Let $F$ be a field and let $B$ be an $F$-algebra. The degree of $B$ over $F$ is the smallest positive integer $\deg_F B = d \geq 1$ such that every element of $B$ satisfies a (monic) polynomial of degree $d$ over $F$ if such an $m$ exists, otherwise $\deg_F B = \infty$.

An algebra of finite degree over $F$ need not be finite-dimensional over $F$: for example, any Boolean ring is an $\mathbb{F}_2$-algebra of degree $2$.

Suppose, however, that $F$ has characteristic $0$ and $D$ is a division $F$-algebra of degree $d \in \mathbb{Z}_{>0}$. Is $D$ finite-dimensional?

By the primitive element theorem, the center $Z(D)$ of such a division algebra $D$ is finite-dimensional over $F$, so one can assume without loss of generality that $D$ is central.

I can prove that the answer is yes when $d=2$: in fact, a division $F$-algebra of degree $2$ is either a quadratic field extension of $F$ or a quaternion algebra over $F$. But the argument is a bit special.

(One can make examples of degree $p$ in characteristic $p$ by considering inseparable field extensions; but the question above is still interesting when $\mathrm{char} F = p > d$.)


1 Answer 1


Yes $D$ must be finite-dimensional over $F$. This follows from a Theorem of Kaplansky's that I found in Herstein's monograph "Noncommutative Rings". The first step is to show that an algebraic algebra of bounded degree satisfies a polynomial identity (see Lemma 6.2.3 in Herstein), i.e., is a P.I. algebra. Note, however, that the degree of this polynomial identity may be larger than the degree of the algebra, as defined in the question. Then since a division algebra $D$ is primitive, one can use:

Theorem 6.3.1 (Kaplansky) If $A$ is a primitive algebra satisfying a polynomial identity of degree $d$ then $A$ is a finite dimensional simple algebra over its center, of dimension at most $\lfloor d/2 \rfloor^2$.

The theorem would still apply in the characteristic $p$ case, but as suggested in the question, it would not follow that the center of $D$ is finite-dimensional over $F$.

Edit: Kaplansky's paper can be found here: http://www.ams.org/journals/bull/1948-54-06/S0002-9904-1948-09049-8/home.html


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

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