5
$\begingroup$

Suppose that a 2-category $\mathcal{C}$ has strict pullbacks and one has maps $f:F\to C$, $g_0,g_1:G\to C$ and a natural transformation $\gamma:g_1\implies g_0$. Is there a good notion of a pullback transformation $f^*\gamma$. If so, I would expect its codomain to be $$f^*g_1\times_G f^*g_0\to f^*g_1\to F$$ and similarly for the domain.

If there is, is the need for the extra pullback here (over $G$) connected to the second layer of pullbacks in a 2-categorical descent diagram (relative to quotients in a 1-cat, which only require a kernel pair=1 layer of pullbacks)?

$\endgroup$

1 Answer 1

6
$\begingroup$

There is a good notion of a pullback transformation, but its domain and codomain aren't what you guessed; rather one asks for a map from $f^*(g_0) \to f^*(g_1)$ and a 2-cell filling the resulting triangle over G. The existence of such a pullback transformation is also not automatic from the existence of pullbacks, but requires $f$ to be a fibration (and the existence of all such pullbacks is equivalent to $f$ being a fibration). This characterization of fibrations is studied in Peter Johnstone's paper "Fibrations and partial products in a 2-category".

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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