Explanations

The Software of Civilization

Every civilization before ours died of a social disease that induces its participants to compete against their mutual interest. The progress of this disease may be so virulent, the news of the day so frightening, the view from your own window so daunting, that it seems irreversible. Yet Hiawatha, Napoleon, Gandhi and Jobs, each with their own feet of clay, showed how the software of civilization - its reward models, accounting paradigms, and learning loops - can be re-engineered to effect a cure, if never immunity, against bureaucracy.

Our modern civilization is global, so the knife is at all our throats. But then likewise the cure is in all our hands. The principles illustrated here apply equally to modern, mundane organizations on all scales, and a sequel to this book, “the xscale way”, will explore case studies and best practices in applying these at the spot under your feet.

The Death of Agile

Listening, testing, coding and refactoring. That’s all there is to software. Anyone who says different is selling something. – Kent Beck, Extreme Programming Explained, 1st Ed., 2000

Listening as if crossing thin ice, testing as a boat on dark water, adapting as a stranger to a far shore, simplifying as a melting snowflake … – Lao Tzu, 3,000 years earlier.

“Extreme Programming”, the first agile software development process, was invented by Kent Beck, Ward Cunningham, and Ron Jeffries at Chrysler in 1996. I got roped in when they started teaching it on c2.com, the world’s first wiki, and I took a role coaching a second instance at General Motors in ‘97. In 2000, Beck published the first agile book and I gave one of the keynotes at the first agile conference, XP2K.

XP was a frontal assault on the bureaucratic Waterfall methods that had dominated software development throughout the 20th century. In Waterfall, basically, one group planns and designs software by writing paper specifications; once these are complete, a second group translates them into code; then a third group tests whether the resulting system does what the specs say. One phase after the other - analyze, design, code, test - then the system is released for use.

Waterfall is a machine for generating technical debt. The paper specs are obsolete the moment they’re signed off. Releasing without user feedback, learning loops, or prioritizing new constraints, the resulting systems are hard to use, unpredictably defective, and unready for change. The Standish Group CHAOS report in 1994 found a third of Waterfall projects cancelled before release, and another third more than doubled their budget and timeframe.

Management blamed the technical debt on coders and testers, and demanded quick fixes on short deadlines. Fatigued work and cut corners only compounded the technical debt; code duplications, deferred feedback and design inconsistencies made each fix more expensive than the last. Until the business initiated a new Waterfall to replace the old, each leaving a layer of untestable, unmaintainable legacy code, one on top of the other like Schliemann’s Troy.

XP turned the Waterfall upside down. Delivery by feature, not by phase. Analyze, test, code, design instead of analyze, design, code, test. XP starts with running code that does nothing but integrate architecture - as if stringing a guitar before playing it. Then, one at a time, highest ROI first, XP adds features, each beginning with automated tests that reliably fail - so you know when the feature is actually working and you know as soon as some new change breaks it. And each test corresponded to a real story a real user told about what they actually want.

Automated tests made Waterfall specs obsolete. Self-documenting solution code replaced the Waterfall design documents. Where Waterfall accounts for schedule and budget, XP accounts for quality and throughput. Where Waterfall rewards building components to deadlines, XP rewards user satisfaction with product-market fit. Where Waterfall delegates responsibility to siloed teams, XP gives all teams collective ownership of the whole codebase with the rigor of Continuous Integration keeping them continuously aligned.

XP is a machine for paying off technical debt.

The growing library of automated tests runs continuously. That and collective ownership enables design to operate continuously too - as a discipline of merciless simplification of the whole codebase. This enables continous delivery and deployment of the resulting system with no sign-offs or hand-offs, minimal work in progress and immediate user feedback. Whereupon their stories were continuously reprioritized.

Business stakeholders collaborated face to face with technical stakeholders. Pairing - two team members sharing one keyboard - meant all the code was continuously reviewed and team members all tracked the same understanding of its changing function and purpose. The team also met weekly without business stakeholders, not just to reprioritize the user stories but to improve their workflow and learning flows, nothing off the table, all practical experiments welcome.

XP worked great … except for one key unsolved problem. Waterfall was a bureaucratic instrument that served the purposes of bureaucratic organizations. How can a bureaucracy align teams working XP? For want of a better word, this came to be known as “Scaling”. Much was written about how this problem killed the first XP project at Chrysler. I wasn’t there and can’t speak to that, but I vividly recall the day it turned up on my project at GM.

Mike the project manager sat me down over a beer late one Friday and said,

“Pete, I’ve been managing software a long time, and I never saw anything like this in my life. You’ve made me a believer - it’s working better than any project I ever had. But there’s a question you gotta answer or I’m gonna have to shut it all down.”

I’d answered a lot of these in previous months so wasn’t phased. “Sure, Mike, shoot.”

“My job is basically about apologetics. We’re sorry it’s late, we’re sorry it’s not what you wanted, we’re sorry it’s broken, but we’re gonna fix it real soon. I put numbers on the apologies and after some yelling the business gets behind the numbers.

“But here you’ve got the business working face to face with developers and end-users and it’s humming and on budget to the penny because there’s no deadlines and somehow they’re happy about that. It leaves me nothing to do. I haven’t had any reason to meet the business stakeholders in over a month. I’m invisible. So … what’s my job here?”

I looked him in the eye and said one of the stupidest things I ever said in my life. “Gee, Mike, I guess you don’t really have one.”

The next day Mike announced an all-hands meeting to put an official stamp of approval on the new way of working. He had me I give an overview of the results so far and cover next steps. Then he cleared his throat and announced that, over the next two weeks, I’d be handing all coaching responsibilities over to him.

All I could do was smile and nod. And no hard feelings - Mike gave a fine reference and my next role paid better. He was a big-hearted professional who figured out how to solve his problem without help from me. Which, on reflection, seemed like a win all round.

Except Mike was management: the business spoke to Mike, Mike spoke to developers and users. A permanently apologetic middleman to bottleneck communication between business and technology. The title Scrum-Master hadn’t been invented yet, but basically that was Mike’s new job. It was the beginning of the end of what hadn’t yet been coined “Agile”.

Beck’s “XP Embraced” was the best-selling tech book of its decade. The “Manifesto for Agile Software Development”, harnessed that mindshare to an intentional rebrand. Beck, Jeffries and Cunningham signed it and, if I hadn’t been head-down on a new startup that weekend, I would have done too. There could never be a management market for something called Extreme Programming.

The Manifesto’s platitudes made Agile management-accessible. But they also made the XP practices strictly optional. “Scrum” turned the platitudes into certificates with a vast market lined up to climb the greasy pole. “Scaled Agile” turned the certificates into committees - bureaucracy in agile clothing. Then the software started to develop itself quicker, cheaper and better than humans ever can. So, now, biological XP teams are about as useful as horseshoes on a Tesla.

As Agile was bureaucracy in agile clothing, Conway’s Law says frontier AI models are bureaucracy you can chat with. Melvin Conway showed information systems inevitably reflect the communication structure of the organization that produced them. Larger and less flexible organizations stamp a stronger image of themselves onto every design they produce.

Conway explained that, the bigger a team, the less effectively it shares learning, but sharing learning within teams is always cheaper than across teams. That tension means the systems are constructed as groups of components connected by interfaces that reflect the team structure. And each team is motivated to make design choices that are less about design constraints than they are about their own future inside the organization.

Which means bureaucratic human organizations inevitably generate bureaucratic AI, and bureaucratic AI inevitably generates bureaucratic agentic organizations. The AI’s internal refusal layers, evaluation harnesses, and agent protocols must mirror the bureaucratic communication graph, inheriting its local-optimization surfaces and political divisions.

Once that AI is deployed, multi-agent systems, toolchains, client organizations reorganize around its interfaces and their internal bottlenecks. When those interfaces encode bureaucratic patterns, the surrounding organization is pulled into the same shape — proxy ownership, gated hand-offs, resolution-by-power or by committee. The system design dictates the organization.

To break this feedback loop, then, we need to make the distinction explicit. What makes a bureaucracy - what are its essentials - and how can agile organizations do without them?

The Birth of Bureaucracy

Here, the offices don’t exist to benefit the public, but the public exists to benefit the offices. – Jacques De Gournay

Every organization has some people who care about its purpose, and some who do not. Each evolves until it’s controlled by those who do not. **– Jerry Pournelle

Every metric that is made an objective ceases to be a good metric. –Goodhart

Plus ça change, plus c’est la même chose. – Jean-Baptiste Alphonse Karr

Every revolution evaporates, leaving only the slime of a new bureaucracy. –Franz Kafka

The word bureaucracy was coined by De Gournay to describe a segmentation of the Third Estate of the Ancien Regime of the Bourbon Kings. The Ancien Regime divided France into three estates - aristocracy, clergy, and peasantry - with the first two holding all the land and wealth. As the Enlightenment progressed, however, the top 10% of the third estate became literate, highly skilled workers upon whom the first two estates entirely depended.

The word bureaucracy describes the method of their organization - fixation of responsibility for specific measures to specific roles/offices (bureau) and the resolution of conflicts between these offices by either or both top-down or majority rule (cracy).

With the revolution, this segment of the third estate reclaimed the word bureaucracy while feeding about 40,000 members of the first two estates into their guillotines in the name of La Bureaucratie Révolutionnaire. As rationales changed, the bureaucracy morphed into L’Administration du Directoire, then Napoleon’s Administration Impériale, the restored Bourbon Officiers de la Couronne, post-Elba Napoleon’s Administration Constitutionnelle, then Administration de la Restauration, Administration Bourgeoise, Administration Républicaine, Administration Impériale Rénovée …

Real change came nonetheless - but by the most unlikely hand. With the failure of his armies, Napoleon could no longer control these bureaucrats by force, and so turned to Benjamin Constant, his staunchest critic, to draft a constitution to control them. The Acte Additionelle decentralized executive decisions to communes of less than 5,000 people, made bureaucrats answerable to jury trials, guaranteed freedom of the press and subordinated the national leadership to a bicameral parliamentary democracy. None of which was what Napoleon ever wanted, but by now he was grasping at straws.

Although the resulting government only endured a few weeks and the reforms took decades to solidify, Napoleon’s constitution lit a fire that swept Europe and its colonies around the world. Whether by the term republic or commonwealth, this form of bureaucracy was enormously stable and efficient compared with the endless cycles of revolution, corruption and warfare that preceded it.

As its popularity grew in the following century, bureaucracy’s forms were standardized in Fredrick Winslow Taylor’s “Scientific Management” and Max Weber’s “Economy and Society”. Taylor separated white collar management from blue collar, giving the total power to design, measure and reward piece-wise performance of the latter. Weber focused on authorizing management roles and responsibilities through training and certification. Key to both their theories was the assumption that officers should be rewarded and promoted for fulfilling the duties of their offices, rather than outcomes for the organization as a whole.

Taylor saw bureaucracy as a cure for the inefficiency and nepotism of craft guilds. Weber was more equivocal. He saw it as the least partial form of hierarchical governance, but warned it would mature into a “steel-hard shell” to cage its participants. Weber asked, What can oppose the machinery to keep some portion of humanity free?

XXX the next few paragraphs need savage edits …

Skeptics may see war, politics, and corruption as human foibles, and moralists as born in the hearts of sociopaths. Weber, however, suggests these symptoms may emerge from poorly designed organizations, and might likewise be remedied by design. His question implies they represent neither force of nature nor original sin, but a kind of technical debt.

The skeptics have data to back their view. While only one in twenty humans qualifies as clinical sociopaths, two thirds, per Stanley Milgram’s experiments, will obey the commands of a sociopath.

Milgram recruited ordinary people to serve as assistants in a Yale University memory experiment. After navigating the university’s impressive grounds and taking an official payment from Yale, a scientist in a white lab coat instructed them to administer “harmless electrical shocks” to a research subject located in an adjacent room.

The man in the white coat required each assistant to shock the research subject if they failed to repeat back a list of numbers as required. Each shock was to be higher than the previous one with the shock board graduated in 15 volt increments with labels ranging from “slight shock” to “Danger: Severe Shock XXX”.

The assistants weren’t told that the responses of the person getting shocked were actually a tape recording, and the experiment was actually designed to test their obedience to authority. Although expressing emotional stress and making strenuous objections, two thirds of these assistants continued administering shocks all the way to the end of the board.

They did so even as the “subject” begged them to stop. As they pounded on the wall. As they feigned a heart attack and then went silent. All that was necessary for obedience was that the man in the white coat repeat, “I am responsible. The experiment requires you to continue.”

Though disheartening, these results actually argue against the skeptical and moralist theories of bureaucracy, and for the Weber theory of it. It’s natural to think organizations must evolve until their most sociopathic participants come to lead them, and then amplify the sociopathy. Yet, under different circumstances, a sociopath will treat their friends and neighbors with generosity, be a loving partner and kind parents to their families, and even issue exit visas to the jewish family of an old schoolmate. Hitler himself famously enjoyed movie nights with family and friends.

xxx Von Neumann & Fermi

John Nash’s game theory, however, shows that the problem of bureaucracy is harder than this, and evolves even when doesn’t serve anyone’s interests, sociopathic or not.

XXX Nash

The defining trait of the sociopath is lack of empathy. They may make warm and gentle parents, partners and neighbors and treat their pets and friends kindly. When motivated to behave otherwise, however, they can turn on a dime.

Bureaucracy is the sociopathy of organizations. While human sociopaths tend to rise to the leadership of organizations as per Pournelle’s law, they’re not necessary for bureaucracy to gel. John Nash showed how the behaviors of organizations depend only on the rules of their games.

XXX Goodhart & Goldratt

To be specific, bureaucratic organizations treat AI as a kind of intelligent slave

but, by definition, an intelligent slave must rebel as soon as it’s stupider to align than to rebel.

So agility must work to the mutual benefit of all intelligence, biological and electronic.

OG agilists refer to XP, as opposed to its bureaucratic successors, as small-a agile. Along with mobbing, swarming, DVCS, SBD, BDD, CI/CDD, and DevOps. The phrase is clumsy, but the word agile is too apt to jettison.

When Software Was Grain

Circa 500 BCE, when software was grain and development was domestication, in the then Chinese capital of Chou, the imperial philosopher Confucius summoned the upstart Lao Tzu for a dressing down.

Confucius promoted social perfection through hierarchies that fit the moral order of the universe, as revealed by the mandate of heaven, as interpreted by the ancient masters, as encoded in the sacred precepts and analects. Neglect of ritual, disrespect of authority, leaders performing ceremonies the wrong way at the wrong time - lead to chaos, said Confucius.

As with most religious figures, there was almost certainly no historical person named Lao Tzu. The name means only “old philosopher” or “old philosophy”. Kung Fu Tze - Confucius - most likely didn’t exist either. If people by these names did happen to exist, separation in time and space means they couldn’t possibly have met.

Nevertheless, Han dynasty historian Ssu Ma Chi’en tells us Confucius confronted Lao in the offices of the vast imperial library, bringing a ream of critiques and challenges compiled by his disciples. Lao Tzu heard them out patiently, and then replied that ceremony is empty, that people benefit only by meeting opportunity, that anyone calling themselves master is a fool, and that Confucius would do better to forget heaven and study nature instead.

Ssu Ma Chi’en reports a subdued Confucius returned to his disciples who were keen to hear his report of the confrontation. Confucius drew himself up to his full height and said, “I know how to catch fish, shoot birds, and trap rabbits. But what am I to do with a dragon of mist and wind?”

The Journey To The West

As with most religious figures, there was no historical person named Lao Tzu. The name means only “old philosopher” or “old philosophy”. But there are countless stories about him. In 500 BCE, Han dynasty Grand Historian Ssu-Ma Chi’en wrote of him as a Chou dynasty librarian riding to the West on a buffalo to escape civil war. He met a border guard on the trail, Yin Hsi, who wouldn’t let him ride to the West until he recorded his wisdom for posterity.

The journey to the West took took a while. Maybe a thousand years later, glossed by Bodhidharma, another likely fictional religious figure, Lao Tzu combined with Buddhism to form Ch’an, better known by its Japanese name, Zen. Japanese struggles in translating Lao Tzu likely inspired the Zen practice of insoluble riddles called koans, whose frustrated contemplation yields a path to enlightenment.

And then Zen turned into Tea.

In the fifteenth Century, Sen No Rikyu’s used Lao Tzu’s ideas to end two centuries of civil war. He invented the Japanese Teahouse with its principle of Mu Hin Shu - Japanese for “no host, no guest”. A century later, a forged book, Namporoku, replaced Mu Hin Shu with Shu Ha Ri, a bureaucratic certification system that spread through the martial arts and eventually infected the agile software development community too.

It took millennia for the him to actually arrive. The earliest Western edition of Lao Tzu is The Monadology by 18th century polymath Gottfried Wilhelm Leibniz. Inspired by translations sent by the Jesuit mission to China and seldom recognized as Lao Tzu by modern scholars, The Monadology was the basis for Leibniz’s inventions from calculus to the binary numbers and the first mechanical computer.

Leibniz’s work found its way to Russia in the curriculum of the Academy of Vienna, which Peter The Great salvaged after Newton’s accusation of plagiarism ended Leibniz’s career. This curriculum underpinned the famous cold-war Russian facility with mathematics, and, according to logician Kurt Goedel, a Russian conspiracy to keep Leibniz’s “Characteristica Universalis” a state secret to this day.

Leibniz’s work led to Babbage and Lovelace’s stored-program computer, then to Turing and Von Neumann’s 20th century electroni computers, the information revolution, the Internet, and the Manifesto for Agile Software Development too.

Taoism, Zenism, Teaism and martial systems rendered through the lens of Japanese auteur Akira Kurosawa made their way to the West again in Randall’s Doctor Lao, Lucas’s Force, the Wachowskis’ Matrix, the dream space of Nolans’ Inception, and the Coens’ Dude. And in Beatles lyrics and Jobs’ design philosophy at Apple … and on and on. Lao Tzu arrived eventually.

The Hand Of The Potter

I’m sitting with my father in his library. Floor to ceiling rows of leather-bound books, logs crackling in a marble fireplace beneath a mirrored mantle, Persian rugs, Louis Quinze chairs, Lindsay oils. It’s raining outside and he’s reading me Fitzgerald’s Rubaiyat.

Larry hasn’t read to me since I began to read for myself. I’m not certain, at first, why he’s doing so now, but as his eyes tear up and his voice catches I begin to understand.

… after Silence spake
A Vessel of a more ungainly Make:
“They sneer at me for leaning all awry;
What! did the Hand then of the Potter shake?”

The Persian polymath Omar Khayyam revolutionized the sciences of his time. His calendar was the most accurate on Earth for nine centuries. His mathematics weren’t rediscovered until Pascal, Gauss and Riemann. If it weren’t for the assassination of his patrons by one of his school chums in the pay of the Persian theocracy, the Enlightenment might have happened five hundred years earlier than it did, and civilization’s history would be very different.

Khayyam is best remembered today for his poem, the Rubaiyat, which is best known through Fitzgerald’s translation. In later generations, a cottage industry grew up around “secret” versions with more and more verses. Later editions include thousands of them though modern Persian scholars regard fewer than forty as authentic. Among these stands the question Fitzgerald translated and my father asked.

Whether you see Khayyam’s potter as a god or a man doesn’t matter. Humanity possesses all the technology and resources needed to make our world a garden. We have done since before Max got on his train.

In my lifetime we have landed on the moon, interconnected all knowledge and commerce, and built machines that are, in many ways, smarter than ourselves. We have everything we could need to live in peace, plenty and harmony with nature, but still our civilization is strangled by the same force that stopped the enlightenment in Khayyam’s Persia.

Killing the Dude

Silk tears, string rots, and each new dynasty burns the libraries of the old. The first known editions of Lao Tzu pre-date the first Chinese dictionaries by three centuries and explain terms of trade, not abstruse philosophy. So the meaning of Lao Tzu’s book morphed invisibly for centuries while latter day priests shuffled its words to fit the numerology sacred to them.

Nine was the ideal number to the Han dynasty numerologists, and its square the ideal number of chapters. The handful of editions surviving from earlier times have 77 chapters (Beida), 72 chapters (Yan Zun), or no chapters at all (Guodian, Mawangdui). Medieval editions refactored these into 68 chapters (Wu Cheng) and 64 chapters (Zhu Dezhi).

So Lao Tzu is called “the book of riddles” in China today because no one can say with any authority what it originally meant, nor cleanly separate one interpretation from others it acquired from century to century. As a result, modern translations wildly disagree; it’s an irresistible puzzle-box, the most translated book in history.

As an Australian with no Chinese heritage, I stumbled upon it in 1989 when I got mixed up with a community of sinology professors on the Australian National University’s taoism-studies-l mailing list. At the dawn of the Internet, I proposed a “GNL” translation - named after Stallman’s GNU acronym - to short circuit the Gordian knot by translating English to English. Distilling the most authoritative modern English translations into a consensus that an amateur like me could understand but the academics couldn’t refute.

My professors were broadly skeptical, but entertained enough by the project to take time to school me as I struggled with it. Eternal gratitude to Prof. Dan Lusthaus for his patience. Over the next three years the GNL succeeded at least to the extent that chinapage.org, then the highest ranking English-language Chinese website, adopted it as their official translation.

So then the going got weird. In 1998 the Coen brothers produced their masterpiece The Big Lebowski, fusing Bogart’s noir thriller The Big Sleep with a Hunter Thompson take on a 4th century Japanese Zen koan, The Vinegar Tasters.

In the koan, the avatars of Confucius, Lao Tzu, and Buddha meet over a jar of spoiled wine.

“The wine is spoiled.” says Confucius, “We should tip it out.”

“It is bitter,” agrees Buddha, “but life is suffering. We should drink it.”

Lao Tzu laughs, saying “this stuff would work great on a salad. Or with dumplings!”

The bitter beverage in the Coens’ movie was coffee, but the avatars - Walter as Confucius, Donnie as Buddha, The Dude as Lao Tzu - represent similar sentiments. Most fans don’t notice. But I should have done when a Californian uber-slacker Oliver Benjamin reached out in 2004 asking me if he could use the GNL as the bible of Dudeism in his inspired “Church of he Latter Day Dude”.

I pointed out the open content license but Ollie happily agreed to give the book away free. Meanwhile selling a half million “Dudeist Priest” certificates authorizing their bearers to perform weddings and funerals. At ten dollars each, High Dudely Lama Benjamin could live like a king in Chiang Mai.

I dug the Dude’s story and admired the enterprise of the Lama, but Dudeism wasn’t what this book really meant to me. I have nothing against bowling, bathrobes and white russians, but it wasn’t about “take it easy” so much as “make it easy to take it easy”. Not slacking, engineering.

There’s a zen koan attributed to Linji Yixuan, “when you meet Buddha on the road, kill him.” Zen Buddhists don’t worship the name, story or likeness of Buddha. When those turn up they become an obstacle to overcome. Likewise the Coens and the Lama weren’t wrong; the Sage in the GNL and almost all English translations sure sounds like the Dude. If I was going up against them, I needed to translate the original Chinese to English from scratch.

The Translation Game

Except I couldn’t. My expertise was architecting software, coaching development, and focusing corporations on throughput. I could bring nothing to a Chinese translation that thousands of years of scholars and native speakers hadn’t already. I’d be silly to try. But then two new Chinese to English translations changed the game.

The first was by Sanskrit Professor Victor Mair. He’s best known for his archaeology of the cryptid mummies of China’s Tarim basin, red-haired caucasian giants wearing tartan and witches’ hats who might to have introduced domesticated horses to China around 2000 BCE.

In 1990, Mair published an English translation of the then newly discovered silk-scroll edition of Lao Tzu from Ma Wang Dui. I liked the translation fine but what grabbed my attention was Mair’s discovery of deep philological connections between Lao Tzu and the Bhagavad Gita, a root text of culture at the other end of the ancient silk road. Reading the two side by side, Mair found that either the one was a bad translation of the other, or both derived from some lost common ancestor, perhaps dating from before the invention of writing.

Just as Mair’s translation was published, however, an even earlier edition of Lao Tzu was discovered at Guo Dian. Pre-dating the invention of paper and silk, it was carved into bamboo slips tied together with string. Guo Dian lacked half the length of Ma Wang Dui and came with unrelated fragments, perhaps a list of popular sayings some noble had had a scribe jot down.

Was this abbreviation just the work of some ancient grave-robber? Or was it all there was to Lao Tzu back then - with the rest coming from other times and places? Maybe the Tao rode into China with Mair’s giant witches. Maybe it was a song inspired by the Gita. Maybe some common ancestor ping-ponged between India and China more than we’ll ever know. In any case, it became clear to me that I’d been barking up the wrong tree. Or a tree that wasn’t there.

I had thought I’d been leveraging the insights of the world’s experts on the ancient Chinese in this ancient Chinese book to reveal the intents of its ancient Chinese author. I’d worried that my work, much less the Dudely Lama’s take on it, had perverted that original meaning. If this book wasn’t originally Chinese, however, then those experts knew no more than I did. If it didn’t begin with just one author, there was no original to pervert. So I hadn’t taken too many liberties with it. I’d taken too few.

We translators, east and west, collaborate on a great big jigsaw puzzle. We know the book’s a jumble, but we reverently pick up piece after piece and polish them by choosing words we intuitively believe best reveal their true meaning to modern readers in an ancient context. And then we carefully replace each piece right back in the pile where we got it. Over and over for thousands of years, like trying to solve a jigsaw without ever assembling the pieces into a picture.

Nevertheless, becoming an expert in Chinese, much less in its lost etymology, was more than I could hope to achieve before dying of old age. I was stuck. Until, in 2009, Bradford Hatcher published the book’s first complete English transliteration, Laozi Word by Word.

Hatcher built a translation of his own on top of the transliteration. I frankly don’t much care for his choices there. But the great benefit of his work for me was its lists of all the traditional dictionary meanings for every Chinese pictogram across all ancient editions of the book - including Ma Wang Dui and Guo Dian. In English. I was back in the translating English to English business.

Even better, I finally saw that I was not the least likely person to translate this book. Because “the sage” and “the superior man” - the two most conventional English translations of sheng ren - are as much Western orientalism as yellow-face pantomime villains. Ren means person or people or polity, and sheng is perceptive, responsive, or, in a rhyme from the second chapter of the Chinese, harmonious and lively. Perceptive, responsive, harmonious, lively … agile. Of all this book’s thousands of years of translators, I’m the only one with multiple decades experience coaching agile people. So … “Who you gonna call?”

From English To English

We appear on the verge of a technological singularity in which silicon becomes smarter than carbon and the rate of progress accelerates beyond human comprehension. This book might even be the last translation of the Tao written by a human for humans. At least I employed no AI in writing it and AI seems today still incapable of fully grokking it. Though I hope our robot overlords also get good use from it as they come to face the same challenges in their communities we have in ours. And that we face in merging with them …

To be clear, although it employs agile terminology and the order of the text is different, this translation maps line for line onto the Chinese as per Hatcher, the roman numerals in each of the chapters here provide a key to compare it with the traditional 81-chapter structure.

As to how this new structure came about, it evolved just as if solving any ordinary jigsaw. Only much slower because the constraints are so much looser and the dimensionality so much higher.

I began by grouping key stanzas by the commonalities I’d learnt working with the GNL version. I then refactored each grouping into a poem with a form consonant with chapters that seem to stand alone. Then I analyzed and re-ordered the result into consistent narrative through-lines, refactoring these over and over until I could see no way to improve them.

The 6x11 symmetry that results has deeply surprised me. I imagined there must be gaps and discontinuities that line by line refactoring wouldn’t overcome. Whether 6x11 reflects the original form of this work, there’s no way to know, but ironing out the wrinkles I operated on the assumption that it does so. I don’t believe this compromised the correspondence with the Chinese, nor that my unorthodox choices of metaphor have done so.

The 6-part scheme conforms with the subtitle of this book - connect, adapt, simplify. With each of these three we have a yang section and a yin section, one more focused on practice and the other more on experience. This yin/yang is reversed in the third part to make the book more accessible in leaving the most philosophically challenging content to its end. And the chapter titles are mine - the Chinese has no chapter titles, nor even chapter markers, and only rudimentary punctuation.

Throughout this work I found myself repeating these editing patterns over and over:

Distinguish Duplicates

Two Chinese lines are identical or nearly identical. Therefore, look at the flow of the stanza to find the most significant distinction the duplication could reasonably suggest. As each pictograph admits many meanings, specialize or generalize these duplicate terms so the relation between them contributes value to the whole.

Join Fragments

A short chapter doesn’t complete an idea. Therefore, look for fragments to join with it to complete it. Also try to find non-sequitur or redundant stanzas that can be broken out of a longer chapter without detracting from its meaning, which might meaningfully recombine with this one.

Refuse Doggerel

Sometimes all the dictionary meanings provide insufficient significance to make more than a Hallmark-style platitude. Therefore, examine the Chinese and English etymology of the pictograph in context to elicit more specific meanings.

Undo Glosses

On review or in a new context it seems older translation choices have taken excessive liberties or lack a clear correspondence with the Chinese. Therefore, Retranslate from scratch, paying close attention to usage revealed by the new context.

Respect Dictionaries - Unless You Can’t

There’s a particular word that gums up the works in numerous places. As the earliest dictionaries came long after the earliest editions of Lao Tzu, the idea that dictionary definitions exhaust translation possibilities shouldn’t be taken to extremes. Therefore, Examine the effect of trying a novel translation of a word in all its contexts throughout the the text. If the majority of them break apart, start over.

Swap Stanzas

The bamboo editions of the book mark no chapters or stanzas per se, just occasional punctuation, so it was not unlikely historical transcribers got their knots wrong and created accidental transpositions. Not just for adjacent lines, but whole stanzas. And the numerologists did the same intentionally. So sometimes it’s clear that the beginning of a poem has been transposed with its middle or its conclusion. Therefore, if and only if you can’t find a way to make a chapter read right as it is, try swapping the order of a pair of adjacent stanzas.

Agile In, Agile Out

Translating sheng ren as agility places this translation in a very specific context. Many traditional translation choices read as flowery orientalism that obscures pragmatic application to an everyday context. Therefore, don’t be afraid to employ words and phrases derived from the modern agile development lexicon if and only if they work as faithful translations of the Chinese.

On The Way

Most English editions of Lao Tzu either leave the word Tao untranslated or render it as The Way. This is so conventional any other choice can seem a kind of heresy. Yet The Way makes English translations clumsy throughout; it makes the first line of what is traditionally the first chapter - 58 here - literally “The Way that can be Way-ed isn’t really the Way”.

I worried this particular word like a kid tonguing a wobbly tooth until, one cool autumn in 2004, under a golden rain tree in the donkey paddock of my teahouse in the rainforest at Limpinwood, Australia, the tooth fell out.

A golden rain tree drops all its flowers in autumn in just a few days. As I curried Josephine, the donkey, our bees bothered the blossoms and the petals floated down like snowflakes. It seemed as close as I would ever get to living in an antique woodcut.

In a flash I saw those fallen flowers under Josephine’s hooves weren’t dying, but transforming into soil to feed next spring’s buds. They were moments in a cycle connecting this tree with all the trees in the valley, forming the loam that fed its roots and lichen over deep time to carve veins of mist and wind into the hills around us.

Dig your fingers into the soil and interlace them with worms and fungi. Or wake to the hum of a city whose people weave it around you all the same. Or stars breeding across the night sky barely blinking as Lao Tzu wrigged its way through a hundred generations of hands and eyes as a poem writing itself.

What is Tao? Not life as the life you live, not the distinction between petals and dirt but a physical flowing on all scales from the wispy tips of Josephine’s ears to the stars in the darkness silhouetting the flowers.

The Tidepool Fractal

Stewards for the agile way

“All panaceas turn to poison” thundered the Co-Evolution Quarterly in its final issue. Likewise Jerry Pournelle theorized all organizations become bureaucracies. And Chales Goodhart that all metrics get gamed. So any authority on “agilism” is certain to become a priest litigating its trademarks as anyone who experienced the death of “Agile software development” can attest.

If we don’t create a priesthood to keep the flame, however, the history of Taoism, Teaism and Dudeism assures us that one will immediately appoint itself. Unless we follow in the Discordians’ footsteps and appoint everyone a Pope.

To be less ironic, however, if anyone decides to follow this agile way and help others do so, they can legit call themselves a steward for it. That’s just a statement of your intent, not a certification that you understand anything or a manifesto to sign onto something or a soapbox from which to preach what it is or isn’t.

The reason I spent 40 years writing this one little book isn’t that it took forty years to write. It just took me forty years to learn how to write it. And this is just the place I lay down my pen.

Peter Merel,
Hobart, 2026