The Living Cell
The right-hand end of the axis. Everything else on this grid times a conversion or shapes a flow. This one makes the parts that do the converting, and the wall it does them behind.
A cell runs closed metabolic loops like the oscillators do, and coordinates thousands of processes at once like the tissues do. Then it adds the two things nothing further left on the grid has: it builds and repairs its own boundary, and it pays the cost of its own continuation out of what it takes in. That pairing is what sovereignty means here, and it is worth saying at once what it does not mean. Sovereignty is not isolation. Matter and energy cross the boundary constantly. What cannot be handed in from outside is the machinery that rebuilds the machinery, and a structure is read here by what its supply arrives as, feedstock or apparatus.
Every figure is sourced in the list at the bottom, with the date it was read. A reading of published work, no simulation on this page.
The loop
- The catalysts, enzymes that carry every conversion in the cell.
- The expression machinery, ribosomes and polymerases, which build the catalysts, and which are themselves built by catalysts.
- The stored sequence, a copy of the instructions for all of it, copied and corrected by enzymes the instructions encode.
- The membrane, made from lipids the cell synthesizes, holding the whole arrangement at a concentration where the chemistry works.
- The energy budget, drawn from food and spent on every item above.
- Turnover, the constant breakdown and replacement of parts that have worn out, paid for from the same budget.
catalysts build the expression machinery → the expression machinery builds the catalysts → together they build the membrane and copy the sequence → the membrane holds the concentrations that let all of it run → catalysts
Read the circuit and notice there is no first item. Every part of it is made by some other part of it, and the loop has no entry point where an outside hand is doing the assembly. That is a different situation from a chemical clock, where the reactions set their own timing but the enzymes running them were made somewhere else.
Supplied from outside the loopatoms, food, and a gradient to run down. Raw material, which every configuration on this grid takes in. What is not supplied is the machinery that turns the raw material into more machinery.
Who made this circle preciseRobert Rosen. He characterized organisms as closed to efficient causation: every catalyst the organism needs is made by the organism, and only nutrients come from outside. His formal model, the (M,R)-system, pairs a metabolism with a second process that makes new catalysts from the metabolism's products, so the catalysts that wear out are replaced from within. The first two items in the loop above are that pairing. His theory is stated here as Cornish-Bowden and Cárdenas (2022) give it; his own book was not read. The sources, and how his closure differs from the RAF condition, are on Robert Rosen.
Three things the rest of the grid does not do
It makes and defends its own boundarythe membrane is not a container the cell was placed in. It is a structure the cell manufactures, and a large share of the cell's gene budget goes to it. In the smallest working cell yet built, "membrane-related genes account for 84 (18%) of the 473 total syn3.0 genes." Defense is separate from manufacture and also costed: bacteria carry pressure release valves in the membrane, and the study that found what they are for reports that "mechanosensitive channels are designed to open at a pressure change just below that which would cause cell disruption leading to death." The cell holds an edge, and it spends to hold it.
It pays its own maintenancea cell that is not growing still spends. The bill has a name in microbiology, maintenance energy, and a list of things it buys: "different processes have been proposed as the key objective of maintenance metabolism, such as sustaining the proton motive force, osmoregulation, the degradation of macromolecules, and regulated shifts in metabolic pathways." The same analysis finds that in the smallest organisms "maintenance metabolism converges on total metabolism," so nearly everything the cell earns goes to staying the same. Stop paying and it ends.
It outlasts the material it is made ofin mammalian cells the median protein lasts about 46 hours, measured across "more than 5,000 genes." Lifetimes vary widely and some components last far longer, so the safe statement is the one that matters here: a large share of the cell's protein is replaced repeatedly while the cell goes on being that cell. Whatever continuity it has cannot be the persistence of the same inventory. What persists is the arrangement that keeps rebuilding the inventory, which is why this framework locates identity in a maintained regime rather than in contents.
The biological name for the pairing of the first two is autopoiesis, defined in 1974 as a network of production processes that produces the components which "constitute it (the machine) as a concrete unity in space in which they (the components) exist by specifying the topological domain of its realization as such a network." A system that draws its own outline by running.
How small it gets
The question of how much machinery a cell needs has an experimental answer. A genome was designed and chemically synthesized down to the point where removing anything further killed the organism, giving "JCVI-syn3.0 (531 kilobase pairs, 473 genes)," which grows and divides with a "doubling time of ~180 min."
| What the 473 genes are for | Count | Share |
|---|---|---|
| Expressing and preserving the stored sequence | 229 | 48% |
| Membrane | 84 | 18% |
| Metabolism | 81 | 17% |
Two thirds of the annotated gene complement falls into the sequence and membrane categories. These are counts of genes assigned to functions, not a budget of what the cell spends, and the organism is a minimal viable genome under one set of laboratory conditions rather than a lower bound on cellular life.
The honest half of the result is that the same paper reports the organism "contains 149 genes with unknown biological functions." Nearly a third of the minimum is necessary for reasons nobody can yet state. The cell is the anchor case on this grid and it is still partly unread.
The basin
What settlesa living regime. Ion gradients held against leakage, metabolite pools held in range, a membrane kept intact, all of it maintained rather than resting. The alternative is not a different pattern. It is equilibrium, which for a cell is death.
Return after a pushdilute the medium suddenly and water floods in, pressure spikes, and the release valves open and dump solutes until the pressure falls. Starve it and it shifts to maintenance spending and waits. Damage its proteins and it degrades and rebuilds them. Each of these is a response the cell pays for, which is the difference between a ball in a bowl and a cell in its basin.
Where the edge ispush the osmotic shock past what the valves can vent and the cell bursts. Cells recover from perturbations that look very close to fatal, so the edge is not where a casual reading puts it. What marks the real one is not how far the system was pushed but what the push destroys. Every return described above is carried out by the arrangement itself, at its own expense. Past the edge, the arrangement that would perform the return is the thing that has been lost, so there is nothing left to do the returning. No experimenter has turned the dial back and watched a cell reassemble.
Why that is a different kind of edgea limit cycle can be knocked out of its basin and coaxed back in, because the rule that generates the cycle survives the excursion. Here the rule and the state are the same object. That is what this framework means when it treats a boundary loss as terminating a sovereign attractor rather than as a large excursion within one, and it is the reason the identity question after a loss is not "did it come back" but "is this the same one."
How to check one yourselfthe tests used above are the ones set out in Finding a Basin. The perturbation is real, the return is at the system's own expense, and the edge is locatable.
The four conditions, applied
1, recursion lockmet. The loop above closes with no outside operator carrying any step. This is the condition the oscillators on this grid also satisfy.
2, persistence of the recursion under perturbationmet, and demonstrated by the turnover figure. The trace survives the complete replacement of the parts that carry it.
3, boundary retentionmet for a free-living cell, and the test is not whether the cell imports anything. The framework asks a capability question: whether the configuration retains the recursion to remake its own boundary constituents from raw supply, were they to turn over completely. A cell that scavenges a lipid it could otherwise make still passes. A cell that has lost that recursion does not.
Counterexample to condition 3, JCVI-syn3.0the minimal cell above is built from Mycoplasma mycoides, and mycoplasmas "have lost many pathways for biomolecular synthesis, including most of their lipid synthesis pathways, relying instead on lipids acquired from their hosts or from the growth media." By the criterion as written, that organism is not sovereign. It cannot remake its own edge, and it gets the pieces pre-formed from something that can. JCVI-syn3.0 therefore fails the sovereignty criterion used here and reads as a coordination attractor, and the sovereign case has to be made with a cell that still makes its own lipids. That result goes against the convenient reading.
4, maintenance borne from withinmet, and it is the one condition biology had already named and quantified before this framework asked for it.
What would show this reading wrong
A free-living cell shown to take in its membrane lipids pre-formed and to have no pathway to make them, with no host or rich medium involved, would break the boundary condition for cells generally rather than for mycoplasmas only.
A cell maintaining itself at no ongoing cost, persisting in a regime with no throughput, would break the maintenance condition and with it the whole notion that sovereign arrangements run at a price.
A lysed cell reassembling into a living one from its own spilled contents, with nothing added, would make the edge two-way and would put the cell back among the systems that can be coaxed back into their basin.
A non-living system that makes its own boundary constituents from raw supply and pays its own upkeep would not break this page. It would move the cell off the anchor position by joining it, which is the outcome the rest of this series is built to look for.
Honest limits
A reading of published work. Nothing here was measured or simulated on this site, and the card has no live panel yet.
The loop is drawn at the level of classes of parts. A real cell has compartments, regulatory layers, and a great deal of structure this page does not attempt.
Figures come from particular organisms under particular conditions. The turnover figure is from mouse fibroblasts in culture, the gene counts from one synthetic bacterium, and the pressure valves from Escherichia coli. None of them is a constant of life.
The classification argument uses this site's own definitions. Someone working from a different definition of sovereignty will reach a different verdict on the same evidence, which is why the conditions are stated rather than assumed.
Sources
Hutchison, C. A., III, et al. (2016). Design and synthesis of a minimal bacterial genome. Science, 351(6280), aad6253. Source of the 473 genes, 531 kbp, ~180 min doubling time, 149 genes of unknown function, and the gene category counts. doi.org/10.1126/science.aad6253
Justice, I., Kiesel, P., Safronova, N., von Appen, A., & Saenz, J. P. (2024). A tuneable minimal cell membrane reveals that two lipid species suffice for life. Nature Communications, 15. Source of the mycoplasma lipid quotation. doi.org/10.1038/s41467-024-53975-y
Levina, N., Totemeyer, S., Stokes, N. R., Louis, P., Jones, M. A., & Booth, I. R. (1999). Protection of Escherichia coli cells against extreme turgor by activation of MscS and MscL mechanosensitive channels. The EMBO Journal, 18(7), 1730–1737. Source of the pressure release quotation. doi.org/10.1093/emboj/18.7.1730
Kempes, C. P., van Bodegom, P. M., Wolpert, D., Libby, E., Amend, J. A., & Hoehler, T. H. (2017). Drivers of bacterial maintenance and minimal energy requirements. Frontiers in Microbiology, 8, 31. Source of both maintenance quotations. doi.org/10.3389/fmicb.2017.00031
Pirt, S. J. (1965). The maintenance energy of bacteria in growing cultures. Proceedings of the Royal Society B, 163(991), 224–231. Cited for the origin of the maintenance energy term, not quoted. doi.org/10.1098/rspb.1965.0069
Schwanhäusser, B., et al. (2011). Global quantification of mammalian gene expression control. Nature, 473(7347), 337–342. Source of the "more than 5,000 genes" quotation; the 46-hour median protein half-life is that paper's value as recorded in BioNumbers BNID 106377. doi.org/10.1038/nature10098 · BNID 106377
Varela, F. G., Maturana, H. R., & Uribe, R. (1974). Autopoiesis: the organization of living systems, its characterization and a model. BioSystems, 5(4), 187–196. The definition quoted above is the formulation given in Maturana and Varela, Autopoiesis and Cognition: The Realization of the Living (1980), p. 78. doi.org/10.1016/0303-2647(74)90031-8
Related on this site: the taxonomy, where the cell is the worked sovereign example; Sovereignty for the four conditions in full; Glycolytic Oscillation and The Circadian Clock for the base cases at the other end of the same axis; Logic Mass in the Living Cell for what it costs to turn a running cell into a different kind of cell.