Skip to main content
added 176 characters in body
Source Link
Jeff Strom
  • 12.5k
  • 3
  • 48
  • 76

If $X$ has torsion in its cohomology, there’s nothing you can say. If $X$ is simply-connected with torsion free (co)homology then it has a homology decomposition, showing that its cone length, hence category, is at most the number of dimensions with nonzero (co)homology. Thus we can say $$ \mathrm{cat}(X) \leq \dim ( H^*(X;\mathbb{Q}). $$ More can be said with information about the distribution of the nonzero groups (if there's clumping).

If $X$ has torsion in its cohomology, there’s nothing you can say. If $X$ is simply-connected with torsion free (co)homology then it has a homology decomposition, showing that its cone length, hence category, is at most the number of dimensions with nonzero (co)homology.

If $X$ has torsion in its cohomology, there’s nothing you can say. If $X$ is simply-connected with torsion free (co)homology then it has a homology decomposition, showing that its cone length, hence category, is at most the number of dimensions with nonzero (co)homology. Thus we can say $$ \mathrm{cat}(X) \leq \dim ( H^*(X;\mathbb{Q}). $$ More can be said with information about the distribution of the nonzero groups (if there's clumping).

deleted 1 character in body
Source Link
Jeff Strom
  • 12.5k
  • 3
  • 48
  • 76

If $X$ has torsion in its cohomology, there’s nothing you can say. If $X$ is simply-connected with torsion free (co)homology then it has a homology decomposition, showing that it’sits cone length, hence category, is at most the number of dimensions with nonzero (co)homology.

If $X$ has torsion in its cohomology, there’s nothing you can say. If $X$ is simply-connected with torsion free (co)homology then it has a homology decomposition, showing that it’s cone length, hence category, is at most the number of dimensions with nonzero (co)homology.

If $X$ has torsion in its cohomology, there’s nothing you can say. If $X$ is simply-connected with torsion free (co)homology then it has a homology decomposition, showing that its cone length, hence category, is at most the number of dimensions with nonzero (co)homology.

Source Link
Jeff Strom
  • 12.5k
  • 3
  • 48
  • 76

If $X$ has torsion in its cohomology, there’s nothing you can say. If $X$ is simply-connected with torsion free (co)homology then it has a homology decomposition, showing that it’s cone length, hence category, is at most the number of dimensions with nonzero (co)homology.