# Using a Parent's Method Explicitly

## Calling the Parent

Our penguin class is getting cramped! To clean things up, lets make a child for it called emperor-penguin. It can do everything that a penguin does, except that when it eats, an emperor-penguin says '(bon apetit) before eating food. Does the following definition work?

(define-class (emperor-penguin name)
(parent (penguin name))
(method (eat)
(print '(bon apetit!))


Let's say we define napoleon as follow:
(define napoleon (instantiate emperor-penguin 'napoleon))
What happens when we call (ask napoleon 'eat)?

## Usual

The correct way to call a parent's method is to use the usual keyword.

(define-class (emperor-penguin name)
(parent (penguin name))
(method (eat)
(print '(bon apetit!))
(usual 'eat)))

usual takes one or more argument, the first being the message, and the others being any arguments the message needs. This message and necessary arguments are then passed to the parent. In this way, an emperor-penguin object will refer to penguin's eat method.

Calling usual is just like saying (ask self ...) with the same arguments, except that only methods defi ned within an ancestor class (parent, grandparent, etc.) are eligible to be used. It is an error to invoke usual from a class that doesn't have a parent class.

## Naming Intuition

You may be thinking that usual is a funny name for this function. Here's the idea behind the name: We are thinking of subclasses as specializations. That is, the parent class represents some broad category of things, and the child is a specialized version. (Think of the relationship of checking-accounts to accounts in general.) The child object does almost everything the same way its parent does. The child has some special way to handle a few messages, different from the usual way (as the parent does it). But the child can explicitly decide to do something in the usual (parent-like) way, rather than in its own specialized way.