Why can you add an attribute to an instance of a subclass of object, but not to an instance of object?

Question

Why does this raise an error:

o = object()
o.i = 1

But this does not:

class A(object):
    pass

a = A()
a.i = 1

?

Answer 1

Because built-in types don't have dictionaries associated with them to hold added attributes:

>>> o = object()
>>> dir(o)
['__class__', '__delattr__', '__doc__', '__format__', '__getattribute__', '__hash__', '__init__', '__new__', '__reduce__', '__reduce_ex__', '__repr__', '__setattr__', '__sizeof__', '__str__', '__subclasshook__']

See? No __dict__.

But adding a subclass gives the attribute somewhere to go:

>>> class A(object):
....    pass
....
>>> a = A()
>>> dir(a)
['__class__', '__delattr__', '__dict__', '__doc__', '__format__', '__getattribute__', '__hash__', '__init__', '__module__', '__new__', '__reduce__', '__reduce_ex__', '__repr__', '__setattr__', '__sizeof__', '__str__', '__subclasshook__', '__weakref__']

Saying that it's "because they're defined in C" isn't a "why". You could certainly define a type in C with an instance dictionary.

Answer 2

object is defined in C. You cannot add arbitrary attributes to instances of types defined in C unless you fill the appropriate slots in the type definition.