From: "Daniel Giaimo" Subject: Re: Set Theory Question Date: Mon, 3 Jan 2000 14:42:37 -0800 Newsgroups: sci.math Summary: regular cardinals and cardinalities of small subsets Fred Galvin wrote in message news:Pine.LNX.4.10.9912311851000.22743-100000@titania.math.ukans.edu... > On Fri, 31 Dec 1999, Daniel Giaimo wrote: > > > Let S be any infinite set. Let T be the set of all subsets P > > of S such that card(P) < card(S). Is it true, and if so how would > > you prove, that card(T) = card(S)? > > Not necessarily, as I pointed out in my earlier reply. On the other > hand, assuming GCH, it is true for most values of card(S); namely, it > is true when card(S) is a regular cardinal. More generally, if S is > any infinite set, and if T' is the set of all subsets P of S such that > card(P) < cf(card(S)), then card(T') = card(S). I'm not sure I see why card(S) >= card(T'). (Clearly, card(S) <= card(T') as T' contains the singleton subsets of S.) --Daniel Giaimo ============================================================================== From: "Brian M. Scott" Subject: Re: Set Theory Question Date: Mon, 03 Jan 2000 21:17:23 -0500 Newsgroups: sci.math Daniel Giaimo wrote: > Fred Galvin wrote [...] > > Not necessarily, as I pointed out in my earlier reply. On the other > > hand, assuming GCH, it is true for most values of card(S); namely, it > > is true when card(S) is a regular cardinal. More generally, if S is > > any infinite set, and if T' is the set of all subsets P of S such that > > card(P) < cf(card(S)), then card(T') = card(S). > I'm not sure I see why card(S) >= card(T'). (Clearly, card(S) <= > card(T') as T' contains the singleton subsets of S.) Well-order S in type card(S), S = {s(i) : i < card(S)}. Let k = cf(card(S)), and let (m(i) : i < k) be an increasing sequence of cardinals cofinal in card(S). If P is in T', then P is a subset of S(j) = {s(i) : i < m(j)} for some j < k. By GCH we have card(powerset(S(j))) = m(j)+ <= m(j+1) < card(S), so card(T') <= k card(S) = card(S). Brian M. Scott