Essay on Science:

Science can have a powerful hold on our imagination and fill us with wonder at the ways of
nature.

We
are all storytellers; we make sense out of the world by telling stories. And
science is a great source of stories. Not so, you might argue. Science is an
objective collection and interpretation of data. I completely agree. At the
level of the study of purely physical phenomena, science is the only reliable
method for establishing the facts of the world.
But
when we use data of the physical world to explain phenomena that cannot be
reduced to physical facts, or when we extend incomplete data to draw general
conclusions, we are telling stories. Knowing the atomic weight of carbon and
oxygen cannot tell us what life is. There are no naked facts that completely
explain why animals sacrifice themselves for the good of their kin, why we fall
in love, the meaning and purpose of existence, or why we kill each other.
Science
is not at fault. On the contrary, science can save us from false stories. It is
an irreplaceable means of understanding our world. But despite the verities of
science, many of our most important questions compel us to tell stories that
venture beyond the facts. For all of the sophisticated methodologies in
science, we have not moved beyond the story as the primary way that we make
sense of our lives.
T
o see where science
and story meet, let’s take a look at how story is created in the brain. Let’s
begin with an utterly simple example of a story, offered by E. M. Forster in
his classic book on writing, Aspects of the Novel: “The king died and
then the queen died.” It is nearly impossible to read this juxtaposition of
events without wondering why the queen died. Even with a minimum of
description, the construction of the sentence makes us guess at a pattern. Why
would the author mention both events in the same sentence if he didn’t mean to
imply a causal relationship?
Once
a relationship has been suggested, we feel obliged to come up with an
explanation. This makes us turn to what we know, to our storehouse of facts. It
is general knowledge that a spouse can die of grief. Did the queen then die of
heartbreak? This possibility draws on the science of human behavior, which
competes with other, more traditional narratives. A high school student who has
been studying Hamlet, for instance, might read the story as a micro
synopsis of the play.
The
pleasurable feeling that our explanation is the right one—ranging from a modest
sense of familiarity to the powerful and sublime “a-ha!”—is meted out by the
same reward system in the brain integral to drug, alcohol, and gambling
addictions. The reward system extends from the limbic area of the brain, vital
to the expression of emotion, to the prefrontal cortex, critical to executive
thought. Though still imperfectly understood, it is generally thought that the
reward system plays a central role in the promotion and reinforcement of
learning. Key to the system, and found primarily within its brain cells, is
dopamine, a neurotransmitter that carries and modulates signals among brain
cells. Studies consistently show that feeling rewarded is accompanied by a rise
in dopamine levels.
This
reward system was first noted in the 1950s by two McGill University
researchers, James Olds and Peter Milner. Stimulating electrodes were placed in
presumed brain reward areas of rats. When allowed full unrestricted access to a
lever that, when depressed, would cause the electrodes to fire, the rats
quickly learned to repeatedly depress the lever, often to the exclusion of food
and water. Realizing that our brains are capable of producing feelings so
intense that we choose to ignore such basic drives as hunger and thirst was a
first step toward understanding the enormous power of the brain’s reward
circuitry.
Critical
to understanding how stories spark the brain’s reward system is the theory
known as pattern recognition—the brain’s way of piecing together a number of
separate components of an image into a coherent picture. The first time you see
a lion, for instance, you have to figure out what you’re seeing. At least 30
separate areas of the brain’s visual cortex pitch in, each processing an aspect
of the overall image—from the detection of motion and edges, to the register of
color and facial features. Collectively they form an overall image of a lion.
Each
subsequent exposure to a lion enhances your neural circuitry; the connections
among processing regions become more robust and efficient. (This theory, based
on the research of Canadian psychologist Donald O. Hebb (1904-1985) a pioneer
in studying how people learn, is often stated as “cells that fire together wire
together.”) Soon, less input is necessary to recognize the lion. A fleeting
glimpse of a partial picture is sufficient for recognition, which occurs via
positive feedback from your reward system. Yes, you are assured by your brain,
that is a lion.
An
efficient pattern recognition of a lion makes perfect evolutionary sense. If
you see a large feline shape moving in some nearby brush, it is unwise to wait
until you see the yellows of the lion’s eyes before starting to run up the
nearest tree. You need a brain that quickly detects entire shapes from
fragments of the total picture and provides you with a powerful sense of the
accuracy of this recognition.
One
need only think of the recognition of a new pattern that is so profound that it
triggers an involuntary “a-ha!” to understand the degree of pleasure that can
be associated with learning. It’s no wonder that once a particular
pattern-recognition-reward relationship is well grooved into our circuitry, it
is hard to shake. In general—outside of addiction, that is—this
“stickiness” of a correlation is a good thing. It is through
repetition and the sense of familiarity and “rightness” of a
correlation that we learn to navigate our way in the world.

S
cience is in the business of making up stories called hypotheses and
testing them, then trying its best to make up better ones. Thought-experiments
can be compared to storytelling exercises using well-known characters. What
would Sherlock Holmes do if he found a body suspended in a tree with a note
strapped to its ankle? What would a light ray being bounced between two mirrors
look like to an observer sitting on a train? What do the lines on the nautilus
shell mean? Once done with their story, scientists go to the lab to test it;
writers call editors to see if they will buy it.
People
and science are like bread and butter. We are hard-wired to need stories;
science has storytelling buried deep in its nature. But there is also a
problem. We can get our dopamine reward, and walk away with a story in hand,
before science has finished testing it. This problem is exacerbated by the fact
that the brain, hungry for its pattern-matching dopamine reward, overlooks
contradictory or conflicting information whenever possible. A fundamental
prerequisite for pattern recognition is the ability to quickly distinguish
between similar but not identical inputs. Not being able to pigeonhole an event
or idea makes it much more difficult for the brain to label and store it as a discrete
memory. Neat and tidy promotes learning; loose ends lead to the “yes, but” of
indecision and inability to draw a precise conclusion.
Just
as proper pattern recognition results in the reward of an increased release of
dopamine, faulty pattern recognition is associated with decreased dopamine
release. In monkeys, the failure to make a successful prediction (correlation
between expected and actual outcome) characteristically diminishes dopamine
release exactly at the time that the predicted event is anticipated but fails
to occur. Just as accurate correlations are pleasurable, lack of correlation
produces the neurotransmitter equivalent of thwarted expectation (or worse).
Once
we see that stories are the narrative equivalent of correlation, it is easy to
understand why our brains seek out stories (patterns) whenever and wherever
possible. You may have read or heard about the famous experiment in which
University of Illinois psychology professor Daniel Simons asked subjects to watch a video
and count the number of times a ball is dribbled by a basketball team. When
focused on counting, the majority of viewers failed to see a woman in a gorilla
suit walk across the playing area. In effect, well-oiled patterns of
observation encourage our brains to compose a story that we expect to hear.
Because
we are compelled to make stories, we are often compelled to take incomplete
stories and run with them. With a half-story from science in our minds, we earn
a dopamine “reward” every time it helps us understand something in our
world—even if that explanation is incomplete or wrong.
Following
the Newtown massacre, some experts commented on the killer having Asperger’s
syndrome, as though that might at least partially explain his behavior. Though
Asperger’s syndrome feels like a specific diagnosis, it is, by definition,
nothing more than a constellation of symptoms common to a group of people. In
the 1940s, Austrian pediatrician Hans Asperger noted that a number of patients
had similar problems with social skills, eccentric or repetitive actions,
unusual preoccupation rituals, and communication difficulties, including lack
of eye contact and trouble understanding facial expressions and gestures. The
recent decision by the American Psychiatric Association to remove the diagnosis
of Asperger’s syndrome from its guidebook for clinicians, the Diagnostic and
Statistical Manual of Psychiatric Disorders (DSM-V), for failing to conform to
any specific neuropathology, underscores the all-too-common problem of
accepting a clustering of symptoms as synonymous with a specific disease.
Syndromes are stories in search of underlying causes.
Similarly,
studies of psychopaths have shown a diminished volume of gray matter in
specific regions of the prefrontal cortex. But these findings aren’t the sole
explanation for violent acts. Because it is impossible to stimulate a specific
brain region to produce complex and premeditated acts, we are left to conclude
that while certain brain conditions can be correlated with a complex act, they
are not necessarily causing it. Likewise, brain scans that reveal abnormalities
in mass murderers may help us understand what might have contributed to their
behavior. But the abnormalities are no more the sole explanation for violence
than childhood neglect or poor nutrition are. They are stories, albeit with a
detailed neurophysiological component, but stories nonetheless.
W
hen we make
and take incomplete stories from science, there are often moral consequences.
How much personal responsibility should we assign to an individual with a
damaged or malfunctioning brain? What is the appropriate punishment and
possibility of rehabilitation for such a person? Only when we openly
acknowledge the degree to which science is presenting its observations in the
form of story can we address this moral dimension. We must each work out our
own guidelines for when we think scientific data has exceeded its bounds and
has morphed into the agenda and bias of story. Of course this is always going
to be a challenge in the absence of a full array of scientific data.
But
we can begin by being aware of the various ways that storytelling can insinuate
itself into the presentation and interpretation of data. Good science is a
combination of meticulously obtained and analyzed data, a restriction of the
conclusions to those interpretations that are explicitly reflected in the data,
and an honest and humble recognition of the limits of what this data can say
about the world.
As
members of the public, we need to ensure that any science we accept as truth
has passed through the peer-review process. We should also understand that even
peer-reviewed data is not always accurate. In 2011, Nature reported that
published retractions had increased by a factor of 10 over the last 10 years,
while the number of papers published rose only 44 percent. Also in Nature,
scientists C. Glenn Begley and Lee M. Ellis wrote last year that their
colleagues at the biotechnology firm Amgen could reproduce only six of 53 landmark
hematology and oncology studies from the scientific
literature. Similarly, scientists from Bayer reported in 2011 that they could not consistently
reproduce about two-thirds of oncology studies
relevant to their work.
When
reading science reports, we should also search for information on the limits of
the data. Were assumptions made? What do the “error bars,” or graphic
representations of variable data, say? We may not always understand the data
limits, but we should be worried when some discussion of them is completely
absent.
In
the end, scientists have the tools, language, and experience to tell us
informed, engaging, and powerful stories. In turn, we should judge their
studies in the same light in which we judge other artistic forms. Like a
literary critic, we should assess the preciseness of language, the tightness of
structure, the clarity and originality of vision, the overall elegance and
grace of the study, the restraint with which they present moral issues, how
they place their studies in historical, cultural, and personal context, and
their willingness to entertain alternative opinions and interpretations.
The
methodology of science remains one of the great advances of humankind. Its
stories, properly told, are epic poems in progress, and deserve to stand
alongside the great stories of history.

AI:

For all its speed, precision and computational power, AI remains incapable of the one
thing that makes moral reasoning truly possible: the ability to question not only what is
right, but why.