1 Definition and Basic Properties
1.1 Absolute continuity on an interval
Let \(f:[a,b]\to \mathbb{R}\). The function \(f\) is absolutely continuous on \([a,b]\) if for every \(\varepsilon>0\) there exists \(\delta>0\) such that for any finite collection of pairwise disjoint intervals \((x_k,y_k)\subset[a,b]\), \[
| \sum_k (y_k-x_k)<\delta \quad \Longrightarrow \quad \sum_k | f(y_k)-f(x_k) | <\varepsilon. |
|---|
\] Intuitively, the total amount of “change” in \(f\) across a collection of intervals whose total length is small can be forced to be arbitrarily small.
1.2 Modulus of absolute continuity (ε–δ formulation)
The definition can be viewed through a modulus of absolute continuity. For each \(\varepsilon>0\), one may choose a corresponding \(\delta\); often one considers the best (largest) admissible \(\delta\) or, equivalently, the smallest guaranteed variation over sets of given total length. This perspective is useful when comparing families of functions, estimating continuity under perturbations, or proving closure properties.
1.3 Immediate consequences: continuity and boundedness on compact sets
| Absolute continuity implies ordinary continuity. Indeed, if \(x_n\to x\) in \([a,b]\), then the intervals \((\min\{x_n,x\},\max\{x_n,x\})\) have lengths tending to \(0\), forcing \( | f(x_n)-f(x) | \to 0\). Moreover, absolute continuity yields boundedness on \([a,b]\): the function cannot exhibit arbitrarily large oscillations on sets with small total length, and partitioning \([a,b]\) into sufficiently many small intervals provides a uniform bound. |
|---|
1.4 Relationship to functions of bounded variation
Functions of bounded variation satisfy a global control on total variation over the whole interval. Absolute continuity is stronger: every absolutely continuous function has bounded variation, but not conversely. The difference is that bounded variation allows singular components (in the sense of decomposition), while absolute continuity eliminates such singular behavior by forcing the variation over small sets to be small.
2 Equivalent Characterizations
2.1 ε–δ formulation in terms of sets
A formulation using arbitrary measurable sets is equivalent to the interval-based one. Roughly, absolute continuity can be expressed as: for every \(\varepsilon>0\) there exists \(\delta>0\) such that whenever a set \(E\subset[a,b]\) has small Lebesgue measure, the total variation of \(f\) across \(E\) is small in an appropriate sense. One convenient way to implement this uses covering by disjoint intervals and then passes to the limit.
2.2 Measure-theoretic formulation (null sets)
Another equivalent viewpoint is measure-theoretic: if a measurable set \(E\subset[a,b]\) has Lebesgue measure zero, then the function’s induced variation on \(E\) vanishes. In particular, an absolutely continuous function behaves regularly with respect to null sets, contrasting with bounded-variation functions that can change on sets of measure zero in a nontrivial way.
2.3 Integral representation via derivatives
A central characterization is that an absolutely continuous function admits a representation in terms of its derivative:
- \(f'\) exists almost everywhere on \([a,b]\),
- \(f'\in L^1([a,b])\),
- and for all \(x\in[a,b]\),
\[ f(x)=f(a)+\int_a^x f'(t)\,dt. \] This ties absolute continuity directly to the fundamental theorems of calculus in an “almost everywhere” setting.
2.4 Fundamental theorem of calculus for absolutely continuous functions
For absolutely continuous \(f\), the fundamental theorem of calculus takes a robust form: \(f'\) is integrable, and integrating \(f'\) reconstructs \(f\) up to the initial value. Conversely, if one starts with an integrable function \(g\in L^1([a,b])\) and defines \(F(x)=F(a)+\int_a^x g(t)\,dt\), then \(F\) is absolutely continuous and \(F'=g\) almost everywhere.
3 Differentiability and Derivatives
3.1 Existence of derivative almost everywhere
Absolute continuity ensures that the derivative exists almost everywhere. While the derivative may fail to exist at exceptional points, those points form a set of Lebesgue measure zero. This places absolutely continuous functions between purely continuous ones and differentiable ones, capturing “differentiability with respect to measure.”
3.2 Recovery formula: f(x) = f(a) + ∫ f'(t) dt
The recovery property states that the function can be recovered by integrating its almost-everywhere derivative. For any \(x\in[a,b]\), \[ f(x)=f(a)+\int_a^x f'(t)\,dt. \] In this sense, the derivative is not merely a local slope notion; it acts as an integrable density governing the global change of \(f\).
3.3 Comparison with classical differentiability and Lipschitz continuity
- Classical differentiability everywhere implies absolute continuity only under additional control (e.g., integrability of the derivative and suitable behavior). In contrast, absolute continuity allows non-differentiability on measure-zero sets while still permitting the calculus machinery.
| - Lipschitz continuity implies absolute continuity, since controlling \( | f(y)-f(x) | \le L | y-x | \) automatically gives the interval-sum condition with \(\delta=\varepsilon/L\). |
|---|
Thus, the hierarchy is typically: Lipschitz \(\Rightarrow\) absolutely continuous \(\Rightarrow\) continuous, with absolute continuity offering a stronger structural conclusion than continuity alone.
3.4 Derivative as an L¹ function
For an absolutely continuous \(f\), the derivative \(f'\) belongs to \(L^1([a,b])\). This integrability is essential for reconstruction and for stability under operations like limits. The \(L^1\) nature reflects that absolute continuity controls total accumulated change rather than pointwise smoothness.
4 Variational and Integrability Aspects
4.1 Absolute continuity vs. uniform continuity
Absolute continuity strengthens uniform continuity. Uniform continuity only requires small changes in input to force small changes in output, without accounting for how changes accumulate over multiple disjoint intervals. Absolute continuity quantifies accumulated variation: many small oscillations across separated regions are still globally controlled when their total length is small.
4.2 Absolute continuity and integrability of the derivative
The connection between absolute continuity and integrability is two-sided:
- If \(f\) is absolutely continuous, then \(f'\) exists almost everywhere and is integrable.
- If one starts with an integrable function \(g\) and integrates it, the resulting primitive is absolutely continuous and has derivative \(g\) almost everywhere.
This equivalence makes absolute continuity a natural bridge between variational behavior and integration theory.
4.3 Total variation bounds and monotone components
Absolutely continuous functions have finite total variation. Moreover, they can be connected to monotone components through variation measures: while general functions of bounded variation decompose into absolutely continuous, jump, and singular parts, the absolute continuity condition suppresses the singular and jump contributions. As a result, the variation is governed by the integrable derivative.
4.4 Interaction with indefinite integrals
Indefinite integrals of \(L^1\) functions produce absolutely continuous primitives. This is particularly useful in analysis because it allows treating solutions of differential equations in integral form: once an integrand is known to be integrable, the resulting accumulated quantity inherits absolute continuity.
5 Stability, Closure, and Convergence Results
5.1 Closedness under addition and scalar multiplication
The class of absolutely continuous functions is a vector space. If \(f\) and \(g\) are absolutely continuous on \([a,b]\), then so are:
- \(f+g\),
- \(\alpha f\) for any scalar \(\alpha\),
and the corresponding \(\varepsilon\)–\(\delta\) controls can be combined using triangle inequalities.
5.2 Composition rules (under suitable assumptions)
Compositions preserve absolute continuity when the outer and inner functions have compatible regularity. Typical statements assume:
- \(f\) is absolutely continuous,
- the outer function is Lipschitz on the range of \(f\), or has controlled variation,
- and continuity of the relevant composition holds.
Under such conditions, the absolute continuity condition passes through via estimates on increments and the modulus of absolute continuity.
5.3 Uniform limits and preservation of absolute continuity
Absolute continuity is not automatically preserved under uniform convergence without further assumptions. However, it is stable under uniform convergence together with uniform integrability-type control on derivatives (or suitable uniform absolute continuity bounds). In practice, one often uses criteria involving convergence in \(L^1\) of derivatives or boundedness of associated variation measures.
5.4 Convergence theorems: dominated convergence and subsequence behavior
When \(f_n\) are absolutely continuous and their derivatives \(f_n'\) satisfy integrability conditions, one can obtain convergence of primitives. A common route is:
- prove pointwise or almost-everywhere convergence of derivatives,
- use dominated convergence or related compactness ideas in \(L^1\),
- conclude convergence of the integrals \(\int_a^x f_n'\).
This yields subsequence behavior and limits that remain absolutely continuous, provided the limiting derivative is integrable.
6 Examples and Non-Examples
6.1 Smooth functions as absolutely continuous functions
If \(f\) is continuously differentiable on \([a,b]\), then it is absolutely continuous. More generally, if \(f'\) is integrable and \(f\) satisfies the fundamental theorem of calculus in the classical sense, then absolute continuity follows. Smoothness provides strong enough control so that small interval lengths lead to small increments.
6.2 Functions defined by integrals of L¹ densities
Let \(g\in L^1([a,b])\) and define \[ F(x)=F(a)+\int_a^x g(t)\,dt. \] Then \(F\) is absolutely continuous. These functions are standard test cases because their regularity is directly tied to integrability of the density \(g\).
6.3 Cantor-type examples and failure of absolute continuity
There exist continuous functions of bounded variation whose derivatives vanish almost everywhere yet the functions are not constant; these are often constructed from singular measures (e.g., Cantor-type functions). Such examples fail absolute continuity because they exhibit nontrivial change concentrated on sets of measure zero. They demonstrate why absolute continuity cannot be inferred merely from almost-everywhere derivative behavior.
6.4 Piecewise smooth functions and jump discontinuities
If \(f\) is piecewise smooth on \([a,b]\) with only jump discontinuities, then \(f\) is generally not absolutely continuous, because absolute continuity implies continuity. If instead the function is piecewise \(C^1\) and matches continuously at the junction points, then it becomes absolutely continuous, since its derivative is integrable and the primitive representation holds.
7 Connections and Applications
7.1 Absolutely continuous paths in analysis and probability
In probability and stochastic analysis, one often studies trajectories that satisfy deterministic integral constraints. Absolutely continuous paths are natural because they allow interpreting increments via an integrable “velocity” (the derivative almost everywhere), while still permitting exceptional nondifferentiable times.
7.2 Role in ordinary differential equations (integral form of solutions)
Many existence and uniqueness frameworks for ordinary differential equations use integral formulations. If a solution \(x(t)\) satisfies \[ x(t)=x(a)+\int_a^t v(s)\,ds \] for some integrable \(v\), then \(x\) is absolutely continuous and \(x'(t)=v(t)\) almost everywhere. This connects analytic solvability with regularity properties.
7.3 Change of variables and chain rule under absolute continuity
Absolute continuity supports versions of the chain rule. If \(f\) is absolutely continuous and another function is differentiable in a compatible sense, then the composition often admits an almost-everywhere derivative given by multiplying the relevant derivatives. The precise assumptions ensure that the resulting derivative is integrable, allowing the integral representation to remain valid.
7.4 Extension toward Sobolev spaces (W^{1,1}) and related notions
In Sobolev space theory, absolute continuity is closely related to membership in \(W^{1,1}\) on an interval. In one dimension, \(W^{1,1}\) functions admit absolutely continuous representatives, and their weak derivative coincides with the almost-everywhere derivative of that representative. This creates a practical bridge from measure-based regularity to weak differentiability.
8 Further Topics
8.1 Absolute continuity on subintervals and localization
Absolute continuity is local in nature: if \(f\) is absolutely continuous on \([a,b]\), then its restriction to any subinterval \([c,d]\subset[a,b]\) is absolutely continuous. Conversely, under mild compatibility conditions, absolute continuity on a covering by subintervals can yield absolute continuity on the whole interval.
8.2 Absolute continuity with respect to measures
The phrase “absolute continuity” also appears in measure theory. One function can be considered absolutely continuous with respect to another measure when changes of the reference measure being zero force changes of the function-associated measure to vanish. This concept aligns with the Radon–Nikodym theorem and provides a common language for densities and derivatives.
8.3 Links to Lipschitz and Sobolev regularity hierarchies
Absolute continuity occupies a position in a regularity ladder: stronger than mere continuity and weaker than uniform Lipschitz control in general. It also interfaces with Sobolev hierarchies, where integrability of derivatives determines which functions belong to which function spaces. These relationships help classify regularity needed for particular analytic arguments.
8.4 Radon–Nikodym perspective (derivatives of measures)
A conceptual framework views an absolutely continuous function as associated with a measure whose “derivative” with respect to Lebesgue measure exists. The integrable derivative \(f'\) plays the role of a density in the Radon–Nikodym sense. This perspective unifies absolute continuity of functions with the absolute continuity of measures and explains why singular behavior corresponds to failure of absolute continuity.