JC, Jamaican Creole, transcripts, transcription, conversations, the prosecution, linguists, Prosecution Translator, native speakers, Linguist, translation, Patrick, Prosecution, linguistic expertise, Caribbean English Creoles, Amsterdam, London, native speaker, John Benjamins, Cambridge University Press, Pidgins and Creoles, Prosecution Linguist, Defense Linguist, expert witness, L. Patrick, Warwick University Caribbean Studies, American Dialect Society, University of Alabama Press, Jamaican Creole English, KEYWORDS Creoles, Aboriginal Law Bulletin, Standard American English, social dialects, standard language, Creole Linguistics, the British legal system, Recent developments in forensic linguistics, Blackwell, Academic Press, University of California Press, non-speakers, British Black English, University of New South Wales Press, Transcripts Labov, London Jamaican, Jamaican Creole syntax, linguistics, Southwest Educational Development Laboratory, data selection, Torres Strait Creole, United States Department of Justice, linguistic evidence, phone calls, natural speech, native knowledge, Donald George Brown, creole language, creole languages, African American English, American English, telephone calls, Patrick Patrick, Peter L. Patrick, natural languages, Jamaica, criminal prosecution, recordings, video recordings, transcription and translation, data reduction, expert witnesses, English Creole
Competing Creole Transcripts on Trial Peter L. Patrick Dept. of Language and Linguistics, University of Essex [email protected]
http://privatewww.essex.ac.uk/~patrickp/ Samuel W. Buell U.S. Dept. of Justice, District of Massachusetts [email protected]
ABSTRACT A criminal prosecution of Jamaican Creole (JC) speaking `posse' (=gang) members in New York included evidence of recorded speech in JC. Clandestine recordings (discussions of criminal events, including narration of a homicide) were introduced at trial. Taped data were translated for prosecution by a non-linguist native speaker of JC. Defense disputed these texts and commissioned alternative transcriptions from a creolist linguist, who was a non-speaker of JC. Prosecution in turn hired another creolist, a near-native speaker of and specialist in JC, to testify on the relative accuracy of both sets of earlier texts. Differing representations of key conversations were submitted to a non-creole speaking judge/jury, both linguists testified, and defendants were convicted. The role of linguistic testimony and practice (especially transcription) in the trial is analysed. A typology of linguistic expertise is given, and effects of the language's Creole status and lack of instrumentalization on the trial are discussed. KEYWORDS Creoles, Jamaican, transcription, translation, types of expertise, wiretap
Patrick & Buell INTRODUCTION1 Linguists appearing as expert legal witnesses in court have as yet rarely been asked to testify about pidgin or creole languages.2 The opportunity for such testimony generally arises in jurisdictions where a standard language (and its dialects) are dominant, and a pidgin or creole is perceived as exotic and not native to the jurisdiction; where pidgins and creoles are routinely spoken, they frequently enter into courtroom proceedings, but are apparently not highlighted as subjects for expert linguistic testimony. Creoles are natural languages which are independent i.e., not dialects of the (standard) languages they are related to.3 Many creolists hold creoles are not directly genetically descended from them by the usual processes of language evolution (Thomason and Kaufman 1988). Creole languages are typically widely-disrespected, unwritten vernaculars which have not undergone standardization or been formally admitted to public use in institutional discourse, even in their home settings. Consequently, many of their own native speakers
are uncertain whether they exist as "real languages". Presenting expertise on creoles to an audience of nonlinguists (e.g., judges, juries and attorneys) who may be disinclined to credit them with the status or complexity afforded to recognized standard varieties, and even to their regional or social dialects, thus raises interesting problems. In the case discussed here, although gaining admission of linguistic testimony to the courtroom was not difficult due to the nature of the evidence, these problems are compounded by the difficulties associated with the interpretation of transcribed evidence that is drawn from recordings made in difficult mechanical or discourse conditions.
Competing Creole Transcripts... This discussion shows the impact linguistic expertise can have on a type of criminal case that is increasingly common in the USA: one involving tape-recorded conversations among individuals speaking non-English languages or dialects. The use of linguists in such cases is likely to gather steam, not only because of the statistically increasing significance of non-Standard-English speaking populations, but also because newly assimilating and economically disadvantaged immigrant groups bear a disproportionate brunt of the overall impact of crime in society. While immigrants from Atlantic English Creole-speaking areas have long been resident in such (post-)colonial metropolises as New York City
, London, Toronto and Miami, their encounters with the courts have only recently, and inconsistently, been recognized as episodes involving cross-cultural communication
.4 The questions of whether intervention by a language expert is required, and what the qualifications for expertise are, thus remain unresolved. The special linguistic relationship between creoles and dominant versions of their lexifier languages (e.g., Jamaican Creole and Standard American English) comes into focus in this account of a criminal trial in which recordings and transcriptions of highly vernacular Jamaican speech played a significant role as evidence. We first describe the general linguistic context and the background of the case briefly, and then discuss two general issues raised in considering the linguistic portion of the evidence: · The role of data annotation, reduction and selection, especially the transition from tape recording to transcription, and · The types and sources of expertise and knowledge required for effective interpretation of such speech data.
Patrick & Buell JAMAICAN CREOLE Nearly all Jamaicans speak a Creole language which is structurally distinct from English. (Most Jamaicans can make themselves understood to non-speakers of JC when desired, and many also speak a standard variety of English, some of them natively.) Jamaican Creole (hereafter JC), or Patwa as its speakers name it, probably crystallized as a separate language in the last quarter of the 17th century.5 The major input of lexical items and historical sound and grammar patterns was from the English of the day, but West and Central African substrate languages also made significant contributions, especially to syntax, phonetics and pragmatics. This rapid process of creolization, in which a language may evolve from typologically diverse source materials in a century or less, is thought to be radically different from language change
and genesis in the normal course. Thus broad varieties of JC are unintelligible today to speakers of American or British English; even speakers of other Caribbean English Creoles may sometimes find Jamaicans difficult or impossible to understand. JC is known both to linguists and West Indians in general as one of the "deepest" varieties of English
Creole in the region. It is also a language of very strong ethnic identification: people of Jamaican ethnicity are strongly expected to be conversant in the Patwa, and nearly all are; while conversely no outsider is expected to be competent in the language, and almost none are. Because this case pitted a close-knit group of Jamaican-born Creole speakers against US-native law ENFORCEMENT OFFICERS
who do not speak it, competence in JC played a crucial role in both the criminal activity of this case, and in its investigation and prosecution. It is important to note that JC is essentially an unwritten and non-standardized language. It is in fact one of the best-studied of all creoles certainly Caribbean English
Competing Creole Transcripts... Creoles: linguists have composed a phonemic orthography (Cassidy 1961), dictionary (Cassidy and Le Page 1967), and grammar (Bailey 1966), and are deeply involved in the teaching of language arts in Jamaica.6 It is common for popular and literary writers to use JC for dialogue and even for narration (Cooper 1993, Lalla 1996). Nevertheless, no standard orthography is yet popularly recognized and used, the language receives no official recognition, and its very existence is routinely denied and denigrated by many of its own speakers. The official language of news media, education, administration, and law courts in Jamaica is a standard English that few Jamaicans learn natively at home. `OPERATION ISLAND GREEN' The case under discussion, US v. Derrick Riley, et al., was the culmination of Operation Island Green, an investigation of Jamaican `posse' activities in urban areas
across the US, conducted by the Federal Bureau of Investigation (FBI). Jamaican posses (members are often known as `yardies') are organized crime units which originate and are grounded geographically in specific neighborhoods of Kingston, Jamaica's capital city, though their activities have spread far beyond it (Gunst 1995). The case was tried in federal court in Brooklyn, New York in Spring of 1998. It charged the commission of grave acts of violence, including multiple homicides, by an organized group of professional criminals. The victims were Jamaicans, members of posse members' own cultural group. Many posse victims are illegal immigrants, distrust the police as a result of experiences both in Jamaica and abroad, have family members
in Jamaica who cannot be protected from reprisal by US law enforcement, and generally have little or no recourse. Two men were charged with numerous crimes under the RICO act, which punishes conspirators acting for organized crime groups, including:
Patrick & Buell · credit card fraud · money laundering · conspiracy to commit robbery · narcotics conspiracy · armed robbery · attempted murder in aid of racketeering, and · murder in aid of racketeering.7 The prosecution noted that language itself was at the core of the challenge facing law enforcement. The posse members had escaped law enforcement efforts in part by conducting their activities in JC, which appears to the average police officer to be an impenetrably dense form of an unwritten foreign language. The difficulty outsiders have in understanding everyday JC was compounded by an overlay of urban `street talk', 8 as well as the highly specialized coded terminology of a rapidly-changing criminal argot, which is often intentionally obscure and unfamiliar to normal native speakers. Intensive work was required to collect and process large amounts of such speech. Theoretically, trial at a lower level was possible; but without the level of resources available to the FBI and federal prosecutors, the case could never have been investigated, much less successfully prosecuted in court. It required the efforts of dozens of officials (including an FBI agent who has specialized in Jamaican posses for over a decade), most working full-time for a period of years, to gather sufficient evidence largely in the form of posse members' own utterances and to make that evidence usable in the highly restrictive environment of a criminal trial.9
Competing Creole Transcripts... DATA SELECTION, ANNOTATION AND REDUCTION The investigation began five years before trial, with wiretaps as early as March 1993. The FBI gathered large amounts of wiretap, video and body-wire linguistic evidence against defendants for several years before the Justice Dept. gave permission to prosecute. The prosecution then spent most of a year, full-time, examining telephone logs and reviewing draft transcripts of the tapes. These data were of two primary types: 1) Wiretap: a) on cellular phone of Derrick Riley for 60 days in New York (May-June 1995) b) on cellular phone of Kirk "Scarry" Lyons for 90 days in Los Angeles (MarchMay 1995), and 2) Consensual recordings: tapes made by a confidential informant of conversations he participated in between mid-1993 and late 1995. The wiretaps in (1a) generated about 75 tapes (roughly 1/day), containing 3-40 phone calls each; those in (1b) amounted to about 250 tapes, each with roughly 10-30 phone calls varying widely in length. Source (2) generated about 1,000 recordings: 75% of them were telephone calls and 25% were in-person meetings recorded by body-wire or hidden videocamera at undercover locations. Although it is difficult to generalize over diverse data-types, the quantity of information collected and processed by the government is massive by the usual standards of linguistic analysis
. A conservative estimate gives 6-7,000 wire-tapped telephone calls made without the knowledge of participants, plus 700 calls and 200 in-person audio or Video Recordings
made with the knowledge of at least one participant.10
Patrick & Buell These impressive amounts of data speak to the importance of the sampling or selection process, and the key role of data reduction
before analysis. The initial steps taken by the Justice Dept. were similar to those empirical linguists might follow: · creation and review of logs detailing the participants and topics of each phone call/conversation; · monitoring of tapes with native Creole speakers when possible;11 · commissioning of draft transcripts by native speakers of JC (non-linguists); · elimination of the many calls not relevant to the investigation; · careful review of remaining tapes and revision of transcripts for use at trial. The point of selection, however, diverged from linguists' usual goals. Instead of sampling with the quantitative aim of representation, i.e. to derive a picture of what activity is typical of speakers, the investigators had a qualitative one: to locate precisely that evidence required to prove serious criminal activity at trial, as well as evidence which might falsify or damage their claim that such activity was committed. Ultimately, only a tiny fraction (1-5%) of all recordings collected and reviewed were brought to trial: ca. 115 wiretap conversations, and 40-50 consensual recordings. The prosecutors screened them in advance under the evidentiary rules and offered only relevant excerpts, not entire conversations. Of this selection, the defense chose to contest approximately 10%, about a dozen transcripts. (To this point we have used the term `transcript' loosely; below we will distinguish transcription from translation.)
Competing Creole Transcripts... TRANSLATION VS. TRANSCRIPTION Though many people worked on the first phase of data selection, subsequently 3 language professionals were principally involved in the reduction of data from tape recordings to transcripts and translations which forms the linguistic core of the case: · `Prosecution Translator', a native speaker of Jamaican Creole (but not a linguist). She produced the documents the prosecution initially introduced at trial, some of which the defense contested.12 · `Defense Linguist', a creolist who listened to the tapes and contested the validity of the prosecution's account of what was on them. He produced a transcription and translation in side-by-side columns. · `Prosecution Linguist' (=Patrick), employed by prosecution to assess the validity of the other two's work. He too worked directly from the tapes and produced his own transcription and translation. Crucially, the Prosecution Translator did not render her product as written Jamaican Creole i.e. did not transcribe the variety as it was spoken but rather produced a translation into colloquial standard English. Possible reasons for this include time pressure, training, the prosecution's view of what was required, and language ideology. It is far more time consuming to produce a transcript plus translation than just the latter; especially when no accepted standards for orthography or transcription of an essentially unwritten language are commonly recognized, as they must first be worked out. The Prosecution Translator's training probably did not include such linguistic-analytic practices. The prosecution did not explicitly request transcription, focusing on accuracy
Patrick & Buell of translation as the crucial issue at stake (and needing to analyse thousands of conversations). Finally, it is common among Jamaicans to attempt to render oral JC directly as written Standard Jamaican English a practice probably learned in the classroom, and formed under the pressure of a powerful ideology which denies JC status as a language (Roberts 1988). Especially in a formal context
such as the judicial system, the effort to translate from JC into SJE is an almost automatic response (regardless of the speaker's abilities in SJE). Thus, when the Defense Linguist was retained to contest the prosecution's interpretation of the recorded evidence, his task was not one of contesting meanings derived from an already agreed-upon text. Rather, he had the opportunity to place before the court an entirely new, and more basic, text his version of the words actually spoken by the defendants as well as his translation of it, which differed significantly from Prosecution Translator's. Given that precision in transcription is the foundation of interpretation, and that apparently tiny differences of form may be the source of crucial differences in meaning, most empirical linguists would unhesitatingly agree with Defense Linguist on the necessity of a transcript. The very nature of the product he delivered thus tends to lend a certain amount of credibility to its contents, principally by making them accessible and falsifiable. On the other hand, the absence of established standards for method of transcription, choice of orthography, assessment of accuracy and qualifications for expertise increases the likelihood that a product which is idiosyncratic, incomplete, or even incompetent will be accepted into evidence, while decreasing the likelihood of comparability between the output of competing experts.
Competing Creole Transcripts... Faced with such disagreement on critical evidence, one option is to solicit further expertise, ideally from an expert with decisive credentials. Patrick was retained as the Prosecution Linguist to evaluate the differences alleged between the work of Prosecution Translator and Defense Linguist.13 Having at his disposal copies of both the initial prosecution translation, and the defense transcription, he worked directly from the primary evidence: the tapes.14 He prepared new transcripts and translations of 9 wiretap calls and 2 consensual recordings of conversations, in a format deliberately comparable to Defense Linguist's. In the event, very little of this material was considered during testimony but the small amount of speech analysed in court assumed considerable significance. TYPES OF EXPERTISE We considered four different types of knowledge or expertise to be relevant to this case: native-speaker competence, explicit structural knowledge, native cultural knowledge, and experience in transcription. 1) Native speaker competence in Jamaican Creole It will be readily recognized by linguists that a native-like command of lexis, grammatical structure, phonetic and phonological patterns, discursive devices and styles, and pragmatic conventions can contribute greatly to optimal comprehension and translation of taped evidence. Prosecution Translator possessed this type of knowledge. So did Prosecution Linguist: having grown up in Jamaica, including Kingston, Patrick
Patrick & Buell conservatively identified himself as a near-native speaker of JC. The defense linguist could not claim to be a native speaker of any Creole. 2) Explicit LINGUISTIC KNOWLEDGE
of JC structure and grammar This is knowledge of the sort linguists typically gain by systematic study
of a variety, and demonstrate in publications. For example, it may be achieved through reading the linguistic literature, analysing a corpus of natural speech, conducting fieldwork in a speech community, or working intensively with native-speaker informants. It is extremely useful in transcribing difficult or ambiguous data, e.g. in identifying (un)grammatical, (un)likely, or (im)possible utterances, as it may cumulatively approximate to a native speaker's intuitive knowledge of the elements and rules underlying the construction of well-formed utterances. While Prosecution Translator lacked such knowledge, both Defense Linguist and Prosecution Linguist commanded it. However, although Defense Linguist has worked with at least one other Caribbean English Creole language, he claimed no special acquaintance with Jamaican Creole per se, has not published any work on JC, or done fieldwork in Jamaica. He may have transcribed and translated JC data for the court via knowledge gained through reading the literature, or second-hand study of recordings, or perhaps by analogy with another Caribbean Creole language that he actually knew better.15 Here it is relevant to note a widespread theoretical position in the field of contact languages: that creoles may be defined by a commonality of linguistic structure and, specifically, that Caribbean English Creoles essentially share a common structure (e.g. Winford 1993). This might
Competing Creole Transcripts... incline a creolist to assimilate aspects of JC data to other creoles perhaps incorrectly a difficulty against which knowledge of type (1) provides a check.16 3) Native cultural knowledge relevant to the speech and data in question This type of expertise is not strictly linguistic, though it relates to the concept of communicative competence (Hymes 1972): knowledge that enables one to function as a competent participant in a culture's everyday speech events and acts. As with nativespeaker knowledge, it is normally acquired through extended membership in a speech community from youth. It is especially important when the task includes identifying or interpreting speech segments: · under difficult mechanical conditions (e.g., against background noise, static, or poor recording quality) · or difficult discourse conditions (over-lapping, latching, polyphonic speech) · or where paralinguistic elements are distinctive or unique to the variety17 · or where the nature of the speech situation confidential conversations among intimate acquaintances with illicit business affairs in common calls for implicit references to shared knowledge. All these conditions apply to the data reviewed in US v. Riley. The best camouflage for posse members has simply been their participation in Jamaican urban culture, which makes it practically impossible for US law enforcement to infiltrate posses from the outside. Each posse is grounded in a specific area of Kingston, often composed of networks of people who have largely known each other as friends or even family since a young age, and are intimately familiar with the popular culture of their peers. Both the
Patrick & Buell Prosecution Translator and Linguist naturally acquired such knowledge during childhoods spent in Jamaica; Defense Linguist, however, did not possess it. A similar type of knowledge is relevant here: 3a) Subcultural knowledge of `posse' practice, argot and history Such information and experience (e.g. knowledge of firearms, nicknames, the status of current and rival gang members, network links to originating neighborhoods) derives from membership in a small, closed criminal group. Most Jamaicans do not possess it, nor probably any linguists, nor any of the three principals who performed transcription/ translation in this case. An FBI agent who worked on Riley possesses a thorough knowledge of aspects of urban Jamaican life gained by tracking posses since 1986; still, it would be impossible for him to pass either as culturally Jamaican or a posse member. 4) Training and experience in the decoding and transcription of speech recorded under non-ideal conditions This type of expertise is also not linguistic in nature. A professional skill shared by many empirical linguists, it is acquired through time spent behind the reels of tapetranscribers, working with recorded data. Prosecution Linguist has over 10 years and thousands of hours worth of experience recording, listening to, transcribing, and analysing tapes of urban JC speech.
Competing Creole Transcripts...
Types of expertise possessed by language professionals in the trial
1 Native speaker 2 Linguistic expertise 3 Cultural knowledge
Translator 9 9
Linguist 9 9 9
3a Subcultural `posse' knowledge
To summarize, Prosecution Translator possessed knowledge of types (1) and (3), at least, and perhaps (4); Defense Linguist possessed complementary expertise of type (2), and it may also be, of type (4). Prosecution Linguist possessed knowledge of types (1-3, 4): a near-native speaker of JC, he has explicit knowledge
of Jamaican grammar, extensive cultural experience, and training in working with natural speech recordings.
ADMITTING MURDER Of the 11 conversations contested by the defense for which Prosecution Linguist prepared alternate transcripts, the prosecution submitted three for jury review. Prosecution Linguist ultimately testified about only a single critical tape, N-65, a long consensual recording for which he transcribed a 20-minute segment (spending ca. 16 hours). Of that segment, only 7:30 mins. came under discussion. Part of Prosecution Linguist's testimony involved pointing out multiple inaccuracies in Defense Linguist's transcripts and translations, based on this material. The import of these defense errors for the case ranged from trivial to grave (though at the time of analysis and testimony,
Patrick & Buell Prosecution Linguist was not concerned, and indeed unable, to judge this). The most important testimony centered on just four disputed lines, lasting 15 seconds. In the key sentence, Defense Linguist's version differed crucially from Prosecution Translator's and Prosecution Linguist's by one phoneme, in fact by a single phonetic feature: voicing of a sibilant. In Table 2 we give this evidence, which received the most attention in court, in all three text versions from the trial, beginning with the Prosecution Translator's, following with Defense Linguist's, and then Prosecution Linguist's.
Competing Creole Transcripts...
3 accounts of murder: Translation & transcription of taped evidence
2a. Prosecution Translator
Is he the one who got licked at the dance?
It is he who (unintelligible) at the dance?
4 DG Brown:
It's Scarry's gun I used too, you know.
Ha, how you mean, man?
6 DG Brown:
Mash up, man ... (pause) When I'm dealing with
those things there, I don't leave till I see that,
boy, the breath has left the body, you know
2b. Defense Linguist
DT: How di, how di him get lickin at
DT: How did, how did he get a licking
1 at the dance?
DGB: Yeah. 2
DT: You been ta da dance?
DT: You been to the dance? 3
DGB: Scarry gun me used to, ya know.
DGB: Scarry's gun, I am used to, you
Patrick & Buell
DT: Huh. How ya mean man?
DT: Huh. How do you mean man? 5
DGB: Mash up man. Any a dem ting, me na lef till me da see da boy who had a lef da body ya know. 2c. Prosecution Linguist (Patrick)18 TRANSCRIPT (11:45) DT: [Aala, aala] even get l- lick here a de dance?
DGB: Smashed up man. With any of those 6 things, I don't leave till I see the boy who had left the body you know. TRANSLATION (11:45) DT: [A., A.] even got h- hit here at the 1 dance?
DGB: Yeah. 2
DT: .. A in di miggle a de dance?
DT: .. Was it in the middle of the dance? 3
(4 second pause)
(4 second pause)
DGB: Scarry gun me use, too, y'know.
DGB: It was Scarry's gun I used, too,
Competing Creole Transcripts...
DT: A-ah, how- how you mean, man?
DT: A-ah, what- what do you mean,
DGB: .... Mash up, man .... Me a deal
DGB: .... (Things got) messed up, man.
in dem t'ing deh, me no lef till when me 6 .... (When) I'm dealing in those things, I
see seh bwoy, breat' lef de body,
don't leave till I see that boy, breath has
left the body, y'know.
As the last witness on the final day of the 7-week trial, Prosecution Linguist was asked to focus on this exchange, in which the prosecution charged murder was admitted. In this segment of tape N-65, defendant Donald George Brown (DGB) speaks to another man (DT), who is recording the conversation. The subject is a fatal shooting in a dancehall. The prosecution charges that Brown is admitting it, or rather boasting of it. The defense contends that his conversation implicates another man as the shooter. (Recall that the first version only consists of a translation.)
ANALYSIS OF TRANSCRIPTS In all the above transcripts there are six turns at talk, consisting of three questions asked by DT followed by what appear to be three answers from DG Brown. All versions agree in turn (1) that DT raises a question concerning someone who got lick at a dance. This JC verb is ambiguous, possibly meaning `hit', `assaulted', `shot', or `killed'; but in the context, the meaning that emerges in all accounts is `shot and killed'. In each version, it
Patrick & Buell is clear that (1) is a topic shift or initiation: this particular dancehall shooting has not previously been raised in this conversation, or at any rate not for many minutes. The thrust of the question in (1) is different in each. All agree, however, that the response in (2) by the defendant, DG Brown, is positive. Prosecution Translator and Prosecution Linguist's accounts both feature a yes/no question answered by a yes. DT attempts to elicit information about the identity of the victim of the shooting. The question is met with a minimal response by DG Brown, who confirms the victim's identity, but does nothing to extend the exchange. By contrast, in the defense account, a WH-question a how-question pre-supposing both the shooting, as a live topic, and the victim's identity is met by an infelicitous response, the non-answer `Yeah'. All versions disagree about the nature of the question that constitutes turn (3), though in each version it serves to elicit more narration about the incident. (Recall that this is a clandestine recording consented to by DT an informant, and the party asking all the questions.) This turn is followed by a long pause lasting 4 seconds, a fact not noted by any account but Prosecution Linguist's. The pause works to delay the giving of information information which both sides agree is crucial. DG Brown's response in (4) is the most important statement in the disputed tape recordings. It is nearly identical in all versions, yet the defense assigns it a very different interpretation from the prosecution. According to Defense Linguist's account, Brown says he is "used to", as in "accustomed to, familiar with", the gun belonging to Scarry (another posse member). That is, Defense Linguist contends he employs the past participle
form used (to) which, in American English, is one of only two forms of the verb use to require the unvoiced sibilant /s/.19 Presumably if Brown is so accustomed to
Competing Creole Transcripts... Scarry's gun, he could recognize its report when it was used (by someone other than Brown) in the shooting. In this account DG Brown's response is, on its face, neither relevant to nor coherent with question (3) which provoked it, or the immediately following Q/A pair (56). In all accounts, question (5) is a request for DGB to clarify his statement by expanding it. Similarly, all versions agree on Brown's response in (6): in such cases, he says, it is not his custom to leave the scene immediately until some resolution is reached regarding a dead body. As an innocent clarification of (4), under the defense's reading, this statement of practice is rather unexpected. By contrast, in both the Prosecution Translator's and Prosecution Linguist's versions, turn (4) has Brown replying to DT's question by admitting that he used Scarry's gun. DT here makes a second attempt to elicit facts about the shooting, and Brown rewards his persistence after a long pause with the most-reportable fact (Labov 1997:406).20 To DT's request for expansion in (5), Brown explains that he does not typically leave the scene of a shooting until he is certain that the victim is dead. If Brown has just admitted (boasted of?) a murder then the response in (6), and indeed the entire sequence, is perfectly coherent. These two very different interpretations are separated by the slenderest of threads: the voicing of the sibilant in used. In both prosecution accounts, Brown employs the transitive main verb use in the past tense with Scarry['s] gun as its object, followed by the adverb too.21 Like all other forms of this verb except the two noted above, the sibilant is voiced /z/ in English. In support of this interpretation, Prosecution Linguist made the following four arguments to the jury:
Patrick & Buell BASIS FOR THE INTERPRETATION OF TURN (4): A. Prosecution Linguist heard / yuuz / with voiced sibilant the first time, and every time, on the tape. He listened to this segment of tape one hundred times or more. B. Used to as a past participle meaning "accustomed to, familiar with" is a common expression in American English, but Prosecution Linguist knew of no evidence to indicate that it is so used by Jamaican speakers, and in his opinion it is much less common in JC if it occurs at all. C. A final / z / in use / yuuz / is not ambiguous and can only have the main-verb meaning given by the prosecution. However, even if there was an unvoiced / s / as in the participle / yuus /, as Defense maintained, it would be ambiguous either of the 2 possible meanings may apply. This is due to the rule of phrasefinal phonetic devoicing, which "turns off" voicing towards the end of a phrase. Thus even if Defense Linguist had correctly heard / s /, he could not be certain of the utterance's meaning it might just as easily have arisen from the sentence heard by Prosecution Linguist, with devoicing applied. D. Coherence of Q/A pairs favors the prosecution's interpretation. If Defense Linguist is correct, Brown replied to three questions in a row infelicitously and incoherently. If Prosecution Linguist is correct, however, he first identified the victim, paused, admitted to shooting them, and finally added a dramatic coda describing his usual practice at murder scenes.
Competing Creole Transcripts... Argument A highlights relative credibility of the two expert linguistic witnesses, rather than their methods, as each claimed to have listened to the tape many times and heard contrasting segments (though see expertise type 4 above).22 Argument B is quantitative and negative, thus relatively weak, but it goes to the issues of nativespeaker knowledge and explicit linguistic expertise raised earlier. In particular, if Defense Linguist relied on his native knowledge of American English when gaps in his knowledge of JC arose, he might well have imagined he heard the American phrase used to where a Jamaican would be unlikely to utter it (Wald 1995). Argument C undermines any claim by Defense Linguist to be certain that Brown's turn (4) was not an admission of murder (though of course the burden of proof for charges remained with the prosecution). In presenting Argument D to the jury, Prosecution Linguist appealed to Grice's (1975) Co-operative Principle, putting the jury in the position of a co-operative hearer who relies on assumptions such as relevance and coherence in order to construct an understanding of conversation (Prince 1982: 2-3) a `common-sense' approach that may have been the most accessible of the linguistic arguments. CONCLUSION Donald George Brown was convicted of numerous criminal charges including six murders, prominent among them the one just described, and sentenced to multiple, consecutive mandatory life sentences
with no parole. We believe this case has significant implications for the use of linguists as expert witnesses, particularly in criminal cases. In the prosecutor's post-trial opinion, Defense Linguist's work was compromised. Defense counsel overlooked and attempted
Patrick & Buell to explain away weaknesses in their expert's qualifications, and involved him too deeply as a member of the defense team (including having Defense Linguist meet alone with defendants to discuss the tapes). The prosecutor reports that defense attorneys went so far as to file a motion for a new trial after conviction, in part on the ground that they allegedly had been hoodwinked by their own expert, claiming to have been ignorant of the flaws in his work until his cross-examination and the subsequent testimony of Prosecution Linguist. In effect, defense's motion endorsed the practices and standards advocated here rather than those of their own witness. As linguists become more frequently involved as expert witnesses in the legal system, it is increasingly common that they find themselves on opposing sides at trial. It is crucial to the "battle of the experts" in such a case that prosecution be able to show their linguist has followed sound, objective practices in analyzing the taped materials, and has expressed careful, independent conclusions in court. Prosecution promoted these tendencies in this case by deliberately screening off Patrick from other evidence, from trial tactics and strategy, and even to some degree from such fundamental premises as which speakers on the tapes would be defendants at the trial. The fact that evidence against the defendant included natural speech in a creole language was also important. Since JC served as a quasi-secret code and a language of strong in-group identification
, it made law-enforcement efforts very difficult and required extensive resources for investigation, prosecution and defense. Creoles are also languages of little regard, even among some linguists who, like the general public, continue to think of them as simple and quaint. Unwritten and little-understood, they are spoken by people without power. Jamaican Creole's non-official status, combined with lack of a publicly-accepted orthography or transcription conventions, and the massive
Competing Creole Transcripts... phonological reduction that characterizes vernacular JC, may promote a cavalier attitude towards representing it in transcripts, and even contribute to a belief that the language is sloppy or imprecise in meaning, or at any rate that care and precision in representing meaning is unnecessary. It would be well to remember that most people who speak creole languages have been badly served by experts, and have known little justice, from the cultures that have dominated them. Their speech is the record of their human acts, and when we encounter it in the courts it deserves nothing less than our best efforts at objectivity, attention to detail, intellectual honesty, and indeed our utmost professionalism, whether as linguists or as lawyers, for the defense or the prosecution.
Patrick & Buell NOTES 1 We thank Bethany Dumas and Diane Eades for commenting on an earlier version of this paper given to the Society for Pidgin & Creole Linguistics (Patrick 1999a) and encouraging subsequent presentation to the International Association of Forensic Linguists (Birmingham, June 1999). Buell, who was at the time of trial an Assistant US Attorney for the Eastern District of New York in Brooklyn, NY, tried the case and retained Patrick as an expert witness; discussion of the case and of Patrick's testimony between the two authors, however, mostly took place after trial, for reasons noted below. The views expressed herein are solely the authors' and do not necessarily reflect the views of the United States Department of Justice. 2 A query sent to the CreoList (Siegel 1998), an international email exchange list for contact language specialists, in late 1997 produced reports of only 7 instances, including 3 by the first author before the one reported here (but see also Alleyne 1980:6). References include Blackwell (1996) for Jamaican Creole in England, Shuy (1993) for Hawai'ian Creole English, Trezise (1996) for Torres Strait Creole in Australia; other languages reported include Sierra Leonean Krio in Washington DC and Jamaican Creole again in Maryland, North Carolina and New York State
s, USA, and Toronto, Canada. All cases reported involve English-related creoles spoken by immigrants to the USA, Australia, the UK or Canada. In only 3 of the 7 was linguistic expert testimony included in court proceedings. This list almost certainly underestimates extant cases. 3 Though there is considerable controversy as to the nature and origins of creole languages and their identity as a group, most creolists will agree with this point, and agree that all the languages mentioned in the preceding note are creoles. Pidgins are not discussed here. See Arends et al. (1994), Holm (1989). 4 On JC in urban England see Sutcliffe (1982), Sebba (1993). For a post-colonial situation in Australia which involves not a creole but a dialect of English see Eades (1995), Walsh (1999). 5 Patrick (1999:91) cites Historical Data
showing that the population went from rough parity between white settlers and African slaves in 1673 to predominance of Africans by nearly 20:1 in 1703; if not in this generation, then the next, JC must have been formed. D'Costa & Lalla (1989) are conservative, only noting that JC is firmly in place by the first attested text of 1740. 6 The orthography used in this article is closely modelled on Cassidy's, which is used by most linguists for JC and has been widely influential among other Atlantic Creoles.
Competing Creole Transcripts... 7 `RICO' stands for the Racketeering and Corrupt Influenced Organizations Act. E.g, apparently in order to advance his standing in the organization, one man shot and killed his girlfriend, who had no involvement in the underlying criminal activity, because he believed she was fraternizing with members of a rival posse. 8 Urban creoles are rarely studied by creolists, who typically concentrate on archaic, rural speech for its conservatism and historical value; but see Patrick (1999b). 9 In the US criminal justice system
the state is the prosecution, and is bound by a more stringent set of rules than the defense. In order to ensure the reliability of criminal convictions, the principle of "error-deflection" is established in the direction of mistaken exoneration of the guilty, rather than wrongful conviction of the innocent. The prosecution bears the burden of proof, must produce evidence sufficient to dispel any reasonable doubt about a defendant's guilt, and is required to obtain a unanimous jury verdict in order to convict. 10 The legality of consensual recording, and the number of participants who must be made aware of the process, vary from state to state in the USA; good discussions by linguists of surreptitious recording in the US and Canada include Shuy (1986), Larmouth (1992) and Murray and Ross-Murray (1992, 1996). See also Patrick (1999c) for a brief list of sources on the ethics of clandestine linguistic recordings. 11 The FBI agents and Justice Department
staffers were not speakers of JC, and in some locations it proved difficult to recruit native JC-speaking staff who could pass the FBI's security clearance standards. 12 Only Prosecution Linguist (the first author) is here identified. The accurately-gendered pronouns are useful in distinguishing the other two. 13 Defense also consulted Patrick on his availability as expert witness just after he had agreed to consult for prosecution but, apparently, before defense learned of his identity as the Prosecution Linguist. 14 We refer here to the initial defense transcripts, created before Patrick was retained. After Patrick's transcripts had been made available to Defense Linguist, a second version of the defense transcripts, containing many revisions, was submitted to the court.
Patrick & Buell 15 Close perusal of contrasts in the transcript sets suggests the possibility that Defense Linguist analogized JC in part to African American
English, as elements native to the latter but not the former (e.g. negative copula ain't) appear in his transcriptions though not on the tapes. 16 Wald (1995), reviewing a published collection of transcripts and analyses of African American English recordings by creolists and other scholars, discusses types of theoreticallyinduced bias and their effect on the transcription process. 17 Non-lexical but conventional sounds and intonations conveying, for example, disagreement, agreement, skepticism, sincerity, sarcasm, humor and negative affect. 18 Conventions for transcription and translation codified by Patrick for this task note: "Where the exact identity of a word is uncertain but a plausible guess has been made, the word is enclosed in square brackets "[ ]" in both the original and the English translation
columns." Since the bracketed portion of turn 1 is believed to be a name, it is translated simply with an initial `A.'. Further, relevant to italicized portions of turn 6, the conventions note: "Words in the English translation which are necessary to accurately translate the Jamaican meaning, but which are omitted or have different elements corresponding to them in Jamaican, are given in the English column (in italics and enclosed in parentheses)." Pause length and timing details also appear in italics and parentheses, on separate lines. 19 The other is intransitive past auxiliary used to + Verb, denoting habitual or repeated action. 20 "The semantic and structural pivot on which the narrative is organized" (Labov 1997: 406). 21 Possession in JC does not use English post-nominal 's; it occurs either by juxtaposition [possessor possessed] or by prenominal free morpheme huufa or fi (Bailey 1966, Patrick fc.). 22 Readers may question why instrumental phonetic evidence was not introduced at trial. Prosecution Linguist offered to perform such analysis on the tape, noting that it might still be inconclusive for such a small segment. The prosecutor decided against introducing a new and highly technical mode of analysis of speech data at the trial's end, not wanting to further inflate the relative importance of linguistic evidence. On the very issue described here, for example who shot the gun the prosecution also brought to bear ballistic evidence and testimony from an accomplice.
Competing Creole Transcripts...
REFERENCES Alleyne, M.C. (1980) Comparative Afro-American, Ann Arbor: Karoma. Arends, J., Muysken, P. and Smith, N. (eds.) (1994) Pidgins and Creoles: An introduction, Amsterdam: John Benjamins
. Bailey, B. L. (1966) Jamaican Creole syntax, Cambridge: Cambridge University Press. Blackwell, S. (1996) `Corrective measures: Some aspects of transcription in the British legal system', in H. Kniffka (ed.), Recent developments in forensic linguistics, Frankfurt: Peter Lang
, 255-76. Cassidy, F. G. (1961) Jamaica talk: Three hundred years of the English language in Jamaica, London: Macmillan. Cassidy, F. G. and Le Page, R. B. (eds.) (1967, 1980) Dictionary of Jamaican English, 2nd ed., Cambridge: Cambridge University Press. Cooper, C. (1993) Noises in the blood: Orality, gender and the `vulgar' body of Jamaican popular culture, (Warwick University Caribbean Studies), London: Macmillan. D'Costa, J. and B. Lalla (eds.) (1989) Voices in Exile: Jamaican texts of the 18th and 19th centuries, Tuscaloosa: University of Alabama Press. Eades, D.(1995) `Aboriginal English on trial: The case for Stuart and Condren', in D. Eades (ed.) Language in evidence. Linguistic and legal perspectives in multicultural Australia, Sydney: University of New South Wales
Press, 147-74. Grice, H. P. (1975) `Logic and conversation', in P. Cole and J. L. Morgan (eds), Syntax and semantics. III: Speech Acts, New York: Academic Press, 41-58. Gunst, L. (1995) Born fi' dead: A journey through the Jamaican posse underworld, New York: Henry Holt. Holm, J. A. (1989) Pidgins and Creoles, Cambridge: Cambridge University Press. Hymes, D. (1972) `Models of the interaction of language and social life,' in J. Gumperz and D. Hymes (eds.) Directions in sociolinguistics, Oxford: Blackwell, 35-71.
Competing Creole Transcripts Labov, W. (1997) `Some further steps in narrative analysis', Journal of Narrative and Life History 7:395-415. Lalla, B. (1996) Defining Jamaican fiction: Marronage and the discourse of survival, Tuscaloosa: University of Alabama Press. Larmouth, D. W. (1992) `The legal and ethical status of surreptitious recording in dialect research: Do human subjects guidelines apply?' Publication of the American Dialect Society 76: 1-14. Murray, T. E., and Ross-Murray, C. (1992) `On the legality and ethics of surreptitious recording', Publication of the American Dialect Society 76: 15-75. Murray, T. E., and Ross-Murray, C. (1996) `Under cover of law: More on the legality of surreptitious recording' Publication of the American Dialect Society 79: 1-82. Patrick, P. L. (1999a) `Applied creolistics in court: Linguistic, methodological and ethical dimensions of expert testimony', paper presented to the Society of Pidgin and Creole Linguistics meeting in Los Angeles, CA, 9 January 1999. Patrick, P. L. (1999b) Urban Jamaican Creole: Variation in the mesolect, (Varieties of English Around the World G-17), Amsterdam: John Benjamins. Patrick, P. L. (1999c) http://privatewww.essex.ac.uk/~patrickp/ethics.html Patrick, P. L. (Fc.) `Jamaican Creole English', in J. A. Holm and P. L. Patrick (eds, to appear), Comparative Creole Syntax: Parallel outlines of 18 Creole grammars, London: Battlebridge Press. Prince, E. F. (1982) `Language and the law: A case for linguistic pragmatics', Sociolinguistic Working paper
94, Austin, Texas
: Southwest educational development
Laboratory. Roberts, P.A. (1988) West Indians and their language, Cambridge: Cambridge University Press. Sebba, M. (1993) London Jamaican, London: Longman. Shuy, R. W. (1986) `Ethical issues in analyzing FBI surreptitious tapes', International Journal
of the Sociology of Language 62:119-28.
Patrick & Buell Shuy, R. W. (1993) Language crimes: The use and abuse of language in the courtroom, Cambridge MA: Blackwell. Siegel, J. (1998) `Language and the law (summary)', posting on the CreoList discussion list: [email protected]
, www.ling.su.se/Creole/CreoList.html, Mon. 9 Feb 1998, from [email protected]
Data also collected by Siegel from the forensic-linguistics discussion list: [email protected]
Sutcliffe, D. (1982) British Black English, Oxford: Blackwell. Thomason, S.G. and T. Kaufman (1988) language contact
, creolization and genetic linguistics, Berkeley CA: University of California Press. Trezise, P. (1996) `Use of language and the Anunga rules: R v Jean Denise Izumi', Aboriginal Law Bulletin 79(3), 17-18. Wald, B. (1995) `The problem of scholarly predisposition', review of G. Bailey, N. Maynor, and P. Cukor-Avila (eds), The emergence of Black English: Texts and commentary (Benjamins, 1991). Language in Society 24(2): 245-58. Walsh, M. (1999) `Interpreting for the transcript: Problems in recording Aboriginal land claim proceedings in northern Australia.' Forensic Linguistics 6(1), 161-95. Winford, D. (1993) Predication in Caribbean English Creoles, (Creole Language Library 10), Amsterdam: Benjamins. CASES CITED United States v. Derrick Riley, et al., Criminal Docket No. 96-149 (S-5) (RJD) (EDNY)
PL Patrick, S Buell