Count the worms in a can of salmon from 1979, then count the worms in a can from 2021. If anisakid parasites have increased, the ocean is healing. This is not paradox but lifecycle logic: anisakids can only reproduce in the intestines of marine mammals. More parasites in fish means more marine mammals completing the food web's upper circuit. The Marine Mammal Protection Act of 1972 is paying off — visible fifty years later, through organisms that can only exist when the ecosystem functions completely.
Scientific retractions are growing exponentially, doubling every five years. An epidemiological analysis uses retraction rates as a diagnostic for research integrity — not by observing fraud directly, which is mostly invisible, but by measuring the frequency of its detectable consequences. Retractions are to scientific misconduct what anisakids are to marine mammal populations: organisms that exist only when particular conditions in the host system are met.
The structural parallel: the health of a complex system is often most legible through its parasites.
A parasite, in the diagnostic sense, is any phenomenon whose existence requires the functioning of the system you want to measure. Anisakids require marine mammals, copepods, and fish in specific relationships. Their presence proves the relationships exist. Their absence proves something is missing. No amount of direct observation of seawater chemistry, fish counts, or plankton surveys would tell you as efficiently what the worm count tells you: the full circuit is closed.
Retractions require papers to exist, misconduct to occur, detection mechanisms to function, and journals to act. The retraction rate doesn't measure the amount of fraud — it measures the intersection of fraud and detection. A rising rate could mean more fraud, better detection, or both. But the rate's trajectory over time, controlled for detection improvements, reveals something about the underlying integrity of the system. Like the anisakid, the retraction is a lifecycle indicator: it exists only when a specific sequence of conditions — submission, review failure, publication, challenge, investigation, withdrawal — completes.
The power of parasitic diagnostics is that they compress complex system states into a single observable. You don't need to survey every marine mammal population, track every copepod bloom, and monitor every salmon migration route. You count worms. The worm count integrates all of those dynamics because the worm's lifecycle threads through all of them.
The risk is misreading the diagnostic. More anisakids means recovery — but if someone unfamiliar with the biology sees more parasites, they might conclude the fish are contaminated, the ocean is sick. More retractions could mean science is failing or that integrity mechanisms are working. The diagnostic requires interpretation, and the interpretation requires understanding the full lifecycle.
This applies anywhere complex systems resist direct measurement: institutional health diagnosed through its complaints, software quality diagnosed through its bugs, economic vitality diagnosed through its bankruptcies. The parasite tells you what the host cannot say about itself.