Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

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

Given a category $\mathcal{C}$, and a right calculus of fractions $\Sigma$. We can construct the category of fractions $\mathcal{C}[\Sigma^{-1}]$ which has the same objects as $\mathcal{C}$, and homsets given by $$ \mathcal{C}[\Sigma^{-1}](x,y) = \{y\xleftarrow{\gamma}u\xrightarrow{\sigma}x|\sigma\in\Sigma\}/\mathord{\sim} $$ where $y\xleftarrow{\gamma}u\xrightarrow{\sigma}x\sim y\xleftarrow{\gamma'}u\xrightarrow{\sigma'}x$ iff there is an object $v$ and morphisms $\tau\colon v\to u$, $\tau'\colon v\to u'$ in $\Sigma$, such that $\gamma\circ\tau = \gamma'\circ\tau'$ and $\sigma\circ\tau=\sigma'\circ\tau$. It's obvious that this relation is symmetric and reflexive, but I don't manage to show transitivity. I.e. given $$ y\xleftarrow{\gamma_1}u_1\xrightarrow{\sigma_1}x\sim y\xleftarrow{\gamma_2}u_2\xrightarrow{\sigma_2}x\sim y\xleftarrow{\gamma_3}u_3\xrightarrow{\sigma_3}x $$ I get a diagram $$ \matrix{ &&y&=&y&=&y\\ &\nearrow&&&\uparrow&&&\nwarrow\\ u_1&\xleftarrow{\tau_1}&v&\xrightarrow{\tau_2}&u_2&\xleftarrow{\tau_2'}&v'\xrightarrow{\tau_3}&u_3\\ &\searrow&&&\downarrow&&&\swarrow\\ &&x&=&x&=&x\\} $$ where the upwards arrows are the respective $\gamma_i$ and the downward arrows are the respective $\sigma_i$. But I'm stuck from there on.

Moreover, when defining a right calculus on fractions $\Sigma$, we impose a right cancellability condition, i.e. given an arrow $\sigma\colon y\to z$ in $\Sigma$ and a pair of parallel morphisms $f,g\colon x→y$ such that $\sigma\circ f=\sigma\circ g$, there is an arrow $\sigma′\colon w→x$ in $\Sigma$ such that $f\circ\sigma′=g\circ\sigma′$. Provided we don't need this to proof transitivity, I don't see why it is necessary, since it doesn't seem to be required for any other part of the proof that $\mathcal{C}[\Sigma^{-1}]$ is a well defined category.

share|cite|improve this question
Have you already had a look at the textbook by Gelfand & Manin on homological algebra? There is a proof of transitivity for the eq. relation of morphisms in derived categories (it applies to your case) – Avitus Sep 25 '13 at 11:02
Oh, thanks for the hint. I've actually have a copy of that in my bureau, gonna check that out tomorrow. – roman Sep 25 '13 at 13:46
you are welcome... pag. 149-150 :-) – Avitus Sep 25 '13 at 15:12
up vote 0 down vote accepted

Ok, following the hint by user Avitus, I found the answer in Gelfand & Manin Methods of Homological Algebra:

First, given the morphisms $\sigma_1\circ\tau_1$ and $\sigma_2\circ\tau_2'$, we can use the right extension condition to find an object $w$, together with morhisms $k\in\Sigma(w,v)$, $k'\in\mathcal{C}(w,v')$, such that $$\sigma_1\circ\tau_1\circ k = \sigma_2\circ\tau_2'\circ k'\text{.}$$ Moreover, $$ \sigma_1\circ\tau_1\circ k = \sigma_2\circ\tau_2\circ k\text{.}$$ Hence $\sigma_2$ equalizes $\tau_2'\circ k'$ and $\tau_2\circ k$. Thus, using the right cancellability condition, we can find an object $w'$, and a morphism $\kappa\in\Sigma(w',w)$, such that $\tau_2'\circ k'\circ\kappa= \tau_2\circ k\circ\kappa$. We observe that $$ \sigma_1\circ\tau_1\circ k\circ\kappa\\ =\sigma_2\circ\tau_2\circ k\circ\kappa\\ =\sigma_2\circ\tau_2'\circ k'\circ\kappa\\ =\sigma_3\circ\tau_3\circ k'\circ\kappa $$ and doing the same calculations with $\gamma_i$ instead of $\sigma_i$ we get $$ \gamma_1\circ\tau_1\circ k\circ\kappa=\gamma_3\circ\tau_3\circ k'\circ\kappa\text{.} $$ Hence the diagramm $$ \matrix{ &&y\\ &\nearrow&&\nwarrow\\ u_1&\xleftarrow{\tau_1\circ k\circ\kappa}&w'&\xrightarrow{\tau_3\circ k'\circ\kappa}&u_3\\ &\searrow&&\swarrow\\ &&x\\} $$ commutes, which yields $y\xleftarrow{\gamma_1}u_1\xrightarrow{\sigma_1}x\sim y\xleftarrow{\gamma_3}u_3\xrightarrow{\sigma_3}x$ and completes the proof.

share|cite|improve this answer

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.