Is a *surface*-line intersection rational if the line passes through another rational point on the surface?

58 Views Asked by At

I've seen some interesting geometric interpretations of Diophantine equations before. For instance, if we wish to find Pythagorean triples, we can take $a^2 + b^2 = c^2$, divide through by $c^2$, and get $x^2 + y^2 = 1$, which is the unit circle. We then choose an obvious rational point on the circle, e.g., $(-1, 0)$.

Now, any line with a rational finite slope that passes through $(-1,0)$ will intersect the circle at another point with rational coordinates. Then $a$ and $b$ are the numerators of the coordinates, and $c$ is the denominator (after bringing them to a common denominator).

Of course, something similar happens on elliptic curves as well.

I've been looking into a similar problem, but it requires three-variables--and therefore a surface more complex than the unit circle. It's also a cubic:

$$z^3 - z = xy(x^2-y^2)$$

(This is in relation to the order-3 magic square of squares problem.) If a line passes through one rational point on the surface, must any other intersections also be rational points? Or is there a restriction of some sort?

1

There are 1 best solutions below

3
On BEST ANSWER

No, the other two points need not be rational. In general they will be defined over a quadratic extension of $\mathbb{Q}$ and conjugate.