Question about a function of sets

In summary: I don't know how to express this better )I don't really understand what you're saying, but I think the point you're trying to make is that bijections are "maximally non-bijective" or somehow the "opposite" of partial functions. I do not find these characterizations very useful, because they blur the difference between a function and its inverse. But that's just me.What I would say is that the results you need to prove are a measure of how useful the concept of inverse function is. It's a bit like multiplication and division. If you work with numbers, multiplication is totally well defined, while division isn't even defined for many pairs of numbers. Still, as a mathematical
  • #1
ubergewehr273
142
5
Let a function ##f:X \to X## be defined.
Let A and B be sets such that ##A \subseteq X## and ##B \subseteq X##.
Then which of the following are correct ?
a) ##f(A \cup B) = f(A) \cup f(B)##
b) ##f(A \cap B) = f(A) \cap f(B)##
c) ##f^{-1}(A \cup B) = f^{-1}(A) \cup f^{-1}(B)##
d) ##f^{-1}(A \cap B) = f^{-1}(A) \cap f^{-1}(B)##

My attempt-
For option a, let an element ##x \in f(A \cup B)##
##\Leftrightarrow## ##f^{-1}(x) \in A \cup B##
##\Leftrightarrow## ##f^{-1} (x)\in A## or ##f^{-1}(x) \in B##
##\Leftrightarrow## ##x \in f(A)## or ##x \in f(B)##
##\Leftrightarrow## ##x \in f(A) \cup f(B)##
##\Rightarrow## ##f(A \cup B) = f(A) \cup f(B)##

A similar analogy can be applied to options c and d as well.
However, option b doesn't seem to fit into this argument. Even though this approach seems rationale, a counter example can be given to disprove option b. It goes as follows :
Let ##X=\left\{1,2,3 \right\}##, ##A=\left\{1 \right\}## and ##B=\left\{2 \right\}##
Let function ##f## be defined as ##f(1)=3## and ##f(2)=3##
Clearly ##A \cap B = \phi## and hence ##f(A \cap B)## becomes undefined.
Therefore disproving option b.
But option d is correct even though option b is incorrect. Can somebody clarify this for me ?
In simple terms, if option a is right then why not option b (surely there must be some flaw in the above proof when applied for option b but what is it) ? And since option b is incorrect how can option d be correct ?

NOTE: The above question appeared in an exam and the correct answers are options a,c,d.
 
Last edited by a moderator:
Physics news on Phys.org
  • #2
ubergewehr273 said:
Let a function ##f:X \to X## be defined.
Let A and B be sets such that ##A \subseteq X## and ##B \subseteq X##.
Then which of the following are correct ?
a) ##f(A \cup B) = f(A) \cup f(B)##
b) ##f(A \cap B) = f(A) \cap f(B)##
c) ##f^{-1}(A \cup B) = f^{-1}(A) \cup f^{-1}(B)##
d) ##f^{-1}(A \cap B) = f^{-1}(A) \cap f^{-1}(B)##

My attempt-
For option a, let an element ##x \in f(A \cup B)##
##\Leftrightarrow## ##f^{-1}(x) \in A \cup B##
##\Leftrightarrow## ##f^{-1} (x)\in A## or ##f^{-1}(x) \in B##
##\Leftrightarrow## ##x \in f(A)## or ##x \in f(B)##
##\Leftrightarrow## ##x \in f(A) \cup f(B)##
##\Rightarrow## ##f(A \cup B) = f(A) \cup f(B)##

A similar analogy can be applied to options c and d as well.
However, option b doesn't seem to fit into this argument. Even though this approach seems rationale, a counter example can be given to disprove option b. It goes as follows :
Let ##X=\left\{1,2,3 \right\}##, ##A=\left\{1 \right\}## and ##B=\left\{2 \right\}##
Let function ##f## be defined as ##f(1)=3## and ##f(2)=3##
Clearly ##A \cap B = \phi## and hence ##f(A \cap B)## becomes undefined.
Therefore disproving option b.
But option d is correct even though option b is incorrect. Can somebody clarify this for me ?
In simple terms, if option a is right then why not option b (surely there must be some flaw in the above proof when applied for option b but what is it) ? And since option b is incorrect how can option d be correct ?

NOTE: The above question appeared in an exam and the correct answers are options a,c,d.

What you did is correct, besides one detail. ##A \cap B = \emptyset## means that ##f( A \cap B) = \emptyset##. The latter set is not "undefined".

As for why (b) isn't correct and (d) is, this is just because preimage and image are very different concepts. For example, the image of a singelton will be a singelton, while the preimage of a singelton can be the entire domain (if the function is constant).

Preimages behave better than images with respect to set operations, as this exercice shows.
 
  • Like
Likes PeroK
  • #3
If ##f^{-1}## exists, then ##f## (and ##f^{-1}##) must be one-to-one, which rules out your counterexample.

PS except, of course, ##f^{-1}## means a preimage in this context.
 
  • Like
Likes ubergewehr273
  • #4
ubergewehr273 said:
Let a function ##f:X \to X## be defined.

How is the notation "##f:X \to X##" defined in your course materials? Does this signify that ##X## is the domain of ##f## ? - or does it include the possibility that the domain of ##f## is a proper subset of ##X##?

If it includes the possibility that the domain of ##f## can be proper subset of ##X## then can we say an empty set of ordered pairs is a function "##X \to X##" ?
 
  • #5
Stephen Tashi said:
How is the notation "##f:X \to X##" defined in your course materials? Does this signify that ##X## is the domain of ##f## ? - or does it include the possibility that the domain of ##f## is a proper subset of ##X##?

If it includes the possibility that the domain of ##f## can be proper subset of ##X## then can we say an empty set of ordered pairs is a function "##X \to X##" ?

I don't know what the course's definition is, but I would say that ##f## is a function of signature ##X \to X##, then it means that the domain of ##f## is a superset of ##X## (you can further restrict it to say that the domain of ##f## is exactly ##X##, but for most purposes, it doesn't matter whether ##f## is defined for a larger domain that ##X## or not). If you wanted to allow the possibility that ##f## has a domain that may be smaller than ##X##, I would call that a "partial function" on ##X##.
 
  • #6
As a somewhat-trivial result that helps me with these arguments ( hopefully to you too) , all nice properties hold when you have a bijection, so, in a sense, the results you need to prove are somewhat of a measure of how non-bijective a function is.
 
  • Like
Likes ubergewehr273
  • #7
Yeah, I got it.
Thanks
 

FAQ: Question about a function of sets

What is a function of sets?

A function of sets is a mathematical concept that maps elements from one set to another. It assigns each element in the first set to a unique element in the second set, and can be thought of as a rule that relates the two sets.

How is a function of sets different from a relation?

A relation is a broader concept that describes a general association between two sets, whereas a function of sets must follow certain rules, such as having a unique output for each input. In other words, a function is a type of relation, but not all relations are functions.

What is the domain and range of a function of sets?

The domain of a function of sets is the set of all possible inputs, while the range is the set of all possible outputs. The domain and range can be finite or infinite, and they can be the same or different sets.

How is a function of sets represented?

There are several ways to represent a function of sets, including using tables, graphs, or equations. In a table, the inputs and outputs are listed, while a graph shows the relationship between the two sets visually. An equation can also be used to represent a function, where the input is the independent variable and the output is the dependent variable.

What is an inverse function of sets?

An inverse function of sets is a function that "undoes" the original function. In other words, if the original function maps from set A to set B, the inverse function maps from set B to set A. It is denoted by f-1(x) and can be used to solve equations and find missing values.

Back
Top