Multiplying vectors is a fundamental operation in linear algebra that appears in physics, engineering, and machine learning. Unlike multiplying ordinary numbers, vector multiplication has several distinct forms, each with its own rules and interpretations. Whether you’re computing forces in a 3D simulation or measuring similarity between data points, knowing these methods is essential. In this guide, you’ll learn the five key ways to multiply vectors, along with practical examples and common pitfalls to avoid.
🔑 Key Takeaways
- Scalar multiplication changes a vector’s magnitude but not its direction (unless the scalar is negative).
- The dot product outputs a scalar and measures how parallel two vectors are.
- The cross product outputs a vector perpendicular to both inputs, following the right-hand rule.
- Triple products combine three vectors and are used in physics for torque and volume calculations.
- Avoid mixing dot and cross products in the same expression without proper grouping because the operations are not commutative.
1. Scalar Multiplication: The Simplest Way of Multiplying Vectors
The simplest way of multiplying vectors is to multiply a vector by a scalar. If $\vec{v} = (v_1, v_2, v_3)$ and $c$ is a real number, then $c\vec{v} = (c v_1, c v_2, c v_3)$. This operation scales the vector’s length without changing its direction—unless $c$ is negative, which reverses the direction. For example, a force vector multiplied by a positive scalar represents a stronger force in the same direction.
Consider $\vec{a} = (2, -1, 5)$. Then $3\vec{a} = (6, -3, 15)$ and $-2\vec{a} = (-4, 2, -10)$. If the scalar is zero, the result is the zero vector $\vec{0}$. Scalar multiplication is commutative: $c\vec{v} = \vec{v}c$. A common use in physics is scaling a velocity vector by time to get displacement, or in computer graphics to scale an object’s position uniformly. In machine learning, gradient steps are exactly scalar multiplication of the gradient vector by a learning rate.
Edge case: If the original vector is the zero vector, scalar multiplication always yields the zero vector regardless of the scalar (including negative and zero). This is consistent with the geometric interpretation—the zero vector has no direction, so scaling cannot change it.
c * numpy.array([v1, v2, v3]). For speed, use numpy.multiply(v, c).2. Dot Product (Scalar Product): Multiplying Vectors to Get a Scalar
The dot product, also called the scalar product, is the most common way of multiplying vectors that yields a scalar (a single number). Given two vectors $\vec{a} = (a_1, a_2, a_3)$ and $\vec{b} = (b_1, b_2, b_3)$, the dot product is:
$$ \vec{a} \cdot \vec{b} = a_1 b_1 + a_2 b_2 + a_3 b_3 $$
Geometrically, $\vec{a} \cdot \vec{b} = |\vec{a}| |\vec{b}| \cos \theta$, where $\theta$ is the angle between them. This means the dot product tells you how much $\vec{a}$ projects onto $\vec{b}$. If the vectors are perpendicular, $\cos 90^\circ = 0$, so the dot product is zero. If they are parallel, the dot product is the product of their magnitudes.
Example: Let $\vec{p} = (3, 4, 0)$ and $\vec{q} = (2, -1, 5)$. Then $\vec{p} \cdot \vec{q} = 3 \cdot 2 + 4 \cdot (-1) + 0 \cdot 5 = 6 – 4 + 0 = 2$. This positive value indicates the angle is acute (less than 90°). If we had $\vec{p} = (1,0,0)$ and $\vec{q} = (0,1,0)$, the dot product is 0, confirming orthogonality.
Where Dot Product Is Used
In machine learning, the dot product measures similarity between feature vectors (e.g., cosine similarity). In physics, it computes work done by a force along a displacement. Dot products also appear in the dot product of a vector and a matrix for linear transformations. For a deeper geometric interpretation, check out this Wikipedia article on dot products (external resource).
3. Cross Product (Vector Product): Multiplying Vectors to Get a Perpendicular Vector
The cross product is a way of multiplying vectors in 3D space to get another vector that is perpendicular to both inputs. For $\vec{a} = (a_1, a_2, a_3)$ and $\vec{b} = (b_1, b_2, b_3)$:
$$ \vec{a} \times \vec{b} = \begin{vmatrix} \mathbf{i} & \mathbf{j} & \mathbf{k} \\ a_1 & a_2 & a_3 \\ b_1 & b_2 & b_3 \end{vmatrix} = (a_2 b_3 – a_3 b_2,\; a_3 b_1 – a_1 b_3,\; a_1 b_2 – a_2 b_1) $$
The magnitude is $|\vec{a} \times \vec{b}| = |\vec{a}| |\vec{b}| \sin \theta$, which equals the area of the parallelogram spanned by the two vectors. The direction follows the right-hand rule. For example, if $\vec{a}$ points along the x-axis and $\vec{b}$ along the y-axis, the cross product points along the z-axis (positive or negative depending on order).
Example Cross Product
Take $\vec{u} = (1, 0, 0)$ and $\vec{v} = (0, 1, 0)$. Then $\vec{u} \times \vec{v} = (0\cdot0 – 0\cdot1,\; 0\cdot0 – 1\cdot0,\; 1\cdot1 – 0\cdot0) = (0,0,1)$. The resulting vector points along the z‑axis, as expected. The magnitude is $1 \cdot 1 \cdot \sin{90^\circ}=1$, so the parallelogram area is 1. If the vectors are parallel (e.g., both along x-axis), the cross product is the zero vector because $\sin 0^\circ=0$.
Cross products are essential in physics for torque and angular momentum, and in computer graphics for normal vectors, lighting calculations, and camera orientation. They also appear in advanced topics like the linear algebra for machine learning when dealing with 3D rotations. For a visual explanation of the right-hand rule, refer to this Khan Academy video (external resource with dofollow).
numpy.cross(a, b) for 3D vectors.“The cross product turns two 3D vectors into a third vector that is orthogonal to both—a powerful tool for creating coordinate systems.”
4. Triple Products: Multiplying Vectors in Groups of Three
When multiplying vectors with three vectors, you encounter two triple products: the scalar triple product and the vector triple product. These operations are invaluable in physics and geometry.
Scalar Triple Product
The scalar triple product combines three vectors via dot and cross: $\vec{a} \cdot (\vec{b} \times \vec{c})$. It gives the volume of the parallelepiped formed by the vectors and is equal to the determinant of the matrix with the vectors as rows. This product is zero if the three vectors are coplanar (linearly dependent).
For example, $\vec{a} = (1,0,0), \vec{b} = (0,2,0), \vec{c} = (0,0,3)$ gives volume $6$ (determinant = $1 \cdot 2 \cdot 3 = 6$). If we set $\vec{c} = (1,1,0)$, the three vectors lie in the xy-plane, so the scalar triple product is zero.
Vector Triple Product
The vector triple product is $\vec{a} \times (\vec{b} \times \vec{c})$. It results in a vector that lies in the plane spanned by $\vec{b}$ and $\vec{c}$. The identity $\vec{a} \times (\vec{b} \times \vec{c}) = \vec{b}(\vec{a}\cdot\vec{c}) – \vec{c}(\vec{a}\cdot\vec{b})$ is useful for simplification. For instance, let $\vec{a}=(1,0,0)$, $\vec{b}=(0,1,0)$, $\vec{c}=(0,0,1)$. Then $\vec{a} \times (\vec{b} \times \vec{c}) = \vec{a} \times \vec{i} = 0$ (since cross product of parallel vectors is zero). Using the identity: $\vec{b}(\vec{a}\cdot\vec{c}) – \vec{c}(\vec{a}\cdot\vec{b}) = (0,1,0)\cdot0 – (0,0,1)\cdot0 = 0$.
A mistake I often see is forgetting that cross product is not associative: $(\vec{a} \times \vec{b}) \times \vec{c} \neq \vec{a} \times (\vec{b} \times \vec{c})$ in general. Always use parentheses. In computational work, be explicit with parentheses to avoid ambiguity.
Common Mistakes When Multiplying Vectors
- Mixing dot and cross without grouping: $\vec{a} \cdot \vec{b} \times \vec{c}$ is ambiguous. Use parentheses: $(\vec{a} \cdot \vec{b}) \times \vec{c}$ (though that is invalid because dot gives a scalar) or $\vec{a} \cdot (\vec{b} \times \vec{c})$.
- Forgetting that cross product is anti-commutative: $\vec{a} \times \vec{b} = -\vec{b} \times \vec{a}$.
- Assuming associativity for cross products: As mentioned, the order matters.
- Confusing dimensions: Dot product requires both vectors same size; cross product works only in 3D.
- Ignoring the zero vector: Both dot and cross product with the zero vector yield zero (scalar or vector respectively).
Frequently Asked Questions
What is the difference between dot product and cross product?
The dot product gives a scalar (number) and measures how parallel two vectors are; the cross product gives a vector perpendicular to both and measures how perpendicular they are. They are fundamentally different operations.
Can you multiply more than two vectors?
Yes, you can use the triple product (scalar or vector) which combines three vectors, or you can multiply a sequence of vectors using dot and cross products – but always with careful parentheses.
What happens when you multiply a vector by a scalar?
Scalar multiplication changes the magnitude of the vector (stretches or shrinks it) without changing its direction, unless the scalar is negative (which reverses direction).
Why is vector multiplication important in machine learning?
Vector multiplication underpins linear transformations, neural network weight calculations, and similarity measures between data points (e.g., cosine similarity via dot product).
What is the easiest way to remember the cross product formula?
Use the determinant method with a 3×3 matrix: i, j, k in the first row, components of the first vector in the second row, and components of the second vector in the third row. Then expand.
What is the difference between scalar multiplication and vector multiplication?
Scalar multiplication involves multiplying a vector by a scalar (a number), producing a scaled vector. Vector multiplication usually refers to operations like dot product or cross product, which combine two vectors to produce a scalar or another vector. Both are forms of multiplying vectors but serve different purposes.
📚 Keep reading
This article covered the essential methods of multiplying vectors. For a deeper dive, explore our linear algebra guide and the dot product of a vector and a matrix for practical applications in machine learning. Understanding multiplying vectors is key to advancing in these fields.
Image alt suggestions: “multiplying vectors dot product example”, “multiplying vectors cross product illustration”, “vector triple product diagram”