I think it's just the terminology they use.
Agree with that.
In my experience I have always used every pin on a PIC as per its particular functionality described in the corresponding Datasheet.
In adressing each particular pin and depending on Programming langauage/compiler you end up using syntax as portaA.1, portE.2 etc but it is only a convention.
Perhaps the only advantage on using the "port" concept is when you simultaneously toggle pin states on a particular port using syntax as portE = %0100 ( again depending on programming platform used ) , acomplishing the modification of the state of 4 pins ( in this example ) in only 1 line of code.
The only "port" concept on a PIC coming from hardware design side I have encountered, is the SPP ( Streaming paralell port ) you find in for example 18F45550 ( uses the 8 bits of "portD" ) but then again , settings depend on those of other "pins" of other " ports"...
I would say you stay with the "unitary" approach and treat every pin as a its own, Datasheet on the other hand (always) provided.