Indeed. The horrible phrase is 'copyright transfer agreement', which is typically a condition of being published. My experience is that the better journals are more of an arse about copyright. I've just written two book chapters and spent maybe a day getting permissions to reuse published figures, sometimes including my own (!) in other works.
On the other hand, publishers have very little power when it comes to enforcing their copyrights when the author of an article decides to make it available on their personal web page. I have never met a researcher who refused to provide a copy of a paper when asked, despite technically violating a copyright by doing so. Academic publishers know that if they start threatening researchers they are playing with fire -- their business model is already obsolete and is only kept alive by institutional inertia, and the last thing they want is for researchers to find the motivation needed to ditch the publishing companies (the technology is widely available, now it is just a problem of politics and of organizing a community to change).