Professional Documents
Culture Documents
Introduction To Abstract Mathematics MA 103
Introduction To Abstract Mathematics MA 103
MA 103
Solutions to exercises 2
1
(a) (i)
(a) Your answers here must engage with the definition of the set X. Im not just looking for
a proper description of what it means for A to be an element of a general set X; I want
to know what it means for A to be a subset of the specific set X defined in the question.
All you have to do is read the definition of X:
To say that A is an element of the set X = { A | A { 0, 1 } } means that A { 0, 1 }.
It really is that simple, but its worth pausing to digest this. The elements of the set X
are themselves sets: specifically, they are subsets of the two-element set {0, 1}.
To say that B is a subset of X means that B is a set, all of whose elements are elements of
X, i.e., B is a subset of X if all the elements of B are subsets of {0, 1}.
(b) If youve understood the answer to (a), then you can now list the elements of the set X.
These are: , {0}, {1} and {0, 1} itself. In particular, {0} X.
Alternatively, to test whether {0} is a member of X, we need to decide whether the
statement {0} { 0, 1 } is true or not. It is, and so {0} is an element of X.
(c) The set B = {0} is a subset of X if every element of B is an element of X. As B has only
one element, there is only one thing to check: we have to ask whether 0 is an element of
X. It isnt: the elements of X are sets. So B is not a subset of X.
(d) We listed the four elements of X: , {0}, {1} and {0, 1}. Weve seen that {0} is not
a subset of X, and similarly {1} and {0, 1} are not subsets of X. The only remaining
possibility is , and this is a subset of X is a subset of any set as well as an element
of X.
n N, m N, n m.
(b) A counterexample for the first statement is x = 1. For this x, we have x Z, but
2x = 2, and there is no z Z so that 1 z 2.
(c) The second statement is true. The natural number n = 1 does have the property that,
for all natural numbers m, 1 m. ( The statement says that there is a natural number
n with a certain property. To prove it, you have to name a natural number n, namely
n = 1 in this case, and argue that it has the required property. )
(d) The negation of a statement of the form for all elements in some set, property A holds,
is the statement there is an element in the set for which property A does not hold . So
the negation of the first original statement is
There is an integer x such that
it is not the case that there is an integer z with x z 2x.
()
The second line is the negation of there is an integer z with x z 2x. The negation
of a statement of the form there is an element in some set such that property A holds
is the statement for all elements in the set, property A does not hold . So the negation
of there is an integer z with x z 2x is the statement for all integers z, it is not the
case that x z 2x. So we can rewrite the negation in () and obtain
There is an integer x such that, for all integers z, it is not the case that x z 2x.
Now the final part of the previous statement is the negation of x z 2x. Note
that this is in fact a combination of two statements : x z and z 2x. And such an
and-statement is false if at least one of the two parts is false. So the negation of x z
and z 2x is the statement x > z or z > 2x.
So as the final version of the negation of the original first statement we might write :
There is an integer x such that, for all integers z, either x > z or z > 2x.
In mathematical notation we can write this as :
B = { b, c, d, f }
and
C = { d, e, f , g },
The shaded areas are not the same, so you strongly suspect that the sets they represent are
not the same. But these pictures do not constitute a proof ! For that you need to construct an
explicit counterexample. Of course, the pictures can be very helpful for that.
The counterexample at the beginning was obtained by making sure that each of the areas in
the Venn diagram contains one element.
However, many other answers are possible. For instance, looking back at the Venn diagrams
for A ( B C ) and ( A B) C, we see that the subset of the set A that consists of elements
in A but not in B or C ( the top-left part of the diagram ) is a subset of the set A ( B C ), but
not of ( A B) C. So, provided our example includes an element that is a member of A but
not of B or C, we should expect to have a counterexample.
For instance, set A = {1}, B = and C = . Then we have
A ( B C ) = A = A = {1};
( A B) C = A = .
Very important: In this question, you are asked to prove certain things for all sets A, B and
C. If you start your answer with: Let A = {1, 4, 5}, B = {2, 3, 4}, C = {1, 2, 6} (or any other
sets that you happen to have chosen), and then work out A ( B C ) and ( A B) ( A C )
for this example, finding them to be equal, then all you will have proved is that these three
sets do not form a counterexample. You can never prove a statement about all sets by looking
at an example. Of course, you might be able to disprove such a statement by giving one
counterexample, as in the previous question.
(a) In order to prove that, for two sets X, Y, we have X Y, we need to prove that every
element in X is also a member of Y. So we proceed as follows :
To prove :
A ( B C ) ( A B ) ( A C ).
A ( B C ) = ( A B) ( A C )
( A B ) ( A C ) A ( B C ).
Hence to prove : for all x ( A B) ( A C ), we also have x A ( B C ).
Proof : Take any x ( A B) ( A C ). That means that x A B or x A C.
In the case that we have x A C, then we have x A and x C. So in particular we
have x A and x B C, hence x A ( B C ).
If we are in the case that x A C, then x A and x C. Again we find x A and
x B C, hence x A ( B C ).
So, in both cases, we have x A ( B C ), as was required.
2
For this kind of question, for which at first you even might have trouble understanding what
is going on, the first rule is dont panic. Take a pen and an empty piece of paper. And read
carefully the information in the question to see what is going on.
Often the best way to get a feeling for what is happening, and what kind of things you are
supposed to be looking at, is by making up an example. And in fact, parts (a) and (b) ask
you to do just that. Quite often such a requirement to write down some examples is not
part of the question, but it is a good way ( if not the only way ) to start thinking about the
problem.
So lets start by doing (a) first.
(a) The question in (a) asks you to write down a specific example of the type of object we
are looking at ( a 5 5 array of the numbers from 1 to 25 ). Following the question, we
are supposed to write down
1 2 3 4
6 7 8 9
11 12 13 14
16 17 18 19
21 22 23 24
5
10
15
20
25
Then you have to find the greatest number in each row. For the array above that gives
the sequence 5, 10, 15, 20, 25. And then the least number of those is s; so we get s = 5 for
the array above.
Next you are required to find the least number in each column, which leads to the sequence 1, 2, 3, 4, 5. And from these you pick the greatest one, leading to t = 5.
So, as was predicted, the example array above is one in which s = t.
At this point, if you obtained s and t that were not the same, you should not just continue
and ignore that. Clearly, in that case, something is going wrong : either there is an error in the
question ( not impossible, but not that likely ), or you have made a mistake ( unfortunately
for you, that is much more likely ). So carefully read the question again, did you take the
right example, did you obtain s and t in the right way, etc.? It doesnt make much sense to
continue with parts (b) and (c) until youve found out ( and corrected ) your error.
OK, suppose you did part (a) as above. If there was no part (b), we would be tempted to
think that for all arrays of the type we are looking at we would get s = t. So if at this point
you were asked to prove a relation between s and t for any array, you would be tempted to
try to build a proof for s = t. And you would fail.
One reason you should be suspicious about the possible truth of the fact that s = t for
all arrays, is that it is based on one very particular example ( namely the very regular one
above ). To get somewhat more evidence, you should consider examples without such a nice
structure and see if you still find s = t. If that would always be the case, then that would
strongly suggest that maybe it is true that s = t for all arrays.
But, alas, it will appear that something different is the case.
(b) In this part you are asked to find an example of an array with a specific property ( s 6= t ).
The one example we saw so far ( in (a) ) didnt have this property, so we need to do
something different. And the question gives no further hints what things to look for.
Now there are a couple of things you can do. One is to see how to make some small
changes to the array you already found. Or just start with a completely new example.
You can try the first thing for yourself ( its not too hard ); I decided to just make a completely new example. This time I tried to avoid making it too organised, and randomly
filled in the numbers to get:
1 13 21 5
20 6 2 15
17 3 12 11
7 18 4 8
25 9 10 23
24
14
16
19
22
For this example the greatest numbers in the rows give the sequence 24, 20, 17, 19, 25.
This gives for the least one s = 17. And the smallest numbers in the columns are
1, 3, 2, 5, 14. The greatest of these is t = 14. So indeed this second example satisfies
s 6= t.
( In fact, the first example was one of the very few unlucky ones in which s = t. Almost
every way you make this array you will get s 6= t. )
Note that in the first example above we have s = t and in the second example we have s > t.
So it seems that indeed s t might be true for any array found this way.
But examples are not proof.
For a proof ( as required in part (c) ) we need to make sure that s t is true for any example
we could make up. In principle, we could investigate all the possible arrays, but there are
far too many of them to contemplate this approach.
OK, so how do we get a proof ? We are asked to prove something about two numbers s and t.
So its a good idea to find out as much as possible about these two numbers.
First lets see how s is obtained from a given array. We first find the greatest number in each
row, and then s is the least among those. But that still means that s is the greatest value in a
certain row. In other words, there is a row in the array so that for each number a in that row
we have s a.
Is that all we can deduce about s ? No, s is the smallest among the greatest. So it follows that
each row that is not the special row from above contains a number b so that s b.
That seems to be about all we can say about s.
We can reason similarly for the number t. This leads to the observation that there is a column
in the array so that for each number x in that column we have t x.
And each column which is not this special column contains a number y so that t y.
So now we should see if this gets us any closer to a proof of s t.
Looking back at what we know about s and t, and in view of having to prove s t, the
two observations involving s a ( for certain a ) and t x ( for certain x ) seem to be the
most useful. If only we could find a and x so that a = x, or even a x, we would be done.
But in fact we can do just that. The statement s a holds for all a in a certain row. And
the statement t x holds for all x in a certain column. But whatever row and column we
take, there is always one number which is both in that row and that column. So let k be that
number for the special row and column related to s and t. Then we know that s k and
t k. In other words, s k t, which immediately gives s t.
And now the real answer
The long story above is not the answer to this question. Its a long description of how to
work on this kind of problem, including things that you only think but not write, and things
you only would write on your rough work but not in the final answer. Below is more or less
what we would expect to see as a real ( and correct ) answer to part (c).
(c) To prove : For all arrays we have s t.
Proof : The number s is obtained by taking first the greatest number in each row, and
then taking the least value among those. So there is a row for which s is the greatest
value. Call this row A. So for all a in row A we have s a.
A similar argument gives that there is a column B so that for all numbers x in column B
we have t x.
Now let k be the number written at the intersection of row A and column B. Then k is
in row A, hence s k; and k is in column B, hence t k. From these two inequalities
we can conclude s k t, and so s t.
2
This was a very challenging problem, and if you solved it then you deserve congratulations!