The Sufficiency of Relaxed, Low Key, Natural Ontology

I have been reading Sellars' Naturalism and Ontology. The discussion veers between the direct and the oblique. This leaves me wondering whether a direct Wittgensteinian approach would be sufficient. Consider the ontological confusions concerning attributes.
How are we to explain "Yes, there are attributes" when it is voiced in response to a question like "Are there attributes?" (where the person asking, 99.999% of the time it's going to be a philosopher, is fishing for possible commitments to abstract objects or some such)? Can we not simply respond: "If we are able to attach a practical sense to "there are attributes" (meaning we know what to do with the phrase - can fit it into intelligible contexts, put it to communicative use, and so forth), then we need not invoke ontological talk and consequent puzzles about what it is for an attribute to exist, etc."?
Isn't the temptation to resist this sub-philosophical approach one that conflates reluctance to indulge in certain imaginative extrapolations (e.g. one imagines attributes floating around in metaphysical space waiting to get hooked up to something appropriate) with a supposed limitation of intelligence, whereas, clearly, intelligent insight is required to see that such imaginative extravagances are redundant - rather as one realises that angels are not real beings and then automatically feels relieved of any need to imagine what they look like, where they reside, and so on?
But then, can we not apply this to "there are infinities"? How does that square with my previous posting? Is it simply that we can again talk about infinities without dragging in the imagined realist underpinnings?