You cannot assume a priori that
- moral properties are basic and so cannot be reduced to anything else ( lets call this "basicness")
- moral properties are non-natural and so cannot be reduced to natural properties (lets call this "non-naturalness")
- moral properties cannot be reduced to other natural properties
Well basicness and non-naturalnes might be explanations or reasons of why supposedly all candidates fail the Open Question argument - but only if this argument works, they cannot be assumed, that would be question-begging.
This leaves (3). However (3) highlights the problem that reduction of moral terms is somehow different or special from anywhere else - there is no Open Question argument anywhere else, what is the difference that makes it apply here - what makes it special - in line with the labelling of (1) as basicness and (2) as non-naturalness we can call this "specialness". However the Open Question argument cannot assume this either without again being question-begging.
So what remains of the Open Question argument if one cannot show nor assume basicness, non-naturalness or specialness which are all instead candidate explanations of why the argument succeeds - if it does. We still need to know if the argument - in principle - can work.
All that is left is an appeal to intuition, that there that there are two different "things" not one. However this can be explained by what is technically called "referential opacity" - it is a mistake in the way we think about or perceive these "things", referential opacity leads us to think there are two when there are in fact one. For example we used to think that the Evening Star and the Morning Star were two different stars, in fact they are not, indeed they are not two stars but one planet - Venus, to make such a mistake is to suffer from referential opacity in that regard. How do we know we are not here with respect to moral value?
The irony is that Moore's version was meant to be an argument for moral intuitionism. So because our intuition tells us that moral properties are different from others we must and can only use intuition to discern moral properties! Yes another piece of circular reasoning! (Of course many today use Moore's argument not arguing for moral intuition but usually some form of subjectivism, but the underlying point below still stands)
It is important to note also that the fact that it might succeed for many candidates does not mean it is a good argument, since it is quite possible to use fallacious reasoning to arrive at a correct conclusion. However one needs some independent means for showing this is a correct conclusion and a fallacious argument alone would not only be an unreliable means it would also mean one could not know that the answer is correct, seeking explanations for such conclusions would be highly dubious as one does not know the accuracy or reliably of those conclusions. In other words, the fact that many candidates are caught by the Open Question argument and we all agree that they are incorrect does not support the Open Question argument as a valuable method, since it would reject all candidates correct or not. For any failed candidates one should be able to use other methods to show why they fail.
So this all this seems to be an elaborate way to cover up that this is an Argument from Intuition. Of course, it is this very intuition and related common-sense morality that is challenged by various reductive models such as most types of utilitarianism. So one cannot use this argument to dismiss those reductive models without question-begging yet again!
To sum up Moore's Open Question Argument and his three types of naturalistic fallacies is window dressing - rhetoric - to cover up that one is making an Argument from Intuition - that one is saying that if any theory conflicts with one's intuition it must be wrong, that one's intuition trumps any empirical technique!