Search All of the Math Forum:
Views expressed in these public forums are not endorsed by
NCTM or The Math Forum.



Re: Reducing Incomparability in Cardinal comparisons
Posted:
Apr 7, 2013 12:42 AM


On Mar 13, 3:09 pm, Zuhair <zaljo...@gmail.com> wrote: > On Mar 12, 10:46 pm, Zuhair <zaljo...@gmail.com> wrote: > > > > > > > On Mar 12, 2:49 pm, Zuhair <zaljo...@gmail.com> wrote: > > > > On Mar 11, 11:17 pm, Zuhair <zaljo...@gmail.com> wrote: > > > > > Let xinj>y stands for there exist an injection from x to y and there > > > > do not exist a bijection between them; while x<bij> means there > > > > exist a bijection between x and y. > > > > > Define: x=y iff x<bij>y > > > > > Define: x < y iff xinj>y Or Rank(x) inj> Rank(y) > > > > > Define: x > y iff y < x > > > > > Define: x incomparable to y iff ~x=y & ~x<y & ~x>y > > > > > where x is defined after Scott's. > > > > > Now those are definitions of what I call "complex size comparisons", > > > > they are MORE discriminatory than the ordinary notions of cardinal > > > > comparisons. Actually it is provable in ZF that for each set x there > > > > exist a *set* of all cardinals that are INCOMPARABLE to x. This of > > > > course reduces incomparability between cardinals from being of a > > > > proper class size in some models of ZF to only set sized classes in > > > > ALL models of ZF. > > > > > However the relation is not that natural at all. > > > > > Zuhair > > > > One can also use this relation to define cardinals in ZF. > > > > x={y for all z in TC({y}). z <* x} > > > > Of course <* can be defined as: > > > > x <* y iff [x inj>y Or > > > Exist x*. x*<bij>x & for all y*. y*<bij>y > rank(x*) in > > > rank(y*)]. > > > > Zuhair > > > All the above I'm sure of, but the following I'm not really sure of: > > > Perhaps we can vanquish incomparability altogether > > > If we prove that for all x there exist H(x) defined as the set of all > > sets hereditarily not strictly supernumerous to x. Where strict > > subnumerousity is the converse of relation <* defined above. > > > Then perhpas we can define a new Equinumerousity relation as: > > > x Equinumerous to y iff H(x) bijective to H(y) > > > Also a new subnumerousity relation may be defined as: > > > x Subnumerous* to y iff H(x) injective to H(y) > > Better would be > > x Subnumerous* to y iff H(x) <* H(y) > > however still this won't concur incomparability completely > > However if we define recursively H_n(x) then we can define > the above relations after those. However still incomparability > would persist, although the above is still a strong approach > against incomparability. > > > > > > > This might resolve all incomparability issues (I very highly doubt > > it). > > > Then the Cardinality of a set would be defined as the set of all sets > > Equinumerous to it of the least possible rank. > > > A Scott like definition, yet not Scott's. > > > Zuhair Hide quoted text  > >  Show quoted text  Hide quoted text  > >  Show quoted text 
Once again, just like I (and others now) said  you throw out stuff that isn't ready, you make mistakes, you change it, you ponder how to fix it, you debate possible solutions. It is a total moving target. It is presented as being a solution but it is just a big problem  trying to keep it straight as you move things around in a desperate attempt to clean up the mess.
CB



