Standard(s) of proof for making/keeping a connection

Question

A critical issue that I don't think is currently addressed by WT guidelines is the standard of proof for making or keeping a parent-child or spousal connection. Both WT guidelines and project guidelines talk about using "reliable sources", but the reliability of sources is only a part of the issue. The overall criteria for whether a connection should be made or maintained is the standard of proof required for a connection.

Note that the "genealogical proof" is not a standard of proof, or at least is not a standard of proof that can be used to make the binary decision of whether or not two profiles should be connected. Genealogical proof is about the work you must do to make your work credible. After one had done work sufficient for genealogical proof, a separate standard is needed to determine whether to make/keep a connection. 

The standard of proof that I think most people assume should apply is the same one used in civil suits - preponderance of the evidence (as it is phrased in the US) aka balance of the probabilities (as it is phrased in the UK) - which means determining whether, after considering all the evidence presented (including its reliability and relevance), whether the claimed relationship is more likely that not correct (i.e., is there more than a 50% chance it is correct).

I believe that is the standard that people are commonly applying on WT, but I don't think it is stated anywhere. I think it should be.

The current guidance on Uncertain Parents, however, raises the question of whether a different standard of proof should apply to post-1700 profiles. The guidance gives its blessing to connections on post-1700 profiles that are supported only by a "highly uncertain" theory, unless the theory is "disproven" or there is a "more likely" theory. This seems to support having connections even if they are not more likely than not correct. No standard of proof for "disproven" is given, but it seems it must be higher than preponderance of the evidence. Is "beyond a reasonable doubt" the standard for "disproving" a relationship in post-1700 profiles?

Regardless of the standard(s) desired, I think it is important to make them explicit. Without an agreed upon standard(s), it is difficult to know whether a connection should exist or not.

Comments

Good point, Chase, but I think that, in the absence of DNA from both, unless you were in the bedroom as an observer 9 months before the child was born, you can't *REALLY* know who the father is no matter how many or what quality of written records say so.

---

@Gaile - Of course not. I am not saying you need 100% certainty. I am suggesting we make explicit that we use the customary civil case standard of proof - preponderance of the evidence. Currently, we just talk about using reliable sources, not how the evidence should be weighed and what the outcome of the weighing needs to be to justify making/maintaining a connection.

---

Just a comment about "preponderance of evidence": while technically it's 51-49, in practice it usually ends up being at least 60-40.  It's difficult for people to perceive small differences in what is very close to a coin flip...and if forced to make a call when it's that close, it probably means they simply choose the outcome they like the best.

---

I don’t think “preponderance of evidence” is the same as “balance of the probabilities.” There are lots of examples where one side of a conflict is supported by many sources, while the other side has just a few, quality sources and it is the latter side that is accepted. Having more evidence doesn’t necessarily make that theory more likely once the analysis, correlation, and resolution of conflict steps of the GPS are performed.

(unless you understand “preponderance” to entail a notion of quality, but my understanding of the word is that it is about quantity)

---

@Barry - I agree that the common meaning of "preponderance" suggests more evidence. However, "preponderance of the evidence" is a term of art and means "considering all of the evidence (including its reliability and relevance) is it more likely than not that the claim is true." Juries are expressly instructed that "preponderance of the evidence" does not mean numerically more evidence.

Answer 1

issues have been raised a number of times about unlikely relationships and conflicts between sources. I was pretty sure there was general guidance on WT and I have now found it. It was staring me in the face. It applies to all information relating to profiles, not just relationships, and to both pre- and post-1700 profiles. It gives explicit support to what probably most of us do automatically.

Briefly, unlikely information should not be given credence, and, where there are conflicts of evidence, we are asked to go with the better evidence if that can be determined (with explanations if appropriate on the profile). It may be helpful to quote the guidance at length.

From https://www.wikitree.com/wiki/Help:Uncertain: ”Uncertain is generally not for unlikely information. [The bold typeface is in the help page - the statement is emphasised.] If one set of facts is more likely than another it should be replaced in the data fields. If facts are unlikely but there is no more-likely information, it can be removed…."

From https://www.wikitree.com/wiki/Help:Disagreements_about_Certainty, to which there is a link in the preamble to the help page https://www.wikitree.com/wiki/Help:Uncertain_Parents:

Should you replace uncertain information added by others?

If there is more evidence for one set of facts over another, the information should be replaced in the data fields.

Ideally, we want the best available information in database fields, and we want all the evidence for and against all facts that have been associated with the person fully explained in the text. Explanations that debunk common myths can be especially valuable…

This may well cause controversy if you weren't the one who added the earlier information. See Communication Before Editing for tips.

Even though it can lead to conflicts, please improve upon information when you can. Resolving conflicting information is the height of genealogy collaboration on WikiTree. It's what our community is all about. Remember that the other member has also signed the Honor Code that says, among other relevant points: II. We care about accuracy. We're always aiming to improve upon our worldwide family tree and fix mistakes.

Should you remove uncertain information added by others without replacing it? 

Removing information that someone else added when you don't have demonstrably better information is especially delicate.

As explained on Uncertain, you can remove uncertain information added by others if it is unlikely to be true. However, if the other member is responsive and is trying to follow WikiTree rules and policies, just like you, do your best to reach an agreement with them. Don't just remove the information. Try to convince them that the information is highly unlikely. Show them your sources and discuss them, in a patient, friendly way. You both want the same thing: a profile that's the best it can be.

See Communication Before Editing and Resolving a Disagreement for tips. See Problems with Members if necessary.

Comments

Thanks, Michael. I was looking at the same Uncertain guideline today. I agree that it sets a general standard that you should not enter information in a field at all (even if you mark it Uncertain) if the information is disproven and and that you should "generally" not enter it if it is unlikely. I read "unlikely" as binary with "likely" and thus equivalent to "more likely than not incorrect" or "less than 50% chance of being correct." That standard works for me.

However, since the specific overrides the general, I do worry about the following language in the Uncertain guideline regarding names and parents:

An Uncertain name or parent may be speculative but it should not be a guess. If you're only guessing at a name or parent you should not use the data fields. Instead, explain it in the narrative. You can link to highly speculative parents in the text. See Help:Uncertain_Parents.

Do you read the endorsement of speculative and highly speculative parents as (1) consistent with the general guideline of not entering "unlikely" information (i.e., parents that are not unlikely can still be considered "highly speculative", but you shouldn't enter parents that are "highly speculative" if they are unlikely), (2) overriding the general guidance so that, in the case of parents, "highly speculative" parents are OK even if they are unlikely, or (3) overriding the general guidance only to the extent provided in the referenced Uncertain Parents guideline, which I would read as only possibly overriding the general "unlikely" standard (and allowing highly speculative parents so long as they are not "disproven") for profiles that are not pre-1700 and not project protected? I think either (1) or (3) would be the right reading.

Re "disproven" - To my knowledge, this isn't defined anywhere in WT. While I find most of Helen Leary's 1998 Article on how the Genealogical Proof Standard relates to the "preponderance of the evidence" standard muddled (since GPS , at least as currently interpreted, is all about process, not making a binary yes/no decision), I am inclined to read "proven" for genealogical purposes to require, as she suggests, something more than "preponderance of the evidence" (i.e., likely vs unlikely) and less than "beyond a reasonable doubt", which I would say gives "proven" a meaning of "after considering all the evidence found after reasonably exhaustive research, the conclusion is that the assertion is highly likely/highly probable" and, more importantly for our purposes, "disproven" a meaning of "after considering all the evidence found after reasonably exhaustive research, the conclusion is that the assertion is highly unlikely/highly improbable/not reasonably plausible." This standard would only need to be used, of course, if "unlikely" was not deemed to apply because the situation was arguably an exception to the guidance that information should "generally" not be entered if it is unlikely or if the specific guidance permitting "highly speculative" parents was deemed to override the general "unlikely" guidance.

---

[Edited for greater precision and for typos]

On "speculative":

• Interpretations 2 and 3 are clearly wrong. What the guidance is saying, is don't connect up parents if they are just guesses. It supplements the general guidance about unlikely suggestions, and does not override it. And the general guidance is so obviously common-sense and rational that it would be very very silly to take any other view. To do so would be a contradiction of the Honor Code and the emphasis on aiming for accuracy that pervades lots of WikiTree guidance. 
• It follows that 1 is the right interpretation.

On ”disproven" I am not sure how much definition of words will satisfy you. One can go on for ever defining words which have a simple meaning. But in my mind "disproven" means what it says. Something is disproven if it has been proved wrong. That is the plain meaning of the word in the English language. For my part, I see no point whatever in attempting a more precise definition. "Disproven" means more than ”very unlikely".

But if you want marginally more precision, look at the wording on the Disproven Existence Project page, https://www.wikitree.com/wiki/Space:Disproven_Existence: for existence to be disproven, there should be "substantial evidence that the person did not exist", or, slightly later, "research has shown that the person never existed". Substitute references to relationships not existing.

As always, each instance has to be looked at on its own facts.

---

Chase, are you in a legal profession perchance?  I agree with Michael, one can go around forever trying to define words more clearly.

---

@Danielle - Gave myself away, did I?

Answer 2

There is explicit guidance for pre-1700. Reliable sourcing is required, as it is generally for pre-1700 profiles. For post-1700 profiles, the issues were discussed to some extent back in December - see https://www.wikitree.com/g2g/1503552/should-we-create-a-research-note-box-for-disputed-parents?show=1503552#q150355 - and people may want to glance over that discussion.

Comments

Reliability of sources is one factor taken into consideration in a standard of proof, but it is only part of the equation. It doesn't address, e.g., relevance of evidence or what to do with conflicting evidence. At some point, you need to have the standard of proof concepts of weighing the evidence and where the balance has to be.

---

I take it as axiomatic that evidence has to be relevant. I doubt that needs spelling out.

Conflicting evidence can be discussed on profiles, often in research notes. For pre-1700 profiles, where reliable evidence is required, a judgement needs to be reached case by case on how to handle things if there is a conflict of evidence, possibly with discussion in G2G or in a Project Google Group or Discord channel, and in consultation with profile managers in the usual way. Each instance of a conflict between reliable sources will be specific to the particular relationship, and I can see little possibility of setting out any general rules. In some instances, more recent research will have disproved what has been said in reliable sources; in others there will be genuinely a degree of uncertainty and it will be a matter of case-by-case judgement how to handle things.

---

Yes, every case is fact-specific, but the standard of proof should be the same. Should it be (1) the totality of the evidence shows it is possibly correct, even though it may be unlikely, (2) the totality of the evidence shows it is more likely than not correct, (3) the totality of the evidence shows that it is correct beyond a reasonable doubt, or (4) some other standard?

---

Chase, I started an idea of an app to assist in subjective evaluation for each source as it relates to the facts at hand, because you could have 10 sources that say one thing, but are written 200 years after the fact, and 1 contemporary primary source that refutes them all.

It fell to the wayside in my zeal to over burden myself, as I am wont to do.

Dont suppse you want to drive that train?

---

@Jonathan - Re overburdening - I hear you. But I'm curious, how you were thinking of going about it?

---

Was not all my idea, Ronnie Grindle had a framework and I was going to build on that. Basically coming up with a way to "standardize" through notation what is essentially a subjective analysis of the strength of each source, and put a personal maintenance category on the profile that indicates how many sources of each tier exist, which could be interpreted to show the difference between "this profile is sketchy" and "this one is rock solid". We couldnt find a project who was willing to use this as project categories and so were stuck with personal ones. (Which would still work fine, just not be universally standardized)

And before anyone asks, had no intention of applying this site-wide, just a tool for members to add to their own work if they wanted.

I'll even got so far as to talk Greg Clarke into maybe being willing to use the format he built for the dna citation app and replace the decision tree for this thing, but then I dropped the ball.

I will send you what I talked through with Ronnie so far, see what you think

Answer 3

The Genealogical Proof Standard (GPS) was produced as a replacement for the “preponderance of the evidence” standard. (See for example, https://bcgcertification.org/ethics-standards/ and https://bcgcertification.org/skillbuilding-evidence-revisited-dna-poe-and-gps/ : both accessed 1 June 2023.) The GPS is supposed to be a stronger requirement than ‘preponderance of the evidence’, but weaker than ‘beyond a reasonable doubt’.

The only reference to ‘preponderance of the evidence’ in Elizabeth Shown Mills’ Evidence Explained (3rd edition) is in the glossary: ‘a legal standard acceptable in civil cases, whereby evidence on one side of an argument outweighs, at least slightly, evidence on the the side of an argument; insufficient when used for historical research’ (pp.827f.).

So, shouldn’t the standard we’re aspiring to meet be the GPS, not ‘preponderance of the evidence’?

Comments

We probably shouldn't call both POE and GPS standards of proof. They are apples and oranges. Here's what BCS says what GPS is:

Genealogical Proof Standard (GPS)

To reach a sound conclusion, we need to meet all five components of the GPS.

1. Reasonably exhaustive research.

2. Complete and accurate source citations.

3. Thorough analysis and correlation.

4. Resolution of conflicting evidence.

5. Soundly written conclusion based on the strongest evidence

POE has nothing to do with research, citations or writing. POE is a standard for making a binary decision for or against a claim. It's not clear to me that GPS even requires a yes/no decision. It seems to me you can satisfy all the components of GPS and conclude that X was probably the son of Y, which makes sense. If you are writing a bio or genealogical article, your conclusions should accurately reflect the evidence and you shouldn't force them into yes/no if your conclusion after following all the steps is "maybe". I think this is where the statement comes from that "proving" something under GPS is a higher standard than POE. If you are going to say yes or no rather than maybe in a bio or article, your level of certainty should be significantly higher than more likely than not.

I don't think we want to adopt a requirement for a connection that is higher than more likely than not, but I leave that up to discussion. I don't think a higher level is something to aspire to, since the evidence is what it is. You can't force it to be more definitive than it is.

I think we should aspire to meet GPS when we do profiles, but I don't think GPS is where we want to go for the criteria of whether or not to make/keep a connection.

---

Agreed. I find the GPS to be rather loosely written.

---

I think that in general standards of proof don’t need to be binary; evidence can prove claims to greater or lesser degrees. Here, of course, the connection must be made or not. But for parent-child connections it can be uncertain, confident, or confirmed with DNA, which gives us four degrees. Perhaps, for example, the appropriate standard for uncertain connections is preponderance of the evidence, but for confident connections is the GPS. (The “Uncertain Parents” page you linked to above points out even more degrees of proof, of course, )

With historical (including genealogical) questions, the right answer is sometimes that the evidence doesn’t let us answer the question. That won’t work, for example, in a dispute about who owns property. So, preponderance of the evidence works as a standard for (some) legal questions, but not for historical questions. Of course, if the evidence doesn’t let us decide whether the people are connected we shouldn’t connect them. But is a mere preponderance of the evidence enough to let us decide?

The other point of the change in the standard of proof from preponderance of the evidence to the GPS, as I understand it, was that for genealogical questions we need to consider issues that the ‘preponderance’ standard doesn’t require us to consider. For example, instead of simply asking whether the evidence for one side is greater, we need to explain the evidence against the conclusion we’re defending. And we need to evaluate the process of collecting evidence, to ensure that we haven’t missed relevant evidence. (I assume that in the US & UK judicial systems, the adversarial nature of trials is supposed to do that. But even if that works in the legal systems, the nature of genealogy isn’t inherently adversarial.) So I’m inclined to think that even for the issue of whether to connect people, preponderance of the evidence is too weak a standard.

The other issue is that the explicit standard for creating post-1700 profiles is very weak. (See https://www.wikitree.com/wiki/Help:Sources_FAQ ) Should the standard for connections be higher than the standard for creating profiles? That’s why I suggested we should be aspiring to meet the GPS, not requiring it (though I didn’t say that).

Answer 4

WikiTree's policies aren't based on legal concepts, such as "burden of proof" and "standard of proof". As a collaborative community, we work together to improve profiles and resolve any conflicting data -- rather than placing responsibility on individual members to "prove their case".

As you have noted, there is guidance about how to handle uncertain parents in Help:Uncertain Parents. There is also guidance for dealing with disagreements about the content of profiles, including parent/spouse/child relationships, in Help:Resolving a Disagreement. This help page says:

Disagreements about Content

Citing sources

Carefully citing appropriate sources can resolve most content disagreements.

If the profile is Open simply edit the narrative and add the source. If the profile is not open for editing, add the source information as a comment on the profile page. ...

Insufficient sources and certainty

Sometimes there are insufficient sources to directly prove a fact. Vital records may have been destroyed, or perhaps never existed, requiring analysis of a variety of pieces of information.

Resolving conflicting information in the face of insufficient sources is the height of genealogy collaboration on WikiTree. It is our greatest challenge.

To facilitate collaboration, consider attaching a G2G discussion to a profile to discuss the evidence. Focus on the content, not on the personality or characteristics of the other member. ...

Whether or not agreement can be reached through such a discussion, add a Research Notes section to the profile and summarize the issue there.

Note that it is not WikiTree policy to disconnect uncertain or unproven parents. This is why we have the Uncertain indicator for mothers and fathers. See Uncertain Parents.

If a Profile Manager or Trusted List member makes no response to any of these types of queries see Unresponsive Profile Managers.

[Edited to fix typo.]

Comments

The Help: Uncertain Parents is insufficient guidance because it mostly only applies to post-1700 profiles. When the guideline was proposed, the consensus was that there should be tighter standards for pre-1700 profiles, but no guidance on what to do with them was provided. I note that the Uncertain Parents guideline itself contains two standards: A connection should be disconnected if it is "disproven" (unclear what degree of certainty is required for that. Maybe GSP?) and a connection should be disconnected if an alternative theory is "more likely" (although I think what is meant is that there is a more likely theory that the parent is a different identified person, than a more likely theory that the parent is someone else who has not yet been identified).

Disagreements are easier to resolve if both sides at least agree on the standard that must be met. If the person defending the connection thinks the appropriate standard is that the connection should remain unless it is 100% proven false, while the person looking to make a disconnection thinks the standard is that the connection should be removed if it is highly unlikely, even if still theoretically possible, there may be no way for them to agree. They can fully agree on the facts, but reasonably disagree on whether the connection should remain.

Here's a common example for a profile of a John Smith of Devon, yeoman, who was probably born in the early 1500s: The profile of John Smith of Devon, yeoman, has Sir John Smith of Yorkshire, knight, connected as his father. The "reliable source" provided is a pedigree in a Visitation that shows that Sir John Smith of Yorkshire had a son named John. No further information relevant to what happened to son John can be found. Should the connection be kept? It's theoretically possible that Sir John's son John was kicked out of the house, moved hundreds of miles away to Devon, and became a humble farmer. However, the parentage is much less likely than not because (1) it is highly unlikely that the son of a noble would end up as a humble farmer, (2) Sir John lived hundreds of miles away from where John the yeoman lived, and (3) there were no doubt lots of commoner John Smiths born much closer to where John the yeoman lived who were much more likely to be him. Should the connection remain or be removed? There is currently zero guidance on this issue.

---

In your example, I would link to the theory in the narrative, but disconnect the parents for the numbered reasons you list.

---

@Jillaine - I totally agree that should be the result. But what WT guideline can you point to that makes it clear that is the right outcome as opposed to leaving it attached because the connection is theoretically possible (although highly unlikely) and there is no better candidate that has been identified to attach? If it were a post-1700 profile, I think that is what the Uncertain Parent guideline says the outcome should be. I think we should have a pre-1700 guideline that says detachment is the right outcome if the parentage is "more unlikely than not" or maybe "highly unlikely", even if we don't have a more likely candidate we think should be attached.

---

The answer is simple, Chase. The requirement for reliable sourcing. There is absolutely no reliable sourcing in the example you give. End of matter as far as I am concerned.

By definition and all normal use of the English language, an unlikely relationship with no good clear sources to demonstrate it cannot be reliably sourced.

---

@Michael - In the example I gave, the profile provided a citation to a Visitation, which is on the list of "reliable secondary sources" for the England project. But we could substitute a Yorkshire parish baptism record (a reliable primary source) and have the same situation.

I think you are using "reliable sourcing" to mean the same thing I am by "preponderance of the evidence". I.e. you are saying a connection is supported by "reliable sourcing" if, considering the totality of the evidence presented (including its reliability and relevance), the evidence indicates that the proposed parentage is more likely than not correct. If that's what we mean, I think we should say it. Both "reliable sourcing" (as you use it) and "preponderance of the evidence" (as used in civil cases) are both being used as terms of art, which many people will not understand from the phrase itself.

---

In the example you gave there is absolutely no reliable source for identifying John the yeoman of Devon as the John named in the Visitation as son of John the Yorkshire knight. Therefore the relationship is not reliably sourced, even if the Visitation is accurate. Of course if someone comes up with a reliable source to demonstrate the identification, that can be considered. But subject to that, the answer is clear under WikiTree guidance. The relationship should be severed. I would argue for adding a research note to explain briefly why this has been done.

---

What if the Visitation refers to son John Smith of Devon? Or son John Smith of Devon, farmer? Or we find that John Smith of Devon, farmer, has land holdings in Yorkshire next to Sir John's? There is no way to get around the fact that we consider all the evidence (including its relevance and reliability) and determine how likely it is that the claimed relationship is correct and, depending on how likely it is that it is correct, either connect the profiles or not. I'm just suggesting that we make the balancing of the evidence and the threshold of probability for making the connection explicit. If you say a relationship is "reliably sourced", that's just short-hand for saying that there is reasonably reliable evidence that, collectively, supports the connection and that, notwithstanding whatever countervailing evidence there may be, the threshold of probability has been passed.

---

I have said what I have to say elsewhere in this series of conversations so I will not repeat myself.

Answer 5

Chase, I respectfully disagree. I hear people use the terms "reliable", "proof", "verifiable", etc., and these are all subjective terms. They don't mean the same thing to every person. The same would go for "preponderance of the evidence" - it is what you believe.

The purpose of a source is simply to tell the user where you got the information. That is why "my mamma told me" (not in citation form, just for illustration purposes) is a source.  Now, mamma may be wrong, and I might find another record later that calls into question what mamma told me. I then may choose to cite that record and explain why I'm choosing it over mamma. Other researchers may stick with mamma over my new source.  That's okay. The question isn't about who is right, because that can't be proven (at least in many cases), but it lets the viewer make their own decision.

Now, we come to the question of other genealogical records. If that is where you got the information, then I have no problem saying it.  In the absence of another source contradicting it, there is no reason to believe it is wrong.  Now, if you don't like a source or believe it is not up to the standards you want to apply, you can find alternative source(s), but until then the genealogical record has value. In some cases, you may have not known the names of the parents or details about when or where they were born. When I see "genealogical" as a source, I may think I don't know how they arrived at this decision. So I'm going to research it further, and when I find a record, add to the profile and explain.

I know we all don't agree on this subject, but I want to explain for some of us why we don't have a problem. As long as the source is documented in a manner consistent with policy, I'm fine with it. It is not to say you can't use those terms mentioned above in your thought process, and this is where the collaborative process comes into play. When discussing it with others, we might use those terms to make decisions. I believe this record is more reliable because..., or I have found multiple sources that I believe prove..., or we all agree Mamma was there and first-hand knowledge is better than information gathered second-hand later.

Comments

Right. People can disagree whether the evidence meets a certain criteria for connecting/disconnecting, but at least there should be an official criteria. There is for post-1700 profiles, although not entirely a clear one.

Here's an example of why we need a guideline for pre-1700 profiles (same example as I gave in a comment above): There is a profile of a John Smith of Devon, yeoman, who was probably born in the early 1500s, who has Sir John Smith of Yorkshire, knight, connected as his father. The "reliable source" provided is a pedigree in a Visitation that shows that Sir John Smith of Yorkshire had a son named John. No further information relevant to what happened to son John can be found. Should the connection be kept? It's theoretically possible that Sir John's son John was kicked out of the house, moved hundreds of miles away to Devon, and became a humble farmer. However, the parentage is much less likely than not because (1) it is highly unlikely that the son of a noble would end up as a humble farmer, (2) Sir John lived hundreds of miles away from where John the yeoman lived, and (3) there were no doubt lots of commoner John Smiths born much closer to where John the yeoman lived who were much more likely to be him. Should the connection remain or be removed? And what guideline can you point to that supports your conclusion?

---

I think we have the same goal, but different paths to get there. The questions you ask about the pre-1700 profile will never be known for certain. Any record we use could be wrong. All the questions you asked are logical and appropriate. There is still room for different opinions or conclusions. Such as the example "farmer". The record saying he was a farmer doesn't necessarily mean he tilled the land himself. He may have owned a farm and had serfs working the land, but at that time he was considered a farmer.

The meaning of words, and the time lapse, make it difficult to prove one way or the other, but to solve the issues you raise. That is why the "[Pre-1700 help profiles (https://www.wikitree.com/wiki/Help:Pre-1700_Projects) states you should seek help with a project, and these issues would be worked out.

I think some people get a little overzealous in the terms we have been discussing, and it discourages others from reaching out or collaborating. This creates more situations you describe. The policy is certainly intimidating to new users, because it sounds so authoritarian. I say this because collaboration should be a happy process where everyone walks away, feeling better. Its purpose is to empower (teach, learn, grow). Throughout Wikitree, we use terms such as "we rarely forbid" or "never" because they discourage collaboration.

In the end, although we may be on different paths, I believe our goals are the same, and the questions you raise will be worked out in the process.

Answer 6

Uncertain Parents page has fairly large options, and although I can understand your point, I don't think revising the page will assist us in getting a better tree.  Rules of evidence in law and in genealogy are indeed not identical, but the ''GPS'' as presented doesn't appear to me to resolve the issue of defining them well.  (Wish they had picked another acronym  Global Positioning System anybody?)

It's always a case by case scenario when going up or down a tree, so trying to make a ''one size fits all'' rule is difficult at best.

Comments

I think the process detailed on that page can be summed up in a fairly simple flow chart. For the record, I side with those who support the policy as it is.

---

I could have added the parts about documenting competing theories, but that chart is just for determining whether to connect and whether to mark uncertain.

Answer 7

Reviewing answers and comments in this rather lengthy and in places repetitive set of exchanges, I believe it is possible there may be some basic misunderstanding among some members about the help page https://www.wikitree.com/wiki/Help:Uncertain_Parents.

It is very important to read the preamble to it. This says:

• Connecting uncertain parents is never required. It's a decision made by the editors of the profile. For pre-1700 and project-managed profiles, parents are generally only connected if they appear in reliable sources, but are sometimes connected temporarily. For modern profiles, if there is a disagreement about whether to connect them, the following guidelines should be used.

There are three things to emphasise here. 

• First, for pre-1700 profiles, the general rule is that parents should be connected only if they appear in reliable sources. That means that the reliable sources should indicate the relationship between the two specific individuals. 
• Second, the guidance says that the decision on whether to have parents connected is for those who edit a profile. The Honor Code, with its commitment to accuracy and fixing mistakes, and quite a bit of other guidance, is relevant to decisions on whether relationships should be severed.
• Third, the detailed guidance given on this help page for post-1700 profiles is meant to come into play only when there is a disagreement between WikiTree members over whether two profiles should be connected. In other words it is not guidance which applies for all post-1700 relationships. In most cases, it is possible to reach agreement on whether it is sensible to attach a parent or spouse, or whether an existing relationship should be severed, without having to think about this detailed guidance.

I am taking it as read that the normal WikiTree guidance on  communication before editing will be followed.

Comments

"for pre-1700 profiles, the general rule is that parents should be connected only if they appear in reliable sources." Yes, that's what it says, but it is extremely unclear what that was supposed to mean.

"That means that the reliable sources should indicate the relationship between the two specific individuals." A reasonable interpretation. But do the sources need to just indicate a possible relationship? Indicate a probable (ie more likely than not) relationship? What if there is some evidence from reliable sources suggesting that the relationship may be incorrect? Shouldn't we be weighing all the evidence and determining whether the likelihood of the relationship passes a certain threshold of probability?

---

Chase, that's where more research is needed.  I deal a lot with pre-1700 profiles, and have come across multiple instances where attached parents only came from a tree, with no supporting documentation, like neither marriage contract nor marriage act naming them.  And unlike Anglophone traditions, up here the majority of marriage acts name parents.  But when we don't have such a record, then it's a compilation of people present at baptisms, relationships named in subsequent acts, etc etc etc.

---

[Turned into an Answer]