Elements of set theory: operations, De Morgan and difference
Why this is asked: A free minute at the start of Part B if you rewrite every difference as A ∩ Bᶜ, and a two-minute case analysis if you do not. The recurring question is A \ (B \ C), and the recurring wrong answer moves the brackets.
In one sentence
Rewrite every set difference as , and the whole subject collapses into Boolean algebra where De Morgan does the rest.
Why the exam asks it
Because it is a free minute at the start of Part B for anyone who knows the rewrite, and a two-minute case analysis for anyone who does not. The identities look interchangeable on the page, so the question is really testing whether you can compute rather than recognise.
The idea
One move does almost all of the work:
Difference is not a third operation to memorise rules for. It is intersection with a complement, and once written that way every question becomes union, intersection and complement — where the laws are the ones you already trust.
De Morgan, the only law you need to hold
Complementation swaps union and intersection. Everything else follows.
The worked case the papers keep setting
. Apply the rewrite twice and push the complement inwards:
Four rewrites, no diagram. The last step is distributivity of over .
Read it in words: something survives if it is in and not in the part of outside . So either it misses altogether, or it is in , where cannot reach it.
Where intuition breaks
Difference is not associative. in general. Take and . The left side is ; the right is . The brackets are not decoration, and an option that has silently moved them is the standard wrong answer.
Difference is not commutative either, which is obvious stated plainly and easy to lose inside a longer expression.
is not . The term has to be cut down to first. An option offering the uncut is testing exactly this.
Symmetric difference is associative, unlike ordinary difference — behaves like addition mod 2 on indicator functions, so brackets genuinely do not matter there. Knowing which of the two is associative is worth a mark on its own.
The exam's angle
Two methods, and use whichever is faster for the shape in front of you:
-
Rewrite and push complements. Best when the expression is short and the options are algebraic. Three lines, as above.
-
Test a small example. Best when the options are concrete or the expression is nested. Take , , . One element sits in each pairwise overlap and nowhere else, so it is quick to evaluate and it kills almost every wrong option.
It has a blind spot. Here is empty, and so is the part of each set outside the other two. An option that goes wrong only on the triple overlap therefore survives. When two still stand, switch to , , . That fills all seven regions, and stays outside for complements.
One counterexample kills an option outright — you never need to prove the survivor, only eliminate the rest.
Method 2 is underrated under time pressure. An identity that fails on a three-element example is finished, and checking costs seconds.
The night before
- . Always start here.
- De Morgan swaps and under complement.
- — note the .
- is neither associative nor commutative; is associative.
- To disprove an identity, one small example is enough — provided the region the options disagree on is actually occupied in it.
In plain English
Think of a set difference as a filter with two conditions at once. "In but not in " means "in " and "outside ". That is the whole content of .
Once every minus sign is rewritten that way, only three operations are left: and, or, not. De Morgan is the rule for pushing a "not" through, and it swaps and with or.
Brackets are not decoration. Taking away "the part of outside " is a different instruction from taking away and then adding back.
Where this shows up
Every database query is this algebra. WHERE active AND NOT banned is , and SQL's EXCEPT is set difference outright. Query planners rewrite it into an anti-join using exactly the identity above.
Search filters on a shopping site do the same work: brand, price band, in stock, not discontinued.
Digital circuits are the same algebra under different names. AND, OR and NOT gates obey these laws, and De Morgan is what lets a chip designer build every circuit out of one kind of gate.
More in The Real Line
Take it further
- The mistake this prevents: execution slip → real wrong options, and why each one tempts
- The mistake this prevents: boundary and endpoint → real wrong options, and why each one tempts
- Previous-year questions set on this — every option and the answer, free:June 2025 · Part BJune 2025 · Part BDecember 2025 · Part BDecember 2025 · Part BJune 2025 · Part BDecember 2024 · Part B
Last revised . Found a mistake? Tell us — corrections are the fastest thing we act on.
Next: Completeness, sup/inf, Archimedean property
Create a free account to keep your place and have this feed your study plan.
Open this in the full syllabus view · Unit 1