Do we really need gold open access if publishers permit self-archiving and sharing?
What you're describing is open access. It's simply a different form of it to the one Springer want you to pay for...
There are effectively two-and-a-half routes to open access, with a lot of subtle variants between journals -
- Gold - the article is available immediately, freely, and in perpetuity, via the journal's website, probably with a permissive copyright license. The big open access journals (PLOS One, Scientific Reports, etc) are generally all-gold. It is usually paid for, but not always - many journals are gold but don't charge.
- Hybrid gold - a special case of the above but for a single article in an otherwise subscription-access journal, rather than an all-gold-OA title. Usually costs more than an all-gold journal and almost always paid for.
- Green - a version of the article (usually but not always the accepted manuscript) is available through a repository (or personal website, etc), often after a delay of around a year, and usually without a permissive copyright license. Usually no charge (other than standard publication costs if relevant).
Almost all mainstream journals allow green open access, though the rules are complicated, and they often refer to it as something like "permitted distribution" rather than "open access". Most journals from the big commercial publishers offer hybrid open access, though exactly how common it is for people to take this option is still a bit of an open question. (Finally, most all-gold journals also permit green open access, but obviously it's less important in that case!)
So, the question becomes - assuming my journal allows green OA, as almost all Springer and Elsevier titles do, what is the point of paying for hybrid? (This is a question I answer a lot for people looking at OA option forms, and most of them decide not to...)
Possible reasons to take gold over green, assuming you've already decided on committing to that journal, might be -
- Audience. If you expect the title to be of wide public interest (= lots of non-academics want to read it) or valuable to particular audiences in regions without great journal subscriptions (Africa, South America, etc) then hybrid gold may make sense.
- Timeframes. If you know it is likely to be of particular interest now, and making it publicly available a year or two down the line would be much less valuable, for whatever reason, then hybrid gold is more useful.
- Distribution. If you know you'll want to make a lot of copies of your paper and distribute them widely and publicly, having a permissive copyright license through hybrid gold may help do this.
- Discoverability. Hybrid OA papers are marginally more discoverable than green OA ones, as you can find them through the journal website, the DOI points to them, etc. A repository or personal website may not always be as easy to find.
- Policy requirements. In some cases, funder or institutional policy will require that a hybrid OA option be taken if available. This is rarer now, and most of them are starting to pedal back on hybrid spending, but it may apply.
In practice, most people don't take the option. And if you want those specific benefits, you can get most of them from using an all-gold OA journal in the first place...
There are several strands to the question, that I will try to answer separately. But I want first to stress that one should not make the confusion between Open Access and Paying author charges. These are distinct notions, the first being about who can read the paper, the second about the business model. There are several way to make one's article open access without paying any charge (let me take this opportunity to advertise the Fair OA Principles) and there are journals that have authors pay charges without making their papers open access (e.g. color figure, excess pages charges, front cover seem to be commonly paid for in chemistry; one of the top economy journal needs all authors to be paying members of a learned society to publish a paper).
The opposition you frame between Open Access and self-archiving is misleading: there are several ways toward open access, and self-archiving is one (called "green OA"). So, in the most common terms, when you put your paper in a public repository, you are actively engaging in Open Access.
Many institutions have their own repository, and several fields have world-wide field-specific repositories, but this does not cover everyone's research. For those who have no good repository available (good meaning in particular not owned by a legacy publisher who can shut it down any moment if it threatens its revenues), putting papers on one's web page is nice but not sufficient for many purposes (such papers go dead often, e.g. when the author changes institutions or retires). In such case, publishing Open Access with the publisher (sometimes called "gold OA", but here the precise meaning may vary) can be a plus.
In most cases (including Springer's and Elsevier's policy in the matter), subscription journals only allow the postprint (author-formated version of the paper, after peer-review) to be shared in repositories and web pages. This may be an issue (e.g. Theorem numbers may not match between the published version and the postprint, which can be an issue). It would be preferable to have a clear, unique version of record, properly identified (with a DOI) and available to all to read. Thus having journal publish OA is a benefit.
Once we look at the problem not author-side, but reader-side, then OA policies matter a lot, because much fewer papers are actually deposited in repositories than could be. So access still is an issue, even in fields that are assumed to use repositories a lot such as maths (to the best of my knowledge, high-energy physics is probably the sole field where nearly 100% of papers are actually available OA).
That said, to answer the question in your text (rather than the title question), I would not advise paying OA fees (often named APC for "Article Processing Charges") unless maybe if you are in a field where this is common and in a position where this is easy (e.g. you have access to specific founds); I would strongly advise against paying OA fees in hybrid journals (journals that mostly run on subscriptions but make individual papers OA for a fee), since the revenue from OA articles seems not to be really offset from subscription prices, and I would strongly advise against paying high fees (high usually start at more $500 or more than $1500 depending on who sensible you ask - note that Nature would say that $30,000 is not that high). Be wary that predatory OA publishers usually charge moderate fees (a few hundred dollars), which are nonetheless extortionate given they do no work and no peer-review, so publish only in journal for which you can establish they have a serious reputation (this is also important for non-OA journals).
Don't pay them for refraining from restricting access - which is like paying them for not beating you up. And do make your papers available online - in a place that's robust and visible to search engines (e.g. ArXiv, or your institute's repository of online publications).
We are in a state of transition - with copyable work in general - between the age of copy restrictions and the age of free dissemination and sharing.
The previous age has its roots in the church in the middle ages with monks transcribing forbidden texts while the masses can't even read; or with English printers getting the crown to forbid independent printing. Publishers like Elsevier and Springer have these behaviors and this mindsets as their heritage.
In the new era, we only hide private information (or rather: we only let the NSA and the other state agencies who spy on us see our private data); but anything else - like scientific and artistic creations - are publicly accessible, to be copied and used for whatever cause people might have.
As more and more people expect to see the latter in the world rather than the former, the publishers are trying to have sort of a compromise which people would be willing to stomach. Let's put more pressure on them.