6
$\begingroup$

Question: Is there any name for the natural algebraic structure of the projective line?

Algebraically, a projective line over a field is a set $L$ endowed with two binary operations $+$ and $\cdot$ and three distinct constants $0,1,\infty\in L$ satisfying the following axioms:

  1. $\forall x,z\in L\;\forall y\in L\setminus\{\infty\}\;\;\big(x+(y+z)=(x+y)+z\big)$;

  2. $\forall x,y\in L\;\;(x+y=y+x)$;

  3. $\forall x\in L\;(x+0=x=0+x)$;

  4. $\forall x\in L\;\exists y\in L\;(x+y=0=y+x)$;

  5. $\forall x,z\in L\;\forall y\in L\setminus\{0,\infty\}\;\;(x\cdot(y\cdot z)=(x\cdot y)\cdot z)$;

  6. $\forall x,y\in L \;\;(x\cdot y=y\cdot x)$;

  7. $\forall x\in L\; (x\cdot 1=x=1\cdot x)$;

  8. $\forall x\in L\;\exists y\in L\;(x\cdot y=1=y\cdot x)$;

  9. $\forall x\in L\setminus\{0,\infty\}\;\forall y,z\in L\;\;\big(x\cdot(y+z)=x\cdot y+x\cdot z\big)$.

  10. $0\cdot 0=0$ and $\infty\cdot\infty=\infty$.

Added in Edit: One of my collegues told me that there exists a related algebraic structure, called a wheel, whose purpose is to express the algebraic structure of fields extended by infinity.

There was a discussion in comments why I force the equalities $0\cdot\infty=1$ and $\infty+\infty=0$ and do not leave that product and sum undefined. For $0\cdot\infty=1$ the reason is simple: when we calculate the cross ratio of the quadruple $[a,b,c,d]$, we use the formula $$\frac{(c-a)\cdot(d-b)}{(d-a)\cdot(c-b)}.$$ If one of the numbers $a,b,c,d$ is infinity, then we use exactly those laws: $x+\infty=\infty$ for $x\ne\infty$, and $\frac{\infty}{\infty}=1$. The latter equality is equivalent to $0\cdot\infty=\frac1{\infty}\cdot\infty=\frac{\infty}{\infty}=1$. So, in order to make such calculations legal, in is better to define the addition and multiplication everywhere but do not require the associativity and distributivity in degenerated cases because they crack at those degenerate cases anyway.

Now I have a proof (a bit long though) that for the algebraic structure $L$ satisfying the above ten axioms, we have the equalities: $$0\cdot\infty=1,\quad \infty+\infty=0,\quad\mbox{and}\quad x\cdot\infty=\infty$$ for any non-zero $x\in L$. Moreover, if $L$ contains more than four elements, then we also have $x+\infty=\infty$ for all $x\in L\setminus\{\infty\}$ and $L\setminus\{\infty\}$ is subfield of $L$.

For the cardinalities $n\in\{3,4\}$ there exist two counterexamples: just take the ring $\mathbb Z_n$ and proclaim the element $2$ to be the infinity. Then modify the multiplication in order to guarantee that $0\cdot \infty=1$ and $x\cdot \infty=\infty$ for every non-zero $x$. Then the obtained algebraic structure satisfies my ten axioms but $\mathbb Z_n\setminus\{\infty\}$ is not a subfield of $\mathbb Z_n$. But those seem to be the only counterexamples.

$\endgroup$
12
  • 1
    $\begingroup$ How do you define $0\cdot\infty$, $\infty\cdot 0$, and $\infty+\infty$? $\endgroup$
    – YCor
    Commented Apr 9 at 7:15
  • 1
    $\begingroup$ @YCor This is a very good question. At the moment I am trying to deduce from the axioms that $0\cdot\infty=1=\infty\cdot 0$ and $\infty+\infty=0$. The equality $0\cdot \infty=1$ indeed can be deduced from the axioms, but with $\infty+\infty=0$ I have troubles. $\endgroup$ Commented Apr 9 at 8:04
  • 2
    $\begingroup$ OK, maybe you need to check that all axioms are fulfilled then. Although the choice $0\infty=1$ seem to be an artefact and probably it should better not be defined. $\endgroup$
    – YCor
    Commented Apr 9 at 9:13
  • 5
    $\begingroup$ In classical projective geometry over a field, the projective automorphism group acts transitively on the points defined over that field, so no point has a special role. But your notion of projective line does not have transtive automorphism group. So your geometry is very different. Is there some motivation for your definition? $\endgroup$
    – Ben McKay
    Commented Apr 9 at 15:50
  • 2
    $\begingroup$ I am familiar with the Riemann sphere (which is the complex projective line). In that case we leave $0\cdot\infty$ undefined; so technically $\cdot$ is not an operation. Similarly we leave $\infty + \infty$ undefined, so $+$ is not an operation. $\endgroup$ Commented Apr 9 at 17:45

0

You must log in to answer this question.