Is it legal to add your publications to ResearchGate?
Is it legal to upload published journal articles on public servers, where everyone can download it for free?
It depends on the publishing agreement you signed, which should specify which rights the authors and the publisher hold. The allocation of rights depends on the journal (and the field you are working in), and sometimes there are modifications or addenda for a specific paper.
You should keep copies of the publishing agreements for your papers, precisely so you can refer to them at times like this. If you don't have them, you can get an approximation by looking on the journal's website to see what their current policies are. The policies could have changed over time (and in principle the publisher could still enforce a stricter policy you agreed to in the past), but it's probably safe to post a paper online if the publisher allows it as part of the current publishing agreement. If you're feeling nervous, you could always ask the publisher whether you could follow the terms of the current publishing agreement for your paper that was published long ago.
First of all, as far as I know, ResearchGate does not require uploading PDF files in order to add your research work to your profile in this website. However, it is very important to explore the possibility of uploading them, as it gives more visibility to your research and makes it available to more people. That is why I think your question is very interesting.
As @AnonymousMathematician and @gawdzilla already said, it all depends on the policy and agreements established by the publisher. You should ask the publisher. Period.
Now, I want to add some extra information about the alternatives we have to share our research without breaking any publishing agreement or rule.
Normally, there are several versions of a paper till it gets finally published:
a. The editor version: This is the final version of the paper, as it is published in the journal, with all the copyright stuff and so on. It is the final PDF, let us say, the one you normally keep as a record of your research.
b. The author version: This normally refers to an accepted version of the manuscript. It is the version that is finally qualified by the reviewers and accepted to publish by the editor. It is just your manuscript, just raw text (and tables and figures and so on), not usually formatted with the final journal style. Sometimes it is also called a postprint version.
c. A preprint version: It is also a sort of author version, but it usually refers to the submitted manuscript version before the reviewers do their job. It is the first (submitted) version of your manuscript.
So, in fact, I've listed them in a sort of reverse chronological order. :D
Now, if you want to make your research as visible and accessible as possible in ResearchGate (or other scientific social networks such as Academia.edu, etc.), which possibilities do you have? Let us see:
- You can upload the final (editor) version of your paper, as long as the publisher allows you to do that, but I do not think this happens very often. [Actually, there are some journals that follow an open-access policy, meaning that the authors keep the copyright of their work. But, anyway, you have to follow and stick to the final publishing agreement and see if you are allowed to distribute your paper by yourself, etc.]
If not, you do have other options:
You can upload the author version. You should also check with the publisher if you are allowed to do so. Sometimes there is an embargo or vesting period and, after that period, you are allowed to publish or distribute the author version of your paper on your own.
You can upload the preprint version. Publishers normally allow authors to publish or distribute preprints of their manuscripts, normally as technical reports. Anyway, you should ask the publisher, just in case.
If none of the above options apply, then you always have the option of adding this text at the end of the 'abstract' field in the ResearchGate record for your paper:
[ Full text available at http://... ]
Now, I want to add some practical examples from my own experience. Hope it helps.
In my case, I usually publish the preprint version of my papers as technical reports in the official repository of my university department. It is a quite common practice among some researchers in my field. Among other benefits, it prevents evil reviewers or editors to copy or plagiarise your research. Anyway, as I told you above, you should check with the publisher whether you are allowed to do so or not.
Now, this is an example of case (3): https://www.researchgate.net/publication/273349919_Optimal_Design_of_Pre-control_Plans, which contains a preprint version of this article that was already published as a technical report.
And here you have an example of case (4): https://www.researchgate.net/publication/45101373_Predicting_the_duration_of_chemotherapy-induced_neutropenia_new_scores_and_validation, which contains the abstract for this article I have coauthored. I will upload the author version in February 2017, when the 36-month embargo period finishes, according to the publisher rules. [Well, it is not actually an example of case (4), as you can see a full version of a preprint in this ResearchGate record. But the preprint is quite different from the final version in this case, so when I am allowed to upload the author version, it will replace the current preprint version.]
As a final remark, please, notice that it is not technically the same publish a research paper than just disseminate, distribute or propagate it. However, I have used them as synonyms in this answer, as sometimes it makes no real difference in the context of your question. [Actually, I think (I am not sure of this) that what you do in ResearchGate when you upload an author or a preprint version of a paper you have coauthored is not techincally publishing it...]
Hope it helps. Any comments to this answer will be welcomed. It is always good to learn new things... :)
As mentioned in the other answers, it depends on your contract with your publisher.
http://www.howcanishareit.com/ indicates given a DOI whether a paper may be shared on ResearchGate or similar websites:
Also, as Benoît Kloeckner mentions in a comment, the RoMEO Journals database contains thousands of journals, labeled with their archiving policy (preprint/postprint/publisher's version):
Except of the database:
Of course, your contract is the reference, whereas those two resources are just approximations.
PS: ResearchGate has quite a bad reputation due to its spam/impersonation/etc. tendencies e.g. see Should I send a "cease-and-desist" letter to ResearchGate?