What to do about missing source attributions?

Question

Copying, Linking, Attributions, and Plagiarism

I have two related questions that I would like to see some discussion on:

1. Do we need to improve our Help Center’s text to remind people that they have to cite their sources by name?

   UPDATE: I’m marking this first question completed as of 2014-07-08 04:38:38 ᴜᴛᴄ.

2. What should we do about postings that lack proper attribution?

   • Leave a comment.
   • Downvote.
   • Edit in the attribution if known.
   • Delete the posting.
   • Flag the posting. (and if so, as what?)
   • Add a blue moderator note.
   • Something else.

---

Edit: Apparently, ours is not the only site with these problems. Most of the issue mentioned there apply here, notably including this quote from that posting of Shoggoth’s:

Do not tolerate answers consisting primarily of text copied from other sources

[. . .]

We require that answers consist primarily of the words of their author, and that all quotes be clearly marked as such and attributed to their respective authors.

A bare link next to verbatim, copied-in text is not “attributed”, and answers that are just text copied from elsewhere provide no words from that author. If you read his posting referenced above, there are specific steps given as guidance.

---

Background

In our Help Center’s section on “How to reference material written by others”, it reads in part:

When you find a useful resource that can help answer a question (from another site or in an answer on English Language & Usage Stack Exchange) make sure you do all of the following:

• Provide a link to the original page or answer
• Quote only the relevant portion
• Provide the name of the original author

Example:

According to Ernest Hemingway - Biographical on Nobelprize.org, Hemingway saw combat when he was a teenager. It says:

After the United States entered the First World War, he joined a volunteer ambulance unit in the Italian army. Serving at the front, he was wounded, was decorated by the Italian Government, and spent considerable time in hospitals ....

[other sources, quotes, explanations, etc. necessary to complete the answer]

However, this is not happening. I believe that it should be happening, and I would like for the community to decide how to approach missing attributions.

This omission is especially noticeable in answers to questions tagged single-word-requests, where the answerer will copy in a dictionary definition without saying where they got the definition from. However, it occurs throughout our site.

Sometimes a link is given, and sometimes it is not. Now in some cases, one cannot provide a link because the source is from a printed book. In those cases, posters are a bit better with supplying the name of the work they are quoting.

However, in the case of links, posters are very bad at this. In some cases, it’s so bad that their would-be answer is nothing but a word that’s hyperlinked to some online resource. Link-only answers aren’t real answers.

But even when there is more, actual text copied out, a link by itself is not an attribution. It does not include in plain text the name of the work linked to. This is burdensome; how can one judge the authority of the cited source if there is no source given for the citation? There is a world of difference between:

1. Citing formally curated resources like the OED, the American Heritage Dictionary, or even CGEL.
2. Citing crowd-sourced resources like Wikipedia, Wiktionary, the Free Dictionary, Etymonline, or Urban Dictionary.
3. Citing John Q. Public’s random private web page.

Even within each of those three types, there is obviously a hierarchy of “trustworthiness”, but if the source is not named, readers of the posting will not be aware of which one it is.

I would like to see the source named so that we can tell how good of a source it is. I believe that all citations, link or no link, need to provide the actual name of where the text has been copied from. That is what our Help Center says, but people are not doing it.

You should not have to punch through a link to find out how trustworthy the citation is, or even where it has come from. Telling people to “hover” won’t work not just because it is an undue burden, but also because it does not work on the mobile interface to SE.

---

The Good, the Bad, and the Ugly: Examples Galore

Edit: Some of the examples listed below as being in one of three buckets seem to have gotten unwittingly tossed into the wrong bucket. In particular, some posts were erroneously tossed in the “Ugly” bucket and do not deserve to be there. My apologies.

Here are examples of good, bad, and completely absent attributions:

• The Good: examples of complete attributions: 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20.

• The Bad: examples of links missing attributions: 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41.

• The Ugly: examples of completely unattributed copying (plagiarism): 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28.

Answer 1

The help text at How to reference material written by others is common to all sites. Here’s Math.SE’s version for comparison — it still has the Hemingway quote.

That means that it’s a site-wide norm and expected of all posters on every Stack Exchange site.

It’s certainly no more difficult to reference a dictionary with its name than it is to create a link on the headword itself: you still have to create a link — just create the link on the text “[ODO]” or whatever instead of the headword.

It’s a service to those who come after you: it immediately identifies the source of the quotation and the reader may or may not feel the need to click through to check it. They certainly don't have to do that (or hover, if that’s available) to find out where the text came from.

It’s good manners towards the source of the quote. They hold the copyright, and while Fair Use may allow a quotation for the purpose of reference or argument, the person or organisation who did the work to provide the material should merit an explicit referencing citation.

It creates a more scholarly look and feel to answers to have even a minimal citation. For an online link a full MLA citation isn’t necessary, but a nod in that direction can’t be bad, especially since it’s not difficult.

A reason that Stack Exchange likes the actual text identifying the origin to appear in the reference (and thus why the How to reference pages mandate it) is that the content of SE is made available via APIs for display elsewhere. Links may not survive. Including the citation in plain text ensures that it is also included wherever SE content is shown elsewhere. This helps increase the value of the content, and of the site as a reliable resource; and SE itself can’t be accused of plagiarism.

I certainly don’t recommend going through and altering thousands of answers to make them match SE policy. But it’s not unreasonable to expect future answers to follow that policy especially since it isn’t particularly onerous. If you can edit in a citation on a post without one when you are tidying things up, so much the better.

This post used to mention “unattributed material may be deleted”, which — while not inaccurate — isn’t the whole story and is open to misinterpretation. Plagiarised material may be deleted summarily without warning. If you mark a quotation as a quotation, then it’s obviously not plagiarised. In this case, it’s unlikely to be deleted without warning. However, if a citation isn’t added after a reminder, or an edit to add one is rolled back, then the quote or the post containing it might be subject to deletion. Each case will be dealt with individually.

Note that a simple link out to another site isn’t normally quoting external material. It's normally original text containing a link. This sort of link isn’t covered by the quotation policy at all.

There are some worked examples in another answer.

Answer 2

Just to add to the chorus: if you quote a dictionary, it is perfectly sufficient to link to the definition - the link itself becomes the author citation. Explicitly citing the dictionary's name would imply that that particular dictionary is somehow to be preferred for this definition, which 99% of the time is not the case. (I have a list of dictionaries bookmarked, and when I need "a definition, any definition", I will randomly choose one of those bookmarks.)

If your answer involves comparative definitions ("dictionary A says [x] but dictionary B says [y]"), then naturally, the names of the dictionaries must be part of the answer. But otherwise, adding the dictionary's name would be worse than unnecessary: it would be possibly misleading.

I do agree that answers that are 99% quoted, or are quoted with no attribution, need to be addressed somehow: fixed if possible, downvoted into oblivion otherwise. But please, don't let's start offending good contributors with counterproductive nitpicking.

Edit This is becoming ridiculous, folks. A LINK IS AN ATTRIBUTION!!!!!!!!! If this were a print medium, then forcing people to add awkward, possibly misleading verbiage about the source of a quote would be understandable, but last I checked, this is the effing INTERNET.

Answer 3

I want to disagree with you strenuously. But I can't. That doesn't mean I can't disagree with you mildly.

I wholly agree with your discussion about The Ugly. We should strongly discourage quotes, including definitional quotes, that do not give the reader the ability to check out the source. A quote that provides neither the name of the source nor a link to that source, at the very least least, borders on plagiarism.

I also agree that an answer that consists of little more than a linked word is also substantially less than helpful, and outside the realm of how we provide answers. It fails because the information that is most important to the reader, the actual definition, is not on this site.

I am troubled by the view that an answer, often a single word, accompanied by a written definition (and sometimes even an example of use), coupled with a link to an established online dictionary or other standard online reference, is an unacceptable response. I am flattered that the first example of a Good answer is one in which I name the source, provide a link, and quote the applicable definition. I am chagrined that this answer was edited by me to read this way as a sarcastic riposte to @tchrist 's chiding.

I get the problem. Someone can post a quote and a link to a site of limited or even spurious repute, and the unwary may simply take the data on our site without vetting the underlying source.

And yet, the most critical information is often the surfacing of the word. The provided definition may clarify and confirm that the answer suits. In many cases, whether the particular definition is from ODO or American Heritage (that sounds so much better than yahoo) is often (but not always) immaterial. That some authority lists this definition as fitting is often enough.

For the reader who is trying to determine the trustworthiness of the source, the link is a better guide than a mere reference to the name of the source. The link can provide context (alternative definitions, examples, etymology) that the textual cite cannot.

Ironically, the section @tchrist cites in this site's Help Section reads

According to this biography, Hemingway saw combat when he was a teenager. It says:

After the United States entered the First World War, he joined a volunteer ambulance unit in the Italian army. Serving at the front, he was wounded, was decorated by the Italian Government, and spent considerable time in hospitals .... [other sources, quotes, explanations, etc. necessary to complete the answer]

[Supplemental note: The above section and the reference to it in the above question have been modified since this answer was written. I think this actually supports my point as noted in comments below.]

The name of the source is not given, just a link to it.

The difference between a fairly complete answer that provides a word, or term, coupled with a definition and a link to the source, and that same answer with the name of the source added seems to parallel the difference between footnotes and end notes. Footnotes are clearly more informative during the process of reading. You know exactly who said what. But they are often distracting (have you ever read a law journal? Save me!). End notes provide the same information, but some of the material is not immediately present. However it is readily accessible if you need it.

I can't disagree that a simple textual cite to the name of the source (along with other critical elements) is a bad idea. I'm not sure that it should be a rule.

Answer 4

Are you sure that you are not misinterpreting this rule? (for the part about giving the exact name of the link as a text)

I started seeing your comments about this in my answers. (I think you should have waited for our opinions first.)

For example: Adjective to describe someone who is knowledgeable, resolute, and calm

Why would I write the exact name of the link as a text in my answer? I'm copying the relevant part and attributing to the link. This covers the first two rules in Help Center:

• Provide a link to the original page or answer
• Quote only the relevant portion

And the third rule does not apply: "Provide the name of the original author"

Should I provide the name of the author of the online dictionary here then? I do not think so.

I provide the name of the authors only when I give citations from books. Even the example in Help Center looks like my answer.

[Edit: Hemingway example in Help Center is updated after this answer]

I see a lot of answers that attributes to the link but that does not give the exact name of the link as a text. Even moderators have these kind of answers.

Why would it be a problem now after all this time?

Note: You are right about the answers that copy from a source but that does not provide any link or source.

Note2: If there is a really big problem about this, more details can be added to the instructions in Help Center.

Answer 5

Downvotes are for poor answers so while some postings may also need to be downvoted I don't think downvoting is appropriate for missing citations.

Commenting gives the best hope of fixing the answer, but there's no guarantee. In the meantime we're hosting content that's essentially stolen; this is not the academic way.

Editing to add attribution sounds fine, but won't always apply. If it can be done then I say go for it.

Flagging may create a lot of work, and what exactly is accomplished? The question/answer/comment hasn't been improved, and the content is still there.

I think deletion is the most appropriate. We may loose content, but was it quality content? It's not that hard to cite your source.

In regards to adding the Author's name. I think that's overkill. The point of a citation is give credit. So long as the source can be found credit has been given. However I think leading by example would be useful here since including the author's name is good practice; I just don't think it needs to be required.

Answer 6

Since I am especially interested in EL&U questions that involve the historical development and alteration of word meanings, or the first occurrence of a word or phrase, providing source information is crucial to the practical value of my answers. A bit of advice I received (some time ago) in a comment from Hugo about the usefulness to others of linking to (as well as quoting) online references has encouraged me to try to be thorough in documenting my sources.

The drawback of this approach is that such information adds an element of stiffness, formality, and physical length to answers, which some answerers may find incompatible with their writing style.

I also wonder whether, for answerers who embrace the "competition for reputation" aspect of this site, the desire to be quick with an answer in order to get it posted (and collecting Up votes) as soon as possible doesn't serve as a constant (though probably in most cases low-level) motive not to show their work. To judge from some of the scores earned by answers with unattributed quotations, many voting readers don't have a problem with the lack of annotation.

Ultimately, I think, it's important to believe that good and useful content drives out (or at least outperforms) bad and useless content. To the extent that it does, we can all contribute by endorsing good questions and answers, and by constructively criticizing—or at least refraining from approving—deeply flawed ones; to the extent that it doesn't, we may simply be dealing with an imperfect website in an imperfect world.

Answer 7

TL;DR

I believe providing links to quotes can be enough to be attribution. Asking for the source to be apparent from the post's text is more likely to prompt inconclusive debates about style rather than good attribution practices. Three Ugly questions are examined with the intent of supporting that claim.

Intro

Growing up I remember a debate among fellow players in a sport about how focused our rules should be. One side said they ought to clearly delineate what is and is not allowed. The other side agreed in principle but cautioned that such a task was difficult (if not impossible) to do -- particularly without coming across as legalistic. Some even thought it might promote shifty behavior, intentionally finding loopholes, perhaps to use as tactical advantage. And so those who said nay to further clarification took to using a phrase that I would like to invoke here: the spirit of the rules.

A Tale of Two Citations

The spirit of ELU's rules about citation, as I understand them, is primarily about academic honesty: giving credit where credit is due, or (more bluntly) not implying the work of another is one's own. These reasons motivate the opening and closing thoughts of the ELU Help Center’s page on How to reference material written by others:

Plagiarism - posting the work of others with no indication that it is not your own [emphases added] - is frowned on by our community, and may result in your answer being down-voted or deleted.

...

And always give proper credit to the author and site where you found the text, including a direct link to it.

Not mentioned, but just as important a hallmark of academic honesty, is to try to present cited material in the way its original author intended. This Meta question covers a serious corruption of the first hallmark, plagiarism. And it has not even touched on the corruption of the second, misinformation:

false or inaccurate information that is spread unintentionally. It is distinguished from disinformation, which is intended to mislead.

I have structured the above citations to display a point. The quote from the Help page goes above-and-beyond the thoroughness I think necessary. In it I:

• Maintain the cited work's hyperlinks and emphases, regardless of how irrelevant I think doing so may be to the topic at hand. The cited work I thought was important thought they were important, they're related (or at least not spammy), and I know how to include them. So why not?
• Use ellipses to indicate omitted information, indicating to the reader implicitly that:
  1. The given information may not comprise all the link's material on the topic
  2. Any loss-of-flow in the cited passage is my fault and not the original author's
• Indicate when I have tweaked the source, using the plural emphases to indicate that both the bold and italics are not necessarily the original author's.

And yet this first example has not even met the standards I was taught in academia! Because of the ephemeral nature of online information, I was asked to provide the times I accessed a webpage when learning the MLA citation format. But posting a timestamp "7 July 2014 20:30:16 MST" to all (or even to the most-likely-to-change) webpage accesses, while official and precise, is not something I would ask every poster to take time to do. (For my purposes, the "upper bound" to when the poster could have accessed the information, automatically provided by the edit/answer capabilities of SE, is more than sufficient.)

I posit that the second citation, while not as rigorously thought-out as the first, still ought to be regarded as fine, even though:

• It does not include the links to the False, Information and Disinformation wikipages that the original citation now links to.
• Not including ellipses or words in the beginning may be seen as a misattribution of source. The wikipages don't go around starting fragmented sentences with lowercase letters like that! :)
• The added emphasis on unintentionally is not explicitly noted.

I think these oversights should be pardoned because they adhere to the spirit of the rules. By virtue of having a hyperlink on the word misinformation that links to a page that reasonably could have contained the information the poster claims it does and by using the site's "< quote" markdown, I think the poster has put enough effort toward academic honesty, that a link alone can be considered attribution -- the poster is pointing to where the credit should go.

But is it properly used attribution or is it misinformation? In this case I would say "used properly enough." I think the poster's emphasized text is consistent with the article's point, so it doesn't matter to me if it's notated because it's being used as a tool to highlight the poster's point.

I agree that posts that present a source's material verbatim without crediting the source ought to be handled, preferably edited to include the source. "Bad" and "Ugly" posts that can include these changes will likely be improved. But I think it's more likely enforcing these changes will invite debates about style. To demonstrate, here are three cases from the Ugly section above.

Case #1 (Ugly #5)

Poster answers Is there a word for telling the truth (technically) in order to misguide? with

Equivocate: To make a statement that is capable of being taken in more than one way, with the aim of exploiting the ambiguity.

The answer does not appear to have been edited since its posting Nov 14 '11 at 15:40. The link on Equivocate leads to an Answers.com page that after two years still has the cited text, attributed to The Oxford Dictionary of Philosophy.

Admittedly, it would have been better to draw the word from a publicly available or hard copy. But I see little difference between trusting a web service's citation of such a source versus an ELU member citing a hard-copy. Though reputation of either helps me intuit what the situation is in both cases, I wouldn't count either as invalid until I had seen the hard copy or disputing sources. Unless disproved, asking for the reference name here seems more like a stylistic choice than keeping with the spirit of proper attribution.

Case #2 (Ugly #7)

Poster answers Single-word synonym for a “pedantic rule-follower”? on Oct 3 '12 at 21:56 with

I've encountered a few people that you describe. Often, they were bureaucrats:

An official who is rigidly devoted to the details of administrative procedure.

The answer is later edited by another user on Jan 2 '13 at 21:32 for the given reasons of "readability (typography), footnote" to

I’ve encountered a few people that you describe. Often, they were bureaucrats:

An official who is rigidly devoted to the details of administrative procedure.¹

The footnote is easy to miss in the edit, but stylistically I agree with the decision. The page leads to the singular bureaucrat, so OP's decision of linking the more obvious bureaucrat (minus s) in the link makes sense. But the editor's change to a footnote is more aesthetically pleasing, as is the change of bold text to less-heavy italics. After about a year-and-a-half, the link to TFD in the original, which stayed through the edits, still contains the cited text but without emphases. I believe any changes to this answer's attribution would be stylistic.

Case #3 (Ugly #13)

Also answering Is there a word for telling the truth (technically) in order to misguide? but on Nov 14 '11 at 15:06:

I would use the word prevaricate:

to speak falsely or misleadingly; deliberately misstate or create an incorrect impression; lie.

Another commonly used word for this same behavior is to fudge, meaning to disingenuously avoid or talk around an issue.

Because prevaricate follows the patterns I defend above, I will focus on to fudge. This appears to be a paraphrase, because my Google search of the quoted phrase leads primarily to this answer.

I suspect OP meant to tag the answer currently above it, which neither cites nor defines its suggested word of obfuscate. A source will certainly improve this answer in my book. But if that's not the case, I think adding in that the cited definition for prevaricate comes from Dictionary.com is mostly a stylistic choice.

End

Because these three cases which I believe have a fair case for legitimate attribution, have been lumped in the Ugly category with the worst offenders because they don't meet a rule about providing the author's name, I think it's better to go with a looser spirit of the rules philosophy. Does it seem like a person is farming rep without attribution? That can't go on. But anything aside from verbatim-without-citation probably ought to be handled on a case-by-case basis.

Answer 8

It is nice to see that even though the rule now is that "moderators are instructed to delete on sight without further warning any content that is not properly attributed", practice shows that editing extra text is still an option, as in this ELU question.

The moderator's comment suggests this is a one-of show of leniency...

I am waiting for this delete-on-sight policy to be put into practice.

I have so far flagged a number of answers, including very good and useful ones, for moderator attention (that is, instant deletion).

I once read something about SE being moderated by the community, but now it seems that the community is being told not to attract any moderator attention if they do not want to see potentially useful information being deleted.

That seems strange. If moderators are being instructed to delete on sight, then certainly users should be encouraged to assist moderators in their work, as they have always been.

If certain content (like an obvious link to an on-line dictionary without an explicit mention to the name of the specific on-line dictionary) is something that the powers that be deem so unwanted on any SE site that it deserves immediate deletion without reprieve (the same way spam is treated), it is hard to understand why users should not be encouraged to treat those terrible offensive questions and answers in exactly the same way as spam.

Telling users to look the other way while instruction moderators to execute on sight is a contradiction. Are users, in order to preserve useful information, now urged to comment on posts, urging the author to "quickly add correct references before a moderator sees it"? I remember those things from the playground... but since we are all over 13, it strikes me as strange.

If such a paradox is the only way to preserve useful content on this site, there is something very wrong in the rules that give rise to this strange construction.

---

How can the attribution in this answer be sufficient, by the way?

The images are, or seem to be, well-attributed:

All images provided by the respective Wikipedia article referenced immediately above each one, with one obvious fair-use exception.

However, there are no Wikipedia articles referenced! Yes, there are several links to Wikipedia articles, however, a link is not sufficient attribution. I guess there should at least be yet another addition to the post to state that the several unattributed links refer to Wikipedia.

Or, as another obvious option, I simply fail to see the difference between those links and links to dictionaries with just the referenced word in the link text.

Answer 9

I like the direction that ermanen and Erik Kowal are going in. I also agree with the comment Erik Kowal said about nitpicking. (His comment can be read under emanen's answer--linked above.)

---

I'm unfortunately on this list in a rather unflattering light (thankfully not Ugly, but The Bad #34). I've since corrected my answer, and hopefully it's better now. _{*Please note: I have screenshots in case linked-to points are tampered with.}

If commenting is an action that should be taken, is there a comment that should be copy/pasted to stay uniform (and to make sure everyone is responded to fairly)?

I agree that flagging is not the answer. Flagging will cause mods to lose sight on their main purpose to SE. Deleting posts seems too rash. While they may not be appropriately cited, who is to say that their post isn't beneficial to SE? I think those posts should just be edited--either by another person or by the one to post the atrocity.

There is a blatant lack of consistency: OP has gone on a rampage and is rather unfair to some while being somewhat fair to others. Sure, they made it to The Bad list. However, the rhetoric isn't the same. (I find this to be more than relevant to the question. The question talks about posts not being uniform to the Help Center and a way to correct this. This response is relevant because if we're going to nitpick on consistency, then perhaps notifying people with vigilante comments should be consistent and polite.)

Not everyone is treated fairly when getting a nasty-gram about their questionable plagiarism or line-bordering plagiarism. There is a non-uniform rhetoric (sometimes with grammar issues). Then there is this case where it is overly friendly #26. Then there is a scenario where there is a blatant disregard for notification for the dastardly plagiarism, but instead a playful comment (listed on The Bad #7).

---

Also, consistency is lacking from OP. Where is your mention of the source for the link you provided? After all, quite a few of those on The Bad list did exactly what you did and, well, they are on your hit list.

---

SUPER ULTRA POINT TO BE MADE HERE: OP NEEDS TO SEE THAT THIS HAS ALREADY BEEN ADDRESSED ON META SE. (link)

Answer 10

To the question "What should we do about postings that lack proper attribution?": when the source is Wiktionary, the good news is that the post can easily be edited to respect copyright.

Legally speaking, a link to Wiktionary is sufficient attribution, because all authors are either stated in references or retrievable from the page history. Additionally, Wiktionary preserves all revisions, hence the link is never going to get outdated though it might become a bit tedious to retrieve the exact version the author was referring to.

So, in order to make the post compliant with copyright, just edit the ELU post to add a link to any quoted Wiktionary entry and mention the CC-BY-SA license.

Permalink is not required, though it can be helpful to understand any changes. If one wants to avoid any copyright concern, a mere link to a Wiktionary entry is totally fine, unless the material referenced is hard to find in the entry or the post is discussing/suggesting the need to edit the entry, in which case a quotation of the (current) relevant passage is most handy. See also my answer on how StackExchange and Wiktionary help each other.

For more detailed information, refer to the terms of use and the license (sections 4(a), 4(c)), which are the only legally binding documents.