Tree reality doesn't necessary conform to textbook theory. I agree that "mostly" we need to limit pruning cuts in the early years to get to fruiting sooner. This is why Mike asked the question "what rootstock?" Heading the leader on a trellised dwarf tree is simply not done. But with freestanding trees--growing a structure intended for decades of fruit production--all depends on the circumstances.
You head at a freestanding whip at planting around 40 inches high. Ethan seems good on this point and I'll assume others are as well. (Look up the "crow's foot response" in
Holistic Orchard, if not.) Nursery stock that comes with feathered beginnings at a reasonable height probably are not headed that first season.
Things proceed apace. Only some trees get downright ornery. There are no lower branches appearing for 4, even 5, even 6 feet up the trunk. I would head such a tree in year two or three unless seriously planning to graze water buffalo in that orchard. Kidding aside, this would be appropriate for a classic full-sized grandmother tree. Only I tend to like fruiting on lower branches for twenty years or so ... after which the eternal form can be found by removing those lower branches.
Another scenario sees a good start to first tier branches but then the leader shoots for the sky on its ownsome. The decision to "head high" to develop second tier branches should take into account extreme lankiness. Don't want to make this cut unless two goals can be attained: popping more laterals for an envisioned midway scaffold AND wanting more girth to better support a lanky leader. Take note: some growers choose not to have a permanent midway structure as a means of keeping tree height down and less interior shading.
One good way to think of overall tree structure is that scaffolds are permanent extensions of the trunk and the "leader zone" is managed more as a dwarf tree atop of all that. So go our journeys in the time-space continuum called pruning.
Lost Nation OrchardZone 4b in New Hampshire