The American Mathematical Monthly

ISSN: 0002-9890 (Print) 1930-0972 (Online) Journal homepage: https://www.tandfonline.com/loi/uamm20

A Determinant Proof of a Generalized Pythagorean

Theorem

Suk-Geun Hwang & Ki-Bong Nam

To cite this article: Suk-Geun Hwang & Ki-Bong Nam (2020) A Determinant Proof of a
Generalized Pythagorean Theorem, The American Mathematical Monthly, 127:3, 269-272, DOI:
10.1080/00029890.2020.1693228

To link to this article: https://doi.org/10.1080/00029890.2020.1693228

Published online: 24 Feb 2020.

Submit your article to this journal

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=uamm20
 A Determinant Proof of a Generalized
Pythagorean Theorem
Suk-Geun Hwang and Ki-Bong Nam

Abstract. A generalized Pythagorean theorem is an equation relating the squares of the vol-
umes of faces of a particular k-simplex in n-dimensional Euclidean space. There are many
proofs of this theorem. This note presents yet another very short elementary proof of the gen-
eralized Pythagorean theorem that is no more than the direct expansion of the determinant of
AT A where A is a matrix determined by the simplex.

1. INTRODUCTION. Given a tetrahedron with vertices O, A, B, C, if the edges

OA, OB, and OC are mutually orthogonal, then the sum of squares of the areas of
the orthogonal faces OAB, OBC, OAC is equal to the square of the area of the oppo-
site face ABC. This is a 3-dimensional generalization of the Pythagorean theorem. In
what follows we present a very short elementary proof of the following well-known
n-dimensional generalization of the Pythagorean theorem.

Theorem (n-dimensional Pythagorean theorem). Let A0 , A1 , . . . , Ak be points

in the n-dimensional Euclidean space Rn . Let  be a k-simplex with vertices A0 ,
A1 , . . . , Ak . For each = 0, 1, 2, . . . , k, let Si denote the face of  whose vertices are
all of A0 , A1 , . . . , Ak except Ai . If the edges A0 A1 , A0 A2 , . . . , A0 Ak of  are mutually
orthogonal, then


k
vol(Si )2 = vol(S0 )2 , (1)
i=1

where vol(Si ) denotes the (k − 1)-volume of the face Si of .

Several proofs of this theorem with k = n are found in the literature [1, 3, 5–8]. The
main tools used in some of the known proofs are the fact that

1
vol() = (vol(S0 ) × h),
k
where h is the distance from the point A0 to the face S0 [5, 7], and the extension
of the cross product of two vectors in R3 to that of n − 1 vectors in Rn [3]. There
are also some proofs depending on somewhat more advanced mathematics such as the
divergence theorem [6] or the concept of exterior algebra [1,4]. In this note, we present
a very short elementary proof of the theorem using the following well-known property
of the volume of a parallelepiped found in [2, pp. 328–329].

Lemma ([2]). The square of the volume of the parallelepiped spanned by k linearly
independent column vectors v1 , v2 , . . . , vk in Rn is equal to the determinant of AT A,
where A = [v1 , v2 , . . . , vk ], the n × k matrix whose ith column is vi , i = 1, 2, . . . , k.
doi.org/10.1080/00029890.2020.1693228
MSC: Primary 51A15, Secondary 51A04; 51A05

March 2020] NOTES 269

2. THE PROOF. Before going further we present a more precise description of
a simplex and a parallelepiped together with the relationship between their vol-
umes. Given m linearly independent vectors v1 , v2 , . . . , vm in Rn , the parallelepiped
and the simplex spanned by these vectors mean the sets [v1 , v2 , . . . , vm ] and
[v1 , v2 , . . . , vm ] respectively defined as

m  

[v1 , v2 , . . . , vm ] = θi vi  0 ≤ θi ≤ 1, i = 1, 2, . . . , m ,

i=1

m  
m 

[v1 , v2 , . . . , vm ] = θi vi  θi ≥ 0, i = 1, 2, . . . , m, θi ≤ 1 .

i=1 i=1

The volumes of these two m-dimensional solids are related as follows.

1
volm ([v1 , v2 , . . . , vm ]) = volm ([v1 , v2 , . . . , vm ]), (2)
m!

where volm denotes the m-volume. Equation (2) can be proved by induction on m. This
equation is true when m = 2, because the area of the triangle [v1 , v2 ] is half the area
of the parallelogram [v1 , v2 ]. Let m > 2, and let h be the distance from vm to the
hyperplane containing [v1 , v2 , . . . , vm−1 ]. Then

1
volm ([v1 , v2 , . . . , vm ]) = ( volm−1 ([v1 , v2 , . . . , vm−1 ]) × h)
m
 
1 1
= volm−1 ([v1 , v2 , . . . , vm−1 ]) × h
m (m − 1)!
1
= volm ([v1 , v2 , . . . , vm ]),
m!

where the second equality is due to the induction on m.

Let A0 , A1 , . . . , Ak and S0 , S1 , . . . , Sk be the points and the faces given in the
−−→
theorem. For i = 1, 2, . . . , k, let xi = A0 Ai . Then

Si = [x1 , . . . ,
xi , . . . , xk ],

where 
xi indicates the deletion of xi . Therefore, by (2), it follows that

1
vol(Si ) = vol([x1 , . . . ,
xi , . . . , xk ]).
(k − 1)!

Similarly, we also get

1
vol(S0 ) = vol([x2 − x1 , x3 − x1 , . . . , xk − x1 ]),
(k − 1)!

because

S0 = [x2 − x1 , x3 − x1 , . . . , xk − x1 ].

270 
c THE MATHEMATICAL ASSOCIATION OF AMERICA [Monthly 127
Therefore (1) is equivalent to


k
vol([x1 , . . . ,
xi , . . . , xk ])2 = vol([x2 − x1 , x3 − x1 , . . . , xk − x1 ])2 . (3)
i=1

We are now ready to give our proof.

−−→
Proof of the theorem. It suffices to prove (3). Recall that we set xi = A0 Ai . Since
x1 , x2 , . . . , xk are mutually orthogonal, we may assume that xi = ai ei , where ai is
the norm of xi and ei is the ith column of the identity matrix of order n. Then

vol([x1 , . . . ,
xi , . . . , xk ]) = a1 · · ·
ai · · · ak , (i = 1, 2, . . . , k). (4)

Let A = [x2 − x1 , x3 − x1 , . . . , xk − x1 ]. Then

⎡ ⎤
−a1 −a1 −a1 ··· −a1
⎢ a2 0 0 ··· 0 ⎥
⎢ ⎥
⎢ 0 a3 0 ··· 0 ⎥
⎢ ⎥
⎢ 0 0 a4 ··· 0 ⎥
⎢ . .. .. .. ⎥
⎢ ⎥
A = ⎢ .. . . . ⎥
⎢ 0 ··· ak ⎥
⎢ 0 0 ⎥
⎢ 0 ··· 0 ⎥
⎢ 0 0 ⎥
⎢ . .. .. .. ⎥
⎣ .. . . . ⎦
0 0 0 ··· 0

so that
⎡ 2 ⎤
a1 + a22 a12 a12 ··· a12
⎢ ⎥
⎢ ⎥
⎢ a12 a12 + a32 a12 ··· a12 ⎥
⎢ ⎥
AT A = ⎢
⎢ a2
⎥.
⎥
⎢ a12 a12 + a42 ··· a12 ⎥
⎢ .1 .. .. .. ⎥
⎣ .. . . . ⎦
a12 a12 a12 ··· a12 + ak2

By subtracting the first row from every other row, AT A is transformed into the follow-
ing matrix:
⎡ 2 ⎤
a1 + a22 a12 a12 ··· a12
⎢ ⎥
⎢ ⎥
⎢ −a22 a32 0 ··· 0⎥
⎢ ⎥
⎢ ⎥.
⎢ −a 2 a42 ··· 0⎥
⎢ 0 ⎥
⎢ .2 .. .. .. ⎥
⎣ .. . . .⎦
−a22 0 0 ··· ak2

March 2020] NOTES 271

Expanding the determinant of this matrix along the first row, we get


k
T
det(A A) = (a12 + a22 )a32 a42 · · · ak2 − a12 (−a22 )a32 · · · ai2 · · · ak2
i=3


k
= a12 · · · ai2 · · · ak2 .
i=1

Thus we get (3) from (4) and the lemma, and the proof is complete.

ACKNOWLEDGMENTS. This note is dedicated to Professor Richard A. Brualdi at the University of

Wisconsin-Madison on the occasion of his 80th birthday. The authors thank the reviewers and the editor for
their valuable comments.

REFERENCES

[1] Bhatia, R. (1991). Letter to the editor. Linear Multilinear Algebra. 30(1–2): 155. doi.org/10.1080/030810
89108818098
[2] Birkhoff, G., Mac Lane, S. (1977). A Survey of Modern Algebra, 5th ed. Natick, MA: A K Peters, Ltd.
[3] Cook, W. J. (2013). An n-dimensional Pythagorean theorem. College Math. J. 44(2): 98–101.
doi.org/10.4169/college.math.j.44.2.098
[4] Darling, R. W. R. (1994). Differential Forms and Connections. New York, NY: Cambridge Univ. Press.
[5] Donchian, P. S., Coxeter, H. S. M. (1935). An n-dimensional extension of Pythagoras’ theorem. Math.
Gaz. 19(234): 206. doi.org/10.2307/3605876
[6] Eifler, L., Rhee N. H. (2008). The n-dimensional Pythagorean theorem via the divergence theorem. Amer.
Math. Monthly. 115(5): 456–457. doi.org/10.1080/00029890.2008.11920550
[7] Quadrat, J.-P., Lasserre, J.-B., Hiriart-Urruty, J. B. (2001). Pythagoras’ theorem for areas. Amer. Math.
Monthly. 108(6): 549–551. doi.org/10.2307/2695710
[8] Yeng, S., Lin, T., Lin, Y. -F. (1990). The n-dimensional Pythagorean theorem. Linear Multilinear Alge-
bra. 26(1–2): 9–13. doi.org/10.1080/03081089008817961

Department of Mathematics Education, Kyungpook University, Taegu 702-701, Korea and Department of
Mathematics, University of Wisconsin-Whitewater, Whitewater, WI 53190, USA
[email protected]

Department of Mathematics, University of Wisconsin-Whitewater, Whitewater, WI 53190, USA

[email protected]

272 
c THE MATHEMATICAL ASSOCIATION OF AMERICA [Monthly 127