334 JOURNAL OF LAW AND POLICY
This article focuses upon the forensic computational
linguistic approach and contrasts this approach to the forensic
stylistics and stylometric computing approaches. In Section II,
best practices for forensic linguistics are presented. The best
practices provide an evaluative framework for the forensic
computational linguistics approach, discussed in Section III; the
forensic stylistics approach, discussed in Section IV; and the
stylometric computing approach, discussed in Section V. In each
section, a discussion of admissibility is included, since best
practices should guide both judicial reasoning as well as
scientific practice.
II. BEST PRACTICES FOR FORENSIC LINGUISTICS
Best practices in forensic linguistics are essential to propel
the field of authorship identification from an academic or law
enforcement sideline consultancy to a real forensic science that
is useful to the judicial system. Best practices include factors
from both the legal standards for evidence, so as to be useful
and address admissibility concerns, and scientific standards for
research, so as to be reliable, replicable, and respectable.
Scientifically respectable and judicially acceptable methods
for author identification should be:
a. developed independent of any litigation;
b. tested for accuracy outside of any litigation;
c. tested for accuracy on “ground truth” data;
d. able to work reliably on “forensically feasible” data;
e. tested for known limits correlated to specific accuracy
levels;
f. tested for any errors of individual testing techniques that
could cause accumulated error when combined with other
techniques;
g. replicable;
h. related to a specific expertise and academic training;
i. related to standard (“generally accepted”) techniques
within the specific expertise and academic training; and
Idiosyncrasies for Authorship Attribution, PROC. IJCAI’03 WORKSHOP ON
COMPUTATIONAL APPROACHES TO STYLE ANALYSIS & SYNTHESIS, 2003.