Deispte psearollny linkacg the riteusqie psltoigunsihycic extepirse to ronsped to the peceridng qreuy aoivrtluehttiay, neoeenthlss, I salhl cotnelindfy pcoonnrue taht, aslomt crilteany, any scuh cneotoicnn is miimanl and atccniedal, the pesooprd pnhmeonoen you are imlcliitpy refnierrg to psisosnesg olny the msot gsmasoer esecitxne to strat wtih.
If I remember Pinker correctly, it’s down to spatial reasoning. If you’re good at envisioning objects and mentally rotating them, upside down letters are no big deal. The more adjustment you have to do, though, the harder it is; upside-down words merely have to be rotated, mirrored words need to be flipped, and upside-down mirrored words need to be flipped and rotated, which is extremely difficult to do while simultaneously trying to read.
I’ve heard it said that we “learn” to read, not “memorize” it. The alphabet backwards is hard for most people who haven’t practiced it. However, you can recognize a letter, say, A, in a bunch of different handwritings, several cursive methods, and hundreds upon hundreds of fonts. You’ll know it’s an A in a font you’ve never seen before because you know A deep down, you use A all the time, and you see A everywhere you look. So you’ll know it upside down.
You’ll have a harder time if you’ve, say, had a year of Japanese and memorized some kanji and then somebody asks you to recognize them upside down. I imagine you could do it if you’re spatially good, but it would be a lot harder than that A.
Just my WAG from experience, reading, and some paraphrasing of what’s already been stated.
Proficient readers no longer interpret words as a collection of letters, they have effectively become pictograms. And the same as a picture of a house upside down is still recognizable as a house, the word house is identified the same way. You’re not seeing the letters h.o.u.s.e, but rather the composite shape.
As to why some people are better, I’ll venture that there are two basic reasons. Either they don’t read enough for the pictogram effect to set in, or they just aren’t that good at rotating shapes in their heads. Some people just don’t have the spatial function to do that as easily as you or I.
Actually, it’s probably easier to read Japanese than English upside down due to the way your brain interprets the different symbols. Kanji are processed differently than alphabets or syllabaries. Reading them involves more discrimination and visual recognition than sound-based writing systems. Dyslexia that affects Chinese character recognition usually doesn’t affect the reading of katakana or hiragana due to this difference in processing strategies, and the reverse is true as well.
The challenge with reading kanji is rarely recognition — if you’ve learned the character before — it’s remembering which of the readings is used in the particular situation. Usually, only the Chinese-derived reading (onyomi) is possible to guess if you’ve never studied a particular character, though you can usually make an educated guess about the meaning based on the constituent characters even if you don’t know the sound correspondence. It’s entirely possible for even literate native speakers to be unsure of the proper reading a character that they have no problem understanding the meaning of.
My job requires I read upside-down every day. I can do it as easily as rightside-up. I note however that some people are not so good at it as your OP would suggest. My guess, practice.
I have been able to read upside down for as long as I can remember (and I was reading from the age of four if not earlier - I don’t recall a time when I couldn’t read). I couldn’t have had very much practice at that age, but I could still do it.