02/27/2011 02:58 pm ET | Updated May 25, 2011

Julia Roberts Isn't a Roberts

Earlier this week, at the US-Ireland Alliance's "Oscar Wilde: Honoring the Irish in Film" event, Julia Roberts presented Paul Rudd with an "Honorary Irishman" award. Showing a touch of Irish-envy, she remarked, "Maybe I can become Irish. My middle name is Fiona, which is a step closer to being Irish than Paul Rudd was five minutes ago." That's essentially an engraved invitation to a genealogist like myself (a half-Irish one at that) to go digging.

Ms. Roberts, it turns out, is a classic American mutt. Her father's side is mostly Southern of British Isles stock. Her roots on this half of her family tree extend back deeply enough that it would take extensive research to nail down the specifics, but the surnames suggest a mix of Irish, English, Scottish and Welsh. By contrast, her mother's side crosses the pond more quickly back to Germany and Scandinavia.

Though I began my quest seeking evidence of Irish heritage, much like the child who's captivated by a shiny object, I found myself distracted by an unexpected discovery. Ms. Roberts apparently isn't really a Roberts, though it's highly unlikely she would be aware of this.

If you trace her Roberts line back in time, you'll find that her great-grandfather, John Pendleton Roberts, was supposedly the son of Willis R. and Rhoda "Rodie" (Suttles) Roberts. Rhoda's tombstone makes the relationship of this couple clear, as does their 1857 marriage and her 1891 application for a pension as the widow of a Confederate soldier. But it was this same pension record that caught my attention. Both this file and his service records indicate that Willis R. Roberts died on March 25, 1864 (of pneumonia and fever, if you're curious).

And that leads to a little problem: John was born in 1878, roughly 14 years after his father's death. In both the 1870 and 1880 census, the widow Roberts shows up with her well-spaced children -- two presumably the soldier's born around the 1858-1861 time-frame, another pair born around the turn of the next decade, and then John born as yet another decade approached. The obvious question, then, is who was really John's father?

I have a theory, but the only way to convincingly prove or disprove it would be DNA testing. In fact, this is a classic genealogical conundrum that cries out for genetic investigation, but I'll leave it to Ms. Roberts to decide if she wants to learn what her surname would have been.