0

As an undergraduate student, who wants to solidify his mathematical skills, I want to understand what exactly determines if a proof is rigorous.

amWhy
  • 209,954
Saikat
  • 2,461
  • I'm not sure there's a short and useful answer to this - one can more or less easily describe a formal proof (which is undoubtedly rigorous), but that excludes the sort of proof that humans actually write. And one can try to describe the sort of proof humans actually write, but that's no easier than explaining how to speak a foreign language. It's somewhat of a skill one has to practice. – Milo Brandt Jan 21 '16 at 03:29
  • I'm a little weak in terminology. What is a formal proof ? – Saikat Jan 21 '16 at 03:30
  • It's what proof theorists work with - basically, the most bare bones notion of a proof, where one writes down a series of statements, each related by a fixed set of inference rules to the previous statements and axioms. It's sort of the structure that displays every step from axiom to theorem (and is, for instance, possible to verify computationally). It's not really what one would first learn as proof since it itself takes some mathematical machinery to understand at a useful level. – Milo Brandt Jan 21 '16 at 03:39
  • Note: a rigorous proof and a formal proof are not the same thing. A formal proof is a very well defined structure, it is a directed tree of expressions where every single node is a proposition, every edge is a rule of inference, the sources are axioms, and the sink is your theorem. You've probably never actually seen a formal proof, it can be like looking at assembly language. A rigorous proof, on the other hand, is just a proof that is trustable. Formal, like a dinner party, every thing proper down to the last mind numbing detail. Rigor, like a bridge, able to withstand criticism. – DanielV Jan 21 '16 at 05:56

2 Answers2

1

To me, a proof is rigorous if I understand every step and the conclusion follows without any gaps from the premises.

This means, of course, that some proofs that other consider rigorous are not considered such by me because of my lack of understanding or knowledge.

marty cohen
  • 107,799
1

A rigorous proof is understood to be a proof that you typically find in mathematical papers and mathematical analysis books. In other words, it's a proof that you can't poke holes in, so the conjecture must necessarily be true because you have covered all possible objections one might come up with. My analysis prof made the analogy of being a lawyer, and arguing before a jury, whom you are trying to convince that what you are saying is true, and that the opposing counsel is lying.

In practical terms for the undergrad student and/or newbie, there are two broad types of formal/rigorous proofs

(i) constructive proofs

(ii) proofs by contradiction

With proof by contradiction, you suppose that the conjecture is false and consequently arrive at a contradiction.

With constructive proofs, you build your arguments and show why necessarily a conjecture must be true.

These types of proofs are common in mathematics because conjectures are written to lend themselves to such style of proof - in the sense that conjectures make an assertion about the conditions under which a given statement is true.


In a more concrete sense, lets take the example of continuity, specifically a continuous function.

Informally, a function is continuous on an interval if we can draw the graph from start to finish without ever once picking up our pencil. In other words, a function is continuous if its graph has no holes or breaks in it.

Formally, we have to go into more detail, meaning:

  • define a set $A$ as a subset of the set of real numbers $\mathbb{R}$, i.e. $S$ $\subseteq$ $\mathbb{R}$
  • define the function $f$ as being a map from the set $A$ to $\mathbb{R}$, i.e. $f:A \rightarrow \mathbb{R}$
  • define a constant number $c$ as being an element of $A$, i.e. $ c \in$ $A$
  • $f(x)$ is continuous at point $c$ if and only if for every number $\epsilon>0$, there exists a number $\delta>0$ such that $|x-c|<\delta$ $\Rightarrow$ $|f(x)-f(c)|<\epsilon$ (with $\epsilon \in \mathbb{R}$ and $\delta\in \mathbb{R}$)
  • if the above statement about $f$ holds for all points in the set $A$, then the function $f$ is continuous on the set $A$.

So you see in the formal sense, we have expanded our "intuitive" definition and provided a constructive algorithm for how you can determine if a function is continuous or not. The things that you should especially notice is that first we picked a single point in the set $A$ and then expanded to the entire set $A$ (which has infinitely many points). Quantifiers (statements like for all, for one, etc.) are key to getting the grip on how to prove things.

A more detailed write-up of how to write rigorous proofs is written by Prof. Hutchings at UC Berkley here: https://math.berkeley.edu/~hutching/teach/proofs.pdf. The book "How to Read and Do Proofs: An Introduction to Mathematical Thought Processes" by Daniel Solow is also a great read to get you used to thinking about rigorous proofs.

rocinante
  • 517
  • Thank you for taking the time to write this answer. I will surely read the book. I do have one question though. Are there any defined rules a proof must for or should it just be enough to convince a well-formed human mind ? I ask because, in theory, something obvious could not be obvious to another species, who may find potholes on it. My point is, is it sufficient to say a proof is rigorous if no human mind can find a loophole in its steps ? – Saikat Jan 21 '16 at 04:43
  • The "unspoken rules" are that you have to DEFINE everything and make sure you remember definitions. A whole bunch of theorems are easily proved by simply writing down the definitions and following through the logic of their consequences. If you set up everything correctly in at the definition stage, you account for all possible objections, so you block all loopholes :) – rocinante Jan 21 '16 at 04:50
  • @user230452 And another thing: reading and participating in MathSE is great practice in learning how to write formal proofs, and understanding the intuition behind theorems. The book "Proofs without Words" by Roger Nelsen is a clever book of proofs, but I would not recommend using that style in your courses. The argument "it's obvious if you look at it" generally doesn't fly with math profs. – rocinante Jan 21 '16 at 04:56
  • Thank you for your tip, Roci. But, my point is the same. We don't really block out all loopholes. We just block out all the ones humanely visible. For example, there would be some proofs that a child would do that would elude all loopholes to him, but not to a mathematician.Similarly, if there is a intelligentmathematically-inclined species out there, would it be possible for them to find loopholes in our proofs ? What I'm asking is if there are some defined standards after which we decide a proof is rigorous or if it is just enough to convince a well formed human mind ? – Saikat Jan 21 '16 at 07:19
  • I think it's a matter of opinion if there are "defined standards". There is no top mathematical body that decrees mathematical standards; it's a community decision, led by influential mathematicians who have done extensive work and convinced a lot of other mathematicians that their ideas make sense. Therefore, the only standard is a "well formed human mind". – rocinante Jan 21 '16 at 07:49