2
$\begingroup$

I am looking for a previous reference to the following very simple geometric lemma, which I use in my paper [arXiv:1503.03462]:

Let $P$ be the parabola $y=x^2$. Let $a,b,c,d$ be four points on $P$ sorted from left to right, and let $z$ be the point of intersection of the segments $ac$ and $bd$. Define the horizontal distances $p=b_x-a_x$, $q=d_x-c_x$, $r=z_x-b_x$, $s=c_x-z_x$. Then $p/q=r/s$.

This is easily proven algebraically (though I wonder whether a nice geometric proof exists...)

EDITED TO ADD: This lemma can be stated in a "power of a point"-like way: Let $P$ be the parabola $y=x^2$, and let $a$ be a point not on $P$. Let $\ell$ be a line through $a$ intersecting the parabola at points $b$ and $c$. Then the product $|a_x-b_x|\cdot |a_x-c_x|$ is independent of the choice of $\ell$.

$\endgroup$
3
  • 1
    $\begingroup$ I don't have enough rep to comment... 2 ideas : 1. It is like intersecting chords ratio, may be a good start 2. On the site www-home.math.uwo.ca/~mdawes/woma/potw_back.html prob. #980421. About colinears points. But i am not sure that is a good start. $\endgroup$
    – user70598
    Apr 16, 2015 at 14:14
  • $\begingroup$ The parabola lemma is a limit case of the intersecting chords theorem: If you take 4 points on a circle very close to the bottom part of the circle, and you affinely stretch the picture vertically, then you get almost a parabola. Since the chords were almost horizontal, their length was very close to their $x$-projection. Unfortunately, taking a limit does not count, in my opinion, as being simpler than a straigntforward algebraic calculation, so this is not a "simple" proof. $\endgroup$ Apr 17, 2015 at 6:01
  • $\begingroup$ BTW, this limit argument can translate any length property on a circle to an $x$-projection property on a parabola. E.g.: Let $p$ be a point below the parabola, and let the two tangents to the parabola through $p$ pass through points $a$ and $b$ on the parabola. Then $|p_x-a_x| = |p_x-b_x|$. $\endgroup$ Apr 17, 2015 at 6:11

1 Answer 1

2
$\begingroup$

Not only power of point, but many things about circles may be defined for parabolas with vertical aces. For example, we may define the angle between lines $y=ax+b$ and $y=cx+d$ as $a-c$. Then parabola passing two points $A$, $B$ is a locus of points $P$ for which angle between lines $PA,PB$ is fixed. More advanced example: parabola passing midpoints of the sides of triangle does touch inscribed parabola of this triangle.

This analogy is sometimes called Galilean geometry. There is a book by Yaglom about it, I guess that correct English reference is

A simple non-Euclidean geometry and its physical basis: an elementary account of Galilean geometry and the Galilean principle of relativity. (Springer 1979)

$\endgroup$
2
  • $\begingroup$ Thanks! Very interesting. I'll try to take a look at the book... $\endgroup$ Jun 4, 2015 at 17:39
  • $\begingroup$ I see that the angle-as-slope definition also makes sense in light of the limiting argument that I mentioned above: For small $\alpha$ we have $\alpha \approx \tan\alpha$ $\endgroup$ Jun 4, 2015 at 17:44

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.