The JLPT Paraphrase Question (言い換え類義): Why the N2 and N1 Synonym Questions Are Harder Than They Look

The JLPT synonym question does not test whether you know what a word means. It tests whether you can tell two words apart that mean almost the same thing, and at N2 and N1, "almost" is the entire exam.

On the official test specifications this question type is called 言い換え類義 (iikae ruigi), listed in English as "Paraphrases" in the section composition for every level. The format is simple: a short sentence with one word or phrase underlined, and four options. You pick the option closest in meaning to the underlined part. You can see the real format in the official sample questions.

Simple format, brutal question. Here is why, and how to study for it properly.

What the question looks like on paper

A typical N1 paraphrase item looks like this:

彼の説明はあやふやだった。

1. はっきりしなかった
2. 間違っていた
3. 長すぎた
4. 珍しかった

The underlined word is あやふや (あやふや, vague, noncommittal). The answer is はっきりしなかった (it was not clear). Options 2 through 4 are wrong not because they are absurd, but because each one describes something a vague explanation might also be. A vague explanation could easily be wrong, or too long. The question forces you to isolate what あやふや itself asserts, and nothing more.

At N2 and N1 this question type sits inside the Language Knowledge section alongside kanji reading, contextually defined expressions, and usage questions. It is only a handful of items, but they are fast points if you have the skill and pure guesswork if you do not, because every option is grammatical, plausible, and roughly on topic.

Why definition matching fails

Most learners prepare for vocabulary the way flashcards train them to: word on one side, English gloss on the other. That approach collapses on the paraphrase question, because the test writers choose distractors that share a gloss with the correct answer.

Take さすが (さすが, as one would expect, impressively so) and 当たり前 (あたりまえ, natural, obvious, only to be expected). A flashcard learner has both filed under "as expected." A paraphrase question will happily exploit that:

彼の対応はさすがだった。

Does that mean his handling of the situation was "as expected, impressive" or "as expected, nothing special"? The words point in opposite emotional directions. さすが is praise: the person lived up to a high reputation. 当たり前 is dismissal: of course it happened, no credit earned. 「さすがプロだ」 admires the professional. 「プロなんだから、できて当たり前だ」 refuses to admire them. Same English gloss, opposite tone, and only one survives the paraphrase.

同義語 versus 類義語: the distinction the exam lives in

Japanese has two words for "synonym," and the difference between them is the difference between passing and failing this question type.

Every worked pair below is a 類義語 pair, and each one shows a different axis on which "same meaning" quietly breaks.

Six pairs that show how the question really works

These pairs come from JLPT Mastery's confusion data, the same dataset behind our per-level commonly confused word pages. They were selected because learners genuinely mix them up, which is precisely the property the exam selects for.

あやふや vs 不明: fuzzy versus unknown (N1)

• あやふや (あやふや, vague, uncertain, noncommittal)
• 不明 (ふめい, unknown, unidentified)

Both translate as "unclear," which is why learners confuse them. The split: 不明 states an objective gap in knowledge, while あやふや describes a wavering, half-formed quality in what a person says or remembers.

「事故の原因はまだ不明だ」 (the cause of the accident is still unknown) works because the cause is a fact nobody has established. Swap in あやふや and the sentence turns strange, as if the cause itself were being indecisive. Conversely, 「彼の記憶はあやふやだ」 (his memory is fuzzy) cannot take 不明: his memory is not missing, it is blurry. On the exam, あやふや paraphrases to はっきりしない (not clear), never to わからない (not known).

たやすい vs あっけない: easy versus over too soon (N1)

• 容易い (たやすい, easy, simple)
• 呆気ない (あっけない, disappointingly quick, anticlimactic)

Both can describe a contest that ended without a struggle, and our data shows learners collapse them into one "it was easy" concept. But たやすい evaluates difficulty, while あっけない evaluates duration and emotional payoff. あっけない carries a note of letdown: you expected drama and got none.

「決勝戦はあっけなく終わった」 (the final was over disappointingly fast) is about the spectator's deflated expectation. 「彼にとって、その計算はたやすい」 (that calculation is easy for him) is about skill against task, with no letdown anywhere. A paraphrase question underlining あっけなく wants something like 意外なほど簡単に早く (surprisingly easily and quickly), and 簡単に alone, the pure たやすい reading, is exactly the distractor waiting for you.

しきりに vs 強いて: repeatedly versus if forced to (N1)

• 頻りに (しきりに, frequently, repeatedly, incessantly)
• 強いて (しいて, daringly, forcedly, if one must)

These two get tangled together because insisting on something usually involves doing it repeatedly. But the words sit on different axes: しきりに counts repetitions, 強いて marks pushing against resistance or reluctance.

「彼はしきりに時計を気にしていた」 (he kept glancing at his watch) is pure frequency. 「強いて言えば、青のほうが好きだ」 (if you force me to say, I prefer the blue one) is pure reluctance: the speaker would rather not choose at all. Swap them and both sentences die. On the exam, しきりに paraphrases to 何度も (many times); 強いて言えば paraphrases to あえて言えば (if I dare to say). Knowing which axis the underlined word lives on is the whole question.

やがて vs そろそろ: the story's clock versus the speaker's clock (N1)

• やがて (やがて, before long, in due course)
• そろそろ (そろそろ, soon, about time to)

Both mean "soon," and this pair is a clean example of context deciding everything. やがて narrates: events unfold and, in due course, something happens. そろそろ judges: the speaker senses the right moment for an action approaching.

「やがて日が暮れた」 (before long, the sun set) is a narrator watching time pass. 「そろそろ帰ろうか」 (about time we headed home, no?) is a person deciding to act. You cannot narrate with そろそろ or nudge someone with やがて. When the exam underlines やがて, the answer is something like しばらくして (after a while); when it underlines そろそろ, the answer leans toward もうすぐ〜する時間だ (it is almost time to). Same gloss, different clock.

さすが vs 当たり前: praise versus dismissal (N2)

• さすが (さすが, as one would expect, impressively so)
• 当たり前 (あたりまえ, natural, obvious, ordinary)

Covered above, and worth restating as a rule: this pair shows that two words can agree on propositional content ("the outcome matched expectations") and still be non-interchangeable because they disagree on attitude. さすが admires, 当たり前 shrugs. The paraphrase question is one of the few places the JLPT directly tests attitude, and learners trained on glosses alone never see it coming.

みすぼらしい vs 浅ましい: shabby versus shameful (N1)

• 見すぼらしい (みすぼらしい, shabby, seedy)
• 浅ましい (あさましい, wretched, despicable, base)

Both describe a person in a miserable state, and learners blur them for that reason. But みすぼらしい is visual and materially poor, while 浅ましい is moral. 「みすぼらしい格好で現れた」 (he showed up in shabby clothes) judges the coat, not the character. 「金のためなら何でもするとは、浅ましい」 (doing anything for money, how despicable) judges the character and says nothing about the coat.

A paraphrase question underlining みすぼらしい wants 貧しそうな、ぼろぼろの (poor-looking, worn out). The trap option will be something in the moral register, and if your mental entry for both words is just "wretched," the trap works.

Test yourself: five paraphrase questions

Each question below is built from one of the pairs you just read, in the real exam format: a sentence with one word marked in bold, four options, pick the closest in meaning. Answers and reasoning are in a separate section further down the page, after the distractor analysis.

Question 1

事件の詳細はまだ不明だ。

1. わかっていない
2. あいまいだ
3. 発表されていない
4. 間違っている

Question 2

決勝戦はあっけなく終わった。

1. 意外なほど早く
2. とても簡単に
3. 静かに
4. 予定通りに

Question 3

彼はしきりに時計を気にしていた。

1. 何度も
2. 無理に
3. こっそり
4. たまに

Question 4

話しているうちに、やがて日が暮れた。

1. しばらくして
2. もうすぐ
3. 急に
4. とっくに

Question 5

その男はみすぼらしい格好で現れた。

1. 貧しそうな
2. 下品な
3. 派手な
4. 変わった

How the distractors are chosen, and what that tells you

Look back at those six pairs and a pattern emerges. Every wrong option on a paraphrase question is a word that shares the correct answer's territory on one axis and quietly leaves it on another.

Test writers do not need to invent these traps. They already exist in learners' heads, because every learner who studies through gloss-based flashcards manufactures these confusions at scale. The distractors are simply a mirror of how vocabulary is commonly mislearned. We wrote about the broader phenomenon in our guide to commonly confused JLPT words; the paraphrase question is that phenomenon distilled into an official question type.

You cannot study for 言い換え類義 by learning more words. You study for it by learning distinctions.

Answers and reasoning

The answer to every question is option 1. That is deliberate: the interesting part is not which option wins but why the specific trap in option 2 loses, because option 2 is always the other half of the confusion pair.

Question 1: わかっていない. 不明 (ふめい) states an objective gap: what happened has not been established. Option 2, あいまいだ (vague), is the あやふや trap; it describes fuzziness in how something is expressed or remembered, an impression, not a missing fact. 発表されていない (not announced) implies someone knows and has not told you, which 不明 does not say. 間違っている (wrong) asserts the details exist and are incorrect.

Question 2: 意外なほど早く. あっけない measures the gap between expected drama and actual duration: over sooner than the build-up promised. Option 2, とても簡単に (very easily), is the たやすい trap, pure ease with no expectation and no letdown. 静かに (quietly) is a different sense entirely, and 予定通りに (as scheduled) contradicts the word, since あっけない means the ending arrived before you were ready for it.

Question 3: 何度も. しきりに counts repetitions, and glancing at a watch again and again is exactly its home ground. Option 2, 無理に (forcibly), is the 強いて trap; it marks pushing against resistance, and nothing in this sentence resists him. こっそり (secretly) adds a stealth the word does not contain, and たまに (occasionally) inverts the frequency.

Question 4: しばらくして. やがて is the narrator's clock: the conversation continues and, in due course, the sun sets. Option 2, もうすぐ (any moment now), is the そろそろ trap, the speaker's clock; try inserting it, 「話しているうちに、もうすぐ日が暮れた」, and the sentence breaks, because もうすぐ cannot narrate a completed past event. 急に (suddenly) asserts abruptness, and とっくに (long since) puts the sunset before the conversation started.

Question 5: 貧しそうな. みすぼらしい judges the coat: visual, material shabbiness. Option 2, 下品な (vulgar, low in taste), is the moral-register trap the 浅ましい pair predicts, judging manner and character rather than the appearance of poverty. 派手な (flashy) points the opposite direction visually, and 変わった (unusual) is neutral oddness with no poverty in it.

A study method that builds the actual skill

The method that works is learning near-synonyms in contrast, deliberately, instead of hoping the distinctions accumulate on their own.

Timing on the day

The Language Knowledge and Reading sections share one timed block (105 minutes at N2, 110 at N1, per the official section composition), so every minute lost on vocabulary is a minute stolen from reading passages that cannot be rushed.

Where to drill this

JLPT Mastery records which wrong option learners actually pick and builds drills from those specific mix-ups. To work through the pairs your own brain confuses, the free commonly confused word pages for N2 and N1 cover 200 pairs per level; that is the same skill 言い換え類義 tests, practiced directly.