I have received an old genealogy report from a relative and it reads thusly (names changed for brevity):

1. Parent b. Ireland, d. Ireland. Notes about parent.
    2.   i Child.
    3.  ii Child.
    4. iii Child b. abt xxxx.
        iv Child b. Ireland, d. Ireland, single.
    5.   v Child b. xxxx.
        vi Child ? some accounts.

For comments like "some accounts" for the last child, how do I represent that on Ancestry.com? (Presumably as if to say, "this person's existence is debated"?)

  • In the report I received the arabic numbers are references to person descriptions later in the report. For those individuals with no additional information they are listed as a child (via the roman numerals) but not an arabic one.
    – fbrereto
    Dec 14 '12 at 1:05
  • 1
    I hope this becomes one of our most upvoted questions. Thanks you for sharing.
    – GeneJ
    Dec 14 '12 at 2:41
  • It would, for me, depend entirely on what those 'some accounts' were. If just other family trees I'd ignore both them and and the alleged child. If, on the other hand, those accounts were were publications of actual records- I'd do a bit more research but I'd still hesitate to add that child to anything online at this point. Dec 14 '12 at 2:49
  • @AndyHatchett: The above record is all the evidence I have to date.
    – fbrereto
    Dec 14 '12 at 3:44

User fbrereto asks about how to use Ancestry.com to share a conclusion that is based not on proof, but on "some accounts." I interpret "some accounts" to mean there exist some sources (including some family tradition) that reports about this sixth child, but that proof is lacking.


See Elizabeth Shown Mills, Evidence Explained ..., 2007 (p. 17) for "Conclusion: Hypothesis, Theory & Proof," in which she explains how conclusions can be categorized as one of three types, "each of which carries a different weight." Her categorization types follow; I have added my own summary as to the meaning of each.

  • Hypothesis: Something is "possible" (might be true) but has yet to be tested.
  • Theory: Something is "tentative." The hypothesis has advance, looks promising, but may remain to be tested against other theories in the real world of a "reasonably exhaustive search" and accompanying written proof.
  • Proof: The home run. Genealogically/historically speaking, you are able to convey and document your notions together with all of your logic and reasoning in a written genealogical proof.

During the research process, you may frequently ask yourself, "is this him/her?" You are probably forming a hypothesis. After extensive research, you may still be uncertain. You have a theory that is as yet unproven ...

If you use a genealogical database or online web service, you have the option of sharing your "research-in-process," or only sharing that which you have proven.

If you share your work-in-process, then you probably face the question fbrereto is asking. In his case, he has an old genealogy into which the author narrated about notions about a relation that fell short of "proof."

Ancestry.com trees

To my knowledge, Ancestry.com is like many database systems, it provides the ability to add a related individual (spouse, parent, child), but there are not obvious mechanisms to communicate whether that conclusion is based on a hypothesis, theory or proof.

(a) While many of us would like more overt solutions, a proof (and/or proof-in-process) can be communicated by a comment to the individual's Ancestry.com profile.

(b) Notions that fall short of proof ("hypothesis" or "theory") can be communicated in the related event "description" field (for parent/child, this would be a birth event; for spouse, this would be the marriage event).

These are only suggestions; I doubt either of these work arounds are widely used or recognized.


I would not want my words to undermine the importance of the distinctions between genealogical hypothesis, theory and proof.

From About.com, Kendra Cherry, "Introduction to Research Materials: [part 2] Theory and Hypothesis"

[T]he difference between a theory and a hypothesis is important when studying experimental design ...

  • A theory predicts events in general terms, while a hypothesis makes a specific prediction about a specified set of circumstances.
  • A theory has been extensively tested and is generally accepted, while a hypothesis is a speculative guess that has yet to be tested.

Lacking direct evidence, a genealogical proof is something that is written. Historically speaking, the work considers the full range of evidence relevant to the problem.

My own family history follows a mostly narrated style. As with the author of fbrereto's material, there are relationships in my tree that can be documented (with a source about something), but those relationships are yet to be proven (genealogically speaking).

At least I find these pose the greatest challenge in database presentation. As part of current work on "one-tree" systems, I've been trying to advance my own approaches to database presentations about notions of hypothesis, theory and proof. An example follows about the comment I add to a conclusion that I deem a hypothesis:

Dr. Charles Banks (History of Martha's Vineyard) concluded that Dorcas Bessey, daughter of Anthony, was the wife of John Presbury (d. 1679 of Saco). Others either dispute the notion or find material unresolved conflict. This includes that Robert Charles Anderson, The Great Migration Project (for Anthony Bessey), did not report any spouse for Dorcas. Lacking a genealogical proof, I conclude that the wife of John Presbury (d. 1679) is only possibly, not probably, Dorcas Bessey.

I consider this a hypothesis because even if I have the materials they worked with, I do not have the experience of Banks, Anderson, etc. On the other hand, I probably do have more information about the "whole" persons/families generally associated with John Presbury (d. 1679). In other words, I may have information or context that Banks, Anderson, etc. did not have; there is probably more that could/hopefully will be discovered.

It's possible that I could further develop those various materials in such a way that it would constitute a theory. Until then, I consider conclusions about Dorcas Bessey and John Presbury (d. 1679) to be just a hypothesis.

  • Sometimes one field of study (such as genealogy) "borrows" terms from a more prestigious one (such as science). When that happens, genealogists are obliged to use the terms in a scientific manner. The idea that a scientific theory is a promoted hypothesis but falls short of proof is UTTER NONSENSE. If we wish to discuss claims about which we have different levels of confidence, then by all means do so. Call them "horses", "cows" and "sheep" but do not grossly misuse the language of science to shore up your argument.
    – Fortiter
    Dec 14 '12 at 5:22
  • 1
    @Fortiter, it's my understanding that the terms hypothesis, theory and proof are also common currency in historical studies, so the 'borrowing' (which is Shown Mills terminology, not Genej's) is likely from there and quite justified.
    – user104
    Dec 14 '12 at 8:40
  • 1
    @Fortiter, I don't want to push this thread into a war of words but I don't understand the strength of your comment here. Terms like proof also have definitions in other fields, such as the legal one, and so I wouldn't claim they were borrowed from science. As it happens, a "scientific proof" is virtually a contradiction in terms if you consider proof to be absolute. Mathematics is the only field that affords such a stringent definition of proof.
    – ACProctor
    Dec 14 '12 at 14:10
  • @Fortiter Awkward. "Claims about which we have different levels of confidence" doesn't quite get at the point about which I was trying to communicate.
    – GeneJ
    Dec 14 '12 at 14:25
  • 1
    Speaking as one with a degree in maths, I'm pulling rank, for once. The terminology is all a bit of a muddle. Tony took the words out of my mouth - I could, tongue in cheek, say "Mathematicians produce proofs, the rest just wave statistics around". More to the point - maths distinguishes "theory" and "theorem" and only theorems are mathematically proven. Unfortunately, scientists and the outside world differ in their use of the word "theory". Fact.
    – AdrianB38
    Dec 14 '12 at 21:58

While this is not a direct answer to your question (@GeneJ has done that), I thought it was worth mentioning that some desktop software (such as Legacy) allows you to formally include a certainty level in the details of source citations. It's a little obscure to find it, but as you can see from this image, you can pick from five levels:

  • Have not decided yet
  • Marginal evidence
  • Probably conclusion
  • Almost certain conclusion
  • Convincing evidence

Snapshot of the Add Source Detail dialog from Legacy 7.5

Ancestry's Family Tree Maker 2012 software implements a similar capability with a star rating and comment:

FTM 2012 Rating dialog

Another version of the same dialog permits more structured annotation of source quality:

FTM 2012 Standardized source detail dialog

So it is possible to represent this information explicitly; just not on the web site.

  • Nice Gene, thanks - I'll have to find the star equivalent on the Mac version. Do you know if FTM transmits the stars to the web site with a synced tree? I would imagine they do, and if so then where they are up there (and how to edit them)...
    – fbrereto
    Dec 14 '12 at 5:48
  • I just checked, and it does not seem to show up at all on the online version, even though I just synched the trees Dec 14 '12 at 5:59

Not an answer per se, but I'd like to point out that uncertainty of this sort generally covers 4 cases:

  • child A might not even exist;
  • child A exists and definitely has a parent named B according to the records. B may or may not be this person of that name in this document, who, e.g., married C;
  • parent B definitely exists and definitely married C. They definitely had a child named A according to the records. A may or may not be this person of that name in this other document;
  • B and C definitely exist. A definitely exists. A may or may not be the child of B and C but we don't have any direct evidence either way.

Summing them all up as "A may be the child of B and C" doesn't do the evidence justice.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.