|
|
Line 1: |
Line 1: |
− | The branch of [[Number theory|number theory]] that investigates properties of the integers by elementary methods. These methods include the use of divisibility properties, various forms of the axiom of induction and combinatorial arguments. Sometimes the notion of elementary methods is extended by bringing in the simplest elements of mathematical analysis. Traditionally, proofs are deemed to be non-elementary if they involve complex numbers.
| + | <!-- |
| + | d0336801.png |
| + | $#A+1 = 44 n = 0 |
| + | $#C+1 = 44 : ~/encyclopedia/old_files/data/D033/D.0303680 Division algebra |
| + | Automatically converted into TeX, above some diagnostics. |
| + | Please remove this comment and the {{TEX|auto}} line below, |
| + | if TeX found to be correct. |
| + | --> |
| | | |
− | Usually, one refers to elementary number theory the problems that arise in branches of number theory such as the theory of divisibility, of congruences, of arithmetic functions, of indefinite equations, of partitions, of additive representations, of the approximation by rational numbers, and of continued fractions. Quite often, the solution of such problems leads to the need to go beyond the framework of elementary methods. Occasionally, following the discovery of a non-elementary solution of some problem, one also finds an elementary solution of it.
| + | {{TEX|auto}} |
| + | {{TEX|done}} |
| | | |
− | As a rule, the problems of elementary number theory have a history going back over centuries, and they are quite often a source of modern trends in number theory and algebra.
| + | An algebra $ A $ |
− | | + | over a field $ F $ |
− | From preserved ancient Babylonian cuneiform tablets one may deduce that the Babyloneans were familiar with the task of factoring natural numbers into prime factors. In the 5th century B.C. the Pythagoreans established the so-called doctrine of even and odd numbers and justified the proposition that the product of two natural numbers is even if and only if at least one of the factors is even. A general theory of divisibility was created, in essence, by Euclid. In his Elements (3rd century B.C.), he introduces an algorithm for finding the greatest common divisor of two integers and on this basis he justifies the main theorem of the arithmetic of integers: Every natural number can be factored in one and only one way into a product of prime factors.
| + | such that for any elements $ a \neq 0 $ |
− | | + | and $ b $ |
− | After C.F. Gauss, at the beginning of the 19th century, constructed a theory of divisibility of complex integers, it became clear that the study of an arbitrary [[Ring|ring]] must begin with the construction of a divisibility theory in it.
| + | the equations $ ax = b $, |
− | | + | $ ya = b $ |
− | All properties of the integers are connected in one way or another with the prime numbers (cf. [[Prime number|Prime number]]). Therefore, questions on the disposition of the prime numbers in the sequence of natural numbers evoked the interest of scholars. The first proof that the set of prime numbers is infinite is due to Euclid. Only in the middle of the 19th century did P.L. Chebyshev take the following step in the study of the function <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e035/e035350/e0353501.png" />, the number of prime numbers not exceeding <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e035/e035350/e0353502.png" />. He succeeded in proving by elementary means inequalities that imply
| + | are solvable in $ A $. |
− | | + | An associative division algebra, considered as a ring, is a skew-field, its centre $ C $ |
− | <table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;"><img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e035/e035350/e0353503.png" /></td> </tr></table>
| + | is a field, and $ C \supseteq F $. |
− | | + | If $ C = F $, |
− | for all sufficiently large <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e035/e035350/e0353504.png" />. Actually, <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e035/e035350/e0353505.png" />, but this was not established until the end of the 19th century by means of complex analysis. For a long time it was considered impossible to obtain the result by elementary means. However, in 1949, A. Selberg obtained an elementary proof of this theorem. Chebyshev also proved that for <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e035/e035350/e0353506.png" /> the interval <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e035/e035350/e0353507.png" /> contains at least one prime number. A refinement of the interval containing at least one prime number requires a deeper study of the behaviour of the function <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e035/e035350/e0353508.png" />.
| + | the division algebra $ A $ |
− | | + | is called a central division algebra. Finite-dimensional central associative division algebras over $ F $ |
− | In the 3rd century B.C. the sieve of Eratosthenes (cf. [[Eratosthenes, sieve of|Eratosthenes, sieve of]]) was used to select the prime numbers from the set of natural numbers. In 1918 V. Brun showed that a modification of this method can serve as a basis for the study of a set of "almost primes" . He proved the [[Brun theorem|Brun theorem]] on prime twins. The [[Brun sieve|Brun sieve]] is a special case of general sieve methods (cf. [[Sieve method|Sieve method]]) which give estimates for collections of almost primes not exceeding <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e035/e035350/e0353509.png" /> and belonging to a sequence <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e035/e035350/e03535010.png" /> of natural numbers. Sieve methods can be used when "good" approximations in the mean are known for the amount of integers <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e035/e035350/e03535011.png" /> belonging to a progression the modulus of which grows with <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e035/e035350/e03535012.png" />. Among the sieve methods developed after Brun, the [[Selberg sieve|Selberg sieve]] is of special significance. The strongest results are obtained by a combination of sieve methods and analytic methods. A sieve method in conjunction with the [[Shnirel'man method|Shnirel'man method]] made it possible to effectively find a number <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e035/e035350/e03535013.png" /> such that any natural number <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e035/e035350/e03535014.png" /> can be represented as a sum of at most <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e035/e035350/e03535015.png" /> prime numbers.
| + | may be identified, up to an isomorphism, with the elements of the [[Brauer group|Brauer group]] $ B( F ) $ |
− | | + | of the field $ F $. |
− | Two integers <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e035/e035350/e03535016.png" /> and <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e035/e035350/e03535017.png" /> are said to be congruent modulo <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e035/e035350/e03535018.png" /> if they have the same remainder on division by <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e035/e035350/e03535019.png" />. Gauss (in 1801) introduced the notation <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e035/e035350/e03535020.png" /> <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e035/e035350/e03535021.png" />. This form of writing, which brings about the analogy of congruences and equations, turned out to be convenient and instrumental in the development of the theory congruences (cf. [[Congruence|Congruence]]).
| + | Let $ [ A: F ] $ |
− | | + | denote the dimension of $ A $ |
− | Many results obtained previously by P. Fermat, L. Euler, J.L. Lagrange, and others, and also the [[Chinese remainder theorem|Chinese remainder theorem]], can be stated and proved simply in the language of the theory of congruences. One of the most interesting results of this theory is the [[Quadratic reciprocity law|quadratic reciprocity law]].
| + | over $ F $. |
− | | + | If $ A \in B( F ) $ |
− | The ancient Babylonians knew a large number of "Pythagorean triples" . Apparently they knew some method for finding arbitrary many integer solutions of the indeterminate equation <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e035/e035350/e03535022.png" />. The Pythagoreans used the formulas <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e035/e035350/e03535023.png" />, <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e035/e035350/e03535024.png" />, <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e035/e035350/e03535025.png" /> to find solutions of this equation. Euclid indicated a method that allows one to find in succession integer solutions of the equation <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e035/e035350/e03535026.png" /> (a special case of the [[Pell equation|Pell equation]]). In his Aritmetika Diophantus (3rd century B.C.) made an attempt to setting up a theory of indeterminate equations (see [[Diophantine equations|Diophantine equations]]). In particular, for the solution of equations of the second and some higher degrees he used systematically a device that enabled him to find from one rational solution of a given equation other rational solutions of it. Fermat (17th century) discovered another method, the "method of descent" , and solved by it a number of equations, but the so-called [[Fermat great theorem|Fermat great theorem]], which he declared to have solved, has turned out to be beyond the power of elementary methods.
| + | and if $ L $ |
− | | + | is the maximal subfield in $ A $ ($ L \supseteq F $), |
− | Fermat solved the problem of representing natural numbers by sums of two squares of integers. As a result of research by Lagrange (1773) and Gauss (1801) the problem of the representation of integers by a definite [[Binary quadratic form|binary quadratic form]] was solved. Gauss developed the general theory of binary quadratic forms. The solution of the problem of representing numbers by forms of higher degree (for example, the [[Waring problem|Waring problem]]) and by quadratic forms in several variables usually go beyond the framework of elementary methods. Only certain special cases of such problems can be solved elementarily. An example is Lagrange's theorem: Every natural number is the sum of four squares of integers. It should be mentioned that Diophantus in his Aritmetika repeatedly used the possibility of representing a natural number as the sum of four squares of integers.
| + | then $ [ A: F ] = {[ L: F ] } ^ {2} $. |
− | | + | According to the [[Frobenius theorem|Frobenius theorem]], all associative finite-dimensional division algebras over the field of real numbers $ \mathbf R $ |
− | To elementary number theory one also reckons the theory of partitions, the foundations of which were laid by Euler (in 1751). One of the basic problems of the theory of partitions is the study of the function <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e035/e035350/e03535027.png" />, the number of representations of a natural number <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e035/e035350/e03535028.png" /> as a sum of natural numbers. Other functions of similar type are also treated in the theory of partitions. Continued fractions (cf. [[Continued fraction|Continued fraction]]) appeared in connection with problems of approximate computations (extraction of the square root of a natural number, search for approximations of real numbers by common fractions with small denominators). Continued fractions are applied in solving indefinite equations of the first and second degree. Using the apparatus of continued fractions J. Lambert (1766) was the first to establish that the number <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e035/e035350/e03535029.png" /> is irrational. In the solution of various problems in the approximation of real numbers by rational numbers one uses in elementary number theory the [[Dirichlet principle|Dirichlet principle]], besides continued fractions.
| + | are exhausted by $ \mathbf R $ |
− | | + | itself, the field of complex numbers, and the [[Quaternion|quaternion]] algebra. For this reason the group $ B( \mathbf R ) $ |
− | In number theory it is easy to state many problems that can be formulated elementarily and have so far remained unsolved. For example; Is the set of even perfect numbers finite or not? Is there at least one odd perfect number? Is the set of Fermat prime numbers finite or not? Is the set of Mersenne prime numbers finite or not (cf. [[Mersenne number|Mersenne number]])? Is the set of prime numbers of the form <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e035/e035350/e03535030.png" /> finite or not? Is it true that there is at least one prime number between the squares of two consecutive natural numbers? Is the set of incomplete fractions of the expansion of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e035/e035350/e03535031.png" /> into a continued fraction bounded or not?
| + | is cyclic of order two. If the associativity requirement is dropped, there is yet another example of a division algebra over the field of real numbers: the [[Cayley–Dickson algebra|Cayley–Dickson algebra]]. This algebra is alternative, and its dimension over $ \mathbf R $ |
| + | is 8. If $ A $ |
| + | is a finite-dimensional (not necessarily associative) division algebra over $ \mathbf R $, |
| + | then $ [ A: \mathbf R ] $ |
| + | has one of the values 1, 2, 4, or 8. |
| | | |
| ====References==== | | ====References==== |
− | <table><TR><TD valign="top">[1]</TD> <TD valign="top"> B.A. Venkov, "Elementary number theory" , Wolters-Noordhoff (1970) (Translated from Russian) {{MR|0265267}} {{ZBL|0204.37101}} </TD></TR><TR><TD valign="top">[2]</TD> <TD valign="top"> I.M. Vinogradov, "Basic number theory" , Moscow (1981) (In Russian)</TD></TR><TR><TD valign="top">[3]</TD> <TD valign="top"> A.O. Gel'fond, Yu.V. Linnik, "Elementary methods in the analytic theory of numbers" , M.I.T. (1966) (Translated from Russian) {{MR|201368}} {{ZBL|0142.01403}} </TD></TR><TR><TD valign="top">[4]</TD> <TD valign="top"> A.Ya. Khinchin, "Continued fractions" , Univ. Chicago Press (1964) (Translated from Russian) {{MR|0161833}} {{ZBL|0117.28601}} </TD></TR><TR><TD valign="top">[5]</TD> <TD valign="top"> C.F. Gauss, "Disquisitiones Arithmeticae" , Teubner (1801) (Translated from Latin) {{MR|2308276}} {{MR|1876694}} {{MR|1847691}} {{MR|1356001}} {{MR|1167370}} {{MR|1138220}} {{MR|0837656}} {{MR|0638487}} {{MR|0479854}} {{MR|0197380}} {{ZBL|1167.11001}} {{ZBL|0899.01034}} {{ZBL|0585.10001}} {{ZBL|0136.32301}} {{ZBL|21.0166.04}} </TD></TR><TR><TD valign="top">[6]</TD> <TD valign="top"> H. Davenport, "The higher arithmetic" , Hutchinson (1952) {{MR|0050598}} {{ZBL|0049.30901}} </TD></TR><TR><TD valign="top">[7]</TD> <TD valign="top"> E. Trost, "Primzahlen" , Birkhäuser (1953) {{MR|0058630}} {{ZBL|0053.36002}} </TD></TR><TR><TD valign="top">[8]</TD> <TD valign="top"> G.E. Andrews, "Theory of partitions" , Addison-Wesley (1976) {{MR|0557013}} {{ZBL|0371.10001}} </TD></TR><TR><TD valign="top">[9]</TD> <TD valign="top"> , ''The history of mathematics from Antiquity to the beginning of the XIX-th century'' , '''1–3''' , Moscow (1970–1972) (In Russian)</TD></TR><TR><TD valign="top">[10]</TD> <TD valign="top"> H. Wieleitner, "Die Geschichte der Mathematik von Descartes bis zum Hälfte des 19. Jahrhunderts" , de Gruyter (1923)</TD></TR><TR><TD valign="top">[11]</TD> <TD valign="top"> L.E. Dickson, "History of the theory of numbers" , '''1–3''' , Chelsea, reprint (1971) {{MR|1793101}} {{MR|1720467}} {{MR|0245501}} {{MR|0245500}} {{MR|0245499}} {{MR|1520248}} {{MR|1519706}} {{MR|1519382}} {{ZBL|1214.11003}} {{ZBL|1214.11002}} {{ZBL|1214.11001}} {{ZBL|0958.11500}} {{ZBL|60.0817.03}} {{ZBL|49.0100.12}} {{ZBL|48.1114.03}} {{ZBL|48.0137.02}} {{ZBL|47.0100.04}} </TD></TR><TR><TD valign="top">[12]</TD> <TD valign="top"> G.H. Hardy, E.M. Wright, "An introduction to the theory of numbers" , Oxford Univ. Press (1979) {{MR|0568909}} {{ZBL|0423.10001}} </TD></TR><TR><TD valign="top">[13]</TD> <TD valign="top"> W. Sierpiński, "Elementary theory of numbers" , PWN (1964) (Translated from Polish) {{MR|0227080}} {{MR|0175840}} {{ZBL|0122.04402}} </TD></TR><TR><TD valign="top">[14]</TD> <TD valign="top"> K. Ireland, M. Rosen, "A classical introduction to modern number theory" , Springer (1982) {{MR|0661047}} {{ZBL|0482.10001}} </TD></TR></table> | + | <table><TR><TD valign="top">[1]</TD> <TD valign="top"> A.G. Kurosh, "Lectures on general algebra" , Chelsea (1963) (Translated from Russian) {{MR|0158000}} {{ZBL|0121.25901}} </TD></TR><TR><TD valign="top">[2]</TD> <TD valign="top"> A.A. Albert, "Structure of algebras" , Amer. Math. Soc. (1939) {{MR|0000595}} {{ZBL|0023.19901}} {{ZBL|65.0094.02}} </TD></TR><TR><TD valign="top">[3]</TD> <TD valign="top"> I.N. Herstein, "Noncommutative rings" , Math. Assoc. Amer. (1968) {{MR|1535024}} {{MR|0227205}} {{ZBL|0177.05801}} </TD></TR><TR><TD valign="top">[4]</TD> <TD valign="top"> J.F. Adams, "On the non-existence of elements of Hopf invariant one" ''Ann. of Math.'' , '''72''' : 1 (1960) pp. 20–104 {{MR|0141119}} {{ZBL|0096.17404}} </TD></TR></table> |
− | | |
− | | |
| | | |
| ====Comments==== | | ====Comments==== |
− | A translation of [[#References|[5]]] into English is [[#References|[a1]]].
| + | Over a finite field every finite-dimensional central division algebra is automatically commutative. For infinite-dimensional division algebras the situation is quite different, because a result of Mokar–Limonov states that such an algebra contains a free algebra in two variables. |
− | | |
− | For Chebyshev's result mentioned above see also [[Chebyshev theorems on prime numbers|Chebyshev theorems on prime numbers]]. | |
− | | |
− | A Pythagorean triple is a triple <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e035/e035350/e03535032.png" /> of natural numbers satisfying <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e035/e035350/e03535033.png" /> (cf. also [[Pythagorean numbers|Pythagorean numbers]]).
| |
| | | |
− | For representations of numbers by special (quadratic) forms see also [[Quadratic form|Quadratic form]]; [[Goldbach problem|Goldbach problem]]. More on partitions can be found in [[Combinatorial analysis|Combinatorial analysis]].
| + | If a finite-dimensional central division algebra $ D $ |
| + | contains a maximal commutative subfield $ L $ |
| + | which is a [[Galois extension|Galois extension]] of $ F $, |
| + | then $ D $ |
| + | is a [[Cross product|cross product]] of $ L $ |
| + | and $ G = \mathop{\rm Gal} ( L/ F ) $ |
| + | in the sense that $ D $ |
| + | is the free $ L $-module generated by $ \{ {u _ \sigma } : {\sigma \in G } \} $ |
| + | with product determined by: |
| | | |
− | Concerning the questions stated at the end of the main article above see also [[Perfect number|Perfect number]]; [[Prime number|Prime number]]; [[Twins|Twins]].
| + | $$ \tag{a1 } |
| + | \left . |
| + | \begin{array}{ll} |
| + | u _ \sigma u _ \tau = c ( \sigma , \tau ) u _ {\sigma \tau } &\textrm{ for some } c ( \sigma , \tau ) \in L ^ {*} , \\ |
| + | u _ \sigma \lambda = \lambda ^ \sigma u _ \sigma &\textrm{ for } \lambda \in L ,\ \tau \in G . \\ |
| + | \end{array} |
| + | \right \} |
| + | $$ |
| | | |
− | Fermat numbers are the special case of Mersenne numbers, which are of the form <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e035/e035350/e03535034.png" />, when <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e035/e035350/e03535035.png" /> itself is of the form <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e035/e035350/e03535036.png" /> (cf. [[Mersenne number|Mersenne number]]); here <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e035/e035350/e03535037.png" />, <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e035/e035350/e03535038.png" /> are natural numbers.
| + | Associativity of $ D $ |
| + | entails that $ c : G \times G \rightarrow L ^ {*} $ |
| + | represents an element of $ H ^ {2} ( G , L ^ {*} ) $ (the second [[Galois cohomology|Galois cohomology]] group). One of the basic problems in algebra was formulated by A. Albert (1931): Is every finite-dimensional central division algebra necessarily a cross product? In 1972, S. Amitsur provided a counter-example using properties of generic division algebras resulting from the theory of PI-algebras (see [[PI-algebra|PI-algebra]], [[#References|[a2]]]). Other examples of division algebras were obtain by F. van Ostaeyen (1972 Thesis, cf. [[#References|[a3]]]), i.e. generic cross products, a notion generalized by Amitsur and D. Saltman (1978), describing all cross product division algebras for a given group $ G $ |
| + | over the field $ F $ |
| + | as reductions of a generic division algebra. |
| | | |
| ====References==== | | ====References==== |
− | <table><TR><TD valign="top">[a1]</TD> <TD valign="top"> C.F. Gauss, "Disquisitiones Arithmeticae" , Yale Univ. Press (1966) (Translated from Latin) {{MR|0197380}} {{ZBL|0136.32301}} </TD></TR><TR><TD valign="top">[a2]</TD> <TD valign="top"> D. Shanks, "Solved and unsolved problems in number theory" , Chelsea, reprint (1978) {{MR|0516658}} {{ZBL|0397.10001}} </TD></TR><TR><TD valign="top">[a3]</TD> <TD valign="top"> A. Weil, "Number theory: an approach through history: from Hammupari to Legendre" , Birkhäuser (1984)</TD></TR></table> | + | <table><TR><TD valign="top">[a1]</TD> <TD valign="top"> A.H. Schofield, "Representations of rings over skew fields" , London Math. Soc. (1986) {{MR|0800853}} {{ZBL|}} </TD></TR><TR><TD valign="top">[a2]</TD> <TD valign="top"> N. Jacobson, "PI algebras. An introduction" , Springer (1975) {{MR|0369421}} {{ZBL|0326.16013}} </TD></TR><TR><TD valign="top">[a3]</TD> <TD valign="top"> F. van Oystaeyen, "Prime spectra in non-commutative algebra" , Springer (1975) {{MR|}} {{ZBL|0302.16001}} </TD></TR></table> |
An algebra $ A $
over a field $ F $
such that for any elements $ a \neq 0 $
and $ b $
the equations $ ax = b $,
$ ya = b $
are solvable in $ A $.
An associative division algebra, considered as a ring, is a skew-field, its centre $ C $
is a field, and $ C \supseteq F $.
If $ C = F $,
the division algebra $ A $
is called a central division algebra. Finite-dimensional central associative division algebras over $ F $
may be identified, up to an isomorphism, with the elements of the Brauer group $ B( F ) $
of the field $ F $.
Let $ [ A: F ] $
denote the dimension of $ A $
over $ F $.
If $ A \in B( F ) $
and if $ L $
is the maximal subfield in $ A $ ($ L \supseteq F $),
then $ [ A: F ] = {[ L: F ] } ^ {2} $.
According to the Frobenius theorem, all associative finite-dimensional division algebras over the field of real numbers $ \mathbf R $
are exhausted by $ \mathbf R $
itself, the field of complex numbers, and the quaternion algebra. For this reason the group $ B( \mathbf R ) $
is cyclic of order two. If the associativity requirement is dropped, there is yet another example of a division algebra over the field of real numbers: the Cayley–Dickson algebra. This algebra is alternative, and its dimension over $ \mathbf R $
is 8. If $ A $
is a finite-dimensional (not necessarily associative) division algebra over $ \mathbf R $,
then $ [ A: \mathbf R ] $
has one of the values 1, 2, 4, or 8.
References
Over a finite field every finite-dimensional central division algebra is automatically commutative. For infinite-dimensional division algebras the situation is quite different, because a result of Mokar–Limonov states that such an algebra contains a free algebra in two variables.
If a finite-dimensional central division algebra $ D $
contains a maximal commutative subfield $ L $
which is a Galois extension of $ F $,
then $ D $
is a cross product of $ L $
and $ G = \mathop{\rm Gal} ( L/ F ) $
in the sense that $ D $
is the free $ L $-module generated by $ \{ {u _ \sigma } : {\sigma \in G } \} $
with product determined by:
$$ \tag{a1 }
\left .
\begin{array}{ll}
u _ \sigma u _ \tau = c ( \sigma , \tau ) u _ {\sigma \tau } &\textrm{ for some } c ( \sigma , \tau ) \in L ^ {*} , \\
u _ \sigma \lambda = \lambda ^ \sigma u _ \sigma &\textrm{ for } \lambda \in L ,\ \tau \in G . \\
\end{array}
\right \}
$$
Associativity of $ D $
entails that $ c : G \times G \rightarrow L ^ {*} $
represents an element of $ H ^ {2} ( G , L ^ {*} ) $ (the second Galois cohomology group). One of the basic problems in algebra was formulated by A. Albert (1931): Is every finite-dimensional central division algebra necessarily a cross product? In 1972, S. Amitsur provided a counter-example using properties of generic division algebras resulting from the theory of PI-algebras (see PI-algebra, [a2]). Other examples of division algebras were obtain by F. van Ostaeyen (1972 Thesis, cf. [a3]), i.e. generic cross products, a notion generalized by Amitsur and D. Saltman (1978), describing all cross product division algebras for a given group $ G $
over the field $ F $
as reductions of a generic division algebra.
References
[a1] | A.H. Schofield, "Representations of rings over skew fields" , London Math. Soc. (1986) MR0800853 |
[a2] | N. Jacobson, "PI algebras. An introduction" , Springer (1975) MR0369421 Zbl 0326.16013 |
[a3] | F. van Oystaeyen, "Prime spectra in non-commutative algebra" , Springer (1975) Zbl 0302.16001 |