Wikipedia talk:Classification of sources

Very rough start
This is currently a very rough start to get a basic idea, and it needs a lot of work. Be bold, or discuss here, as you feel appropriate. SamBC(talk) 02:39, 15 August 2007 (UTC)

Distinction based on what the author does --> how the author is situated
I think that defining primary sources as being "experience based", and secondary sources as being "interpretive" is pretty close to the way the terms are generally used, but not quite. I'm coming at this from a generally historiographic point of view, and I think that the main distinction historians make has to do not so much with what the author is doing as how the author is situated.

There are many instances, where a primary source author is doing something interpretive. For example, an eye-witness to a bank robbery might say that his interpretation of the robber's behavior leads him to think that he was probably drunk, and a foreigner. That's still primary source information. Similarly, a scientist may relate her scientific conclusions based on reams of data: certainly the data would be a primary source, but so would her observational conclusions based on the data and her general expertise in the field, as reported in a journal article.

I would propose that the distinction be thus:
 * primary source material is information related by someone who was in a position to authoritatively know the facts, either from personal observation or from some personal, contemporary connection with someone who was in a position to know the facts.
 * secondary source material are conclusions and interpretations by an author who was not directly and personally connected to the facts.
 * tertiary sources are secondary sources such as textbooks, treatises, and dictionaries that are intended to be balanced and authoritative overviews of primary and secondary source material.


 * in "hard" (as opposed to social) sciences, which is where my experience mostly lies, a journal article wouldn't generally necessarily be considered a primary source, in my experience. We need to find a balance that is equally applicable across all fields, even if it disagrees with the way some (or all) fields use the term, if we want a domain-specific definition for the whole of wikipedia (which is the point behind trying to do this). SamBC(talk) 23:50, 15 August 2007 (UTC)


 * By not necessarily being a primary source, I mean it depends on the content of the study or work document. Postulative papers would be primary, while conclusive papers would be secondary, and everything in between open to case-by-case interpretation. The terms aren't widely used in hard sciences anyway, this is partly a backward interpretation of the meaning attached to the terms. SamBC(talk) 23:52, 15 August 2007 (UTC)


 * I've never really heard the terms used in the context of hard sciences, so I don't know what the convention is, or if there even is a well-defined convention. Do you know of any sources we could look at that terms in relation to hard science? Maybe the distinction is that a journal article acts as a primary source when it relates experimental results, but as a secondary source when it makes conclusions based on those results that anyone else looking at the data could also have made. As to papers postulating a new theory based on prior published data, they would be primary sources to the extent the citation concerns the theory, but secondary sources to the extent the citation concerns the author's review of the prior published data. CO GDEN  22:04, 17 August 2007 (UTC)


 * That sounds about right. I don't know any sources, and the terms aren't widely used (IME) but the concepts are used under any number of terms. It's sort of implicitly discussed. I'll keep my eyes peeled for some sort of relevant source, though. SamBC(talk) 22:09, 17 August 2007 (UTC)


 * I just found an online source describing the difference that also includes hard science in the definition: (from Ithaca College library):
 * "A primary source is an original object or document -- the raw material or first-hand information.… In the natural and social sciences, the results of an experiment or study are typically found in scholarly articles or papers delivered at conferences, so those articles and papers that present the original results are considered primary sources.


 * "A secondary source is something written about a primary source.…Secondary source materials can be… articles found in scholarly journals that evaluate or criticize someone else's original research.


 * Here's another source from U.C. Berkeley library, which classifies "Research reports in the sciences and social sciences" as primary sources, but it doesn't make any fine distinctions about journal articles that might be secondary sources. CO GDEN  22:51, 17 August 2007 (UTC)

(outdent) It does make distinctions:


 * Emphasis added. To me that is clear. Regarding journal articles, if the author of the research is the only person proposing a theory supporting his research findings (it's his research so he was there), for example, and there is no other source for the theory such as a discussion in a textbook or discussion in a review article that analyzes competing perspectives, or other supporting evidence independent of the original researcher, then it is a primary source.   -- Mattisse  02:03, 21 September 2007 (UTC)

COGDEN's major change
There's one or two things I'm unsure of in this new version, but the most salient is that the general use of the terms does not require a secondary source to only collate information. Analysing the information and drawing conclusions, making comparisons, etc, doesn't create a primary source. That's why NOR ended up with the distinction, because we have to cite secondary sources in order to have analysis that isn't OR. SamBC(talk) 08:53, 11 September 2007 (UTC)


 * I agree. -- Mattisse  02:04, 21 September 2007 (UTC)