In the bad old days if you as a scientist had something worth saying, a journal would (after vetting through a mainly fair confidential review system) publish it. If you had good things to say, whether or not you had grants, your ideas were heard, and you could make a career on the basis of the depth of your thought, your careful results, and so on.
If you needed funds to do your research, such as to travel or run a laboratory, well, you needed a grant to do your work. This was the system we all knew. You had to have funding, but you couldn't just pay your way through to publishing. Also, if you were junior, start-up funds were typically made available if you needed them, to give you a leg up and a chance to get your career going.
Publishing has always had costs, of course, but the journals survived by library and personal subscriptions, often based on professional society memberships, where the fees were modest, especially for the most junior members.
Now what we have is a large pay-to-play (PTP) industry. Pay-to-play journals are almost synonymous with corruption. The mass of nearly-criminal ones prey on the career fears of desperate students, post-docs, and faculty (especially junior faculty, perhaps). Even the honest PTP journals, of which there are many, essentially prey on investigators, and taxpayers, but the horde of dishonorable ones are no better than highwaymen, robbing the most vulnerable. A story in the NY Times exposes some of the schemes and scams of the dishonorable PTPers. But it doesn't go nearly far enough.
How cruel is this rat race? Where does the PTP money come from?
We have every moral as well as fiscal right to ask where the PTP subscriptions are coming from. Are low-paid, struggling post-docs, students, junior or even more senior faculty members using their own personal funds to keep in the publication score-counting game? How much taxpayer money goes, even via legitimate grants, to these open-source publishers rather than to the research costs for which these grants were intended. In the past, you might have had to pay for color figures, or for reprints, and these costs did come generally from grant funds, but they were not very expensive. And of course grants often pay for faculty salaries (a major corruption of the system that nobody seems able to fix and on which too many depend to criticize).
The idea of open-source journals sounded good, and not like a private-profiteering scam. But too many have turned out to be the latter, chickens laying golden eggs even for the better journals, when there is profit to be made. The original, or at least more publicly proclaimed open-source idea was that even if you couldn't afford a subscription or didn't have access to a university library--especially, for example, if you were in a country with a paucity of science resources--you would have access to the world's top science anyway. But even if the best of the open-source organizations are non-profit, non-predatory PTP operations, and how would we know?, we are clearly preying on the fears of those desperate for careers in heavily oversubscribed, heavily Malthusian overpopulated science industries.
There is no secret about that, but too many depend on the growth model for there to be an easy fix, except the painful one of budget cuts. The system is overloaded and overworked and that suggests that even if everyone were doing his/her best, sloppy or even corrupt work would make it through the minimal PTP quality control sieve. And that makes it easy to see why many may be paying with personal funds or submitting sloppy (or worse) work--and too much of it, too fast.
There isn't any obvious solution in an overheated hyper-competitive system. We do have the web, however, and one might suggest shutting down the PTP industry, or at least somehow closing its predatory members, and using the web to publicize new findings. Perhaps some of the open review sources, like ArXiv, can deal with some of the peer reviewing issues to maintain a quality standard.
Of course, Deans and Chairs would have to actually do the work of evaluating the quality of their faculty members' works (beyond 'impact factors', grant totals, paper counts, and so on) to reward quality of thought rather than any quantity-based measures. That would require the administrators to actually think, know their fields, and take the time to do their jobs. Perhaps that's too much to ask of a system now sometimes proudly proclaiming it's on the 'business model'.
But what we're seeing is what we deserve because we've let it happen.