序章：千灯之城——汴京为何值得被重新讲述

第一章　千灯之城——汴京为何值得被重新讲述

一、一个被压扁的王朝

In the spring of 1127, when the Jurchen cavalry crossed the Yellow River and stormed the gates of Bianjing (汴京), the Northern Song capital burned for three days and three nights. The fall of Kaifeng — of a city that, by contemporary European standards, simply should not have existed at that scale — became one of those hinge moments around which Chinese historical memory has long organized itself. We remember 靖康之难 (the Jingkang Disaster) as a wound; we remember 文天祥 crossing zero south to die for a dynasty already dead; we remember 岳飞 and his 精忠报国 carved on the back. What we rarely remember, however, is the city that burned. We rarely ask: what kind of place was Bianjing that its loss could be felt, across East Asia, as a civilizational rupture rather than a routine dynastic transfer?

This book begins with a question that has bothered me for years, and that I suspect has bothered anyone who has stood in the wind on the old city walls of Kaifeng watching the Yellow River creep along its modern floodplain: Why has the Song dynasty (960–1279) — by almost any measurable index the most economically, technologically, and demographically advanced polity of the eleventh century — been compressed in our collective memory into a "weak" interlude between the glories of Tang and the traumas of the Mongol Yuan?

The standard Chinese high-school history textbook, in its heroic-narrative mode, treats the Song as a parenthesis. Tang is given to you as 盛唐 (High Tang) — Chang'an with its million souls, the cosmopolitan bazaar of foreign merchants, Li Bai drunk under the moon. The Yuan and Ming are given to you as 崖山之后无中华, or as Zheng He's treasure fleets, or as the Yongle Emperor moving the capital north. The Song sits in the middle, a dynasty whose territory was smaller than the Tang's, whose armies lost to the Khitan Liao and the Jurchen Jin, whose emperors were eventually kidnapped by nomads. The textbooks call it 积贫积弱 — "accumulated poverty and accumulated weakness." The phrase is so sticky that even scholars sometimes forget it was coined by a particular 19th-century reformer (王夫之 was an early user; the formulation hardened during the late Qing, when the lesson of the Song was being repurposed to warn the Qing) and treat it as a self-evident verdict of history.

But here is the inconvenient fact. In 1100, the territory controlled by the Northern Song housed roughly 100 million people — somewhere between a fifth and a quarter of the entire human species at that moment. Its capital, Bianjing (modern Kaifeng), was probably the largest city on earth, with credible estimates of 700,000 to over a million residents within the walls and another floating population in the suburbs. By comparison, contemporary London had perhaps 18,000 inhabitants; Paris, 50,000; even Constantinople, the heir to Rome, was smaller. The Song economy was the first in world history to recover its per-capita output after the Roman collapse and to grow it substantially; estimates by economic historians like Angus Maddison and more recently by Stephen Broadberry and B. Zorina Khan place Song China's GDP per capita at roughly the same level as the Netherlands in 1600 or England in 1700 — eight hundred years later.

So: why has this world-historical achievement been forgotten? Why does the average educated reader, Chinese or otherwise, know more about the boulevards of Baroque Rome than about the canals of Song Bianjing?

The short answer, which this book will unfold chapter by chapter, is that the Song does not fit the script. It is not a dynasty of conquest. It does not have a charismatic founder or a dramatic last stand. It does not fit the 夷夏东西 binary of pre-1937 nationalist historiography, nor the 阶级斗争 framework of post-1949 historiography, nor the 现代化 narrative that Westernizers like to project onto the Ming and Qing. To see the Song clearly, we need a different instrument. We need to look at it not as a state that fell, but as a system that worked — a complex adaptive system (复杂适应系统, CAS) of commerce, technology, culture, and institution that produced, for nearly three centuries, an urban civilization of startling density and creativity. And we need to begin not with the war memorials, but with the city itself. With the lamps.

---

二、量化奇迹：一份让历史学家沉默的数据清单

Before we step into the streets of Bianjing, we need to lay down the numbers. Numbers are not poetry; they are, however, the closest thing we have to a photograph of an economy that left no GDP statistics. The Song did have a remarkable bureaucratic culture of record-keeping — the surviving 宋会要辑稿 and 宋史 contain millions of characters of economic data — and modern scholars, beginning with the Japanese and now extended by a transnational community, have spent a century converting those records into the kind of quantitative series that make the Song's achievements visible.

Let me give you four sets of figures that, taken together, should make you pause.

2.1 人口：一亿、二十个百分点与一场静悄悄的革命

The most striking figure is also the most contested. 宋代人口峰值 (the peak Song population) is one of the great unsettled questions in global historical demography. The traditional 户口 figures from the Song administrative system range from 20 million 户 (households) in the early Northern Song to over 20 million 户 by 1100, with household sizes conventionally multiplied by five to give a population of around 100 million. Modern demographic historians — Cao Shuji (曹树基), Ge Jianxiong (葛剑雄), and the Anglo-American demographers who have engaged with their work — argue for a higher multiplier and a peak of perhaps 110 to 125 million around the turn of the 12th century. Either way, Song China at its height contained more human beings than all of Christian Europe combined, and probably more than the Abbasid Caliphate at its Abbasid peak.

The structural shift, however, is not the absolute number but the distribution. In Tang China, perhaps 80% of the population lived in rural villages. By the Song, the urbanization rate (城市化率) — the share of people living in cities of 10,000 or more — had crossed 20%, a level that England would not match until the 1850s and that the world as a whole would not match until the late 19th century. The Song had at least a dozen cities with populations over 100,000: Bianjing, Hangzhou, Chengdu, Fuzhou, Guangzhou, Jingdezhen, Yangzhou, Suzhou, Mingzhou (modern Ningbo), and so on. Jingdezhen alone, the porcelain capital, may have housed half a million people by 1100, almost all of them in some way connected to the ceramics industry.

The implication of these numbers is profound. A society in which one in five people lived in cities is not a peasant society with a few big towns. It is, in modern economic terms, a society that has crossed the Lewis turning point — the point at which labor has become scarce enough, and urban wages high enough, that agriculture can no longer feed the cities by simply exporting surplus grain. The Song, in other words, was the first large pre-modern economy to face the productivity problem that the Industrial Revolution would later solve. It solved it, in part, through the Champa rice (占城稻) imported from Vietnam, a fast-ripening strain that allowed double-cropping in the Yangtze valley and added perhaps 30% to total grain output. It solved it, in part, through the massive expansion of cash crops — tea, sugar, cotton, mulberry for silk — that drove the integration of regional markets. It solved it, in part, by moving the demographic center of gravity from the dry north to the wet south, an economic-geographic shift that, once made, was irreversible.

2.2 铁与煤：被提前了八百年的工业曲线

Now the numbers that, for me, are the most haunting. **In 1078, the iron output of the Song state-controlled sector alone is recorded at roughly 13,000 metric tons per year; adding private production, Robert Hartwell and his students estimated total output at perhaps 35,000 to 40,