Wrong. Copyright is widely misunderstood. Particularly the role of copyright in science publishing. First of all, there is this idea that some journals and publishers don't require copyright transfer, but 'just' the exclusive dissemination and exploitation rights. To all practical intents and purposes, that is exactly the same, and 'copyright' is just shorthand for 'exclusive dissemination and exploitation rights'! So if it helps to drop the word 'copyright' then that should, and easily can, be done.
Secondly, transfer of exclusive rights to a publisher is a form of 'payment'. Payment for the services of a publisher. The publisher subsequently uses these exclusive rights to sell subscriptions and licences in order to recoup his costs, in a rather roundabout way. This form of payment – as opposed to cash – has advantages and disadvantages. The advantage is seemingly for the author, who (mistakenly) has the feeling that he doesn't have to pay for the services of formal publication of his article, but who seldom realises why he is asked to transfer exclusive rights. The disadvantage is that payment in the form of exclusive rights limits access, because it needs a subscription/licence model to convert this form of 'payment' into money. And subscriptions/licences are by definition restrictive in terms of dissemination. Article fee supported open access publishing, where the transfer of exclusive rights is replaced by the transfer of money, consequently doesn't have the need for subscriptions and can therefore abolish all restrictions on dissemination.
Stevan Harnad c.s. will argue that none of this matters, because there is 'green', meaning that whatever 'exclusive' rights have been transferred, authors can still disseminate their articles via self-archiving in open repositories. In that model, having transferred 'exclusive' rights is meaningless, and that implies that the 'payment' that exclusive rights transfer actually is, has become worthless. In mandates with embargos, the 'payment' may not be completely worthless (depending on the length of the embargo) but is at least severely devalued.
I am a great fan of open access, but not a great fan of 'green'. 'Green' is a kind of appeasement by publishers (some of who, it must be said, themselves didn't – sometimes still don't – realise the 'payment' nature of exclusive rights transfer). Appeasement is often regretted with hindsight. Instead of allowing the nature of exclusive rights transfer to be compromised, publishers should much earlier have offered authors the choice of payment – either transfer of exclusive rights, or cash. The appeasement, the 'green', now acts as a hurdle to structural open access, perhaps even an impediment.
Harnadian orthodoxy will dismiss this. It holds that subscription journals will survive, that they will be paid for by librarians even if the content is freely disseminated in parallel via open repositories, and that it doesn't matter anyway (the guru is tentatively beginning to admit that large scale uptake of self-archiving, for instance as the result of mandates, may indeed destroy journals) because a new order will only come about after the complete destruction of the old order. After all, morphing the old order into the new, without complete destruction, entails a cost in terms of money, which "isn't there", and anyway, the cost that comes with complete destruction of the old order is preferred to spending money on any transition, in that school of thought.
I doubt that a complete wipe-out will come. But there are quite a large number of vulnerable journals and a partial wipe-out as a result of mandated self-archiving is entirely plausible. Although there seems to be a myth that journals are very, even extremely, profitable, the fact is that a great many journals are not profitable or 'surplus-able' (in not-for-profit parlance). In my estimate it is the majority. Within the portfolio of larger publishers these journals are often absorbed and cross-subsidised by the journals that are profitable. Smaller (e.g. society-) publishers cannot do that. Marginal journals do not have to suffer a lot of subscription loss before they go under. Some of these, especially society ones, will be 'salvaged' by being given the opportunity to shelter under the umbrella of the portfolio of one of the larger independent publishers. Others will just perish if they lose subscriptions. They could of course convert to open access journals with article processing fees, but setting those up is no sinecure, and requires a substantial financial commitment, as the experience of PLoS and BMC has shown. Journals that are run for the love of it, by the commendable voluntary efforts of academics, are mostly very small, and are the first to be affected, unless, of course, they do not need any income because they are crypto-subsidised by the institutions with which their editors are affiliated. Such journals have always been there and there are probably more now than ever (and some are very good indeed, or so I'm told), but to imagine scaling them up to deal with the million plus articles per year published as a result of global research efforts seems far-fetched, indeed.
Open access is the inevitable future, and it is worth working on a truly robust and sustainable way to achieve it.