The other day I was writing the characters 忐忑 tan3te4, restless & fidgety, and they brought to mind the 上下 shang4xia4, 乒乓 ping1pang1 (pingpong) and 孑孓 jie2jue2 (mosquito larvae) pairs. But I couldn’t recall any more.
I’m wondering whether there are any other such graphs out there which were clearly created as a matching pair, especially with one portion of the graph inverted like these. Anyone?
Does this count? 彳亍
I believe this one is somewhat contrived, though: 查某
Erm… if you type “58008” into your calculator and turn it upside down, it looks like “BOOBS”.
Those are interesting. Well, the thread might be interesting even if people post pairs that aren’t exactly what I’m looking for.
I’m interested in pairs that are really used as characters, rather than graphic extractions (like 彳 and 亍) which never really had an independent existence except perhaps as bu4shou3 or when borrowed as phonetic symbols. In other words, 彳 and 亍 are TTBOMU graphic extractions, not characters. But that’s definitely in the right direction! 
左右 wouldn’t count either, since they weren’t created as a pair, and didn’t originally mean left and right (rather I believe they coincidentally both meant ‘help’ – now reflected in their derivative forms 佐 and 祐 respectively). But the fact that they may have started out as oral synonyms and ended up as opposites is neat.
God! You fellas were brainy, right enough. We only knew 0553, and we thought we were pretty damned clever, let me tell you.
Maybe this isn’t one of those, but it might be. My wife was trying to say ‘tap dancing’ but she didn’t know how to say it in English, so she said something like ‘ti ta’. I wonder if that is Chinese, and if it is one of those onomatopoetic words.
What’s the story with that character you always see at weddings and stuff? You know, it’s symmetrical down the middle.
You mean this one, right?
囍
Double happiness. Two people in joy. Happy happy!
But it’s one character–a variant of 喜–not a pair of analogous but different characters like DB posted.
How about 凹凸 ?
Yeah, that’s how you say tap dancing in Chinese (ti1ta1wu3), and like pingpong, it certainly does seem onomatopoetic. Compound words like 蝴蝶 which both have the same semantic but different phonetics wouldn’t count. Rather, I’m looking for character pairs which might have been invented simultaneously as a matching pair through variation of a graphic element, e.g., by graphic inversion, like 凹 ao1 and 凸 tu1 (that pair just came to me). I’m sure there must be more! In the oracle bones, pee and poo were similar, with the excretia depicted to the front and back of a standing person respectively, for instance. 
That’s a good one. And I never thought of it before when I saw those characters alone, but as soon as I saw them together in your post I immediately thought of a pair of lovers in a state of pre-copulatory excitement - which is what its meaning ought to be. Then it could be used to say something like: 你們兩個凹凸,馬上去找一個房間,好不好! How about we coin it as new slang?
They should have one for a glass that’s half empty, and then another for a glass that’s half full.
There’s a company called 由甲地產, and a 1979 Hong Kong film called 油脂甲由 (“Disco Fever”), but I guess those don’t really count.
How about 日月? And 森林? 炎焱? 沝淼? 刀刃?
I’d be very surprised if any of these pairs were really “created as a matching pair.” In each case, almost certainly one character came before the other – perhaps long before, perhaps just a little, and perhaps even in a different form. But you’d have better references than I for just when these individual graphs appeared.
I was just guessing. 
有冇
OK, here’s a good one: 曱甴
Cool! A visually interesting pairing. Unfortunately the etymology of both is a bit obscure, but there’s little chance they were created as a pair, no.
甲 is most often theorized to have been a depiction of a helmet or some turtle or bug’s carapace, since it is part of the compound words for armor and fingernail for instance. But that analysis is presumably heavily influenced by the seal form 
. The oracle bone forms are + (or + inside a box): 
and 
. The latter is close enough to the seal form so as not to reject out of hand the helmet hypothesis, but the former is harder to explain. It might just be a simplification, but note that the former + is homographic with the graph for 七 at the time, which meant 切 cut (i.e., 七 qi(e)1 ‘cut’ was borrowed to write its homophone qi(e)1 ‘seven’). It is a line, cut in half, or an x-shaped incision. And if you look at the pronunciations of 甲 jia3 and 切 qie1, they’re so close that it becomes plausible that these homographs (in the OB) are actually cognate. Worth considering, anyway.
由 on the other hand has been theorized (Wieger?) to depict the germination of a large fruit stone or seed, thus explaining the meaning of ‘to proceed from; from; origin’. It could also simply be the depiction of a large, pomelo-like fruit, now 柚. There aren’t any pre-bamboo graphs that I know of, and the bamboo and clerical forms of this are basically identical to the modern one, with the exception of one bamboo form 甴 (see the pair sjcma posted).
The first are merely depictions of the round sun and the crescent or half moon; their similarity now is coincidental. As for the next three, well, there are lots of reduplicated elements, and those are fun, but we should do a separate thread on them (including pronunciations and meanings).
Dāo and rèn, on the other hand, are closer. But I think they fall in a different group, in which a stroke is added to indicate a specific part of a graph or to otherwise alter the meaning of an original, the way 未 wèi and 末mò add to 木 mù. Another example would include 尤 yóu, which appears to add a stroke or dot to the hand-arm graph (the basis of the 手 丑 九 shǒu choǔ jiǔ group), likely to indicate a deformity or wart (now 肬 or 疣 yóu ‘wart’).
Really? Looking at my original list, it’s hard to imagine any of the following being created independently and at different times, then borrowed to form a pair: 忐忑乒乓孑孓. I don’t think any of them ever had an independent existence. How many times do we use 孓 by itself? And even 上 and 下 are likely to have been created at the same time, probably by the same person or small group of people. As for time periods, I don’t have good info on忐忑乒乓孑孓 but上 and 下 are both found in the oracle bones.
I always liked the latter creation. Not created at the same time, and not an inversion, but definitely in the right spirit. 
Ten bucks says you own a Hanyu Dazidian! Nice find! But how are they related? The former is defined as 取物 (?!) and the latter is either 1) zhá, a vulgar version of 甲 jiǎ ‘armor’, or 2) yóu, 同“由”. 