1 Definitions and Basic Ideas
1.1 One-sided limits
For a function \(f\) defined near a real number \(a\), one-sided limits describe the behavior of \(f(x)\) as \(x\) approaches \(a\) from only one side. The right-hand limit \(\lim_{x\to a^+} f(x)\) tracks values of \(x>a\) tending to \(a\). The left-hand limit \(\lim_{x\to a^-} f(x)\) tracks values of \(x<a\) tending to \(a\). If the corresponding limit exists as a finite real number, it summarizes the “incoming” behavior from that side.
1.2 Left-continuity
A function \(f\) is left-continuous at a point \(a\) (provided \(a\) is in the domain) if the function value equals the left-hand limit: \[ \lim_{x\to a^-} f(x)=f(a). \] This condition ensures that approaching \(a\) from smaller arguments matches the stated value at \(a\).
1.3 Right-continuity
A function \(f\) is right-continuous at a point \(a\) if \[ \lim_{x\to a^+} f(x)=f(a). \] It guarantees that approaching from larger arguments yields the value at \(a\).
1.4 One-sided continuity at boundary points
When \(a\) is an endpoint or boundary point of the domain, only one side may be available. For example, if \(f\) is defined on \([a,\infty)\), there is no meaningful approach from the left inside the domain, so right-continuity is the appropriate notion at \(a\). Similarly, for \(f\) defined on \((-\infty,a]\), left-continuity is the relevant endpoint condition.
1.5 Relation to standard (two-sided) continuity
Standard continuity at an interior point \(a\) requires both one-sided limits to exist and agree with the function value: \[ \lim_{x\to a^-} f(x)=f(a)=\lim_{x\to a^+} f(x). \] Thus, two-sided continuity is equivalent to having both left- and right-continuity at \(a\). At boundary points, only the relevant one-sided continuity can be required.
2 Characterizations and Equivalent Conditions
2.1 Epsilon–delta formulations
One-sided continuity admits an epsilon–delta description analogous to ordinary continuity. For right-continuity at \(a\), the requirement is: for every \(\varepsilon>0\), there exists \(\delta>0\) such that \[
| 0<x-a<\delta \quad \Rightarrow \quad | f(x)-f(a) | <\varepsilon. |
|---|
\] For left-continuity, the implication becomes \[
| -\delta<x-a<0 \quad \Rightarrow \quad | f(x)-f(a) | <\varepsilon. |
|---|
\] These formulations formalize that \(f(x)\) must approach \(f(a)\) when \(x\) approaches from the specified side.
2.2 Sequential (one-sided) characterizations
One-sided continuity can also be expressed via sequences. Right-continuity at \(a\) is equivalent to: for every sequence \((x_n)\) in the domain with \(x_n>a\) and \(x_n\to a\), one has \(f(x_n)\to f(a)\). The left-continuous version uses sequences with \(x_n<a\) and \(x_n\to a\). This approach is frequently used in proofs because it converts limit statements into convergence of values along selected sequences.
2.3 Neighborhood-based viewpoints
| In neighborhood language, right-continuity at \(a\) means that values of \(f(x)\) for \(x\) sufficiently close to \(a\) with \(x\) greater than \(a\) lie within any prescribed tolerance of \(f(a)\). Concretely, for each \(\varepsilon>0\), there exists \(\delta>0\) such that for all \(x\) in the intersection of the domain with \((a,a+\delta)\), the inequality \( | f(x)-f(a) | <\varepsilon\) holds. The left-continuity counterpart uses \((a-\delta,a)\). |
|---|
2.4 Connections with existence of one-sided limits
A common viewpoint is: left-continuity at \(a\) is exactly the statement that the left-hand limit exists and equals \(f(a)\). Similarly, right-continuity at \(a\) is exactly that the right-hand limit exists and equals \(f(a)\). This makes one-sided continuity tightly connected to limit existence theorems for functions with jump behavior or piecewise definitions.
3 Examples and Typical Patterns
3.1 Continuous functions and matching one-sided limits
If \(f\) is continuous at an interior point \(a\), then it is automatically both left- and right-continuous there. In such cases the left-hand and right-hand limits exist, are equal, and coincide with \(f(a)\). Typical smooth functions, polynomials, and trigonometric expressions illustrate this behavior, with no mismatch between approaches from either direction.
3.2 Functions with jump discontinuities
Consider a function with different limiting values from left and right. For instance, \[ f(x)=\begin{cases} 0, & x<0,\\ 1, & x\ge 0. \end{cases} \] At \(a=0\), the left-hand limit is \(0\) while the right-hand limit is \(1\). Since neither one-sided limit matches the function value simultaneously from both sides (and the limits differ), the function is not two-sided continuous at \(0\). However, it is right-continuous at \(0\) because \(\lim_{x\to 0^+} f(x)=f(0)=1\), while it is not left-continuous because \(\lim_{x\to 0^-} f(x)=0\ne f(0)\).
3.3 Removable discontinuities seen one-sided
A removable discontinuity can still appear as one-sided mismatch if only one side is incorrect. For example, \[ f(x)=\begin{cases} g(x), & x\ne 0,\\ h, & x=0, \end{cases} \] where \(g(x)\to L\) as \(x\to 0\), but \(h\ne L\). Then neither left- nor right-continuity holds if both one-sided limits equal \(L\). If, however, the limit behavior differs by side, one may have one-sided continuity even when there is a discontinuity overall. This demonstrates that “removable” as a global description can coexist with one-sided continuity depending on the sided limits.
3.4 Piecewise-defined functions
Piecewise formulas are a primary setting for one-sided continuity. For instance, define \[ f(x)=\begin{cases} x, & x\le 1,\\ 2x-1, & x>1. \end{cases} \] At \(a=1\), the left-hand limit is \(\lim_{x\to 1^-} x=1\), matching \(f(1)=1\), so the function is left-continuous at \(1\). The right-hand limit is \(\lim_{x\to 1^+} (2x-1)=1\), also matching \(f(1)\), so it is right-continuous at \(1\) as well, hence fully continuous at that point. If the second branch were adjusted so its limit were different, one-sided continuity would fail on the affected side.
3.5 Endpoint behavior and domain restrictions
When the domain is restricted, one-sided continuity often becomes the natural statement. Suppose \(f\) is defined on \([0,1]\). At the endpoint \(0\), only right-continuity is meaningful because there is no domain region to the left of \(0\). At the endpoint \(1\), only left-continuity is meaningful. In applied contexts, such as models defined only for nonnegative time, enforcing right-continuity at \(t=0\) is standard practice because it ensures the model starts consistently with what happens immediately after time begins.
4 Properties and Theorems
4.1 Algebra of one-sided continuous functions
One-sided continuity behaves well under basic arithmetic operations, mirroring the corresponding results for two-sided continuity.
4.1.1 Sums of one-sided continuous functions
If \(f\) and \(g\) are right-continuous at \(a\), then \(f+g\) is right-continuous at \(a\). The reason is that limits preserve addition: \(\lim_{x\to a^+}(f(x)+g(x))=\lim_{x\to a^+}f(x)+\lim_{x\to a^+}g(x)=f(a)+g(a)\). The same statement holds with “right” replaced by “left.”
4.1.2 Products and scalar multiples
If \(f\) is right-continuous at \(a\) and \(g\) is right-continuous at \(a\), then their product \(fg\) is right-continuous at \(a\), since \(\lim (fg)= (\lim f)(\lim g)\) for finite limits. Scalar multiples follow immediately: for constant \(c\), \(cf\) is right-continuous at \(a\) because \(\lim_{x\to a^+} cf(x)=c\lim_{x\to a^+} f(x)=cf(a)\). Analogous statements hold for left-continuity.
4.1.3 Quotients under nonvanishing conditions
If \(f\) and \(g\) are right-continuous at \(a\), and additionally \(g(a)\ne 0\) (and, in a limit sense, \(g(x)\) stays away from \(0\) sufficiently close to \(a\) from the relevant side), then the quotient \(f/g\) is right-continuous at \(a\). The key requirement is that the one-sided limit \(\lim_{x\to a^+} g(x)\) equals \(g(a)\ne 0\), so division is well-behaved near \(a\). The left-continuity version is analogous.
4.2 Composition with continuous functions
If \(f\) is right-continuous at \(a\) and \(\phi\) is a continuous function (two-sided continuity in its domain), then \(\phi\circ f\) is right-continuous at \(a\). The one-sided limit passes through \(\phi\) because \(\lim_{x\to a^+}\phi(f(x))=\phi(\lim_{x\to a^+} f(x))=\phi(f(a))\). For left-continuity, the same argument applies with the approach restricted to \(x\to a^-\).
4.3 Intermediate value behavior via one-sided continuity
A useful consequence of one-sided continuity appears in limit-based arguments for values near a point. While one-sided continuity alone does not fully reproduce the classical intermediate value theorem (which depends on connectedness and continuity), it supports “approach” versions: if \(f\) is one-sided continuous at \(a\) and its nearby values cross a level along the approach side, one can often conclude the existence of points where \(f\) takes intermediate values, using connectedness of an interval on that side.
4.4 Limits of monotone functions and one-sided continuity
For monotone functions on an interval, one-sided limits exist at every point in the interior and also at endpoints within the domain. This structural fact frequently pairs with one-sided continuity: if a monotone function has a one-sided limit at \(a\), verifying that the function value equals that limit establishes one-sided continuity. Many classical results in real analysis about monotone functions therefore naturally involve one-sided behavior.
4.5 Preservation under uniform convergence (one-sided context)
Uniform convergence can preserve one-sided continuity properties under suitable hypotheses. If a sequence of functions converges uniformly to \(f\) on a domain region near \(a\), and each function in the sequence is right-continuous at \(a\), then \(f\) is right-continuous at \(a\) provided the one-sided limit operations remain compatible with uniform control. A typical proof uses the uniform bound to interchange limiting procedures: the values of \(f_n(x)\) stay close to \(f(x)\) uniformly for \(x\) near \(a\) from the right, while \(f_n(x)\) approach \(f_n(a)\) by right-continuity.
5 One-sided Continuity in Analysis Practice
5.1 Handling discontinuities in proofs
In many arguments, discontinuities do not invalidate reasoning if the needed limit direction is specified. One-sided continuity allows proofs to proceed by selecting the correct approach (left or right) depending on the structure of the problem. For example, when analyzing behavior of a process defined by a rule that switches at \(a\), statements about what happens immediately after or immediately before \(a\) correspond exactly to right- or left-continuity.
5.2 Verifying continuity for piecewise models
Piecewise definitions are often checked at junction points by comparing the formula outputs with the one-sided limits from each side. The verification procedure is systematic: compute the limiting value from the left branch and compare it with the defined value at the junction for left-continuity; then compute the limiting value from the right branch for right-continuity. This prevents confusion caused by global continuity assumptions when only one side is required by the model.
5.3 Graphical interpretation (left/right “limits”)
On a graph, left-continuity at \(a\) can be visualized as the curve’s approach from the left meeting the plotted point at \(a\). Right-continuity is the analogous statement about the approach from the right. For jump discontinuities, the graph reveals two distinct approaching heights, and at most one of them may match the function’s value at the jump point.
5.4 Jump size and one-sided limit mismatch
The magnitude of a jump discontinuity at \(a\) can be measured using the difference between the one-sided limits: \[ \text{jump}=\lim_{x\to a^+} f(x)-\lim_{x\to a^-} f(x). \] If this quantity is nonzero, the function cannot be continuous at \(a\). One-sided continuity may still hold if, for example, \(f(a)\) matches the right limit but not the left limit. In such cases, the jump size quantifies the mismatch between behaviors across the point, while one-sided continuity determines whether the function’s declared value agrees with one side of the approach.
5.5 Using one-sided continuity to extend functions
One-sided continuity is frequently used to define or extend functions at boundary points. Given a function on \((a,b]\) with a right-hand limit at \(a\), one can define \(f(a)\) to equal that right-hand limit to obtain a right-continuous extension to include \(a\). Similarly, left-continuous extension can be produced using the left-hand limit. This technique is common in constructing functions from incomplete data near endpoints.
6 Advanced and Related Concepts
6.1 One-sided differentiability versus one-sided continuity
One-sided differentiability is stronger than one-sided continuity. If \(f\) is right-differentiable at \(a\) (the limit defining the right derivative exists), then \(f\) must be right-continuous at \(a\) because differentiability implies continuity along the same approach direction. The converse is not true in general: a function can be right-continuous without having a right derivative.
6.2 Dini derivatives and related one-sided notions
Dini derivatives provide refined one-sided measurements of growth using limsup and liminf of difference quotients. These quantities capture directional or one-sided “upper” and “lower” slopes without requiring an actual derivative to exist. They are widely used in differential inequalities and regularity questions where standard derivatives fail due to oscillations or nondifferentiability while still maintaining one-sided structural constraints.
6.3 Cadlag/caglad functions and one-sided limits (brief overview)
In probability and stochastic processes, functions with well-structured one-sided limits are categorized by regularity conventions. “Càdlàg” (right-continuous with left limits) means the function is right-continuous everywhere and has finite left limits at every point. “Càglàd” (left-continuous with right limits) reverses the roles. These classes are designed to accommodate jump behavior while maintaining a consistent form of one-sided control.
6.4 Generalizations to topological spaces (left/right via order topology)
The idea of one-sided continuity can be expressed in more abstract settings by using the topology induced by an order. In a linearly ordered set (like the real numbers), one can define one-sided neighborhoods of a point that include elements strictly greater than \(a\) or strictly less than \(a\). A function between such ordered spaces can then be studied for continuity with respect to these sub-neighborhood filters. This generalization retains the core principle: continuity is checked using limits along a specified approach direction determined by the order structure.