Why are the particles "は" (ha⇒wa), "へ" (he⇒e), and "を" (wo⇒o) not spelled phonetically?

Question

As far as I know only three words (or particles) have irregular, non-phonetic spelling in Japanese:

• "は" - The topic particle is pronounced "wa" but the kana is otherwise pronounced "ha"
• "へ" - The movement towards particle is pronounced "e" but the kana is otherwise pronounced "h"
• "を" - The kana is only used for the object particle as far as I know and in the kana table takes the position of "wo" but pronunciation seems to hover somewhere between "o" and "wo"

So did the sounds of these particles change since the spelling was set down? Or were the spellings intentionally chosen for some reason? Are there any other irregular kana spellings I omitted?

(P.S. I found the question Why is は pronounced as わ when used as a topic particle? while working on this question but since mine is broader I decided to go ahead anyway, plus it might give an opportunity to test our duplicate marking strategies here)

EDIT

Some of the comments to Lukman's answer below suggests that these might not seem like irregular sounds or kana to native speakers. I was surprised that people interpreted this question to be about romaji rather than being about kana and particles as I intended it.

So a further question: Are these particles and their kana and pronunciations seen as irregular by native Japanese speakers? Or just to learners of Japanese? And what do Japanese and non-Japanese linguists think?

Answer 1

I'll give you the same one I gave to the other question: Yes, the sounds of these words has changed since their spelling was set down. In general, no matter the language, whenever you see a discrepancy between spelling and pronunciation that is not entirely regular, this is the result of sound changes. And while there are some counter-examples of words that have changed their spelling to something that doesn't reflect any prior pronunciation due to folk etymology (such as the English word island, which was thought to come from the Latin insula so a totally superfluous s was inserted there) , they are rather rare.

In the case of the unruly Japanese particles を, は and へ, they all reflect an earlier pronunciation that was retained in spelling for many words in historical kana usage, but after the major language reforms of 1946, this conservative spelling was kept only for these 3 particles.

To see what happened here exactly, we first have to get acquainted with traditional kana, that had two extra letters (ゑ and ゐ in hiragana). These letters are usually transliterated as WE and WI respectively, and that was indeed their pronunciation in Old and probably Classical Japanese. In the early 20th century however, the entire W-column (ワ行) has merged with the vowel-column (ア行), except for the letter わ itself. As for the H-column (ハ行), it merged even earlier with the ワ行 in some positions (between vowels), as I've explained in my other answer. Since the ハ行 was already merged with the ワ行 for these positions, it also underwent the merger with the ア行 when it came. So to lay it out more clearly, this is the order of the changes:

1. /p/ softens to /f/ and becomes /w/ between vowels.
   1. は (originally pronounced as PA) is now pronounced WA.
   2. へ (originally pronounced as PE) is now pronounced WE.
2. /w/ disappears before all vowels except /a/.
   1. は (previously pronounced as WA) remains WA.
   2. へ (previously pronounced as WE) becomes E.
   3. を、ゑ、ゐ (originally pronounced as WO, WE, WI) become O, E, I.

Up to 1946, you could find, for instance, spellings such as アヲイ (for 青い, the color blue) and アヲヒ (a name for a family of flowers, sometimes translated as Hollyhock). awoi underwent the disapperance of w and became aoi, while awopi became first awowi and then aoi.