Field Notes

In 1871, a civil servant named Maejima Hisoka returned from London with an obsession: postal codes. The British system of numbered streets and sequential addresses seemed impossibly rational compared to Tokyo's labyrinth of unnamed alleys and landmark-based directions ('turn left at the willow tree, third house past the tofu shop'). Maejima proposed a radical solution: number everything.

The chome-ban-go system he helped design is elegant once you understand its logic. Tokyo was divided into ku (wards), then into chome (districts roughly 100-200 meters square), then into ban (building lots within each chome), and finally go (individual buildings). The key insight: buildings within each ban are numbered in order of construction, not position. This is why 2-15-1 might be next to 2-15-8 — the older building got the lower number.

What surprises most visitors is that this system works remarkably well for mail delivery. A postal worker knows that all addresses in 2-15-xxx are in the same visual cluster. The 'random' numbering actually groups buildings by lot, making them easier to find on foot than sequential numbering would. The problem isn't the system — it's that nobody explains it to outsiders.

Maejima went on to found Japan's postal service and what would become Nippon Telegraph and Telephone. His addressing scheme remains essentially unchanged 150 years later, a quiet triumph of urban logic hiding in plain sight on every blue address plate in the city.

Pick up a map at Tofuku-ji, Daitoku-ji, or any major Kyoto temple complex. Notice something missing? There is no scale bar. No '1:500' notation. No indication that 1 centimeter equals 10 meters. This is not an oversight — it is a fundamental statement about the nature of sacred space.

Buddhist temple gishiki-do (worship maps) are designed to guide a spiritual journey, not measure a physical one. The main hall appears largest because it is spiritually most significant, not because it occupies the most square meters. The distance between the sanmon (main gate) and the hondo (main hall) might be compressed or expanded based on ritual importance rather than metric accuracy.

This approach has a name in cartographic theory: topological mapping. Like a London Underground diagram, the gishiki-do prioritizes correct sequence and relational importance over spatial accuracy. You are meant to experience the spaces in order — gate, bell tower, garden, main hall, sub-shrine — not to navigate between them efficiently.

The absence of scale is also philosophical. Buddhist spatial theory holds that sacred space is not measurable in meters. The meditation garden at Ryogen-in occupies perhaps 30 square meters but represents the entire cosmos. To put a scale on such a map would be to reduce the infinite to the finite. The map does not depict space; it prescribes an experience of it.

In a workshop behind his family home in Shizuoka Prefecture, Tanaka Hiroshi, 71, still carves michi-shirube — the stone road markers that guided travelers along Japan's highways for four centuries. He is one of perhaps twelve people in the country who still practices the craft in the traditional manner, and he knows he may be the last.

The stones served as a navigation system before maps were widely available. A typical marker indicates the road name (often the name of the daimyo who funded it), the direction to the next post town, and the distance in ri (roughly 3.9 kilometers) or cho (about 109 meters). Some include the carver's name and the date — a stubborn insistence on provenance that modern road signs rarely match.

Tanaka's method has not changed. He selects granite from a specific quarry in Okazaki for its resistance to moss growth. He sketches the characters with charcoal, then carves with a sequence of chisels progressing from 12mm to 2mm. A single marker takes three weeks. He has carved 847 in his career, including replacements for historical markers damaged in the 2011 tsunami.

When asked why he continues, Tanaka points to a marker in his garden carved in 1847. 'That stone has guided farmers, samurai, tourists, and pilgrims. GPS coordinates change. Satellites fall. But this stone will be here, doing its job, in 200 years.' The last carver, it seems, is also the most patient futurist.

Harry Beck's 1931 London Underground diagram created a revolution in cartography: the topological map. Beck realized that subway passengers care about line connections and sequence, not geographic accuracy. He distorted distances, straightened curves, and aligned routes to 45-degree angles. The result was a map that lied about space but told the truth about navigation.

Tokyo's Metro diagram inherits this tradition and pushes it further. Shibuya and Omotesando are shown as adjacent stations on the Ginza Line. In reality, they are 1.2 kilometers apart — a 15-minute walk. But the diagram compresses them because the critical information is: 'you can get from one to the other without transferring.' Meanwhile, two stations that are physically close may appear far apart if they require a transfer.

This distortion is not deceptive; it is functional. A topological map encodes network connectivity, not spatial relationships. When you look at the Tokyo Metro diagram, you are reading a graph: nodes (stations) connected by edges (lines). The question 'how do I get there?' is answered perfectly. The question 'how far is it?' is deliberately suppressed as irrelevant.

The brilliance lies in what the map omits. It does not show surface streets, building footprints, or geographic features. It does not even show all train lines — JR lines are simplified, private lines may be absent entirely. This ruthless filtering is what makes the diagram readable. In a system with 13 lines and 285 stations, geographic accuracy would produce an incomprehensible tangle. The lie is the feature, not the bug.