Reciprocal polynomial equations: symmetry, substitution, and applications

Original publication: Updated:

What is a reciprocal polynomial?

A polynomial is reciprocal, or palindromic, when its coefficients read the same in ascending and descending order. For example,

x4+2x3+3x2+2x+1x^4+2x^3+3x^2+2x+1

has coefficient list 1,2,3,2,11,2,3,2,1. If PP has degree dd, the symmetry is expressed by P(x)=xdP(1/x)P(x)=x^dP(1/x) for nonzero xx.

This structure can reduce the degree of an equation substantially. It does not guarantee a simple radical formula for equations of arbitrarily high degree, but it provides a systematic first reduction.

Even degree: divide by the middle power

Example. Solve

x4−x3−4x2−x+1=0.x^4-x^3-4x^2-x+1=0.

The constant term is one, so x=0x=0 is not a solution. Divide by x2x^2:

x2+1x2−(x+1x)−4=0.x^2+\frac1{x^2}-\left(x+\frac1x\right)-4=0.

Set X=x+1/xX=x+1/x. Since x2+x−2=X2−2x^2+x^{-2}=X^2-2, the equation becomes

X2−X−6=0,X^2-X-6=0,

giving X=−2X=-2 or X=3X=3.

  • x+1/x=−2x+1/x=-2 gives (x+1)2=0(x+1)^2=0, so x=−1x=-1.
  • x+1/x=3x+1/x=3 gives x2−3x+1=0x^2-3x+1=0, so x=(3±5)/2x=(3\pm\sqrt5)/2.

Thus the distinct solutions are

x=−1,3+52,3−52.\boxed{x=-1,\quad\frac{3+\sqrt5}{2},\quad\frac{3-\sqrt5}{2}.}

The root −1-1 has multiplicity two; indeed the quartic factors as (x+1)2(x2−3x+1)(x+1)^2(x^2-3x+1).

Why the substitution works in general

After a reciprocal polynomial of degree 2m2m is divided by xmx^m, terms occur in pairs xj+x−jx^j+x^{-j}, together with a possible constant term.

Write Tj=xj+x−jT_j=x^j+x^{-j}. Then

T0=2,T1=X,Tj=XTj−1−Tj−2.T_0=2,\qquad T_1=X,\qquad T_j=XT_{j-1}-T_{j-2}.

By induction every TjT_j is a polynomial in XX. The original degree-2m2m equation therefore reduces to one of degree at most mm in XX. This is also an instance of expressing a symmetric polynomial in x,1/xx,1/x through their sum XX and product one.

Odd degree: first remove x + 1

Every odd-degree reciprocal polynomial has P(−1)=0P(-1)=0: symmetric terms cancel in pairs. Hence it has a factor x+1x+1.

For example,

x5−5x3−5x2+1=(x+1)(x4−x3−4x2−x+1).x^5-5x^3-5x^2+1=(x+1)(x^4-x^3-4x^2-x+1).

The remaining quartic is the one already solved. Consequently the distinct roots are again −1,(3+5)/2,(3−5)/2-1,(3+\sqrt5)/2,(3-\sqrt5)/2, with −1-1 now having multiplicity three.

The zero coefficients of x4x^4 and xx are part of the palindromic coefficient list. Including them makes the symmetry visible.

Why the quotient remains reciprocal

If P(x)=(x+1)Q(x)P(x)=(x+1)Q(x) has degree 2m+12m+1, reciprocity gives

(x+1)Q(x)=x2m+1(1+1x)Q(1/x)=(x+1)x2mQ(1/x).(x+1)Q(x)=x^{2m+1}\left(1+\frac1x\right)Q(1/x)=(x+1)x^{2m}Q(1/x).

Canceling for nonzero x≠−1x\ne-1 proves Q(x)=x2mQ(1/x)Q(x)=x^{2m}Q(1/x). Since this is a polynomial identity after clearing negative powers, it establishes the coefficient symmetry of the degree-2m2m quotient.

Application 1: the fifth roots of unity

Problem. Find all complex numbers satisfying z5=1z^5=1.

Factor

z5−1=(z−1)(z4+z3+z2+z+1).z^5-1=(z-1)(z^4+z^3+z^2+z+1).

For the quartic factor, z≠0z\ne0. Divide by z2z^2 and put X=z+1/zX=z+1/z:

X2+X−1=0,X=−1±52.X^2+X-1=0,\qquad X=\frac{-1\pm\sqrt5}{2}.

For either value of XX, solve z2−Xz+1=0z^2-Xz+1=0. The five roots are

1,−1+5±i10+254,−1−5±i10−254.\boxed{1,\quad\frac{-1+\sqrt5\pm i\sqrt{10+2\sqrt5}}4,\quad\frac{-1-\sqrt5\pm i\sqrt{10-2\sqrt5}}4.}

These agree with the more compact trigonometric description

z=cos⁡2kπ5+isin⁡2kπ5,k=0,1,2,3,4.z=\cos\frac{2k\pi}{5}+i\sin\frac{2k\pi}{5},\qquad k=0,1,2,3,4.

The reciprocal substitution also derives the exact sine and cosine values rather than requiring them beforehand.

Application 2: prove a quartic is always positive

Problem. Show that for every real xx,

2x4−x3+x2−x+2>0.2x^4-x^3+x^2-x+2>0.

At x=0x=0, the expression is two. For nonzero xx, division by the positive number x2x^2 preserves the inequality. With X=x+1/xX=x+1/x, it becomes

2X2−X−3>0.2X^2-X-3>0.

First find the allowed range of X

A real nonzero xx satisfying X=x+1/xX=x+1/x exists exactly when x2−Xx+1=0x^2-Xx+1=0 has a real root. Its discriminant gives

X2−4≥0,X≤−2 or X≥2.X^2-4\ge0,\qquad\boxed{X\le-2\text{ or }X\ge2.}

Conversely, for every such XX, the quadratic formula supplies real roots whose product is one, so they are nonzero. This proves the entire range, including attainability.

Now factor

2X2−X−3=(X+1)(2X−3).2X^2-X-3=(X+1)(2X-3).

For X≤−2X\le-2, both factors are negative. For X≥2X\ge2, both are positive. Hence their product is strictly positive throughout the allowed range, proving the original inequality.

The quadratic y=2X squared-X-3, with the relevant outer ranges X at most -2 or at least 2 highlighted.

The important step was to keep the substitution's range. Showing positivity for every real XX would be false and is unnecessary. Bounds from an inequality can help locate a range, but a complete range statement should also explain why every claimed value is attained.

Related: root placement in intervals and conjugate radical powers.

Related articles

About the author

Koki Saito

Physician and graduate of the University of Tokyo Faculty of Medicine. Drawing on his entrance-exam experience, he explains how to choose mathematical methods and construct complete solutions.

Full profile →

Have a question or spotted an error? Contact us.