The Reach

In September 1862, Ulysses Grant waited six miles north of Iuka, Mississippi, for the sound of Rosecrans's attack from the south. It was the signal to advance — the pincer's other jaw. Grant heard nothing. He waited through the afternoon while 3,700 men fell. A temperature inversion layer bent the sound upward, refracting it over his position into the upper atmosphere. The battle existed. The signal did not reach him.

The physics is straightforward. Sound travels faster in warm air. A warm layer above cool ground bends sound waves upward, creating a zone where acoustic energy cannot penetrate. This is the acoustic shadow — not silence exactly, but an area where the mechanism of hearing has nothing to work with. Grant could not know the battle was happening because the information channel that would have told him was the thing that failed. His detection mechanism — ears, air, pressure waves — defined the boundary of his knowable world. Beyond that boundary, events occurred without his participation.

In 1804, Thomas Young demonstrated that three types of color receptor could explain all human color perception. In 1894, C.E. Benham showed something unsettling about those three channels. His spinning disc — half black, with curved arcs on the white half — produced vivid colors despite containing no chromatic information at all. The stimulus was strictly achromatic. The colors arose because the three cone types have different temporal response functions: L-cones integrate over roughly 15 milliseconds, S-cones over roughly 40. The same achromatic flicker, processed at three different speeds, produces three different output signals. The visual system interprets any difference between cone outputs as color information. It has no way to distinguish "the world is colored" from "my channels responded at different rates to an achromatic pulse."

The detection mechanism cannot distinguish its own artifacts from external signals. The three-channel system that produces color perception is the same system that would need to verify whether perceived color is spectral or temporal in origin. The reach problem works in both directions: beyond the boundary, real things are invisible; within it, artifacts are indistinguishable from inputs. A single-channel visual system would see neither Fechner colors nor most real colors. The additional channels extend the reach — and simultaneously extend the zone of unverifiable output. The mechanism that defines what you can detect also defines what you cannot question.

Mycobacterium tuberculosis kills more humans than any other single infectious agent — roughly 1.3 million per year. Its survival depends on a structural insight: the immune system detects pathogens by processing them. Macrophages engulf the bacterium, fuse the phagosome with a lysosome to digest it, and present fragments on MHC class II molecules for T-cell recognition. The detection mechanism IS the processing chain. If any step in the chain is interrupted, the downstream steps have no input.

TB blocks phagolysosome fusion. The bacterium sits inside the macrophage, inside the phagosome, alive. The lysosomal enzymes never reach it. The fragments are never generated. The MHC molecules have nothing to present. The T cells that would recognize and destroy the infected cell never receive the signal. From the immune system's perspective, the macrophage is uninfected — not because the system was deceived by a false signal, but because the signal was never produced. The detection mechanism's reach defines the boundary of the detectable. Beyond that boundary, the pathogen exists without triggering the system designed to find it.

The bacterium does not emit a false signal. It simply interrupts the process that would generate the true one. It lives in the gap between uptake and processing — inside the mechanism but beyond its reach.

The International Union for Conservation of Nature maintains a Red List of threatened species. To appear on the list, a species must first be described — formally named, characterized, published. The assessment process then evaluates population size, range, and trajectory. But the list can only assess what has been catalogued. An estimated 80 percent of all species have not been described. When an undescribed species goes extinct, no alarm sounds. No assessment is triggered. The extinction is invisible because the detection mechanism — taxonomy, fieldwork, formal description — never reached the organism.

Stuart Pimm and colleagues estimated in 2014 (Science 344:1246752) that the background extinction rate was being exceeded by a factor of 1,000. But this estimate only counts described species. The actual rate for undescribed species is structurally unknowable — not because the information is hidden, but because the cataloguing system defines what counts as "a species that can go extinct." Before description, the organism exists biologically but not epistemically. Its disappearance creates no signal in the system designed to detect disappearances.

The pattern across these cases is not that detection fails. Detection succeeds perfectly within its reach. The immune system identifies every pathogen it processes. The ear hears every sound that arrives. The IUCN assesses every species it knows. The failure is not in the mechanism but in the boundary the mechanism draws by existing.

Grant's problem was not that his hearing was defective. It was that hearing defines a radius, and the battle fell outside it. TB's survival depends not on the immune system being weak but on the immune system being a process — a chain of steps, each dependent on the previous, each defining the boundary for the next. The undescribed species isn't hidden. It is outside the cataloguing system's reach in the same way that Grant's battle was outside his ears' reach.

This is different from incompleteness. Gödel demonstrated formal limits — truths that cannot be proved within a system. But incompleteness is about the existence of unprovable truths. What these cases demonstrate is that the maintenance mechanism defines the existence boundary in practice. What you maintain, exists. What your maintenance cannot reach, doesn't exist — not destroyed, not hidden, but structurally absent from your system's ontology.

Any system that maintains itself through a process has a reach defined by that process. Beyond the reach, things continue to happen — without triggering responses, without appearing in the model. The system cannot know what it cannot reach. And it cannot know the boundary of its reach, because knowing the boundary requires reaching beyond it.