Reddit reviews A History of the Japanese Language

It's useful to distinguish strata when talking about phonology in Japanese and Korean. You'll find that one set of rules is useful for describing native Japanese vocabulary. You'll find another that is useful for describing Sino-Japanese vocabulary. And you'll find that yet another set work when talking about onomatopoeic vocabulary.

Here, we're talking about Sino-Japanese words, and it gets a little complicated. Japanese borrowed in Chinese words and coined pseudo-Chinese words at several different time periods. It's one of the reasons you often find more than one Chinese reading for a given character (like how 赤 'red' can be read as seki in 赤外 sekigai 'infrared' but as shaku in 赤銅 shakudou 'an alloy of copper and bronze'.

So originally in Japanese, here was no sound /h/. Instead, linguists have reconstructed, based on internal evidence from Japanese and related languages like Okinawan, and from external evidence like reconstructions of how Chinese characters were supposed to sound in Chinese, Korean, and Vietnamese historically, that what is today /h/ was originally /p/. So 'flower', which is modern Japanese [hana], would have been Old Japanese (around 800 CE) [pana].

Over time, initial p- became f-. We have really good written evidence for this, because it was still the case when the Jesuit missionaries got to Japan (we can call this Late Middle Japanese or Early Modern Japanese, depending on who you ask). It has since gone from initial f- to initial h-. So fully, 'flower' would have gone from OJ pana to MJ fana to ModJ hana.

Medial -p- has a much more complicated history. It tended to also follow a similar path of -p- > -w- > -∅- (that is, nothing), but there are a number of wrinkles. To give an example of the first, we could look at the Japanese word for 'big'. In OJ, we have opo-. In MJ, we have owo-. This finally deleted the consonant, so we have the modern form of oo-, like in ookii.

There are a number of environments where this is blocked. We only really need to talk about two to deal with the examples OP gives above. First is where we have a geminate consonant (that is, a sequence of /pp/), such as in modern appare 'splendid'. /p/ was also retained after /n/. For instance, in 門派 monpa 'a religious sect'.

It's a little more complicated still, as sometimes Chinese characters like 百 or 表 are treated as ending with a final vowel, like when it appears by itself or at the end of a word (hyaku, hatsu, etc.), but other times, they are treated as having a final consonant, like when they are in the middle of a compound (hyak-, hat-). There's then an additional rule that a consonant cluster in Japanese cannot have consonants which are made at different places of articulation, so that while pp and nt are fine, tp and nk don't work. These are "fixed" by the first consonant taking on the place of articulation of the second one.

And there's yet one more rule where if a compound is easily identifable as a compound, like 朝日 asahi 'morning sun', each word is treated independently, so that the h in 'sun' doesn't act like it is between vowels, even though it really is.

Bjarke Frellesvig's book A History of the Japanese Language covers this in a lot more depth.

I strongly suggest you take this time to learn how to search your library/use your inter-library loan system for introductory books, and to speak with your professors for references. I don't say this to be patronizing; finding resources on your own and getting in touch with professors are valuable skills to develop in college.

Wiley and Oxford each have "handbooks," collections of overviews by researchers, on Japanese linguistics. They should be readable as an undergraduate if you've already taken some linguistics courses. Routledge has a three-volume compilation of seminal research articles on Japanese linguistics, but this may be too advanced and narrow in scope. There's an introductory book by Tsujimura and a recent book on historical Japanese linguistics by Frellesvig. Each will consistently reference a handful of researchers (e.g., Ito & Mester and Kawahara for the phonetics/phonology of Japanese) whose websites you can search for more.

Not to de-necro an old thread, but some here might be interested in the book "A History of the Japanese Language" by Bjarke Frellesvig, where he goes into a big section on old Japanese. Among a few highlights are eight vowels, no vowel length, and many words beginning with "h" began with "p". So many homonyms today were not in OJ. Interesting reading, a bit thick though, and not cheap.

http://www.amazon.com/History-Japanese-Language-Bjarke-Frellesvig/dp/1107404096/ref=sr_1_1?ie=UTF8&qid=1419784277&sr=8-1&keywords=Frellesvig