Robert Rosen: Closed to Efficient Causation

The test this site already uses when it says “not sovereign,” and the person who stated it first

This page is overdue, and it is a debt page. Across the Laboratory, when a structure is refused the word sovereign, the reason given is usually some version of it cannot remake its own parts. That test has an author. Robert Rosen (1934–1998) characterized organisms by closure to efficient causation: a system whose catalysts are all produced by the system itself, with only raw nutrients taken from outside. This site has been applying it without naming him, and the autonomy tradition it credits as its nearest kin is built partly on his work. This page puts the name where the idea already was.

How this page is sourced. Rosen’s own book, Life Itself (1991), was not read for this page; it is not openly available. Everything attributed to Rosen here is taken from Cornish-Bowden and Cárdenas, who worked on his theory for years and whose account is openly published, and it is marked as coming through them. Two papers in the dispute over his computability claim are named from their citations and were not read. Nothing on this page is marked for later verification: what could not be read is said plainly instead.

What Rosen made precise

Rosen, R. (1991). Life Itself: A Comprehensive Inquiry into the Nature, Origin, and Fabrication of Life. Columbia University Press. Read here through Cornish-Bowden, A., & Cárdenas, M. L. (2022). The essence of life revisited: how theories can shed light on it. Theory in Biosciences, 141, 105–123.

Cornish-Bowden and Cárdenas state the core of it in one line: for Rosen, “organisms are closed to efficient causation”. They then say what that means in chemical terms. “He considered that all catalysts needed by an organism needed to be products of the organism itself”, and so “a living organism is an emergent system, because nothing apart from nutrients is given from outside”.

CEClosure to efficient causation. Every efficient cause the system depends on, which in a cell means every catalyst, is itself produced by the system. Material inputs may come from outside. The things that make the reactions go may not.

MR(M,R)-system. Rosen’s formal model of such a system: a metabolism that turns nutrients into products, and a second process that makes new catalysts for the metabolism out of those products, so the catalysts that wear out are replaced from within. Rosen called the second process repair. Cornish-Bowden and Cárdenas note that “For Rosen (1991) the R in ‘(M, R)-system’ signified ‘repair,’ but it does not correspond to any modern usage of repair in biology”, and that they “introduced the wording ‘replacement’ (or ‘resynthesis’) as less likely to be misleading”. This page follows them.

The point that matters for this framework is the direction of the dependency. Metabolism needs catalysts. Catalysts are made from metabolism’s products. So the organism is the thing that manufactures the machinery that manufactures it, and Rosen made that circle the definition rather than a feature.

Related to RAF, and different from it

The RAF page treats reflexive autocatalysis as recursion lock made checkable. Rosen’s closure is a related condition with a different demand, and the difference can be put in one sentence of each.

Hordijk and Steel’s reflexive condition reads: “Each reaction r ∈ R is catalyzed by at least one molecule type that is either a product of R or is present in the food set F”. A catalyst is allowed to arrive with the food.

Rosen’s closure does not allow that. Nutrients come from outside; catalysts do not. A RAF whose reactions are sped by something supplied in the food set satisfies RAF and fails Rosen.

RAF asks whether the reactions can run on what the set makes plus what it is fed. Rosen asks whether everything that drives the reactions is made inside. RAF explicitly permits catalysts from the food set; Rosen’s closure does not permit the system’s efficient causes simply to be supplied from outside. So a RAF may satisfy its own closure condition while failing Rosen’s.

That places the two where this framework needs them. RAF is the threshold at which a chemistry ignites and keeps itself going. Rosen’s closure asks whether it has stopped depending on anyone for its catalysts, which is the half of self-production this site’s sovereignty question keeps turning on.

Where the framework agrees, and owes a debt

The test in use. The circadian clock is classed recursive and not sovereign because it cannot remake its own parts: the clock proteins are made by a cell the clock does not build. The self-maintaining membrane names its open question as the catalyst, which is never produced and never decays, so that asked of the catalyst the reconstitution question does not close. Both are Rosen’s test, stated in this site’s vocabulary. The membrane card shows its shape exactly: the links are made inside, from substrate, while the only catalyst is given at the start and never made. Treat the catalyst as supplied and the links close. Ask Rosen’s question of the catalyst and they do not.

The lineage already credited. The autonomy tradition, which this site calls its nearest kin, formalizes organisms as a closure of constraints. That is not the same construct as Rosen’s, but it grew in conversation with it: Mossio, whose work with Montévil is the formal core cited on the autonomy page, is also first author of a 2009 paper giving a computable expression of Rosen’s closure. Crediting the descendants and leaving out the ancestor was the gap this page closes.

The exception that shows where the line really is

Cornish-Bowden and Cárdenas do not state Rosen’s closure as something real organisms meet perfectly. Their fuller sentence is that “all the catalysts (apart from metals, as discussed later) are produced by the organism itself”.

That parenthesis matters. Many enzymes cannot work without a metal ion, and no cell makes iron, zinc or magnesium. They come in with the food. So even the paradigm case of a closed organism takes part of its catalytic machinery from outside, and closure to efficient causation holds only once metal ions are counted as nutrients rather than as catalysts.

What this does for the framework. It shows that where the line between food and catalyst is drawn is a modelling decision, and that the verdict depends on it. That is the same point this site makes about supplied members in the financial loops: the question is always which inputs are counted as food. Rosen’s criterion does not remove that judgment. It makes it explicit, which is what a criterion should do.

The claim this site cannot ignore: simulation is not modelling

Rosen drew a conclusion from closure that bears directly on a site built around a laboratory of simulations. According to Cornish-Bowden and Cárdenas, “Rosen (1991, pp. 182–201) insisted that closure to efficient causation implied that it was impossible to make a functional model of an organism”. They quote him from a later work: “We cannot stress this too strongly, simulation is not modelling.”

They explain the distinction. “Behaviour can be simulated, for example in a computer, without any claim that the simulation replicates the internal structure of the system simulated”, whereas “A model, in contrast, requires a far stronger condition to be satisfied, that this internal structure is reproduced”.

The claim is contested. Cornish-Bowden and Cárdenas record that “His view has been controversial”, citing Mossio, Longo and Stewart (2009) and Cárdenas and colleagues (2010), and give their own judgement that “much of the controversy has been due to a failure to understand what Rosen meant”. This page does not settle it and does not try to.

What this site takes from it, as a discipline rather than a verdict. Every panel in the Laboratory reproduces behaviour. None reproduces the internal structure of a living thing, and none claims to. The honesty boxes on the panels already say so in their own words: a simulation of the model, not a recording of a bench. Rosen’s distinction gives that habit a name and a reason. Whether or not his stronger claim is right, a simulation on this site is evidence about the behaviour of a stated rule, never about whether something is alive.

Where the framework draws its line

The line is the same one this site draws against autopoiesis and the autonomy tradition, and it should be stated with the same caution. Rosen’s closure is relational and categorical: a system is closed or it is not. It carries no account of what the closure costs to keep, no ledger of the supply it runs on, and no event at which it is lost and does not come back. Those are the parts this framework adds, through its fourth condition and its loss account. Whether they add a measurement that closure does not already contain, or re-describe it, is open, and this page does not claim to have shown it.

What the pairing buys

Three things, each small and each usable.

A name for a test the site was already running, so that a reader who knows the field recognizes it and a reader who does not can find where it came from.

A precise difference between two criteria the site uses: reflexive autocatalysis allows catalysts from food; Rosen’s closure allows only nutrients to come from outside. A structure can pass the first and fail the second, and it fails in the way the membrane card does, at a catalyst the system uses and does not make.

A standing rule for the Laboratory, taken from a critic of simulation and applied to a site full of them: say what a panel reproduces, and do not let reproduced behaviour stand in for a claim about structure.

Not read, and named so it can be.

Rosen, R. (1991). Life Itself. Columbia University Press. The primary source, including pp. 182–201 on the impossibility of a functional model of an organism.
Mossio, M., Longo, G., & Stewart, J. (2009). A computable expression of closure to efficient causation. Journal of Theoretical Biology.
Cárdenas, M. L., et al. (2010). Closure to efficient causation, computability and artificial life. Journal of Theoretical Biology, 263, 79–92.
Louie, A. H., on the mathematics of Rosen’s relational biology.

Sources

Cornish-Bowden, A., & Cárdenas, M. L. (2022). The essence of life revisited: how theories can shed light on it. Theory in Biosciences, 141, 105–123. Open access. Source of every statement of Rosen’s position on this page, the page reference to Life Itself, the quotation on simulation and modelling, the metals exception, and the record of the dispute. Read 20 September 2026. doi.org/10.1007/s12064-021-00342-w · PMC8101340

Hordijk, W., & Steel, M. (2018). Autocatalytic Networks at the Basis of Life’s Origin and Organization. Life, 8(4), 62. Source of the reflexive autocatalysis condition used in the comparison with RAF. Read 20 September 2026. doi.org/10.3390/life8040062

Adjacent-work assessments state where Principia Attractum agrees with and departs from neighbouring frameworks. They introduce no constructs and modify no canon; they locate the framework relative to its field. The reading of this site’s “cannot remake its own parts” test as Rosen’s closure to efficient causation, and of the membrane card as closing on its links and not on its catalyst, is this page’s interpretation and is offered as one. No sovereignty verdict is computed or claimed here.

← Back to Relation to Adjacent Work