Sets, subsets and intervals

Every branch of mathematics you will meet from here on — relations, functions, probability, even calculus — is written in the language of sets: well-defined collections of distinct objects. Before you can say "the domain of ff" or "the sample space of an experiment," you need to say precisely what a set is, how to describe one, and how one set can sit inside another as a subset. This article builds that vocabulary from the ground up: the two ways to write a set, the special sets (empty, finite, infinite, equal), the subset and power-set ideas, and the one genuinely ISC-specific skill in this topic — translating an inequality on R\mathbb{R} into interval notation. It closes with a worked example, the traps that cost the most marks, and a quick-fire quiz.

The core insight

A set is a well-defined, unordered collection of distinct objects, described either by listing its elements (roster form) or by a defining rule (set-builder form); AA is a subset of BB when every element of AA lies in BB, a set of nn elements has 2n2^n subsets in its power set, and subsets of R\mathbb{R} bounded by two numbers are written compactly as intervals using round brackets for excluded endpoints and square brackets for included ones.

The core idea

Describing a set. A set can be written in two equivalent ways. Roster (tabular) form simply lists the elements inside braces, e.g. A={2,4,6,8}A = \{2, 4, 6, 8\}. Set-builder form states a rule every element must satisfy, e.g. A={x:x is an even natural number,x≤8}A = \{x : x \text{ is an even natural number}, x \le 8\}. The two describe the same set; roster form is convenient for small, explicit sets, set-builder form for sets defined by a condition (especially infinite ones, which cannot be listed at all).

Special kinds of sets. The empty set ∅\varnothing (also written {}\{\}) has no elements at all — it is not the same as {0}\{0\}, which contains one element, the number 00. A set is finite if the process of counting its elements terminates, and infinite otherwise; N\mathbb{N}, the natural numbers, is the standard example of an infinite set. Two sets AA and BB are equal, written A=BA = B, exactly when they contain the same elements — order and repeated listing never matter, so {1,2,2,3}={1,2,3}\{1,2,2,3\} = \{1,2,3\}.

Subsets and the power set. AA is a subset of BB, written A⊆BA \subseteq B, if every element of AA also belongs to BB:

A⊆B  ⟺  (x∈A  ⟹  x∈B)A \subseteq B \iff (x \in A \implies x \in B)

where: x∈Ax \in A reads "xx is a member of AA", and   ⟹  \implies is logical implication — the definition says every membership test that passes for AA also passes for BB. If in addition A≠BA \neq B, then AA is a proper subset, written A⊂BA \subset B: BB has at least one element AA lacks. Every set satisfies A⊆AA \subseteq A (reflexivity), and ∅⊆A\varnothing \subseteq A for every set AA, vacuously, since ∅\varnothing has no element that could fail the test. The power set P(A)P(A) collects all the subsets of AA, proper or not, together with AA itself and ∅\varnothing; a set with nn elements has a power set of size 2n2^n (proved below). A universal set UU is the fixed background set — everything under discussion in a given problem — relative to which subsets and complements are defined.

Intervals. An interval is a subset of R\mathbb{R} made of every real number lying between two fixed endpoints aa and bb. Four flavours cover every case, depending on whether each endpoint is included:

(a,b)={x∈R:a<x<b},[a,b]={x∈R:a≤x≤b}(a,b) = \{x \in \mathbb{R} : a < x < b\}, \qquad [a,b] = \{x \in \mathbb{R} : a \le x \le b\}

where: a round bracket excludes the endpoint next to it (strict inequality) and a square bracket includes it (inequality with equality allowed); the same logic gives the two mixed cases (a,b]={x:a<x≤b}(a,b] = \{x : a < x \le b\} and [a,b)={x:a≤x<b}[a,b) = \{x : a \le x < b\}. On a number line, an excluded endpoint is drawn as a hollow (open) circle and an included one as a filled (closed) circle, with the segment between them shaded.

Figure
A number line showing the interval a to b included at a and excluded at b, drawn as a filled circle at a, a hollow circle at b, and a shaded segment between them
The interval [a, b) on the real number line a (included) b (excluded) every real x with a ≤ x less than b
Derivation AWhy a set of n elements has 2^n subsets

The goal is to show, purely by counting choices rather than by listing every case, that a set AA with nn distinct elements has exactly 2n2^n subsets in its power set P(A)P(A) — a fact you will use constantly to check whether you have found every subset of a small set, and which underlies counting arguments in probability later in the course.

Step 1. Build any subset S⊆AS \subseteq A by going through the nn elements of AA one at a time and making a yes/no decision for each: "is this element in SS or not?" Every subset — including ∅\varnothing (answer "no" every time) and AA itself (answer "yes" every time) — corresponds to exactly one sequence of nn yes/no decisions, and every such sequence produces exactly one subset.

Step 2. Count the decisions. Each of the nn elements independently has exactly 22 possible answers, "in" or "out," and the elements are decided independently of one another (including one element never restricts the choice for another).

Step 3. Multiply the choices, since independent yes/no decisions combine by multiplication (the same rule used to count outcomes of tossing nn coins): the total number of decision-sequences, and hence the total number of subsets, is 2×2×⋯×22 \times 2 \times \cdots \times 2 (nn times), i.e.

n(P(A))=2nn(P(A)) = 2^n

Step 4. Check it on a tiny case. For C={1,2}C = \{1,2\}, n=2n = 2, so the formula predicts 22=42^2 = 4 subsets; listing them by hand gives ∅,{1},{2},{1,2}\varnothing, \{1\}, \{2\}, \{1,2\} — exactly four, confirming the count.

Worked example

Problem. Let A={x∈N:x2<30}A = \{x \in \mathbb{N} : x^2 < 30\}. (a) Write AA in roster form and state whether it is finite. (b) Find n(P(A))n(P(A)). (c) A set BB satisfies −3≤x<5-3 \le x < 5 for x∈Rx \in \mathbb{R}; write BB in interval notation.

Step 1. Find every natural number xx with x2<30x^2 < 30: testing x=1,2,3,4,5,6,…x = 1,2,3,4,5,6,\dots gives 12=1, 22=4, 32=9, 42=16, 52=25<301^2=1,\ 2^2=4,\ 3^2=9,\ 4^2=16,\ 5^2=25 < 30, but 62=36≮306^2=36 \not< 30, so the search stops at 55.

Step 2. Collect the results in roster form: A={1,2,3,4,5}A = \{1, 2, 3, 4, 5\}. Counting its elements terminates, so AA is finite, with n(A)=5n(A) = 5.

Step 3. Apply the power-set count from the derivation above, with n=5n = 5:

n(P(A))=25=32n(P(A)) = 2^5 = 32

Step 4. For part (c), the condition −3≤x<5-3 \le x < 5 includes the left endpoint −3-3 (square bracket) and excludes the right endpoint 55 (round bracket), giving B=[−3,5)B = [-3, 5) — a filled dot at −3-3 and a hollow circle at 55 on the number line, exactly the picture in the figure above with a=−3, b=5a=-3,\ b=5.

Final answers: A={1,2,3,4,5}A = \{1,2,3,4,5\} is finite with n(P(A))=32n(P(A)) = 32, and B=[−3,5)B = [-3, 5).

Common traps

  • Writing {0}\{0\} when the empty set is meant. {0}\{0\} contains one element, the number 00; it is not empty. The empty set is only ∅\varnothing or {}\{\}, never {0}\{0\} or {∅}\{\varnothing\}.
  • Confusing ⊆\subseteq (subset, allows A=BA = B) with ⊂\subset (proper subset, forbids A=BA = B). Every set satisfies A⊆AA \subseteq A, but A⊂AA \subset A is always false.
  • Swapping round and square brackets in interval notation. (a,b)(a,b) excludes both endpoints while [a,b][a,b] includes both; one misplaced bracket describes an entirely different set.
  • Undercounting the power set. n(P(A))=2nn(P(A)) = 2^n counts every subset, including ∅\varnothing and AA itself — students often list only the "obviously smaller" subsets and forget these two.

Ask the tutor: Stuck telling apart ⊆\subseteq and ⊂\subset, or converting an inequality into interval notation? Ask the AI tutor to walk through a few practice sets with you.

Why does the empty set count as a subset of every set?

The definition of A⊆BA \subseteq B only fails if you can find an element of AA that is missing from BB. Since ∅\varnothing has no elements at all, there is nothing to check, so the condition can never fail — it holds automatically ("vacuously"). This is why ∅\varnothing appears in the power set of every set, including P(∅)P(\varnothing) itself, which is {∅}\{\varnothing\}, a set with exactly one element.

Is A always a subset of itself, and does that make it a proper subset too?

A⊆AA \subseteq A is always true, because every element of AA trivially belongs to AA. But A⊂AA \subset A (proper subset) is always false, because the proper-subset definition additionally demands A≠BA \neq B, and here A=AA = A. So AA belongs to P(A)P(A) but is never one of its own proper subsets — a distinction examiners test directly.

How do I decide whether to use a round or square bracket at an endpoint given as an inequality?

Look at whether the inequality at that endpoint is strict or not. A strict inequality like x<5x < 5 or x>−3x > -3 excludes the endpoint, so use a round bracket there. An inequality with "or equal to," like x≤5x \le 5 or x≥−3x \ge -3, includes the endpoint, so use a square bracket there. Treat each endpoint separately — a set can easily be closed at one end and open at the other, like B=[−3,5)B = [-3, 5).

Before this

Where this leads