Talk:Source Value

Definitions [15 May 2009]

I have seen some of the discussions on definitions and had a few comments for what they are worth. If this is addressed for the general user, I think getting too academic in your terms and definitions will hurt the accessibility of the final document.

Possibly so. The problem is that it needs to be "right". If we get it to the stage where we have it "right", then we can create a more accessible version for the general user. Q 13:26, 14 May 2009 (EDT)
"Right" is relative to your goal. Is your goal to educate the user that doesn't know the importance of sources, or to impress the experienced genealogist? If the latter, the approach would be simple: don't let them edit WeRelate unless they are a certified genealogist. So I assume the former. Therefore, you need to make them sensitive to the basic issues, and they will learn finer points later on. --Jrich 23:26, 14 May 2009 (EDT) [more comments below with similar indentation]
I think there are ultimate two or more objectives here. One is educational, but as you point out this purpose may not be well served by this article. The second purpose is not to impress, but to "get things right, clear, consistent". The devil is always in the details, so somewhere you have to lay out the detail and devil. Once that's done, a version of the article can be crafted that would be more user friendly. Those who read the user friendly version, but want the detail and the devil, could then be sent to the supporting detailed article. But if you don't have the details somewhere, you're just asking for trouble when folks come along and say "Yes, but you haven't considered...." and then mention whatever detail it is that they are concerned with. You need something somewhere that treats the problem comprehensively. Q 12:33, 15 May 2009 (EDT)

I think some of the wikipedia definitions are more concerned with crediting the person who did the work than the value of the source, and may not apply exactly to genealogy.

The wikipedia article seems to be primarily concerned with describing how these terms are used. What's said there seems well thought out, and does reflect how these terms are commonly used in the various fields. If you see evidence of a bias towards crediting the original authors, that's because that's the emphasis in the various fields considered. Considerable emphasis is placed in most of these fields on authorship, and the authority with which an author writes. That's the way these disciplines work. Its the way genealogy should work, but usually does not. I should also add this is the way the Wikipedia works, and in this they are emulating the various disciplines mentioned in the article. Why do they do that? Because its necessary for credibility. Q 13:26, 14 May 2009 (EDT)
As you said, "in various fields." In science one makes a discovery and honest people want to make sure a person is credited with their advancement. By that thinking, the person that first points out a genealogical fact would be the primary source. I would argue that primary source in genealogy has a different meaning: i.e. as close to authoritative as possible. That has nothing to do with who discovered it, much more to do with contemporaneousness and closeness to the person whose genealogy is being investigated.

The BCG definition of a primary or original source seems pretty good: A piece of information is primary when it is recorded by a knowledgeable eyewitness or participant in that event, or by an official whose duties require him or her to make an accurate record of the event when it occurs.

No, I don't believe you have that correct. Certainly for their concept of "original source" (not really theirs, others use it too, but its not emphasized), an "original source" is referring to a document, not a datum. For their treatment of "primary source" and "secondary source", there statement is a bit misleading. What they are referring to is in fact information, but information is not a "source". The distinction between a "source" and a "datum" is a point of common confusion in genealogy. What BCG is doing is referring to data as "primary" and "secondary" implying that it came from a "primary" or "secondary" source.
I think this is splitting hairs. There are no facts. A vital record is a datum. The birth date or other particular that one concludes from it is a hypothesis, as we can never know for sure, given than any datum can marred by human error. A compiled genealogy considering many data becomes information, meaning organized data, and can have more value than the raw data (especially, for example, dealing with marriage records and matching the married person to the correct parents). You point is semantic quibbling. Not worth fussing over in my opinion. We all mean essentially the same thing whether we use the word data, information and source. Exactly what I meant by becoming too academic. If the goal is to address the user that doesn't believe in providing sources, or thinks citing a GEDCOM is adequate, this is not useful.
Words mean what they mean. When people use the same word "source" to mean two entirely different albeit related things, it tends to create confusion. Better, I think, to leave sources as documents, and data as data. Q 12:28, 15 May 2009 (EDT)

There is a distinction between (original and mechanical copy) versus a transcription. In areas of controversy, this becomes important, but in the general case where there is no apparent controversy, one has to assume good-faith efforts to make accurate transcriptions. Wish there was a way in WeRelate for multiple people to mark that they concur with a transcription.

Yes, that would be a helpful thing. That may be one application of the Digital Library, where a document could be presented in transcribed form, but not readily modified later on. That way you could attach a statement to the effect "original transcription by", "confirmed by", or "discrepancy noted by" kind of thing. Q 12:28, 15 May 2009 (EDT)
Yes, that's an issue. Genealogists rely heavily on such sources (transcriptions, extracts, and abstracts) The extracts in Chalkley, 1912 (aka Chalkley's Chronicles) comes to mind as an example. We tend to treat them as primary sources, but as you correctly point out, only the "original source" in the BCG sense of the term, is the primary source. The others are derivative. As someone wrote on APG, "Most genealogists are never, never, never, going to see an original document themselves. Instead they are going to rely on transcriptions, extracts, and abstracts. (That by the way was the original knock on Chalkely, 1912; it wasn't that it was error filled, it was that it "left everything out". Yes, it certainly did. that's why its called an extract. For Chalkley to have done what he was criticized for not doing, would have resulted in exactly zero product being produced. Very hard to send around a complete transcription of the Augusta County records. Chalkely would still be working on it, if someone had been willing to support him doing it.) In anycase, this point will be discussed in subsequent developments, expansions and contractions.

In my mind, secondary sources are compilations of data based on primary sources, either through direct citation or through citation of sources that cite the primary sources. They do represent original research in that they provide integration of various sources. In some cases, a secondary source can be more accurate than a primary source because it draws on multiple sources and presumably follows GPS to resolve discrepancies.

Tertiary sources are presumably everything else. This would include sources that rely on only one or selected sources (no integration/research component, no attempt to identify the primary basis for the data), or no sources (in which case you have to assume only one source).

There are fine gradations of all types of course, for example, death record is often based on testimony of a relative and may be second-hand. Vital records communicated to town clerk after the fact. Probably if would be more useful to avoid too much discussion of such fine distinctions and keep the theme focused on encouraging basing data on primary sources.

Dawes-Gates is an excellent secondary source. The author did much original research some of which is quoted by Great Migration Study in some cases, which itself also seems to be a good secondary source.

GEDCOMS and emails are not inherently tertiary despite the example in the chart. It depends on whether they provide sources for their data and the quality and integration of those sources.

There's another nuance in play here. I think that to be considered a reasonable secondary source, a work has to be revisitable, and revisitable long term. Can't see it, can't evaluate it, its useless. Most GedComs and emails have no potential for being revisted. In some systems such as Ancestry's, and on message boards (email equivalents), there's some potential for someone else seeing them again, and independently evaluating their contents. There's also the problem that a GedCom in particular, is subject to change without notice. What was said there yesterday, may not be there tomorrow, as the person revises and re-evaluates their material. Its not a dependable, revistable resource, even if its uploaded to Ancestry. (Same could be said for WeRelate.) As for emails, you can't really quote them, as you'd be copyright violation. So they may sit on someone's computer, but no one else can see them. (can't forward them since that's also copyright violation.)
I think you can always abstract emails unless you agree to keep an exchange off the record. Also, I think I added my comment below after you wrote this, so I may be repeating myself, but it is not the email that is the source, but what the email references. I agree verifiability is important, but what you want to verify is not the email, but the primary or secondary source that is the basis for the information in the email. In a properly written citation, you can leapfrog the email and verify the primary source itself. The email is only mentioned to give credit to the discoverer.
All quite true. Unfortunately, from what I've seen, most folks just take the email as their source and that's that. Also, my guess is that most emails don't provide the sources anymore than most Gedcom's. So you have an email that says X, and nothing that allows you to pursue whether X is right or not. Something to tuck away for future reference. Here, "personal communication" is probably the better approach. "John Smith told me in a personal communication that ..... but gave no sources to document the statement. This needs to be verified and validated." Q 12:38, 15 May 2009 (EDT)

I didn't notice town histories in the examples. As with all types of sources, you cannot make a generalization about all instances of a type of source material, but town histories often seem to have intimate access to town records (i.e., primary sources), and also a certain distance from family issues, that may make them slightly higher value than family histories as a general rule. However, it all depends on execution of the individual work.

That's a very good point. I'll add that. Q 13:26, 14 May 2009 (EDT)

--Jrich 11:23, 14 May 2009 (EDT)

Let me modify the statement about GEDCOMs and emails a little. Citing such a source, even if that source is of what I call secondary quality, carries no authority since it is not publicly accessible, so people may not inspect the sources it cites. Thus, it is simply another tertiary source...

...unless, enough of an abstract is given to relay the original source's reference to a primary source.

So listing an email from John Doe as a source is tertiary, worthless, really. But listing an email from John Doe, and giving an abstract saying that the death certificate from Cook County, IL gives death as 23 Mar 1969 in Chicago, would be of secondary quality because it cites a primary source and gives enough information that the primary source may be verified. Listing the email would be done to give credit to the researcher, and the abstract could add whether it was verified if appropriate. --Jrich 13:33, 14 May 2009 (EDT)

That, I think, is accurate. In response to your comment, I'd made a few changes in the page, but I think this says it better. Please have at it if you will. Q 13:38, 14 May 2009 (EDT)