Experienced genealogists are always aware that they must verify information by looking at original documents or a microfilm or digital image of an original document. We should know better than to believe a statement on a web site, in a genealogy book, or a verbal statement from Aunt Tilley about the "facts" of our family trees. However, what is the definition of an "original document?"

Let's take one well-known claim of an original document that isn't really accurate: the U.S. Declaration of Independence. Almost all American schoolchildren are familiar with this document; and, if we paid attention in class, we know that the document is on display at the U.S. National Archives building in Washington, D.C. In fact, millions of us, myself included, have visited that building to view the document on display. However, how many of us were ever told that the document displayed in Washington is not the original, hand-written document? Instead, it is one of many copies that were produced on a printing press.

No, this isn't a story plot from a Nicholas Cage movie. In fact, the document displayed at the National Archives building in Washington, D.C. is a copy made by Philadelphia printer John Dunlap, official printer to the Congress, during the evening of July 4, 1776, after the original, hand-written document was given to him. Admittedly, the original and the copies made by John Dunlap had no signatures. The "copy" now on display at the National Archives is the only copy that was actually signed by each delegate and therefore is the one that we can now refer to as the real Declaration of Independence. However, it was produced on a printing press and is not the original, hand-written piece of paper.

The original Declaration of Independence was written by hand by Thomas Jefferson. After making alterations to his draft as suggested by Ben Franklin and John Adams, Jefferson later recalled that, "I then wrote a fair copy, reported it to the Committee, and from them, unaltered, to Congress." 

The committee sent the hand-written manuscript document, probably Thomas Jefferson's "fair copy" of his rough draft, to John Dunlap, official printer to the Congress. Dunlap printed the copies on the night of July 4, 1776. It is unknown exactly how many copies were printed, but the number is estimated at about 200. On the morning of July 5, copies were dispatched by members of Congress to various assemblies, conventions, and Committees of Safety as well as to the commanders of Continental troops. Also on July 5, a copy of the printed version of the approved Declaration was inserted into the "rough journal" of the Continental Congress for July 4. The text was followed by the words, "Signed by Order and in Behalf of the Congress, John Hancock, President. Attest. Charles Thomson, Secretary." 

Twenty-six copies are known to exist today of what is commonly referred to as "the Dunlap broadside," 22 owned by American institutions, 3 by British institutions, and 1 by an unknown private owner. A list of their present locations may be found on Wikipedia at https://bit.ly/2ZqPFCM. 

All of these copies were unsigned as they were printed before approval had been granted by the 13 colonies. Each delegate had to await approval from his home colony before being allowed to sign.

Contrary to popular belief, the U.S. Declaration of Independence was not signed on July 4, 1776. While the document was APPROVED by the delegates on July 4, several weeks were required for the document to be printed and distributed to all 13 colonies for approval, and then some more time to re-assemble all the delegates again in Philadelphia. Delegates were not authorized to sign until after their home colony had approved the document and that required some time back in the days before instant communications.

One of the "Dunlap broadside" copies was signed by all the delegates in attendance on August 2, 1776, and that copy now is on display at the National Archives in Washington, D.C. Therefore, the document that most people think of as the U.S. Declaration of Independence is not the original, hand-written document. It is a copy, although it is the only SIGNED copy. The copy on display in Washington was printed on a printing press, but each delegate signed this one copy by hand.

If Thomas Jefferson's memories were correct, and he indeed wrote out a fair copy which was shown to the drafting committee and then submitted to Congress on June 28, the original document has not been found. "If this manuscript still exists," wrote historian Ted Widmer, "it is the holy grail of American freedom." 

What does this have to do with our searches for accurate genealogy information? A lot.

In all cases, we should strive to look at original documents or a microfilm or digital image of an original document. We then document our efforts by recording a "source citation" that refers to the location of the original document. In recent years, many genealogists also include a digital image of the appropriate part of the original document.

Wikipedia defines a citation this way: "Broadly, a citation is a reference to a published or unpublished source (not always the original source)." Bold text added by myself for emphasis.

Elizabeth Shown Mills, probably the leading expert of today when it comes to recording genealogy source citations, has written no less than two books on the subject for genealogists: Evidence Explained; Citing History Sources from Artifacts to Cyberspace and Evidence! Citation & Analysis for the Family Historian. Mills states that the best source is an original source, one created at the time an event occurred. However, she also states that a source can be either an original or derivative document. Let's focus on the word "derivative."

In the example of the Declaration of Independence, we note that the original, hand-written document has been lost. For all we know, printer John Dunlap may have tossed the original into a local trashcan after he finished making his copies. Of course, that is just a guess. Nobody knows what happened to the hand-written original. However, the existing twenty-six copies still meet Elizabeth Shown Mills' definition of an acceptable source citation. It is a "derivative document" that was made at nearly the same time as the original, probably within a few hours, and apparently is an exact copy of the original. Therefore, it is believable.

Of course, not all derivative documents are exact copies. For instance, let's consider the U.S. Census records. In most cases, the enumerators (census takers) visited homes, asked questions, and wrote the answers in small notebooks or something similar that they carried with them. We can only imagine what the notebooks contained. Can you imagine the words written by an enumerator with poor handwriting, traveling around the countryside on horseback or on a small wagon in the 1800s and recording his words with a quill pen and ink of questionable quality? Some of these enumerators traveled in rain or sleet or snow. We have to assume that some of these pages got wet. Perhaps a few pages became unreadable or were even lost.

At a later time, the enumerator went home or to an office or perhaps to a local tavern, got out the official enumeration pages that we all know and love, and transcribed his findings from the notebooks to the worksheets. He then sent the worksheets to his superiors, where the worksheets became the official record. Most of the worksheets have been preserved while most of the notebooks were discarded. 

What do we see today when we look the census records online? Do we see the enumerators' hand-written notes from their workbooks, made at the time of each visit? Or are we viewing the official forms that were filled out later, also hand-written? You probably already know the answer: what we see online and in microfilms are images of documents made within hours or days AFTER the original visit. These are derivative documents.

There is nothing wrong with using a derivative document. In fact, it is all we have in most cases. A derivative document was made at or shortly after the original event by a person who had full knowledge of the facts involved. Genealogists will generally accept a derivative document as a suitable "original" source citation.

You can find thousands more examples of citing "original" sources that are really derivative documents. Most birth records made by town clerks prior to the twentieth century were recorded by men or women who were not present at the birth. They weren't midwives; they were town clerks! They recorded what was told to them by reliable witnesses, often the mother or father or perhaps the doctor or midwife in attendance at the birth. Marriage records were often the same. Town clerks may not have attended the marriage ceremony; but, in most cases, the clerks recorded information given to them at a later date by the clergyman who performed the ceremony.

The list goes on and on. Death records, military service records, and thousands of other documents were not recorded at the event by the individuals involved and often not within hours. Instead, these documents were recorded by clerks and clergymen and others shortly after the event and were based upon information provided by the principal(s) involved. In some cases the description was verbal while in other cases the clerks transcribed written information created earlier. 

Land transfer records recorded in deed books were rarely written by the individuals who bought or sold land; they were written by clerks who listened to the descriptions provided by the principals involved. Probate records typically were transcribed from original, often hand-written wills, often years after the will was written and always after the person who perhaps wrote the will had died. 

While these may be derivative records, we still accept them as primary source citations.

NOTE: In contrast, secondary sources are generally those records created after a passage of time. Examples include an elderly person recounting events in his or her youth or an author of a genealogy book recording the life events of people who have been dead for many years or other people whom he or she has never met. Secondary sources are never as reliable as primary sources.

Now that we have examined both kinds of primary sources (original and derivative) as well as secondary sources, a question arises: Just how reliable are derivative (primary) sources?

Genealogists generally consider primary sources to be reliable, including derivative sources. After all, these records were made at or shortly after the event and were recorded by eyewitnesses or, in the case of derivative records, by transcribers who were given information by eyewitnesses. Yes, we all know that eyewitness reports occasionally contain errors, but are usually correct. 

How about derivative records where the information was recorded by a third party, using information provided by eyewitnesses? Can we really trust the enumerator's record made some hours or days or even a few weeks after visiting our ancestors? Could he read his own writing, smudged from rain or melting snow? If he was sitting at a fireplace in a warm and cozy tavern, already having consumed a few drinks, can we believe his written recollection of a visit made a few days earlier? Did the residents give him correct information? Or did he obtain his info from a neighbor who may or may not have known all the correct answers? 

When an eyewitness provided information to a clerk, can we always believe that both parties understood clearly what was said and the information recorded by the clerk is a true and faithful recording of the facts provided by the eyewitness?

For an example, I will offer the 1910 census record for my great-grandparents in a small town in northern Maine. The enumerator lived in the same town and recorded his own family on another page of the same census. (That is one record that I would believe!) He wrote his place of birth as "Scotland." Therefore, we can assume that he spoke with a Scottish accent, perhaps a very strong accent.

My great-grandparents were Joe and Sophie Theriault. (My great-grandfather was often listed in many records as Joe, but never listed as Joseph.) "Theriault" is a common Acadian French-Canadian name. They lived in the same town. The enumerator recorded that neither of them was able to speak English. Can you imagine that conversation? A Scotsman with his accent trying to ask questions and obtain answers from someone who could not speak English? It is no wonder that their last name was recorded as "Tahrihult."

To make matters worse, there was another couple in the same small town with the same names: Joseph and Sophie Theriault. However, they were listed with different children and different dates of birth and marriage. The enumerator then spelled their last name correctly! I assume the correct information was recorded for them because the enumerator listed them as being able to speak English. That should have been an easier, and probably more complete, conversation. Yet all of these are derivative records, written by the enumerator a few hours or days after his interview and assisted by the notes he made in his notebook.

Similar errors have been repeated thousands of times in census records and probably elsewhere as well.

Let's return to the original question: How accurate are derivative records where the information was recorded by a third party, using information provided by eyewitnesses?

I believe the only correct answer is: Derivative records made at or shortly after the event are generally correct, but we should always be aware that there are exceptions. All derivative records should be treated as “probably correct”–with a strong emphasis on the word PROBABLY.