3.1 The much-maligned evidential rule of hearsay exclusion has been subject to some interesting challenges in many common law jurisdictions since 2005. An anathema to lawyers of the civil or administrative law system and seemingly largely misunderstood in its complexity by many common law lawyers, the hearsay rule has been supplemented in some respects and undermined in others by various legislative reforms in both civil and criminal proceedings. This chapter does not seek to provide a comprehensive exposé of the hearsay rule. However, in drawing the rule back to its historical foundation we will, in part, consider its relevance in the context of electronic evidence and attempt to demonstrate the application of hearsay to evidence in electronic form.

The rule of hearsay exclusion and its rationale

3.2 We begin with a traditional definition of the hearsay rule. Sir Rupert Cross defined the hearsay rule of evidence in these terms: ‘[A]‌ statement other than one made by a person while giving oral evidence in the proceedings is inadmissible as evidence of any fact stated’.¹ In offering this definition, Sir Rupert Cross intentionally conflates in one definition the rule against self-corroboration with the rule against hearsay in the narrow sense,² to both encompass and contrast the situations where the witness who made the statement is either available to testify (the rule against self-corroboration) or unable to testify (the rule against hearsay in the narrow sense). If a witness’s earlier statement is sought to be admitted of the facts therein, it is at common law inadmissible except as an informal admission provided against that witness or as part of the res gestae rule.³ And where it is so admitted, it is not as evidence of the truth of the assertions contained in them.⁴ If the witness’s earlier statement is in the form of electronic evidence, this does not change the application of this rule, and does not engage the rule against hearsay.

3.3 Turning to the rule against hearsay, the most widely accepted judicial formulation of the rule is as follows:

3.4 The hearsay rule is prompted when the testimony of a witness as to what a declarant said is admitted in evidence to establish its truth, but not otherwise. Hearsay evidence is thus subject to at least four clear vulnerabilities: belief by the declarant that X exists (the risk of impaired perception), justification for the declarant’s statement that ‘X exists’ (bad memory, ambiguity and insincerity), confirmation that the witness correctly heard the declarant as having said ‘X exists’ (impaired perception) and justification for the witness in repeating that statement (the duplicated risks of bad memory, ambiguity and insincerity).¹ Hearsay therefore increases these risks because it has to contend not only with the testimony of the witness, but also with the testimony of the declarant. This is because the usual safeguards that apply in relation to ordinary testimony must also apply to second-hand evidence,² since there is no link between the testimony of the witness and the declarant’s proposition that ‘X exists’ that is sought to be supported.

3.5 Thus, some academics have advanced the argument that ‘[t]‌he basic rationale of the hearsay rule rests on the right of cross-examination’.¹ ‘The central reason for the presumptive exclusion of hearsay statements is the general inability to test their reliability. Without the declarant in court, it may be impossible to inquire into the declarant’s perception, memory, narration or sincerity.’² However, it should also be noted that cross-examination is only one way to test evidence: if it is the rationale for hearsay, evidence which is inherently reliable or which can be evaluated for reliability without the need for cross-examination could be admitted.³

The right of confrontation

3.6 A further, apparently important but ill-defined foundation of the hearsay rule that is related to the right of cross-examination is the right to confront an accuser as an essential element of the right to a fair trial. It was on this basis that the United States Supreme Court in Crawford v Washington held that where cross-examination is not possible for the testimonial statements of witnesses who have since become unavailable, the statements are inadmissible.¹

3.7 The right to confront draws on the notion that the right to humane treatment and procedural integrity both feel undermined by the admission of hearsay evidence.¹ Although expressly established in the Constitution of the United States² and in the European Convention of Human Rights,³ a review shows that in other common law jurisdictions there is no ‘right’ per se of confrontation for hearsay evidence.⁴ As the UK Supreme Court explained in Horncastle, under the common law system of trial by jury, the conditions relating to the admissibility of hearsay evidence (including exceptions to the hearsay rule), combined with the trial judge’s role as gatekeeper in applying them and her general residual discretion to exclude prejudicial or unfair evidence from going before the jury, provide the appropriate assurance in guaranteeing a fair trial.⁵ This greatly reduces the influence of this right as justification for the hearsay rule.

3.8 United States jurisprudence also suggests waning support for this rationale as justification for the hearsay rule. Even though any ‘testimonial’ evidence in respect of which there has been or can be no cross-examination is excluded, it has also been held that hearsay exceptions including dying declarations and non-testimonial evidence such as business records and statements in furtherance of a conspiracy are not excluded by the Confrontation Clause.¹ As the Supreme Court itself recognized, the right of confrontation was not absolute and exceptions to the hearsay rule are valid inroads on the right of confrontation.²

Hearsay and electronic evidence

3.9 For these reasons, we adopt the broader view taken by Stein that the rule against hearsay exists as a rule of evidence that properly allocates the risk of error. Confirmed risks that can be eliminated or avoided¹ are part of the fundamental ‘principle of maximal inferential individualization’, which Stein described as follows:

3.10 Admitting hearsay thus denies to the party adversely affected by the evidence the opportunity to test all of the inculpatory arguments that transform the evidence.¹ Hearsay evidence is excluded, not because a court could not be assured of its reliability, but that its reliability was effectively unknowable due to, among other things, the absence of an ability to cross-examine. Furthermore, in the interest of fairness to the adversely affected party, as part of the exclusionary strategy of the principle of maximal inferential individualization, such a line of inquiry should not be allowed to begin.² This is because admitting hearsay will expose the adversely affected party to the possibility of not being able to test every practical possibility of the hearsay inference to the proponent’s case, thereby introducing ‘value-preferences’ that threaten the overall coherence of the proceedings.³

3.11 Based on the discussion thus far, it could be possible to conclude that electronic evidence will not initiate the hearsay rule because records in digital form are ‘non-testimonial’ and that the hearsay rule focuses only on the ‘human-centric’ nature of statements. This would be wrong. In fact, because of the third (and some would argue fourth) industrial revolution and the digital economy, ever-increasing amounts of information are now in digital form.¹ The digital revolution has forced companies and individuals alike to record, use, transform and generate information in complex ways. ‘Our reality [is now defined] in written records, oral communications, and immersive experiences’ that are in electronic form.² Global interconnectedness and reliance on electronic information, especially to cope with the speed of change, has led to the development of increasingly complex information systems that have outpaced much of existing jurisprudence. It follows that it is necessary to fundamentally reconsider their justification and underlying rationale. For the rules of evidence, the question should be the relevance, if any, of the rule against hearsay in relation to this shift in evidence from physical form to digital form. In reality, the difficult application of the hearsay rule of exclusion to electronic evidence has been considered to be complex and confusing.³

3.12 In this regard, a useful tool for evaluating electronic evidence under the hearsay rule and for sharpening the evidential analysis is, first, to classify the type of device that is used to produce the evidence in question:

3.13 After identifying the device, the next step is to analyse the use that is made of the output of the device to determine if its use is testimonial: that is, if the output is tendered for the matter stated in the output as a fact, or otherwise.¹

3.14 Category 1 relates to the use of data processing devices primarily to store and record content that is written or spoken by one or more persons. Category 1 devices operate primarily to receive human input and store that same input for subsequent retrieval. When such information is retrieved, most of the time it is principally used for the testimonial content of the human input. As an antecedent to the storage and recording functions of these devices, this information may be recorded, digitized, compressed and processed in some way before being stored on the device or some other platform. There may be further processing or ingestion in the form of activation of the device for recording, transcribing or indexing the recording to facilitate its retrieval.

3.15 Evidence produced by Category 1 devices is therefore generally hearsay, even though it is the product of automation, because these devices are used primarily as recording devices for the evidential value of human input. Examples include lists of contacts and their information in mobile devices, text messages and email correspondence between parties – the only difference between paper-based documentation and such records is that their creation, capture and storage are done electronically. The Singapore case of Aw Kew Lim illustrates this point well.¹ In this case, the accused were charged with the copyright offence of possession of illegally copied gramophone records for sale. To prove that the accused were owners of the shop that was searched and found to have the records in question, the prosecution tendered in evidence a computer printout from the official government Registry of Companies and Businesses that identified the name of the firm, other company registration information and the names and addresses of the accused as directors. The court ruled that the information about the names of the accused was inadmissible as hearsay. As the court explained, the printout contained information that was derived from some other source such as a database or an original document, which would be inadmissible hearsay, even if that printout had been correctly certified.² To put it another way, the printout was admitted because it recorded the fact that ‘[the six accused] were directors of [company X]’. This is a hearsay use of the electronic evidence enabled by the recordation function of the device.

Electronic evidence and real evidence

3.16 Category 2 devices (records generated by software that have not had any input from a human), simply put, generate evidence that is substantially the product of automation. Digital cameras and video recorders are the ubiquitous example of Category 2 devices. Fundamentally, after some form of manual setup, these devices work autonomously by collecting light information via digital sensors and converting that via digital processing into digital form. After some additional processing, depending on the specifications and features of these cameras and recorders, the digitized images or video streams are stored or recorded. When the images or videos are retrieved, they are used to replicate the input or environment within which these devices are situated.

3.17 English courts appear relatively sanguine in admitting ANPR (automatic number plate recognition) evidence, without raising any noted hearsay challenges.¹ The primary reason for the absence of challenges is that, typically, photographs and video recordings, such as ANPR evidence made from Category 2 devices, are not used ‘testimonially’: they are considered real evidence. Real evidence is ‘an independent species of evidence as [its] production calls upon the court to reach conclusions on the basis of its own perception, and not on that of witnesses directly or indirectly reported to it’.² The genesis for its modern application to electronic evidence may be found in Re The Statue of Liberty, where the issue was whether the cinematograph film recordings of a navigational traffic radar system were admissible to prove the movements of the vessels that had collided. Sir Jocelyn Simon P held that they were. His Lordship opined, at 195 (emphasis added):

3.18 It is also for this reason that the digital audio recorder is an example of either a Category 1 or a Category 2 device, depending on its use. From a technical standpoint, the analogue sound waves produced by human speech or sounds are captured by the microphones on such recorders, converted into digital form by the use of digital signal processing techniques and then further processed (for instance, compressed, tagged or indexed), before the digitized representation of sound is stored or recorded. Whether the recording made by a digital audio recorder is hearsay will depend on whether it is used as evidence testimonially. For instance, if Amazon Echo recordings of conversations were used as evidence of the incidents that allegedly took place in households being investigated, the recording would be an experiential reproduction of the context of the offences allegedly being committed.¹ In other words, it is used as a surveillance tool – a Category 2 device. On the other hand, if the Echo is used as a dictation tool to record the user as saying ‘X exists’, such a use as evidence that ‘X exists’ would be testimonial and would be excluded as hearsay. In those circumstances, it would be a Category 1 device.

3.19 But while real evidence from Category 2 devices does not amount to ‘assertions’ that are caught by the hearsay rule, this does not mean that such evidence is automatically admissible. All relevant evidence is generally considered admissible only if a proper foundation has been laid for its admission: this will be by way of satisfying the authentication requirements of the laws of evidence. A more detailed discussion of the authentication evidence in general and specific challenges to the reliability and accuracy of electronic evidence can be found in Chapter 6.

3.20 In many circumstances, the proponent of the electronic evidence, usually the prosecutor, would, in the absence of foundation evidence, seek to circumvent the requirement for foundation evidence by relying on the doctrine of omnia praesumuntur rite esse acta or presumption of reliability. However, this presumption is not a suitable substitute for authentication, particularly in relation to complex and sui generis devices and systems. It is often ill-appreciated that foundation evidence must also be sought before there can be reliance on the presumption itself. And the scope and effect of the presumption has often been misunderstood. A more detailed discussion of the presumption of reliability as applied to devices that produce electronic evidence can be found in Chapter 5.

Testimonial and non-testimonial use of information

3.21 A large proportion of electronic evidence, however, will be evidence produced by Category 3 devices (records comprising a mix of human input and calculations generated by software). In this category, the device functions through a mix of human-supplied input and produces data output without further human intervention. The human input could be in the form of the initial calibration of such devices or their internal programming. These are ex facie testimonial input of some form in the absence of further elucidation. Even machine learning systems – the epitome of autonomous systems – have varying levels of reliance and dependence on human-labelled data, choices and configurations to enable their proper autonomous operations.

3.22 The line between Category 2 and Category 3 devices can be hard to draw: the difference is one of degree that represents the relative significance of the level of contribution of human-supplied input and pre-programmed autonomous processes to the eventual output, and the extent to which the testimonial input qualifies the (purportedly) non-testimonial use of the output. In Wood, for instance, the admissibility of chromatograph and spectrometer outputs as real evidence was supported by the oral testimony of the chemists who calibrated and operated the machines, as well as the programmers of the program who analysed the chromatograph and spectrometer outputs, leading the courts to characterize the processes of recording, processing and calculating the testing sample as operating without human intervention. The Court of Appeal in Wood, in characterizing the test results as real evidence, said:

3.23 It could be argued that the phrase ‘it did not contribute to its own knowledge’ (as noted above) is a misleading representation of the testimonial input of the chemists and programmers. When programmers set up instructions (or code) in devices, it is these instructions which are followed by the devices to process the input received. These instructions amount to conditional statements embedded in such devices.¹ But it is clear that the effect of the supporting testimonies of the chemists and programmers is that it is only the non-testimonial parts of the chromatograph and spectrometer outputs – the analysis results – that are being relied on by the courts. In contrast, in Mehesz v Redman² the electronic evidence of blood alcohol concentration was produced by a chromatograph that had to be calibrated using a pre-prepared standard. The chromatograph was in turn coupled to a data analyser which had to be manually configured by following an instruction manual prior to its use to produce the concentration analysis. Only one of the two analysts who ran the tests testified. He also admitted to not following the instruction manual and making changes to the configuration of the analyser. Bereft of the testimony of the second analyst who ran the second test (two tests were run and the results averaged for the final concentration report), the testimony of the programmer of the data analyser, as well as the instruction manual, the court upheld the objection as to the admissibility of the final printout from the analyser. Zelling J of the South Australian Supreme Court pointed out that because the calibration depended on a standard prepared by somebody, the final result was the average of two analyses and the second person was not called, and because the instructions in the manual were not followed, ‘there was no overall evidence called from an expert as to the trustworthiness of the computer itself … under these circumstances the objection as to hearsay should have been sustained’.³ Likewise, in Holt v Auckland City Council,⁴ which concerned a similar setup of a chromatograph with a data analyser (and almost similar lack of evidence), the court held that the final output was ‘inadmissible as incorporating hearsay data outside the field of [the analyst’s] proven competence’.⁵

3.24 There is much to commend in the detailed analyses undertaken in Mehesz and Holt. They attempt to resolve the different components of the electronic evidence into its testimonial and non-testimonial components, identify the programmers as declarants responsible for each of the components, require each declarant to explain how she was responsible for her respective testimonial component, and justify the results that arise from each non-testimonial component. This analysis approach more accurately reflects the reality that modern-day devices and computers are complex systems¹ subject to inevitable faults and errors,² clearly part of the ‘heightened reliability and testability’ regime proposed by Tepper,³ and consonant with the principle of maximal inferential individualization advanced by Stein.⁴ For instance, if the contents of text messages and email correspondence are to be used testimonially and constitute hearsay, some metadata associated with such messages and correspondence, such as the date and time of transmission and receipt and the routing paths, would be generated by Category 2 devices without human intervention. Such metadata is produced as a consequence of software code, and subject to the tests of authentication, reliability and testability, it could constitute real evidence and be independently admissible.

3.25 But we caution against treating the test of reliability as the ultimate aim for this hybridized hearsay-authentication analysis of electronic evidence. The concept of ‘reliability’ is a value-loaded concept that varies depending on, among other things, the specifications of the device, the purposes to which the device has been put and the underlying social concepts that it seeks to protect.¹ For instance, it is well known that devices that operate on neural networks have statistically modelled accuracy, precision and recall measurements of reliability. It is also well known that there are two estimates of such parameters: using the data on hand and using unavailable data. The latter obviously cannot be done (‘you do not know what you do not know’) and has to be an estimate. This explains why a problem or situation that occurs only outside normal operating parameters (called ‘corner cases’), or errors that occur in open environments which subject the device to situations or environments outside normal operating parameters, are so difficult to ascertain. That is not to say that it is impossible to ascertain the ‘reliability’ of devices and systems: we contend that unexplained and unjustified instances where the device or system is not operating as intended is certainly evidence of absence of reliability.

3.26 Therefore, there is much to be said for treating computing systems and devices as ‘the witness’¹ in proceedings. Just as a human witness will be subject to an examination as to his or her experience and qualifications, subjecting the output of a device to the scrutiny of the hearsay rule helps to tease out the embedded human assertions from the results sought to be admitted in evidence – be it the code or its data.² If there is no opportunity for the human assertions to be tested – for instance, if the automatically produced analysis is to be relied on but the programmer who wrote the software that generated the analysis is not called to testify – the analysis becomes hearsay.³ Given that the product of devices will inevitably be based on a multiplicity of, and interplay between, direct and indirect human assertions, not all of which have been validated, let alone completely assessed for their accuracy and correctness,⁴ it will be near impossible to call all contributors of these assertions to give evidence in legal proceedings. Therefore, considering that these models entrench various human assertions and even biases, it is more apt to proceed with caution and subject such electronic evidence to closer scrutiny for the ‘human input’ by placing it into the hybrid Category 3. Of course, this closer scrutiny can be further assisted with a robust approach to authentication of such evidence and to a more effective stance regarding disclosure.

Implied assertions

3.27 Since the decision of Wright v Doe d Tatham,¹ the law in England and Wales has been that implied assertions from testimonial statements are inadmissible under the rule against hearsay. Although difficult issues may arise in relation to distinguishing implied assertions from the use of such statements as grounding a relevant inference (which falls outside the hearsay rule),² this position at common law was later affirmed by a majority of the House of Lords in R v Kearley,³ who held inadmissible, as being irrelevant, calls made to the defendant’s number asking to be supplied with drugs as proof that the defendant had intent to supply drugs, or alternatively, inadmissible as hearsay.

3.28 In the context of electronic evidence, the circumstances in which the strictness of the rule against implied assertions at common law could apply are legion. In the digital economy, electronic records proliferate with an abundance of unique identification numbers that identify records and documents (e.g. hash numbers, login IDs, registration numbers), labels that identify the nature, origin, provenance or ownership of goods, services or other records (e.g. barcodes and QR codes on product packaging), dates of records and documents and the relationships or identities of persons (e.g. organizational unit attributes and organizational codes used for grouping accounts and corporate entities). In the controversial case of Myers (James William) v DPP,¹ the manufacturer’s records, made routinely in the course of production, that identified the cars in question as stolen cars because they recorded identification numbers indelibly stamped inside the cylinder blocks, which corresponded to the numbers found in the cylinder blocks of the cars sold by the defendant, were held to be implied assertions and inadmissible as hearsay. A majority of the House of Lords reasoned that the cogency of the records, maintained on microfilm, depended on hearsay, as each witness could not prove that the records were correct or that the numbers which the records contained were in fact the numbers on the car cylinder when it was made by the unknown workmen.² While the records in question in Myers were microfilm records, even if they were substituted with electronic records, the analysis at common law would remain the same.

3.29 The House of Lords was not unaware of the controversial nature of their decision, but noted that, at common law, the reliability and trustworthiness of records is not an exception to the hearsay rule,¹ and urged legislative reform.² However, the statutory treatment of the rule against hearsay is different, depending on whether electronic evidence is adduced in civil proceedings or in criminal proceedings. It is to the statutory rules against hearsay in civil and criminal proceedings and the complex and unique issues that arise in the treatment of electronic evidence that we now turn.³

Civil proceedings and the requirement to give notice

3.30 In England and Wales, the hearsay rule was abolished for civil proceedings by s 1(1) of the Civil Evidence Act 1995. The Act applies to all civil proceedings,¹ including proceedings in the magistrates’ court.² A party that intends to adduce hearsay evidence in civil proceedings is required to give the other party or parties notice of his intention and, should it be requested, particulars of the evidence.³ This requirement to give notice is not unique to England and Wales. A criticism of hearsay evidence said to justify the existence of a rule of exclusion is that admission of hearsay would amount to an unjustified element of surprise causing delay and unwarranted disruption in a proceeding.⁴ This criticism has largely been addressed through the power given to any other party to the proceedings, with leave of the court, to call as a witness the person whose hearsay evidence is relied on by the proponent but who is not called as a witness by the proponent, and cross-examine him on the statement.⁵ It is noteworthy that pursuant to the notice, hearsay of any degree may be admitted.⁶ This is especially relevant in relation to electronic records made in the course of manufacturing or production by various workers, who may or may not be identified, within the organizational hierarchy or chain of responsibility. However, the court may have regard to, among other things, whether the evidence involves multiple hearsay in estimating the weight to be given to such hearsay evidence.⁷

3.31 The Act includes a number of exceptions to the hearsay rule that are particularly relevant to documents stored in digital form. Published works dealing with matters of a public nature, public documents and public records are all admissible under the provisions of s 7(2) of the Civil Evidence Act 1995. More distinctively, where a document can be shown to be part of the records of a business or public authority, the document can be received into evidence in civil proceedings without further proof in accordance with s 9, subject to certification procedures¹ that may be dispensed by the court.² Subject to the threshold requirement that these records must be part of a ‘business’, which includes any activity regularly carried on over a period of time,³ the wording of this and similar provisions in other jurisdictions⁴ means that the form a technology takes (‘records’ means ‘records in whatever form’)⁵ will not prevent the admission into evidence of data stored in digital form.

3.32 Finally, it is noteworthy that s 9(3) of the Civil Evidence Act enables the absence of an entry in the records of a business or public authority to be proved by affidavit of the relevant officer of that business or authority. This addresses the problem of ‘negative hearsay’, in which evidential significance is attributed to the absence of the requisite records of a business or public authority. Of course, to enable the requisite conclusions to be drawn, especially in relation to electronic records or electronic databases, the necessary foundation evidence as to, among other things, good record-keeping practices must first be established.

Criminal proceedings

3.33 In England and Wales, the enactment of the Criminal Justice Act 2003 repealed the provisions relating to hearsay in the Criminal Justice Act 1988, and by doing so, abrogated most of the common law of hearsay,¹ substituting in its place a statutory regime for admitting hearsay² and multiple hearsay.³ The operative provision is s 114(1), which reads:

3.34 The provisions of s 114 serve as an introductory provision to the other provisions in that chapter. The operative scope of s 114 and the other provisions lies in the definition of a ‘statement’ as a representation of fact or opinion made by a person by whatever means, not made in oral evidence in the proceedings, but sought to be admissible as evidence of any matter stated.¹ A matter stated is something which the maker of the statement intended someone (generally the recipient) to believe or to act upon, or to cause a machine to operate.² Thus defined, s 114 retains the default exclusion of the hearsay rule in relation to statements intended to be acted on (express assertions), and operates to admit them in criminal proceedings within the parameters set out in s 114(1)(a)–(d) (although a number of common law exceptions are retained in statutory form by virtue of s 118(1) Criminal Justice Act 2003).

3.35 In contrast, statements that are not intended to be acted on (implied assertions) fall outside the definition of a ‘statement’ as defined¹ and are admissible pursuant to the abolition of the common law rules governing the admissibility of hearsay evidence in criminal proceedings.² This has the effect of reversing the common law position in R v Kearley,³ but with the proliferation of records of electronic communications and the ubiquity of their use, it has in turn opened up new considerations regarding the scope of the rule as regards implied assertions and their relevance to the issues.

Telephone calls and messages

3.36 We start with the cases that deal with the inclusion of evidence of telephone calls and text messages sent on mobile telephones, especially in relation to cases involving illegal drugs, that have caused some confusion. For instance, in R v Chrysostomou (Mark)¹ the trial judge admitted four text messages, apparently sent to the appellant by someone called ‘John’ who attempted to set up a supply of drugs, as evidence that the appellant was a dealer in drugs. In giving judgment for the court, Aikens LJ agreed that the text messages were not caught by the statutory code on hearsay in the Criminal Justice Act on the basis that the messages were adduced, not to prove, as fact, any matters stated in the messages, but ‘as evidence of an underlying state of affairs, which was the basis on which “John” apparently sent the texts to the appellant, namely that the appellant dealt with drugs and so could meet John’s demands’.² In his commentary, Professor Ormerod agreed with the conclusion reached by Aikens LJ but disagreed with the reasoning, pointing out that the text messages were actually relied upon for the truth of the implied assertion contained in the message that the accused was a dealer in illegal drugs. This, however, did not render the message hearsay because, as Professor Ormerod noted,³ for a statement to be hearsay, the purpose of making the statement must be to cause another to believe the matter or to act on the matter stated,⁴ while ‘the purpose of the texter [“John”] was not to cause [the appellant] C to believe/act on his being a dealer’.⁵

3.37 In contrast, in R. v Leonard (Mark Alan) the members of the Court of Appeal (Criminal Division) determined that two text messages sent by unknown people to the appellant on two separate mobile telephones were hearsay evidence, and should not have been admitted at trial in support of the case for the Crown that the appellant was a drug dealer (with the prosecution conceding that the evidence would not have been admitted following a consideration of all the factors set out in s 114(2)). The messages were set out as follows:

3.38 It was assumed that the content described feedback on the quality of the drugs purported to have been supplied (the first one positive and the second one negative). Professor Ormerod considered the decision by the Court of Appeal to be incorrect because the Crown did not rely on the content of the text messages for the truth of whether the quality was good or bad, or the nature of what had been supplied. The issue was whether the appellant was the supplier of a controlled drug, not the quality of the drugs supplied, which was irrelevant.¹ This must be undoubtedly correct, and the argument illustrates the absurdity of the largely arbitrary line between hearsay and non-hearsay statements.

3.39 The Court of Appeal considered an entire set of electronic messages and their use in inculpating the defendants in R. v Twist (Andrew Terence),¹ where the issue was the admissibility of text messages sent over mobile telephones as evidence against various defendants for a variety of offences. In determining the general approach to take as to whether the hearsay rules in the Criminal Justice Act applied, Hughes LJ set out the following approach:

3.40 Hughes LJ went on, at [18], to indicate that the ‘answers to these questions will be case-sensitive. The same communication may sometimes be hearsay and sometimes not, depending on the matter for which it is relied upon and the fact which it is sought to prove.’¹ While correct, this line of argument emphasizes the largely arbitrary nature of the distinction. A text message commenting on the quality of drugs bought will not be a hearsay statement² (even if they could amount to an implied assertion that the recipient was a drug supplier, the purpose of the sender(s) did not include causing him or anyone else to believe that he was³) and can be adduced in support of a contention that the recipient actually sold drugs. However, a statement to the effect of ‘thanks for selling me those drugs’ will be inadmissible hearsay because it contains a statement that the recipient had sold drugs. An argument might be made that what was sought to be established was the state of mind of the maker of the message, not whether drugs were actually sold by the recipient of the message. Depending on the issue to be proved, this argument may render the statement admissible.

3.41 The tests set out in R. v Twist (Andrew Terence) were considered in the case of R. v Midmore (Billy Nathan),¹ where a WhatsApp message containing an image of a box of sulphuric acid with the caption ‘This is the one face melter’ was considered by the trial judge to be hearsay. The members of the Court of Appeal determined it was not hearsay. One commentator considered this analysis to be unsound,² because in relation to the second question in Twist (the question of relevance of the statement), it does not appear that there was a statement of the matter sought to be proved. McKeown suggested that the image, taken with the caption:

3.42 McKeown suggested that ‘[S]‌ince implied representations may now potentially be hearsay, a question arises as to whether the Twist formula should continue to be followed’¹ and that another approach to identifying hearsay could be as follows:

3.43 To conclude that anything inferred from a statement is not hearsay, whereas anything directly stated is, is to establish a distinction that dances on the head of a pin. The better approach is to treat all types of assertion – express or intended and implied or unintended – as prima facie hearsay and leave their admission to the judge on the basis of an analysis of a list of balancing criteria.

3.44 With the abolition of implied or unintended assertions from the scope of the hearsay rule, not all assertions made with the intention to communicate will be qualifying hearsay statements. The inadmissible hearsay assertion has to be associated with the object for which it is tendered in evidence in support, failing which it could be admissible as an implied or unintended assertion. To amplify McKeown’s point, to enable the admission of implied representations as to the supposed mental state of the maker of the statement presupposes the relevance of such mental state in the first place. If sanctioned, this could allow for all statements to be classified as implied statements and bypass the hearsay exclusion against express statements. Indeed this was the concern that was well recognized at common law.¹ In the same vein, Professor Hirst observed that if there is nothing to prove an established relationship, or an incriminating response or reaction from the defendant, the assertion may be inadmissible, regardless of whether it is hearsay or not.²

3.45 This illustrates the fundamental weakness of the rule. By excluding unintended assertions, there is a possibility that arbitrary limits may arise in that the difference between a hearsay statement and a non-hearsay statement will rest with the question of whether there exists an intention to communicate. The existence of an intention to communicate is of such little value as to render the distinction meaningless. Furthermore, such a distinction exposes the application of the exclusionary rule to the formulation of a clever submission of a lawyer in that the application of the rule might be avoided by classifying the statement as a reflection of the mindset of the maker as opposed to an intention of the maker, an approach made plausible given the ubiquity and accessibility of electronic devices. In such a case, no real distinguishing factor truly exists.

Representations other than by a person

3.46 Of particular relevance to electronic evidence is s 129 Criminal Justice Act 2003. It reads:

3.47 The UK Law Commission considered the admissibility of a computer printout, whether it is hearsay, and whether the printout itself is relevant:

3.48 The effect of s 129 is to treat not as hearsay and as admissible representations of any fact made in a document emanating from a device that is adduced to prove something other than the truth of a statement previously entered into them by a human being.¹ They are admissible subject to proof of the accuracy of any information supplied by a human to such a device,² including its proper setup and calibration.³ But as previously noted, the proper treatment of such representations of facts by devices or machines requires a proper categorization of the function of such devices, and an identification and subsequent characterization of the different parts of the representation, as to whether they emanate from the device or are made by a person or persons.

3.49 This may be illustrated by the Australian case of Hansen Beverage Company v Bickfords (Australia) Pty Ltd.¹ In this case, working television sets in homes were monitored by a meter system that recorded that a person was physically located in the home when he registered his presence by pressing a button when a television was on. This was for the purposes of establishing the size of the audience that might be watching a particular programme. That the evidence was produced on a printout and was automatically recorded by software was not at issue. It was argued that the evidence was not hearsay as there was no previous representation made by a person, as required by s 59(1) of the Australian Evidence Act 1995. Middleton J, it is suggested correctly, identified the evidence as hearsay because it was a representation of fact that a certain number of people clicked on the buttons. The judge commented:

Body-worn camera footage

3.50 Body-worn cameras have rapidly become a normal feature of policing.¹ In McGuinness v The Public Prosecution Service for Northern Ireland,² on appeal by way of case stated from the County Court for the Division of Londonderry, the question arose whether the trial judge was correct to admit the recording from a body-worn video camera, which documented allegations made by the complainant of assault by the appellant, under the hearsay provisions of the Criminal Justice (Evidence) (Northern Ireland) Order 2004, in particular under articles 18(1)(b) and 22(1)(4)(a). Although the complainant withdrew her complaint, the appellant was charged and tried of the offence. The complainant was not called to give evidence, because she was reconciled to the appellant.

3.51 In reaching their decision, the members of the Court of Appeal in Northern Ireland considered the tests set out by Lord Ackner in R. v Andrews (Donald Joseph)¹ under the doctrine of res gestae exception to the hearsay rule where a conversation took place in such circumstances that the possibility of concoction or distortion by the victim could be disregarded:

3.52 In giving the judgment of the court, Weatherup LJ said, at [36]:

3.53 The court noted, at [41], that the prosecution relied on the res gestae exception in order to provide support to the complainant in the changed circumstances brought about by the reconciliation of the parties, while seeking to deal with the alleged previous conduct of the appellant. The prosecution had to balance the competing interests when deciding to prosecute the appellant. The trial judge was correct to admit the recording under the statutory provisions.

3.54 The res gestae exception at common law has been expressly preserved in the Criminal Justice Act 2003.¹

3.55 The case of DPP v Young¹ has a similar set of circumstances. The complainant was not able to appear to give evidence. The prosecution sought the admission of an audio recording of a 999 call and a recording from a body-worn video camera from one of the police officers attending the scene. Both items of evidence overlapped, because the telephone line on which the 999 call was made remained open up to the time the police officers attended the scene, and the telephone recording contained identical evidence to that in the body-worn video camera recording. Both recordings were admitted as real evidence. No consideration was given to the tests set out in R. v Andrews (Donald Joseph). The defence submitted that there was no case to answer, given, it was argued, that the evidence was tenuous in nature, weak and vague. The lay magistrates accepted the submission and dismissed the case against the defendant. The Divisional Court quashed the decision of the justices. Lord Justice Holroyde, in giving the judgment, indicated, at [19]:

Business and other documents

3.56 In this regard, it is useful to start with a review of the cases that considered s 24 of the Criminal Justice Act 1988. This provision, concerning the admission in criminal proceedings of business and other documents, was the predecessor provision to s 117 of the Criminal Justice Act 2003. In Brown v Secretary of State for Social Security,¹ the Secretary of State adduced evidence of statements from computer records by way of two witnesses where the identity of the persons who supplied the information could not be established. It was submitted on behalf of the appellant that the two statements were inadmissible because they did not comply with the terms of s 24. Section 24 was written to enable business documents to be admissible without the need to call the maker where the documents formed part of records about which the maker could not be expected to know anything in detail, and which were created in the course of trade or business. The members of the Divisional Court, Balcombe LJ and Collins J, agreed that the statements were not admissible under s 24(4) of the Criminal Justice Act 1988 ‘as there was no evidence that it was impossible that the makers of the statements would have no recollection of the matters referred to in their statements’.² In comparison, the members of the Court of Appeal (Criminal Division) in the case of R. v Derodra (Kishor),³ rightly it is suggested, admitted the contents of a police ‘CRIS’ report, which was a computerized record of incidents of crime under s 24. In this instance, the person who reported the crime to the police – the lodger of the appellant – could not be found to give evidence of his complaint. It was the statement of the lodger that was to be relied upon testimonially, not that of the police officer who made the relevant entry.⁴

3.57 In Vehicle and Operator Services Agency v George Jenkins Transport Limited,¹ the prosecution had to prove that certain commercial drivers had failed to properly record their journeys with the tachographs in their vehicles, and had worked beyond the number of hours that were permitted without the prescribed rest periods or breaks. To discharge this burden, the prosecution sought to put in evidence a number of drivers’ time sheets pursuant to s 24. On a preliminary point, the trial judge ruled them inadmissible and dismissed all charges against the defendants. The prosecutor appealed, and the appeal raised a number of issues regarding the interpretation of these provisions. First, the provisions in s 24, described by Mackay J at [10] as ‘criteria or gateway’ provisions,² must be satisfied before the second issue is addressed, that is whether the documents in question can be admitted in evidence. Mackay J quoted³ from the judgment of Roch LJ in R. v Foxley (Gordon):

3.58 In this instance, Mackay J and Kennedy LJ agreed that the documents satisfied the criteria provisions, and were admissible and self-proving in evidence.¹ Kennedy LJ also noted the criticisms that Professor Smith made of the decision in R. v Foxley (Gordon), although it was observed that a further analysis of another case² by Professor Smith was capable of applying to the case in hand if it was adjusted slightly.³

3.59 Section 24 is succeeded by s 117 of the Criminal Justice Act 2003. Section 117(1) to (5) reads:

3.60 The provisions of s 117, dealing with the business document exception, are very wide and permit the admission into evidence of multiple hearsay,¹ although the various foundational conditions set out in s 117 must be satisfied. In R. v Humphris (Andrew James),² the Crown sought to adduce evidence of the appellant’s previous convictions under s 117. For that purpose, they relied on a statement of Police Officer Grimes, who retrieved relevant records from the Essex Police computer facility, the contents of which were in turn derived from staff of the Essex Police Force, who acted under a duty to record information and who either had or may reasonably be supposed to have had personal knowledge of the matters dealt with in the records. These records were attached to Police Officer Grimes’ statement. Section 117 provides certain conditions that must be fulfilled before evidence can be admitted. The defence accepted that the provisions of s 117(2)(a) were complied with, but argued that for each record of the appellant’s previous conviction, s 117(2)(b) required the statement to have been obtained from each complainant as the relevant person, rather than the police officer who recorded the information. Although Lord Woolfe upheld the conviction of the appellant, he agreed and held that the necessary foundations for the admissibility of the evidence were not properly laid.

3.61 Where a document is put in under the provisions of s 114 and s 117, care must be taken over any content that is hearsay.¹ In addition, the trial judge must ensure that the members of the jury understand the purpose of admitting the document. In R. v Horncastle (Michael Christopher),² there was an email statement made by an ISP which identified the appellant and his address as being the possible holder of an email account suspected to have been used to send abusive images of children. The ISP acknowledged that this information could have been supplied by the email account holder impersonating the appellant. The prosecution adduced this email to show the address of the place (the appellant’s home) where the police raid took place, but not to prove the fact that the account was that of the appellant or used by the appellant. (In fact, no evidence of abusive images of children was found on the appellant’s computer, although there was evidence that the appellant’s lodger had used the email account.) No directions were given by the trial judge as to the limited purpose for which the ISP’s email was adduced. On appeal, Thomas LJ held that the judge’s failure to explain the use was a material misdirection, as the jury could have used the ISP’s email to link the appellant to the email account. The appellant’s appeal was allowed and his conviction was set aside.

Judicial discretion to include hearsay

3.62 Notwithstanding the other routes of admissibility in s 114(1), one particularly wide¹ route is to admit hearsay evidence ‘in the interests of justice’ under s 114(1)(d),² subject only to the conditions in s 114(2).

3.63 Careful consideration needs to be made of the provisions of s 114(2) regarding evidence in digital form when it is obtained from the Internet and where the evidence relating to the material, such as its authorship and ownership of the website from which it originates, is not known, as in the case of R. v Bucknor (Ashley Dwayne).¹ In this case, the trial judge admitted evidence found by the police on BEBO (https://www.bebo.com), consisting of 46 separate website ‘pages’. The material included a number of photographs of Bucknor that he had taken of himself after he had left prison. The photographs had been placed on the page by someone in such a manner as to portray Bucknor as a member of the Organised Criminals (OC) gang. There was a hyperlink to a YouTube page that portrayed the OC gang as violent. The YouTube page was also shown to the jury in the form of a DVD. The prosecution did not have any evidence of the IP address from which the material was uploaded. The trial judge admitted the evidence as part of the background to the case, but on appeal the appellant argued that the judge failed to give a sufficient direction regarding the ownership of the website in question. The members of the Court of Appeal agreed with the submission. The material was clearly hearsay because it seemed likely that the maker as the source of the material was representing as fact or opinion that Bucknor was a member of the OC gang. In considering the issues set out in s 114(2), Hooper LJ, giving the judgment for the court, said that the judge ought to have considered the reliability of the maker of the statement (sub-paragraph (e)), whom the judge failed to identify.² Failing to identify the maker meant that it was not obvious how many levels of hearsay were involved. The judge also failed to consider the reliability of the statement that the appellant was a member of the OC. Hooper LJ concluded:

Judicial discretion to exclude hearsay

3.64 A trial judge also has the ability to refuse to admit a statement in accordance with s 126(1)(b) where ‘the court is satisfied that the case for excluding the statement, taking account of the danger that to admit it would result in undue waste of time, substantially outweighs the case for admitting it, taking account of the value of the evidence’.¹

3.65 The Law Commission described the necessity of this judicial discretion as the power to exclude superfluous hearsay evidence. The concern is not with evidence that is wholly irrelevant, but evidence which has some relevance, yet ‘where the probative value of the evidence is so slight that almost nothing is gained by admitting it’.¹ Although this discretion is intended to be exercised only in exceptional cases, the existence of the discretion serves as an important counterweight to address concerns about the expansionary rules for the admission of barely relevant hearsay evidence.²

Concluding observations

3.66 Almost everybody now uses digital data, whether this interaction is by way of the ether – a terminal linked by software to a server located in an unknown location – or from a physical device. Software code has become part of the everyday fabric of the majority of people. This means we are all, wittingly or unwittingly, assessing electronic evidence every day: from whether to trust that incoming email from an unknown source, to dealing with the veracity of content from networking sites.

3.67 The digital world is now awash with evidence: direct statements over the Internet; communications between telephones and other devices; messages made by a known author, anonymously or by somebody who cannot be traced. Every day we are dealing with the multiplicity of direct and indirect assertions (whether factually accurate or not), in the form of statements by one person or relayed, correctly or incorrectly, by others, and the interplay between them and the reality of the physical world. For the first time, we are all assessing evidence every day.

3.68 This chapter seeks to demonstrate the importance for lawyers to be aware of the dangers of admitting hearsay evidence and the need to distinguish between the various hearsay representations and ‘representations’ made by devices that may be embedded in electronic evidence. The identification of the nature of the device that stored or produced the electronic evidence and the testimonial or non-testimonial use that is made of the evidence remain important steps for determining the admissibility and relevance of the evidence, even in a statutory regime for admitting hearsay. It is only with a careful understanding of the nuances of both the hearsay rules as well as the nature of electronic evidence that the principle of testing the evidence can be observed, to ensure that any rational inference from evidence admitted against the defendant is objective and fair, and that the law, in its quest to establish the truth of what is contained in the statement, has indeed been true to itself.