If you consider the dog's work as a series of smaller behaviors where one behavior leads to another, you have a behavior chain. Completing one behavior leads to the next behavior and so on until the task is completed. When the dog does something wrong it has departed from the correct behavior chain and has started a different set of behaviors. In theory, any correction for a preceding behavior will be indirect pressure for a behavior following it in the chain.
I am surprised that it works but it does.
Luckily a trainer doesn't have to understand the theory but just has to know correct application.