It's unfortunate that this kind of logic isn't emphasized more in mathematics or science in a pure form. I'm pretty sure in high school I had one semester of this kind of logic associated with geometry. That was it until I hit discrete structures in college.

Failure to understand this logic might be part of the reason there's so much magical thinking about things relating to pseudoscience (e.g. paranormal activity, UFOs, conspiracy theories, New Age anything). A stronger base in logic would help people understand exactly why these things are impossible, stupid, or plain crazy with simple deductive and inductive reasoning techniques taught in logic courses.

I think a stronger base in real probability theory (as opposed to simply frequency analysis) would be more beneficial, especially if you follow Jaynes' route and treat it as an extension to logic. As a commenter noted above, "implication" is easy to grasp, except it's not, because logical implication is not necessarily causal or does not even have to make real-world sense, whereas one of the goals of Bayesian Probability Theory is that it should have a qualitative correspondence with common sense.

I'm not sure logic exclusively helps in stopping pseudoscientific and magical thinking. Really a more holistic approach is required:

Logic: to see how to reason deductively.

Epistemology and/or philosophy of science: to see why common meta-beliefs about knowledge and discovery are wrong.

Direct exposure to either a scientific or logic-driven discipline of study or practice: doesn't matter too much if this is a history book or the demands of a compiler. You have to put the reasoning in practice.

Developing a high-quality bullshit detector relies on much more than just an introduction to logic textbook. It is a bit like immunization: you have to encounter bad reasoning in a safe environment and build up your own defenses. (That's all IMHO of course.)

Newton devoted much of his time to alchemy and theology.

This is a great example of how awesome teaching yourself using the Internet can be. If you read and understood that in 30 minutes then you learned it in about one quarter of the time we spent covering this stuff in my discrete math class--and I'm supposedly going to a really good college.

I don't see why you mention your college is good seemingly surprised it took longer than 30 minutes to cover this: wouldn't you expect a good college to have good teachers and wouldn't you expect a good teacher to go slowly enough that all or most students understand?

I suspect there is an optimum speed at which the greatest number of students are able to understand, and I suspect that it isn't the slowest speed possible.

In late elementary and junior high school I recall getting in trouble for being ahead of the class, particularly in computer-related classes and assignments. A faster pace would've allowed students like me to benefit from the class, but may have left slower students behind.

Well I'd figure that a good college would have smart students and with smart students a teacher could go fast.

In any case, even if the teachers are making the right move by going slowly, that doesn't change the fact that teaching yourself using the internet is awesome for the justification that I gave.

Upvoted especially for his list of suggested reading at the end of the article.

Wonderfully straightforward resource for the basics.

Good article overall, but I think I should make a comment:

> Many understand implication intuitively, yet find its symbolic formulation puzzling.

Yes, and for a reason: material implication (the kind of implication discussed in the article) cannot always explain our intuitive sense of implication. My favourite example, from Priest's Introduction to Non-Classical Logic: (A ⇒ B) ∧ (C ⇒ D) ⊦ (A ⇒ D) ∨ (C ⇒ B) is valid when ⇒ is material, yet the statement:

"If John is in Paris he is in France, and if John is in London he is in England. Hence, it is the case either that if John is in Paris he is in England, or that if he is in London he is in France."

makes no intuitive sense. It is very important to discard intuition when dealing with implication in mathematics -- this is why although I'm a big fan of using English connectives instead of symbolic ones, I try to not use the English "if" and "then" in proofs.

Maybe because I'm a mathematician my idea of "intuitive sense" is warped, but intuitively speaking, either John is in Paris or London exclusively. (This isn't given in the statement, but it holds intuitively.) Therefore, in the conclusion, we can assume that at least one of the antecedents "If John is in Paris..." or "If John is in London..." is false, and therefore at least one of the implications in the conclusion is true.

This is a great point.

For something that seems so "common sense," the meaning of implication is difficult to grasp.

I wondered whether I should spend some time discussing the differences between material and logical implication.

In the end, I decided that was just too much for a short overview on the basics.

I also think it is unfortunate that his first theorem is incorrect (n=1 cannot be written as the product of two distinct integers)

That may cost the author some of the readers for whom this text is the most educational, i.e. the ones that have trouble grasping the logic.

Even if it's not meant to be true, I find it even more disturbing that the symbolic and English formulations doesn't seem to quite match up. (IMHO the symbolic formulation doesn't state that n must be integer.)

It's like the source code and comments being out of sync - you don't know which one is wrong.

Another great point.

I fixed the broken theorem, and I added a blurb on the "universe of discourse."

Good catch!

Will fix. ;)

It makes perfect intuitive sense if you're used to the way the word "if" is used in logic.

It makes perfect intuitive sense if you're used to the way the word "if" is used in classical logic. Remember that it's you who has bent your intuitive sense to fit how material implication works. Most people are not going to think it makes intuitive sense, and the fault doesn't lie with them -- it lies with the logic. Logic is like UI: it's Not the User's Fault he doesn't find implication intuitive.

Logic is not like UI. UI is a translation layer between the way humans think and the way a machine works. Logic isn't a translation layer. It's more like a theoretical machine.

The English terms we use to talk about logic could be considered similar to UI. But I don't think that using the word "if" (instead of making up a new word like "garply") was obviously the wrong design decision for the inventors of English logical terminology. "If" is the word in the English language that most closely maps to the concept the terminology designers wanted to describe.

> Logic isn't a translation layer. It's more like a theoretical machine.

More precisely, it is a model for human thought. The way logic is similar to UI is that logic's philosophical success is a function of how well human thought maps to it and how intuitive theorems in the logic are, just as UI's success is a function of how intuitive it is. The last hundred years of discourse on philosophical logic have been focused on one thing -- to formalize the human notion of implication better than material implication does. That's why logicians came up with non-classical logics of the modal, counterfactual (which captures the very intuitive notion of mutatis mutandis), fuzzy and intuitionistic varieties.

OK, but has anyone managed to construct the real numbers with intuitionist logic? Maybe different logics are like different building materials, each with their own advantages and disadvantages.

> OK, but has anyone managed to construct the real numbers with intuitionist logic?

I think it should be possible to do so. As long as you don't need the law of the excluded middle you should be fine.

> Maybe different logics are like different building materials, each with their own advantages and disadvantages.

I see where you're coming from, but I think logic is more than just a tool.

Sorry, that's more precisely ceteris paribus -- "all else being the same".

Is that still true constructively?

No.

Intuitionistic logic has the pleasing property that whenever G |- P \/ Q is provable, then either G |- P is provable or G |- Q is provable.

I think this gets to the heart of what is confusing about this example: (London -> England) /\ (Paris -> France) |- (London -> France) \/ (Paris -> England) is classically valid, even though neither (London -> England) /\ (Paris -> France) |- (London -> France) nor (London -> England) /\ (Paris -> France) |- (Paris -> England) is

By "constructively" do you mean in a constructivist logic like intuitionist logic? I'm too lazy to work it out right now, but I'd place my money on that statement not being valid in intuitionist logic.

Yes. That's what I meant. That was my feeling too, but I know myself too well than to place any bets on what is provable or not constructively unless I work it out. :)