Ma · Mathematics

How to Prove the Triangle Inequality for the L2 Norm

Published By the 9to5Science editorial team

At a glance

  1. 1. State the setting

    Take two measurable functions whose squared absolute values have finite integrals on the same measure space.

  2. 2. Expand the square

    Write the squared absolute value of their sum and bound its cross term using the absolute value.

  3. 3. Apply Cauchy–Schwarz

    Bound the integral of the product by the product of the two L2 norms.

  4. 4. Take square roots

    The resulting bound is the square of the desired inequality; take nonnegative square roots.

Advertisement

The triangle inequality for the L2 norm says that adding two square-integrable functions cannot make the norm of their sum exceed the sum of their individual norms. The proof is a direct application of the Cauchy–Schwarz inequality, and it works for both real- and complex-valued functions.

Statement and assumptions

Let (X,A,μ)(X,\mathcal{A},\mu) be a measure space, and let ff and gg be measurable functions with finite L2 norms:

∥f∥2=(∫X∣f∣2 dμ)1/2<∞,∥g∥2=(∫X∣g∣2 dμ)1/2<∞.\|f\|_2=\left(\int_X |f|^2\,d\mu\right)^{1/2}<\infty, \qquad \|g\|_2=\left(\int_X |g|^2\,d\mu\right)^{1/2}<\infty.

Then f+gf+g is also in L2(X,μ)L^2(X,\mu) and

∥f+g∥2≤∥f∥2+∥g∥2.\|f+g\|_2\leq \|f\|_2+\|g\|_2.

Here, functions that agree almost everywhere are treated as the same element of L2L^2. The inequality concerns their equivalence classes, so changing either function on a set of measure zero does not affect it.

Proof using Cauchy–Schwarz

For real or complex numbers, ∣f+g∣≤∣f∣+∣g∣|f+g|\leq |f|+|g|. Squaring this pointwise bound gives

∣f+g∣2≤(∣f∣+∣g∣)2=∣f∣2+2∣f∣∣g∣+∣g∣2.|f+g|^2\leq (|f|+|g|)^2 =|f|^2+2|f||g|+|g|^2.

Integrate over XX. The Cauchy–Schwarz inequality gives

∫X∣f∣∣g∣ dμ≤(∫X∣f∣2 dμ)1/2(∫X∣g∣2 dμ)1/2=∥f∥2∥g∥2.\int_X |f||g|\,d\mu \leq \left(\int_X |f|^2\,d\mu\right)^{1/2} \left(\int_X |g|^2\,d\mu\right)^{1/2} =\|f\|_2\|g\|_2.

This also shows that the product ∣f∣∣g∣|f||g| is integrable. Therefore,

∥f+g∥22=∫X∣f+g∣2 dμ≤∥f∥22+2∥f∥2∥g∥2+∥g∥22=(∥f∥2+∥g∥2)2.\begin{aligned} \|f+g\|_2^2 &=\int_X |f+g|^2\,d\mu\\ &\leq \|f\|_2^2+2\|f\|_2\|g\|_2+\|g\|_2^2\\ &=(\|f\|_2+\|g\|_2)^2. \end{aligned}

Both sides are nonnegative, so taking square roots proves the result. In particular, the estimate shows that f+gf+g has a finite L2 norm, as required for membership in L2L^2.

For complex-valued functions, another way to see the key point is to expand ∣f+g∣2=∣f∣2+∣g∣2+2Re⁡(fg‾)|f+g|^2=|f|^2+|g|^2+2\operatorname{Re}(f\overline{g}). Since Re⁡(fg‾)≤∣f∣∣g∣\operatorname{Re}(f\overline{g})\leq |f||g|, the same integral bound follows. The absolute-value proof above covers both cases at once.

Worked example on the unit interval

Take X=[0,1]X=[0,1] with Lebesgue measure, and set f(x)=xf(x)=x and g(x)=1−xg(x)=1-x. Their norms are

∥f∥22=∫01x2 dx=13,∥g∥22=∫01(1−x)2 dx=13.\begin{aligned} \|f\|_2^2&=\int_0^1 x^2\,dx=\frac{1}{3},\\ \|g\|_2^2&=\int_0^1 (1-x)^2\,dx=\frac{1}{3}. \end{aligned}

Thus ∥f∥2=∥g∥2=1/3\|f\|_2=\|g\|_2=1/\sqrt{3}. Since f(x)+g(x)=1f(x)+g(x)=1 throughout the interval,

∥f+g∥2=(∫011 dx)1/2=1.\|f+g\|_2=\left(\int_0^1 1\,dx\right)^{1/2}=1.

The inequality becomes

1≤13+13=23,1\leq \frac{1}{\sqrt{3}}+\frac{1}{\sqrt{3}}=\frac{2}{\sqrt{3}},

which is true. This example also shows that the inequality need not be an equality.

When does equality hold?

The L2 norm comes from the inner product ⟨f,g⟩=∫Xfg‾ dμ\langle f,g\rangle=\int_X f\overline{g}\,d\mu. Equality in the triangle inequality occurs when the two functions point in the same direction in this inner-product space: one is a nonnegative real multiple of the other almost everywhere, with the zero-function cases included. For example, if g=2fg=2f, then ∥f+g∥2=3∥f∥2=∥f∥2+∥g∥2\|f+g\|_2=3\|f\|_2=\|f\|_2+\|g\|_2.

Common mistakes

  • Forgetting the square root: The integral gives a bound on ∥f+g∥22\|f+g\|_2^2. Take square roots at the end to obtain the stated norm inequality.
  • Using Cauchy–Schwarz on the wrong quantities: Apply it to ∣f∣|f| and ∣g∣|g|, giving ∫∣f∣∣g∣≤∥f∥2∥g∥2\int |f||g|\leq \|f\|_2\|g\|_2.
  • Assuming the inequality is always an equality: In general it is strict; equality requires the functions to align as described above.
  • Ignoring the finite-norm assumption: The usual triangle inequality is a statement about functions in L2L^2. Ensure both ∥f∥2\|f\|_2 and ∥g∥2\|g\|_2 are finite before applying it.

The same argument is often called Minkowski’s inequality for p=2p=2. For other exponents, the corresponding Lp triangle inequality is also known as Minkowski’s inequality.

References

  1. Real Analysis: Modern Techniques and Their Applications, 2nd ed. — Gerald B. Folland

Spotted an error? Send a correction. Read how we write and check articles in our editorial policy.