You notice them before you can name them. A red globe the size of a car hanging in a temple gate. A row of small white ones strung along a canal in the evening. A squat stone shape sitting off to one side of a garden path, apparently doing nothing. In English they are all "Japanese lanterns," which is a category so wide it stops being useful the moment you want to say something specific about one.
Japanese does not use one word. It uses at least four, and the differences are not cosmetic — they are differences of material, of era, of what the object was for and who was meant to see it. Knowing which is which changes what a walk through Tokyo looks like, in the same way that a quiet trip to Tokyo through art and architecture changes when you can read a roofline.
The most useful reference in English is the Japanese Architecture and Art Net Users System entry for tourou, compiled from the vocabulary Japanese architectural historians actually use. It sets out the family in one place, and it begins with the origin: the earliest lanterns "were introduced to Japan from China through Korea along with Buddhism in the 6c."
The andon (行灯) is the one that did the ordinary work. JAANUS describes "standing oil lanterns" with "iron or wood frames," in "many different shapes and sizes which burned oil in shallow saucers suspended within a frame covered with paper," and notes that this type "became popular during the Edo period" and was used in private homes.
The light source is a shallow saucer of oil, not a candle, which places the andon inside a specific fuel economy. The Japan Lamp Museum traces it: a method for pressing oil from rapeseed was devised between the end of the Muromachi period and the beginning of Edo, and the museum writes that the spread of rapeseed oil prompted a move from a simple stand with an oil dish to purpose-built devices — the andon among them.
Both accounts agree on only two things: that it is carried by hand, and that the fire sits inside a translucent cover — paper in both, glass in JAANUS's later years, silk as an alternative in the dictionary. Past that, they part. JAANUS makes it hexagonal, with wood or metal frames covered with paper — glass in later years — and poles attached horizontally at the top for carrying. The Seisen-ban Nihon Kokugo Daijiten makes it a hand-held candlestick with a paper or silk cover, or a small andon on a handle and base, its fire-chamber hexagonal and narrowed at the bottom, or cylindrical.
Where they part is worth knowing. JAANUS defines bonbori (雪洞) as "portable lanterns," "distinctively hexagonal, with wood or metal frames covered with paper, or glass in later years," which "generally had poles attached horizontally to the top of the frame for ease of transportation."
The Seisen-ban Nihon Kokugo Daijiten describes something rather different: a hand-held candlestick with a paper or silk cover, or a small andon with a handle and a base, its fire-chamber hexagonal and narrowed at the bottom, or cylindrical. They do not agree on how it is held, or on whether it is always hexagonal.
For a visitor the consequence is small but useful: "bonbori" is a reasonable word for a small, softly glowing, paper-covered lantern, but its edges are genuinely blurred.
The chōchin (提灯) is the lantern the word "Japanese lantern" usually means in English, and the one an English shop listing will call a chochin lantern. JAANUS: paper lanterns "used outside the house and suspended from the eaves of buildings or carried in processions," where "the frame was a collapsible structure of thin bamboo strips covered with paper," with a candle inside, "often decorated with the names or logos of restaurants or inns."
The andon is an oil-lit piece of furniture; the chōchin is a candle-lit piece of luggage. It goes outdoors, it is carried, and it collapses — and that last property is not a convenience feature added later. It is the design.
The tōrō (灯籠) is the heavy one, and the oldest use. JAANUS records that stone lanterns, ishidourou, "were used first as votive lights at temples and shrines," later to light the grounds of those precincts, and that "secular use began in the 16c. when stone lanterns were used by tea masters for gardens surrounding their tea huts." There are, it notes, "about 9 major categories of stone lanterns based on general shapes and over 75 sub-categories."
There is also a metal branch of the family: hanging lanterns, tsuridourou, "usually of bronze or iron," hung from the corner eaves of palatial residences, temples and shrines. Remember that this metal branch exists; it returns later for a reason that has nothing to do with aesthetics.
Of the four, the chōchin is the one that became urban. A chōchin hangs outside izakaya, marks the entrance to shrines, gets carried in processions, and appears on the front of a temple gate at a size that stops traffic.
The paper is not incidental. Gifu — a name that recurs in lantern making — sits next to a papermaking region, and Gifu City's own account of the craft places the two side by side: the Mino area has long been a producer of excellent washi, and lanterns came to be made using "this thin, strong washi" together with good bamboo.
Thin and strong is the whole specification. The Agency for Cultural Affairs, registering the tools and products of Gifu lantern making as a Registered Tangible Folk Cultural Property, describes the object in the same terms: thin paper such as Mino washi pasted onto extremely fine bamboo ribs, finished with refined designs and patterns.
The chōchin's second job — after light — is to say whose place this is. JAANUS notes that they were "often decorated with the names or logos of restaurants or inns," and that habit never stopped. The red lantern outside a small drinking place in Tokyo is doing that same job: a light that is also a name.
That is why writing goes on a chōchin in the first place: a shop name, a logo, a district — a surface meant to be read at a distance, at night.
Start with what the firefighters carried. A lantern is not on the list.
The Tokyo Fire Department's own account of the Edo fire brigades explains that the method of the time was to pull down the surrounding houses so the fire could not spread, that a hand pump called a ryūdosui existed but could not throw water anything like a modern hose, and that the firefighters wore thick cloth half-coats, doused themselves to keep off the sparks, and "used tools such as the tobiguchi and the large saw to pull houses down." The matoi — the standard whirled on the roof of a house near the fire — was there "to let the people of the district know that they had arrived early and had been fighting the fire from early on."
Ryūdosui, tobiguchi, large saw, half-coat, matoi. No lantern.
What the lantern actually did is recorded, precisely, in the collection of the Fire Museum run by the Tokyo Fire Department. Its record for a replica of a late-Edo takahari-chōchin — a lantern raised on a pole — reads: in the Edo period, if the matoi and the banner were the daytime markers at a fire scene, the lantern was the night-time marker, showing the brigade name and the office of the man holding it. There were also yumihari lanterns.
So the lantern was not a tool for putting out fires. It was a tool for identification in the dark — which, at a night fire in a city built of wood and paper, with several brigades converging and the demolition of buildings underway, was not a decorative function. It was how you knew who was on the roof.
In Gifu the frame is a single continuous strip of bamboo wound in a spiral. Gifu City's description of the process lists the step by name — winding the rib along a groove cut into the forming mould, in a spiral — and the Gifu Chōchin Cooperative Association describes winding "an extremely fine bamboo rib in a spiral" onto the assembled mould before the thin washi or silk is pasted on without the slightest deviation.
The technique even has a name where it was formalised. Acros Fukuoka's account of Yame lanterns records that around the Ansei era, Yoshinaga Tahei established the foundations of the Yame chōchin with the ichijō rasen method — "single-strand spiral" — along with the introduction of thin handmade washi.
But it is not universal, and the exception is a famous one. Odawara City's own explanation of the Odawara chōchin says it directly: said to have been born in the mid-Edo period so that travellers on the Tōkaidō could carry it easily over the Hakone pass, it differs from ordinary lanterns in that "the inner ribs are independent rings," which is what lets it fold down small enough to tuck into the breast of a robe.
The practical use of knowing this: "one continuous rib" is a Gifu and Yame fact, not a chōchin fact. If someone tells you a lantern is inauthentic because its ribs are separate rings, they have described an Odawara lantern.
In Gifu the paper goes onto the spiral-wound ribs, and the association's phrase for how is the memorable part: it is pasted "without the slightest deviation." Gifu City's process notes add that the paper is applied to the ribs with one gap left between — which is to say not every rib takes paste.
Electric light did not replace the andon directly. The Japan Lamp Museum puts the intermediate step where it belongs: petroleum came into wide use as a lighting fuel in Japan only after kerosene lamps were imported from abroad around the end of the Edo period. Domestic production of those lamps began soon after, in forms adapted to Japanese ways of living. The museum describes the lamp — brought in with the Meiji-era opening to Western civilisation — spreading across the regions as the light that represented the era.
Electric light came separately, and later. The Japan Electric Association dates the first public lighting to 25 March 1878, at a ceremony for invited guests, and notes that ordinary citizens first saw electric light four years later, at Ginza in 1882. The National Museum of Nature and Science holds an improved model of the Brush arc lamp of the kind lit at Ginza, and describes such lamps as having astonished people as "the light of civilisation."
So the sequence is three steps, not two: rapeseed oil and the andon; then the imported kerosene lamp; then electricity.
One category was named on the wartime collection lists, and it is the metal branch flagged earlier. In 1941 the government issued the Metal Collection Ordinance — Imperial Ordinance No. 835 of 30 August 1941, in the National Diet Library's index of Japanese law, wholly revised in 1943 — requiring metal to be surrendered. The Kamakura study records that it was enacted under the National Mobilisation Law.
Its reach included religious institutions, and the documents are specific. A study published by the City of Kamakura transcribes the implementation schedule applied to temples: the scope covered "temples, churches, and religious corporations belonging to sects and denominations, including cemeteries under their management." Items were sorted into categories for immediate collection — temple bells and the metal fittings of the precinct — and a third category to be collected "so far as no direct hindrance is caused," which lists, by name, tōrō: lanterns, alongside small bells, gongs, lamp fittings and hanging lights. Even the minimum retained for ritual use, which included a pair of gold lanterns, was to be collected once substitutes were supplied.
First, this applied to metal — the ordinance names iron and copper, brass, bronze and other copper alloys. Stone lanterns were not the target, and a claim that "lanterns were requisitioned" without that qualifier is wrong. Second, the Kamakura study is an introduction to documents held by one temple, and its author notes that research on wartime metal collection is not extensive.
The lanterns most often called kasuga take their name from Kasuga Taisha in Nara, and the shrine's own account explains why the name carries weight: about three thousand lanterns stand in its precincts, offerings that began in the Heian period, many of them more than three centuries old, and the shrine states that the custom of lining a shrine or temple approach with lanterns is held to have begun there. Close to seventy per cent of all Muromachi-period lanterns in Japan, it adds, are at Kasuga.
The Japan Tourism Agency's English commentary counts the stone ones specifically — approximately 2,000 line the approach, and one famous lantern by the second torii came from 800 donors. Until the Edo period all of them were lit each evening, consuming 280 litres of oil a night. Today they are lit twice a year, in February and August.
The yukimi — "snow-viewing" — type is a form JAANUS singles out as unusual: a very broad canopy, a gently sloped jewel on top, and three, four or six legs, most often four, with no base. JAANUS calls it "an unusual type of lantern" whose form "is outstanding and it remains popular today," used primarily in residential gardens and tea gardens, with famous examples at Sennyūji, the Katsura Imperial Villa, the Sentō Gosho and the garden of Sanbōin at Daigoji.
There is a rule of placement, and it is narrow. JAANUS records that after the Edo period "it became popular to place this type of lantern near water," which "sets it off to its best advantage," and adds the difficulty: "the proper placement is difficult; if it is set too high or too low it will not be seen at its best."
Near water, and at a height that has to be judged.
What that gives you is a way of looking: a yukimi by water is not decoration placed for symmetry but an object positioned by eye, in relation to a reflective surface — the same order of judgement that governs the spirit of ikebana.
During Obon — the 13th to the 15th or 16th, in July or August depending on the district — the spirits of the dead are said to return. The Jōdo-shū school of Pure Land Buddhism describes the welcome: on the 13th, the first day, the spirits of those who have gone before are held to come back, and the method of meeting them is the mukaebi, the welcoming fire. It is lit at the gate or entrance, and the ancestors are said to come riding the rising smoke, taking the flame as their marker.
The lantern has its own place in this. The school notes that in some districts people go to the grave carrying a lantern to meet the spirits, and that bon-chōchin are set to the left and right of the spirit shelf "as a marker for the ancestral spirits." For the first Obon after the forty-ninth day, a white lantern may be hung under the eaves instead. The lantern is doing what it does elsewhere in this article: telling you who, and where.
The qualification is the school's own, and it is emphatic and repeated: practice differs greatly by region and by household. The account here is Jōdo-shū's.
The sending side of the week is where the lantern goes onto water. Jōdo-shū describes the practice: in some districts, the decorations and offerings are placed on a small wooden boat or wrapped in wild rice leaves and floated on a river to send the ancestors back — the rite called tōrō nagashi, also known as shōryō nagashi. Where the welcoming fire is a marker so the returning spirits do not lose their way, the sending fire is there to watch them go back.
The Japan Tourism Agency's English description of the Eiheiji floating lantern festival states the connection plainly: at the end of Obon, floating lanterns "are often released to guide the spirits back to the other world."
One practical note, from the school itself: in recent years a good number of municipalities have prohibited the practice on environmental grounds. Jōdo-shū advises confirming whether it is being held, and whether it is permitted, before doing it. If you are hoping to see one, that check is the first step, not an afterthought.
Namidabashi (泪橋) means "bridge of tears," and the reason given is grim. Arakawa City's own local history notes record that the bridge crossed the Omoi River, a branch of the Otonashi, linking what are now Minami-Senju 2 and 3 with present-day Taito ward, and that the name is said to come from the prisoners on their way to the Kozukappara execution ground, who wept as they crossed it.
The ward writes to iu — "so it is said."
The surrounding district was a working one. Arakawa City's note on the post town records that the area from the south side of the great bridge to what is now Kotsu-dori contained the towns of Kozukappara and Nakamura, known as the shimojuku or lower station, lined with wholesalers, shops and inns — and that Senju, taking the main and lower stations together, was the longest station street of any post town in Edo.
The ward records the post town itself: a long commercial street on a highway out of the city, lined with wholesalers, shops and inns. What a street of that kind needs, and what the ward does not comment on, is a way for each business to be identified after dark.
The name survives at the crossing. The neighbourhood around it is quiet, low, and largely unvisited by the people who queue at Asakusa, which makes it one of the more rewarding parts of a day trip into Tokyo's old shitamachi.
The ward's own record describes the district as a post town of wholesalers, shops and inns. What is here is a workshop, and the way to encounter one is not to walk past it.
Lanterns fall into two levels of access, and the difference between them is not the view.
The first level needs nothing but a walk. Of all the places to visit in Tokyo, the outer gate of Sensō-ji holds the lantern you will have seen in photographs, and the numbers are worth having. The temple's own page gives it as 3.9 metres high, 3.3 metres wide, about 700 kilograms, and records the most recent rehang: April 2020, the sixth.
The same page records the gate's own history: lanterns have been offered there since around its reconstruction in 1795. The present gate dates from 1960, donated by Matsushita Konosuke — president of Matsushita Electric, now Panasonic — personally.
What is not obvious from below is that this lantern collapses like any other. The Japan Tourism Agency's English commentary on the gate notes that during events like the Sanja Festival "the lantern is collapsed so that participants carrying mikoshi portable shrines can pass through more easily." Seven hundred kilograms of lantern, folded down on the same principle that lets any chōchin collapse.
The second level is the working floor of a lantern maker during a working day, and it exists because someone introduces you. Namidabashi Chōchin Ōshimaya, in Minami-Senju, is one of the makers we work with, and the access we arrange there is the workshop itself, not a demonstration laid on for visitors. We have been in it.
What that level gives you is not better viewing. It is a different question. In a museum or at a gate you ask what a lantern is. On a working floor the question becomes how a rib is wound here, and how long a single lantern takes a pair of hands, and what the person doing it thinks about while doing it.
The lantern hanging there now was not made in Tokyo. Panasonic's own record of the 2020 offering states that the lantern was commissioned from Takahashi Chōchin, in Kyoto, and took about a year to make. The company has carried out the restoration roughly once every ten years since 1960, at the temple's request.
The lantern at Kaminarimon is a bespoke commission from a named workshop, on a build cycle measured in years — which is the opposite of how the object reads from the pavement.
The replacement cycle has shortened, and the reason is recorded. Panasonic states that the 2020 rehang came only seven years after the previous one in 2013, and gives two causes: an increase in large typhoons in recent years meant the lantern was being collapsed more often, so deterioration progressed faster; and the timing was brought forward for the Olympic and Paralympic Games then scheduled for that year.
Only one of those two reasons is weather, but it is worth holding. A folding structure fails at its folds, and more storms mean more folding. Whether seven years becomes the new interval is not something anyone can say yet — Panasonic's own figure for the cycle is still "roughly once every ten years."
The same arithmetic runs through other crafts. Our piece on the world of the koto follows the same clock on the material side: Fukuyama City's own account of koto making gives a year for the paulownia to dry before the work begins, and says nearly every step after that is still done by hand. What continues is what somebody still needs for something. A lantern hung outside a bar in Minami-Senju tonight is not heritage. It is signage, and somebody still winds bamboo to make it.
Four words, and they are not interchangeable. Andon stands on the floor and burned rapeseed oil. Bonbori is the one whose definition genuinely varies. Chōchin folds, is carried, and expects to be read. Tōrō is stone or bronze, began as a votive light at temples and shrines, and moved into gardens in the sixteenth century.
One correction worth keeping. At an Edo fire, the matoi was the daytime marker and the lantern was the night-time one, showing the brigade and the rank of the person holding it. The chōchin was not a tool for putting out fires. It was how you knew who was on the roof in the dark — and every property of the object, from the collapse to the writable surface, reads as a requirement once you know that.
What you can now do is name what you are looking at.
And one place. Namidabashi is a short ride from Asakusa and has almost none of its crowd. The bridge is gone and the name is still on the crossing. Somewhere near it, this week, a person is winding bamboo onto a mould.