Home

                                What is Equality?

Veterans of our SDG (Science Discussion Group – which met regularly in Aberdeen) will recall having constructed the first of the two transformations of a point along a line shown below.  All such transformations are mediated by conics; the first transformation depicted below is mediated by a pair of straight lines—which is one of the two conic sections made by a plane cutting a right, circular cone through its vertex: the other is a point.

ruler_in_perspective.pngThere are always two so-called “invariant” points on the transforming line.  Invariants cannot be transformed (moved).  They can both be real, or both imaginary.  If they are real, they can be either distinct, or, as on the left, co-incident.  In the construction, they are co-incident where the parallel edges of the ruler meet, namely at infinity.  This point happens to be in our field of view.  If this point were as far out of view as possible, then the ruler’s intervals would seem equal in the accustomed way.  (Try this with a physical ruler: look squarely at it, then askance along it.  Do note there is no abrupt discontinuity between these views.)

Now, it would appear that we will get equal steps of translation along a line if, and only if, this co-incident pair of real invariants lies precisely in or at the infinite on that line, as the match of the geometry to the ruler seems to show.

 

HalfProtractor.pngThe second construction is probably new to most of us. It is like the first, in that a point is being transformed along a line via a conic in exactly the same way, but here the conic is an ellipse—and the line bearing the transforming point misses it.  This means that the two invariants are imaginary.

The main effect of this on the transformation is that, because there are no real invariants to ‘get in the way’, all the (real) points of the line can be moved, and iterated indefinitely—in fact, cyclically, round and round, so this style of transformation ideally suits rotation, and can provide a stepping measure of it.  Now, we will by this iteration get equiangular steps only if the transforming point always lies in the infinite, as the match, here, to a photograph of an actual, physically-extant protractor seems to show: outgoing and incoming lines through a point in the line, from and to the directrices, are parallel chords of the conic, and so must meet in the infinite.  And all such points must lie in the one ‘line at infinity’ of the conic’s plane.

Thus it would seem that equality, and with it the ability to calibrate (or, to coin a phrase, “to unitise”) both translation and rotation, depend intimately on certain relationships with infinite elements.  Accordingly, in order for us to be sure that our intervals are absolutely equal, we need to be sure that infinity is absolutely locatable.

                         What is Equality?      page 2                  

Towards locating infinity -

·         I expect that most of us accept that N mutually parallel lines all have the same orientation.

·         And we will accept that all N of them meet in the same point at infinity.

·         And we further accept that another set of such parallels (on the same plane as the first set) with a different orientation must meet in a different point at infinity.

·         Finally, we accept that the line drawn through these two, different points at infinity must pass through the points at infinity selected by pairs of parallels on the plane of those parallels.

So, if we can find two such pairs of absolutely parallel lines,

we can find the absolute line at infinity.

Let us try it!

The photograph on the left, below, of lines drawn on a plane, contains two such pairs of parallels.  Find them, and mark them.

The photograph on the right, below, of conics drawn on a plane, contains two circles.  Find and mark them.

perspective_find_the_parallels.jpgperspective_find_the_circles.jpg

 

It should dawn on us that parallelity and circularity are never preserved under projection.

Since spatial measurement, even physical measurement, always involves projection*, it follows from the foregoing that we can never be sure of finding absolute infinity, and in consequence cannot detect absolute equality.

I venture to suggest that this changes everything!

 

* If you doubt this, try using a ruler to measure something in total darkness!