arnellos and Moreno (2012) in the references, cited for the complexity closure needs (MS p. 18). Distinct time scales among constraints are not mentioned either; “time scale” appears once, in the definition of a constraint (MS p. 14).
- It folds the division of labour into closure. Closure “conveys the idea that self-determination results from the integrated interactions among a set of entities which would not exist in isolation: each component makes a specific and complementary contribution”, and “by virtue of the interplay between division of labour and mutual dependence that it implies, closure captures a fundamental aspect of the idea of ‘organisation’” (MS pp. 10–11). Functions likewise need “different yet complementary contributions” and are “the set of constraints subject to closure” (MS p. 15). This matches the record’s reading of LIT-582 that THEORY-072 states (“that builds differentiation into closure”), and it is the authors’ own words, a year before LIT-582, not the record’s inference.
- It denies dissipative structures closure, not merely differentiation. LIT-570 (2009), as THEORY-072’s step 2 reads it, granted the flame “organizational closure” and drew the line at differentiation. This paper instead says dissipative structures “do not realise closure” and would at most “generate a single macroscopic constraint” (MS p. 19). So by 2014 the 2009 broad closure is gone from the authors’ vocabulary, and “one constraint versus a set of mutually dependent constraints” is where the line falls. That is the single-constraint half of THEORY-072’s title, stated by the authors; the “different time scales” half is not in this paper.
- It does not analyse a flame. It quotes Bickhard’s candle (fn 23) only as a view on self-maintenance, and it leaves open whether dissipative structures self-constrain at all (MS pp. 18–19). So the first test THEORY-072’s promote_when asks for (a dissipative structure analysed under LIT-582’s Definitions 1–4) is not met here.
- On whether chemical systems can realise closure, the question for which LIT-582 cites it: no clear non-biological example is known beyond abstract models such as autocatalytic sets, and a material one would be “at the edge” of biology (MS p. 18). That supports THEORY-072’s protocell clause as a place to look, but supplies no analysis of one, so the promote_when’s second test is not met either.
In sum, this reading settles the question promote_when puts about the authors’ view: they fold differentiation into closure. It does not do either of the two analyses the promote_when requires.
Limitations
- The closure of constraints is asserted, not defined, here. The paper defers the formal account to Mossio et al. 2013 and to the book then in press (fn 17, MS p. 14). It gives no criterion for counting constraints or for when two of them are “mutually dependent”.
- “A single macroscopic constraint” is not argued. That a dissipative structure has one constraint, and not several (a flame’s temperature field, the convection it drives), is stated, not shown (MS p. 19).
- The authors’ own view on dissipative structures is divided. They say they have argued elsewhere that dissipative structures self-constrain, and here present it as one side of an unsettled dispute (MS p. 19).
- Supra-organismal closure is left open. Symbioses and ecosystems “might also possibly be shown to be organisationally closed”, and the authors have “no principled objections” (MS p. 17).
Open questions
- Does a candle flame, analysed constraint by constraint, have one constraint of its own or several? The paper says one and does not show it.
- Does any non-biological chemical system realise closure in material form? The paper knows of none (MS p. 18).
- Is the division of labour a consequence of mutual dependence among constraints, or an extra condition? The paper treats it as implied by closure (MS p. 11) without proving that a closed set must divide labour.