How is that counter-intuitive? Or at least what’s your general intuition then? If it was [x] with x being a complex object, would you expect it to be deep copie

How is that counter-intuitive? Or at least what’s your general intuition then? If it was [x] with x being a complex object, would you expect it to be deep copied? Shallow copied? Copied via IRepeatable? What’s your intuition on a process of repeating a value that isn’t reference-copy? What should a[3] = a[2] = a[1] = a[0] or an equivalent chunk do, assuming trivial __setitem__()? Why that should be different from a * 4? I’m not even a python guy, if you think I’m biased. These questions are natural for a programmer and a programmer has to ask themselves about what they expect, why, is it consistent, is it safe to assume. Making [[]]*4 to create three new lists implies that you add a non-obvious ephemeral re-evaluation point in your grammar. Should [[f()]]*4 call f() three more times? These may be added for convenience. But you have to make sure that syntactic context makes it dead clear. E.g. in a for-i iterator part, loop conditions, function argument init, etc. Locations which are known to be executed more than once. (Python specifically doesn’t even do that for arguments. Every def f(x=[]) shares the same list as a default for x.)