Why can't we decrease the total entropy in the universe? I have a question: why can the total entropy in the universe not be reduced? I know that an open system's entropy can decrease, and a closed system goes to the maximum entropy. But for the universe, why can't the total entropy be reduced?
 A: It is important to know that the universe is always expanding, which also results in expansion of energy. Entropy is a measure of how much energy is spread out. Change in entropy $\Delta S$, is measured on the basis of whether the energy is being spread out or being contracted into a small region. If energy is spreading out, then change in entropy is positive $\Delta S > 0$, but if, energy is contracting into a small region, then change in entropy is negative $\Delta S < 0$.
As, I have already mentioned, the universe is always expanding, hence, it's overall energy is also expanding. Hence, due to this reason, the overall change in entropy of the universe is always positive, and cannot be negative.
Also, it is important to consider i)entire universe (isolated system), ii)Observable universe (open system).
The universe is expanding from inside, that way, it's overall boundary is also expanding, hence with the expansion of overall boundary, it's energy is also expanding/spreading out. Even if you consider the entire universe, having no surrounding doesn't impact it's expansion. Expansion will take place nevertheless.
Now, the observable universe (limit to what we humans can see), is just a subset of the entire universe. It is considered as a open system, where there is exchange of  energy and matter involving the boundary of the observable universe. Again, it favours, expansion of the entire universe as a whole, which again leads to increasing entropy. It is possible to reduce the entropy of the universe (as a whole) only if we are able to compress it from outside and reduce the expansion of energy, making it forcibly contained into a small region. But I don't think that is in our limit(currently).
A: How can we tell whether the entropy of the universe $S(t)$ at some time $t$ is greater than or less than $S(t’)$ at some other time $t’$ ? We have to have some record or memory of $S(t)$ available at time $t’$ (or vice versa). But the interpretation of records or memories (indeed, the very assumption that they are accurate records and not just the result of a random fluctuation) depends on the assumption that entropy when the record or memory was made is lower than entropy when the record or memory is retrieved.
In other words, we naively think that entropy is increasing as time passes, whereas we are actually inferring the direction of time from an increase in entropy.
Another way of putting this is that we think we can infer the thermodynamic arrow of time from the causal arrow of time, whereas in fact the two arrows are always correlated.
