I can't really ask my retired pilot about this, as he retired before EGPWS was a thing (which uses topography information to give much better predictions of, uhm, potential future ground interaction, than earlier and much more basic GPWS), but I did not think the kind of topography information that EGPWS has is used for navigation. Neat if it does? Do you have a ref for that?

Absent of GNSS, aircraft can use VORs - radial radio stations the aircraft can pick up and follow. These are (I gather) very commonly used for precision approaches still, least in Europe, and can be used to navigate over and not-too-far-from land. The USA seems to be decommissioning a lot of VOR stations though - not sure what they intend the backup to be for GNSS failures. For trans-oceanic navigation commercial aircraft historically used INS, and the older electro-mechanical systems could achieve within 10 km accuracy over the atlantic. I assume modern INS systems (someone mentioned laser gyroscopes) are a lot more accurate.

There is also celestial navigation. This was used before GPS. Commercial aircraft had observation domes in olden times for a navigator to take measurements and do manual calculations. The USAF had automated celestial navigation systems even back in the electro-mechanical / electronic-valve days - pretty impressive (e.g., the SR-71 astro-inertial system). With modern tech, this should be much easier and cheaper to fit to aircraft, but I don't think there are any in use in commercial aviation - they rely on GNSS, and INS and VOR as fallbacks. I guess modern INS is more than good enough to not make automated celestial nav systems worth it?

Obviously, there is also the olde magnetic compass. Which hopefully can get you near enough a VOR to pick it up, if you're somewhere where those are sparse. In visual conditions, there is also the mark 1 eyeball, and following landmarks (rivers, lakes, mountains, coasts and such if at altitude; roads, railways, etc.. if lower down).