Is the C language more complicated than building a database?! That's looking to be way more exaggerated.
For open-source projects, being written in a scary[1] language is a liability in the long run.
For many tasks, we used to have no good alternative than using C (or C++ which is even scarier), but this is no longer the case now thanks to Rust.
I'm obviously not saying we should rewrite PostgreSQL in Rust, but the ability to grow/maintain the contributor base is an argument that should not be underestimated.
[1]: no matter how difficult or not it is in practice, the C is scary for many people and most of them will never feel confident enough to suggest changes in a C code base.
Scary is having to run from a lion to save our life.
I think people do have this feeling, but likely it's because they don't actually know much about writing databases, so they see one issue but not the other.
OTOH it might be possible to contribute small things (e.g. improving the database client program, adding a test) if one can write some code in $LANG_MORE_ACCESSIBLE_THAN_C, even without knowing what a btree is or what "cache locality" means.
By my reading for most of the submitted patches language issues aren't in the top three of problems preventing the patches from being merged; although such issues certainly exist. That said, that obviously doesn't say anything about patches never getting submitted.
Personally the biggest issue with growing the contributor base in postgres isn't language related, it's that a) the easy thing have been done long ago, making the learning curve steep b) lack of time from the experienced people/committers. Most of us have other stuff and our own projects to deal with :(