Forensic linguistics: An introduction to language, crime and the law (review)

The term ‘forensic linguistics’ is used in two different, but related, senses. In its broadest usage, it serves as a cover term for the entire field of language and law with its multifaceted endeavors and subfields. This broad meaning is manifested in the journal The International Journal of Speech, Language and the Law: Forensic Linguistics (which began publication in 1994), Sue Blackwell’s forensic linguistics website at Birmingham University (http://web.bham.ac.uk/forensic/), and the e-mail discussion list for ‘Language and the Law’ (forensic-linguistics@jiscmail.ac.uk; see http://www.jiscmail.ac.uk/archives/forensic-linguistics.html).

A much narrower notion of forensic linguistics (which parallels the term ‘forensic medicine’) focuses on the ‘study of any text or item of spoken language which has relevance to a criminal or civil dispute’ (see http://www.thetext.co.uk/info.html). This usage derives from the classic work by Svartvik (1968), which may be the first appearance of the term ‘forensic linguistics’. Olsson’s Forensic linguistics takes a very narrow approach to this already narrow conception and thus has to be a disappointment to anyone who might be looking for a broad introduction to language and law that could, for example, be adopted for a general course on the subject.

Most of O’s book is devoted to describing how authorship inquiry can be carried out through textual analysis, what is now often called forensic stylistics (see e.g. the excellent volumes by McMenamin (1993, 2002)). The essential issue in all cases is whether the text is authentic and whether some individual was or was not the author of the questionable document. This text could be a will, a suicide note, or a piece of hate mail, the forensic text type being a critical factor in the analysis (Ch. 10, 143–69). Variables to consider include linguistic factors such as word choice, morphological formations, and syntax, plus written language features such as spelling, capitalization, and punctuation. O points out that the analysis can be qualitative, that is, identifying idiosyncratic individualistic markers or characteristics, or quantitative, that is, requiring large samples and sophisticated statistical analysis. O is at his best when he illustrates the use of statistical techniques for determining authorship, but unfortunately he invariably backs away when the discussion begins to get really interesting for fear that even simple math might be over the heads of his readers. An essential point that O touches on, but does not emphasize nearly enough, is that the real contribution of forensic linguistics to authorship determination is often in being able to state clearly that some individual could not have been the author of the document in question even when the real author has not been identified.

Ch. 3 (41–48) discusses legal requirements concerning the admissibility of expert evidence in court. This chapter is pointless for at least three reasons. First, as O himself notes, the rules vary from country to country and thus anything he says here (e.g. his discussion of the American Daubert rule) wouldn’t apply elsewhere. Second, in a trial it is up to the attorneys and the judge to understand and apply the rules governing expert testimony, not the forensic linguist. Third, and this important point is often neglected, most legal matters never go to court and thus the potential contribution of a forensic linguist in helping to solve a problem will never be affected by admissibility rules.

Plagiarism is a topic of enormous interest these days both from analytical/theoretical and from practical points of view. It is a topic about which forensic linguistics potentially has a tremendous amount to contribute. Thus, when one finds that O devotes only ten pages to the subject (Ch. 8, 109–18), one wonders why he even bothered, especially since he buys into the totally misguided notion—unfortunately, shared by many university administrators—that plagiarism is a form of theft when the only sensible characterization of...