Right. I want to judge the validity of a paper on the Higgs boson, so I rent time on the SSC to reproduce the experiment. Everyone else who wants to judge does the same thing. Seems like a good use of limited resources. Can you find me a funding agency that will pay for this?
Boy, aren't we exaggerating.
Before thinking about purchasing a particle accelerator, you have a considerable number of things which you can and must actually do by yourself in order to test the paper's validity. One of those things is actually reading the paper, understanding the theoretical hypothesis which were laid out, analyse the data which was used as a basis for the results presented in the paper, check if it holds out, evaluate the results... You know, the peer review process.
In this context, the need for a particle accelerator only enters the equation if you suspect that the results presented in the paper aren't up to par, and you wish to replicate them to see if you aren't being duped. Even in that case, you still need to run the series of tests which I pointed out.
So, if you actually intend to judge the validity of a paper on the Higgs boson then it's safe to say that purchasing a particle accelerator is the least of your concerns. In fact, how many of those actually involved in reviewing that sort of papers have access to their own personal particle accelerator? And does that stop them from doing it? Precisely.
Peer review puts this work in the hands of a few people who are allegedly experts in the field, and their job is to judge the validity of the paper, not necessarily the results of the experiment that it may be reporting on. Was the scientific process followed? Were there controls where necessary? Does the data support the conclusion, whatever it may be? Is the data presented in a logical and reasonable manner? Are the assumptions underlying the paper reasonable? Is there some glaring error of omission or execution? Is the material itself publishable? Is it fresh and new, or simply reworked decades old textbook material? Are there proper citations for previous work, or previous work that should have been cited but was not?
Notice that you actually don't need access to a particle accelerator to do any of those things. You actually only need to have an academic interest in the subject. If you are interested in scientific problems such as the validity of the Higgs boson and you are curious enough to be willing to spend your time rummaging through articles on the Higgs boson then odds are to actually know something about it, and you are actually in a position to judge by yourself at least a portion of those details.
So, unless you intend to avoid having to think about the stuff and instead you want an authority to tell you what you must believe in, whether a paper is published in a journal or is distributed directly from the site of a university or research institution is actually irrelevant, because you are quite able and willing to turn on your brain and actually do science.
You forget, the readers may not be experts in the field. They may be expanding their horizons or looking for new research questions, and expecting every one of them to "test the explanations and predictions" for themselves is silly. Expecting them to know that Smith and Wesson in 1975 did a similar experiment and came up with similar results but a different conclusion, and that the paper they are reading is incomplete because it did not discuss that experiment, is outrageous.
You aren't required to be an expert in the field. If you are interested enough on a given field to actually read the papers then you are certainly already knowledgeable about the subject. Moreover, you certainly are already aware that a paper is supposed to be food for thought, not a textbook.
... because organizations such as universities and research institutions are more than willing to put their logo on the cover of their member's papers,
And this serves the function of peer review and validation how, precisely?
It isn't intended to replace the peer review process. If all a person wants in journals is that they serve as an authoritative seal of approval which gives him enough confidence to place blind faith on a paper then the logo of a big name institution actually does the same thing. Let's not pretend that that doesn't happen right now with how the logo of companies such as elsevier is perceived and used.
The fact of the matter is this: if someone actually reads papers of a given field and actually cares about how the peer review process works in that field then that person is already quite able to digest that information. If, instead, someone only wants to have someone tell him what to believe, then that's an entirely different problem which the existence of journals does nothing to mitigate.