7

The other day, I watched the YouTube video "Can you pass this logic test from Brazil?" by MindYourDecisions.

We are told that Pinocchio always lies, and that Pinocchio says 'All my hats are green'.
We are asked to find which of the following 5 sentences we can 'conclude' from Pinocchio's sentence, considered as a mathematically false statement.

A) Pinocchio has at least one hat.
B) Pinocchio has only one green hat.
C) Pinocchio has no hats.
D) Pinocchio has at least one green hat.
E) Pinocchio has no green hats.

Not a logician or mathematician myself, I sort of figured that the negation of Pinocchio's sentence P = 'All my hats are green' is S = 'I have at least one hat that is not green'.

As S must be true, Pinocchio must have at least one hat, and one is for sure not green.

Watching the video, I was very puzzled by how the sentences were eliminated.
I would have looked for direct contradictions between A-E and S.
This would have ruled out B and C.
That's where I would have been stuck, though. I wouldn't know how to eliminate D and E.
For D my attempt would have been '1 green hat', which is compatible with D but contradicts S, so it is not safe to conclude D from S.
But for E, I don't see the contradiction. '1 blue hat', or '2 blue hats' are all compatible with both E and S.

Instead, in the video, B, D, and E were eliminated by making an example that contradicted them, and showing that this was compatible with P also not being true.
E.g. to eliminate B, the argument was: if Pinocchio had 2 green hats and 1 blue hat (which contradicts B), he could say P and he would be lying, therefore B cannot be concluded from not(P).

I am clearly missing something, as I do not get how this works.

Can anyone please explain how the elimination method in the video works, logically/mathematically?

As extra food for thought, I imagined the case if Pinocchio had said 'None of my hats are green'.

Wouldn't that lead to the same conclusion, A?


EDIT further attempt to elaborate/solve

Assuming that 'what can we conclude' means 'in which case $S \to X, X \in \{A,B,C,D,E\}$ is always true, i.e. can never be false.

$S \to X = \bar S \lor X$ is false only when $S$ is true and $X$ is false. In all other cases it's true.

So $S \to X$ is always true when $\bar S \lor X$ is always true, i.e. when there is no instance of its opposite $S \land \bar X$ being true.

Given that $S = \bar P$, perhaps the method in the video is based on showing that $\bar P \land \bar X$ is true in at least one case (i.e. that $P$ and $X$ can both be false at the same time), which proves that $S \to X$ is not always true.

Am I on the right track or completely off?

  • 3
    Although this question may appear basic, it is thoroughly explained. I don’t think this deserves downvotes – FShrike Jun 30 '22 at 18:59
  • 1
    @insipidintegrator Not so. If I have no hats, the statement "All my hats are green" is a vacuous truth. As a liar, Pinocchio is not allowed to state any truths, including vacuous ones. – Robert Israel Jun 30 '22 at 19:10
  • 1
    Thank you @FShrike . I don't mind downvotes. I have a genuine scientific curiosity about this topic, so if anyone is generous enough to try and help me understand, I will appreciate that very much; those who take some bizarre pleasure in humiliating others for no reason whatsoever, can safely knock themselves off with the downvoting button; I for sure am not going to lose any sleep over it :) – user6376297 Jun 30 '22 at 19:12
  • 1
    Well, downvotes are not intended for humiliation here... but I certainly don't think it was appropriate. I'll see if I can answer your post or at least provide comment – FShrike Jun 30 '22 at 19:14
  • @FShrike , thanks, that'd be great! I really do not get that way of showing that a sentence cannot be a consequence of S by two negatives. As for downvoting = humiliation, you're right of course, it is not intended like that, but that will not stop people from doing it exactly only for that reason and nothing else. Besides, just downvoting without providing an explanation or a suggestion for improvement is like slapping a child unconscious because he tried to cross the road without looking, and never explain to him what he did wrong or what he should do instead. Very mature and useful. – user6376297 Jun 30 '22 at 19:23
  • 4
    I also dislike anonymous downvoting, where no pre-existing comment communicates the downvoter's thinking. Unfortunately, this is a very common practice at MathSE. – user2661923 Jun 30 '22 at 19:27
  • 1
    @user6376297 I hope you’re not a child and can at least try to find simple explanations (though they may not be correct) for downvotes (eg. answering PSQs, no mathjax, wrong answer, low-effort) and try to correct them? There’s a lower age limit on the site for a reason. Tl;dr: I find your comparison with a child incorrect. – insipidintegrator Jun 30 '22 at 19:27
  • @insipidintegrator : I did try! I explained my thought process. And I thought very hard about the video's approach, but I don't get it. Hence my asking for advice in a website designed for exchanging information and insight. I completely disagree that one should guess the reasons for downvotes. Why would an assumption be better than an explanation? Sorry, you seem to be accusing me of being lazy when I actually tried to figure this out, but you justify people lazily clicking buttons without even bothering to argument it. There is no need to have an opinion on / judge everything and everyone. – user6376297 Jun 30 '22 at 19:51
  • @user2661923 : well, clearly someone gets off on such stuff. Let them have their fun. People who are seriously interested in science will carry on trying to help one another, in spite of all haters and trolls. But yes, it is very annoying, admittedly, and IMO it creates a toxic culture that discourages open discussion and collaboration. Such a pity, as there are really brilliant people around, especially on MathSE. Constructive criticism, yes, please! Mindless demotivation and anonymous snipers, what's the point? – user6376297 Jun 30 '22 at 20:01
  • 3
    A general suggestion: don't take YouTube videos too seriously ... – WhatsUp Jun 30 '22 at 22:19
  • @WhatsUp :D good point. But to a non-expert like myself, that guy's videos are very instructive. I learned a lot about maths and logic by watching them - clearly not enough though, if i am stuck on this one :( :) . – user6376297 Jul 01 '22 at 05:43
  • Miiind your decissssionnns, IIIIIIIIIImmm PreshTalwalkar! (Or at least that's how some of his videos start) Big fan of him, although some of his titles look a lot like clickbait, he's one of the channels I would recommend people to watch (as long as they watch the correct videos, and not confusing ones like evaluating $9 \div 3(3+3)$ or whatever). By the way, I saw this video as well, and I actually liked Presh's explanation so I'm happy that someone discussed it in detail here. Good question. – Sarvesh Ravichandran Iyer Jul 01 '22 at 06:38
  • The ratio of good youtubute math videos is too small to give the advice to generally watch them. But finally, everyone must make this decision him/herself. My advice is clearly to read wikipedia and pdf-articles instead. – Peter Jul 01 '22 at 07:23
  • @Peter : sure, I fully agree. I am not planning to learn maths or logic only by watching YouTube videos. As I am not doing this professionally, I start from those videos because they're fun and show me problems that I wouldn't have been exposed to otherwise, and I take it from there for further study in the appropriate places. E.g. my asking the question here is also a way to get pointers to appropriate reading / literature. – user6376297 Jul 01 '22 at 15:29

3 Answers3

4

A nice instructive question indeed —with the proviso that one should not appeal to formal devices.

Our task is to draw a decidedly true conclusion from Pinocchio's false statement. Hence, we want to obtain this position: Pinocchio's statement is false while our statement is true.

It must be relatively trivial to eliminate (B) and (D): There are situations that make Pinocchio's statement be false, but our statement be false as well. Examples: 2 green hats, 1 blue hat for (B) and 2 blue hats for (D).

(E) certainly falsifies Pinocchio. But there are situations that make our statement be false at the same time. Example: 1 green hat, 1 blue hat.

(C) is tricky: In daily speech, we use such statements with existential import, that is, the listener understands also that Pinocchio actually has got hats (since the quantifier 'all' usually connotes plurality). However, we have been warned that we should not regard the question in daily speech conventions. So, we shall seek a falsifying instance. If there is no falsifying instance, then the statement is accepted as true. Hence, Pinocchio may have no hats and that makes whatever he says about his hats be true. In order that what Pinocchio has said to be false, he must have at least one non-green hat as a counter-instance. So, if (C) were our statement, Pinocchio's could be true, too.

We can take (A) as a continuation of (C): If not (C), then (A).

Let us check it formally: Pinocchio's statement can be expressed as $\forall x(Px\rightarrow Gx)$. Its negation is $\exists x(Px\wedge\neg Gx)$. That is, at least, Pinocchio has one hat and it is not green —by formal analysis, we secure more information in a much swifter fashion.

Tankut Beygu
  • 4,412
4

The elimination approach can be thought of as considering a universe of situations $S$. (Or worlds.) For any situation $x \in S$, we say $P(x)$ iff all of Pinocchio's hats are green in $x$.

In order for us to conclude some other predicate $Q$ from $\neg P$ (that is, that Pinocchio is lying), it must not be possible for us to find an $x$ such that $\neg P(x)$ but also $\neg Q(x)$. The existence of even a single such $x$ breaks the implication.

For example, suppose $Q =$ "Pinocchio has no hats." Could $\neg P$ imply $Q$? No: Consider the situation $x =$ "Pinocchio has one red hat and no other hat." We see that $\neg P(x)$ (it is not the case that all of Pinocchio's hats are green, since there is a red one), but also $\neg Q(x)$ (it is not the case that Pinocchio has no hats, since we has exactly one). Thus, it cannot be the case that $\neg P$ implies $Q$.

Brian Tung
  • 35,584
  • Thanks Brian, I accepted your answer as it's the one I understand best. I believe it's what they call a proof by contrapositive, as mentioned by others, too(?). http://math.colgate.edu/faculty/valente/math250/lecturenotes/Sec101supplement.pdf . The thing that still unsettles me is that one makes up a case for $Q$ almost at random, with the only condition that it falsifies $P$. I cannot immediately think of a real-life case where this would make intuitive sense. BTW I like the Brian Griffin icon, with the dry Martini :D – user6376297 Jul 01 '22 at 16:43
  • Wait... I get it now... If $P =$ 'all my hats are green', $\bar P = S =$ 'I have at least one hat, at least one is not green' and $E =$ 'I have no green hats', to prove that $S \to E$ is not always true, I only need to find one case where $S$ is true and $E$ is false (e.g. 1 green hat and 1 blue hat). Instead I was looking for cases where $E$ was always false when $S$ was true. The latter works for $B,C$ because they are directly contradicted by $S$, but not for $D,E$. What confused me was the fact that the case made to falsify $E$ seemed disconnected from $S$. Sorry... – user6376297 Jul 01 '22 at 18:15
  • @user6376297: Yes, that's right, a single instance in which $S$ is true and $E$ is false is sufficient to disprove $S \implies E$. – Brian Tung Jul 02 '22 at 02:42
  • @user6376297: Oh, and yes, I like my avatar too. He can be a bit of an obnoxious blowhard, and I'm afraid to say that that sometimes is true of me too. :-P – Brian Tung Jul 24 '22 at 18:12
0

the negation of Pinocchio's sentence P = 'All my hats are green' is S = 'I have at least one hat that is not green'.

Not necessarily. From sentence $P$'s wording alone
   All my hats are green
   All my hats are green
   All my hats are green
we know only that Pinocchio is lying about at least one of the following: the colour of his existent hats (note that if they don't exist, then whatever claim he makes about them is vacuously true), the portion of his existent collection of hats being green, or the type of possession of which his entire collection is green (in this case, the collection of his existent hats is not entirely green).

In other words, we know that Pinocchio has at least one hat (let's symbolise this as $L),$ and that $$¬X(g) \lor S \lor ¬Y(h).$$

The task is to determine which of the given options is/are (must be) implied by this compound premise.

in the video, B, D, and E were eliminated by making an example that contradicted them, and showing that this was compatible with P also not being true.

Yes, this is a natural approach. For each of options $\textbf A$-$\textbf E,$ let's assume for the sake of contradiction that it is false; if a contradiction is derived from the above premise, then our assumption must have been false, that is, that option must be true (note that that option is not necessarily false if we aren't able to derive some contradiction).

A more direct approach (using option $\textbf C$ as an example) is this: we want to check whether $$L\land\big(¬X(g) \lor S \lor ¬Y(h)\big)\implies \textbf C;\tag1$$ equivalently, by contrapositive, whether $$\lnot \textbf C\implies \lnot L\lor\big(X(g) \land ¬S \land Y(h)\big).$$

Implication $(1)$ is true only for option $\textbf A.$ On the other hand, for each of options $\textbf A$-$\textbf E$ (still using option $\textbf C$ as an example), $$\bigg(L\land\big(¬X(g) \lor S \lor ¬Y(h)\big)\bigg) \land\lnot \textbf C$$ is satisfiable, meaning that implication $(1)$ is false.

imagine the case if Pinocchio had said 'None of my hats are green'.

Wouldn't that lead to the same conclusion, A?

Yes: if $\textbf A$ is false, that is, Pinocchio has no hat, then "None of my hats are green" is true, so Pinocchio wasn't lying, which is a contradiction, which means that $\textbf A$ must actually be true, contrary to our supposition.

EDIT Assuming that 'what can we conclude' means 'in which case $S \to X, X \in \{A,B,C,D,E\}$ is always true

The premise is actually $$L\land\big(¬X(g) \lor S \lor ¬Y(h)\big)$$ instead of merely $S.$

ryang
  • 44,428