What exactly is a multi-index $\alpha$, and what restrictions apply to its components?

A multi-index is an n-dimensional vector $\alpha=(\alpha_1,\dots,\alpha_n)$, where n matches the number of variables in the function $f$. Each component $\alpha_i$ must be a natural number or 0, so all entries are nonnegative integers. Zeros are important because they indicate differentiation does not occur with respect to that variable.

How does $x^\alpha$ work for a vector $x\in\mathbb{R}^n$?

For $x=(x_1,\dots,x_n)$, the notation $x^\alpha$ means $x_1^{\alpha_1}x_2^{\alpha_2}\cdots x_n^{\alpha_n}$. If some $\alpha_i=0$, then the corresponding factor is $x_i^0=1$, so that variable does not appear in the monomial product.

Multivariable Calculus 15 | Multi-Index Notation

Q: What does $D^\alpha f$ mean in terms of partial derivatives with respect to each variable?

$D^\alpha f$ is the mixed partial derivative where each variable $x_i$ is differentiated $\alpha_i$ times. Formally, $D^\alpha f=\frac{\partial^{|\alpha|} f}{\partial x_1^{\alpha_1}\partial x_2^{\alpha_2}\cdots\partial x_n^{\alpha_n}}$. Example: for $n=3$ and $\alpha=(1,0,0)$, $D^\alpha f=\partial f/\partial x_1$. For $\alpha=(1,2,1)$, it becomes $\partial^4 f/(\partial x_1\partial x_2^2\partial x_3)$.

TL;DR

A multi-index $α = (α_{1}, \dots, α_{n})$ is an n-tuple of nonnegative integers (zeros allowed) that encodes how many times to differentiate with respect to each variable.

Briefing Cornell Notes

Briefing

Multi-index notation streamlines multivariable partial derivatives by packaging “which variables to differentiate” and “how many times” into a single object. Instead of writing long strings of partial derivative operators with an order index and evaluation point, the notation uses an n-dimensional multi-index $α = (α_{1}, \dots, α_{n})$ whose entries are nonnegative integers (allowing zeros). The key payoff is compact formulas—especially for higher-order derivatives and later Taylor expansions—without losing precision about derivative order or variable selection.

The starting point is the familiar partial-derivative conventions: one can denote repeated differentiation using an upper order index and a Leibniz-style fraction where the denominator lists variables, each raised to a power indicating repetition. Schwartz’s theorem then justifies rearranging the order of mixed partial derivatives when the function is sufficiently smooth—specifically, when all relevant higher-order partial derivatives exist and are continuous. That commutativity means the notation can ignore the sequence of differentiation and focus only on counts.

A multi-index $α$ provides exactly those counts. Its “length” (also called absolute value) is defined as $∣ α ∣ = α_{1} + \dots + α_{n}$ , which becomes the total order of the derivative. The same multi-index also encodes monomials: for a vector $x \in R^{n}$ , the expression $x^{α}$ means $x_{1}^{α_{1}} x_{2}^{α_{2}} \dots x_{n}^{α_{n}}$ . Factorials generalize similarly: $α! = α_{1}! α_{2}! \dots α_{n}!$ , with the convention $0! = 1$ . These definitions let formulas treat multi-variable polynomials and derivative operators in a uniform way.

The central operator notation is $D^{α} f$ , which stands for applying partial derivatives according to $α$ : $D^{α} f = \frac{\partial ^{∣ α ∣} f}{\partial x _{1}^{α_{1}} \partial x _{2}^{α_{2}} \dots \partial x _{n}^{α_{n}}}$ . Here, the multi-index simultaneously determines which variables appear and the exponent on each variable’s differentiation. For example, when $n = 3$ and $α = (1, 0, 0)$ , $∣ α ∣ = 1$ and $D^{α} f$ reduces to $\partial f / \partial x_{1}$ . When $α = (1, 2, 1)$ , the total order is $∣ α ∣ = 4$ , $α! = 1! 2! 1! = 2$ , and the operator becomes a mixed fourth-order derivative: $\partial^{4} f / (\partial x_{1} \partial x_{2}^{2} \partial x_{3})$ .

The transcript also notes that generalized binomial coefficients can be defined for multi-indices (though they’re deferred for later use). The immediate motivation is clear: multi-index notation sets up the compact, systematic indexing needed for multivariable Taylor polynomials, where many derivatives of different orders and variable combinations must be organized efficiently.

Cornell Notes

Multi-index notation packages repeated partial differentiation in multivariable calculus into a single n-tuple $α = (α_{1}, \dots, α_{n})$ . Each $α_{i}$ is a nonnegative integer (zeros allowed), and its length $∣ α ∣ = \sum α_{i}$ gives the total derivative order. The same $α$ also encodes monomials $x^{α} = \prod x_{i}^{α_{i}}$ and factorials $α! = \prod α_{i}!$ (with $0! = 1$ ). The derivative operator $D^{α} f$ means $\partial^{∣ α ∣} f / (\partial x_{1}^{α_{1}} \dots \partial x_{n}^{α_{n}})$ , and Schwartz’s theorem supports ignoring the order of mixed partials when the function is smooth enough. This compact indexing is designed for later formulas like multivariable Taylor polynomials.

What exactly is a multi-index $α$ , and what restrictions apply to its components?

A multi-index is an n-dimensional vector

α = (α_{1}, \dots, α_{n})

, where n matches the number of variables in the function

f

. Each component

α_{i}

must be a natural number or 0, so all entries are nonnegative integers. Zeros are important because they indicate differentiation does not occur with respect to that variable.

How does $∣ α ∣$ relate to the order of the partial derivative $D^{α} f$ ?

The length (absolute value) of the multi-index is defined by

∣ α ∣ = α_{1} + \dots + α_{n}

. In the operator

D^{α} f

∣ α ∣

becomes the total number of partial differentiations, i.e., the overall derivative order. For instance, with

α = (1, 2, 1)

R^{3}

∣ α ∣ = 4

, so

D^{α} f

is a fourth-order mixed partial derivative.

How does $x^{α}$ work for a vector $x \in R^{n}$ ?

For

x = (x_{1}, \dots, x_{n})

, the notation

x^{α}

means

x_{1}^{α_{1}} x_{2}^{α_{2}} \dots x_{n}^{α_{n}}

. If some

α_{i} = 0

, then the corresponding factor is

x_{i}^{0} = 1

, so that variable does not appear in the monomial product.

What does $D^{α} f$ mean in terms of partial derivatives with respect to each variable?

D^{α} f

is the mixed partial derivative where each variable

x_{i}

is differentiated

α_{i}

times. Formally,

D^{α} f = \frac{\partial ^{∣ α ∣} f}{\partial x _{1}^{α_{1}} \partial x _{2}^{α_{2}} \dots \partial x _{n}^{α_{n}}}

. Example: for

n = 3

and

α = (1, 0, 0)

D^{α} f = \partial f / \partial x_{1}

. For

α = (1, 2, 1)

, it becomes

\partial^{4} f / (\partial x_{1} \partial x_{2}^{2} \partial x_{3})

Why does the notation not track the order of differentiation steps?

Schwartz’s theorem says mixed partial derivatives commute when the function is sufficiently smooth—specifically, when all relevant higher-order partial derivatives exist and are continuous. That allows the derivative to be indexed only by counts (the multi-index) rather than by the sequence of differentiation.

Review Questions

Given $α = (0, 3, 2)$ in $R^{3}$ , what are $∣ α ∣$ and $D^{α} f$ ?
Compute $α!$ and $x^{α}$ for $α = (2, 0, 1)$ .
Explain how Schwartz’s theorem justifies using $D^{α} f$ without specifying an order of mixed partial derivatives.

Key Points

1
A multi-index $α = (α_{1}, \dots, α_{n})$ is an n-tuple of nonnegative integers (zeros allowed) that encodes how many times to differentiate with respect to each variable.
2
The total derivative order is $∣ α ∣ = α_{1} + \dots + α_{n}$ .
3
Monomials in several variables are written compactly as $x^{α} = \prod_{i = 1}^{n} x_{i}^{α_{i}}$ .
4
Multi-index factorials generalize the usual factorial: $α! = \prod_{i = 1}^{n} α_{i}!$ , using $0! = 1$ .
5
The derivative operator $D^{α} f$ means $\partial^{∣ α ∣} f / (\partial x_{1}^{α_{1}} \dots \partial x_{n}^{α_{n}})$ .
6
Schwartz’s theorem supports ignoring the order of mixed partial differentiation when derivatives are continuous enough.
7
The notation is designed to make later multivariable Taylor polynomial formulas shorter and easier to index.

Highlights

Dαf turns a long mixed-derivative expression into a single symbol by using α to specify both variables and repetition counts.

∣α∣ directly gives the total order of the derivative, so α simultaneously controls structure and order.

Zeros in α are meaningful: they remove variables from both xα and the differentiation pattern.

Schwartz’s theorem is the mathematical permission slip for treating mixed partial derivatives as order-independent under smoothness assumptions.

Topics

Multi-Index Notation
Mixed Partial Derivatives
Schwartz's Theorem
Multi-Index Factorial
Multivariable Taylor Setup

Briefing

Cornell Notes

What exactly is a multi-index α, and what restrictions apply to its components?

How does ∣α∣ relate to the order of the partial derivative Dαf?

How does xα work for a vector x∈Rn?

What does Dαf mean in terms of partial derivatives with respect to each variable?