Porisma

Resolve parallel and perpendicular components

Separate nonnegative length from signed vector components.

Resolve parallel and perpendicular components

Start with a nonzero target direction and turn it into a unit vector. Then decide whether the question asks for a signed scalar, a nonnegative length, or a vector component. The same dot product appears in each, but the final objects are different.

Let b^=b/∣b∣\widehat{\mathbf b}=\mathbf b/|\mathbf b|, with b≠0\mathbf b\ne\mathbf0. In the source diagrams, PR→=a\overrightarrow{PR}=\mathbf a and QQ is the perpendicular foot from RR to the line through PP in direction b\mathbf b. Distinguish three different results:

PQ=∣a⋅b^∣(projection length: scalar),PQ→=(a⋅b^)b^(parallel component: vector),QR→=a−(a⋅b^)b^(perpendicular component: vector).\begin{aligned} PQ&=|\mathbf a\cdot\widehat{\mathbf b}| &&\text{(projection length: scalar)},\\ \overrightarrow{PQ}&=(\mathbf a\cdot\widehat{\mathbf b})\widehat{\mathbf b} &&\text{(parallel component: vector)},\\ \overrightarrow{QR}&=\mathbf a-(\mathbf a\cdot\widehat{\mathbf b})\widehat{\mathbf b} &&\text{(perpendicular component: vector)}. \end{aligned}

The signed scalar component a⋅b^\mathbf a\cdot\widehat{\mathbf b} may be negative; a length cannot. The source assumes nonparallel vectors to draw a genuine triangle. The formulas also work for parallel or perpendicular a\mathbf a and for a=0\mathbf a=\mathbf0; only the target b\mathbf b must be nonzero.

Acute case: the projection follows b

Let θ\theta be the angle between a\mathbf a and b\mathbf b. In the acute diagram, it is also ∠QPR\angle QPR. The complete length proof is

PQPR=cos⁡θ,PQ=PRcos⁡θ=∣a∣cos⁡θ=∣a∣∣b^∣cos⁡θ=a⋅b^.\frac{PQ}{PR}=\cos\theta,\qquad PQ=PR\cos\theta=|\mathbf a|\cos\theta=|\mathbf a||\widehat{\mathbf b}|\cos\theta=\mathbf a\cdot\widehat{\mathbf b}.

Since PQ→\overrightarrow{PQ} and b\mathbf b have the same unit direction,

PQ→∣PQ→∣=b∣b∣⟹PQ→=∣PQ→∣b∣b∣=(a⋅b^)b^.\frac{\overrightarrow{PQ}}{|\overrightarrow{PQ}|}=\frac{\mathbf b}{|\mathbf b|} \quad\Longrightarrow\quad \overrightarrow{PQ}=|\overrightarrow{PQ}|\frac{\mathbf b}{|\mathbf b|}=(\mathbf a\cdot\widehat{\mathbf b})\widehat{\mathbf b}.

Obtuse case: the projection opposes b

In the source’s second drawing, b\mathbf b points left while PR→\overrightarrow{PR} points up-right. The angle between the vectors is obtuse θ\theta, while the triangle’s angle ∠QPR=π−θ\angle QPR=\pi-\theta is acute. This distinction corrects the source’s unqualified opening “let angle QPR be θ” for the second case. Now

PQPR=cos⁡(π−θ),PQ=−PRcos⁡θ=−∣a∣cos⁡θ=−∣a∣∣b^∣cos⁡θ=−a⋅b^.\begin{aligned} \frac{PQ}{PR}&=\cos(\pi-\theta),\\ PQ&=-PR\cos\theta\\ &=-|\mathbf a|\cos\theta\\ &=-|\mathbf a||\widehat{\mathbf b}|\cos\theta\\ &=-\mathbf a\cdot\widehat{\mathbf b}. \end{aligned}

The unit directions are opposite, so

PQ→∣PQ→∣=−b∣b∣,PQ→=−∣PQ→∣b∣b∣=−(−a⋅b^)b∣b∣=(a⋅b^)b^.\begin{aligned} \frac{\overrightarrow{PQ}}{|\overrightarrow{PQ}|}&=-\frac{\mathbf b}{|\mathbf b|},\\ \overrightarrow{PQ}&=-|\overrightarrow{PQ}|\frac{\mathbf b}{|\mathbf b|}\\ &=-(-\mathbf a\cdot\widehat{\mathbf b})\frac{\mathbf b}{|\mathbf b|}\\ &=(\mathbf a\cdot\widehat{\mathbf b})\widehat{\mathbf b}. \end{aligned}

The Teacher asks why the length formulas have opposite signs: they are the positive and negative cases of the same absolute-value expression. The vector formula is unchanged in both cases. Finally, the source triangle-law proof gives

PR→=PQ→+QR→,a=(a⋅b^)b^+QR→,QR→=a−(a⋅b^)b^.\overrightarrow{PR}=\overrightarrow{PQ}+\overrightarrow{QR},\qquad \mathbf a=(\mathbf a\cdot\widehat{\mathbf b})\widehat{\mathbf b}+\overrightarrow{QR},\qquad \overrightarrow{QR}=\mathbf a-(\mathbf a\cdot\widehat{\mathbf b})\widehat{\mathbf b}.

Example 15 applies all three formulas and labels the parallel and perpendicular arrows (ii) and (iii). Its Teacher note stresses that projection of a on b follows b, whereas projection of b on a follows a. Their vectors generally differ even if their lengths happen to agree, as they do in this example. Self-Practice 5 keeps a negative projection coefficient; do not put absolute-value bars into that vector answer.

The earlier scalar-product table refers forward to this projection section for a distributivity proof. The supplied projection pages derive these formulas; a separate proof of distributivity is not printed there.

Reverse the direction and track what changes

Let a=(−3,4,0)\mathbf a=(-3,4,0) and n^=(1,0,0)\widehat{\mathbf n}=(1,0,0). Find the signed scalar component, projection length, parallel vector component and perpendicular vector component of a\mathbf a relative to this direction. Predict which results change when n^\widehat{\mathbf n} is replaced by −n^-\widehat{\mathbf n}, then verify.

After your attempt: compare the reasoning

Initially the signed component is −3-3, the projection length is 3, the parallel component is (−3,0,0)(-3,0,0) and the perpendicular component is (0,4,0)(0,4,0). With the reversed unit direction the signed component becomes 3, but

[a⋅(−n^)](−n^)=(a⋅n^)n^.[\mathbf a\cdot(-\widehat{\mathbf n})](-\widehat{\mathbf n})=(\mathbf a\cdot\widehat{\mathbf n})\widehat{\mathbf n}.

Both vector components and the projection length stay unchanged. Also ∣a×n^∣=4|\mathbf a\times\widehat{\mathbf n}|=4 measures the perpendicular length. A signed scalar and a vector component are different objects: reversing the measuring direction changes the scalar’s sign without moving the foot.

Choose the object the question asks for

A signed scalar, a length and a vector component are different results. Start with the same unit direction, then choose the required form.

Read the full method map and examples

Diagrams for this unit

Open 1 diagram

Try these examples

Open an example in chapter practice. Try it first, then read its solution. Use “Learn this step” to return to this unit.