The TACIT team studies authentic monolingual and bilingual witness interview transcripts and recordings in different languages (English, Spanish, Portuguese, Latvian, Lithuanian and Russian) from the UK and the United States. The main goal of these investigations is to detect when, where and why difficulties in communication and translation may occur, how to prevent them and how to resolve them when they happen.
TACIT project is interdisciplinary and addresses a broad spectrum of language and communication issues. It provides useful, real-life-based exemplification of communications problems and offers practical solutions, enabling the police to obtain relevant information and evidence more accurately and efficiently. The insights from this project pave the way towards removing communication barriers in investigative interviewing, making the process less costly and more reliable.
There are currently three interconnected research strands:
1. Interaction in Discourse Strand: Complexities in police interviews: Complex questions and statements, meta-communication and meta-representation
2. Rapport & Emotion Strand: Effects of empathy, politeness, mitigation and aggravation strategies on interview outcomes
3. Translation Strand: Problems in translation stemming from presence vs. absence of word meanings and construction meanings, or different cultural preferences, in two languages.
What we study – An Example
Previous research on multilingual police interview transcripts carried out in the United States (Filipović 2007, 2013) has demonstrated that serious misunderstandings can occur in police interviews, sometimes even leading to a misinterpretation of denial as a confession. It is known that a number of linguistic subtleties, potentially of great importance for the law, are missed or lost in translation or misunderstood even if communication takes place in a single language. It happens because languages vary in the means they have for expressing and translating certain meanings. For example, some languages do not require speakers to specify every event as either intentional or unintentional. The statement She pushed the bottle off the table in English is ambiguous with regard to intentionality – we cannot conclude, based on this sentence, whether the person did it on purpose or not. Some other languages, like Italian, Serbian or Spanish offer different constructions that make it clear which meaning, intentional or non-intentional, is intended.
Speakers can also show different levels of certainty or uncertainty in their statements, which is expressed different meanings of modal verbs such as can, could, may, might, should, etc. in English. Accordingly, speakers are judged to be more or less certain and more or less believable based on the modal verbs they use to modify their statements (Filipović, 2016). There is very rarely a perfect overlap in meaning and use of these modal meanings and modality expressions in other languages, which makes translation difficult, and which can result in different conclusions about what speakers said in the original vs. the translation.
These are just some illustrative examples of the kind of detailed work we do, which are integrated into our training programmes and delivered in the UK and abroad, at universities and to police teams.
Our Research Methodology
This project used a combination of quantitative and qualitative approaches to analyses of police interview transcripts. The database for the project consists currently of over 100 authentic UK and 500 US police interview transcripts of various length and its corresponding audio recordings. The analysis is interdisciplinary (involving cross-cutting themes within linguistics, psychology and sociology) and it is informed by numerous and different theoretical insights (from CDA, Appraisal Theory, Cognitive Linguistics and Forensic Linguistics).
Two features of the project make it quite unique:
i. the same datasets are studied from different theoretical perspectives and
ii. this research includes a substantial quantity of material for detailed analysis (rather than a single or a small number of cases studies as is more usual in the field).