This is actually a very major issue. They are copyrighted to varying degrees.
The fact that feature "z" is located at "x,y" can't be copyrighted, but the data file that describes exactly that can be and is. (There are good reasons for this. It's not as simple as the straw man I set up there.)
The US is very fortunate to require all government-produced cartographic data is solidly in the public domain. This is not the case in most countries.
Now there are several public-domain global datasets, but not very long ago obtaining basic cartographic data for many parts of the world was fraught with copyright issues.
It's rather annoying to have made a map of your study area and then find out that:, "Oh, wait. You can't publish that. Our license for the river locations only allows us to use the data internally. We can't publish a figure with that data on it!"
The US is very fortunate to require all government-produced cartographic data is solidly in the public domain. This is not the case in most countries.
You can make the argument that making these maps costs money, so those that benefit from the maps should actually pay the government for them - thus people who don't need these maps don't have to pay for them through their taxes.
I'm not entirely sure I subscribe to that - often having something publicly available like some kind of a utility is more beneficial to everybody than properly allotting the cost.
But I don't think it's clear cut in the case of maps, and in any case it's a reasonable argument to make.
So, the census needs maps (and best we are able to double check that, as it is used to apportion representatives). But also: the modern nation-state is, in many ways, founded on the notion of maps. A nation exists not just by itself but instead of and in place of other government. So it's pretty reasonable to assume public maps come from a time where e.g. the western border of the country was fuzzy and who controlled what was information of great importance. For instance, if it's 1823 and I'm looking to farmstead, a map both legitimizes, in some way, American rule, and delineates reality about Comanche raids (which were not something you'd want to bump into).
I generally like open government, but I think this particular practice may come from a history where maps were more interesting and relevant to the integrity of the nation-state.
This issue is rapidly coming to a close. The disruption of the mapping companies (and sovereign mapping efforts) is going power law now because of mobile devices.
Five years ago, www.OpenStreetMap.com started getting comparable to commercial maps in many places. And we have not even entered the golden age of OSM, when all of these sensor-laden devices we carry will be automatically feeding the map updates and corrections.
Open data will be the best for nearly all applications on a 20-year timeframe. Google, TomTom, Russia, etc will not be able to make money on the actual map data.
The fact that feature "z" is located at "x,y" can't be copyrighted, but the data file that describes exactly that can be and is. (There are good reasons for this. It's not as simple as the straw man I set up there.)
The US is very fortunate to require all government-produced cartographic data is solidly in the public domain. This is not the case in most countries.
Now there are several public-domain global datasets, but not very long ago obtaining basic cartographic data for many parts of the world was fraught with copyright issues.
It's rather annoying to have made a map of your study area and then find out that:, "Oh, wait. You can't publish that. Our license for the river locations only allows us to use the data internally. We can't publish a figure with that data on it!"