LINGUISTICS IN THE COURTROOM By Penelope O. Pickett, M.F.S Forensic Linguist FBI Laboratory Washington, DC With increasing frequency, the term "linguistics" is being heard in the courtroom as linguists bring their expertise to bear during judicial proceedings. Both prosecutors and defense attorneys realize the effectiveness of linguistic testimony which, oftentimes, turns a case around. Because linguistic analysis and testimony can influence investigations and the outcome of cases tried in court, the law enforcement community may benefit from knowing what to expect from the discipline of linguistics. This article discusses three types of linguistic analysis presented in the courtroom. It shows the differences among the three types by providing examples and explaining the analytical focus of each. LINGUISTIC ANALYSIS Linguistics, the scientific study of language, is a well-established discipline that interlinks with other disciplines. The 1980's saw the beginning of a wave of linguistic activity in civil and criminal investigations, which swelled into the decade of the 1990's. Although various linguistic analysis interrelate with judicial matters, the predominant activity centers around three basic types of examination and testimony--author/speaker comparison, author/speaker assessment, and discourse analysis. All three focus on language usage and involve comparison methodology, each from a different perspective. Author/Speaker Comparison Linguistic examinations can compare a written communication with a voice communication or a typewritten text with a computer printout. These examinations can also be conducted on two or more written documents or two or more voice recordings. In the author/speaker comparison process, linguistic examiners analyze and compare applicable elements of the specific communications. These elements include vocabulary selection, syntax, phraseology, spelling, style, format, sentence length, pronunciation, intonation, pitch, rate of speech, voice quality, etc. When testifying in court, linguistic experts might present computer printouts of word frequency counts and analyses, which show correlations of common word choice or word length between two communications, or words infrequently used by the general population. Experts might also demonstrate comparisons of other items, such as grammatical constructions and errors or speaking characteristics. The following cases illustrate author/speaker comparisons (1) in linguistic examinations. CASE #1: In late 1989, package bombs killed a Federal judge in Alabama and an attorney in Georgia. A linguistic examination by the FBI Laboratory compared the typewritten communiques associated with the bombings to documents known to have been authored by a prime suspect in the case. (Traditional document examination determined that the communiques and the documents were prepared on the same typewriter.) As a result of the linguistic examination, FBI examiners concluded that this suspect was not responsible for the bomb communiques. When the FBI Laboratory received known writings of another suspect, Walter Leroy Moody, Jr., a linguistic examination determined that this suspect most likely authored the bomb communiques. This caused investigators to shift their attention to Moody, who was subsequently identified as the perpetrator and later tried and convicted. CASE #2: A police chief in Pennsylvania received threatening letters in disguised and distorted handprinting. (2) Linguistic examiners were able to compare the threatening letters to letters in normal handwriting written by the suspect. This examination revealed sufficient similarities in vocabulary, grammar, spelling, etc., for examiners to conclude that the suspect most likely composed the anonymous letters. The suspect, the town's former police chief, subsequently confessed. CASE #3: "Dear Sir: I have been involved in espionage for several years...." So began the first in a series of anonymous typewritten letters to the FBI that revealed participation in a spy ring. Later, when John Walker and his family came under investigation for espionage, known writings of his friend, Jerry Whitworth, were printed out from computer memory and submitted to the FBI Laboratory for comparison with these anonymous typewritten letters. Because the questioned documents were typewritten and the documents of known authorship were computer printouts, the Laboratory could not conduct a traditional document comparison examination. Linguistic comparison examinations, however, found strong similarities in word usage, grammar, spelling, and format, concluding that Whitworth most likely authored the anonymous letters. Whitworth was later convicted for his participation in the spy ring. Author/Speaker Assessment In analyzing communications to determine demographic and psychological characteristics of the author/speaker, the linguistic examiner looks at the same features as in the comparison examination, e.g., vocabulary selection, syntax, phraseology, etc. The examiner basically does the same type of comparison examination, but in author/speaker assessments, uses population standards as the comparative material. Unfortunately, a comprehensive, centralized set of population standards does not exist, even though linguists, sociologists, psychologists, and others have conducted studies on the various factors (3) that could be used to identify otherwise anonymous authors. Without a full data bank of such standards, current author/speaker assessments (4) depend on an examiner's own knowledge of standard textbook usage, academic studies, and familiarity with language usage of various demographic groupings, as well as on the examiner's experience. Using knowledge and experience, and the ability to obtain supplemental information, the examiner makes qualified determinations concerning an author's/speaker's age, sex, education, occupation, geographic/ethnic background, veracity, etc. At present, this type of examination serves as an investigative aid and is not intended for court testimony. As comparison standards become more developed, however, the results of linguistic examinations will undoubtedly be the subject of more frequent expert witness testimony. CASE EXAMPLE: In Fayetteville, Tennessee, the body of a teenage girl was found in her bathtub. The investigation of this homicide centered on adult males until the FBI Laboratory conducted an examination of a note thought to be connected to the homicide. Linguistic analysis of the note determined that the vocabulary, grammar, and style were most likely that of a teenage male. Investigators then turned their attention to the teenage population, in which they eventually found the perpetrator. They subsequently learned that the male teen committed a similar crime in a neighboring jurisdiction. Discourse Analysis "[There were] one hundred and twenty-five differences between what I heard on the tape and the transcript." (5) A linguist made this statement while testifying as an expert witness for the defense. With this comment, the linguist set the stage for his analysis, which pointed out major flaws in the State's case and which led to the defendant's acquittal. Discourse analysis, (6) conducted in the judicial system context for court testimony, is the analysis of conversations recorded on audio or videotape that are used as evidence in court. This analysis is offered to facilitate accurate listening, or accurate understanding, of what was said in the recording. Correcting transcripts that may be used as listening guides is one manifestation of discourse analysis. Another involves explaining the dynamics of conversation to show who is more in control of a particular conversation, what underlying agenda each participant has, and so on. To perform a discourse analysis, the linguist first typically prepares a corrected transcript from which to work. Corrections are almost always necessary because most transcribers do not realize that there are different types of transcripts, and therefore, they produce transcripts that do not meet the stringent accuracy requirements of law enforcement. (7) Next, the linguist analyzes the conversation, identifying such things as topics raised in the conversation, who raised them, and the responses to them. Issues of entrapment and inaccuracy in case facts (both transcription and listening errors) are highlighted in this type of analysis. CASE EXAMPLE: After Senator Harrison A. Williams, Jr., had been convicted of bribery in the ABSCAM case, the U.S. Senate debated his censure and expulsion. During the debate, Senators heard a linguist's statement of his analysis of the videotapes in question, concluding, "Despite all of the strategies used by the Government agents to accomplish these aims [of recording a self-incriminatory act or statement by the Senator], the recorded evidence makes it abundantly clear that they fail." (8) APPROPRIATENESS FOR COURT Linguistics, a well-established discipline recognized by the scientific community for decades, has proven helpful in examining evidential communications in case investigations. But, is linguistic analysis appropriate for court testimony requiring an expert witness? Some judges believe so and have allowed the testimony; others have not. A judge's decision to allow linguistic analysis as expert testimony often hinges on whether that judge believes the testimony would aid the trier of fact. It can also depend on whether the judge believes that the benefit to the trier of fact outweighs the influence that scientific testimony may have on the jury's decision. Each case calls for independent evaluation of admissibility. CONCLUSION Linguistic analysis and testimony in criminal cases include three major types-author/speaker comparison, author/speaker assessment, and discourse analysis. Whenever cases involve written or recorded language as evidence, linguistic analysis may become involved in the investigation and court proceedings. Being able to differentiate among these three types of analysis will assist the law enforcement community in understanding how each could affect a particular investigation or trial. ENDNOTES (1) Other terms may be used for author/speaker comparisons. These include forensic linguistics, psycholinguistics, authorship identification, speaker identification, comparative stylistics, forensic stylistics, stylistic analysis, stylometry, forensic phonetics, and disputed authorship. (2) United States v. Clifford, 704 F.2d 86 (3d Cir. 1983). (3) Studies have been conducted on such factors as word frequency counts, dialects, accents, regional expressions, social setting influence on language, sex and age differences in language usage, occupational jargon, word associations, psychological influences on language, and ideational disturbance. (4) Author/speaker assessments may be referred to by other terms, such as forensic linguistics, psycholinguistics, authorship identification, speaker identification, psychological profiling, demographic profiling, stylistic analysis, and threat assessment. (5) State of Texas v. Thomas Cullen Davis, Case 16838 (Criminal District Court, Tarrant County, TX, October 18, 1979). (6) Topic analysis, conversational analysis, linguistics, and sociolinguistics are other terms used to refer to discourse analysis. (7) P.O. Pickett, "Transcripts for Law Enforcement: Special Requirements," Journal of Forensic Sciences, JFSCA, 34, No. 5, September 1989, 1250-1258. (8) 128 Cong. Rec. S1620 (1982) (reading of Exhibit 7).