Years ago I did an internship at a foreign company in Japan whose local branch had the somewhat enigmatic extension “K.K.” Throughout the internship I kept wondering what these two letters were supposed to mean, but found it just too embarrassing to ask.

I’m mentioning this episode because it relates — with some detours — to an interesting phenomenon in Japanese word formation. When two words melt into one, they frequently seal their morphological relationship with a “blurring” of the second part. Take hana-bi (花火, firework), for instance, which is a compound of hana (花, flower) and hi (火, fire). This process, which turns hi into bi, ku into gu, chi into ji, etc., is commonly known as rendaku (連濁), a connection (連) that blurs (濁).

Rendaku is particularly frequent in native Japanese words, as in our hana-bi example. Sino-Japanese vocabulary is known to resist rendaku, though on closer inspection this turns out to be not quite in line with the facts. True, it’s kakueki-teisha (各駅停車, local train) and shōmei-shashin (証明写真, ID photo) rather than kakueki-deisha and shōmei-jashin (no matter how blurred the photo is). But on the other hand, there are quite a number of Sino-Japanese words that do blur: renshū-jiai (練習試合, practice game), doryoku-busoku (努力不足, lack of effort) and wa-gashi (和菓子, Japanese-style confectionery), to name but a few. And rendaku is very common in Sino-Japanese two-character words such as renpai (連敗, succession of defeats), senpai (先輩, one’s senior) and kanpai (乾杯, cheers).

A much more reliable candidate for nonblurring is loanwords. Except for less than a handful of exceptions, this in fact seems to hold water. No matter how hard you try, the second part of terms like danshi-toire (男子トイレ, men’s room) or nama-kuriimu (生クリーム, fresh cream) will never turn into doire or guriimu. When it’s the first part of a compound that’s foreign though, rendaku sets in as usual, as in hāto-gata (ハート型, heart-shape) or hautsū-bon (ハウツー本, how-to book).

Another regularity regarding word type is that rendaku is common in plural formations and other forms of kanji reduplication, but does not normally occur in onomatopoetic vocabulary. Compare the blurred combinations hito-bito (人々, people) and toki-doki (時々, sometimes) to nonblurring kari-kari (かりかり, crunchy) and fuwa-fuwa (ふわふわ, fluffy), which would lose all of their charm if pronounced kari-gari and fuwa-buwa.

And then there is a well-known phonemic constraint that blocks rendaku. Though not part of the Japanese legal code, it’s known as Lyman’s Law. Named after the American mining engineer Benjamin Smith Lyman, who dug out this regularity when working for the Meiji government at the end of the 19th century, Lyman’s Law states that if the second part of a compound already contains a blurred element, no other blurring can occur. An example is naga-sode (長袖, long sleeves), where the blurred de in sode prevents it from becoming naka-zode.

Though not nearly as systematic as Lyman’s Law, even a blurring in the first part of a compound seems to have some overspill potential to not make rendaku happen. At least that would explain the differing readings of family names like Nakajima (中島), which normally does blur, and Nagashima (長島), which does not. It might even help solve the mystery of why there is a station called Ekoda (江古田) on Seibu Ikebukuro Line, and one called Shin-Egota (新江古田) on the Oedo Line.

Phoneticians, with and without additional diplomas in mining, over the years have excavated a larger number of other rules and regularities, including the fact that there are so-called rendaku-lovers such as fune (船, ship) or hako (箱, box), which always give in to blurring, just as there are explicit “rendaku-haters” like kita (北, north) or himo (ひも, string), which will never blur.

One rather strange thing that seems to have gone unnoticed so far is this: When a Japanese term is written in the roman alphabet, there is a strong tendency not to blur — and indeed, even un-blur — a rendaku term, at least visually. Take sushi, which clearly belongs to the group of rendaku lovers: temaki-zushi (手巻きずし, hand-rolled sushi), oshi-zushi (押しずし, pressed sushi), kaiten-zushi (回転寿司, sushi-go-round), etc. Yet when this last term is romanized, it usually becomes “kaiten-sushi.” Similarly, the name of the chain restaurant かっぱ寿司 (read: kappa-zushi) is given on their English home page as “Kappa-sushi.”

And there are more examples. A restaurant in the Yokohama region offers dining that’s “和 cocoro” (wa-gokoro, “with Japan spirit”), a dance school in Tokyo holds a regular event called “夜Sakura会” (yo-zakura-kai, “evening cherry-blossom club”), and a bar in Osaka knows to impress with a greater selection of “地Sake” (chi-zake, local sake). Which brings us back to the K.K. initials from the start: Just in case you have been as ignorant as me, they stand for kabushiki-gaisha (株式会社, Inc.). So, can you see now what the second K represents?

In a time of both misinformation and too much information, quality journalism is more crucial than ever.
By subscribing, you can help us get the story right.