What Knowledge Needs to Stay Alive

Working notes, not prose. This page is a research packet for drafting the chapter: the brief, source material, questions, examples, reader perspectives, and the earlier draft.

Brief

Must: Show how ideas persist through time: how knowledge stays relevant or doesn’t, how words drift in meaning, and how memes pass laterally between people and cultures the way bacteria share genes through lateral gene transfer.
Serves: It shows the information world evolving, so the reader can judge an idea by the path it took rather than by its age or popularity.

Quote options

“Just as genes propagate themselves in the gene pool by leaping from body to body via sperms or eggs, so memes propagate themselves in the meme pool by leaping from brain to brain via a process which, in the broad sense, can be called imitation.” — Richard Dawkins, The Selfish Gene (1976), ch. 11 “Memes: the new replicators”. [verified: https://colinallen.dnsalias.org/Courses/2004-2005_Fall/X755/Readings/dawkins-memes.html] Why: The founding gene/meme analogy that the chapter then extends from vertical to lateral transfer.
“Because of the high levels of HGT, evolution at this stage would in essence be communal, not individual. The community of primitive evolving biological entities as a whole as well as the surrounding field of cosmopolitan genes participates in a collective reticulate evolution.” — Carl Woese, “A New Biology for a New Century,” Microbiology and Molecular Biology Reviews 68(2) (2004). [verified: https://pmc.ncbi.nlm.nih.gov/articles/PMC419918/] Why: Supplies the real biology of lateral gene transfer behind the chapter’s picture of memes passing sideways.
“He who receives an idea from me, receives instruction himself without lessening mine; as he who lights his taper at mine, receives light without darkening me.” — Thomas Jefferson, letter to Isaac McPherson, 13 August 1813. [verified: https://press-pubs.uchicago.edu/founders/documents/a1_8_8s12.html ; https://founders.archives.gov/documents/Jefferson/03-06-02-0322] Why: Shows why ideas can spread laterally without loss, which genes and goods cannot do.
“…sounds are too volatile and subtile for legal restraints; to enchain syllables, and to lash the wind, are equally the undertakings of pride, unwilling to measure its desires by its strength.” — Samuel Johnson, Preface to A Dictionary of the English Language (1755). [verified: https://www.gutenberg.org/files/5430/5430.txt] Why: Word drift described by the man who tried hardest to stop it.
“For a large class of cases—though not for all—in which we employ the word ‘meaning’ it can be defined thus: the meaning of a word is its use in the language.” — Ludwig Wittgenstein, Philosophical Investigations (1953, Anscombe trans.), §43. [verified: https://dourish.com/classes/ics203bs04/12-Wittgenstein.pdf (Anscombe text; that PDF drops “of cases”, which appears in standard editions)] Why: If meaning is use, then meaning drifts as use drifts. This is the mechanism behind words changing over time.

Research

Core argument

Start with the garden sign from the earlier draft. “Water on Tuesdays and Saturdays” is knowledge that has outlived its reason. An idea can survive while its purpose dies, and it can die while its purpose is still urgent. So survival and truth are separate questions, which picks up Chapter 7’s “why does a belief survive, apart from whether it is true?”
Knowledge is a process that has to be carried out again and again. Every piece of knowledge exists only while something copies it: a brain, a practice, a text, a tool, an institution. When the copying stops, the knowledge stops, even if the ink stays on the page (Linear A, the BBC Domesday discs, FOGBANK). This builds on Chapter 18: a tradition is a loop, and loops can learn, stabilize or decay.
Knowledge needs three things to stay alive: a carrier (someone or something to hold it), a use (a reason to call it up again), and a way to correct it (feedback that tells a copy when it has drifted). Take away the carrier and the knowledge is lost (Tasmania). Take away the use and it fades into ritual or trivia. Take away the correction and it survives but goes wrong (the switch from lemons to limes against scurvy).
Words drift because meaning is use (Wittgenstein). Use changes with every speaker. “Nice,” “silly,” “girl” and “literally” show drift that is gradual, runs in no set direction, and mostly goes unnoticed. Johnson’s own admission that “some words are budding, and some falling away” shows that even a dictionary is a snapshot of a moving thing, not a way to fix language in place. Darwin noticed that languages and species evolve in “curiously parallel” ways.
Ideas travel sideways as well as down the generations. Genes mostly pass from parent to child, but bacteria swap genes across lineages (Ochman et al. 2000; Woese’s “communal” early evolution). Ideas behave far more like bacteria than like mammals: numerals, paper, “OK,” Buddhism and antibiotic-resistance plasmids all jumped between lineages. The tree of culture is really a web.
Sideways transfer explains both how fast culture moves and how badly it can go wrong. Borrowing lets a group gain a capability it never developed and skip thousands of years of trial. It also means an idea can arrive without the context that made it work, like a sign with no reason attached, or a practice lifted out of its ecology.
The size and connectedness of a group matters more than genius. Henrich’s Tasmanian model shows that when the pool of learners shrinks, complex skills decay while simple ones hold. How long knowledge lasts is partly a matter of demography: how many people learn from one another, and how well connected they are.
So neither age nor popularity tells you much by itself. An old idea may have survived because it works (the Lindy intuition), because it is easy to copy, because it flatters its carriers, or because nobody ever tested it. A popular idea may be spreading sideways faster than anyone can check it. Planck’s grim remark that new truths win when their opponents die shows that even good ideas travel by demography, not by persuasion alone.
Judge an idea by the path it took. Ask who carried it, what it was used for, what corrected it along the way, what it picked up or lost in each transfer, and whether its original reason is still alive. That is a first-principles check that works without deferring to an authority, and it ties back to the book’s goal of not being cheated.
Close on maintenance as the price of keeping knowledge. The Ise Shrine is rebuilt every 20 years so that the skill of building it never lapses. Ostrom’s commons stay alive through rules that get monitored and revised. Keeping knowledge alive is ongoing care, which sets up Chapter 20’s collective floor and Chapter 21’s inherited practices.

Key questions

If nobody remembers why a rule exists, is it still knowledge, or just a habit that looks like knowledge?
What is the difference between information that is stored and knowledge that is alive? Is a book no one can read knowledge?
Why did the Tasmanians lose the ability to make bone tools, when that ability had been useful? What does that say about us and our smartphones?
If “nice” once meant “foolish,” which of today’s words will mean something else in 300 years, and will anyone notice the change happening?
When a dictionary records a new meaning (“literally” meaning “figuratively”), is it describing drift or approving it? Who gets to decide what a word means?
Is “meme” a useful analogy or a trap? Ideas are not replicated with high fidelity the way DNA is. Does the analogy survive that?
Why do ideas spread sideways so much more easily than genes do? Jefferson’s taper: sharing costs the giver nothing.
What gets lost when an idea jumps cultures (yoga into Western gyms, Zen into business books, “karma” into English)? What gets gained?
Is an idea more trustworthy because it is old? What about because many people hold it? When is either a good signal, and when is it a bad one?
If a scientific truth wins only when its opponents die (Planck), how much of what we “know” is just who is still alive?
What knowledge do you personally carry that would die with you? A family recipe, a shortcut through a city, how to fix a particular machine.
Can knowledge be too alive, changing so fast that no one can check it (rumor, viral misinformation)?
How do you warn someone 10,000 years in the future, when your language will probably be dead? What does that problem reveal about all knowledge?
Why did sailors who knew how to prevent scurvy in 1800 get scurvy again in 1875?
Is the internet making knowledge harder to kill, or easier to lose (link rot, dead file formats)?
Which does more to keep knowledge alive: writing it down, or doing it regularly? (Plato’s Theuth vs Thamus against the Ise rebuild)
If bacteria that trade genes adapt faster, should cultures be as porous as possible? What is the cost of that porousness?
A curious child might ask: “Where do words come from? Who made up the first one? If I make up a word and my friends use it, is it a real word?”
A cynical adult might ask: “Isn’t ‘judge the path, not the popularity’ just a polite way of saying ‘trust my preferred experts’? Everyone thinks their own ideas have a good pedigree.”
What would it look like to keep an idea alive on purpose, tending it the way you’d tend a garden?

Examples

Tasmania’s shrinking toolkit (civilization scale). Rising seas cut Tasmania off from mainland Australia about 10,000 years ago. Over the following millennia, Tasmanian societies appear to have lost bone tools, cold-weather clothing, hafted tools, fishing spears and more. Henrich’s model links the loss to a smaller pool of learners, in which complex skills decay faster than simple ones. The case is contested: some archaeologists argue the losses were adaptive, or doubt they happened at all. [Henrich 2004, American Antiquity 69(2):197–214, https://doi.org/10.2307/4128416]
Lateral gene transfer and antibiotic resistance (cell scale). In late-1950s Japan, researchers found that multi-drug resistance in Shigella could pass to E. coli and back during an infection. It traveled on transferable “R factors” (plasmids), work associated with Akiba, Ochiai and later Tsutomu Watanabe (Bacteriological Reviews, 1963). A capability that took one lineage generations to evolve moved to another in hours. Ochman, Lawrence and Groisman (2000) argue that bacteria get much of their diversity this way. [https://pubmed.ncbi.nlm.nih.gov/10830951/]
Bdelloid rotifers (animal scale). These microscopic animals have apparently reproduced without sex for tens of millions of years. Genome studies estimate that roughly 8 to 10 percent of their genes came from bacteria, fungi and plants (Gladyshev, Meselson & Arkhipova, Science 2008; Boschetti et al. 2012). Lateral transfer is not limited to microbes. [needs link check]
Scurvy: knowledge lost after it was won. In 1747 James Lind ran his citrus trial, and in 1795 the Royal Navy adopted lemon juice. Later the Navy switched to West Indian lime juice, which contains much less vitamin C, and the cure appeared to stop working. The 1875–76 Nares Arctic expedition suffered scurvy, and by the early 1900s the leading theory blamed tainted tinned meat (“ptomaine”). The knowledge survived as a word (“limey”) while its working content decayed, because nothing checked the copy. [Maciej Cegłowski, “Scott and Scurvy,” 2010, https://idlewords.com/2010/03/scott_and_scurvy.htm]
The garden sign, generalized as Chesterton’s fence. G. K. Chesterton (The Thing, 1929) argued that you shouldn’t tear down a fence until you know why it was built. The counterweight is that some fences are still standing only because nobody checks them. Pair it with the Tuesday/Saturday sign from the earlier draft. Paraphrase this; don’t quote it unless the wording has been verified.
The Ise Grand Shrine (institution scale). The shrine has been rebuilt on an adjacent plot about every 20 years for well over a millennium (the 62nd rebuild, Shikinen Sengū, was completed in 2013). Each rebuild hands carpentry and ritual knowledge to the next generation while the elders are still alive. The building is renewed so that the knowledge of how to build it doesn’t have to be. [https://www.isejingu.or.jp/en/]
FOGBANK and the F-1 engine: knowledge lost at modern scale. In the 2000s the US nuclear weapons program had to relearn how to make FOGBANK, a classified material, because the original process depended on undocumented details and the workers who knew them had left. NASA engineers around 2012–13 took apart a stored Apollo F-1 engine to recover know-how the blueprints didn’t capture. [Ars Technica, 2013, “How NASA brought the monstrous F-1 ‘moon rocket’ engine back to life” (needs link check)]
BBC Domesday Project (digital decay). In 1986 the BBC released a multimedia survey of Britain on LaserDisc, 900 years after the original Domesday Book. Within about 15 years it was nearly unreadable, and it was rescued around 2002 through the CAMiLEON emulation project. The parchment from 1086 can still be read. This is Vint Cerf’s “digital dark age” worry in concrete form. [https://en.wikipedia.org/wiki/BBC_Domesday_Project]
Words drifting. “Nice” comes from Latin nescius, “ignorant,” and passed through meanings like “foolish,” “fussy” and “precise” before it came to mean “pleasant.” “Silly” comes from Old English sælig, “blessed, happy.” In Middle English, “girl” meant a young person of either sex. The OED records “literally” used as an intensifier and discusses the objections to it. [https://www.etymonline.com/word/nice ; https://www.etymonline.com/word/silly ; https://www.etymonline.com/word/girl]
“OK”: a word that spread sideways. It began as a jokey abbreviation, “oll korrect,” in the Boston Morning Post in 1839 (traced by Allen Walker Read and Allan Metcalf, OK: The Improbable Story of America’s Greatest Word, 2011). It now appears in most of the world’s languages, much as a plasmid spreads across species.
Hindu–Arabic numerals and paper. Positional decimal numerals developed in India, were written up by al-Khwārizmī in Baghdad around 825, and reached Latin Europe through Fibonacci’s Liber Abaci in 1202. Florence’s bankers resisted them: a 1299 Arte del Cambio rule is often cited as banning them in account books. Papermaking traveled west from China through Samarkand. The story that it began with papermakers captured at Talas in 751 is a legend historians treat cautiously. Both show sideways transfer, resistance, and the stories that grow up around a transfer.
Bartlett’s “War of the Ghosts” (people scale). In 1932 Frederic Bartlett had English participants retell a Native American folk tale, each hearing the previous person’s version. The story became shorter, more conventional and more “English,” and its supernatural elements dropped out. Drift follows the expectations of whoever is carrying the idea. [Bartlett, Remembering, 1932]
Polynesian wayfinding revived. In 1976 the voyaging canoe Hōkūleʻa sailed from Hawaiʻi to Tahiti without instruments, navigated by Mau Piailug of Satawal in Micronesia, one of the few remaining masters. The knowledge passed sideways from a Micronesian lineage into Hawaiian practice, and it now has hundreds of practitioners. Knowledge that was nearly extinct was brought back by a living carrier, not by a book.
Warning 10,000 years ahead (thought experiment). The US Waste Isolation Pilot Plant needed markers that would still say “danger” in 12,000 CE. Proposals included hostile landscapes, Thomas Sebeok’s “atomic priesthood” (a ritual order that re-teaches the warning), and “ray cats” bred to change color near radiation. Sebeok’s idea is essentially that a myth might keep knowledge alive longer than a text, which links to Chapter 21. [https://en.wikipedia.org/wiki/Long-term_nuclear_waste_warning_messages]
Continental drift and Semmelweis (ideas right too early). Alfred Wegener proposed continental drift in 1912–15. It was widely rejected until seafloor-spreading evidence arrived in the 1960s. Ignaz Semmelweis showed in 1847 that chlorinated-lime handwashing cut deaths from childbed fever, but he was largely ignored and died in 1865. Neither idea’s rise depended on its age or its popularity at the time. What mattered was the path: a mechanism to explain it, carriers, and new instruments.
Linear A vs Linear B. Michael Ventris deciphered Linear B in 1952, and it turned out to be an early form of Greek. Linear A, from the same island, is still unread. Both are stored equally well, but only one is connected to a living descendant language.

Source material

“…no dictionary of a living tongue ever can be perfect, since while it is hastening to publication, some words are budding, and some falling away…” — Samuel Johnson, Preface to A Dictionary of the English Language (1755). [verified: https://www.gutenberg.org/cache/epub/5430/pg5430.txt] Use: This is a different Johnson line from the one already in Quote options. It uses the language of growth and decay for words, and could open the section on word drift.
“The formation of different languages and of distinct species, and the proofs that both have been developed through a gradual process, are curiously parallel.” — Charles Darwin, The Descent of Man (2nd ed., 1874), ch. III. [verified: https://www.gutenberg.org/cache/epub/2300/pg2300.txt] Use: Darwin makes the language–evolution analogy about a century before Dawkins coined “meme.” The same passage also says “Distinct languages may be crossed or blended together,” which anticipates lateral transfer.
“The survival or preservation of certain favoured words in the struggle for existence is natural selection.” — Charles Darwin, The Descent of Man (2nd ed., 1874), ch. III, following his quotation of Max Müller (Nature, 6 January 1870). [verified: https://www.gutenberg.org/cache/epub/2300/pg2300.txt] Use: Put this beside Darwin’s preceding line that “mere novelty and fashion” also decide which words survive. Selection doesn’t track merit alone, and that point sets up judging an idea by its path.
“…this discovery of yours will create forgetfulness in the learners’ souls, because they will not use their memories; they will trust to the external written characters and not remember of themselves.” — Plato, Phaedrus 275a, trans. Benjamin Jowett. [verified: https://www.gutenberg.org/cache/epub/1636/pg1636.txt] Use: The oldest worry that storage is not the same as living knowledge. It ties to the garden sign and the Domesday discs. There is an irony worth noting: we know about the worry only because it was written down.
“Unlike eukaryotes, which evolve principally through the modification of existing genetic information, bacteria have obtained a significant proportion of their genetic diversity through the acquisition of sequences from distantly related organisms.” — Howard Ochman, Jeffrey G. Lawrence & Eduardo A. Groisman, “Lateral gene transfer and the nature of bacterial innovation,” Nature 405 (2000): 299–304, abstract. [verified: https://pubmed.ncbi.nlm.nih.gov/10830951/] Use: The precise biological claim behind the chapter’s central analogy. It also marks the contrast: eukaryotes (us) mostly evolve by descent, but our ideas behave like bacteria.
“…over an approximately 8,000-year period, from the beginning of the Holocene until European explorers began arriving in the eighteenth century, the societies of Tasmania lost a series of valuable skills and technologies.” — Joseph Henrich, “Demography and Cultural Evolution: How Adaptive Cultural Processes Can Produce Maladaptive Losses—The Tasmanian Case,” American Antiquity 69(2) (2004): 197–214, abstract. [verified: https://api.crossref.org/works/10.2307/4128416 (publisher abstract via Crossref)] Use: The anchor case for “knowledge needs carriers.” Henrich’s The Secret of Our Success (2015, in references.json) develops the idea for a general audience.
“Building trust in one another and developing institutional rules that are well matched to the ecological systems being used are of central importance for solving social dilemmas.” — Elinor Ostrom, “Beyond Markets and States: Polycentric Governance of Complex Economic Systems,” Nobel Prize Lecture, 8 December 2009, p. 435 of the printed lecture. [verified: https://www.nobelprize.org/uploads/2018/06/ostrom_lecture.pdf] Use: Rules survive when they stay fitted to their conditions. That supports the garden-sign argument and hands off to Chapter 20.
“This place is not a place of honor… no highly esteemed deed is commemorated here… nothing valued is here.” — K. M. Trauth, S. C. Hora & R. V. Guzowski, Expert Judgment on Markers to Deter Inadvertent Human Intrusion into the Waste Isolation Pilot Plant, Sandia National Laboratories SAND92-1382 (1993), pp. F-49–F-50. [verified against secondary transcription: https://en.wikipedia.org/wiki/Long-term_nuclear_waste_warning_messages ; primary at https://digital.library.unt.edu/ark:/67531/metadc1279277/ not checked directly] Use: The purest test of whether knowledge can outlive its carriers. The authors wrote it knowing their language might not survive.
“A new scientific truth does not triumph by convincing its opponents and making them see the light, but rather because its opponents eventually die, and a new generation grows up that is familiar with it.” — Max Planck, Scientific Autobiography and Other Papers, trans. F. Gaynor (1949/1950), pp. 33–34; quoted in Kuhn, The Structure of Scientific Revolutions. [UNVERIFIED: this wording comes from Wikiquote, https://en.wikiquote.org/wiki/Max_Planck, which also gives a differently worded revised translation. Check it against the Gaynor edition or Kuhn ch. XII before use.] Use: Even true ideas spread by demography, not only by argument. Use it with care, since the chapter shouldn’t imply that persuasion never works.

Counterarguments and limits

The meme analogy is weak in precise ways. Ideas aren’t copied with high fidelity. Dan Sperber argues that transmission reconstructs ideas rather than replicating them, pulled toward “attractors.” There is no clear unit of culture and no equivalent of a genome. The chapter should present lateral gene transfer as an illuminating comparison, not a mechanism, which matches how references.json treats Darwin and Dennett (comparison, not ladder).
The lateral gene transfer analogy can be overstated. Most eukaryotic evolution, including human evolution, is still mostly vertical. And ideas also pass vertically, from parents, teachers and within a lineage. Culture is a mix of vertical and sideways transfer, not purely a web.
The Tasmanian case is contested. Some archaeologists (for example Read 2006, and responses in American Antiquity) argue that the losses were adaptive simplification, or that the population-size effect isn’t supported. Treat it as a well-known model, not an established fact.
“Judge by path” may be impractical. Most people can’t trace where most of their beliefs came from. The cynic’s objection is that tracing an idea’s path just moves the trust to whoever writes the history. The chapter may need a lightweight version: a few questions anyone can ask, not a full genealogy.
Lindy is a real heuristic. For non-perishable things, age does carry some information: things that have lasted tend to keep lasting. The chapter shouldn’t dismiss age as a signal. It should say what age is evidence of, which is being easy to copy and staying useful under past conditions, not truth.
Drift isn’t decay. Prescriptivist nostalgia (for “literally,” “decimate”) can sneak in. The chapter should treat drift as neutral or even adaptive, and save alarm for losing the ability to correct.
Survival bias. We only have the stories of knowledge that survived, or whose loss was documented. The dead knowledge we never learned about is invisible.
A political edge. “Knowledge needs carriers” quickly raises the question of who gets to be one (colonial destruction of languages, suppressed traditions). That is an important point, but it could pull the chapter off course.

Connections

Ch. 6 (Second Blade): knowledge carried physically. Here the carriers wear out.
Ch. 7 (Third Blade): why a belief survives separately from whether it is true. This chapter works that question out over time.
Ch. 9 (Cost of the Blade): a measure outliving its purpose is the same thing as the garden sign outliving its reason.
Ch. 10 (Thingification): things that harden into social furniture. This chapter shows how that furniture gets maintained, drifts or rots.
Ch. 11 (Knowledge): the map of epistemology. This chapter adds a time dimension, how long knowing lasts.
Ch. 14 (Learning: Measure, Model, Manipulate): “retrieval keeps knowledge usable” belongs to the same argument at individual scale.
Ch. 17 (Maintenance, Meaning, and Magic): maintenance is the price of keeping knowledge alive.
Ch. 18 (Loops): a tradition is a loop that learns, stabilizes or runs away. This chapter is Ch. 18 applied to culture across generations.
Ch. 20 (Raising the Floor): Ostrom and the commons. Knowledge is a collective floor.
Ch. 21 (Practices Before Patterns): Sebeok’s atomic priesthood and the Ise rebuild are myth and ritual used to carry practical knowledge.
Act III close (Unnatural in Our Loops): loops that run “across individuals, cultures and generations” are this chapter’s subject.

Exercise ideas

Trace one belief’s path. Pick one thing you believe that you didn’t check yourself, such as a health rule, a family saying or a “fact” about history. Write down, as far back as you can: who told you, where they got it, and what it was originally for. Mark each step as vertical (parent, teacher, own tradition) or lateral (friend, another culture, the internet). Notice: where the chain goes dark, and whether the original reason is still around. Why: it puts the chapter’s main test into practice by judging an idea by its path rather than by its age or popularity, and it usually shows how little of the path we know.
Find a sign that has lost its reason. Find one rule in your home, workplace or town that nobody can explain (a posted notice, a form field, a “we always do it this way”). Ask two people why it exists. Then decide: is it Chesterton’s fence (there’s a hidden reason still in force) or the garden sign (the reason died)? What would you need to check to tell the difference? Notice: how the rule has survived, whether by habit, by authority or by being useful. Why: this is the chapter’s opening case in the reader’s own life, and it separates survival from correctness.
Watch a word drift. Pick a word you use that your grandparents used differently (“awesome,” “sick,” “literally,” “gay,” “cloud”). Look it up in an etymological dictionary (etymonline.com) and write its path in three or four steps. Then listen for one day for a word that you and a younger or older person use differently. Notice: nobody decided the change, and each step made sense at the time. Why: it makes Wittgenstein’s “meaning is use” and Johnson’s “budding and falling away” something the reader can see happening.

Open questions for the author

How central should the lateral gene transfer analogy be? Is it a spine that runs through the whole chapter, or one striking image in one section? And how strongly should the chapter hedge it, given the meme critiques from Sperber and others?
Should the earlier draft’s garden, water-infrastructure and Ostrom material stay? It leans toward institutions, and that overlaps with Ch. 20. One option is to keep only the garden sign as the frame and move the Ostrom material to Ch. 20.
Does “judge the path” need a named, reusable checklist (carrier / use / correction / transfers), or should it stay as prose?
Lindy effect: introduce it by name and then qualify it, or leave it out to avoid pop-Taleb associations?
How much should the chapter take up the politics of who counts as a carrier (endangered languages, colonial erasure), and how much should it leave to Ch. 20 and 21?
Is there a personal scene the author wants to use, such as a family recipe, an elder’s skill or a word their family uses? The chapter would be stronger with one.
Before quoting Planck, decide between the Gaynor translation and the revised one, and check the page.

Reader perspectives

Curious young child

First reactions:

“Knowledge can DIE?” Great hook. They’ll want to know if it goes to heaven.
The garden sign (“Water on Tuesdays and Saturdays”) makes sense to them: “It’s like the sign at the pool that nobody knows why.”
Bacteria sharing genes sideways is the weirdest, coolest thing in the brief. They’ll want much more of it and much less water-pipe budget talk.
Jefferson’s candle quote clicks right away: “If I light your candle, mine doesn’t go out. That’s like telling a joke!”

Questions they’d ask:

“How does an idea die? Does it get sick?”
“What’s a meme? Is it the funny pictures my cousin sends?”
“Can bacteria really swap bits of themselves? Is that like trading cards?”
“Why did ‘awful’ used to mean something good?” (and “Why did ‘nice’ used to mean silly?”)
“Who decides when a word changes? Is there a word boss?”
“Why did the dictionary man say you can’t lash the wind? What’s lashing?”
“If nobody remembers why the rule is there, can we just stop doing it?”
“How do songs like ‘Happy Birthday’ get to everyone?”
“Why does Grandma know how to make the dumplings and Mum doesn’t?”
“When you play Telephone, why does the word come out all wrong at the end?”
“If everyone believes something, is it true?”

Where they’d get lost, bored, offended or unconvinced:

The earlier draft is almost entirely about infrastructure (pipes, budgets, technicians with “authority to act”). The brief’s exciting parts (word drift, memes, lateral gene transfer) don’t appear in the draft at all. A child would ask where the bacteria went.
“Its intelligence is distributed” is abstract. They’d need the Telephone game, or a class project where everyone knows one bit.
Ostrom’s commons findings are important, but the paragraph is dense. They’d tune out at “graduated responses.”
The revised garden rules (“posted in two languages; tool access rotates; disputes reviewed at a monthly meeting”) read like a list of chores.
The Serves line says to judge an idea “by the path it took rather than by its age or popularity.” A child would find that surprising and exciting, but the draft never says it plainly. Say it plainly.
The draft ends on “whom the rule is for.” Fine, but they’d want the chapter to end on something alive, like a word or a song still travelling.

Examples they’d bring:

Playground games and clapping rhymes that kids pass to kids with no grown-up teaching them. That’s lateral transfer between children, not handed down from parents.
The Telephone game, which shows a message drifting as it passes along.
Slang that changes every year (“Mum, nobody says that anymore”).
Grandma’s recipe that only lives in her hands. If nobody cooks with her, it could die.
A school rule nobody understands (no running in the hall), and the day someone found out why.
Trading stickers or cards. You swap sideways and both of you end up with new stuff, which is the bacteria picture in kid form.

What would win them over:

A cartoon-simple explanation of bacteria swapping genes sideways, compared with trading cards at lunch.
Real word-drift facts. “Nice” once meant foolish and “awful” once meant full of awe. They’d repeat these at dinner.
The idea that a rule can outlive its reason, told through the garden sign plus a kid-sized rule they can go and investigate.
A clear takeaway they can use: “Ask where an idea came from, not just how many people like it.”

Cynical adult

First reactions:

The brief is about word drift, memes and lateral gene transfer. The draft is about a garden sign and water pipes. Those are two different chapters, and the draft is the less interesting one.
“Judge an idea by the path it took rather than its age or popularity” is a genuinely useful promise. It’s the anti-guru tool I’d want. The chapter should deliver it as a method.
Using Dawkins’ memes in 2026 is funny, because “meme” itself is the best example of word drift: it now means a cat picture with a caption.
Ostrom and the commons also show up in chapter 20. It feels like the book is running out of examples.

Questions they’d ask:

How exactly do I “judge an idea by its path”? What’s the checklist: who said it first, what did they mean, who changed it, who profits from the current version?
Is lateral meme transfer a real mechanism or just a pretty analogy? Biologists can sequence genes. What’s the equivalent evidence for ideas?
Memetics was fashionable in the 1990s and fizzled as a science. Why build on it?
If “meaning is use,” is every drift legitimate? Is “literally” meaning “figuratively” fine now?
Who decides when knowledge is “dead” versus just unpopular?
Don’t old ideas that survived deserve some credit for surviving? What about the Lindy effect?
Jefferson’s taper line says ideas spread without loss. Then why do they get distorted every time they’re passed on?
What’s the difference between an idea “staying alive” and an idea being kept on life support by an institution with money in it?
How does this help me spot a misquote or a fake statistic in my feed?
Is the garden sign about knowledge, or just about bad communication by a committee?

Where they’d get lost, bored, offended or unconvinced:

The water-infrastructure paragraph (“drawings… pumps… people trained to listen for a failing valve…”) is a long list that proves a point everyone concedes. It reads like a municipal report.
“Frozen knowledge” vs “dependent knowledge” is a nice distinction but gets one sentence before the draft wanders back to the committee.
The garden committee revising its sign in two languages with monthly meetings is so wholesome it’s hard to take seriously. Real committees fight.
“This is Learning, Creating, and Becoming at a collective scale, but the labels add less than the questions” is the author admitting the framework adds nothing here. Then why say it?
Lateral gene transfer needs one clear, correct paragraph of biology. If it’s vague I’ll assume the analogy is decoration.

Examples they’d bring:

The “10,000-hour rule”: Anders Ericsson’s research on deliberate practice, popularized and simplified in Malcolm Gladwell’s Outliers (2008), then flattened further on LinkedIn. Ericsson publicly objected to the simplification. It’s a perfect traceable path from study to slogan.
“Survival of the fittest” was coined by Herbert Spencer, not Darwin (Darwin adopted it in later editions of On the Origin of Species). It’s an idea whose popular meaning drifted from “best fitted” to “strongest.”
Citogenesis, named in xkcd #978 (2011): a claim added to Wikipedia without a source gets picked up by a journalist and then cited back into Wikipedia as a source. Lateral transfer producing fake knowledge.
The nuclear-waste warning problem: the U.S. Human Interference Task Force (1980s) asked how to warn people 10,000 years from now. Thomas Sebeok proposed an “atomic priesthood” that would keep the warning alive through ritual and myth. That’s the chapter’s exact question, taken to its limit, and it bridges to chapter 21.
“Decimate” once meaning killing one in ten, and “nice” having once meant foolish. These are cheap but everyone enjoys them.
Corporate “best practices” that nobody remembers the reason for: the job’s garden sign.

What would win them over:

A short, repeatable “provenance check” for any idea: origin, original claim, each hand it passed through, what changed, who benefits now.
One idea traced end to end with real sources, showing exactly where it mutated.
Honesty that memetics is a weak science, and that lateral transfer is being used as an illuminating analogy, not a mechanism.
Cut or shrink the infrastructure material that overlaps chapter 20.

Believer / spiritual reader

First reactions:

This is the chapter where religious traditions are the best evidence the book has. The oldest continuously transmitted bodies of knowledge on earth are scriptures, liturgies and oral commentaries. A believer will be surprised if they aren’t the central case.
The brief’s aim, to “judge an idea by the path it took rather than by its age or popularity,” is exactly what Islamic hadith scholarship has done for over a thousand years with the isnād (chain of transmitters). A Muslim reader will be delighted or annoyed, depending on whether the book notices.
The meme framing (Dawkins) will put religious readers on guard. Dawkins used memes partly to describe religion as a mind virus, and they’ll hear that echo whether the book intends it or not.
The garden sign “Water on Tuesdays and Saturdays,” whose reason nobody remembers, is a perfect picture of a commandment whose reason has been forgotten. A believer will think of Sabbath or kosher laws, and wonder whether the book is setting up to say that such rules should be revised once the reason is lost.
They’d appreciate “knowledge can become unavailable, unintelligible, or impossible to enact.” Every tradition that has lost a language or a sacred site knows that grief.

Questions they’d ask:

Does a rule need its reason remembered to stay alive? My tradition keeps some laws (chukim, in Jewish terms) precisely without a known reason. Is that a failure of knowledge, or a different kind of knowledge?
If an idea is judged by its path, how do you judge revelation? Its path claims to begin outside the loop.
You compare memes to lateral gene transfer. Religions borrow from each other constantly: the flood story, festivals, saints, meditation techniques. Is borrowing a corruption or a sign of life?
When “charity” drifted from meaning love to meaning a donation, did something get lost? Does word drift ever count as decline?
Scripture is frozen text. Commentary is living practice. Isn’t the pairing of a fixed text with a living interpretive community the best answer anyone has found to your chapter’s question?
Who has the key to the shed in a religious community, and what happens when the keyholders stop answering to the people who depend on them?
If a practice survives because it’s useful, and not because it’s true, should I stop believing it’s true?
Does it matter to you that a teaching has been tested by millions of lives over centuries, or does the book see that only as popularity?
The Buddha told the Kalamas not to accept something just because it was handed down. Isn’t that the same caution you’re offering, from inside a tradition?
Who keeps the knowledge alive when a language has only a few speakers left, like liturgical Aramaic or Coptic?

Where they’d get lost, bored, offended or unconvinced:

Offended: If the meme section implies that religious ideas spread the way cold viruses do, regardless of truth, without applying the same point to scientific or secular ideas. The chapter should be explicit that memes cover the book’s own ideas too.
Unconvinced: The book’s goal statement (in the outline) says it will help readers avoid being “cheated by religions.” Here, where traditions are the obvious success story for keeping knowledge alive, a believer will ask why that goal isn’t balanced by credit.
Bored: The water utility passage is long and procedural. A believer wants to see a living tradition of interpretation at work, not a pipe diagram.
Lost: The draft never gets to word drift or lateral transfer, which the brief requires. The Johnson and Wittgenstein quotes sit unused.

Examples they’d bring:

The isnād in hadith science: each saying of the Prophet is transmitted with a chain of named narrators, and scholars like al-Bukhārī (d. 870) graded reports by the reliability and continuity of those chains. It is the most developed pre-modern system for judging a claim by its path. It fits the brief almost exactly.
Pirkei Avot 1:1: “Moses received the Torah from Sinai and transmitted it to Joshua, Joshua to the elders, the elders to the prophets, and the prophets transmitted it to the Men of the Great Assembly.” A tradition describing its own chain of custody, and the Talmud that followed is a record of arguments kept alive, including the losing side.
Vedic chanting: Vedic texts were preserved orally for centuries using interlocking recitation patterns (pada, krama, jaṭā and others) designed to catch errors. UNESCO proclaimed the tradition a Masterpiece of the Oral and Intangible Heritage of Humanity in 2003. It is an error-correcting memory system that doesn’t rely on writing.
The Kālāma Sutta (Aṅguttara Nikāya 3.65): the Buddha tells villagers not to go by oral tradition, lineage, scripture or a teacher’s authority alone, but to know for themselves what leads to harm and what to benefit. It is a religious text making the book’s own point about authority.
Word drift in religious vocabulary: “silly” comes from Old English sǣlig, meaning blessed or happy. “Enthusiasm” once meant possession by a god (Greek entheos) and was an insult in eighteenth-century England. “Charity” comes from Latin caritas, love. Each shows meaning drifting along with use, and religious words make especially clear cases.
Syncretism as lateral transfer: Christmas absorbed older winter customs, Buddhism took on local deities as protectors when it moved into Tibet and Japan, and Sikhism draws on both Hindu bhakti and Sufi devotion. These are real cases of ideas crossing sideways between cultures, and they’re more vivid than an abstract meme.

What would win them over:

Using religious transmission (isnād, Avot 1:1, Vedic chanting) as the chapter’s strongest case of knowledge that stays alive, with its failures (lost texts, schisms, abuse of keyholders’ authority) named just as clearly.
Applying meme logic evenhandedly: science, nationalism, self-help and this book all spread for reasons besides truth.
Treating commandments whose reasons are unknown as a real question, not assuming they should be discarded once no one remembers why.
Showing that some traditions already teach the discernment the chapter wants, as in the Kālāma Sutta.

Skeptical scientist

First reactions:

The brief and the earlier draft are about different things. The brief is cultural evolution: persistence, relevance, semantic drift, lateral transfer. The draft is about infrastructure and maintenance (garden sign, water pipes, Ostrom), which reads more like Chapter 17 or 20. Almost nothing in the draft is about words drifting or ideas crossing laterally.
The lateral gene transfer analogy is good and more accurate than Dawkins’s original vertical meme picture. Ideas really do move sideways far more than they move parent-to-child. But the analogy needs its disanalogies stated: memes don’t have a clear unit, copy fidelity is low, and transmission is heavily biased by who’s talking.
Memetics as a research program largely stalled (the Journal of Memetics stopped publishing in 2005). Cultural evolution survived in a different form: population models of transmission biases (Boyd and Richerson, Henrich). The chapter should use the living field, not the dead one.
“Judge an idea by the path it took” is the best thing in the brief, and there is a directly relevant, real study showing how citation paths manufacture authority.

Questions they’d ask:

What’s the unit here: a word, a claim, a practice, a story? Genes have a physical unit; memes don’t. Does the argument depend on having one?
How faithful is idea transmission? If fidelity is low, why do some ideas stay stable for centuries?
Is lateral transfer in bacteria actually a good model, or just a vivid one? What’s the mechanism correspondence (conjugation, transformation, transduction) and what’s the idea equivalent of each?
Which words drift fastest, and is there a measurable pattern?
How do you know an idea is “still alive” versus preserved in a book nobody uses? Is a text in an archive alive?
What does “relevant” mean operationally? Cited? Practised? Taught? Correct?
Does the chapter distinguish ideas that persist because they’re true from ideas that persist because they’re catchy? What’s the evidence that these come apart?
How does knowledge get lost, and is there a case where we can see it happen?
Doesn’t “judge by path” risk the genetic fallacy? An idea’s origin doesn’t determine its truth. How does the chapter avoid that?
Where is selection happening, and on what? Is it on truth, usefulness, memorability, or the status of the carrier?

Where they’d get lost, bored, offended or unconvinced:

Lost at the water-infrastructure section: it’s well written, but a reader who came for the brief will wonder where word drift and lateral transfer went.
Unconvinced by “Calling such artifacts ‘frozen knowledge’… misses their dependence” as the chapter’s main idea. True, but thin as a thesis for a whole chapter.
Uneasy with the Ostrom paragraph here. It’s accurate and properly hedged, but belongs in the collective-life chapter.
Would want the genetic-fallacy worry answered head-on. Judging by path is about reliability of the process, not origin, and the chapter should say so.

Examples they’d bring:

Lateral transfer you can watch. The plasmid-borne colistin resistance gene mcr-1, first reported by Liu et al. (2016, Lancet Infectious Diseases) in China, was then found in samples on several continents within months. It’s the cleanest picture of a useful “idea” jumping between unrelated lineages because it confers advantage under selection (antibiotic use), independent of ancestry.
Word frequency predicts word stability. Pagel, Atkinson and Meade (2007, Nature) found that frequently used words in Indo-European languages are replaced more slowly. Lieberman et al. (2007, Nature) found irregular English verbs regularize at a rate inversely related to the square root of their frequency (“wed” will go regular before “went”). Drift has laws, not just stories.
Measuring semantic drift. Hamilton, Leskovec and Jurafsky (2016, ACL) used diachronic word embeddings across centuries of text and reported that frequent words change meaning more slowly and polysemous words faster. Gives the chapter numbers instead of anecdotes like “awful” and “nice.”
Citation paths creating authority. Steven Greenberg (2009, BMJ, “How citation distortions create unfounded authority”) traced a claim about β-amyloid in muscle disease through a citation network and showed how citation bias, amplification and invention turned weak data into apparent consensus. This is exactly “judge an idea by the path it took,” with the path drawn.
Knowledge loss. Joseph Henrich (2004, American Antiquity) modeled how a small, isolated population can lose complex skills, using Tasmania after sea-level rise as a case; the interpretation is debated among archaeologists, so present it as a model with a contested case.
Stories with family trees. Jamshid Tehrani (2013, PLoS ONE) applied phylogenetic methods to variants of “Little Red Riding Hood” and related tales; da Silva and Tehrani (2016, Royal Society Open Science) estimated some folktales may date back thousands of years. Shows ideas have measurable lineages, including lateral borrowing between traditions.
Falsehood travelling faster. Vosoughi, Roy and Aral (2018, Science) found false news stories on Twitter spread farther and faster than true ones, largely through people rather than bots. Popularity and truth demonstrably come apart.

What would win them over:

Use the brief’s structure: persistence, drift, lateral transfer, then “judge by path,” each with at least one quantified example.
State where the gene analogy breaks (no clear unit, low fidelity, biased transmission, intentional modification) so the reader trusts where it holds.
Make “path” concrete: who tested it, how many independent routes it arrived by, whether it was ever exposed to disconfirmation. Contrast that with the genetic fallacy explicitly.
Move the infrastructure material to 17 or 20 unless it’s reframed around knowledge persistence specifically.

Earlier draft

From “What Knowledge Needs to Stay Alive”

At the edge of a shared garden stands a weathered sign: Water on Tuesdays and Saturdays. It looks like information congealed into wood. Nobody remembers that the rule was written after a dry summer, when daily watering emptied the tank before the lower plots received any.

The sign carries knowledge, but incompletely. It does not explain its reason. It cannot notice a heat wave. It cannot teach a newcomer how slowly to open the valve. For the rule to remain intelligent, somebody must remember, demonstrate, monitor, question, and sometimes revise it.

Knowledge does not stay inside a head. It becomes a recipe, bus timetable, handrail, keyboard layout, clinical routine, or garden rule. These artifacts let strangers benefit from earlier observation and effort. They also make choices look natural. A timetable directs bodies through a city; a form decides which answer can be recorded; a locked shed decides who can use the tools.

Imagine a small town whose water arrives through pipes most residents never see. The system is not just water and metal. It includes drawings that say where the pipes run, pumps sized for a certain demand, people trained to listen for a failing valve, a budget for replacement parts, tests for what comes through the tap, and authority to close a road when repair is needed. A household can turn a handle because many pieces of stored knowledge, material, skilled labor, maintenance, and permission continue to meet each other.

The arrangement can fail in more than one way. A pipe may crack. The records may be wrong. A technician may notice a problem but lack authority to act. A repair may be postponed until the people with the least margin bear the interruption. None of this means water infrastructure is merely a story. It means its intelligence is distributed: no single diagram, worker, or office contains the whole working system. The useful question is not whether the system is natural or artificial. It is whether its parts can still notice, communicate, repair, and answer to the people who depend on it.

Calling such artifacts “frozen knowledge” catches their durability but misses their dependence. Paper fades. Software becomes incompatible. Crafts disappear when no one practices them. Records can be destroyed; access can be restricted; a language can lose its speakers. History cannot be undone, because even a lost practice once changed lives. But knowledge can become unavailable, unintelligible, or impossible to enact.

The garden makes the same maintenance visible at a size where we can see the hose. Who checks the tank? Who has a key? Who repairs it? Whose complaint changes a rule? These are questions about labor and authority as much as information. A garden and a water network differ enormously in scale and risk, but both reveal that public reliability is made, not simply possessed.

Elinor Ostrom’s studies of shared resources challenged the assumption that a commons must always collapse or require one universal solution. Communities have sometimes built durable arrangements through locally fitted rules, participation by users, monitoring, graduated responses, and ways to resolve conflict. These are findings from varied cases, not a recipe that guarantees justice. A group can cooperate efficiently while excluding a neighbor or assigning invisible work to the same person every week.

Imagine that gardeners revise the sign: during heat advisories, each plot receives a measured share; changes are posted in two languages; tool access rotates; disputes are reviewed at a monthly meeting. The new arrangement may work better. It also creates costs: meetings consume time, monitoring can feel intrusive, and equal shares may ignore different needs. Institutions hold memory by choosing what to repeat. Every choice deserves inspection.

Take one ordinary artifact—the household calendar, a workplace form, a building entrance, or the pipe cover in a pavement. Ask what past judgment it embodies, what upkeep it needs, who understands it, and who bears the mismatch when conditions change. This is Learning, Creating, and Becoming at a collective scale, but the labels add less than the questions.

Once knowledge is built into a shared world, changing it becomes an ethical act. The garden does not only ask whether a rule can learn. It asks whom the rule is for.