When you say nothing at all: Invoking inferences from suspect silence in the police station

Abstract

Drawing on qualitative research with criminal justice professionals, this article explores the practical operation of provisions allowing for inferences to be drawn from silence at the point of police questioning in Ireland. The article examines (1) pre-interview disclosure and the timing of the invocation of inferences within the detention period; (2) the manner in which suspects are informed about the possible consequences of any failure to answer questions or mention certain facts; and, (3) the value and impact of inferences at trial. Findings include the need to reconsider the late disclosure approach adopted to police interrogation in Ireland, while maintaining separate inference interviews; difficulties with the current ‘ordinary language’ examples used to explain inference provisions to suspects; and, a notable distinction between the use of inferences in particular courts in the Irish criminal process.

Introduction

Legislative provisions allowing for inferences to be drawn at trial from the failure of a suspect ¹ to answer certain questions or mention certain matters during pre-trial investigative interviews with police exist in a number of jurisdictions. ² While those in operation in England and Wales are perhaps most well-known, they are not the only operational inference provisions. In Ireland a set of legislative provisions allows for inferences to be drawn in a number of specific circumstances, as fully outlined later. Indeed, the original Irish inference provisions (which have since been extended and amended) pre-dated the introduction of similar provisions in Northern Ireland by four years and in England and Wales by ten years. ³

In each jurisdiction where such provisions exist, the courts have had to consider their operation, both in the police station and at trial. When is it fair to invite the jury to potentially draw an inference from silence? Do specific questions need to be put to the suspect in order to later attach evidential value to a failure to answer? ⁴ Can an inference be avoided if the reason for silence is the suspect's legal advice? ⁵ Can an inference be drawn against a suspect in relation to a particular offence where they were originally arrested and detained in regard to another? ⁶ Can an inference be avoided by a suspect giving any account of matters, no matter how implausible? ⁷

In this article we consider three significant aspects of the practical operation of inference provisions at the point of police detention and questioning in Ireland. First, we examine the timing of the invocation of inferences within the detention period, and the related issue of pre-inference interview disclosure. Secondly, we explore the manner in which suspects are warned or informed that a failure to answer questions or mention certain facts at the pre-trial interview might lead to the drawing of an inference at trial. Finally, we consider the value and impact of inferences at trial, before drawing the article to a conclusion with some observations and recommendations for reform.

The article draws on qualitative interviews conducted with professionals working across the criminal process in Ireland. These interviews were conducted as part of an EU-funded project known as ‘EmpRiSe: Right to silence and related rights in pre-trial suspects’ interrogations in the EU: Legal and empirical study and promoting best practice’ which examined the right to silence in police interrogations in Belgium, Italy, the Netherlands and Ireland. Of those jurisdictions Ireland is the only one with a legislative inference-drawing regime. ⁸ However, this does not mean that trial courts in other jurisdictions are oblivious to suspect reliance on silence in the pre-trial period or carefully avoid attaching any evidential value to same. In fact, research within the EmpRiSe project suggested that pre-trial silence quite often has at least some evidential purpose within criminal trials in a number of jurisdictions, whether that be to undermine the credibility of a defence, to corroborate or support other evidence in the case, to assist the interpretation of other evidence, or to validate the prosecution case in the absence of a counter-narrative from the defence (Daly et al., 2021). In the absence of legislation, however, suspects detained for questioning in such jurisdictions are not routinely warned of the potential trial significance of their pre-trial silence. ⁹ One benefit of the legislative provision for inferences is that their operation is at least transparent, and a relevant caution or explanation must be provided at the pre-trial interrogation stage to allow for their later use at trial. Nonetheless, it should be remembered that inferences are a significant incursion on a traditionally protected fair trial right, and their true evidential value remains questionable. ¹⁰

Before getting into the detail of the three specific aspects of the operation of the inference provisions in Ireland set out above, we outline the methodology adopted for the research, and we set out the Irish legislative provisions. In doing so, we also outline the inference provisions operational in England and Wales and in Northern Ireland, in order to place the Irish laws in comparative context. This serves as a useful backdrop to the empirical findings on the operation of the Irish provisions which follows.

Methodology

The findings presented in this article are based on research undertaken in Ireland as part of the EmpRiSe project, which is described above. A multi-method, qualitative research design was used to explore legal professionals’ experiences and perceptions of the right to silence at the pre-trial investigative stage. This involved carrying out two focus groups with a total of 19 criminal defence solicitors and one-on-one, semi-structured interviews with 10 barristers, 11 staff from the Office of the Director of Public Prosecutions, 4 judges and 6 retired gardaí. The researchers applied through the Garda Research Unit on two occasions requesting research access to currently serving gardaí but these requests were unfortunately denied. Accordingly, the researchers turned to recently retired gardaí in order to ensure that the very important policing perspective on the issue of the right to silence could be captured, and that the perspective offered would be current. Of the six retired garda participants, four had retired within the previous two years, one within the previous four years and one within the previous seven years.

The research took place between March 2020 and April 2021. Focus groups were initially selected, given their suitability to explore experiences and perceptions as well as for participants to interact with each other to bring added nuance to discussions (Acocella, 2012). However, restrictions that were introduced in Ireland to curb the spread of COVID-19 necessitated a switch to conducting remote, individual interviews. Semi-structured interviews are similarly well suited to exploring and unpicking different experiences (Thille et al., 2021).

A total of 50 participants took part in the research. Participants for both the focus groups and interviews were recruited using accredited/professional and organisational networks. Those who were interested in participating in the research project were invited to contact the research team, who used purposive sampling to select participants. Purposive sampling is particularly recommended in order to obtain a sample of research participants who are capable of providing the relevant information (Morgan et al., 1988). Given the lack of empirical research on the inference provisions, this was deemed to be an appropriate sampling strategy to ensure that participants had knowledge and experience of those provisions. To minimise omission and inclusion bias, the research team sought to include as many different groups of professionals who had relevant experience as possible to ensure that a range of experiences, opinions and perspectives were captured achieving a mix of gender, years of experience, different professional roles within relevant organisations and participants who were involved in prosecution, defence or both. This was achieved by pre-screening potential participants using a short questionnaire to gather such details.

Those who had been selected to participate were supplied with participant information sheets, which gave details in plain language about the research aims, the nature of participation, the right to withdraw from the study and of how their data would be managed in a GDPR-compliant manner. Information sheets clearly set out the confidential and anonymous nature of participation, while also highlighting the limits of this. For instance, information that suggested that individuals might harm themselves or that someone else was at risk of being harmed would not have been treated as confidential. It was also explained that focus groups could present difficulties with anonymity, given that it was possible that participants would hear what others said in the focus group and so would be in a position to tell non-participants what was spoken about by whom. The research team requested that participants did not attribute statements to individual participants. The above information was explained verbally to participants and contained in the participant consent forms, which were signed in advance of the focus group or interview. Ethical approval for the study was granted by Dublin City University and Maastricht University.

The researchers prepared a topic list for the focus groups and a schedule for the interviews, which were based on a literature review specific to Ireland and developed in consultation with the wider, cross-jurisdictional EmpRiSe team. The precise questions varied, depending on the nature of the participants’ professional roles and experiences of the right to silence and the inference provisions. Examples of questions included: ‘What are the possible positive or negative consequences of remaining silent?’; ‘In your view, what should be the impact of such inference on a case?’ and ‘Are there any challenges or difficulties associated with the use of the inference provisions?’. Open-ended questions such as these were often followed up with more probing questions, which varied according to the participant's response. Participants were very willing to discuss their experiences, often providing detailed examples to illustrate their views. There were organic discussions and interactive engagement between focus group participants, who added nuance, counter-points and qualifications, to what other participants had mentioned. Interviewees were equally open about their experiences and perspectives, and spoke in detail about their day-to-day encounters with issues relating to the right to silence, and their policy-oriented views on the current operation of the law. The willingness of participants to contribute readily and thoroughly suggested a real desire to engage with issues surrounding the inference provisions, and the right to silence more broadly.

With participant permission, the in-person focus groups and phone interviews were recorded by Dictaphone. Interviews that took place on Zoom also were recorded using a Dictaphone, so as to avoid any concerns about third-party data processing and storage. The research team transcribed the focus groups and interviews and carried out a process of data cleaning to anonymise the transcripts. The transcripts were uploaded to NVivo and analysed following Braun and Clarke (2013) stages of thematic analysis: transcription, reading and familiarisation, coding, searching for themes, reviewing themes, defining and naming themes, and finalising the analysis. In light of the lack of empirical research on this area, thematic analysis offered the flexibility to explore different perspectives and lived experiences between participant groups, as well as facilitating a combination of inductive and deductive coding (Braun and Clarke, 2021). ¹¹

In the presentation of the findings below, a participant's profession is indicated by an assigned letter(s) as follows: solicitor focus group participants (S), barrister interviewees (B), DPP staff interviewees (D), retired members of An Garda Síochána (RG) and judges (J). This letter is followed by a number, to refer to specific participants (e.g. B1, D2, J3 etc). For focus group participants, there is also an additional letter to show which focus group the participant attended (e.g. S1A or S1B).

The Irish law on inferences

The right to silence has been recognised as a constitutional right in Ireland though it is not an absolute right. ¹² The same can be said of the protection for the right, often interchangeably referred to as the privilege against self-incrimination, under the European Convention on Human Rights and the jurisprudence of the European Court of Human Rights. ¹³ Irish law has curtailed the right to silence in the criminal process in two distinct ways: first, by creating a number of offences based on a failure to answer questions or provide certain information, and secondly by the introduction of legislation specifically allowing for inferences to be drawn at trial from the failure of the suspect to account for certain matters or mention certain facts during the pre-trial investigative period. This article is focused on the latter provisions. ¹⁴

The general rule which operates in order to uphold and protect the right to silence within the criminal process in Ireland requires that a jury in a criminal case should not be told about any failure or refusal of a detained suspect to answer garda (police) questions or provide information to gardaí (police officers) during the pre-trial process. ¹⁵ If a suspect maintains silence throughout garda interrogation, the jury at trial will be told nothing about the interrogation. If certain questions were answered and others not, the jury will only be told of the answers, not the failure to answer. ¹⁶ However, when certain legislative provisions are operational a jury can be told about, and invited to draw inferences from, specific failures of the detainee. Three of these inference provisions, applicable to all arrestable offences (i.e. offences with a potential sentence of imprisonment of 5 years or more) are contained within ss. 18, 19 and 19A of the Criminal Justice Act 1984, as amended by the Criminal Justice Act 2007 and provide as follows:

the trier of fact at trial can be informed of the accused's failure (during questioning, at or before the point of charge, or when informed that they might be prosecuted) to account for any ‘object, substance or mark, or any mark on any such object’, that was (i) on his or her person, (ii) in or on his or her clothing or footwear, (iii) otherwise in his or her possession, or (iv) in any place in which he or she was during any specified period (s. 18). A garda must have reasonably believed that the object, substance or mark may be attributable to the accused's participation in the commission of the relevant offence and must have informed the accused of that belief. The trier of fact may ‘draw such inferences as appear proper’ from the failure or refusal to account for the object, substance or mark, though no further guidance is given on what such an inference might be;

in a similar manner, the trier of fact at trial can be informed of the accused's failure to account for their ‘presence at a particular place at or about the time the offence is alleged to have been committed’ and an inference may be drawn from this (s. 19);

and, under the widest of the provisions, the trier of fact at trial can draw an inference from the accused's failure ‘to mention any fact relied on in his or her defence…being a fact which in the circumstances existing at the time clearly called for an explanation’ (s. 19A).

Those familiar with the provisions in England and Wales or in Northern Ireland will recognise the similarities across the three jurisdictions. The operation of ss. 18, 19 and 19A of the Irish Criminal Justice Act, 1984, is almost identical to ss. 36, 37 and 34 respectively of the Criminal Justice and Public Order Act (CJPOA) 1994 in England and Wales, and Articles 5, 6 and 3 respectively of the Criminal Evidence (Northern Ireland) Order 1998. Interesting, while an inference may be drawn in England and Wales ¹⁷ and in Northern Ireland ¹⁸ from a suspect's failure or refusal to give evidence at trial, this is not a part of Irish law.

Two additional legislative inference provisions are operational in Ireland. These are very broad in terms of the grounds on which an inference can be drawn but narrow in terms of the specific offences to which they apply. Section 2 of the Offences Against the State (Amendment) Act 1998 and s. 72A of the Criminal Justice Act 2006, which was inserted by s. 9 of the Criminal Justice (Amendment) Act 2009, both allow for inferences to be drawn at trial from the pre-trial failure of a suspect to ‘answer a question material to the investigation of the offence’. Section 2 applies only to the offence of ‘membership of an unlawful organisation’, ¹⁹ while s. 72A applies in cases relating to participation in or contribution to any activity of a ‘criminal organisation’. ²⁰

Across all of the inference provisions in Ireland, a number of important procedural protections are mandated in the legislation:

the inference may not be the sole or main basis for a conviction but may serve as corroboration of any evidence in relation to which the failure is material;

the suspect must be afforded a reasonable opportunity to consult a solicitor prior to the relevant silence;

the court or jury in deciding whether or not to draw inferences ought to consider when the account or fact concerned was first mentioned by the suspect;

no inference shall be drawn in relation to a question asked in an interview unless either the interview has been electronically recorded or the detained person has consented in writing to the non-recording of the interview; and,

the suspect must be told in ‘ordinary language’ what the effect of any relevant failure or refusal to account for a pertinent matter might be.

Similar, though not identical, safeguards exist in the operation of the inference provisions in England and Wales and in Northern Ireland. For example, a person cannot be convicted on the basis of an inference alone. Furthermore, in order to draw an inference, a jury or court must be satisfied that the reason for the silence was that the suspect had no innocent explanation or none that would stand up to scrutiny. ²¹ For an inference to be drawn under s. 34 CJPOA, where the suspect fails to reveal facts at interview which they rely on in their defence at trial (similar to s. 19A in the Irish context), a number of conditions must be met: the suspect must have been cautioned; they must have been questioned by a constable trying to discover whether or by whom an offence has been committed; the suspect must have failed to mention a fact later relied on in their defence, and; the fact must have been one which they could reasonably have been expected to mention in the circumstances existing at the time. ²²

The caution

Despite the introduction of the above-outlined inference provisions in Ireland, no change was made to the traditional caution administered at the point of arrest and/or the point of police questioning, which derives from the Judges’ Rules 1918. This caution states that

You are not obliged to say anything unless you wish to do so, but anything you say will be taken down in writing and may be given in evidence.

No reference is made to the potential consequences of any failure or refusal to provide certain information.

The cautions in England and Wales and in Northern Ireland were similar prior to the reforms introduced by the CJPOA 1994 and the Criminal Evidence (Northern Ireland) Order 1988, but they were adjusted to account for the inference-drawing provisions introduced, in particular under s. 34 and article 3 respectively. In England and Wales, the police caution now states that

You do not have to say anything. But it may harm your defence if you do not mention when questioned something which you later rely on in Court. Anything you do say may be given in evidence. ²³

The current caution administered in Northern Ireland is largely similar. ²⁴

The modified cautions reflect the fact that while, in principle, the suspect has a right to silence, this right is limited, and invoking that right may lead to an inference being drawn if the suspect relies on facts at trial which they did not tell the police about on being questioned under caution or on being charged, provided that it is reasonable to expect them to have mentioned them.

Because the wording of the Irish caution conflicts with the inference provisions, it must be withdrawn at the point where they are invoked during questioning. In notable contrast to the experience in Northern Ireland and in England and Wales then, the inference provisions are not operational in every garda interview with a suspect. Garda practice, at present, sees a suspect undergoing a number of interviews where the traditional caution is administered, before arriving at an interview in which the inference-drawing provisions are to be invoked. In such an interview, gardaí must withdraw the traditional caution and, as required under the inference provisions, explain ‘in ordinary language’ the potential consequences of a failure or refusal to provide information in response to specific questions that are about to be asked. ²⁵ It is to this aspect of the practical operation of the Irish inference provisions that we now turn.

Procedure, disclosure and logistics

As just noted, current Garda practice involves an initial number of interviews with a detained suspect where no warning about the possible drawing of inferences from silence at trial is administered, and then, potentially, one or more interviews in which the relevant legislation is specifically invoked towards the end of a detention period. This contrasts with the position in England and Wales, particularly in relation to the operation of s. 34, and in Northern Ireland under article 3 of the Criminal Evidence (Northern Ireland) Order 1998. In those jurisdictions, the caution accounts for the fact that it might harm the suspect's defence if they do not mention something at the point of police questioning which they later seek to rely on at trial, and therefore the general inference provision is in play, so to speak, throughout each police interview. In Ireland, the traditional caution still administered conflicts with the inference provisions and has necessitated the evolution of a different process.

Our empirical research indicates a reluctance on the part of gardaí to provide significant pre-interview disclosure, particularly in relation to interviews early in the detention process. The Garda Síochána Interview Model (GSIM), under which gardaí are trained to undertake interviews with suspects, is a late disclosure model. This model of investigative interviewing, which began to be rolled out in 2014, adopts an information-gathering approach, which prioritises active listening, empathy and rapport-building. All gardaí are to be trained in basic interviewing skills under the model, and some can advance to further stages of training appropriate to more serious and sensitive offences. Level 3 is the highest level of training for interviewers, while those in a supervisory role within investigations can go on to Level 4 (Sweeney, 2016: 147).

Under the GSIM, there are a series of steps to be taken in each interview. Gardaí begin the interview by seeking to build rapport with the suspect. They then move on to seek an account of relevant matters from the suspect themselves, without any direct interrogation or cross-examination. The suspect may then be challenged on their account and asked to clarify certain details, in a persistent but patient manner, though the questioning may be robust. Finally, gardaí seek to bring closure to the interview, reviewing the account given by the suspect, outlining the next steps and responding to any questions (Noone, 2015).

As a late disclosure model, the GSIM seeks to have the suspect provide their own account of information prior to the revelation of evidence obtained by the gardaí. While this has the benefit of avoiding the contamination of any inculpatory statement that might be made with information that the suspect might not otherwise have, as well as allowing for more spontaneous responses to the revelation of evidential material, it has a significant impact on the efficiency of garda interviews, and indeed on the period of detention as a whole, and on solicitors’ ability to comprehensively advise suspects.

The gardaí are not alone in adopting such a position. A police preference for late disclosure of evidence was noted in other jurisdictions on the EmpRiSe project also, in particular in Belgium and the Netherlands, two inquisitorial jurisdictions. ²⁶ In the Netherlands, the objective of the late disclosure approach was similarly identified as avoiding contamination of inculpatory statements with information which the suspect might not already have, so called ‘daderinformation’ in Dutch. ²⁷

Solicitors in Ireland informed us that the late disclosure model often leads them to advise their client to remain silent, adopting a ‘wait and see’ (S1A) strategy, until the evidence in existence becomes more apparent as the interviews unfold. ²⁸

A lot of the time we have to sit through one or two interviews…that's how we get our disclosure. So meanwhile you’re…advising your client in most of those cases to remain silent so that you can just get the basics of what the evidence is, what exactly is being investigated, what role they say your client had in the alleged offence and then you can advise your client. That could be several hours into a detention before you can actually advise your client. (S5B)

They [the gardaí] have all the power and all the cards and we wait for them to give us bits and pieces so we can advise…it's very unsatisfactory for a client because all you can say to a client is…‘in our experience they hold back on some information and on that basis, at this stage perhaps you’re best to say ‘no comment’ for the first interview at least’ (S7A)

You kinda do this waiting thing where you sit there for an hour and a half and go to the first interview and the client maintains ‘no comment’. And then they’ll do another one the very same and then – haha, we’ll take out the big surprise, bring out the CCTV – and then eventually you’ll have to have a consultation where now they’ve given you disclosure four hours in…I just have to sit there basically watching paint dry until we go through this motion and that the gardaí want to do this dance… (S10A)

I would certainly be waiting, okay maybe sort of further down the line I’d be saying ‘okay you now know what we’re dealing with, what do you think?’ But I would definitely be waiting to see what evidence there was against the client in this situation before anything is volunteered. (S3B)

According to the Garda Code of Practice on Access to a Solicitor by Persons in Garda Custody, ²⁹ ‘there is no legal requirement to have a meeting with a suspect's solicitor, or to provide information prior to interview.’ The Code warns that

…the premature disclosure of information/details may sometimes impede or interfere with the investigation. It must be remembered that an interview is part of the investigation process and there must be some spontaneity about the actual interview. If information is handed out first, the suspect can make up his/her answers and there is no spontaneity about the matter. ³⁰

However, it does advise gardaí to be as accommodating as possible, without compromising the integrity of the investigation, in order to decrease the chances of a ‘no comment’ interview or multiple requests for consultations throughout the interview by the solicitor or detainee. ³¹

The Garda Code lays out information that should, as a general principle, be disclosed pre-interview, including information relating to the alleged offence, the arrest, when the request for legal advice was made, and any available material evidence which would not prejudice an investigation. ³² Gardaí are also obliged to ensure that any information that is given to a detainee or solicitor is accurate and true and must not misrepresent the strength of the evidence they are presenting. It is arguable whether or not this level of disclosure discharges the state's obligations under Article 6 of EU Directive 2012/13/EU on the Right to Information in Criminal Proceedings, which provides that:

Member States shall ensure that suspects or accused persons are provided with information about the criminal act they are suspected or accused of having committed. That information shall be provided promptly and in such detail as is necessary to safeguard the fairness of the proceedings and the effective exercise of the rights of the defence.

Retired gardaí in our study provided their perspective on the reluctance to provide significant pre-interview disclosure of existing evidence. They suggested that this would lead to a genuine and spontaneous reaction from the interviewee, without the possibility of any false excuse being fabricated in advance.

We want the interview to be spontaneous, not telling the solicitor everything we have, and then the suspect coming in with prepared answers to the evidence the guards have. That would be the garda perspective on that… We like spontaneous answers, not prepared answers… Particularly, when he tells a lie, and you can show him later on that he's telling lies. That's a huge thing for the guards. And it goes against his credibility. Whereas if you told him you had that piece, he might never tell the lie. (RG6)

…the guards have a duty of care by the truth. And what they can't do is prepare the person coming in to put a defence up to what they’re going to see. If they’re telling the truth, that won't be necessary. (RG3)

Solicitors, as noted above, often consider that they cannot properly advise their detained clients on the basis of sparse pre-interview disclosure. Indeed, this is acknowledged by the Law Society of Ireland, and in its Guidance for Solicitors Providing Legal Services in Garda Stations, ³³ it suggests that if investigating gardaí are unwilling to make disclosures, solicitors may find it helpful to inform them that ‘in the absence of full and proper disclosure, clients cannot receive comprehensive legal advice’. ³⁴

The difficulty created for solicitors advising clients was acknowledged by many participants in our study, including both solicitors and barristers:

…we need to get disclosure that enables us to give real advice to our clients, so we’re not advising in a vacuum… (S3A)

…if they’re not willing to give disclosure then you’re unable to advise your client how to proceed. (S6A)

The other problem the solicitor has is that often the solicitor is advising blind because they won't have the full extent of the evidence available to them, they will have…often nothing more than an outline of the accusation. (B8)

As the main method of disclosing the evidence in existence against the detained suspect then is a drip-feed of information in the course of interviews, it can take some time before a solicitor is prepared to advise anything other than remaining silent, and/or before a suspect considers it appropriate to comment or respond to questions.

Part of the impact of the late disclosure approach of the GSIM is that a suspect might well refrain from answering questions through several interviews, in the absence of a full picture of the case against them at that juncture. This can lead to the invocation of the inference provisions, as inference interviews are held only where a suspect has remained silent either entirely or at least in relation to certain material matters.

Yeah, and it's usually the last interview you do. Because the whole basis of the inferences is that you’ve interviewed him and he has made no comment about all those issues and now you’re putting the inferences to him. (RG6)

Gardaí trained to Level 2 of the GSIM will have been given lectures on the inference provisions and those trained to Level 3 will have undertaken a greater level of experiential learning in that regard. Those trained to Level 4 act as supervisors on serious investigations. There was a sense among participants that inexperienced gardaí generally struggled with inferences, displaying a lack of confidence and clarity in what they were doing. In contrast, gardaí with more experience and higher levels of training in conducting interviews are seen as able to administer inferences with more confidence, and with better outcomes in terms of admissibility.

The Special Detective Unit guard who would be doing those interviews are trained to do it correctly. (B3)

Participants noted that the level of training and experience of gardaí involved in certain investigations may influence the invocation of inferences.

The more experienced the guard, the more likely you are to have the inference provisions put to an accused. Sometimes, I dunno, I think guards might be a little bit afraid of them sometimes and don't utilise them (D11)

Relatedly, the seriousness of an offence is likely to impact on the usage of the inference provisions. More specialised gardaí with a higher level of interview training are involved in investigations into serious offences. Suspects are also more likely to remain silent in such cases (Daly et al., 2021: 72, 41–43). Accordingly, participants noted that inferences were more likely to be deployed in serious cases. One significant exception observed by participants is that inferences are rarely invoked in sex offence cases. There are two possible reasons for this: (1) suspects detained on sex offences charges tend not to rely on their right to silence, and are more likely to deny that the encounter was non-consensual; and (2) where the question of consent is central, inferences are of little to no value (Daly et al., 2021: 72, 41–43).

The detention time available for certain offences may also be a relevant factor in whether or not inferences are engaged. The operation of the inference provisions in the Irish context is a relatively lengthy procedure. Accordingly, in some investigations, there simply may not be sufficient time to invoke inferences (Daly et al., 2021: 72–73).

As compared with non-inference interviews, the Garda Code provides for a greater level of pre-interview disclosure where the inferences are to be used, given the evidential value of silence in that context. The Code envisages a pre-interview briefing between the investigator and the solicitor in such circumstances, during which the investigator should furnish the solicitor with enough information to enable them to perform their role without compromising the interview process. The Code specifically states that provision of information in this context is not the same as any post-charge prosecutorial disclosure requirement, but the solicitor should be provided with ‘certain basic facts that contextualise the matters to which the questions are going to relate’ so that they can advise the suspect appropriately on any decision to answer or not answer such questions.

In accordance with the legislative safeguards attaching to the inference provisions, the Garda Code of Practice indicates that, prior to an inference-drawing interview, gardaí should inform the suspect of their intention to invoke the provisions, remind the suspect of their right to remain silent, and explain the effect of the relevant provision(s) in ordinary language in a step-by step-fashion. Gardaí should inform them that the member reasonably believes the facts may link the suspect to the offence, advise them that the interview will be electronically recorded (unless the suspect consents in writing to it not being recorded) and that the recording may be given in evidence at any future trial. ³⁵

In the next section we discuss the challenges associated with explaining the inference provisions to detained suspects in ‘ordinary language’ at the beginning of an inference interview. Before moving on, however, it is interesting to note our findings in relation to the impact of the use of inferences in the garda interview room. It seems that, generally, the invocation of inferences, in and of themselves, does not induce individuals to start answering questions where they had earlier failed or refused to do so.

I would find it very unusual to have a ‘no comment’ interview and then an answer to one of the inference questions. (D10)

If you’ve got somebody who has got to the end of 7 or 8 h of interviews and they’ve made no comment, just because somebody pulls out a piece of paper that says their DNA was on something, they’re not going to change their view. They’re just going to stare at the wall. They will be well advised to do that, in fairness, in reality. (D2)

Retired garda participants had some experience of suspects providing an account of relevant matters in inference interviews, but where this occurred it was often a relatively limited account, addressing an existing piece of evidence.

My experience is you would get a lot of people saying something because they’re now being told that something can happen as a result of them not doing something. You’re not going to break, you’re not going to get the big, ‘we’ve sunk the case’ job out of it. That's not going to happen…it could be very vague…they wouldn't be saying, ‘oh yeah that's the baseball bat I used to smash Michael's head in.’ Again, I speak in my experience. Other people might say, ‘no, they never say a bloody thing.’ It depends on the individual. (RG4)

Where a suspect did answer inference questions, some participants viewed this as being more likely due to the number of interviews that the person was subject to, and the revelation of the evidence the gardaí then held as part of that process, rather than as a result of concerns around the use of their silence at any later trial. It seems that the strength of the evidence against the suspect at a certain point in the interview process might result in them beginning to answer questions, rather than the invocation of the inferences per se.

There has been changes, yes, they have changed their position. But I would put it down to good interviewing technique [rather] than using the inferences. Because your investigation and interview process is a process that you’re going through. And it must be quite robust because it will be tested in the courts and rightly so, that it would be tested. So it has to be very fair and very transparent… So I think in situations like that people have, been like, ‘you know what, at this stage, you might as well come through with something and see’. (RG5)

…so my experience, where accused or suspects tend to change their story is when they’ve done loads of interviews. And at the last interviews, they tend to change, if they change. It's not through the use of inference provisions generally speaking. (B1)

In all the cases that I’ve seen, before the inference interview…the guards have set out their stall: they’ve set out all their information and they don't spring something new at the inferences stage. Of course, statutory requirements allow for legal advice and you could make a very strong case that no inference could be drawn if there wasn't proper disclosure at that point prior to the legal advice. But then they have the opportunity to give their answers in the full knowledge of the totality of the picture. (B2)

There were some cases where solicitors may advise someone to give an account of relevant matters at the inference stage, particularly if they had an exculpatory statement to make.

You maybe have an eye on what you’re going to advise in relation to inferences. Sometimes it’ll be, I think, in case there's something helpful to say, maybe at that stage…but if there's not, a lot of the time I think the advice is the same. (S7B)

If they invoke the inference provisions properly and if they ask the right questions, you nearly have to give an answer I think, if you happen to have an explanation. (B4)

While the current Irish process, involving a separate, stand-alone ‘inference interview’ at the end of a detention period has evolved somewhat organically from the failure to update the caution when introducing the inferences, as well as from the late disclosure approach of the GSIM, we suggest that it represents a more principled approach to the application of evidential value to reliance on the right to remain silent than the ‘always-on’ inference process which operates in England and Wales, and in Northern Ireland. At the point of principle, it seems fairer to attach evidential value to silence only at the point where the suspect is faced with all of the evidence in existence against him. This both ensures that the suspect is then making an informed decision in considering whether or not to respond to garda questions to provide certain information, and it brings the evidence to the threshold at which it could be said to ‘clearly call for an explanation’ or to create a reasonable expectation of a response. ³⁶

One retired garda participant was of the view that it would be preferable if the inference provisions were in operation in all interviews:

…my attitude with rights is that they come with responsibilities as well, and there should be some responsibility there in relation to inferences from the very start in every question that's asked in interview, that's my view. (RG6)

However, in the Irish context at least, very often it is only at a late stage in a detention that the existing case and evidence against the suspect will have crystallised and there will be a case presented for answer. As one barrister participant put it:

…the greater degree of information afforded to an accused, the more prejudice perhaps…could result in their failure to account for a mark or whatever, their presence etc. There must be a commensurate level of information available to them before they are prejudiced for not responding. (B8)

A suspect must also be informed, in ordinary language, of this prejudicial impact of a failure or refusal to respond to certain questions during the inference interview. We turn now to examine the practical operation of this legislative safeguard.

Explaining the inference provisions

As noted earlier, the cautions administered respectively in England and Wales and in Northern Ireland at the point of arrest and prior to any police interview with a detained suspect allow, within their wording, for the concept of certain inferences being drawn from silence. Concerns have been raised in a number of studies about police officers’ actual understanding the terms of the caution, and their concomitant ability to explain its meaning to a suspect who might not understand (Blackstock et al., 2014: 378–380; Clare et al., 1998: 326; Quirk, 2017: 62–67; Skinns, 2019: 118). ³⁷ Other studies have highlighted a lack of comprehension or ability to adequately explain the caution amongst legal advisers (Quirk, 2017: 65), suspects (Pivaty, 2019) and the general public (Clare et al., 1998; Fenner et al., 2002; Shepherd 1995; Weaver, 1994).

In Northern Ireland, it does not appear that the caution is commonly explained or that steps are taken to ensure that suspects understand it. A solicitor in a 2018 study by Kemp (2018: 6–7) stated:

You can find the police rambling through the caution at lightning speed and with no emphasis being placed on what it actually means. This is the most important legal protection and exposition of a suspect's rights, and the police treat it as if it's some form of spiel they have to go through

Similarly, in a study from the mid-1990s there were reports that police officers followed the caution with phrases such as, ‘now is the time to speak’ or ‘you have to speak to us now or we’ll charge you’, which was misleading and over-emphasised the consequences of the 1988 Order (Weaver, 1994: 16).

In England and Wales more recently, Quirk (2017: 63) found that police explained the caution in a manner which made suspects feel like they had no choice but to forgo reliance on their right to silence, and Skinns (2019: 118) noted that the explanation of the caution may emphasise the inferences, compounding the pressure on vulnerable suspects, such as children, those with intellectual disabilities and those inexperienced in the criminal justice system.

Our research brings to light the manner in which inferences from silence are outlined to suspects within the Irish criminal process, and raises specific concerns around current practice. At the beginning of an inference interview, gardaí need to revoke the traditional caution and explain to the suspect that silence in response to specific questions, or in relation to a fact which they subsequently seek to rely on at trial, could potentially result in adverse inferences being drawn at any trial. It is a legislative pre-requisite to the drawing of inferences at trial that the suspect was told about the impact of any relevant silence in ‘ordinary language’. ³⁸ While initially there was no universal approach adopted to this, more recently gardaí seem to be employing two specific scenarios in an effort to explain the very complex issue of inferences from silence to suspects prior to questioning them under one or more of the inference provisions. While efforts to standardise the process, and to provide examples to aid understanding are laudable, participants in our research study, solicitor participants in particular, raised concerns around both the legal accuracy of these scenario examples, and the level to which they assist suspect understanding of the inference provisions.

Participants consistently reported the use of two specific examples.

So they always give two examples, one is there's a child in the kitchen with loads of chocolate on its face and there's a slice taken from a chocolate cake, the inference you can draw is that the child ate the cake…and the other is, in your house someone has come in and they’re soaking wet with a wet umbrella or it's wet after you wipe up, you can draw an inference that it rained last night or whatever. (S5A)

Many participants across the different groups were critical of how effective the examples were, in terms of helping suspects understand what inferences meant and how it applied to them. Part of this difficulty relates to the fact that, in the view of participants, the scenarios do not accurately reflect the process by which an inference is drawn from silence under the legislative inference provisions. One participant, for example, contended that what's happening in the scenarios is that a conclusion is being drawn on the basis of circumstantial evidence, not an inference from a failure or refusal to account for something, or to mention a fact later relied upon:

That is not an inference…that is a conclusion based on circumstantial evidence. An inference is where you don't have any direct evidence and you are inferring what probably took place. Jimmy is in a part of town in which he does not live. A jewellery shop was robbed and Jimmy is apprehended by a guard 200 yards away. The question is put, ‘Jimmy, what were you doing in that part of town?’ If Jimmy fails to answer, then…the court will be invited to infer that Jimmy had something to do with the robbery. There is no evidence, there is nothing to connect him other than the supposition that, because he's in a part of town he doesn't belong, that he has something to do with the crime. (S3B)

I don't think I’ve ever been at an interview in a garda station where they’ve actually explained inferences correctly; instead of explaining a legal inference, they always give the factual inferences which are not the same…factual inferences…make sense, they are rational. Whereas legal inferences are not. Legal inferences are where the fact of you not admitting something or answering a question – they want to use that to convince a jury that you must be guilty. Therefore, the examples they give are not consistent with what a legal inference is. (S1A)

They’re wrong, what they say to them about the chocolate cake makes no sense. (B8)

This point seems fair. In both scenarios there is some evidence from which a conclusion can be drawn. What is happening with inferences, however, is that additional evidence is being created, in a way, on the basis of a failure to account for existing evidence. So, for example, with the chocolate cake scenario, if the child was asked to account for the chocolate on their face but fails or refuses to do so, that is where the inference would be drawn. The chocolate cake court would then have two distinct items of evidence: the chocolate on the face, plus the failure to account for the chocolate on the face. In the same manner, if a suspect is asked to account for their presence in a particular place, which can be established by CCTV evidence for example, but they fail or refuse to so account, then the court has the evidence that they were at the place plus the evidence of the silence in the face of a request to explain what they were doing there. The scenarios employed need to be revised in order to describe in a legally accurate manner the operation of the inference provisions.

Furthermore, in terms of assisting suspects comprehension, solicitor participants highlighted that for some suspects, such as those with intellectual disabilities or for particularly young suspects, the inference provisions were difficult – and perhaps impossible –to understand. The confusion was compounded by the length of the caution and explanations. Retired garda participants also recognised that the meaning and operation of the inference provisions is complicated and can be difficult for suspects to understand.

To tell somebody that inferences will be drawn if you refuse to answer this question, it really means nothing to them. Unless they’re educated. And a lot of people who get caught in crime are not educated. (RG1)

And for juveniles, people with significant issues, there is no way that they can understand the inference provisions. (S6B)

…for some people it's just a concept that is very hard to understand and the way the caution is given is really confusing… It's too long, I mean honestly sometimes by the time the guards are finished, I don't know what an inference is either. (S5B)

Many participants reported that the inference invoking procedure was lengthened by the fact that some gardaí read out the relevant piece of legislation, even though one retired guard specifically noted that this was not necessary.

[I]t doesn't require you to read the legislation over to the suspect, you just have to tell them what it means in ordinary language. So we started doing that and that works out a lot better…probably still lots of guards around the country tying themselves up in knots over that reading the legislation…you usually only have an hour or two left at the end of your detention period to get your inferences interview done and you can be flying through it reading really fast so it has no benefit to his understanding anyway, if you’re reading the legislation. (RG6)

Indeed, while the legislation requires that an ‘ordinary language’ explanation is given, there is no corresponding requirement to ensure that the suspect understands the implications of the inference provisions.

You have guards taking 20 min to read out the legislation and then they’ll turn around to the fella and go, you understand that now? And he says nothing, and everybody moves on. (B8)

The courts have recognised the division of labour as between gardaí and solicitors, suggesting that

It is for the member of the gardaí to inform the accused in ordinary language as to the general effect of a failure or refusal, and it is for the solicitor to advise the accused as to the potential impact in the particular circumstances of his own case. ³⁹

Of course, only a minority of detained suspects access legal assistance in garda custody (Daly and Conway, 2019), and therefore the ordinary language explanation of the gardaí might be the only explanation provided to some. Further study is needed in the Irish context on the ability of suspects to comprehend the inference provisions, and a review of the explanations provided by gardaí is urgently needed.

The value and use of inferences at trial

Views may differ on the evidential value of an inference drawn at trial from the failure of a suspect to respond to certain requests for comment or to mention certain facts later relied on. Prosecutors in our study were generally of the view that the inference, when used, is an additional piece of evidence to add to the prosecution pot, though it might not be enough to tip the scales in favour of conviction.

I don't think it would tip the scale really…it won't be what convicts you but it’ll just go towards it, but it's not the nail in the coffin if you know what I mean… (D9)

From a trial perspective, it's not the magic bullet that everyone thinks it is. (D1)

Inferences are not utilised in all cases. In fact, they form no part of a lot of cases. One prosecutor noted that in cases with strong evidence of guilt, the inferences are largely irrelevant, but where the evidence is more circumstantial they can be important.

The cases where it is invaluable are cases where the prosecution is based on circumstantial evidence. It's not going to be the case where we can prove all of the good stuff straight away on the basis of solid evidence because if we had that evidence we wouldn't need to rely on the inferences in the first place. So, it's going to be almost exclusively confined to very serious cases which rely on certain types of evidence which is inevitably going to be cases involving organised crime or something like that. (D2)

Even amongst the professionals across our study there were varying views on the real value of an inference, and the impact that it might have on a jury, those being two separate things. One retired garda speculated that inferences may have been significant in a case they had dealt with, but still underlined their weakness as a form of evidence:

…there was a whole heap of evidence suggesting it was him and the inferences probably convinced the jury ‘yeah, he's not answering questions about that, so it has to be him’ or whatever. We don't know the way that juries look at, and that's why, I suppose that's why we have to be so careful about using the inferences, the jury could take them to be a lot more… They’re not really strong corroboration of anything the inferences. (RG6)

One of the judges in our study also noted the impact that the judge telling the jury that they may draw an inference might have on them.

So, where you do have a judge saying you are now entitled to draw an adverse inference because the accused has failed to mention something that he is relying on in his defence, I think it has a very strong impact on a jury… (J4)

Speaking to the more specific inference provisions under which an existing piece of evidence grounds the request for an account, some barrister participants also viewed silence in the face of such evidence as impactful on a jury.

I believe personally that an inference interview where, in the face of a very strong account that requires an explanation, that's being met by silence by an accused, I believe in front of a jury that's powerful evidence. (B8)

I would have thought that the rhetorical question, ‘why wouldn't he give an explanation if he had one, why didn't he say something at the time?’ – I think that is potent, potentially potent material from the point of view of a juror. (B2)

On the other hand, many participants simply referred to inferences as the ‘icing on the cake’ (D11) or the ‘jelly on the ice-cream’ (J2) of existing evidence. A number of the judge participants queried the real value of adding the inference in an already strong prosecution case:

… At that stage…the case is almost made against the accused…I always feel that you’re so far ahead of the game at that stage that I don't know what additional icing they in fact give… I’ve never seen a case where it's actually the inference is what ultimately hangs the accused. I’ve never seen that. I mean…a lot of circumstantial evidence has already come and you’re really giving someone the opportunity to explain their situation but the circumstantial evidence is already there. (J1)

Really the inference doesn't come from the silence, the inference comes from the silence in response to what is in anybody's book a highly relevant question that ought to be answered…but the point about that is that if you’ve reached that stage where you have a highly pointed piece of evidence, the highly pointed piece of evidence is probably going to be good enough by itself. So I always think, I sometimes think, not always but a lot of times I think there's a bit of surplusage or circularity about the inference provisions because you’re invoking the inference provisions where you have a good bit of evidence anyway and you just really want to sort of drive it home. (J2)

One wonders then if all of the effort attaching to the proper administration of the inference provisions, both in the garda station and at trial, is worthwhile. And, indeed, if the level of interference in the right to silence occasioned by the inferences is justified. Research with real juries to determine the impact of inferences at trial would be interesting, though is currently not possible.

In the Irish criminal court structure there is one non-jury court which hears serious offence trials; the Special Criminal Court. This three-judge court, initially established under the Offences Against the State Act, 1939, deals largely with offences linked with paramilitary or organised criminality, though it can hear other cases also in certain circumstances. Two of the judges in our study had experience of both this court and the more usual courts in the Irish system. They noted that inferences can be extremely important in cases before the Special Criminal Court, particularly those relating to the very specific offence of ‘membership of an unlawful organisation’.

Usually in ‘membership’ cases there might be 7 interviews, 3 of them would be all ‘no comment's, then the inferences are invoked in section 2 and then they know to say ‘I’m not nor have I ever been a member of the IRA’ and that's the mantra. And then you have to go back through the things that called for an explanation and they’re [the inferences are] very helpful in those cases, you wouldn't have convictions is the truth of it, otherwise (J1)

There are several reasons why inferences are likely to be employed in cases before the Special Criminal Court more often than in the ‘ordinary’ courts. These include the fact that detained suspects in such cases are more likely to give ‘no comment’ responses because of fear of reprisal from organised crime associates, as well as distrust of gardaí; the offences are more likely to be investigated by specialised and highly trained gardaí (e.g. from the Special Detective Unit); and longer detention periods apply, giving greater time for gardaí to invoke the inference provisions.

Interestingly, in ‘membership’ cases, the Special Criminal Court may accept the evidence of a Garda Chief Superintendent that he believes the suspect was at a material time a member of an unlawful organisation. ⁴⁰ This evidence is often based on privileged information and, as such, is unchallengeable by the defence. While that belief evidence in and of itself cannot ground a conviction, an inference drawn under s. 2 of the Offences Against the State (Amendment) Act 1998 (from a failure to answer a question ‘material to the investigation of the offence’) can be used as corroboration, and these two items of unusual evidence can collectively lead to a conviction.

The fact that the Special Criminal Court gives a written judgment in each case provides an opportunity to see what impact inferences have in individual cases but, while a matter of public record, these judgments are not readily accessible, and, furthermore, the impact of inferences on a professional judge in an organised crime case may be quite different to that on a lay juror in a more ‘ordinary’ case. It is somewhat worrying to consider though that inferences might be as impactful as it seems they can be before the Special Criminal Court.

Conclusion

While Ireland is a member of the European Union and has opted in to a number of the EU Procedural Rights Directives, such as Directive 2012/13/EU on the Right to Information in Criminal Proceedings, mentioned above, it has not opted in to all of those measures. Most notably, in the context of this article, Ireland has not opted in to Directive 2016/343/EU on the strengthening of certain aspects of the presumption of innocence, or indeed to Directive 2013/48/EU on the right of access to a lawyer in criminal proceedings. The Irish Department of Justice appears to believe that the existence of Ireland's legislative inferences makes it impossible for the state to fulfil the terms of Directive 2012/13/EU (Daly, 2021), but this is not without question. Commentators (Pivaty et al. 2021) have observed that the final wording of Directive 2012/13/EU allows for a number of possible interpretations such that its asserted prohibition on the use of adverse inference might not be as definitive as originally foreseen when the Directive was first considered. In discussing the need to continue those provisions in operation the Irish Minister for Justice and Equality stated in 2019 that such provisions are ‘a key tool in the investigation and prosecution of serious crime’ ⁴¹ (Daly, 2021).

In this context, it seems an uphill task to row back on the very existence of inferences within the Irish criminal process, or indeed within the criminal processes of Northern Ireland or England and Wales. ⁴² They have become embedded in current criminal procedure and arguments for legislative repeal of these measures are likely to fall on deaf ears. Indeed, as noted in the introduction to this article, the EmpRiSe project research suggests that inferences from silence are a feature in other jurisdictions across Europe too, even in the absence of legislation allowing for same. At least where inferences are legislatively established there is transparency around their operation and safeguards attached to ensure the provision of related procedural protections, such as access to legal assistance specific to the drawing of inferences, audio-visual recording of interviews and an explanation of the operation of the inferences.

Having said that, it remains important to repudiate any suggestion that further, more expansive provisions could or should be introduced, and, as the inferences we already have appear to be here to stay, we must ensure that they are operating in a procedurally fair manner.

While the lack of an updated caution in Ireland may initially have been viewed as a negative aspect of the operation of the inference provisions, in fact this led to the organic development of a process whereby separate, stand-alone inference interviews are conducted toward the end of a detention period. In our view this is significantly preferable, from the perspective of fairness, to the situation whereby inferences can be drawn from any and all suspect interviews. In the context of the late disclosure approach adopted within the GSIM, this in fact seems to be the only method which makes sense. The slow revelation of existing evidence across a number of interviews with suspects means that it is only at quite a late point in proceedings that they will be in a position to consider the strength of the evidence against them and their attitude to responding or otherwise to garda questions. From a principled perspective, drawing on the foundational presumption of innocence, it is only at the point where the police hand has been shown that any prejudice could or should be attached to a suspect's failure to respond.

Nonetheless, the late disclosure model causes inefficiencies in the system, affecting suspects, solicitors and gardaí. A suspect, and their solicitor, if present, find themselves sitting through several interviews over a number of hours before the existing case against them is clear. Our research indicates that one of the greatest push factor towards a suspect responding to garda questions is the strength of the evidence. ⁴³ Of course, in some cases there may be a particular need to withhold disclosure so as to observe the spontaneous response of the suspect, but this is unlikely to be necessary in all cases. We advocate a review of the approach to pre-interview disclosure from both an efficiency and a fairness perspective, and additional training for gardaí to assist them in determining the situations in which more comprehensive disclosure might be expedient, and those situations where late disclosure is specifically preferable.

As noted above, the ordinary language explanations and examples being employed by the gardaí need to be reviewed and replaced, to ensure both legal accuracy and suspect comprehension. While we have discussed throughout this article the impact of late disclosure on legal advice to remain silent or otherwise, the need for a legally accurate and understandable explanation of the inference provisions is almost more important where a suspect does not have a solicitor, who would advise them further on the potential operation of the inference provisions in their specific case. While our research is the first of its kind to provide qualitative insights on the right to silence in the Irish criminal process, further studies would be very valuable. For example, qualitative research on suspects’ understanding of the inference provisions and the garda explanation thereof would be very useful to shed further light on this complex issue.

Research on the impact of inferences on jurors would also be very beneficial. Our research indicates that while the use of inferences is increasing somewhat in the ‘ordinary’ courts, they currently feature in only a small proportion of trials. This is partly explained by the fact that they are not always invoked by gardaí during the interview period, there may have been some procedural error in their usage, or the suspect may have provided an account and therefore they were not invoked. They feature more often and appear to be more impactful, in cases coming before the non-jury Special Criminal Court. We are particularly concerned about the use of an inference from silence as corroboration of the belief evidence of a Garda Chief Superintendent that an individual is a member of an unlawful organisation, and therefore guilty of the offence of such membership: two weak forms of evidence, before an unusual court, leading to conviction on an unusual offence. We believe that additional evidence should be necessary for conviction in such cases, and we question the legitimacy of unchallengeable belief evidence of this nature. Again, further research on the judgments of the Special Criminal Court to ascertain the level of impact of inferences on decision-making therein would be very valuable.

We have advocated herein for a number of procedural improvements, but there are further adjustments which could also be made to ensure that the provisions operate as fairly as possible, and do not lead to miscarriages of justice. For example, in our view, inferences should only be allowable in relation to a failure to respond to evidence which exists at the time of questioning and is disclosed to the suspect, and this requirement of grounding evidence should be clarified in the legislation relating to all inference provisions. Other associated procedural and rights-oriented protections must also be working effectively in order to ensure that inference provisions do not operate to create unfairness. For example, the right of access to legal advice must be practical and effective; police, prosecutors and judges must remain alive to the fact that silence is not always indicative of guilt and allow inferences only to be drawn where the trier of fact is satisfied beyond a reasonable doubt that there is no innocent explanation for the relevant silence; judicial charges to the jury on the operation of inference provisions should be carefully constructed so as not to place undue emphasis on this form of evidence; and gardaí should at all times be cautious not to directly or indirectly induce suspects to waive their right to remain silent (Daly et al., 2021).

Inferences represent a significant incursion on the right to silence, and a potential threat to the presumption of innocence. Convictions in criminal trials ought not to be lightly reached, and we must ensure that the operation of inference provisions is fair and appropriate in individual cases. The ongoing existence of inference provisions should not result in their strength being elevated beyond their real value, through normalisation and repeated usage. We must not lose sight of the fact that attaching evidential value to a meaning drawn from silence is, in truth, a very weak form of evidence.