In lieu of an abstract, here is a brief excerpt of the content:

  • Accidents Happen
  • John Cheney-Lippold (bio)

ROBINSON V. MANDELL

Sylvia Ann Howland died on July 2, 1865, in New Medford, Massachusetts. A rich woman at her time of passing, she fell on poor health at the close of the United States Civil War. In the years preceding, she served as the mother figure for her niece, Hetty Robinson, and in typical high-society dramatic form, a series of conflicts over the beneficiary of her $2-million estate emerged immediately after her death.

First, Howland's will had originally allocated half her fortune to various recipients, with the remainder earmarked for her niece, Robinson, in the form of a trust. Then, following Howland's death, Robinson mysteriously supplied a second, "earlier" will. According to this will, allegedly signed by Howland herself, the estate would be given in full to Robinson. And attached to this "earlier" will was an addendum: any subsequent will was deemed invalid, voiding all prior potential distributions of the Howland fortune.

I assume the reader concurs with the executor of Howland's estate, Thomas Mandell, and audibly scoffed, accusing Robinson of a thinly disguised forgery. The timing seemed odd; the future-oriented nullification even more odd. And most odd was the manner by which the actual case, Robinson v. Mandell. 1868. 20 F Cas 1027, C.C.D., would proceed. In this case, Mandell asserted that two of the three signatures (called signatures no. 10 and no. 15) in Robinson's earlier will were copies of the third (signature no. 1). Robinson, he argued, had traced the former two signatures according to an actual, authenticated signature. Like a carbon copy of a check, the contours of the letters appeared as impossibly exact mirrors of the source, sounding the alarm about fraud. [End Page 513]

But how could such subjective larceny be proved? Mandell mobilized the help of Benjamin and Charles Peirce, the respective grandfather and "father of pragmatism," who analyzed the disputed signatures according to the burgeoning field of mathematical statistics. The Peirces' method contrasted the commonalities between the signature no. 10 authenticating future wills' nullification and 42 other signatures written by Sylvia Ann Howland in the final years of her life (e.g., see figure 1). In writing her full name, Howland would use 30 discrete downstrokes. Signature no. 10, it appeared, matched flawlessly with one of the three signatures (signature no. 1) that appeared on the amended page (Meier and Zabell 1980).

Benjamin Peirce regarded this perfect fidelity as absurd. The probability that each of the 30 downstrokes in Howland's name would be exact was, he calculated, 1 in 2.6 x 10^21 (over two and a half sextillion). This "vast improbability" was "practically an impossibility … such evanescent shadows of probability cannot belong to actual life" (Meier and Zabell 1980, 499). The signatures were so similar that Peirce deemed it statistically inconceivable for the same person to have penned them both. The facsimile nature of no. 1 and no. 10 was, according to Charles Peirce, "an infallible evidence of design" (499). Charles would continue his testimony with an even bolder claim: "the mathematical discussion of this subject has never, to my knowledge, been proposed, but it is not difficult; and a numerical expression applicable to this problem, the correctness of which would be instantly recognized by all the mathematicians of the world, can be readily obtained" (499).

I offer the case of Robinson v. Mandell as an introductory paradox of statistics and algorithmic logics, a paradox that undergirds how I will later describe the concept of the accident. In the above account, two signatures were construed as too similar to be signed by the same person. While Hetty Robinson was clearly not the first individual who dared to forge a signature for financial gain, Robinson v. Mandell marked one of the first instances when a subject's authenticity was fully reconsidered, and thus redefined, on the grounds of statistics.


Click for larger view
View full resolution
Figure 1.

The unquestioned original signature (no. 1) and the two disputed signatures (nos. 10 and 15) American Law Review 1870, 641.

The Peirces' method, which amounted to mathematically assessing 30 downstrokes in relation to 42 other signatures...

pdf

Share