[net.math] a logic brain teaser

joseph@hp-pcd.UUCP (joseph) (08/14/84)

  This is a VERY well known puzzle. I didn't immediately post the 
  solution that I knew, for the benefit of readers who may have
  the satisfaction of seeing it the first time. There was a reason
  for the poster asking for personal mail. The above response
  just spoiled it.

james@ur-laser.uucp (James Chavin) (08/20/84)

    You are Dorothy on the way to the Wonderful Wizard of Oz.  One day, while 
gaily skipping along the yellow brick road, you come to fork, one path leading 
north and another one west.  Unfortunately, you don't know which path leads to 
Oz and which path leads to /dev/null. 
    Between the fork lies a quaint old cottage.  In this cottage there live two
identical witches.  One witch always tells the truth while the other witch
never tells the truth.  Both witches know the way to Oz, and have graciously
granted you one question to ask one of them.  You must consider your choice of 
question carefully, for there must be no doubt as to the correct way to Oz.  A 
question such as 'Which way to Oz?' is obviously useless, as you will not know 
whether you asked the truthful or the deceitful witch.  The problem is not to 
guess the correct way, but to ask the correct question so that regardless of
which hag is asked, the same path would be indicated.


    Mail your questions to James Chavin
                           {allegra,seismo} !rochester!ur-laser!james
                           Gary Krakower
  
    We will reply promptly, and post the correct question to the net Monday, 
August 27.

    Have fun!!!!!!!!!!!!!
 
--------------------------------------------------------------------------------

    X-rays let you see your soul 

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

keith@seismo.UUCP (Keith Bostic) (08/21/84)

If my next question was "Is this the way to Oz", would you answer "yes"?

Q1: Would you answer "yes" to Q2?    (The answer "no" works just as well,
                                      you just have to reverse the meaning.)
        It is         It isn't       Liar:     always lies
Liar:    YES            NO           TT:       always tells the truth
                                     It is:    it's the way to Oz
TT:      YES            NO           It isn't: it isn't the way to Oz


Q2: Is this the way to Oz?

        It is         It isn't
Liar:    NO             YES

TT:      YES            NO

                Keith 
                        ARPA: keith@seismo 
                        UUCP: seismo!keith

kaufman@uiucdcs.UUCP (08/23/84)

#R:ur-laser:-21900:uiucdcs:28200042:000:128
uiucdcs!kaufman    Aug 23 15:32:00 1984

Another possible solution:

"Which way would the other witch say led to dev/null?"  The direction indicated
would be toward Oz.

ken@ihuxq.UUCP (ken perlow) (08/25/84)

--
Martin Gardner hacked this 2 roads, 2 persons (1 liar) problem
to death years ago in one of his puzzle books.  The usual answer
given is correct IF the liar is a logical liar, that is, tells
logically false statements, as opposed to a deceitful person.
The devious liar will see through the awkwardly loaded question
and deliberately mislead you.  Gardner's best question for the more
realistic type liar was something like, "I hear there's free beer
at <correct destination>.  Shall we go?"
-- 
                    *** ***
JE MAINTIENDRAI   ***** *****
                 ****** ******    24 Aug 84 [7 Fructidor An CXCII]
ken perlow       *****   *****
(312)979-7261     ** ** ** **
..ihnp4!ihuxq!ken   *** ***

markn@ios.UUCP (Mark Nudelman) (08/27/84)

>Another possible solution:
>
>"Which way would the other witch say led to dev/null?"  The direction indicated
>would be toward Oz.

Not true.  If "dev/null" means a place which does not exist, the
question "Which way leads to dev/null?" would not be answerable
by the truthful witch; the lying witch could answer with either
direction.  So the proposed question would not get you any information.

Raymond Smullyan has a delightful book full of variations and elaborations
on logic problems of this type, called "What is the Name of This Book?"
I recommend it to anyone interested in logic puzzles.  It has more
variants on the old liar/truth-teller puzzle than you would believe.
At the end he constructs a puzzle whose solution is nearly isomorphic
to the proof of Godel's incompleteness theorem.

	Mark Nudelman
	..decwrl!ios!markn