I'm not entirely sure about this as I am a wx user not a gtk user, however, it looks as though gtk_widget_destroy is an event/signal. Of course the first argument is self, while the second (as micmast said) is usually an instance of the widget that called it.
This is useful for various things, for example if the user entered some data in a dialog, with this method, as it is destroyed you could look up the value of the data entered using the second argument. Make sense? Of course you don't always need to access the calling widget, so the second argument might not come into play.
To summarize, the second (widget instance) argument is there only if you need it, otherwise just ignore it! :)
Hope this helps.
P.S. Here is the doc for
gtk.Dialog (Google is your friend.)
