Around + │ Si │ −, how many fields? Not two, not three. There is one electric field, with two poles (+ and −), structured across three regions (p │ depletion │ n) — and the one field lives entirely in the inert gap between the planes. The two doped sides are nearly neutral; the field is in the middle. Same shape as gravity: one field, sources, structure.
The + and − are not two separate fields — they're the two poles of one field, the way a bar magnet isn't "a north field and a south field" but one field with two ends. And the three regions aren't three fields — they're the spatial structure the single field is shaped across. So the count is: one field, two poles, three zones.
The crucial fact: the field is not spread evenly through the device. The p-side and n-side bulk are nearly neutral and field-free. The built-in electric field exists only in the depletion region — the inert middle. The field is the gap. Remove the gap and there is no field, no junction — just a wire. This is why the inert middle is the device: it's literally where the one field lives.
And here is the parallel that ties it to relativity: gravity is also one field. There aren't "two gravities" — there's one gravitational field (the spacetime metric), with sources (masses) shaping it. So neither the junction nor gravity is "2 or 3 separate fields." Each is one field, structured by its sources — the diode's one EM field concentrated in its gap, gravity's one field curved by its masses. One field, structured, concentrated where the geometry bends.
(And no — there is no meaningful gravity in the junction itself. Gravity exists wherever mass does, but it's ~10³⁶ times too weak to matter here; the junction is purely electromagnetic. The relativity link is the structure — one field, geometry-as-law — not literal gravity in your diode.)