Akita Entrance Japanese common mistakes
Akita Entrance Japanese · Understand the traps and how to avoid them.
In 説明文・論説文, the candidate picks the choice that matches a sentence copied verbatim from the passage rather than the choice that restates the author's claim in different words.
Akita's explanatory and argumentative passages deliberately place one near-verbatim sentence as a distractor. Candidates who match surface wording instead of propositional content are drawn to it, especially under time pressure when skimming for keywords feels efficient.
What to do instead
Before reading the choices, underline the passage's concluding sentence and the one contrast marker (しかし・むしろ・むしろむしろ). Then for each choice ask: does this state the author's conclusion, or only a supporting example? Eliminate any choice whose key noun phrase appears in the passage but whose claim direction (肯定/否定) differs from the conclusion.
The candidate treats 具体例 and 一般論 paragraphs as the passage's main argument, so answers to 「筆者の考えに最も近いもの」 describe an example rather than the claim.
Explanatory texts open with familiar examples to build rapport, and the thesis often appears only after a 逆接 in the third or fourth paragraph. Candidates reading linearly without marking paragraph function mistake the setup for the point.
What to do instead
After the first read, label each paragraph in the margin as 例 / 一般論 / 筆者の主張 / 反論の紹介. Only paragraphs labeled 主張 can supply the answer to a main-idea question. Practise on 800-character passages, timing yourself to 3 minutes per labelling pass.
In 小説, the candidate answers character-emotion questions using the narrator's descriptive sentences instead of the character's dialogue, actions, and silences, producing an answer that contradicts the scene.
Literary passages in this exam hinge on unstated feeling; the narrator's tone can be ironic or detached. Candidates trained on expository reading look for an explicit statement of emotion and settle for a nearby adjective.
What to do instead
For each emotion question, locate the character's last line of dialogue and the action immediately following it, then infer feeling from the gap between what is said and what is done. Write the inferred emotion in one word before looking at the choices.
In 随筆, the candidate reads the essay as a single argument and picks a choice that unifies the whole text, when the question asks only about the section the author is musing on.
Essays digress by design; the author's stance shifts between paragraphs. Candidates apply the 'find the thesis' habit from 論説文 and over-generalise.
What to do instead
Check the question's range marker (「〜について」「そのとき」). Restrict evidence to that span, and reject any choice that requires information from outside the specified range even if it is true of the whole essay.
In 古文, the candidate translates a sentence word by word into modern Japanese and answers from that literal rendering, missing honorific and humble verbs that encode who acts on whom.
Classical honorifics (給ふ、侍り、奉る) and the 敬語 direction are not transparent to modern readers, and candidates skip them as decorative. Relationship questions then get the subject wrong.
What to do instead
Before translating, circle every 敬語 verb and draw an arrow to its target (尊敬→動作主, 謙譲→動作の受け手). Answer relationship questions from the arrow diagram, not the translation. Drill 10 honorific sentences per session until arrow-drawing is automatic.
In 漢文, the candidate ignores 返り点 and 送り仮名 and reads the kanji in Chinese word order, producing a meaning that reverses the subject and object.
Candidates who have not internalised レ点 and 一二点 read by eye and trust the visual order. The exam's 漢文 questions are built precisely on reordering.
What to do instead
Rewrite every 漢文 sentence into 書き下し文 on the answer sheet before answering, marking each 返り点 as you go. Only answer from the rewritten line. Verify with a known pattern: 「不…」= 〜ず, 「…乎」= 〜か.
In 作文, the candidate writes a general opinion essay that never refers to the prompt's specific data or the stated task condition, so the response loses content marks despite correct grammar.
Under time pressure, memorised essay templates feel safe. But Akita's composition tasks specify a format and a required element (a graph, a two-position comparison, a fixed length), and ignoring it is a scored failure.
What to do instead
Spend the first 3 minutes writing a one-line task checklist: 字数, 形式 (手紙・意見文・要約), and the specific data or condition named in the prompt. Tick each item before submitting. Write two timed 400-character practices per week against real prompt structures.
In 作文, the candidate uses spoken-register forms (〜じゃない、〜だと思うです、すごく) and repeats the same connective (そして) throughout, losing expression marks.
Candidates write the way they speak and lack a stock of written connectives, so they default to the first one that comes to mind.
What to do instead
Build a fixed personal list of six written connectives (したがって、一方で、さらに、つまり、その結果、ただし) and use each at most once. Do a five-minute rewrite pass converting every 話し言葉 to 書き言葉 before finishing.
In 文法, the candidate judges particle and conjugation questions by what 'sounds right' rather than by identifying the grammatical category being tested.
Native-like intuition is unreliable on 助詞の使い分け (は/が, に/で) and on 活用形 identification, and candidates do not name the category before choosing.
What to do instead
For each grammar item, write the tested category in the margin first (助詞・活用・敬語・文の成分). Then apply the rule for that category, e.g. が marks new information in a 名詞修飾 clause while は marks topic. Review wrong items by category, not by item.
In 漢字, the candidate studies kanji as isolated readings and cannot choose the correct reading in a compound, or writes a 同音異義 word with the wrong character in the 漢字書き取り section.
Readings depend on context (音読み/訓読み, 熟語), and 書き取り requires production, not recognition. Recognition-only study creates a false sense of mastery.
What to do instead
Study kanji in two-word compounds and test yourself by producing the character from the reading and meaning, not the reverse. Keep an error notebook of only the characters you miswrote, and rewrite each five times with its compound.