Real Analysis 37 | Uniform Convergence for Differentiable Functions [dark version]

Q: Why isn’t uniform convergence of $F_n'$ by itself enough to conclude that $f$ is differentiable?

Uniform convergence of $F_n'\to G$ controls derivatives, but it does not guarantee that $F_n$ converges to a specific limit function $f$ pointwise (or at all). The theorem needs pointwise convergence of $F_n$ to identify the candidate limit $f$. Without that, there may be no well-defined function $f$ to differentiate, or the limit could fail to match the derivative behavior encoded by $G$.

Q: What role does pointwise convergence of $F_n\to f$ play in the difference-quotient argument?

Fix $x_0\in I$ and consider $x\to x_0$. Pointwise convergence ensures that $F_n(x)\to f(x)$ and $F_n(x_0)\to f(x_0)$. That lets the difference quotient built from $F_n$ approximate the difference quotient built from $f$, so one of the error terms can be driven to zero as $n\to\infty$.

Q: How does the proof conclude $f'(x_0)=G(x_0)$?

For any $\varepsilon>0$, the three-term bound is arranged so the total absolute error is $\le \varepsilon$. As $n\to\infty$ and $x\to x_0$, the error terms vanish, leaving the difference quotient limit equal to $G(x_0)$. Since $x_0$ was arbitrary, the equality holds for every point in $I$.

Q: What is the practical takeaway for later examples like power series?

Power series often produce sequences of partial sums $F_n$ that converge pointwise to a function $f$. If the derivatives of those partial sums converge uniformly to $G$, the theorem justifies differentiating term-by-term and guarantees that the resulting derivative matches $G$.

TL;DR

If $F_{n}$ are differentiable on a common domain $I$ , and $F_{n} \to f$ pointwise while $F_{n}^{'} \to G$ uniformly, then $f$ is differentiable on $I$ .

Briefing Cornell Notes

Briefing

Uniform convergence of derivatives is the key condition that preserves differentiability when a sequence of differentiable functions converges. Start with a sequence of functions $F_{n}$ defined on a common domain $I$ . If every $F_{n}$ is differentiable, and the derivative functions $F_{n}^{'}$ converge uniformly to some function $G$ , then the limit function $f$ is differentiable and its derivative satisfies $f^{'} = G$ —but only if $F_{n}$ also converges pointwise to $f$ .

Without that extra pointwise convergence, differentiability preservation can fail: uniform convergence of $F_{n}^{'}$ alone does not guarantee that $F_{n}$ converges to a well-defined limit function $f$ in the first place. The theorem therefore combines two convergence modes: a weaker one (pointwise convergence of $F_{n}$ to $f$ ) to identify the limit, and a stronger one (uniform convergence of $F_{n}^{'}$ to $G$ ) to control how derivatives behave across the entire domain.

The proof strategy fixes an arbitrary point $x_{0} \in I$ and studies the derivative of the limit function using the difference quotient. For $x \neq = x_{0}$ , the expression $\frac{f ( x ) - f ( x _{0} )}{x - x _{0}} - G (x_{0})$ is rewritten by inserting and subtracting corresponding difference quotients and derivative-related terms built from $F_{n}$ . This creates three pieces whose sizes can be bounded separately using the triangle inequality.

Two of those pieces can be made small: one term shrinks because $F_{n} \to f$ pointwise (so values of $F_{n}$ near $x$ and $x_{0}$ match those of $f$ ), and another term shrinks because $x \to x_{0}$ in the difference quotient. The remaining piece is the only potential obstacle, and it is precisely where uniform convergence of $F_{n}^{'} \to G$ is needed. Uniform convergence ensures that the derivative-related term does not misbehave for large $n$ at points near $x_{0}$ .

Once these bounds are arranged, the difference quotient of $f$ at $x_{0}$ is forced to converge to $G (x_{0})$ . Since $x_{0}$ was arbitrary, $f$ is differentiable everywhere on $I$ , and the derivative matches the uniform limit of the derivatives: $f^{'} = G$ . The practical payoff is that this result becomes a powerful tool for constructing and differentiating functions defined by power series in later work, where uniform convergence of derivative sequences can be checked to justify term-by-term differentiation.

Cornell Notes

A differentiability-preservation theorem links three ingredients: (1) each $F_{n}$ is differentiable on a common domain $I$ , (2) $F_{n}$ converges pointwise to a limit function $f$ , and (3) the derivatives $F_{n}^{'}$ converge uniformly to a function $G$ . Under these assumptions, the limit function $f$ is differentiable on $I$ . Moreover, the derivative of the limit equals the uniform limit of the derivatives: $f^{'} = G$ . The proof controls the difference quotient at an arbitrary point $x_{0}$ by splitting it into terms that vanish via pointwise convergence, the limit $x \to x_{0}$ , and—crucially—uniform convergence of $F_{n}^{'}$ .

Why isn’t uniform convergence of $F_{n}^{'}$ by itself enough to conclude that $f$ is differentiable?

Uniform convergence of

F_{n}^{'} \to G

controls derivatives, but it does not guarantee that

F_{n}

converges to a specific limit function

f

pointwise (or at all). The theorem needs pointwise convergence of

F_{n}

to identify the candidate limit

f

. Without that, there may be no well-defined function

f

to differentiate, or the limit could fail to match the derivative behavior encoded by

G

What role does pointwise convergence of $F_{n} \to f$ play in the difference-quotient argument?

Fix

x_{0} \in I

and consider

x \to x_{0}

. Pointwise convergence ensures that

F_{n} (x) \to f (x)

and

F_{n} (x_{0}) \to f (x_{0})

. That lets the difference quotient built from

F_{n}

approximate the difference quotient built from

f

, so one of the error terms can be driven to zero as

n \to \infty

Where exactly does uniform convergence of $F_{n}^{'} \to G$ become essential?

After rewriting the difference quotient error using triangle inequality, one term remains that depends on how

F_{n}^{'}

behaves near

x_{0}

. That term can only be forced small uniformly in the relevant points if

F_{n}^{'}

converges uniformly to

G

. This prevents the derivative-related error from staying large for some

n

even when

x

is close to

x_{0}

How does the proof conclude $f^{'} (x_{0}) = G (x_{0})$ ?

For any

ε > 0

, the three-term bound is arranged so the total absolute error is

\leq ε

. As

n \to \infty

and

x \to x_{0}

, the error terms vanish, leaving the difference quotient limit equal to

G (x_{0})

. Since

x_{0}

was arbitrary, the equality holds for every point in

I

What is the practical takeaway for later examples like power series?

Power series often produce sequences of partial sums

F_{n}

that converge pointwise to a function

f

. If the derivatives of those partial sums converge uniformly to

G

, the theorem justifies differentiating term-by-term and guarantees that the resulting derivative matches

G

How does this theorem relate to the earlier fact that uniform convergence preserves continuity?

Both results use uniform convergence to preserve a “local-to-global” property. Uniform convergence of

F_{n}

preserves continuity of the limit. Here, uniform convergence of

F_{n}^{'}

preserves differentiability of the limit, but it must be paired with pointwise convergence of

F_{n}

so the limit function

f

is properly identified.

Review Questions

State the assumptions needed to conclude that $f$ is differentiable and that $f^{'} = G$ .
In the difference-quotient proof, which convergence assumption controls which error term?
Why does the theorem require pointwise convergence of $F_{n} \to f$ in addition to uniform convergence of $F_{n}^{'} \to G$ ?

Key Points

1
If $F_{n}$ are differentiable on a common domain $I$ , and $F_{n} \to f$ pointwise while $F_{n}^{'} \to G$ uniformly, then $f$ is differentiable on $I$ .
2
Under those assumptions, the derivative of the limit satisfies $f^{'} = G$ everywhere on $I$ .
3
Uniform convergence of derivatives alone does not determine the limit function $f$ ; pointwise convergence is needed to identify $f$ .
4
The proof fixes an arbitrary $x_{0} \in I$ and analyzes the difference quotient for $f$ at $x_{0}$ .
5
Triangle inequality splits the difference-quotient error into multiple terms, each controlled by a different convergence step.
6
Uniform convergence of $F_{n}^{'}$ is the crucial ingredient that prevents derivative-related error from staying large near $x_{0}$ .
7
The theorem sets up rigorous term-by-term differentiation for settings like power series when uniform convergence of derivative sequences can be established.

Highlights

Differentiability survives the limit when Fn′​ converges uniformly—provided Fn​ converges pointwise to the function being differentiated.

The conclusion is exact: f′ equals the uniform limit G of the derivatives.

The difference-quotient proof hinges on triangle inequality and on using uniform convergence to control the only stubborn error term.

Pointwise convergence is not decorative; it identifies the limit function f so the derivative claim has a target. 

Real Analysis 37 | Uniform Convergence for Differentiable Functions [dark version]

Briefing

Cornell Notes

Why isn’t uniform convergence of $F_{n}^{'}$ by itself enough to conclude that $f$ is differentiable?

What role does pointwise convergence of $F_{n} \to f$ play in the difference-quotient argument?

Where exactly does uniform convergence of $F_{n}^{'} \to G$ become essential?

How does the proof conclude $f^{'} (x_{0}) = G (x_{0})$ ?

What is the practical takeaway for later examples like power series?

How does this theorem relate to the earlier fact that uniform convergence preserves continuity?

Review Questions

Key Points

Highlights

Topics

Get summaries like this for any content

Briefing

Cornell Notes

Why isn’t uniform convergence of Fn′​ by itself enough to conclude that f is differentiable?

What role does pointwise convergence of Fn​→f play in the difference-quotient argument?

Where exactly does uniform convergence of Fn′​→G become essential?

How does the proof conclude f′(x0​)=G(x0​)?

What is the practical takeaway for later examples like power series?

How does this theorem relate to the earlier fact that uniform convergence preserves continuity?

Review Questions

Key Points

Highlights

Topics

Get summaries like this for any content

Why isn’t uniform convergence of $F_{n}^{'}$ by itself enough to conclude that $f$ is differentiable?

What role does pointwise convergence of $F_{n} \to f$ play in the difference-quotient argument?

Where exactly does uniform convergence of $F_{n}^{'} \to G$ become essential?

How does the proof conclude $f^{'} (x_{0}) = G (x_{0})$ ?