The cross product and the triple product
Three edges leave one corner of a slanted box. Is there one number that gives the box’s volume and also says whether the edges are arranged like the coordinate axes or like their mirror image?
Where you are 5 in Vectors and space curves
Why this exists
The cross product defined \mathbf{u} \times \mathbf{v} by components and checked its properties one expansion at a time. The right-hand rule stayed a picture, and products of three vectors were not touched.
One observation settles both. The dot product of a vector with a cross product is a 3 \times 3 determinant. This triple product is the signed volume of the box on three vectors, its sign is the exact meaning of “right-handed,” and the algebra of the cross product follows from the algebra of determinants in Properties of determinants: row operations, products, transposes.
Definition
For \mathbf{u} = \langle u_1, u_2, u_3 \rangle and \mathbf{v} = \langle v_1, v_2, v_3 \rangle the cross product is the formal determinant
expanded along its first row, as in The cross product. Write \det(\mathbf{u}, \mathbf{v}, \mathbf{w}) for the determinant of the 3 \times 3 matrix with rows \mathbf{u}, \mathbf{v}, \mathbf{w} in that order.
The triple product of \mathbf{u}, \mathbf{v}, \mathbf{w} is the number \mathbf{u} \cdot (\mathbf{v} \times \mathbf{w}). The vector triple product is the vector \mathbf{u} \times (\mathbf{v} \times \mathbf{w}).
An ordered triple (\mathbf{a}, \mathbf{b}, \mathbf{c}) is positively oriented, or right-handed, when \det(\mathbf{a}, \mathbf{b}, \mathbf{c}) > 0, and negatively oriented (left-handed) when it is negative. Since \det(\mathbf{i}, \mathbf{j}, \mathbf{k}) = 1, the axes of Vectors in space are right-handed.
The parallelepiped on \mathbf{u}, \mathbf{v}, \mathbf{w} is the set of points a\mathbf{u} + b\mathbf{v} + c\mathbf{w} with 0 \le a, b, c \le 1: a slanted box whose six faces are parallelograms.
Intuition
Stand the box on its face spanned by \mathbf{v} and \mathbf{w}. The base has area \lVert \mathbf{v} \times \mathbf{w} \rVert, and \mathbf{v} \times \mathbf{w} points straight out of it. The height is how far \mathbf{u} reaches along that perpendicular, the Vector and scalar projection of \mathbf{u} on \mathbf{v} \times \mathbf{w} up to sign. Base times height multiplies the two lengths and divides one away, leaving \lvert \mathbf{u} \cdot (\mathbf{v} \times \mathbf{w}) \rvert.
The sign is positive when \mathbf{u} is on the same side of the base as \mathbf{v} \times \mathbf{w}, negative on the other side, and 0 when the box is flat. It is the signed volume that What the determinant means attaches to a 3 \times 3 determinant.
Results
Throughout, \mathbf{u}, \mathbf{v}, \mathbf{w}, \mathbf{x} are vectors in \mathbb{R}^3 and c is a number.
Theorem 1 (triple product as a determinant). \mathbf{x} \cdot (\mathbf{u} \times \mathbf{v}) = \det(\mathbf{x}, \mathbf{u}, \mathbf{v}). Consequently
swapping any two of the three vectors changes the sign, and \mathbf{u} \cdot (\mathbf{v} \times \mathbf{w}) = (\mathbf{u} \times \mathbf{v}) \cdot \mathbf{w}.
Theorem 2 (algebraic laws). \mathbf{v} \times \mathbf{u} = -(\mathbf{u} \times \mathbf{v}); (c\mathbf{u}) \times \mathbf{v} = c(\mathbf{u} \times \mathbf{v}) and (\mathbf{u} + \mathbf{x}) \times \mathbf{v} = \mathbf{u} \times \mathbf{v} + \mathbf{x} \times \mathbf{v}, and likewise in the second slot; \mathbf{u} \times \mathbf{u} = \mathbf{0}; and \mathbf{u} \times \mathbf{v} is orthogonal to \mathbf{u} and to \mathbf{v}.
Theorem 3 (vector triple product).
Theorem 4 (Lagrange’s identity). For all \mathbf{a}, \mathbf{b}, \mathbf{c}, \mathbf{d},
In particular \lVert \mathbf{u} \times \mathbf{v} \rVert^2 = \lVert \mathbf{u} \rVert^2 \lVert \mathbf{v} \rVert^2 - (\mathbf{u} \cdot \mathbf{v})^2, and for nonzero \mathbf{u}, \mathbf{v} at angle \theta \in [0, \pi], \lVert \mathbf{u} \times \mathbf{v} \rVert = \lVert \mathbf{u} \rVert\,\lVert \mathbf{v} \rVert \sin\theta.
Theorem 5 (volume and coplanarity). The parallelepiped on \mathbf{u}, \mathbf{v}, \mathbf{w} has volume \lvert \mathbf{u} \cdot (\mathbf{v} \times \mathbf{w}) \rvert. The triple product is 0 if and only if \mathbf{u}, \mathbf{v}, \mathbf{w} lie in one plane through the origin.
Theorem 6 (the cross product without coordinates). If \mathbf{u} and \mathbf{v} are not parallel and \theta is the angle between them, then \mathbf{u} \times \mathbf{v} is the only vector \mathbf{n} with
- \mathbf{n} \cdot \mathbf{u} = 0 and \mathbf{n} \cdot \mathbf{v} = 0,
- \lVert \mathbf{n} \rVert = \lVert \mathbf{u} \rVert\,\lVert \mathbf{v} \rVert \sin\theta,
- (\mathbf{u}, \mathbf{v}, \mathbf{n}) positively oriented.
Theorem 7 (distance from a point to a line). The distance from a point Q to the line through P with direction \mathbf{v} \neq \mathbf{0} is
Why it is true
Lemma. If \mathbf{x} \cdot \mathbf{a} = \mathbf{x} \cdot \mathbf{b} for every \mathbf{x}, then \mathbf{a} = \mathbf{b}. Take \mathbf{x} = \mathbf{a} - \mathbf{b} to get \lVert \mathbf{a} - \mathbf{b} \rVert^2 = 0.
Theorem 1. Expand \det(\mathbf{x}, \mathbf{u}, \mathbf{v}) along its first row by Determinants by cofactor expansion:
which is \mathbf{x} dotted with the components of \mathbf{u} \times \mathbf{v}. A row swap changes the sign of a determinant (Theorem 1 of Properties of determinants: row operations, products, transposes). Moving the rows \mathbf{u}, \mathbf{v}, \mathbf{w} to \mathbf{v}, \mathbf{w}, \mathbf{u} takes two swaps, so the cyclic forms agree, and one swap negates. Finally (\mathbf{u} \times \mathbf{v}) \cdot \mathbf{w} is \mathbf{w} \cdot (\mathbf{u} \times \mathbf{v}), a cyclic form.
Theorem 2. By Theorem 1 and a row swap, \mathbf{x} \cdot (\mathbf{v} \times \mathbf{u}) = -\det(\mathbf{x}, \mathbf{u}, \mathbf{v}) for every \mathbf{x}, and the Lemma gives anticommutativity. Expansion along the second row shows the determinant is linear in that row, so (c\mathbf{u} + \mathbf{y}) \times \mathbf{v} and c\,\mathbf{u} \times \mathbf{v} + \mathbf{y} \times \mathbf{v} have the same dot product with every \mathbf{x}; the second slot follows by anticommutativity. A determinant with two equal rows equals its own negative, so it is 0. Hence \mathbf{u} \times \mathbf{u} has dot product 0 with every \mathbf{x}, and \mathbf{u} \cdot (\mathbf{u} \times \mathbf{v}) = \det(\mathbf{u}, \mathbf{u}, \mathbf{v}) is 0, likewise for \mathbf{v}.
Theorem 3. The first component of \mathbf{u} \times (\mathbf{v} \times \mathbf{w}) is
since the terms \pm u_1 v_1 w_1 cancel. Renaming the indices 1 \to 2 \to 3 \to 1 carries each component of the cross product to the next and leaves dot products unchanged, so the same computation gives the other two components.
Theorem 4. By the last part of Theorem 1, with \mathbf{c} \times \mathbf{d} as the third vector, then Theorem 3:
which is the right side. Put \mathbf{c} = \mathbf{a} = \mathbf{u} and \mathbf{d} = \mathbf{b} = \mathbf{v} for the squared length. With \mathbf{u} \cdot \mathbf{v} = \lVert \mathbf{u} \rVert\,\lVert \mathbf{v} \rVert \cos\theta (The dot product and the angle between vectors) it becomes \lVert \mathbf{u} \rVert^2 \lVert \mathbf{v} \rVert^2 \sin^2\theta, and \sin\theta \ge 0 on [0, \pi].
Theorem 5. Let \mathbf{n} = \mathbf{v} \times \mathbf{w}. If \mathbf{n} = \mathbf{0}, then \mathbf{v} and \mathbf{w} are multiples of one vector, the box is flat, the volume and the triple product are both 0, and all three vectors lie in a plane through the origin containing \mathbf{u} and that vector. Otherwise the base has area \lVert \mathbf{n} \rVert (Theorem 4) and \mathbf{n} is perpendicular to it (Theorem 2), so the height is h = \lvert \mathbf{u} \cdot \mathbf{n} \rvert / \lVert \mathbf{n} \rVert. Every slice parallel to the base is a copy of it, so by Cavalieri’s principle (Volumes with known cross sections) the volume is \lVert \mathbf{n} \rVert h = \lvert \mathbf{u} \cdot \mathbf{n} \rvert.
For coplanarity with \mathbf{n} \neq \mathbf{0}: if \mathbf{u} = a\mathbf{v} + b\mathbf{w}, linearity and Theorem 2 give \mathbf{u} \cdot \mathbf{n} = 0. Conversely, let \mathbf{u} \cdot \mathbf{n} = 0. Theorem 3 gives
and for any \mathbf{y}, Theorems 2 and 3 give
a combination of \mathbf{v} and \mathbf{w}. With \mathbf{y} = \mathbf{n} \times \mathbf{u}, the vector \mathbf{u} = -\mathbf{n} \times (\mathbf{n} \times \mathbf{u}) / \lVert \mathbf{n} \rVert^2 is such a combination. So the triple product is 0 exactly for dependent vectors, as Theorem 3 of Properties of determinants: row operations, products, transposes also says.
Theorem 6. Orthogonality is Theorem 2 and the length is Theorem 4. For orientation, Theorem 1 gives \det(\mathbf{u}, \mathbf{v}, \mathbf{u} \times \mathbf{v}) = \lVert \mathbf{u} \times \mathbf{v} \rVert^2 as a cyclic form, positive because \sin\theta > 0 for non-parallel vectors.
Now let \mathbf{n} have the three properties. The identity just displayed, with \mathbf{y} = \mathbf{n}, gives (\mathbf{u} \times \mathbf{v}) \times \mathbf{n} = (\mathbf{n} \cdot \mathbf{v})\mathbf{u} - (\mathbf{n} \cdot \mathbf{u})\mathbf{v} with both coefficients 0. Both \mathbf{u} \times \mathbf{v} and \mathbf{n} have positive length, so they are parallel by The cross product: \mathbf{n} = t(\mathbf{u} \times \mathbf{v}). Equal lengths force \lvert t \rvert = 1, and \det(\mathbf{u}, \mathbf{v}, \mathbf{n}) = t \lVert \mathbf{u} \times \mathbf{v} \rVert^2 > 0 forces t = 1.
So the right-hand rule is a theorem, and since the conditions mention only lengths, angles and orientation, any right-handed perpendicular unit axes yield the same \mathbf{u} \times \mathbf{v}.
Theorem 7. Let \mathbf{w} = \overrightarrow{PQ}. The point F = P + \operatorname{proj}_{\mathbf{v}} \mathbf{w} of the line has \overrightarrow{FQ} perpendicular to \mathbf{v} (Vector and scalar projection), so it is the nearest point. The Pythagorean theorem in triangle PFQ and Theorem 4 give
In the figure, d is the height of the parallelogram on \mathbf{w} and \mathbf{v}.
Worked examples
The volume of a unit cell
In Amara’s mineralogy lab, a crystal is modeled as copies of one parallelepiped, its unit cell, with edges \mathbf{a} = \langle 3, 1, 0 \rangle, \mathbf{b} = \langle -1, 4, 1 \rangle and \mathbf{c} = \langle 1, 1, 5 \rangle in ångströms. Find the volume of the cell, decide whether (\mathbf{a}, \mathbf{b}, \mathbf{c}) is right-handed, and find the volume of the tetrahedron with edges \mathbf{a}, \mathbf{b}, \mathbf{c} from one corner.
SolutionThe components of \mathbf{b} \times \mathbf{c} are 20 - 1, 1 + 5 and -1 - 4, so \mathbf{b} \times \mathbf{c} = \langle 19, 6, -5 \rangle and
By Theorem 5 the cell has volume 63 cubic ångströms. Check by the cyclic form of Theorem 1: \mathbf{a} \times \mathbf{b} = \langle 1, -3, 13 \rangle and \mathbf{c} \cdot (\mathbf{a} \times \mathbf{b}) = 1 - 3 + 65 = 63. Since \det(\mathbf{a}, \mathbf{b}, \mathbf{c}) > 0, the triple is right-handed.
The tetrahedron is a pyramid on half the cell’s base with the same height, and a pyramid has one third of base times height, so it has \displaystyle \frac{1}{6} of the cell’s volume: \displaystyle \frac{63}{6} = 10.5 cubic ångströms.
Four points in one plane
For which k do A(1, 0, 2), B(3, 1, 1), C(0, 2, 3) and D(2, k, 4) lie in one plane?
SolutionThey do exactly when \overrightarrow{AB}, \overrightarrow{AC}, \overrightarrow{AD} lie in one plane through the origin, which by Theorem 5 means their triple product is 0. Here \overrightarrow{AB} = \langle 2, 1, -1 \rangle, \overrightarrow{AC} = \langle -1, 2, 1 \rangle, \overrightarrow{AD} = \langle 1, k, 2 \rangle, and
Then \overrightarrow{AD} \cdot \langle 3, -1, 5 \rangle = 3 - k + 10 = 13 - k, which is 0 only for k = 13. For any other k the four points are corners of a tetrahedron of volume \displaystyle \frac{\lvert 13 - k \rvert}{6}.
The vector triple product, two ways
For \mathbf{u} = \langle 1, -1, 2 \rangle, \mathbf{v} = \langle 2, 0, 1 \rangle, \mathbf{w} = \langle 0, 3, 1 \rangle, find \mathbf{u} \times (\mathbf{v} \times \mathbf{w}) directly and by Theorem 3, then find (\mathbf{u} \times \mathbf{v}) \times \mathbf{w}.
SolutionDirectly: \mathbf{v} \times \mathbf{w} = \langle -3, -2, 6 \rangle, and
By Theorem 3, with \mathbf{u} \cdot \mathbf{w} = -1 and \mathbf{u} \cdot \mathbf{v} = 4, the product is -\mathbf{v} - 4\mathbf{w}, again \langle -2, -12, -5 \rangle.
For the other grouping, anticommutativity and Theorem 3 give
With \mathbf{w} \cdot \mathbf{u} = -1 and \mathbf{w} \cdot \mathbf{v} = 1 this is -\mathbf{v} - \mathbf{u}, which is \langle -3, 1, -3 \rangle. The two groupings differ: the cross product is not associative.
How far a flaw is from the beam
Kofi aims a laser through a crystal sample along the line through P(1, 2, 0) with direction \mathbf{v} = \langle 2, -1, 2 \rangle, in millimeters. A flaw sits at Q(4, 3, 5). How far is the flaw from the beam, and where on the beam is it closest?
SolutionWith \mathbf{w} = \overrightarrow{PQ} = \langle 3, 1, 5 \rangle,
of length \displaystyle \sqrt{90} = 3\sqrt{10}, and \lVert \mathbf{v} \rVert = 3. Theorem 7 gives \displaystyle d = \sqrt{10} \approx 3.16 millimeters.
The nearest point is F = P + \operatorname{proj}_{\mathbf{v}} \mathbf{w}. Since \mathbf{w} \cdot \mathbf{v} = 15 and \lVert \mathbf{v} \rVert^2 = 9,
Check: \displaystyle \overrightarrow{FQ} = \left\langle -\frac{1}{3}, \frac{8}{3}, \frac{5}{3} \right\rangle has dot product \displaystyle \frac{-2 - 8 + 10}{3} = 0 with \mathbf{v}, and squared length \displaystyle \frac{1 + 64 + 25}{9} = 10.
A proof: crossing two cross products
Prove that (\mathbf{u} \times \mathbf{v}) \times (\mathbf{u} \times \mathbf{w}) = \big(\mathbf{u} \cdot (\mathbf{v} \times \mathbf{w})\big)\,\mathbf{u}.
SolutionLet \mathbf{a} = \mathbf{u} \times \mathbf{v}. By Theorem 3, with \mathbf{a} in the first slot,
Here \mathbf{a} \cdot \mathbf{u} = 0 by Theorem 2, and \mathbf{a} \cdot \mathbf{w} = (\mathbf{u} \times \mathbf{v}) \cdot \mathbf{w} = \mathbf{u} \cdot (\mathbf{v} \times \mathbf{w}) by Theorem 1. Substituting gives the claim.
Volume from lengths and angles alone
Crystallographers record a unit cell by its edge lengths and the angles between edges, not by coordinates. Let M be the matrix with rows \mathbf{a}, \mathbf{b}, \mathbf{c} and G = M M^{T}, the 3 \times 3 matrix of the nine dot products of the edges. Prove that the cell has volume V with V^2 = \det G, and use it on Yuna’s cell, whose three edges have length 2 and meet pairwise at 60^\circ.
SolutionThe entry in row i, column j of M M^{T} is row i of M dotted with row j, which is why G holds the dot products. By the product and transpose rules of Properties of determinants: row operations, products, transposes, \det G = \det M \cdot \det M^{T} = (\det M)^2, and \det M = \mathbf{a} \cdot (\mathbf{b} \times \mathbf{c}) = \pm V by Theorems 1 and 5.
For Yuna’s cell each edge dotted with itself gives 4, and each pair gives 2 \cdot 2 \cos 60^\circ = 2:
so \displaystyle V = \sqrt{32} = 4\sqrt{2} \approx 5.66, with no coordinates chosen at all.
Boundaries
- Three dimensions. A bilinear product perpendicular to both factors with length \lVert \mathbf{u} \rVert\,\lVert \mathbf{v} \rVert \sin\theta exists, apart from the zero product on \mathbb{R}^1, only in \mathbb{R}^3 and \mathbb{R}^7, a theorem beyond this course, and the seven-dimensional one fails Theorem 3.
- No associativity. Theorem 3 and (\mathbf{u} \times \mathbf{v}) \times \mathbf{w} = (\mathbf{u} \cdot \mathbf{w})\mathbf{v} - (\mathbf{v} \cdot \mathbf{w})\mathbf{u} differ in general. The expression (\mathbf{u} \cdot \mathbf{v}) \times \mathbf{w} means nothing: a number has no cross product.
- Orientation depends on the axes. In left-handed axes, or after a reflection, the component formula returns the negative of the vector of Theorem 6.
- Theorem 6 needs non-parallel vectors, and Theorem 7 needs \mathbf{v} \neq \mathbf{0}. For parallel vectors the product is \mathbf{0} and orientation is undefined.
Where this goes next
The normal to the plane through three points is a cross product, the first use of this lesson in Lines and planes in space. Differentiating \mathbf{u}(t) \times \mathbf{v}(t) obeys a product rule that keeps the order of the factors, proved in Vector-valued functions. A small patch of a parametric surface has area \lVert \mathbf{r}_u \times \mathbf{r}_v \rVert\,du\,dv in Parametric surfaces and surface integrals, and the curl in Divergence and curl is the formal cross product \nabla \times \mathbf{F}.
Exercises
Triple products and volume
- Compute \langle 2, 1, 0 \rangle \cdot (\langle 0, 3, 1 \rangle \times \langle 1, 0, 4 \rangle).
- Find the volume of the parallelepiped with edges \langle 1, 2, -1 \rangle, \langle 3, 0, 2 \rangle, \langle -1, 4, 1 \rangle, and say whether the triple in that order is right-handed.
- Find the volume of the tetrahedron with vertices A(1, 1, 1), B(3, 2, 1), C(1, 4, 2), D(2, 2, 5).
- Given \mathbf{u} \cdot (\mathbf{v} \times \mathbf{w}) = 5, find \mathbf{v} \cdot (\mathbf{u} \times \mathbf{w}), (\mathbf{u} \times \mathbf{v}) \cdot \mathbf{w}, (2\mathbf{u}) \cdot (\mathbf{v} \times (\mathbf{w} + \mathbf{u})) and (\mathbf{u} + \mathbf{v}) \cdot (\mathbf{v} \times \mathbf{w}).
- Yuna models a crystal by the cell with edges \mathbf{a} = \langle 5, 0, 0 \rangle, \mathbf{b} = \langle 2, 4, 0 \rangle, \mathbf{c} = \langle 1, 2, 6 \rangle. Kofi uses \mathbf{a}, \mathbf{b} and \mathbf{c}' = \mathbf{c} + 3\mathbf{a} - \mathbf{b} = \langle 14, -2, 6 \rangle instead. Find both volumes and explain why they agree.
Orientation and coplanarity
- Is (\langle 1, 0, 1 \rangle, \langle 1, 1, 0 \rangle, \langle 0, 1, 1 \rangle) right-handed or left-handed?
- Find every real t for which \langle 1, t, 0 \rangle, \langle 0, 1, t \rangle, \langle t, 0, 1 \rangle lie in one plane through the origin, and the values of t for which the triple is right-handed.
- Do P(0, 1, 1), Q(2, 3, 0), R(1, 0, 2), S(5, 5, 1) lie in one plane? If not, find the volume of the tetrahedron PQRS.
Identities in use
- Find \langle 1, 0, 2 \rangle \times (\langle 0, 1, 1 \rangle \times \langle 3, 1, 0 \rangle) by Theorem 3, and check it directly.
- Use Lagrange’s identity to find (\mathbf{u} \times \mathbf{v}) \cdot (\mathbf{w} \times \mathbf{x}) for \mathbf{u} = \langle 1, 2, 0 \rangle, \mathbf{v} = \langle 0, 1, 1 \rangle, \mathbf{w} = \langle 2, 0, 1 \rangle, \mathbf{x} = \langle 1, 1, 1 \rangle, then check by computing both cross products.
- \lVert \mathbf{u} \rVert = 2, \lVert \mathbf{v} \rVert = 5 and \mathbf{u} \cdot \mathbf{v} = -6. Find \lVert \mathbf{u} \times \mathbf{v} \rVert without an angle, then find the angle.
- A cell has edges \mathbf{a}, \mathbf{b} of length 3 at 60^\circ to each other and an edge \mathbf{c} of length 4 perpendicular to both. Find its volume from the matrix G of the last worked example, and check it as base times height.
Distance to a line
- Find the distance from Q(2, 0, 3) to the line through P(0, 1, 1) with direction \langle 1, 2, 2 \rangle.
- Kofi spins a sample about the axis through the origin with direction \langle 1, 1, 1 \rangle, in millimeters. A corner of the sample sits at (3, 0, 0). Find the radius of the circle the corner traces and the length of that circle.
Proofs
- Prove the Jacobi identity:
- Prove that (\mathbf{u} + \mathbf{v}) \cdot \big((\mathbf{v} + \mathbf{w}) \times (\mathbf{w} + \mathbf{u})\big) equals 2\,\mathbf{u} \cdot (\mathbf{v} \times \mathbf{w}).
- Prove that (\mathbf{u} \times \mathbf{v}) \times \mathbf{w} = \mathbf{u} \times (\mathbf{v} \times \mathbf{w}) if and only if (\mathbf{u} \cdot \mathbf{v})\mathbf{w} = (\mathbf{v} \cdot \mathbf{w})\mathbf{u}.
- Prove that if \mathbf{u} \neq \mathbf{0}, \mathbf{u} \times \mathbf{v} = \mathbf{u} \times \mathbf{w} and \mathbf{u} \cdot \mathbf{v} = \mathbf{u} \cdot \mathbf{w}, then \mathbf{v} = \mathbf{w}.
- Deduce the Cauchy-Schwarz inequality \lvert \mathbf{u} \cdot \mathbf{v} \rvert \le \lVert \mathbf{u} \rVert\,\lVert \mathbf{v} \rVert in \mathbb{R}^3 from Lagrange’s identity, with equality exactly when \mathbf{u} \times \mathbf{v} = \mathbf{0}.
- Using only Theorem 2, show that a triple product with two equal vectors is 0, then expand (\mathbf{u} + \mathbf{v}) \cdot ((\mathbf{u} + \mathbf{v}) \times \mathbf{w}) to show that swapping the first two vectors negates the triple product.
- Prove that if \mathbf{u} \cdot \mathbf{v} = 0 then \mathbf{u} \times (\mathbf{u} \times \mathbf{v}) = -\lVert \mathbf{u} \rVert^2 \mathbf{v}. What does this say about the map \mathbf{v} \mapsto \mathbf{u} \times \mathbf{v} on the plane perpendicular to a unit vector \mathbf{u}?
Answers
- 25. \langle 0, 3, 1 \rangle \times \langle 1, 0, 4 \rangle = \langle 12, 1, -3 \rangle, and 24 + 1 + 0 = 25.
- Volume 30; left-handed. \langle 3, 0, 2 \rangle \times \langle -1, 4, 1 \rangle = \langle -8, -5, 12 \rangle, and the triple product is -8 - 10 - 12 = -30 < 0.
- \displaystyle \frac{23}{6}. \overrightarrow{AB} = \langle 2, 1, 0 \rangle, \overrightarrow{AC} = \langle 0, 3, 1 \rangle, \overrightarrow{AD} = \langle 1, 1, 4 \rangle; \overrightarrow{AC} \times \overrightarrow{AD} = \langle 11, 1, -3 \rangle, and \overrightarrow{AB} \cdot \langle 11, 1, -3 \rangle = 23.
- -5, 5, 10, 5. One swap negates; Theorem 1 moves the dot; \mathbf{u} \cdot (\mathbf{v} \times \mathbf{u}) = 0 leaves 2 \cdot 5; and \mathbf{v} \cdot (\mathbf{v} \times \mathbf{w}) = 0.
- 120 cubic ångströms both. \mathbf{b} \times \mathbf{c} = \langle 24, -12, 0 \rangle and \mathbf{a} \cdot \langle 24, -12, 0 \rangle = 120. By linearity in the third slot the second volume adds 3\,\mathbf{a} \cdot (\mathbf{b} \times \mathbf{a}) and subtracts \mathbf{a} \cdot (\mathbf{b} \times \mathbf{b}), both 0.
- Right-handed. The determinant is 1 \cdot (1 - 0) - 0 + 1 \cdot (1 - 0) = 2 > 0.
- Coplanar only for t = -1; right-handed for t > -1. Expanding along the first row gives 1 \cdot 1 - t(0 - t^2), that is 1 + t^3, whose only real zero is t = -1, and 1 + t^3 > 0 exactly when t > -1.
- No; volume \displaystyle \frac{7}{6}. \overrightarrow{PQ} \times \overrightarrow{PR} = \langle 2, 2, -1 \rangle \times \langle 1, -1, 1 \rangle = \langle 1, -3, -4 \rangle, and with \overrightarrow{PS} = \langle 5, 4, 0 \rangle the triple product is 5 - 12 + 0 = -7 \neq 0.
- \langle -6, 1, 3 \rangle. With \mathbf{u} \cdot \mathbf{w} = 3 and \mathbf{u} \cdot \mathbf{v} = 2 it is 3\langle 0, 1, 1 \rangle - 2\langle 3, 1, 0 \rangle. Directly, \langle 0, 1, 1 \rangle \times \langle 3, 1, 0 \rangle = \langle -1, 3, -3 \rangle and \langle 1, 0, 2 \rangle \times \langle -1, 3, -3 \rangle = \langle -6, 1, 3 \rangle.
- 1. (\mathbf{u} \cdot \mathbf{w})(\mathbf{v} \cdot \mathbf{x}) - (\mathbf{u} \cdot \mathbf{x})(\mathbf{v} \cdot \mathbf{w}) = 2 \cdot 2 - 3 \cdot 1. Directly, \langle 2, -1, 1 \rangle \cdot \langle -1, -1, 2 \rangle = -2 + 1 + 2 = 1.
- 8, and \theta \approx 126.9^\circ. \lVert \mathbf{u} \times \mathbf{v} \rVert^2 = 4 \cdot 25 - 36 = 64. Then \displaystyle \cos\theta = -\frac{6}{10}, so \theta is obtuse with \displaystyle \sin\theta = \frac{8}{10}.
- \displaystyle 18\sqrt{3} \approx 31.18. G has rows \displaystyle (9, \frac{9}{2}, 0), \displaystyle (\frac{9}{2}, 9, 0), (0, 0, 16), so \displaystyle \det G = 16\left(81 - \frac{81}{4}\right) = 972 and \displaystyle V = \sqrt{972}. Base \displaystyle 9 \sin 60^\circ = \frac{9\sqrt{3}}{2} times height 4 agrees.
- \displaystyle \frac{\sqrt{65}}{3} \approx 2.69. \overrightarrow{PQ} = \langle 2, -1, 2 \rangle, \overrightarrow{PQ} \times \langle 1, 2, 2 \rangle = \langle -6, -2, 5 \rangle of length \displaystyle \sqrt{65}, divided by 3.
- Radius \displaystyle \sqrt{6} \approx 2.45 mm; circle \displaystyle 2\pi\sqrt{6} \approx 15.39 mm. \langle 3, 0, 0 \rangle \times \langle 1, 1, 1 \rangle = \langle 0, -3, 3 \rangle has length \displaystyle 3\sqrt{2}, divided by \displaystyle \sqrt{3}.
- By Theorem 3 the first term is (\mathbf{u} \cdot \mathbf{w})\mathbf{v} - (\mathbf{u} \cdot \mathbf{v})\mathbf{w}, and the other two come from it by cycling \mathbf{u} \to \mathbf{v} \to \mathbf{w} \to \mathbf{u}. Since the dot product is symmetric, each vector appears twice with opposite coefficients (\mathbf{v} gets \mathbf{u} \cdot \mathbf{w} and -\mathbf{w} \cdot \mathbf{u}), so the sum is \mathbf{0}.
- By Theorem 2 and \mathbf{w} \times \mathbf{w} = \mathbf{0}, the cross product is \mathbf{v} \times \mathbf{w} + \mathbf{v} \times \mathbf{u} + \mathbf{w} \times \mathbf{u}. Dotting with \mathbf{u} leaves \mathbf{u} \cdot (\mathbf{v} \times \mathbf{w}) (the others are perpendicular to \mathbf{u}). Dotting with \mathbf{v} leaves \mathbf{v} \cdot (\mathbf{w} \times \mathbf{u}), which is the same triple product by Theorem 1. The sum is twice it.
- By Theorems 2 and 3, (\mathbf{u} \times \mathbf{v}) \times \mathbf{w} = -\mathbf{w} \times (\mathbf{u} \times \mathbf{v}) equals (\mathbf{u} \cdot \mathbf{w})\mathbf{v} - (\mathbf{v} \cdot \mathbf{w})\mathbf{u}, while \mathbf{u} \times (\mathbf{v} \times \mathbf{w}) equals (\mathbf{u} \cdot \mathbf{w})\mathbf{v} - (\mathbf{u} \cdot \mathbf{v})\mathbf{w}. Their difference is (\mathbf{u} \cdot \mathbf{v})\mathbf{w} - (\mathbf{v} \cdot \mathbf{w})\mathbf{u}, which is \mathbf{0} exactly under the stated condition.
- Let \mathbf{x} = \mathbf{v} - \mathbf{w}. By linearity \mathbf{u} \times \mathbf{x} = \mathbf{0} and \mathbf{u} \cdot \mathbf{x} = 0. Lagrange’s identity says \lVert \mathbf{u} \rVert^2 \lVert \mathbf{x} \rVert^2 equals \lVert \mathbf{u} \times \mathbf{x} \rVert^2 + (\mathbf{u} \cdot \mathbf{x})^2, which is 0. As \lVert \mathbf{u} \rVert \neq 0, we get \mathbf{x} = \mathbf{0}.
- Lagrange’s identity gives (\mathbf{u} \cdot \mathbf{v})^2 = \lVert \mathbf{u} \rVert^2 \lVert \mathbf{v} \rVert^2 - \lVert \mathbf{u} \times \mathbf{v} \rVert^2, which is at most \lVert \mathbf{u} \rVert^2 \lVert \mathbf{v} \rVert^2. Take square roots of these nonnegative numbers. Equality holds exactly when \lVert \mathbf{u} \times \mathbf{v} \rVert^2 = 0.
- A repeated vector gives 0: in the last two places because \mathbf{v} \times \mathbf{v} = \mathbf{0}, in the other two pairs of places by orthogonality. Now (\mathbf{u} + \mathbf{v}) \cdot ((\mathbf{u} + \mathbf{v}) \times \mathbf{w}) = 0, and expanding by linearity gives four terms, two of which repeat a vector. What remains is \mathbf{u} \cdot (\mathbf{v} \times \mathbf{w}) + \mathbf{v} \cdot (\mathbf{u} \times \mathbf{w}) = 0.
- Theorem 3 gives (\mathbf{u} \cdot \mathbf{v})\mathbf{u} - (\mathbf{u} \cdot \mathbf{u})\mathbf{v}, and the first term is \mathbf{0}. For a unit \mathbf{u} the map sends the perpendicular plane to itself (Theorem 2), keeps lengths (\sin\theta = 1), and applied twice gives -\mathbf{v}: it is a quarter turn of that plane about \mathbf{u}.