Empathy as analogy: Inferring other minds &
the LISA architecture
In any complex socio-cultural system agents are faced
not only with ethical dilemmas such as altruism, but also with the more general
epistemological problem of understanding other minds. Recently, Alvin Goldman
argued that an appreciation of empathy was necessary to these issues. He
suggested a simulation-theory (ST) view of folk psychology intended to replace
the reigning view of theory-theory (TT). In this view, we do not understand
others by applying any “theory of mind” to them, but instead use the contents
of our own minds to simulate them, a process from which empathy emerges.
Like Goldman, I hold a defined view of empathy as essential to the understanding
of how we attribute beliefs, desires and other mental states to our fellows.
However, I support the recent view of empathy as a form of analogical thinking
that includes emotion and describe how this view can end the debate between
ST and TT, revealing them to be necessarily compatible. I describe LISA,
a symbolic-connectionist model of relational inference and generalization.
I maintain that the LISA architecture is uniquely suited as a starting point
for a discussion of empathy as analogy. Furthermore, the capacity limits
inherent in the model mirror those of human WM and suggest the possible integration
of reflexive and reflective processes in human thinking. Finally, I appeal
to the neural imaging studies of false memories to shed light on these distinct
processes, explore their possible relationship to ST and TT and discuss these
findings within the framework empathic-analogical understanding
of other minds.
Charlotte Nolan
University of California in Los Angeles
cmnolan@ucla.edu