From: jpcolm@aol.com (JPCOLM) Subject: Re: Question Date: 21 Jul 1999 02:48:46 GMT Newsgroups: sci.math Keywords: proving commutativity in rings from x^n=x jvoight wrote: >Let $R$ be a ring in which $x^3=x$ for every $x\in R$. Prove that $R$ >is a commutative ring. > >Any takers? > The analogous theorem is true for any n >= 2: if n is a fixed integer and R is a ring for which x^n = x for all x, then R is commutative. I think it is due to Jacobson. The only proof that I know of relies on fairly involved structural considerations (any such ring can be embedded in a product of finite such rings, which can in turn be seen to be commutative since finite division rings are commutative.) It would be of some interest to come up with purely equational proofs of these theorems (one theorem for each n). Such proofs must exist by Birkhoff's completeness theorem for equational logic but for n greater than 3 or 4 they are still (I think) unkown. Is anyone familar with any work in this area? - John Coleman ============================================================================== From: Robin Chapman Subject: Re: Two other questions Date: Wed, 21 Jul 1999 10:49:54 GMT Newsgroups: sci.math To: jvoight34@my-deja.com In article <7n2bu0$2ks$1@nnrp1.deja.com>, jvoight34@my-deja.com wrote: > > > > And: suppose in a ring that x^3=x for all x. Show that the ring is > > > commutative. > > > > It's a theorem of Jacobson that a ring R with the property that > > for each x in R there is m(x) > 1 with x^{m(x)} = x is commutative. > > The proof I've seen is a bit difficult.... > > This is an early problem in a chapter on ring theory, so the proof > can't be that difficult if not proven in its full generality. OK. As (2x)^3 = 2x then 8x^3 = 2x or 8x = 2x for all x in R. Thus 6x = 0 for all x in R. Thus R is the direct sum of the ideals 3R and 2R. Both of these are rings in their own right, both have x^3 = x for all x in them and they have characteristics 2 and 3 respectively. Also 3R and 2R commute since the product of elements from two of them is zero (either way). We can reduce to two special cases (i) R of characteristic 2 and (ii) R of characteristic 3. Expanding out the relations (x+y)^3 - x^3 - y^3 = 0 and (x+y+z)^3 -(x+y)^3 -(x+z)^3 - (y+z)^3 + x^3 + y^3 + z^3 = 0 gives x^2y + xyx + yx^2 + xy^2 + yxy + y^2x = 0 (1) and xyz + xzy + yxz + yzx + zxy + zyx = 0 (2) for all x, y and z. Putting z = x in (2) gives 2(x^2y + xyx + yx^2) = 0 and in characteristic 3 we get x^2y + xyx + yx^2 = 0. Thus in characteristic 3 we have 0 = x(x^2y + xyx + yx^2) - (x^2y + xyx + yx^2)x = x^3y - yx^3 = xy - yx and so R is commutative in characteristic 3. Assume now R has characteristic 2. Then x = -x for all x. Putting z = xy in (2) gives xyxy + x^2y^2 + yx^2y + yxyx + xyxy + xy^2x = 0 or x^2y^2 + yxyx + xy^2y + yx^2y = 0. (3) Swap x and y in (3) and add to 3 to get x^2y^2 + y^2x^2 + xyxy + yxyx = 0. (4) From (1) we get 0 = x(x^2y + xyx + yx^2 + xy^2 + yxy + y^2x) +(x^2y + xyx + yx^2 + xy^2 + yxy + y^2x)x = x^3y + yx^3 + x^2y^2 + xyxy + yxyx + y^2x^2 = xy + yx by (4). Thus xy = yx and R is commutative. -- Robin Chapman http://www.maths.ex.ac.uk/~rjc/rjc.html "They did not have proper palms at home in Exeter." Peter Carey, _Oscar and Lucinda_ Sent via Deja.com http://www.deja.com/ Share what you know. Learn what you don't. ============================================================================== From: "G. A. Edgar" Subject: Re: Question Date: Wed, 21 Jul 1999 08:04:09 -0400 Newsgroups: sci.math > > Let $R$ be a ring in which $x^3=x$ for every $x\in R$. Prove that $R$ > > is a commutative ring. > > This is (word-for-word) in Herstein, Topics in Algebra, Ch. 3 sec. 4, > exercise 19. I once heard Herstein comment that this one problem generated more mail than the entire remainder of the book. -- Gerald A. Edgar edgar@math.ohio-state.edu Department of Mathematics telephone: 614-292-0395 (Office) The Ohio State University 614-292-4975 (Math. Dept.) Columbus, OH 43210 614-292-1479 (Dept. Fax) ============================================================================== From: Robin Chapman Subject: Re: Two other questions Date: Wed, 21 Jul 1999 14:33:39 GMT Newsgroups: sci.math In article <7n4laf$tds$1@nnrp1.deja.com>, chri0562@my-deja.com wrote: > > > OK. As (2x)^3 = 2x then 8x^3 = 2x or 8x = 2x for all x in R. Thus > > 6x = 0 for all x in R. Thus R is the direct sum of the ideals 3R > > and 2R. Both of these are rings in their own right, both have > > x^3 = x for all x in them and they have characteristics 2 and 3 > > respectively. Also 3R and 2R commute since the product of > > elements from two of them is zero (either way). We can reduce > > to two special cases (i) R of characteristic 2 and (ii) R of > > characteristic 3. > > I've just got a little query - what do you mean by 3R and 2R ? > R doesn't have a 1, right? I wasn't assuming that R was unital, since the result doesn't need this. By 3R I mean the set {x+x+x: x in R} etc. -- Robin Chapman http://www.maths.ex.ac.uk/~rjc/rjc.html "They did not have proper palms at home in Exeter." Peter Carey, _Oscar and Lucinda_ Sent via Deja.com http://www.deja.com/ Share what you know. Learn what you don't. ============================================================================== From: Robin Chapman Subject: Re: Two other questions Date: Wed, 21 Jul 1999 20:27:39 GMT Newsgroups: sci.math In article <7n48kg$or4$1@nnrp1.deja.com>, Robin Chapman wrote: > In article <7n2bu0$2ks$1@nnrp1.deja.com>, > jvoight34@my-deja.com wrote: > > > > > > And: suppose in a ring that x^3=x for all x. Show that the ring > is > > > > commutative. > > > > Assume now R has characteristic 2. Then x = -x for all x. I am being really obtuse today. This case is much better dealt with as follows. From (x + x^2)^3 = x + x^2 we get 3x + 3x^2 = 0 or x = x^2 for all x. From this it's easy to see that R is commutative (R is a Boolean ring) as then (x + y)^2 = x + y gives xy + yx = 0. -- Robin Chapman http://www.maths.ex.ac.uk/~rjc/rjc.html "They did not have proper palms at home in Exeter." Peter Carey, _Oscar and Lucinda_ Sent via Deja.com http://www.deja.com/ Share what you know. Learn what you don't. ============================================================================== From: "pascal ORTIZ" Subject: Re: Two other questions Date: 22 Jul 1999 08:31:38 GMT Newsgroups: sci.math Robin Chapman a �crit dans l'article <7n48kg$or4$1@nnrp1.deja.com>... ... > > Expanding out the relations (x+y)^3 - x^3 - y^3 = 0 and > (x+y+z)^3 -(x+y)^3 -(x+z)^3 - (y+z)^3 + x^3 + y^3 + z^3 = 0 > gives > x^2y + xyx + yx^2 + xy^2 + yxy + y^2x = 0 (1) > and > xyz + xzy + yxz + yzx + zxy + zyx = 0 (2) > for all x, y and z. Putting z = x in (2) gives > 2(x^2y + xyx + yx^2) = 0 > and in characteristic 3 we get x^2y + xyx + yx^2 = 0. > Thus in characteristic 3 we have > 0 = x(x^2y + xyx + yx^2) - (x^2y + xyx + yx^2)x = x^3y - yx^3 = xy - yx > and so R is commutative in characteristic 3. > (following hints from a problem in a french textbook) : by expanding out (x+y)^3 - (x-y)^3 you get xyx + x^2y + y^2x = 0 and multiplying by x on the left and on the right, you get xy = yx ( = - x^2yx - xyx^2). ============================================================================== From: Robin Chapman Subject: Re: Two other questions Date: Thu, 22 Jul 1999 07:58:34 GMT Newsgroups: sci.math In article <7n2bu0$2ks$1@nnrp1.deja.com>, jvoight34@my-deja.com wrote: > > > > And: suppose in a ring that x^3=x for all x. Show that the ring is > > > commutative. Here's a better (I think) argument than the one I posted yesterday. For all x and y we have 0 = (x + y) - x - y = (x + y)^3 - x^3 - y^3. Expanding this out gives x^2 y + xyx + yx^2 + y^2 x + yxy + xy^2 = 0. (1) Setting y = x in (1) gives 6x^3 = 0 so that 6x = 0 for all x. Setting y = x^2 in (1) gives 3x^4 + 3x^5 = 0. Using x^3 = x we get 3x^2 + 3x^3 = 0 = 3x + 3x^2. Hence 3x^2 = - 3x = 3x. Hence 0 = 3(x + y) - 3x - 3y = 3(x + y)^2 - 3x^2 - 3y^2 = 3xy + 3yx. Thus 3xy = -3yx = 3xy. Replace x by -x in (1) to get x^2 y + xyx + yx^2 - y^2 x - yxy - xy^2 = 0. (2) Adding (1) and (2) gives 2(x^2y + xyx + yx^2) = 0 and so 0 = 2x(x^2y + xyx + yx^2) -2(x^2 y + xyx + yx^2)x = 2x^3 y - 2yx^3 = 2xy - 2yx. Thus 2xy = 2yx. Subtracting this from 3xy = 3yx gives xy = yx. -- Robin Chapman http://www.maths.ex.ac.uk/~rjc/rjc.html "They did not have proper palms at home in Exeter." Peter Carey, _Oscar and Lucinda_ Sent via Deja.com http://www.deja.com/ Share what you know. Learn what you don't. ============================================================================== From: mareg@primrose.csv.warwick.ac.uk (Dr D F Holt) Subject: Re: Question Date: 22 Jul 1999 08:47:19 GMT Newsgroups: sci.math In article <210719990804094467%edgar@math.ohio-state.edu.nospam>, "G. A. Edgar" writes: > >> > Let $R$ be a ring in which $x^3=x$ for every $x\in R$. Prove that $R$ >> > is a commutative ring. >> >> This is (word-for-word) in Herstein, Topics in Algebra, Ch. 3 sec. 4, >> exercise 19. > >I once heard Herstein comment that this one problem generated >more mail than the entire remainder of the book. > I remember spending many hours on this as an undergraduate. Here is the most concise proof that we managed to come up with: An element x is called central if xy=yx for all y in R. Note that the central elements form a subring of R. 1. xy = 0 => yx = 0. (Proof: yx = (yx)^3 = y (xy)^2 x = 0.) 2. x^2 = x => x central. (Proof: x(y - xy) = xy - x^2y = xy-xy=0, so (by 1) (y - xy)x = 0, and yx = xyx. Similarly, (y - yx)x = 0 => x(y - yx) = 0 => xy = xyx.) 3. x^2 is central for all x in R. (by 2, because (x^2)^2 = x^4 = x^2). 4. If x^2 = nx for an integer n, then x is central. (Proof: x = x^3 = qx^2, which is central by 3.) 5. x + x^2 is central for all x in R. (Proof: By 4, because (x + x^2)^2 = 2(x + x^2).) 6. By 3 and 5, x = (x + x^2) - x^2 is central, completing the proof. Derek Holt.