"On Cian Dorr's veiw, composition never takes place. There may be partless particles (simples) arranged in shapes of teacups and turkeys, but there are no teacups or tukeys. On David Lewis's view, composition always takes place. so, not only are there teacups and tukeys, but also teacup-turkeys; spatially scattered objects consisting of of one-part dishwater and one part bird." Metaphysics, new essays in ontology page 2 David Manley.
Both these extremes, and views in-between, result from not taking the plain man's view, that reality is independent of humans seriously enough. If we try and take this view seriously; reality can't depend on our descriptions! We are not trying to describe things, but are trying to avoid describing things! (trying not to describe things is not the same as not describing things, but there are strategies we can employ to try and avoid placing on the subject anything that is not there; avoid describing it.) But any 'one part dishwater and one part bird' depends on not taking things as we find them and trying to avoid drawing conclusions beyond them. It depends on linking things that in our experience do not present themselves together let alone as linked. In order to link them we have to suppose that there might be some theoretical point missed by how they are presented, which would be got by linking them. But does that supposed theory depend on us or is it independent of us? If its independent of us it doesn't depend on our descriptions, so it should be arrived at, not by trying to describe the subject, but by trying to avoid describing it!!--Which returns us to how things are presented, at least as our starting point. But also, are tea cups and turkeys conceptually dependent on us? If they are they can't be part of the independent world, but, if we try to avoid drawing conclusions beyond objects and as a result wind up with something indistinguishable from THIS (which is a teacup, or turkey) then those objects are sufficient for this; they can establish the object sufficiently. The conceptual associations of the classification ' a teacup' or 'a turkey' that are 'linguistic' and look beyond the objects existence in itself are another matter, not being judged. The object it self's existence is established sufficiently. This later position seems analogous to the case of being in love or not being in love when we are making a judgement about addition. The experience of being in love is transformed from not being in love, but the addition comes out just the same either way because the transformational aspect is not relevant to what is being judged in the addition. Similarly there may be all sorts of conceptual or poetic dimensions to turkeys or teacups but they are not relevant to what is being judged when we wonder if a turkey or teacup exists.
Quine says we must employ some conceptual scheme or other, we can't stand outside all perceptual schemes. Davidson suggests that the ridiculous position that we can know reality independently of language should always make us reject the premiss, and recognize we can't think without language. But if we try to stick to what is experienced, and try avoiding drawing any conclusions beyond it, as the situation develops, then this doesn't show we have added anything to the situation; instead we've tried to avoid it. So, in as far as we succeed, there is no evidence we are applying 'our conceptual scheme' to the situation. We don't need to apply language in attempting to make these judgments, we attempt to compare the situation at various times and avoid drawing conclusions beyond objects as the situation develops. This doesn't REQUIRE language, even if it may be helped by language. Suppose we employ tally marks to make the comparison, and then transform this procedure by using an ordered system of words instead of the tally marks, we are still not obviously adding anything to the subject by such a procedure to accomplish our comparisons. This seems to illustrate how unclear it is to say we must use language, what do you mean by language? You might employ a system of marks to aid making comparisons and avoid drawing conclusions beyond objects? And similarly what do you mean by thought when you say we can't think without language? We can compare things and realize that e.g. this gear moving like that will produce such and such effect on that situation, it is only logical deduction that makes it seem impossible we can do this directly.
But, to a large extent, we don't succeed; so, on the supposition that the contents of the subject must, somehow or other produce what occurs, and, it seems the only way we can think of they could do that is if they are themselves sufficient to it, which means we can't go beyond them as the subject develops; if we fail to avoid drawing conclusions beyond objects there must be something objectively faulty with our view. This may be true, and so evidence that the objectivity of the view we have arrived at must be faulty somewhere. Still this will not show WHERE it might be faulty and so not entirely objective. So, as far as this abstract argument goes, any particular part of our handling might be truly objective in that it entirely depends on the subject as it is in itself. Our procedure is not to try and add anything to our subject, and to base our understanding on that, our failure doesn't show we are adding anything to any PARTICULAR bit of the subject, it just shows we must be going wrong SOMEWHERE OR OTHER. And then it only shows this on the supposition that the contents of the situation must themselves be sufficient to produce what occurs (which is what we have failed to fully find).
Apart from this, it seems to me that Davidson is in danger of being charged with the same sort of confusion he charges conceptual schemers with when he says we must employ language. There is a difference between language and objects, and between knowing language and knowing our environment, if you say such things as we can't understand or think about our environment without language what environment are you referring to? You have collapsed the difference, so your point can't be made anymore. This seems to go back to the point made in the Tractatus that everything Wit. says is nonsense, but, supposedly, it is like climbing a ladder so that it may be kicked away in the end; this does not save the position but just shows there is something completely wrong with it.