Home › Lectures › 3Blue1Brown — Essence of linear algebra
Dot Products and Duality
The lecture connects the dot product’s projection interpretation to linear transformations from vectors to numbers, introducing the duality between vectors and such transformations.
Free, sign in with Google. A ready quiz does not use your hourly limit.
What the lecture covers
The lecture reviews the dot product as a coordinate-wise multiplication followed by addition, and explains its geometric meaning: project one vector onto the direction of the other, then multiply the projection’s signed length by the other vector’s length. This accounts for positive, zero, or negative values depending on whether the vectors point in similar, perpendicular, or opposite directions. Although the projection description seems asymmetric, scaling either vector changes the result in the same way, consistent with the dot product’s symmetry.
The lecture then considers linear transformations from two-dimensional space to the number line. Each is represented by a 1×2 matrix, and applying it to a vector amounts to a dot product with a corresponding vector. A projection onto a diagonal copy of the number line illustrates the connection: the matrix entries are the coordinates of the unit vector along that line. Scaling that vector scales the transformation’s output. Thus every linear transformation from a vector space to one dimension corresponds to a unique vector, and vice versa. This correspondence is an example of duality: a vector can be understood not only as an arrow, but also as the linear transformation it encodes.
Key ideas
The dot product pairs corresponding coordinates, multiplies them, and adds the results.
Geometrically, the dot product is the signed length of one vector’s projection onto the other, multiplied by the other vector’s length.
The dot product is positive for vectors pointing generally in the same direction, zero for perpendicular vectors, and negative for vectors pointing generally opposite ways.
A linear transformation from a multidimensional space to the number line maps vectors to numbers and is represented by a 1×n matrix.
A 1×2 matrix acts on a two-dimensional vector in the same numerical way as a dot product with a vector.
Projecting vectors onto a diagonal copy of the number line defines a linear transformation whose matrix entries are the coordinates of the unit vector along that line.
Taking the dot product with a non-unit vector corresponds to projecting onto its direction and scaling the projection by that vector’s length.
Every linear transformation from a vector space to one dimension corresponds to a unique vector, illustrating duality between vectors and transformations.
Sample questions
For two vectors v and w, the dot product can be interpreted geometrically by projecting w onto the direction of v. Which quantity does this interpretation use?
AThe length of w’s projection onto v, multiplied by the angle between the vectors
BThe distance between the tips of v and w
CThe signed length of w’s projection onto v, multiplied by the length of v
DThe area of the parallelogram formed by v and w
Show answer
Correct answer: C. The dot product combines the component of w along v with the length of v; the projection’s sign reflects whether that component points with or against v.
For two vectors of equal length, the dot product can be computed by multiplying the length of one vector by the signed length of the other vector’s projection onto it. Why does exchanging the vectors give the same result?
AExchanging the vectors reverses the angle, which changes the sign of both calculations.
BThe two projections have equal signed lengths, and the vector lengths used as factors are equal.
CEach vector projects to itself, so both calculations use the same vector.
DProjection lengths are always equal for any pair of vectors, regardless of their lengths.
Show answer
Correct answer: B. With equal vector lengths, the projection-based products match because the projections onto the other vector’s direction have equal signed lengths.