A student proves that Fourier partial sums Sₙ(f) satisfy ‖f − Sₙ(f)‖₂ → 0 for some f ∈ L²([0, 2π]), then concludes Sₙ(f)(x) → f(x) for every x. What is wrong with this conclusion?
AL² convergence of Fourier series is not actually guaranteed for all f ∈ L²; the student's premise is false
BL² convergence means the integral of squared error goes to zero — it says nothing about behavior at individual points; pointwise convergence requires additional smoothness or regularity of f
CPointwise convergence is weaker than L² convergence, so it would follow automatically if the student's claim were about almost every x
DThe conclusion is valid for continuous functions but the student should have verified continuity of f first
L² convergence (norm convergence) and pointwise convergence are fundamentally different modes of convergence. A sequence can converge in L² while diverging pointwise at specific x-values — the 'typewriter sequence' of sliding indicator functions is the canonical example. Moreover, L² functions are equivalence classes: they are defined only up to modification on measure-zero sets, so 'the value of f at x' isn't even well-defined as a concept in L². Pointwise convergence requires f to have additional regularity — Hölder continuity, bounded variation, or similar conditions.
Question 2 Multiple Choice
Why is L²([0, 2π]) considered the 'right' space for Fourier analysis compared to L¹([0, 2π])?
AL¹ functions cannot be integrated on [0, 2π], making Fourier coefficients undefined in L¹
BL² has an inner product structure that makes Fourier coefficients orthogonal projections and guarantees norm convergence; L¹ lacks an inner product and Fourier series can fail to converge in L¹ norm
CL² contains strictly more functions than L¹ on [0, 2π], making it the more general and powerful space
DFourier series converge pointwise for L² functions but not for L¹ functions
The inner product ⟨f, g⟩ = (1/2π)∫fg̅dx on L² makes the complex exponentials {eⁱⁿˣ} an orthonormal basis, and the Fourier coefficient cₙ = ⟨f, eₙ⟩ is exactly the orthogonal projection of f onto eₙ. Parseval's theorem then says ∑|cₙ|² = ‖f‖² — the infinite-dimensional Pythagorean theorem — and completeness of the system guarantees L² convergence for all f ∈ L². L¹ lacks an inner product, so there's no natural notion of 'projection,' and Fourier series can fail to converge in L¹ norm. The Hilbert space structure of L² is what makes p = 2 so special.
Question 3 True / False
If f ∈ L²([0, 2π]) and ‖f − Sₙ(f)‖₂ → 0, then the Fourier partial sums Sₙ(f)(x) converge to f(x) for nearly every x ∈ [0, 2π].
TTrue
FFalse
Answer: False
L² norm convergence does not imply pointwise convergence at every x. L² functions are equivalence classes (defined up to sets of measure zero), so 'the value at x' isn't even well-defined. Pointwise convergence is a separate question that requires f to have additional regularity. Without further hypotheses, you can only conclude that the approximation is good 'on average' in the L² sense — the squared error integrates to zero — but the series may misbehave at particular points.
Question 4 True / False
In L²([0, 2π]), the Fourier coefficient cₙ = (1/2π) ∫f(x)e⁻ⁱⁿˣ dx is the inner product of f with the nth basis element eₙ(x) = eⁱⁿˣ, and equals the orthogonal projection of f onto that basis direction.
TTrue
FFalse
Answer: True
With the standard inner product ⟨f, g⟩ = (1/2π)∫fg̅ dx on L², we have ⟨f, eₙ⟩ = (1/2π)∫f(x)e⁻ⁱⁿˣ dx = cₙ exactly. This is the projection formula for orthonormal bases in Hilbert spaces. Parseval's theorem — ∑|cₙ|² = ‖f‖² — is then the infinite-dimensional Pythagorean theorem: the squared norm of f equals the sum of squared magnitudes of its projections. The entire Fourier theory in L² is just Hilbert space geometry applied to a concrete function space.
Question 5 Short Answer
Why does the Hilbert space structure of L² make Fourier series natural, and what does Parseval's theorem say in abstract Hilbert space terms?
Think about your answer, then reveal below.
Model answer: L²([0, 2π]) with inner product ⟨f, g⟩ = (1/2π)∫fg̅ dx is a Hilbert space in which the complex exponentials {eⁱⁿˣ} form a complete orthonormal system. A Fourier coefficient cₙ = ⟨f, eₙ⟩ is the coordinate of f in the eₙ direction — the amount of f lying in that basis direction. Parseval's theorem is the infinite-dimensional Pythagorean theorem: ∑|cₙ|² = ‖f‖². This says the squared norm of f equals the sum of squared coordinates, exactly as in finite-dimensional Euclidean space. Completeness of the system then guarantees that the partial sums Sₙ(f) converge to f in L² norm — recovering f from its coordinates — with no additional hypotheses on f beyond f ∈ L².
The key payoff of the Hilbert space perspective is that it makes Fourier theory a special case of a completely general abstract result. Any Hilbert space with a complete orthonormal basis has the same properties: projections give coordinates, Parseval holds, and the series converges in norm. L² is special only in being the concrete function space where this abstract structure naturally lives for Fourier analysis. This perspective also immediately explains why p = 2 is special — the inner product is a p = 2 feature with no analogue for other values of p.