Goal

Emotions are foundational for human flourishing. Concepts such as happiness, love and peace constitute much of what it means to flourish; concepts such as grief, anxiety and anger constitute much of what it means not to flourish. Yet these emotions have eluded any clear investigation because nobody can agree on how to operationalize them, or even what discipline to assign them. An overarching goal is thus to bring together psychologists, neuroscientists, philosophers and theologians to clarify our thinking about emotions; to use interdisciplinary approaches to study emotions; and to target specific topics for focused experiments to increase human flourishing and well-being.

Opportunity

What has been done is piecemeal. However, the piecewise progress that there is can be assembled to take stock and to move forward. Recent theoretical work in emotion, as well as progress in the tools to study them scientifically, offer an opportunity to bring together psychological and neuroimaging work in humans, circuit-level studies in animals, and a revision of historically entrenched theories. Adversarial experiments could test current theories.

Roadblocks

The main challenge to an integrative science of emotion is that the topic has long-standing theories that are outdated and difficult to revise or reject. We should start with a broad critical discussion to clear the slate, and then take a more data-driven approach that permits revision of all of our prior ways of thinking about emotion.

Breakthroughs Needed

Stage 1: Theoretical and conceptual. Philosophers, neuroscientists, and psychologists need to have structured debates and meetings to generate a list of features of emotions and some working definitions. This stage should produce clearly defined questions: what are the necessary features of emotions? Are there emotion categories or dimensions? What are the best examples of states that are emotions, and states that are not?
Stage 2: Data-driven discovery. No theoretical considerations alone can give us an answer to what emotions are: we need data. Once Stage 1 has cleaned up and put forth a more refined list of emotions and their properties to tackle, we need large-scale data to begin quantifying those properties and looking for relationships between them. Modern machine-learning and dimensionality reduction methods could be applied.
Stage 3: Focused experiments. Once we have better models for emotions, these will almost certainly be non-unique. How do we decide among them? Experiments that manipulate specific variables while holding others constant will be needed, and likely neuroscience data will be required. In general, these should be adversarial experiments testing competing theories.

Key Indicators of Success

The only way to assess success is by engagement with the entire community of scientists, philosophers, and theologians. Proceedings from meetings should be published rapidly; all data from experiments should be made publicly available.