Most if not all Japanese lessons start with a sentence in the like of: 私はXXです。 (Watashi ha XX desu, I’m XX.) This is bad pedagogically for at least three reasons — we may dive in that issue in later posts — in addition to bringing the inevitable question: “why is は (transcribed ha) pronounced wa instead of ha” as one would expect?
That legitimate questioning is brushed over by the teacher with an expeditious “for historical reasons” non-answer, the topic never to be addressed again. I certainly was very frustrated by this, on my first lesson nonetheless! The truth is, my teacher probably didn’t know the details herself.
Before writing this article, I checked what are the proposed “answers” on the web, and I found a single correct one among what is a sea of misleading, wrong or incomplete yet sometimes lengthy content.
Fortunately, I came accros the answer while reading various papers. First, we’ll see why is the topic marker written は, and then how it evolved to be pronounced wa.
The Topic Marker Orthography
Old Japanese is the oldest attested state of the Japanese language with records dating from the 8th century. The Kojiki (712), the Nihon Shoki (720) and the Man’yōshū (759) are the three largest and famous texts from that period. At this time the topic marker already existed and is reconstructed as being pa.
In the 8th century, Japanese language was painfully written using Chinese characters, using either their semantic value denoting a native Japanese word, or for the sound they had in Chinese, heard through a Japanese ear (and a Korean one before this in the chain of transmission). Particules such as topic marker didn’t have a Chinese counterpart and were thus always written using that second mode of representation (that is phonographically).
Looking at the following extract from Man’yōshū, we see the topic marker used just after the word ima, which have the same meaning as today: now.
伊麻波伊可尓世母
ima pa ika ni se mo
“Whatever shall I do now?”
The character used to write pa there is 波, now read /pwɔ/ in Standard Chinese. This kanji’s reading also starts with an initial /p/ in multiples Chinese languages like Cantonese and Hakka, and /pʰ/ in Taiwanese which is how linguists reconstructed the Middle Chinese prononciation /puɑ/. Since Japanese language internal rules (phonotactic constraints) doesn’t allow for a /u/ between that initial and the main vowel, it’s easy how to see Japanese could have use this character to write pa.
With time, a set of characters used for their pronunciations and written in cursive style give birth to the hiragana syllabary. The kana は spawned from the kanji 波.
Now, we’ve seen why the topic marker is nowadays written は: stemming from 波 read as pa to note the Old Japanese particule pa. Yet it doesn’t explain why it’s now read wa.
Languages Change
Languages change all the time, in terms of vocabulary, grammar rules but also phonemes (sounds) they use. Often a change arise to make the language easier to speak but that’s not always the case. One remarkable case of sound changes in English is the Great Vowel Shift. It explain a lot of “weirdness” in English orthography.
Japanese is of course not immune to that phenomenon. When a sound change, it normally changes in all words containing that change, with the catch that context matters: depending on the position of the sound as well as the preceding and following phoneme, the effect can be different.
The phonetic change that interest us occurred between 950 and 1000, when /p/ is between to vowels, which is called the intervocalic position. The phoneme /p/ changed to /w/ in front of all vowel except /u/ (in that case it disappeared). This is how modern kawa (river) is stemming from the earlier kapa.
Few centuries later /p/ would also change when in initial position to /f/ (as recorded by Christian missionaries) then /h/, but this is not the case applicable to は as the topic marker: in effect, particules in Japanese are always appended to word or phrase, and this is the resulting whole unit that is taken into account for other phenomena. Notice how speakers, especially on TV, will make pauses after ga, wo, ni, ha, etc. but never in between the particule and the preceding word.
So, because は is always following a word, and because a Japanese word is always ending in a vowel save for the nasal consonant noted by ん, ha is in effect almost always in intervocalic position. Which is why pa changed to wa.
From this historical explanation plus the one about orthography, it should be clear why は is read wa. In addition, I recommend keeping a critical mind when reading things on the internet (including this blog) because a lot of content is of questionable veracity. In the case of languages, native speakers explanations are often equally if not more lacking because they never had to question their knowledge.
References
The Man’yōshū extract, transcription (adapted) and translation is taken from [1]. The image is taken from Wikipedia at this address. Note that the post content is exclusively based on the changes listed by Frellesvig’s book. More hypothesis are mentioned in Labrune’s books.
[1] Kupchik, J. (2012). Morpheme-Based Rendaku as a Rhythmic Stabilizer in Eastern Old Japanese Poetry. Acta Linguistica Asiatica, 2(1), 9-22.
[2] Frellesvig, B. (2010). A history of the Japanese language. Cambridge University Press.
[3] Labrune, L. (2006). La phonologie du japonais (Vol. 90). Peeters Publishers.
[4] Labrune, L. (2012). The phonology of Japanese. Oxford University Press.
Why is the Japanese topic marker は read 'wa'?
Is it just me or as non-Japanese I actually find it _easier_ to say "wa" in that position than "ha"?! Probably something rooting in neuro-linguistics and/or mouth's muscle memory.