MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

The question is in the title.

I am trying to apply the Mitchell (Freyd-Mitchell?) embedding theorem, which states that for every small abelian category $A$, there exists a ring $R$ such that A embeds into the category $R$-mod. The derived category is not abelian, of course, but I have a particular subcategory that is abelian, and life would be easiest if the derived category was smal, so that the subcategory was small and abelian.

share|cite|improve this question
Of course it is, you can easily find a (transfinite) upper bound for the 'number' of objects – Fernando Muro Feb 13 '12 at 20:18
Presumably small should be understood up to equivalence. I agree with Fernando, although writing down the details would entail some work: Choose a finite affine cover $\{Spec A_i\}$. Then a bounded complex of coherent sheaves is given by a collection of finitely presented $A_i$ modules $M_i^\bullet$ plus patching data. A morphism is ... – Donu Arapura Feb 14 '12 at 0:03
The term you actually want is "essentially small" (equivalent to a small category). – Qiaochu Yuan Feb 14 '12 at 2:52
Another term for essential small is "svelte": – B. Bischof Feb 14 '12 at 3:53
out of curiosity: how does the Freyd-Mitchell embedding make life much easier? – Yosemite Sam Feb 14 '12 at 11:59
up vote 4 down vote accepted

The answer to your question really depends on what you mean by the word "the". An unhelpful answer is that the coherent sheaves over any variety form a proper class (hence "no"). A more useful answer is (as mentioned in the comments) that there exists a small category that is equivalent to any category that can be reasonably called the bounded derived category of coherent sheaves (hence "yes").

Furthermore, the construction of such a category can be accomplished without the use of replacement. In particular, the category lives in the same ZC universe (i.e., $V_\alpha$ for $\alpha$ a not-necessarily-inaccessible limit ordinal greater than $\omega$ - see e.g., Wikipedia or the set theory section of the Stacks project) as the defining field.

share|cite|improve this answer
Thank you. Do you know of a reference for the proof? – David Steinberg Feb 14 '12 at 3:22
No, but it is straightforward. If you set $\kappa$ to be the smallest infinite cardinal containing your base field, then any variety is covered by finitely many spectra of finitely generated rings, also of cardinality $\kappa$. Each complex of coherent sheaves is then encoded by a finite collection of finitely generated modules and arrows describing gluing data and differentials. Up to isom. all modules can be realized by subquotients of a fixed infinite direct sum of copies of a ring (size $\kappa$), and the arrows are maps of such sets ($\leq 2^{2^\kappa}$ total choices). Done. – S. Carnahan Feb 14 '12 at 6:56
Oops, not quite done. For any pair of complexes of size bounded by $\kappa$, the set of ordinary homs is bounded by $2^\kappa$. Localization yields equiv. classes of finite zig-zags of maps, so the hom group between two objects in the derived category is also bounded in size by $2^\kappa$. Total maps in category is then bounded by $2^{2^\kappa}$. – S. Carnahan Feb 14 '12 at 8:14

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.