Imagine an exotic "universe", which does not have deterministic laws, but does have a notion of discrete time. At each step in time, the state of the universe — its "material" content — is given by a set of objects. It has no conservation laws as such.
What happens is at each step, the set of material elements is replaced by either the power-set of its previous contents, or by any one of the elements of the power-set. In particular, at any point in time, all objects in the universe might disappear leaving nothing behind, and the universe may stay that way for any length of time, only to eventually produce non-empty sets of potentially exponentially growing size. We may suppose that these are all of the dynamics (can we?): while one can derive empirical probabilities of various transitions from any given history of such a universe, we may suppose that the histories are of a sort that cannot be described well even by randomised theories of finite size, simply by supposing that the dynamics are adversarial to any given ensemble of theories that you might propose.
In this universe, it will happen that something may emerge from nothing, and there are no symmetries which I am aware of to give a name of "potential" to any entity in the description, except the transition rule itself.
Are mechanics a "something"? If so, then there by definition can never be even a seeming 'nothing' according to a mechanistic premise, in which there is at least some intelligibility to the world. Otherwise, something can emerge from nothing, simply because the mechanism may provide a way for this to happen.
Furthermore, reflecting on what I could really mean by a "mechanism" or by "intelligibility", it seems really that we're talking about whether there must be a subject of discourse. Can something emerge from — well, from a state to which I cannot refer, because it cannot be a subject? Or is it the case that by abstraction and indirection that I can always refer to a state of affairs, so that there is always implicitly a 'something'? Or is this just our conceit of the same class as the Ontological Argument: that because we can vaguely imagine something, that it must be a some-thing?
The matter is that we mistake the map for the territory. "Mechanism" is our explanatory tool for what things happen and why, and as with the exotic universe I describe above, we are only ever right to the precision we can see because nature is too lazy to be constantly adversarial. That is our good luck, but also it is an intuition which may turn false at any moment. But even 'subjectifying' it by calling it 'nature' presumes too much. Things merely happen, and we seek the pattern.
Mechanism is not a thing in itself — unless we suppose it to be, but this tells us more about what we mean by 'thing' and our prejudices than it does about the world. Anyone who takes this for granted should be warned about what happened to the notion of 'position', or indeed arguably of 'thing', with the advent of the discovery of quantum mechanical phenomena, ie. complicated happenings the likes of which we previously would not even have expressed with naive conceptions of these words.