How police investigators seek to secure that suspects speaking a second language understand their rights in investigative interviews

ABSTRACT

This study presents an empirical investigation of 54 police interviews in Norway conducted with suspects speaking Norwegian as a second language. Using Conversation Analysis as our method, we examine how investigators seek to ensure that suspects understand their rights. We find a range of practices for making rights more accessible compared to how they are formulated in legal acts. Investigators divide their turns into smaller segments and monitor the suspect for displays of understanding along the way. They simplify the syntax and substitute technical terms with plain vocabulary. Sometimes they supplement the wording of the prosecution instructions with explanations or reformulations. In a few instances, they also check understanding by asking a simple yes/no question. However, such questions only generate claims of understanding and no actual evidence of what has been understood. While there are isolated cases of requests to retell the rights, even these practices are problematic in generating strong evidence of understanding. This study discusses the importance of linguistic education and training in police forces for dealing with potential or emerging understanding problems during investigative interviews.

KEYWORDS:

• Investigative interviews
• communication of rights
• simplification
• understanding check
• conversation analysis

1 Introduction

At the beginning of an investigative interview in Norway, police officers are required to inform suspects about the allegation and their rights and obligations in the interview, especially the right to silence and to get assistance from a legal counsel. To do so, Norwegian police do not use a scripted formula such as the Miranda warning in the U.S.A. or the caution in England and Wales but leave it to the individual investigator to find the most appropriate and efficient formulation. As a result, the wording of cautions may vary significantly from interview to interview. Yet, in practice, Norwegian investigators tend to rely on official wording from the law which they normally read out as they proceed through the introductory phase (Urbanik and Pavlenko, 2021). The legal basis for the obligation to inform suspects about their rights is The Prosecution Instructions (undefined), § 8–1, where it is stated that:

Before an investigative interview with a suspect is carried out, he shall be informed about the nature of the case and a possible charge. He shall be informed that he has no obligation to give a statement.

The suspect shall in addition be informed that he has the right to be assisted by a defence counsel of his choice at each stage of the investigation, including the police investigative interview of him. The charged should be asked whom he wishes to be appointed as his public defender when he is entitled to such a defender.

(Our translation)

Regardless of whether they stick to the official wordings or provide individual adaptation, the investigators bear the responsibility to give ‘correct, adequate and sufficient’ information and to ‘ensure that the person who is to be interviewed has understood their rights’ (Riksadvokaten [Attorney General], 2016, p. 4). Textbooks on investigative interviewing instruct investigators to (1) present the rights in a ‘simple and clear way’ and (2) ascertain that they have been understood correctly (Bjerknes & Fahsing, 2018, p. 219). There is, however, no further guidance on how to secure and ascertain understanding, especially when the suspect is a second language (or partial) speaker of Norwegian (L2), which refers to someone who indicates in any way that Norwegian is not their first language (L1). Thus, in this study, we investigate the following question:

How do police investigators seek to secure reasonable understanding of the right against self-incrimination and the right to defence counsel by partial speakers of Norwegian?

We divide this question up into two interrelated questions:

1. How do they preempt understanding problems in formulating the rights?

2. How do they check the suspects’ understanding of their rights?

The aim of our study is to provide an overview of the variety of practices investigators employ to make their cautions more understandable. We particularly focus on investigators’ treatment of less frequent and potentially difficult legal terms and phrases and on questions asked to check whether the suspect has understood his rights. By investigating how police investigators deal with official wording, we show how they orient to two partially opposing concerns of legal correctness, on the one hand, and comprehensibility, on the other hand. Through our study, we not only wish to draw practitioners’ attention to different practices used in seeking to secure the suspects’ understanding of their rights but also wish to point out some of the practices’ interactional limitations. We argue that working with transcripts of police interviews during training courses may help practitioners to become aware of these limitations and consciously select from a range of practices.

2 Previous research

2.1 Communication of rights to L2 suspects

Most research on communication of rights in police interviews has so far had three main points of focus. Early studies were occupied with the linguistic complexity of cautions, paying particular attention to the accumulation of grammatical and lexical forms that made the content less intelligible (Gibbons, 2003; Shuy, 1998). Thus, such features as syntactic complexity, combinations of advanced concepts, passive voice, hidden negation, and low-frequency words were considered hindering the comprehension of cautions. One of the recommendations from these studies has been to simplify the language. Similar conclusions have been made in research testing the comprehension of cautions. It has been shown that clarity decreased with syntactic complexity and sentence length, as well as uncommon words, higher level of abstraction and legal terms (Chaulk et al., 2014; Rogers et al., 2007, 2011). However, this research has problematized the issue of structural simplification by pointing out that although short and simple phrases are easier to memorise, they do not guarantee correct understanding of crucial concepts, such as ‘right’ (Rogers et al., 2007, 2011). This is because such concepts are underspecified and their meanings may have wider or narrower scopes than assumed by the suspect or they may carry different lay connotations. Consequently, suspects may presume that the right to silence is a risky choice that can be used by the police as incriminating evidence or that the legal rights presented do not have continuing power, i.e., they cannot be used later in the interview (Rogers et al., 2010). These findings demonstrate that understanding of rights goes far beyond wording and encompasses details that may easily be blurred in too long formulations or omitted in too short deliveries (Rogers et al., 2007, also; Rock, 2007).

These issues have been more prominent in research that focuses on police discursive practices. Several studies have uncovered challenges in how cautions are delivered to suspects. Researchers have found relatively great variation in terms of language choice, ways of structuring the caution, the intelligibility of paraphrase, attempts to secure understanding, opportunities for suspects to claim their rights, and police officers’ attitudes to formalities (Berk-Seligson, 2016; Cotterill, 2000; Diepeveen et al., 2022; Heydon, 2005; Leo & White, 1999; Nakane, 2007; Rock, 2007; Russell, 2000). In her comprehensive study, Rock (2007) has examined most of these issues systematically, pointing out that suspects tend to avoid displaying their lack of understanding (cf. Russell, 2000; Snook et al., 2010), and police officers, lacking linguistic training, do not know how to assess and check understanding (cf. Cotterill, 2000; Nakane, 2007). Moreover, suspects are rarely interested in explanation, while the police do not receive guidance on how to reformulate caution and what to do when reformulation fails. Rock (2007) also names typical practice investigators employ when cautioning detainees, such as repeating the official formulation, using metadiscourses, downplaying communication of rights, checking understanding, changing the structure of the caution, or breaking it down into smaller parts (see also Heydon, 2005; Leo & White, 1999; Nakane, 2007; Russell, 2000). Her study clearly demonstrates that understanding of rights is not a straightforward, either/or question, and that the communication of rights is difficult to standardize. Being an interactional activity, cautioning is sensitive to contextual features and dynamic assessments made by investigators in each individual case.

The issue of rights communication to second language speakers has received increased attention in recent years (Berk-Seligson, 2016; Bowen, 2019, 2021; Dumas, 2020; Innes & Erlam, 2018; Pavlenko, 2021; Pavlenko et al., 2019). Studies from the U.S.A. and New Zealand have shown that L2 speakers score significantly lower on comprehension tests than L1 speakers (Innes & Erlam, 2018; Pavlenko et al., 2019). These differences may even be dramatically stark when it comes to understanding of spoken warnings (Pavlenko et al., 2019). Moreover, for L2 speakers even simple and short sentences may be problematic if they contain legal terms (Pavlenko et al., 2019). In contact with the language of cautions, L2 suspects may rely on their own first-language categorizations, phonological similarity of words, or contextual information to make inferences and reconstruct the sense of delivery (Bowen, 2019; Gibbons, 1987; Pavlenko et al., 2019; Roy, 1990). Several researchers also mention other factors that play an important role in L2 suspects’ decision-making, such as culture-specific politeness norms, which may make suspects more likely to give positive responses to questions and understanding checks (Roy, 1990), known as ‘gratuitous concurrence’ (Lieberman, 1981; Eades, 2010, cf.; Bowen, 2019). Finally, understanding may be impeded by the investigators’ omissions, poor lay translation into the L2, trivialisation of the procedure, or inaccurate paraphrasing (Berk-Seligson, 2016; Bowen, 2021; Dumas, 2020; Urbanik and Pavlenko, 2021).

A step forward in addressing the issue of L2 speakers and their understanding of rights in police interviews is the Guidelines for communicating rights to non-native speakers of English in Australia, England and Wales, and the U.S.A. (CoRG, 2015) that provide recommendations on how to avoid standard mistakes that obscure the core message of cautions and how to secure understanding of this message. In Norway, where there are no similar guidelines and standardized formulations in plain Norwegian, investigators rely on their own experience and intuition when they seek to secure understanding. In practice, they have to respond to two different and often competing requirements, on the one hand, of keeping the procedure ‘legally watertight’ and, on the other hand, of making it understandable to suspects (Gibbons, 2001). While previous research has paid more attention to what is or may be the reason for intelligibility challenges among L2 speakers, the current study focuses on what police officers do to reduce this risk and preemptively avert understanding problems.

2.2 Preemptive practices in conversations with L2 speakers

Ever since Ferguson (1975) coined the term foreigner talk, linguists have shown an interest in how first language users speak to second language users with a limited proficiency. Especially in the field of second language acquisition, this has been a topic of interest in that it has been thought to influence the learners’ acquisition of the target language (see overview in Fischer, 2016). But also in the field of interactional sociolinguistics, researchers have investigated how L1 speakers adapt their speech and use interactional procedures to achieve mutual understanding. In an overview of such strategies, Bremer et al. (2013) observe that L1 speakers may raise the expectability of their contributions by making metadiscursive comments and marking topic shifts clearly. Furthermore, they may raise the accessibility by speaking in short utterances, by using high-frequency vocabulary, by segmenting information into smaller parts, and by articulating words clearly. Finally, they may raise explicitness by using full forms of words and clauses instead of elliptical constructions, pronominalisation, and contractions.

In the field of Conversation Analysis, the concept recipient design is used to describe the practices by which speakers adapt their talk to particular addressees, especially their background knowledge (Sacks et al., 1974). In line with this, some studies have sought to describe how speakers address potential understanding problems by engaging in preemptive self-repair, that is, modifications to their own utterance in progress in order to address and avoid potential sources of trouble. One field of research where this has been particularly important, is the establishment of reference to persons and places. Two main practices used for preempting reference problems are described in the literature: one is presenting information that will help identify the referent, the other is checking whether the interlocutor knows the referent. Speakers may present information by either reformulating the referring expression, for instance, from a name to a description (‘Joey – my brother’) or by adding clues to identifying the referent (‘Joey – the guy you met yesterday’) (Heritage, 2007).

In the field of second language interaction, conversation analytic research has also addressed how speakers preempt understanding problems related to the recipient’s potential lack of familiarity with a lexical item. This is especially relevant concerning specialized or technical vocabulary in institutional interaction (Greer & Leyland, 2018) and in cases where L1 speakers talk to L2 speakers with a limited proficiency. For instance, Kim (2019) has studied knowledge check questions (‘do you know what X is?’), which may be used prospectively in introducing a potentially unfamiliar lexical item, and thereby preempt recognition problems. Svennevig (2010) describes how L1 speakers may modify and expand an utterance in progress to insert explanations or check the interlocutor’s knowledge of a word. Of special relevance here, the study notes that potentially unfamiliar lexical items may be marked as such by inserting metalinguistic frames such as ‘something called an X’, thereby projecting a subsequent explanation of the term introduced.

At the level of more extended turns at talk, Svennevig (2018) has described a preemptive practice for making complex information more accessible to L2 speakers. This involves dividing an extended turn up into smaller constituent parts and presenting them one at a time in instalments with pauses in between. The speakers thereby invite their interlocutors to display their understanding along the way by producing minimal responses or by nodding. Alternatively, the interlocutors may exploit the opportunity to display a lack of understanding, thereby identifying the source of trouble at an early stage (Robinson, 2014). In this way, this turn design is tailored for making complex information more manageable to process by L2 speakers and for establishing joint understanding step by step.

Going beyond the mere description of practices, Svennevig et al. (2017) sought to measure the efficiency of such strategies. To do this they set up an experiment where L1 speakers were to give instructions to L2 interlocutors. The analysis showed that when the L1 speakers used a greater number of preemptive strategies, there were fewer manifestations of understanding problems by the L2 interlocutors. The study thus concluded that the use of pre-emptive strategies is effective in reducing the likelihood of understanding problems.

3 Data and method

The data for the current study consist of 54 audio and video recordings of police interviews with L2 suspects conducted without an interpreter in Oslo Police District between 2016 and 2020.¹ Most of the interviews are conducted at police stations, while some are conducted on-site. The suspects (S) are juveniles and adults (mostly males) with various first language background interviewed about different types of offences, from minor ones (e.g., underage sale of alcohol) to major ones (e.g., rape or child sexual abuse). The police investigators (PI) are both males (n = 27) and females (n = 25) with Norwegian as their first language and varying extent of experience as interviewers.² Compared to previous research in this field, our data consist of a sizable corpus of authentic investigative interviews. The representativity of the corpus is strengthened by the fact that the sample was randomly selected (apart from the language background criterion). However, a limitation is that it is collected at a single police district.

In this analysis, we focus on the introductory part of the interviews, where police officers collect personal information from suspects and inform them about their legal status and their rights and obligations before proceeding over to the next phase where suspects can give a free statement. When informing about the rights, investigators usually support the delivery with official formulations to which they have access on their computer screens displaying an electronic interview protocol (cf. Bjerknes & Williksen, 2015, p. 298, see also Section 1). As a result, they often read out the wordings (or their parts) as these are formulated in the law.

The introductory phase has been transcribed according to conversation-analytic conventions (Jefferson, 2004, see Appendix). The analysis has consisted in a qualitative and inductive process of identifying preemptive strategies and practices for checking understanding. The goal of the analysis was to identify the various linguistic and conversational practices used as attempts to secure understanding. The analysis section below presents the main types of practices that were identified and exemplifies them with extracts that illustrate their linguistic and interactional characteristics. The study focuses on the officers’ strategies for securing understanding and the feedback they get about the suspects’ understanding in the interview itself. We do not have any independent sources of knowledge about the suspects’ actual understanding of their rights.

This is a qualitative study of the array of practices used in the data. The aim is not to account for the distribution of these practices, and consequently we have not undertaken a quantitative analysis of their use. However, we occasionally present some rough estimates of the relative frequency of the different practices. These characterizations are based on a qualitative assessment made by the three researchers involved. A more rigorous quantitative analysis of their distribution will have to await a future study.

4 Analysis

In this section, we provide an overview of the common practices Norwegian investigators employ in seeking to secure understanding of the right to silence and the right to defence counsel. We first focus on a variety of preemptive practices that display investigators’ orientations to suspects’ potential troubles with specific words, phrases, and sentences. Then, we look more closely at how investigators seek to secure understanding by checking it post hoc.

4.1 No use of preemptive strategies

Before analysing preemptive practices, it is useful to note that not all investigators make use of them. Some merely read the rights more or less verbatim as they are stated in the interview protocol without modifying the wording or seeking feedback from the suspect along the way (cf. Rock, 2007). This is exemplified by the following extract (see Appendix for transcription conventions):

The rights are here presented in a complex and dense style characteristic of bureaucratic and legal documents. It includes technical, low-frequency terms such as forsvarer (legal counsel, line 1) and straffbar handling (criminal (literally ‘punishable’) act, line 10) abstract verbs and nominalisations, such as pådra siktelse (‘inflict charge upon’, lines 9–10). The syntax is complex due to multiple coordination, passive voice (det vil bli tatt lydopptak - ‘there will be made an audio recording’), and reflexive constructions, as in la deg bistå av forsvarer (‘let yourself be assisted by a legal counsel’, line 1). Finally, the information density is high due to an accumulation of phrases within each sentence, for instance: la deg bistå [1] av forsvarer [2] etter eget valg [3] på ethvert trinn av saken, også [4] under avhør (lines 1–3). Such features contribute to making the formulations inaccessible to second language speakers even those at a high level of proficiency (Gibbons, 1987, 2001; Pavlenko et al., 2019). The police officer does not do anything to simplify the language or to check the comprehensibility of his utterances along the way.

With this example as a background, we now turn to cases where the officers engage in interactional work to secure understanding. One way to do this is to decompose the utterance into smaller units and allow the suspect to respond along the way.

4.2 Decomposition

The cautions are cited more or less verbatim from the prosecution instructions. Some officers may yet be considered orienting to potential understanding problems by dividing the cautions up into instalments and presenting them one at a time, leaving room for the suspect to confirm their understanding along the way (or, potentially, to initiate repair) (Svennevig, 2018). This can be observed in the following extract.

After presenting the first right (line 1), the officer leaves a pause, allowing the suspect to produce a response (line 3), thereby claiming understanding of the turn so far. The next right is presented in two instalments, in lines 5 and 8. Each instalment is produced with slightly rising intonation, projecting continuation of the turn and inviting merely a minimal response from the recipient. Also here, the officer pauses and the suspect produces a minimal response in the gap between the two turns. In this way, the officer makes it possible for the suspect to confirm his understanding of the rights along the way and potentially signal a problem of understanding if needed. As noted above, previous research shows that this practice may slightly improve understanding of each component presented (Fenner et al., 2002) and contributes to less manifestations of understanding problems (Svennevig et al., 2017).

4.3 Simplification

Another way of orienting to potential vocabulary problems is to substitute technical terms with more common and informal words. The wording in the legal acts of the right to legal counsel includes both a formal bureaucratic term, bistå (‘aid, assist’), and a low-frequent legal term, forsvarer (‘legal counsel’). Some officers simplify the wording by substituting or omitting these terms (cf. Bowen, 2021). Here is an example:

In this case, the officer says ha en advokat (‘have a lawyer’) rather than la deg bistå av en forsvarer (literally ‘let yourself be assisted by a defence counsel’). He thus uses common and everyday words that are more likely to be understood by an L2 speaker than the words in the prosecution instructions. In addition, he changes the syntax by using a simplified verbal phrase (‘have’ rather than ‘be assisted by’ a lawyer). Finally, we may also note that he subsequently specifies that this right also applies ‘when you talk to the police’ (line 7), which echoes the specification in the prosecution instructions (‘including the police investigative interview with him’) but in simpler words.

The other right may also be difficult to understand for an L2 speaker, especially the phrase ingen plikt (‘no obligation’). In Extract 4, we see that the investigator does not use this formulation but instead chooses simpler words, trenger du ikke å (‘you don’t need to’) prefaced by the conditional ‘if you do not want to’, which emphasizes the volitional character of the right. Importantly, the complex syntax is additionally unweighted by the practice of decomposition mentioned above, which provides an opportunity for the suspect to mark the receipt of the first piece of the warning before the second one is delivered (line 3).

4.4 Term explanation

In some cases, legal terms and other potentially difficult words and expressions are explained. We find an example of this in the following extract.

Here, the original formulation, which is a verbatim quote from the protocol referring to the Criminal Procedure Act (Straffeprosessloven [Criminal Procedure Act], 1981), in lines 1–4 has several technical terms (siktelse – ‘charge’, straffbar handling – ‘criminal act’). The officer orients to this wording as potentially problematic for the interlocutor by additionally explaining the terms. He rephrases it in a markedly different style, using everyday vocabulary and simple, active syntax. Especially telling is the use of the word gæernt (‘wrong’) which is a colloquial form of the more standard word form galt.

4.5 Reformulation

A final strategy is to start by presenting the right using the formulation in the prosecution instructions but then to reformulate the right in simpler words. This may involve single words that are considered potentially problematic, or more extended constructions. The first example is a reformulation of the technical term ‘forsvarer’ (defence counsel).

In the midst of a verbatim quote from the prosecution instructions, the investigator here inserts a reformulation in which he substitutes the technical term for the more everyday term advokat (‘lawyer’). This constitutes self-repair and is introduced by a pragmatic particle (altså - roughly translatable as ‘that is’) commonly used in self-repair generally and reformulation more specifically (Heinemann & Steensig, 2018, Svennevig, in press). This practice has common traits with the practice called simplification (above), but as we shall see, it may have important implications that the investigator actually produces the original formulation instead of merely simplifying it from the start.

Other reformulations are more concerned with making the general statements relevant to the concrete here-and-now situation, thus relating to problems of understanding the procedural implications of what is said. An example of this type of contextualising reformulation can be found in the next excerpt.

In addition to formulating the right in simpler words, the reformulation also spells out the practical implications of the right, namely that it is the suspect’s choice to exercise it by speaking up and asking to talk to a lawyer. This practical implication for the suspect is especially salient in the use of so-called free indirect speech, that is, presenting the hypothetical request by the suspect in the words he would then use, namely the first-person pronoun (jeg – ‘I’).

Common to reformulations and term explanations is that the investigators start by presenting the rights using the formulations (more or less verbatim) from the legal acts. Only subsequently do they simplify or explain, thereby making the formulations more accessible and understandable to the suspects. In this way, they seem to orient to a concern for legal correctness, that is, presenting the rights in the judicially prescribed formulations so that they cannot be accused of having misrepresented them. This reveals what Gibbons (2003) calls a ‘two audiences dilemma’ that some investigators try to deal with. By using wordings that are ‘legally watertight’, they meet the statutory requirement and thus secure the interview as evidence for the courtroom audience. This, however, may weaken the intelligibility of the caution, undermining its basic function and – paradoxically – poses a risk of inadmissibility in the court. Therefore, in the long-run reformulations and explanations also serve to prevent accusations of miscommunication with the lay audience, which is particularly crucial in interviews with vulnerable individuals, such as L2 suspects.

4.6 Understanding checks

As mentioned in Section 1, Norwegian investigators are required to ensure that the suspect has understood the rights. One way of doing this is to ask an explicit understanding check question. Here is an example of how this may look.

The investigator’s explanation of rights is potentially complete in line 3. There is no response from the suspect, and the investigator adds an increment restating (and reformulating) the right to silence (line 5). After this, there is still no response from the suspect in the ensuing pause (line 6) or during the investigator’s extended hesitation sounds in line 7. At this point, the investigator asks an explicit question about understanding, thereby explicitly pursuing a response. The suspect now responds by producing a three-fold positive response token, emphatically claiming understanding. As the response is delayed by a 0.5-s pause, the investigator adds yet another increment in overlap with the response, recompleting the question and specifying its import (line 10). This is again confirmed by the suspect. We may thus see that the inquiry about understanding comes at a point where the suspect has not produced minimal responses or other claims of understanding. The response it generates is a claim of understanding, which, however emphatic, does not display what this understanding amounts to.

Yes/no-questions about understanding, such as ‘do you understand?’, involve a response bias in favouring confirming responses (Tranekjær, 2018). Indicating that one does not understand is a more sensitive thing to do, as it displays a lack of knowledge on the part of the speaker and imposes an additional burden on the interlocutor. Here is an instance where the suspect does not answer affirmatively the question of whether he has understood his rights.

The police officer’s question in line 1 is an explicit question about understanding. It is formulated as a yes/no-question with a preference for a confirming response. The suspect answers only after a rather lengthy pause of 1.0 s, which is a marker that the upcoming response may not be the projected (or ‘preferred’) one (Sacks, 1987). The following answer is indeed not a confirmation of understanding but instead a somewhat oblique answer in which the suspect comments on his limited Norwegian skills, which seems to imply a negative answer to the question. This leads the investigator to inquire further into whether the suspect finds it difficult to speak in Norwegian, possibly orienting to a potential need for an interpreter, which is also one of the questions that investigators are required to assess in the initial stage of an investigative interview (Attorney, 2016). The suspect confirms with a positive response (ja) and elaborates his answer by observing that he has a problem with ‘complicated’ words. This seems to imply problems with the type of language used in the formulation of rights and thus a further indication that the information about rights may not have been understood. One would expect that this would be a clear signal to the investigator that the rights need to be explained again, or possibly that an interpreter should be called in, following the legal provisions to do so when someone is not able to communicate adequately with a public body (Tolkeloven [Interpreting Act], 2021; undefined). Instead, he proposes that they should proceed with the interview and rely on the suspect asking questions if he does not understand. This extract follows immediately after the previous one.

We can note that the suspect eagerly accepts the suggestion, as can be observed by the early onset and multiply repeated acknowledgement tokens (line 24). He furthermore completes the officer’s unfinished utterance, a so-called collaborative completion, in line 27, displaying understanding and acceptance of the investigator’s suggestion. This emphatic acceptance displays an orientation to not creating problems for the interlocutor, thereby presenting oneself as cooperative. And when the investigator repeats the question whether the suspect has understood (line 31), the latter produces an affirmative answer, contrasting rather significantly with his previous answer. This answer is a claim of understanding but does not provide any evidence of what was understood. The officer accepts this confirmation at face value and proceeds with the interview.

What this example illustrates is that questions about understanding may be problematic since positive answers may well be a face-saving way to avoid bringing understanding problems to the surface (cf. Eades, 2010; Lieberman, 1981; Snook et al., 2010). And even when suspects admit that they have problems, police officers may avoid investigating the nature or the extent of these problems by resorting to solutions which put the burden of ensuring understanding on the suspect. This is contrary to the interview instructions and regulations (Attorney, 2016; Bjerknes & Fahsing, 2018), which state that it is the obligation of the investigator to secure understanding.

In our data, we also find isolated cases where investigators do not merely accept a claim of understanding but actively pursue a more elaborate demonstration of understanding, as in the next excerpt.

The claim of understanding in line 3 is not merely accepted as evidence of understanding. Instead, the investigator makes a new request to the suspect, this time to tell what he has understood, thereby eliciting a more encompassing demonstration of understanding. And as we can see, it succeeds in getting the suspect to formulate one of the rights in his own words (lines 14–15), thereby providing positive evidence of understanding. Yet, we may also note that the suspect’s answer only provides evidence of understanding of one of the two rights that have been presented. This leads the investigator to repeat the other right (to silence) once more in the following (data not shown). Thus, even in this rare case of elaborate work to check understanding, only a partial confirmation is achieved.

Another observation that arises from this excerpt is that asking such a question may constitute a sensitive action. Rather than answering the question immediately, the suspect hesitates and asks for clarification (line 8), thereby displaying trouble in responding. The investigator adds an account for asking, namely that he wants to check that he has understood (line 12). Accounts are generally added when a social action is not fully ‘self-explanatory’ or expected (Scott & Lyman, 1968). The account here preempts certain possible interpretations of the question, such as it being an expression of doubt about the sincerity of the suspect’s claim or a test of his language skills. In this sense, the account functions as a disclaimer, seeking to avert potentially face-threatening interpretations of the request.

What this example alerts us to, is two things. First, that asking suspects to ‘tell back’ the rights in their own words may be a very useful practice in that it engenders very substantive feedback on the suspects’ actual understanding. In fact, this practice of asking for rephrasing is recommended in the guidelines for communicating rights to non-native speakers (CoRG, 2015). The current study thus provides empirical evidence that this practice may have a positive effect on securing understanding. However, the other thing that this example shows is that such questions may be sensitive and face-threatening and require explicit accounting in order to be understood as a mere understanding check.

5 Discussion

In the Norwegian legal system, investigators are required to inform suspects about their legal rights and to ensure that they have understood them. However, these requirements are formulated in general and vague terms and do not include guidance on how to do this (cf. Bowen, 2021). As the rights are not scripted, investigators have the opportunity to adapt the formulation of them to the needs of the individual suspect. Until recently, police officers have not been trained in how they can adapt official formulations to L2 speakers or other individuals who may not understand the legalese. Thus, they need to rely on their own intuition, experience, and contextual cues while working on minimising the risk of miscommunication. In this study, we have provided an overview of practices they use to preempt misunderstanding and to secure understanding.

Addressing the study’s first research question about the use of preemptive practices, we found three types to be most common in the data. First, the investigators presented the rights in instalments, leaving room for the suspect to signal his understanding along the way, or alternatively, to initiate repair at an early stage. This practice reveals investigators’ orientation to the structural complexity of the caution as potentially problematic for suspects. Second, they oriented to legal terms as a potential challenge by avoiding them and choosing simpler and more common ones instead, or by explaining their meaning in plain language. Third, they stated the caution’s official wording and then reformulated it, thereby manifesting their orientation to its whole content as a potential source of misunderstanding.

The employment of one or several of these practices shows what in the content of the warnings investigators recurrently consider as potentially problematic and requiring modification. Thus, although not having been trained in how to secure understanding, they do identify similar aspects of the caution content as challenging for suspects and needing preemptive action to avoid misunderstanding. Based on previous research on what causes understanding problems for L2 speakers, we can ascertain that some of these are indeed oriented to as problematic by the officers, especially the use of legal terms and specialized vocabulary. Furthermore, the variation between individual officers reveals different orientations to the two contrasting concerns of legal correctness, on the one hand, and comprehensibility of the formulation of rights, on the other hand. Some officers seem to orient primarily to the former by not changing the wording of the cautions to any extent, while others orient more to the latter by substituting legal terms with more colloquial ones (the practice of simplification). Finally, many investigators balance the two concerns by providing both the official wording and a simplified version. This is manifested in the practices of term explanation and reformulation.

The second aim of the study has been to show how investigators check suspects’ understanding of their rights. We have found that by presenting the rights in instalments they invite displays of understanding along the way. However, such displays may be misleading, as they merely represent claims of understanding and do not provide evidence of what it is the suspects have understood. Another practice common in the data is asking for confirmation of understanding (‘do you understand?’), but such requests involve a similar concern. Simple confirmations do not display what that understanding amounts to. Requests for confirmation of understanding are problematic also because they involve a preference for confirmation. Our study shows that investigators do little to pursue more substantial displays of understanding. On the contrary, they sometimes even treat a lack of confirmation as not requiring further action to assure that the suspect has understood. Only in a couple of cases in the data do we find that the suspect is asked to retell the rights in his own words (cf. Pavlenko, 2021). This way of checking understanding is recommended in the guidelines for communicating rights to non-native speakers (CoRG, 2015). Our study supports this recommendation by showing that this does in fact generate more substantial displays of understanding. Yet, our example also shows that it does not necessarily engender extensive accounts of all the communicated rights even when the suspect claims to have understood. Furthermore, our analysis indicates that the request to retell the rights in one’s own words may be challenging on the interpersonal level, as it may be heard as a ‘test question’ casting doubt about the suspect ’s claim of understanding.

Following from the observations above, we can see that investigative interviewing practices with L2 suspects often display a need to find a balance between interactional and institutional interests. The consequences of this balancing may seem somewhat contradictory. On the one hand, going beyond the legal and procedural correctness displays orientation to suspects’ understanding as an important aspect of rapport building and in the long run secures the interview’s admissibility in court. On the other hand, it often results in both qualitative and quantitative variations (cf. Bowen, 2021; Rock, 2007), which raises the question of suspects’ equal opportunities to understand their rights. Thus, the strength of preemptive practices in that they respond to local circumstances may as well be seen as their weakness because they run the risk of non-uniformity. Yet, employing decomposition, reformulation, simplification, and term explanation is beneficial to both the investigator and the suspect to a greater degree than merely reciting the prosecution instructions or the interview protocol. While the current study did not aim to provide evidence of the suspects’ understanding of their rights, previous research has shown that such strategies are in fact beneficial and are associated with an increased chance of successful understanding in police interview subjects (e.g., Snook et al., 2010) and in L2 speakers (e.g., Svennevig et al., 2017).

In recent years, the Norwegian courts have rejected interviews as evidence on the basis of the defence’s claims that suspects (both L2 and L1) had not been informed about their rights in a sufficiently understandable way, and/or that there was no proof that they had understood the rights.³ This shows that incorrect and unintelligible communication of rights to suspects in general and second language speakers in particular may involve considerable risks to the legal procedure. One way to address the great variation we found in our data might be to develop more scripts and standardized formulations that are more understandable than those that are written in the law and interview protocols (following also guidelines formulated in CoRG, 2015). Seeing that many investigators in our data cite from those legal sources, it may be useful to offer an alternative text in plain language to function as a starting point for communicating rights. Prescribing standardized formulations does not, however, seem to be a solution in itself, as this may lead to careless automatization of paraphrasing, resembling ‘tick-box consent talk’ (Rock 2016). It could lead investigators to rely on the plain language text as necessarily understandable and not requiring further adaptation to the individual suspect in a specific context. In our opinion, a way of dealing with L2 suspects is not necessarily to add more scripts to the procedural policies, but to train police officers in methods for securing understanding in interviews. This can be done by involving more linguists in police education and training and working with transcripts of naturally occurring interactions between investigators and their interviewees. This will allow investigators to critically reflect on their own practices and discuss and explore what strategies they may use to explain legal rights in an understandable way.

Seeing that our study has focused specifically on suspects who are L2 speakers, it is also relevant to consider whether and how this particular group is and should be given specific attention in interview training and in legal regulations. Currently, specific mention of language issues in relation to the communication of rights may be found at an international level in the European Convention of Human Rights (European Convention for the Protection of Human Rights and Fundamental Freedoms, 1950, Art. 5 §2), as well as in the recently adopted Norwegian Interpreting Act (Interpreting, 2021). Both treat the L2 speaker as someone who either speaks the language or not – in which case s/he should preferably get an interpreter. Such formulations lack the recognition that in many cases L2 speakers interacting with the police may find themselves somewhere between having no and having full competence in a language. Also, many L2 speakers may come across as quite fluent in talking about everyday matters but experience difficulty in understanding the complex technical language of legal regulations (Pavlenko et al., 2019). Finally, access to (qualified) interpreters may sometimes be highly limited, for instance, when the interview is conducted on-site. In such cases, knowledge of comprehension strategies is of crucial importance for both securing understanding and guaranteeing the legal admissibility of the interview in court.

Acknowledgments

We would like to thank Oslo Police District for making these data available to us and especially Håkon Johannessen for invaluable assistance in selecting and exporting the relevant audio and video files. This work was partly supported by the Research Council of Norway through its Centres of Excellence funding scheme, project number 223265.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

The work was supported by the Norges Forskningsråd [223265].

Notes on contributors

Jan Svennevig

Jan Svennevig is Professor at the University of Agder and Research Professor at MultiLing – Center for Multilingualism in Society across the Lifespan at the Uni¬versity of Oslo. His research deals with conversation analysis in multilingual set¬tings, with an emphasis on practices used by first-language speakers for securing mutual understanding in interaction with L2 users. He has also published studies on sources of misunderstanding and hostility in intercul¬tural and multilingual conversations in various institutional settings.

Paweł Urbanik

Paweł Urbanik is Associate Professor at the Norwegian University of Science and Technology in Trondheim, Norway. His research focuses on language in everyday and institutional interactions. Since 2019 he has been studying police interviews and training Norwegian police officers and instructors on securing understanding.

Aafke Diepeveen

Aafke Diepeveen is a PhD research fellow at the Department of Languages and Literature Studies at the University of South-Eastern Norway. Her research focuses on the analysis of interaction in police investigative interviews.

Notes

1. Access to the recordings has been granted by the Oslo Police District (Oslo politidistrikt) with the permission of the Director of Public Prosecutions (Riksadvokaten). Ethics approval has been given by the Data Protection Officer (Personvernombudet) at the University of Oslo and Norwegian Agency for Shared Services in Education and Research (ref. no. 474082).

2. Yet, we do not possess information about the exact of the investigators’ experience.

3. In one case (LB-2021–63748), an interview was ruled inadmissible as evidence by the Court of Appeal (Nor. Borgarting lagmannsrett) based on the claim that an L2 suspect had not been informed in a sufficient and sufficiently understandable way of his right to legal counsel and specifically of the fact that he qualified for having appointed a public defender. In another case, (HR-2022-866-U) an interview was ruled inadmissible as evidence by the Appeals Selection Committee of the Supreme Court (Nor. Høyesteretts ankeutvalg) due to ‘insufficiently clear information’ and the suspect ‘not equivocally renouncing the right’ to legal counsel.

Appendix

Table

Symbol	Description
(.)	A micropause – a pause of no significant length.
(0.7)	A timed pause – long enough to indicate a time, in seconds.
[]	Square brackets show where speech overlaps.
>word<	Arrows showing that the pace of speech has quickened.
word-	A dash indication a cut-off, typically a glottal stop in phonetic terms.
(word)	Unclear/uncertain word or section; no plausible candidate if empty.
underlining	Emphatic stress on syllable
↑word	Rise in pitch
↓word	Drop in pitch
. , ?	Markers of final pitch direction: Final falling intonation (.) Slight rising/continuing intonation (,) Sharp rising intonation (?)
=	Placed at the end of one turn and the start of the next, indicating that there was no pause between them (latching).
: ::	Colons indicate a stretched sound: the more colons, the longer the sound stretch.
.hhh	Inbreath. Duration indicated with fewer or more letters.
hhh	Outbreath. Duration indicated with fewer or more letters.
#	Hash sign indicates creaky voice.