How Emotions Are Made: The Secret Life of the Brain

It’s sort of like sausage-making. You know what emotions are, but you’re not sure you want to know what goes in them. How Emotions Are Made: The Secret Life of the Brain takes you through the journey where emotions aren’t consistent across cultures – or even people. The journey, if you’re willing to believe it, flies in the face of the thoughts of dozens of researchers. I’m not convinced that Lisa Barrett has all the right answers with How Emotions Are Made – but at least there are some things to think about.

Ekman FACS

I’ve been a fan of Paul Ekman’s work for some time. (See Emotional Awareness, Telling Lies, and Nonverbal Messages: Cracking the Code.) Even the Pixar movie Inside Out is a brilliant story around emotions and how our rational and emotional selves co-inhabit the same body. Fundamental to Ekman’s work is the belief that microexpressions reveal what someone is thinking very quickly and briefly. The premise of his work is that the microexpressions response is a recognizable pattern that happens before your conscious mind has the ability to stop it.

From a neurological point of view, there are some reasons to support this thinking. There are separate sensory pathways that get differing levels of processing, some of which allow you to trigger fight or flight reactions very quickly. However, Barrett says that the research doesn’t support Ekman and his perspective that microexpressions are real and consistent.

The problem is that I’ve looked at the research – including the research cited by Barrett – and though it identifies a set of problems with the theory, including the problem of emotions not being distinct, the research is far from saying that the entire model is bad. Even research that was unable to replicate the findings of microexpressions universally across the face indicates that there is something happening.

Ekman himself says that you must exercise caution in that, even if you feel strongly that a microexpression indicates the presence of an emotion, you can’t determine why that emotion was triggered. There may be reasons why other experiments partially replicated the findings. It can be that microexpressions aren’t sufficiently distinct or that the experimental controls didn’t contain unexpected variables.

There are training programs that have been created from Ekman’s work on microexpressions – including his Facial Action Coding System (FACS) and training programs created for the Department of Homeland Security (DHS). The criticisms Barrett levels seem squarely pointed at the lack of efficacy of the training program for DHS. Having created training programs for many years, I know that there is a winding road between the lab and the lieutenant. Who knows whether the concept is bad, or the training wasn’t effective?

System 1 and System 2

Daniel Kahneman’s work Thinking, Fast and Slow is a powerful explanation of how we think. He uses an idea of System 1 and System 2: System 1 takes the most common, everyday tasks and refers to the more calorically more expensive System 2 when it can’t handle the job. Once something is habitualized, it ends up in System 1. It’s automatic. System 2 is deep and rational thought.

Kahneman is careful to say that he views these as mental models rather than specific indications of brain regions, but many have said that System 1 is the basal brain, including the amygdala, and System 2 is the prefrontal cortex. Barrett asserts that this localization isn’t true, that emotions are composed from all over the brain, and that the lack of an amygdala isn’t enough to suppress emotion.

Here, there are many pieces. Kahneman’s work is based off earlier work of researchers who have moved on to a three-part model instead of a two-part model. My friend Paul Culmsee finds this distinction particularly interesting while I do not. For me it’s just a refinement of a model.

What’s more interesting to me is the nuanced nature of our learning of the brain. Barrett speaks about how Broca’s area isn’t the only component of the brain necessary for speech. As luck would have it, I just recently finished The Tell-Tale Brain, which speaks extensively about how speech works from a neurological point of view. At one level, Barrett is correct that speech isn’t exclusively processed in Broca’s area. However, syntactic structure does appear to be centered in Broca’s area – and that is most commonly used in the creation of speech.

Because there is still so much that we don’t know about the brain, I think it would be premature to indicate that there isn’t a space where emotions are rooted or triggered.

Degeneracy

Many neurons can create the same outcomes. It’s a process that builds redundancy into our brains. For instance, there may be multiple pathways today that can trigger the same fear response. It’s the principle of degeneracy – multiple paths leading to the same result. This concept seems to be inefficient; however, the inefficiency results in resiliency, something that evolution may have needed more. Donella Meadows cautions in Thinking in Systems about over-optimization and the lack of resiliency. Nassim Taleb is slightly more direct in his criticisms of over-optimizing the system in The Black Swan and Antifragile.

The degeneracy and plasticity of our brains allow us to recover from seemingly unrecoverable brain damage. Neuroscience has found that brains will rewire themselves to support the needs of the individual, whether it’s the larger hippocampus for London cab drivers (who need more spatial memory), or it’s how the normal workload of a damaged portion of the brain is taken up by other parts of the brain.

Simulations

Barrett notes that one of the brain’s primary purposes – if not the primary purpose – is prediction. It makes models of the world around us and then uses those models to predict what will happen next. In the book Incognito, it is made clear that we’re not perceiving reality. We’re perceiving some made-up idea that our brains have concocted.

We use these concoctions as models and run simulations – what Barrett calls predictions. Gary Klein was clear in his belief that we develop these models unconsciously, and we use them to provide predictions. (See Sources of Power for more.) Ultimately, we adjust our predictions when we perceive that our predictions are incorrect.

The problem is that sometimes our ego prevents us from accepting the mismatch between reality and our predictions, because it’s unwilling to give up its grip on the perceptions that it holds. Sometimes we become blinded to the discrepancies in the world and our beliefs. This sometimes manifests in disorders like schizophrenia. In these situations, it’s difficult for the person to continue to realign themselves to reality and to prevent their perceptions from drifting too far. Somehow the mechanisms, like humor, aren’t sufficiently effective to hold perception and reality into relative alignment. (See Inside Jokes for more on humor as a correction mechanism for predictions.)

Interpreting Interoception

Most people can describe bodily sensations which match their moods. Before taking a test or performing on stage someone might discover they have “butterflies in their stomach.” This is perceived as an indicator of anxiety. If you ask someone how they felt the first time they were kissed or when preparing for a special date, they may report the same feelings. In fact, you’ll infer the emotion that I’m trying to convey with “butterflies in my stomach” based on the context. You’ll assume I’m excited or anxious depending upon the context. We have learned to listen to our bodies and evaluate how we feel about something based on the way that our bodies are reacting. Our heart rate quickening may be a result of a spark of fear.

The problem that Barrett raises is that we can – and do – interpret our bodies signals incorrectly. We believe that we’re becoming attracted to someone – when we really have the flu. (Her example from the book.) Sometimes the things that we interpret as coming from our mental state are just side effects of our body doing its normal processing.

Our brain can and does make wild predictions about what is happening to our bodies on very little evidence.

Affect, Arousal and Valence

Affect, as described by Barrett, is a general sense of feeling – not an emotion, but simply a feeling. Other definitions say that affect is the expression of emotion. Affect is described as having two dimensions – arousal and valence. Arousal is how alert or relaxed you are – effectively, the relative state of the competing sympathetic (aroused) and parasympathetic (non-aroused) systems. Valence is how pleasant or unpleasant you feel.

I can say that I use an exercise in my information architecture workshops where I hand folks emotion words and ask them to categorize them. The intent is to indicate the degree of difficulty that sometimes occurs with content that you’re unfamiliar with. One of the most common dimensions that students attempt to categorize emotions into are positive and negative. Interestingly, Buddhists believe that emotions are afflictive or non-afflictive rather than positive or negative. (See Emotional Awareness.) Their point is that anger, for instance, can be non-afflictive if it motivates you to address the disappointment (anger is disappointment directed) and afflictive if it paralyzes you or causes you to ruminate.

Categorization

Barrett points out that, when we create categories, we’re not discovering similarities in the world, but rather we’re creating them. When my students are categorizing, they’re creating mental structures that allow them to simplify objects into categories that they can work with. These categories don’t objectively exist in the world, but they exist inside the heads of my students.

For each category, there’s a representative prototype. That is, for “furniture,” you’ll see a specific object. For most folks, it’s a chair, a couch, or a table. (I’ve done this exercise a few times.) If you put rugs into the furniture category, it will be difficult for other folks to find it there because the specific item looks and “feels” nothing like the prototype. Therefore, developing information architecture is difficult. You must recognize that not everyone sees things the same way, and there will be some items that don’t fit the prototype. Once you’ve created a category, you’ll fall into the trap of the curse of knowledge (see The Art of Explanation) and be unable to think that others don’t know about the category.

Concept Development and Prediction

It’s a simpler model to think about information being processed linearly, from individual sensations up through concepts and into our perception. However, the reality is that the process isn’t linear. Things don’t flow only in one direction, from the many sensory neurons to the neuron clusters making up concepts. Instead, as we learned in The Tell-Tale Brain, the path is bidirectional. The information is fed upward, and as concepts are formed and predictions are made, that information is fed back into the neurons that are working on less-processed data.

We push data back from our concept into what we see, in some sense distorting it by amplifying the attributes that match our expectations. This isn’t a desirable situation but rather an adaptation. It is how evolution allowed us to be successful with so little processing power. We swing the spotlight of attention to the areas that are the most interesting. We identify ways of recognizing things that don’t require further processing and ways of identifying those that do.

We like to believe that the concepts we see are independent of our beliefs or predictions, but this isn’t the case. We see, as Chris Aryris said, what we expect to see. (See Organizational Traps.)

Words, Collective Intentionality, and Emotions

Barrett argues that emotions are social realities. They are, in a sense, a way of communicating our inner state to others. She further argues that having a word – or a name for the emotion — makes this substantially easier than having an emotion for which we have no word. (Actually, in some places, she implies it can’t be done, and in others, concedes it’s possible, just difficult.)

So, I’ll concede that our emotions are more complex than we’d like to believe. There’s more going on than a simple amygdala hijack. (See Emotional Intelligence.) There are ways of reducing sensitivity to amygdala hijack. There are factors – like adverse childhood events (ACE) that make it more likely. (See How Children Succeed.) So are our emotions constructed from our previous experience and our skills? Yes. However, I’m not convinced that this means that we have to “throw out the baby with the bathwater.”

I often think of Jonathan Haidt’s Elephant-Rider-Path model. (See The Happiness Hypothesis and Switch.) The rider can’t control the elephant when the elephant really wants something. However, I think of the rider-elephant relationship, and how your emotions can be calmed when the elephant (emotion) trusts that the rider (rational reason) will respect it and keep both the rider and the elephant safe. Building that metaphoric trust is the way that I believe we get better at managing our emotions.

I still don’t believe that I’ve figured out How Emotions are Made – and though I’ve got dozens of issues with Barrett’s sometimes skewed data and logic, the challenge of our assumptions about how emotions work makes it worth the time. Who knows, maybe you’ll figure out How Emotions Are Made.