When factorial(4) calls factorial(3), the first call isn't finished: it's waiting for the answer so it can multiply by 4. Python keeps a stack of these waiting calls, called the call stack. Each call gets its own fresh set of local variables, so the n in each call is separate:

factorial(4)
  4 * factorial(3)
        3 * factorial(2)
              2 * factorial(1)
                    returns 1
              returns 2 * 1 = 2
        returns 3 * 2 = 6
  returns 4 * 6 = 24

The calls pile up until the base case answers, then the answers flow back down, each waiting call finishing in turn. You can watch it happen:

def factorial(n, depth=0):
    indent = "  " * depth
    print(f"{indent}factorial({n})")
    if n <= 1:
        result = 1
    else:
        result = n * factorial(n - 1, depth + 1)
    print(f"{indent}-> {result}")
    return result

factorial(4)