Contribution
Before this paper the function debate faced a dilemma. Etiological (selected-effect) theories explained why a trait exists and grounded malfunction, but said nothing about what the trait does now. Dispositional theories said what it does now but could not say which contributions are proper functions. The paper offers one definition that does both, by grounding teleology and normativity in the current organization of a self-maintaining system, and it adds a criterion, differentiation, that keeps functions out of flames and hurricanes.
Key insight
Circular self-maintenance makes a process’s existence depend on its own effects, which gives teleology and norms without history. But every dissipative structure is circular in that sense. Functions start only when the system produces several distinct structures that each contribute in a different way, so that one can ask what each part is for.
Assumptions
- Self-maintenance as far-from-equilibrium closure. A macroscopic pattern exerts “at least one ‘top-down’ constraint that contributes to the maintenance of the FFE flow of energy and matter enabling the persistence of the microscopic dynamics” (p. 8, citing Ruiz-Mirazo 2001).
- Context dependence. Self-maintenance holds “with respect to a given environment and context” (endnote v), so the functional/indispensable and primary/secondary distinctions are context-relative too.
- Classes and regimes. Self-maintaining systems fall into classes, each admitting several regimes of self-maintenance; functions are relative to a regime (§5.1).
- Biological traits, not artefacts. Artefacts are deferred; endnote xi suggests they are produced by, but not maintained by, an organization.
Key results
- Closure grounds teleology and normativity (§4.1). “To the question ‘Why does X exist in that class of systems?’ it is legitimate to answer ‘Because it does Y’”. The system’s conditions of existence “are the norms of its own activity”.
- Closure is not sufficient (§4.1–4.2). Bénard cells and a candle flame are worked examples of closure (pp. 8–9). Their components have no functions because they make “a unique contribution” through one pattern. Material complexity (many molecular species, autocatalytic sets) does not change this; endnote viii shows how a single-catalyst autocatalytic set becomes functional once it produces a membrane.
- Definition (§5). C1 contribution to O; C2 produced and maintained under constraints of O; C3 S differentiated. The heart pumps blood: C1 through circulation, C2 because the organism makes and keeps the heart, C3 because the organism makes many other contributing structures.
- Indispensable versus merely functional (§5.1). The heart is indispensable in every human regime; the eye’s light transduction is functional but not indispensable (blind people survive by shifting regime).
- Primary and secondary functions (§5.2). The nose’s primary function is to warm and humidify air; supporting eyeglasses is secondary, subject to closure only in a more complex (socio-technical) regime.
- Useful, dysfunctional, irrelevant, accidental (§6.1–6.2). Useful: C1 and C3, not C2. Dysfunctional: C2 and C3, not C1. Irrelevant effects: C2 and C3, no C1, not interfering with function. Frequency is irrelevant.
- Against selected-effect theories (§6.3). Primary functions usually coincide with proper functions, because basic regimes are evolutionarily older, but diverge for traits selected and now useless (SE: function; OA: had one) and for current contributions with no selection history (OA: function; SE: none).
- Reproduction (§6.4) is conceded not to be functional for the individual. It may be functional in a second-order self-maintaining system (lineage or ecosystem), which the authors say “still lacks adequate justification”.
Claims
| id | claim | strength | support |
|---|---|---|---|
| C1 | Organizational closure naturalizes teleology and normativity in current organization | moderate | conceptual argument §4.1 |
| C2 | Dissipative structures (flames, Bénard cells, hurricanes) have closure but no functions | moderate | worked examples pp. 8–10; endnotes iii, vii |
| C3 | Organizational differentiation, not material complexity, is what functions require | moderate | §4.2, endnote viii thought experiment |
| C4 | The OA integrates etiological and dispositional accounts | moderate | consistency with Wright’s schema (§5); §6.3 |
| C5 | The OA distinguishes function from usefulness, dysfunction and accidental contribution | moderate | definitions §6.1–6.2 |
| C6 | Reproductive traits are functional in second-order self-maintaining systems | weak | sketch §6.4, conceded as unjustified |
Concepts
- Organizational closure. “A circular causal relation between some macroscopic (or higher-level) pattern or structure and the microscopic (or lower-level) dynamics and reactions” (p. 9). Note: this is broader than the closure of constraints of LIT-582.
- Organizational differentiation. The system “produces different and localizable patterns or structures, each making a specific contribution to the conditions of existence of the whole organization”.
- Regime of self-maintenance. A specific organization a member of a class can adopt without ceasing to exist or leaving the class.
- Primary function. The contribution subject to closure in the most basic regime of self-maintenance that produces T.
Connections
The paper’s targets are the systemic approach (Cummins, Craver, Davies), the goal-contribution approach (Boorse, Adams), the propensity view (Bigelow & Pargetter) and selected-effect theories (Wright, Millikan, Neander, Godfrey-Smith). Its allies are Christensen & Bickhard (2002) and McLaughlin (2001), whom it claims to improve on by giving explicit conditions. Its physics is Glansdorff & Prigogine and Nicolis & Prigogine, represented in the record by LIT-516. Montévil & Mossio (LIT-582) replace this paper’s broad closure with closure of constraints and keep the function thesis. Barandiaran et al. (LIT-566) cite it for organism-generated normativity.
Bearing on the record
- The demarcation question (LIT-192, NOTE-094; LIT-526; LIT-516). NOTE-094 records that closure accounts face the candle-flame problem. This paper does not deny the flame closure; it denies it functions. So the line between flame and cell, in the 2009 account, is differentiation. In the 2015 account (LIT-582) closure itself is narrower: each constraint must be both produced by and productive of another constraint in the set, with direct dependence, and a cycle of processes such as the water cycle is said not to qualify. Whether a flame’s single pattern can meet that is not discussed in either paper; my reading is that a single macroscopic constraint cannot form a closed set by itself, so the 2015 closure likely excludes the flame where the 2009 closure did not. That is an inference, not a result of either paper.
- Baedke (LIT-167) argues closure is too permissive to individuate organisms. This paper never uses closure to individuate; it uses it to ground functions. The objection bears on LIT-582 §6, not here.
- THEORY filed as THEORY-072 (Proposed): what separates organisms from dissipative structures is closure together with organizational differentiation, not closure alone.
- Biological Autonomy (LIT-536) chapter 3 should contain the book’s version; a reader of it could check whether differentiation survives.
- No instruction for machine-learning practice.
Limitations
- Artefacts and reproduction are deferred, and reproduction is the obvious counterexample to a self-maintenance account of function.
- “Differentiation” is not given a measure; “different and localizable” is left informal.
- The relation between primary functions and selected effects is said to need “an independent paper”.
Open questions
- Is there a quantitative criterion of differentiation that a model could be checked against, so that a flame, a protocell and an autocatalytic set are sorted without intuition?
- Does the 2015 closure of constraints make differentiation redundant as a demarcation criterion, or are both still needed?