The hardest short questions on Paper 2 ask not what a passage says but how it says it, or what it implies without stating. These reward a different kind of reading: slower, attentive to word choice, and disciplined about naming what the writer does rather than what the writer argues.
Some questions ask you to suggest a reason, or to explain why something is surprising, where the passage does not say it outright. These test inference: a reasonable conclusion drawn from what the passage gives you, with the textual basis shown. The key word is reasonable. You are not guessing freely; you are reading between the lines and pointing to the lines you read between.
A good inference answer does two things: it states the inferred idea, and it grounds it in something specific from the passage. An answer that floats an idea with no anchor is a guess. An answer that quotes a detail but never says what it implies has stopped halfway. You need both halves.
Attitude questions ask how the writer feels about something, or how the tone shifts between two parts of the passage. The mistake is to name a feeling and leave it bare: the writer is critical. That is an assertion, not an answer. The mark is for tying the feeling to the words that carry it.
Work from the evidence to the label, not the other way round. Suppose a writer describes a new policy as another grand promise destined to gather dust. The words grand promise and gather dust carry weary scepticism: the writer expects the policy to fail and has seen this before. Name the attitude, then show the words that prove it. A tone shift works the same way: point to the change in word choice between the two paragraphs, not just the change in subject.
The most demanding short question asks how the writer uses language to convey a stance or effect. The phrase uses language is a gate: it tells you the object of analysis is the writing, not the content. The single most common failure here is to answer the argument instead, explaining what the writer thinks rather than how the words work.
Test yourself before writing each point. Am I about to explain what the writer says, or what the writer does with words? If your answer would survive even if the writer had used completely different words, you are explaining content, and it scores nothing. The question wants the words themselves.
Every use-of-language point needs three parts, and the mark depends on all three: name the device, quote a short piece of evidence, and explain the effect tied to the stance the question named. Drop any one and the point fails. A device with no quotation scores zero. A quotation with no named device and no explained effect scores zero.
Here is the method on an invented sentence. Suppose a writer attacks a company by saying it gorges on its customers data while feigning concern for their privacy. Point one: the writer uses the verb gorges, a word associated with greedy, animal overconsumption, which casts the company as predatory rather than businesslike. Point two: feigning concern accuses the company of pretence, suggesting its privacy assurances are a deliberate act rather than a real commitment. Two devices, two quotations, two explained effects, each tied to the writer's disapproval.
If a question asks for three ways, it wants three genuinely different techniques, not the same technique spotted three times. Calling three separate words strong language is one point made three times, and it earns one mark. The marker is looking for range: a loaded verb, then a metaphor, then a rhetorical question, for instance.
So when you scan a paragraph, deliberately reach for different categories of device. Word choice and connotation is one. Imagery and metaphor is another. Sentence structure, repetition, listing, or a rhetorical question is another again. Pulling from different categories guarantees your three points are distinct rather than three labels for the same effect.
A variant asks how a particular sentence, often the final one, works as an effective conclusion. This is still a use-of-language question, so you still explain what the sentence does, not what it says. The difference is that you are analysing its function in the passage as a whole.
Look for things like a return to an image from the opening, which creates a sense of the argument coming full circle; or a shift in phrasing that turns a claim into an invitation the reader is more likely to accept; or a touch of wit that makes the closing line memorable. Each of these is a distinct function, explained, not a summary of what the sentence states.
Book a free trial lesson and see how we teach GP from the exam backwards, with marked practice against the band descriptors.