跳到主要內容

Across Taipei's Invisible Wall

Most of you will know that I recently moved, from Yuquan to Guangfu. A few streets, from one Taipei into another.


For years I lived in Yuquan (玉泉里), on the southern edge of Dadaocheng (大稻埕). I now live in Guangfu (光復里), inside the old walls of Taipei City (臺北城). No river crossed, no hill climbed: only a few streets east.

Yuquan is warehouses, hardware, spare parts, wholesalers and old family firms. Its streets still obey the Tamsui River (淡水河): goods arrive, change hands, acquire credit and move on. Guangfu is gates, government offices, public markets, hotels, bookshops, banks, schools and pork-chop rice served to people whose lunch break has an ending.

Both belong to western Taipei (臺北西區); both are sold as "the old city". They are not old in the same way. Yuquan grew around goods and private capital; Guangfu around files, public authority and salaries.

The map shows a short walk. The city shows a river port and a city gate, merchants and mandarins, inherited property and the state payroll. One of the shortest border crossings I know.

Ian Nairn distrusted the flattening effects of modern planning. He defended the contradictions that give a place its charge. Western Taipei would have pleased him. Its disorder is not a failure to become one city. It is evidence that several cities, built by rival systems of power, still occupy the same ground.

Names that remember, names that command

The distinction begins with the names.

"Yuquan" suggests a jade spring. Its origin is less romantic: Izumicho (泉町), or "Spring District", imposed during the Japanese colonial administration's renaming of Taipei in 1922.

Before 1922, the names described the ground. Hegoutou Street (河溝頭街) marked the head of a drainage channel; Ximenwai Street (西門外街), the space beyond the West Gate; Jiangbin Street (江瀕街), the road beside the river. Mud and river water remained inside the words.

Izumicho flattened those local biographies beneath the colonial surveyor's plan: several places, one neat Japanese administrative name.

After 1945, power changed hands. The Nationalist government abolished the Japanese district system, erased one colonial vocabulary and installed the auspicious language of Chinese tradition. It kept "spring" and placed "jade" before it. Izumicho became Yuquan - Jade Spring.

The boundaries were redrawn repeatedly. Today's Yuquan contains fragments of Minatocho (港町), Eirakucho (永樂町) and Taiheicho (太平町). It is not Izumicho preserved intact. Yet however often the city has been cut apart and reassembled, the spring remains on the map: a geographical ghost, a final act of defiance.

Guangfu follows another logic, and another arrogance of power. The word means "restoration" or "recovery", the Nationalist term for the end of Japanese rule. It describes no river, street, trade or position. It announces how the state has chosen to interpret history.

After 1945, Restoration, Nation-Building, People and China appeared on roads, schools and neighbourhoods across Taiwan. Their political value lay in their local emptiness: they could be placed anywhere because they belonged nowhere.

The mechanism is familiar beyond Taiwan. Regimes rename streets because possession of the map rehearses possession of memory. Guangfu's bland correctness conceals unusual density. It covers fragments of Kitamoncho (北門町), Kyomachi (京町), Yamatocho (大和町), Sakaemachi (榮町) and Nogicho (乃木町). Within a few blocks stand the North Gate (北門, formally Cheng'en Gate 承恩門), Futai Street Mansion (撫臺街洋樓) and Zhongshan Hall (中山堂), formerly the Taipei Public Auditorium (臺北公會堂). Qing fortification, Japanese empire, post-war authoritarianism, commerce and transport speak at once.

The name supplies one official answer. The buildings continue to ask questions.

The losers who inherited the port

Dadaocheng began not with a plan, but a commercial feud and a violent defeat.

In 1853, settlers from Fujian fought for control of the river trade in Bangka (艋舺), now Wanhua (萬華). The "upper guilds", associated with three counties of Quanzhou (泉州府), attacked merchants from Tong'an (同安), the "lower guilds", after passing through the Qingshui temple precinct (清水巖祖師廟). The Tong'an side lost. Its leader, Lin Youzao, fled north with his people and the statue of the Xiahai City God (霞海城隍). They settled on an open threshing ground called Dadaocheng and built another port.

The winners secured Bangka. The losers secured the future.

As Bangka's river silted up, larger vessels struggled to berth. Dadaocheng was better placed when the port of Tamsui (淡水港) opened to foreign trade in 1860. Tea, camphor, textiles, sugar, preserved foods and imports moved through its quays. British and American firms established trading houses along Guide Street (貴德街) and today's Nanjing West Road (南京西路). Foreign commerce and consular activity drew in compradors - brokers between foreign firms and Taiwanese trade - together with tea merchants, warehouse owners, boatmen, accountants and labourers.

This was more than a market. It was a system. Cargo required storage; storage required land; long-distance trade required credit, information and trusted intermediaries. Profit moved from shipment to shop, from shop to property.

The colonial Railway Department (鐵道部), station and rail network later redirected movement from water to land without immediately displacing the river. For a time, wharf, railway, gate and market worked together. Dadaocheng's southern edge became a hinge between transport regimes.

In Yuquan, goods did more than move. They learnt to reproduce themselves as capital.

When a trading house outlives the trader

The Zhuang family makes the process visible.

Its ancestors arrived from Tong'an in 1746 and settled at Bali (八里坌), across the Tamsui. After the fighting of 1853, the family established Yifang Trading House (義芳行) in Dadaocheng, trading with Hong Kong and Xiamen in sugar, Eagle Brand condensed milk, Kirin beer, olives, flour and other imports.

In 1921, the business became Zhuang Yifang Trading Company (莊義芳商事株式會社), adding property sales and rentals. In 1932 it helped establish the Eirakucho Shop Building Credit and Purchasing Co-operative (永樂町店舖建築信用購買利用組合), an early attempt to remake Taipei's commercial fabric. Three years later the family joined tea merchant Chen Tianlai to finance the First Theatre (第一劇場), one of only two major theatres operated by Taiwanese proprietors.

The business survives, transformed, as the Lung Yuan Group (隆遠集團). Eight generations after the family's arrival, its holdings extend across metropolitan Taipei.

A cargo of tea yields one profit. A trading house yields profits repeatedly. Land, rent and shares continue after the person who struck the bargain is dead. Wealth detaches itself from one body and acquires a life of its own.

Dadaocheng was never uniformly rich. A capitalist city requires clerks, porters, tea pickers, bookkeepers, artisans and servants. A few families controlled goods, land and credit; many supplied labour. Merchant houses and poverty did not contradict one another. Each made the other possible.

The street shows both. Yihe Tai trading house (怡和泰商行), with yellow-green tiles, arched windows and floral medallions, turned its facade into a declaration of solvency. Chen Tianlai's residence (陳天來故居) became a Japanese model of the Taiwanese home. Beneath such facades, the arcade remained a workplace: boxes against columns, handcarts and lorries competing for space, merchandise spilling into the street.

Yuquan retains this structure. Hardware, spare parts, wholesalers, warehouses and old shop-houses still operate. The street is not arranged for the camera. Doors must open, lorries unload, rent be paid. Heritage aesthetics come later.

This history is not under museum glass. It sweats, stains the pavement and double-parks.

Capital beyond money

Dadaocheng's greater achievement was not private fortune, but a public sphere financed by it.

Trade carried information with tea. It supported literacy, newspapers, doctors, lawyers and local notables. Li Chunsheng, a comprador who made his fortune through foreign firms, expanded the tea trade, assisted the reforming governor Liu Mingchuan and mediated religious and commercial disputes as a Christian in a largely non-Christian society.

In the 1920s, Chiang Wei-shui opened Taian Hospital (大安醫院), later an important base for the Taiwanese Cultural Association (臺灣文化協會). Reading rooms, lectures, newspapers and modern theatre required ideas; they also required halls, printing presses and networks. Merchants and intellectuals met at Shanshuiting restaurant (山水亭), the First Theatre and the district's cafes.

Capitalism produced brutal inequality. It also paid part of the cost of civic culture and political opposition. The port that connected Taiwanese tea to the world gave Taiwanese people the means to argue about their place in it.

Through the North Gate

Walking east from Yuquan, the real boundary is not the line between Datong (大同區) and Zhongzheng (中正區). It is the North Gate.

When the Qing completed Taipei's walls in the late nineteenth century, the gate divided two urban orders. Outside: Dadaocheng, river, wharves, private trade. Inside: governor's offices, provincial administration, barracks, state.

In 1892, Governor Shao Youlian spent 20,000 silver taels on the Qinchai Xingtai (欽差行臺), a reception compound for imperial commissioners. Red tiles and swallowtail roofs made political theatre from timber and brick. Japanese troops entered through the North Gate in 1895; the compound became the first headquarters of the colonial Governor-General, serving until the new Governor-General's Office - today's Presidential Office Building (總統府) - was completed in 1919.

The Japanese administration demolished most of the walls and replaced them with broad boulevards. Railways and the station changed Taipei's scale and direction. Today the surviving gate takes seconds to cross. Inside and outside have merged into asphalt. Their institutions have not.

Futai Street Mansion combines local Qilian sandstone (唭哩岸石), an upper timber storey and a Mansard roof cut by dormer windows. Its classical arcade and European Renaissance references belong to a city staging imperial modernity.

Nearby Sakaemachi, around Hengyang Road (衡陽路), was the "Ginza of Taipei". Restaurants, cafes and fashionable shops served officials, professionals and businessmen. In 1932, Eiji Shigeta opened the seven-storey Kikumoto Department Store (菊元百貨店), known to Taiwanese customers as the "seven-storey heaven". It had Taiwan's first publicly accessible commercial passenger lift and was, for a time, the island's tallest privately owned building. The Taipei Public Auditorium followed in 1936, giving the colonial capital a monumental civic stage.

This Taipei was prosperous and modern. The question is not whether it had money, but where the money came from.

The city of salaries

Chengzhong (城中區), the old centre containing Guangfu, prospered through concentrated authority, public budgets, banks, infrastructure and salaried elites. Officials drew stipends; police and soldiers received pay; professionals depended on institutional positions. Department stores, hotels, bookshops, foreign firms and cafes served a population whose income arrived at regular intervals.

Dadaocheng belonged to capital. Chengzhong belonged to salary.

The distinction is structural, not demographic. Dadaocheng was never uniformly rich, nor Chengzhong uniformly poor. The difference is the source and durability of wealth. Capital makes land, companies, rent and dividends work in the owner's place. Salary, however generous, still requires the sale of time and expertise.

A title cannot collect rent. A memorandum does not appreciate for the benefit of one's children.

For people without a family business, shop-house or land, the civil service converted education into office, office into status, status into reliable cash flow. A clear staircase of social mobility, with an equally clear ceiling. Dignity, security and pension were attainable; wealth that continued working after its owner stopped was harder.

Family money changed the equation. For the already wealthy, office supplied status, respectability, networks and access to power. Two officials at the same grade, on the same salary, could belong to different classes: one with an inherited flat in Xinyi (信義區), family land in Da'an (大安區), a housing deposit supplied by parents or an existing investment portfolio; the other with salary alone.

The class divide hides behind an equal payslip.

Power and pay are also unequal. A mid-ranking official in Guangfu may determine whether a plot in Yuquan can be developed, when a valuable licence is issued, who wins a public contract or how costly an inspection becomes. The economic consequence may be immense; the legal income remains confined to a salary scale.

This does not excuse corruption. It explains the need for transparent procedure, credible oversight and pay proportionate to responsibility. Administrative power acquires a price when it alters another person's costs and profits. Unless law prevents that price from becoming private income, it returns as patronage, influence-peddling or bribery.

The buildings make authority monumental. The person exercising it remains exposed to inflation, school fees and housing costs. Power without capital is not wealth. Unless there is family money behind the salary.

That mismatch is Guangfu's central tension.

Private dining, pork-chop rice

Food makes the distinction visible.

Dadaocheng offers reservation-only dining, elaborate tea services, mullet roe and sweets sold with history attached. Ziyang (滋養和菓子), a post-war maker of Japanese-style wagashi that later moved into Dadaocheng, is not a surviving Qing merchant house. It fits the district because craft, memory and place can be converted into distinction, then price.

Wealthy merchants always needed banquets and places where hospitality could soften business. The audience and vocabulary have changed. Old shop-houses are now stages; provenance, technique and family memory justify the premium. Capital extracts another return from the past.

Guangfu's pork-chop rice, hand-cut noodles, breakfast shops and inexpensive clothing serve another economy. Civil servants, office workers, students, travellers and market traders need a known price, adequate portion and predictable time. The meal is not culturally inferior. It is designed for salaried time.

Dadaocheng sells time arranged as an experience. Chengzhong sells a meal that will not delay the next obligation. One turns history into margin; the other restores labour for tomorrow.

History: premium or repair bill

Dadaocheng's revival is described through creative industries, tourism and conservation. It was possible because the district already possessed assets that could be repriced: shop-houses, land, trading names, tea, temples and family histories.

Capital can wait. A building remains neglected until the market learns to desire "old Taipei", then returns as restaurant, gallery or boutique. Decay becomes scarcity; memory becomes brand.

Planning law helps. In Dadaocheng's designated historic district, transfer of development rights moves unused building potential from protected sites to other plots. Conservation carries costs, but can also create tradable value. Continuous street fronts, surviving trades and an established tourist route make the past legible, then saleable.

Guangfu's past is harder to sell. It remains inside offices, hotels, markets and buildings with ordinary work to do. The North Gate, Futai Street Mansion and Zhongshan Hall bring protection, height limits and development constraints. Ownership is fragmented; repairs expensive; assembly difficult. Daily operations cannot stop while someone invents a heritage narrative.

The same old building sits on two balance sheets. In Dadaocheng, age adds a premium. In Chengzhong, age adds maintenance. One district makes history visible; the other keeps the system running.

Tax data prevent rich merchants and poor officials from becoming a cartoon. In Yuquan and neighbouring Yongle (永樂里), average declared household income stands far above the median: a small number of asset owners lift the mean. Guangfu is even more dispersed, combining high professional incomes, rent, business holdings and post-war property. A bureaucratic city does not mean every resident lives on one government salary.

The numbers refine the distinction. Class is not one year's average income. It is whether income can detach itself from continued labour, and whether wealth can pass intact to the next generation.

The state changes language, not address

Japan's defeat in 1945 did not reverse the old city's political character. It replaced the occupants.

The Governor-General's Office became the Presidential Office. The governor's residence became Taipei Guest House (臺北賓館). Japanese official housing passed to the Republic of China's military and political elite. The former naval officers' residences (海軍將官官舍) later housed senior commanders including Ma Jizhuang and Li Yuxi.

Across Taipei, the state created military dependants' settlements for the population arriving with the Nationalist government. Beitou's Heart Village (中心新村) reused Japanese military medical facilities. Chenggong New Village (成功新村), built by the army, later became modern public housing. Allocation followed bureaucratic rank and institutional affiliation, not the market.

Chengzhong continued to depend on ministries, schools, postal administration, transport agencies and state-linked finance. Japanese colonial officials gave way to civil servants speaking Mandarin in different provincial accents. Kikumoto became the China Products Company (中華國貨公司), then housed the Friends of the Armed Forces Association (軍人之友總社). The signs changed. The salaried city endured.

The state never left the old centre. It changed uniform.

Across the invisible wall

From the North Gate, I can look west to Yuquan and east to Guangfu.

West: lorry exhaust, machine oil, goods, risk, accumulated private wealth. East: office workers and bank employees, one form, one meeting, one lunch break at a time.

Jonathan Meades wrote that Nairn treated architecture as "deflected autobiography": an argument with oneself through buildings. These streets now read that way to me. Not an abstract urban history, but a map of where I have lived, what I inherited and which forms of security I recognise.

The walk from Yuquan to Guangfu crosses no physical barrier. It crosses two philosophies of survival. Dadaocheng turns disorder into opportunity, profit and inheritance. Chengzhong turns education, discipline and institutional service into status and security.

The systems need each other. Capital needs permits, infrastructure and law. The state needs taxes, commerce and salaried obedience. Each distrusts the other's claim to legitimacy: the merchant sees an official controlling value without creating it; the official sees a merchant converting private advantage into hereditary power.

A few streets apart, they face each other across a city wall that no longer exists and never entirely disappeared.


留言

這個網誌中的熱門文章

當法律跑在能力前面,執法就變成賭局

When law outruns capability, enforcement becomes a gamble. 歐盟最新提出的「數位綜合方案」,將原本預計自 2026 年起陸續落地的 AI Act 高風險義務整體延後,並對 GDPR 的若干適用標準作出調整。這一系列時間表與技術條文的修改;從法制運作角度看,其實是歐盟試圖重新修正過去幾年高度前傾的監管節奏。核心訊息很直接: 當規範在行政能力、標準體系與產業準備都尚未到位時提前生效,制度本身便會成為新的風險來源,同時對 法律確定性(Rechtssicherheit)與可執行性(Vollzugstauglichkeit)產生損害。 過去十年,歐盟在數位領域採取的是一套高度主動的立法模式:先以框架性規範設定邊界,再透過技術標準、指引與執法實務慢慢填補細節。GDPR、DMA、DSA 乃至 AI Act 無不如此。這樣的作法在政治上具有明顯的宣示效果,也強化了歐盟作為「規則輸出者」的角色。但在 AI 與資料治理領域,這種先立架構、後補能力的路線,逐步暴露出其結構性限制: Regelungsdichte(規範密度)可以很高,Vollzugskapazität(實際執行能力)卻未必能跟上。 AI Act 的高風險義務便是一個典型例子。法條要求涵蓋技術文件完整性、訓練資料可追溯性、模型行為監測機制、風險管理流程等多重層面,每一項都假設存在一套成熟的標準體系與行政審查機制。然而,相關技術標準仍在制定過程中,各國主管機關的準備度明顯不一,企業端也尚未形成穩定的 best practice。在這樣的條件下,法規若在原定時程強行生效,實務上極易出現「義務已存在,但合格標準與審查方式未臻明確」的狀態。 對企業而言,這意味著法規遵循被迫建立在猜測之上:不知道做到何種程度才足以被認定為合規,卻必須提前調整內部結構與資源配置。對主管機關而言,則是在執法時缺乏穩定的判準,不同成員國之間的差異難以避免。這種情形直接侵蝕了 Rechtssicherheit,使法律本身成為一種額外的不確定性,而不是降低不確定性的工具。從這個角度看,延後義務並非削弱監管,而是試圖讓規範重新落在與現實能力大致相稱的水位,回到 Verhältnismäßigkeit(比例原則)可接受的範圍之內。 GDPR 的調整呈現出相同的邏輯,只是焦點從 AI 行為,轉移到資...

The price of waiting: what Taiwan’s AI law reveals about regulatory uncertainty

As Taiwan advances its proposed Artificial Intelligence Basic Act, the debate has largely focused on familiar themes: ethics, principles, and the need for “responsible AI”. These questions matter. But they are not the most consequential ones. The more important issue is economic rather than moral. It concerns how law structures expectations, how uncertainty is distributed, and how delay becomes a rational response when judgement is deferred. Taiwan’s AI legislation offers a revealing case study in the political economy of regulatory uncertainty — and in the costs of asking markets to decide first. At a symbolic level, the Basic Act marks a clear shift. Artificial intelligence is no longer treated merely as a technical input or an industrial productivity tool, but as an object of public governance. Its deployment is recognised as having implications for legal responsibility, administrative authority and decision-making frameworks. Yet symbol and structure are not the same. The law’s...

REGULATE ONLY WHAT YOU CAN ENFORCE

The EU’s Digital Omnibus signals a critical pivot: admitting that premature regulation endangers the very certainty it seeks to create. The European Union’s introduction of the "Digital Omnibus" package—which proposes delaying key high-risk obligations under the AI Act and recalibrating GDPR standards—is more than a mere adjustment of timetables. From the perspective of legal operations, it represents a structural correction to a regulatory rhythm that has become dangerously front-loaded. The core message from Brussels is blunt: when regulations enter into force before administrative capacity, technical standards, and industrial readiness are established, the institution itself becomes a source of risk. Instead of fostering order, premature regulation simultaneously damages Legal Certainty ( Rechtssicherheit ) and Executability ( Vollzugstauglichkeit ) . The Structural Deficit: Density vs. Capacity For the past decade, the EU has pursued a hyper-active legislative model: set...