An Area-Based Proof

There are numerous proofs of the Pythagorean Theorem.

The one attributed to president James Garfield is notably compelling:

Having examined his Area-Based approach, we present the following:

Consider a right triangle \Delta ABC where AC \perp BC and DE \perp BC (Fig. 1)

Fig. 1

It can be affirmed that \frac{BE}{BC}= \frac{DE}{AC} = \frac{BD}{AB}. Furthermore, \frac{CE}{BE} = \frac{AD}{BD}.

From Fig. 1,

S_{\Delta ABC} = S_{\Delta BDE} + S_{\text{trapzoid}\;ACDE}

\implies \frac{1}{2} BC \cdot AC = \frac{1}{2}BE \cdot DE + \frac{1}{2}(DE+AC)\cdot(BC-BE)

\implies BC\cdot AC = BE\cdot DE + DE\cdot BC -DE\cdot BE +AC\cdot BC - AC\cdot BE

\implies DE\cdot BC=AC\cdot BE.

So,

\frac{BE}{BC} = \frac{DE}{AC}.

Moreover, letting

\frac{BE}{BC} = \frac{DE}{AC} = k\quad\quad\quad(1)

gives

BE = k \cdot BC and DE=k \cdot AC.\quad\quad\quad(2)

By Pythagorean Theorem,

AB = \sqrt{BC^2+AC^2}\quad\quad\quad(3)

and,

BD = \sqrt{BE^2+DE^2}

\overset{(2)}{\implies} BD = \sqrt{(k\cdot BC)^2+ (k\cdot AC)^2}

\implies BD = k\cdot \sqrt{BC^2+AC^2}

\overset{(3)}{=}k\cdot AB.

That is,

\frac{BD}{AB} = k.

Thus,

\frac{BE}{BC} =\frac{DE}{AC} \overset{(1)}{=} \frac{BD}{AB}

or

\underline{\frac{BC}{BE}}= \frac{AC}{DE} = \underline{\frac{AB}{BD}}.

It follows that

\underline{\frac{BC}{BE}} -1 = \underline{\frac{AB}{BD}}-1 \implies \frac{BC-BE}{BE} = \frac{AB-BD}{BD} \overset{BC-BE=CE, AB-BD=AD}{\implies} \frac{CE}{BE} = \frac{AD}{BD}.

See also “cos(x), sin(x)/a and 1 (Part 1)” and “A Proof without Calculus“.


Exercise-1 Given any triangle \Delta ABC where DE \parallel BC:

Show that \frac{BC}{DE}=\frac{AC}{AE}=\frac{AB}{AD}, \frac{AD}{DB} = \frac{AE}{EC} and \frac{AD}{DB} = \frac{DE}{BC}.

Analytic vs. Synthetic

An isosceles right triangle ABC is depicted blow where M is the midpoint of the hypotenuse AB, P is any point on AB, PE\perpAC and PF\perpBC. Show that MF\perpME. 


Consider the triangle within a coordinate system, as illustrated in Fig. 1:

Fig. 1 AC = BC, PE\perp AC and PF \perp BC

Let k_1 and k_2 denotes the slopes of MF and ME, respectively. We express them as follows:

k_1= \frac{a-y}{a-0} and k_2 = \frac{a-0}{a-x}.

This means that

MF \perp ME \implies k_1\cdot k_2 = -1

\implies \frac{a-y}{a-0}\cdot \frac{a-0}{a-x} = -1

\implies \frac{a-y}{a-x } = -1

\implies a-y = x-a

\implies x + y =2a

\implies \frac{x}{2a} + \frac{y}{2a} = 1

Certainly, this holds true for P(x, y) as it lies on the hypotenuse AB, which can be represented by the equation

\frac{x}{2a} + \frac{y}{2a} = 1.\quad\quad\quad(1)

Now, let’s provide a step-by-step proof:

P(x,y) lies on the hypotenuse of AB

\implies \frac{x}{2a} + \frac{y}{2a} = 1

\implies x + y =2a

\implies a-y = x-a

\implies \frac{a-y}{a-x}=-1

\implies \frac{a-y}{a}\cdot\frac{a}{a-x}

\implies\frac{a-y}{a-0} \cdot \frac{a-0}{a-x}\cdot=-1

\implies k_1\cdot k_2= -1

\implies ME \perp MF.

An alternative proof is:

MF^2 + ME^2 = \underbrace{(a-y)^2+a^2}_{MF^2} + \underbrace{(a-x)^2+a^2}_{ME^2}

= 4a^2+x^2+y^2-2a(x+y)

\overset{(1) \implies x + y = 2a}{=} 4a^2+x^2+y^2-2a\cdot 2a

=x^2+y^2.

From the right triangle \Delta CEF, we observe that

x^2+y^2= EF^2.

Therefore,

MF^2+ME^2=EF^2.

This implies that \Delta ME is a right triangle. i.e.,

MF \perp ME.

Treating MF and ME as vectors and utilizing the vector equation of line AB, we have yet another proof:

Fig. 2

Let

\overrightarrow{MF} = \begin{pmatrix} 0-a \\ y-a \end{pmatrix} and \overrightarrow{ME} = \begin{pmatrix} x-a \\ 0-y \end{pmatrix}.

The inner product of \overrightarrow{MF} and \overrightarrow{ME}

<\overrightarrow{MF}, \overrightarrow{ME}>=(0-a)(x-a) + (y-a)(0-a) = 2a^2-a(x+y).\quad\quad\quad(2)

The vector equation of the line AB

\overrightarrow{CP} = (1-t) \overrightarrow{CB} + t \overrightarrow{CA}, \quad t \in \mathbb{R}

is

\begin{pmatrix} x \\ y \end{pmatrix} = (1-t)\begin{pmatrix} 0 \\ 2a \end{pmatrix} + t \begin{pmatrix} 2a \\ 0 \end{pmatrix}.

Thus, combining the components:

x + y = \underbrace{(1-t)\cdot 0 + t\cdot 2a}_{x} + \underbrace{(1-t)\cdot 2a + t\cdot 0}_{y}= 2a.

Submitting x +y=2a into (2) gives

<\overrightarrow{MF}, \overrightarrow{ME}> = 2a^2 - a\cdot 2a = 0 \implies MF \perp ME.


Let’s also prove it synthetically.

Fig. 3 MH \perp BC, MG \perp AC

Given

MG \perp AC, HC \perp AG \implies HC \parallel MG \implies \angle FMG = \angle MFH\quad\quad\quad(3)

and

\frac{MH}{AC} = \frac{BH}{BC} = \frac{1}{2}, \frac{MG}{BC} = \frac{AG}{AC} = \frac{1}{2}\implies\frac{MH}{AC} = \frac{MG}{BC} \overset{AC=BC}{\implies}NH = MG.\quad\quad\quad(4)

Similarly,

HF = GE.\quad\quad\quad(5)

It follows that

(4):MH=MG, (5):HF=GE, \Delta MHF and \Delta MGE are right triangles

\implies \Delta MFH \cong \Delta MGE.

Thus,

\angle MFH = \angle MEG.\quad\quad\quad(6)

Consequently,

FME = \angle FMG + \angle GME

\overset{(3):\angle FMG = \angle MFH}{=} \angle MFH + \angle GME

\overset{(6):\angle MFH=\angle MEG}{=} \angle MEG + \angle GME = \frac{\pi}{2} \implies MF \perp ME.

See also “Deriving the vector equation of line”.

Being Effortless

In a triangle, the line segment that joins a vertex and the midpoint of the opposite side is called a median. Show that the medians of a triangle intersect at a common point that is two-thirds of the distance from the vertices to the midpoint of the opposite sides.


Let \Delta ABC be a triangle with vertices (x_1, y_1), (x_2, y_2) and C(x_3, y_3). The medians of the triangle are denoted as AE, BF and CD where E(\frac{x_2+x_3}{2}, \frac{y_2+y_3}{2}), F(\frac{x_1+x_3}{2}, \frac{y_1+y_3}{2}) and D(\frac{x_1+x_2}{2}, \frac{y_1+y_2}{2}) are the midpoints.

Expressing the medians in determinant form, we have:

AE: \begin{vmatrix} x_1 & y_1 & 1 \\ \frac{x_2+x_3}{2} & \frac{y_2+y_3}{2} & 1 \\ x & y & 1\end{vmatrix}=0\quad\quad\quad(1)

BF: \begin{vmatrix} x_2 & y_2 & 1 \\ \frac{x_1+x_3}{2} & \frac{y_1+y_3}{2} & 1 \\ x & y & 1 \end{vmatrix}=0\quad\quad\quad(2)

CD: \begin{vmatrix} x_3 & y_3 & 1 \\ \frac{x_1+x_2}{2} & \frac{y_1+y_2}{2} & 1 \\ x & y & 1 \end{vmatrix}=0\quad\quad\quad(3)

Solving \begin{cases} (1) \\ (2) \end{cases} (see Fig. 1) shows that AE and FB intersect at the point

G: (\frac{x_1+x_2+x_3}{2}, \frac{y_1+y_2+y_3}{3}).\quad\quad\quad(4)

Substitute (4) into (3), we see that (4) satisfies (3). Therefore, AE, BF and CD intersect at G. i.e.,

the medians of a triangle intersect at a common point. 

Fig. 1

Now, we define

U = \underbrace{\sqrt{\left(x_1-\frac{x_1+x_2+x_3}{3}\right)^2+\left(y_1-\frac{y_1+y_2+y_3}{3}\right)^2}}_{AG}

and

V = \frac{2}{3}\cdot \underbrace{\sqrt{\left(x_1-\frac{x_2+x_3}{2}\right)^2+\left(y_1-\frac{y_2+y_3}{2}\right)^2}}_{AE}.

Consequently,

U^2-V^2=0\quad\quad\quad(5)

(see Fig. 1).

Since U^2-V^2=(U+V)\cdot(U-V) and U+V>0, (5)\implies U-V=0. That is, 

AG = \frac{2}{3}AE.

Similarly, BG = \frac{2}{3}BF and CG =\frac{2}{3}CD (see Exercise-1).


Exercise-1 Show that BG = \frac{2}{3}BF and CG =\frac{2}{3}CD.

Easy Does It

Prove: Given any quadrilateral, the lines connecting the midpoint of two opposite sides bisect each other.


The quadrilaterals are illustrated in Fig. 1 with E, F, G and H representing the midpoints of the sides.

Fig. 1

In a rectangular coordinate system:

For points A(x_1, y_1), B(x_2, y_2), C(x_3, y_3) and D(x_4, y_4), the midpoints are:

E(\frac{x_1+x_2}{2}, \frac{y_1+y_2}{2}), F(\frac{x_2+x_3}{2}, \frac{y_2+y_3}{2}), G(\frac{x_3+x_4}{2}, \frac{y_3+y_4}{2}), H(\frac{x_1+x_4}{2}, \frac{y_1+y_4}{2}).

The line EG has a midpoint given by:

  EG_{mid} : \left(\frac{\frac{x_1+x_2}{2} + \frac{x_3+x_4}{2}}{2}, \frac{\frac{y_1+y_2}{2} + \frac{y_3+y_4}{2}}{2}\right).

Similarly, the midpoint of line FH‘s is giveny by:

FH_{mid} : \left(\frac{\frac{x_2+x_3}{2} + \frac{x_1+x_4}{2}}{2}, \frac{\frac{y_2+y_3}{2} + \frac{y_1+y_4}{2}}{2}\right).

Therefore,

EG_{mid} = FH_{mid}

which implies

EG and FH intersect at each other’s midpoint.

In other words,

EG and FH bisect each other.