The high-profile trial of George Zimmerman, a Florida neighborhood watch volunteer accused of second-degree murder in the shooting death of teenager Trayvon Martin in February 2012, last week featured what be one of the most unsettling moments ever seen in a criminal case. First, Martin's mother, Sybrina Fulton, identified a voice crying out in a three-second segment of a 911 call as definitely that of her son Trayvon. Shortly afterward, she was followed on the stand by Gladys Zimmerman, who insisted, with just as much seeming conviction, that the voice was that of her son George, who has claimed that he was attacked by Martin.
While both women appeared sincere in identifying the voice as that of their sons, they can't both be right. That leads us to wonder: How adept are we humans at recognizing the voices of our offspring? And what might cause a parent to mistake another voice for his or her own progeny?
It might seem like a no-brainer that parents would be able to pick out the voices of their own flesh-and-blood. After all, studies show that a wide range of animal species, from penguins to rhesus monkeys, possess the ability to aurally identify their offspring.
An Australian study published in PLOS ONE in 2012, for example, found that sea lions begin to recognize the distinctive cries of their pups within 48 hours of birth, and at 72 hours will turn to respond to their own pups' vocalizations more quickly than they will to the cries of other pups.
British researchers reported in a 2012 article in Proceedings of the Royal Society B that pygmy goat mothers not only recognized the calls of their kids at five weeks, but were able to recognize recordings of them a year to 18 months after weaning. The scientists say that parental ability to recognize voices for such an extended time is useful, because, among other benefits, it helps keep mother goats from mating with their sexually mature sons.
You'd think that we would be pretty good at it, too. While humans may not be able to hear as wide a range of frequencies as some of these animals, we actually are more adept at detecting fine differences between frequencies, according to work by U.S. and Israeli researchers.
And like other primates, humans have complicated anatomical structures for making sounds, with plenty of individual variations -- such as the length and flexibility of the vocal tract -- that can make one person's voice sound different from another, according to a 2006 article by University of Glasgow psychology professor Pascal Benin.
We also have big brains that devote a lot of their processing power to speech. A 2001 study by Japanese researchers, published in the journal Neuropsychologia, found that familiar voices activate different areas of our brains -- in particular, the right frontal pole and right temporal pole -- than the sound of people we don't know.
Not surprisingly, research shows that human parents are pretty good at recognizing their infants' cries. A study published in Developmental Psychobiology in 1983 found that when subjects were asked to listen to recordings of four different month-old infants, 80 percent of mothers were able to recognize their own children, as were 45 percent of fathers.
Another 1994 study by the same researchers found that humans were still able to identify an infant's cries, even when the sound was altered by recording it from a distance outdoors.
But pretty good isn't perfect. University of Florida researcher Harry Hollien, in his 2001 textbook "Forensic Voice Identification," notes that in a 1970s study, he found examples of babies whose voices were hard to tell apart.
"When two babies sounded similar, their mothers not only correctly identified their own child, but also identified the other one as theirs," he writes.
And recognizing the voice of an adolescent or adult in a stressful situation is a trickier task.
Hirotaka Nakasone, an FBI voice analysis expert, testified for the prosecution that an extreme emotional state can significantly alter a person's pitch, increasing the vibration of the vocal cords from 120 to 220 hertz to 500, 600 or even 700 hertz.
Additionally, a person's resonance -- the ability to project, a key part of what makes one person's voice sound distinctive -- can also be affected, he testified. Nakasone testified that it was beyond the ability of present scientific methods to identify the voice.
George Papcun, a New Mexico-based audio forensics expert who has worked for Zimmerman's defense team, told Discovery News that it's even more difficult to identify the voice on the 911 recording because the sound is in the background of a cell phone call.
"A cellphone takes your voice and turns it into a bunch of numbers, then sends it up to a tower and then to another tower, and then it's reconstituted at the other end," Papcun said. "That removes some of the data -- all kinds of distortions are possible."
But given the difficulty of making sense of that brief, distorted sound, how is it that each of the two mothers seemed equally convinced that they could hear their own child's voice on the recording?
Alan J. Lipman, an attorney and psychologist who is an associate clinical professor of psychiatry and behavioral sciences at George Washington University Medical Center, believes that it may come down to confirmation bias -- that is, humans' tendency to interpret information in a way that bolsters our existing beliefs.
"Faced with an ambiguous stimulus, we tend to project what we would want to find," he told Discovery News. "We pick out the aural characteristics that we want to hear, and filter out the characteristics that would contradict our conclusion."