I think it's because of all of those transition mechanisms and fallback code added over the years. IPv6 fails the same way IPv4 does, but because of the terrible bullshit ISPs do to IPv6 connections, you end up with tons of software triggering obscure timeouts and fallback mechanisms that lead to a system of almost working networking code.

If the absence of IPv6 would've been treated the same way absence of IPv4 is, troubleshooting would've become a lot clearer. In fact, it probably would've been easier because ISPs can't just ignore and disable ICMP on IPv6 so you can actually get a hunch where in the network the problem is rather than seeing traffic vanish into the void.