Lesson 4, Paraphrase & Prediction

The answer options never use the speaker’s words. This lesson is the bridge, and the armour.

The same rule as Reading, with less to hold onto

If you’ve done our Reading course, you know the paraphrase principle: the correct answer restates the idea in different words, while wrong answers often echo the exact wording. Listening runs the identical design, with one aggravation:

In Reading, the passage stays on screen. In Listening, the audio is gone.

You can’t re-check the source. All you have is your memory of the meaning and your notes of the facts - which is exactly why the previous two lessons built those. The options will be four sentences you’ve never heard, one of which restates what a speaker meant, and three of which are engineered from fragments of what they literally said.

The echo trap, listening edition: an option containing the vivid phrase you remember, “$75,000,” “a nightmare,” “without even asking”, feels like recognition, and recognition feels like truth. But the number you remember, welded to a claim nobody made, is the test’s favorite wrong answer. Your notes say “75K”; the trap says “she felt she was worth more than $75,000.” The keyword matches. The meaning was never said.

The defense is the same as Reading’s, sharpened for audio: familiar wording = slow down. The right answer usually feels less familiar, because it shares no vocabulary with your memory of the clip, only its meaning.

Predict before you look

The single most protective habit in the section:

When the question appears, answer it from memory, in your own words, before reading the four options.

The mechanics of why: the options are one answer and three engineered traps. If you read them cold, all four enter memory and start competing, and the traps are built to feel plausible against a ninety-second-old recollection. If you arrive holding your own answer, you’re not choosing the most familiar-feeling stranger; you’re matching each option against a standard.

The routine, per question:

1. Read the question. (“How did the wife feel about the purchase?”)
2. Answer it yourself, eyes off the options. (“Furious, mostly that he didn’t ask her first.”)
3. Now scan the options for your meaning in their words. (“Upset about not being consulted” → match.)
4. If nothing matches your prediction, then re-think, but you’ve re-thought from a position, not from a void.

This is the payoff of the explain-it-to-yourself drill (Lesson 2): prediction is re-explanation, done at question time. If you’ve practiced restating clips in your own words daily, step 2 takes three seconds.

Fine emotional shades: where predictions need precision

Many Listening questions turn on distinctions between neighboring feelings, and the options are built from exactly those neighbors:

How did she react? - (a) angry (b) disappointed (c) understanding (d) indifferent

All four are negative-to-neutral; a vague prediction (“she wasn’t happy”) can’t split them. Two tools that can:

1. Weigh the airtime. Which emotion got the most evidence in the clip? A speaker who spent forty seconds explaining why the decision hurt and five seconds saying “but I get it” is disappointed, not understanding, the brief concession doesn’t outweigh the sustained complaint. The test loves offering the five-second emotion as an option.

2. Respect the qualifiers. Doesn’t care is not isn’t offended. Cares what her friends think doesn’t mean obeys them. Sometimes is not always - recall Clara’s cousin, sick after HIIT only when it runs over an hour: the trap option says “HIIT always makes her cousin sick,” and only the qualifier in your notes kills it. When predicting a feeling or a claim, predict its strength and conditions, not just its direction.

Attribution: who actually said it?

The reattribution trap, a true statement pinned on the wrong person, runs throughout the section and peaks in Parts 5 and 6. Three versions:

• Speaker A’s claim in a question about Speaker B: your memory confirms the fact; the question asked about a different person
• The narrator’s framing pinned on an interviewee: “the debate has divided the town for months” was the reporter, not the campaigner; this is why Lesson 3 brackets narrator lines
• A quoted opinion mistaken for the quoter’s own: “critics say the plan is reckless, but I disagree” contains the word reckless while asserting its opposite

The two-check rule, listening edition: before accepting an option, confirm both that it was said and who said it. Your prediction should already carry a name: not “someone thought it was too expensive” but “the wife thought so.”

The compute traps: when the answer was never spoken

A distinct family of questions gives you the pieces and asks for the assembly, the correct answer is a number or fact no one said aloud:

Heard: “75% of crabs with both claws removed don’t survive.” → Asked: roughly what share survive? → Answer: about 25%. Heard: “They’re opening their seventh location next week.” → Asked: how many operate today? → Answer: six.

The trap options are, of course, the numbers you actually heard, 75, seven, sitting there radiating familiarity. The spoken figure is the raw material; the answer is one inference away.

Defense: when a question involves any number, pause one beat and ask, is this asking what was said, or what it implies? Tense is the usual pivot: opening next week means not-yet-open; don’t survive means the survivors are the remainder. Your notes hold the input; the arithmetic is yours to do, deliberately, not by reflex-matching.

When you’re torn: the subjective vs objective test

Endgame situation, two options left, both plausible, audio long gone. The tiebreaker:

Choose the one you can point to evidence for, in your notes or your explicit memory of a line. Not the one that “seems right.”

Seems right is your model of the world talking, plausibility, stereotype, what a person like that would probably think. Can point to is the clip talking. The test deliberately builds one option to flatter your assumptions and one to match the audio; under uncertainty, the assumption-flatterer usually feels slightly better. That feeling is the tell. Go with evidence, every time.

(Recognize this? It’s the pointing test from Reading Lesson 4, rebuilt for audio, with your notes standing in for the passage.)

On Celpora, each explanation quotes the moment from the clip. Listening practice sets reveal the sentence the correct answer paraphrases, and mark the exact spot each distractor twisted, so you can hear what your notes should have caught. Over five sets, the paraphrase habit stops being effortful.

Grade your own question technique

After any practice set:

• I answered from memory before reading options, every question
• My predictions carried names: who felt/said/did it
• Familiar wording in an option made me slower, not faster
• Emotion answers weighed airtime, not just direction
• Qualifiers (only if, sometimes, unless) survived into my answer
• Number questions got the said-or-implied check
• Ties broke on evidence, not plausibility

6+ ticks = the options are working for you now, as a lineup to identify your prediction in, not a maze to get lost in.

Your practice

Open one set on Celpora and put today’s lesson to work.

Practice Listening. Guess the answer in your own words before you look at the options.