Verbal Playbook: The Passage Is the Universe

The shape of the section

23 questions, 45 minutes — the tightest per-question budget on the exam (~1:57). Two formats only:

📚 Format	What you get	What's measured
Reading Comprehension (RC)	A passage + several questions	Understanding structure, drawing supported inferences
Critical Reasoning (CR)	A short argument (under ~100 words) + one question	Evaluating logic: strengthen, weaken, assumptions, flaws

One law governs both: the passage is the entire universe. Every correct answer is provable from the given text; every tempting wrong answer recruits something you brought with you — domain knowledge, plausibility, real-world experience. Verbal is less a reading test than a discipline test.

Reading Comprehension: read for architecture

Passages span science, history, business, social commentary — and expertise in the topic is deliberately useless. What pays is reading for structure over substance:

• 🏗️ On first read, track the job of each paragraph: claim → evidence → counterpoint → verdict. A four-word margin note per paragraph ("theory", "objection", "data", "so what") outperforms highlighting everything.
• Locate the author: are they reporting a debate, taking a side, qualifying a popular view? Half the question set hangs on voice, not facts.
• Don't speed-read. The section rewards comprehension; passages are short enough that one careful read beats two skims.

The question types and their tells:

Type	Stem sounds like	Winning move
Main idea	"primary purpose…"	Match the whole arc, not the loudest paragraph
Detail	"according to the passage…"	Find the line; verify wording, not vibes
Inference	"can be inferred / suggests…"	The answer is one small certain step away — never a leap
Application	"most analogous to…"	Abstract the principle first, then map it
Structure	"in order to…"	Answer the function ("concede, then rebut"), not the content

Wrong-answer factory settings — RC distractors are manufactured in four moulds: too extreme (always/never where the passage hedges), half-right (true clause welded to a false one), true-but-unasked (correct fact, wrong question), and outside knowledge (true in the world, absent from the passage). Name the mould as you eliminate; it gets mechanical with practice.

Critical Reasoning: find the conclusion, then the gap

Every CR stimulus has the same anatomy: evidence → (assumption bridge) → conclusion. Your routine:

1. 🎯 Read the question stem first — knowing the task (weaken? assumption? evaluate?) changes how you read the argument.
2. Underline the conclusion. It's often disguised mid-paragraph behind a "therefore", "clearly", "should".
3. Articulate the gap between evidence and conclusion before touching the options. The argument leaps from "sales rose after the campaign" to "the campaign caused it"? The gap is causal. Nine times out of ten, the correct answer lives exactly there.

Task playbook:

• Strengthen / Weaken — add a fact that props up / kicks out the bridge. Beware options that strengthen a different conclusion (scope shift).
• Assumption — what must be true for the leap to hold? Test with negation: if negating the option demolishes the argument, it's your answer.
• Flaw — name the reasoning error (correlation→causation, unrepresentative sample, percentage vs absolute number, false dilemma).
• Evaluate — the right option is a question whose two possible answers swing the verdict in opposite directions.
• Inference / conclusion — pick what is guaranteed, however boring. Bold claims lose to timid certainties.
• Plan / proposal — judge fitness for the stated goal; an option can be a great idea and a wrong answer.

You are never asked whether the evidence is factually true — only whether the reasoning from it holds. Argue the logic, not the world.

Pacing the section

• CR (~10 of the 23) runs on a tight loop: stem → conclusion → gap → options, about 90 seconds when the routine is grooved.
• RC sets amortize the reading: a careful 2-minute read typically buys three or four sub-minute questions. Protect the read; rushing it taxes every downstream question.
• Same endgame as everywhere: eliminate, commit, 🔖 bookmark doubts for the three-edit review. Blanks are the only unforgivable answer.

Building the muscle

Drill by question type in the question bank — type-pure sets first (ten assumption questions in a row teaches the pattern faster than mixed practice), then mixed timed sets, then full mocks. In review, your error log should record which mould trapped you, not just "got it wrong". Verbal improvement is the most log-driven of the three sections.