1. Definition of \(L^p\) spaces

1.1 Measure spaces and measurable functions

An \(L^p\) space is built from a measure space \((X,\Sigma,\mu)\). Its elements are (equivalence classes of) measurable functions \(f:X\to\mathbb{R}\) or \(f:X\to\mathbb{C}\) whose “\(p\)-power size” is integrable in a sense made precise below. The measurable structure \(\Sigma\) and the measure \(\mu\) determine which subsets are considered and how large they are, and therefore what it means for \(f\) to be integrable at exponent \(p\).

1.2 The \(L^p\) norm for \(1 \le p < \infty\)

For \(1\le p<\infty\), the \(L^p\) norm of a measurable function \(f\) is \[

\|f\|_p=\left(\int_Xf^p\,d\mu\right)^{1/p},

\]

provided the integral is finite. The defining condition for membership in \(L^p(X)\) is exactly that \(\int_Xf^p\,d\mu<\infty\).

1.3 The essential supremum norm for \(p=\infty\)

For \(p=\infty\), the relevant size is not an integral but an essentially bounded requirement. The essential supremum norm is \[

\|f\|_\infty=\operatorname*{ess\,sup}_{x\in X}f(x),

\]

meaning the smallest number \(M\) such that \(f(x)\le M\) for all \(x\) outside a set of measure zero.

1.4 The role of equivalence classes (almost everywhere)

Because norms for \(p<\infty\) depend on integrals and for \(p=\infty\) on an essential supremum, changing a function on a set of measure zero does not affect its \(L^p\) class. Therefore, \(L^p\) is formed from equivalence classes where \[ f\sim g \quad \text{if and only if}\quad f=g \text{ almost everywhere (a.e.)}. \]

1.5 Modifications of functions on null sets

Given a representative \(f\) of an equivalence class, one may alter \(f\) arbitrarily on any null set \(N\) with \(\mu(N)=0\) without changing its \(L^p\) norm or its membership in \(L^p\). This flexibility is frequently used to enforce convenient pointwise properties (such as choosing versions that agree with limits) when working with convergence, while retaining the same functional-analytic behavior.

2. Basic properties and examples

2.1 Norm axioms and when they hold

For \(1\le p\le \infty\), the quantity \(\|f\|_p\) defines a norm on \(L^p\) (on equivalence classes). The nonnegativity and homogeneity properties follow directly from the definition. The triangle inequality is a consequence of inequalities such as Minkowski’s inequality for \(p<\infty\), and for \(p=\infty\) it reduces to the triangle inequality for essential supremum.

2.2 Completeness: \(L^p\) as a Banach space

With the norm \(\|\cdot\|_p\), each \(L^p\) space is complete: every Cauchy sequence in \(L^p\) converges in \(L^p\) to an element of \(L^p\). This makes \(L^p\) a Banach space, allowing one to use standard results about bounded operators, series, and fixed-point methods in analysis.

2.3 When \(L^p\) is separable

Separability depends on the measure space and \(p\). On many common \(\sigma\)-finite spaces, \(L^p\) is separable for \(1\le p<\infty\), meaning it contains a countable dense subset. Separability is an important feature for approximation, for describing dual spaces, and for building functional-analytic arguments that rely on countable constructions. In contrast, \(L^\infty\) is generally not separable.

2.4 Common examples

2.4.1 Simple functions and indicator functions

Indicator functions \(\mathbf{1}_E\) of measurable sets \(E\subseteq X\) satisfy \[

\|\mathbf{1}_E\|_p=\mu(E)^{1/p}\quad (1\le p<\infty),

\qquad

\|\mathbf{1}_E\|_\infty=1\ \text{if}\ \mu(E)>0.

\] More generally, simple functions—finite linear combinations of indicator functions—are typically dense in \(L^p\) under mild assumptions, and they form a convenient starting point for constructing approximations.

2.4.2 Finite measure vs infinite measure behavior

The measure of the whole space affects inclusion relations and convergence. For example, when \(\mu(X)<\infty\), functions in \(L^q\) for larger \(q\) often belong to \(L^p\) for smaller \(p\). On infinite-measure spaces, these implications can fail because integrability at one exponent does not control behavior over sets of unbounded measure.

3. Relations between \(L^p\) spaces

3.1 Inclusion theorems for different \(p\)

Inclusion between \(L^p\) spaces depends on both the exponent and the measure. Roughly, in finite-measure settings higher integrability (larger \(p\)) implies lower integrability (smaller \(p\)), while on infinite-measure spaces neither inclusion direction is guaranteed without further assumptions.

3.1.1 Finite measure case: monotonicity with \(p\)

If \(\mu(X)<\infty\) and \(1\le p<q\le\infty\), then \(L^q\subseteq L^p\). A standard way to see this is that the factor \(f^p =f^q \,f^{p-q}\) can be controlled using that \(f^{q}\) is integrable and the region where \(f\) is small is manageable on finite measure spaces. The result yields a continuous inclusion with an inequality relating the norms.

3.1.2 Infinite measure subtleties

When \(\mu(X)=\infty\), inclusion can break in both directions. One may find functions that are integrable at a lower exponent but not a higher one, reflecting large-measure regions where decay is insufficient. Conversely, functions may belong to \(L^q\) but not \(L^p\) if their decay is too slow relative to the larger exponent’s requirement.

3.2 Hölder-type comparisons via embeddings

Many inclusion inequalities can be derived from Hölder’s inequality. For example, on finite-measure spaces one often uses the estimate \[

\|f\|_p \le \mu(X)^{\frac{1}{p}-\frac{1}{q}}\|f\|_q \quad (p<q),

\] showing how the norm at exponent \(p\) is bounded by the norm at exponent \(q\), with the dependence on \(\mu(X)\) made explicit.

3.3 Counterexamples and sharpness of inclusions

Sharpness refers to whether stronger inclusions can be expected without additional hypotheses. Typically, without constraints such as finiteness of measure, moment conditions, or weights, one cannot replace a correct inclusion by its converse or by an inclusion at a different threshold exponent. Counterexamples are usually constructed by choosing functions supported on sets with carefully chosen measures or by using power-like decay profiles.

4. Convergence and continuity in \(L^p\)

4.1 Almost everywhere vs \(L^p\) convergence

Almost everywhere convergence is pointwise and disregards what happens on null sets. \(L^p\) convergence is stronger in the sense that \[

\|f_n-f\|_p \to 0

\] forces a particular kind of integrability control on the difference. While almost everywhere convergence may occur without \(L^p\) convergence, many classical theorems provide conditions under which a.e. convergence plus an integrable domination yield convergence in \(L^p\).

4.2 Norm convergence and completeness implications

Norm convergence in \(L^p\) implies convergence in the Banach-space sense, and completeness ensures that limits of Cauchy sequences exist within \(L^p\). This is essential for passing to the limit in analysis: one can interpret convergence of approximating sequences as convergence to an \(L^p\) function rather than merely to a formal expression.

4.3 Convergence theorems

4.3.1 Dominated convergence theorem

The dominated convergence theorem states that if \(f_n\to f\) almost everywhere and there exists an integrable function \(g\) such that \(f_n\le g\) almost everywhere, then \(\|f_n-f\|_1\to 0\). Variants apply to \(L^p\) by considering \(f_n-f^p\) under an appropriate dominating function.

4.3.2 Monotone convergence theorem

If \(f_n\) increases pointwise to \(f\) and all \(f_n\ge 0\), then integrals converge: \[ \int f_n\,d\mu \to \int f\,d\mu. \] This theorem is useful when approximating a function by truncations of increasing sets or by monotone sequences of nonnegative quantities.

4.3.3 Fatou’s lemma and consequences

Fatou’s lemma provides an inequality: \[ \int \liminf_{n\to\infty} f_n\,d\mu \le \liminf_{n\to\infty}\int f_n\,d\mu \] for nonnegative measurable functions. It underlies lower semicontinuity properties and is often used to show that limiting objects preserve integrability bounds.

4.4 Uniform integrability viewpoint (optional framing)

A useful way to think about convergence in \(L^1\) and beyond is in terms of controlling the “mass” of functions on sets where they are large. Uniform integrability captures this idea: it prevents sequences from carrying significant integral mass into rare but increasingly large spikes, enabling convergence in \(L^1\) under weaker hypotheses than pointwise domination.

5.1 Hölder’s inequality

5.1.1 Conjugate exponents \(p\) and \(q\)

For \(1&lt;p&lt;\infty\), the Hölder inequality uses exponents \(p\) and \(q\) satisfying \[ \frac{1}{p}+\frac{1}{q}=1, \] with \(q\) the conjugate exponent. If \(f\in L^p\) and \(g\in L^q\), then \(fg\in L^1\) and \[

\int_Xf g\,d\mu \le \|f\|_p \|g\|_q.

\]

5.1.2 Applications to product functions

Hölder’s inequality is the standard tool for bounding integrals of products. It converts pointwise multiplication into a bound using norms, enabling one to estimate bilinear forms and to prove boundedness of many operators derived from multiplication or convolution-type structures.

5.2 Minkowski’s inequality

Minkowski’s inequality is the triangle inequality in \(L^p\) form: \[

\|f+g\|_p \le \|f\|_p+\|g\|_p \quad (1\le p<\infty).

\] It justifies norm properties and is repeatedly used in establishing estimates for solutions of integral and differential equations expressed in \(L^p\) norms.

5.3 Young’s inequality (integral form)

Young’s inequality relates products of real numbers to sums involving conjugate powers. In integral settings it is used to control expressions like \(f g\) by quantities involving \(f^p\) and \(g^q\), which is helpful in establishing energy-type bounds and in handling nonlinear estimates.

5.4 Generalizations and equality cases

Several refinements exist, including versions for sequences and for weighted exponents. Equality in Hölder typically occurs when \(f^p\) and \(g^q\) are aligned in a specific proportional way (up to sets of measure zero), reflecting optimality of the inequality under the chosen exponents.

6. Duality and functional analytic structure

6.1 Dual space of \(L^p\) for \(1&lt;p&lt;\infty\)

For \(1<p<\infty\), the dual space \((L^p)^*\) is isometrically isomorphic to \(L^q\), where \(q\) is the conjugate exponent. This identification translates continuous linear functionals on \(L^p\) into integration against an \(L^q\) function.

6.2 The pairing \(\langle f,g\rangle\) via integration

Under the duality, a functional determined by \(g\in L^q\) acts as \[ \Lambda_g(f)=\int_X f(x)\,g(x)\,d\mu(x). \] Hölder’s inequality ensures that \(\Lambda_g\) is bounded on \(L^p\), and the norm correspondence makes the identification precise.

6.3 Duality for \(p=1\) and \(p=\infty\)

The cases \(p=1\) and \(p=\infty\) differ from the reflexive range \(1<p<\infty\). While \(L^1\) has a dual that can be described using essentially bounded functionals, the identification is not simply \(L^\infty\) in a norm-isometric sense without additional structure. Similarly, \((L^\infty)^*\) is larger than the “naive” pairing by functions in \(L^1\), reflecting the more complex nature of the dual of \(L^\infty\).

6.4 Reflexivity and its criterion

A Banach space is reflexive if the natural embedding into its double dual is surjective. For \(L^p\) spaces, reflexivity holds precisely for \(1<p<\infty\). This property supports the use of weak compactness arguments, which are central in variational problems and compactness methods.

6.5 Uniform convexity and uniform smoothness

For \(1<p<\infty\), \(L^p\) spaces exhibit uniform convexity, which implies strong geometric control over how sequences behave under weak convergence. Closely related is uniform smoothness, which influences differentiability of norms and the stability of optimization procedures expressed in \(L^p\) settings.

7. Approximation and density results

7.1 Density of simple functions

Simple functions are built from finitely many values on measurable sets and are well suited for approximation. Under standard assumptions on the measure space, simple functions are dense in \(L^p\) for \(1\le p<\infty\). This means every \(L^p\) function can be approximated arbitrarily well in \(L^p\) norm by a sequence of simple functions.

7.2 Density of continuous functions under assumptions

When the measure space has additional topological structure (such as locally compact spaces with suitable measures), continuous functions with compact support can be dense in \(L^p\). Such results depend on the interplay between measure regularity and the available approximation by continuous functions.

7.3 Approximation by truncations and cutoffs

A common technique is truncating a function to control large values and cutting it off on regions of large measure. For instance, define \(f_N = \max(\min(f,N),-N)\). Then \(f_N\to f\) in \(L^p\) as \(N\to\infty\) for \(p<\infty\), provided \(f\in L^p\). Similarly, multiplying by indicator functions of sets with finite measure approximates on spaces where global integrability is subtle.

7.4 Density of \(L^p\cap L^q\) in many settings

In many classical measure spaces, the intersection of two \(L^p\) spaces can be dense in either one, enabling approximation by functions with better integrability properties. This is useful because such functions often serve as test objects for inequalities and operator arguments that require membership in multiple \(L^r\) classes.

7.5 Convergence in norm via truncation schemes

Truncation schemes provide a practical framework: combine local control (via cutoffs) with control of tails (via truncations) to obtain \(L^p\) convergence. The approach is especially common in proofs that start from bounded or compactly supported approximants and then pass to the general case using dominated-convergence-type reasoning.

8. Linear operators on \(L^p\)

8.1 Bounded linear operators and operator norms

A linear operator \(T:L^p\to L^p\) is bounded if there exists \(C\) such that \(\|Tf\|_p\le C\|f\|_p\) for all \(f\). The smallest such constant is the operator norm \(\|T\|\). Boundedness ensures continuity with respect to the \(L^p\) norm and enables the use of standard functional-analytic theorems.

8.2 Hölder-bounded multipliers (pointwise multiplication)

Pointwise multiplication by a function \(m\) defines an operator \(M_m f = mf\). If \(m\in L^\infty\), then \(M_m\) is bounded on every \(L^p\) with \(\|M_m\|=\|m\|_\infty\). More refined multiplier statements exist for other classes of \(m\), depending on additional structure and on which operator family is studied.

8.3 Integral operators with kernels

Integral operators of the form \[ (Tf)(x)=\int_X K(x,y)f(y)\,d\mu(y) \] arise frequently in analysis. Boundedness on \(L^p\) often depends on estimates for the kernel, such as bounds in suitable mixed-norm spaces or conditions ensuring that the integral operator maps \(L^p\) functions to \(L^p\) functions without uncontrolled growth.

8.4 Composition operators and basic criteria

Composition operators, where \((Tf)(x)=f(\phi(x))\) for a measurable transformation \(\phi\), reflect how measure changes under \(\phi\). Determining boundedness typically involves comparing \(\mu\) and the pushforward measure \(\mu\circ\phi^{-1}\), yielding criteria expressed in terms of Radon–Nikodym derivatives or related measurable Jacobian-type quantities.

9. Special cases and variants

9.1 \(L^1\) and \(L^\infty\) characteristics

\(L^1\) emphasizes integrability of absolute values and is closely tied to probabilistic expectations and conservation-type quantities. \(L^\infty\) measures essentially bounded magnitude, relevant to uniform control and stability. Their dual relationship is central, yet their geometric and compactness properties differ markedly from \(L^p\) with \(1<p<\infty\).

9.2 Hilbert space structure of \(L^2\)

The space \(L^2\) becomes a Hilbert space with inner product \[ \langle f,g\rangle=\int_X f(x)\overline{g(x)}\,d\mu(x). \] This inner product structure yields orthogonality concepts, projections, and expansions that underlie many methods in Fourier analysis and in the study of linear operators.

9.3 Weighted \(L^p\) spaces

Weighted versions replace \(\mu\) with a weighted measure \(w\,d\mu\), where \(w\) is a nonnegative measurable function. For \(1\le p&lt;\infty\), \[

\|f\|_{L^p(w)}=\left(\int_Xf^p w\,d\mu\right)^{1/p},

\] and for \(p=\infty\), \[

\|f\|_{L^\infty(w)}=\operatorname*{ess\,sup}f.

\] Weights are used to capture non-uniform behavior, often matching the needs of differential equations or non-homogeneous measures.

9.4 \(L^p\) spaces on sigma-finite measure spaces

On \(\sigma\)-finite measure spaces, \(X\) can be decomposed into countably many measurable pieces of finite measure. This allows many foundational results—such as density of simple functions and workable approximation arguments—to extend beyond finite-measure settings.

10. Tools and common techniques

10.1 Truncation and splitting into regions

Many proofs begin by splitting the domain according to level sets of \(f\): for example, separating where \(f\le N\) from where \(f&gt;N\). Truncation handles large values, while cutoffs restrict to regions where the measure is controlled. This technique converts difficult global statements into manageable local estimates.

10.2 Rearrangement inequalities (overview level)

Rearrangement techniques replace a function by one that is equimeasurable but arranged in a monotone fashion. This can simplify extremal problems involving norms by reducing them to properties of distribution functions. The approach is particularly relevant in harmonic analysis and in inequalities where the measure of level sets matters more than the precise location of values.

10.3 Interpolation ideas (overview level)

Interpolation provides a way to deduce bounds at intermediate exponents from bounds at extreme exponents. Complex and real interpolation methods are classical tools in functional analysis, allowing one to estimate operator norms on \(L^p\) spaces without re-proving the estimate for every \(p\) from scratch.

10.4 Testing convergence using subsequences

In non-compact settings, convergence may be studied through subsequences. Weak convergence, almost everywhere convergence along subsequences, and compactness criteria can be combined to extract convergent subsequences whose limits inherit integrability properties. This method is commonly used in compactness and variational arguments.

11. Further topics (survey pointers)

\(L^p\) spaces underpin many harmonic analysis results, including estimates for Fourier transforms, maximal functions, and singular integral operators. These applications frequently rely on inequality frameworks such as Hölder, Minkowski, and interpolation.

11.2 Connections to PDE energy estimates

In partial differential equations, solutions and their derivatives are often measured in \(L^p\) norms. Energy estimates, a priori bounds, and stability arguments typically use \(L^p\) structure to control nonlinear terms and to propagate integrability across time or space.

11.3 Probabilistic interpretation via moments

In probability, when \(X\) represents a random variable’s underlying space, \(L^p\) norms correspond to moments: for instance, \(\|f\|_p^p=\mathbb{E}f^p\) when \(\mu\) is a probability measure. This viewpoint clarifies how tail behavior influences convergence and integrability.

11.4 Compactness criteria in \(L^p\) (high-level)

Compactness in \(L^p\) is subtler than boundedness. Results such as Rellich-type theorems, compact embeddings, and criteria based on tightness or equicontinuity in appropriate senses can determine when sequences have strongly convergent subsequences. These themes connect \(L^p\) theory to modern variational methods and to the analysis of limiting processes.