QuickLesson 24: Evaluating DNA as Evidence

Of course, DNA is evidence. Prosecutors and defense attorneys use it daily to build cases for guilt or innocence. Forensic genealogists and police use it to build cases for the identity of human remains. Some historians and millions of genealogists use it to build cases for historical identity and kinship.

However …

Evidence is not just a fancy word for “something that supports what I want to believe.”  Yes, evidence is a mental concept. It’s our decision that a piece of information is relevant and our interpretation of what that specific information actually means. Yet, even though it is mental rather than physical, evidence still has definable forms and boundaries we have to understand to build solid cases.

This QuickLesson explores DNA as a class of evidence and how to use it for case building. To do that, we first need to lay a foundation by defining evidence in terms of the traditional sources and situations more familiar to most readers of this lesson.

The Foundation on Which We Build

Before we can turn any piece of information into evidence, we must decide one thing: What is the purpose for which we want to use this information?  We can express that purpose in one of two ways:

As a research question. For example:

Was John Watts of Washington County, Georgia, who (a) was appointed in 1793 to select the site of the new Washington County courthouse, the same John Watts of Washington County who was (b) simultaneously appointed to select the courthouse site for the new offshoot county of Montgomery?¹

As an hypothesis—i.e., a declaratory statement of possibility to be explored. For example:

John Watts of Washington County, Georgia, who was appointed in 1793 to select the site of the new Washington County courthouse, was not the same John Watts of Washington County who was simultaneously appointed to select the courthouse site for the new offshoot county of Montgomery.²

We have sound sources for the two appointments. They explicitly name “John Watts.” But, without defining the question we want to answer or the hypothesis we are trying to prove, we cannot begin to say whether a piece of information is relevant to the case we are trying to build. A piece of information could answer one question, but not another. It could strongly support one hypothesis and demolish another. It could also be right or wrong in either case. That last point introduces a value judgment that requires us to understand the three different types of evidence we work with.

The Materials with Which We Build

Evidence offers three basic building blocks that we call direct, indirect, and negative.³  As with any building materials, different types of materials are used for different purposes. If we build with this-or-that type used in the wrong way, our case will collapse.

Direct evidence is obvious. We have an information statement that explicitly addresses the issue we are wrestling with. For example:

Research Question: Who chose the site of the Washington County courthouse?

Information:  “John Watts, John Stokes, Owen Fort, Solomon Bechum and John Marcus, are hereby appointed commissioners for building and fixing on a proper place, as nearly central as may be convenient, for the court house and gaol in the county of Washington.”⁴

Evidence?  Direct.The record plainly states that John Watts and four other men were to pick the site. It doesn’t give us personal identifiers for this John Watts. It doesn’t tell us whether he was the same man appointed to select the Montgomery site. No one record ever tells us all we want to know. But this one’s information does explicitly answer the research question that we defined above: John Watts and four other men were appointed to choose the site.

Indirect evidence is not obvious. We discover it by following the old-timers’ advice to “read between the lines.” In the spaces between the words that actually appear, we see clues that are not actually stated and a way to use the information to help us build the case. Because indirect evidence doesn’t explicitly address our research question, we will have to combine it with other evidence to reach a conclusion.  For example:

Hypothesis: The John Watts who was appointed to pick the site of the Montgomery County Courthouse likely lived in the part of Washington from which Montgomery was cut.

Information: “And be it enacted by the authority aforesaid, That a new county shall be … taken from the county of Washington, in the following manner: First by a line beginning at Carr’s Bluff, on the Oconee river, and running along the Uchee path … to the Ogechee river; thence … to the Effingham line; thence … to the line of Liberty County; thence … to the Alatamaha river; then up the said river to the confluence of the Oconee and Oakmulge rivers; thence up the Oconee river to the beginning.”⁵

Evidence?  Indirect. John Watts is not mentioned at all in this legislative act. However, a mental bell still rings an alert: Ding-a-ling! Think about this! This information defines the limits of the new county from which John Watts would help choose the seat of government. This information is relative to him in a roundabout way. Logically, the defined region should be one with which the appointee was especially familiar, hence the selection of him to choose the best site. By extension, this region is likely to be the one in which he lived.  

However, because this is indirect evidence, we cannot just assert it as proof that our interpretation is valid. We need other evidence to support or disprove it. For example, we might  identify all Washington County land previously granted to someone named John Watts and see whether any of that land fell within the bounds of the new county—and whether other land granted in that name remained in the old county where it might be owned by a different man of that same name who stayed in that county after John of Montgomery was cut away. Then, by combining the information from all these and other sources, we might build a case on indirect evidence that there was—or was not—two men of that name in Washington County shortly after it was created.

Negative evidence is silent evidence in a situation in which there should be noise. QuickLesson 13, drawing from Sherlock Holmes famed story, “The Adventure of Silver Blaze,”⁶ defines negative evidence as

“the sound of the dog not barking … the absence of what should happen under a given set of circumstances. A watch dog is expected to bark. If it doesn’t, then its very silence attests the likelihood of certain things.”⁷

Negative evidence should not be confused with negative findings, in which research turns up no evidence at all, pro or con.⁸  With negative evidence, relevant sources survive that, logically, would provide relevant information; but those sources are totally silent on our question.  In other words: A dog was there, but the dog did not bark.

Among the evidence developed to disambiguate between the two John Wattses of the Georgia Wiregrass, one piece of negative evidence was this:

Question: Did John Watts of Washington and Montgomery Counties serve in the American Revolution, as alleged?

Logical source of information: State of Georgia land grant records, resulting from a process by which residents of the state could apply for Revolutionary War bounty land, if they were eligible, or for headright land. John Watts of Montgomery applied for and received two tracts of land in what became Montgomery and then Tattnall County. In both cases, he applied for headright land. Never did he request bounty land.⁹

Evidence? Negative. The dog was there. Relevant bounty land records were created and still exist. John Watts showed up in the presence of the dog. But the dog did not bark Bounty land! Bounty land!  Where the dog did bark was amid the set of headright grants. For the question of John’s alleged service to the Patriot cause, the evidence is negative.

Up to this point, we have used traditional records, evaluating them in traditional ways, to develop three pieces of evidence for the identity and activities of John Watts of Montgomery County, Georgia. The core principles we have defined here also apply to the use of DNA test results as evidence. For that, we will continue with DNA evidence relevant to John.

Evaluating DNA as Evidence

Historical researchers currently use four types of DNA for historical research:

• Autosomal (atDNA), being that found in our 22 numbered pairs of chromosomes.
• Mitochondrial (mtDNA), being that found in the energy-producing organelles that fuel our cells.
• X-DNA, being the female sex chromosome, of which females inherit two and males inherit just one.
• Y-DNA, being the male sex chromosome that females do not inherit.

Let’s use these to examine how DNA test results can yield direct, indirect, or negative evidence.

Direct Evidence (Using Y-DNA as an Example)

Research Question:  Were the two John Wattses of Washington County, Georgia (who have now been separated into two separate men¹⁰) part of the same Watts family?

Information: According to the Watts Y research project, male Y-line descendants of Col. John Watts, Esq., of Washington County (whose descents are acceptably proved) fall into haplogroup E-L117. Meanwhile, male Y-line descendants of Rev. John Watts, Esq., of Washington and Montgomery (whose descents are soundly proved) fall into haplogroup I-M223.¹¹

Evidence: Direct. We have soundly conducted genetic tests that tell us the two Watts Y-lines are radically different.

Indirect Evidence (Using mtDNA as an Example)

Information: My children, my maternal aunt and uncle, my sibling, and I all have matching mtDNA with Gwen Holloway Morgan.¹²

Research question: Given that Gwen has sound documentary evidence that her third-great-grandmother in the mt-line was one Nancy (Watts?) Parks,¹³ does this mean that I also descend from Nancy?

Evidence? Indirect. The test results are relevant to the question. The possibility exists. However, this DNA evidence alone cannot provide an answer to the research question because other possibilities also exist. Our common mt-ancestor could be a woman in the much-more distant past. We needed to find other compatible evidence to combine with this mtDNA evidence to prove or disprove the possibility.

Indirect Evidence (Using the X Chromosome as an Example)

Information:  A sizable section of my X chromosome is also shared by Harold Terry. Our kinship is well-proved with traditional documentation. He shares mtDNA with both Gwen and I, but he descends from Nancy’s sister Elizabeth.¹⁴

Research question: Does this X chromosome data support an hypothesis that I descend from Nancy?

Evidence?  Indirect.The test results are relevant to the question. The possibility exists because Terry’s descent from Elizabeth follows one of his X-chromosome paths and my hypothesized descent from Nancy also follows one of my X-chromosome paths.¹⁵ We could inherit that matching segment of the X from the same woman: the mother of Nancy and Elizabeth. However, the DNA results, alone, do not allow us to make this claim. To do so—or to disprove the hypothesis—we must find and correlate other relevant evidence.

Negative Evidence (Using atDNA as an Example)

Research problem: Documentary evidence has established that the Montgomery County John Watts is the great-grandson of an early-1700s settler of Albemarle County, Virginia, named William Mills.¹⁶  Documentary evidence also points strongly to the fact that my late husband descends from that same William Mills. This matters doubly to me, your teacher-of-the-moment, because I have now soundly established my descent from John Watts.¹⁷

Question: Does DNA evidence support a relationship between me and the Mills family into which I married?

Information: atDNA test results for me and four of my siblings-in-law report no shared segments.

Evidence: Negative. In this case, the test is the dog. If we have measurable matching segments then the dog should bark. However, at this distance, the likelihood of—say—7th cousins having shared DNA ranges from 3.2% (Ancestry's calculation) down to 1.1% (23andMe).¹⁸

However, as with indirect evidence, negative evidence cannot stand alone. To reach a conclusion, we must couple it with other forms of evidence that support or rebut the negative evidence.¹⁹ In the case of atDNA, experienced users of genetic tests know that other explanations exist for a negative result—two common ones being these:

• The chances of inheriting DNA from any specific ancestor is halved at each generation. After the passage of eight-to-ten generations, as in this problem, we might not carry any segments at all from a specific forebear.
• Reported matches are based on algorithms created by each testing company. Typically, they require a certain number of segments of an arbitrarily set minimum size, or a large segment of a certain length, and/or a total centimorgans length of a certain size for all segments combined. When segments fall short of these criteria, they are not reported, even though matching segments may still exist.²⁰

For these and other reasons, negative DNA evidence cannot, alone, disprove a point.

Bottom Line:

Whether we are working with DNA evidence or traditional documentary evidence, no one piece of evidence can serve as valid proof. Direct evidence can be wrong, even DNA evidence when false matches—aka IBS/IBC situations—exist.²¹ Indirect evidence is but one piece of a puzzle that will necessarily have many pieces; once assembled, a body of indirect evidence could actually serve to disprove direct evidence. Similarly, negative evidence is just one indicator—one suggestion—that a certain hypothesis might be proved or disproved. Sound research uses all types of evidence. It strives to find and correlate all possible pieces of evidence.

In the end, as researchers we must set forth an argument for why we believe that x, y, and z (as well X-, Y-, mt- and at-) all add up to a sound conclusion. To do that, we must understand the three different kinds of evidence, their strengths and weaknesses, and how we can use each type to arrive at that sound conclusion.