Rufus pretty much summed up the issue I'm grappling with. So permaculture aims to meet our needs sustainably, but half the human population relies on bulk calories delivered to the supermarket. So what can be sustained using PC is a much smaller population and the transition is a nightmare for humans and wildlife.
But where I feel there is a gap between theory and action is in how much of the Earth's resources do we appropriate for ourselves in permaculture. Globally there's 0.2 hectares per person of arable land (that's 0.5 acre). If we appropriate more land than that, or occupy that land and are inefficient in converting incoming energy into food energy, then there is less to go around, either for nature or other humans. If we have larger properties and set aside anything in excess of 0.2ha/person for nature then that's a value choice, that we're valuing nature over people. If we appropriate more for our own use, similarly, that's valuing ourselves over nature and other people.
Bill Mollison characterises permaculture as using "the inherent qualities of plants and animals combined with the natural characteristics of landscapes and structures to produce a life-supporting system for city and country, using the smallest practical area"
This last part often seems lacking in practice. This attitude probably originates here in Australia where people tend to have large tracts of land and you often can't subdivide below, say, 40ha. And the last state to allow land-sharing has just stopped allowing it.
It does seem inherent in the intersection of the three ethics that we'll be concerned about the impacts of our choices on others and on nature, but this doesn't necessarily translate into concern about the off-site implications of our land-use efficiency in terms of energy productivity (and nature productivity - if there is such a thing!).