A brief introduction to authorship analysis

            We encounter phrases without knowing the author of them on a daily basis, but rarely do we wonder about the person from whom they originated. In this essay my goal is to provide an outline on what are some of the possible methods for analysing the authorship of a given text. The essay will begin with a brief overview on the field of forensic linguistics; this will be followed by an elicitation of different clues to be sought for in trying to establish the possible authors. Also, examples will be provided where possible.

Forensic linguistics is a branch of Applied linguistics and as such it deals with knowledge about language. Its peculiarity is that it basically does this in a judiciary context as it can be used to establish authorship of texts. It possesses great importance since by forensic linguists cooperating with the police and the court of justice there is a higher possibility of catching criminals who previously would have a narrow escape.

In order to be able to determine the author of a given text, a thorough and widespread investigation needs to be conducted the first stage of which is the evidence gathering. An important clue to be sought for is the presence of those peculiarities that are recurring all the way through of the text. The highest importance of them are the investigation of marked and unmarked terms; the latter referring to the dominant form and the former to the more rare form (Olsson, 2008). Though as it is stated in the introduction of An Introduction to Forensic Linguistics: Language in Evidence, “cases raise new and exciting questions for descriptive linguistics, which require basic research, such as how can one measure the ‘rarity’ and therefore the evidential value of individual expressions, or how can one assess the reliability of verbal memory” (Coulthard & Jones, 2007, p. 6). There is a scale of importance for the different types of structures: of the highest importance with a distinctive feature are the structures of longest matches, whereas spelling mistakes are the least suitable in determining the authorship. For instance, in the case of a man who was the suspect of a rape, a few witnesses received threatening and malicious letters by a ‘friend’ of the suspect claiming that he did not commit the rape. These were compared to the statements made by the suspect himself and a tendency to similar disordering was found: past-present formulations are more frequent than present-past ones (“[…] as I am an was impotent” (suspect) and “[…] that he is and was impotent” (‘friend’)) and what can be called a logical disorder (in both the known text written by him and the unknown ones: “the proof and evidence was given to the police” or “in the past years and months”).

Style is a good lead because it has an unconscious relation as well that cannot be controlled by the individual and this makes it usable and identifiable. For instance, in a case where child pornography was involved, the likelihood of a suspect being the author of an email to a website dealing with child pornography was established with this technique. There were two men involved, a businessman with the computer from which the email was sent and a plumber who was accused by the businessman of sending the given email. By analysing other written texts by them, the forensic linguist was able to establish his claim that was in favour of the businessman as perpetrator.

Here the notion of ‘linguistic fingerprint’ could enter the picture: this is the set of markers in a text or speech that is characteristic of an individual (Olsson, 2008). Useful as it may sound, it can hardly be beneficial since it is unproven and the term is accepted rather in those circles of which the main interests are related to financial gain. It is important to note that although forensic linguistics is rather science than art as it is claimed by Olsson (2008), it cannot be stated confidentially that from two persons one is definitely the author; only the likelihood of that person being the author can be accounted for. Comparing texts in order to establish the identity of the author, there are several clues that can be used. 

Based on Olsson (2008), a list is provided here. First, if there is a text proven to be written by the suspect and the aim is to find out whether a certain text is written by him, several things can be done. It is important to pay attention to the fact whether the known and the unknown texts are compatible and comparable to each other. The next step is looking for differences and similarities. This is followed by the identification of linguistic levels to which these similarities and differences belong after which relationships are sought among them. It is important to test the frequency and importance of the given features in a corpus, but alternatively at a first attempt, using Google can be fine. Here, statistics come in handy, because the findings have to be quantified. Then as the last step, the determination of the likelihood of the suspect being the author takes place; for this, a scale of judgement is used (p. 43).

If a single text is received without any possible suspect, the question “What kind of person wrote the text?” can be asked as stated by Grant (2008). However, since forensic linguists are not interested in the personality of the perpetrator, the question needs further specification: the question “What kind of linguistic person wrote the text?” is more applicable. According to Grant (2008), one of the factors to which attention is paid is the educational background. There is a possibility of attempting to disguise one’s level of competence. This can be made by making deliberate spelling mistakes or using low status informal expressions. Another factor is the origins of cultural items; however, the influence of accommodation has to be taken into consideration here.

As it may be seen from this short essay, forensic linguistics is a diverse field in which there are several methods to turn to.

References

Coulthard, M., & Johnson, A. (2007). An introduction to forensic linguistics: Language in evidence. Abingdon, England: Taylor & Francis Group.

Gibbons , J., & Turell, M. T. (Eds.) (2008). Dimensions of forensic linguistics. Amsterdam, Netherlands: John Benjamins Publishing Company.

Olsson, J. (2008). Forensic linguistics: Second edition. New York, NY: Continuum International Publishing Group.