A common refrain heard across the internet is that Japanese is a phonologically impoverished language. In fact, they say, it is so phonologically impoverished that the language is inundated with an unwieldy amount of homophones. As a result, it is impossible to write Japanese using a purely phonetic writing system. Kanji are the only way to disambiguate words.
The truth is much more complicated. Japanese does in fact have an unusually large amount of homophones but the cause is only tangentially related to its phonology. Additionally, kanji may actually be exacerbating the problem.
First, let’s unpack the notion that Japanese is just too phonetically simple. Japanese has 10 vowel (5 short and 5 long) and roughly 20 consonants (14 phonemes and 6 allophones). Japanese syllable structure is of medium complexity. On the one hand, it does not allow for consonant clusters. However, it does allow for a consonant followed by a glide in a syllable onset, and the sound <ŋ>, written with the kana ん・ン, as the only consonant allowed in a syllable coda. Thus, Japanese syllable structure can be described as (C(j))(V)(ŋ). This is much simpler than English syllable structure, which is (C)3V(C)5, meaning that a syllable can have at most 3 consonants clustered in an onset and up to 5 consonants clustered in the code. However, it is more complex than languages like Hawaiian, which only allow (C)V syllables.
Given the number of phonemes in Japanese and its restricted phonotactics, we can easily compute the exact number of possible mora in the language. The answer is 103 (excluding the obsolete moras ゐ and ゑ). Here are all of them handily summarized in this nice chart from Wikipedia.
However, a mora is a distinct concept from a syllable and although there are only 103 possible mora, the total number of possible syllables is much higher.
So, what is a mora? Japanese is a rhythmically timed language, and a mora is essentially a unit of speech that constitutes one beat. A syllable on the other hand, can be comprised of multiple moras. This is because long vowels are two moras, the syllable coda ん is considered a mora in its own right and the glottal stop, represented by a small つ is also a mora. As a result we can have words that each have a different number of syllables but the same number of moras.
Tōkyō -とうきょう – 2 syllables – 4 mora
Ōsaka -おおさか – 3 syllables – 4 mora
Nagasaki -ながさき – 4 syllables – 4 mora
So given that explanation, there are at least several hundred possible syllables in Japanese, conservatively put at around 400. It’s easy to see how I reached the number 400. First, 100 of the mora in Japanese are allowed as syllable onsets, and these can easily be extended by the addition of a long vowel. So きゅ becomes きゅう, に becomes にい, etc. This alone brings us up to 200 possible syllables. Next, each of these 200 syllables can further be used to create another syllable by the addition of the mora ん. So きゅう would become きゅうん、にい would become にいん, etc. This doubles our list of possible syllables from 200 to 400. I cap my count here at 400, because it is debatable whether other sequences like わいん where the vowels in the middle are not identical constitutes one syllable or two (わ and いん).
So given 400 potential syllables, even if Japanese consisted of nothing but two syllable words, it would be theoretically possible to create 160,000 distinct, non-homophonous words (400 x 400). The average adult native speaker of English or Japanese has an active vocabulary of between 20,000 and 30,000 words. So even with the supposedly simple and limited phonology of Japanese, we have by far more than enough room to make every single word distinct. And this is only with limiting ourselves to 2 syllable words. Since Japanese words can be quite long, oftentimes being anywhere from 4-8 mora or more, there are potentially billions of distinct possible words.
So given the huge potential of Japanese, why are there still so many homophones? That question will be explored further in part 2 of this series.