Phonology & Connected Speech · 8 min read · Published
Rendaku, Sokuon, and Chōon: Decoding Connected Speech Shifts in Japanese
Master sequential voicing (連濁), choked consonant stops (促音), and vowel elongation (長音) to understand fast, native-speed Japanese.
The acoustic reality of natural Japanese speech
In classroom Japanese, words are articulated with textbook clarity. But in natural conversation, speech sounds undergo systematic morpho-phonological modifications to maximize articulatory ease and rhythmic flow.
Failing to recognize these modifications makes learners feel that speakers are swallowing words, when in truth they are adhering to well-defined phonological rules.
Sequential voicing: The logic of Rendaku (連濁)
Rendaku is the phenomenon where the initial unvoiced consonant of a second compound element becomes voiced (k->g, s->z, t->d, h->b/p).
- Example: 'hito' (person) + 'bito' -> 'hitobito' (people); 'te' (hand) + 'gami' (paper) -> 'tegami' (letter).
- Lyman's Law: Rendaku is strictly blocked if the second element already contains a voiced consonant (e.g., 'kuro' + 'kumo' -> 'kurokumo', NOT 'kurogumo', because 'gumo' already has voiced 'g').
- Sino-Japanese constraint: Pure Katakana loanwords and Chinese compounds rarely undergo Rendaku compared to native Yamato words.
Sokuon (促音) and Chōon (長音): Timing is meaning
Japanese is a mora-timed language where each mora occupies equal rhythmic duration. The small 'tsu' (っ / sokuon) is a full mora of silent glottal or consonant closure. Skipping this beat alters meaning: 'kita' (came) versus 'kitta' (cut).
Similarly, long vowels (長音) require holding the vocalic sound for two full beats. Confusing 'ojisan' (uncle) with 'ojiisan' (grandfather) is an iconic trap stemming from rushed mora timing.
Training rhythmic mora perception
When doing Japanese dictation, count the rhythmic beats with fingers or tap along. Ensuring every small 'っ' and prolonged vowel occupies a full rhythmic slot guarantees accurate Romaji and Kana transcription.