User talk:Victoriaearle/Archive 16

I wonder

From my wishlist: I'd like to find several gold-plated reliable sources that explicitly address the differences between "secondary" sources and "independent" sources, with a particular emphasis on standard stories that have appeared in recent newspapers.

It's easy to find sources (most university library websites, for example), but I want something "fancier". A scholarly journal or a reputable textbook would do. Do you have any ideas for me? WhatamIdoing (talk) 20:43, 23 June 2016 (UTC)

Yes, you're right, every college library has a webpage defining secondary vs. primary sourcing, but keep in mind those are meant to be used for college writing, which is much different that what we do here. Even though it's a college website, Purdue's Online Writing Lab (OWL) is very respected. Here's a link that might help. The last time I looked the MLA Style Manual, 7th edition wasn't available online, but says on page 3: "Secondary research is the examination of studies that other researchers have made of a subject. Examples of secondary sources are articles and books about political issues, historical events, scientific debates, or literary works."

Basically anything that reports is a secondary source. Minutes of a meeting, say a library board meeting, would be primary. A newspaper article reporting about the meeting is secondary. If Wiesel had subtitled Night as A Memoir, then the title page would be primary. If a newspaper reports he says it's a memoir, that's secondary. I might have to dig around a bit more for this, because we do more than academic work here. My Chicago Manual of Style is out-of-date and so I'd have to get up and go search for it, but if it's different, I'll add here.

What do you mean by "independent"? I haven't seen that term used. Victoria (tk) 22:04, 23 June 2016 (UTC)

Still watching this page, I've seen "independent" usually to mean a "Third-party source". That article can be compared with the article "Secondary source" for good insight, however I'm not familiar with a RS comparing the two. Perhaps look at the refs on those articles. ɱ (talk) · vbm · coi) 22:16, 23 June 2016 (UTC)

One of my hopes for Someday, When I Have Some Spare Time™, is to merge WP:3RDPARTY and WP:INDY. It would make me very happy if you all kept both of those pages on your watchlist, just in case Someday™ ever gets here. WhatamIdoing (talk) 00:26, 24 June 2016 (UTC)

@What—my working approach is that 'independent' and 'secondary' exist on separate axes. Secondary sources may or may not be independent. For example—a blog covering an event from afar is secondary, is not WP:RS, and may or may not be independent. I'd guess you may find it difficult to find a concise, reliable statement expressing what you ask. With news organizations, editorial control is a two-edged sword; it can enforce the aims of a publisher while promoting accuracy. In the U.S., broadcast media have governmental restraints, the fairness doctrine and the license renewal process (though revocation is extremely rare). If I run across a useful WP:RS, I will post on your talk page. — Neonorange (talk) 23:25, 23 June 2016 (UTC)

Neonorange, I completely agree with your statement that "'independent' and 'secondary' exist on separate axes". And so long as the blog in your question is 'transforming' primary sources, as opposed to regurgitating their contents, I agree that it's a secondary source (because it's not just WP:LINKSINACHAIN, and because secondary and second-hand are not synonyms.)

But I find that many Wikipedia editors struggle with this. To be fair, the English Wikipedia's policies and guidelines did use the terms interchangeably until a few years ago, so it's "our fault" that editors get confused, but they do get confused. WhatamIdoing (talk) 00:26, 24 June 2016 (UTC)

Purdue's Online Writing Lab (OWL) lists "journals, magazines, and books", but not daily newspapers, as examples of typical secondary sources. Of course, it depends upon the contents (some academic journals run exclusively primary or exclusively secondary sources, especially in the sciences), but I think that's a fair description. An article in a political magazine, for example, usually contains critical analysis. But I cannot agree that "anything that reports is a secondary source", because "Secondary research is the examination of studies that other researchers have made", not "reporting simple facts without commentary, analysis, critical thought, or any other intellectual exercise", e.g., in a routine newspaper report of the basic facts about what happened during a meeting. The definition you give requires (at least the way that I read it) two key components:

1. "examination" (the intellectual exercise that transforms other works), and
2. "studies that other researchers have made" (some other, previously created source for you to examine).

A standard newspaper article contains neither of these key components. A planned meeting might have had an agenda, but that agenda usually isn't "examined" in a news report. Other typical newspaper contents don't even have that much claim to the existence of prior sources or any kind of "examination". Think about what fills the bulk of any daily newspaper: nearly verbatim reproduction of the police blotter, the story about the house fire, the results of the high school ball game, a human interest story about local resident with a disease, a list of local events. Those stories aren't secondary sources. There's no "examination" and very few are based on "studies that [anyone has] made" before the journalist started writing about it.

Properly speaking, an "independent" source is one without a conflict of interest or similar bias-inducing problems. For example, if Paul Politician says that he's running for office again, then he is not an independent source for the claim that he's running for office. (He is an authoritative source for that statement, but he is not an independent one.) If the daily newspaper says that Paul is running for office again, then the paper is an independent source – but not necessarily an authoritative one. "Third-party" sources take their name from a concept in contract law. Usually, a source that is independent is also third-party (or, according to the lawyer-types, not really any party at all – that language drives some of them nuts). It is, however, possible for a third-party source to not be independent. Wikipedia:Party and person#Doesn't "third party" mean "independent"? gives some examples. WhatamIdoing (talk) 00:48, 24 June 2016 (UTC)

Ok, to back up a little. Using college/university definitions and/or standards is difficult because their purpose is different than ours. Universities and colleges spend enormous amounts of money on libraries (books and digital) and, particularly at research institutions, there's an emphasis on teaching the proper method of research, which trickles through most of the disciplines.

The other thing that's important is that we have to consider the purpose for any piece of writing - almost always either informative or persuasive/argument in nonfiction. Analysis is simply one of the many modes of rhetoric available to build an argument. An argument is always meant to persuade. WP writing is always informative.

To take the example of the library board meeting: a library board has an agenda item to increase its budget; during the meeting they vote to increase the budget. Now someone comes along and wants to add to a WP article that such and such a library increased its budget in 2016. The statement can either be sourced to the meeting minutes or the write-up in the local newspaper. Presumably the local newspaper simply reports what happened at the meeting. That's good reporting, and a good secondary source. But let's say there's a follow-up story the next week, (with analysis) about how irresponsible the library board is for increasing their budget because now the local taxes will have to be raised. This piece of reporting goes on to explain that the raising of the local taxes is suboptimal for the village for a variety of reasons and so on. That's an argument - and not necessarily true. The taxes might not have to be raised. People might prefer books over having their streets plowed in the winter (doubtful, but I'm already in the realm of the absurd, so ...). Anyway, in my opinion that first report, without analysis, is a fine source to use and preferable to the meeting minutes.

These types of examples get scaled up and down in all sorts of ways. Almost all academic writing is argument and that's why, particularly for the humanities, so much has to be sifted through to find some sort of a consensus. Good reporting is just that: whether the event happened today, yesterday or so on. The farther we get from the event the less value newspaper reporting has and the more value is found in other sources. I honestly don't think we can find a gold-plated reliable source for this, not in the brave new world of digitization. That said, national newspapers and the really good regional papers, i.e. the LA Times, Chicago Tribune, Boston Globe, to name a few, are definitely secondary sources, as are journals, (and their associated online versions), in my view. Of course Neonorange is correct re secondary and independent. If the head librarian writes a piece in the newspaper about the benefit of the budget increase, the article in the secondary source is no longer independent. Victoria (tk) 02:33, 24 June 2016 (UTC)

I find this discussion engaging—more meta of this kind could improve collaboration, and perhaps slow the degeneration of featured articles. From following a few FA I see additions and changes made with no comparison of the quality of research, judgment, and skill that produced such a high quality article.versus the that of the changes. Wikipedia is de facto the encyclopedia of record; that's a responsibility requiring concentrating on quality (quantity will take care of itself—vide sv.wikipedia).

As for newspapers, I agree with Victoriaearle. I'd add The Oakland Tribune and, for coverage of Latin America and the Caribbean, The Miami Herald (and the NYT—but neither your list nor mine are meant to be exhaustive). For broadcast television, I'd only consider ABC, CBS, and NBC—even the largest local stations, even those owned by the networks, are deficient in editorial control. Which brings up another point. In the highest tier newspapers, and in the three networks, there are layers of editors and researchers (before and after) between the reporter and publication. Hundreds of hours of work may go into a story, including discussion.

Thank you WhatamIdoing, for the initiating question. — Neonorange (talk) 03:37, 24 June 2016 (UTC)

Is it really sensible to take a general newspaper and declare that every single thing in it is "secondary"? Flip through yours and tell me if you really believe that every piece in it is secondary. My bet is that it's a mix of primary and secondary pieces.

Also, I think that you might be interested in WP:NOTGOODSOURCE. I wouldn't want anyone to think that I believed that "quality" and "primary/secondary" were the same axis. WhatamIdoing (talk) 15:35, 24 June 2016 (UTC)

Hi WhatamIdoing, I edit conflicted, lost my post (that happens sometimes) and wandered away to look at a few more links. So, yes, I do think this is an important discussion, and it's interesting. Essentially I agree with Neonorange who makes a good point about the layers of researchers, editors, etc., which is particularly germane today. I intentionally chose an example about a small town story so as to avoid today's Big Story.

Anyway, I've spent time to see what I can find and the results are interesting, because they are very contradictory. I don't want to put up a lot of links because it muddies the waters and then we get into the reliability of the institution and so on, but yes, a lot of university libraries define newspapers as primary. This makes sense in the context of writing about historical events - if something happened fifty or a hundred years ago, then the researcher can use newspaper accounts and those are considered primary. I've relied, sparingly, on medieval chroniclers for some of the articles I've written and I would call those primary sources. But in the example of the reporter who was not a library board member, did not have an opinion, and simply reported the vote, I'd consider that a secondary source. I do like this link from Brown University.

I also took a look at our own guidelines (which I haven't looked at in a long time - years, to be honest) and I see our definitions have changed since this 2012 version of WP:NOR and today's version. When I get the chance I might trawl through the history there to find the change and look at the talk page. If it were up to me, I stick with the earlier definition; and this is another example what worries me about about what were putting out to the internets, because I am finding a lot of high school and college libraries using our definition. That makes wonder which came first.

No, not everything in a newspaper, or online version, is secondary. Op-ed columns come to mind as an example. Victoria (tk) 16:15, 24 June 2016 (UTC)

Break one

I'm sorry that we lost your previous post. Also, we should bring User:SmokeyJoe into this conversation, as he's more knowledgeable about the subject than I.

A new thing to think about: The Brown source says "are used as both primary and secondary sources", which I think is an interesting formulation. Do you[1] think of "primary" (etc.) as being primarily a characteristic inherent in the source, or does it say more about how you're using it?

[1] That's "you personally", in the kind of editing that you like best; I'm not really trying to find the One True Answer That All True Editors Should Agree With™ here. WhatamIdoing (talk) 20:14, 24 June 2016 (UTC)

This article in today's NYT (24JUN16, on Brexit) is, I think, a good example of a secondary source. The article has clearly been in the works for months, involving interviews, archive searches, and quite a bit of discussion. My contention is that the source is objectively primary or secondary. Classifying a source based on how it is used is a trip down the rabbit hole (aside from Victoria's comment on historical sources which were once secondary but now are primary; a view I support). So, WhatamIdoing I pick door #1. Meanwhile, I must cook dinner, and later will continue collecting sources on journalistic standards—or lack thereof. — Phil — Neonorange (talk) 21:49, 24 June 2016 (UTC)

+ These three handbooks may be useful: The Association of American University Presses Best Practices in Peer Review, The New York Times Ethical Journalism handbook, and NPR Ethics Handbook. Still looking for a copy of a similar CBS News document. — Phil — Neonorange (talk) 06:27, 25 June 2016 (UTC)

++Thanks for the Brown link, but doesn't work for me—my 1,000,000 population county's library system saves money by not joining the University of Georgia Internet access research system. — Phil — Neonorange (talk) 06:37, 25 June 2016 (UTC)

Thanks Neonorange for the links. I'll take a look later. That Brown link doesn't always work for me either. I agree about the rabbit hole and there are lot of rabbit holes! WhatamIdoing I don't really spend much time thinking in terms of primary or secondary but that's maybe because of the areas I edit in. I almost always lean on scholarly or academic sources. Victoria (tk) 13:07, 25 June 2016 (UTC)

Neonorange, I've taken a look at that those links and read most the NYT one. It's fascinating. Thanks so much for digging it up. I thought I'd maxed out the number of stories the NYT will allow me to read in month, but it let me read that story. It's very well done. Victoria (tk) 18:51, 25 June 2016 (UTC)

Victoriaearle I have a digital subscription to the NYT, and access to all the issues from 1851 to date, should you ever need a NYT excerpt or search. I know that even in the videotape era, CBS News saved every second of tape shot, no tape was erased or reused. The tape library did not have a very sophisticated retrieval system, but the librarians in NY had magical powers. Digital video storage is now dirt cheap, and much easier to retrieve — Neonorange (talk) 20:53, 25 June 2016 (UTC)

Victoria, I hope you don't mind if I jump in here.

Wikipedia regards newspaper articles as secondary sources for contemporary events, assuming the authors are not involved in those events. This is why editors search during AfDs for secondary sources in the form of news reporting.

WAID argues that a newspaper reporting the existence of a petition against a professor, to use a recent example, and doing so without analysis, has simply regurgitated the material in the primary source (the petition), and the newspaper article is therefore a primary source itself. This ignores that independent editorial judgement is required to decide how to handle the primary source. No matter how bland the report, no matter how lacking in analysis, the newspaper does not simply reproduce the petition. It adds something secondhand, something uninvolved, including the decision to mention the petition at all.

The primary/secondary confusion may have arisen because whether a newspaper article is primary or secondary has in part to do with its age relative to the event and relative to the period in which it is used. On Wikipedia, almost all our newspaper articles are current articles reporting on contemporary events, written by those not involved in the events, which makes them secondary sources. But older newspapers reporting on contemporanous events would be regarded as primary sources for the period in which the events occurred and for how things were handled at the time.

In case they're helpful, I've mentioned some of these sources before in discussions about this.

• Ithaca College library: [1]

"In the humanities, age is an important factor in determining whether an article is a primary or secondary source. A recently-published journal or newspaper article on the Brown v. Board of Education Supreme Court case would be read as a secondary source, because the author is interpreting an historical event. An article on the case that was published in 1955 could be read as a primary source that reveals how writers were interpreting the decision immediately after it was handed down."

• Suraiya Faroqhi, Approaching Ottoman History: An Introduction to the Sources, Cambridge University Press, 1999, 2:

"For while the available primary sources condition the kinds of question an historian might usefully ask, it is also true that we read secondary sources, including non-scholarly ones such as newspapers and magazines, long before we ever embark on specialised training."

• Willie Thompson, former Professor of Contemporary History, Glasgow University, What Happened to History?, Pluto Press, 2000, 7:

"Leaving the remote past for any other century, including the twentieth, the same principles apply no less. Newspaper reports, having been subject to editing, are secondary sources for the events they record; the original reporter's notebook, if it had survived, would be a primary source. A document can, as we have seen, be in different dimensions, either primary or secondary."

SarahSV ^(talk) 21:43, 25 June 2016 (UTC)

Thanks SlimVirgin, those are really useful too. We're getting a nice collection here and I think the point you make about AfDs is important (I don't work on the deletion part of the project so that always slips my mind). The example you give about the petition is similar to my fictional example about the library board voting to increase funds (and hypothetically there could have been a petition). The person/s who reports these things, from the seemingly inconsequential to the very Big Stories (Brexit), is one step removed, and that's how WP:NOR was written four years ago. I think it's important to try to get that back, by the way. Yesterday's story (Brexit) is significant. WP has an article about, there's been a ton of media reporting and in the long run there will be newspaper articles, journal articles, books (maybe lots of books), it will make its way into history books and textbooks, and so on. But a little more than 24 hours on, we have to rely on newspaper reporting, and yes, today I think the newspaper reporting is secondary and independent. With time it will become primary. Neonorange's links do a very good job of demonstrating that the NYT practices independent journalism, and as a non-journalist it's very interesting to me to read that document.

Also, Neonorange, the way this year is going I'll probably break down and buy a subscription to the NYT and maybe a few others. Or I suppose I could be old-fashioned and go sit in the library and actually read the papers. The info about the video library is interesting too; I have a six-degree-of-separation sort of connection (to do with someone else) when it comes to old CBS news tapes. I might shoot you an email when I get a chance.

I don't know whether WhatamIdoing has more to add or not; if so I hope she does. I'm happy to host this conversation, because it's interesting. When it comes to the issue of sourcing I'm always up for a discussion. P.s. apologies for the late replies. I was deep in an article and surrounded by sources :) Victoria (tk) 00:38, 26 June 2016 (UTC)

SarahSV, I'm glad that you jumped in, because I wasn't sure how many people I could invite to Victoria's page before she'd start wishing that we'd all go away. ;-)

I think your first quotation (from Ithaca College) speaks to my recent question on whether the classification inherent to the source, or if it says more about how we use it. It says that something written at the time of an event "could be read as" primary; something written decades after the event "would be read as" secondary. This suggests that when it's Year X, and the subject of the article is an event in Year X, then the source should be treated as a primary source, regardless of whether Year X is 1955 or 2015, and regardless of whether the subject is Supreme Court rulings or self-published petitions. On the other side, when the event is in Year X and the source is from Year "X+50", then it's (probably) secondary. ("Probably" because it's possible to generate brand-new primary sources, even if that's not likely, and especially not likely in a newspaper.)

I'm thinking that classification is both inherent to the document and also a matter of our use. For example, a meta-analysis, which is practically defined as a secondary source, doesn't stop being a meta-analysis merely because it's a century out of date, but you really ought to treat it like a primary source after a while. On the other side, every source could be a primary source for something, e.g., its self-reported publication date. What do you all think?

Victoria, I think your focus on non-historiography issues is good. In the end, we want a reliable source, not a secondary one. A scholarly primary source is normally better than a secondary blog post.

I think that it's a mistake to classify all newspaper articles as being either way. A source with critical analysis is IMO a secondary source regardless of whether it's published in a newspaper or a journal or a website. So I'd cheerfully use any good "news analysis" piece about Brexit (or Supreme Court rulings, or petitions, or anything similar). WhatamIdoing (talk) 00:55, 26 June 2016 (UTC)

WAID, your posts mixes up several issues (e.g. you mention reliable v. secondary, but reliability can adhere to both primary and secondary sources; and the point that all sources are primary sources for their own views, or for their date of publication, reduces to tautology).

The key issue is that the age of a source matters, as does its position relative to the user of the source. As the Ithaca College author wrote, the concept of primary/secondary depends on the vantage point.

A 1950 newspaper article about an event in 1950 is a secondary source for someone in 1950. But it is a primary source for someone in 2016. For Wikipedians writing about historical issues, their encounters with contemporaneous newspaper reports (reports from that era) are encounters with primary sources. They help us to look into the past. For the same reason, a meta-analysis performed in 2016 is likely to be viewed as a primary source by historians in 2116.

Similarly, a 2016 newspaper article about an event in 2016 is a secondary source for someone in 2016 (though it may contain material that is a primary source for that reader, such as an interview with a participant in the event). But that same article – the interview and the article that presented it – will be a primary source later on for historians; how much later will depend on the issue.

Most Wikipedians write about issues soon after they happen, so this use of contemporary newspaper articles is the one they most often encounter. In those instances – for people from this era using newspaper articles from this era – the articles are secondary sources, because they are written by those uninvolved in the event, were subject to editing, and are no closer to the era than we are. SarahSV ^(talk) 01:45, 26 June 2016 (UTC)

The Ithaca College website doesn't say that "A 1950 newspaper article about an event in 1950 is a secondary source for someone in 1950", and I'm not certain that it's a good idea to assume that they meant it. Additionally, as you say, not everything in a newspaper is a secondary source, no matter what year it's in. WhatamIdoing (talk) 02:21, 26 June 2016 (UTC)

.

(edit conflict recovery)

Thanks for the quotes SlimVirgin. I am especially interested in the Ithaca quote as it intersects with WhatamIdoing's meta-analysis comments. "...an article on the case (Brown v Board) that was published in 1955 could be read as a primary source that reveals how writers were interpreting the decision immediately after it was handed down." For that particular subject, the historical process that generated current stances make it difficult to not take a meta-view. However, contemporaneous articles puublished by, say, the NYT, remain secondary sources that can be used as input to meta-treatment. (I'm not convinced that my distinction is meangful, nor that I am expressing it clearly.) Low quality sources decay rapidly to grist for meta.< Disclaimer: I am using an iPad for this—any weird typos can not be blamed solely one me. > — Phil — Neonorange (talk) 03:15, 26 June 2016 (UTC)

I had one too, so we must have edited conflicted each other. Then I went off looking for comic strips (I like to decorate my talk page) and accidentally closed the window and now the comment is completely gone. It's late and I can't recreate the post, but it went something along the lines that the newspaper article about Brown vs. Brown written in 1955 is primary in 2016, but in 1955 it was secondary. I'm not convinced the instructor worded it as well as could be, otherwise there wouldn't be any confusion. Also, I had a question for WhatamIdoing. What wouldn't you consider secondary in a newspaper? (this is the short version - I had some examples but am now too lazy to write them out). Victoria (tk) 03:46, 26 June 2016 (UTC)

Breaking news, eyewitness reports, reports of simple facts, (edited) transcripts of interviews and similar contents are primary sources. So: "This just in: The high school football team won the state championship" is primary – but not a piece that analyzes various factors that resulted in the win: "The football program received a major boost when they hired Coach Chris two years ago, and the result of his endurance training program and focus on sub-skills can be seen in the trophy case" is secondary. WhatamIdoing (talk) 06:35, 28 June 2016 (UTC)

I would agree about interviews because all too often we see in WP articles material that's sourced to a newspaper interview when in fact it's essentially the subject speaking about her or himself. The others I'm not sure I agree with.

I spent some looking through college textbooks and the results were interesting. Essentially the issue isn't even addressed in "older" books, i.e from the 1990s but it is addressed after about 2000 and the advent of Google. I found a small section in the McGraw Hill Reader giving the example of a speech: the text of the speech is primary, the newspaper report secondary. They don't make a distinction about age, analysis, etc. Another example they use is that when writing about a movie, the movie is primary, any analyses are secondary. That sounds about right to me.

I meant to ask yesterday: can you link to the essay you wrote that's mentioned downthread? Victoria (tk) 00:40, 29 June 2016 (UTC)

Still in broad agreement with everything....

Interviews, which can be primary sources or first-person secondary sources (the subject is commenting on himself), are usually to be first tested at AfD for being probably non-independent. That is, the interviewer and publisher have decided to interview the subject because there is a deal, payment, purchase of advertising space, etc.

On "the text of the speech is primary, the newspaper report secondary", the rule of thumb I use at AfD is whether the news report includes at least a single additional adjective. If the report is just a reprint of the speech, or just a statement that the speech happened, and it doesn't even say that it was a good or bad speech, well-received or entertaining, etc, then it is just a photocopy of the primary source (and is not journalism). This is not just academic, AfD often sees non-notable (WP:PSTS-failing) subjects attempting to be propped up by multiple mentions (isolated primary source repeats) in newspaper articles. --SmokeyJoe (talk) 01:05, 29 June 2016 (UTC)

Interviews can go either way. When you interview Joe Film about his latest film, then it's primary. However, when you interview the local history prof about the history behind this weekend's holiday, that's actually a secondary source. On Wikipedia, we see a lot more primary-source interviews, but the format itself isn't defining.

Victoria, that was probably a reference to WP:USEPRIMARY, but it might have meant WP:Secondary does not mean independent. I've also started WP:USESPS on self-published sources, but I haven't made much progress on it. The merge of WP:INDY and WP:3PARTY will happen first. WhatamIdoing (talk) 01:05, 30 June 2016 (UTC)

WhatamIdoing, thanks for the links, I hadn't seen those pages. WP:Secondary does not mean independent is a redirect? Or did I land on the wrong page? I'm not sure I'm in agreement with that page, particularly with the point that secondary = "a magazine article based on previous media reports".

I'm not involved in the deletion process, nor do I create new pages, so it's an area I really can't speak to. But your example of the football team win got me thinking. First, how do we know the win is the because of the coach (causality is difficult to determine). What if there's a another news story saying the team won because of a new crop of freshman, or for other reasons? And so on. Or what if we simply have a single story (without analysis) reporting the win? I think the story about the win would be fine – certainly it would be fine in the short run. Also, I've thought for a long time that we're still at the point in this project where getting in sources, any sources, is better than none. Victoria (tk) 23:45, 30 June 2016 (UTC)

That example ("a magazine article based on previous media reports") may be a case of "it's a bit more complicated than that". But at the simplest level, if you take prior sources and create new material out of them, then it's secondary (or at least not purely primary). It may be too simple – but, then, getting into complications is what the WP:PRIMARYNEWS section is for.

If other sources (news media or otherwise) have different opinions about the cause of the football win, then we report them in WP:DUE proportion to their prominence.

I agree that a news article that simply mentions the fact of the win is "fine". It's just "fine, in the sense of being an independent, reliable, primary source with good editorial control" rather than "fine, in the sense of being an independent, reliable, primary secondary source with good editorial control".

(Yes, it's a redirect. After I wrote it, I decided that another title would be better, but at this point, it would need an admin to move it.) WhatamIdoing (talk) 06:46, 1 July 2016 (UTC)

Ah, thanks for explaining about the redirect (for some reason I thought you were an admin). I'm not quite following your point about "It's just, 'fine, in the sense of being an independent, reliable, primary source with good editorial control rather than 'fine, in the sense of being an independent, reliable, primary source with good editorial control'".

I think we can add newspaper sources to the secondary column. That seems to be where the disagreement comes in. Victoria (tk) 16:02, 2 July 2016 (UTC)

Many of the non-local stories that appear in newspapers are already secondary—provided by the Associated Press. In many cases these stories can be the product of a stronger editorial structure than that at the newspaper. For the tv organizations with the budget, AP stories can trigger independent coverage (as can stories in the NYT, LA Times, etc.) — Phil — Neonorange (talk) 18:02, 2 July 2016 (UTC)

That's a good point. I keep forgetting to mention wire services. I wonder if that's the distinction we want to make. Pinging Sarah for her opinion. Victoria (tk) 18:20, 2 July 2016 (UTC)

Nothing like an confusing typo. I've fixed the sentence.

I agree that many non-primary sources get published in newspapers. As a rather imperfect rule of thumb, the longer the story, the more likely it is to be secondary. WhatamIdoing (talk) 02:41, 3 July 2016 (UTC)

Victoria, wire-service reports, like any other, would depend on whether they were first- or second-hand accounts. A wire service report from someone involved in an event is a primary source. A report from a journalist who witnessed some aspect of 9/11 is a primary source. But a wire service report that describes a White House press release is a secondary source. That's assuming we're talking about contemporary reports about current affairs. Age changes things. SarahSV ^(talk) 05:07, 3 July 2016 (UTC)

SarahSV Perhaps you used involved in a more narrow sense, but a journalist is not a naive observer. I would add: the training and experience of the observer can make a story (wire service, newspaper, magazine, tv news) a secondary source. Especially when the editorial process necessary before publication is considered. (Example narratives available upon request.) If one substitutes blogger (untrained and not subject to the editorial process) for journalist, then your hypothetical would be a primary source. I am unclear how training other than as a journalist (military expert, for example) affects this particular (observer) secondary vs. primary question. Pinning down tertiary is simple. Secondary vs. primary is contingent, and on more than just lapsed time. (May I ping another editor who may be interested, DGG?)— Phil — Neonorange (talk) 04:20, 4 July 2016 (UTC)

Whether someone is trained in something is irrelevant. What matters is whether they are involved. I write a diary entry about my day. It's a primary source. You write a story about my diary entry. That's a secondary source. Things get more complicated because texts are primary/secondary depending on how they are used, and can contain elements of both, and how old a source is changes our evaluation of its proximity to the event. But people are getting confused here about the basic distinction, which is based on proximity and involvement, so it's worth repeating it. SarahSV ^(talk) 07:09, 4 July 2016 (UTC)

Referring repeatedly to "first- or second-hand accounts", which is a generalized factor that correlates well but is not the definition, is more confusing than nearly anything else mentioned. "secondary sources are second-hand sources" is only generally true, even if it was written by someone in a history department; it has exceptions and its converse has many exceptions. In the absence of transformation, a second hand source is a primary source with chain of custody. The most relevant article to read would be secondary source. "How to classify a source is not always an obvious decision". When push comes to shove, "Second hand" is not the criterion, it is an introductory generalisation. --SmokeyJoe (talk) 05:14, 5 July 2016 (UTC)

I think it's interesting that, with the amount of brain-power here, and the many years of Wikipedia editing represented, this seems to be a contentious topic. I spent some time again looking to see what various institutions post on the net, and am linking one that I like. It's an organization affiliated with George Mason University, and essentially tells us that, indeed, age is important, and it mentions the concept of eyewitness, which SarahSV is getting at a few posts above. I think it represents a simple definition, but note that it's written for primary and secondary school students.

I will admit that it's not at all difficult to find links that define primary/secondary somewhat differently, i.,e this from Rochester University. They categorize newspapers as primary, yet in the chart below (distinguishing disciplines) they tell us a review of a play is secondary, which I think is confusing. My suspicion is that there's a shift, but it's not fully realized yet. These things happen. The question is how should Wikipedia define newspaper articles.

Do we want to have instruction creep and a layer of many if-then statements in regards to secondary sources? Isn't it simpler to stick with the WP:NOR definition we had in January 2012?

To address WhatamIdoing's original question: a fancier source would be the McGraw Hill Reader (it is a text book) that I linked somewhere here. I looked at about 20 textbooks and that was the best I could find. Victoria (tk) 00:31, 6 July 2016 (UTC)

Wikipedia editors have decided that reviews are primary sources. The usual reason given is that they're "subjective". (NB that I'm merely reporting what editors have said in these discussions; I'm not agreeing with them. This item is definitely on the "choose your battles – and this one isn't worth it" list for me.) WhatamIdoing (talk) 06:22, 6 July 2016 (UTC)

WhatamIdoing, what kind of reviews have editors decided are primary sources? I can't offhand think where that has been argued. SarahSV ^(talk) 04:16, 11 July 2016 (UTC)

Sorry, that must be very confusing. Thanks for asking for clarification, SarahSV. I'm thinking of restaurant reviews, book reviews, and similar reviews. I'm not thinking of review articles in the sciences. WhatamIdoing (talk) 04:34, 11 July 2016 (UTC)

WhatamIdoing, thanks for the clarification. I can see that these would hover in a twilight zone between primary and secondary. SarahSV ^(talk) 04:40, 11 July 2016 (UTC)

I agree with you. You can make a good-faith case either way, and many reviews contain some elements of each type.

I'm thinking that "primary versus secondary" is entirely the wrong way to think about reviews. What you really want is a good source, for the statement (and article) in question. A good source has a well-informed person, completely independent of the film/opera/restaurant/whatever, writing in a publication with decent editorial control. And if you've got that, then who cares whether it's primary or secondary? "Secondary" is not the definition of a reliable source.

(One more thing: I don't hear people making the same claim about product/computer reviews. A film review is "primary" because it's "subjective", but a software review is "clearly secondary and proves notability". We may have a bit of WP:ITSIMPORTANT informing people's view of the historiographical classification.) WhatamIdoing (talk) 18:59, 11 July 2016 (UTC)

I have trouble with the essay WP:THIRDPARTY (Third-party sources), as does WhatamIdoing. It is unnecessary; even the essay acknowledges this:

Although there is technically a small distinction between a third-party source and an independent one, most of Wikipedia's policies and guidelines use the terms interchangeably, and most sources that are third-party also happen to be independent.

What does Wikipedia lose by concision if clarity is gained? If the phrase 'third–party' disappears from Wikipedia, except when needed In the text of an article, is there any loss? Perhaps chopping verbiage will encourage editors to actually read guidance. I agree with Victoriaearle: when every possibility can be covered by 'if statements', then robots can do the editing. — Phil — Neonorange (talk) 02:33, 6 July 2016 (UTC)

I hope to make 3PARTY go away (via merge). Maybe this will finally be the year.  :-)

In the meantime, anyone who wants to replace links to 3PARTY with links to WP:INDY is welcome to do so. WhatamIdoing (talk) 06:18, 6 July 2016 (UTC)

Secondary and involvement

I agree with SmokeyJoe that it's wrong and unhelpful to conflate involvement with classification as primary or secondary. Here's a simple example:

• I run some experiments. I publish them. These are primary sources, as very widely agreed by scientists.
• Later, I decide to do a meta-analysis of all the previously published reports on the same subject, including mine. This is a secondary source, as very widely agreed by scientists.

But I'm involved! Some (or possibly all) of that original data came from my own notebooks!

Well, who cares? A meta analysis is inherently a secondary source – no matter who wrote the original papers, no matter how biased I am, no matter who's paid by whom, no matter whether I'm trained, no matter how old it is, no matter if it's well (or badly) done, no matter any of these things. That's why Secondary does not mean "involved". These are completely separate considerations. There's some good correlation, but it's actually a separate consideration.

I think that the proximity-and-involvement notion comes from taking a source that is "inherently" a secondary (or tertiary) source, and explaining to first-year history students why they're not allowed to use it as a secondary source. A century from now, my hypothetical meta-analysis – although defined as a secondary source – should not be used as a secondary source. That is, you shouldn't use it to say "The whawhojits is fleebled in the frobbins"; instead, you would use it to say, "Over a century ago, WhatamIdoing published a meta-analysis claiming that the whawhojits is fleebled in the frobbins" (followed, one hopes, by a statement about how wrong that claim is now considered). The same is true for century-old political analysis (whether published in newspaper, magazine, or book form), which "is" a secondary source, but should be "treated like" a primary one after it reaches its expiry date.

(Oh, and you can certainly have second-hand reports from highly involved people. Think "officials deny abuse allegations against their staff". Nobody's going to agree that the officials are truly "uninvolved" in the situation now, even if they weren't personally engaged in any incidents.) WhatamIdoing (talk) 06:27, 6 July 2016 (UTC)

I don't think that it happens in the classroom. Peter Damian explained it below: "Wikipedians shouldn't be adding 'transformative', but should rather be relying on reliable secondary sources which do this work for them". During the intervening time since the event, more is written and these are the sources where the analysis occurs. Students in the humanities are taught to how evaluate primary and secondary sources for their research, and to a great extent those rules apply here. There are gray areas though, i.e, in obscure topics when we have few sources available, or in current events, which is outside a historian's purview and hence the issue of whether current news stories can be used.

Anyway, I'm thinking this has gone around in circles for a while now and I'l be closing it up soon. My sense is that there are some gray areas, some areas that are unnecessarily confusing (particularly for new editors), but I don't think those problems will get solved here. I'll be archiving this in a few days. Victoria (tk) 11:37, 10 July 2016 (UTC)

Victoria, thanks for hosting. I continue to see length as a problem for Wikipedia guidelines and essays in the area of this discussion. Accretion explains some, but not all. How about a quick-start for new editors? I've been reviewing pending changes, looking in depth at (or searching for) sources. and I wonder

• There is no way to enforce read this guide (or even read the article) before editing. But read a single page and take a 'pop quiz' before editing might be attainable—and eventually enforceable.
• How much incivility is bred by confusion over reliable sources, in the broad sense, for Wikipedia. Recognizing that some people move to incivility more quickly than others, may not have patience for eventualism, and have been trained by using consumer technical support.

— Phil — Neonorange (talk) 19:55, 10 July 2016 (UTC)

• Yes, Phil, essentially I agree with you. I'm not sure we can enforce pop quizzes, but I think we have to have strong unambiguous definitions, and, as Sarah says below, definitions that aren't at odds with what is learned outside of Wikipedia. Victoria (tk) 23:50, 11 July 2016 (UTC)

• This is becoming hard to follow, so sorry if this is in the wrong place, but in brief we should rely on the definitions used by academics and other publishers and not make up our own. Primary sources are widely regarded as firsthand accounts. Secondary sources are widely regarded as secondhand accounts. Age affects what we regard as firsthand because old sources are closer to an old event than we are. Therefore, something that today we would regard as a secondary source will be a primary source in 50 years' time. This isn't a difficult idea, and I'm not sure what the benefit is of trying to make it more complex. SarahSV ^(talk) 04:22, 11 July 2016 (UTC)
  • The problem with "firsthand is primary, secondhand is secondary" is that it's not true, according to the very academics and publishers that you think we should follow. That simplistic formulation falls in the category of lies to children. It's a place to start, but it's not 100% correct.
    The benefit of making it more "complex" is "getting it right", e.g., not using breaking news to support contentious matter about BLPs.
    And, yes, Victoria, thanks for letting us take over your talk page for a while. It's been interesting, pleasant, and fun. WhatamIdoing (talk) 04:38, 11 July 2016 (UTC)

• WhatamIdoing, what Sarah says is correct. Academics agree that a firsthand account is primary. It's quite simple in literature: the text is the primary source, the criticism (including reviews) are secondary. The same rule applies in history: primary texts are letters, journal, proclamations, etc., etcs., and the commentary about them secondary.
• Somehow we have to get this right on Wikipedia; I didn't realize (as I've already mentioned) that there were different approaches, but that has to be confusing, not only to new editors. I don't mind having this conversation continue, but it's getting a little messy and hard to follow. I've been thinking about taking all the links here, sandboxing them, and maybe continue the work there. If I do, I'll post here. Victoria (tk) 23:50, 11 July 2016 (UTC)
  • The problem isn't "firsthand is primary". The problem is "secondhand is secondary". "Secondhand" does not mean, or even imply, "criticism (including reviews)". "Secondhand" can equally mean "12-year-old Mary told 12-year-old girl Jane that 12-year-old Jacob said he likes Jane". Jane now has "secondhand" information (aka gossip), but Mary is not a "secondary source". WhatamIdoing (talk) 01:40, 12 July 2016 (UTC)
    • Yes, I see what you mean. That happens sometimes, i.,e on the Hemingway article where someone posted that he'd heard in a podcast that a gunsmith in Idaho said Hemingway's wife asked him to bury the gun Hemingway used to shoot himself, the gunsmith kept some of the parts, and claims it's not the same make as the one mentioned in the coroner's report (primary source) and the many biographies since written (secondary sources). That's why we have to be careful about reliability, but essentially to me secondhand is a good term. It might be a matter of semantics or even of Engvar though. Victoria (tk) 20:52, 12 July 2016 (UTC)
      • Editorial process divides secondary sources from the merely secondhand? Secondary sources aggregate, compare, and evaluate primary sources.

        When reporting on events we did not witness personally, we seek multiple independent perspectives to get a sharper, more accurate understanding of what happened.
        — NPR Ethics Handbook

        Neonorange (Phil) 22:56, 12 July 2016 (UTC)

Break two

I can't find any disagreement among the sources that contemporary news reports are secondary sources but that older reports can be primary. The following is from SUNY Albany. Something like this used to be in the NOR policy. It would be worth adding it because it makes the issue clear:

"Primary sources are first hand sources; secondary sources are second-hand sources. For example, suppose there had been a car accident. The description of the accident which a witness gives to the police is a primary source because it comes from someone who was actually there at the time. The next day's newspaper story is a secondary source because the reporter who wrote the story did not actually witness the event. The reporter is presenting a way of understanding the accident or an interpretation.

• From North Park University, History Department

"However, the distinctions between primary and secondary sources can be ambiguous. It is important to remember that you cannot determine whether a source is primary or secondary solely based on the document type. An individual document may be a primary source in one context and a secondary source in another. For example, the movie Love, Marilyn is a secondary source when the topic is Marilyn Monroe; it would be considered a primary source if the topic of research was the works of Liz Garbus (the film's director).

"Additionally, time can be a defining element. For example, a recent newspaper article is not usually a primary source; but a newspaper article from the 1860’s may be a primary source for United States Civil War research."

• From CBB Library and IT Consortium

SarahSV ^(talk) 21:06, 26 June 2016 (UTC)

These are all very good Sarah. I actually thought of using an account about a car accident as an example. It's an example that makes a lot sense: the police report is the primary source, the news report the secondary. I also like the one from the CBB Library. We discourage use of newspaper accounts when writing about historical events and instead encourage the use of "secondary", scholarly or academic research. I would like to see the wording at WP:NOR simplified. I was surprised when I read it and if we're at odds with the how these sources are conventionally defined then we should change. Victoria (tk) 23:24, 26 June 2016 (UTC)

I like the CBB Library quote. I've been skimming NYT articles from the Civil War. Time past and an accumulating weight of scholarly examination split these "secondary sources" and they become dual in nature—secondary for the more basic, undisputed facts, primary when there is significant weight of recent examination. (I think the Brown v Board example stung me since I was reading news papers at the time of the event.) I like the brevity of the statements in the NYT Ethical Journalism, and the given rational. The WP:NOR policy is way to long—making it less accessible to those editors who need it most. (Time to stop—my iPad is now auto correcting 'and' to 'no'. If no one objects, I'll continue to dredge up handbooks is an policies from the areas I know best. I continue to enjoy the exchange, and love the stress free environment. — Phil — Neonorange (talk) 01:33, 27 June 2016 (UTC)

Hi User:WhatamIdoing. (Hello User:Victoriaearle, you have a very pleasant talk page here.)

I have never seen a gold-standard source for the terms “secondary source” and “independent”. I don’t think they exist, as these terms are matters of definition, and different fields have used and defined them differently.

Instead, I take the view that the gold standard for terminology is the definitions taken from the field of historiography, and the justification for this is the assertion that Wikipedia, like any encyclopedia but more-so, is primarily an historiographical work.

Every field is prone to some differences in practice, but I have found it useful to talk in terms of three big ones encountered commonly on Wikipedia. One is historiography, which I maintain is the one to be applied to Wikipedia. Another is journalism, which almost dominates current affairs, including politics, international affairs, and BLPs. The third is science. Fourth, which I prefer to ignore, would be sports.

In historiography, a secondary source is well-defined as a source of material that adds transformative information to primary source material. The use of the big words “historiography” and “transformative” conveniently impresses those not seriously interested. Few disagree that an encyclopedia is an historiographical document. “Transformative” by nice co-incidence has the same meaning as the legally significant word that defines the difference between copyright infringement and fair use. Transformation of the primary source material into secondary source material is always fair use, and fair use always involves transformation of the primary source material. The transformation can be described as any of: comment; analysis; satire; ridicule or many other words.

In journalism, “second hand” is frequently confused with “secondary” in “secondary source”, mixing into and muddying the accepted historiographical term. SarahSV and I have previously discussed this, and while I don’t think we every fully agreed on definitions, we immediately agreed that good journalism sources are good historiography sources, and I have never believe that people with different perspectives should be forced into a common perspective. One thing we fought on is “second-hand sources are secondary sources”, which I reject citing a relay messenger service, even though the converse has an academic source.

Science rarely causes us problems. The scientist’s use of “primary” is often taken extremely literally as the one first source, such as the lab book. By the time the student and advisor write their paper, they don’t consider the information to be primary, the information have been analyses, processed and selectively presented. And in science, the primary source is readily doubted, and is re-verified by repeating the experiment and observation. Scientists readily understand that a scientific review is a secondary source, because it is not supposed to any new data, but they don’t really make use of the term at all.

When Victoria says “Basically anything that reports is a secondary source”, I say Hah! Journalist! The murals of Pompeii report on life in ancient Rome. Primary source. Newspapers usually attempt to report on events accurately and without bias. Primary sources. Go count the number of people on the electoral role in the towns of Assam, and report back to me. The report facts will be mere “primary source material”. Counting of names is not creative, it is not transformative. Mere copying, or routine non-intellectual processing, does not transform primary source material into secondary source material.

When it comes to journalism, the very usefully discriminating words are “report” (as a noun) versus “story”. A newspaper “report” on the library board meeting probably resembles the minutes reformatted for readability, and is primary source material. A newspaper “story” on the library board meeting probably makes comments or adds something to make the “story”, as the meeting itself was probably not much of a story. I think the word “article”, as in newspaper article, is best avoided, as it implies neither “report” nor “story”. The typeset editor uses the word “article”.

Newspapers contain a mix of primary source material and secondary source material, absolutely. The front pages are primary source reporting. The editorial section is secondary source material. The form guides and classifieds are primary source stuff, and at the back there is sport. If you want to get more specific, then the devil is in the details, the rules wash away, and it depends on what you are using it for. Are you using the source for facts (primary source material), or for opinions (secondary source material). Rabbit holes if you like, they exist, and they will break your leg if you don’t look out for them.

“Third party” versus “independent” is not something I have much opinion on. Neither imply “secondary source”. I accept them as rough synonyms for each other, although I appreciate WAID’s essays on these terms. --SmokeyJoe (talk) 05:01, 27 June 2016 (UTC)

The historiographical take is interesting, and very helpful. It takes a direction that is new to me, and I will do some reading. I see your view of journalism as reductive, or not, depending on the quality of the editorial process. My comments will probably apply to the three long established broadcast network news organizations, and be based on personal observations and some well founded assumptions. I'll start with an example, showing the transformative nature of the production of a CBS News, for its evening news program.

I'll set a limited stage, and use series of events far back enough for there to be a weight of scholarly examination of the events: the U.S. forced intervention in Haiti (non-lethal invasion is a better fit), Operation Uphold Democracy. My starting point is the free election of Jean Aristide that was abrogated by a military coup led by General Cedras.

1. The stories covering the run-up to the elections and the elections involved three to fifteen or more CBS News staff (or freelance) personnel, plus local hires (translators, drivers, "fixers") on site in Haiti. The standard group might consist of a producer (the manager, with the local editorial responsibility), a correspondent, an editor, and a two person camera crew, and two or three driver-translators. The local work in producing the story is vary collaborative, requiring good news judgement all around.
2. In the news headquarters, editorial discussions plan what stories to produce, and how much time, if any, is available in the broadcast schedule. Bureaus in the U.S. are assigned support duties—interviewing experts to gain insight and provide "sound bites" to fold into the story. In this case, thr Pentagon and State Department correspondents will be working their contacts for supporting information.
3. If a story is scheduled for the day, the on site producer will discuss the expected events and how to cover local actions, and will have calls with NY to request other support—interviews, research, graphics (maps, etc.)
4. As events unfold, the story will take shape, and video editing begins. The script (including description of the images and sound) will be sent to NY for approval. After discussion and amendments, the story takes shape, and is sent by satellite transmission to NY for further video editing. This process can become very compressed by rapidly developing events as deadlines approach.
5. After broadcast, the story is reviewed and a critique is developed. Kudos or "sweat-seeking rockets" are sent to those involved.
6. In this story, after a supporting UN resolution, Jimmy Carter, Colin Powell, and Sam Nunn were sent by the U.S. to negotiate the departure of Cedras.By this time, CBS News had around fifty staff in Haiti, including five camera crews, several editors, multiple producers and correspondents, and the evening news anchor/managing editor. Thirty-five or so U.S. warships appeared in the Port-au-Prince

Producing this ongoing story qualifies as transformative. The extensive capture of events and interviews required extensive editorial judgement to synthesize a coherent set of stories. (CBS News makes a distinction; reporter or correspondent, among their on air staff. New staff without previous network experience may start as a reporter, and be promoted to correspondent when retained with a new contract.

Disclaimer: this was a big story, covered by network news organizations in completion with each other. The quality of coverage varied. Local tv news organization don't have the resources for this kind of effort. Don't expect to see much useful transformation there. Keep in mind this rule of thumb: "for every "real" local tv news story, local stations three slots to fill need to fill three slots, mainly from the local area—for each network news slot, three stories are competing, drawn from the whole world. Economic pressures are degrading the performance of television news. CBS News was not profitable, nor expected to be, until the early 1980's. That changed.

In Wikipedia, I'd never use local tv news stories for anything but the most basic facts. I try to use only the NYT for cites when using newspaper sources. A high quality, extensive editorial process is what makes the secondary source. OpEds are not the only transformative process. Note that network tv news has no OpEds. — Phil — Neonorange (talk) 08:51, 27 June 2016 (UTC)

—Addendum for television journalism—

From the Code of Ethics I found:

Journalism requires more than merely reporting remarks, claims or comments. Journalism verifies, provides relevant context, tells the rest of the story and acknowledges the absence of important additional information.

I think that in choosing to source a Wikipedia article to a television story, that story should be judged against the the above standard before use (not be given an automatic pass as WP:RS). — Phil — Neonorange (talk) 19:08, 27 June 2016 (UTC)

I roundly agree Neonorange, except I am quick to ignore commercial TV as biased and entertainment. The RTDNA code quote I read as aspirational. Its writer has an agenda. I suggest that it should be read with an implied "good" appended to "journalism". Sometimes, the report is no more than merely reporting remarks, claims or comments. Not even context. Such a report is a primary source, regardless of how many hands, if it includes no "story", not even a base story on which a good journalist would add to. Such a source may be reliable, for example, breaking news on a wire service. --SmokeyJoe (talk) 07:13, 28 June 2016 (UTC)

SmokeyJoe, we are in agreement here. I do not watch local tv news for the same reason I do not read local newspapers. I do read on a regular basis a de facto national newspaper, and watch a national television networks news.

• I expect bias as a fact in all sources.
• The RTNDA code is aspirational—there is no enforcement mechanism. The individual handbooks of news organizations are enforceable documents. — Phil — Neonorange (talk) 16:47, 28 June 2016 (UTC)

I numbered your bullet points so I could refer to them succinctly.

1. Number of people involved: This is irrelevant. It's secondary if primary sources get transformed; it's not secondary if previously existing sources don't get transformed. It doesn't matter whether this is done by a lone person or by a supporting cast of thousands.
2. Editorial control: Editorial control is extremely important for the question of whether it's a reliable source, but it's irrelevant for the question of whether it's a secondary source. A meta-analysis is still a secondary source even if there are no editors involved. If I write and post a meta analysis on my talk page, it would be an unreliable source, but it would still be a secondary source.
   Also, the "casts of thousands" idea is present in this item. It doesn't matter how many people collect or review the source material.
3. Number of people involved, again: It's still irrelevant.
4. Editorial control again: Still extremely important for reliability. Still not what makes something a secondary source.
5. Post-publication review: Also irrelevant. By the time this happens, the story's main character as either primary or secondary has already been determined.
6. Results: Irrelevant. Also, that's at least the fourth appearance of the "cast of thousands" red herring in a six-point list.

Having said that: I believe that much of the material produced here actually is a secondary source. I'd exclude live interviews, eyewitness reports, and simple factual statements, but much of the overall content probably is secondary. The process of collecting and transforming primary source material (e.g., all the interviews and other sources that weren't shown on air, but whose contents were critically appraised, compared, and then summarized in a sentence or two on air) makes that coverage (mostly) a secondary source. It wouldn't matter whether that thoughtful transformation had been done by a lone superhuman or by this cast of thousands, but it does matter that the thoughtful transformation happened.

Additionally, I think that it's important to recognize the distinction between "the total process of creating this source" (which happens to involve a lot of 'video editing') and "the process of Capital-E Editing this source", which is the much smaller part during which an editor (and optionally the publisher) decides that the source meets the appropriate standards and is worth publishing. Capital-E Editing is extremely important for reliability. But you can Capital-E Edit a primary source just as readily as a secondary one. The involvement of a Capital-E Editor does not change the classification of the source. WhatamIdoing (talk) 01:42, 30 June 2016 (UTC)

My fault for not being more clear. My little narrative has a limited purpose—to show the methods of tv news production at the network level. Recognizing the numbers of people involved in that process is one way of evaluating the quality of a story as a source. The time pressure during such production is unique among secondary sources, I'd say, and my post was an attempt to show the necessary ways of dealing with this constraint. And a hint at the financial risk. The post-publication period is an important part of the editorial process, especially for an on-going story. Both the numbers of people and the post-publication process is important to the reliability of a source and to this particular way to produce a secondary source. Because the symboytic nature of newspapers and television network news organizations may be more involved than observers may realize, large numbers are significant to both kinds of transformative processes. I can see the academy process from outside, as a user—my guess is that there are some goals in common, but differences in paths. Perhaps you could post, with similar purpose, a short narrative of the academic process? — Phil — Neonorange (talk) 05:06, 30 June 2016 (UTC)

You could put all this work (and financial risk, and anything else you want) into it, and still end up with a primary source. A cast of thousands, and millions of dollars, and all of these other pieces do not make a secondary source (cf., many Hollywood films, including many non-fiction/documentary films).

I don't think it's really pointful to talk about newspaper or academic publishing processes, because the publication process is not what determines the historiographical classification. The publication process has a lot to do with the quality – and therefore the reliability – of a source, but that's a separate (and IMO more important) thing. WhatamIdoing (talk) 05:50, 30 June 2016 (UTC)

The problem we have, and that I think Neonorange (Phil) is trying to highlight, is that these days Joe Blogger (or anyone, anywhere on the net), is able to report what they wish – from a current event to having a website about a famous author. I think that the number of people involved and the dollars spent is important. I suppose Joe Blogger could have lots of money and people too, and that's what makes it difficult to determine reliability these days. In terms of academia (just to answer Phil's question), I'm not published, but the process involves a fair amount of time (research and whatnot), and people (journal editors, peer reviews, copyeditors, etc.). That's what makes a high quality scholarly source. But we won't don't always have the luxury of finding those. There's an American woman author whose article I'd like to expand some day, (she was short-listed for the Pulitzer Prize, I think more than once, coming in second to Faulkner and Hemingway), yet all I can find for references are newspaper articles from the 1950s. Very little scholarly research has been done on her, which, of course, shows a systemic bias. In the end, I'll probably just use the newspaper stories I've found, because it's better than nothing. Victoria (tk) 00:06, 1 July 2016 (UTC)

I agree that the number of people (or, more precisely, the number of separate roles) involved is important for reliability (which is roughly a synonym for quality).

It's just not important for "secondary-ness".

Also, I think we have all three agreed that it's generally more important to have a high-quality source (which might also happen to be a secondary source) than to have just any old secondary source. WhatamIdoing (talk) 06:50, 1 July 2016 (UTC)

Break three

Welcome to the discussion SmokeyJoe. Yes, I've read that an encyclopedia can be considered historiography. That's not an area that I feel qualified to speak to, but I can speak to the subject of discourse (to some extent). Some years ago I read a fascinating paper, to do with literature and the nature of encyclopedic texts. To build her argument the author traced the history of encyclopedias and defined what she considers encyclopedic discourse. This passage stuck with me:

"… the encyclopedia retrieves and shapes knowledge as a whole, whereas other forms of discourse draw upon and shape specific branches of knowledge. The encyclopedia potentially accesses the very frames, scripts, schemata by which we know. It does not have the problem of ordinary discourse – that of weeding out material, of 'restricting knowledge only to the relevant details' needed in understanding texts and new situations (Brown and Yule, 244). (In encyclopedic discourse, nontheless, ideology may influence what is included and excluded)".

Clark, Hilary A. "Encyclopedic Discourse." SubStance 21.1 (1992): 95-110.

http://www.jstor.org/stable/3685349, page 105

I think that last sentence is the one that struck me and the one we need to have some care for. So, to back up. When I wrote that "Basically anything that reports is a secondary source", I should have taken more care with my wording. I would expect to see a masthead, editorial control, the concept of Aristotlean rhetoric's ethos. Of course anyone could report about the putative library board meeting, but in my view if the reporter for a newspaper is there on the local beat, reports it, then it's "one step removed" as our WP:NOR policy used to say, and so can be considered secondary.

It's somewhat ironic, I think, that this conversation landed on a page of an editor who's really a content editor, in the purest sense. As such I lean almost exclusively on secondary sources, and those, I suppose, by definition are transformative. The areas where I edit the most, literature and art, are disciplines that require a great deal of analysis, but I would argue that every analysis must be examined, that we shouldn't simply vacuum/hoover up information because it exists, and then regurgitate it. Furthermore, I'd argue that many analyses must be attributed rather than presented in Wikipedia's voice. This is tedious work, but it's the way I edit.

If I were to edit in areas that relied on news reports, I suspect I'd steer clear of op-eds, or certainly I'd attribute anything from an op-ed. The transformation, in my view, is so great that the distance should be indicated. To digress, libraries are on my mind because I recently went to my small local library to check out books for an article I'm working on only to find that since I was last there about half the stacks have been emptied. I wish they would have a board meeting to increase the budget, or least bring back the books they did have! But, anyway, to go back to that example. I think the reporter's story in the local paper (or posted on the paper's online website) is both secondary and neutral. I think the follow-up analysis (if there were one), claiming taxes would have to be raised is overly transformative, and I think the head librarian's op-ed or letter to the editor is neither independent nor neutral. Again, as I wrote earlier, this is an example that can be scaled up, but I'd prefer to avoid writing about some subjects and chose an example that seemed relatively safe.

To speak to Neonorange's points: that example about how news is gathered is fascinating and gives a good perspective. Thanks for taking the time to write that out. I like the RDA quote, and of course the ethics speaks to ethos. Yes, I agree that WP:NOR is overly long and not as accessible as it could be.

This has gotten longer than I wanted, but one more point. The issue of primary vs. secondary is pounded into students in secondary and higher education in the US. Essentially students are told that a primary source is a document, whether the Declaration of Independence, John and Abigail Adams' letters, a court document, etc., etc. They are told that secondary is one step removed (though that's a bit of an oversimplification). So, yes, it is transformative, but I think the degree of transformation is the issue. At least in my mind it is. In terms of historiography, I think I'd like to ping Jbmurray to the conversation (if he's around or interested), because, if I remember correctly, he's much more conversant than I am on the subject. Victoria (tk) 23:07, 27 June 2016 (UTC)

SmokeyJoe, there's no reason for us to invent definitions. A secondary source is a secondhand account. These are of little or no interest to journalists and historians, who prefer primary sources (material produced by those involved in an event).

I'm pinging Peter Damian, who has a relevant background. Peter, in case you're willing to comment, the issue is whether news reports are primary or secondary sources. Any input would be very helpful.

Victoria and I say that contemporary news reports (a 2016 NYT article about a 2016 traffic accident) are secondary sources, whether or not they contain analysis, so long as written by someone not involved in the event. But older contemporaneous news reports (a 1944 article about D-Day) are often regarded as primary sources because they are close to the event relative to us. So far as I know, this is a standard way of looking at the distinction. See some quotes in green above for sources, such as at 21:06, 26 June and 21:43, 25 June.

WhatamIdoing says that news reports are primary sources, regardless of age, unless they contain analysis. Smokey Joe also says they are primary sources, even when they're clearly secondhand accounts, unless they contain something "transformative." This is a term from copyright law, and I don't know how he's defining it. SarahSV ^(talk) 04:48, 28 June 2016 (UTC)

Hi SarahSV. No, no need to invent definitions, I'm just saying that the best definitions to use are those used within historiography. I think they are the most applicable, and are well-defined. You and I have long since had an almost silly difference on the putative secondhand-but-still-primary-source news report. Such a thing would not be worthy of any proper "reporter". Person A tells person B; person B runs to person C and repeats the message verbatim. The first concern would be reliability. In the modern world, person B has been replaced with a technological communication service, but in principle human messengers still exist, and if they add nothing to the message, if they don't create a "derivative work", then they don't change the nature of the message. While the word "transformative" is critical and defined in copyright law, historians don't take it as seriously, and "how the source is used" quickly becomes more important. In any case, I would love to hear opinions of Jbmurray and Peter Damian. --SmokeyJoe (talk) 05:44, 28 June 2016 (UTC)

SmokeyJoe, I agree that the best definitions for us are those used by historians; see above for some quotes. Journalists don't use the terms any differently. SarahSV ^(talk) 06:32, 28 June 2016 (UTC)

I think that analysis is a type of transformation (and, specifically, it is the type of transformation most likely to be found in the news media). I would accept any sort of transformation of primary sources, and I would accept it regardless of whether the resulting source is news media. You've probably seen me say before, "Whatever the game, whatever the rules, the rules are the same for both sides." If a type of intellectual transformation of previous sources results in a secondary source in an academic journal or a scholarly book, then the same type of transformation results in a secondary source in news media, blogs, or any other source. WhatamIdoing (talk) 02:08, 30 June 2016 (UTC)

Analysis is the most important type of scholarly transformation. Others include "placing in context", "comparison/justaposition/contrast" (weak versions of analysis). Non scholarly transformation includes ridicule. Satire is borderline scholarship. With regards to analysis, the analysis is the secondary source material, and the data analysed is the primary source material. It starts to separate the information from the document. It is possible to take a secondary source, extract primary source material, and misuse it just like any other misuse of primary sources. --SmokeyJoe (talk) 02:18, 30 June 2016 (UTC)

Thanks for pinging me Sarah. It’s all too difficult for me (and I am writing a book on the subject). The classic definition is pretty hard to apply. Take Gibbon’s account of Islam in the Decline and Fall. Gibbon, who professed his ‘total ignorance of the Oriental tongues’ relied extensively on Jean Gagnier’s 1723 Latin translation of a biography by the Syrian Abu'l-Fida (1273—1331), writing more than 600 years after Muhammad’s death . His work is little more than a transcript of of Ibn Al-Athir’s biography, who in turn relies almost entirely on that of al-Tabari (d. 923), who in turn relies on Ibn Sa'd (d. 843) who relied on his master al-Waqidi (d. 823), and Ibn Ishaq (d. 768). These biographies were probably based on tradition or oral accounts, so which was the primary source? Generally in Wikipedia we try to distinguish sources which require significant interpretation and specialist understanding, or which would involve synthesis of some kind. Thus Wittgenstein's Tractatus would be a primary. But we do allow newspaper reports, so long as they are referenced as such and are essentially 'factual'. The basic rule is to avoid sources which could be slanted in some way to some POV, and to prefer sources which are self-evidently neutral. Peter Damian (talk) 19:54, 28 June 2016 (UTC)

PS I quite like the definition above: 'In historiography, a secondary source is well-defined as a source of material that adds transformative information to primary source material.' The key is the 'transformative information'. Wikipedians shouldn't be adding 'transformative', but should rather be relying on reliable secondary sources which do this work for them. Peter Damian (talk) 19:58, 28 June 2016 (UTC)

Hi Peter, thanks for stopping by. I agree. Completely. Thanks for being so succinct. Victoria (tk) 00:44, 29 June 2016 (UTC)

I also think that the concept of "transformative information" is key to the subject and perhaps an easy way for someone to grasp the core idea. It's not that I took the primary sources and put them through some sort of "process"; it's that I took the primary sources and transformed them into something new. WhatamIdoing (talk) 02:03, 30 June 2016 (UTC)

An aside comment: this edit isn't meant to be pointy given the discussion we're having but rather, I realized, some of the people we've pinged in know me by my former name and might want to know whose page they're being invited to, fwiw. Victoria (tk) 01:07, 29 June 2016 (UTC)

WhatamIdoing, Victoriaearle, SlimVirgin, PeterDamian: does anyone wish to bring this home to Wikipedia? In skimming Wikipedia articles and talk pages, I see many cases where sources are evaluated primarily in terms of winning a competive argument rather in determing quality. If the distinction wasn't learned through formal instruction, how important a role can Wikipedia play, and how? Does Wikipedia encourage the former? — Phil — Neonorange (talk) 01:26, 30 June 2016 (UTC)

In practical terms (which may be premature), I have two goals for Wikipedians' knowledge. I want them to know that secondary is not a fancy way of spelling reliable (see WP:NOTGOODSOURCE), and I want them to know that secondary is not a fancy way of spelling independent (see WP:Secondary does not mean independent).

The inclusion of "secondary" in WP:GNG is one of the main (remaining) structural sources of this battleground mentality. If you want the English Wikipedia to include information about businesses, events, or living people below the had-books-written-about-them level of fame, then you "need" to define all normally reliable sources as being "secondary". Otherwise, we'll be deleting half a million articles for failing the GNG.

I'm not sure that the GNG should require "secondary" sources. It was written back when secondary and independent were frequently used interchangeably by editors, and we probably didn't mean to be demanding secondary sources. I wrote about this problem a few years back (see here), but it's neither a popular view (with anyone who wants to write about any subject whose highest quality sources are the news media) nor tactfully expressed. WhatamIdoing (talk) 02:03, 30 June 2016 (UTC)

Neonorange, can we bring it home? Not sure we're ready yet, but it's a subject that bears looking at. Yes, lots of talkpage disputes are about sourcing and about winning, and that's a problem. I'm wondering if we can just simplify our definitions? SmokeyJoe is correct in saying that "historiography" and "transformative" are big words. It seems that using little words often works the best (though I have to laugh as I write this, looking at the wall of text above!). Can Wikipedia play a role in formal education? Absolutely. I use it as a teaching tool and for the most part we look at sourcing in random articles. It's a fun activity (actually the students love it) and I don't think I've ever done it when I haven't found at least one mistake. Which I then change and everyone "oohs" and "ahs". It's amazing the power this place has, when seen from the other side. Victoria (tk) 00:24, 1 July 2016 (UTC)

Break five

I won't pretend to have read all this, but I've never been too happy with WP:PRIMARY, according to which it would seem to me that the Anglo-Saxon Chronicle, not to mention all Roman historians, are secondary sources. Johnbod (talk) 03:34, 30 June 2016 (UTC)

Thanks for stopping in Johnbod. That's a good point. I finally had to take the Battle of Poitiers off my watch (one example off the top of my head), because it's mostly sourced to medieval chronicles. It's not as if a lot of ink hasn't been spilled over that single battle in 500 years, but WP relies on Froissart and others. I'm sure you can think of many many such examples. And then there's this edit that just showed up on my watch. A 2010 scholarly source replaced with chronicles. On an FA. Either I can have a fight, or I can ignore it. These days, if I'm here at all, I've learned to ignore. But it doesn't make me happy. Victoria (tk) 00:48, 1 July 2016 (UTC)

I did once raise this point but got very crossly and firmly sat on by some power admin. Of course I was forgetting Julius Caesar, primary even on WP! Hope you are well, Johnbod (talk) 02:29, 1 July 2016 (UTC)

Caesar should be primary (diaries, letters, etc.) but really you'd think, as Peter mentions above, that someone has done this work for us in the past 2000 years! I just noticed your Core Contest entry btw (I didn't even know another was going on!). Very nice work. Victoria (tk) 16:06, 2 July 2016 (UTC)

I've just noticed that buried deep in Wikipedia:Administrators'_noticeboard/Incidents#CurlyTurkey_unilaterally_deleting_good_article_nomination (first section) is a claim, made in all seriousness, that the Nihon Shoki (completed 720) is a secondary source for the History of Japan. Whether it is or isn't is discussed at some length by User:Nishidani (against) and others. One to note if we ever try to improve the policy wording. Johnbod (talk) 09:49, 17 July 2016 (UTC)

I never read wiki rules, otherwise I'd never be able to edit content in, apart from apathy when reading any bureaucratic prose (though wikipedias's is certainly quite well written for laymen). But when it was drawn to my attention, that def. of Primary source seemed fantastistically stupid, since it makes no room for the distinction that is normal in historical scholarship. It seems to be based on a 'reality-text' idea in legal works: primary is eyewitness or close to the event, secondary is everything derivative of that. That doesn't work in historiography: Livy's history drew on Polybius extensively: Polybius in turn drew on Fabius Pictor, who in turn drew on Diocles of Peparethus, who in turn almost certainly composed his lost history by basing it on unknown earlier sources. According to the wiki guidelines, the primary source is unknown except through the secondary source of Diocles, as cited by the tertiary source of Polybius which an editor might quote from the fourth source Livy, etc. Sheesh. All written works surviving from antiquity are primary sources, even if they draw on other, lost, sources. All commentaries on them are secondary sources. My working rule is, avoid primary sources for history wherever possible, and if you cite them, preferably do so as they are quoted in the secondary scholarship. Thanks, Johnbod, for the notification of this discussion.Nishidani (talk) 10:21, 17 July 2016 (UTC)

Hi Nishidani, and welcome to this very long conversation. Good luck if you try to read through. I just had a look at the thread Johnbod linked and had to raise my eyebrows when I found the claim that the Nihon Shoki isn't close to events and thus equals secondary. Your replies are good. I should have pinged you when I needed someone to speak about historiography. I don't do big words very well. Victoria (tk) 13:39, 17 July 2016 (UTC)

Nishidani, good to see you here. I wish I'd thought to ping you. I've made the point several times that age is a factor in determining what is primary and secondary, because of proximity to the event relative to us. The discussion here was about whether a (say) 2016 newspaper article about a 2016 traffic accident, written by someone not involved in it, is a primary or secondary source for that event. Historians and journalists would say it's a secondary source because it's secondhand and edited. The key issue is proximity. See the three quotes above in my post of 21:43, 25 June. SarahSV ^(talk) 16:50, 17 July 2016 (UTC)

As is nearly (just because I don't believe in Papal Infallibility) always the case, your examples illuminate and your judgement is spot on. I'm not worried about the modern stuff or how we classify it and I avoid technical wiki issues because it is beyond the usual reach of my natural radar sensitivities, and just trust that people like yourself, and many others, will get it right. All I can add is that it would seem to be necessary to adjust the policy to make it clear that in speaking of preprinting age historical works in the broad deep sense, a simple distinction between primary (the basic written materials) and secondary (how scholars interpret and edit them) holds. In that terrible discussion Johnbod alluded to, despite my adducing over a dozen examples of scholarly sources referring to the Nihon Shoki as a primary source, the recalcitrant POV-pusher (in my view of course) grasped at this policy to deny what the scholarly consensus maintains. Don't read that thread, it's just a recipe for a headache or a nightmare. But it does show that our failure to make a simple, widely accepted distinction, at least with regard to ancient history, causes endless confusion. I've been pleading for years for some policy that would allow citation of primary texts (Bible, Assurbanipal's archives, Rg-Veda, the Sishu wujing etc. only if the citation is accompanied by a secondary scholarly source which uses that passage. There's an obvious reason for this. All such primary histories are themselves not 'reliable' history, but must be understood in terms of how the relevant scholarship reads each passage. Cheers Slim.Nishidani (talk) 18:02, 17 July 2016 (UTC)

I find I agree with Nishidani and Johnbod here - and would extend Nish's "ancient history" to medieval history also. Bede is a primary source - as is almost any medieval chronicle or writer. We should not be using them except as quotes, pretty much. Unfortunately, I get a lot of charters/chronicles/etc used as sources in medieval articles... And Johnbod will recall the Bulgarian controversy on Middle Ages, so even secondary sources aren't without problems. Who knew that there would be POV pushing in medieval topics! Ealdgyth - Talk 18:16, 17 July 2016 (UTC)

Hi Ealdgyth, good to have you here too. I wonder whether we could all put something together about this for the policy. SarahSV ^(talk) 18:54, 17 July 2016 (UTC)

Hi Ealdgyth, and welcome. Yes, Sarah, I think we should maybe put something together. I remember when years ago someone asked why in my refs I added "qtd in" to indicate that a secondary source was quoting the primary. That's really how it should be done. As Peter Damian mentions up page, we should be letting the scholars do the work for us. In terms of medieval chronicles, also mentioned up page, I'll use Froissart or Michel Pintoin but only if quoted in a secondary scholarly source. In fact even for more recent material, such as Hemingway, I very rarely use his letters but rather use what the scholars tell me in the secondary sources and quote accordingly. Nishidani, I did read the thread linked above using the "find" command for Nihon Shoki and scanned through. I saw, to paraphrase, that someone thought it was ok to use as a secondary Nihon Shoki because the "eyewitness" clause doesn't apply. But, in the many centuries since so many scholars have written and written again about the material, that we really should be using the most recent sources we can - all the time. I'd like to see something worked beyond what Nishidani wants: instead of adding a recent scholarly source alongside a primary, I'd say the primary should always be cited as "quoted in 'such and such'". Victoria (tk) 19:08, 17 July 2016 (UTC)

It was worse than that. The same editor objected to the Nihon Shoki as a primary source when I cited it, and then objected to it as a secondary source later. In both instances, he was drawing on the policy guideline. If a source can be classified as primary and secondary by the same editor, depending on his mood, then evidently the policy as written gives rise to confusions. I think there are enough of you to work out a brief line or two correcting our guideline to avoid such misprisions. Perhaps Sarah could make a proposal for fine-tuning?Nishidani (talk) 19:55, 17 July 2016 (UTC)

She probably could. What this discussion here has amply demonstrated is that there seem to be more than one approach to the primary/secondary issue (I thought it was very simple!), with a soupçon of "independent" and "third-party" thrown in to confuse matters. I'm not sure why it says at WP:Primary: "Primary sources may or may not be independent or third-party sources." That just seems to confuse the matter - either it's a primary document or not. Yes, I thought I was reading that the Nihon Shoki was characterized as both primary and secondary; then my head began to spin. Victoria (tk) 20:16, 17 July 2016 (UTC)

I agree about independent/third-party being confusing there. I'd like to see those words removed. SarahSV ^(talk) 20:59, 17 July 2016 (UTC)

I believe that those words are important, because some editors have trouble remembering that WP:Secondary does not mean independent. An eyewitness journalist is independent of a house fire; that same journalist's television report, saying "Firefighters were called to 123 Main Street tonight for a house fire" while standing on the street in front of the blazing home, is not a secondary source. WhatamIdoing (talk) 02:06, 23 July 2016 (UTC)

Perhaps I could simplify my point. Leave the larger issue of newspapers, and the modern world, and just add a specific line re anything down to medieval times, i.e.

'In topic areas regarding antiquity down to late medieval times, primary sources are the original works that furnish the material for modern scholarship, whose works are secondary sources. The former should, ideally, be cited through the latter.'

I'd better catch a movie. My prose is always suspect, no more than when trying to pretend I know how to make suggestions for a policy. Cheers. Nishidani (talk) 20:30, 17 July 2016 (UTC)

Enjoy your movie. I don't do policy either, but this has been an interesting conversation (albeit very long), and it seems important. Victoria (tk) 20:48, 17 July 2016 (UTC)

Nishidani, that's nicely expressed. I'd be happy to support it. SarahSV ^(talk) 20:57, 17 July 2016 (UTC)

I'd support that too. It is nicely worded. Victoria (tk) 23:25, 17 July 2016 (UTC)

I'd prefer something like: "Sources written before 1700, and works of history written before 1880, should almost always be regarded as primary, regardless of their distance from events. Where possible, they should only be used via references to them in modern secondary sources." Dates rather arbitary of course. Johnbod (talk) 01:29, 18 July 2016 (UTC)

That sounds generally better in wiki's voice. The dates are problematical, perhaps. Edward Gibbon's Decline and Fall, and Theodor Mommsen's Römische Geschichte as a primary sources? Nishidani (talk) 07:35, 18 July 2016 (UTC)

Gibbon's pretty much not a useful source any more - he belongs to the history of history, rather than being a historian in common use now. I wouldn't consider him a primary source but I wouldn't consider him a useful secondary source either. Too much has changed in what is available to the historian and too much has changed in how we approach history too. Unfortunately I don't have any better idea of dates to offer for "end points". I think I'd place it a bit earlier than 1700, though. Maybe 1500 or 1400? In my own "specialty" ... we start having good number of sources (for a medievalist's values of "good") in England around 1300. 1200 is still much better off than 1100, which is immensely better than 800. By 1400 we have letters, original documents, original deeds, etc, and no longer have to rely on stuff at a remove from the events they are describing. But I'm sure the dates for this explosion of documents is different for continental historians - I just don't know when it occurs. And of course, it'll differ greatly for other regions. I've only dabbled in Chinese or Japanese history, but I've read enough general historical works to know that the sources available for historians in those fields are of different orders of magnitude than what's available in medieval European history and that they are available in quantity for different periods. I'm not sure it's possible to write a simple "rule" that covers all periods and regions. I know as a medievalist in training, I was taught that "secondary" was anything produced after the Reformation, but that much of stuff between 1600 and 1900 was suspect, but not always - it depended on how much had been written about the specific subject more recently. When I first started in college, there was little in print on William II of England, we still studied Edward Augustus Freeman's work on Rufus (1882) because the only other book on Rufus was Barlow's recent (1983) biography of the king. Now, the only reason a medievalist would study Freeman's work on Rufus was to study how views of him have changed over time - we have many more monographs and journal articles available so there is no need to rely on Freeman's work. There are still some things where nothing recent (post-1900) is available so we must rely on pre-1900 works, which makes defining rigid rules difficult. I don't KNOW that this is the case in other regions/periods, but I suspect it probably is. Ealdgyth - Talk 17:18, 18 July 2016 (UTC)

Maybe "treated" rather than "regarded" as primary would be better - the point is that, however classified, they should not be able to be taken as RS without considerable scrutiny, including Mommsen, and indeed Freeman (who is popping up a lot just now at Talk:Anglo-Saxon settlement of Britain). There are (in some respects, not others) more early sources available in East Asian history, but if anything they need to be treated with more caution than those in the West, as more prone to push an official line that may bear little relation to reality. There must be very few areas indeed in "straight" history that have been so neglected by the scholarship industry in recent decades that anything 19th-century is still required reading at a Wikipedia level. In art history many early monumental surveys of particular types of object have never been redone - they tend to be things with ominous biblio details like "Leipzig, 3 vols, 1898-1907". Johnbod (talk) 17:44, 18 July 2016 (UTC)

The only possible examples I can think of are a few areas of family history/biography. J. Horace Round is still citable for some information as what he did hasn't been redone - but these are mostly for his monographs/articles. Some local history will still cite the early Victoria County History volumes. And of course, the Complete Peerage is still citable, but only to the 2nd edition which started in 1910. Some of the Sanctorum stuff is also still citable - since often no one else has done anything with some of the more obscure stuff. Some of the text societies stuff from the 19th century is still useful, too. I'd say a general rule of, for English medieval subjects, is most stuff published more than a 100 years ago is getting suspect, and 150 years is way outdated, in general. That puts us to 1866, which is a good safe date. I look closely at anything cited to much before 1900 as a general rule, trying to make sure it hasn't been studied more recently. Touching briefly on my other "specialty" - horses - generally anything about horse care or horse anatomy/physiology/evolution needs to be within the last 20 years, at best. Horse bloodlines, on the other hand, will very often require you to cite things that are very old. We routinely cite to stud book entries and newspaper accounts, which I would consider primary, because often there is not a lot of coverage in the secondary literature. You base the notability on the secondary coverage, but details of ancestry and race records are safe to source to primary accounts. Just shows how different things can be in different subjects. Ealdgyth - Talk 18:02, 18 July 2016 (UTC)

I agree basically with all of this. This is all very tricky, unless one just cuts the Gordian knot somewhere. I think the essential point is the juncture where historical writing becomes sufficiently self-aware of a metacritical dimension, which goes way back, perhaps to as you suggest, the Victorian period, when at least the basics of rigorous source analysis, archival investigation and textual criticism (although present by the Renaissance cf. Joseph Scaliger et al) more or less joined hands (Niebuhr, Leopold von Ranke and his 'nur sagen, wie es eigentlich gewesen', Lachmann) . On the other hand, the Kojiki is a primary source.There is a long commentary on it by a rough contemporary of Gibbon's, Motoori Norinaga. This is a wonderfully acute, if highly nationalistic, work. It would be a secondary source on the Kojiki in some sense, but no longer treated as such nowadays. These days we live in the aftermath ofHayden White's Metahistory, and the 19th century historians look far too embedded in tropes, in their age, to be of much value. Indeed metacritical history makes even last year's masterpiece look, if not dated, then framed within a tradition or school of thought.

My own suggestion would be to treat all pre-modern texts as primary sources, consider 16th-19th century historical works as secondary but mostly dated, and restrict historical sourcing to specialist books and articles produced within the last half century. My reasoning is that, in writing encyclopedic articles, we should try, wherever, to draft them on the basis what the modern, overwhelmingly post-WW2 academic or specialist literature argues. One example: there was a running POV battle over the word 'Palestine'. A standard meme says that it became current after the destruction of the Temple, the expulsion of the Jews, and the crushing of bar Kochba's revolt by the Romans. This is important because the name is intensely disliked in one variety of popular narrative which wants the native Biblical toponymy to prevail. Well that was corrected by citing its use in Herodotus's Histories. It's ancient. But you can't cite Herodotus alone, since the 2 passages have various interpretations topographically. You would once cite the primary source, Herodotus, together with Walter Wybergh How & Wells's 1912 commentary. Now, however a conscientious editor wouldn't use that because it is dated and rather would titrate Herodotus through David Asheri, Alan B. Lloyd and Aldo Corcella's A Commentary on Herodotus I-IV, (2007) aor specialist articles on the relevant passages (just to restrict oneself to the English-speaking world) which supersede Wells and How.

You can't expect most editors to be aware of this kind of thing. But a general hint to use Google Books and privilege those on the topic of history after 1970 would be enough to clarify how you can assure yourself you are getting quality sourcing.Nishidani (talk) 20:21, 18 July 2016 (UTC)

Yes, titrate is the right word. Basically if a scholar has done his/her job then the earlier works will be distilled and can be picked out from a more recent secondary source, if needed. And yes, basically I agree with all of this. I was thinking about this last night and my inclination is to push the cut-off into the 20th century. Anything before 1923 is PD and that might be a date we can use, or if we want earlier, then 1911 could work since much of WP was built on the EB 1911 edition in the early days. I don't do policy, so how to word it and present it beyond me, but I think it should be done. Furthermore PRIMARY should be simplified - the business of independent/third party should be removed. I'm almost inclined to suggest that all these policies should be more vague and less specific because the more specific they become the more they are cited with the relevant buzz words, but with zero evidence that they're really understood. I think the long ANI thread amply demonstrates that. As sort of an aside: I've read it. I've followed the links. I looked at the article in question, I've pulled sources, I've looked at the article history, I've read the talk page and ad nauseam. It represents a behavior problem that we see here all to often when an editor or group of editors spout policy without understanding it simply as a means to bludgeon in their edits. That's a problem I have no solutions to, but as an aside to Nishidani, if you do rewrite (my suggestion is to send it to AfD partially because articles such as List of National Treasures of Japan (archaeological materials) exist, are all feature lists, and include much of that material in summary form), I'd be willing to pitch in. Victoria (tk) 00:18, 19 July 2016 (UTC)

I am following this all very carefully, it is very interesting.

(from way above) Secondary source ascription for (1) historical documents, and (2) newspaper article on this week's car crash, are two extremes of dissimilarity.

Sorry, I have to disagree that "titrate" and "distill" are good words. These are school chemistry terms, used metaphorically, and are dangerous because many people know the chemistry but may not appreciate the metaphor. All metaphors break down on analysis. --SmokeyJoe (talk) 00:35, 19 July 2016 (UTC)

I like titrate, but you have a valid point. We do have to be careful with our words and metaphors. About your point re two extremes of dissimilarity: isn't that one of the issues here? We cover such an enormous range, from classical history to updates on the most recent news story, that I think it's really difficult to pin down. Unless I'm misunderstanding you? Victoria (tk) 01:53, 19 July 2016 (UTC)

I myself have learned to keep these sorts of metaphors mental-only. I like them too, not "reasoning" by analogy, but exploring by analogy, brainstorming with analogy, but have learned that explaining the metaphors/analogy to someone unfamiliar with the main concept is unhelpful.

The extremes of dissimilarity, I think we are on the same page. I can see very easily that some of the disagreements on this page can be ascribed to one person thinking of historical documents, and the other thinking of current journalism. I think this goes to my original point on your talk page, that historiography and journalism have points of conflict with language, with definitions. --SmokeyJoe (talk) 02:45, 19 July 2016 (UTC)

My main real-life interest is in analyzing how a poem functions, Joe, which means I think mostly in terms of analogy and metaphor. I rarely edit on the subject, because it is not what encyclopedias do: it is language used in its most extreme and compact sense of semantic depth: the reader throws his lures, like a trout angler, around those rock ledges where the object of his quest survives elusively under the still waters of the stream of consciousness (the historic resonances of words) to flush out and fish up the poet's intentions. I switch off this mode of thinking when editing, but not wholly if I have to argue on a talk page. 'Titrate' was loosely used of the process whereby a scholar takes a thick composition of words, like a primary text, and by using all of the analytical agents at his disposal, tries to get the work to yield up the core substance of facts from all of the other elements that are bound into the texture of what he studies. Perhaps 'leach' would have been le mot juste, but in the quick thrust and parry of talk page exchanges, one doesn't always 'strain' after the perfect speech form. If one did, no would be writing a poem, which is in the public sphere, a game of dare-devilry with the linguistic complacencies that inform conversion, and journalism. I concur that the issue is basically the difference between journalism and historiography. Everything breaks down under analysis, even ourselves! Nishidani (talk) 07:15, 19 July 2016 (UTC)

I too very much enjoy analysing with analogy and metaphor. My quick 00:35 comments, which looks in hindsight negative and terse, comes from my experience of failure when trying to use an analogy to explain something. Your last sentence made me smile, so true. --SmokeyJoe (talk) 07:33, 19 July 2016 (UTC)

I like the word titrate to describe how primary sources should be viewed through the lens of the secondary literature, and I'd like to retain Nishidani's phrase that "primary sources are the original works that furnish the material for modern scholarship."

SmokeyJoe, on what do you base your argument that historians and journalists approach primary/secondary any differently? They really don't so far as I can see. Primary sources are the original works that furnish the material for journalism too. SarahSV ^(talk) 21:59, 19 July 2016 (UTC)

I too like the statement "primary sources are the original works that furnish the material for modern scholarship."

SarahSV, that question isn't quite right. There is a difference between historians and historiographers, and my view on journalists, or rather on journalism, is that they do not formally approach or use the term "secondary source". In fact, it seems that only in historiography is "secondary source" well defined, and it is as per our article secondary source.

On the difference between historians and historiographers, historians have and use the facts, and historiography comes to the fore where the facts are sparse or nearly all lost and they scrounge the best as can be done. Historiography is about the process, source analysis of non-ideal sources, and carries the assumption that the facts per se can never be known.

Journalists, as I understand them, are far more focused on source reliability. --SmokeyJoe (talk) 22:46, 19 July 2016 (UTC)

SmokeyJoe, can you say what makes you believe that journalists and historians use or categorize sources differently? This is an argument you've made for years during these discussions, but you haven't said what it's based on, or why it would matter. (I'm not sure what you mean about journalists focusing on source reliability. Journalists and historians look for reliable primary sources.)

All formal students of journalism are taught the basics of reporting, source use and fact checking. People who enter journalism with university degrees in other areas will often come from history, classics, philosophy, English, and will be familiar with source categories. What is it that makes you think they handle source types differently once they start practising journalism? If you could say what you see as the difference, and where you're taking it from, we might find it easier to reconcile our positions. SarahSV ^(talk) 23:24, 19 July 2016 (UTC)

SarahSV, my beliefs about journalism have largely developed from talking to you, and then offline trying to work out why you would say seemingly unexpected things. My working hypothesis is that you have a journalism background. I do not, and I would not tell you how to write about current affairs, or biographies, the sorts of articles you write so well. I know a little more about historiography, and note that that historiography has easier to find well-defined terms, and assert (as a matter of choice) that the encyclopedia is an historiographical document.

I know little about how academic journalists categorise sources, except that I cannot find any reference for academic journalism defining the term "secondary source".

The impetus for my comments on journalism comes from having discussions assuming historical documents, clearly in the historiographical in nature, and then you might say something like "The discussion here was about whether a (say) 2016 newspaper article about a 2016 traffic accident, written by someone not involved in it, is a primary or secondary source for that event", for which much of the discussion is inapplicable. Thus, I began to consider that journalists, if they truly use the term "secondary source" (do they??), that they seem to be using it differently, and extended form, compared to historiographers. I am wondering whether source classification advice should be divided, advice for historical sources, and advice for news sources. I which to avoid the term "historian" in favour of "historiographer", for the point of clear contrast, as a historiographer is far more dissimilar to a journalist than is an historian. --SmokeyJoe (talk) 23:45, 19 July 2016 (UTC)

Okay, thanks. Well, I have never said or implied that journalists use sources differently from historians, so if I'm your source, we can put that to rest. (I'm also not sure what you mean by historiographer rather than historian; historiography is just the study of history (meta-history).) Journalists and historians don't use primary/secondary differently. The difference between history and journalism is that journalists write about current affairs and therefore rely on more recent sources. They will also tend to use eyewitness accounts more, whereas historians prefer documents. (Journalists like documents too: they're just hard to get hold of for current issues.) SarahSV ^(talk) 00:12, 20 July 2016 (UTC)

"Well, I have never said or implied that journalists use sources differently from historians"

No you didn't, but I have inferred that "journalists use classify sources differently from historians".

Historiography, which I take to mean history scholarship where reliable sources are scarce, I do think classify sources differently to journalists working on traffic accident on a daily time scale, because the sources are very different in nature. --SmokeyJoe (talk) 02:17, 20 July 2016 (UTC)

I don't know what "history scholarship where reliable sources are scarce" might refer to. Historiography is just the study of history, the way it's written. The historiography of a subject is the way it has been written and sourced, which issues have been included, which causes identified, etc. SarahSV ^(talk) 02:41, 20 July 2016 (UTC)

Analysis of the uses and depth of the collection of the Ancient Library of Alexandria, or of the contents of the Q source, are examples of historiography that should not be called "history", because it is speculative modern opinion. (somewhat speculative, but not mere speculation). We don't have the library catalogue, or a copy of the document. The Q source is not even known for a fact to have existed. --SmokeyJoe (talk) —Preceding undated comment added 03:46, 20 July 2016 (UTC)

Break six

In journalism, how are secondary sources explicitly used? In network level television news, concretely—a sound bite (with image) from a recognized expert, taken from an interview. News bureaus will be tasked to interview experts located in their geographical areas to support a story that may originate elsewhere. In some cases more than a sound bite will be transcribed to use as research for the script. Sometimes the interviewee is knowledgeable as an academic, sometimes through experience (professor of international affairs vs political operative). As an aside, the vast majority of those with journalism degrees don't work as journalists. — Neonorange (Phil_tk)

Thanks, Phil, those are good examples. It might be worth quoting Willie Thompson, former professor of history at Glasgow, at length on the primary/secondary distinction (What Happened to History?, 2000, 6–7). Note: I've added two paragraph breaks:

"The first and central principle is that the past must be reconstructed, retrieved or whichever term one wants to employ, only on the basis of evidence which was generated (by human agency or otherwise) at the time in question or by individuals who participated in events under scrutiny. These are termed primary sources, generally written evidence. Historiography is distinguished by the key significance it assigns to the documentary record or written sources – as distinct from other explorations of a historical nature, especially archaeology. (The distinction of course is artificial so far as the content of historical development is concerned; it is purely one of method and in many instances documentary and archeological evidence can be profitably combined.)

"By contrast, secondary sources are accounts which have been compiled on the basis of primary (or original) sources – or of other secondary sources – and cover everything from the most profound exposition of new knowledge to the routine monographs and articles written by historians, and at the further end, the merest ephemeral trivia.

"Not that the status of a primary source is by any means clear cut. As an example we may take the gospel narratives of the New Testament, composed at least several decades after the events they purport to record, by individuals who lacked first-hand knowledge of them. These represent primary evidence for the outlook and thinking of the early Christian communities among whom they were compiled, but only secondary evidence for the personality and career of Jesus of Nazareth. In many instances of course, particularly for remote areas, secondary sources are all that historians possess and they simply have to make the best of them. In those cases even where the primary source does exist (again the biblical texts are a good example) the original version of the document may very well not do so (the Dead Sea Scrolls are a renowned exception) and what remains extant are copies of copies of copies.

"Leaving the remote past for any other century, including the twentieth, the same principles apply no less. Newspaper reports, having been subject to editing, are secondary sources for the events they record; the original reporter's notebook, if it had survived, would be a primary source. A document can, as we have seen, be in different dimensions, either primary or secondary. The notoriously mendacious history of the Soviet communist party compiled in 1938 under Stalin's editorship could scarcely qualify even as a secondary source for the CPSU's actual history, but it provides an indispensable window into the Stalinist mentality."

SarahSV ^(talk) 16:06, 20 July 2016 (UTC)

Lucid, Sarah. But I think the key word in the citation is 'artificial':Willie Thompson makes a distinction between historiography and other disciplines used to reconstruct the past (archaeology) only to undermine it immediately (proof of a fine mind, rules are heuristic not set in stone) that we are dealing with a rule of thumb. His definition is one that reflects a the working principles of a specialist in modern history, where you have contemporareity or a brief time gap between events and reportage (archives etc). It doesn't quite work for ancient works - he cites the NT and the point is correct, but it does not alter the status of the NT as, for editors here, a primary source, as is all of the Bible, and all ancient historical works. Genesis (primary source) mentions a Tid'al, 'king of the nations' (ṯiḏ‘āl meleḵ gōyim) whom some consider a reflection of one of the Hittite Tudhaliyas, attested in cuneiform archival correspondence - primary sources, written at least a 1,000 years before the final recension of Genesis, which though for us a primary source is, like the NT, a 'secondary source' according to the common distinction Thompson makes. If so, contemporary scholarship must be shunted aside from its secondary source category, since it is a comment, on a secondary source, that in turn derives from a primary source.

I think it important to keep in mind that any rule is a guideline for general editors who may not have, in the majority of cases, do not have that acquired 'nose' for all these refined distinctions and all their qualifications and, when they seek illumination in written policy, derive different interpretations as the polic y now stands. When I spokes of cutting the Gordian knot, this is basically, (it might sound sniffishly condescending to many -stiff cheddar!) I had this in mind, and suggested something along the lines of:'All written works surviving from antiquity are primary sources, even if they draw on other, lost, sources. All modern scholarly works on them are secondary sources. (The former should be cited via their use in the latter)'

I've changed 'commentaries' to modern scholarly works, because there are a lot of popular books that draw on secondary scholarship but are written by authors who read the secondary scholarship and present it for a general readership. In short we must keep uppermost in our definitions the requirement that general editors be given clear guidelines. It would avoid a lot of talk page argufying and board disputes. Best regards.Nishidani (talk) 19:21, 21 July 2016 (UTC)

I wrote Virginia (Virgin=Sarah) for Victoria, confusing two editors. Apologies.Nishidani (talk) 19:26, 21 July 2016 (UTC)

I agree with Nish (sorry if that contraction offends, please let me know so I can stop if it does) here - Thompson is writing mainly from a modern historian's perspective. Unfortunately, that lovely overload of sources isn't available for the ancient historian or the medievalist (or for many non-European specialists either). Projecting a modern view of primary/secondary into the ancient world is one of the main problems we face as wikipedia editors - it's one of the problems Nish faced with the Korean influence on Japan spat that blew up recently. Attempting to use something like Bede or Orderic Vitalis as secondary sources is not something that really fits with the spirit of the primary/secondary usage on Wikipedia. For Wikipedia purposes in history, a source we should be using is ideally a professional historian (or independent scholar) from the last 100 years or so (the boundary is fuzzy, but a good safe choice) ideally published by an academic press or highly reputable mainstream publishing company. We should not rely on mainstream press sources or similar - because they aren't really trained as historians, but as journalists. What to do about historical societies, etc is more difficult, but in general they shouldn't be used unless there is no academic scholarship available. Professional historians will occasionally put out useable work in non-academic presses - I'm thinking the Oxford Dictionary of National Biography or similar specialist encyclopedias. Of course, this is going to only apply to things classified as "history" - journalists are much more useful for something recent. (This would have the advantage of helping weed out some of the worst examples of fancruft/trivia - Sarah will probably remember the FAC on a model that was filled with trivia because some editors think of newspaper/media accounts as "secondary" and thus every. single. thing. in. the. media. must. be. in. the. article. ) Ealdgyth - Talk 20:20, 21 July 2016 (UTC)

Touching on an earlier point, I'd really prefer to avoid any reference to Google books - too many people think they can research articles only using Google book snippet and/or partial book searches. We really should discourage it's use as much as possible. Ealdgyth - Talk 20:20, 21 July 2016 (UTC)

I think we could use the Thompson's quote Sarah posted as a starting place for a wholesale rewrite of WP:PRIMARY and WP:SECONDARY. I think those two policies need to be looked at word by word, to decide what stays in, what is tossed, or whether to begin anew, or whether to do nothing. I can't speak to Nishidani or Ealdgyth's points about ancient history - not my area - but I think their points underscore the differences in disciplines. When I write about literature I always use a., the most recent scholarly sources (academic presses and journals), and b., the critics who have done the most work in that field. You do get a nose for these things and it becomes second nature. What's confounding me is that I can't think of a way to present it as Wikipedia policy. The analogy I keep coming up with is that it's like trying to teach someone how to drive a standard shift car; I know how but can't put into words what's necessary to keep the car from stalling. Equally I know exactly what to do when I begin to research an article, but it's instinctive and I think hard to teach. But - it is important that we teach it. What struck me about the Korean influence article is that that Farris's book is a good source, and I suspect Donald Shively would be too (though I didn't bother to look), but on Jstor there is a journal article that does seem to be perfect for that page. Closer examination shows it's not at all perfect, but that's where the disconnect lies, because the Jstor article satisfies policy. Re the recent FAC about the model - I commented on the most recent iteration. Re google books - I wish we could get rid of them. But that would be worth another long thread. Victoria (tk) 21:01, 21 July 2016 (UTC)

The difficulty with writing policy is that it has to be understood by editors who have no relevant background and therefore haven't developed a feel for what is primary and secondary. So the policy has to spell it out, briefly, and has to cover everything, from ancient history to articles about models. It isn't easy to do that.

Nishidani, Thompson's point that the gospels are primary sources for certain aspects but secondary accounts for events about which the authors had no direct knowledge, is a crucial one. We treat the NT as a primary source without qualification because of its proximity to the events for which we use it, relative to us, and that means we use it through the lens of (titrate it through, as you said) modern secondary scholarship. But the question arises nevertheless, of any given event: does that source offer a firsthand or secondhand account? This discussion started because of claims that newspaper articles in 2016 about current affairs, written by people not involved in the events, are primary sources, and that's incorrect, as Thompson notes. The older the source, the more the firsthand/secondhand distinction breaks down, because of the source's distance from us. Our position in time is part of the equation. SarahSV ^(talk) 21:45, 21 July 2016 (UTC) (clarified SarahSV ^(talk) 06:01, 22 July 2016 (UTC))

• The WP:NOR policy, as SarahSV mentions, needs to be written for the generic editor, including newcomers to the project with no experience in the topics they edit. Really! WP:NOR is one of the first pages that newcomers are urged to read, especially newcomers that are adding material poorly. It is a poor document for expert advice on nuances applying to narrowly defined topics, though its statements much be correct at multiple levels. Currently, I think the policy is pretty good. What problem is being alleged? Should the nuances of source use in ancient history, remote regions, modern history, current affairs, and BLPs, be detailed separately?

• The way I look at the gospels and a second hand account of a 2016 traffic accident is that for both it is a similar matter of source classification. Neither are a primary source to a scientist. Both are second hand accounts (of Jesus, of the driver's actions) to a journalist. Both could be primary/secondary depending on usage, but this "usage" out involves a dramatic change of perspective. The Gospel is always a secondary source on Jesus, but a primary source on 2nd century Christianity, and these are two separable topics (arguably, ignoring bible literalist believers). The accident report is secondary if a summary/retelling of a primary report, but is primary if the topic is 2016 attitudes to car accidents.

What I occasionally wonder is "why do we care?" Is it for a WP:GNG-based deletion decision? A WP:DUE decision? Is it for a decision on how to choose the best sources to make the best content? These are different questions. Is there confusion along the lines of "primary=bad; secondary=good"? I think WAID has written an essay on that, and I think WP:NOR/WP:PSTS does not imply it. Obviously, no secondary sources is bad, at the extreme, but the policy does not mandate a ratio.

Also, does primary/secondary source classification distract from the question of reliability? I think I observe that the contorted examples of second hand primary sources, which I maintain theoretically exist, should be tossed aside as unreliable. On the topic of Jesus' childhood travels to Egypt, is the gospel a primary source, or a secondary source? I think it is an unreliable source, making the primary/secondary aspect irrelevant. Extraordinary care is needed to write acceptably the article covering it, Flight into Egypt. --SmokeyJoe (talk) 02:17, 22 July 2016 (UTC)

A few small points: You're thinking of the WP:NOTGOODSOURCE section of WP:USEPRIMARY; NOR actually does mandate a ratio of secondary to non-secondary ("Wikipedia articles should be based on reliable, published secondary sources and, to a lesser extent, on tertiary sources and primary sources", which means >50% of the article should be based upon secondary sources; yes, fighting over primary vs secondary distracts editors from thinking about reliability, just like thinking about reliability in a general sense ("this kind of article is reliable") distracts people from thinking about reliability for a specific statement ("this exact source is reliable for this exact sentence"). WhatamIdoing (talk) 02:38, 23 July 2016 (UTC)

A simple rule would be. Fuck the distinctions - in citing anything premodern, any primary or secondary source, NT/Bible/Sima Qian etc., must be accompanied by a scholarly work that discusses it. That would immediately oblige work on several thousand articles or sections where Josephus/the Bible/Livy/Herodotus/Rg Veda/Kojiki/Nihon Shoki etc.etc., are referenced.Nishidani (talk) 10:00, 22 July 2016 (UTC)

Nishidani, yep, I wholeheartedly agree.

SmokeyJoe I think WP:NOR can be streamlined, updated, and made more user friendly. Instead of quoting its lead here, for brevity's sake will just say that if a new editor were only to read those first few paragraphs and not go beyond, the impression they'd gain is that any material added to Wikipedia has to attributable to a reliable source. So, ok, I have a couple of examples: take a look at the references for Mayflower Compact. There is a book published in 2006 but it's self-published, and there is a book published by a reliable publisher in 1966. The rest of the sources are 19th and in some cases 17th century. I chose this article because it's part of a suite (have a look at the nav template at the bottom). Almost all the articles in the suite are the same. As it happens for years that area was a walled garden of sorts and though I used to keep some of those articles on my watchlist, long ago took them off, although I would like to work in that area. It's in important period in America's history. How many such walled gardens exist? And how many of our 5 million plus articles are sourced to pre-WWII sources? My sense is that with the increased availability of digitized materials, either through HathiTrust, Internet Archive, or simply g-books, that we'll being seeing more rather than less of that type of referencing. I can give many many other examples, i.e, when last year I complained about an article trying for FAC that relied on 19th century newspaper coverage (my complaint wasn't well received, to say the least). Again, this is just my sense, but I think we're at the phase of the project where adding references, swapping existing references, is of more importance than creating new articles. We've been through the stage when lots of articles were created without referencing, or low quality references, and I strongly believe that it behooves the project to strengthen the referencing. This is easily achieved with the addition of a single sentence to NOR stating that sources published before World War Two are considered primary for Wikipedia. I've never edited a single policy page, rarely speak up, but since you all have landed here on my talk page, those are thoughts I've had for a long time. I not only think but know that our shoddy referencing weakens the project. And, as Nishidani says, asking for modern sources would oblige work on many articles, and it would have the benefit (one can only hope) of cutting down on endless talkpage discussions about whether a source is or is not primary. We've all seen these types of discussions spin out of control, become a timesink, add to editor attrition, get people blocked, etc. All that said, (and I'm aware this is a mini rant), I think getting such a small change through would be extremely difficult - which is depressing, because it makes so much sense. Victoria (tk) 23:34, 22 July 2016 (UTC)

Changing NOR will not solve any problems easily. At best (partly because WP:Nobody reads the directions, partly because it's always a matter of whose ox is getting gored), you could hope that in a couple of years, the general belief about the rules will have changed. WhatamIdoing (talk) 02:38, 23 July 2016 (UTC)

I've collapsed my comment. I do think we can do a better job, but I'll keep to content creation and butt out of making suggestions for policy changes. Clearly I'm lacking in that respect. Victoria (tk) 03:46, 23 July 2016 (UTC)

Break 7

IMO the main problem with using Willie Thompson's idea is that his claim that newspaper reports are secondary basically puts him in the tiny-minority camp. I further doubt that even Thompson would agree that every single thing in a newspaper was secondary – especially if he realized that there's no such thing as the "original reporter's notebook" for many common types of stories. "Car wreck without injuries on Main Street" often gets written (or filmed) directly, without anyone bothering to first find pen and paper and write down what they heard over the police scanner before typing it up and submitting it for editorial review.

The Ithaca College source that Sarah provided earlier is a much more typical POV: a 20th-century news report should be treated as a primary source, not a secondary one. WhatamIdoing (talk) 02:39, 23 July 2016 (UTC)

That begs the question of what Wikipedia is: an online repository about car crashes or an online repository about knowledge. Furthermore, in my experience outside of Wikipedia Thompson is in the majority. I've honestly never heard of a newspaper article as a primary source in the way it's currently described here, and my experience extends beyond this place. We could make that rule; apparently we have; but I think it's detrimental and muddies the waters considerably. Victoria (tk) 03:46, 23 July 2016 (UTC)

That would depend upon the car crash. We certainly do include information about some car crashes. It would presumably be nice to have those supported by sources that engage in analysis and critical thinking, rather than police blotter reports – which, as the name suggests, really are just copies of what the police sent the newspaper, reproduced without even any effort to fact-check them, much less to think about the contents or transform the contents into something original. I still believe that a copy of a primary source, even if it has been re-typed and re-formatted to fit into newspaper columns, is still a primary source. WhatamIdoing (talk) 20:48, 2 August 2016 (UTC)

This project, unlike the print encyclopediae of the past, has no editorial structure (WP:NOR, anyone can edit, etc.) Thus the emphasis on secondary sources. To deem 'news reports' primary and deprecated has existential implications for Wikipedia. The focus should be on choosing reliable sources. An editorial process produces a secondary source. How to divide reliable from unreliable? Quality of editorial process. Same holds true of any secondary source—academic journal, book, or news story. Neonorange (Phil_tk) 03:18, 1 August 2016 (UTC)

Is it your contention that peer-reviewed journal articles that report original experimental results are all secondary sources, or that they do not have an editorial process? Because those are primary sources, and if you believe that a good editorial process produces "secondary" sources, rather than "reliable" ones, then logically you must either agree that all scientific journals have no editorial process, or that your assertion is wrong. WhatamIdoing (talk) 20:50, 2 August 2016 (UTC)

In thinking about your reply, WhatamIdoing, I'm moved to ask how you define 'editorial process'. Here is how JAMA explains their process to authors.

JAMA' editorial and peer review:

...evaluated according to the following criteria: material is original and timely, writing is clear, study methods are appropriate, data are valid, conclusions are reasonable and supported by the data, information is important, and topic has general medical interest. Other manuscripts are sent to expert consultants for peer review...

JAMA editing:

Accepted manuscripts are edited in accordance with the AMA Manual of Style, 10th edition, and returned to the corresponding author (or his/her designee) for approval. Authors are responsible for all statements made in their work, including changes made during editing and production that are authorized by the corresponding author.

Is this an editorial process that transforms the material (which has been mentioned earlier in this thread)? My view is no, there is a gatekeeper function but no substantive transformation of the original paper. No material not from the original experiment is added. The paper is still an original report from the original authors of the original experiment—a single source report.

This, of course, brings up news stories, and the 'police blotter' report (which I consider a straw man). Wikipedia should not use a 'police blotter' report for BLP, it is not an indication of anything except an arrest, which may or may not be valid. If an arrest is of genuine interest, then there will be a article that is a secondary source. And the same applies to the 'standing in front of a blazing fire' story. — Neonorange (Phil_tk) 16:46, 3 August 2016 (UTC)

The peer-review process sends one or (usually) more written reports back to the authors, with demands that the paper be changed to meet the reviewers' expectations. If you spend an hour reading what authors say about these reviews, it will more than rival any anything you've heard about newsrooms, and it has a more direct and significant effect on the source's contents. (This may partly be due to the anonymous and semi-random nature of reviews; as a journalist, you get to know your editor's expectations [so you usually meet them in advance], and you've generally gotten approval to write the story in advance.) JAMA also downplays the main role of the editors, which is to reject papers that they don't want to run.

The police blotter is not a straw man: we have had editors claim that everything in a newspaper is a secondary source. If "secondhand" is the definition of "secondary", then "Drunk driving arrest in the 100 block of Main Street", copied straight off the police blotter, is "secondhand" information and therefore "secondary" (according to that inadequate definition). I don't agree that it's secondary, and it appears that you don't agree that mindlessly repeating secondhand information from a police blotter, phone call, or press release makes a source be secondary, but this position does appear to still be held by a few editors.

If we could all agree that at least some things in a newspaper are primary, and that at least some other things in a newspaper are secondary, then I would consider that to be progress. WhatamIdoing (talk) 20:25, 3 August 2016 (UTC)

WhatamIdoing, it seems to me that a paper presented in scientific, peer reviewed journals consist of two parts; observed data and conclusions, both from the same authors. Is the peer review process transformative? Does it change the data points? Does it change the conclusions? On the other hand, for journalistic product, the editorial process can add additional information gathered from additional participants, experts, public records, etc. This process can use the organization's archives to validate and expand. It can use access accumulated through being a journalistic organization. The resulting story will not necessarily resemble the original reporting (which can be divided among correspondents).

Does the hard science vs humanities divide plays a role here? If it does, there will be a spectrum: Physical Review → JAMA → The Southern Review on one hand and The Oakland Tribune → CBS News → The New York Times on the other. (The examples shown here are illustrative, not definitive).

"Mindlessly repeating" a police blotter is not a product of journalistic endeavors with a strong editorial process. A television network news organization does not have the space for filler, neither does a national newspaper. Part of the editorial process is access to information that is 'not' on the public record,

Unfortunately, rules aren't a substitute for good judgement. Perhaps examples of specific documents that are recommended sources will help editors who want to learn. For those who don't, even iron-clad rules will be of limited value. — Neonorange (Phil_tk) 22:15, 3 August 2016 (UTC)

If your peer-review reports insist that you change your data points, then it's called fraud – every bit as fraudulent as a television interview being edited so that a person who said "twenty-seven thousand" is presented on air as saying just "twenty". On the other hand, demanding that you consider and cite other sources, that you re-test your samples, that you run additional experiments, or that you reduce the scope of your conclusions to what the peer reviewers believe can be justified by your data is just normal.

"Mindlessly repeating" the police blotter and other basic facts in which the public has a legitimate interest is an important journalistic endeavor. It's not glamorous, but it's important. Stories like "Here's why you thought you heard gunshots last night" are literally bread and butter that make it possible for journalists to do the occasional glamorous story. What matters for an article is whether that article is a "reliable" story (for whatever sentence it's supporting), and, for our discussion here, whether that is "primary" on the grounds that no transformative thought went into it, or "secondary" or even "tertiary" on the grounds that the journalist is repeating information third-hand (i.e., from the bystanders to the police, and from the police to the journalist). I say it's primary. What would you say? WhatamIdoing (talk) 17:56, 6 August 2016 (UTC)

• WhatamIdoing, you wrote above: "a 20th-century news report should be treated as a primary source, not a secondary one." That is a mistake (unless it's an older source; see above for the age factor). The Ithaca College source that you yourself cite says: "A primary source provides direct or firsthand evidence about an event, object, person, or work of art." That is the issue, firsthand or secondhand access.

On Talk:Thomas Pogge, you implied that anything that simply reports facts is a primary source, [2] and wrote that the Buzzfeed article that first reported allegations about Pogge is a primary source. [3] While it may not be an RS, that article is clearly a secondary source. It is reporting the primary sources.

Had there been a dispute about the Buzzfeed article (for example, if someone had complained that they had been misrepresented), Buzzfeed's editors and lawyers would have asked to see the primary sources: the reporter's notebook, any audio or video recordings of interviews, signed statements, legal documents, emails, original tweets, Facebook and other posts by those involved, the first posting of the academics' petition (if that is mentioned in the story; I forget whether it was). Those are the primary sources.

The article that was based on those primary sources was written by someone who had had only secondhand access. It was subject to the reporter's view on how the primary sources should be combined and summarized; the newsdesk's view on how the article should be copy edited; and the lawyer's view on what should be left out. It is a secondary source. SarahSV ^(talk) 09:24, 1 August 2016 (UTC)

If "A primary source provides direct or firsthand evidence about an event", then don't you agree that "Here I am, a local 'news personality', standing in front of a blazing home and telling you that it's on fire" is "direct or firsthand evidence about an event"? Or "Here I am, a professional reporter attending a political debate, and telling you that I saw this Republocrat smack that Demican with his fist"? Do we at least agree on that much? WhatamIdoing (talk) 18:40, 2 August 2016 (UTC)

WhatamIdoing, the only way to explain this is that the person who stands in front of the house and reports on the fire, or the person who made their way to report on the Normandy Landings, or the person who puts him or herself in danger whenever reporting from a warzone, about an earthquake, on politics, whatever it might be, as opposed to the bystander who "reports" via Twitter, iphone, what-have-you, is trained, has agreed to a code of ethics, works for an organization that follows a code of ethics, etc., etc. Basically WP:Primary is written as it is; change appears impossible. Some people who've commented here disagree with Primary, but what you find here is a hardcore group of non-paid volunteers who tend to dig into content and have a good grasp of sources. My sense is that when it's impossible to find a reference for something - i.,e a gold standard that newspapers are always primary, regardless of point of time (i,e today, last century, a millennia) then it probably doesn't exist. We've discussed how this situation is treated in various professions and on Wikipedia. Generally it's standard to accept recent reporting as secondary. With the passage of time a secondary source will eventually become primary. In the meantime, I've been fairly involved with content work and thank you and all for commenting, apologize for replying late and so on. I do think I need my page back so that we can hash out what needs to be hashed out content-wise, so, if you haven't any objections or last comments, I'll be archiving this later today. Victoria (tk) 12:48, 7 August 2016 (UTC)

Last comment: A source – any source, gold-plated or not – that actually says "An eyewitness newspaper story written yesterday is a secondary source" is exactly what we've never found. (Even Willie Thompson, who falls into the tiny-minority camp by claiming that century-old newspaper stories are secondary sources, never addresses the question of the eyewitness reports or stories for which no 'reporter's notebook' ever existed.) We've got sources that say "Newspapers are primary" (yes, exactly the opposite of what some editors believe), and we've got sources that say "All old newspaper stories should be treated as primary" (which all editors agree upon). Some editors take this last phrasing and assume that the converse is also true (that every single item in today's newspaper is secondary). We don't have a single source that actually says that. We only have editors' assumptions that if a source says "all old ones are X", then "all new ones must be the opposite" rather than "all new ones are also X (and we didn't mention it, because our focus was old ones)" or "new ones are complicated (and we didn't mention it, because it's too complicated to explain)".

And: Thank you very, very much for hosting this conversation. I'm sure that it's been wearing to have the "new messages" button light up for weeks and weeks now. But I deeply appreciate this opportunity. May I suggest a separate archive page for this, both due to length and so that it will be easier to link to it later, if someone wants to? WhatamIdoing (talk) 08:01, 8 August 2016 (UTC)

I found [Research101: Primary and Secondary Sources] at the Ithaca College Library website.

Primary Sources

A primary source provides direct or firsthand evidence about an event, object, person, or work of art. Primary sources include historical and legal documents, eyewitness accounts, results of experiments, statistical data, pieces of creative writing, audio and video recordings, speeches, and art objects. Interviews, surveys, fieldwork, and Internet communications via email, blogs, listservs, and newsgroups are also primary sources. In the natural and social sciences, primary sources are often empirical studies—research where an experiment was performed or a direct observation was made. The results of empirical studies are typically found in scholarly articles or papers delivered at conferences.

Secondary Sources

Secondary sources describe, discuss, interpret, comment upon, analyze, evaluate, summarize, and process primary sources. Secondary source materials can be articles in newspapers or popular magazines, book or movie reviews, or articles found in scholarly journals that discuss or evaluate someone else's original research.

There will be a [quiz]. — Neonorange (Phil_tk) 08:49, 11 August 2016 (UTC)

Note: I've moved most of the discussion to archives but decided to keep the last section here because I've not had time to read through. Once I have, I might make some more replies. Victoria (tk) 19:57, 12 August 2016 (UTC)