The Cave Keeps the Note

A clap enters

A cave does not repeat a sound. It keeps spending it.

Clap once. The first wave reaches the ear by the shortest road. Copies follow after striking limestone, crossing the chamber, and returning from ledges, domes, and side passages. Each trip costs energy. Each surviving fragment joins what remains. The cave turns one sharp event into a history.

Cave researchers make that history visible with an impulse response. A brief test signal exposes the direct arrival, the first separate reflections, and then the late, crowded tail. Measurements in decorated and tourist caves show that chambers, passages, surface shapes, and source position produce sharply different acoustic responses.^1–3

A low ceiling may send quick reflections. A long passage can return a late echo. Rough walls scatter some frequencies; broad rock faces preserve stronger mirror-like returns. One reverberation number can therefore hide several acoustic neighborhoods. In three Portuguese caves, measured average reverberation times lay between 1.3 and 1.7 seconds, but the spatial and frequency details mattered.²

A cave is not a container for sound. It is a machine for giving sound a past.

The accounting trick

A clap is almost an impulse: brief enough to reveal what the enclosure does next. Call that response h(t). For any source s(t), the received sound can be represented by convolution. Every instant of the source leaves a shifted copy of the cave’s tail. Speech in a reverberant chamber is not a row of separate beads. Each syllable drags a veil across the next.⁴

The clean textbook tail is exponential. If stored acoustic energy E loses a constant fraction per unit time, then:

Real caves need not grant us one neat slope. Chambers couple. Absorption changes with frequency. The sound field may never become perfectly diffuse. The simple curve remains useful because its failure has a shape: a kink, a double slope, a stubborn low note. The miss tells us where to look.

Ten equal sparks

Pain laboratories use a different impulse train. Heat or mechanical stimuli arrive at controlled intervals. Temporal summation of second pain is associated with repeated C-fibre input and central processing in the spinal dorsal horn. It is related to neuronal wind-up, but a spoken pain rating is not an electrode recording from a dorsal-horn neuron. Keeping those levels separate matters.^5,6

In a 2007 experiment, Staud and colleagues used repeated heat pulses and examined both build-up and maintenance. The fibromyalgia group showed enhanced maintenance of temporal summation and greater aftersensations at 15 and 30 seconds.⁵ A later experiment found abnormalities tied to temporal features of C-fibre pain rather than a uniform rise in every pain measure.⁶

The gap between pulses is the hinge. If input arrives slowly enough, much of the prior response fades. Move the pulses closer and residue accumulates. For equal impulses separated by Δ seconds, the leaky model becomes:

The multiplier is memory written as a number. Near zero, the system clears its desk. Near one, yesterday’s paper remains under today’s.

The interval is evidence

The model can create a rising sequence without making any individual stimulus stronger. It can also produce an aftersensation after input stops. That makes the decay—not only the peak—a load-bearing measurement.

Protocol matters. Castelo-Branco and colleagues compared tonic and phasic tests in 52 people with fibromyalgia. Fewer than 30% met that study’s heat-based criterion for temporal summation at individualized pain-60 temperatures.⁸ The result blocks a lazy claim that every nervous system with fibromyalgia must wind up in the same way.

Other work finds that baseline heat sensitivity and temporal-summation magnitude interact.⁹ An end-point rating is hard to read without the starting temperature, pulse spacing, body site, medication context, and analysis rule. A cave microphone also needs coordinates. Move it behind a rock shoulder and the chamber seems to acquire a different memory.

Reverberation is often mistaken for echo. An echo returns as a distinct copy. Reverberation is the crowd: reflections packed so closely that the ear hears a tail rather than separate arrivals. Temporal summation is likewise not merely repeated pain. It is the change across repetitions when nominal input is held steady.

Listen to the state

The cleanest bridge between cave and physiology is not loudness. It is state. Both experiments strike a system, wait, strike it again, and inspect what remains. Both can be described through an impulse response, a decay scale, and the overlap of successive inputs. Only the cave is close to linear. The nervous system can alter its response while the train is still running.

This difference is where the comparison earns its keep. If a single exponential fails to fit an acoustic decay, physicists look for coupled chambers or frequency-dependent loss. If one time constant fails to fit a pain sequence, investigators must consider facilitation, inhibition, adaptation, expectation, and changing gain. A bad fit is information.

Imagine two caves with the same peak sound level. One swallows a clap at once. The other lets it hover under the roof. Peak level misses the difference; the tail reveals it. In pain testing, a single threshold or rating can also miss the trajectory. A sequence asks a richer question: not only how strongly did the system answer, but how did each answer alter the conditions for the next?

The cave keeps no diary. It has surfaces, distances, and losses. The nervous system does keep history, though not in one place or by one mechanism. Still, the discipline of listening to the tail is useful. Strike, wait, measure. Then resist the urge to call the first loud moment the whole event.

At the chamber mouth, daylight cuts a clean shape. Deep inside, the clap has stopped at the hands but not yet at the walls.