International Journal of Speech Language and the Law, Vol 22, No 1 (2015)


doi : 10.1558/ijsll.v22i1.27073

The Language of Bribery Cases

Roger W. Shuy (2013)

Oxford University Press 276pp

Reviewed by Ronald R. Butters

Roger Shuy is widely acknowledged as the dean of American forensic linguistics (FL) and one of the principal founders of the field in the world. Scholars, linguistic consultants, and the system of justice have benefitted greatly from his guidance as author, teacher, mentor – and expert witness. He also was a founder of the American Association of Applied Linguistics, and his work in the field of language and law places FL squarely within the useful branches of science. His numerous articles and books (including Linguistics in the Courtroom: A Practical Guide [2006], Fighting over Words [2008], The Language of Defamation Cases [2009] and The Language of Perjury Cases [2011]) are at heart practical works, designed to instruct linguists and legal professionals about how to use linguistics in assisting defendants, prosecutors, juries and judges, all of whom must come to conclusions about matters of language and law in the courtroom. Shuy’s books do their job well. Any linguist who finds herself tempted to accept a request for expert assistance from a lawyer on behalf of a client would be unwise indeed not to read Shuy’s relevant publications. Her lawyer would profit as well from consulting them.

The Language of Bribery Cases (LBC) is no exception in its usefulness and general good sense. As is typical of Shuy’s books, it begins with a brief historical introduction that touches both the legal concept of bribery and Shuy’s own consulting experience, sketching out also his view of what a linguist may contribute within the context of American bribery law. Chapter 3 (39–69) is particularly important because it details the linguistic methodology that Shuy has developed in his own practise of analysing evidentiary recorded speech of various kinds (e.g., defamation, libel, perjury, money laundering, drug trafficking). Chapters 4–13 describe the specifics of a dozen bribery cases in which Shuy advised lawyers (and often testified in courts), showing how his methodological framework can be profitably applied. Finally, Shuy returns in chapters 14 and 15 to discussions of ‘The Role of Linguistic Analysis in Bribery Cases’ and ‘The Legal Context of Bribery Language and Law’, referring to the details of the case studies. Those who are familiar with Shuy’s earlier books will find that, while LBC adds little that is new to the methodology that he has discussed repeatedly in earlier works (even the analysis of bribery is sketched out in Language Crimes [1993] and Creating Language Crimes [2005]), it is nonetheless useful to see the methodology applied so extensively to so many specific cases.

At the most elementary, a linguist can assist the court in cases that involve evidentiary recorded speech by ensuring that the data itself is correctly presented – preparing written transcripts or editing faulty ones. This can be important, as Shuy notes in several places:

  • The prosecutor’s transcript of this passage emphasized that McDonald said, ‘you’re right on the team’, but took this out of his overlapped and ongoing sentence, including the beginning word, ‘if’, which clearly introduced a conditional clause that gave no evidence that McDonald knew whether or not the two men were on ‘the right team’ (99–100).
  • The defense transcript, which differs significantly from the prosecution’s … show[s] … important differences between them … [e.g.,] Government transcript: … ‘No, I would take it today, wouldn’t you?’ … Defense transcript: ‘No, I wouldn’t take it today, would you?’ (176–177)
  • My first task was to correct the government’s faulty transcript, in which I found 143 errors, many of them crucial. (183)

By training and experience, it is obvious that a linguist will be better able than anyone else to compare a recording of speech with a written text and determine the more accurate transcriptions with respect to phonology, morphology, syntax and punctuation. If necessary, she can even subject the recording to laboratory acoustical analysis.

But correction of transcripts is only a small part of the forensic linguist’s task. Shuy’s basic theoretical donné is that a linguist schooled in sociolinguistic theory (including conversation analysis and pragmatics) is uniquely and professionally qualified to recognise what evidentiary passages are indeed the most salient for a jury to consider and what the passages may mean semantically and pragmatically. Unlike the average judge or juror, the linguist has knowledge that can be brought to bear upon the data that can assist in determining the crucial question of the defendant’s alleged intent to commit a crime. (I would also add that the linguist has immensely more time to examine the entire set of data than the jury does.) Of course, Shuy writes (16, echoed passim), ‘it is virtually [sic] impossible to get inside the minds of people to determine their intentions or understandings’ (by considering recordings of their conversations) except through means such as ‘mind reading’ (58). I take this to mean that forensic linguistics is also not dependent on truth serums, lie-detector machines, or well-honed devices of physical torture administered to the conversants in pre-trial investigation.

In LBC, however, FL’s inability to ‘get inside the minds of people’ is also somehow related to the fact that the linguist must be wary of subverting the jury’s job of making the ultimate decision about guilt or innocence (emphasis added):

It should be clear that discovering clues to intentions does not mean that these intentions can be firmly established. As in the work of detectives, clues do no more than lead investigators to the next steps of understanding. Linguists find the clues, put the clues together, and point out what these clues can mean and what they can’t mean, based on the structural rules of language. It is always up to the trier of fact, however, to deliberate and conclude exactly what these clues mean for the case at hand. But discovering all of the language clues is an important task that most people, including triers of … fact, are not trained to do, which is why linguistic analysis can play a much-needed role in law cases such as bribery. Carefully examining the speech event, schemas, agendas, speech acts, and conversational strategies of the participants can reveal many clues to intention. (52)

Shuy appears here to be drawing an extremely fine distinction without a real difference, while at the same time blurring the distinction between crime ‘investigators’ and ‘triers of fact’. No one would disagree that the linguist in a bribery case cannot legitimately opine as to whether a particular passage legally shows the intent of the defendant to commit the crime. That is indeed the job of the jury, as Shuy reminds us elsewhere (e.g., 244). However, it is not clear how it is possible for a linguist to decide which passages are ‘clues’ that ‘point out’ something relevant to a jury without having ‘firmly established’ in her own mind that the passage is relevant to the question of intent to commit the crime. Nor is it clear how she may also opine, ‘what these clues can mean and what they can’t mean’ (often contradicting what the police and prosecutors may opine about them) without communicating her conclusions about intent to the jury.

Of course, there may be considerations other than the linguistic that a jury must consider – motive, for example, or the testimony of alleged co-conspirators. Even so, it is not clear in what sense Shuy means us to understand that the matter of intent that a ‘clue’ ‘points’ to is not to be considered ‘firmly established’ if at the same time the salience and interpretation of that ‘clue’ is solid enough to warrant pointing out to a jury. Shuy continues to generate what appear to be ‘firmly established’ conclusions throughout the book, as for example in the following passage:

I reported to him that … Kelly danced around the bribe offer for a while but … he eventually took it. In short, there was no way that any linguistic analysis could be helpful for Kelly’s defense … [T]his ended my relationship with the … case. [¶] When Kelly went to trial, his defense theory was that he was highly suspicious of the agent and took the money only in an effort to establish proof that the agent was crooked. That didn’t sit any better with the jury than Congressman Myers’s defense that he was only playacting. (82)

A reader may note that the inductive conclusions about the salience and meaning of the data that he reports on in the passage seem to be ‘firmly established’ in Shuy’s mind; ‘danced around’ at least implies strong doubt about the Myers’ innocence. Elsewhere, Shuy has this to say about Kelly’s intent, ‘[In this portion of the text], Kelly tried to distance himself from the agent’s offer of paying a hundred thousand dollars’ (76). Of course, these passages only report his findings to the reader; he is not testifying before a jury. But it is not clear where or how Shuy draws the line.

LBC attempts also to confront the larger theoretical question of how to link FL methodology to principles of rigorous scientific inquiry in the study of language. The principles that Shuy prescribes for recognizing ‘clues’ and determining their meanings are summarised in the 30 pages of chapter 3, in which Shuy briefly relates those principles to the work of other established linguists as well as to his own earlier books.

He emphasises (56) that FL does not offer the putative evidentiary certainty of DNA identification and comparison. Unlike FL, (a) DNA testimony offers meanings and known error rates that are generally accepted within the scientific community, and (b) judges and juries could not even begin to carry out DNA analysis on their own. Although FL in general, and the science of linguistic analysis of language data in particular, are inductive scientific modes of inquiry, the methodology that LBC presents has no known (or perhaps even knowable) error rates, and the significance and meanings of the portions that the linguist has identified as germane may be contested in court by other qualified experts who claim to be using the same methodology. In addition, unlike the work of the DNA analyst, the ultimate inductive task of the forensic linguist is essentially the same as that of the defendant and the defendant’s lawyers, the prosecutors, the judge and, ultimately, the jury: to determine what is the meaning and significance of the data.LBC’s methodology asks, ‘What portion of the data are the “clues” that should be analyzed?’ ‘What is the meaning of these “clues” with respect to the question of intent to commit a crime?’ The analysis of the data in a bribery case thus begins at a place that is roughly comparable to asking, in DNA analysis, (a) what portion of the crime scene should even be tested for DNA (is the gunk on the floor blood or cranberry juice?) – and, once it is tested, (b) if there is one other living creature that the tested DNA uniquely matches (the suspect? the murder victim? a dog with an injured paw?).

One might well ask, then, ‘What is there about linguistic science in general (and Shuy’s prescribed methodology in particular) that privileges the linguist’s introspective observations to the extent that they can be of assistance to a judge or jury?’ One answer lies in the fact that, while ordinary people know that they cannot perform DNA analysis, they do have a basis for making judgments about linguistic data (after all, they have conversations themselves every day) – they are just so naïve about language that they are likely to be mistaken – particularly about such conversations as are typically studied in the analysis of evidentiary speech, where both defence and prosecution may draw the jury’s attention to putatively revealing passages that are taken out of context. Evidentiary conversations are highly unnatural because only one of the speakers knows that the conversation is being recorded, and the one who is in the know has a hidden duty to make the other speaker look as guilty as possible. A linguist can point all this out to the trier of fact, to good end.

Shuy is not claiming that forensic linguistic analysis can be derived automatically from his prescribed methodology (or anyone else’s). While the data of FL themselves are empirical rather than subjective, and while specialised linguistic knowledge may draw the linguist’s attention to aspects of the data that an ordinary juror or judge would not notice or understand, the linguist is, in the end, inferring that, for example, X means Y and not Z, and that the undercover agent was deliberately distorting the conversation to make the mark look guilty, and that the significance of some recycled arguments is different from that of others. Just as sociolinguists have argued (persuasively, I believe) that using data that results only from introspection runs the inherent risk of creating analyses that support the analyst’s thesis, so, too, is it legitimate to observe that finding and interpreting ‘clues’ in empirical data also presents an inherent risk of confirmation bias – unconsciously highlighting some ‘clues’ and finding meanings in them that are favourable to the client, while avoiding those that another analyst might say points in the opposite direction. Indeed, it frequently happens that two expert linguists testify on the opposite sides in the same case, and one cannot with any justification conclude that one of them must be corrupt, insincere or less insightful than the other. Indeed, remembering that another linguist may find major flaws in one’s selection and interpretation of ‘clues’ is a healthy inducement to objectivity and conservative claims.

One of the great values of Shuy’s books is that they continue to show a brilliant forensic scientist in action, one who brings high ethical standards to the struggle to avoid such confirmation bias. One need look no further for practical validation of the scientific merit of the application of Shuy’s methodology than the forensic linguistic analyses detailed in the dozen or so cases that Shuy reports on in the middle chapters of the book. They clearly offer analysis that is more linguistically sophisticated than what a jury could have found on their own. The rigor with which Shuy applies the linguistic methodology is obvious, and he even reports of having formed conclusions that were contrary to what the attorneys who employed him would have liked, as in the passage from p. 82 quoted above.

It is also worth emphasising, though Shuy does not specifically do so in this book, that, while linguists must struggle to maintain objective perspective in interpreting data, the duty of attorneys is advocacy. In addition, law-enforcement officers who participated in the undercover recording of their encounters with defendants are often allowed to testify at trial about what they meant and what they feel certain that the defendant meant in the conversations at issue; as LBC demonstrates, they are especially susceptible to confirmation bias. Defendants, however, can testify only if they are willing to give up the right to immunity from testimony, thus opening themselves up to the pitfalls of cross-examination. The forensic linguist thus has a considerable role to play in balancing justice: if the linguistic analysis warrants it, she can be the one to offer what is in effect an alternative scenario, the one that the defendant would like to have given. The work of Roger Shuy has thus contributed considerably to the righting of the imbalance to justice brought about by the invention of mechanical recording of speech.


  • There are currently no refbacks.

Equinox Publishing Ltd - 415 The Workstation 15 Paternoster Row, Sheffield, S1 2BX United Kingdom
Telephone: +44 (0)114 221-0285 - Email:

Privacy Policy