(False) Confessions Become Compelling at Trial
Chapter 4
(False) Confessions Become Compelling at Trial
Storytelling has a pragmatic efficacy. In pretending to recount the real, it manufactures it. It renders believable what it says, and it generates appropriate action. … The voices of narration transform, reorient and regulate the space of social interaction. (de Certeau 1986, 200)
False Confessions as Evidence
Confessions are given inordinate weight at every stage of a criminal case. Interrogators are trained to pursue them, but interrogation can induce innocent people to confess falsely. And false confessions are counterintuitive: most jurors cannot imagine that they would ever confess to a crime they did not commit (Kassin 2002). In two large studies of proven false confessions in the United States, false confessors who pled not guilty and took their cases to trial were convicted by juries 73–81 percent of the time (Leo and Drizin 2004).
Academic studies of wrongful conviction cases show that 15–20 percent contained confessions in evidence (Leo and Drizin 2004; Gross et al. 2005). Of over 250 people exonerated by DNA testing, more than 40 confessed falsely to rapes and murders (Garrett 2009).
Studies with mock jurors have shown that confessions are powerful even when they are known to have been elicited under pressure, are not consistent in their chronology or compatible with forensic or witness evidence, and are retracted soon after they are given. These studies have consistently shown that “confession evidence boosts conviction rates even among jurors who believe the confession was coerced” (Kassin et al. 2009).
Neither police nor prosecutors nor judges expect the post-interrogation confession to be complete, coherent or completely true. Police interrogators expect suspects to withhold and distort information. They expect “lying and gauging” (cf. the Hayat rebuttal page 63 of this chapter). Prosecutors tell the story that way: defendants’ admissions are incomplete or inconsistent because defendants are wily, lying, hiding key information, protecting themselves.
How do suspects come to admit after initially denying involvement in the crime? What makes them stop telling what we know in retrospect to be the truth, and be willing instead to take the officers’ cues, imagine, hypothesize or simply play along and “lie”?
There is now a large and growing literature by psychologists and linguists which attempts to explain this phenomenon. They have studied the way false confessions arise from the use of tactics and ploys used by interrogators (see, for example, Ofshe and Leo 1997a; Shuy 1998), the effects of different kinds of interrogation questions (Tully and Cahill 1984; Oxburgh et al. 2010) and the psychological vulnerabilities of the suspects who make them (Gudjonsson 1992; Kassin and Gudjonsson 2004). They have determined that children and juveniles are especially vulnerable to interrogation, especially when it is accusatory (Redlich and Goodman 2003; Drizin and Colgan 2004; Redlich et al. 2004), as are mentally retarded and mentally ill suspects (Cahill et al. 1988).1 Non-native or non-English speakers interrogated without effective interpreters are also particularly vulnerable to interrogation (Berk-Seligson 2002; Berk-Seligson 2009).
Proven false confessions show that police officers can alight on the wrong suspect, sometimes in ways that reveal racial or cultural bias. At times they misinterpret behaviors and demeanor. Interrogating officers may assume a suspect is guilty, or be under pressure to solve a crime, and in an accusatory, guilt-presumptive interrogation drive for an admission, and whether intentionally or inadvertently, supply critical details of the narrative (Hill 2003; Garrett 2009).
Police-induced False Confessions
Research has shown that invoking a suspect’s rights as prescribed in Miranda v. Arizona2 does little to protect against “unknowing” rights waivers (Leo 2008; Ainsworth 2010), uphold the right to an attorney (Ainsworth 1998; Ainsworth 2010; Solan and Tiersma 2005), prevent self-incrimination or protect against false confessions.
According to Miranda, a suspect should be able to control the interaction during police interrogation. The suspect should be able choose whether or not to speak, how much to say, which questions to answer, when he or she needs a break, and when to end the interrogation.
Interrogators, however, are trained to dominate the interaction (see Inbau et al. 1986; the John Reid website;3 Walkley 1987) and they do this by various verbal and non-verbal means, including bombarding the suspect with large quantities of talk, interrupting continually, controlling the topic, questioning accusatorily, shouting and ignoring the suspect’s denials.
Some innocent suspects cannot maintain their innocence in the face of this coercive interaction. They reach an emotional breaking point and begin to admit. After this breakdown (or, according to the prosecution, “breakthrough”), the false confessor finds relief, a semblance of synchrony (Gumperz 1982) and a degree of interactional harmony by agreeing to tell the story, to cooperate, to work with the officer to construct an account of a crime.
The final push by the officers and the breakdown by the suspect are almost never recorded. And even when the entire interrogation has been recorded, we can never hear those crucial comments made in the car, in the hallway, in the restroom or when the tape is being changed, though we can sometimes see what Richard Ofshe calls their “footprints” (Ofshe and Leo 1997a; Ofshe and Leo 1997b).
When the line is crossed from denial to admission, real world consequences flow: from false confession to charge to custody to trial to conviction. There is no more pivotal moment in the entire criminal justice process. And yet interrogators, like jurors, believe that innocent people do not stop denying. As the Reid interrogation training says:
It is very rare for an innocent suspect to move past this denial state. He remains steadfast in the assertion of his innocence. The sincerity with which he presents his case is a clear indication of his truthfulness.
Research has also shown that “actual innocence” does not protect suspects, but may put them at risk (Kassin 2005). Suspects waive their rights, “naively believing in the transparency of their innocence,” and once armed with a confession, police investigators often proceed with “tunnel vision,” excluding alternative lines of investigation and other possible suspects (Findley and Scott 2006; see also Solan 2010).
Once the suspect has “admitted” and the videotape is on, the interrogator struggles to elicit the kind of details that confirm and corroborate the “admission” or confession. It is this post-admission interrogation that needs more study (Leo 2008).
The purpose of the post-admission phase of interrogation is to confirm, corroborate and fill in the details of the admission. In Reid interrogation training, this is Step Eight, and here “the investigator’s objective is to obtain a brief oral review of the basic sequence of events involved in the commission of the crime, while obtaining sufficient detail to corroborate the suspect’s guilt.”
Haworth (2010) writes of the “interview as evidence.” She cites Baldwin (1993), saying that “instead of a search for truth it is much more realistic to see interviews as mechanisms directed towards the ‘construction of proof.’ At trial it is de and recontextualized” (Briggs 1996). As Cotterill (2002) puts it: “each part of the (re)constructive process of the crime narrative is potentially the site of influence from previous interviews and hearings and, equally significantly, has an eye on potential future appeal proceedings.”
Mysteriously, false confessors tell about people and places, actions and events they never actually observed. They describe conversations they never heard and feelings they could not have felt. They seem to draw on memories we know they do not have. And yet these details, co-produced during the post-admission portion of interrogation and all unfounded, are transformed by the prosecution at trial into a compelling story. Impugned by the defense, the confession is revived and relegitimized by the prosecution as emotive and highly probative evidence.
Leo and Davis (2009) point out an important gap in knowledge: too little is known, they say, about what leads inexorably from a police-induced confession to a conviction. To fill this gap, the “post-admission” stage of interrogation needs more study (Leo 2008). Beyond this, we have to understand the way the confession and interrogation are repackaged as evidence and used strategically at trial. It is at trial that both the pre- and post-admission interrogation become the grist for “adversarial storytelling” (Amsterdam and Bruner 2002), and the defendant is characterized through them and other evidence, as well as by his or her demeanor in court (Levenson 2007), as the kind of person who is capable of committing the crime in question.
We look at the confession’s repackaging at trial because that is where the jury meets it. Jurors may have been exposed to the confession by the media when it was repackaged into an affidavit or charge long before they became jurors. But the only real “confession” the jurors take to the deliberation room is the one that has been presented to them during the examination and arguments of a trial.
At trial, jurors may be influenced by the believability, verisimilitude and compellingness of the prosecution story, persuaded by what clinical psychologist Donald Spence (in Ochs and Capps 2001) calls narrative truth. Narrative truth depends on “continuity and closure and the extent to which the fit of the pieces takes on aesthetic finality” (Spence, in Ochs and Capps 2001, 285). Historical truth, on the other hand, aims to come as close as possible to what “really” happened. Is the confession reliable? Does it “fit” what else is known about the alleged crime (Ofshe and Leo 1997b)? Are there inconsistencies in it (Shuy 1998)?
Discovering the answers to these questions should be the goal of jurors considering confession evidence that has been retracted, is weakly or not at all corroborated, and according to information presented by the defense, likely to be false. As Heffer (2002; 2010) points out, a jury trial is a complex genre whose opening address and closing arguments are in story mode. Jurors make decisions to a large extent through story-making, transforming narratives which emerge in testimony to “reconstruct reality in the courtroom” (Bennett and Feldman 1981; Pennington and Hastie 1994; Amsterdam and Bruner 2002).
Laurence Rosen (2006) describes the process of collective decision-making that juries undertake as they strive for a verdict. Individual jurors “put together a narrative as they individually listen to the trial, matching each new piece of information to the story they’ve been telling themselves along the way.” When they adjourn to discuss the evidence they work towards a “collective narrative, one that necessarily depends on the ways in which narratives are generally created in their culture” (Rosen 2006, 148).
The “confession” the jurors take to the deliberation room, then, is the one that has been shaped for them during the trial, and that they then shape again together during deliberation. The prosecution rebuttal argument is the final filter through which they hear about the confession, and it can pack an “emotional punch” (Read 2007). This final narrative doesn’t “simply describe ready-made events,” but provides “central means by which we create notions as to what took place and how the action unfolded” (Briggs 1996, citing Bauman 1986).
Jurors adjourn for deliberation after the prosecution rebuttal argument and after receiving instructions from the judge about how much weight to give the confession. These instructions may also play an important role in the jurors’ decision-making.
Human memory does not distinguish among various sources of information; it doesn’t distinguish who said something, but only the fact that it was said, (the fundamental attribution fallacy); it doesn’t consider the conditions under which something was said. The rebuttal plays on these weaknesses. The jury will remember the confession through this filter.
False Confessions at Trial
Although we have constructed an elaborate system of evidentiary rules and courtroom procedures, an American criminal trial is much more than a mere sum of its evidentiary parts. Rather, it is a theater in which the various courtroom actors play out the guilt or innocence of the defendant for the trier of fact to assess. (Levenson 2007)
The fabrication of legal evidence out of a false confession is effective even though it is based on fiction. It is constructed with the same attention and intentionality as any piece of (acknowledged) fiction, striving for verisimilitude. However, because the confession on which it is based is not identified as fictional but rather as sincere, serious and true, it is not critically analyzed.
Political scientist Kristin Bumiller, writing as “an observer of spectacles of interracial gang rape,” describes the way the prosecutor in the 1991 Central Park Jogger trial builds a narrative out of what we know now to be false confessions. According to Bumiller, the prosecutor is author of a story about the crime that “has the explicit purpose of making the case for the guilt of the defendants.”
In this important case, one black and two Hispanic teenage boys were charged with the sexual and physical assault of a young professional woman in Central Park. The prosecution case was based on forensic photography of the woman’s body and the written and videotaped confessions of the defendants. Bumiller says that in this trial, the prosecutor’s primary role was as a “scenographer” who created “narrative structure” by arranging and placing objects, persons and events within a “landscaped vision.” In closing argument, “the prosecutor tells the story in a way which suggests that the jogger is ‘picked’ as the victim; as she runs North on East Drive, she becomes ‘illuminated for the defendants and others The boys took ‘positions’ behind trees and ambushed her: she fell to the ground and then was dragged into the woods and darkness off the roadway.”
Bumiller’s interpretation focuses on how prosecutors produce stories that do more than satisfy “love of justice.”
In a recent Stanford Law Review, Brandon Garrett (2009) looked at 42 cases of proven false confessions amongst 252 DNA exonerations. Rather than join scholars who look at “the psychological techniques that can cause people to falsely confess and document instances of known false confessions,” he takes a different approach. Garrett examines the substance, or content, of false confessions, including what was said during interrogations, and then, importantly, how the confession statements were litigated at trial and afterwards.
In this chapter I take a similar approach, although I look at the life of only one problematic confession: its “production” at FBI headquarters, and the ways it was strategically repackaged and commented on at trial. We see how the defendant’s own original story was rejected by his interrogators and replaced by the government story, how that story went to trial as a probatively powerful confession, and the strategic repackaging of the confession in closing argument. This retelling used a number of narrative devices to create a compelling narrative.
The Hayat case, like other anti-terrorism cases in the United States after September 11, 2001, is one which the prosecution characterized as pre-emptive: its purpose is to prevent a crime. No crime had actually been committed. In fact, it may have been wholly fabricated, a case built on entrapment and self-incrimination, where actual truth may have been trumped by narrative truth. And yet the repercussions for the defendant, his family and the Moslem community in Lodi are pervasive.
There are instances of true stories suppressed and supplanted (under duress) by false stories which are then so compellingly told (by the state) that they convince decision-makers of their reliability. This is what happens when a person is convicted on the basis of a false confession, and Hayat’s trial may be an example. In the closing arguments, the original interaction on which the evidence was co-produced (both between the cooperating witness Khan and the defendant, Hamid Hayat, and between the police officers and Hayat) is erased while other, imagined relationships are created and new interpretations made. By using as evidence words that were spoken by the defendant, these closing arguments gain a superficial air of authenticity (Komter 2002) especially with regard to the so-called confession – a speech act which always seems to generate an expectation of sincerity and seriousness (Brooks 2000).
The Case
Surveillance and Arrest
After September 11, 2001, the United States government began to identify and investigate Pakistani American youths who had been to Pakistan to study at religious schools (madrasas) affiliated with banned militant groups. Twentytwo-year-old Hamid Hayat, born in Stockton, California in 1983, was the first to be arrested.
Hayat is part of a small, largely working-class Pakistani community whose families began to arrive in Lodi during the early 1900s to work on the railroads and in the fields. The FBI’s Joint Terrorism Task Force, set up in Sacramento after September 11, had been investigating the Lodi Moslem community since 2003. Cooperating witness Nassim Khan had infiltrated the community in 2002, and begun secretly tape-recording his conversations with Hamid Hayat in March 2003. In February 2006 the Director of National Intelligence, John Negroponte, told the Senate Intelligence Committee that federal authorities had uncovered a network of Islamic extremists in this San Joaquin County farming town.
Hamid Hayat was first interviewed on May 30, 2005, when his flight from Pakistan was rerouted to Japan. He was interviewed by an FBI agent and deemed safe to return to the United States.
When Hayat was eventually charged with three counts of lying to the FBI, the interview in Japan was the basis of the first count. The second count was based on an interview on June 3, 2005 at his home. At this time FBI agents took various publications, Hamid’s scrapbook and a tawiz (traveler’s prayer) from his wallet.4 Hamid and his father, Umer Hayat, were asked to come to FBI headquarters the next day to help with the investigation. When they arrived, Hamid was taken into an interview room to be questioned by Agent Harry Sweeney. Sweeney read Hayat his rights at 12.30 p.m. The third count was based on the first three and a half hours of a four-hour interrogation during which Hamid again answered questions about his time in Pakistan, telling the agents that he played cricket, helped his mom, and spent time with his cousins.
Interrogation and Admissions
A false confession raises a challenging question: what does it take to get an innocent suspect to put his or her own reality on hold, to move into the frame of guilt insisted upon by the interrogator, and then to produce a fictional account in apparent cooperation with this powerful “co-teller” (Ochs and Capps 2001, 259e)?
At 4 p.m. Hamid Hayat stopped denying his presence at a training camp and made what Agent Sweeney considered his first “admission,” saying that he had been to a training camp in 2000. Forty-five minutes later he made his second admission: that he had been to a training camp in 2003.
Hayat’s confession has all the risk factors that are by now well known from the analysis of proven false confessions: the defendant was young, not exceptionally bright, with a compliant personality. His language abilities in English were weak, and at the time of interrogation he was exhausted because of the time difference between Pakistan and California, and from working the night shift. We do not know whether he understood his rights, and he certainly did not understand the implications of waiving them. Although he was interrogated for over thirteen hours, he never seemed to grasp the fact that he was a suspect; he thought, and was allowed to think, that he was helping the agents with their investigation.
This crucial first three and a half hours of the thirteen-hour interrogation of June 4 and 5 were not videotaped, so our only way to evaluate the voluntariness of the confession is through Agent Sweeney’s notes and the things he said during cross-examination. Hayat’s initial admission came after repeated denials and the use of several well-known interrogation tactics.
Before what Sweeney calls Hayat’s first admission, Sweeney used a “minimization” tactic (Inbau et al. 1986), suggesting to Hayat that perhaps he had been to a religious rather than a military training camp, mitigating the serious of the accusation. This tactic of Sweeney’s also framed the admission as hypothetical, another familiar interrogation tactic. It took repeated questioning, but when Hayat finally answered that yes, it was possible, perhaps all his subsequent answers also referred to the possible and not the real.
To secure Hayat’s second admission, Sweeney used two other familiar tactics: he told Hayat, falsely, that there were satellite photos of him at a training camp, and later, that he had failed his polygraph test.
The Post-admission Interrogation
Once FBI Agent Sweeney considered that Hayat had made admissions, he made ready to videotape and handed Hayat on to Agent Schaaf, whose job (with the occasional help of Agent Leticia Lucero) it was to seek details to corroborate, verify and bring to life the sparse admissions. Two further interrogations were conducted that same evening, and both were videotaped. During these nine hours, agents and suspect worked together until the early hours of the morning to shape “the unfolding tale” (Ochs and Capps 2001), the joint production of an admission narrative. What began as two opposing stories was transformed into one—the FBI version—and became the key evidence in the trial.
At the beginning of the first tape-recorded interrogation, Special Agent Gary Schaaf encourages Hamid Hayat to take off where he’d left off (GS is Agent Schaaf, HH is Hayat, and LL is Agent Leticia Lucero, another FBI interrogator):
GS: Now my understanding uh Hamid is that you uh you attended some camps.
HH: OK.
GS: Um and that’s that’s true right? (HH nodding)
In a coerced-compliant false confession, the interrogator sometimes leads a suspect to his first admission by asking him for a hypothetical description. This seems to have happened during the first interrogation. During the second interrogation, the agent invites Hayat to imagine a scene and tell what he sees:
GS: So, all right. And uh, so tell me, what put, put me in the front seat of the bus. What am I seeing as you’re coming into the camp? Uh here I am, I mean through your eyes tell me what you’re seeing as you come out.
HH: Um just go inside and you’re not gonna go try to take bus inside.
When the agent tries the same approach again a little later, Hayat does not seem to understand and fails to come up with anything:
GS: So that’s, do you go up the main route into the camp?
HH: I mean, on walk we go?
GS: Yes.
HH: Yeah, we walk.
GS: OK.
HH: We walk.
GS: OK. And so tell me, alright, take me on the walk. What happens on the walk?
HH: Nothing happens.
When the agent tries to get a description of the camp setting, this is the exchange:
GS: You said this was almost in the jungle.
HH: (unclear)
GS: What, well, like jungle ah to me means like ah, you mean like pine trees?
HH:(unclear)
GS: Like, like the mountains?
HH: Regular trees, yeah like that ….
GS: Ah, so you like, like just lots of tall, very tall trees.
LL: Like a forest.
HH: Yeah.
LL: Or just trees.
HH: Trees.
When dominant co-tellers such as police interrogators insist that the truth is a lie and “contend that certain unremembered events transpired” (Ochs and Capps 2001), the story that emerges often has signs of what linguists call the language of uncertainty—it is full of vagueness, hedging, pauses and self-repair. It may contain the subjunctive and conditional formulations indicative of what Ofshe and Leo (1997b) call the “grammar of confabulation.”
Throughout Hayat’s interrogation, we hear the interrogating officer, who is trying to ask questions about something he knows little or nothing about, and the defendant, who is trying to cooperate by answering his questions, both using this language of uncertainty. Agent Schaaf is trying to help Hayat describe a place where Hayat may never have been and which in fact may not even exist. Here he suggests that there are usually hundreds of people at a training camp:
GS: So, um during the course of the time that you’re there um how many people did you ob— I mean, usually these camps have, you know hundreds of people coming and going so that I assume that this is, the way you’re describing it, would sound similar in that regard. Is that, is that right?
HH: Like hundreds of people.
GS: Yeah, yeah, you know.
HH: I didn’t see that much like uh over 50 people I’ll say. I didn’t see over 50 people there, sir.
Schaaf persists. Perhaps Hayat means he saw 50 people at a time but there could be more, and Hayat complies. The number rises:
GS: At any one time?
HH: One time you know I see, I see like you know all together all the time I like you know, about like 70.
In the early hours of the morning, when Hayat was asked again about possible targets of terrorist action, he was still vague:
HH: Like buildings and I’ll say buildings.
TH: What kind of buildings?
HH: Bigger building, you know, buildings.
TH: Buildings? Commercial buildings?
HH: You know commercial, projects and like those kind of buildings. I’ll say.
And a little later, asked yet again,
TH: But I need you to tell me details about targets, what they said you know. And, this is where I need your memory to come back.
HH: Like I said sir, you know, big buildings and you know hospitalities and you know, finance buildings, banks and what’s it called ah, hmmm maybe like you know uh stores, stores.
TH: What kind of stores?
HH: Stores, like food stores, anything like that.
This exchange was cited in the first affidavit for Hayat’s arrest, which read: “Hamid advised that he specifically requested to come to the United States to carry out his jihadi mission. Potential targets for attack would include hospitals and large food stores” (affidavit, June 2005). Agent Timothy Harrison offers to help Hayat in exchange for his cooperation. One example of several is in this excerpt from his interrogation. It’s 2 a.m.:
TH: OK I’ll meet you, I’ll meet you halfway, alright.
HH: What’s that mean, halfway? You want me to cooperate.
TH: Cooperate.
HH: Cooperate, yeah I get that, sir.
TH: Yeah, you gotta give me something, in order for me to go help you.
HH: OK.
TH: You know, to tell my bosses that you’re cooperating, that you’re working with us. You know, you’re in a bad situation, you know, but it could be a lot worse, and there’s a lot we can do to help you. But you got to help us, you gotta work with us. (HH nodding) All right.
Like people who have made what we know to be false confessions, Hayat did not seem to understand the implications of his admissions. Like others, he may have been aiming simply for a short-term result, an end to the interrogation. Hamid repeatedly asked to see his father, he offered to return another day, he offered to take the agents’ numbers and report anything suspicious to them. At about 6 p.m., after two hours of interrogation by Agent Gary Schaaf, Hayat seems to think the interview is over. He asks Schaaf:
HH: Do you have a card, sir?
GS: I will. I don’t have one on me.
HH: I contact you (unintelligible).
GS: Yeah.
HH: How ‘bout beeper to (unintelligible) remember anything to contact you.
Later, after nearly thirteen hours of interrogation, Hamid still did not seem to understand that he was a suspect. He offered to return the next day. He was surprised that he was about to be arrested and that he was going to jail. He had stopped asking for his father. Now all he wanted was sleep:
HH: So I come back here tomorrow? Again.
TH: No, no. You’re not leaving here tonight, no.
HH: No, I mean ah, tomorrow. I’m going to be here tonight. Staying here? In the building?
TH: No, no, you’re going to go, you’re going to go to jail.
PA: Hamid, you’re going to jail.
HH: Yeah, so am I going to get a place to sleep over there like that?
PA: It’s jail, Hamid, you understand that?
HH: Yeah, I know, I know it’s a jail, but can I lay down because my head (HH points to his head) is hurting, I want to sleep.
The Confession at Trial: Adversarial Storytelling
At the trial we hear about three aspects of the confession: the interrogation itself (Was it given voluntarily, is it reliable?), the crime (Did Hamid Hayat go to a training camp and get trained there?) and the character and intent of the defendant based on “his own words” and actions.
First, we will look at the confession’s evidential value. If the apparently remembered experience of being at a training camp is not reliable, then there is no evidence with which to charge Hamid Hayat. As the defense attorney said at trial:
And so that’s what you have, Ladies and Gentlemen, you have a meaningless confession. You have Hamid Hayat being intimidated into saying things that the FBI wanted him to say …. The entire interview was meaningless, and it does not prove that Hayat went to a camp. And because [Hayat’s] statements come from a meaningless confession and they are completely unreliable, the government ladies and gentlemen, is left with nothing.5
But the prosecution case depends on a jury beginning its deliberations with the confession intact. Because the government has the burden of proof, the prosecution gets the last word with a “rebuttal” argument. David Deitch uses this “extremely powerful tool for the plaintiff” (Read 2007, 279–80) to great effect.
In rebuttal, Deitch discredits the defense closing argument.