The more general idea is transformation, which includes a frame of reference and typically rotation. The frame of reference with respect to position could be another object ie. 10 feet above earth's surface at lat X, long Y, or it could be implied to be the world origin (0,0,0). For rotation, we usually use a canonical rotation that is the result of whatever format, say, a model is in, and then the rotation is implied to be offset from that.
Unity actually takes frame of reference from parented objects, such that each item has a local position that is offset from its parent objects. For example, a driver's position is local to their car, and thus the world position of the driver is relative to the car's rotation and position.
This is an important concept because position at any given time can depend on attached objects' current rotation and position. An especially complex example:
Local Cluster -> Galactic Center -> Sun -> Earth -> Car -> Driver -> Driver's Hat -> Spider on Hat
With this scale of simulation, we need to know the driver's hat's position relative to the driver, the position of the driver relative to the car, the rotation and position of the car relative to earth, the location and rotation of the earth, and so on.
Of course, unless you're working on something like Kerbal Space Program, this amount of parenting is unnecessary. Most often we use this to attach clothes and props to bodies and bodies to vehicles.