8
$\begingroup$

We know the general figure of Venn diagrams for two or three distinct sets.

There are many formulas related to two or three sets.

For example, one of Distributive Law is

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

We can visualize it by using Venn Diagram, and guess that it is true.

And also, we can prove it to show that each side is contained in the other side.

It is my question. For rigorous proof, I know we should only use mathematical logic and theorem. Nonetheless, I want to check that Venn Diagram proof is also available for some easier cases.

Can Venn Diagram be one method of proof?

Can we prove that all proof by using Venn Diagram method for two or three sets is true?

If we prove that, then all statements for two or three sets can be strictly proved by using Venn Diagram.

$\endgroup$
  • 1
    $\begingroup$ Indeed, the Venn diagram is one method of proof. $\endgroup$ – Wuestenfux Dec 27 '18 at 9:40
  • 1
    $\begingroup$ Related question: under what circumstances is a Venn diagram for more than three sets drawable? $\endgroup$ – timtfj Dec 27 '18 at 12:31
10
$\begingroup$

Venn diagrams are not a formal proof, nor a substitute for it, just an illustrative tool that can be useful as a guiding tool for your narrative/proof.

If writing a formal proof for this law, you will need to show

$$A \cup (B \cap C) \subseteq (A \cup B) \cap (A \cup C) \;\;\; \text{and} \;\;\; (A \cup B) \cap (A \cup C) \subseteq A \cup (B \cap C)$$

and then use the fact that if $X \subseteq Y$ and $Y \subseteq X$, then $X = Y$.

If you don't need formality, then in the appropriate context it can be used, I suppose. And, up to your ability to produce said diagrams, you could use a Venn diagram of $n$ circles, depending on what you're proving, but it gets messy quick so I wouldn't recommend it for more than $3$ sets.

In short, it depends on the level of formality that is expected of you. There's no denying that Venn diagrams in contexts like these are super, super helpful in illustrating concepts, and can be taken as a sort of heuristic proof, but they are not a substitute for formal proofs.


I say this in light of the assumption that you are probably encountering this in a class of some sort like a number of questions here. Classes in set theory, generally, will expect formality, not Venn diagrams, for example. In research, publications, journals, etc., things are much, much murkier depending on the context.

$\endgroup$
  • $\begingroup$ Technically, you don't want circles; there are no Venn diagrams of more than three sets with circles. $\endgroup$ – NoLongerBreathedIn Dec 29 '18 at 1:56
6
$\begingroup$

I think that this is a case of "consider the audience".

If the proof is intended for research-level mathematics, I can't imagine a scenario where a Venn diagram would be necessary because any statement that can be sensibly encoded in a Venn diagram is probably elementary enough to not require a proof.

If the proof is from a student studying a topic, then this depends on the teacher's expectations. If they are teaching you formal proof techniques, then a Venn diagram would probably not be sufficient because the student would be missing the point of the exercise. This might depend on the level of the course though.

If I were teaching a proof writing course to math majors, I would not consider a Venn diagram to be sufficient. If I were teaching an intro to discrete math for CS or science majors, then I would accept a Venn diagram if it was accompanied by a sentence or two explaining why it shows what the author is claiming it shows.

$\endgroup$
4
$\begingroup$

I think the second part of the question amounts to "can we formally define a set of equivalences between features of a Venn diagram, and symbols and operators of set theory, such that every Venn diagram proof is equivalent to a formal proof using the operators and symbols?"

It seems to me that this might be possible in a very restricted way, but will run into difficulties because of things like needing to distinguish between open and closed sets, $\subset$ and $\subseteq$, so on. So we'd have to be very careful about which relations can be included in the system, and would probably find rather quickly that there weren't enough of them for our purposes.

I think it could probably be made to work for a defined class of simple identities that met certain preconditions—but that the preconditions would be so restrictive that we wouldn't find a "formal Venn diagram method" particularly useful.

Example of the problem

What does this diagram tell us about $A$, $B$ and $x$?

A set, a subset and an element of the subset.

We probably want this:

  • $x$ is represented as a point inside the boundary of the area representing $A$; this is defined to mean $x\in A$.
  • Similarly, $x\in B$.
  • The boundary of the area representing $A$ is wholly inside that of the area representing $B$: this is defined to mean $A\subset B$.

Already some limitations are apparent:

  • We've defined the boundary of $A$ being inside that of $B$ to mean $A\subset B$. That's fine, but how will we represent $A\subseteq B$?
  • There's an area on the drawing which represents $B-A$. Is it meant to contain any elements? If not, how do we avoid being misled by it?
  • What about points on the boundary between $A$ and $B$? Which set do they belong to? (Equivalently, is $A$ open or closed?)

Ultimately, the issue is this: the diagram is a region of $2$-dimensional space, of which areas on the diagram are subsets. However, we can't necessarily give these subsets the same properties as the sets they represent. In particular, we can't necessarily avoid giving them extra, unwanted properties—such as all subsets being represented by proper subsets.

$\endgroup$
  • $\begingroup$ My gut says this could be worked on and turned basically into something equivalent to truth tables. It's reminiscent of Lewis Carroll's symbolic logic (before modern logic was standardized). $\endgroup$ – Mark S. Dec 27 '18 at 19:13
  • $\begingroup$ I think the general rule would be that a region in the figure may or may not represent any elements of the sets (unless an element such as $x$ is explicitly shown) and that elements never occur on the boundaries between regions. You can still go wrong by forgetting these rules, but then you can go wrong even using the most formal methods if you forget the rules. $\endgroup$ – David K Dec 27 '18 at 20:08
  • $\begingroup$ @MarkS.when I saw the question I immediately thought of Karnaugh maps, which I had a lot of dealings with at university (in connection with digital circuit design). They're used mainly as a graphical way of simplifying Boolean expressions, which makes them pretty close to set theory. ("Mainly" since they're also used to find "overlapping" expressions, which is important in preventing momentary erroneous outputs when an input changes.) $\endgroup$ – timtfj Dec 28 '18 at 22:06
  • $\begingroup$ @DavidK The no points on a boundary rule would actually allow open and closed sets, wouldn't it—it makes the drawn boundaries into symbolic ones in a "cut" that can be as wide as we like. There's nothing to say that the representation of $[0,1)$ has to be right next to the representation of $ [1,2]$ $\endgroup$ – timtfj Dec 28 '18 at 22:19
  • 1
    $\begingroup$ @DavidK I've also a sneaking suspicion that iin the process of proving the system worked, we'd already have done most of the proofs anyone would want to use it for. (And a suspicion that someone in the 19th century probably had a go!) $\endgroup$ – timtfj Dec 28 '18 at 22:52
3
$\begingroup$

That the two sets are equal is equivalent to showing that an arbitrary element $x$ is a member of the left hand side if and only if it is a member of the right hand side. Now that will depend only on which of the statements $x\in A$, $x\in B$, $x\in C$ hold. Therefore it suffices to check eight cases. This corresponds to looking at the eight regions of a Venn diagram for three sets. So, yes, a proof by Venn diagram is valid in this case.

$\endgroup$
2
$\begingroup$

I would say that Venn diagrams are not good for formal proofs, but great if you have to say whether a formula is a tautology or not - just check for any counter-example and if you find one, then the formula is not a law.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.