Japan’s lost morphemes?

Today Namiko Abe of answers the following question: “Why there are two ways to write “ji” and “zu” for hiragana and katakana?”

I was very interested in this, because I’ve noticed it too. One obvious example is Great Teacher Onizuka, which some romanize as Great Teacher Onidzuka.

I always assumed that there was a pronunciation difference between じ and ぢ and ず and づ; namely, /ji/ versus /dji/ and /zu/ versus /dzu/. (Note that that really isn’t a /d/. The difference is that of a fricative versus an affricate. The /d/ is just the easiest way to write it without using the IPA.)

There may be a difference, but apparently if there is it has no bearing whatsoever on the meaning of the word. Abe indicates that the alternate spellings occur as an artifact of assimilation. I therefore suspect that the Japanese don’t differentiate between the two sounds.

I wonder if this has always been the case, or if the writing system reveals a past difference?