Lets say @something is at memory location 0x12345

$ref= \@something; 
print $ref; 
# would print 0x12345, right?
$ref2= \@something;
print $ref2;
# would print 0x12345 as well
[download]

Obvious, right? You take a reference to the array, it will be always the same. Now the other case:

$ref= [ @something ]; # would be the same as 
$ref= \( @something ); # if that syntax where possible, or
my @x= ( @something ); # with correct syntax
$ref= \@x;
[download]

i.e. [] is just () with an additional reference-taking. And by doing ( @array ) you are creating a new array (at a new memory address) with the contents of @array filled in

( @array ) is not the same as @array. This is obvious when you look at @newarray= ( @array1, @array2 );. It can't be the same

$ref1 = \@arr;
$ref2 = \@arr;
[download]

The two references both point to the same underlying data structure.

$ref1 = [ @arr ];
$ref2 = [ @arr ];
[download]

The two references each point to different anonymous arrays, which have the same values as @arr, but are different structures.