Hmm, this looks like a pretty misleading case of parallel evolution.
To account for the first lot, the “lacking” ones, is comparatively easy: it’s the familiar privative α that managed to survive virtually unscathed for a respectable number of millennia. There are plenty of occurrences also in everyday speech and not only in the initiatic jargons they probably leaked out of.
As for the second lot, the “being” ones, I’m as puzzled as you: the easy (and probably wrong) hypothesis would assume that an “at” followed by a verb or an adjective were eventually joined into a single word by the vox populi - “at blaze” → “ablaze”, “at float” → “afloat” and so on.
But that’s just a shot in the dark. Would any true etymologist provide something less airy-fairy?