Connectedness and intermediate values
Express the absence of a separation and use continuous images to explain why intermediate values cannot be skipped.
On this page
The question for today
What you’ll understand
- Recognize a separation using relatively open subsets.
- Prove that the continuous image of a connected space is connected.
- Apply the intermediate value principle without confusing existence with uniqueness.
Connectedness captures an inability to split a space into two separated open pieces. On the real line this becomes the familiar property of an interval: whenever it contains two numbers, it contains every number between them. A continuous real-valued measurement on a connected domain inherits that no-skipping property. This is the structural reason for the intermediate value theorem, and it works even when the original domain has no meaningful derivative or straight-line coordinate description.
A separation is relative to the space
A separation of X consists of two disjoint nonempty subsets that are open relative to X and together cover X. If no separation exists, X is connected. Because the two sets are complements in X, each is also relatively closed. For X=[0,1] union [2,3], intersecting with the ambient open sets (−1,1.5) and (1.5,4) produces a separation. The two resulting pieces do not need to be open in the real line; the relevant openness is measured within X. This is another reason to state the ambient or relative space carefully.
A connected space has no nontrivial subset that is both open and closed: such a subset and its complement would provide a separation. Conversely, a separation immediately supplies such a subset. The empty set and the whole space are always both open and closed, so their existence says nothing about disconnection. In a discrete space with at least two points, a singleton is a nontrivial open-and-closed subset and separates the space. By contrast, an interval's arbitrarily close neighboring points prevent this kind of clean split.
Continuous images cannot create a separation
Suppose a continuous image f(X) were separated into two nonempty relatively open pieces. Their inverse images would be open in X by continuity, nonempty because the pieces contain image points, disjoint, and together equal X. They would separate X. This contradiction proves that the continuous image of a connected space is connected. The proof parallels the compact-image argument but transfers a different property. A compact space can be disconnected, and a connected space can be noncompact, so the two conclusions must not be treated as interchangeable.
Connected subsets of the real line are exactly intervals, with singletons and the empty set included under the usual convention. If a subset contains a and b but omits a number c strictly between them, its intersections with the two open half-lines separated by c form a separation. The converse uses the order completeness of the real line: a proposed separation of an interval would force a boundary point that cannot belong consistently to either relatively open piece. This order argument is what the informal instruction to draw without lifting a pencil leaves unproved.
Paths provide a sufficient condition, not every notion of continuity
A path from x to y is a continuous map from [0,1] into X whose endpoint values are x and y. A space is path connected when every pair can be joined this way. Path connectedness implies connectedness, because a separation would pull back along a path to separate the interval. In general metric spaces the converse can fail, so connectedness should not be defined by the existence of paths. Convex subsets of Euclidean space are path connected because the straight segment between any two of their points stays inside the set.
The intermediate value conclusion guarantees existence, not uniqueness, numerical stability, or an efficient root-finding algorithm. A continuous function may cross the same level many times or remain at that level on an interval. A sign change across endpoints of a real interval guarantees a zero, but absence of a sign change does not rule one out: a square can touch zero and stay nonnegative. A discontinuous function can jump from negative to positive without taking zero. Each use should therefore name both the connected domain and the continuity assumption.
Work it through
Certify a root before approximating it
Show that h(x)=x³+x−1 has exactly one root in [0,1]. Separate the existence argument from the uniqueness argument.
The polynomial h is continuous on the connected interval [0,1]. Its values are h(0)=−1 and h(1)=1, so the intermediate value theorem supplies at least one zero in the interval.
For y>x, subtract the function values: h(y)−h(x)=(y−x)(y²+xy+x²+1). The second factor is positive because y²+xy+x² is nonnegative for real x,y.
Thus h is strictly increasing, so it cannot take the value zero at two distinct inputs. This additional monotonicity argument establishes uniqueness; it was not part of the intermediate value theorem.
Exactly one root lies in [0,1]. Continuity and connectedness prove existence, while strict increase proves uniqueness.
The living notebook
Explore the model
Interactive lab · function
A level may be attained more than once
Set coefficient a=1 and move x between −1 and 1 in the quadratic lab. Predict the two positions with height one, and compare their opposite tangent slopes. This illustrates a continuous interval image and repeated level values for ax²; it does not display the cubic from the worked example or automatically certify roots.
- y at chosen x
- 0.608
- Tangent slope · 2ax
- 1.35
- Signed area · ax³/3
- 0.182
This lab uses only y = ax². Playback traces the curve within the fixed integration interval from 0 to x; it does not estimate numerical error or change the chosen point. Parameter changes pause playback and recompute the model at the current elapsed time. Reset restores both parameters and returns time to zero. Playback starts paused. Adjust either parameter, then choose Play or Step.
Look forIntermediate values are guaranteed by continuity on a connected domain, while uniqueness needs another argument.
Pause & reason
Check your understanding
Your turn
Try it yourself
Let X=[−2,−1] union [1,2] and f(x)=x. The function is continuous and takes negative and positive values. Explain why the intermediate value principle does not force a zero.
Take a moment to make a prediction or work it out on paper. Open a hint when you need one.
Give me a hint
Identify a separation of the domain and inspect the image directly.
Compare with the solution
The domain has two relatively open nonempty pieces, its negative and positive intervals, separated by the gap around zero. Thus X is disconnected. The identity function is continuous, but its image is X itself and omits zero. The intermediate value theorem requires a connected domain, so its hypotheses are not satisfied. Continuity alone cannot fill a gap that was already present in the domain.
Keep the idea
Take it with you
- Connectedness excludes a separation into two nonempty relatively open pieces.
- Continuous images preserve connectedness and therefore cannot skip intermediate real values.
- Root existence, root uniqueness, and numerical root finding are distinct claims.
Read further
Original sources and deeper explanations to follow at your own pace.
Optional progress, saved in this browser when available. Revisit any lesson at any time.
Lesson not marked complete.