What Does It Mean for a Sequence to Be Bounded? Explained
A sequence is bounded when all of its terms stay inside some fixed interval — that is, no term ever escapes beyond a certain lower and upper limit. Formally, a sequence $(a_n)$ is bounded if there exist real numbers $m$ and $M$ such that $m \le a_n \le M$ for every index $n$. So when you ask what does it mean for a sequence to be bounded, the shortest honest answer is: its values are trapped between two finite walls forever, no matter how far down the list you go.
That single idea is the whole definition. The subtleties come from what it does and does not guarantee.
The Core Definition, Precisely
A sequence is a function from the positive integers to the real numbers: $a_1, a_2, a_3, \dots$. Boundedness concerns the range of that function — the set of values it actually takes — not the order in which they appear.
If there is a number $M$ with $a_n \le M$ for all $n$, the sequence is bounded above. If there is a number $m$ with $a_n \ge m$ for all $n$, it is bounded below. It is bounded (or bounded on both sides) when it is bounded above and below.
An equivalent phrasing uses absolute value: the sequence is bounded if there is some $K > 0$ with $|a_n| \le K$ for every $n$. The two versions say the same thing — an interval $[m, M]$ can always be widened symmetrically into $[-K, K]$ by choosing $K$ large enough.
A few quick readings of the definition:
- $a_n = \frac{1}{n}$ is bounded: every term lies in $(0, 1]$, so $m = 0$ and $M = 1$ work.
- $a_n = n$ is not bounded; it grows without limit.
- $a_n = (-1)^n$ is bounded, taking only the values $-1$ and $1$. Oscillating wildly is not the same as growing.
- $a_n = (-1)^n n$ is unbounded, because the magnitudes $1, 2, 3, \dots$ have no ceiling.
Notice that boundedness says nothing about convergence. A bounded sequence may still refuse to settle down.
The Bounds Do Not Have to Be Tight
Beginners often assume the bound must be the exact largest or smallest term. It need not be. For $a_n = \frac{1}{n}$, the value $M = 1$ happens to be attained, but $M = 5$ is also a perfectly valid upper bound. Boundedness asks only whether some finite wall exists — not whether it is the best possible one.
This mirrors the epsilon-style reasoning used throughout analysis, where you routinely pick a constant that is convenient rather than optimal. The Encyclopedia of Mathematics treats boundedness as a property of the set of values, which is exactly why slack in the constants is allowed.
Bounded Above, Bounded Below, and Neither
It helps to separate the two one-sided conditions, since plenty of sequences satisfy only one.
| Sequence | Bounded below? | Bounded above? | Bounded overall? |
|---|---|---|---|
| $a_n = n$ | Yes ($m = 1$) | No | No |
| $a_n = -n$ | No | Yes ($M = -1$) | No |
| $a_n = (-1)^n$ | Yes ($m = -1$) | Yes ($M = 1$) | Yes |
| $a_n = \sin(n)$ | Yes ($m = -1$) | Yes ($M = 1$) | Yes |
| $a_n = n\sin(n)$ | No | No | No |
The monotone case is the cleanest to reason about. If a sequence is increasing and bounded above, or decreasing and bounded below, it must converge — this is the monotone convergence theorem, a cornerstone result in real analysis. Boundedness on the relevant side is precisely what forces the limit to exist rather than drift off to infinity. MIT OpenCourseWare’s analysis materials develop this theorem carefully as the bridge between “stays in a range” and “actually has a limit.”
Why Boundedness Matters
Boundedness is one of the most reusable facts in analysis because it converts a vague question — “could these values run away?” — into a concrete, checkable inequality.
Bounded Does Not Imply Convergent
The sequence $a_n = (-1)^n$ is the standard counterexample. It is bounded, yet it never converges: it alternates between $-1$ and $1$ forever. So two separate questions must be asked of any sequence — is it bounded, and does it converge? Boundedness alone answers only the first.
What is true is a one-way implication: every convergent sequence is bounded. If the terms must eventually crowd around a single limit $L$, then beyond some index they all lie in, say, $(L-1, L+1)$, and the finitely many earlier terms can only push the overall bounds out finitely far. This is a standard early result in any real analysis course.
The Bolzano–Weierstrass Theorem
Boundedness buys something even when convergence fails. The Bolzano–Weierstrass theorem states that every bounded sequence of real numbers has a convergent subsequence. The alternating sequence $(-1)^n$ has no limit, but the subsequence at even indices is constantly $1$ and converges to $1$.
This is the reason bounded sequences are so valuable: even when the full sequence misbehaves, you can always extract an orderly piece. The Encyclopedia of Mathematics records this as one of the defining features of compactness in the real line.
A Worked Example
Take $$a_n = \frac{3n + 2}{n + 1}.$$
To check boundedness, rewrite it: $a_n = 3 - \frac{1}{n+1}$. The fraction $\frac{1}{n+1}$ is positive and never exceeds $\frac{1}{2}$ for $n \ge 1$. Therefore every term satisfies
$$\frac{5}{2} \le a_n < 3.$$
So the sequence is bounded below by $\frac{5}{2}$ and above by $3$. It is also increasing and bounded above, so by the monotone convergence theorem it converges — in fact to $3$. One manipulation delivered both the bound and the limit.
Checking Boundedness in Practice
For most sequences you meet, a short argument suffices:
- Constant or finite range: if the sequence takes only finitely many values, it is automatically bounded.
- Rational expressions: divide numerator and denominator by the highest power of $n$, or rewrite as a constant plus a vanishing fraction.
- Trigonometric terms: $\sin(n)$ and $\cos(n)$ always lie in $[-1, 1]$, so any sequence built only from them is bounded.
- Products that mix: $n\sin(n)$ is unbounded because the growing factor $n$ overwhelms the bounded one. Look for an unbounded factor when a sequence combines both kinds of behavior.
- Alternating signs: check the magnitudes. Signs alone never make a sequence unbounded, and they never make it convergent either.
Key Takeaways
- A sequence is bounded if there exist finite numbers $m$ and $M$ with $m \le a_n \le M$ for every term.
- Equivalently, $|a_n| \le K$ for some positive constant $K$; the bounds need not be tight.
- A sequence can be bounded above only, bounded below only, or bounded on both sides.
- Every convergent sequence is bounded, but bounded sequences such as $(-1)^n$ need not converge.
- By the Bolzano–Weierstrass theorem, every bounded sequence of real numbers contains a convergent subsequence.
- A monotone sequence that is bounded on the side it moves toward must converge.
