It's a simple point: if you take a piece of fish and then marinade and bake it, the taste of the resulting meal will be determined not only by the process by which you prepared it (the ingredients in the marinade, the baking), but also by the type of fish you had there to begin with (whether it was cod or halibut or haddock). This simple point, however, is always made light of or obscured by those who engage in the following kind of argument - the argument, in the present case, of Stanley Fish. Evidence, he assures us, is never a plain given, never just raw:
Evidence, understood as something that can be pointed to, is never an independent feature of the world. Rather, evidence comes into view (or doesn't) in the light of assumptions... that produce the field of inquiry in the context of which (and only in the context of which) something can appear as evidence.
Fine. So how we see and understand evidence isn't just by being hit in the eye - as by a ball. We interpret. To do this we have frameworks of meaning, of interpretation, and so forth. But necessary as these are, they don't simply create what we call evidence out of nothing; there has to be some raw material there for them to work on. Fish, swimming one way, seems to grasp this:
[T]here is no such thing as "common observation" or simply reporting the facts. To be sure, there is observation and observation can indeed serve to support or challenge hypotheses. But the act of observing can itself only take place within hypotheses (about the way the world is) that cannot be observation's objects because it is within them that observation and reasoning occur.
Soon enough, however, he gets to swim the other way, as the proponent of this postmodern trope always does. From evidence being something with more than a single determinant, it becomes the creature of the simple evidence-constructing mind.
If there is no thought without constraints (chains) and if the constraints cannot be the object of thought because they mark out the space in which thought will go on, what is noticed and perspicuous will always be a function of what cannot be noticed because it cannot be seen.
No, only partially a function of that, because also still a function of what is 'out there', of which it is (if it indeed is) evidence. Otherwise I might just as well say that the taste of the meal could have been produced without any fish, merely by the process of preparing and cooking it. 'It'? What? Aaah, the fish.
Try this. Someone is throwing me a ball, and I'm attemping to catch it. First, I move my hands on the basis of what I see of the ball's trajectory and speed, anticipating where my hands need to be... NOW. Next, I adopt a framework of interpretation according to which the ball will always travel to the spatial point at which I randomly place my hands. Two different systems of constructing evidence. One of them is better. The example is crude, but more sophisticated ones are available. Evidence isn't a function of interpretative meaning unless it is also (always) a function of something else as well. That at least goes for good evidence. (Via the man in the margins.)