The scene of the crime was the small two-bedroom house at 8767 Birmingham Dr. in South Austin. The victim was Sylvia Reyes Holt, a gregarious 36-year-old mother of one and employee of computer chipmaker Spansion Inc.
Two days before Christmas 2010 Sylvia was found dead in her home by an Austin paramedic, dispatched after a friend of hers called 911. He'd seen her just after her shift ended on Wednesday evening, Dec. 22, but uncharacteristically, she never arrived at work the next morning. He drove to her house just before 5pm on Dec. 23; her Suzuki SUV was in the driveway. No one answered his knock – could she have had an accident, or fallen ill?
Paramedics arrived four minutes later. Medic Pete Israel tried the front door, then hopped the fence and slipped into the living room through an unlocked sliding door. He let in his partner, and in the bedroom they found Sylvia, dead on the floor and naked except for black panties, white socks, and a black scrunchie around her wrist.
According to the medical examiner's report, Sylvia had been strangled by something long and thin. Curved abrasions crossed her face and neck. Near her body, police found a pair of green panties, the straps of which were stained and stretched and appeared roughly the same width as Sylvia's abrasions, said Charlie Baird, a former state district judge who represents Sylvia's husband, Jonathan Holt.
Nearly five hours later, Austin Police Detective Richard Jennings traveled roughly four miles from Sylvia's house to the home she used to share with Holt and their then-8-year-old daughter, Jade, to make the death notification. Sylvia and Jonathan Holt had been living apart since late summer 2009, and had nearly completed their divorce. When Jennings arrived, accompanied by another detective and a victim services counselor, Holt had just sat down to a late dinner after putting Jade and her friend, a neighbor, to bed.
The detective's news stunned Holt. "How did she die?" he asked. "What would he tell Jade?" he recalls thinking. Holt says he doesn't remember much about the conversation, but is certain that something he said must've been taken the wrong way. "I asked ... what happened. At first I was thinking car accident, because they didn't say where and they didn't say how. So, in my mind I'm thinking accident – did something happen at Spansion? The last thing I'm thinking is murder," Holt recalled in a lengthy interview with the Chronicle.
According to Baird and Holt, whether it was something he said – or didn't say – during that late-night encounter, or merely the fact that, as Sylvia's husband, he would be a default suspect, it didn't take long for Jennings to focus on Holt as the only suspect. In less than two weeks – after numerous interviews and a marathon day of interrogation – Holt confessed to the crime, was charged with first-degree murder, and was jailed on a $1 million bond.
On its face, it was an open-and-shut case: Estranged husband snaps and murders a wife he can't imagine living without – at least that's how it played in January 2011 on a wildly speculative episode of the inflammatory TV crime show, Nancy Grace.
In retrospect, the picture is much more complicated. By the time Jennings was finished interrogating him, Holt had confessed to murdering his wife multiple times in multiple ways – none of which matched physical evidence from the crime scene. Baird and Holt both say that Holt was pressured into confessing by an investigator possessed by a hypothesis, but without evidence, and who abandoned other, promising leads and manipulated the interrogation process, coercing Holt into a narrative that ended with a murder charge. In short, Baird says, this is a classic false confession scenario – a problem that has plagued the Austin Police Department over the years, notably in two other notorious and high-profile murder cases. But this isn't just a local problem; nationally, false confessions have been implicated in 25% of all DNA exonerations.
Following his confession, Holt was duly charged with murdering Sylvia – but the charge was ultimately dismissed in 2013, "pending further investigation," after state-requested DNA testing revealed an unknown male profile on the green panties police had identified as the likely murder weapon. More than three years later, Sylvia's murder remains unsolved, and Holt struggles to put his life back together.
Jonathan Holt met Sylvia Reyes in 1996 through a mutual friend, Abel Castillo. Everybody liked Sylvia, Castillo says; she was sociable and vivacious, always ready for fun. The attraction between Holt and Sylvia was intense, and they married in 1998.
But while they got along well as friends and lovers, under the same roof, they soon found differences. The couple started therapy. "It just didn't work," Holt recalls, because of "her not willing to change some things, me not willing to change some things – and the counselor not being really helpful. And we both agreed on the counselor part." They decided to part ways, and in 2001 were finalizing their divorce when Sylvia learned she was pregnant. Because of the pregnancy, they decided to try again. By 2005, they were living in separate areas of their large home, having agreed to do things as a family to support Jade, but otherwise to have separate lives. That worked for a while – and then didn't, and ultimately Sylvia decided to move out, Holt says; the two agreed it was time to divorce, and they readily worked out a child-custody agreement.
As was their pattern, says Holt, as soon as he and Sylvia parted company, their relationship began to improve – and their mutual attraction reignited. Holt regularly visited Sylvia's home to help with yard work, to pick up and drop off Jade, and, on occasion, for sex. The couple's on-again, off-again intimacy lasted until roughly a month before her death. "It was just that kind of relationship," Holt says. It finally ended, he recalls, because that's how Sylvia wanted it. During a phone call, he says he got the feeling that Sylvia was dating someone who didn't like that she and Holt still saw each other regularly. Holt was also in a relationship, and so Sylvia's decision to end things for good was fine by him.
So it went, says Holt, throughout December 2010 – including the week that Sylvia was murdered. Sylvia met Holt that Tuesday evening with Jade. Sylvia had to work Wednesday and Thursday and the plan, to the best of Holt's recollection, was that the two would meet again on Friday, so Jade and Sylvia could travel to San Antonio for Christmas.
Holt says he spent Wednesday working and taking care of Jade and her friend, Castillo's daughter. Early that evening, while the girls were watching a movie, Holt got on his computer to do some Christmas shopping for Jade, but decided instead to drive to a local retail store called Science Stuff. He called Castillo, who lived with his wife just up the street, and asked if they could watch the girls for a while. Holt drove north and then up and down Guadalupe, but couldn't find the store, and eventually gave up and headed back home. He called Castillo, parked his car, and walked over to Castillo's to collect the girls. At home, Holt returned to the computer and purchased several gifts for Jade. Eventually, they all went to bed.
It was Holt's last good night of sleep.
The night that Detective Jennings came to inform him that Sylvia was dead, Holt couldn't sleep. The interaction with the detective had made him uncomfortable; Jennings had attacked him from the get-go, Holt says, peppering him with questions about where he'd been Wednesday evening, while Holt was preoccupied trying to process what was happening.
Holt met with Jennings the next day, Christmas Eve, to provide a statement. For nearly four hours, Holt answered each of Jennings' increasingly pointed questions about his relationship with Sylvia and his whereabouts the night she was killed. He offered everything he could to help Jennings, including his cell phone, a sample of his DNA, consent to search Sylvia's Suzuki (at the time still a joint possession). Jennings moved into intimate questions about Holt and Sylvia's relationship, and eventually accused Holt of lying for not revealing details of their sex life. From there, according to a transcript of the interview provided by Holt, Jennings' attitude turned confrontational.
Holt readily complied with a request to remove his shirt for photographs (there were no scratches or other marks on his body or hands), offered to ride with Jennings along the route he drove on his search for Science Stuff, and agreed to return for a polygraph exam. "I mean, I'll give you whatever you need," Holt told Jennings. "I've got nothing to hide."
When Holt returned for the polygraph, he still had no idea what had happened to Sylvia. Holt and Castillo had spent the two days before Sylvia was buried cleaning the rented home on Birmingham Drive. A police crime-scene warning was still taped across the sliding glass door and fingerprint dust remained on the handle. There was a piece of carpet missing from Sylvia's bedroom floor. Otherwise, the place looked fairly normal – cluttered, but without obvious signs of violence.
On Jan. 4, 2011, Holt sat down for the polygraph with APD's civilian employee, Ray Montague. Because the case has not been closed, the Chronicle was not able to review the interrogation tapes, but in detailed interviews, Holt and Baird described what happened that day.
For nearly two hours the questioning was fairly routine, Baird said, until Montague directed Holt's attention to a list of six possible causes of death written on a paper tacked up on the wall of the small interview room – "stabbed," "poisoned," "hit in head," "strangled," "shot," "electrocuted," the list read. In a series of questions, Montague asked Holt whether Sylvia had been murdered in each of the six manners of death; Holt was instructed to say "no" each time, Baird and Holt recall. Holt did as he was told.
Montague informed Holt that the test indicated he was not being truthful, and that Holt, in his mind, correctly identified the manner of death. Holt was stunned, he says – "She was poisoned, right?" he asked Montague. Montague avoided the question, and instead offered his opinion: Holt had gone to talk to Sylvia at her home the night she was murdered and things got out of control. Holt repeatedly insisted that he was not there and that he did not kill Sylvia, he and Baird say.
After Montague questioned Holt for several minutes, Jennings entered the room. The test results speak for themselves, Jennings told Holt – that Holt was lying, not that he was nervous, and that he needed to explain his lies. Jennings told Holt that the only way to right a wrong would be to come clean. No matter what he said, Jennings would not listen, Holt recalls.
For whatever reason, Holt decided to confess – to make up a story that Jennings would determine was a lie, and that would make Jennings realize Holt was innocent. "I was thinking ... if I tell him a story he's going to know it's not the right story – because I don't know the right story," Holt recalls. Holt told Jennings that he went to Sylvia's house that evening to talk to her. As she sat on the couch watching TV, he walked behind her into the kitchen and, after getting himself a glass of ice water, grabbed two fresh needles, filled them with Sylvia's insulin, and stabbed her in the neck, sending her into a diabetic coma. He then left, he said.
There was no evidence that Sylvia had been injected with anything, and it was clear that the confession was untrue, Baird said. But instead of stopping there, assessing the actual evidence, and determining why Holt would offer a demonstrably false tale, Montague and Jennings pressed on – and in the process provided Holt a series of details about the killing. Montague suggested that perhaps Holt had actually strangled her. Holt initially said no, but then, in a second confession told police he'd injected her with insulin and then strangled her, leaving her slumped on the couch.
That was also inaccurate, but the police pressed further: "What if I told you we didn't find her on the couch?" Jennings asked Holt, Baird said. Holt told Jennings that he knew Sylvia was on the floor, because that had been on the news. "So, maybe she rolled off the couch," Holt told the detective. "You know where she was," Jennings replied. The detective pushed on: "How do you strangle someone?" he asked Holt. "With your hands," Holt recalls saying. Jennings said that's not what he thinks of when he thinks of strangulation, and asked Holt to name other ways that someone could be strangled, Holt recalls. If Holt would just show compassion for his wife by telling the police what happened, Jennings told Holt, that would help Holt later.
Eventually Holt asked to speak to Jennings privately and the two left the interview room. Tell me what you want me to say, Holt implored Jennings, and I'll say it. Jennings rebuffed Holt's request and eventually Holt told the detective that he was leaving; he'd told the truth and now he was ready to go home. Since Holt was not officially in custody, Jennings let him leave.
By the time Holt got home, the place was surrounded by police, who were in the process of getting a search warrant. He was not allowed to go inside, Holt said. Jennings arrived on the scene and Holt got into the car with the detective and continued talking to him for several more hours, Baird said. During that conversation, Holt again changed his story: With prompting from Jennings, he finally said that he had simply strangled Sylvia. Then he said that he went over there, saw Sylvia through the sliding glass door window with another man. In yet another version, Holt said he went to Sylvia's home before she got home from work and laid in wait for her, strangling her in the hallway after she came out of the bathroom.
More than 10 hours after he initially went in for the polygraph exam, Holt was arrested and charged with murder.
Holt still doesn't understand why the police would think he'd killed Sylvia. "It didn't matter what I said, what kind of story I gave. It always came back to: 'I know it was you,'" Holt recalls of his interrogation. Holt said he really doesn't know why he kept talking – and why he didn't request a lawyer – except that he was overwhelmed by everything that was happening, that he was trying to be helpful, and that he felt trapped by Jennings' insistence that he was responsible for Sylvia's death. "They basically gave me a path to go down – I just didn't know they were giving me the path. ... So here I'm thinking, 'If I give him a story, he's going to know I'm not telling the truth,'" he continued. "I just kept talking, and I don't know why I kept talking."
In fact, none of Holt's stories matched the physical evidence connected to Sylvia's murder. She was not injected with insulin. She was not strangled manually, nor with any of the items Holt eventually told Jennings could be used – in one confession, Holt says he used rope, and although police note this in an affidavit for Holt's arrest, they never did find any that matched Sylvia's wounds. In their arrest affidavit, police also alleged that Holt was "completely unable to verify his alibi as to his whereabouts" the night Sylvia was murdered, and that he had provided to Jennings "details about the crime scene that have not been released to the public."
According to Baird, those statements are false. Jennings never checked Holt's alibi; although Jennings initially said he wanted to ride with Holt on the route he said he drove the night of Sylvia's murder, he never did so. Baird said he investigated Holt's story – checking the route Holt said he took, discovering that the website for Science Stuff at that time did list an outdated address, and that Holt that evening searched for and purchased gifts online for Jade. It appears that no one bothered to look into the details of Holt's story after Jennings received basic cell phone data reflecting that on the night of Sylvia's murder, at around 10:30pm, Holt's cell phone used a cell tower near her home – although the tower in question is roughly equidistant between both Sylvia and Holt's home. And there is no indication that Jennings or anyone else sought additional data that would better pinpoint Holt's location that evening, Baird said.
Also inaccurate is the assertion that Holt revealed to Jennings details of the crime only a suspect would know, Baird said. Jennings asserted that no one but the police knew that Sylvia was found on the floor, or that the back door was unlocked. But both of those details were contained in a TV news report that aired just days after she was found.
Perhaps more important, says Baird, is the evidence that police appear to have ignored. For starters, there was the car. The night that Sylvia's body was found, two different neighbors reported to police that Sylvia's Suzuki was missing that morning, from just after 8am until about noon. Moreover, Baird said, in searching Sylvia's Suzuki the day after her body was found, police found no fingerprints at all, anywhere in or on the vehicle. Baird said neither of those pieces of information was ever sufficiently pursued. Neither was a tip from neighbors who said that there was an unfamiliar SUV parked on the street the night Sylvia was killed.
Also important, Baird believes, is information about Sylvia's likely use of Craigslist personals, which was brought to Jennings' attention by several sources, but again, never pursued. Not only did Holt and a friend of Sylvia's draw Jennings' attention to her possible use of Craigslist personals, but the detective who helped to process the crime scene also noticed that the site was included among a file of favorites cataloged on Sylvia's computer. Although police seized Sylvia's computer – as well as computers belonging to Holt and to Jade – there is no indication that forensic work was ever completed on any of the seized machines, Baird said. Those computers remain in APD custody.
It appears that no one at APD ever looked beyond the presumption that the husband did it, Baird says. "And ... I bet you that 80 percent of the time, it is the husband. And I think that they thought, this is an 80 percent chance, if we can get this guy to confess real quick we don't have to worry about the legwork or anything else," said Baird. "I just think they got tunnel vision [and] totally focused on him. ... They never did any good, solid police work to follow up on those leads – which certainly would have been leads if Jonathan had said, 'I've got an iron-clad alibi: I was in California when this happened.'" If that had been the case, perhaps they would've found Sylvia's murderer, Baird says. Instead, he says, they extracted from Holt a false confession, and simply wrapped up the investigation without accounting for any of the contradictory physical evidence.
It wasn't until two-and-a-half years later – after DNA testing revealed an unknown male profile on the stretched-out panties police collected near Sylvia's body, and an unknown profile of a female found on a water bottle next to Sylvia's bed – that prosecutors finally dropped the charges against Holt. But Baird said he doesn't believe APD is actually looking for the real killer. Instead, he says, prosecutors have indicated that, DNA evidence notwithstanding, they still believe Holt is guilty – they just can't prove it.
The problem with false confessions, experts say, is that many people – even people within the criminal justice system – find it hard to comprehend that a person would actually admit to something that he didn't do. But it happens, and not rarely. According to the Innocence Project, 25% of wrongful convictions overturned by DNA evidence involve a false confession. And many of those confessions actually contain details that match the crime – details that were not in the public domain, says Saul Kassin, a social psychologist and expert in false confessions who is a professor at John Jay College of Criminal Justice in New York. The problem is that all of those details are already known to police, and through the process of interrogation those details can, and do, shape a person's confession, with police – purposely or inadvertently – divulging details of the crime.
Once a confession is made, it's incredibly powerful, say Kassin and others who study false confessions and the tactics that lead to them; jurors often consider a confession infallible evidence, he explains, even though research clearly demonstrates that is not always the case. Among the most vulnerable suspects are those actually innocent, Kassin points out – in many ways, innocence is their enemy. "When I ask innocent people, 'First of all, why didn't you lawyer up?' they tilt their head and look at me like I'm crazy ... and say, 'Well, I didn't need a lawyer, I didn't do anything wrong,'" he says. "They volunteer to take lie detector tests, they volunteer their keys, their shoes, their cars, and they just don't apprehend that there is any risk. And their explanation is as simple as, 'I didn't do anything wrong, I have nothing to do with it, I have nothing to fear, I have nothing to hide – in fact, the more they get to know me, the more they'll see that I didn't do it,'" he continues. "People believe in the transparency of their innocence."
The problem can be compounded by the fact that police generally believe they have a good sense for determining which of the people they encounter during an investigation should be elevated to suspect status. In fact, according to research done by both Kassin and by Iowa State University psychology professor Christian Meissner and others, police are actually no better at determining who is lying to them than are untrained laypeople. Indeed, some of the signs police are trained to believe indicate deception, and techniques they're taught to use in interrogations, may negatively impact their ability to determine lies from truth.
The dominant method of interrogation used in the U.S. is known as the Reid Technique, named after its originator, former Chicago cop John Reid. It's an adversarial and persuasive technique designed to yield compliance from suspects – police are taught to assess in a first interview whether a person should be considered a suspect and then, once that determination is made, to not allow the suspect any ability to deny their involvement in the crime. "It is an excellent ... psychological approach to getting confessions from criminals," Kassin says. "I think the problem begins with this. I always say this: If every person interrogated was the criminal, the Reid Technique would be perfect. The problem is [police] often bring innocent people into the interrogation room." (In fact, during one early use of the technique by Reid himself, a false confession was extracted from a man whose wife had been murdered; it took more than five decades for the man to be exonerated.)
Police have long been taught, for example, that nonverbal cues, like hair touching, leg bouncing, and shifting eye movements, indicate deception. The research says otherwise. "These things ... are in the folklore of police training," says Meissner, but "there is plenty of research to show that they're not reliable." Also problematic are techniques police are allowed to use inside the interrogation room – including techniques that "manipulate the perception of consequences," says Melissa Russano, a professor of criminal justice at Roger Williams University. Specifically including lying to a person about evidence – saying that a fingerprint or DNA ties a person to a crime, when in fact no such evidence exists – or implying leniency, "stressing the importance of cooperation [with police] ... [suggesting] if you cooperate, things will be better for you and if you don't, things are going to get worse," Russano details.
Taken together, investigation detail leakage, faulty assumptions about who is lying, and heavy-handed, perception-altering interrogation tactics increase the likelihood of a false confession.
According to Russano, current research reflects that the heavy-handed tactics traditionally taught to police investigators are less effective and riskier than more information-based approaches, in which detectives use investigative interviewing and rapport-building techniques – open-ended questioning, timeline-building, and fact-checking. "Really, the science is showing that it's about how people tell their stories and how they remember events" that can reveal whether a person is lying, says Meisser. "It's about the cognitive properties of how they tell their stories that tend to be very important." This less adversarial approach is widely used elsewhere in the world, but U.S. police retain their decades-old reliance on the abrasive, unscientific, old-school approach.
Regardless of the technique, however, each confession should be tested for reliability by seeking corroborating evidence. And the very best form of corroboration, says Kassin, is when a suspect provides a detail or fact that police did not already know. "So, if he leads the police to the body, or to the murder weapon, or to whatever was stolen, that is gold-standard, corroborated evidence."
Although Austin Police declined to discuss Holt's case, homicide Sgt. Brian Miller says that his unit approaches each investigation carefully, with attention to every detail: "There's really no rush to a homicide investigation, so the big thing to do is to gather all sources of information that you have" before determining which way to move and which suspect to pursue, he explains. "That's sort of a supervisor's job, to slow everybody down, make sure everything gets covered." And Miller's investigators – many of whom have been in the squad far longer than the two years he's been there – are seasoned, he says, so much of that work happens organically. "People up here know quite a bit about working [a homicide investigation] so it kind of just happens."
Miller says he and his team are aware that false confessions occur, and he believes that his investigators take pains to avoid problems in the interrogation room. In the time he's been there, he says, he's never seen a "confrontational" interview, and he's not aware of a detective ever lying about evidence to a suspect. In fact, he says his investigators don't often take confessions and that they rely much more heavily on tried-and-true gum-shoeing to solve crimes. And when there is a confession that needs to be taken, he says it is monitored by a supervisor or another investigator, it is recorded, and it needs to be corroborated – he says he's never seen a confession that lacked corroboration. "If you can corroborate it, that's good."
While Miller says his detectives don't use heavy-handed tactics, it seems clear the department still relies heavily on the Reid Technique and on other aspects that confession experts find troubling. In a March presentation to the city's Public Safety Commission, APD Detective Trent Watts, who teaches interrogation techniques, told commissioners that the department uses Reid, but that it's not the "Bible to us," and noted also that he teaches investigators to pay attention to "subconscious communication" cues, like eye movements and body language – the very discredited tools that experts warn against. "There's science behind that," Watts said, but it shouldn't be relied upon. Whether the department also teaches evidence-based practices that reduce the incidence of false confessions is unclear; repeated requests to interview Watts for this story went unanswered by the department.
Unfortunately, says Baird, all the worst aspects of police interrogation collided in Holt's case – and he notes that it isn't the first time the APD has secured a demonstrably false confession from a defendant they thought was certainly guilty, and in some cases still insist is guilty. APD homicide detectives secured a detailed confession from Christopher Ochoa about how he and his friend Richard Danziger raped and murdered 20-year-old Nancy DePriest inside an Austin Pizza Hut restaurant in 1988. The pair spent 12 years in prison – during which time Danziger was beaten so badly by another inmate that he sustained permanent brain damage – before they were finally exonerated in 2002, some six years after the real murderer, Achim Marino, first contacted Austin authorities to say they'd convicted the wrong men.
And then there are the notorious and as yet unsolved yogurt shop murders, which happened nearly eight years before Austin detectives, in 1999, collected lengthy, detailed confessions from two men, Robert Springsteen and Michael Scott, each of whom, after hours of Reid-style interrogation, confessed to the crime. Unfortunately, the details each man offered were uncorroborated beyond what police already knew – or thought they knew – about the crime. It wasn't until 2008 that DNA from two unknown males was found – DNA evidence that directly contradicts Springsteen and Scott's confessions.
Although that evidence was enough to see Springsteen and Scott released from prison, it apparently isn't enough to convince either District Attorney Rosemary Lehmberg or the APD that the confessions were false. Law enforcement officials have suggested, in essence, that the DNA will eventually be explained, and that the confessions are the better evidence. The fact that the D.A.'s office, in particular, hasn't embraced the science in either the yogurt shop or Holt cases – either the DNA science that has proven confessions wrong in both instances, or the science that has demonstrated there are best practices for interrogation – frustrates Baird, who believes Lehmberg should be working with APD to ensure that their investigators receive the best training possible.
"I think they have the responsibility. I think the D.A.'s office, in theory, is supposed to be kind of like the adult in the room. I can see the police running in and saying 'We've got the case on this guy,' and the D.A. is supposed to take a step back and review it. APD should look at it, but if their protocol doesn't catch it, the D.A.'s office surely should. Because they have had past problems with false confessions," he said. "The problem with confessions generally, and false confessions in particular, is that once they have that statement they stop doing the legwork."
That's what appears to have happened in the case of Jonathan Holt. Sylvia Holt's murder remains unsolved, though leads still exist that might prove fruitful – including a possible forensic examination of the computers APD has in evidence. Baird says he requested funding to have the computers examined, but that request was denied by the court, and as far as Baird can tell, no meaningful examination of that evidence has ever been conducted. Unless the police get a hit on the unknown DNA, he says he fears Sylvia's murder will remain unsolved.
Meanwhile, Holt has struggled to put his life back together. While the charges have been dismissed, he remains under the thumb of the D.A. – not charged, but not exonerated. And until he can put the case to rest, he fears he will remain estranged from Sylvia's family and, most importantly, from Jade, who now lives with Sylvia's sister in San Antonio. He hasn't seen her since just after her mother's murder. He wishes the police would have acted more deliberately, seeking DNA testing first, before throwing his and his daughter's life into further chaos. "I mean ... [Jennings] could have shut this whole thing down" with that testing, he said. "It would've been over. But he didn't do that. It would've saved a lot of dollars, a lot of time, and a lot of heartache."
Copyright © 2015 Austin Chronicle Corporation. All rights reserved.