Ordinary comprehension asks what the passage says. Critical reading asks what the author is doing and how they feel about it. Those are different questions with different evidence, and a candidate who treats them as the same will keep choosing the option that is true but irrelevant.
Three operations decide almost every such question: separating what is stated from what is only implied, reading tone from the words the author chose, and following the structure of the argument. Each has a testable procedure, and together they cover the hardest verbal items on the paper.
For every passage, ask three questions and keep the answers separate.
What the model omits: it cannot supply outside knowledge, and it will not tell you the author's purpose unless you notice how the paragraphs are arranged. It also omits that a single word can carry the tone, so tone must be read from individual word choices, not from the overall subject.
Purpose is a verb. Ask what the author is trying to do, and match that verb to the passage's shape.
| Purpose | What the passage does |
|---|---|
| Inform | Presents facts and findings without judging them |
| Persuade | Argues for a position and answers objections |
| Describe | Builds a picture through sensory or concrete detail |
| Criticise | Identifies faults in a theory, policy or work |
| Analyse | Breaks a topic into parts and shows how they relate |
| Compare | Sets two things side by side to reveal a difference |
Tone lives in word choice. The same event can be reported as neutral, cautious, sceptical, critical, approving or ironic, and the clue is a small set of evaluative words.
| Tone | Words that signal it |
|---|---|
| Cautious | suggests, may, appears, preliminary |
| Sceptical | claims, allegedly, so-called, supposedly |
| Critical | flawed, neglects, misleading, overlooks |
| Approving | elegant, rigorous, promising, valuable |
| Neutral | reports, notes, documents, measures |
Inference has a hard boundary. The correct inference is the option that must be true given the text. An option that is merely plausible, or true in the world but not argued here, is wrong. Extreme words — always, never, all, only, eliminate — are the most reliable signal that an option has overreached.
Structure and function. When a question asks why a sentence or paragraph is present, name its job: it introduces, it qualifies, it gives an example, it concedes a point, it turns the argument. The first sentence of the paragraph and its link to the previous one usually give the answer.
A decade-long study of irrigation in the Indus basin found that fields watered at night lost roughly a third less water to evaporation than fields watered by day. The authors caution, however, that the saving depends on soil type and that night watering is impractical where electricity is rationed to daytime hours.
An inference question asks: which conclusion is supported? The tempting wrong answer is “night watering would reduce total water use on every farm”. The passage found a saving on the fields studied, but adds two qualifications; every reaches past the evidence and the word however warns that the claim is limited.
The supported inference is that the saving is not guaranteed for all fields, because the authors themselves tie it to soil type and to the availability of night electricity. That is one step from the text, uses the author's own qualifiers, and stops where the evidence stops.
In the passage above, what is the authors' tone towards the finding?
Which option is the strongest inference from the passage?
What is the function of the second sentence, beginning The authors caution?
An option to an inference question contains the word only. What does that most likely mean?
Without looking back: name three purpose verbs, and state the boundary rule for a valid inference.
The answer is: inform, persuade, describe, criticise, analyse and compare. A valid inference must be one step from the text and must be something the text requires to be true.
Fill in the blanks from memory:
A claim you can point to a line for is , while a claim that follows in one step is . An option containing an extreme word such as only or always should be suspected of failing on . A writer who merely presents facts without judging them aims to .
Take one passage and write three short lists before answering anything: the sentences that state facts, the claims that are only suggested, and the evaluative words that reveal tone.
Writing the three lists forces the separation the questions are testing. When a question asks for an inference, you answer from the second list, not the first, which prevents the common error of choosing a stated fact for an inference question.
The tone list is deliberately mechanical. You are collecting evaluative adjectives and verbs, not deciding how the passage feels in general. A passage about a serious subject can still be written in a cautious or even ironic tone, and only the word choices reveal which.
Finally, when an option survives, check it against the scope of the passage. Ask whether the passage argues this exact claim, or something narrower. The correct answer sits inside the passage's scope; the wrong ones either repeat a stated fact or step outside the text.
Critical reading closes the teaching of verbal technique. The next lesson puts everything in the module to work under a clock: a thirty-question mixed drill at the real paper's pace, with a full answer key and a repair plan.