1

Suppose that we have a Group $(G,.)$, where . is binary operation on $G$. Prove that for any $a_1,a_2,\cdots, a_n\in G, a_1.a_2.a_3.\cdots a_n$ is independent of how it is bracketed.

My attempt: First of all for any $k\in \mathbb N,a_1.a_2.a_3.\cdots a_k$ must be defined in a way. Let's define it like this
$a_1.a_2.a_3.\cdots a_k=a_1.(a_2.(a_3.(\cdots a_k))\cdots) \tag 1$

Now we have to prove that howsoever $a_1.a_2.a_3.\cdots a_n$ is bracketed, it evaluates to form $(1)$, that is $a_1.a_2.a_3.\cdots a_n$ has a unique value which can be given by the form $(1)$.

For $n=1$, $a_1=a_1.e=e.a_1$, where $e$ is identity of the group is a unique element in $G$ as $.$ being a binary operation associates $a_1=a_1.e$ to a unique element in $G$.
For $n=2, a_1.a_2$ is a unique element in $G$ as $.$ is a binary operation on $G$.
For $n=3$, the hypothesis is true by associativity of Group.

Induction hypothesis: Let our hypothesis be true for $r\le n-1$, that is $a_1.a_2.a_3.\cdots a_{r}$ is independent of how it is bracketed and has a unique value which can be given by the form $(1)$.

Now the problem that I'm facing is lies in breaking $a_1.a_2.a_3.\cdots a_{r+1}$ into two parts. I assumed $1\lt i\lt r$ and considered $a_{i+1}.a_{i+2}.a_{i+3}.\cdots a_{r+1}$. As per our definition above,
$a_{i+1}.a_{i+2}.a_{i+3}.\cdots a_{r+1}=a_{i+1}.(a_{i+2}.(a_{i+3}.(\cdots (a_r.a_{r+1}))\cdots )$. RHS is composition of two elements of $G$ under binary operation $.$ and hence unique.
Now $a_1.a_2.a_3.\cdots a_{i}=(a_1.a_2.a_3.\cdots a_{i})$
We have $a_1.a_2.a_3.\cdots a_{i}.(a_{i+1}.a_{i+2}.a_{i+3}.\cdots a_{r+1})=a_1.a_2.a_3.\cdots a_{i}.(a_{i+1}.(a_{i+2}.(a_{i+3}.(\cdots (a_r.a_{r+1}))\cdots )$. Now I'm stuck.

Please see bracket on LHS here, I don't understand how to remove it and once this is removed how do I show that finally the product (compositio under binary operation $.$) evaluates to a unique value given by the form $(1)$.

Please help. Thanks.

Arturo Magidin
  • 398,050
Koro
  • 11,402
  • 1
    Does this help? I don’t want to mark as duplicate, because that will close the question. By the way: you aren’t asking for “solution verification”, since you are not presenting a full solution. The tag is inappropriate. – Arturo Magidin Mar 23 '21 at 02:23
  • @ArturoMagidin: Thanks a lot. I'll have a look at the link. – Koro Mar 23 '21 at 02:25
  • If the link does do the trick, let me know and I’ll vote this as a duplicate; but because I have a group-theory gold badge, that vote by itself closes the question. That’s why I’m holding off. – Arturo Magidin Mar 23 '21 at 02:28
  • @ArturoMagidin: Sure! I'll let you know. – Koro Mar 23 '21 at 02:29
  • @ArturoMagidin: I have gone through your answer in the link. My question still remains. The question is: when induction hypothesis was made for $ r\lt n$, it was applicable to $a_1,a_2,\cdots, a_r$ (starting from subscript $1$ and ending at $r$). But in your answer and in other proofs as well, it is considered that since $w_2(a_{k+1}...a_n)$ (terminology from your answer) is a product of "less than $n$" elements , by induction hypothesis it can be bracketed as we please. Here subscript starts at $k+1$ so my question is how induction hypothesis is applicable here? – Koro Mar 23 '21 at 03:43
  • @ArturoMagidin: If that is clear to me, then my question is solved. – Koro Mar 23 '21 at 03:45
  • @ArturoMagidin: If instead I choose my induction hypothesis to be suppose for any $r\le n-1$ , it is true that $r$ nos. elements of $G$ can be bracketed the way we please, then also my question is solved. But in your answer, you have clearly mentioned that $a_1,a_2,\cdots, a_n$ are in order. Please clarify that. – Koro Mar 23 '21 at 04:04
  • 1
    Your question makes absolutely no sense to me, and I suspect it rises from being careless about setting up both your induction hypothesis and what things mean. When you start, you need to assume you have some bracketing. And you are trying to show that bracketing is equal to the “standard” one. And what is the complaint about “order”? We are proving associativity, not commutativity. The order of the factors is fixed, it’s the way we put in the parentheses that changes. – Arturo Magidin Mar 23 '21 at 10:25
  • 1
    The real induction hypothesis is not “the first $r$ can be bracketed any way you want”, the correct induction hypothesis is “any product of $r$ elements can be bracketed any way you want.” If you make too weak an induction hypothesis, then of course you can’t conclude what you want. If that is the problem, then it’s because your approach is flawed. And if that is indeed the problem, then your question should be closed. – Arturo Magidin Mar 23 '21 at 10:27
  • @ArturoMagidin: Any product of $r$ elements! and not necessarily the first $r$ elements.Yes. That answers my question. Thanks a lot. – Koro Mar 23 '21 at 10:29

0 Answers0