In White Collar Crime 2001. Chicago: American Bar Association, 2001pp. F 19 to F 38.
It has become increasingly common for law enforcement agencies to surreptitiously tape-record conversations between undercover agents and suspected money launderers. Some of the targets self-generate their own guilt quite clearly, openly talking about how the money will be "washed." There is no magic that linguistic analysis can offer in such cases. But in many other cases, targets seem unaware of the direction of these conversations and suddenly find themselves as defendants in a money laundering trial. It is these cases that I wish to discuss here.
In such cases, the crucial information is usually the language used by the target, making linguistic analysis very important to both the prosecution and the defense. Elsewhere I have referred to such cases as, language crimes (Shuy 1993).
Language crimes
Undercover tape recording, of course, is not limited to efforts to uncover money laundering, but whether the alleged crime is bribery, solicitation to murder, fraud, or other language crimes, it pays to have a linguistic professional to analyze the language evidence. There is an uncanny similarity of undercover techniques in attempts to capture crimes on tape. The alleged crime may be different, but the conversational strategies of the agents are very similar.
Linguistic analysis is especially critical in money laundering cases brought under 18 USC 1956A3, which reads in part as follows:
(3) Whoever with the intent--
(A) to promote the carrying on of specified unlawful activity;
(B) to conceal or disguise the nature, location, source,
ownership or control of property believed to be the proceeds
of specified unlawful activity; or
(C) to avoid a transaction reporting requirement under State
or Federal law, conducts or attempts to conduct a financial
transaction involving property represented by a law
enforcement officer to be the proceeds of specified unlawful
activity, or property used to conduct or facilitate specified
unlawful activity......
Note that the ways to promote, conceal, disguise, and avoid here are all accomplished through language use and are, therefore, subject to linguistic analysis. If there is a crime here, it is a language crime. Was an offer ever really made? Did they have a quid pro quo? What exactly did the target promise, if anything? Did the target really agree or was he non-committal? The defense can't let the prosecution pull words out of context to prove these things. Human communication is much more complex than that.
The language of the undercover agent
An even more critical issue, also subject to linguistic analysis, focuses on the way that the undercover agent uses language. One of the guidelines for FBI undercover operations states that the agent must be crystal clear about the proposed illegal activity. The exact wording of this guideline, articulated by then Assistant Attorney General Philip Heymann before the House of Representatives Committee on the Judiciary, is as follows:
...making clear and unambiguous to all concerned the illegal nature of any opportunity used as a decoy. This provides the strongest possible protection against any unwitting involvement by individuals brought in by intermediaries or who are encountered directly. We attempt to structure our undercover decoy transactions by requiring overt participation on the part of all individuals... (Heymann 1980) Linguistic analysis of the agent's effort to represent the unlawful activity can be crucial to a case involving charges of money laundering.
Involve a linguist early in the case
I have been called upon both as an linguistic expert witness or as a consultant in hundreds of cases where tape recorded conversations provide the major evidence. For the defense it is best to consult a linguist for initial impressions and potential analysis at the earliest possible time in the case. Regular face to face meetings are crucial. For example, the BCCI money laundering case in the early 90's involved a thousand tape recorded conversations. Larry Barcella and his colleagues met with me weekly, going over the findings I made as I examined the tapes. Each week I reported the good parts along with the bad parts, ultimately contributing to their decision to take the best plea they could get.
There is little that the field of linguistics can do for a defendant to clearly self-generates his own guilt on tape. In most cases, early and regular meetings between attorney and linguist have proved beneficial to the defense. This was true in the case of US v Richard Silverman in 1990, when I met regularly with attorney James Brosnahan and his associates and testified at trial about the way language was used by both Silverman and the undercover agents, ultimately leading to a favorable jury verdict.
Four step process
Linguistic analysis in a case involving tape recordings is pretty much the same whether or not money laundering is the issue. There are normally four stages of such analysis:
1. Prepare an accurate, jury-ready transcript of the tapes
2. Analyze the allegedly hard parts of the conversations
3. Contextualize these hard parts with the rest of the conversations
4. Focus of the parts of the conversation that are good for the defendant
Step 1: Transcript preparation
One thing I have learned from working in hundreds of tape cases is that the government transcripts are never totally accurate. Once the defense is given discovery materials of tapes and transcripts, it is not unusual to find many inaccuracies. My usual approach is to correct the government's version, using high quality equipment and earphones, often listening to much of the tapes many times. It is also not unusual for the government to correct its own earlier transcripts shortly before trial, sometimes matching the corrections made by the defense. This would seem to indicate that no harm is done; however, if the defense had been relying on the earlier government transcript version, its case might suddenly need to be drastically revised. Time is the culprit here. The defense simply cannot afford to wait for the government's revisions. The only viable solution is to produce its own.
One can argue that the tape is the evidence, not the transcript, and the courts view transcripts as only a guide to such evidence. As harmless as this may seem, the effect of the written word, especially when it is inaccurate, in front of jurors causes them to hear what is on the tape what they see on the printed page.
Unlike other kinds of evidence, tape recorded language goes past the listener very quickly, and many listenings are sometimes necessary to produce an accurate transcript. In practice, it is common to get a secretary or court reporter to make a transcript. Such persons, however, generally lack the time, the high quality equipment, and the special language skills necessary to do the job accurately. Still another common practice is for the government to ask its agents and cooperating informants to go over the transcript draft to correct it from their own perceptions of what occurred. As logical as these procedures may seem, they are fraught with dangers of perceptual bias. People who participate in the taped events are often less able to determine what is actually on the tape than are well-trained total outsiders to the event. Agents or surrogate agents who wore the microphone are easily influenced by their original goal in taping a suspected criminal. That is, they may think they hear a suspect utter a given word or phrase simply because it is their goal to capture such words on tape. For example, in a Reno, Nevada case in 1981, the prosecution and the defense were at total odds about one crucial sentence:
Prosecution version: "I would take a bribe, wouldn't you?"
Defense version: "I wouldn't take a bribe, would you?"
Although the words were only semi-audible, the rhythmic beats and intonation of the sentence were perfectly clear. The speaker uttered six syllables, paused slightly, then produced two more syllable beats. The prosecution's version didn't match this, having five syllables, a short pause, then three more syllable beats. The obvious issue was the two "woulds" in the sentence, in particular which one of them contained the negative contraction. Syllable and intonation analysis clearly favored the defense version.
The point here is that the prosecution began with a schema that the defendant would take a bribe and consequently heard it that way, ignoring the physical evidence to the contrary. Hours of painstaking effort must be spent listening to the tape itself before a transcript is made. For much the same reason, defendants are usually no better than the prosecution at accurately remembering and recording their conversations long after they took place. Their perceptions, schemas, and memories also tend to skew what they hear.
Even the most accurate transcript cannot tell us many things. For example, even though the words may be audible on tape, we can't always know for certain whether or not they were actually heard or listened to by the target. Nor can we always tell exactly to whom they were directed, especially when several people are in the conversation. Other transcript problems miss important conversation clues as well, such as the stress placed on certain words, the pauses (which carry their own form of meaning), interruptions (which can be an important indication of speaker control, intention and manipulation), and many other things that linguistic analysis can provide.
Step 2: Attacking the difficult passages
Discourse topic analysis
In most tape cases there is a vast quantity of conversation for the jury to listen to and process. And in most cases they simply aren't up to such a daunting task. Commonly they get to hear each conversation once at trial. Occasionally the prosecution will try to use only some of the tapes or, worse yet, only parts of some of the tapes. In doing this, the prosecution recognizes the problem of swamping the jury with information. Unfortunately, this is a real danger. But the danger for the defense is greater if the prosecution has the final say about what parts of what tapes will be played. Obviously it will play only those parts that show the defendant in the worst possible light.
The defense is then faced with a difficult problem. How can the entire context of the conversations be shown without swamping the jury and boring it with lots of apparently extraneous information?
One way is to get a linguist to carry out a topic analysis of the entire body of evidence. The product of such analysis is a chart that shows every topic introduced by each speaker, as illustrated by the first twelve of the 77 topics in one tape of the DeLorean case, as follows:
Conversation Between John DeLorean and James Hoffman
September 4, 1982
L'Enfant Plaza Hotel
Washington, DC
topic page DeLorean Hoffman
1 1 Nice to see you
2 1 Had lunch?
3 2 Nice building
4 2 alternate is in Rockville
5 2 want a drink?
6 3 I thought we'd be together
7 3 Where do you stand with
the company?
8 4 we've had delays
9 4 Prior to interim financing?
10 5 group has ability--30 million
11 5 They have an interest?
12 6 Colombian folks
(for a complete topic analysis in this case, see Shuy, Language Crimes, 1993)
This type of analysis gives a skeletal view of the entire conversation and shows the points at which the crucial passages must be addressed. It also provides a clear picture of which one of the participants introduces the most topics, which topics are recycled, which participant asks questions showing his lack of knowledge, and, most important of all, it gives important clues to the intentions of each speaker since the topics one brings up are the best indication of his agenda.
Discourse response analysis
Once the topic analysis is presented, the areas of disputed interpretation are set in the context of the entire conversation. If the target has not self-generated any guilt, the next step is to examine his responses to the overtures either explicitly or indirectly presented by the agent. Response analysis examines these passages in the context of the conversation and charts the responses next to the agent's overtures. The following example is from the 1980 bribery trial, Texas v Billy Clayton, referred to as the Brilab operation. The agent, Joe Hauser, posed as a representative of Prudential Insurance Company, trying to convince Clayton to let Prudential bid on the state's insurance.
C.I. Joe Hauser's topics Billy Clayton's response
There will be a savings of Any time you can save the
approximately a million dollars state a buck by God I'm for it
We want to make a contribution Let's get this thing, and
to your campaign try to take care of it first...
and uh, uh, then, then, uh,
then let's think about that.
Could I give you a contribution? Any time you can show me
I will, in your whatever you want where you can save the
to run. $100,000 going in and we state money, well, by God,
can prepare to put in a half a I'll go to bat for you...I think
million. that's what my job is, try to
save the state.
There's $600,000 every year. I'm Our only position is we don't
keeping 600 and 600 whatever want to do anything illegal
you want to do with it to get the or anything to get anybody
business. in trouble and you don't
either. And this is as
legitimate as it can be
because anytime somebody
can show me how we can
help save the state some
money. I'm going to bat for
it. But you know it'll be
reported.
Why do you have to report it? Well I don't want to get into
no damn tax--(interrupted)
You can report it later on, a year No, no, no.
from now.
The agent here cleverly tries to mix two topics: his insurance bid to save the state money and his offer of a campaign contribution. At the end of the meeting, Hauser left an envelope containing money with Clayton, who put it in his desk drawer. Clayton was then arrested for bribery before he had time to report the money as a campaign contribution. His responses to Hauser show that his mind was on keeping the two topics separate and that when Hauser escalates the amount of money, Clayton becomes suspicious and gives his exculpatory statements. The government did not put the topics and responses together in this way but after this response analysis was demonstrated at trial, Clayton was acquitted.
The importance of context
Clayton's indictment carefully selected isolated passages out of context, making the case against him look much worse than it was. It is commonly not too difficult to locate the parts of the conversation that are alleged to cause the defendant problems. Usually the government has already made this clear in its indictment. At this point it is too early to become discouraged. The major task of the defense is to contextualize these passages and not to let the prosecution isolate them as though they had no context at all.
For example, in a recent money-laundering case the prosecution pinned much of its case on the following passage, spoken by a corrupt IRS undercover agent to the target, an attorney:
This money we're doing right now is part of a ten kilo deal
that we did together a couple of weeks ago...I want to make
sure that if this guy gets busted...if they start looking at stuff
that he's done, I wanna make sure my name isn't in there with
him.
On the surface, this passage might seem inculpatory, but the conversation must be set in the context of what came before and after it. The target is an attorney who was originally approached by the agent, posing as an investor, for future representation if any possible criminal case should ever develop against him. Months later the agent sought out this same attorney to help him establish a trust account. The agent testified that, in his mind, his above statement had made perfectly clear that the money for that trust was generated from drug sales. For our attorney to have understood this, however, he would have had to do considerable inferencing.
The question here concerns just exactly how clear and unambiguous was the agent's statement. He did not, in fact, state explicitly that his money for the trust is drug money. Here is where linguists use discourse analysis and syntax analysis to resolve this issue. The specific analytical procedures are called 1) discourse focus and 2) referencing.
Discourse focus
A review of the entire conversation revealed that the agent made a lengthy effort to convince the attorney that he needed to have a trust set up as a procedure for purchasing oil stock with money from his offshore corporation. The attorney offered a number of other suggestions, none of which met the agent's need to involve and implicate the attorney in something illegal. Perhaps in desperation, the agent finally explained that the man from whom he was purchasing oil stocks told him to have his money placed in a trust for distribution rather than sent to him directly. The agent had expressed doubts about the character of this man. He had done business with him before and he worried about what might happen if "this guy" got arrested. The discourse focus of this conversational topic was clearly on how to best protect the confidentiality of the agent's name in this transaction so that he would not be implicated if the oil man ever got arrested. A trust, he said, would accomplish this. The discourse focus of the agent's speech was not on representing the illegal nature of the currently planned transaction but, rather, on some unspecified future danger that might be avoided by having a trust.
Thus, for the target to have understood that the agent was representing the illegality of this transaction, he would have had to infer that the agent's request to help protect his confidentiality was equivalent to a request to help him launder money, a rather large inferential leap. This leap is not consistent with the FBI guideline of making the illegal nature of the act clear and unambiguous.
Syntax referencing
The way people refer to past events or to other people is probably the single greatest contributor to confusion and ambiguity in everyday conversation. When the event or person is not referred to explicitly, the listener is forced either to infer or guess at the reference. Alternatively, he can request clarification. So why don't we request clarification? There is a very simple rule of conversation which virtually everybody follows but which is consistently ignored by law enforcement in their effort to capture crime on tape. It is called the cooperative principle. This rule tells us to be cooperative to the extent that we don't keep interrupting our conversation partners with demands that they be explicit and clear. It is considered more polite to hold off with such requests and hope that we will eventually understand what is going on.
It is uncooperative and impolite to say, "Stop, I don't know what you are talking about." It is even more impolite, if not dangerous, to say, "Stop, what you're doing is illegal and I don't want to have anything to do with this." For one thing, saying this is an accusation, something that can end a conversation, a business transaction, or a friendship with great rapidity. Undercover agents tacitly know that if they make inferences about illegality, the target is not likely either to request clarification or, even if they get the inference, risk accusing them of doing something illegal. Whole criminal cases often rest on the skills of agents in this matter.
Why then don't the agents make clear and unambiguous the nature of the illegal activity? An insight to the answer can be found in my experiences in training DEA undercover agents on how to use language effectively to capture street crime. After I explained to them that they have to be clear and unambiguous about the drug sale they are involved with or risk losing their case at trial, many of the DEA agents objected strongly, explaining that if they did that, the targets would not cooperate and would just walk away. It was clear to me from this that these DEA agents were considerably less interested in justice than in netting targets.
Some undercover agents and most cooperating witnesses, who need to catch their prey as part of their cooperation, are often no different from those recalcitrant DEA agents. For them the target is their opponent in a rather high stakes linguistic poker game. They are less interested in police intelligence analysis than in winning the game with a linguistically stacked deck.
In the example of the attorney and the IRS agent, the agent used several unclear references. "This money we're doing right now," for example is unclear for several reasons. The prosecution said that it referred to the oil stock transaction for which the attorney was asked to set up a trust. From the context, however, this interpretation was not all that clear. The preceding topics were about several ongoing ("right now") deals that the agent had been transacting with "this guy." It is ambiguous, at best, as to whether "this money we're doing right now" referred to the current oil stock transaction or to other current deals between the agent and "this guy."
If the agent wished to follow the FBI guidelines, it is odd that he chose the verb, "do," to express this action. "Do" is what linguists call a dummy verb, a substitute for virtually any specific or explicit verb, much in the same way that a pronoun is used to replace its more specific noun reference or antecedent. Speakers use dummy verbs for a number of reasons. They may not be able to think of the exact verb they want or they may want to avoid a sensitive word, such as "kill." In the Texas v Davis solicitation to murder case, for example, the cooperating witness consistently said, ("I'll do Priscilla" instead of "I'll kill Priscilla"). Or the dummy verb can be used to be deliberately ambiguous. In this case it is not possible to know the exact motive of the agent for using a dummy verb, but the result of the ambiguity is the same no matter what his motivations were. If the agent had used the verb, "washed," or "laundered" here, the representation of illegality would have been explicit and clear.
In addition to "do," the agent's use of "we" is also ambiguous. Pronouns in English are extremely vulnerable to multiple meanings which can often be clarified through syntactic analysis, which operates at the sentence level to disambiguate grammatical relationships. For example, when a speaker says, "I want you to do it," "you" is potentially ambiguous, since "you" can refer to one person present (the one being addressed), to more than one person ("you" is also the plural pronoun), or to some combination of persons present or absent with whom the addressee is associated (his company, his family, or his group). The same can be said, of course, for "your," "yours," "our," "ours," and "we." Even the pronoun, "I," has potential ambiguity, particularly when a speaker uses "I," "me," "my, or "mine" to refer to his/her role as employee or officer in a company rather than to his/her personal self.
When the agent refers to "this money we're doing right now" and continues that it is part of a ten kilo deal that "we did together a couple of weeks ago," even the prosecution agreed that the two uses of "we" here referred to different people. They claimed that the first "we" referred to people in that room while the second "we" included only the agent and the man from whom he was buying stocks. The crucial question was who the first "we" referred to. Since other transactions between the agent and the oil man were represented as current and ongoing, the target, our attorney, could have easily understood the agent to be saying that he and the oil man had engaged in several recent transactions, including one involving that ten kilo deal, and furthermore, that the ten kilo deal was different and separate from the oil stock purchase for which the attorney was being asked to set up a trust. This interpretation is supported by the agent's repeated concern that his name could be associated with that of the oil man because he feared that the oil man might get arrested and drag his name into the process.
Conversation strategies used by the agent
No matter what type of case it is, undercover agents employ surprisingly similar conversation strategies. The following strategies, contaminating the tape, securing the appearance of agreement, coaching the target, camouflaging the illegality, blocking exculpatory statements, and using language and culture differences, are the most common.
The Strategy of Contaminating the Tape
One of the tricks commonly used by undercover agents in tape cases has been called the principle of contamination. Agents, who wear the microphone and are the only ones who know that the conversation is being taped, have amazing power and control over what the final tape product will look like. In this respect, their power is much like that of making a documentary film. For example, if you overhear a conversation in which one person is telling a dirty joke, you may remember that event as both people telling dirty jokes.
An important topic in money laundering cases is that of confidentiality. Confidentiality is a common banking concept, not to be confused with terms such as "hidden," or "secret." In the Abscam case of Senator Harrison A. Williams, Jr. of New Jersey, the senator consistently specified that a blind trust would be appropriate for any potential ownership he might have in the planned business venture. It was the FBI agent, Tony DeVito, who consistently referred to this as a "hidden agreement." He sprinkled the tapes with words such as "hidden," and "secret," but never, of course, in the senator's presence. But the damage was done. The prosecution and the jury remembered the words but not who spoke them. The tapes were contaminated by the FBI agent, not by Williams. In the process of creating a scam operation, government agents or their surrogates (such as Mel Weinberg in Abscam, Joe Hauser in Brilab, or James Hoffman in the John DeLorean case) used language to create the impression of illegal behavior.
In the case of our attorney charged with money laundering, there is a significant contrast between the expressions used by the attorney and those used by the agent when referring to confidentiality, as the following table demonstrates:
Table 1
Comparison of selected expressions used
expression used Attorney Agent
confidentiality 14 0
my name won't appear 0 4
protection 10 8
can't trace my name 0 2
people start looking at me 0 5
insulate myself 0 4
Of particular interest here is that only the attorney, the target, uses the term, "confidentiality," of a trust agreement. The agent refers to it as a way to avoid having people start to look at you, of insulating himself, and to avoid being traced. Both speakers use "protection," but careful examination of the context in which this expression was used is revealing. In each of the ten times the attorney used the word, the discourse context makes it clear that he was referring to protecting the agent's money from being lost or stolen. In contrast, when the agent uses the expression, the discourse context makes it clear that he meant to protect himself from visibility. This issue was not particularly central to the trial, but it provides an example of how the principle of contamination can work against the defendant.
The Strategy of Securing the Appearance of Agreement
In all conversation there is a requirement for feedback. When one person talks, the other person is obliged to say something that
indicates that he is attending. Without such feedback, speakers stop and ask if the other person is listening. Common feedback markers include "uh-huh," "alright," "yeah," "okay," and even "right" when said with the proper intonation. The important point here is that feedback markers signal only "keep talking" rather than agreement with what was said by the other person.
Like all other conversations, those recorded in undercover operations contain hundreds of examples of such feedback markers, especially in taped telephone calls where it's necessary to let the other person know that you are still on the line. Even though such markers seldom, if ever, signify agreement, prosecutors, courts and juries often confuse them with consistent agreement. On the other hand, if the target should utter words such as, "I agree," "I'm with you," or "it's a deal," we can be much more certain that he has actually agreed. Agents should be versed in how to achieve true agreement from their targets and not accept feedback markers as evidence of guilt.
The Strategy of Coaching the Target to be Illegal
In some cases, there are two or more covert agents who speak with the target. In such cases one of the agents may coach the target in what to say when he speaks with the other agent. The Abscam case of Senator Williams provides a classic example. In one meeting confidential informant Mel Weinberg scripted Williams with explicit instructions about exactly what words to use in the following meeting with an FBI agent posing who was posing as an Arab sheik, previously corrupted mayor Angelo Errichetti, and FBI agent Tony DeVito. Interestingly, Williams used none of these expressions in the follow-up meeting, as the following chart shows:
Weinberg's coaching topics introduced by:
Errichetti DeVito Williams
how high you are X X
you're 5th in the Senate X X
who you know, control X
chairman of whatever X
how important you are X X
no deal without you X X
you put this together X
get government contracts X
your influence for contracts X
who you are X X
you open doors X
you use your influence X X
you guarantee this X
you can produce X
blow your own horn X X
you can move this X
sell him like mad X X
you're the boss X
throw names around X
come on strong X X
This is a classic example of contamination as well as failed coaching. In the follow-up meeting Williams used none of the expression coached by Weinberg. The resulting tape recording, however, gave the impression that Williams had actually said what he was coached to say. During his Senate impeachment hearing, several of Williams' fellow senators came up to him and asked why he did all that bragging about himself. As the above chart clearly shows, he did none of it; the agents did. Coaching is the equivalent of a leading question and should be frowned upon for its contaminating effect even if, as in the Williams case, the target ignored it.
The Strategy of Camouflaging Illegality
When an agent uses the camouflaging strategy, he tries to make something illegal or important seem legal or unimportant. They can do this in a number of ways. One strategy used in the Abscam operation was to declare the conversation "unofficial," much in the same way that interviewees sometimes claim that their exchanges with reporters are "off the record." No conversation in undercover cases is ever "unofficial" or "off the record." In the Brilab case of Representative Billy Clayton, the agents tried to make the conversation appear to be about giving Clayton a campaign contribution while their real goal was to get him to accept a bribe. As seen in the above example, after first defining it as a contribution they then tried to redefine it, although somewhat ambiguously, by implying that it was a quid pro quo for Clayton to open up the state's insurance contract bidding. When Clayton replies that he would report the campaign contribution, the agents then try to discourage him from doing so, leading to Clayton's statement that neither he nor they wanted to do anything illegal. The camouflage strategy is common in government taping. It is not unambiguous, despite the FBI's claim that all such overtures must be clear and unambiguous. In contrast, it thrives on deliberate, created ambiguity and confusion.
The Strategy of Criminalizing the Tape
By criminalizing the tape I mean instances in which the agent peppers the tape with concepts, references, and vocabulary that seem to fit the atmosphere of illegality. In the Williams Abscam case for example, there was considerable discussion about how the Senator would handle any profit he might gain from the proposed business venture. Note the contrast between the way Williams and the agent refer to a blind trust.
FBI Agent DeVito Senator Williams
Keep it a secret. Pay the taxes.
Everything is going to be hidden. I'm going to find a way to
protect myself with some
kind of declaration.
Everybody can declare...you can't. I'm going to have to go
public with something or
other.
You were gonna declare it but in
some other way. We can blind trust me.
When the Senator said he wanted to A blind trust, that's the way
declare...he was trying to protect for my purposes.
himself by coming up with some
kind of gimmick. We have it under the trust.
Despite the clear differences in the terms used by the Senator and the agent, the prosecutor insisted that Williams was after a hidden, secret, gimmick instead of a blind trust. He, along with the court, the jury. and even the Senate Ethics Committee were somehow able to overlook the Senator's objections to the agent's language. They were all contaminated by the agent's criminalizing of the tapes, even though the agent's agenda of trying to get Senator Williams to hide his interest in the proposed venture failed, the damage was done by the criminalizing itself.
Other examples of gratuitous criminalizing the tapes include the agent's vague or oblique references to things that, in some contexts, might suggest illegality. Undercover agents trying to tiptoe around the requirement to represent drug money sources in a clear and unambiguous manner, for example, often resort to references such as, "the money's from down south," or "these men are Colombians."
In the case of one stock broker indicted for money laundering, the agent was characteristically unclear about the source of the money he proposed for the purchase of bonds. As it turns out, the broker was earlier led to believe that when the agent's said his money was "from down south," this referred to the fact that the money had Mexican flight capital origins. The agent's many references to "down there," "the hot place," and "down south" were consistent with that understanding. Then the agent tried to up the ante and referred to Columbia, a fact that the prosecution relied on as a clear representation of money generated from cocaine sales. The government ignored the possibility that the target could have understood that the Colombians may well have resettled in Mexico and now were fleeing to the US. And since when did the word, "Colombian," become synonymous with "cocaine?" It stretches the imagination of anyone with South American friends and contacts to believe that all Colombians are in the drug business.
The negative effect of such criminalizing can be devastating. It might be argued that the FBI is not overstepping its ground by appearing criminal to the target. But when the ultimate effect of this appearance leads to misreadings or misunderstandings by jurors, this strategy must be seriously questioned. The person who does not know that a taped record is being made of his conversation focuses on the content of the message, not the manner in which it is said.
The Strategy of Blocking Exculpatory Statements
It cannot be denied that law enforcement's major goal of recording conversations in the first place is to capture illegal acts. In the effort to do this, however, agents are often severely tempted to block statements in which the target starts to clarify his position or reveal his rejection or denial of any suggested or implied illegality. The most common way to do this is to interrupt at the point when the target appears to be denying involvement. This strategy is common in sales pitches and many other areas of everyday life.
In fact, most conversations include interruptions, making this strategy seem more innocuous than it is. It is not the fact of interrupting that is dangerous here as much as the placement of the interruption.
In video taped conversations, the telephone interruption is commonplace. While one agent talks with the target, other agents monitor the conversation in another room. If things are not going well for the agent in contact with the target, the observing agents call him to the phone and offer advice. If the phone call happens at the right time, the target's responses can be mutilated by the telephone interruption. A classic example of this is the Williams Abscam case when an agent disguised as a Arab sheik offered Williams money as a quid pro quo:
Agent Farhart: I, I will for, for your help, er, assistance and
assistance. I would like to give you, er, you know, some
money for, for permanent residence.
Sen. Williams: No, no, no, no. This is when I work in that area,
that kind of activity, it is purely a public, not, er, no....
Within my position, when I deal with law and legislation,
it is, it is, it is not on, it's, it's, er, not with, within--
At this point the telephone rings and interrupts the rest of the senator's sentence. After the sheik hangs up, he immediately introduces a new topic, leaving the senator's possible explanation about how he cannot accept such money unsaid and despite the fact that the senator never took the money, he was eventually both accused and convicted of accepting a bribe (for other examples of the interruption strategy, see The Language of Confession, Interrogation and Deception, Shuy 1998).
The Strategy of Using Language and Culture Differences
In some money laundering operations the government has employed savvy and convincing street people to do the taping. For example, in a large money laundering case in the DC area a few years ago, a Black agent tried to use large amounts of cash to purchase automobiles at several local dealerships. He spoke vernacular English and attempted to give the appearance of a drug dealer without saying as much. Perhaps from the target's perspective, many wealthy people in the sports and the entertainment industries might well talk and look like this man. The undercover agent used imprecision and feigned or real inarticulate language to send his signal, thus not only creating vague and ambiguous messages but also ones the targets were disadvantaged to overcome. One simply does not insult clients or prospective customers by criticizing or correcting their behavior. One even runs the risk of offending by requesting clarification too frequently.
The culture and language difference strategy involves people who are presumably from a different culture and different language background. Since their words are often inexplicit and their sentences garbled, the listener has to infer a great deal. The conditions of politeness are heightened and the target is forced to tolerate strangeness in behavior. When standard English speakers converse with those who speak a non-standard variety, the latter are given certain accommodations. It is not good taste to tell a person with a foreign accent that they should talk and act like us. Some of Senator William's many critics told him that he should have simply informed that sheik that it was dishonest and illegal to offer him a bribe. But would they have told a foreign dignitary, whose culture may operate under entirely different rules from ours, that he is a scumbag? And would the DC auto dealers tell a well-heeled customer that they didn't want his money? It is simply not that easy in a cross cultural situation. And the agents, knowing this, often use this strategy to nail their targets.
The culture and language difference strategy puts the target at the disadvantage not only of trying to decipher the meaning of the unusual behavior, but also at the even greater disadvantage of trusting that later listeners, such as juries, will also be able to understand the meaning that they got out of the event. This strategy can lead, as it did in the cases of Senator Williams, the DC car dealers, Billy Clayton, and many others, to an inaccurate and unfair perception of the event, even though the objective data which are immediately available should lead to a totally different conclusion.
Conclusion
To this point I have outlined the steps for using linguistic analysis in cases involving tape recorded evidence, alleged language crimes, including money laundering--producing a jury-ready transcript, analysis of supposed hard parts, contextualizing, and finding passages that support innocense. I have also pointed out seven of the common language strategies used by undercover agents. Each of these strategies departs from the FBI guidelines. Even the techniques used in videotaping can contribute to the appearance of wrongdoing. If poor quality videotape is used in a poorly lighted room, using fixed angle cameras, a magnificent hotel can be made to appear like a sleazy flophouse. By focusing the camera on the wrong person, an important conversational statement can be totally missed or an ambiguity can be created about who was actually speaking to whom. Such ambiguities work for the prosecution unless the defense works hard to point it out, usually with the help of a linguist.
Another crucial task is to analyze the conversational strategies of the target. Using topic analysis and response analysis, the target's agenda or conversational goals can be objectively identified. No science, including linguistics, can reveal the actual intentions of anybody, but the topics that people introduce and recycle, along with the patterns of their responses to topics introduced by the agent, provide very strong clues to such intentions.
Some twenty years ago, Paul R. Michel, then Associate Deputy Attorney General of the U.S., observed: "The honest man simply rejects the offer and departs." But for the target unlucky enough to be indicted for using a feedback marker "uh-huh," rather than a true agreement, Mr. Michael's statement has a hollow ring. If the offer has been camouflaged into looking like something very different and the target is indicted, Michael's statement does not hold water. If the target is coached or scripted to say something other than what he intended and then is indicted, Michael's statement is problematic. If the target is isoloated from the information that others have or if he succumbs to pressure to go along with others who have that information, Michael's observation is untrue. If the target is confused by the garbled language of a non-standard English speaking agent, does not catch the subtleties of that garbled speech, and is then indicted, there is no basis for truth in Michael's statement.
An honest target certainly can reject the offer and depart as long as the offer is clear, the option is open, the conversational stragegies are fair, and the deck is not stacked. The seven strategies outlined here are all found in recent government surreptitious tape recorded conversations. They are not unique to the cases cited. But they certainly do bring indictments and convictions.