By Marc Debono | 11 December 2020 | Opinion Articles
In this article, Marc Debono, Associate Professor in Language Sciences at the University of Tours, challenges us to consider the role of academic linguistic experts in court, arguing that academic expertise on matters of language, while it can be useful in court, must not be allowed to encroach on judges’ freedom to make their own decisions, which, he argues, is a democratic imperative.
Who can say (and how) if in a specific case submitted to a judge, there is defamation, moral harassment, public insult or glottophobic discriminations? The purpose of this reflection is to examine the issues involved in the intervention of sociolinguists as “experts” before the courts.
In French sociolinguistics today, the most frequent way the linguistic expression of conflicts is examined is through the analysis of speech acts considered ‘violent’ (verbal violence, harassment, insult, defamation, etc.). B. Fracchiolla, C. Moïse, C. Romain and N. Auger have contributed to the emergence of this particular field of research in France.
This group of researchers tend to focus their work on ‘verbal violence’ in the workplace, and sometimes – more specifically – in the context of the university, for example in the use of email.
In 2016, the question of linguistic discrimination – particularly in relation to ‘regional accents’ – received considerable media coverage in France, due to the publication of Philippe Blanchet's book [Discriminations: fight glottophobia]. The term ‘glottophobia’ is used to describe this type of illegitimate (if not illegal) linguistic discrimination, and is defined by Blanchet as ‘contempt, hatred, aggression, rejection, exclusion of persons, negative discrimination actually or allegedly based on the fact of considering some linguistic forms used by these persons as incorrect/inferior/bad’ (p. 45).
Of course, studying such discrimination is not new in sociolinguistics, but the advantage of the term ‘glottophobia’ is that it ‘reintegrates linguistic discrimination into all forms of discrimination against people rather than restricting it to discrimination based on language’ (p. 44). Thus, it can be argued that the fight against ‘glottophobia’ – as well as the fight against ‘verbal violence’ – are part of a more global project of French-speaking sociolinguistics: the fight against social inequalities and discrimination.
These two subjects are favorable to the transposition into criminal law of a "scientific" fight against inequalities. Indeed, verbal violence such as glottophobic discrimination is particularly suitable for interventions of sociolinguists in courts. P. Blanchet also writes his book because, from a legal point of view, in France, such [linguistic] discrimination is considered "acceptable, discriminatory but not condemnable by law" (p. 40 : des « différenciations acceptables, discriminatoires mais non pas discriminantes ») ; and the work on insult or verbal violence has been transposed into judicial cases, especially by academics such as D. Largogette (see 2010) who acts as an expert witness in various criminal cases concerning insults, discrimination at work, freedom of speech, etc. The aims of such interventions are certainly generous: to fight against inequalities, discrimination and domination.
Nevertheless, the purpose here is to ask the question: is this type of intervention in court such a good idea?
W. Labov, commenting on his work as a (socio-)linguist expert in the courts in a famous 1988 text (translated the following year into French), wrote, rather sententiously:
"A serious look at the world around us makes us understand that what matters is the reality of the facts. There are immense factual realities: the origin of the universe, the drift of continents, the evolution of the human species. There are more modest factual realities: the innocence or guilt of a given man" (Labov, 1989: 114; my italics).
The text from which this quotation is taken is an important one in sociolinguistics, laying the foundations for the "intervention" of sociolinguists in the social field. As French sociolinguist J. Boutet (in 2002), one of the founders of the "Language and Work" research group in the 1980s, recognized, Labov here expressed “very strong views on[the] question of social demand". We might even wonder if these views are "too" strong to constitute an ideal for sociolinguistic intervention?
Is "The innocence or guilt of a given man" a factual reality as Labov suggests? Before we continue, we have to examine the question of legal qualification: an intellectual (and experiential) operation which consists in apprehending a fact according to a category of law. This operation – carried out by the judge – aims to answer, for example, the following question: in the light of the facts and criminal law, can it be said that in the specific case submitted to the judge, there is defamation, moral harassment or public insult?
There is a theoretical difficulty here. The understanding of this act of interpretation (“qualification”) in legal theory inevitably raises important questions: relying on Gadamer's philosophical hermeneutics, the French academic A. Papaux considers “qualification” an act of micro-political power exercised by judges, who will choose between the various possible meanings of the act under their examination (cf. Papaux 1999). For example, the judge may qualify a mail or a verbal interaction as an insult, a fact constituting harassment, a defamation... or decide that the incriminated interaction does not constitute such offences. There will therefore always be "an irrepressible part of sovereignty and therefore of a-legal freedom in any jurisdictional act of decision", as Béchillon writes (2002: 48). This act of micro-political power, partly "free", is also a form of "bet": "the interpretation of facts is a bet or ‘hypothesis’: it is a provisional categorization given to the object in relation to a moral judgment and for a practical resolution" (Chappe, 2010: 544-545).
A concrete example of that is provided by the news: in 2018, to protect freedom of speech, an Austrian judge decided that for an extreme left-wing group to say "fuck" to an extreme right-wing leader is not a public insult deserving condemnation. The Vienna Court of Appeal argued that the "rejection of a political leader" can be "expressed in a provocative and shocking way" (‘En Autriche, dire “fuck” à un homme politique n’est pas condamnable’, Le Monde, 01/03/2018).
If we adhere to this hermeneutical theory of interpretation/"qualification", the "bet" it constitutes is therefore moral and teleological, which is extremely different from the determination of the "reality of the facts" in science: I refer here to Labov's quotation of course, which seems to put the two operations on the same level.
Conceptualisations of ‘truth’ differ between the legal and social sciences fields (see Supiot, in 2014), which makes it difficult for linguists to operate in legal settings.
However, the confusion of the two types of truth (scientific and legal) and interpretation, suggested by Labov, is – and despite all the ethical precautions – largely continued by forensic linguistics as it is currently developing in France, based on the model operating in the United States and Canada for over 40 years.
There is a real problem of encroachment by the expert on the judge’s office. My hypothesis is that this potential encroachment is often a problem of understanding, by the researchers of disciplines solicited by the judicial institution, of some fundamental democratic guarantees such as "intimate conviction" and "free will".
The judge's "intimate conviction" and "free will": expertise in the face of fundamental democratic guarantees
What is the judge's intimate conviction in the interpretation of facts, and why is it so important? And can scientific expertise undermine it?
These are the questions raised by F. Fezzani (in 2017), starting from a particular type of expertise (in neuroscience), which recently appeared in French courts. The danger is clearly presented: expert testimony tends to become more and more binding for the judge, thus compromising their intimate conviction and reducing their independence.
However, decision-making according to the judge's "intimate conviction" was established by the legislature during the French Revolution as a guarantee of the judge’s independence (and therefore of the proper functioning of democracy) and set out in the Code of Criminal Procedure (arts. 427 and 353). This principle leaves "complete freedom to the judge to decide" (Fezzani, 2017: 175), and this freedom humanizes justice by allowing the judge to interpret evidence, "so that evidence is not used as an automated truth-setting mechanism" (id.). The risk of expertise is an "overconfidence bias" that potentially affects all types of expertise: it can very easily become prescriptive and "blinding" rather than "enlightening", and thus endanger the democratic institution: the independence of judges being one of its pillars. The prescription of justice solutions by scientific authority is therefore explicitly condemned by the French highest criminal court (Cour de Cassation). This problem is neither fictional nor new: with judicial expertise, we often move from "resources to coercion", as Dumoulin writes (2000: 202).
Recently, the emergence in the US of "predictive justice" has raised the question in slightly different but equally worrying terms: based on data sciences, legaltech developed by private and successful start-ups, predictive justice aims to exploit big data to offer lawyers or judges statistically probable solutions in a given case - with particular reference to previous decisions taken in comparable cases (see also recent events in Estonia regarding the possible automatisation/robotisation of – a part of – justice).
Similar risks exist in some ways of thinking about legal expertise and forensic linguistics. Without in any way denying the necessary contributions of linguistics and social sciences to legal theory and to the judicial process, some reflection is required on the modalities of their integration.
My conclusion is short and simple: the independence of the judge in evaluating situations of conflict or linguistic discrimination must be preserved as a democratic imperative.
And whatever one may think of:
The particular judge or case (there are cases of famous miscarriages of justice where this free will may have played a role, of course);
The way that judges are trained (this is an important level for action)
The dominant ideology within this professional group (their supposed ‘conservatism', for example, may have been problematic in the past: here, again through training is possible)
Or whatever other grievance can be made against the institution or the judicial system,
…the protection of the judge’s free will is an essential guarantee of the democratic system and of a justice that must remain human and humane, with all its imperfections.
To come back to the initial question of whether it is such a good idea to fight inequality, discrimination, violence by introducing linguistics in a courtroom: before stepping into a courtroom, sociolinguists must be well aware of the importance of democratic principles/guarantees; and must not brandish the alleged ‘empirical factuality’ of their analysis of a given situation as a standard for the recognition of the social utility of their discipline.
Thus, this "analysis" can only be modest: an open, incomplete and imperfect interpretation of a researcher in the humanities field, who is also a human. Frequently recalled in our disciplinary fields, this necessity seems sometimes to contradict the researcher's idea of the interpretation of facts by a judge and evidence in law.
Finally, to guarantee this human imperfection of any – very poorly named – "expertise", ethics is not enough, and it is necessary to adopt a consistent epistemology with such an objective: an epistemology conscious of what are "facts according to a linguist"… what are "things of linguists” (I refer here to Martin Heidegger’s quotation about “lizard things”/”things of lizard” - see excerpt). The "language part" of conflicts, discrimination, and social inequalities can be a "fact according to a linguist"/a "thing of linguist", such as a "thing of judge": "things" that can be compared – of course and fortunately. But understanding them is based on experiences, insights and goals that ethics alone cannot make transparent, neither scientifically nor legally.
Excerpt: "The lizard is not just on the stone heated in the sun ... The rock on which the lizard extends is certainly not given to the lizard as a rock, rock he could question the mineralogical constitution. The sun to which he is warming is certainly not given to him as sun, sun about which he could ask questions of astrophysics and answer them. [...] The lizard has a relationship to the rock, to the sun and other things. We are tempted to say that what we meet there as rock and sun are for the lizard, precisely lizard things/things of lizard. When we say that the lizard is lying on the rock, we should scratch the word 'rock' to indicate that what the lizard is lying on is certainly given in one way or another, but is not recognized like rock" (Heidegger, 1929: 294-295; my translation).
Boutet, Josiane. 2001. ‘La part langagière du travail : bilan et évolution’, Langage et Société, 98: 17-42
Ismard, Paulin. 2015. La démocratie contre les experts (Paris: Seuil)
Lagorgette, Dominique. 2010a. ‘Linguistique légale et demande sociale: les linguistes au tribunal (Présentation) ’, Langage et Société, 132: 5-14
Lagorgette, Dominique. 2010b. ‘“Le crime est dans l'oeil de celui qui regarde le dessin”: l'analyse linguistique pour les tribunaux dans les procès Siné (2009)’, Langage et Société, 132: 77-99
Leclerc, Olivier. 2005. Le juge et l’expert. Contribution à l’étude des rapports entre le droit et la science (Paris: L.G.D.J.)
Moïse, Claudine, Emmanuel Meunier and Christina Romain. 2015. Analyses et solutions: La violence verbale dans l’espace de travail (Paris: Bréal)
Van de Kerchove, Michel. 2000. ‘La vérité judiciaire : quelle vérité, rien que la vérité, toute la vérité ?’, Déviance et société, 24.1: 95-101
International Association of Forensic Linguists. 2013. Code of Practice, https://www.iafl.org/wp-content/uploads/2018/07/IAFL_Code_of_Practice_1-1.pdf