The Japanese Verb System as One Morphological Machine

Core examples: 書く/書かない/書いた, 食べる/食べられる, する/しない, 来る/来ない, 書かせられる.

Conjugation charts hide the machine

Japanese learners often memorize forms one by one:

dictionary form masu form nai form te form past form potential passive causative causative-passive

The list becomes intimidating. Each form feels like a separate rule.

But Japanese verbs are more organized than they first appear. A verb has a class. The class determines stems. Stems combine with endings and auxiliaries. Sound changes occur predictably. Larger forms are built by stacking operations.

The key principle is:

Japanese verb forms are produced by a morphological machine: identify class, choose stem, attach operation, adjust sound.

Once you see the machine, conjugation becomes less like memorization and more like controlled assembly.

The three practical verb classes

For learners, Japanese verbs fall into three main groups:

1. Godan verbs 書く, 読む, 話す, 行く, 飲む

1. Ichidan verbs 食べる, 見る, 起きる, 教える

1. Irregular verbs する, 来る

Godan verbs change the final kana across vowel rows. Ichidan verbs drop る and attach endings more directly. する and 来る must be learned separately.

Dictionary form is the input form

The dictionary form is the base form used for lookup:

書く 食べる する 来る

But the dictionary form is not the only important form. The machine uses different stems for different operations.

Nai form: negative base

Godan verbs shift to the a-row before ない:

書く → 書かない 読む → 読まない 話す → 話さない

Ichidan verbs drop る:

食べる → 食べない 見る → 見ない

Irregular:

する → しない 来る → 来ない

This is not a separate magic form. It is class + negative operation.

Masu stem: polite and connective base

The masu stem is used for polite forms and many derived expressions.

Godan:

書く → 書きます 読む → 読みます 話す → 話します

Ichidan:

食べる → 食べます 見る → 見ます

Irregular:

する → します 来る → 来ます

This stem appears in many nouns and compounds too:

書き方 読み物 話し合う

Te-form and past form: sound-change families

Godan te/past forms involve sound changes:

書く → 書いて / 書いた 読む → 読んで / 読んだ 話す → 話して / 話した 行く → 行って / 行った

Ichidan:

食べる → 食べて / 食べた 見る → 見て / 見た

Te-form and past form are closely related. Learn them as a pair.

Voice and modality operations

Potential, passive, causative, and causative-passive are not random. They are operations added to verb stems.

Potential:

書く → 書ける 食べる → 食べられる

Passive:

書く → 書かれる 食べる → 食べられる

Causative:

書く → 書かせる 食べる → 食べさせる

Causative-passive:

書く → 書かせられる 食べる → 食べさせられる

A long form such as 書かせられる looks frightening until you see its layers:

書く → 書かせる → 書かせられる

cause + passive.

Auxiliaries stack

Japanese verbs often stack with auxiliaries:

書いている 書いておく 書いてある 書いてしまう 書かなければならない 書かせられている

The machine keeps building. A complex form is usually a sequence of smaller operations.

Example bank walkthrough

書く / 書かない / 書いた

Godan verb showing negative and past.

Learner action: identify stem changes.

食べる / 食べられる

Ichidan verb with potential/passive form overlap.

Learner action: use context to distinguish can eat vs be eaten.

する / しない

Irregular verb.

Learner action: memorize core forms early.

来る / 来ない

Irregular verb with reading changes.

Learner action: learn as high-frequency special case.

書かせられる

Causative-passive.

Learner action: break into layers, not panic.

Verb-machine workflow

For any verb form:

1. Find the dictionary form.
2. Identify verb class.
3. Identify the stem.
4. Identify operation: negative, past, te, polite, potential, passive, causative.
5. Check sound changes.
6. Check auxiliary stacking.
7. Interpret meaning in context.

The machine has stems, slots, and auxiliaries

Japanese verb morphology feels overwhelming because textbooks often present forms one by one: masu-form, te-form, nai-form, ta-form, potential, passive, causative. The better model is mechanical.

A verb has a class. The class determines which stem or base appears before an ending. Endings and auxiliaries add meaning.

For example, 書く can produce:

書かない 書きます 書いて 書いた 書ける 書かれる 書かせる

These are not random spellings. They are outputs of a system.

Godan verbs: consonant-stem behavior

Godan verbs shift across the あ, い, う, え, お rows.

書く 書かない 書きます 書く 書ける 書こう

The kana before the ending changes because the verb stem interacts with the suffix. This is why knowing kana rows helps grammar.

Other examples:

読む → 読まない / 読みます / 読める 話す → 話さない / 話します / 話せる 待つ → 待たない / 待ちます / 待てる

Ichidan verbs: stable stem behavior

Ichidan verbs drop る and attach endings:

食べる → 食べない / 食べます / 食べた / 食べられる 見る → 見ない / 見ます / 見た / 見られる

This is easier, but learners must still know whether a verb is actually ichidan. Some verbs ending in る are godan:

帰る → 帰らない / 帰ります 走る → 走らない / 走ります

Do not classify by final る alone.

Irregular verbs: する and 来る

The two major irregular verbs are essential.

する → しない / します / して / した / できる in potential meaning for many contexts

来る → 来ない（こない） / 来ます（きます） / 来て（きて） / 来た（きた） / 来られる（こられる）

来る is especially tricky because the kanji stays the same while readings change.

Auxiliaries stack

Japanese builds complex meanings by stacking auxiliaries.

書かせられる be made to write

食べられなかった could not eat / was not eaten, depending on context

見ていなかった was not watching / had not seen, depending on context

A learner should parse from the verb base outward:

1. core verb,
2. voice/potential/causative,
3. negation,
4. tense/aspect,
5. politeness,
6. modality.

Verb parsing example

読ませられませんでした。

Break it down:

• 読む: read,
• 読ませる: make/let read,
• 読ませられる: be made to read,
• 読ませられません: polite negative,
• 読ませられませんでした: polite negative past.

Meaning: “was not made to read” or “could not be made to read,” depending on context. The exact translation depends on sentence frame, but the morphology is recoverable.

Learner workflow

When seeing a long verb form:

1. Remove politeness if present: ました, ません.
2. Identify negation: ない, ぬ, ず, ません.
3. Identify tense/aspect: た, ている, ていた.
4. Identify voice/potential/causative: れる, られる, せる, させる.
5. Recover dictionary form.
6. Rebuild meaning outward.

This turns conjugation from memorized tables into a working parser.

A strong tool for this article would generate forms by operation.

Suggested functions:

1. Verb input: 書く, 食べる, する, 来る.
2. Class detector: godan, ichidan, irregular.
3. Operation buttons: negative, past, te, potential, passive, causative.
4. Layer view: 書く → 書かせる → 書かせられる.
5. Sound-change warning: 行く → 行って.
6. Context examples: potential/passive ambiguity.
7. Reverse parser: identify dictionary form from complex form.

Final rule

Japanese verbs are not a pile of unrelated charts. They are a machine.

Identify the class. Choose the stem. Attach the operation. Account for sound changes. Then interpret the result in context.

Once you learn the machine, long verb forms become readable.