Wikipedia's servers are located in the United States, and so we are bound by US copyright law, although in an effort to keep our content free wherever possible we also attempt to respect copyright laws of other countries. (See Wikipedia:Non-U.S. copyrights.) If material is not automatically excluded from copyright (see Stanford's summary of some of the exclusions), public domain for age or other reasons, or released under a license compatible with ours (that would be CC-BY-SA for all text, with most text co-licensed under WP:GFDL), we can only use it if it meets our non-free content guidelines. In order to keep our content as freely distributable as possible, this is a deliberately more narrow range than fair use. For text, this means we can utilize limited quotations of copyrighted material—so long as that material is clearly marked, cited and used verbatim—when appropriate to illustrate a point, establish context, or attribute a point of view or idea. (Some contributors are under the mistaken belief that if the source is cited, exact duplication of text is not a copyright infringement. This is not the case. United States copyright law does permit limited use of copyrighted text under fair use, but makes very clear that "acknowledging the source of the copyrighted material does not substitute for obtaining permission."[1]) Since copyright covers the creative expression of ideas, not facts, we can paraphrase external sources (though we should still acknowledge them to avoid plagiarism and meet verifiability policies). But we do have to be careful when paraphrasing that we do not too closely follow the original in structure and language.

The US government utilizes a "substantial similarity" test intended to determine if infringement exists. Melville Nimmer produced subcategories of "substantial similarity" for which the courts search. In the first, they look for "fragmented literal similarity", checking for phrases and passages copied or closely paraphrased from the original text. Unless such phrases are defensible as fair use, their presence is a strong indicator of infringement. (Note that on Wikipedia, such phrases must always conform to our non-free content guideline.) This is a definition of copyright with which most contributors are familiar.

But in the second, courts look for "comprehensive non-literal similarity." Even if there is no verbatim duplication of the copyrighted original, infringement can be found if the new version follows so closely on the structure of the original that copying is clear. As the US Court of Appeals noted in discussing Artica v. Palmer, et al. (970 F.2d 106, 1992): "A plaintiff succeeds under this doctrine when it shows that the pattern or sequence of the two works is similar."[2] Wikipedia's contributors are cautioned here against utilizing great detail in summarizing or analyzing, to avoid creating a derivative work, as only the original copyright holder has the legal right to license derivative works.

This can be a challenge in practice, but we do need to be careful to comply, since word-for-word duplication is not the limit of copyright infringement. Basically, what this means is that you can't read an article in The Fabulous Encyclopedia of Everything and reproduce it here, not even if you tweak the language a bit so that there is no "literal similarity." A close paraphrase of another source (whether comprehensive or fragmented) may be a derivative work, which is actionable unless it meets the fair use doctrine. If substantial similarity exists, you (yes, you, if you added it here) could be in trouble. Wikipedia could potentially be in trouble along with you. The best way to avoid this is to not only substantially restructure the article, but also to incorporate additional sources. If you're drawing on multiple sources, you're less likely to be taking too much from one.

To read more about summary, I heartily recommend the following:

• Wikipedia:Close paraphrasing
• Wikipedia:FCDW/Plagiarism
• Tweedledum and Tweedledee: Plagiarism and Copyright.
• This guide from Purdue University

With regards to addressing copyright problems, I am guided by Wikipedia:Copyright problems/Advice for admins. In terms of articles what this means (briefly) is that if copyright infringement is substantial and there is no non-infringing version in history, the article is likely to be deleted. It may be deleted immediately under criterion WP:CSD#G12 of the speedy deletion criteria if it meets certain conditions. It may be deleted after a seven day process if it does not. This deletion is necessary for copyright protection.

Sometimes, copyright problems are not extensive or have not been present through the entire history of the article. In that case, I may delete the "history" of the violation, leaving an earlier version of the article where the problem did not exist. Alternatively, I may remove the material or revise it, leaving a warning in edit summary and at the article's talk page against the restoration of that text. In all of these cases, content can be restored to the article (or a new article written) as long as the material is sufficiently revised. You also may restore the material if you can verify that the material is free for Wikipedia to use.

Sometimes, material may be already free for use, and it can be a simple matter of providing evidence of that (for one example, by providing a link to show where on the site the material is already placed in public domain). If it is released under free license, it may come down to a question of whether the licenses are compatible with CC-By-SA. (Source: [3])

License Compatibility
Creative Commons Licenses
License/Compatible		License/Compatible
CC-By-NC	N	CC-By 2.0	Y
CC-By-NC-ND	N	CC-By 2.5	Y
CC-By-ND	N	CC-By 3.0	Y
CC-By-NC-SA	N	CC-By-SA 1.0, 2.0, 2.5, 3.0	Y
CC-By-NC-SA	N	CC-By-US 3.0	Y
Other Licenses
Any GNU only license	N	GFDL & CC-By or CC-By-SA	Y

If there is no indication of free license at the source page, you will need to verify permission. See the next section.

Sometimes the material is not already free for use, but copyright owners are willing to release their material in a form we can use. In that case, we will need external verification of permission to duplicate previously published material on Wikipedia or even to very closely summarize it. We need this even if you have claimed to be the copyright owner of the external source, because we do not require verification of identity at account creation.

You can verify permission in one of two ways, though licensing requirements depend in part on who authored the material. The licensing requirement is as follows:

• If you are the sole copyright holder of previously published material that you, yourself, are placing on Wikipedia, it must be co-licensed for use under CC-BY-SA and GFDL (both of which licenses allow commercial and non-commercial reproduction and modification so long as the material is properly attributed and as long as subsequent versions are licensed the same).
• If you have coauthored the material with somebody else, or if somebody else owns the copyright, it may be co-licensed as above or released under any single license compatible with CC-BY-SA (including CC-BY-SA itself).

When transitioned to Wikipedia, it will be released under CC-BY-SA and, if co-licensed, GFDL.

The two methods of verifying permission are:

• Note on the original source of publication: If you (as the copyright owner) choose to do the former, it is sufficient to place at the site "The contents of this website (or page, if you are specifically releasing one section) are available for modification and reuse under the terms of the Creative Commons Attribution/Share-Alike License 3.0 and GNU Free Documentation License unversioned, with no invariant sections, front-cover texts, or back-cover texts)." Make a note at the article's talk page linking to that release, and the administrator who evaluates the article at the end of the 7-day listing period will restore the contents. Note, please, that the release must remain displayed, or the material may be challenged later and removed. (If you are not the sole copyright holder, reference to GFDL can be omitted.)
• Communication directly to the Wikimedia Foundation verifying the release, via e-mail, postal letter or FAX: If you choose to send a letter to the Wikimedia Foundation, it is easiest by far to do so via e-mail. The address that sends the e-mail must be associated with the original publication. There is a boilerplate release form at Wikipedia:Declaration of consent for all enquiries which can be helpful. For text, the release must either put the text into public domain or license it under CC-BY-SA 3.0 and (if you are the sole copyright holder) GFDL. Please provide a clear link to the website in your e-mail and specify by name the article on Wikipedia in which the material is being used. Once your e-mail is received and processed by a member of the Communications Committee, the article's contents will be restored if your release is legally sufficient. Please make a note that you've done this on the talk page of the article to help guard against premature deletion of the page. You can compose a note or very simply paste the following on the talk page, brackets and all: {{OTRS pending}} (For more information, including the address to which the release should be sent, please see Wikipedia:Requesting copyright permission or Wikipedia:Donating copyrighted materials.)

Occasionally a contributor sends a letter, but forgets to note that at the article talk page, or the Communications Committee becomes backlogged and the letter is not addressed in a timely fashion. Sometimes the letter did not address all concerns, and correspondence does not complete in time. The article may be deleted. When the verification is received by the Communications Committee, it should be restored. If this has happened to you, I would recommend waiting a few days to see if the process completes. If it does not, it may be that your permission letter did not reach the Committee, as e-mail is sometimes lost. I would recommend resending it if no action has been taken 10 days after you originally sent your note. My observation suggests all requests should have been processed by then.

Wikipedia:Image use policy offers some guidance on image copyright and public domain. If the image is not public domain—or if you are not the original author of it—it still may be usable. Again, our non-free content guideline can help you, as it explains when we can sometimes use images that do not have a compatible license. There are a number of Category:Non-free image copyright tags that can be used in conjunction with a properly filled-out Template:Non-free use rationale to assert when an image is so usable. Please keep in mind through all of this that you cannot claim authorship of a derivative work. For example, you can't photograph a copyrighted painting and release the photograph into public domain. You can't take a screen capture of a television show and release the screencap into public domain. You can't crop a copyrighted image and release the crop into public domain. Only the copyright holder has the legal right to make and release derivative works.

I may sometimes delete images or tag them for speedy deletion under one of the processes set out at our guide to image deletion, but only if I find the concerns clear-cut. Typically, I will list them for evaluation at our possibly unfree image forum or the forum for non-free content review.

For questions about image copyright not explained above, please visit the helpful folks at Wikipedia:Media copyright questions.