The only value a scientist gets from publishing a journal article is another publication they can add to their CV. A long list of publications is important for any scientist who wants to get funding, and even grad students need to have that long list of publications on their CV if they hope to get a research job later in life. What the journal provides is its name, and nothing else -- because as I said, everything else that goes into a journal is done by volunteers.

Now, once upon a time, journals served a secondary purpose, which was to distribute scientific results on a global scale. Back then, there was no Internet, and printing enough copies of a journal to satisfy the world's needs required industrial equipment. Fortunately those days are done and over with, but unfortunately we are still dealing with the relics of that bygone age in the form of academic publishing companies. Worse, in fact, since today's publishers are far greedier and for more profiteering than many of the publishers of the past.

Really, were it not for copyright, we could cut publishers out of the equation without any ill effects and have a net gain for society by removing all paywalls from scientific articles.

See my post (GP of yours). What scientists get is the prestige of having an accepted article in one of the better journals. The more prestigious journals can then trade on that prestige to force concessions that hurt everyone but themselves and undermine the whole public-funding, knowledge-sharing system. The scientists do it because, often, there is tenure and/or promotion riding on it. Plenty of scientists want open access, but few want to risk their careers to push for it.