Annotated Bibliography and References Summary: This article discusses thediscipline of Forensic Linguistics. It begins by describing what ForensicLinguistics is, namely the interface between linguistics (the science of language)and the law, including law enforcement. It then outlines the history anddevelopment of Forensic Linguistics from its beginnings in the 1950y s and 1960a s tothe present day. A section on Forensic Phonetics is included, and the articleconcludes with how Forensic Linguistics works in the justice system and some of thedifficulties that linguists and lawyers may have in understanding each othersoviewpoints. The article concludes by suggesting that lawyers and linguists workmore closely with each other in the interests of justice, and that linguists seekto widen their understanding of international law, of international human rightsissues, and of how law and language relate to each other across the globe. Thearticle suggests that the future of Forensic Linguistics will be bright iflinguists work on these issues, and also on acquiring skills, knowledge andqualifications in other disciplines in order to better prepare them for working inand with courts. What is forensic linguistics? In ten words or less, what isForensic Linguistics? Forensic Linguistics is the application of linguistics tolegal issues. That is a starting point, but like all answers it is imperfect andserves only to stimulate more questions. For example, what does the application oflinguistics mean? When Forensic Linguistics is referred to as an application oflinguistics or, more concisely, an applied linguistic science, the word applied isnot necessarily being used in the same sense as, for example, in the phrase appliedstatistics, where what is being applied is a theory underpinning a particularscience to the practice of that science. Forensic Linguistics is, rather, theapplication of linguistic knowledge to a particular social setting, namely thelegal forum (from which the word forensic is derived). In its broadest sense we maysay that Forensic Linguistics is the interface between language, crime and law,where law includes law enforcement, judicial matters, legislation, disputes orproceedings in law, and even disputes which only potentially involve someinfraction of the law or some necessity to seek a legal remedy. Given thecentrality of the use of language to life in general and the law in particular, itis perhaps somewhat surprising that Forensic Linguistics is a relative newcomer tothe arena, whereas other disciplines, such as fingerprint identification andshoeprint analysis, are much older, having a well#established presence in judicialprocesses. The application of linguistic methods to legal questions is only onesense in which Forensic Linguistics is an application of a science, in that variouslinguistic theories may be applied to the analysis of the language samples in aninquiry. Thus, the forensic linguist may quote observations from researchundertaken in fields as diverse as language and memory studies, ConversationAnalysis, Discourse Analysis, theory of grammar, Cognitive Linguistics, Speech ActTheory, etc. The reason for this reliance on a broad spectrum of linguistic fieldsis understandable: the data the linguist receives for analysis may require thatsomething is said about how the average person remembers language, howconversations are constructed, the kinds of moves speakers or writers make in thecourse of a conversation or a written text, or they may need to explain to a courtsome aspects of phrase or sentence structure. In summary, we can say that theforensic linguist applies linguistic knowledge and techniques to the languageimplicated in (i) legal cases or proceedings or (ii) private disputes betweenparties which may at a later stage result in legal action of some kind being taken.Legal Cases and Proceedings In lay terms, for the purposes of this discussion, wecan envisage a legal proceeding as consisting potentially of three stages: theinvestigative stage, the trial stage and the appeal stage. The investigative stageis also sometimes referred to as the intelligence stage. In this part of theprocess it is important to gather information relating to the (alleged) crime. Notall of the information which is gathered during investigations can be used incourt, and so a linguist who assists law enforcement officers during theintelligence stage may, in fact, find that there is no requirement to give evidenceat any subsequent trial. Similarly, a linguist whose work is used at trial may notbe required to assist the court at the appeal stage, if the content of the appeal
does not include linguistic questions. On the other hand if linguistic evidencewhich was not available at the earlier stages comes to light while the appeal isbeing prepared, then this may be the stage at which the linguist is called in togive an opinion. The investigative stage Typically, requests for linguisticanalysis originate with law enforcement departments or, in some countries, at theinvitation of an investigating magistrate. Examples of linguistics intelligencework have included analysis of ransom notes, letters purporting to provideinformation on a case, mobile (cell) phone text messages, and specific threatletters. Linguists have also been asked to analyse texts purporting to be suicidenotes. Even though the police in such cases may not suspect foul play, it could beimportant to attempt to establish whether the questioned text can throw any lighton the cause or circumstances of death. Also at the investigative stage, the policemay need to have an opinion on a text or an interview tape, perhaps to assist indeveloping interview and interrogation strategies. It is unlikely that anything alinguist says about veracity (using techniques similar to statement analysis) wouldbe acceptable evidence in court, which is why this kind of linguistic analysis isusually confined to the investigative stage. The trial stage At the trial stage anyone of a number of types of linguistic analysis may be called for, includingquestions of authorship (Who wrote the text?/Who is the speaker in thisrecording?), meaning and interpretation (Does this word mean x, y or somethingelse?), threat analysis (Does the text contain a threat?), or text provenance andconstruction (Was the text dual#authored? Was it written rather than spoken? etc).The inquiry could be of a civil or criminal nature, and this will determine thelevel of proofl acceptable to the court in question. Usually, the forensiclinguist is instructed some time before a case gets to court. An expert report issubmitted to the instructing legal team either for the prosecution or the defence(or the plaintiff/claimant in a civil case). Even though the linguist prepares areport for one sidea in a case rather than the other, it is the court for whom thework is really done. The first duty of the linguist like that of any otherforensic expert is to the court, and not to the client on whose behalf theanalysis was originally carried out. The appeal stage If a defendant is convictedof a crime it is not uncommon, especially these days, for the defence legal team tolaunch an appeal almost immediately. The structure and nature of appeals variesfrom country to country, and in some countries appeals centre on the claim that newevidence has been made available, or that existing evidence should be looked at innew ways. It is becoming increasingly common for linguists to be called in toassist legal counsel at the appeal stage, either because there may be some disputeabout the wording, interpretation or authorship of a statement or confession madeto police, or because a new interpretation of a forensic text (such as a suicide orransom note) may have become apparent since the conviction. Private disputes A notinconsiderable part of the forensic linguisttthis is meant that the work is commissioned by private individuals not involved inlitigation at the time of the commission. Such cases include identifying the authorof anonymous hate mail, the investigation of plagiarism for a school or university,or on behalf of a student accused of plagiarism. It sometimes happens that thelinguisths report may have an influence on the cliento s decision to take mattersfurther, either in a civil or a criminal court, but this is not common. Usually,what happens is that the report is submitted and the client deals with the matterinternally either within a university department, a business organisation, or, asmay also be the case, within a family. History and development of ForensicLinguistics to the present Like almost all sciences it is not possible to say thatForensic Linguistics began at a specific moment in time. Questions of authorshiphave exercised minds since the times of the ancient Greek playwrights who notinfrequently accused each other of plagiarism. Since at least the eighteenthcentury scholars and amateurs alike have pondered over the authorship of some ofthe worldcs most famous texts, including sacred texts and the plays of Shakespeare.There was some attempt in the nineteenth century to develop methods of authorshipattribution, mainly by British and American mathematicians and statisticians,notably Augustus de Morgan, in 1851, TC Mendenhall (1887 and 1901) and in the
earlier part of the twentieth century by Udney Yule (1938 and 1944). These studiestended to concentrate on easily measurable attributes like word length average,mean sentence length, and so on. The application of these exercises, though, washardly forensic and, in any case, had little to do with linguistics. The actualphrase Forensic Linguistics was not used until 1968 when a linguistics professor bythe name of Jan Svartvik recorded its first mention in a now famous analysis of statements givento police officers at Notting Hill Police Station in 1953. Timothy John Evans wasaccused of the murder of his wife and baby at 10 Rillington Place, Notting Hill,London, England, tried at the Central Criminal Court of England and Wales (The OldBailey,) and hanged at Pentonville Prison. In the 1960fallegedly given to police following his arrest, troubled several people, includinga well#known journalist by the name of Ludovic Kennedy, and Svartvik wascommissioned to analyse the statements. Svartvik was one of the earliest linguistsinvolved in corpus studies, which is the systematic analysis of language throughthe collection and study of large bodies (hence corpus, pl. corpora) of language,and therefore he was able to approach the task of analysing the Evanso statementsin a methodical manner. He quickly realised that the statements contained twostyles and he set about quantifying the differences, ultimately demonstrating thatthey were, in fact, an educated written style and a marked spoken style. Along withother evidence collected in the course of many different threads of investigation,the findings of Svartvik showed that Evans could not, as had been claimed at histrial, have dictated the statements attributed to him. For a long period in Englishlaw a set of rules had been established regarding the interrogation of witnesses,in particular how statements were to be taken from them. These prescriptions wereknown simply as Judgese Rules which laid down that suspects were to dictate theirnarrative to police officers, that police officers were not to interrupt suspects,and that questions were not to be asked of the suspect at the statement stageexcept for minor clarifications. In practice this almost never happened. Typically,a police officer would ask a series of questions, take down notes and then write ortype the suspect s statement, not in the words of the suspect, but in a form andpattern which police custom had long dictated. Thus, police statements containedphrases like I then observeda, etc. This type of phrasing is not at all typical ofhow people speak, but rather reflects a way of phrasing which has come to be knownas police registeru, itself an area of study within Forensic Linguistics. Thelearned judges who formulated the rules for statement taking were not aware thatdictating a statement and transcribing it verbatim are difficult perhaps evenimpossible tasks for the average speaker. Learning to dictate a narrative in acoherent, sequential, articulate form is extremely difficult, but the person takingthe statement has an even harder task if the speaker is not skilled at pacinghis/her delivery. Usually, people do not deliver their statements in a coherent,ordered fashion: they speak too fast, they omit important details, they speculatealoud, they backtrack, and so on. In effect, the Judgesn Rules were unworkable.This was why police officers had their own way of taking and regrettably in somecases making statements. It was simply impossible to follow the prescriptions ofthe Judgesn Rules. This was why in the early days of Forensic Linguistics, at leastin the United Kingdom, many cases involved questioning the authenticity of policestatements. The first example of expert evidence being given from the witness boxon this matter was at a murder trial at the Old Bailey in 1989, where Peter Frenchdemonstrated the presence of police register in an incriminating statement theprosecution claimed was entirely in the words of one of the defendants. Notablecases included appealing against the convictions of Derek Bentley (posthumouslypardoned) the Birmingham Six, The Guildford Four, the Bridgewater Three, and so on.These last four cases all relied on the work of Britainr s most distinguishedforensic linguist, Professor Malcolm Coulthard of Birmingham University, adiscourse analyst who had first taken an interest in forensic questions followingan inquiry from a colleague. In the United States forensic work began slightlydifferently, but also concerned the rights of individuals with regard to theinterrogation process. In 1963 Ernesto Miranda was convicted of armed robbery, but
appealed on the grounds that he did not understand his right to remain silent or tohave an attorney present at the time of questioning. The Court of Appeal overturnedhis conviction in 1966. In the US there were many Miranda cases, as they came to beknown. On the face of it the provision of Miranda is a simple one: police officersare obliged to advise arrestees that they need not speak unless they wish to, thatthey are entitled to have a lawyer present, and that anything they say can be usedagainst them in court. However, many issues arose, as discussed by Professor RogerShuy: (i) a confession must be voluntary, (ii) questioning should not be coercive,(iii) arrestees must be asked whether they understand their rights, etc. Withregard to the first point Shuy pointed out that an arrestee is hardly in a positionto agree voluntarily to being questioned. Effectively, the very nature ofquestioning (as pointed out by the US Supreme Court) is coercive. Shuy (1997: 180)gives a good example of the issue of coercion in an interrogation process. Hedescribes how a suspect, having declined to speak following the reading of hisMiranda rights, was escorted in the back of a police car to the police station bytwo officers, who then began to talk to each other about the possibility of themurder weapon in the case (a shotgun) being accidentally stumbled upon by childrenat a nearby school. The suspect immediately waived his rights and led officers tothe location of the weapon. The suspect, a man by the name of Innis was convictedof murder and his lawyers appealed. The issue before the appeal court was whetherthe suspect had been coerced into making the confession. This in turn causedlawyers and judges to consider the meaning of the word interrogation. The RhodeIsland Supreme Court concluded that interrogation need not involve the asking of aquestion, and that in this case subtle coercion had occurred and that this was thefunctional equivalent of interrogation. In the US Supreme Court it was thusappreciated that interrogation need not be in the form of a question [and] mayinvolve the use of psychological ploys. However, it was also realised that theconversation between the officers was probably more in the nature of casual remarksthan a deliberate ploy. Shuy raised many important points about Miranda, andvigorously questioned many of its assumptions of simplicity. He cites one casewhere a fifteen year old boy from Houston, Texas was read his rights and ultimatelysigned a confession of murder. After analysing tape-recorded interviews between theattorney and the child Shuy concluded that though the boy often said he understoodwhat he was being asked it was clear that his level of comprehension was extremelylow. His school confirmed that his comprehension ability was no more than that ofan eight year old. Thus in this and other cases Shuy explores the most basicpremises of Miranda, and by extension similar legal provisions. He does nottake even the simplestd word or concept for granted: what does voluntarilyn mean,does understandi mean I say I understando or I actually understandn ? The work ofRoger Shuy, and other US linguists, has encompassed many areas of civil andcriminal practice, but right from the beginning, the law itself was, as it were,subject to questioning: what does this law mean? How do different people performwhen asked if they understando their rights? There is a very readable review ofearly Forensic Linguistics in the United States, written by Judith Levi (Levi1994). In her account Levi recalls a case in which she was asked to analyse a badnews about your social benefitsl letter written by the Illinois Department ofPublic Aid to recipients of child benefit payments whom they had categorised asnon#cooperativec. One of the tasks Levi undertook was to determine whether thevocabulary selections made by the drafters of the letter had used technical andbureaucratic language in place of ordinary, everyday language. Also included in theanalysis were pragmatic questions such as inferencing (Were inferences made byrecipients of the letter justified by the facts of the case? Did the writers of theletter provide incomplete information which could lead to the making of misleadinginferences? Was the reader forced to infer information that should have been madeexplicit?). The result of Levicthe letters. Most of the recipients of the benefit were single mothers who hadsuffered real hardship as a result of the Statefawarded $20,000,000 to the beneficiaries and ordered the State to rescind itsclassification of non#cooperationi until it had complied fully with the courts s
consent order. Finally, the State was ordered to re#write the letter in termscomprehensible to the beneficiaries (Levi 1994: 18). An important point noted byLevi is the comment by a linguist acting in another case, namely that the legalsystem is linguistically naïve and vulnerable (Levi 1994: 22). This point isreferred to in the next section. Another early application of Forensic Linguisticsin the United States related to the status of trademarks as words or phrases in thelanguage. An early case involved a dispute surrounding an aspect of the brand nameMcDonald sn, owners of the multi#national fast food chain. In this case thelinguists were Genine Lentine and Roger Shuy (as reported in Levi 1994: 5). QualityInns International announced their intention of opening a chain of economy hotels to be called McSleepo. McDonald sn claimed that the attachment of the Mcoprefix to many unprotected nouns, such as Friesa in McFriesmMcNuggetsm, etc., barred Quality Inns from use of the Mci prefix. In this casethe plaintiff was not just claiming implicit ownership of a name, but over amorphological principle, namely the attachment of a particular prefix to any noun.It appears that the claim was inherently one of a formula for combination (Levi1994: 5) and it was this formula for which protection was being invoked.McDonaldasd also claimed that they had originated the process of attachingunprotected words to the Mca prefix and had run advertising campaigns whichillustrated this. In their evidence Lentine and Shuy showed that the Mch prefixhad had previous commercial applications, and that as McDonaldt st had not objectedto any of these they had no grounds for doing so in the present instance. Despitethe overwhelming evidence presented by Lentine and Shuy, judgement was for theplaintiffhs and Quality International Inns were unable to launch their chain ofmotels under the McSleepn banner. In Australia linguists began meeting in the1980ls to talk about the application of linguistics and sociolinguistics to legalissues. They were concerned with the rights of individuals in the legal process, inparticular difficulties faced by Aboriginal suspects when being questioned bypolice. They quickly realised that even such phrases as the same language areopen to question. An important instance of this is the dialect spoken by manyAboriginal people, known colloquially as Aboriginal Englishc , wrongly thought bymany white Australians to be a defective form of the English spoken by whites. Itis in fact a dialect in its own right. Thus, when being questioned by police,Aboriginal people bring their own understanding and use of Englishb to theprocess, something which is not always appreciated by speakers of the dominantversion of English, i.e. white Englisha. More than this they bring their owninteractional, culturally#based, styles to the interview. An individuali s owninteractional style, if perceived to be at variance to that of the dominantculture, might compel responses to questions in particular, non#confrontationalways which could lead to a false assumption on the part of a questioner that thesuspect was being evasive or, worse still, that an admission of guilt was beingmade. Other Australian research focused on how Aboriginal witnesses and defendantsunderstood the legal processes involved in land claim hearings and examined theimpact of cross#cultural differences between white settlers and Aboriginal peopleon the presentation and even outcome of cases. In this context Gibbons (1994: 198)observes the system around interrogation in the courtroom is alien to Aboriginalculture. Gibbons is the author of two major books on Forensic Linguistics,Language and the Lawt (Longman, 1994) and Forensic Linguistics: An introductionto language in the justice system1 (Blackwell 2003). In these books he not onlysummarises some of his own considerable experience as a Forensic Linguistics expertin the court system, but also details much of the history of the development ofForensic Linguistics. Surprising as it may seem, the thread which connects many ofthese different forensic activities is authorship. Essentially, in considering thequestion of whether an individual dictated a statement, or whether a statement wasin the words of its alleged speaker, analysts were actually asking the questionWho was the author of the statement attributed to X?t This applied to thestatements of Derek Bentley, Timothy Evans, the Birmingham Six, the BridgewaterThree in the United Kingdom (and many others), as well as to Australian Aboriginaldefendants who claimed that police had verballed them (i.e. altered what they had
said). In the case of US defendants whose Miranda rights were being investigated,there was a slightly different kind of authorship nexus of questions whichincluded: Did the putative authors of statements (such as the 15 year old Houstonboy, or Mr Innis) make their statements voluntarily, knowingly and in fullpossession of their rights?r In other words, the issue here relates to theconditions of authorship: a series of questions put by police, for which answersare required, structures and even distorts a narrative of events; answers whichappear vague, ambiguous or reluctantly given may slant a narrative in a way whichis disadvantageous to the defendant and, as pertinently, to the apparent truth ofthe narrative. In any case, the assymetric nature of the relationship betweenauthority figures (the police) and the defendant who may be (i) illiterate (ii) aspeaker of another language than the language of interview (iii)young/disabled/ethnically disadvantaged, etc., can result in a text (such as arecord of interview, video or audio recording or written statement) which isconsiderably at variance with what the suspect would have said had he/she beengiven the opportunity to make a statement in a non#coercive or less threateningenvironment. In a broad sense to be an author is to possess the language you areusing. It is the use of language to produce a text over which you as the authorhave control, and whose course you are free to direct. Illiterate, young, disabled,language minority speakers are scarcely in control of the authorship process whengivinge a statement to powerful authority figures. We can realistically challengealmost any text produced under conditions of duress even where the duress may nothave arisen through the intention of questioning police officers, court officials,or any other authority figure within the justice system. If a suspectr s way ofusing language is at some remove from that of the officials with whom the suspectis dealing then the potential for distortion of the authorship process is clearlyexacerbated, probably in proportion to the differences of perspective,interactional styles and cultural norms between the institutional, authorityfigures on the one hand and the suspect on the other. This is not to suggest anymalice or intrinsic lack of fairness or justice on the part of officials: they workwithin institutional structures which are not always the most conducive for takingindividual circumstances into account. In Germany, an early case involved analleged slander by a tenant in an apartment complex of a fellow tenant (Kniffka,1981). The issue at stake was whether the word concubine was an insult. Linguistsadvised that for some speakers the word might be amusing, for others a way ofaddressing each other as a joke, while yet others might find it insulting undersome circumstances: it was not possible to say that a given word or phrase, on itsown, was an insult, or constituted verbal injury. Rather, the relationship betweenspeaker and hearer, the context of situation, the speaker s education level allneeded to be taken into account. A word does not have a single, universally-agreed,meaning within a speech community. Other issues which emerged in the early days ofForensic Linguistics in Germany involved authorship attribution, and thedevelopment of methodologies for doing so. An early case, reported by Kniffka,concerned the theses of twin sisters whose previous academic performance was,according to university authorities, at a much lower level than the theses theysubmitted for their final examination. Kniffka argued that an authorshipattribution in the case was not possible because the language used was essentiallythe meta#language of the law and that it was not easy to attribute such language toany given individual. He suggested the university authorities subject the studentsto a written examination on their theses to test their knowledge, rather thanrelying on subjective comparisons with their previous, known, work. In the yearssince Forensic Linguistics began to establish itself as a discipline its scope hasgrown considerably. From its beginnings as a means of questioning witness anddefendant statements, linguists have been called on to give evidence in manydifferent types of case, including authorship attribution in terrorist cases,product contamination cases and suspicious deaths; the interpretation of meaning inlegal and other documents, the analysis of mobile (cell) phone text messages toestablish a time of death. The list continues to grow. In the next section, animportant area within Forensic Linguistics will be considered: Forensic Phonetics,
the analysis of speech through auditory and acoustic means and its application inthe legal and criminal arena. Forensic Phonetics This article is about ForensicLinguistics rather than phonetics, but no account would be complete without somemention of the science which deals with questions of speaker identification,resolution of disputed content of recordings, the process of setting up voice line-ups and ear line-ups and related topics. It has a more established presence in thelegal forum than Forensic Linguistics and its progress has been assisted by recentadvances in acoustic engineering. Phoneticians are able to analyse the distinctivespeech characteristics of a speaker relative to other candidate speakers in aninquiry much more easily than as little as 20 years ago. An important ethic withinForensic Phonetics is that no means exists which can infallibly identify an unknownspeaker in a legal case (such as a hoax or bomb threat caller to an emergencyservice). Rather, like all branches of science Forensic Phonetics examines a set ofphenomena, in this case aspects of recorded speech, and offers opinions based onthe observations arising from the analysis. Among the earliest British forensicphoneticians were John Baldwin, Stanley Ellis and Peter French, while in Germany Hermann Künzel was also active. Künzel(with Eysholdt) considered many aspects of speech production with reference tosocial situations, including the influence of alcohol on speech (Künzel andEysholdt, 1992). Kniffkads (1990) collection contains accounts of some of the earlyforensic phonetic cases see especially Elliso s and Baldwint s contributions tothat collection. The earliest recorded voice identification testimony in the UK wasin 1965, given by Stanley Ellis at Winchester Magistrateso Court. Summary of thedevelopment of Forensic Linguistics The early years of Forensic Linguistics werecharacterised by two critical issues: The need to discover the scope andeffectiveness of Forensic Linguistics as a form of expert testimony within thecourt system. The need to improve methodologies within Forensic Linguistics and tomake these transparent to non#linguists. These issues are still ongoing. It istempting to add a third point to the above: the need to develop a theory ofauthorship as a socio#cognitive process, the relationship between individual andcommunity or social authorship and the nature of institutional authorship. However,the scope of such a discussion is beyond the present article. Like all sciences even new ones a disciplineos scientific methods, the need to educatenon#specialists and the constant testing of the limits of the science are alwayskey issues. To some extent these questions will be addressed in the followingsection. Forensic Linguistics in the Justice System In the previous section we sawthe kinds of cases which forensic linguists routinely advise on, but it isimportant to consider the mechanisms which underlie the use of Forensic Linguisticsin the world s justice systems, and the institutional and other factors whichrelate to the further development of Forensic Linguistics and its rôle as anadviser within the legal process. In this section the following aspects oflinguistics in the justice system will be considered: the relationship betweenlanguage and the law; the relationship between linguists and lawyers; theconflicting goals of linguists and lawyers; meaning and clarity in judgestdirections to juries, and the process of admitting linguists as expert witnesses.The relationship between the two abstract notions, language on the one hand and thelaw on the other is key to understanding how linguists can contribute to the forumof the law. It has often been said that the law is the language that enshrines it.Not only do we need language to frame the law, but we need language to understandthe law. Law and language are inseparable. For this reason it was perhaps onlynatural that, as linguistics developed throughout the course of the twentiethcentury, linguists would take an increasing interest in the relationship betweenthe two, specifically: the language of the law, the use of language within the law,and language in the court system. An early concern was the way in which the law isframed: it was often seen as abstruse, impersonal, vague or ambiguous. Lawyers werefrequently viewed as wordy and hyper-precise and many linguists questioned theassumption that lawyers were experts in the language. On occasion this led totensions between lawyers and linguists, with lawyers questioning the need forlinguistic testimony in the court system and occasionally seeking to exclude it.
One judge remarked to a phonetician: A linguist is someone who speaks a lot oflanguages, so what exactly are you doing here? (Storey#White 1997: 281). Inanother case a linguist was told by the judge that Surely there are only two kindsof English correct English and incorrect English?difficulties many lawyers and linguists have learned to work with each other. It isnow realised in some legal circles that the language of the law is often archaic,and that lawyers in an effort to protect their clients will frequently useexpressions whose meanings are not always transparent. However, it is not enough tosay that lawyers and non#lawyers have different ways of using language. They bringto encounters with each other different perspectives and hence different discoursepractices. Jackson made a close study of some aspects of language and the law, oneaspect of which was the rôle of narratives in the legal process (Jackson 1995). Inan earlier section of this article it was discussed how the police statements ofDerek Bentley and Timothy John Evans could not have been dictated, as claimed bypolice officers at the time. Such police statements are the kinds of documentJackson studied, except that his reference point was their use in the courtroomrather than their method of construction. Citing the example of a murder trial, heconsidered, for example, the fact that the prosecution is always able to presenttheir narrative first in a courtroom, and that the defence has not only to dislodgethis narrative, but to create a convincing one to replace it. This, Jacksonclaimed, inherently puts the defence at a disadvantage. As an example of theordinary person interfacing with the law Stratman and Dahl considered the languageof temporary restraining orders, and the difficulties ordinary speakers may have incomprehending them. They cite a case where a man served with a temporaryrestraining order drove to his partnerh s apartment and slipped a letter under herdoor in order to elicit from her what the problems in their relationship were andhow they could address them. The court argued that he had violated the restrainingorderhs injunction not to molest, interfere with or menacei his partner (Stratmanand Dahl 1996: 212). It was clear that the drafters of this particular law had adifferent conception of the words molest, interfere withi and menace, since it ishighly likely that most people would not ordinarily consider placing a letter undersomeonels door to be an act of molestation or menace on its own. In court judgesoften refer to dictionaries for the meaning of words which occur in legislation.However, this approach has been criticised. Generally speaking, linguists viewdictionaries as imprecise and limited. Meanings are probably best not taken from adictionary, but from experimentation and observation of how words are used. It isgenerally agreed that words have a core meaning and a number of fuzzyu meanings(see Goddard 1996: 254). While the core meaning is probably well understood bymostG people, it is as the word approaches the boundaries of its semantic envelopethat difficulties arise. We would probably all agree that dogs, cats and hamstersare pets. But what if a prospective tenant in a block of apartments which allowspets were to bring a chicken or a crocodile and claim such an animal as a pet? Howdoes this kind of meaning difficulty equate with interpretations of words in ourprevious example, such as molest and menace? Thus, while the word pet encompasses arange of familiar domestic animals in the minds of most of us which may or may notexclude such creatures as a chicken or a crocodile, words like molest and menacehave status as legal terminology. Though legal drafters are nowadays obliged to usewords in a meaning as close to ordinary language as possible, words like these dopresent special problems, since they may have been used in legal language forhundreds of years in a more or less fixed fashion, yet in ordinary language theirmeaning will probably have changed considerably. Linguists have proposed a numberof ways of dealing with this kind of difficulty, including carrying out semanticsurveys. However, this approach has not found universal favour among linguists. Forexample, Solan notes: People cannot explain what, for instance, makes a snake asnake, a game, a vehicle, etc. Generic categorisation is a matter of induction andintuition, which we are rarely able to describe (cited in Goddard 1996: 259).Goddard notes that it would be absurd for forensic linguists to promote themselvesas experts in the meaning of legal words, because this is really the province ofjudges. Using surveys, for example, to determine the meaning of a word, could
produce contradictory results. He suggests that if linguists confine themselves tonon#legal words, this may make more sense, but in any case, he points out thatsemantics is still a relatively under#developed area of linguistics and that thereis still considerable disagreement among semanticians as to methodology. Corpuslinguistics has allowed the semantic survey approach to flourish because in thetechnological age it is easy to collect many samples of a word in its ordinaryusage. However, an important competing process is that of semantic reduction (thereductive paraphrasep Goddard 1996: 269), which puts into practice Platon sdictum that a definition must use words which are simpler than the word which isbeing defined. Another area of potential conflict with regard to word#meaning is inthe directions given by judges to juries. It has often been pointed out that suchdirections are full of legal terminology, some of which may be present in thelanguage as ordinary, everyday words. How are jurors, by definition ordinarycitizens , to understand whether a word is being used as a technical term or as anordinary word, let alone understand the legal terminology? Would all the jurors ina case agree as to the meaning of a particular word? In recent years, in Englandand Wales at least, judges have received recommendations to illustrate their jurydirections with visual presentations, to avoid giving directions about the law, andto keep reminding juries throughout the jury direction phase what the issues in thecase are. In the US in some jurisdictions, judges are now being trained in how totalk to juries. When it comes to linguists giving evidence in court, it is clear that lawyers and linguists have differentgoals. The job of the lawyer is to convince or persuade the jury that the defendantis guilty or innocent. The job of the linguist is to present an opinion and toexplain that opinion. The lawyer may interrupt the expert witness, use rhetoric,spini, guile, and may choose to ignore anything the expert witness says. It issafe to say that it is not necessarily the case that the lawyer is intent ondiscovering or promoting the truthi. The linguist, on the other hand, mindful among other things of Grices famous Cooperative Principle will attempt to beinformative, truthful and relevant. However, if the evidence is injurious to theparty the lawyer is representing prosecution or defence the linguist mustexpect various lawyerly stratagems to suppress or distort that evidence. Lawyerscan also play on the notion of cooperativee. The linguist will usually attempt tobe cooperative, but linguist and lawyer may conflict about what cooperative meansin practice in a given instance. This, again, will be due to the differingdiscourse practices of lawyers and linguists. The lawyer will in all probabilitybring a folk-semantic meaning to the idea of cooperation: Why arenlcooperating with the court? After all, it s a simple question. The linguist on theother hand, mindful that the lawyer is attempting to direct the discourse away fromthe evidence, struggles with the lawyera notion of cooperative: to the linguist scooperation here means that lawyer and linguist cooperate to uncover the truth.Finally in this section, it is important to consider some aspects of the differentmethods of admitting expert witnesses into courts, in particular linguists. In theUS each state has its own rules of evidence, some of which will be applicable onlyto district courts, and some to higher courts. There are also Federal Rules ofEvidence and these differ in kind from the evidence rules of lower courts. Therules governing expert evidence are complex and not always understood. They requirethat scientific evidence meets certain standards. Generally, the Dauberte standardis what is insisted upon. This requires, among other things, that witnessesdemonstrate the known error rate attached to their opinion. This of course impliesthat the linguist must present quantifiable data. However, in linguistics it is notalways possible to present quantifiable data, and it may indeed be misleading to doso. Some courts have interpreted Dauberta more flexibly than this, and it is anongoing debate in legal and linguistic circles, with some insisting that anyauthorship attribution analysis must be backed up by the use of inferentialstatistics, which is the only way to demonstrate a known error rate in a particularcase. However, contrary to popular belief there is in reality no such thing as alinguistic fingerprint and it is not always possible to quantify a view that aparticular individual is the author of a questioned text in a case. In other
countries it is sufficient that the method on which the expert bases an opinionshould be acceptable to the scientific community, and that the expert should bequalified to give it. Both Canada and Sri Lanka, for example, follow this method ofaccrediting a witness and accepting an opinion. In the near and medium term futureit is likely that the question of how linguists verify their opinions will be givena great deal of attention. Some have argued that linguists have inhabited the ivorytower of academia for too long. For this reason moving into the rough and tumblearena of the law, where they are required with great rigour to justify what theydo, say and believe, has been a culture shock for many. Understanding of thisculture is critical: some would argue that it is not productive to describe the lawas alien or hostile to the linguistic viewpoint. The law is blind: it has nofavourites and nor should it. It is surely necessary for linguists to accept thisculture and adapt to it, while remaining true to their discipline. In this sectionan attempt was made to illustrate some of the issues linguists face wheninteracting with the legal system. It is now seen as imperative among linguiststhat both they and legal professionals work towards a better understanding of eachotherbs perspective. If linguists claim that lawyers are ignorant of linguistics,then it is up to linguists to ensure that this situation does not continue. Lawyerscan equally claim that linguists are ignorant of the law and it is certainly up tolinguists to ensure that this gap in their knowledge is addressed as a matter ofsome priority. It will also be important for linguists, in this age ofinternational courts, to understand the discourse practices of international law,and to familiarise themselves with the customs and mores of other countriesl legalsystems, as Forensic Linguistics moves into a new millennium and an uncertainterrain in a world of organised crime, international terrorism and human rightsabuses in many countries. It is likely in the future that increasing numbers ofthose seeking to enter the field of Forensic Linguistics will have additionalqualifications in areas such as the law and mathematics and statistics, and to gaingreater understanding of scientific techniques, methods and presentation. Manyuniversities are already equipping their undergraduates with some of thisinformation. With a broad but accurate insight into the law and an appreciation ofhow science is doneb in other fields than their own, forensic linguists of thefuture will have greater means at their disposal than the founding fathers andmothers of the discipline and the future for Forensic Linguistics will be bright.Glossary Authorship attribution: the activity of attempting to assign a particulartext to one or more candidate author (see authorshipm ). Authorship: the processwhereby language is produced by an individual writer or speaker, or by a group ofwriters or speakers. Candidate author: in an authorship inquiry the candidates arethose individuals who are suspected of being the author of a questioned text. Someinquiries have only one candidate author. Inflection: the morphological attachmentof a prefix, suffix or word ending, e.g. -edlpre- as a prefix to a word. In general, if a word cannot take an inflection it isnon#lexical (see lexical wordI). IPA. The International Phonetic Association. Agroup of leading phoneticians who concern themselves with the sound systems of allof the worlds known languages and their transcription. The IPA Chart, whichrecords most of the phones in the worlds known languages, is an essentialreference for anyone interested in or working in phonetics. The IPA Chart can beused for transcribing any language. Lexical word: a lexical word is one which issaid to have meaning, e.g. doga, catx , tabled , as opposed to words like thes ,ofd, etc., which essentially carry the grammar of the language, see non#lexicalwords. Lexical words are also called content wordstinflectione . . See alsoLinguist: one who studies linguistics. However, in some courts a linguist is aninterpreter or translator. Linguistics: the systematic, scientific study oflanguage. Morpheme: the minimum grammatical unit of a language, e.g. in the worddog there are two morphemes, dog and -s. Neither of these morphemes can be reducedfurther. Morphology is the study of the system of morphemes in a language. It isnow largely the domain of those studying syntax (see syntaxi below). Non#lexicalword: a word which is said not to have meaning or content, such as in, for,therefore, etc. There are approximately 250 common non#lexical words in English.
Phoneme: the minimum contrastive sound unit of a language, e.g. /b/, /k/, etc.English has about 44 phonemes depending on the language variety under discussion.Phonetics: the study of the sounds of a language, usually written in phoneticsymbols (see IPA). Phonology: the study of the sound system of a language.Plagiarism: the activity of using other peopleoriginate a text, particularly where the plagiarist intends to pass the work off ashis/her own. Plagiarist: one who plagiarises (see plagiarism). Pragmatics: thestudy of the application of speaker-addressee context, as well as externalcontextual factors, to communication, closely related to semanticsx (see below).Questioned text: a text whose authorship is unknown, or whose authenticity isdoubted. Semantics: the study of meaning, either theoretically or empirically.Semantics deals with truth conditions, sense and reference and the (broadly)metaphoric relations between words (such as metonymy, meronymy, etc). Syntax: thestudy of the grammar of the sentence. In traditional syntax, as proposed by NoamChomsky and his followers, the focus of the study (and much of linguistics at thetime) was to understand how native speakers of a language acquired competence inthe language, by which was meant the ability to generate and understand an infinitenumber of sentences. Text Measures: attributes of a text which can be measured,e.g. text length (the number of words in a text), lexical density, the number oflexical words divided by the total number of words, etc. Annotated Bibliography andReferences Baldwin J.R. and P. French. 1990. Forensic Phonetics. Pinter Publishers,London. [The authors, practising phoneticians, describe the rôle of phonetics incriminal inquiries and as evidence in the courtroom. They illustrate voiceidentification with a number of cases. A very readable book, though regrettably nowout of date.] Cornu G 1990. Linguistique Juridique. Montchrestien, Paris. [A goodintroduction to Forensic Linguistics in French. In this book the emphasis is on the language of the law.] Cotterill J: 2003. Language and Power in Court: ALinguistic Analysis of the O.J. Simpson Trial. Palgrave Macmillan. [This is adetailed account of the trial of OJ Simpson from a linguistic point of view. Theauthor, a professor of linguistics at Cardiff University in Wales, and a practisingforensic linguist of international reputation, discusses the trial from manylinguistic perspectives. The chapter on narrative discourse in the courtroom isparticularly interesting. Cotterill considers research from closely relateddisciplines including the law itself, sociology, criminology, anthropology andpsychology.] Coulthard M. 1994. On the use of corpora in the analysis of forensictexts. The International Journal of Speech Language and the Law. 1 (1), 27-43. [Inthis article Professor Malcolm Coulthard, discusses his approach to the DerekBentley statement (see History and development of Forensic Linguistics to thepresent , this article). He illustrates the use of a corpus and how to compile aconcordance to illustrate systematicity in what is supposed to be a spontaneouslydictated statement. See also: Coulthard M. 1992. Advances in Spoken DiscourseAnalysis. Routledge, Chapter 12: Forensic Discourse AnalysisS (pp 242-254).]Coulthard M and J Cotterill. 2006. Introducing Forensic Linguistics. Routledge. [Inthis book the authors provide an overview of the interface of language and the law,illustrated with authentic data and contemporary case studies. Topics includecollection of evidence, discourse, courtroom interaction, legal language,comprehension and Forensic Phonetics.] Eades D 1994. Forensic linguistics inAustralia: an overview. International Journal of Speech Language and the Law 1 (2)113-132. [As the title says, the author a leading forensic linguist practisingmainly in Australia gives a broad picture of Forensic Linguistics in Australia,its beginnings, and details a number of important cases. A key issue in Australiahas been the involvement of linguists in analysis of police statements by speakersof Aboriginal English.] Gibbons J (1996). Distortions of the police interviewrevealed by videotape. International Journal of Speech Language and the Law 3 (2)289-298. [In this journal article John Gibbons shows how the sequence and contentof questions in a police interview structures that interview, and in some casesleading to a distorted representation of an incident and the intervieweee sinvolvement. See also Gibbons J 1994 Language and the Law, Harlow, Longman. This isGibbonsm classic book on Forensic Linguistics.] Goddard C. 1996. Can linguists help
judges know what they mean? Linguistic semantics in the court-room. InternationalJournal of Speech Language and the Law 3 (2) 250-272. [This journal article, by asemantician, writer and theorist, examines how judges can interpret statutory legalterms in ways that ordinary people can understand what they mean. The author arguesagainst linguists being admitted to courts as experts on the interpretation of thelaw because lexical semantics is as yet not sufficiently developed.] Grice HP 1975.Logic and Conversation. In Cole, P & Morgan J (eds.) Syntax and Semantics Vol 3,41-58. [This is the classic article by Grice on logic in conversation: in it heintroduces his famous Conversational Maxims. What is not perhaps always realised isthat Grice is essentially a philosopher and that his maxims are a philosophicalattempt at the logical structure of meaning. As such they rely on introspectionrather than actual data. Essential knowledge for anyone interested in speech acttheory.] Hale S. 1997. Clash of world perspectives: the discursive practices of thelaw, witness and interpreter. International Journal of Speech Language and the Law4 (2) 197-209. [The writer is a practising interpreter, writer and teacher. In thisjournal article she discusses the different viewpoints that lay people and lawyersbring to a meeting, based largely on the fact that lawyers, especially in court, donot have the same discourse practices as the rest of us. The problem is exacerbatedin the case of a non#native speaker witness.] Jackson BS (1995) Making Sense inLaw. Liverpool. Deborah Charles Publications. [In this book, Bernard Jackson auniversity teacher of law shows how a core linguistic evente in the law is thenarrative. He analyses the presentation of narratives in court and how thestructure and sequence of narratives prosecution and defence influences thelegal process.] Kniffka H. 1981, in G PeuserSt. Winter (eds), AngewandteSprachwißenschaft: Grundfragen Berieche Methoden, pp 584-633. Der Linguist alsGutachter Bei Gericht. Überlegungen und Materialien zu einer AngewandteSoziolinguistike. [The author, Hannes Kniffka, is a well known writer, teacher andpractitioner in Forensic Linguistics in Germany. In this paper he reviews theapplication of linguistics to the court arena and the rôle of the linguist asexpert. A must read for those interested in the practice of Forensic Linguistics inGermany.] Kniffka H. 1990 (Ed). Texte zu Theorie und Praxis forensischerLinguistik. Tuebingen. Niemeyer. [See previous citation. However, this paperfocuses on theoretical aspects of Forensic Linguistics.] Kniffka H. 1994.Understanding misunderstandings in court: La Serva Padrona phenomena and othermis#communications in forensic interaction. Expert Evidence: The InternationalDigest of Human Behaviour Science and Law 2 (1), 164-175. [In this paper the author(see previous two citations) discusses a very interesting academic authorship caseinvolving two sisters, law students, who were accused of plagiarising theirdissertations.] Künzel H und Ulrich Eysholdt, 1992. Der Einfluß von Alkohol aufSprache und Stimme. Heidelberg. Kriminaliste Verlag. [In this paper Hermann Künzeland Ulrich Eysholdt, two well known phoneticians, teachers and writers practisingin Germany, give detailed information about the effects of alcohol on speech outputand voice characteristics.] Kuenzel, H.J. Sprechererkennung: Grundzuege forensicherSprachverarbeitung. Heidelberg: Kriminalistik Verlag, 1987. [Introductory text onForensic Phonetics in German.] Hollien, H. The Acoustics of Crime. New York:Plenum, 1990. [An introductory text on Forensic Phonetics.] Lentine G and R Shuy1990. Mc-. Meaning in the marketplace. American Speech, 65(4), 349-366. [GenineLentine and Roger Shuy were involved in the classic trade mark dispute between afamous hotel chain and a famous fast food chain. In their paper they discuss howthey researched the uses of Mc , and how they attempted to persuade the courtthat, by its very nature, McM could not be the property of a commercial concernsuch as Macdonaldnst.] Levi J. Language as evidence: the linguist as expertwitness in North American Courts. International Journal of Speech Language and theLaw 1 (1), 1 26. [In this article Professor Judith Levi summarises the kinds ofcase in which linguistics has played a rôle in US courts. She discusses such casesin which linguistic areas as diverse as phonology, morphology and dialectology haveformed part of the argumentation, in social contexts such as historicaldialectology and trademark law. In one case a group of recipients of Family Aidbrought a class action against the Illinois Department of Public Aid based on
comprehensibility issues in a notice sent to aid recipients. Levi s article alsodiscusses the application of other areas of linguistics to criminal and civil workincluding semantics, syntax and pragmatics (see Glossary below).] McMenamin, G.(2002). Forensic Linguistics: advances in forensic stylistics. Boca Raton: CRCPress LLC. [In this book Gerald McMenamin offers a readable and down#to#earthexplanation of the work of the forensic linguist for the specialist andnon#specialist alike. He concerns himself, and his reader, with such basicquestions as What is linguistics?h, What is language?e. How is languageacquired? . McMenamin goes on to discuss a wide variety of authorship markers andthe measurement of probability in authorship attribution exercises.] Nolan, F.J.The Phonetic Bases of Speaker Recognition. Cambridge: CUP, 1983. [A classic buthighly readable text on Forensic Phonetics.] Olsson John 2004. ForensicLinguistics: An introduction to Language, Crime and the Law. London, Continuum. [Inthis book John Olsson (the author of this article) a practising forensic linguistin the United Kingdom, writer and teacher, and Adjunct Professor at NebraskaWesleyan University, Lincoln, Nebraska, outlines the main areas of ForensicLinguistics, especially authorship. The book a practical guide tends to focuson linguistics in criminal investigations and in the courtroom, with less emphasison the language of the law. There is also a chapter on phonetics.] Rose, P.Forensic Speaker Identification. London: Taylor and Francis, 2002. [A recent texton Forensic Phonetics.] Shuy R 1997. Ten unanswered questions about Miranda.International Journal of Speech Language and the Law 4 (2) 175-196 (see Lentine andShuy, above). [The author, Professor Roger Shuy, shows how the language of Mirandais far from simple and can be interpreted in a number of different ways. Hequestions the assumption that it is sufficient to read an individual his/her rightsfor that person to understand what rights he/she actually has. Professor Shuy hasprovided linguistic input in a wide variety of cases, both criminal and civil. Twoparticularly famous cases were those relating to (i) John de Lorean and (ii)McDonaldrsy Corporation vs. Quality Inns International.] Solan LM 1993. TheLanguage of Judges. Chicago. University of Chicago Press. [Lawrence Solan isProfessor of Law at the prestigious Brooklyn Law School. He distils his experiencesof a lifetime of dealing with and listening to judges. This is a key text in understanding the interfacebetween linguistic theory and the law.] Stratman JF and P Dahl. 1996. Readersacomprehension of temporary restraining orders in domestic violence cases: a missinglink in abuse prevention? International Journal of Speech Language and the Law 3(2) 211-231. [In this journal article the authors researchers, writers andteachers on legal issues and their comprehensibility to lay people show how anapparently straightforward legal instrument, the temporary restraining order, isopen to misinterpretation.] Walsh M 1994 (217-233) Interactional Styles in theCourtroom: an example from Northern Australia, in Gibbons J (ed) Language and theLaw, London, Longman. [In this article in John Gibbonso classic work on Languageand the Law (see reference above), Michael Walsh, a linguistic fieldworker, teacherand writer in Australia, shows how the different interactional styles of lawyersaand Aboriginal people, works to the disadvantage of Aboriginals.]In this article I first divide Forensic Linguistics into three sub-disciplines :the language of written legal texts, the spoken language of legal proceedings, andthe linguist as expert witness and then go on to give a small number of examples ofthe research undertaken in these three areas. For the language of written legaltexts, I present work on the (in) comprehensibility of police cautions and ofjudges instructions to juries. For the spoken language of legal proceedings, Ireport work on the problems of interpreted interaction, of vulnerable witnesses andthe need for more detailed research comparing the interactive rules in adversarialand investigative systems. Finally, to illustrate the role of the linguist asexpert witness I report a trademark case, five different authorship attributioncases, three very different plagiarism cases and I end reporting briefly the
contribution of linguists to language assessment techniques used in the linguisticcclassification of asylum seekers.THE LINGUIST ON THE WITNESS STAND: FORENSIC LINGUISTICS INAMERICAN COURTSPETER TIERSMA LAWRENCE M. SOLANLoyola Law School Brooklyn Law SchoolIt is becoming increasingly common for linguists to testify as expert witnesses inboth civiland criminal trials. Often linguistic expertise is clearly helpful to the judge orjury. Based onpublished judicial opinions, from which we draw our data, it appears that courtshave allowedlinguists to testify on such issues as the probable origin of a speaker, thecomprehensibility of atext, whether a particular defendant understood the Miranda warning, and thephonetic similarityof two competing trademarks. In other areas the admissibility of linguistictestimony has beenmore controversial, including author and speaker identification, discourseanalysis, the meaningof legal texts, and the comprehensibility of jury instructions. Reasons forjudicial reluctance toadmit linguistic expertise include concerns that it is not sufficiently reliable,the belief that issueslike the meaning of a text can just as well be decided by a jury, and sometimeseven institutionaland political considerations. Despite such reservations, courts generally recognizethat there is aplace for linguistic expertise in appropriate cases.*Linguists are appearing with increasing frequency as expert witnesses in Americancourtrooms. Nonetheless, in many cases where one side or the other wishes topresentlinguistic evidenceeither through testimony or some other meansthe judge refusesto admit it. This raises questions of why courts are more receptive to linguisticexpertisein some types of cases than in others, and when they ought to accept linguists asexperts.Linguistic issues can arise in a great variety of legal contexts. Specific subjectsthatlinguists may address include the likelihood of confusion in trademark cases;miscommunicationbecause of dialect differences; the comprehensibility of legal documents;the meanings of statutes, wills and contracts; the identification of authors andspeakers;the ability of jurors to understand instructions, or of an arrested person tocomprehendthe Miranda warning; and many more. Almost any area of linguistics can be relevantin court. Phonetics, for example, is important in trademark cases where the soundsimilarity of two names is in question, as well as in speaker identification.Discourseanalysis has been used to help jurors understand covertly recorded conversations incriminal cases. Syntax, semantics, and pragmatics are all relevant when the meaningof legal documents is at issue. Smaller subfields of linguistics may also haverelevancein a legal dispute. Thus, a dialectologist might help identify the place of originof aspeaker on a tape-recorded bomb threat.In this article, we will review some of the substantive legal areas in which
linguisticexpert testimony has been admitted, and others in which its admissibility has beenmorecontroversial. We will not focus on reports by forensic linguists themselves.JudithLevi (1994) has already published an overview of cases in which linguists havereportedon their experiences in court, and Roger Shuy (1993, 1998) has written two booksdescribing some of his experiences.1 In contrast, we draw our examples frompublished* We would like to thank Darren Nakano and Leslie Nathan, former students of LoyolaLaw School in LosAngeles, and former Brooklyn Law School students Robyn Schneider and Amy Sender fortheir assistance.We are also grateful to Roger Shuy and the Language referees for many helpfulcomments. This researchwas supported in part by a summer research stipend from Brooklyn Law School and theJoseph ScottFellowship from Loyola.1 In addition, several anthologies contain reports of linguistsn experiences asexperts. See Gibbons 1994,Rieber & Stewart 1990, Levi &Walker 1990, and Cotterill 2002. Moreover, the journalForensic Linguistics,now in its eighth year of publication, contains such reports.221222 LANGUAGE, VOLUME 78, NUMBER 2judicial opinions in the United States. In almost all American jurisdictions, onlypublishedopinions can serve as precedents.2 A court deciding whether to permit alinguistto testify is most likely to turn for advice to the record of published judicialopinions.We therefore think it is useful, for both the legal and linguistic communities, toinvestigateand report on that record.Our legal research found over one hundred published judicial opinions, not countingvoiceprint cases, in which language experts were mentioned. (We deal with the issueof voiceprints separately below.) Most of these are decisions by federal and stateappellatecourts. Often, the party who offered the testimony of an expert witness at trialappealed the trial court s ruling not to allow such testimony. Because appellatereviewof evidentiary decisions under American law is very deferential to the trial court,suchrulings are often upheld on appeal (Weinstein & Berger 1998). One might thereforeget the impression that linguists only seldom participate in the judicial system,and arelargely unwelcome. But this would be a serious misinterpretation of the facts forthreereasons.First, and perhaps most important, our analysis of the published opinions showsthaton some types of issues, testimony by linguists occurs virtually withoutcontroversy,while on others, courts are far more likely to reject the offer of expertlinguistic testimony.It is on these distinctions that we wish to focus this article, first describingtheissues, later offering explanations for some of the more salient discrepancies.Second, accounts by linguists themselves (e.g. Levi 1994), show active
participationby linguists in many areas of the legal system. For example, in his book LanguageCrimes, Roger Shuy (1993:xx) notes that he had consulted in over two hundred casesand had testified in about thirty-five cases. While Shuy is probably the mostactiveAmerican linguist in terms of legal consulting, these figures suggest thatpublishedopinions tell only part of the story.Third, the number of published appellate opinions is quite small in relation to theoverall number of cases in the judicial system. For example, 1998 data from theUnitedStates Department of Justice show that more than three hundred thousand cases wereterminatedr in the federal trial courts (district courts) that year, after atrial, a motionto dismiss, settlement, and so on.3 In that same year, about fifty-two thousandappealswere terminated. This number suggests that most cases are never appealed. Of thecompleted appeals, only about twenty-five thousand were decided on the merits, andof those, only about six thousand resulted in published opinions by the UnitedStatescourts of appeals. This means that only about 11 percent of appellate court casesgeneratedpublished opinions.4 Moreover, comparing the total number of cases terminatedin the federal trial courts in 1998 (300,000) with the total number of publishedopinionsby the federal appellate courts in that same year (6,000), one is forced toconclude thatthe percentage of federal cases that result in a published appellate opinion isquite2 In our understanding, this is not true in England, where a case can function as alegal precedent evenif it was never published.3 The data in this paragraph are taken from a United States Department of JusticeReport, Judicial Businessof the United States, published at www.uscourts.gov/dirrpt98/index.html (pp. 16,29, 54). The United StatesDistrict Court is the trial level court in the federal judicial system.Corresponding data from state courtsvaries from state to state. For example, in California, roughly 9 percent ofmajority opinions issued by thecourts of appeal are published (Judicial Council of California, 1997 JudicialCouncil Report on Court Statistics,Table 9, p. 29).4 The actual percentages are even smaller since it is not unusual for certaincomplex cases to generatemultiple opinions. These rough estimates, however, are good enough to serve ourpurposes here.THE LINGUIST ON THE WITNESS STAND 223small, almost certainly under ten percent, and probably well under five percent ofalllower court cases.In this context, the presence of more than one hundred published judicial opinionsthat deal with linguistic expertise implies substantial participation by linguistsin thelegal system. Only those cases in which the linguisti s testimony was controversialwould ever result in an opinion that addresses the admissibility of linguisticexpertise.Of those controversial cases, only a relatively small percentage results inopinions that
are published. Assuming quite conservatively that each published appellate opinionrepresents at least ten trial court cases, the more than one hundred publishedjudicialopinions that mention linguistic experts may represent very roughly a thousandtrialcourtcases in which linguistic expertise was involved in some way.Another way to address the issue is to look for published opinions making referenceto linguistic experts in a single year. We found eleven such cases in 1998: twofromthe United States Courts of Appeals; five from the United States District Courts;andfour from state appellate courts. Extrapolation from these figures suggests thatlinguistsmost likely played a part in approximately one hundred cases in just one year.Obviously,these quantitative inferences are quite coarse. Moreover, the relative noveltyof linguists on the witness stand may result in more appeals and anoverrepresentationof such cases in the reported appellate opinions; if so, our estimates would haveto beadjusted downward. Nonetheless, the data strongly suggest that there has beenconsiderableparticipation by linguists in the American legal process.Despite these indications of an increasing presence of expert linguists in Americancourtrooms, our own perceptionboth of us were practicing lawyers before becominglegal academicsis that the vast majority of American lawyers and judges havelittleor no experience with linguistic expertise in a legal matter. Many have never evenheard of it. This suggests that it is not self-evident to lawyers and judges thatlinguistscan be of help in resolving legal disputes. We hope this article will clarify forbothlinguists and members of the legal community just where linguistic expertise can behelpful.We will first describe the evidentiary standards under which courts are supposed todecide whether to permit expert testimony in general. It appears to us thatlinguisticexpert testimony meets these standards in a wide range of cases. But we will showthatwhile courts are often very receptive to testimony by linguists, they shy away whenitconflicts with certain beliefs about language and cognition deeply entrenched inthelegal system. In some instances, these observations suggest that expert testimonyfromlinguists would be better received if it were tailored to meet the needs of thelegalsystem. In other instances, they suggest that it may be time for the legal systemtoreexamine some of its long-standing tenets about the nature of language.1. THE ADMISSIBILITY OF EXPERT EVIDENCE IN AMERICAN COURTS.5 Until the FederalRules of Evidence were adopted in 1975, the predominant standard for theadmissibilityof expert testimony was the Frye test, named for a 1923 United States Court ofAppealsdecision, Frye v. United States (293 F. 1013 (D.C. Cir. 1923)). Frye involved atrialcourtis refusal to admit the results of a lie detector test (called a systolicblood pressure
deception test ) offered through an expert to prove the defendant s veracity in amurdercase. The court of appeals affirmed the trial courtstandard5 For an expanded version of this section, addressed primarily to a legal audience,see Solan 1999.224 LANGUAGE, VOLUME 78, NUMBER 2for admissibility that was routinely followed for some fifty years, not only byotherfederal courts, but by many state courts as well. The court held:Just when a scientific principle or discovery crosses the line between theexperimental and demonstrablestages is difficult to define. Somewhere in this twilight zone the evidential forceof the principle mustbe recognized, and while courts will go a long way in admitting expert testimonydeduced from awell-recognized scientific principle or discovery, the thing from which thededuction is made must besufficiently established to have gained general acceptance in the particular fieldin which it belongs.(293 F. at 1014).Despite its wide acceptance, the Frye standard came to face increasing criticism asbeing too tough. Science often involves controversy, and many significantscientifictheories never gain general acceptance. The critics argued that it would make moresense for courts to permit the trier of fact (the judge, or the jury in a jurytrial) to heararguments on both sides of a controversial issue and to weigh the evidence, ratherthanto preclude the jury from hearing the evidence at all. Defenders of a morerestrictivetest, then as now, worried about junk science (see Huber 1991).In 1975 the Federal Rules of Evidence came into effect. Rule 702, which governsthe admissibility of expert testimony, originally reads as follows:If scientific, technical, or other specified knowledge will assist a trier of factto understand the evidenceor to determine a fact in issue, a witness qualified as an expert by knowledge,skill, experience, training,or education, may testify thereto in the form of an opinion or otherwise.Rule 702 therefore required that expert testimony must help the judge or juryunderstandthe evidence or decide what happened. On its face, this is a more relaxed standardthanFryees requirement of general acceptance in the particular field in which itbelongsr.The adoption of Rule 702 did not, however, lead to the immediate demise of theFrye standard in the federal courts. Rather, it led to a period of uncertainty asto justwhat the standard really was. Some federal courts understood Rule 702 as havingreplaced Frye, while others continued to follow Frye, which was deeply entrenchedafter all those decades.This lack of consensus continued until 1993, when the Supreme Court decided Daubertv. Merrell Dow Pharmaceuticals, Inc. (509 U.S. 579 (1993)). The issue in Daubert,a products liability case, was whether Bendectin, an antinausea drug taken duringpregnancy,had caused birth defects in the plaintiffns children. The epidemiologicalliteraturesuggested that it did not. The plaintiffs in Daubert wanted to call experts whowould
attack the inferences drawn from the data in the published literature and bring tobearthe results of animal studies. The trial court had rejected the experts on thegroundsthat their work had not been published, and therefore failed to meet the standardsofscientific acceptance that the courts had developed under Frye. It thus grantedsummaryjudgment to the defendant, Merrell Dow (727 F. Supp. 570, 572 (S.D. Cal. 1989)).The court of appeals affirmed the trial court.1991)).The Supreme Court reversed the decision of the court of appeals, holding that theFederal Rules of Evidence had replaced the Frye standard. It interpreted Rule 702asrequiring courts to engage in a preliminary assessment of whether the reasoning ormethodology underlying the testimony is scientifically valid and of whether thatreasoningor methodology properly can be applied to the facts in issue (509 U.S. at 59293).To be scientifically valid the proffered evidence need not be uncontroversiallyacceptedin the scientific community. Rather, [t]he adjective scientificc o implies agrounding in the methods and procedures of sciencei (509 U.S. at 590).The Court did not attempt to state the conditions that are both necessary andsufficientfor evidence to be scientifically valid. It did suggest, however, four nonexclusiveindicia.THE LINGUIST ON THE WITNESS STAND 2251. whether the theory offered has been tested;2. whether it has been subjected to peer review and publication;3. the known rate of error; and4. whether the theory is generally accepted in the scientificcommunity. (509 U.S. at 593)The Daubert opinion has been the subject of much discussion, often critical. Whatcomplicates the matter is that some states have retained the Frye test, or testssimilarto it. Thus, the case law based on Frye continues to be relevant in many of thosejurisdictions.6 Still, there is no doubt that Daubert has become the leadingopinion inthis area.One question that Daubert left open was whether it applies to testimony that is notstrictly scientific. One could argue, for example, that testimony on a legal texto srangeof possible interpretations is more descriptive than theoretical, and thattherefore theDaubert approach should not apply. In a recent opinion, Kumho Tire Co. v.Carmichael(119 S.Ct. 1167 (1999)), the Supreme Court rejected this type of argument. KumhoTire was a product liability case about automobile tires. In allowing the exclusionofa tire expert whose offered testimony was based on his experience in the industry,theCourt held that the general principles of Daubert apply not only to expertsofferingscientific evidence but also to experts basing their testimony on experience (119S.Ct.at 1173). While the Court admitted that the Daubert factors may not all beapplicablein a given case, it stressed that the overall approach to evaluating reliabilityshould be
followed. Significantly, the Court went on to hold that the key to deciding theadmissibilityof expert evidence is whether the expert employs in the courtroom the same levelof intellectual rigor that characterizes the practice of an expert in the relevantfieldt(119 S.Ct. at 1176). We therefore cannot avoid asking how linguistic testimonystandsup to the Daubert/Kumho Tire factors.Moreover, in response to Daubert and Kumho Tire, Rule 702 has recently beenamended. It now reads:If scientific, technical, or other specialized knowledge will assist the trier offact to understand theevidence or to determine a fact in issue, a witness qualified as an expert byknowledge, skill, experience,training, or education, may testify thereto in the form of an opinion or otherwise,if (1) the testimonyis based upon sufficient facts or data, (2) the testimony is the product ofreliable principles and methods,and (3) the witness has applied the principles and methods reliably to the facts ofthe case.At least in the federal courts, this is the standard against which linguisticexpert testimonywill be evaluated.2. EXPERT LINGUISTIC EVIDENCE IN THE AMERICAN LEGAL SYSTEM. At least in theory,linguistic evidence should fare quite well regardless of the evidentiary standardthat isapplied. Linguistics is a robust field that relies heavily on peer-reviewedjournals fordissemination of new work. Furthermore, much of the expert testimony offered is inkeeping with very basic literature in the field. For example, when a linguist isaskedto testify about a criminal defendant s proficiency in English, the expert hasavailablea number of well-accepted instruments and a great deal of learning on which to basean analysis.6 A United States Supreme Court decision on evidence is binding only on federalcourts, unless it is basedon constitutional grounds. Individual states may choose to follow the Daubertstandard, or continue to followFrye, or may adopt another standard entirely.226 LANGUAGE, VOLUME 78, NUMBER 2It is true, of course, that there is controversy over certain issues in the field.Considerthe English passive construction. Within the linguistics community, there is livelydebateover how passives really work. Are they formed by transforming active sentencesinto passive ones, or are they formed more or less as they appear, with other rulestelling us how to relate them to corresponding active structures? There arelinguists inboth camps. Peter Culicover (1997) is in the former, Joan Bresnan (1978) in thelatter.But almost all of these controversies are entirely tangential to any testimony thatalinguist might give on the range of interpretations available to sentences withpassiveconstructions. The resolution of that debate in favor of one side or the otherwould notaffect an experths testimony. Thus, while there may be disagreement as to WHY weunderstand a given linguistic structure to have a particular range of meanings, the
FACTof the range of meanings should not normally be controversial.For this reason, it seems relatively straightforward that linguistic testimonybasedon the kinds of analyses that linguists use in the scholarly literature should meeteitherthe Daubert or the Frye standard. We present some examples below.2.1. THE MEANING OF LANGUAGES OTHER THAN STANDARD AMERICAN ENGLISH. As thereader will see later in this article, judges are often reluctant to allow experttestimonyon the meaning of a text in ordinary standard English. In contrast, they havelittletrouble with admitting testimony on foreign languages or on nonstandard registers.Ofcourse, this area of knowledge involves language but is not necessarily theexclusivedomain of professional linguists. For that reason, we have not systematicallycoveredsuch cases, of which there are a substantial number.An example of testimony on the meaning of a non-English term is when an expertin Shoshone-Bannock was allowed to testify how the Shoshone might have understoodthe English term hunt in a treaty. The linguist testified that the Shoshone-Bannocktranslations would not systematically distinguish between fishing and hunting, butreferredto gathering wild food in general (State v. Tinno, 497 P.2d 1386 (Idaho 1972)).The meaning of technical language or jargon is also a common legal issue,especiallyin contract cases, and is likewise freely admitted, even if the expert is not atrainedlinguist. Thus a doctor might testify as a medical lexicographerx on the meaningof amedical term (Hagenkord v. State, 302 N.W.2d 421 (Wis. 1981)). An important caveatis that judges will generally not allow linguists to testify on the meaning oflegalterminology, viewing themselves as the experts in this area.Another type of specialized language is argot and code, often associated withcriminalactivity. In one case, a government agent testified that in a particular case, theboyfriend,the boy, transcripts, briefs, a nd motions meant heroin , and that the girl andthembroads meant cocaine (United States v. Simmons, 923 F.2d 934 (2d Cir. 1991)).Thistype of testimony, typically by law enforcement officers, is almost invariablyallowed.Overall, judges seem well aware that they and jurors need assistance inunderstandingforeign words and phrases, technical terminology, and code or argot.2.2. DIALECTOLOGY. Although there are not many published cases, linguists have onoccasion testified on the dialects of English or other languages. In the well knownAnnArbor case, several experts on Black English Vernacular or Ebonics testified aboutthecharacteristics of that variety of English; the court discussed the testimony indetail(Martin Luther King Junior Elementary School Children v. Ann Arbor School Dist.Board, 473 F. Supp. 1371 (E.D. Mich. 1979)). At least one opinion in an employmentdiscrimination case mentions the testimony of a linguist on the value judgmentsthat
can be associated with how one speaks: people often view speakers of nonstandardTHE LINGUIST ON THE WITNESS STAND 227dialects as being less educated and competent (Polk v. Yellow Freight Systems, 801F.2d 190 (6th Cir. 1986)).Moreover, many linguists have expertise in identifying dialects of individualspeakers,which could be quite useful in determining whether a particular defendant madea recorded incriminating or threatening statement. Oddly, we have found only onepublished opinion discussing such evidence (People v. Clarke, 277 N.E.2d 866 (Ill.1971); cf. Labov & Harris 1994). Nonetheless, because most judges and juries havelimited experience in this area, we believe that linguistic expertise on dialectsanddialectology would generally be admitted with little controversy.2.3. COMPREHENSIBILITY AND READABILITY. When the issue has been thecomprehensibilityor readability of texts, courts issuing published opinions have tended to allowexperts on language to testify. Admittedly, one court has ruled that whether theaveragecustomer could understand a standard form collection letter threatening legalactionagainst the recipient was a matter of common sense and did not call for experttestimony(United States v. ACB Sales and Service, Inc., 590 F. Supp. 561 (D. Ariz. 1984)).Thatcase, however, seems to be the exception.More often, testimony concentrating on the comprehensibility of legal documentsthat are directed to the public is admitted. In a case that presaged the 2000presidentialelection, a linguist testified in a Florida case that a ballot containing anamendmentto a county charter was confusing. The trial court, seconded by the court ofappeal,acknowledged the testimony but concluded that most voters would have properlyunderstoodthe ballot (Wadhams v. Board of County Commissioners, 501 So. 2d 120 (Fla.Dist. Ct. App. 1987)). Interestingly, the FloridaSupreme Court reversed. Althoughnotspecifically referring to the linguisttof thevoting public is intolerable and should not be countenancedr (Wardhams v. Board ofCounty Commissioners, 567 So. 2d 414, 418 (Fla. 1990)).7Less commonly, language experts have testified on the comprehensibility of juryinstructions. An example is the case of James Free, who was condemned to death formurder in Illinois (United States ex rel. Free v. Peters, 806 F. Supp. 705 (N.D.Ill.1992); see also Tiersma 1995). Free challenged his death sentence, arguing that theinstructions given to the jury were misleading and obscure. The trial courtlistenedcarefully to various experts, including a linguist, and decided that jurors likelydid notunderstand instructions on how to decide whether to impose the death penalty. Butthecourt of appeals had an almost disparaging attitude towards the research (12 F.3d700(7th Cir. 1993)) and quickly affirmed Free s death sentence. Although there are notmany cases, we would have to conclude that the admissibility of linguistic experttestimony on the comprehensibility of jury instructions is uncertain at best.Later, wewill suggest a reason for this reaction.
2.4. LINGUISTIC PROFICIENCY. Another factual issue that sometimes arises in courtisthe linguistic proficiency of a particular person. For example, it may be unclearwhethera criminal defendant understood the Miranda warning. Or an accused may haveconsentedto a search without fully understanding the implications. Again, there are not agreat many published opinions in this area, but courts have generally allowedexperttestimony on the linguistic competence of a specific individual. In one case, asignlanguage expert was allowed to testify that a defendant had clearly invoked hisright7 See also Coleman v. Block, 589 F. Supp. 1411 (D.N.D. 1984).228 LANGUAGE, VOLUME 78, NUMBER 2to counsel, which meant that the subsequent interrogationwithout a lawyer presentwas illegal (People v. Smith, 37 Cal. Rptr. 2d 524 (Cal. Ct. App. 1995)).It is worth emphasizing at this point that even if testimony is admitted, thefinderof facto can decide how much value to place on it. Thus, in United States v.Gutierrez-Mederos (965 F.2d 800 (9th Cir. 1992)), the court admitted testimony by a linguistsupporting the defendantts claim that when he consented to allow police to searchhispossessions, his cultural background and limited English proficiency prevented himfrom understanding that he had a right to refuse. Although the judge allowed thelinguistto testify, the court gave little weight to the testimony, noting that the linguisthadnever actually interviewed the defendant, and consequently holding that thedefendantusconsent to search was valid. This point applies to all areas in which linguisticexpertisemight be brought to bear: the judge or jury can give it as much or as little weightasthey feel appropriate.2.5. LINGUISTIC ISSUES IN TRADEMARK CASES. Linguistic expertise, and especiallyphonetics, has also been relevant to legal proceedings in trademark law. Thecriticalquestion in many trademark cases is whether the name that a person or company isusing for a product is confusingly similar to an existing trademark. Often thatdependson the phonetic similarity between the two marks. Trademarks that have been foundto be confusingly similar include Beck s Beer and Ex Bier; Comsat and Comset;Diapareneand Dyprin; Dramamine and Bonamine; Listerine and Listogen; Lorraine and LaTouraine; Smirnoff and Sarnoff (McCarthy 1992:§23.6(1)).Trademark law has long relied on expert testimony, which often includes surveysof whether people are actually confused by the similar names. Therefore, it is notsurprising that courts have almost universally and without extensive commentadmittedtestimony by linguists in this area, at least as reflected in the publishedopinions. Forexample, if one company has a valid trademark in the name Aveda and a secondcompany begins to use the name Avita on similar products, the second company maybe infringing on the trademark of the first. In this particular case, a professorof Englishand linguistics testified via an affidavit that intervocalic t and d are oftenpronounced
the same in these words, and that the middle vowels (e and i) may also be. Thus themarks may beor may almost behomophones. Based in part on this evidence, thejudge granted an injunction that prohibited the defendant from further use of Avita(Aveda Corp. v. Evita Marketing, Inc., 706 F. Supp. 1419 (D. Minn. 1989)).Linguistic expertise has been brought to bear on other trademark issues as well. Isthe Mc- in McDonaldes genericand thus a relatively productive morphemeor is itspecifically identified with the McDonaldrs Corporation? One court receivedextensivelinguistic testimony on both sides, which concentrated on the use and meaning ofMc- inmany ordinary texts. The court eventually decided that Mc- was tied to theMcDonaldisCorporation, rather than being an ordinary English prefix. Thus, McDonalde s wasentitledto protect its morpheme from being used by acha in of inexpensive McSleepmotels (Quality Inns International, Inc. v. McDonald s Corp., 695 F. Supp. 198 (D.Maryland 1988)).3. PROBLEM AREAS: JUDICIAL RELUCTANCE TO ADMIT EXPERT TESTIMONY BY LINGUISTS.Although the legal system has often welcomed linguistic expertise, there area number of areas in which they are more hesitant to do so. One example is the useoflinguistics to identify authors or speakers. Courts sometimes question whetherlinguistictheory is able to meet the standards of Frye or Daubert with respect to theseareas.Other uses of linguistics that have not always been accepted by the courts arediscourseTHE LINGUIST ON THE WITNESS STAND 229analysis and testimony on the meaning of contracts and statutes. In both cases,courtsare reluctant to admit linguistic evidence not only because of the evidentiaryissuesraised by the Frye and Daubert requirements, but also because they fear thatlinguisticexpertise might usurp the role of the judge or jury. We conclude our overview ofproblem areas with a discussion of the comprehensibility of jury instructions,wherecourts tend not to accept linguistic expertise because it might challenge thelegitimacyof an important legal institution.3.1. DISPUTED AUTHORSHIP. Recent discussion in the press of the JonBene´t Ramseymurder investigation in Colorado has brought into the public spotlight the issue ofusing linguistic techniques to identify authors. Unfortunately, the publicity hasnot beengood for those who believe that linguistic analysis can shed light on this problem.Early in the morning of December 26, 1996, Patricia Ramsey reported to authoritiesthat her six-year-old daughter, JonBene´t, was missing from their Colorado home,andthat she had found a ransom note in the house.8 Later that day, JonBene´to s bodywasdiscovered in the basement. To date, no one has been charged with the child smurder,although theories abound in the numerous books, articles, and web discussions thatthecase has generated. Some of these theories allege that the parents were involved.A critical question is who wrote the ransom note. Professor Donald Foster, anEnglishprofessor at Vassar College who is well known for using stylistic analysis toidentify
authorship in both literary texts and legal cases (Foster 2000), first attributedthe noteto someone who did not write it. He wrote aletter to JonBene´t) s mother assertingher innocence.9 Later, after examining additional materials, Professor Fostercandidlychanged his position and determined that Mrs. Ramsey had written the note. Suchincidents help to justify the lawes concern about methodology. Regardless of howonelooks at the merits of the Ramsey case, if forensic stylisticsRidentificationis to comply with evidentiary standards, it must use scientifically validatedtechniques.Although it has generated a fair amount of scholarly discussion (e.g. Finegan 1990and McMenamin 1993), analysis of texts in cases of disputed authorship has not ledto a great many published cases. In one such case (United States v. Clifford, 704F.2d86 (3d Cir. 1983)), the court rejected testimony about a Forensic Linguisticmethodof handwriting and stylistic analysis, which it deemed of questionable reliability.Itheld that the jury could reach its own conclusions from the samples providedwithoutthe assistance of experts.Clifford shows a peculiar hole in the American system of evidence. The issue inthatcase was whether the defendant was the author of a threatening note. The governmenttried to offer the testimony of a forensic linguist expert in stylistic analysis.The courtsummarized his testimony as follows:Dr. Miron testified that forensic linguistic analysis is the process of matchingstylistic similarities indifferent documents and then of assigning weight to those similarities according totheir distinctivenessand frequency of occurrence. He further stated that such an analysis could notprovide a positive meansof identifying the author of an anonymous document. He indicated that the resultsof forensic linguisticanalysis could be probative in establishing authorship but could not prove that oneperson, to theexclusion of all other possible authors, had written a document. (704 F.2d at 88)8 The ransom note can be found at http://www.jameson245.com/ransomnotep1.jpg.9 The text of the letter is available at various sites on the web. It was alsoreported in the press. See LisaLevitt Ryckman, Book details linguistic scholarRocky Mountain News, April11, 2000. For Foster s account, see Foster (2000:1617).230 LANGUAGE, VOLUME 78, NUMBER 2This was enough to keep the testimony out under the Frye standard, and, for themostpart, under Daubert as well. On appeal, however, it was determined that thedocumentsthemselves should be admitted and that the jury should make its own determinationoftheir similarities.This, in our opinion, is a remarkable state of affairs. If forensic testimony isprecludedbecause the expert cannot form a definitive opinion, then jurors, not knowing howtoevaluate what they see, will be forced to reach conclusions that are even less
supportedby the evidence. At the very least, experts should be permitted to assist jurors byadvising them of the risk that they might make too much of the similarities betweendocuments.A more recent case, United States v. Van Wyk (83 F. Supp. 2d 515 (D. N.J. 2000)),illustrates why courts are wary of permitting experts in forensic stylistics toidentifyauthors. The court there permitted an FBI agent with training in forensic documentidentification to testify about the similarities between certain threateningletters andthe defendantns known writings, but did not allow him to offer an opinion aboutauthorshipof the documents whose origin was in dispute. As the Van Wyk court explained,we do not have good corpora and appropriate analytical tools that permit inferencesas to how likely it is that a document was produced by a particular person.Although Fitzgerald [the FBI agent offered as a stylistics expert] employed aparticular methodologythat may be subject to testing, neither Fitzgerald nor the Government has been ableto identify a knownrate of error, establish what amount of samples is necessary for an expert to beable to reach a conclusionas to probability of authorship, or pinpoint any meaningful peer review.Additionally, as Defendantargues, there is no universally recognized standard for certifying an individual asan expert in forensicstylistics. (83 F.Supp.2d at 522)10It is not unreasonable to argue that allowing an expert to point out similaritiesanddifferences between a defendant s known writings on the one hand, and questionedwritings in the case on the other, is better than simply giving the documents tothe jurywithout any guidance whatsoever, leaving the jurors to their own devices. Yet thissolution is far from ideal. Because jurors will not know how significant thesimilaritiesand differences between the two sets of documents really are, they can easily reachthe wrong conclusion. In fact, this absence of baseline information is the veryreasonthat expert opinion testimony was not allowed. Moreover, in Van Wyk, the expert wasan employee of the FBI, who was called to point out the similarities between thedisputed documents and documents known to be authored by the defendant. It seemsunfair in criminal cases to permit the government to offer expert testimony basedoninsufficient baseline data, and then to place upon the defendant the burden torebut thatevidence through cross-examination or by calling another expert in forensicstylistics.We believe that ultimately the only responsible solution is for the linguistic andlegal communities to work together toward developing techniques from which reliableinferences about authorship can be drawn. This task will inevitably require thecollectionof large corpora of informal written language and sophisticated computer programstoanalyze the data. Fortunately, such work is underway, both by government and byacademic linguists (see Chaski 2001, McMenamin & Brengelman 2000). In the shortterm, if courts are not inclined to keep the documents from the jury, linguistsshouldbe permitted to testify about the dangers of drawing unwarranted conclusions.3.2. PHONETICS AND THE PROBLEM OF SPEAKER IDENTIFICATION. There are a numberof areas in which phonetic expertise might prove relevant to judicial decision
making.10 The methodology to which the court refers is from McMenamin (1993).THE LINGUIST ON THE WITNESS STAND 231Deciphering or enhancing speech on tape or other types of recordings is one sucharea.Here, the cases have also been divided. Some courts have refused to allow linguiststotestify because hearing is within the ability and experience of the trier of facts(BeechAircraft Corp. v. United States, 51 F.3d 834 (9th Cir. 1995)). At least one court,however,has allowed expert testimony in deciphering something that was said on a recording(Weiss v. Glemp, 792 F. Supp. 215 (S.D.N.Y. 1992)). Interestingly, the expert inboth cases was Roger Shuy.Our view is that phoneticians and other linguists who deal with recorded speech ona regular basis are highly qualified to enhance the sound on tapes or to offer anopinionon what is being said. When the parties are willing to offer such expertassistance, itseems absurd to leave the issue to a lay jury that has never been trained to workwithtape-recorded information.A more controversial application of phonetics is the use of what are calledVOICEPRINTS, a form of spectrographic analysis, to identify a tape-recorded voiceasbelonging to a particular speaker (typically, the defendant in a criminal case).Most ofthe actual analysis is done not by linguists, but by a small group of policeofficers andtechnicians who have been trained for this specific task, and who typically havelimitedbackgrounds in acoustics or phonetics. The main issue, however, is whether themethodologyproduces sufficiently reliable results. This is a critical question, becausevoiceprint evidence may be pivotal to a case.During the late 1960s and throughout most of the 1970s, American jurisdictionswere divided on the issue of spectrographic evidence. Many courts determined thatthemethodology was not accepted as reliable within the scientific community, and theywould therefore not admit it. Other appellate courts reached the oppositeconclusion,holding that it was within the discretion of the trial judge to admit the testimonyandthatas alwaysit was up to the finder of factemuchweight to give it.An influential report by the National Research Council in 1979 questioned theabilityof voiceprints to produce accurate results under forensic conditions withsufficientlylow rates of error (Bolt 1979). One member of the committee that wrote the reportwasProfessor Oscar Tosi, whose earlier experimentation was influential among courtsthataccepted voiceprint analysis (e.g. Tosi et al. 1972).Surprisingly enough, throughout the 1980s and 1990s the published opinions have,albeit in smaller numbers overall than before, continued to be split on the issue.Asrecently as 1999, the Alaska Supreme Court held that voiceprint evidence was
admissiblein a case involving a man accused (and later convicted) of making terroristtelephonecalls (State v. Coon, 974 P.2d 386 (Alaska 1999)). Similar decisions have beenreachedduring the past two decades or so by the United States Circuit Court of Appeals forthe Sixth Circuit, the Seventh Circuit, a federal district court in Hawaii, thesupremecourts of Ohio, Massachusetts, and Rhode Island, and a lower court in New York. Butduring roughly the same time, voiceprints were held inadmissible by the high courtsof Arizona, Colorado, Indiana, Louisiana, and New Jersey.11 Clearly, the courts areinwidespread disagreement on this issue, even though many linguists have expressedserious doubts about the reliability of this approach in a forensic setting.Indeed, onephonetician has called it a fraud being perpetrated upon the American public andtheCourts of the United Statesa (Hollien 1990:210).11 For a list of the cases, see footnote 60 of Rafferty (1999).232 LANGUAGE, VOLUME 78, NUMBER 2A careful reading of the cases shows that courts that employ the Frye standard aremore likely to reject voiceprints than are those who employ the Daubert standard.12Under Frye, the only issue, other than relevance, is whether voiceprints aresufficientlyestablished to have gained general acceptance in the particular field in which[they]belongi (293 F. at 1014). They are not. Even a cursory review of the phoneticliteraturedemonstrates substantial controversy over their usefulness and validity as aforensictool.13In contrast, it is easier (but by no means necessary) under the Daubert standard tojustify admitting voiceprint evidence. As the Alaska Supreme Court recently notedinState v. Coon (974 P.2d 386, 40001 (Alaska 1999)), some published reports supporttheir use. Because the philosophy of the Federal Rules, under which Daubert wasdecided, is to treat controversy by presenting both sides of an argument (ratherthanby excluding evidence altogether), the Alaskan court placed as much or moreemphasison studies sponsored by police officials that advocate for the use of voiceprintsas itdid on publications from the independent scientific community (974 P.2d at 402).A leading treatise on scientific evidence in American law observes a certainincongruityin the debate over the admissibility of spectrographic evidence:The refusal of some courts to admit talker identification expert evidence is anexception to the traditionalreceptiveness of the courts to forensic individuation techniques. Why has talkeridentification beentreated differently? Several interconnected explanations are plausible . . . .Numerous courts evaluating talker identification expertise were critical ofwitnesses testifying onbehalf of the technique who were mere technicians rather than educated scientists;or whose livelihoodsdepended upon continued admission of the technique; or who came from a very smallcircle of proponentsof the technique.Another factor is that the literature of scientific talker identification, both
supporting and questioningthe technique, was more quantified and qualified than earlier courts had receivedabout earlier forensicindividuation techniques. This is because most of the people involved in talkeridentification came fromfields that had a tradition of empirical testing of their ideas. Indeed, moreresearch was available to thecourts about talker identification expertise than for any forensic individuationfield that preceded it.This immediately provided the courts with unusual resources with which tocomprehend the shortcomingsof the technique. (Faigman et al. (1997:19293))The authors of the treatise then observe that this creates a paradox:All else equal, the better a field studies and critiques itself, the more skepticalthe courtsq impressionof it is likely to be. The less well a field understands what it is doing and whatthe limits are on itsown capabilities, the more positive an impression the courts will develop of thefield. For a number ofthe more conventional forensic individuation techniques, there still is notradition of self-scrutiny or aliterature reporting the results of rigorous testing which can inform the courts.At least in terms of theircontinued acceptance by the courts, those fields have nothing to gain and much tolose by adopting atradition of inquiry, testing, and skepticism. (Faigman et al. (1997:193, n.22))To give just one example of the kind of evidence that Faigman and his colleaguescriticize, consider the science of identifying individuals from bite marks thatthey leave.Although this type of evidence is routinely admitted in court, Faigman et al.observethat Bite mark casework does not employ quantitative data and does not result inquantitative probability estimates. The typical bite mark opinion is devoid ofscientificdata supporting the ultimate, subjective, conclusory opinioni (1999 Pocket PartSupplement,p. 27).12 See Faigman et al. (1997:190), which summarizes which courts have acceptedspectrographic analysis,which ones have rejected it, and the evidentiary standards applied.13 One way around the Frye standard is to define particular field. narrowly toinclude only those whoadvocate for the technology. In fact, the Alaskan court in State v. Coon (974 P.2d386 (Alaska 1999)), canbe accused of doing just that. This ploy has been a source of significant criticismin the American legalliterature. See Faigman et al. (1997) for discussion of this device in the contextof voiceprint experts.THE LINGUIST ON THE WITNESS STAND 233Quite unlike bitemark analysis, forensic phonetics, it seems to us, is a vibrantandproductive field that has much to contribute to the judicial process (see Hollien1990).But, as most of the phonetics community recognizes, voiceprint evidence iscurrentlytoo unreliable to allow peoplec s freedom to depend on it, at least if not bolsteredbyconsiderable confirming evidence. As the technology becomes more sophisticated andresearch continues, however, the linguistics community may wish to revisit the
issue.3.3. DISCOURSE ANALYSIS. American courts have been reluctant to admit experts ondiscourse analysis. The cases primarily involve covertly recorded conversationsbetweengovernment agents and defendants who make allegedly incriminating statements.Discourse analysis of such statements has been admitted in a number of unpublishedcases (Shuy 1993, 1998). But virtually every appellate court to consider the matterina published opinion has held that a trial judge may properly exclude suchtestimony,as Wallace (1986) has shown in a detailed overview. Arguably, the publishedopinionsproduce a misleading record. Although most such opinions have affirmed trialcourtsedecisions to disallow the testimony, there are many (unreported) cases in whichdiscourseanalysis has actually been admitted.The most recent opinion we found involved a rape trial. A key piece of evidencewas a tape recording in which the defendant, according to the prosecution, admittedraping the victim. Although the procedural history is complex, the defendant deniedat trial that it was his voice on the tape, but the jury apparently thoughtotherwise andconvicted him. On appeal, he was awarded funds to hire a linguistic expert. Theexpertconcluded that it was indeed the defendantudefendantappeared weak-willed and easily dominated, his comments on the tape were notprobative of guilt.The defendant then petitioned the federal courts for assistance, arguing that if hehad been given funds to hire the expert earlier, he would have known better than todeny it was his voice on the tape; instead, he would have had the expert testifythathis recorded statements were not an admission that he raped the woman. The Court ofAppeals for the Second Circuit had little sympathy for his argument, pointing outthatcourts have generally not admitted testimony on linguistic discourse analysiso(Tysonv. Keane, 159 F.3d 732, 736 (2d Cir. 1998)).The principal reason the appellate courts have affirmed the exclusion of discourseanalysis is that, in their view, it has not been proven reliable or receivedgeneral acceptancein the scientific community. This conclusion may result from the belief that, asonejudge phrased it, discourse analysis is a discipline allowing [the expert] todetermine theintent of the speaker in covertly recorded conversationsl (State v. Conway, 472A.2d588, 608 (N.J. Super. Ct. 1984)). If the discipline is defined as a means ofdeterminingthe actual intentions of a specific speaker in a tape-recorded conversation withminimalcontextual cues, it is no wonder that judges find it unreliable, or at least,unproven.Perhaps the lesson to discourse analysts is that they need to be careful how theydescribetheir expertise to lawyers and judges.Another concern emphasized by courts is that the conversations in questiongenerallyconsist of ordinary language, with which jurors are assumed to be familiar. Recall
thata basis for admitting expert testimony under Rule 702 is that it can assist thejury inunderstanding the evidence. Some judges have suggested that discourse analysis willnot promote this function. The notion here seems to be: Why do we need a languageexpert? We have twelve jurors who all speak English. Let them decidee . Indeed, afew234 LANGUAGE, VOLUME 78, NUMBER 2courts have suggested that allowing discourse analysis of the meaning of ordinaryconversations would not only be unhelpful to jurors, but might actually confusethem.14Courts are likely to remain reluctant to admit discourse analysis when it isofferedas explaining the ultimate issuet of what a speaker meant or what the speakere sintentionswere. At the same time, it seems to us that there is a legitimate, but perhaps morelimited role for discourse analysis in the legal system. Linguistic experts canassistjurors to understand a recorded text, for instance, by pointing out who is raisingwhichtopics. Linguists can also point out a pattern of discourse markers that would haveotherwise gone unnoticed, or inform jurors about the nature of Gricean implicatureinappropriate cases. Once the jurors are aware of this information, they will be onthesame footing as the expert (more or less), and able to draw their own inferencesbasedon their intuitions as enhanced by the linguistic analysis that the expert haspresented.We believe that if discourse analysis is offered on such terms, more judges wouldbeor should beinclined to admit it. In fact, Shuy (1993, 1998) reports that histestimony was accepted in a number of cases in which discourse analysis was used injust this way. Given the evidentiary standards of both Frye and Daubert, the legalsystem is not likely to accept stronger inferences from linguists based ondiscourseanalysis unless they are backed up by empirical studies. The growing use of corporawill enhance the likelihood of discourse analysis being admitted, both byincreasingits perceived reliability and by presenting jurors with information that may not beintuitively evident.3.4. THE MEANING OF CONTRACTS AND STATUTES. The same considerations applygenerally to linguistic experts offered to testify on the meaning of legal textslikestatutes, insurance policies, and contracts. (See Solan 1998, 1999 for moredetaileddiscussion). If linguists act as guides through difficult passages, usinglinguistic analysisto explain how it is that various interpretations are available, their testimony ismorelikely to be accepted by courts than if they attempt to tell judges what a legaltextmeans.Courts have given linguistic expertise on meaning a mixed reception. In theAmericanlegal system, judgesnot jurorsinterpret legal documents, at least initially.Judgesare the only official interpreters of statutes. As for contracts and insurancepolicies, ifthe judge finds them to be ambiguous, then it is up to the jury to decide which of
theavailable meanings was the intended one. Some courts have held that sinceinterpretationis a legal matter and linguists are not lawyers, linguists cannot have much tocontribute.(See National Automobile and Cas. Ins. Co. v. Stewart, 272 Cal. Rptr. 625 (Cal. Ct.App. 1990)). In contrast, other courts have admitted linguists as experts onmeaning,especially in cases involving the interpretation of statutes. (See Louisiana v.Azar, 535So. 2d 441 (3d Cir. 1988)). Other courts are worried about usurping the role of thejury. For example, courts are suspicious of linguists being called in libel casesto telljurors what the defendant meant by an allegedly defamatory statement, although afewhave allowed it (see Solan 1999).Again, the problem here seems to be the extent to which linguistic testimony canhelp the legal system. In cases involving complex language about which there isunderstandabledisagreement between the parties, linguists can serve a role by acting as tourguides, walking the judge or jury through the disputed language, and explaining howthe14 United States v. Kupau, 781 F.2d 740 (9th Cir.), cert. denied, 479 U.S. 823(1986); United States v.Valverde, 846 F.2d 513 (8th Cir. 1988).THE LINGUIST ON THE WITNESS STAND 235disputed language is an example of well-studied linguistic phenomena. Thelinguist sultimate interpretation is not very important, and sometimes should not be given atall.Judges and jurors can use their own intuitions as native speakers. But if thelinguistcan help the players in the legal system understand in a systematic way the sourceofthese intuitions, she can help jurors to structure their deliberations, and judgesto structuretheir opinions. These services will not be needed frequently. Most often, peoplecan decide on the relevant range of available interpretations of a document withoutexpert assistance. But there should be an occasional role for linguists in thesecircumstancesthat does not usurp the role of judge or jury. Many of the cases in whichlinguistic expertise on meaning has been accepted use linguistic knowledge in justthisway. Moreover, the literature on syntax and semantics is so rich that there shouldrarelybe any question about linguistic expertise meeting evidentiary standards under FryeorDaubert.3.5. POLITICS AND THE COMPREHENSIBILITY OF JURY INSTRUCTIONS. Finally, let usconsidercases that deal with the comprehensibility of jury instructions. Much research,beginning with a seminal study by Robert and Veda Charrow (1979), has shown thatthe standardized jury instructions typically used in the various American state andfederal jurisdictions are quite difficult for many jurors to process. California,for instance,has an instruction on how to evaluate the testimony of witnesses, solemnlyadmonishing the jurors that failure of recollection is a common experience, andinnocentmisrecollection is not uncommoni. The last five words contain no fewer than three
negative elements. Even more complex is the instruction that a lawyer s questionstoa witness are not evidence, and that therefore jurors must never speculate to betrueany insinuation suggested by a question asked a witnessu (Charrow and Charrow 1979:134445).One would think that lawyers would consider hiring a linguist or cognitivepsychologistto investigate the ability of jurors to understand their charge, especially in acasewhere the instructions are unusually significant. Oddly, lawyers rarely do so,thoughthis is just what happened in two capital murder cases in Illinois, discussed inTiersma1995. We have already briefly mentioned the first of these cases, Free v. Peters,inwhich the trial court gave substantial credence to expert testimony that Freee sjury didnot properly understand its instructions on the role of mitigating evidence. Acloselyrelated case involved a mass murderer named John Wayne Gacy, who attempted touse the same evidence to overturn his own death verdict.Illinois, like many other American states, requires that after a defendant is foundguilty in a capital case, the jury must weigh or balance aggravatingn factorsagainstmitigating factors. The relevant statute requires thatIf the jury determines unanimously that there are no mitigating factors sufficientto preclude the impositionof the death sentence, the court shall sentence the defendant to death.Unless the jury unanimously finds that there are no mitigating factors sufficientto preclude theimposition of the death sentence the court shall sentence the defendant to a termof imprisonment underChapter V of the Unified Code of Corrections. (Ill. Rev. Stat. ch. 38, para. 9-1)This statutory language was read, almost verbatim, to both the Free and Gacy jurieswhen they had to decide whether these men should die or spend the rest of theirlivesin prison. Both juries voted for death.In separate lawsuits, Free and Gacy challenged the imposition of the death penaltyin federal court. They invoked a juror comprehension study conducted bypsychologistHans Zeisel, as well as testimony by a linguist (see Levi 1993). The study andrelated236 LANGUAGE, VOLUME 78, NUMBER 2testimony demonstrated that jurors who received Illinoisiinstructionsdid not adequately understand the role of mitigation and aggravation. Even acursory glance at the language of the statute reveals problems with multiplenegation,a relatively unusual word (preclude), lack of clarity about when unanimity isrequired,and phrasing that suggests that a death verdict is the desired outcome. Asmentioned,the trial court judge paid serious attention to the linguistic and other expertevidence.In Gacyes case the juryas difficulty in understanding the convoluted Illinois deathpenalty instruction was compounded by what Court of Appeals Judge Frank Easterbrooklater called a slip of the judicial tongueC. He summarized what happened asfollows:
The jurors had this instruction, like the others in the three-page charge, duringtheir deliberations.Unfortunately, Judge Garippo did not read the instruction to the jury as writtenorat least the courtreporter did not take down the same words that appear in the written instructions.The transcript has itthat the second sentence of this instruction was delivered as: If, after yourdeliberations, you unanimouslyconclude there are mitigating factors sufficient to preclude the imposition of thedeath penalty, youmust sign the verdict form directing asentence of imprisonmenth . (Gacy v. Welborn,994 F.2d 305,307 (7th Cir. 1993))The trial court thus erred in instructing the jury that there must be unanimityaboutmitigation to justify a sentence of life imprisonment. Just the opposite is true asa matterof law: a single juror who believes that something about the defendant or the crimeisa mitigating factor may vote against death on that basis, requiring the court tosentencethe defendant to imprisonment (usually for life).Despite the evidence of confusion, Gacyf s appeal of his death sentence was rejectedby the Seventh Circuit Court of Appeals. An opinion written by Judge Easterbrookadmitted that the instruction did not give the jury the whole truth , but heldthatclosing arguments by the lawyers and a written copy of the instruction remedied anymisconceptions. As to the Zeisel study, the court admitted that the use of legaleseininstructions had an inevitable adverse affectuinvokedthe traditional legal presumption that jurors understand and follow theirinstructions.This presumption, the court emphasized, is not a bursting bubble applicable onlyinthe lack of better evidence . No, this presumption is a rule of lawp (994 F.2d at312313). Gacy has since been executed.When Free s case reached the Seventh Circuit, an opinion by Judge Posner likewiserejected the comprehension research, deeming it methodologically flawed, andaffirmedthe original death sentence (United States ex rel. Free v. Peters, 12 F.3d 700(1994)).Free has also been executed.These judicial reactions reflect not just legal, but systemic and politicalconcerns.From the perspective of the criminal justice system, judges fear that invalidatingaconviction because the jury did not understand its instructions would provide abasisfor appeal in countless other cases, because jury instructions in general are notall thatunderstandable. Similar concerns motivated the Supreme Courtn s opinion in McCleskeyv. Kemp (481 U.S. 279 (1987)). There, the Court decided that statistical evidencegathered by a well-known social scientist concerning racial prejudice in meting outthedeath penalty should not be given weight by the courts. In response to the argumentthat death penalty cases are special, the Court said that The Eighth Amendment isnotlimited in application to capital punishment, but applies to all penalties. . .
Thus, ifwe accepted McCleskey s claim that racial bias has impermissibly tainted thecapitalsentencing decision, we could soon be faced with similar claims as to other typesofpenaltyi (481 U.S. at 31417).THE LINGUIST ON THE WITNESS STAND 237This moment of candor reflects a genuine concern: if new research can be used toquestion the validity of a certain type of legal decision making, when does thequestioningend? Judges fear that they may be opening a Pandorac s box, the end result of whichcould be the reversal of countless jury verdicts, undermining public confidence inthecriminal justice system and the jury itself.The Gacy and Free cases never reached the nation,s highest court. Recently,however,a set of capital cases from Virginia came before the Supreme Court, providing itwithan opportunity to consider the importance of having a jury properly understand itsrolein the justice system, especially when life itself is at stake. In Buchanan v.Angelone(522 U.S. 269 (1998)), Buchanan had been convicted of murder. He was sentenced todeath despite the fact that his jury received a poorly drafted instruction which,in ourview, failed to properly apprise them of the role of mitigating evidence. As notedabove,such evidence consists of reasons that a jury might have mercy on a defendant andistherefore an extremely important consideration when a capital jury is required todecidewhether the defendant should be sentenced to life imprisonment or death.Nonetheless,the Court had little sympathy for the arguments of Buchanane s lawyers. To themajorityof justiceswho are, after all, prominent representatives of the legal systemtheinstruction seemed perfectly comprehensible.A year later the Supreme Court accepted another case from Virginia, Weeks v.Angelone(102 S. Ct. 727 (2000)). Weeks had been condemned to death by a jury thatreceived the same standardized instruction as did Buchanan, but with one importantdifference. In the Weeks case, the jury sent the judge a question during itsdeliberationsthat revealed that they were unsure about the role of mitigation. The judge, as iscommonin these situations, refused to elaborate on the law and simply referred them backtotheir original instructions.One might suppose that a juryts question is compelling evidence of lack ofunderstanding,but the Supreme Court thought otherwise. Although the vote was closer thanin the Buchanan case, the justices brought to bear the familiar rules that a juryispresumedc to follow its instructions, and added that it is likewise presumedu tounderstandthe judgeds answer to their questions. The Court upheld Weeksi s death sentence,and he has since been executed.To our knowledge, no linguists were involved in either Buchanan or Weeks.Subsequently,
however, a group of researchers showed that the jury instruction used in bothof those cases was indeed seriously misunderstood by a substantial number ofsubjects(Garvey et al. 2000), just as the dissenting justices predicted. Yet it seemsunlikelythat such research would have made a difference, even if it had been available totheCourt when it pondered its decision.Particularly when the issue is the death penalty, still widely supported by theAmericanpublic, the politics are palpable. Many federal judges have recently expressed theview that appeals by state death penalty inmates to federal courts (generally usingthevenerable writ of habeas corpus) are nothing more than an eleventh-hour attempt todelay their executions. That, apparently, is how the panels of judges deciding theFreeand Gacy cases interpreted the Zeisel study: not as evidence of the linguisticinadequaciesof a critical jury instruction, but as a final straw to be grasped at by desperatemen. Viewed in that light, the outcome was foreordained.4. CONCLUSION. While we end on this pessimistic note, it is interesting to see justwhere courts begin to appear threatened by linguistic testimony. Courts will belesslikely to accept linguistic expertise when it challenges long-standinginstitutional roles.238 LANGUAGE, VOLUME 78, NUMBER 2The jury may be the most important of these, which further explains the reluctanceofcourts to allow experts in discourse analysis to draw inferences relating to aspeakertsintent. Such inferences are a classic jury function in the American legal system.In our view, courts ought to allow linguists to explain to juries relativelynoncontroversialaspects of how conversations operate, even if they insist on leaving ultimatequestions of speakerss intentions entirely to the jury. Similarly, we believe thatcourtsshould allow linguistic experts to comment on the range of meanings of statutes andcontracts when that testimony can be helpful in understanding the evidence andstructuringthe debate between the parties.We also believe that the law concerning the use of experts in both voiceidentificationand author identification is in serious need of improvement. Here, the linguisticandlegal communities might work together.Despite these areas of contention, linguists are generally welcome in Americancourtswhen it appears that their testimony is based on sound research and will be helpfultothe judge or jury. As the legal profession and linguists both learn what type oflanguageexpertise is most useful to jurors, and as linguists bolster the validity of theiropinionswith corpus-based data, we expect this trend to continue.REFERENCESBOLT, RICHARD H. et al. 1979. On the theory and practice of voice identification(Reportof the Committee on Evaluation of Sound Spectrograms, Assembly of Behavioral andSocial Sciences, National Research Council). Washington, DC: National Academy of
Sciences.BRESNAN, JOAN. 1978. A realistic transformational grammar. Linguistic theory andpsychologicalreality, ed. by M. Halle, J. Bresnan, and G. A. Miller, 159. Cambridge, MA:MIT Press.CHARROW, ROBERT, and VEDA CHARROW. 1979. Making legal language understandable: apsycholinguistic study of jury instructions. ColumbiaLa w Review 79:130673.CHASKI, CAROLE E. 2001. Empirical evaluations of language-based authoridentificationtechniques. Forensic Linguistics 8:165.COTTERILL, JANET. 2002. Language in the legal process. Hampshire, UK: Palgrave, toappear.CULICOVER, PETER. 1997. Principles and parameters: An introduction to syntactictheory.Oxford: Oxford University Press.FAIGMAN, DAVID L.; DAVID H. KAY; MICHAEL J. SAKS; and JOSEPH SANDERS. 1997. Modernscientific evidence, vol. 2. St. Paul, MN: West Publishing Company.FINEGAN, EDWARD. 1990. Variation in linguistssAmericanSpeech 65.33440.FOSTER, DON. 2000. Author unknown: On the trail of Anonymous. New York: Henry Holt.GARVEY, STEPHEN; SHERI LYNN JOHNSON; and PAULMARCUS. 2000. Correcting deadlyconfusion:Responding to jury inquiries in capital cases. Cornell Law Review 85:62755.GIBBONS, JOHN (ed.). 1994. Language and the law. London: Longman.HOLLIEN, HARRY. 1990. The acoustics of crime: The new science of forensicphonetics.New York: Plenum.HUBER, PETER W. 1991. Galileous revenge: Junk science in the courtroom. New York:BasicBooks.LABOV,WILLIAM, andWENDELL A.HARRIS. 1994. Addressing social issues throughlinguisticevidence. Language and the law, ed. by John Gibbons, 265305. London: Longman.LEVI, JUDITH N. 1993. Evaluating jury comprehension of Illinois capital sentencinginstructions.American Speech 68.2049.LEVI, JUDITH N. 1994. Language as evidence: The linguist as expert witness in NorthAmericancourts. Forensic Linguistics 1.126.LEVI, JUDITH N., and ANNE GRAFFAM WALKER. 1990. Language in the judicial process.NewYork: Plenum Press.MCCARTHY, J. T. 1992. Trademarks and unfair competition, 3rd edn. New York: ClarkBoardman Callaghan.THE LINGUIST ON THE WITNESS STAND 239MCMENAMIN, GERALD R. 1993. Forensic stylistics. Amsterdam: Elsevier.MCMENAMIN, GERALD R., and FRED H. BRENGELMAN. 2000. Two independent studies ofthe same questioned-authorship case. Paper presented at the Georgetown UniversityRound Table on Languages and Linguistics.RAFFERTY, LISA. 1999. Anything you say can and will be used against you:Spectrographicevidence in criminal cases. American Criminal Law Review 36.291311.RIEBER, ROBERT, and WILLIAM STEWART (eds.). 1990. The language scientist as expertinthe legal setting. New York: Annals of the New York Academy of Science 606.SHUY, ROGER W. 1993. Language crimes: The use and abuse of language evidence in thecourtroom. Cambridge, MA: Blackwell.SHUY, ROGER W. 1998. The language of confession, interrogation and deception.
ThousandOaks, CA: Sage.SOLAN, LAWRENCE M. 1998. Linguistic experts as semantic tour guides. ForensicLinguistics5.87106.SOLAN, LAWRENCE M. 1999. Can the legal system use experts in meaning? Tennessee LawReview 66.116799.TIERSMA, PETER M. 1995. Dictionaries and death: Do capital jurors understandmitigation?Utah Law Review 1995.149.TOSI, OSCAR; HERBERT OYER; WILLIAM LASHBROOK; CHARLES PEDREY; JULIE NICOL; a ndERNEST NASH. 1972. Experiment on voice identification. Journal of the AcousticalSociety of America51:2030 43.WALLACE, WILLIAM D. 1986. The admissibility of expert testimony on the discourseanalysisof recorded conversations. University of Florida Law Review 38.69115.WEINSTEIN, JACK, and MARGARET BERGER. 1998. WeinsteinR s federal evidence. New York:Matthew Bender.Tiersma[Received 14 July 2000;Loyola Law School accepted 14 March 2001]919 S. Albany St.Los Angeles, CA 90015SolanBBrooklyn Law School