The fact that copyright used to last for N years in 1909 is irrelevant. Whether or not a law used to exist should not determine whether or not a law should exist, for obvious reasons (slavery, universal suffrage, mandatory military service, etc.).
It would be cool to explain explicitly why things should go in the public domain after X years instead of N years.
There needs to be a judgment call about the explicit rule to have. That said, I think we have made the wrong tradeoff.
The principle embodied in the US Constitution is that the purpose of intellectual property is to maximize the development of the common public domain that we can all benefit from. Therefore copyright should have a long enough horizon to provide a sufficient incentive to encourage innovation, but should not be longer than that.
What fraction of the economic value of the average copyrightable work is captured in the first 30 years of its term? Most of it. In the first 70 years of its work? I saw an estimate of over 99% of it in one of the amicus briefs in the Eldred case. Therefore the real difference between current copyright law and perpetual ownership is less than 1% of possible value. That extra 1% is not material to people's decisions about whether to create new works, and therefore increasing the term of copyright further should lose compared to the economic value of having more material enter the public domain. (After the case was lost I saw Lessig mention that one of his mistakes was not pushing this line of economic argument more strongly.)
Anyways, given how the court ruled, this is a moral argument regarding future copyright laws and not a legal argument based on which one could overturn further bad laws.
It would be cool to explain explicitly why things should go in the public domain after X years instead of N years.