Module 1: Mathematical Foundations · Theory · assumes Foundation 4
The first way to multiply two vectors — and the one that turns geometry into energy.
Mahavakya
The dot product asks one question of two vectors: how much of one lies along the other? Everything else it does follows from that.
On this page
Learning objectives
After this reading you should be able to:
- State the dot product geometrically and read off its sign from the angle alone.
- Compute it from components in two and three dimensions, and explain why that shortcut is valid.
- Find the scalar and vector projection of one vector onto another.
- Extract the angle between two vectors given only their components.
- Recognise work and power as dot products, and state the conditions under which those formulas hold.
Notation. Vectors in bold: A, B. Magnitudes in italics or bars: A, |A|. The dot is never optional — A · B and AB mean different things. Unit vectors carry a hat: î, ĵ, k̂, and b̂ means the unit vector along B.
1. The geometry of alignment
Foundation 4 showed how to add vectors and take them apart. Mechanics also needs to multiply them — and since direction is part of the quantity, ordinary multiplication has nothing to say. There are two useful ways to define a product of vectors, chosen because they turn out to describe things we care about. This module covers the first: the scalar product, or dot product.
A · B = AB cos θ
θ is the angle between the vectors when drawn tail to tail, 0° ≤ θ ≤ 180°
Look at what comes out. A is a number, B is a number, cos θ is a number. Two vectors go in and a scalar comes out — the direction has been entirely spent on measuring how well the two line up.
Projection is not symmetric. The dot product is.
That diagram is worth sitting with, because it contains a fact students routinely get backwards. The shadow B casts on A is a different length from the shadow A casts on B. Yet A(B cos θ) and B(A cos θ) are the same number. Projection is asymmetric; the dot product built from it is not.
The sign follows entirely from the angle:
- θ < 90° — cos θ is positive, so A · B > 0. The vectors broadly agree.
- θ = 90° — cos 90° = 0, so A · B = 0, no matter how large the vectors are.
- θ > 90° — cos θ is negative, so A · B < 0. One opposes the other.
A solar panel
Let A point from the panel toward the sun, and B point straight out of the panel’s face. Aim the panel at the sun and θ = 0°: you capture everything available, AB. Tilt it and you capture AB cos θ — a fraction, set entirely by the angle. Turn the panel edge-on and θ = 90°: nothing at all, even though both the sunlight and the panel are still very much there. The dot product is that fraction, made precise.
2. The component form, and why it works
Measuring angles in three dimensions is awkward. Written in components, the dot product needs no angles at all — and the reason is worth following rather than accepting.
Start with the basis vectors themselves. Each has length 1, each is parallel to itself and perpendicular to the other two, so the geometric definition gives:
î · î = ĵ · ĵ = k̂ · k̂ = (1)(1) cos 0° = 1
î · ĵ = ĵ · k̂ = k̂ · î = (1)(1) cos 90° = 0
Now expand two general vectors. Multiplying out A · B term by term produces nine products — every component of A paired with every component of B. Six of those nine pair different axes, and every one of them carries a factor of î · ĵ or its equivalent, which is zero. They vanish. Only the three matching pairs survive:
A · B = AxBx + AyBy + AzBz
The angle has not disappeared. It is hiding inside the components, which were themselves built from angles back in Foundation 4. Choosing perpendicular axes is what lets six terms cancel — which is one more instance of Foundation 3’s point that the frame you pick decides how much work the algebra costs.
Worked example
Given A = 3î − 2ĵ + 4k̂ and B = 2î + 5ĵ + k̂, find A · B.
Multiply matching components: (3)(2) = 6, (−2)(5) = −10, (4)(1) = 4.
Add them: A · B = 6 − 10 + 4 = 0.
Zero, and neither vector is zero — so these two are perpendicular. Notice what you have just done: established a right angle in three dimensions without drawing anything or computing a single angle.
3. Properties
The dot product behaves like ordinary multiplication in two respects and unlike it in one.
- Commutative. A · B = B · A, because cos θ does not care which vector you measured from.
- Distributive. A · (B + C) = A · B + A · C. This is the property that made section 2’s expansion legal.
- Not associative — and not in the sense that it fails, but in the sense that (A · B) · C cannot even be written down. The bracket is a scalar, and a scalar has nothing to dot with.
Dotting a vector with itself gives a result used constantly:
A · A = AA cos 0° = A²
That innocuous line does more than it appears to. Take the magnitude of a sum and square it:
|A + B|² = (A + B) · (A + B)
= A · A + 2(A · B) + B · B
= A² + B² + 2AB cos θ
That last line is the parallelogram formula from Foundation 4, arrived at without drawing a triangle or invoking the law of cosines. What was a geometric rule to be remembered is now an algebraic consequence of a definition. This is worth noticing as a pattern: a good definition does not merely describe things, it makes earlier results fall out of it.
Key insight
The dot product is not an extra fact bolted onto vectors. It is the machinery that makes Foundation 4’s cosine formula a consequence rather than a rule — and once it is in place, questions about angles, lengths and perpendicularity all become the same question.
4. Where it appears in mechanics
The dot product shows up wherever a scalar quantity is produced by two vectors acting in concert.
Work
W = F · d = Fd cos θ. Only the part of the force lying along the displacement transfers energy. Drag a sled with a rope at angle θ and the vertical part of your pull does no work at all — it lightens the sled, but the sled never moves upward.
Boundary: this form assumes a constant force and a straight displacement. When either varies, work becomes an integral — Foundation 9.
Power
P = F · v = Fv cos θ, the rate at which a force delivers energy. The same geometry, with velocity in place of displacement — which is no accident, since velocity is displacement per unit time.
The negative case is worth pausing on. A braking force opposes the motion, so θ = 180° and the work is negative: energy leaves the object rather than entering it. The sign is not a bookkeeping convention. It is telling you which way the energy went.
Contemplation
Why should work be a scalar? Energy has no direction. A body holds some number of joules and there is no sense in which that number points anywhere. Yet the causes of energy transfer are thoroughly directional: a force pointing one way, a displacement pointing another. Something has to carry us from two directed quantities to a single undirected one, and it cannot be arbitrary — it has to keep exactly the part of the force that matters and discard the rest. That is what the dot product does. It is less a piece of algebra than a hinge between two kinds of physical quantity.
5. Projections and angles
Two standard tasks fall directly out of the definition.
Scalar projection — how far
A · b̂ = (A · B) / B
A signed length. Negative when the angle is obtuse, which is meaningful rather than an error — it says the projection points backwards along B.
Vector projection — how far, and which way
A∥ = (A · b̂) b̂ = [(A · B) / B²] B
The scalar projection, pointed back along B. This is Foundation 4’s resolution, generalised: instead of resolving onto a coordinate axis, you resolve onto any direction you like.
And rearranging the two definitions of the dot product gives the angle between any two vectors, in any number of dimensions:
cos θ = (A · B) / (AB)
Worked example
Find the angle between A = 2î − 3ĵ + 6k̂ and B = î + 4ĵ + 2k̂.
A · B = (2)(1) + (−3)(4) + (6)(2) = 2 − 12 + 12 = 2.
A = √(4 + 9 + 36) = √49 = 7. B = √(1 + 16 + 4) = √21 ≈ 4.58.
cos θ = 2 / (7 × 4.58) ≈ 0.062, so θ ≈ 86.4° — very nearly, but not quite, perpendicular.
Compare this with the quadrant trouble in Foundation 3. There, tan−1 could not distinguish opposite directions and needed correcting. Here, cos−1 returns a value between 0° and 180° — which is exactly the range an angle between two vectors can occupy. No correction is possible and none is needed. The formula gives the answer outright.
Common misconceptions
1. The dot product gives a vector
It gives a scalar — a number with units and possibly a sign, but no direction and no unit vectors. Writing an answer as “5î” when the question asked for F · d means the whole operation has been misunderstood, not merely mis-stated.
2. A zero dot product means one of the vectors is zero
Almost never what it means. Two perfectly substantial vectors give zero whenever they are perpendicular — as the worked example in section 2 did. In ordinary arithmetic ab = 0 forces one factor to vanish; here it does not, and that difference matters.
3. You can cancel a vector from both sides
From A · B = A · C it does not follow that B = C. All you may conclude is that A · (B − C) = 0, meaning B − C is perpendicular to A. Infinitely many different vectors share the same projection onto A.
4. Negative work is a sign error
It is a physical statement. Negative work means the force removed energy from the object — friction, braking, gravity on something rising. Students often take the modulus to make it “look right” and destroy the physics in doing so.
Reflection
- The dot product throws information away — two vectors go in, one number comes out. What exactly has been discarded, and why is discarding it the right thing to do when computing work?
- The component formula never mentions an angle. Where did the angle go?
- Carrying a heavy bag along a level floor is exhausting, yet the physics says you do no work on the bag. Is the physics wrong, or is “work” not what the everyday word means?
Key takeaways
- A · B = AB cos θ = AxBx + AyBy + AzBz, and the result is always a scalar.
- Zero dot product means perpendicular, not zero — the fastest perpendicularity test there is.
- A · A = A², from which Foundation 4’s cosine formula follows in three lines.
- Projection is asymmetric; the dot product built from it is symmetric.
- W = F · d holds for constant force and straight displacement; otherwise it becomes an integral.
- cos−1 needs no quadrant correction, because its range already matches the range of angles between vectors.
Prerequisite: Foundation 4 · Reading: 16 min · Practice: 20 min · Difficulty: Beginner