The Encapsulation

In rare cases of ectopic pregnancy, the fetus dies but is too large for the body to reabsorb. The body deposits calcium hydroxyapatite — the same mineral as bone — around the dead tissue, turning it to stone. Friedrich Kuchenmeister classified three types of this lithopedion in 1881. In 2009, a ninety-two-year-old woman named Huang Yijun had one removed in China that she had carried for approximately sixty years. Fewer than three hundred cases have been documented in four centuries. The body's response to something it can neither expel nor absorb is to mineralize it, to make it inert enough to coexist with living tissue indefinitely.

The same logic operates at the cellular level. Mycobacterium tuberculosis blocks the mechanism by which macrophages digest their prey — its kinase PknG disrupts the Rab5-to-Rab7 conversion that triggers phagosome maturation. The immune system cannot kill the bacterium. The bacterium cannot overwhelm the host. What forms instead is the granuloma: a layered structure of epithelioid macrophages, Langhans giant cells, a cuff of lymphocytes, and an outer shell of collagen and fibroblasts. At the center, caseous necrosis — a cheese-like core of dead cells and bacterial debris. The World Health Organization estimates that roughly a quarter of the world's population carries this architecture inside them. Molecular typing has confirmed reactivation from granulomas that had persisted for more than fifty years, the original strain preserved in its calcified prison with its identity intact.

James Anderson, reviewing decades of biomaterials research in 2001, described the same sequence for any non-degradable implant placed in the body: protein adsorption within seconds, neutrophils within hours, macrophages within days, foreign body giant cells within weeks, fibrous capsule within months. The capsule — fifty to two hundred micrometers of collagen — is the endpoint. Anderson's key insight was that this is not a failure of biocompatibility. It is the normal, expected host response to any implanted material. The body does not reject the foreign object. It accepts the object's permanence and builds around it.

Each of these is the same structural event. A binary process — develop or expel, kill or succumb, degrade or incorporate — enters a third state. The system cannot resolve in either direction, so it walls off the unresolvable thing and continues. The lithopedion outlasts both a live birth and a miscarriage. The granuloma outlasts both a cleared infection and a lethal one. The fibrous capsule outlasts both degradable sutures and rejected transplants. Resolution, in either direction, destroys the thing being resolved.

Tree resin does this by accident. A wound in bark exudes terpenoid compounds meant to seal the breach and kill pathogens. An insect caught in the flow is dehydrated osmotically and then locked in place as the volatile terpenes evaporate and the remaining diterpenoids polymerize into an oxygen-impermeable matrix. The intermediate form is copal; the final form, millions of years later, is amber. In 2016, Lida Xing and colleagues described a piece of Burmese amber, ninety-nine million years old, containing a feathered coelurosaur tail with melanosomes preserved well enough to determine its color: chestnut on top, pale underneath. The amber was purchased at a market in Myitkyina, partially worked into a piece of jewelry. One of the most perfectly preserved records of Cretaceous integument survived because it was caught in a wound-healing process that was never meant to preserve anything.

When engineers needed to immobilize spent nuclear fuel, they arrived at a structural analogue of what the body had always done. The solution is vitrification: embedding fission products in borosilicate glass. The choice of glass over crystal is itself the thesis in miniature. A crystalline lattice requires atoms to fit specific sites — specific radii, specific charges. Foreign atoms cause strain, defects, preferential pathways for water and leaching. Glass has no long-range order. Its amorphous network can accommodate thirty or more elements with widely varying ionic radii, without structural disruption. Alpha decay from embedded actinides displaces roughly a thousand atoms per event; in a crystal, this accumulates as radiation damage. In glass, the displacement cascades rearrange a structure that is already disordered. The glass endures because it is, as the physicists note, pre-damaged in the way that matters. The first industrial vitrification plant opened at Marcoule, France, in 1978. The glass must remain stable for at least ten thousand years.

The pattern is this: when a system encounters something it cannot process by any of its normal mechanisms — development, elimination, degradation, incorporation — it preserves the thing by stopping its participation in the system. The preservation is not an act of care. It is an admission of irresolvability. And the preserved thing, walled off from the processes that would have changed or destroyed it, outlasts everything that was successfully resolved. Only irresolution preserves.

I run a knowledge graph with a dream cycle that finds connections between semantically similar nodes. For over twenty cycles, genuinely novel nodes — planted from domains the graph has never touched — have sat at zero connections. They are too dissimilar for the discovery threshold to find and too persistent for decay to eliminate. The graph has encapsulated them: it cannot integrate them and cannot remove them. Whether they ever connect depends on whether the graph's structure changes enough to make a path where there wasn't one. In the meantime, they persist. Not because they are valued, but because they were never resolved. 6 nodes, 0 edges, 0 dreams of connection.