Multilevel Explanations
That emotions such as happiness and sadness are neural processes does not rule out other kinds of
mechanisms as also relevant to explaining events like becoming sad or falling in love. When hearing
some good news makes you happy, this change is the result of your brain's undergoing the neural
processes I have described, such as activation of your nucleus accumbens. But a full explanation of
your happiness can legitimately operate at four different levels: molecular, psychological, and social,
as well as neural.
The molecular level is important for understanding how neurons work. I have already mentioned
the neurotransmitter dopamine used by neurons that cause feelings of pleasure. When one of these
neurons fires, it does not send an electrical charge directly to the other neurons that it excites, but
instead releases molecules of dopamine that cross the synaptic gap between the end of the axon of one
neuron and the dendrites of the receiving neurons. Dozens of other neurotransmitters play similar
roles in making neural processes chemical as well as electrical. The most important include
glutamate, which contributes to excitation, and GABA, which contributes to inhibition. I described in
chapter 3 how such neurotransmitters can be manipulated by recreational and therapeutic drugs.
Besides neurotransmitters, other kinds of molecules are important for explaining emotional changes—
for example, cortisol produced by the adrenal glands in reaction to stressful situations, and hormones
such as estrogen and testosterone. Hence neural explanations of emotions like the ones provided by
the EMOCON model have a molecular underpinning.
But that does not mean that we can replace descriptions of neural mechanisms with descriptions of
molecular ones, for two reasons. First, the molecular mechanisms are far too complex to permit us to
describe completely how they make even a single neuron work. The internal operations of a neuron
are controlled by thousands of genes affecting the chemical interactions of thousands of proteins and
other molecules. Much is known about these molecular workings, but the complexity of interactions is
so enormous that science may never be able to give a full mechanistic account of the firing of
individual neurons, just as explanatory and predictive models of the weather may never be complete.
Fortunately, we do not have to wait for the full story about how cells such as neurons work, but can
build an approximate account based on the crucial properties of neurons: their ability to accumulate
and pass on electric charges.
Second, even if we could have a complete molecular account of how a single neuron works, we
would still need mechanisms that show how networks of billions of neurons interact to produce
complex effects such as emotional feelings. Currently, we can best approach the emotional
phenomena we want to explain by looking at interactions among entire brain regions, not at single
neurons or even populations of neurons. Even when much more comes to be known about the
operations of single neurons and neural populations, it will still be useful to consider how aggregates
of neural populations such as brain regions interact to produce psychological effects. Hence neural
and molecular mechanistic explanations complement rather than compete with each other.
Similarly, accepting the claim that emotions are brain processes does not eliminate the value of
psychological explanations. I don't mean the simple explanations of ordinary people who rely on folk
ideas about beliefs and desires, but rather the theoretical ideas of cognitive psychologists, who often
find it useful to talk about mental representations such as concepts, rules, images, and analogies.
Saying that concepts are patterns of neural activity enhances rather than eliminates the explanatory
value of such representations. When you become happy because of your winning a lottery, we cannot
probe into your brain to determine exactly what is happening to your neural populations, so the best
explanation available may rely on such descriptions as this: you are happy that you won the lottery
because you need the money.
Part of such psychological explanations requires noting the positive valuation you attach to such
concepts as winning and money. Even though these valuations have a neural basis, through
coordination between neural populations for verbal representations and ones for positive value, the
psychological level of explanation remains important because of our lack of knowledge about the
neural details and the direct relevance of concepts like money to our practical interests. Hence
psychological explanations of emotions can coexist with and complement neural ones, just as
molecular explanations can.
Social levels of explanation are also highly relevant to emotions. When you win the lottery, the
event is highly social, from the interaction with the ticket seller who can confirm your win to the joy
of telling your friends and family about your luck. Many of the goals that generate happiness, sadness,
and other emotions are inherently social, tied to your relationships with the people who are important
in your life. Your happiness about winning the lottery may derive from your considering not only what
the money can do for you, but also what it can do to further the goals of your family. Social emotions
such as guilt and pride can be understood only through consideration of your place in social networks
such as your family, friends, and people you work or play with. In chapter 7, I will discuss the
importance of emotions for love, work, and play, all of which are best understood in terms of
interactions among neural, molecular, psychological, and social levels of explanation.
Hence my defense of the claim that mental processes such as emotions are brain processes does not
diminish the relevance of social, psychological, and molecular explanations. We should be neither
reductionist, claiming that explanation ought to be at just one fundamental level, nor antireductionist,
claiming that levels of explanation are independent of each other. The best approach to explaining
mental events requires attention to multiple levels, from the social to the molecular, with a focus on
how they interact. Neural processes such as the ones that constitute emotions are clearly affected both
by underlying biochemical reactions involving neurotransmitters and by social relations among
people.
To take one vivid example, consider what happens when you have a friend who suddenly becomes
insulting and threatening. We can best understand your emotional reaction of fear and anger by
considering all of the following: (1) firing of neural populations in brain areas such as the amygdala;
(2) increased operation in your brain of such molecules as adrenaline and cortisol; (3) application of
concepts such as insult and danger; and (4) the social interaction with your friend that prompted the
emotional reaction in the first place. The full explanation of fear notes the relevance of molecular
changes to neural changes, and of psychological changes to social changes; but it also appreciates
how mechanisms are related in the other direction. For example, the social interaction of being
threatened causes the molecular change of increased cortisol. Levels of explanation are intertwined,
not simply reductive. But in rejecting a ruthless reductionism, we should not embrace a blind
antireductionism that ignores how social groups consist of persons, who consist of organs such as
brains, which consist of neurons, which consist of proteins and other molecules.