2

I came up with a way to do set theory using polynomial arthimetic.

To explain, consider two polynomials $P(x)$ and $Q(x)$, then we can observe the following analogies:

1. In Polynomial addition

$$ |P(x)|+|Q(x)|$$

Is analogous to set intersection of sets because the above equation is only true for $x$ values which make both $P$ and $Q$ zero at the same time.

2. In Polynomial multiplication

$$|P(x)| \cdot |Q(x)|$$

Can be thought of as set union because the above equation is true for the numbers which satisfy either $P$ or $Q$

3. In polynomial division

$$ \frac{|P|}{|Q|} $$

The above expression can be thought of a set difference, we remove the zeros of Q which exist also in P.


Analogies to set complements:

Polynomial reciprocal is similar to set inverses, for example

$$ |P(x)^{-1}| = \frac{1}{|P(x)|}$$

Is defined for all values of $x$ except where $P(x)=0$ where it is undefined. To get the whole universal set back, we multiply inverse with regular one:

$$|P(x)| \cdot \frac{1}{|P(x)| }=1$$


Proving set identities using polynomials

  1. $ A \cup (B \cup C)= (A \cup B) \cup C$

This is directly analogous to associativity of polynomial multiplication , considering three polynomials $P,Q,R$:

$$ ( |P| \cdot |Q|) |R| = |P|( |Q| \cdot |R|)$$

  1. $A \cup B = B \cup A, A \cap B , B \cap A$

This is directly analogous to communality of polynomial multiplication.

Demorgan's laws

  1. $$ A^c \cup B^c = (A \cup B)^c$$

Easy relation to understand with polynomial way, consider two polynomials $P(x),Q(x)$, then it is trivial that:

$$ \frac{1}{|P|} \cdot \frac{1}{|Q|} = (|PQ|)^{-1}$$

  1. $$ (A \cap B)^c= A^c \cup B^c$$

This one just turns out weird, but it works well for most identities where there isn't addition and multiplication on both sides


My question: Has this idea been studied before? Where there any new insights from thinking of sets like this?

[Note: I know of generating functions already]

Asaf Karagila
  • 393,674
  • Your first property doesn’t work: You can easily have $P(x)+Q(x)=0$ for $x$ which makes neither polynomial zero (e.g., if $P=-Q$, then any value of $x$ will work). Either way, it seems like you’ll need to allow infinite series if you want any infinite sets – boink Feb 11 '21 at 20:44
  • @boink fixed that – tryst with freedom Feb 11 '21 at 20:46
  • yes, in the generating function book, infinite sequences was modelled using polynomials albeit in a different way. – tryst with freedom Feb 11 '21 at 20:47
  • I think it should still not work: In general, for any $x$ where $P(x)=-Q(x)$, we have $P(x)+Q(x)=0$, so take, for example, $P(x)=x$ and $Q(x)=x^2$. Then at $-1$, the sum is zero – boink Feb 11 '21 at 20:51
  • 1
    @boink I think I found a way to fix that problem – tryst with freedom Feb 11 '21 at 20:54
  • 1
    You're sort of getting at algebraic geometry - see here. – Noah Schweber Feb 11 '21 at 21:15
  • 1
    It's an analogy, but how would it help you in DOING set theory? Compare to indicator functions, which are much simpler, satisfy similar properties and are used everywhere. – Sam Feb 11 '21 at 21:51
  • @Sam Indicator functions are new to me >.< . I was just playing around and thought of this abstract way of doing set theory and wanted to see if someone else had thought of this before me. Nonetheless, I'm adding Indicator functions and algebraic geometry to the stuff I wish to learn later on :P – tryst with freedom Feb 11 '21 at 21:55
  • How do you get transfinite recursion, the constructible universe, forcing, and large cardinals using polynomials? – Asaf Karagila Feb 12 '21 at 08:15

1 Answers1

2

A rather elaborate theory related to this was worked out over a period of several decades by the Russian mathematician Vladimir Logvinovich Rvachev (1926-2006) (biography), called called $R$-functions. See Semi-analytic geometry with R-functions by Vadim Shapiro (2007; footnote on p. 3 discusses the origin of the name "R-function").

  • Also the fact that polynomials correspond to formulas (in the language of rings), and so their zero sets correspond to definable sets (with parameters), and both model theory in general and algebraic geometry in particular have studied these sort of behaviour. And also the fact that there are many different algebras, that students don't normally get told about, and certainly not about their interconnections throughout mathematics. So maybe universal algebra should also be mentioned. – Asaf Karagila Feb 12 '21 at 09:34