"If you want the new trees to produce the same way as the parent tree, you'll have to graft them onto rootstock."
Not so. If you can get pear cuttings to root - and some root quite readily, while others do not - the resulting tree will be every bit the 'clone' of the original that a graft on seedling or clonally-propagated rootstock would be.
Granted, grafts on seedling pear (callery or communis) will *possibly* grow faster, early on, and *might* be longer lived, but if you think about it, the clonal rootstocks (OHxF for pears, EMLA, Geneva, etc., for apples) are, for all intents and purposes, a rooted cutting. No reason why a rooted cutting of a pear or apple, 'on its own roots' would be different - with regard to fruit characteristics - from that produced by a similar cutting grafted onto an appropriate rootstock.