In the first chunk of code you are overloading the method for the different arguments. In the second chunk of code there is not way to determine which method is to be called and so you need to over load the return method arguments.

For instance you can try:
Code:

 int getVal(int i){
 return intVal;
 }
 
 float getVal(float f){
 return floatVal; 
 }  

which you would then call by placing a random int or float val in the method arguments:
Code:

intValue = getVal(0);
floatValue = getVal(0f);
 

The only issue about doing it this way is that it then becomes a pain to keep track of which value you are calling at any given time. In java there is a degree of auto casting and so you may end up with problems when you compile. It's far more common to have different methods to call the different values. For instance:
Code:

 int getIntVal(){
return intValue;
}
float getFloatVal(){
return floatValue;
}