Building Fields from Polynomials
A field containing is an extension of , written . If is irreducible over , then — polynomials over , computed modulo — is a field, and the class of is a root of in it. So every irreducible polynomial has a root in some extension field (Kronecker's theorem).
Working "modulo " is like clock arithmetic with polynomials: every polynomial is replaced by its remainder on division by , so the elements are the polynomials of degree less than . Addition is ordinary, and products are reduced using the rule . This is how arises from : computing with modulo is exactly computing with , since the rule becomes .
Over , is irreducible. In the elements are , with the rule (since ). The nonzero elements multiply by
Every nonzero element has an inverse (the appears in each row), so this is a field with elements — and it is not , where .
In write for the class of . Then , using — the familiar complex multiplication. So .
The polynomial must be irreducible. In , the class of satisfies although : a zero divisor, so this quotient is not a field. That happens because over .
In , compute .
Show the answerHide the answer
. So has order in the group of nonzero elements.
Degree and Bases
An extension of is a vector space over ; its dimension is the degree . If is irreducible of degree , then , with basis . In particular, over this gives a field with elements.
The degree measures how much bigger is than . has degree over (basis ); has degree over (basis ); has infinite degree over . For a finite field the degree pins down the size: a degree- extension of has exactly elements, one for each choice of coefficients.
with basis . with basis , because is irreducible. , since is irreducible (Eisenstein, ).
Over , has no root ( are the values takes, so ), so it is irreducible, and is a field with elements: all with .
Adjoining two square roots does not give degree : has degree , since the minimal polynomial of is . And the degree of equals only when is irreducible.
How many elements does have, given that has no root in ?
Show the answerHide the answer
A cubic with no root is irreducible, so this is a field of degree over , with elements.
Algebraic Elements and Minimal Polynomials
An element of an extension of is algebraic over if it is a root of some nonzero polynomial in . Its minimal polynomial is the monic polynomial of least degree with as a root; it is irreducible, divides every polynomial in vanishing at , and equals its degree.
The minimal polynomial is the most efficient description of over . To find it: produce some polynomial that satisfies, typically by isolating radicals and squaring, then check irreducibility. Numbers like and are not algebraic over (they are transcendental), and has infinite degree.
Let . Then , so , i.e. . This quartic is irreducible over , so it is the minimal polynomial and .
satisfies but . Since , is a root of , which has no rational root and so is irreducible: the minimal polynomial of over is .
A polynomial that satisfies need not be minimal. is a root of , but its minimal polynomial is . Also the minimal polynomial depends on the base field: has minimal polynomial over but over .
Find the minimal polynomial of over .
Show the answerHide the answer
If then , so . This quadratic has no rational root ( is irrational), so the minimal polynomial is .
The Tower Law
If are fields with finite degrees, then . Consequently divides , and the degree of any element of over divides .
The tower law is the field-theory twin of Lagrange's theorem. It is the main tool for computing degrees: adjoin one element at a time, and multiply. It also proves impossibility results — if , then cannot contain , whose degree does not divide .
. Is ? If then , forcing , and then or , impossible in . So stays irreducible over , giving , with basis .
. If were in it, then would give . So — proved without any computation.
, and does not contain , so is irreducible over it: . Total degree .
Do not multiply degrees blindly. has degree , not , because is already in . Always check whether the new element is already in the field built so far.
What is ?
Show the answerHide the answer
, and (the same argument as for ), so the second step also has degree : total .
Why It Works
Let be a polynomial not divisible by . Since is irreducible, , so the Euclidean algorithm gives with . Modulo this says : the class of has inverse the class of . Every nonzero class is invertible.
If the minimal polynomial of factored as with smaller degrees, then in a field, so or — a polynomial of smaller degree with root , contradicting minimality.
Let be a basis of over and a basis of over . Every element of is with , and each with . So the products span over ; reversing the argument shows they are independent. Hence .
If has degree , divide any polynomial in by : only the remainder, of degree less than , matters. So span , and they are independent because a dependence would be a nonzero polynomial of degree less than vanishing at .