BEST PRACTICES 345
the late 1990s, I had developed a method now known as SynAID
(Syntactic Author Identification) within ALIAS (Automated
Linguistic Identification and Assessment System).^31 This research
has played a role in adjudicated cases in 1998, 2001, and 2008,
discussed later.
Litigation-independent validation testing on forensically
feasible ground-truth data is a core feature of the forensic
computational linguistics approach. Implementation in software
that is responsive to messy data is central to the forensic
computational linguistics approach for both replicability and
error control. Linguistic theory plays a central role in the
forensic computational linguistics approach. These features
distinguish the forensic computational linguistics approach in
sometimes obvious, sometimes subtle ways from forensic
stylistics and stylometric computing.
A. Linguistic Theory Does MatterIn linguistic theory, language is divided into levels for
analytical purposes.^32 These levels are sound, word, and word
combinations.^33 These levels, respectively, are analyzed in
phonetics and phonology; morphology and the lexicon; syntax;
semantics and pragmatics; and prosody.^34 These levels have
determining authorship, Grant ID 1995-IJ-CX-0012, Visiting Fellowship,
Linguistics Methods for Determining Authorship.
(^31) See Chaski, Empirical Evaluations, supra note 16; Chaski, Empirically
Testing, supra note 16; Carole E. Chaski, Recent Validation Results for the
Syntactic Analysis Method for Author Identification, International Conference
on Language and Law (2004) [hereinafter Chaski, Syntactic Analysis Method
Identification]; Chaski, Who Wrote It?, supra note 1; Chaski, Who’s at the
Keyboard?, supra note 16.
(^32) This division of language into analytical levels is commonplace in
standard textbooks in linguistics. See e.g., RICHARD AKMAJIAN ET AL.,
LINGUISTICS: AN INTRODUCTION TO LANGUAGE AND COMMUNICATION (6th
ed. 2001); EDWARD FINEGAN, LANGUAGE: ITS STRUCTURE AND USE (6th ed.
2012); VICTORIA FROMKIN ET AL., AN INTRODUCTION TO LANGUAGE (10th
ed. 2013).
(^33) See AKMAJIAN ET AL., supra note 32; FINEGAN, supra note 32;
FROMKIN ET AL., supra note 32.
(^34) See AKMAJIAN ET AL., supra note 32; FINEGAN, supra note 32;