NipponDailyNipponDaily

Patterns

What the vocabulary is made of

The same entries, counted across words instead of read one at a time: which layer of the vocabulary they come from, how the JLPT levels differ, and which processes keep turning up together.

Words open
7830
Taken into parts
1455
With a changed reading
127
  • 和語 Native Japanese
  • 漢語 Sino-Japanese
  • 外来語 Loanword
  • 混種語 Hybrid
  • Not stated

Layers

Each JLPT level, by layer

Every bar is one level scaled to 100%, so you can compare their mix rather than their size.

Processes

How many words carry each process tag. A word can carry several, so these add up to more than the number of words. The colours show which layers the process turns up in.

Processes that travel together

Three or more together

Sets of three or four tags, counting a word's layer as one, that the same words carry. A word that also has other tags is counted too. Each set opens in Explore with “All of them” on.

Readings that change inside a word

In 127 of the 7830 words, the “Taken apart” row records a part whose reading differs from its own — 日 is ひ alone but び in 日曜日. This is what these entries show, not a rule of the language: it is a small sample, and the parts come from parsing Wiktionary's text. Browse the rendaku words.

Ending in っ

The part's last kana became っ in the word.

How the pitch falls, by layer

Every accent the entries show, sorted by where it falls and grouped by the word's layer. Each bar is one layer scaled to 100%. A word with two accents counts in both, and a word with none shown is left out.

  • 和語 Native Japanese2446 wordsflat (heiban) 38% · head-high (atamadaka) 13% · middle-high (nakadaka) 39% · tail-high (odaka) 10%
  • 漢語 Sino-Japanese3391 wordsflat (heiban) 64% · head-high (atamadaka) 28% · middle-high (nakadaka) 6% · tail-high (odaka) 2%
  • 外来語 Loanword479 wordsflat (heiban) 14% · head-high (atamadaka) 57% · middle-high (nakadaka) 28% · tail-high (odaka) 0%
  • 混種語 Hybrid138 wordsflat (heiban) 50% · head-high (atamadaka) 14% · middle-high (nakadaka) 27% · tail-high (odaka) 9%
  • Not stated151 wordsflat (heiban) 50% · head-high (atamadaka) 15% · middle-high (nakadaka) 27% · tail-high (odaka) 8%

How long the words are, by level

The mean length of each level's words so far: the beats (morae) in the reading and the kanji in the spelling.

Level Words Morae per word Kanji per word Written without kanji
N570431.1154 (22%)
N46553.41.3121 (18%)
N320883.21.4243 (12%)
N217273.61.5274 (16%)
N126563.61.5370 (14%)

Words that sound the same

566 readings are each shared by two or more words so far, 1315 words in all. In 486 of them every word shows a pitch accent, and in 94 of those no two words share one, so the pitch alone tells them apart. The readings with the most words are below; only the readings are compared, not how the words are related.

These count the tags on the entries so far — what KANJIDIC2 and the quoted Wiktionary text establish for each word — not the Japanese language as a whole. The words are drawn from the JLPT N5–N1 lists, and a layer or process is counted only where a source establishes it, so the mix reflects what the sources hold. A layer is left unstated where neither source settles it.

© 2025 - 2026 NipponDaily. Released under the Apache-2.0 License.

Origins quoted from Wiktionary (CC BY-SA 4.0) via Kaikki.org; meanings & kanji from JMdict, KANJIDIC2 & KRADFILE (EDRDG, CC BY-SA 4.0) — how we check it