Great Progress

Metrics that discipline the argument

Child mortality is among the most unforgiving of statistics. It measures the fraction of children who die before their fifth birthday — a brutal, if incomplete, index of how well a society protects its most vulnerable. Historical estimates suggest that in many populations around 1800 roughly one child in three, and sometimes more, never saw age five. The estimates are reconstructed from incomplete records, but the scale of loss is not in doubt: childhood death was common across continents, empires, and faiths.

Redraw the map for 1950 and the picture fractures. The industrialized world — Western Europe, North America, Japan, Oceania — had cut child mortality below five percent. Antibiotics, vaccines, sanitation, and rising prosperity insulated their children from the epidemics and malnutrition that once claimed millions. But the rest of the globe still bled: across Africa, India, the Middle East, and Latin America, between a fifth and two-fifths of children still died young. The division was stark, and it was not racial or cultural or providential. Modernity saved lives, and its absence cost them.

Redraw it once more for 2015 and the transformation is unmistakable. In wealthy nations child mortality sits below one percent. Latin America, East Asia, and the Middle East approach survival rates once reserved for kings’ heirs. Even Sub-Saharan Africa, still the most burdened region, has fallen to a range of five to twelve percent — down from one-in-three only two generations before. From one-in-three to under one-in-a-hundred across most of the earth, in the span of two centuries: millions of lives spared, families unbroken, futures restored.

I lead with this number because the underlying outcome is hard to trivialize. The historical rate is still a modeling result: records are incomplete, definitions vary, and uncertainty grows backward in time. Unlike a preference-weighted index, however, survival before age five tracks a concrete capability with unusually broad relevance. The long-run improvement is enormous without needing to call the estimate assumption-free or monotonic in every place and year.

The causes are neither mysterious nor ideological. Vaccines, antibiotics, oral rehydration therapy. Clean water, sanitation, nutrition. Maternal education and economic growth. Each step a practical intervention; together, a revolution in human destiny. The story is not finished — Sub-Saharan Africa and parts of South Asia still carry the heaviest burdens, where poverty, weak health systems, and conflict conspire against further gains. But humanity has demonstrated that child mortality is not fate. It is policy, technology, and will. For centuries death claimed children as a tax on life itself; in most of the world that tax has been abolished. Civilization’s scorecard is disputed on almost every line, but here the verdict is clear: progress is real, measurable, and monumental.

One Number for Civilization

Grant the child-mortality collapse and a broader question follows: how do you measure the level of a civilization as a whole? The usual answer is a dashboard — GDP per capita, literacy rates, technological capability, political stability, happiness indices — and the dashboard drowns you. Every metric offers something; their multiplicity obscures the trend and invites cherry-picking, which is precisely what the pessimists do for a living.

I propose a single robust proxy: life expectancy at birth.

Life expectancy at birth is a powerful summary, not a uniquely sufficient civilizational score. It integrates mortality across ages and responds to healthcare, nutrition, sanitation, prosperity, education, violence, and institutional capacity. That breadth is its virtue. It is also strongly affected by infant mortality, epidemics, migration, and measurement choices, and it says little directly about the distribution of health or the quality and freedom of the lives preserved.

Sustained high life expectancy usually indicates effective public health and material capacity, but an authoritarian society can deliver some of those goods while denying agency. Freedom, disability, happiness, cultural richness, and unequal exposure therefore remain independent dimensions. Use life expectancy as one robust outcome in a dashboard, not as a proxy that dissolves the dashboard.

The Pharaoh’s Missing Luxury

Metrics measure progress from the outside. From the inside, the strange thing about progress is that its greatest products are invisible to the people who enjoy them.

We flatter ourselves that modernity’s luxuries are smartphones, jet travel, and AI, and we overlook the true apex of modern luxury: the long, hot, private, cheap shower. No king or emperor ever had it. The richest tyrants of history could summon armies but not hot water on demand in a private chamber. That miracle belongs to you.

Take it apart. Heat: your ancestors shivered through winters and feared cold as death’s herald; you step into a chamber where fire has been tamed and piped into water. Privacy: bathing was once communal, humiliating, or ritualized; now it is solitude at will, dignity preserved by tile and lock. Convenience: the richest pharaoh could not roll out of bed and be under cleansing warmth in thirty seconds, and you can — true luxury is not rare but repeatable, frictionless. Cheapness: you need not calculate whether one more minute under the water is decadent excess. The indulgence is guiltless because the infrastructure makes it scalable. Luxury, democratized.

You do not see the miles of pipe, the energy grid, the pumps and boilers, the generations of engineers and laborers who built this invisible empire for your convenience. All you see is water that obeys. That is what civilization is: miracles normalized into background noise. Lose the shower for a week — camping, a war zone, a broken heater — and you will remember what it was. Not hygiene: sanctuary.

This is wealth in the proper sense — accumulated capacity to satisfy human purposes, almost none of it appearing as anyone’s line item — and its invisibility is why the progress argument has to be made at all. Nobody feels grateful for the absence of cholera. The baseline resets within a generation, the miracle becomes plumbing, and the beneficiaries of the greatest enrichment in human history conclude that the system that produced it has failed them. Poverty is the default condition of mankind; wealth is the anomaly that needs explaining. So explain it.

The Rome Test

Here is a thought experiment that forces the explanation into the open. Suppose you could send a single document back to the Roman Empire in 27 B.C. — a codex of roughly a hundred pages, no smartphone, no magic, everything buildable with Roman materials, labor, and social structures. What is the smallest, most catalytic subset of modern knowledge that could trigger an industrial revolution nearly two thousand years early?

Working out what goes in the codex is a precise way of asking what causes progress, because every page spent on one thing is a page not spent on another. And the exercise delivers a sharp verdict: almost none of the budget goes to gadgets. It goes to four kinds of recursion.

First, and above everything, the method. Without a reliable way of generating and testing ideas, any single invention is brittle — a marvel that decays into ritual within a generation. The keystone pages teach the scientific method: truth comes from repeated, controlled observation, not from authority or tradition; experiments must be documented and reproducible; predictions that fail are discarded no matter who proposed them. Rome already had literate elites and superb engineers. What it lacked was the engine that turns tinkering into a systematic discovery process — the one technology that manufactures all the others.

Second, concentrated energy. Progress is, at bottom, a thermodynamic project: more work done per human muscle. The codex sketches the upgrade path from water wheels to heat engines — coke from coal, hotter furnaces, steel consistent enough for pressure vessels (a few pages suffice; the details are refinements the method will supply) — arriving at a low-pressure atmospheric steam engine that Roman metallurgy could actually build: boiler, piston, and beam, pumping water from mines, then a crank and flywheel to turn reciprocation into rotation, then line shafts driving mills, saws, and lathes. The lathes cut better engine parts, and the loop closes. Alongside it, a page of basic thermodynamics and a page of germ theory and sanitation — boil water, wash hands, separate sewage from drinking water, quarantine the contagious — because industrial growth requires a healthy, numerous workforce, and lower child mortality is the first dividend any civilization collects.

Third, information reproduction. A printing press — a screw press adapted to movable type, well within Roman carpentry — accelerates the spread of everything else. Mass-produced technical manuals, standardized weights and measures across the empire, results from Alexandria criticizing results from Gaul. Knowledge that cannot be cheaply copied dies with its keeper; knowledge that can be copied compounds.

Fourth, capital formation. The quietest pages matter most: double-entry accounting, the joint-stock company, an early patent system, broad literacy. Inventions do not fund, build, and deploy themselves. There must be institutions that let strangers pool resources for ventures none could attempt alone, ledgers that make ventures legible, and property rules that let an innovator capture enough of the value to bother. This is the machinery of turning wealth into capital, and without it the steam engine remains a temple curiosity — which is, roughly, what Hero of Alexandria’s actual steam toy remained.

The sections could amplify each other: method improves engines, engines power presses, presses spread method, and capital institutions fund the next round. They could also fail against Roman metallurgy, energy prices, slavery, political incentives, literacy, transport, or elite resistance. The exercise is a dependency map, not a historical forecast with a defensible two-century clock.

That is the answer to what causes progress. Not any single gadget, and not resources — Rome had resources, and so does every poor country today. Progress is a self-amplifying stack: a method for finding truth, concentrated energy to act on it, cheap reproduction of information, and institutions that form capital. The Industrial Revolution happened when those four finally coincided, and the child-mortality map turned from red to pale wherever the stack was allowed to operate. The hot shower is what the stack looks like from inside, two centuries in.

The Curve Ahead

A trend is not a law. Anyone can plot two centuries of rising life expectancy and rule a line onward; the question is what Credence the extrapolation deserves, given everything that could break it — nuclear war, engineered pandemics, severe climate disruption, unaligned artificial intelligence.

Existential risks provoke exaggerated responses in both directions: dismissal, because the catastrophes have never happened, and paralysis, because their consequences would be unbounded. The framework I defend in deciding under uncertainty — Effective Decision Theory — cuts between them: condition on scenarios with realistically estimable probabilities, and refuse to let unresolvable tail cases dominate the calculation. Applied here, the reference class is instructive. Civilization has absorbed world wars, pandemics, depressions, and totalitarian regimes, and the life-expectancy curve records each as a notch, not a reversal; the underlying stack — method, energy, information, capital — kept compounding through all of them, because knowledge, once cheaply copied, is very hard to unlearn.

As a dated personal forecast, I assign 75% Credence that global life expectancy at birth will be higher in 2051 than in 2026. The event needs a specified data series and revision rule; wars, pandemics, climate effects, aging, measurement changes, and unequal regional trends remain model uncertainty rather than a decorative residual. The number records my judgment and makes it falsifiable. It is not estimated by a published forecasting model.

Optimism of this kind is not a temperament; it is a calibration. The same discipline that licenses it also polices it, and the place to see the policing in action is the doomsday prediction that already ran to completion — the peak-oil panic, the next chapter’s field guide to how confident catastrophism goes wrong, and the questions that tell you when it might finally be right.