{{first_name | みんな}}さん、こんにちは ☕

There's a section in the JLPT listening test that even strong readers tank on. It plays one short line. Three replies. Nothing is printed except a number. By the time your brain finishes translating, the options are already gone and you're picking on a vibe.

You finish the mock and check the sub-scores. Every other listening block looks survivable. This one looks random.

Readers describe it the same way every cycle: I just blank out. If that's been your worst sub-score, it's the worst sub-score for a lot of readers — including the ones who pass the rest of the test comfortably.

The section is called 即時応答そくじおうとう — "quick response." At N1 it's eleven of the thirty listening questions. That's about 37% of your N1 listening items riding on a sub-format almost nobody trains for explicitly.

At N2 it's around nine to ten. At N3 it sits next to 発話表現はつわひょうげん for roughly eight to ten between them.

Standard prep materials mention the section exists and then move on to longer-passage strategy. That's the gap.

Here's the part the gap is hiding. 即時応答 isn't really a listening test. It's a pragmatic reflex test wearing a listening costume.

Audio plays once. Options are spoken, not printed. You get two or three seconds. The section rewards a specific reflex (pragmatic pattern-matching) that anime, podcasts, and general shadowing do not train directly.

Which is the good news, if you're 24 days out from the July test. Reflexes are trainable on short timelines. Vocabulary cramming isn't.

We're going to walk through the actual cognitive procedure (three steps, in order) that you can run in real time during the test. Then a ten-minute daily drill that makes it stick.

The whole section lives inside small contrasts like this one.

Prompt: 「明日あしたの会議かいぎ、何時なんじからだったっけ？」 — "What time was tomorrow's meeting again?"

All three are grammatical. Only one is a natural reply to a colleague performing a 確認かくにん — a confirmation-seeking question, marked by 〜だったっけ.

The difference: 確認 wants a confirmation answer, not a fresh proposal. c proposes a new time. b confirms the time the speaker is half-remembering. b is the reply.

The trap is that on the page, you'd catch this in five seconds. At spoken speed, with options playing once, you don't.

Most JLPT listening tips say "practice more" or "read the options first." The second one is impossible here — options aren't printed. So we need the actual cognitive procedure, the one you can run while the audio is still playing.

Three steps. In order.

Step 1 — Identify the speech act in the first one or two seconds.

A speech act is what the utterance is doing, not what it's saying. Japanese pragmatics has a standard inventory, the same one Japanese textbooks and SLA papers use. Eight or nine categories cover almost everything you'll hear on the test.

That's the working set. Memorize the Japanese terms. They're the labels actual JLPT prep materials use, and they'll show up in your future reading. Half the credibility of this section comes from learning the field's actual vocabulary instead of made-up English glosses.

Step 2 — Who acts next? 話はなし手て or 聞きき手?

This is the pivot. One question collapses about a third of the section's apparent ambiguity.

The speaker (話し手) is the person whose voice you just heard in the prompt. The listener (聞き手) is whoever they're talking to (usually the test-taker, narratively). The right reply depends on who's about to take action.

Step 3 — Match to a stored response chunk by function.

Once you know the speech act and who acts next, you don't have to compose a reply. You reach for one already on the shelf. Native speakers do this. Test-takers can too.

A starter chunk bank:

You don't need a hundred of these. You need ten or twelve that fire automatically. Speed of retrieval beats library size.

This is where the section gets ugly for plateau fighters.

〜ましょうか vs 〜てくれますか — speaker-offers vs request-of-listener. Two grammatically clean responses are wrong. One is the reply.

〜してください vs 〜させてください — one syllable changes who acts.

〜だったっけ vs 〜にしましょうか — confirmation vs fresh proposal.

In print, you'd catch every one of these. At 0.8× cognitive speed in real time, with three replies playing in a row, your parser misses the cue and your hand picks the grammatically-correct-but-pragmatically-wrong option.

That's the failure mode. It's not a Japanese gap. It's a reflex gap. You need to build a parser that catches the cue before the meaning resolves.

The 10-Minute 即時応答 Reflex Drill — do this 5 of the next 7 days.

Open any JLPT 聴解ちょうかい past-paper audio, 新完全しんかんぜんマスター, or a YouTube playlist of 「JLPT 即時応答 練習」. Your SRS system might have these as cards too. Find one 即時応答 set. Then run three short passes.

Pass 1 (3 minutes) — label only. Play only the prompt utterance. Skip the options. After each prompt, say out loud: the speech act is ___, the next actor is 話し手 / 聞き手. Don't pick a reply. Just tag.

Pass 2 (4 minutes) — generate. Replay each prompt. Speak your own natural reply aloud before the official options play. Don't pick. Generate. This is the chunk-retrieval rep.

Pass 3 (3 minutes) — pick under timer. Play prompt + three options at normal speed. Pick under a hard 3-second timer. Then check the script and grade your error type: was it perception (you misheard), parsing (you missed the grammar), or pragmatics (you picked a grammatically-fine but socially-off reply)?

(No literal answer key — this is a reflex drill. Pattern reveal drops in next week's brew.)

You'll know you did it when: You can name the speech act of each prompt within 2 seconds, even on the ones you eventually get wrong on the reply. The reflex is the win, not the score.

Last week's answers (The Reading-Speed Trap): No literal key — that was a measurement drill. Early replies are clustering around three patterns (fuller results when more numbers land). Most N3 readers report somewhere between 150 and 200 chars/min on first measure, which is the textbook plateau and exactly what the drill is for. N2 readers tend to cluster around 220–280, just under the 250–350 target. Speed gap, not comprehension. And almost everyone who reported a "comprehension fine, time bad" mock matched the diagnostic profile: drop the new graded reader, pick up the stopwatch.

If you only do Pass 1 for three minutes tomorrow morning (label the speech act, name the next actor, move on) you've trained more 即時応答 reflex than 95% of test-takers will between now and July 5. The bar is on the floor. You only need to label one prompt.

For: Plateau fighters who want one structured source of 即時応答 prompts that match real test pacing.

Use: 新完全マスター聴解 N2 or N1. Open the 即時応答 chapter, run the three-pass drill on one set per session. Ten minutes total. Don't binge.

Why it helps: The prompts are written at JLPT pacing, the speech acts are intentionally varied across the chapter, and the audio is the closest you'll get to the real test without a past-paper download. If you don't have it, a YouTube playlist of 「JLPT 即時応答 練習」 works for the same drill.

これからも一緒いっしょに頑張がんばりましょうね〜 💪