The negation test: how to be certain on Necessary Assumption questions

Necessary Assumption questions ask which unstated thing the argument requires. Most students answer them by feel — which answer sounds most like something the author believes — and feel is exactly what the wrong answers are built to exploit.

The negation test replaces the feeling with a check you can actually run.

The test

Take the answer choice and negate it. Then ask: does the argument still hold?

• If negating it destroys the argument — the conclusion no longer follows — the assumption was necessary. That is your answer.
• If negating it leaves the argument standing, it was not necessary, however reasonable it sounds.

That is the whole method, and its power is that it converts a question about what an author probably thinks into a question about whether a specific piece of reasoning survives. One is a matter of taste; the other has an answer.

Negating correctly is the hard part

Almost every failure of this technique is a negation done wrong, and the error is nearly always the same: negating too hard. The negation of a claim is its logical opposite, not its reversal.

• "All X are Y" negates to "not all X are Y" — which means at least one is not. It does not negate to "no X are Y".
• "Some X are Y" negates to "no X are Y".
• "X is the most important factor" negates to "X is not the most important factor" — not to "X is unimportant".
• "X will happen" negates to "X will not necessarily happen", not to "X will never happen".

Overshooting the negation makes weak answers look devastating, because almost any argument collapses if you assume the extreme opposite of one of its parts. If every choice seems to destroy the argument when negated, you are negating too hard.

When to use it

Not on all five choices — that is slow, and on a 35-minute section slow is expensive. Read all five first, eliminate the obviously irrelevant ones, and run the test only on the two or three that survive. Used that way it is a tiebreaker rather than a procedure.

It is also the single best tool for blind review of assumption questions, where you have unlimited time and the goal is certainty rather than speed. Running it properly on a question you got wrong usually shows you exactly which word you were reading past.

What it does not work on

Sufficient Assumption questions, which ask for something that would make the conclusion follow rather than something the argument needs. The two question types look similar and the test only applies to one of them, so identify which you are looking at before you reach for it.