[erlang-questions] List Comprehensions permutation

[contact@REDACTED]

Hi all,

I recently bought the book "Programming Erlang" by Joe Armstrong and I
have a little problem of comprehension about the logical of an
algorithm.

Here is the sample from the book : 

perms([]) -> [[]];
perms(L) ->
	[[H|T] || H <- L, T <- perms(L--[H])].

I would like to know why we return [[]] and not [] ?

I have try some tests : 

6> [[H|T] || H <- [1], T <- []].           
[]
7> [[H|T] || H <- [1], T <- [[]]].
[[1]]

For me it's not logical, in my brain its should be something like : 

8> [1|[]].
[1]
9> [1|[[]]].
[1,[]]

(all the sample are paste from erlang shell).

Someone can "light" me about this ?

Thanks in advance.