Dot Products in Triangle and Parallelogram Geometry

From Geometric Sides to Vectors

A triangle or parallelogram is described by points and line segments, but angles and perpendicularity are often easier to analyze using vectors. The key is to turn each relevant side or diagonal into a directed vector, then compare the vectors with a dot product.

Direction matters. An angle at a vertex is formed by two vectors that both start at that vertex. If a side vector points toward the vertex instead, it must be reversed. This choice can change the sign of a dot product and can change an acute angle into an obtuse one.

For nonzero vectors \(\mathbf{u}\) and \(\mathbf{v}\), the angle \(\theta\) between them lies in the interval from \(0\) to \(\pi\) and satisfies the dot-product angle formula. In particular, a zero dot product means the vectors are perpendicular.

The angle formula requires both vectors to be nonzero, so their magnitudes in the denominator are positive. For a triangle, the two vectors used to find an interior angle must be the two sides directed outward from that angle’s vertex.

Angles in a Triangle

Let a triangle have vertices \(A\), \(B\), and \(C\). To find the angle at \(A\), use \(\overrightarrow{AB}\) and \(\overrightarrow{AC}\), because both begin at \(A\). The angle at \(B\) uses \(\overrightarrow{BA}\) and \(\overrightarrow{BC}\); the angle at \(C\) uses \(\overrightarrow{CA}\) and \(\overrightarrow{CB}\).

These vectors are constructed directly from the vertex coordinates. Once they are written in components, compute their dot product and magnitudes, then apply the angle formula. If the dot product is zero, the chosen angle is a right angle. If it is positive, the angle is acute; if negative, the angle is obtuse.

Why the Triangle Formula Fits the Side Lengths

The dot product at a vertex can also be related directly to the three side lengths. This gives a useful derivation of why the vector calculation agrees with the law of cosines.

Write \(\mathbf{u}=\overrightarrow{AB}\) and \(\mathbf{v}=\overrightarrow{AC}\). The third side is \(\overrightarrow{BC}=\mathbf{v}-\mathbf{u}\), since it runs from the endpoint of \(\mathbf{u}\) to the endpoint of \(\mathbf{v}\). The squared magnitude identity for a difference gives

Rearranging and applying the angle formula at \(A\) yields the following relationship. Here \(a=\|BC\|\), \(b=\|CA\|\), and \(c=\|AB\|\), using the usual convention that each lowercase side is opposite its corresponding uppercase vertex.

This derivation relies only on the dot-product definition and the identity for the squared magnitude of a difference. It also shows why a triangle is right-angled at \(A\) exactly when \(a^2=b^2+c^2\): that equality makes \(\overrightarrow{AB}\cdot\overrightarrow{AC}=0\).

Diagonals of a Parallelogram

Let \(ABCD\) be a parallelogram, with adjacent side vectors \(\mathbf{u}=\overrightarrow{AB}\) and \(\mathbf{v}=\overrightarrow{AD}\). Assume the parallelogram is nondegenerate, so these adjacent sides are not parallel. Starting at \(A\), the opposite vertex is reached by taking both side steps, so \(\overrightarrow{AC}=\mathbf{u}+\mathbf{v}\).

The other diagonal runs from \(B\) to \(D\). Since \(B=A+\mathbf{u}\) and \(D=A+\mathbf{v}\), subtracting the starting point from the endpoint gives \(\overrightarrow{BD}=\mathbf{v}-\mathbf{u}\). These vectors point in the directions of the two diagonals, so their dot product tests whether the diagonals are perpendicular.

To justify the result, expand the diagonal dot product using distributivity and the symmetry of the dot product:

The mixed terms cancel because \(\mathbf{u}\cdot\mathbf{v}=\mathbf{v}\cdot\mathbf{u}\). The diagonals are perpendicular if and only if this dot product is zero, which holds if and only if \(\|\mathbf{v}\|^2=\|\mathbf{u}\|^2\). Since magnitudes are nonnegative, this is equivalent to \(\|\mathbf{v}\|=\|\mathbf{u}\|\). Thus the perpendicular-diagonal condition is exactly that the parallelogram is a rhombus, including the special case of a square.

Common Direction Mistake

A frequent error is to use a side vector that points toward the angle’s vertex rather than away from it. Reversing one vector changes the sign of the dot product, so the resulting angle can be mistaken for its supplement.

For a concrete example, let \(A=(0,0)\), \(B=(2,0)\), and \(C=(3,1)\). At \(B\), the correct vectors are \(\overrightarrow{BA}=(-2,0)\) and \(\overrightarrow{BC}=(1,1)\). Their dot product is \((-2)(1)+(0)(1)=-2\), so the angle at \(B\) is obtuse. If one incorrectly uses \(\overrightarrow{AB}=(2,0)\) instead, the dot product becomes \((2)(1)+(0)(1)=2\), falsely suggesting an acute angle.

A Reliable Geometric Procedure

When analyzing a triangle or parallelogram, keep the geometric construction explicit before computing. The following sequence prevents most sign and endpoint errors.

These methods matter because coordinates can make a geometric relationship testable without relying on a drawing. A sketch may suggest that an angle is acute or that diagonals meet at a right angle, but a dot product supplies an exact criterion. The vector representation also works in three dimensions, where a diagram may not show the true spatial angle.