I trying to understand the proof of Proposition 2.10.5 from the book: TENSOR CATEGORIES, by P. Etingof, S. Gelaki, D. Nikshych, and V. Ostrik, See http://www-math.mit.edu/~etingof/egnobookfinal.pdf.

Proposition 2.10.5. If $X\in \mathcal{C}$ has a left (respectively, right) dual object, then it is unique up to a unique isomorphism.

Something I did not understand is why the three small squares of the following diagram commute? enter image description here

where $e_1, c_1, e_2, c_2$ the corresponding evaluation and coevaluation morphisms


Let's take the top left square. We can rewrite it as $$\require{AMScd} \begin{CD} X_1^*\otimes 1\otimes 1 @>{id\otimes id\otimes c_1}>> X_1^*\otimes 1\otimes (X\otimes X_1^*)\\ @V{id\otimes c_2\otimes id}VV @V{id\otimes c_2\otimes id}VV \\ X_1^*\otimes (X\otimes X_2^*)\otimes 1 @>{id\otimes id\otimes c_1}>> X_1^*\otimes(X\otimes X_2^*)\otimes(X\otimes X_1^*) \end{CD}$$ where $1$ is the unit object. Commutativity of this square is now immediate from the fact that $\otimes$ is a bifunctor.

The other two squares are similar: once you explicitly write out the unit objects rather than implicitly eliminating them as is done in the original diagram, you are just using the functoriality of $\otimes$.

  • $\begingroup$ Thank you Eric! This is very nice explanation. I think there is a typo: the right vertical arrow $id \otimes c_2 \otimes c_1$ should be $id \otimes c_2 \otimes id$. $\endgroup$ – Dan Aug 17 '15 at 6:54
  • $\begingroup$ Oops, yes, corrected now. $\endgroup$ – Eric Wofsey Aug 17 '15 at 7:41
  • $\begingroup$ Could you tell me why the triangle in the upper right corner commutes by axiom (2.43) applied to X_2 ^* (that what the authors say)? $\endgroup$ – Dan Aug 17 '15 at 8:45
  • $\begingroup$ The upper right triangle is just (2.43) tensored with $id:X_1^*\to X_1^*$ on both sides. If it helps, the diagonal map can be labelled as $id\otimes id\otimes id$ rather than just $id$. $\endgroup$ – Eric Wofsey Aug 17 '15 at 8:54
  • $\begingroup$ Perfect! . Thank you Eric! $\endgroup$ – Dan Aug 17 '15 at 8:58

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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