Here is an example of using a zero reference point that will hopefully help clear up any confusion:
An architect is drawing up a set of plans for a multistory building with an underground parking deck below the building. He needs to draw some elevation views to show the distances between floors, ceiling heights, window sill heights, etc. And he needs a reference point (zero point) that he can calculate the various elevation distances from. Lets say he selects the main (ground) floor as his reference point. He can designate the finished ground floor as zero and show the elevations of the other items as a distance from his zero reference point. In this case the underground items would have a negative number for their elevations.
However, he could have selected, for example, the floor of the lowest level of the underground parking deck to be his zero reference point, designated the distances to the other items accordingly, and still ended up with, for practical purposes, the same resulting building. Provided of course that the builder followed the plans properly.
So, in a sense, the selection of what to use for the zero reference point is arbitrary. And I believe that that is what Dale Meyers is trying to say in that sentence.