Linguistic Confusion in Court: Evidence from the Forensic Sciences
Journal of Law and Policy, 21, 515-539, 2013
26 Pages Posted: 3 Mar 2013 Last revised: 3 Jun 2013
Date Written: March 3, 2013
This paper, which was presented at the Authorship Attribution Workshop at Brooklyn Law School, addresses issues related to how scientific and technical information should and should not be communicated in court. This issue has special relevance for authorship attribution testimony and forensic linguistics more generally. Because confusion in the DNA and fingerprint areas has been documented and is relatively common, the paper focus largely on linguistic problems in DNA and fingerprint expert testimony in hopes that forensic linguists can avoid the testimonial traps and errors that plague these forensic scientists. Section I examines DNA match statistics and describes the confusion that legal actors experience when dealing with conditional probabilities. Section II examines statistical inverse errors in the 2010 U.S. Supreme Court case McDaniel v. Brown. Section III examines a seductive, but faulty, statistical assumption that commonly arises in paternity cases. Section IV examines the role of error rates in forensic sciences and concludes that identifying those error rates is particularly important in fields that rely on highly discriminating statistical techniques. Section V offers an illustration of the crucial role ill-defined language can play in a legal proceeding. Standard and precise terms are recommended. The paper concludes with a section identifying implications for the forensic linguistics and authorship attribution communities. This paper is forthcoming in the Journal of Law and Policy.
Keywords: forensic science, statistical evidence, forensic linguistics, authorship attribution, DNA evidence, fingerprint, Bayes, error rate
JEL Classification: C11, D81, K10, K19, K40, K49
Suggested Citation: Suggested Citation