Voice identification and ear-witnessing in the dock

• Jurisdiction: South Africa
• Date: 16 August 2019
• Author: Lirieka Meintjes-van der Walt
• Published date: 16 August 2019
• Pages: 391-409

Voice identication and ear-

witnessing in the dock

LIRIEKA MEINTJES-VAN DER WALT*

ABSTRACT

This art icle denes ear-witness identi cation and discuss es voice

identication as a mechan ism for legal fact-nding. It is contended that

voice identication evidence proffered by a lay person i s highly problematic.

While the gu idelines for the treatment of eyewit ness testimony are clearly

established in a number of ju risdictions, t he treatment of ear-witness

testimony is somewhat less regu lated and specic issues pe rtaini ng to ear-

witness evidence are investigated i n this article. A recent case, S v Mahlungu

2018 (2) SACR 64 (GP), which dealt with non-technical voice identi cation, is

discussed and the a rticle suggests th at it might have been pertinent for t he

court to have taken cognisa nce of the guidelines for ear-wit ness evidence

set out by Lord Justice Cage in R v Flynn and St John [20 08] Crim LR 799. The

article interrogate s factors pertinent to e ar-witness evidence such as voice

identication parades; fam iliarit y; time delay between the hear ing of the

voice during the commission of t he crime and identic ation; interviewing

ear-witnesses; suggestibi lity and falli bility. The article concludes th at as

voice identication in the legal sphere is s till in its infancy, establishing best

practice standard s for conducting voice line-ups is needed. It is furthe rmore

contended that ear-witness evid ence is generally weak evidence and should

be treated with great caut ion and that this complex matter shou ld not be

considered on the basis of int uition, common sense and accepted b eliefs

but that decisions should be in formed by empirical resear ch. Finally, the

article briey refers to the N AS Report and the PCAST Report bot h of which

strongly warned again st scientically u nsubstantiated forensic evide nce

being accepted in court.

1 Introduction

Ear-witness identication evidence can be dened a s:

the process of a witness hearing the voice of a target person or persons,

retaining that information in memory, retrieving that information later when

called to identify the suspect(s) either in a 1-person voice line-up or a many-

person voice line-up, and nally, testifying or communicating this decision to

a police investigator, trial judge, or jury.1

* BJuris LLB (UPE); LL M (Rhodes); DJuris (Rijksuniver siteit Leiden), Adjunct Professor,

Law, Science and Justice Research Niche A rea Leader, Universit y of Fort Hare,

Advocate of the High Cour t of South Africa.

1 AD Yarmey ‘Earwi tness speaker identi cation’ (1995) 1 Psych’y, Public Pol’y, L 792

at 795.

391

(2018) 31 SACJ 391

© Juta and Company (Pty) Ltd

‘Ear-witness identication’ is “otherwise known as ‘lay’ or ‘naïve’ speaker

identication’.2

Voice identication could be of benet in cases such as where the

perpetrator is not recognisable, where the victi m’s ability to see,

is consciously obstructed by the perp etrator, or where telephonic

communicatio n is involved.

Individual voices differ in import ant ways (not including contextual

differences and speech disorders) that could be used to identif y a

speaker and what is being said. Each human being has a voice that

is distinct and di fferent from everyone else’s. It therefore seems to

be intuitive that a person should be able, relatively easily, to identify

another person on the basis of his or her voice. The following factors

could affect voice identication: i) vocal quality (timbre and clarity)

ii) dynamics and style (intonation, rhythm, st ress, timing and speed)

iii) accent and dialect iv) loudness (intensity) and pitch (frequency)

and v) vocabulary (formal, regional and slang).3 The following three

processes inform human memor y: encoding, storage and retrieval.4

Encoding refers to the process whereby the brain captures a nd converts

information.5 Storage refers to the process whereby the information is

kept in the brain for potential utilis ation at a later stage.6 Retrieval

refers to the process when the stored information is recal led in order

to be utilised .7 These three processes are interlocked. As a result of

this interrelatedness t he encoding of new information is facilitated by

previously stored knowledge. The retrieval process in tur n, is affected

by the nature and the qualit y of the process of indexing when the

information was stored in the brai n.8

Non-technical speaker identication refers to the abil ity of almost all

humans to identify a speaker simply by listen ing to his or her voice.

If a person receives a telephone call from another person well-known

to the receiver, it is often relatively easy to recognise the voice of the

caller even when the caller does not identify h imself or herself.

There is a common perception that human bei ngs are more

competent at identifying voices than is actua lly the case. The reasons

why the quality of ear-witness evidence is of ten inadequate are not

2 D Watt ‘The identicat ion of the individual through sp eech’ ch 8, available at http: //

www-users.york.ac.uk/ ~dw539/watt2009.pdf, accessed on 22 October 2018.

3 AK Dunn ‘Unders tanding ear wit ness testimony’ (2013) Psych’y Rev, available at

http://irep.ntu.ac.uk/ id/epr int/18508/1/217444_PubSub729_Dunn.pdf, accessed on

5 October 2018.

4 D Reisberg Cognition: Ex ploring the Science of the Mind 6ed (2015) 187.

5 Ibid.

6 Ibid.

7 Ibid.

8 Ibid.

392 SACJ . (2018) 3

© Juta and Company (Pty) Ltd