2.1 Homomorphisms from simple extensions.

Let FF be a field, and let EE and EE^{\prime} be fields containing FF. Recall that an FF-homomorphism is a homomorphism φ:EE\varphi\colon E\rightarrow E^{\prime} such that φ(a)=a\varphi(a)=a for all aFa\in F. Thus an FF-homomorphism φ\varphi maps a polynomial

ai1imα1i1αmim,ai1imF,αiE,\sum a_{i_{1}\cdots i_{m}}\alpha_{1}^{i_{1}}\cdots\alpha_{m}^{i_{m}},\quad a_{% i_{1}\cdots i_{m}}\in F,\quad\alpha_{i}\in E,

to

ai1imφ(α1)i1φ(αm)im.\sum a_{i_{1}\cdots i_{m}}\varphi(\alpha_{1})^{i_{1}}\cdots\varphi(\alpha_{m})% ^{i_{m}}.

An FF-isomorphism is a bijective FF-homomorphism.

An FF-homomorphism EEE\rightarrow E^{\prime} of fields is, in particular, an injective FF-linear map of FF-vector spaces, and so it is an FF-isomorphism if EE and EE^{\prime} have the same finite degree over FF.

Proposition 2.1

Let F(α)F(\alpha) be a simple extension of FF and Ω\Omega a second extension of FF.

  1. (a)

    Let α\alpha be transcendental over FF. For every FF-homomorphism φ:F(α)Ω\varphi\colon F(\alpha)\rightarrow\Omega, φ(α)\varphi(\alpha) is transcendental over FF, and the map φφ(α)\varphi\mapsto\varphi(\alpha) defines a one-to-one correspondence

    {F-homomorphisms F(α)Ω}{elements of Ω transcendental over F}.\{F\text{-homomorphisms }F(\alpha)\rightarrow\Omega\}\leftrightarrow\{\text{% elements of }\Omega\text{ transcendental over }F\}.
  2. (b)

    Let α\alpha be algebraic over FF with minimal polynomial f(X)f(X). For every FF-homomorphism φ:F[α]Ω\varphi\colon F[\alpha]\rightarrow\Omega, φ(α)\varphi(\alpha) is a root of f(X)f(X) in Ω\Omega, and the map φφ(α)\varphi\mapsto\varphi(\alpha) defines a one-to-one correspondence

    {F-homomorphisms φ:F[α]Ω}{roots of f in Ω}.\{F\text{-homomorphisms }\varphi\colon F[\alpha]\rightarrow\Omega\}% \leftrightarrow\{\text{roots of{\ }}f\text{\ in }\Omega\}.

    In particular, the number of such maps is the number of distinct roots of ff in Ω\Omega.

Proof.

(a) To say that α\alpha is transcendental over FF means that F[α]F[\alpha] is isomorphic to the polynomial ring in the symbol α\alpha. Therefore, for every γΩ\gamma\in\Omega, there is a unique FF-homomorphism φ:F[α]Ω\varphi\colon F[\alpha]\rightarrow\Omega such that φ(α)=γ\varphi(\alpha)=\gamma (see 1.5). This φ\varphi extends (uniquely) to the field of fractions F(α)F(\alpha) of F[α]F[\alpha] if and only if nonzero elements of F[α]F[\alpha] are sent to nonzero elements of Ω\Omega, which is the case if and only if γ\gamma is transcendental over FF. Thus we see that there are one-to-one correspondences between (a) the FF-homomorphisms F(α)ΩF(\alpha)\rightarrow\Omega, (b) the FF-homomorphisms φ:F[α]Ω\varphi\colon F[\alpha]\rightarrow\Omega such that φ(α)\varphi(\alpha) is transcendental, (c) the transcendental elements of Ω\Omega.

(b) Let f(X)=aiXif(X)=\sum a_{i}X^{i}, and consider an FF-homomorphism φ:F[α]Ω\varphi\colon F[\alpha]\rightarrow\Omega. On applying φ\varphi to the equality aiαi=0\sum a_{i}\alpha^{i}=0, we obtain the equality aiφ(α)i=0\sum a_{i}\varphi(\alpha)^{i}=0, which shows that φ(α)\varphi(\alpha) is a root of f(X)f(X) in Ω\Omega. Conversely, if γΩ\gamma\in\Omega is a root of f(X)f(X), then the map F[X]ΩF[X]\rightarrow\Omega, g(X)g(γ)g(X)\mapsto g(\gamma), factors through F[X]/(f(X))F[X]/(f(X)). When composed with the inverse of the canonical isomorphism F[X]/(f(X))F[α]F[X]/(f(X))\rightarrow F[\alpha], this becomes a homomorphism F[α]ΩF[\alpha]\rightarrow\Omega sending α\alpha to γ\gamma.

We shall need a slight generalization of this result.

Proposition 2.2

Let F(α)F(\alpha) be a simple extension of FF and φ0:FΩ\varphi_{0}\colon F\rightarrow\Omega a homomorphism from FF into a second field Ω\Omega.

  1. (a)

    If α\alpha is transcendental over FF, then the map φφ(α)\varphi\mapsto\varphi(\alpha) defines a one-to-one correspondence

    {extensions φ:F(α)Ω of φ0}{elements of Ω transcendental over φ0(F)}.\{\text{extensions }\varphi\colon F(\alpha)\rightarrow\Omega\text{{\ of }}\varphi_{0}\}\leftrightarrow\{\text{elements of }\Omega\text{{\ % transcendental over }}\varphi_{0}(F)\}.
  2. (b)

    If α\alpha is algebraic over FF, with minimal polynomial f(X)f(X), then the map φφ(α)\varphi\mapsto\varphi(\alpha) defines a one-to-one correspondence

    {extensions φ:F[α]Ω of φ0}{roots of φ0f in Ω}.\{\text{extensions }\varphi\colon F[\alpha]\rightarrow\Omega\text{{\ of }}\varphi_{0}\}\leftrightarrow\{\text{{roots of }}\varphi_{0}f\text{ in }\Omega\}.

    In particular, the number of such maps is the number of distinct roots of φ0f\varphi_{0}f in Ω\Omega.

By φ0f\varphi_{0}f we mean the polynomial obtained by applying φ0\varphi_{0} to the coefficients of ff. By an extension of φ0\varphi_{0} to F(α)F(\alpha) we mean a homomorphism φ:F(α)Ω\varphi\colon F(\alpha)\rightarrow\Omega whose restriction to FF is φ0\varphi_{0}. The proof of the proposition is essentially the same as that of the preceding proposition (indeed, it is essentially the same proposition).