John Searle endorses "external realism", the view that there exists a mind-independent reality, in spite of the fact that our access to that reality is necessarily mediated by conscious experience.
The thesis of external realism is in some sense not a theory at all, but rather a presupposition of our being able to have theories in the first place.
One can show that this or that claim corresponds or fails to correspond to how things really are in the 'external world,' but one cannot in that way show that the claim that there is an external world corresponds to how things are in the external world, because any question of corresponding or failing to correspond to the external world already presupposes the existence of an external world to which the claim corresponds or fails to correspond. External realism is thus not a thesis nor an hypothesis but the condition of having certain sorts of theses or hypotheses (Searle, 1995, 178).
External realism isn't a thesis or representation but rather a "background condition" for having a thesis in the first place. Why? The idea seems to be that very idea of truth or falsity seems to imply a mind-independent reality.
Rather than evaluate the merits of Searle's argument, I will rather presuppose my own version of the distinction between theories and background conditions.
Following Bas van Fraassen, explanations are answers to why-questions. Theories help us adjudicate plausible answers to why-questions from implausible answers. In a Sherlock Holmes story, a prized racing horse has been stolen. Why is the horse missing? Answer: The trainer stole the horse. Holmes is able to infer this from "the curious incident of the dog in the night-time"— the dog did not bark. The argument in favor of the explanation, of course, rides on a tacit theory: namely, that guard-dogs bark at strangers. This theory helps the trainer-story explain the missing horse. Notice that if dogs were more like cats in this respect, Holmes inference would be bunk.
I want to suggest that external realism is a background condition rather than a thesis because there is no why-question it can be called on to help answer. We cannot exhaustively specify the background understanding that goes into the evaluation of even simple why-questions, such as the one Holmes is addressing. Reasons for or against one a rival explanation can thus only be expected to supplement someone's understanding just enough to get them on board with the speaker. It's difficult to imagine a situation where pointing out the existence of a mind-independent reality would be sufficient bridge the gap in understanding which separates speaker and interlocutor; indeed, in a Monty Python parody of investigation, Holmes' pointing out the existence of a mind-independent reality could only serve to underscore the unbridgability of the interlocutor's world-views.
My claim is that theories have to be able to make a difference to the selection of rival possible explanations. Background conditions—including the claim there exists a mind-independent reality—are factors which do not make a difference to the adjudication of rival answers to a request-specification. These conditions change from explanatory puzzle to explanatory puzzle, and depending on who the audience is, but there certain things that will almost never be helpfully said—unless one is wishing to end the conversation.