Lesson 4 · Vocabulary
Eighty words, two letters each
The forty-six characters are readable. That is not the same as reading. This lesson turns them on real words — eighty of them, every one exactly two kana long, every kana drawn from the forty-six you already have.
Fifty of the eighty have already walked past you as example words in lessons 1 to 3. They were props then — something for a character to sit inside. Here they are the point.
1.Two beats, not two letters
Before the words, the thing that makes them sound Japanese.
Each kana is one mora: one beat, and every beat gets the same length. A two-kana word is therefore two beats, evenly spaced, with neither one leaning on the other. Nothing is stressed, nothing is swallowed.
This is where beginner pronunciation goes wrong, because English does the opposite — it picks a syllable to stress and squeezes whatever is left.
| Word | The two beats | What an English mouth does instead |
|---|---|---|
| にく | に · く | neek — one long beat where there should be two even ones |
| すし | す · し | SOO-shi — the first beat stretched, the second thrown away |
| ほん | ほ · ん | hone — the ん loses its beat entirely |
ん is a beat of its own. ほん is ho-n, held as long as みみ or にく — not the single syllable an English ear hears in "hone". It is the one kana that is not a consonant plus a vowel, and it is the one most often robbed of its time.
Arabic already counts length
Arabic makes duration carry meaning: عَلَمٌ is a flag, عَالَمٌ is a world, and the whole difference is how long one vowel is held. Nobody who reads Arabic lets those two collapse into each other. Japanese asks for the same discipline and applies it more strictly — it counts beats rather than vowels, and it counts them in every word, not only the ones where a mistake would change the meaning.
So read every word below in two. Tap the speaker if a rhythm will not settle; the synthesised voice is imperfect on pitch but the timing is right.
2.The words, by meaning
Nine groups. Grouping by meaning rather than by first character is deliberate: you will want やま when you are talking about a mountain, not when you are working through the や row.
Each card gives the word in kana, its romaji, its English sense, and the Arabic — Modern Standard, fully vowelled — with a play button for each language. Read the kana first and use the romaji only to check yourself.
What the rule leaves out, and why
Two constraints decide what is here: exactly two kana, both from the forty-six. That second one is stricter than it sounds. Perfectly ordinary words are excluded by it — みず (water), まど (window), そば (buckwheat noodles), かぎ (key), ゆび (finger) — because ず, ど, ば, ぎ and び all need a mark you have not been taught. Two small strokes turn す into ず and open up several hundred more words. They are not here yet, so neither are those words — lesson 5 adds the marks, and those five words are the first five on it.
The word cards need JavaScript. The same eighty words, in the same nine groups, are on the two-kana word list reference sheet.
3.Read them for meaning
Lesson 3 drilled shape to sound. This drill assumes the sound and asks for the meaning, which is the part that makes a word a word. It runs in three directions, and they are not equally hard:
- かな → English — reading. You will be best at this.
- English → かな — production. Much harder, and much more useful: this is the direction speaking needs.
- かな → العربية — the same word, straight into Arabic, with no English in between.
The three wrong answers are drawn from the same meaning group as the right one. Offered "mountain, cat, to buy, chair" against うみ, anyone could pick the right one without knowing the word; offered "mountain, river, sky, forest", you either know it or you do not. Answer by tapping or by pressing 1–4. Pick a single group from the menu when eighty in a row is too many.
The drill needs JavaScript. Every word it asks about is in the list above.
A correct answer moves on by itself. A wrong one stops and shows you the whole card — kana, romaji, English, Arabic, both play buttons — because a multiple-choice question that leaves a wrong answer standing will happily teach you the wrong answer.
One pass through a group, then the misses again, is a session. Do not sit there re-running it to a clean round; a clean round twenty minutes after a dirty one proves very little. Come back tomorrow and switch the direction — the words you can read are not yet the words you can produce.
4.When two words share a spelling
Six of the eighty carry a note saying they also mean something else. That is not bad luck in the selection. Japanese has a small inventory of sounds and an enormous number of two-beat words, so collisions are unavoidable — and the commonest words collide most, because they have been in the language longest.
| Word | Here it means | It also means |
|---|---|---|
| はな | nose | flower — the sense lesson 1 taught |
| あめ | rain | sweets, hard candy |
| くも | cloud | spider |
| はし | chopsticks | bridge, and also edge |
| かみ | paper | hair, and also god |
| した | tongue | below, under |
Two things pull them apart, and neither is your job yet.
In speech, pitch. はし meaning chopsticks starts high and drops; はし meaning bridge does the opposite. Japanese pitch accent is real, it is regional, and it is a later problem — say the beats evenly and let context carry the rest, which is what it does for the many pairs pitch does not separate either.
In writing, the characters. Written out properly these are not the same word on the page at all: rain and sweets are 雨 and 飴, nose and flower are 鼻 and 花. Those are kanji, and there is nothing to learn about them here — only the reassurance that the ambiguity you are looking at is an artefact of writing everything in kana, and that Japanese itself is not as ambiguous as this page makes it look.
Meanwhile, treat the collisions as free vocabulary. Learning that くも is a cloud has already taught you the word for spider.
5.What eighty words buys you
Not sentences — those need particles and verb endings that have not been taught. What it buys is the thing that comes before sentences: enough words in the head that a real one, met in the wild, is mostly known rather than mostly unknown.
It also buys eighty pieces of Arabic on the same hook. That is the whole arrangement of this course — one set of meanings, reached from two directions — and it is at its cheapest here, where each meaning is a single concrete noun with one obvious translation. It gets more expensive later.
6.Sources
- Jisho.org, serving the JMdict/EDICT dictionary project — the source for every reading, every gloss and every "also means" on this page. A word was allowed in only if JMdict marks at least one of its entries common, which is what keeps the list free of dictionary curiosities. The one tool to keep open while studying: it takes kana directly, so you can look up anything on this page without knowing a single kanji.
- NHK World — Easy Japanese — free lessons with native recordings. The best available check on the even two-beat rhythm of section 1; speech synthesis gets the timing right but the pitch wrong, and only a real voice fixes that.
- Vance, T. J. (2008), The Sounds of Japanese, Cambridge University Press — the standard English-language account of the mora as Japanese's unit of timing, and of why ん counts as one. A library book rather than a link to read tonight; the chapter on moras and syllables is the only part section 1 draws on.
- Tofugu, "Learn Hiragana" — the mnemonic set behind lesson 3, and still the place to go when a character stops being readable under time pressure.
- Madinah Arabic and Almaany — the Arabic forms and their harakat. Neither settles the harder question of whether a word is the one a speaker would actually reach for; that needs a native ear.
- Two-kana word list — all eighty in one page, laid out to print and keep beside the paper.