Crafting Proof Arguments & Citing Ourselves in the Process


In another forum, a thoughtful researcher asked for advice on writing proof arguments. Questions and answers revolved around two issues:

  1. Should a proof argument be part of a research report or a stand-alone piece of writing?
  2. Is it appropriate to cite ourselves when writing a proof argument, considering that we will have done most or all of the underlying research on which the argument is based?

The answers to both questions require us to understand what a proof argument represents and how it differs from others types of writing.


A proof argument is a documented essay—be it a few paragraphs or umpteen pages—in which we lay out the evidence for a certain conclusion, along with our reasoning. Typically a proof argument is a stand-alone piece of writing.

Proof discussions do frequently appear within research reports but, there, they typically represent a preliminary analysis of a record, a situation, an identity, or a kinship. That preliminary analysis will be considerably different from a final proof argument.

Proof arguments are needed when our research problems have complicated answers.1 They are valid only when the answer we propose has met the first four criteria of the Genealogical Proof Standard. In that case, the proof argument becomes the final criteria of the GPS. Those 5 criteria are these:

  1. Reasonably exhaustive research
  2. Use and identification of the best possible sources
  3. Correlation and analysis of all evidence
  4. Resolution of any conflicting evidence
  5. A soundly reasoned, written, assembly of the evidence2

A research report is a step-by-step, source-by-source account of one block of research. We've detailed its essentials in QuickLesson 20, "Research Reports for Research Success." Because time limitations always exist, whether we are doing personal or professional research, it's rarely possible for any research report to meet even the first criteria of the GPS: reasonably exhaustive research.  Almost always, meeting that criteria requires multiple blocks of research (and multiple research reports) before we get to the point that we have reasonably exhausted the sources that exist for the problem.

Criteria 3, correlation and analysis, are done amid each of those research efforts. When we use new sources and compare their information against what we had already gathered, we see issues that need to be discussed right there in that research report. We see flaws in the source or its assertions. We see clues to pursue. We develop hypotheses that need to be recorded, along with our reasoning. But these focused discussions are not proof arguments.

Amid the correlation and analysis that we discuss in a research report, we frequently see conflicting evidence that has to be resolved. Resolving that conflict usually requires additional research—which means one or more additional blocks of research and one or more research reports to detail the results.

We also play Devil’s Advocate with our evidence. We test other theories and analyze why they can’t be accurate. That, too, requires additional research; and that also is work we document within a research report. Or multiple reports.

Once our correlation and analysis convinces us that we have a sound hypothesis, once we’ve played Devil’s Advocate and resolved all conflicts, then we are ready to draft the proof argument. If we choose, we may do this as a standalone document, or we might create a research report that analyzes all past findings and lays out our hypothesis to be investigated.3

At that point we assemble all the evidence we've found to date, we cull it down to essentials, and we explain our reasoning on each point. This writing process usually spotlights holes in our research that we had not seen before, as well as problems with our logic. Filling those holes will also require additional research.

A sound proof argument is the end product of all of the above. On its face, we can see that all of this will rarely be done within the confines of one research report. Thus, the proof argument ends up being a standalone piece of writing.

Yes, proof arguments can be written at all stages of our research. However, these are preliminary arguments, rather than final products. We write them to organize our evidence and our thoughts—and to see how well our hypothesis of the moment will hold together and what additional glue it might  need.

As for the other major concern ...



Can we or should we cite ourselves? That depends upon three issues

  • Is it necessary to cite ourselves as part of our proof—if so, why?
  • What exactly will we be citing?
  • Is that prior work of ours publicly available so that readers of our proof argument can evaluate it?

Citing ourselves can be appropriate—or not—depending upon the circumstances. Citations to self are frowned upon in academic works, unless

  • that earlier work of ours has been published in a peer-reviewed forum;  and
  • there is no sound work by others that could  be cited instead;
  • (or for instructional manuals or articles, as with this one) we are demonstrating how to handle specific situations.

Citing our underlying research is acceptable within bounds. When works of scholarship are the products of complex projects, researchers frequently post online their unpublished underlying work, or their raw data, not only to support their final publication but also to invite scrutiny and feedback. That is the context in which we would cite ourselves in a proof argument.

When we do organized research, when we make our results available in thorough and well-documented reports where we analyze evidence as we gather it, these underlying reports can strengthen our proof argument. But it can do so only if these reports are accessible to others. Some researchers maintain websites where they post reports of past or ongoing research. Others post their work at Scribd or some similar public venue.

Citing ourselves within a proof argument also has another critical restriction: our proof argument must identify and analyze all the evidence needed to make our case. When we cite a piece of direct evidence in the proof argument, we don’t cite some prior work of ours in which a document is abstracted or discussed. We cite the actual document.

That said, if our underlying research meets that first criteria of the GPS—reasonably exhaustive research—our reports will contain valuable context and considerably more supporting evidence than we can fit into a cogent, well-honed proof argument. In that circumstance, citing our detailed and publicly available "raw" research can not only be appropriate but useful.

A convenient example of the essential use of self-citation—and its limits—would be my own article "Frontier Research Strategies—Weaving a Web to Snare a Birth Family: John Watts (ca. 1749–ca. 1822."4 As a foundation for the task covered by this paper, it was necessary to first separate this man from another same-named man with whom he had long been confused. Both research problems could not be handled in the confines of one article. Therefore, the published proof article confined itself to the issue of parentage. Then in footnote 7 (out of 159) my editors and I cited three underlying research reports that I posted at my website Historic Pathways.5


Whatever personal work we cite has to be a formal piece of writing, it should present well-done and well-documented research, and it needs to be publicly available.



1 When working with historical records, we usually document our assertions in one of three ways, depending upon how complicated matters are:

A. If we have one or two pieces of direct evidence provide identical information, we only need a reference note with a basic citation to each source.

B. When we have a goodly number of sources providing direct evidence—with each providing a partial answer and none of it conflicts—we may write a proof summary to identify all the sources and what each shows. Within genealogical work, these are most often used as cover sheets for the documents attached to lineage applications.

C. When we are working with conflicting evidence, indirect evidence, or negative evidence, identifying our "proof" requires a proof argument that resolves the conflict or builds the case. It's called an argument because we actually are arguing for a certain conclusion that is not obvious on its face.

For deeper dives into this subject, see "Proof Argument vs. Proof Summary" and QuickLesson 13, "Classes of Evidence: Direct, Indirect, and Negative."

2Board for Certification of Genealogists, Genealogy Standards, 2d ed. (Nashville, TN: Ancestry imprint, Turner Publishing Co., 2019), 2–3.

3 As an example of this process by which a proof argument is developed: my website Historic Pathways, under the "Research" tab and the subhead "Cooksey," presents seven, related, research reports. The purpose of the research was to identify an elusive woman and her last husband whose identity has been widely mangled. The first and second reports lay out findings from two blocks of research in 2008 and 2009. These findings suggested an hypothesis for the birth identity of that female. The third report in 2010 ("Possibility to Pursue: Were John & Judith [—?—] Watts the Parents of Zilpha [—?—] Price Cooksey?") was a preliminary argument. It summarized, for Cooksey researchers, the key evidence found up to that point, demonstrated why counterclaims were not reliable, and presented an hypothesis for future research. Four additional reports followed, before evidence was sufficient to construct a credible proof argument.

The final argument was published in 2014 as “Testing the FAN Principle Against DNA: Zilphy (Watts) Price Cooksey Cooksey of Georgia and Mississippi,” National Genealogical Society Quarterly 102 (June 2014): 129–52; PDF download at Mills, Historic Pathways (

4 Elizabeth Shown Mills,  “Frontier Research Strategies—Weaving a Web to Snare a Birth Family: John Watts (ca. 1749–ca. 1822),”National Genealogical Society Quarterly 104 (September 2016): 165–90.

5 A similar approach is seen in the same journal's publication of Rachal Mills Lennon's "Southern Strategies: Merging Identities by Mapping Activities and Linking Participants—Solomon Harper of South Carolina's Lowcountry," National Genealogical Society Quarterly 107 (September 2019): 165–84.  That article's note 8 (of 67) refers the reader to the underlying research in Lennon's Harper Reports 1, 2, 4, 5, and 19, which she (with her client's permission) posted at her website Finding Southern Ancestors (, to provide much supporting evidence and context that could not be included in the article.



CanStockPhoto ( : downloaded 13 February 2020), "Public Defender," uploaded 20 February 2015 by innovatedcaptures; used under license.


Elizabeth Shown Mills, "Proof Arguments & Citing Ourselves in the Process," QuickTips: The Blog @ ( : posted 13 February 2020).