Paper presented at the Second Pan Paper presented at the Second Pan- - - PowerPoint PPT Presentation

Witnesses’/Victim’s Recognition of Once-Heard Voices g Professor Ray Bull, Forensic Psychology, University of Leicester, UK

Paper presented at the Second Pan Paper presented at the Second Pan-

• American/Iberian Meeting on Acoustics

American/Iberian Meeting on Acoustics Organised by the Acoustical Society of Organised by the Acoustical Society of g y y g y y America America Cancun, Mexico November 2010 Cancun, Mexico November 2010

3aSC8 Special Session on Forensic Voice Comparison and Forensic Acoustics @ 2nd Pan-American/Iberian Meeting on Acoustics, Cancún, México, 15–19 November, 2010 http://cancun2010.forensic-voice-comparison.net

1.

In 1974 in response to media focus on some criminal cases In 1974 in response to media focus on some criminal cases involving what turned out to be false convictions based on involving what turned out to be false convictions based on mistaken witness testimony, the Government in England mistaken witness testimony, the Government in England d W l t ffi i l C itt f I i th t d W l t ffi i l C itt f I i th t and Wales set up an official Committee of Inquiry that and Wales set up an official Committee of Inquiry that sought to better understand how honest witnesses can sought to better understand how honest witnesses can sometimes give incorrect testimony in court This sometimes give incorrect testimony in court This sometimes give incorrect testimony in court. This sometimes give incorrect testimony in court. This Committee was chaired by Lord Devlin and its report was Committee was chaired by Lord Devlin and its report was published in 1976. published in 1976.

2

In In the the early early 1980 1980s s I I conducted conducted (with (with Brian Brian Clifford) Clifford) for for the the Government Government a programme programme of

• f research

research studies studies that that was was Government Government a programme programme of

• f research

research studies studies that that was was given given an an impetus impetus by by the the 1976 1976 publication publication of

• f this

this ‘Devlin ‘Devlin Report’ Report’. . p The The Report Report stated stated that that as as far far as as the the Committee Committee members members were were concerned concerned no no research research had had been been conducted conducted on

• n voice

voice identification identification (i (i. .e e. . human human recognition recognition of

• f a voice

voice heard heard only

• nly
• nce
• nce before)

before) but but that that “research “research should should proceed proceed as as rapidly rapidly ibl ibl i t i t th th ti lit ti lit f i d as as possible possible into into the the practicality practicality of

• f voice

voice parades parades ... ... or

• r any

any

• ther
• ther appropriate

appropriate methods” methods”. .

3

In In a a 1984 1984 book book chapter chapter in in which which we we reviewed reviewed our

• ur research

research studies studies (and (and those those of

• f others)
• thers) we

we concluded concluded that that studies studies (and (and those those of

• f others)
• thers) we

we concluded concluded that that “Until “Until future, future, more more realistic realistic studies studies argue argue to to the the contrary contrary we we would would recommend recommend that that prosecutions prosecutions based based solely solely on

• n a

p y witness’ witness’ identification identification of

• f a

a suspect’s suspect’s voice voice (if (if the the suspect suspect is is a a stranger) stranger) ought

• ught not

not to to proceed, proceed, or

• r if

if they they do do proceed proceed they they should should fail fail We We say say this this because because we we are are of

• f the

the they they should should fail

• fail. We

We say say this this because because …. we we are are of

• f the

the

• pinion
• pinion

that that ear ear-

• witnessing

witnessing and and eye eye-

• witnessing

witnessing are are similarly similarly and and considerably considerably error error prone prone. . This This is is not not to to say say h i id ifi i id ifi i h ld h ld b d id id h that that voice voice identification identification should should not not be be used used as as an an aid aid to to the the prosecution prosecution or

• r the

the defence, defence, but but it it should should not not form form any any major major part part of

• f the

the evidence evidence presented presented in in court court.” major major part part of

• f the

the evidence evidence presented presented in in court court.

4

Five Five years years later, later, in in 1989 1989, an an overview

• verview on
• n ear

ear-

• it

it id tifi ti id tifi ti ( itt ( itt b fi fi witness witness identification identification (written (written by by five five respected respected North North American American psychologists psychologists – – D ff b h D ff b h t l ) bli h d bli h d th t th t Deffenbacher Deffenbacher et et al al.) was was published published that that examined examined all all the the published published research research on

• n the

the i h i h hi h hi h l (i (i accuracy accuracy with with which which people people (in (in experiments) experiments) are are able able correctly correctly to to identify identify a a i h h d h d i l i l I h i h i voice voice they they heard heard once

• nce previously

previously. . In In their their concluding concluding paragraph paragraph they they stated stated that that

5

“Inasmuch “Inasmuch as as the the results results we we have have reported reported are are optimal

• ptimal in

in that that witnesses witnesses were were not not stressed stressed and and there there was was no no attempt attempt at at witnesses witnesses were were not not stressed stressed and and there there was was no no attempt attempt at at voice voice disguise, disguise, recognition recognition accuracy accuracy at at realistic realistic delays delays and and speech speech sample sample durations durations was was so so low low that that we we would would agree agree ith ith B ll B ll and and Clifford’s Clifford’s (1984 1984) concl sions concl sions Depending Depending with with Bull Bull and and Clifford’s Clifford’s (1984 1984) conclusions conclusions. Depending Depending

• n
• n the

the parameters parameters involved, involved, recognition recognition of

• f an

an unfamiliar unfamiliar voice voice may may have have a a sufficient sufficient probability probability of

• f accuracy

accuracy that that it it y p y p y y could could be be of

• f use

use in in a a police police investigation investigation. . Unless Unless further further more more ecologically ecologically valid valid studies studies argue argue to to the the contrary, contrary, however however ear ear-witnessing witnessing is is so so error error prone prone as as to to suggest suggest that that however, however, ear ear-witnessing witnessing is is so so error error prone prone as as to to suggest suggest that that no no case case should should be be prosecuted prosecuted solely solely on

• n identification

identification evidence evidence involving involving an an unfamiliar unfamiliar voice voice. .” ”

6

A later later overview

• verview of
• f research

research on

• n voice

voice identification identification was was published published in in 1995

• 1995. In

In that that chapter chapter a Canadian Canadian professor professor of

• f

published published in in 1995

• 1995. In

In that that chapter chapter a Canadian Canadian professor professor of

• f

psychology psychology (Dan (Dan Yarmey) Yarmey) reviewed reviewed not not

• nly
• nly

12 12 publications publications of

• f his

his own

• wn but

but also also some some 22 22 publications publications by by

• ther
• ther people

people on

• n the

the topic topic of

• f
• ice
• ice identification

identification This This

• ther
• ther people

people on

• n the

the topic topic of

• f voice

voice identification identification. This This

• verview
• verview stated

stated that that “One “One of

• f the

the myths myths still still held held by by many many laypersons laypersons and and officials

• fficials in

in the the criminal criminal justice justice system system is is yp yp j y the the belief belief that that eyewitness eyewitness memory, memory, including including voice voice recognition, recognition, is is merely merely common common knowledge” knowledge” and and that that “Most “Most voice voice identification identification issues issues of

• f concern

concern to to the the court court

• f
• f

voice voice identification identification issues issues of

• f concern

concern to to the the court, court, of

• f

course, course, are are for for voices voices of

• f strangers

strangers… …identification identification for for unfamiliar unfamiliar voices voices must must by by treated treated with with caution” caution”. .

7

In In December December 1998 1998 I I was was invited invited by by the the British British A d A d f F i F i S i S i t t Academy Academy of

• f Forensic

Forensic Sciences Sciences to to present present a paper paper

• n
• n ear

ear-

• witness

witness testimony testimony. . In In August August 1999 1999 the the ti l ti l C i i l C i i l C t C t f A l A l (i (i th th f national national Criminal Criminal Court Court of

• f Appeal

Appeal (in (in the the case case of

• f

Roberts Roberts) reported reported in in its its written written judgement judgement that that the the l f th th ll t ll t (i (i th th i t d i t d lawyers lawyers for for the the appellant appellant (i (i.e e. . the the convicted convicted man man who who was was appealing appealing the the conviction) conviction) had had placed placed b f b f it it th t th t 1998 1998 d th th C t C t t d t d th t th t before before it it that that 1998 1998 paper paper and and the the Court Court noted noted that that among among the the points points I I made made were were the the following following: :

8

• voice

voice identification identification is is more more difficult difficult than than visual visual identification identification; identification identification;

• voice

voice identification identification of

• f a

a stranger’s stranger’s voice voice is is a a very very difficult difficult task, task, even even where where the the opportunities

• pportunities to

to listen listen to to the the voice voice are are task, task, even even where where the the opportunities

• pportunities to

to listen listen to to the the voice voice are are relatively relatively good good;

• voice

voice identification identification is is more more likely likely than than visual visual identification identification y to to be be wrong wrong; ;

• ordinary
• rdinary people

people seem seem as as willing willing to to rely rely on

• n identification

identification by by ear ear-

• witnesses

witnesses as as they they are are on

• n identification

identification by by eye eye-

• witnesses

witnesses; ;

9

• in

in the the light light of

• f the

the above above points, points, the the warning warning given given to to jurors jurors of

• f the

the danger danger of

• f a miscarriage

miscarriage of

• f justice

justice in in relation relation jurors jurors of

• f the

the danger danger of

• f a miscarriage

miscarriage of

• f justice

justice in in relation relation to to witnesses witnesses who who are are identifying identifying by by voice voice should should be be even even more more stringent stringent than than that that routinely routinely given given by by judges judges in in England England to to j rors j rors in in relation relation to to the the e idence e idence of

• f e e

e e England England to to jurors jurors in in relation relation to to the the evidence evidence of

• f eye

eye- witnesses witnesses. . It It should should be be brought brought home home to to jurors jurors that that there there is is an an even even greater greater danger danger of

• f the

the ear ear witness witness believing believing g g g him/herself him/herself to to be be right right and and yet, yet, in in fact, fact, being being mistaken mistaken; ;

• ear

ear witness witness identification identification is is so so prone prone to to error error that that it it h ld h ld t b li d li d f i ti i ti l should should not not be be relied relied upon upon for for a a conviction conviction unless unless some some

• ther
• ther supporting

supporting or

• r confirming

confirming evidence evidence is is available available. .

10

In In the the light light of

• f these

these points points the the Court Court of

• f Appeal

Appeal d id d d id d i th th ti l ti l b f b f it it th t th t “W “W decided, decided, in in the the particular particular case case before before it, it, that that “We “We do do not not think think that that the the identification, identification, which which rested rested l t l t h ll h ll th th i f th th ll t ll t h almost almost wholly wholly on

• n the

the voice voice of

• f the

the appellant appellant as as he he spoke spoke to to the the police police officers,

• fficers, was

was good good enough enough to to bl bl t th t th t thi thi i ti i ti f d enable enable us us to to say say that that this this conviction conviction was was safe safe and and consequently consequently we we quash quash this this conviction” conviction”. .

11

In In some some criminal criminal trials trials judges judges do do not not agree agree with with requests requests from from the the defence defence lawyers lawyers that that ear ear witness witness evidence evidence may may be be so so the the defence defence lawyers lawyers that that ear ear-witness witness evidence evidence may may be be so so error error prone prone that that such such evidence evidence should should not not be be allowed allowed to to form form part part of

• f the

the prosecution prosecution case case. . Instead, Instead, they they sometimes sometimes p p , y allow allow an an ‘Expert ‘Expert Witness’ Witness’ (such (such as as myself) myself) to to testify testify (e (e. .g g. inform inform the the jury) jury) (i) (i) about about research research findings findings on

• n the

the general general li bili li bili f i i i i ( h ( h h i d i d b ) b ) reliability reliability of

• f ear

ear-

• witnessing

witnessing (such (such as as that that mentioned mentioned above) above) and and (ii) (ii) on

• n factors

factors directly directly relevant relevant to to the the ear ear-

• witness

witness evidence evidence being being presented presented in in that that particular particular trial trial Regarding Regarding evidence evidence being being presented presented in in that that particular particular trial

• trial. Regarding

Regarding the the latter latter I I have, have, for for example, example, conducted conducted experiments experiments for for and and testified testified in in a a number number trials trials concerning concerning: g

12

My My ‘expert’ ‘expert’ evidence evidence in in example example case case one

• ne was

was concerned concerned with with whether whether people people could could tell tell which which one

• ne (actually

(actually the the with with whether whether people people could could tell tell which which one

• ne (actually

(actually the the suspect’s) suspect’s) of

• f several

several voices voices in in the the ‘voice ‘voice parade’ parade’ played played by by the the police police to to the the rape rape victim victim was was the the only

• nly one
• ne that

that was was an an dit d dit d

• ice
• ice sample

sample (being (being from from a police police inter ie ) inter ie ) the the edited edited voice voice sample sample (being (being from from a police police interview), interview), the the

• thers
• thers speaking

speaking in in a a monologue monologue (that (that is, is, unedited) unedited) (all (all the the speakers speakers said said different different things) things). p g ) g ) The The defendant defendant and and his his lawyers lawyers were were of

• f the

the view view that that if if his his voice voice ‘stood ‘stood out’

• ut’ in

in this this way way from from the the other

• ther voices

voices played played t th th i ti i ti h i d i d i ht i ht h hi hi i (i (i f to to the the rape rape victim victim her her mind mind might might choose choose his his voice voice (i (i.e. . for for the the wrong wrong reason) reason). . In In order

• rder to

to test test this this I I conducted conducted a a simple simple experiment experiment. . p p

13

In In this this experiment experiment I I played played the the ‘voice ‘voice parade’ parade’ to to a a number number of

• f people

people and and asked asked them them which which voice voice sample sample was was from from an an interview interview with with the the p police police. . The The vast vast majority majority of

• f the

the listeners listeners chose chose the the suspect’s suspect’s sample sample. . At At trial trial I I was was allowed allowed to to present present my my experiment experiment and and its its results results to to the the jury jury jury jury. In In this this trial trial the the jury jury could could not not agree agree a a verdict verdict (even (even by by a a majority) majority). . A A year year later later there there was was a a re re-

• trial

trial and and the the suspect suspect was was convicted convicted (by (by a a majority majority t ) t ) vote) vote). In In response response to to this this case case (and (and similar similar cases cases that that I I was was involved involved in) in) the the police police service service in in England England began began to to draw draw up up guidelines guidelines (in (in the the light light of

• f

p g g p g ( g each each case case and and my my related related expert expert reports) reports) on

• n how

how ‘voice ‘voice parades’ parades’ should should be be constructed constructed. .

14

I l t ( hi h ( hi h t k t k l l t ) l t ) In In example example case case two two (which (which took took place place some some years years later) later) my my evidence evidence concerned concerned the the extent extent to to which which the the (arson (arson and and murder) murder) suspect’s suspect’s voice voice stood stood out

• ut from

from the the other

• ther voices

voices in in murder) murder) suspect s suspect s voice voice stood stood out

• ut from

from the the other

• ther voices

voices in in the the voice voice parade parade as as better better matching matching a a major major aspect aspect of

• f the

the brief brief voice voice description description originally

• riginally given

given to to the the police police by by the the witness witness of

• f the

the perpetrator’s perpetrator’s voice voice (in (in terms terms of

• f it

it being being “high “high pitched”) pitched”). .

15

In In this this case case the the police police had, had, quite quite rightly, rightly, taken taken the the trouble trouble to to consult consult an an expert expert in in phonetics phonetics to to assist assist with with the the choosing choosing of

• f the

the non non-

• suspect’s

suspect’s p p g p voices voices that that would would appear appear in in the the ‘voice ‘voice parade’ parade’. . This This expert expert had had chosen chosen from from the the several several dozen dozen voice voice samples samples that that the the police police had had given given to to him, him, those those that that best best matched matched the the suspect’s suspect’s in in terms terms of

• f accent

accent recording recording those those that that best best matched matched the the suspect s suspect s in in terms terms of

• f accent,

accent, recording recording quality, quality, and and the the way way in in which which the the several several (short) (short) samples samples of

• f each

each person’s person’s speech speech were were put put together together into into a a longer longer sample sample for for that that speaker speaker This This expert expert then then took took the the trouble trouble to to conduct conduct a a ‘mock ‘mock witness’ witness’ speaker

• speaker. This

This expert expert then then took took the the trouble trouble to to conduct conduct a a mock mock witness witness test test in in which which he he could could demonstrate demonstrate that that people people who who listened listened to to the the ‘voice ‘voice parade’ parade’ he he had had constructed constructed were were not not likely likely to to pick pick out

• ut the

the t’ t’ i th th b i b i i t i d i t i d (b (b th th li ) li ) b t b t suspect’s suspect’s voice voice as as the the one

• ne being

being interviewed interviewed (by (by the the police) police) about about an an arson arson attack attack. .

16

Unfortunately, Unfortunately, the the police police did did not not inform inform this this expert expert that that when when briefly briefly describing describing to to them them the the voice voice he he overheard

• verheard (through

(through briefly briefly describing describing to to them them the the voice voice he he overheard

• verheard (through

(through a closed closed door) door) planning planning the the arson arson attack attack the the witness witness had had mentioned mentioned that that the the voice voice was was “ “… …high high pitched pitched… …”. . I I cond cted cond cted a small small st d st d in in hich hich people people listened listened to to the the conducted conducted a small small study study in in which which people people listened listened to to the the voice voice parade parade that that the the police police had had played played to to the the only

• nly witness

witness. . I asked asked them them to to indicate indicate (on (on a response response sheet sheet involving involving a ( p g seven seven-

• point

point scale) scale) for for each each voice voice “how “how high high pitched” pitched” it it was was. . Eighty Eighty per per cent cent of

• f them

them indicated indicated that that ‘voice ‘voice G’ G’ was was the the highest highest pitched pitched voice voice and and the the remaining remaining 20 20% indicated indicated the the highest highest pitched pitched voice voice and and the the remaining remaining 20 20% indicated indicated that that voice voice G G and and another another voice voice were were higher higher pitched pitched than than all all the the other

• ther voices

voices. . Voice Voice G G was was the the suspect’s suspect’s voice voice. .

17

At At trial trial (in (in December December 2002 2002) ) I testified testified to to the the effect effect that that the the defendant’s defendant’s voice voice could could have have inappropriately inappropriately stood stood out

• ut

defendant s defendant s voice voice could could have have inappropriately inappropriately stood stood out

• ut

from from the the others

• thers because

because the the witness witness had had originally

• riginally said

said to to the the police police that that it it was was “high “high pitched” pitched”. . p g p The The jury jury convicted convicted the the defendant defendant of

• f murder

murder (a (a woman woman died died in in the the arson arson attack attack on

• n her

her home) home) but but probably probably largely largely because because the the co co-

• accused

accused (her (her ex ex-

• boyfriend)

boyfriend) during during the the trial trial changed changed from from his his ‘not ‘not guilty’ guilty’ plea plea to to testifying testifying that that he he had had asked asked the the defendant defendant to to carry carry

• ut
• ut the

the arson arson attack attack. .

18

Largely in the light of the cases I have briefly Largely in the light of the cases I have briefly d ib d ( d l t d id d d d d ib d ( d l t d id d d d described (and related guidance produced and described (and related guidance produced and updated by the police in response to each of the updated by the police in response to each of the l i hi h I t tifi d) th G t l i hi h I t tifi d) th G t several cases in which I testified) the Government several cases in which I testified) the Government (in England and Wales) decided to issue official (in England and Wales) decided to issue official id i id tifi ti ( l t t id i id tifi ti ( l t t guidance on voice identification (relevant aspects guidance on voice identification (relevant aspects

• f which are presented in the following three
• f which are presented in the following three

lid ) lid ) slides). slides).

• 19. ADVICE ON THE USE OF VOICE IDENTIFICATION

PARADES Home Office December 2003

PREPARATION OF MATERIAL PREPARATION OF MATERIAL 6 The identification officer in charge should obtain a detailed statement from the witness This 6 The identification officer in charge should obtain a detailed statement from the witness This

• 6. The identification officer in charge should obtain a detailed statement from the witness. This
• 6. The identification officer in charge should obtain a detailed statement from the witness. This

should contain as much detail and description of the voice as is possible should contain as much detail and description of the voice as is possible. All descriptions of the voice All descriptions of the voice given by the witness must be included in the material supplied to the relevant forensic phonetics/ given by the witness must be included in the material supplied to the relevant forensic phonetics/ linguistics expert. The statement and any ‘first description’ of the suspect's voice should also be the linguistics expert. The statement and any ‘first description’ of the suspect's voice should also be the subject of disclosure to the suspect/ solicitor prior to any identification procedure. subject of disclosure to the suspect/ solicitor prior to any identification procedure. subject of disclosure to the suspect/ solicitor prior to any identification procedure. subject of disclosure to the suspect/ solicitor prior to any identification procedure.

• 8. The identification officer should obtain a representative sample of the suspect's voice... Experts in
• 8. The identification officer should obtain a representative sample of the suspect's voice... Experts in

the field state that under no circumstances should the suspect be invited to read any set text, as the the field state that under no circumstances should the suspect be invited to read any set text, as the the field state that under no circumstances should the suspect be invited to read any set text, as the the field state that under no circumstances should the suspect be invited to read any set text, as the speech/rhythm/tone may be unnatural and may well be altered by a person reading aloud from speech/rhythm/tone may be unnatural and may well be altered by a person reading aloud from prescribed written material. prescribed written material.

• 9. The identification officer should obtain no less than 20 samples of speech, from persons of similar
• 9. The identification officer should obtain no less than 20 samples of speech, from persons of similar
• 9. The identification officer should obtain no less than 20 samples of speech, from persons of similar
• 9. The identification officer should obtain no less than 20 samples of speech, from persons of similar

age and ethnic, regional and social background as the suspect. A suitable source of such material may age and ethnic, regional and social background as the suspect. A suitable source of such material may be other police recorded interview tapes from unconnected cases. be other police recorded interview tapes from unconnected cases.

• 20. The expert
• 11. The identification officer should request the services of a force approved expert
• 11. The identification officer should request the services of a force approved expert

witness in phonetics/ linguistics… to ensure the final selection and compilation of the witness in phonetics/ linguistics… to ensure the final selection and compilation of the l i t h ith th t' t l ibl l i t h ith th t' t l ibl sample voices match with the suspect's as accurately… as possible. sample voices match with the suspect's as accurately… as possible.

• 12. The tape containing the sample of the suspect’s voice, together with the batch of
• 12. The tape containing the sample of the suspect’s voice, together with the batch of

‘similar voices’ tapes should be passed to the commissioned expert witness ‘similar voices’ tapes should be passed to the commissioned expert witness similar voices tapes should be passed to the commissioned expert witness. similar voices tapes should be passed to the commissioned expert witness.

• 13. The expert should be commissioned to take selected samples of speech from the
• 13. The expert should be commissioned to take selected samples of speech from the

batch of tape sources … A total of nine samples should be selected (i.e. the suspect’s batch of tape sources … A total of nine samples should be selected (i.e. the suspect’s plus 8 others). plus 8 others).

• 16. It is strongly advised that the expert and identification officer conduct a number of
• 16. It is strongly advised that the expert and identification officer conduct a number of

t t h i tili i k it Th i di id l h ld b i b i f t t h i tili i k it Th i di id l h ld b i b i f test hearings, utilising mock witnesses… These individuals should be given a brief test hearings, utilising mock witnesses… These individuals should be given a brief resumé resumé of the case. They should then be asked to listen to the series of samples under

• f the case. They should then be asked to listen to the series of samples under

controlled conditions and asked to try and pick out the suspect for the offence (which controlled conditions and asked to try and pick out the suspect for the offence (which they will only be able to do on a random basis…). they will only be able to do on a random basis…). y y ) y y )

• 21. CONDUCT OF AUDIO/VOICE PROCEDURE
• 22. The suspect's solicitor must be given the opportunity to be present when the voice
• 22. The suspect's solicitor must be given the opportunity to be present when the voice

identification procedure is conducted. identification procedure is conducted.

• 23. The identification procedure should be videotaped and the suspect given the
• 23. The identification procedure should be videotaped and the suspect given the
• pportunity to review it at a suitable time after the procedure has taken place.
• pportunity to review it at a suitable time after the procedure has taken place.
• 25. The witness must be instructed by the identification officer that the voice of the
• 25. The witness must be instructed by the identification officer that the voice of the

suspect may, or may not be on one of the samples played during the procedure. The suspect may, or may not be on one of the samples played during the procedure. The witness must be instructed to listen to each tape at least once before he/she makes a witness must be instructed to listen to each tape at least once before he/she makes a

• selection. The witness must be allowed to listen to any or all the samples as many times
• selection. The witness must be allowed to listen to any or all the samples as many times

as they wish. as they wish. 27 Following the procedure a statement must be taken from the witness recording the 27 Following the procedure a statement must be taken from the witness recording the

• 27. Following the procedure a statement must be taken from the witness, recording the
• 27. Following the procedure a statement must be taken from the witness, recording the

events and their selection. events and their selection.

22

Thank you for listening to Thank you for listening to Thank you for listening to Thank you for listening to my voice. my voice. my voice. my voice.