symmetric monoidal (∞,1)-category of spectra
Classically, the fundamental theorem of algebra states that
Since every non-zero polynomial could be made a monic polynomial by dividing by the leading coefficient, it could also be expressed as
Many proofs of this theorem are known (see the references below); some use complex analysis (the reciprocal of a polynomial function cannot be bounded), some use algebraic topology (the degree of a map is invariant with respect to homotopy), and some use advanced calculus (polynomial functions on the complex numbers are open mappings). All of these proofs involve, at some level, the fact that the real numbers are Dedekind complete, which has as a consequence the fact that the real numbers are archimedean.
Despite its name, the fundamental theorem of algebra makes reference to a concept from analysis (the field of complex numbers). However, the analytic part may be reduced to a minimum: that the field of real numbers is real closed. This has been known essentially forever, and is easily proved using (for example) the intermediate value theorem.
The rest of the proof is algebraic and, unlike the other proof methods, applies to all real closed fields, which need not be archimedean. It is due to Emil Artin, and forms a basic chapter in the Artin–Schreier theory of real closed fields.
We recall that a real closed field is an ordered field such that every positive element has a square root, and every polynomial function of odd degree has a root. Note that the polynomial function $x^2 + 1$ cannot have any root in a real closed field — or in fact in any ordered field, since we always have $x^2\ge 0$ and hence $x^2+1 \ge 1$.
If $F$ is real closed, then $K = F[\sqrt{-1}]$ is algebraically closed.
We must show that any irreducible polynomial $p$ of degree greater than $0$ with coefficients in $K$ has a root in $K$. Since $F$ has characteristic $0$, it is a perfect field.
Thus the splitting field of $p$ is a finite Galois extension $L$ of $F$, with Galois group $G$. If $G(2)$ is the Sylow 2-group of $G$, then the fixed field? $E$ of $G(2)$ is an odd degree extension of $F$. Any $\alpha \in E$ must then have an irreducible polynomial function $q \in F[x]$ of odd degree. But since $F$ is real closed, $q$ has a root in $F$; by irreducibility, $\deg(q) = 1$ and $\alpha \in F$, forcing $E = F$ and $G = G(2)$. We have ${|G|} \gt 1$ since the splitting field contains $K$.
So $G$ is a $2$-primary group. But for any prime number $p$, a nontrivial finite $p$-group has nontrivial center (see here), and is therefore solvable by an inductive argument. Therefore the extension $L/F$ arises from a tower of non-trivial quadratic extension?s
By the quadratic formula, the first field $L_1$ arises by adjoining roots to $F$ of a polynomial function $x^2 + a x + b$,
where $a^2 - 4b$ is negative. Since $F$ is real closed, the positive element $4b - a^2$ has a square root in $F$, so that the roots displayed above belong to $K = F[\sqrt{-1}]$. So $L_1 = K$. But $K$ has no nontrivial quadratic extensions by the lemma that follows, so in fact $L_1 = L_n = K$ and the theorem is proved.
Every element of $K = F[\sqrt{-1}]$ has a square root in $K$.
The proof is most easily apprehended by analogy with polar coordinate representations of complex numbers and half-angle formulas, where a square root of $r e^{i\theta}$ is given by $r^{1/2}e^{i\theta /2}$. Let $i$ be a fixed square root of $-1$, and let $a + b i$ be an arbitrary element of $K$, with $a, b \in F$. We must solve $(x + y i)^2 = a + b i$, i.e., find $x, y \in F$ that solve
Since $a^2 + b^2$ has a square root in $F$, we may assume by homogeneity in $x, y$ that $(a, b)$ is on the unit circle: $a^2 + b^2 = 1$. By interchanging $x$ and $y$ if need be, we may assume $0 \leq a \leq 1$; replacing $y$ by $-y$ if need be, we may assume $b \geq 0$. Taking $x, y \geq 0$ such that
we obtain a solution (since $x^2 - y^2 = a$ and $4 x^2 y^2 = b^2$).
As noted above, many proofs of the fundamental theorem are known. The following proof, ultimately rooted in the fact that polynomial mappings on $\mathbb{C}$ are open mappings, has the advantage that it requires very little machinery. From what I (Todd Trimble) understand, it is close to the method used by Argand to give his proof (1814)^{1}.
Let $f\colon \mathbb{C} \to \mathbb{C}$ be a nonconstant polynomial mapping, and suppose $f$ has no zero.
Let $s$ be the infimum of values ${|f(z)|}$; choose a sequence $z_1, z_2, z_3, \ldots$ such that ${|f(z_n)|} \to s$. Since $\lim_{z \to \infty} f(z) = \infty$, the sequence $z_n$ must be bounded; by the Bolzano-Weierstrass theorem it has a subsequence $z_{n_k}$ that converges to some point $z_0$. Then ${|f(z_{n_k})|}$ converges to ${|f(z_0)|}$ by continuity, and converges to $s$ as well, so ${|f(z)|}$ attains an absolute minimum $s$ at $z = z_0$. By supposition, $f(z_0) \neq 0$.
The polynomial function $f$ may be uniquely written in the form
where $g$ is polynomial function and $g(z_0) \neq 0$. Put
and choose $\delta \gt 0$ small so that
$F$ maps the circle $C = \{z : {|z - z_0|} = \delta\}$ onto a circle of radius $r = {|g(z_0)|}\delta^n$ centered at $f(z_0)$. (This uses the fact that any complex number has an $n^{th}$ root, which one can prove using polar coordinate representations. We omit the details.) Choose $z' \in C$ so that $F(z')$ is on the line segment between the origin and $f(z_0)$ (we can always choose $\delta$ so that also $r \lt {|f(z_0)|}$). Then
We also have
according to how we chose $\delta$ in 2. We conclude by observing the strict inequality
which contradicts the fact that ${|f(z)|}$ attains an absolute minimum at $z = z_0$.
Many proofs rely explicitly on the double negation rule by first supposing that a polynomial function $p$ has no root and deriving a contradiction.
In constructive mathematics, integral closure and algebraic closure of a field are not the same, because not every nonconstant polynomial function has a well-defined degree. This is still the case even if we take “nonconstant polynomial function”, we mean “apart from every constant polynomial function”. In general, every nonconstant polynomial function has a well-defined degree if and only if the field is a discrete field.
Thus, the fundamental theorem of algebra is usually expressed in terms of integral closure, which classically is the same as algebraic closure for any field:
There is a second problem, namely with Lemma . It is usually impossible to prove that the monic quadratic function $x \mapsto x^2 + c$ has a root for all complex numbers $c \in \mathbb{C}$. The problem here is once again that any constructive notion of complex numbers is only a Heyting field, rather than a discrete field. The function $x \mapsto x^2 + c$ has a root for $c = 0$ and for invertible $c$, but it is not provable that every complex number $c \in \mathbb{C}$ is either invertible or equal to zero. In fact, this fails in certain topoi, such as sheaves over $\mathbb{C}$, because that the complex numbers have a continuous square root over $\mathbb{C}$ can only be proven in the presence of weak countable choice.
However, the following weaker theorem is still true in constructive mathematics for the modulated Cantor complex numbers:
Every unramifiable monic polynomial function with coefficients in $\mathbb{C}$ has a simple root in $\mathbb{C}$.
A fully choice-free constructive proof of this theorem by Wim Ruitenberg (Ruitenberg 1991) exists for the modulated Cantor complex numbers (which agree with the Dedekind complex numbers by weak countable choice or the limited principle of omniscience). It is unknown if this weaker FTA is true of the Dedekind complex numbers as well.
Fred Richman (1998) has proposed that, in the absence of $WCC$, the FTA should be interpreted as a statement about sets of roots rather than about individual roots. He constructs a complete metric space $\hat{M}_n(\mathbb{C})$ which, classically, is the space of $n$-element multisets of complex numbers (and constructively is the completion of that space) and proves that every complex polynomial function $p$ of degree $n$ may be associated with a point in this space in such a way that the $n$ elements of that point (when viewed as a multiset, if possible, and morally in any case) are the $n$ roots of $p$.
Assuming classical logic, but weak foundations, it can be shown that FTA is true in the reverse mathematics system $RCA_0$ (Tanaka-Yamazaki 2005).
The proof was attempted many times before Gauss gave what is accepted as the first proof in his dissertation (Gauss 1799), although this was not without issues (Gauss ‘fixed’ this proof almost 50 years later, but the last gap was not filled until the 20th century).
All proofs of this fact (of which there are many) require something analytic, in the sense that ordinary algebra will not suffice: one needs to know that the real numbers (or the complex numbers) ‘have no algebraic gaps’. For instance, the rational numbers famously don’t contain the square root of $2$. The cleanest proof I know, due to Artin, that isolates this analytic germ, uses the step-ladder result that the real numbers form what is called a real closed field. This is essentially saying that non-negative real numbers have square roots, and odd degree polynomial functions have roots (anyone who has plotted a cubic can appreciate this fact). Alternatively, one can characterise real closed fields as those for whom the Intermediate Value Theorem (IVT) holds for polynomial functions. Accepting this result (which does need proof), the FTA follows using pure algebra (although not of the high-school sort).
However, it is of interest, partly theoretical, partly for the sake of finding the bare minimum needed to prove the FTA, to know an elementary proof, namely one that minimises the use of analytic techniques (for instance, the IVT for polynomial functions follows from the IVT for continuous functions, but that is like killing a mosquito with a bazooka). Gauss’ second proof (Gauss 1866) is elementary (and predates Artin’s by a long time). Since Gauss lacked modern algebraic techniques, some of his proof is laborious, but (Taylor 85) gives a modern gloss. (With some amusing side notes: as Taylor puts it – ‘Gauss takes the opportunity [to] be rude to his inferior contemporaries’.) Gauss’ proof, in modern language, takes up less than a page and a half, but this presupposes familiarity with some of the theory of fields (but which is pure algebra). Artin’s proof, by comparison, drawing on major theorems can be given in half a page.
It should be noted, in the context of the last statement, that proofs of the FTA can be given, relying on analytic ‘bazooka’ theorems, that are one sentence. However, to spell out the proofs of the necessary theorems, one needs a course in analysis, of some variety, so one is merely sweeping a lot under a very small rug.
Carl Gauss, Demonstratio nova theorematis functionem algebraicam rationalem integramunius variabilis in factores reales primi vel secundi gradus resolvi posse, Dissertation, Helmstedt (1799); Werke 3, 1–30 (1866) (English transl. pdf))
Carl Gauss, Demonstratio nova altera theorematis omnem functionem algebraicamrationalem integram unius variabilis in factores reales primi vel secundi gradus resolviposse, Comm. Soc. Reg. Sci. Göttingen 3, 107–142 (1816); Werke 3, 33–56 (1866)
Michael Eisermann. An Elementary Real-Algebraic Proof via Sturm Chains. pdf
Another new proof of the theorem that every integral rational algebraic function of one variable can be resolved into real factors of the first or second degree translated by Paul Taylor and B. Leak (1983) (web)
Paul Taylor, Gauss’ Second Proof, Eureka 45 (1985) 42-47 (pdf)
A proof that Cauchy sequences of rational numbers satisfy the fundamental theorem of algebra:
The Reverse Mathematical treatment is given in
A constructive algebraic proof of the fundamental theorem of algebra for the modulated Cauchy real numbers without choice princples such as weak countable choice
A full formalization in the Coq proof assistant is in
See also
Despite the credit given to Gauss for his demonstration of 1799, Argand’s proof is often credited as the first one that is fully rigorous. The proof given here also uses the Bolzano-Weierstrass theorem, first proven by Bolzano in 1817, making it somewhat contemporaneous. Argand is also widely credited as the one who introduced the cutting-edge idea of viewing complex numbers and their operations geometrically, which the proof here also uses (the complex plane $\mathbb{C}$ being also known as the Argand plane). ↩
Last revised on May 20, 2023 at 15:00:21. See the history of this page for a list of all contributions to it.