The entropy of the signal went up
In the last century the environment a human body has to interpret changed character. Tens of thousands of novel synthetic molecules entered ordinary life — in food, packaging, textiles, water, air, cosmetics, construction. In parallel, the informational environment acquired an entirely new statistical profile: continuous novel stimuli, engineered for salience, arriving faster than any inherited pattern for sorting them.
Both changes have the same abstract shape. The number of distinguishable inputs rose sharply, and the amount of inherited context for classifying those inputs did not. In information-theoretic terms, the entropy of the signal increased while the prior distribution stayed roughly where evolution left it.
A system in that position has a decision to make about thresholds, and it is worth taking seriously that the decision is not obviously a malfunction when it goes the way it has gone.
Hypersensitivity as threshold setting
Consider the classification problem an immune system solves. Every molecule it encounters must be sorted into threat or not-threat. The two errors are not symmetric. A missed threat — a false negative — can be fatal. A false alarm — a false positive — is costly: inflammation, tissue damage, energy expenditure, discomfort. Costly is not fatal.
Under that asymmetry, a system facing a flood of inputs it cannot reliably classify has a rational response available: lower the threshold. Accept more false positives to avoid false negatives. Reacting to a peanut protein, a pollen grain, or a detergent enzyme is what a low threshold looks like from inside a life.
On this reading, allergy is not a failure of the filter. It is the filter operating at a setting appropriate to high uncertainty. The hygiene hypothesis says something adjacent — that reduced early exposure leaves the system poorly calibrated — and the entropy framing is not a rival to it so much as a generalization: calibration depends on the ratio of familiar to novel inputs, and that ratio has moved sharply in one direction.
STATUS — Analogy carried deliberately, with a stated cost. This is a framing, not a mechanism. Chapter Nine states what would distinguish it from the hygiene hypothesis empirically; if nothing does, it should be discarded as a metaphor.
Protective at a price
The reframing does not make allergy benign. Anaphylaxis kills. Chronic inflammation shortens lives. Autoimmunity is not a strategy anyone would choose. Nothing here should be read as consolation, and it certainly should not be read as advice.
What the reframing supplies is a cost structure. A low threshold pays off in an environment dense with genuine novel threats and ruins a body in an environment dense with harmless novel molecules. Which environment a person is in depends on where and when they live, what they are exposed to, and what their population's recent exposure history has been.
That is a claim with distributional consequences. Populations meet the same informational and chemical noise unevenly — by geography, occupation, income, and increasingly by the cultural sorting described in Part II. If the costs and benefits of a low threshold are unevenly distributed, then hypersensitivity becomes another axis along which populations differentiate.
The body as the site of metabolization
This chapter is where the arc's abstractions become physical. Cultural change is metabolized into biological difference somewhere, and the somewhere is a body: an immune threshold, a cortisol curve, a sleep architecture, a gut ecology.
Differential survival, fertility, and mate choice around hypersensitivity are not hypothetical. Severe allergy shapes diet, occupation, geography, and partnership. Chronic inflammatory disease shapes reproductive timing. These are ordinary demographic effects with ordinary genetic consequences, and they are currently increasing in prevalence in exactly the populations most exposed to environmental novelty.
There is a public-health corollary that this chapter offers carefully. Framings that treat hypersensitivity purely as pathology may be missing a functional logic worth measuring. That is an argument for adding a question, not for withholding treatment.