The first paragraph of Kishi Fumikazu 岸文和's Edo no enkinhō: ukie no shikaku 江戸の遠近法—浮絵の視覚 ("Edo perspective: visions of uki-e"):

In the mid-Edo period, during the reign of the eight shogun Yoshimune, an unusual kind of ukiyo-e called uki-e 浮絵 became popular. What was unusual about them was that they were drawn using Western geometrical perspective, creating a sense of deep space that "appears to sink inwards towards the background" [むかふへくぼみて見ゆ] against which foreground figures seemed to float, which was quite astonishing to the average Edoite of the time. How astonishing? As astonishing as the news that a man had been killed by an elephant, apparently [...]

That's pretty astonishing! It turns out that what Kishi is referring to is a contemporary scandal sheet that had "Uki-e published" (浮絵出版行事) listed as a headline alongside "Man returned from desert island tells tale of eating bizarre bird" (無人島帰国者喰於異鳥物語之事) and, yes, "Elephant kills man" (象殺人事).

Interesting factoid: Kishi argues that the Western perspective techniques probably arrived via China, specifically Suzhou prints 蘇州版画 using the technique, rather than directly from the west.

Popularity factor: 13


What, I have to find the links to these awesome things myself? Grumble grumble...



I dunno about mid-Edo, but late Edo, the Dutch studies folks were trying their hands at perspectival paintings. It was like the Nintendo 3DS of the Sakoku era.

Leonardo Boiko:

Carl is right, they even used an analogue of the infamous 「◯◯発売!!」 copytext template that the videogame industry uses for every single announcement ever.


Hm, I wonder how Kishi's work and Screech's <i>The Lens within the Heart</i> compare. Since Screech's book is later as far as I can tell (although also available in Japanese), he should mention Kishi....


If Google Books search is to be believed, he quotes a monograph of Kishi's but not this book. Scandal!

language hat:

Talking about scandal, how come your comment box doesn't enable italics?


Mostly to shame those who use the "i" tag instead of "cite".


Bah, ageist! In my day, cite wasn't a standard tag supported by browsers! And semantically it only makes sense up there, not if I wanted to italicize for being a foreign word or for emphasis!

... There really is not tag for italicizing for being a foreign word, is there? Or have you established one on the stylesheet here, <cite>tono</cite>? (<em>...that feels so wrong</em>)


.... And it doesn't even work! Well sir I never!

Leonardo Boiko:

(X)HTML4 purists wanted everyone to stop using <b> and <i> and use <emph> and <cite> and such, to better separate semantics from styling. This was much criticized by typographers, designers, editors and other non-computer people, who pointed that italics and bold in typography have their own specific meaning that’s wider than simply “emphasis”. So the new HTML5 standard rehabilitates <i> and <b>, now with a “semantic” definition as “everything that’s traditionally typeset as italics (bold)” (e.g. ship names, foreign words &c). Quite a genius move if you ask me.

In the new framework I suppose <cite> and <emph> are subsets of <i>. The way to mark foreign text would probably be <i lang=ja>kono you ni</i> (quotes for attribute values are now optional in HTML5). This brings a problem with kanji+kana, that shouldn’t EVER be italicized. You could use the generic tag <span lang=ja>この様に</span>, or use CSS classes to make sure only rōmaji is italic.

Marking with lang= attributes has the potential of helping browsers spellcheck, screen-read, break lines, choose fonts (e.g. Chinese vs. Japanese–style), &c. Most current software ignore it though.


Oh, you can use "i" for non-citing italics, even romaji (although actually I have "span class=romaji" for that -- doubt it works in comments though).

I prefer "cite" for "i" when actually citing things because this blog probably has a higher-than-usual rate of actually citing works that I might want to look at again, and it's handy for searching. And maybe one day it will help some rogue AI learn about the Meiji calendar.

Leo: I remember you discussing this in the Language Log comments section, too. I'm not sure about how much of a genius move it really is, although I agree with the general conclusion reached at LL that it's hard and probably not worth it to try to parse out and separate all the possible reasons for italicizing something. Still, "cite" in particular feels very relevant and useful to me.

As for why none of them work in comments: because I wrote my comments system myself and couldn't be bothered handling escaping and so on properly...

language hat:

You are a bad person and should feel bad.

Leonardo Boiko:

Errata: it’s <em> not <emph>. And I agree, I use <cite> too (but was relieved to be free from <em> and <strong>).

My own personal style at the moment looks like this:

<i lang=ja>rōmaji</i>
<span lang=ja>漢字</span>
<span lang=ja class=r>rōman-tai no rōmaji</span>
<span lang=ja><i>nihongo</i> 日本語</span>

Then I CSS proper fonts for those and equivalently for Chinese hànzì and pīnyīn. A bunch of Emacs macros save typing.

Comment season is closed.