How to test someone's logic skills?

[The idea]

What's a good way to test someone's logic skills? The following approach occurred to me.

Provide a series of alleged proofs, half of which are valid and half of which are invalid. The valid proofs will be as unpersuasive as possible and the invalid proofs will be as persuasive as possible. The person will be required to specify, for each alleged proof, whether it is valid or invalid.

What are the advantages and disadvantages of this approach?

What other approaches might one use?

[Bodhi Dharma Zen]

Can you be more specific? Are you talking about syllogisms?

[The idea]

Originally posted by Bodhi Dharma Zen
Are you talking about syllogisms?
I suppose one could start with syllogisms, but I'm talking about actual reasoning that people actually use when they are arguing with each other. I presume that some of it is valid.

[Soapy Sam]

Can anyone demonstrate to me that logic generates right answers. Not true answers, mind you.
Right answers.

I have a dark suspicion that there are few real-world situations more complex than whether a switch is open or shut, which are actually amenable to logical analysis.

For example-
Which is better, a dog or a camera?
What's so bad about being over forty?
Why me?

[Jeff Corey]

And, "Of what use is a philosopher, unless he can get that football off the roof from behind the chimney?"

[Kaylee]

Originally posted by Jeff Corey
And, "Of what use is a philosopher, unless he can get that football off the roof from behind the chimney?"
LOL. I think those college students "flunked" an insight test... (http://forums.randi.org/showthread.php?s=&threadid=56057) Guess they should have given that problem to some lab monkeys. :)

[Iamme]

Originally posted by Jeff Corey
And, "Of what use is a philosopher, unless he can get that football off the roof from behind the chimney?"

----------------------------------------------

:dl:

[Iamme]

Originally posted by The idea
What's a good way to test someone's logic skills? The following approach occurred to me.

Provide a series of alleged proofs, half of which are valid and half of which are invalid. The valid proofs will be as unpersuasive as possible and the invalid proofs will be as persuasive as possible. The person will be required to specify, for each alleged proof, whether it is valid or invalid.

What are the advantages and disadvantages of this approach?

What other approaches might one use?

-----------------------------------------------------

Please give us one real life example. Thanx.

[Iamme]

*I* know a test that will determine if you have high speed brain processing skills. Here is how: (Note, I am going to submit this idea to the Wisconsin Board of Education.)

Are you ready? Okay. Think fast. Whoever posts the answer here the fastest, wins. But also, be honest in saying about how fast it took you to come up with the answer.

Who would my grandfather's boy's son be, who has an uncle that is the married husband of my blood aunt, who is not my brother?

[T'ai Chi]

Originally posted by The idea
What's a good way to test someone's logic skills?


Hmm, I'd give them some mathematics problems, but without equations and etc. so it doesn't really look like a math problem.

[Jeff Corey]

Originally posted by Iamme
*I* know a test that will determine if you have high speed brain processing skills. Here is how: (Note, I am going to submit this idea to the Wisconsin Board of Education.)

Are you ready? Okay. Think fast. Whoever posts the answer here the fastest, wins. But also, be honest in saying about how fast it took you to come up with the answer.

Who would my grandfather's boy's son be, who has an uncle that is the married husband of my blood aunt, who is not my brother?
Depending upon the state you live in, it could be an lamb, named Arthur.

[LW]

Originally posted by Soapy Sam
I have a dark suspicion that there are few real-world situations more complex than whether a switch is open or shut, which are actually amenable to logical analysis.

Well, logic is useful in cases where you can abstract the problem suitably. (That is, you can take in all relevant information and still have the theory small enough to allow you reason about it.)

One real-world example where I would use logic would be creating a seating arragement for a largish formal dinner (say, about 100 participants). A nice seating arrangement has the property that everybody has at least a couple of friends in immediate vicinity, sits opposite his or her avec, men and women alternate in such a way that a woman sits in the rightmost seat of the table (to the extent permitted by numbers of men and women), persons who dislike each other are seated far apart, etc.

[LW]

Originally posted by The idea
I suppose one could start with syllogisms,

Well, one could if one likes to feel pain. (If one was a true masochist he or she would start with Hilbert's axioms or some equivalent construct). It is true that syllogisms were the best way to apply logic for about 2000 years. However, that also holds for Roman numerals (and equivalent systems) and mathematics but nobody thinks that they would be a suitable starting point for mathematics anymore.

I would suggest using some puzzles or Raymond Smullyan, perhaps from books What is the name of this book? or The Lady and the Tiger. His puzzles have the nice property that solving them needs progressively more logical abilities. Almost everybody can solve the first few ones but some the advanced ones are really devious. I have to admit that I couldn't solve the one about the "universal question" that can be used to solve any puzzle that involves asking "yes/no" questions in an optimal number of questions.

[Bill Burke]

To: The Idea

You ask a paradoxical question.

If you need to ask for good tests, it suggests you may not have the requisite logic skills to tell a good answer to your question from a bad one.

But if you have such skills, why aren't you able to already see what a good test would be?

How would you resolve this paradox?

Bill

[Iamme]

Jeff, your lamb/Arthur thing flew over me like a crow in flight. Explain. ;)

[Iamme]

Originally posted by LW
Well, one could if one likes to feel pain. (If one was a true masochist he or she would start with Hilbert's axioms or some equivalent construct). It is true that syllogisms were the best way to apply logic for about 2000 years. However, that also holds for Roman numerals (and equivalent systems) and mathematics but nobody thinks that they would be a suitable starting point for mathematics anymore.

I would suggest using some puzzles or Raymond Smullyan, perhaps from books What is the name of this book? or The Lady and the Tiger. His puzzles have the nice property that solving them needs progressively more logical abilities. Almost everybody can solve the first few ones but some the advanced ones are really devious. I have to admit that I couldn't solve the one about the "universal question" that can be used to solve any puzzle that involves asking "yes/no" questions in an optimal number of questions.

---------------------------------------------------------

LW, some dryer, furnace, and other appliance manufacturers use the Yes/no 'puzzle" thing, to help guide you through a quick troubleshooting procedure. This works very nicely.

Example: Power at switch? Yes/no

Yes? Power at door swirtch? Yes/no

No? Bad door switch. Replace.

Yes?...and so it goes. These things are always in easy to read large print and only take up one page and cover everything in just a few short yes/no questions. They cover more ground than you think would be possible by the questions they ask.

[Jeff Corey]

What I hate are the instructions written by Kumar.
"For instructions to format your chip, see page 29.
Page 29, "First, format your chip.."

[Lucky]

From LW:
the one about the "universal question" that can be used to solve any puzzle that involves asking "yes/no" questions in an optimal number of questions[I have several of Smullyan’s books, but don’t recall that problem. What is it, please? Perhaps post it in Puzzles forum?

[LW]

Originally posted by Lucky
I have several of Smullyan’s books, but don’t recall that problem. What is it, please? Perhaps post it in Puzzles forum?

For each liar-truthteller puzzle domains there is possible to form a statement X such that you can combine X with your yes/no question to force correct answers. (Not to spoil too much, I'll use 'X + Q' to denote combining the questions). For example, if you had the standard case of two doors and two guards one of whom lies, you could form the question 'X + "Is the left door safe?"' and the answer would be "Yes" if it was safe and "No" if it wasn't, no matter which guard you asked. (In fact, in this case you wouldn't even know whether the guard lied or not).

Of course, the precise formulation of 'X' depends on the domain, but there is a systematic way of constructing it. The particular problem domain in What is the name of this book? was Transylvania where all people are either humans or vampires and either sane or insane. The idea is that humans tell what they think is the truth while vampires lie, but insane people believe in falsehoods. So, the groups are:

- sane human: tells the truth
- insane human: lies but believes that he tells the truth
- sane vampire: lies
- insane vampire: tells the truth but believes that he lies.

[ingoa]

Originally posted by Soapy Sam
Can anyone demonstrate to me that logic generates right answers. Not true answers, mind you.
Right answers.

I have a dark suspicion that there are few real-world situations more complex than whether a switch is open or shut, which are actually amenable to logical analysis.

For example-
...
Why me?

Why not?!

That was easy. :D

[The idea]

Originally posted by Bill Burke
If you need to ask for good tests, it suggests you may not have the requisite logic skills to tell a good answer to your question from a bad one.

A proposal has to be conceived of before it can be evaluated. I don't see why the fact that I haven't conceived of something would suggest that I am incapable of evaluating it.

Also, since we are talking about testing human beings, to evaluate the quality of a test of logic skills would probably require actual experiments to be performed. A person hasn't yet performed some particular experiments--how does that suggest that the person may not have the requisite logic skills?

Finally, I would point out that there is a difference between a general approach for a test of logic skills and a particular test based on that approach. (Analogy: The Hindenburg disaster doesn't indicate that all lighter-than-air gases are unsafe for getting lift. We have to distinguish between hydrogen and helium.)

[Bill Burke]

Originally posted by The idea
A proposal has to be conceived of before it can be evaluated. I don't see why the fact that I haven't conceived of something would suggest that I am incapable of evaluating it.

Also, since we are talking about testing human beings, to evaluate the quality of a test of logic skills would probably require actual experiments to be performed. A person hasn't yet performed some particular experiments--how does that suggest that the person may not have the requisite logic skills?

Finally, I would point out that there is a difference between a general approach for a test of logic skills and a particular test based on that approach. (Analogy: The Hindenburg disaster doesn't indicate that all lighter-than-air gases are unsafe for getting lift. We have to distinguish between hydrogen and helium.)
------------------------------------------------
Your response is too unclear for me to comment on, perhaps -- in all fairness -- because my comment to which you respond was equally or even more unclear.

Let me start again from scratch. I think it's difficult to test a person fairly and accurately with a smattering of specific tests chosen haphazardly, even if the tests are many and are individually sound.

Therefore the tests would have to be chosen very carefully so that they reflect the logic overall, or they have to test basic, generic skills. Either way, the tester needs a good idea of the fundamentals, I would guess.

But to test a person's fundamental skills at logic, the tester needs to have a clear and correct idea on what fundamental logic skill actually is.

Although I'm not a professional tester or logician, it seems to me that a fundamental test could be based on the conditional logic of the if/then type.

It seems no exaggeration to say that if facts A and B are true, then logically C is also true. As in if 1+1, then 2.

And if we know that 1+1 is the evidence, then we can know that 2 is the answer, if we know how to infer validly. And this is the mental process behind all of our true knowledge. How can it not be?

Furthermore, not just the conclusions are attainable by this basic kind of inference, but even the observed facts are put together this way by assembling their
atomic" parts into a whole bite-sized fact. Think, for example, about the observational bits by which we identify a cat or a dog or Stan Laurel as an actual observed fact.

All fallacies of argument and logic are merely variations on the violation of basic if/then logic.

And a simple test for if/then skills is to examine the opinions of assertions about facts that people make and determine if they violate this logic either by asserting unproven facts or unproven inferences from the facts.

The most common, pandemic violation is to arrive at certain conclusions from uncertain or inconclusive facts.

Of course your views on what basic logic is may be different, but my point is that you may need to have a clear idea of your views in order to carry out tests useful to you.

[Jekyll]

Originally posted by Iamme

Who would my grandfather's boy's son be, who has an uncle that is the married husband of my blood aunt, who is not my brother?
A cousin who's on the side of your family tree where your brother didn't marry that parents sister?:D

Ignoring the ambigious grammer unless you're female this doesnt have enough data. Its either you or the son of a blood uncle.

[Jyera]

Originally posted by The idea
What's a good way to test someone's logic skills? The following approach occurred to me.

Provide a series of alleged proofs, half of which are valid and half of which are invalid. The valid proofs will be as unpersuasive as possible and the invalid proofs will be as persuasive as possible. The person will be required to specify, for each alleged proof, whether it is valid or invalid.

What are the advantages and disadvantages of this approach?

What other approaches might one use? The Advantage and disadvantage, as well as the approaches depends on the practical intend of the test.

Eg. 1. We want to design a test to label a person as "logical" and "not logical", based on the result of the test.

Eg. 2. We want to design a test to measure the logic "skill" level of a a person. (Nil or Basic-Level/Intermediate-Level/Advance-Level)

Eg. 3. We want to design a test to measure the tendency to use the logic-skill. In your case, your might be trying to measure how they react when faced with odd situation.(unpersuasive-but-valid vs persuasive-but-invalid)

The intend helps define the approach, and allow the advantage and disadvantage to be managed for optimal effectiveness.

[Jyera]

Hi The Idea,

While "Label" and "Tendency" are important.
I think "Skill" is what you wanted to measure.

On this basis, I think a syllabus and definition of Basic/Intermediate/Advance-level logical thinking "Skill" is needed.
To measure, create the ruler and the unit of measure.

Short of a formal definition of what is basic/intermediate/Advance,
LW's recommendation Puzzle books with "Progressive" difficulty is an option.

However "Tendency" to be logical is important.
You may have all the skill but somehow do not always use it.

Labelling has it's simplicity, practicality and "unfairness".

But hopefully "labelling" based on a combination of "Skill" and "tendency" will be practically useful.

[skepticality]

Who would my grandfather's boy's son be, who has an uncle that is the married husband of my blood aunt, who is not my brother?

Umm, since there is some issue with structure there...

On the information given.

It would be *you*, I figure.

That took me like... the first time I read it. But I tried to see if I missed something in the grammar a few times. It feels like you missed something or made a mistake in the question for some reason.