Art Forensics #4:

ntroduction

In the first two articles regarding forensic science and its application to fine art authentication we focused on mechanical print and handwriting analysis as telltales of document authorship. In this article the focus will be on authorship determination as found through the process of composing the document by an author rather than the actual structure of letters and words. Document analysis here will focus on vocabulary used, punctuation and grammar identifiers, horizontal and vertical alignment of writing, colloquialisms, verb/noun ordering, sentence complexity, and active versus passive sentence presentation. Commonly referred to as forensic linguistics, it is also known as literary forensics or stylistics. It is both quantitative and qualitative; it uses statistics to analyze data and derives conclusions based on experience and intuition.

Practitioners of forensic linguistic analysis argue that establishing document authorship is more of a convergent process than a one step cause and effect: authorship is determined by a convergence of evidence factors leading to a logical and credible conclusion. Just because the vertical alignment of the questioned document is the same as that found in the exemplar, it does not follow that the document was authored by the exemplar author. It is the combination of structures and style which lead one to conclude that the writing is by the same author.

Case Studies

McMenamin argues in his Forensic Linguistics: Advances in Forensic Stylistics that studies of linguistics (stylistics) as predictors of authorship go as far back as the middle of the 19th century. The practice of establishing authorship via analysis of literary style in documents goes back to the 1930s case of the Lindbergh Kidnapping in New Jersey. Here, a match between the kidnapper's writing structure and the words used linked Bruno Hauptmann to the kidnapping. The writing samples were later used to convict Hauptmann. In 1968 Swedish linguist Jan Svartvik was able to demonstrate that a murder confession from Timothy Evans did not match other samples of Evan's writings. It was unfortunate that Svartvik's work took place years after Evans was executed. The notion of literary footprints was legitimized by Don Foster in his study of the authorship of a Shakespeare sonnet, Joe Klein as the author in Primary Colors, the Unabomber Trial, and the Hatfill Anthrax case following 9/11.

The Process of Forensic Linguistics

To the extent that our cultures and education determine how we behave, speak, and dress similarly they influence and shape how we write. The ultimate consequence of such conditioning is that coupled with our individual make-up or our genes each person has a distinct way of expressing him or herself via the written word. The determination of authorship looks at those factors identifying the written word in terms of geographical origination (such as Germanic background indicators in the writings of the Lindbergh kidnapper), or the use of male versus female words in the Israeli study (Magazine Desk, 2003) indicating that it is possible to identify with an 80% accuracy whether a male or female authored a given document. Taken together with an exemplar writing sample it is possible to construct data subject to statistical analysis to provide a given conclusion. Thus, by identifying cultural, educational, geographical, gender, word usage, punctuation and idiosyncratic writing patterns of a questioned document and comparing it with an exemplar, an investigator may draw a conclusion of a match or no match. Don Foster used a computer program called "SHAXICON" to compare word usage of the, A Funerall Elegye in memory of the late Vertuous Maister William Peeter and Shakespeare's writings to reach a match conclusion.

A Composite Example

The field of forensic linguistic analysis has not reached a point where a given methodology operates to enable an investigator to arrive at a sound conclusion. Forensic linguistics moves in different directions using an array of structures and methods. A composite example is instructive. Imagine that you have just purchased a Michelangelo drawing for $2,000,000. With the drawing you received documents written by the artist asserting that the drawing was given to the Medici family as a gift for their generosity during his early years. You call in a forensic expert to examine the document to determine if in fact the written style of the document matches that of Michelangelo.

The expert begins his analysis by first securing an exemplar/s of the artist's writings. Michelangelo was noted for his many writings, letters, and poetry. The expert begins counting the number of articles used, particular words, active or passive voice, misspellings, length of sentences, use of punctuation, level of complexity of words, and grammar. He develops data sets from the provenance document in terms of words, etc. provided to the purchaser, and compares data sets from existing documents. The expert inputs the data sets into a statistics software program. The program will help determine whether the data sets are the same or different. The expert may only offer an opinion as to the style of the document based on the analysis. Style analysis does not include handwriting analysis or the structure of the writing in terms of slant of words, alignment etc.

Forensic Linguistics and its Problems

Although forensic linguistics has been around for some time, it can best be described as having only reached its infancy. Similar to computer analysis of artistic style provided by Farid discussed in New Technologies of Fine Art Authentication (FAR) much work still has to be accomplished before it can be considered as a mechanism of scientific analysis.

The problems forensic linguistics must overcome are as follows:

1. In any scientific undertaking, analysis demands reliability. Except for a few successes such as the Lindbergh case (although there are some critics who would argue that the kidnappers were framed), many cases exist which turn out to be unreliable. Steven Hatfill sued the forensic linguist for his findings in an anthrax case investigated by the FBI. The detective who wrote a book on the findings of a forensic linguist in the JonBenet Ramsey case was sued for using the findings in his book.
2. It is not clear how the field wants to structure itself. Is it literary forensics, stylistics, or what? It is a field searching for a definition.
3. To the extent that handwriting samples change as a result of age, infirmity, drugs, and so on, it is reasonable to assert that those factors or telltales in a document will change also. The "you all" from the South would certainly be impacted by spending time in a northeast city working as a manager in a large company.
4. Although the telltales of style may be present in a certain culture, how significant are such telltales or identifiers in predicting where a person came from?
5. Just because a person uses complex words does not mean he or she is highly educated. Dictionaries are readily available to supply complex words.
6. The use of quantitative analysis to ground forensic linguistic conclusions has been taken to task by those who use it. Foster noted that using statistics in linguistic analysis is not grounded in sound science.
7. The Daubert court rulings have demanded that experts providing testimony must demonstrate that they are experienced, educated, trained, and that their endeavors have been reviewed by their peers. Except for one course provided by a forensic linguist there are not a lot of areas where a forensic linguist could demonstrate that he or she satisfies the requirements or standards of expertness.

Summary

It is not altogether clear what or where forensic linguistics is moving towards. Its ambiguous nature grounded on the subjectivity of stylistics makes it more apparitional than concrete. I do not think that it has demonstrated that merely because one writes in a certain fashion today, that same fashion will be duplicated tomorrow or next year. Its reliability is weak, and it seems to be trying to connect to a field of science when in fact its literary content is somewhat beyond the realm of science. The potential is there as long as one can extract data from its sources and at the same time keep the skin on the skeleton.

Conclusion

From the above analysis of forensic linguistics it is reasonable to assert that at this stage of its development its use as a tool in provenance determination or analysis leaves much to be desired. Although it has potential, particularly when used in conjunction with other forensic approaches to authentication, that potential is not close at this point: very few practitioners, little in the way of available experience and education providers, few identified peer groups, and comparatively little in the way of journals acknowledging its structures, systems, and processes seem to be grounding its reviews and conclusions.

— by Dr. John Daab CFE, CFC | August 31, 2009