01The same paper, in several states
A research article passes through stages, and copies survive at each of them. The author's original submission is the earliest, written before any reviewer has seen it. The accepted manuscript comes next: identical in content to what was published, reviewed and revised, but set in the author's own formatting rather than the publisher's. Then the published version of record, typeset, paginated and behind the subscription.
The accepted manuscript is the one that matters to you. Its findings, its numbers and its argument are the same as the published article, and very many publishers permit the author to place a copy somewhere public after a waiting period. It is legitimately obtained, freely available and, for the purposes of understanding and citing a study, sufficient.
One caution, and it is the reason to know these apart. A submission that was never accepted has not been reviewed, and its numbers may differ from anything that reached print. Check whether the copy in front of you carries a note saying it was accepted for publication and in which journal, and always cite the published article rather than the copy you happened to read. If page numbers are needed and your copy lacks them, the record in your library catalogue will supply them.
02Where the free copies actually sit
Open copies are scattered rather than centralised, which is why a single search rarely finds them and four short ones usually do. None of these takes more than a minute once you know it exists.
Work down this list and stop as soon as something opens:
- The repository of the university where the author works. Most institutions run one, most require their researchers to deposit accepted manuscripts, and almost no student thinks to look there.
- The archive run by whoever funded the study. Public funders across several countries require deposit as a condition of the money, and those archives are open to anyone.
- The author's own page: a departmental profile, a personal site, or a scholarly network profile. Researchers post their own work and are pleased when it is read.
- A preprint server for the field, which will hold the early version and often a link to the final one.
- The journal itself, if it is open access or has passed the point at which it releases older issues. A surprising number of paywalls are only a year or two deep.
03Interlibrary loan is quicker than its name
The service still called interlibrary loan is the most underused thing in any library, largely because the name suggests parcels and fortnights. For a journal article it is neither. Another library scans the pages and a copy arrives in your inbox, frequently within a day or two and sometimes the same afternoon.
It is free to you at almost every institution, there is normally no limit worth worrying about, and online students are entitled to it exactly as campus students are. The one thing that slows it down is a bad request, so give them the digital identifier if the record has one, plus the journal, year, volume and pages. A request naming only an article title makes somebody do detective work before anything can be sent.
Two practical notes. Ask early in a project, not the night before, and ask for several at once rather than one at a time; a batch request costs the same effort as a single one. And if a request comes back declined, ask why. Occasionally the item is held in a form they can point you at directly, and the decline was a routing problem rather than an answer.
04Writing to the author
This succeeds a good deal more often than people assume, and it is the route almost nobody tries. Researchers are permitted to share their own work with individuals who ask, most are glad somebody wants to read it, and a reply frequently arrives within a day.
Keep the message to four sentences: your name, your programme and the project it is for; the exact title you are after; one specific line about why their study bears on it; and a plain request for a copy. The specific line is what separates a message that gets answered from one that reads as an automated request. Use your institutional email address if you have one, and write to the corresponding author, whose address is printed on the paper.
All of this sits upstream of the writing. Once the sources are in hand, whether a source deserves to be in your reference list at all is a separate question, taken up in predatory journals and how to spot one. If the reading is where a term is disappearing rather than the writing, a tender can carry that stage as part of work on a single assignment or across a longer research project, and every source handed back is one you can open yourself.
FAQQuestions to control
Is it acceptable to cite a free version I found online?
Yes, provided it is the accepted manuscript of a published article and you cite the published article rather than the copy. The content is the reviewed content. Give the journal, year, volume and pages from the catalogue record, and if you quote directly, take the page numbers from a version that has them. A copy that was never accepted for publication is a different matter and should be treated as unreviewed.
How long does interlibrary loan take for an article?
For a scanned article, often a day or two and sometimes the same day, because nothing physical travels. Books take longer since they are posted. Requests submitted with a digital identifier, journal, year, volume and pages move fastest, as nobody has to identify the item first. It is free at most institutions and available to online students on the same terms as everyone else.
Will an author really send me their paper?
Frequently, and quickly. Sharing a copy with an individual who asks is permitted, and most researchers are pleased that somebody outside their field wants to read the work. Write four sentences: who you are, the exact title, one specific sentence about how it bears on your own work, and the request itself. Send it to the corresponding author from your institutional address.
What if the article does not exist anywhere for free?
Then put a request in with your library instead of buying it, since the price a publisher asks of an individual is not a price anybody should pay for a single reading. While you wait, check whether the study is described in detail in a later review article, which will often give you enough to decide whether the original is worth the wait at all.
