Theoretical and Methodological Problems of Defining the ‘idiolect’

Obligatory Assignment 1

ENG3423: Language, Literature and Crime

Due 25th February, 2013

Ragnhild Sollid

Theoretical and methodological problems

of defining the ‘idiolect’.

‘Idiolect’ is a term that has caused a lot of discussion in forensic linguistics. Linguists have tried to develop theories and methods to safely determine the authorships of texts. Such theories and methods would of course be of considerable significance in solving criminal cases involving textual evidence. These theories and methods have not, however, gained sufficient acknowledgement concerning reliability to be considered safe, and should therefore merely be used to give a hint at the probability of a texts’ authorship. I will in this assignment start by defining the word ‘idiolect’, and then briefly explain the theories and methods that are used in attempts to define the idiolect, and lastly discuss in what way these theories and methods are flawed and unreliable.

Idiolect is originally a term used to describe a person’s unique speech pattern, involving vocabulary, grammar and pronunciation. The term is now more widely used about language, including written language. It is also called the linguistic fingerprint, and some theorists claim that the authorship of a given text can be determined by thorough analysis because the linguistic fingerprint, or the idiolect, gives away the author. The notion is that each individual, consciously and unconsciously, uses language in a particular and unique way, and has certain features that mark their use of language that is identifiable. While the idiolect concerns the individual, the sociolect is the unique pattern of language used by a specific group of people. The reason for mentioning the sociolect is that it influences the idiolect, and vice versa. Whenever an individual picks up new words and makes changes to the idiolect, these changes quickly merge into the group and change the sociolect, and changes in the sociolect influence the idiolect.

Authorship analysis of the bible were among the first contributions to this field of study, and mainly average sentence length was used as a marker to identify the authors, but the results did not prove the method as reliable. Later statistical research have experimented with different parameters, such as average number of syllables per word, article frequency, typetoken ratio and use of punctuation, and though there can be traced patterns that indicates that authors have different styles, there is still no statistical theory or method that can produce absolute and reliable results in author identification (Olsson, pp. 19). In 1990 Morton and Michaelson claimed in their book The Qsum Plot that the Cusum (cumulative sum) method always gave reliable results, although they could not explain exactly how the method works (Olsson, pp. 21).

As computers became increasingly important in all fields of science throughout the 1980s, programs were also developed to analyse texts. Computational linguistics, as it is called, is still trying to find algorithms for identifying authorship, and some linguists firmly believe that it in this way is possible to get absolutely reliable results. These methods have shown some promising results, but only as long as the amount of accessible data is extremely large.

This is one of the biggest challenges when adopting these theories into practical use. There is just not enough data. In most criminal cases, forensic linguists have only a very limited amount of data available, rarely more than 750 words (Crankshaw, 2010). This makes the comparison of idiolect to that of the fingerprint (linguistic fingerprint) rather inappropriate, as we know we only need one fingerprint to identify an individual, and a fingerprint remains the same throughout a whole lifetime and does not change. This is not true of the idiolect, which will undoubtedly change continually through our lives. The idiolect is influenced by age, culture, education, geographical location, social class, setting, sex and so forth, and will develop in dynamic interaction with the individual’s surroundings and experiences. New words are constantly added to a person’s vocabulary, and other words are forgotten or devaluated (atrophy). This means that there will be what we call intra-author variation that may vary a great deal over a longer time interval due to the above-mentioned factors. This adds to the inconsistency of an individual’s idiolect. We also know that it is common to vary the use of language in order to conform with certain settings or genres, although the vocabulary can be manipulated only to a certain extent. Some theorists claim that parts of our language, in terms of grammar and style, are used unconsciously, and that these unconscious parts are not available for us to manipulate. Still it is possible to vary the language considerably within the constraints of the conscious parts of a person’s idiolect, and this might also open for different ways of deliberately disguising the idiolect by altering distinctive traits. Plagiarism or imitation of someone else’s idiolect are other ways of disguising an individual’s identity, which again adds to the notion of the linguistic fingerprint as unreliable and flawed. In addition to this, we know that every single person have their own interpretations of the lexicon, which means that people have different ways of using words in correspondence to what the specific words means to them. This may also change over time, and words may be used differently in different situations. All this considered there is no doubt that using the idiolect is no reliable way of determining authorship of texts.