Prove the ideal $I = \left<X^2,3\right> \space $of$ \space \mathbb{Z}[X]$ is not a principal ideal.

The solution I have been given is the following:

Assume for contradiction that I were a principal ideal, i.e., $I = \left<f(X)\right>$ for some $f(X) \in \mathbb{Z}[X]$. This means that

$$f(X) = 3 \cdot g(X) + X^2 \cdot h(X), \tag{1}$$

$$3 = f(X) \cdot k(X), \tag{2}$$

$$X^2 = f(X) \cdot s(X), \tag{3}$$

for some polynomials $g(X), h(X), k(X)$ and $s(X)$ in $\mathbb{Z}[X]$. As $\mathbb{Z}$is an integral domain, the equation $(2)$ implies that $0 = \deg(3) = \deg f(X) + \deg k(X)$, whereby $\deg f(X) = 0$, i.e., $f(X)$ is a constant polynomial, say $f(X) = n \in \mathbb{Z}$. Next, the equation $(1)$ implies that $n = f(0) = 3g(0)$, whereby $n \in 3 \mathbb{Z}$. Thus, all the coefficients of $n \cdot s(X)$ are in $3\mathbb{Z}$, contradicting the fact that this is supposed to equal $X^2$ by the equation $(3)$.

Could someone help me break this down a bit please. I am struggling to come to terms with the definition of a degree. These are the current definitions I have in simple terms:

Ideal - A subring which agrees with the relevant axioms

Principal ideal - If $I$ is generated by a set with only one element

Apologies if this seems quite vague. TIA

  • 1
    $\begingroup$ What exactly is your question? You just want the definition of the degree of a polynomial? $\endgroup$
    – D. Brogan
    Apr 18 '19 at 21:18
  • $\begingroup$ Possible example of applying the definition of degree to which I would understand as having just come across it, and to help break down the solution to this question. Thanks $\endgroup$
    – Ben Jones
    Apr 18 '19 at 21:21
  • $\begingroup$ Another way: $3$ and $x$ are nonassociate primes so $(x^2,3) = (f)\,\Rightarrow\, f = \gcd(x^2,3) = 1,\,$ so $\,1 = 3 g + x^2 h \,\overset{\large x=0}\Longrightarrow\, 1 = 3 g(0)\,$ in $\Bbb Z,\,$ contradiction, $\endgroup$ Apr 18 '19 at 22:40

The degree of a polynomial $f(x)=a_nx^n+\cdots+a_0$ (where $a_n\neq 0$) is by definition $n$, the highest exponent of $x$ in the expression of $f$. If $f(x)=a_nx^n+\cdots+a_0$ and $g(x)=b_mx^m+\cdots+b_0$ are polynomials, then we have $$f(x)g(x)=a_nb_mx^{n+m}+\cdots+a_0b_0,$$ so that $\deg(fg)=\deg f+\deg g.$ Observe also that the polynomials $f$ with $\deg f=0$ are exactly the constant polynomials.

Now let's take a look at the proof you've presented. Assume for a contradiction that $(3,x^2)=I=(f(x)).$ Since $3,x^2\in I=(f(x)),$ we can write $$3 = f(x)\cdot k(x)$$ and $$x^2 = f(x)\cdot s(x)$$ for some $k(x),s(x)\in\mathbb Z[x].$ Now, since $3 = f(x)\cdot k(x),$ we have $0=\deg 3 = \deg(fk)=\deg f+\deg k.$ This forces $\deg f=\deg k=0,$ since degrees are always nonnegative. [Your proof says "since $\mathbb Z$ is an integral domain." This may as well be good justification, but I my justification is just as good.] Hence $f$ is constant, and in fact equals $\pm 3$ since the equation $3=f\cdot k$ means $f$ divides $3.$ However we also have $x^2=f(x)\cdot s(x)$. Since $f=\pm 3$, this means all the coefficients of $f(x)s(x)$ are divisible by $3.$ But this is not true of $x^2.$ This is a contradiction, and it completes the proof.

You'll notice we never used the equation $f(x)=3\cdot g(x)+x^2\cdot h(x).$

Hope this helps.

  • $\begingroup$ Worth emphasis: the product of the leading coef's remains nonzero because $\Bbb Z$ is a domain (otherwise the product could have lower degree, e.g. $\,(2x-3)(2-3x) = x\ $ over $\Bbb Z_6 =$ integers $\bmod 6\ \ $ $\endgroup$ Apr 18 '19 at 22:30
  • $\begingroup$ Actually, you can only conclude that $f=\pm3$ or $f=\pm1$. You concluded correctly for the former case; but the latter case should be analyzed as well and is where $f(x)=3g(x)+x^2h(x)$ comes into play. $\endgroup$
    – egreg
    Apr 18 '19 at 23:18
  • $\begingroup$ Both good comments. I'll revise when I get home (if I remember). $\endgroup$
    – D. Brogan
    Apr 19 '19 at 17:15

I'll assume you are happy with the first line.

Equation (1) is saying $f$ must be a combination of the generators given. Similarly, equations (2) and (3) say the generators given must be recoverable from $f$. Thus, the equations together give equality.

The degree of a univariate polynomial in $x$ is simply the highest power of $x$ occuring (so with nonzero coefficient). This is a useful tool allowing us to make the deductions in the next paragraph. That is, if $g = fk$, then $\deg g = \deg f + \deg k$, as used in the case $\deg g = 0$ to force, in particular, $\deg f = 0$.

Finally, substituting $X = 0$ in equation 1 forces $n$ to be a multiple of 3, deriving a contradiction in equation 3.

Is it now clear? If not, what in particular could I explain better?


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.