# Is it beneficial to think about what natural numbers/object of sets are?

The question is certainly a bit philosophical but there might be interesting answers/views mathematically on this.

## Background

At first, one encounters natural numbers as a set $$\mathbf{N}:=\{0,1,2,3,4,...\}$$ and that a natural number is any element of this set. With a bit of curiosity one might doubt, whether the "..." notation can ever be formal. Of course, everyone knows what is meant by this set, but a suggestive notation is still only that and not more. More reading then leads to the Peano-Axioms (see for example: https://en.wikipedia.org/wiki/Peano_axioms) that are supposed to be satisfied by what we think of as the natural numbers. This still does not answer any question such as "Do natural numbers exist?" (I don't necessary mean "in the real world" but rather as a set) and "what are natural numbers? (again, I don't necessarily mean what they really are, but rather how to think of them)

Hence, as far as I understand, one of the Axioms of ZFC is the axiom of infinity, which ensures that there is an inductive set, meaning a set that contains the empty set as well as the successor of every object in the set. One can then define $$\mathbf{N}$$ as the intersection of all inductive sets and prove that what we defined as $$\mathbf{N}$$ actually satisfies the Peano-Axioms. Hence, it appears quite reasonable to think that these are what we mean by natural numbers (or rather meant by it before we defined them in this way, ensuring that this is a suitable definition). In order to have the names we want to have for the objects, we define $$0:=\emptyset, 1:=s(0), 2:=s(1)$$ and so on. Now, one way to visualize the natural numbers might be points that are equally distanced and to think of the objects of $$\mathbf{N}$$ as those points (it will hopefully become clear later why this might be important) or perhaps one might not visualize them at all.

I am self reading a logic text in which it says that $$\forall$$-statements can be proven by assuming that one has an arbitrary object. This is motivated by everyday examples, such as people: if I know want to prove that every person has some property, I can assume that $$x$$ denotes a person and if I can prove that $$x$$ satisfies this property, I can infer that every person has that property. This is known as universal generalization as far as I know. Now all of this makes sense in this real life example and it also makes sense in general I think (again, I will get to that in my question).

## The Question

Is it any good to think/question what the objects of a set are? I chose the natural numbers because this is a pretty basic example. What I mean by this more precisely is that, of course, one has $$0 \in \mathbf{N}, 1 \in \mathbf{N}, 2 \in \mathbf{N}$$ and so on in this case. But when proving a $$\forall$$-statement about the natural numbers one may prove this by assuming that $$n$$ denotes any natural number. Now, if I visualize a set as some points as mentioned above, this is fairly easy, $$n$$ is just any of those points and that is what a natural number is as well here. But this feels more like a model to escape the abstraction a bit (and thus a bit like cheating?). Hence I wondered if this is a common thing to do, which is still more abstract than using numbering such as $$0,1,2,3,...$$ or if one really thinks of natural numbers as just that, objects, whatever they are and thus $$n$$ is just any of that abstract object, whatever that is.

This can analoguously be done for arbitrary sets: visualize any points as the objects of a set. It also feels more acceptable to me, that the universal generalization holds if one visualizes the objects as sets. It really feels difficult to me to think about objects as something more abstract, but perhaps there are good ways to do so (if there are, I appreciate any comment). I don't know if it does any "harm" to think of objects less abstractly like I do here, but it just does not feel right which is why I wanted to ask what other views on this are. Thank you in advance for any comment.

Edit: Thinking more about this, I think that the more interesting question is why the logical rules that we agree upon (as far as I see it, based on real life intuitions) should also apply to abstract objects. This is probably where the need to visualize it comes in to me and why I mentioned the universal generalization above. If I just think about objects as objects and not more, then I wonder why we can apply the logical rules to them too. What is "for every natural number" even meaning then? This might be because these rules appear to be the most plausible we have, but again, I don't know, this is just a speculation.

• Asking what mathematical objects do rather than what they are has historically proved helpful. This is explained in detail in Chapter 2 of this accessible book. If that doesn't address what you're asking, feel free to reflect that in an edit explaining why.
– J.G.
Dec 26, 2021 at 23:15
• In formal set theory, every object is a set.
– Karl
Dec 26, 2021 at 23:25
• A safeguard I learned from Wittgenstein for this kind of questions is to ask yourself first: What kind of answer could I even expect / assess? You say you don't know "what natural numbers are". Well, are there any other abstract objects where you think you know what they "are"? (If not, how could you assess any answer?) Or you ask why the logical rules "also apply to abstract objects". Well do you think you know why they apply to anything? Or is there anything you know about abstract objects which leads you to accept that something applies to them? If not, then how could anyone answer? Dec 27, 2021 at 21:04
• @Karl Though watch out for junk theorems.
– J.G.
Dec 28, 2021 at 7:26
• You might also consider asking on the philosophy stack exchange, as it's a well-explored concern in philosophy.
– J.G.
Dec 28, 2021 at 14:16

## 2 Answers

All that matters about mathematical objects is their properties and how they relate to each other. Their “nature” is irrelevant. The idea of the “essence” is a bit nonsensical, even if applied to real objects. Anything is just the sum of its properties. When you use a hammer, all that matters is it’s solidity, weight, rigidity, etc. Is there anything beyond that? Mathematicians are just very honest people and do not pretend to know the final reality, as others do.

• Thanks for your answer. I made an edit to the post containing a related issue. If you find the time, I would be interested in hearing your views on that as well! Dec 27, 2021 at 9:45
• "Anything is just the sum of its properties" Hooray! The metaphysics of objects sorted in a sentence!! Dec 30, 2021 at 10:25
• @PeterSmith I understand that this is a long debated question. I just expressed my opinion. Honestly, I think anything beyond manifestations of objects is an illusion. I may be biased by Mathematics. In Mathematics, the "essence" is certainly irrelevant. And I have the impression that the same is true about the "real" world. Dec 30, 2021 at 20:31

I am not sure I understand what your question is, so let me know if this does not approach an answer to your question. Here's how I see it:

Objects are whatever you define them to be, and to prove that an object has a property, you don't really have to assume the object itself exists or imagine how it can be visualised. Instead, you can show that anything satisfying the definition of the object also satisfies the property.

In a way this is even necessary, because we cannot prove that a model of $$\mathsf{PA}$$ or $$\mathsf{ZFC}$$ even exists, without assuming stronger axioms. We can prove what kind of properties things like natural numbers have if they were to exist, by reasoning with the description of the object instead of with the object itself. This way you don't have to contemplate the actual nature of natural numbers, nor their actual existence, but you can still agree on how they behave.

• Thanks! If you have any recommendation how I can clarify the question, make sure to mention it and I will gladly try to improve it. I made an edit to the post, perhaps you find the time to include an edit in your answer too, commenting on that. Again, thank you. Dec 27, 2021 at 9:41
• I believe your edit is sufficiently different to ask a separate question, although I'm fairly certain it has been asked before. Basically you seem to be asking why we accept first-order logic as a basis, and not some other logic, right? You can take a look at math.stackexchange.com/questions/683072 or math.stackexchange.com/questions/141759 or math.stackexchange.com/questions/176263 , and any related links you can find through there. Dec 27, 2021 at 15:00
• Yes, that is what I am asking. I read the links you provided, but unfortunately I did not find a satisfying answer for that question. Dec 27, 2021 at 16:40