The Field Remembers

In September 1981, Nature reviewed a first book by a young Cambridge biochemist under a headline no scientific journal of record had used before and none has used since: "A book for burning?" Sir John Maddox, the editor, was not being playful. He wrote that Rupert Sheldrake's A New Science of Life was "the best candidate for burning there has been for many years," and the review's own discomfort shows through the provocation, because the book was not crank work in the usual way. Sheldrake had a Royal Society research fellowship, a decade of published cell biology, and a hypothesis stated carefully enough to be tested. That was precisely the problem. The hypothesis was that nature has habits rather than laws. Memory, Sheldrake proposed, is inherent in things: once a process has happened somewhere, it happens more easily everywhere afterward, carried by what he called morphic resonance, a kind of collective memory pooled across every system that shares a form.

The examples were chosen to irritate. New chemical compounds are notoriously hard to crystallize the first time and get easier afterward, including in laboratories that never spoke to one another. Rats appeared to learn standard mazes faster in the decades after other rats had learned them. The rat evidence had a stranger pedigree than most people who cite it know. In the 1920s William McDougall ran a Lamarckian experiment at Harvard, breeding trained rats for generation after generation to test whether learned behavior could be inherited. Over more than thirty generations, later generations learned the maze faster, and McDougall took this as proof that acquired traits pass down. When W. E. Agar's team in Melbourne spent twenty-five years trying to replicate the result, they failed to confirm the inheritance, but their final report in 1954 carried an anomaly nobody explained away: every line of rats, trained and untrained alike, was learning the maze faster than its predecessors had. The inheritance story died. The improvement did not. Sheldrake's reading is that the field of rat-maze-running had been strengthened by every rat that ever ran it, and that the field, not the germline, is where the memory sits.

Modern psychology has since produced a cleaner version of the same shape, measured at population scale and beyond dispute. James Flynn documented in 1984 that Americans gained fourteen IQ points between 1932 and 1978, roughly three points a decade, a gain large enough that test makers have to renorm the instruments so the average stays at one hundred. Nobody serious believes human genes changed that fast. The standard explanations reach for nutrition, schooling, and test familiarity, and all of them may be true; what the Flynn effect demonstrates either way is that performance on a fixed form improves across generations at a rate no record-keeping theory of the individual can account for. Something outside the file is doing compounding work.

The fury at Sheldrake was never really about the rats. It was about the founding deal of the modern world, the agreement that the universe runs on fixed laws written nowhere, and that all memory lives in records: in genes, in archives, in databases, in the file. A hypothesis that locates memory in the system rather than the record is a threat to everyone whose authority rests on keeping the records. Sheldrake understood this and said so in the talk TED later pulled from its site in 2013, after two bloggers objected: "There's a conflict within science," he told the Whitechapel audience, "between science as a method of inquiry based on reason, evidence, hypothesis, and collective investigation, and science as a belief system or a worldview." The method had produced the anomaly. The worldview suppressed it. Burning is what an archive does to a memory it cannot file.

Whatever the rats eventually settle, the founding deal is worth examining because organizations sign it too. Every company operates on the belief that its memory lives in records: the wiki, the CRM, the process document, the data room. Then the company loses a cohort and learns what the records actually were, an index to a library that lived in habit. The Society for Human Resource Management puts the cost of replacing a salaried employee at six to nine months of salary, and that figure is not the price of reprinting documents. It is the price of regrowing the habits the documents only gestured at: who to call, which exception is real, how the approval actually gets signed. The merger market runs the same experiment at full scale. Companies spend more than two trillion dollars a year on acquisitions, and the Harvard Business Review's reckoning puts the failure rate between seventy and ninety percent. The acquirer buys the records, the contracts, the code, the customer lists, and then watches the thing that made the records valuable, the accumulated habit of the firm, walk out the door or evaporate in the integration. You can photocopy a playbook. The habit travels with people or not at all.

This is also why industries converge without conspiring. Every sector ends up with the same org chart, the same funnel metrics, the same Tuesday standup, and the carriers are not difficult to name: consultants, job-hoppers, investors pattern-matching across portfolios, each one moving the habit from firm to firm the way Sheldrake's crystals supposedly teach one another to form. Best practice is morphic resonance with an invoice attached. The observation cuts both ways. A habit that spreads this easily is a weak moat, and a firm whose memory lives only in what can be copied has already told you what it is worth to a competitor.

What Sheldrake offers, used carefully and without signing up for the metaphysics, is a name for the part of a company that compounds without ever being booked. Culture is the field: invisible, unlisted on any balance sheet, measurable only by what it lets people do without being told. Capital can buy every record a firm owns and none of its memory, which is the mechanism underneath a divorce everyone in markets has watched happen in slow motion, capital drifting ever further from the culture that gives it somewhere productive to land. The spreadsheet sees the acquisition price. It cannot see the field, and the field is what the price was for.

Maddox died in 2009, the review his obituaries all mentioned. The rats, if the anomaly is real, are still getting faster, and so is everyone else, three points a decade, with the test makers quietly resetting the average so nobody has to discuss it. Something is doing the remembering, and it is not the archive. A company that treats memory as records will keep buying back, at six to nine months of salary a head, what it already knew.