Summer 2016 Bulletin

What Evidence Should We Trust?

Jerome Kagan

When forced to decide between a career in biochemistry or psychology in the spring of 1950, I chose the latter because of a gnawing puzzlement provoked by the observation that apparently sane people living in the same community held different beliefs about love, honesty, and whom was entitled to respect and whom to scorn. It would take another twenty-five years before I appreciated that the answer to my adolescent confusion was that people trusted different sources of evidence when establishing their beliefs. Some trusted their feelings; others their observations; and some relied on the statements of respected authorities. Millions of Americans of my generation who reached adulthood during the first half of the last century relied on a feeling when they decided that Freud’s explanation of the causes of their doubt, anxiety, and guilt were probably right. A majority of equally anxious Americans, reflecting on different evidence in 2016, find the same explanations deeply flawed.

Before the invention of machines that allowed humans to observe phenomena hidden to their senses, everyone understood that A’s statements about B’s fears, honesty, prejudices, or talents were based on observations of B or descriptions of B by those who knew him. Following the invention of new forms of evidence for human properties – questionnaires, the speed of a motor response, changes in the amount of sweat on a finger, and brain activity – many investigators borrowed the same words to describe different evidence. Adolescents who report on a questionnaire that they are among the most popular youths in their class are often observed to be consistently rejected by their classmates. A fair number of neuroscientists believe that a particular pattern of blood flow in the brain of a person looking at a face displaying the wide eyes and open mouth characteristic of fear means that the individual is in a state of fear, even if the person vehemently denies any semblance of fear or anxiety. The neuroscientists implied that the meaning of anxiety had not changed despite the change in the evidence.

Contemporary populations have become more dependent on the claims of scientists than earlier generations. Unfortunately, the average citizen is unable to judge the validity of the evidence behind most declarations. Most adults who learn that the universe is 13.8 billion years old assume that this statement must be based on impeccably trustworthy evidence. A majority do not appreciate that the temperature of the universe the moment after the Big Bang, which is necessarily an estimate that is subject to revision, is one of the critical pieces of evidence. Lord Kelvin, the most respected physicist of the nineteenth century, intimidated Charles Darwin by declaring that his evidence on the cooling of the earth’s interior implied that the planet could not be older than 25 to 30 million years – too short a time for the changes that Darwin’s arguments required. Lord Kelvin’s evidence turned out to be flawed, as was the evidence for an aether as the medium through which light traveled.

Linda Bartoshuk provides a persuasive example of the dependent relation between the truth of a statement and its source of evidence.1 She was surprised to discover that adults who had many taste buds for sweetness on their tongue and those with few taste buds gave the same judgments when they evaluated the sensation of the sweetness of a sip of soda on a scale that went from “not sweet” to “very sweet.” But when she asked these adults to adjust a lever controlling the loudness of a sound so that the intensity of the sound corresponded to the sweetness of the soda, those with more taste buds for sweetness selected a loudness corresponding to a train whistle. Those with fewer taste buds chose a sound resembling a dial tone. Hence, the answer to the question, “Do individuals with more taste buds for sweetness experience a food as sweeter?” is “It depends on the evidence.” This principle, which the physicist Niels Bohr articulated a century ago, applies to every scientific statement.

New sources of evidence often alter earlier conclusions. The observations of the European voyagers of the fifteenth, sixteenth, and seventeenth centuries as well as the later discovery of fossils made evolutionary ideas possible. Declarations about the nature of anger, fear, sadness, and joy were changed after the invention of the camera, which allowed more detailed measurements of the muscle patterns in facial expressions. Genetic evidence, which is replacing fossil evidence, suggests that dogs evolved from wolves more than 120,000 years ago. It is possible that optogenetics, which allows more precise activation of neurons with light, will replace the measurement of blood flow to sites containing many thousands of neurons. Each new source of evidence is accompanied by the replacement of some old conclusions with new ones.

Contemporary American and European scientists, and a fair number of non-scientists, place greater trust in biological measures than behaviors when deciding on the truth of claims about psychological states. A woman who typically feels anxious when talking with strangers learns about a new drug that reduces anxiety. She assumes that patients’ reports saying that the drug muted the intensity of their anxious feelings comprised the evidence. She would be surprised to learn that a major basis for the claim was the drug’s ability to tempt mice, who naturally prefer to remain in a dark chamber, to explore a brightly lit chamber. The scientists who argued that a mouse’s avoidance of light was a sensitive index of anxiety assumed that because humans avoid experiences that make them anxious, it must be the case that a mouse who avoids a brightly lit place must be afraid of this experience. These scientists forgot, or chose to deny, the obvious fact that humans who avoid spicy foods, large cities, or country music are not afraid of these events. They simply do not like them.

Statements about human consciousness have, until recently, relied on a person’s behaviors or subjective reports. Consciousness assumed a new meaning when investigators discovered that neurons in a motor region of the brain of a woman who had been in an uncommunicative, vegetative state for more than five months became active when a scientist asked her to imagine she was playing tennis.2 Advances in genetics may alter the truth of statements about illness. Is a woman who feels perfectly healthy ill if she possesses genes that are known to place women at risk for breast cancer?

The current bias favoring the truth of statements about human traits that are based on biological evidence has serious implications when judges and juries are deciding on the guilt or innocence of a defendant accused of a violent crime. Respected experts have argued that the immaturity of the adolescent brain is a major reason for the high prevalence of homicides committed by American youths. These scientists often fail to acknowledge that poor black youths living in one of the Southern states committed the majority of violent crimes in the United States in 2014. There is no evidence to suggest that the brains of these adolescents were less mature than the brains of black adolescents from affluent families living in the Northeast. One measure of an immature brain is incomplete elimination of the synapses that connect neurons. This process, which is a normal phase in maturation, is accompanied by a thinner level of gray matter in the cortex. Measures of gray matter thickness in individuals aged 10 to 50 years revealed that some 10 year olds had thinner cortices than some 30 year olds and most youths between ages 12 and 18 had similar thickness levels.3 Equally important, men between ages 25 and 45, most of whom possess a mature brain, commit more rapes and murders than adolescents.

Predicates that are appropriate in sentences that have an animal or human as the noun are inappropriate when a brain profile is the noun. Sentences with the predicates see, feel, plan, or remember require an animal or person to be the agent. This issue was the theme of a debate between Maxwell Bennett, a neuroscientist, and Peter Hacker, a philosopher, who objected to using predicates that implied psychological processes in sentences in which the brain was the noun. Philosophers Daniel Dennett and John Searle saw nothing wrong with this practice.4 Dennett and Searle argued that it was legitimate to write, “The hippocampus remembers a name” because neuroscientists understood this sentence to mean that activity in the hippocampus contributes to and is necessary for remembering a name. Because neuroscientists agree on the special meaning of remember in the sentence above, it is not a serious error to attribute psychological properties to the brain. That conclusion is reasonable when the language community consists only of neuroscientists. It has the potential for misunderstanding when others read the same sentence. That is why neuroscientist Joseph Le Doux5 suggested that the predicate fear assumes a conscious human and should not be used to describe a brain profile.

Social scientists studying human traits, symptoms, or past experiences typically rely only on a person’s verbal reports and assume that the verbal statements are faithful proxies of the evidence that would have been found if these individuals had been observed. Sadly, this premise is flawed. An adult’s description of a psychological state, trait, past behavior, or experience often fails to correspond to what direct observations would have revealed. This claim applies to how well one slept, the quality of memory, a feeling of well-being, penis length, prior use of cocaine, or the stress of public speaking.6 For example, a mother’s description of her child as restless, fearful, or confident does not always match what observers record when they film the child in a laboratory or at home. Almost two thirds of New Zealand mothers whose four-year-old children had stayed in a hospital at least one night did not remember that hospitalization when asked about it one or two years later.7 If a parent cannot recall an event as emotionally salient as a child’s prior hospital admission, it is likely that many memories of 20 year olds recalling less salient childhood events are of questionable accuracy. Current explanations of relations between childhood experiences and later traits assume that the early events occurred, not that the individual thought they occurred.

The inability to judge the trustworthiness of most scientific claims allows the average citizen to dismiss statements he or she prefers to forget or deny. As a result, the sentiment of the majority exerts a palpable influence on legislation. The evidence implicating the dangers to human health posed by the molecule bisphenol A, which is used in the manufacture of plastics, is as firm, some would say firmer, as the evidence implying the dangers of passively inhaling cigarette smoke. But neither the public nor Congress is ready to outlaw the use of this molecule. Facts and the relevant evidence join reasoning, intuition, and community sentiment in determining the beliefs that a society is prepared to accept as a basis for action.

Jerome Kagan is Daniel and Amy Starch Professor of Psychology, Emeritus at Harvard University. He has been a member of the American Academy of Arts and Sciences since 1968.

© 2016 by Jerome Kagan



1. Linda Bartoshuk, “The Measurement of Pleasure and Pain,” Perspectives on Psychological Science 9 (2014): 91 – 93.

2. Adrian M. Owen et al., “Detecting Awareness in the Vegetative State,” Science 313 (2006): 1402.

3. Jerome Kagan and Norbert Herschkowitz, A Young Mind in a Growing Brain (Mahwah, N.J.: Erlbaum, 2005).

4. Maxwell Bennett, Daniel Dennett, Peter Hacker, and John Searle, Neuroscience and Philosophy: Brain, Mind & Language (New York: Columbia University Press, 2003).

5. Joseph E. Le Doux, “Coming to Terms with Fear,” Proceedings of the National Academy of Sciences 111 (2014): 2871 – 2878.

6. David Veale et al., “Beliefs about Penis Size,” Journal of Sexual Medicine 11 (2014): 84 – 92; David Torrents-Rodas et al., “No Effect of Trait Anxiety on Differential Fear Conditioning or Fear Generalization,” Biological Psychology 92 (2013): 185 – 190; Kate M. Scott, Don A. R. Smith, and Pete M. Ellis, “A Population Study of Childhood Maltreatment and Asthma Diagnosis,” Psychosomatic Medicine 74 (2012): 817 – 823; Glenn J. Landry, John R. Best, and Teresa Liu-Ambrose, “Measuring Sleep Quality in Older Adults,” Frontiers in Aging Neuroscience 7 (2015): 166, doi: 10.3389/fnagi.2015.00166.

7. Nataliia Burakevych, Christopher J. McKinlay, Jane M. Alsweiler, Jane E. Harding, and CHYLD Study Team, “Accuracy of Caregivers’ Recall of Hospital Admissions: Implications for Research,” Acta Paediatrica 104 (2015): 1199 – 1204; Sean P. Wojcik et al., “Conservatives Report, but Liberals Display, Greater Happiness,” Science 347 (2015): 1243 – 1246.