In the course of my research, I have begun to come across highly detailed references to which time of day events took place. I don’t know yet whether I will end up having anything to say about the ritual, political, or social significance of the fact that such-and-such an event took place at such-and-such a time, but at least on a basic logistical level, just trying to visualize whether people were going up to the castle in the morning or in the evening, how long an audience with the shogun lasted, and so on, so long as my sources are telling me what time it was, I might as well make a note of it, and make an effort to understand what time that was. Over the years, I’ve skimmed over quite a few different explanations or guides to telling time in early modern Japan. Perhaps one of the best online is that posted by the late Anthony Bryant on his website, SengokuDaimyo.com. Even so, I never quite managed to grasp it, until this week.
Above: A Japanese clock from 1678, on display at the British Museum. Not quite as impressive as this other one also at the BM, but I think the face and mechanics are a bit more visible here, because of the size. Photo my own. Apologies for the graininess that emerged as I brightened the photo.
I think a large part of what makes it so difficult to grasp is simply because our own timekeeping culture is something we learn from such a young age, and use so ubiquitously in our lives, it is so deeply ingrained, that it seems almost natural, and so it can be hard to conceive of other systems. Of course, the fact that there were several different ways of telling time in early modern Japan (by numbers that don’t correspond to our 12pm, 1pm, 2pm system, and by zodiac symbols which I have never managed to memorize the order of), and that Japanese “hours” shortened and lengthened with the seasons, doesn’t exactly help.
Thanks to Japanese blogger Chihuahua Luke for this diagram.
The day was divided into six koku (刻 – though often referred to as “hours” 時 or 時分 in the documents), and the night another six koku, for a total of 12 koku corresponding to each of our 24-hour days. You can see on the above diagram, the six “hours” of night on the top half, and the six “hours” of day on the bottom half. Midnight is at the very top, and noon at the bottom, with sunset at the left and sunrise at the right.
So, since there are 12 koku in each day+night cycle, each is roughly equivalent to two hours in our modern 24-hour reckoning. Kind of. The thing is, daylight was always six koku long, and night was another six koku. So, depending on the seasons, as daylight grew longer and shorter, so too did the koku. As this diagram below shows, in winter, there is less daylight each day, so the daytime (昼) koku are shorter, and the nighttime (夜) koku longer. In summer, this is reversed. When mechanical clocks were first introduced to Japan by Europeans in the 16th or 17th century, their mechanisms – designed in Europe to tell regular time, one hour per hour, 24 hours per day, like clockwork (literally) – had to be modified to allow for these shifts in the “hours” (or koku) with the seasons. Basically, the small weights which drove the clockwork (and which you can see under the bell on the image at the top of this post) had to be adjusted every day, or every few days, to accommodate the days growing longer or shorter. If you’re interested in further details on how these clocks worked, wristwatch company Seiko has a nice description on their website.
Another diagram from Chihuahua Luke. Thank you! This one shows how daylight hours shifted across the year. The small 1-24 numbers on top and bottom are our modern hours, while the numbers given in kanji are the bell system I describe below. You can see on top how in summer, with sunrise around 4am and sunset around 7pm, the six daylight hours (from 明け六ツ to 暮れ六ツ) were lengthened. And the reverse in the winter, shown on the bottom.
Still with me? It gets a little more complicated. If you read Edo period documents, or look at Edo period clocks, you won’t see the hours identified in a simple progression from one to six, or one to twelve. Nanatsu-toki 七つ時 or nanatsu-jibun 七つ時分, which we might call “7 koku” is not the seventh one of the day, and it does not come after six. Rather, each koku was assigned to one of the twelve “zodiac” animals, progressing from Hour of the Hare at dawn, to Hour of the Horse at noon, Hour of the Cock at dusk, and Hour of the Rat at midnight. These “animal” names for the hours can be seen in numerous sources, including in Utamaro’s ukiyo-e woodblock print series “Twelve Hours in the Yoshiwara” – twelve prints depicting courtesans at various hours of their day. The print for the Hour of the Hare shows a courtesan presenting her client with his jacket, as it is dawn and it is time for him to go.
The time was also announced in the big cities by networks of belltowers, which rang nine bells at noon or midnight, progressing down to eight, seven, then six bells at dawn or dusk, then five, and four, before jumping back up to nine. I have pasted a copy of a chart of this up on the wall by my desk, and have been consulting it frantically, as I was just a little too overwhelmed with the complexity, was having a really hard time remembering which numbers corresponded to which time of day, and just didn’t think I was going to be able to memorize it. As I made my way through my sources, I took meticulous notes of the corresponding times – for example, where the source says 七ツ時 (7 bells), I wrote “3-5am,” as it says directly on my chart.
But, then my advisor reminded me that it really doesn’t correspond directly to our regular hours; rather, it shifted over the course of the seasons. (EDIT, 3/13: Besides, let us not forget the idiosyncrasies of our own system, which includes setting our clocks forward or back by an hour each spring and autumn.) Oy gevalt. But, complicated as this all is, I had a sort of “aha!” moment today, and realized two things, which spurred me to be writing this post.
First, no one had wristwatches or anything like that at the time, and in an age before railroad timetables, very few things were done strictly according to schedule (i.e. directly on-time). So, really, it’s the rough time relative to dawn or dusk, or relative to noon or midnight, that is perhaps most relevant – and this gives us a stronger sense of the actual look/feel of the day. 七ツ時 (7 bells) is shortly before dawn, so that means it’s dark out. People would have put out paper lanterns to help light the way; these will be extinguished right around dawn. Are people up yet? Are they milling about? Are they just sort of first starting their day, starting to get things ready? The source tells us it was snowing that day… So, I think I may simply change all my references to “3-5am” to instead read something like “shortly before sunrise.” While this is vaguer, it is also less inaccurate, and arguably perhaps more directly indicative of the time of day relative to dawn, dusk, etc.
Second, while I do think I’ll be leaving the chart up for reference, I think once you manage to learn/remember that six bells is always sunrise or sunset, that nine bells is always noon or midnight, and that the bells count down from nine to four, and then jump to nine again, everything else falls into place. Five bells (五つ時) is the early morning or the early evening, four bells (四ツの時分) is late morning or late evening, and then we jump back up to nine bells for the time around noon, or midnight. Eight bells is either early afternoon, or very early morning (i.e. the hours after midnight), seven bells is either late afternoon (approaching dusk) or the hours approaching dawn. And that’s actually about it.
People milling about, possibly getting their day started? Or, perhaps it’s closer to sunset, and closing time? A model of the Echigo-ya, one of Edo’s most major department stores, at the Edo-Tokyo Museum. Photo my own.
It was quite common for people in early modern Japan to rise during the period of seven bells (七つ時), that is, within the last koku before dawn (the last hour or two before sunrise in our modern conception), to begin to get ready for the day. While on the road, we find that Ryukyuan missions very often departed a town around dawn (thus implying they’d already been awake for a bit, to pack up and prepare for departure), and arrived places by around dusk. Still, there were many occasions when they arrived considerably after dusk, and fure were circulated around the town ordering that homeowners & shopowners put up paper lanterns (chôchin), taking the lanterns down at dawn.
When traveling up to Edo castle for formal audiences, the missions generally got prepared around 8 bells (that is, two koku before dawn) – as, one supposes, there were a lot of preparations to be done – and then departed the mansion for the castle shortly before sunrise, arriving at the castle after daybreak (6 bells). It’s certainly something to think about, that they would have been marching through the streets, in their colorful costumes, banners, palanquins, and everything, and blasting street processional music, at such an early hour – and in the faint light of dawn. One supposes the popular crowds came out more when the missions came back down from the castle later in the day? But, then again, we should not presuppose based on modern-day conceptions of what feels too early in the day according to our own modern lifestyles…
As for how time was actually calculated in order to know when to ring the bells, I’m not actually sure. But, both for individuals and institutions (e.g. castles, temples), there were a number of other ways in which time was counted as well. Perhaps one of the most obvious is to simply look at the sun – I haven’t actually read up on it, but I wouldn’t be surprised if this was the most common method out in the countryside. Shuri castle maintained a water clock – water was allowed to drain out of a large tank in a controlled manner, such that the level of water could be used to tell how much time had passed. This was used in combination with a sundial to tell the time, which was then announced to the castle and the city by drums. Though this is a Ryukyuan (Okinawan) example, I wouldn’t be surprised if something like it were used in Japan as well. So, various kinds of water clocks and sundials. Candles could also be burnt to tell the time – just keep track of how far down the candle has burned, or how many candles you’ve gone through. In the Yoshiwara, a client’s time with a courtesan was measured based on how many incense sticks had been burned, and he was charged on that basis.
For more on timekeeping in Edo period Japan, check out Dissertation Reviews’ review of Yulia Frumer’s recent PhD dissertation, “Clocks and Time in Edo Japan.” The dissertation itself is sadly embargoed until November 2016. Hopefully Frumer will be getting her work published as a book in the near future; I’ll be looking out for it.