This to me seems like a really stupid doubt I have but here goes anyway.
Problem:Find the probability of choosing 1 AND 2 from the numbers 1,2
Here, we take 2 possible routes-choosing 1 then 2 OR 2 then 1.
For either route the equation becomes (1/2)*(1/1)=(1/2) However, taking into account both routes, we double to get (1/2)*2=1 which is trivially true.
Take another problem:Find the probability of choosing 2 AND 2 from the numbers 2,2.
Following a similar method to the first problem, we take 2 routes-choosing 2 then 2 OR 2 then 2(because even though they have same magnitude they are still different numbers)
For either route the equation becomes (2/2)*(1/1)=(1)
However, taking into account both routes, we double to get 1*2=2 which is trivially wrong.
Why is this?
You confused the notions and ended up with a conundrum.
Here is a restatement of your second problem to make it clear. I have 2 balls in a bag, both marked as $2$, but one is green and one is yellow. Find the probability of choosing these 2 balls if we draw 2 balls without replacement.
There are 2 paths to drawing the 2 balls: green + yellow or yellow + green, with the probability of taking each path individually at $1/2$ you end up with $$\frac12 + \frac12 = 1$$ as in your previous case.
Here, you differentiate between the two kinds of 2-labels correctly...
An elaboration of your mistake. If you take $2$ and $2$ in the bag to be same magnitude but different numbers (i.e. one is yellow and one is green, as in my example above), your probability of getting the particular path stays $1/2$ as the ball example illustrates -- since you are treating them as different then there are 2 paths available still, and nothing changes from your original setup.
You said they were same in 1 place but treated them different in the end, and you have to be consistent to get reasonable results, you cannot change the model midway...