I am not trying to be argumentative. I've read most your posts and agree with your approach and what you say. Its just the whole breed=type that's throwing me. I'm new to chickens, other than the few chickens I had as a kid, but I've got a degree in biology so I'm not new to the concepts of genetics. Maybe that's what's throwing me for a loop. I'm applying a formal concept and equating breed with sub species. It just blows my mind that breed is essentially a collection of phenotypes regardless of genetic makeup and could possibly be arrived at through various cross breeds.
Breeds in all livestock are quite a bit different from subspecies. It's not a bad mental map but biologically they're not equivalent.
If you're familiar with horses, it's not unlike the european warmblood approach. Animals there are closely pedigreed but they are not managed to be "purebred" as many mammal livestock breeds are. They focus on phenotype as expressed in the shape of the animal and in its ability to perform certain tasks. The "breed" assigned is more about where the animal is born than about who its parents are.
You see something similar in the club lamb circuit also - mostly unregistered animals with some combination of Hampshire and Suffolk, bred and selected for a particular shape with no worries about how the animal got that way genetically. ("Club lambs" are typically shown by 4H/FFA and are raised for meat. But perversely, a good show lamb can cost thousands of dollars.)
In poultry, in part the culture is different because of the difficulty in tracking individual animals and the relatively low unit value of each one.
But, the notion of "purity" in breeds is perhaps overblown and harmful itself. The notion of the pedigree being more important than the type and functionality of the animal creates its own kind of harm. I think as we have more information about the genetic code, that perhaps it will create another shift in breeding, in all species of animals.
Just as an example: it is apparently now economic for meat sheep breeders to flush embryos from newborn lambs, run genetic tests, create embryos, and then implant embryos from the best baby ewes in recipient ewes. That is simply stunning to me. I suspect it will not be that long before genetic testing is an option in poultry as well, especially for something relatively simple like color pattern.
Kind of back to the topic: it's very much true that for example bantams and LF of the same breed are generally produced independently from totally different stock, and it can be true within varieties. But, once established, it's not common to do much outcrossing. If anything, poultry breeders tend to do very close inbreeding/linebreeding much more so than in mammals. When you want offspring that are essentially identical and you don't mind eating your mistakes, this approach can work pretty well.
An appeal of crosses to many people new to poultry is that they create more distinct individuals which can make it easier to have a relationship with particular birds. With production poultry, that's not an asset, especially not for animals that are intended to be eaten. (Both because of the emotional considerations and because it makes processing more challenging.) The standard breeds were mostly produced with that practical utility in mind.