I recently attended the annual conference of the Society for the Science of Motivation (SSM) in Barcelona. Joyce was presenting a paper about our recent work automating subcategory coding for implicit motives, and I and another co-author were able to attend.
Location and participation
SSM is a small-ish conference, and is generally held at the same time as the Association for Psychological Science (APS), often at the same venue, as many presenters/members attend both conferences. This year it was at a different hotel but in the same general part of Barcelona, which actually marks the first time that the main conference has been hosted outside of the USA.
Barcelona is a lovely city and it would have been nice to explore a bit. Unfortunately we were only able to stay for the duration of the conference itself, but the hotel was very good and we were able to visit La Sagrada Familia. We had some nice tapas at local cafes and evening dinners with good food and conversation.
Automating implicit motive coding
It was really great to meet other people who work on motivation science. Some folks are also trying to automate these processes, with various approaches, using machine learning and large language models. I think our approach is currently the best one, for various reasons (maybe I’ll write more about the reasons in another blog post), but of course I’m somewhat biased.
There was a lot of interest in our models, as well as the other models that automatically code for motives. We provided a QR code linking the blog post which illustrates how you can run all the models for general category coding and many of the audience members took pictures of that slide, so it clearly struck a chord.
There were a number of interesting talks and I found myself taking quite a few notes. Many talks dealt with theories of behavior that I’m not familiar with, so I have a lot of reading up to do. In general, though, I found there were good points of connection between various psychological theories and behaviors related to speech. Linguists are observing behavior too, after all!
The automaticity of wishful thinking
One of the keynotes that stood out to me was by David Melnikoff, titled “The Automaticity of Wishful Thinking”. He examined how different experimental settings involving lawyers resulted in similar outcomes regarding how people updated their beliefs. Not being a specialist in the field, I may have gotten this wrong, but essentially it seems that if lawyers were defending a client, they were highly likely to assert their client’s innocence, even when they were explicitly told that the client was guilty.
This makes me wonder a bit about the situation being examined - does the context influence how participants respond? Does the fact that this is a western legal context, with all the requirements and presumptions (“innocent until proven guilty”) impact the findings? If this study was conducted in non-western context would it give a different result? And what are the stakes? If the stakes are not related to the identity of the participants, would the results not be as significant?
Strengthening Theory
Another keynote that stood out to me was by Susann Fiedler, titled “Strengthening Theory: From Verbal Accounts to Testable Propositions”. She reported on a project aimed at making psychological theories testable, by having multiple psychologists re-formulate existing theories using logical propositional statements. It reminded me quite a bit of the Natural Semantic Metalanguage approach in linguistics pioneered by Anna Wierzbicka, where lexical items are reduced to semantic primitives. But instead of lexical items, it is theories that are reduced to their most basic aspects, in order to test them. It’s quite an interesting approach, and you can see one attempt (for Self-Determination Theory) here.
Final thoughts
Overall the conference gave me a lot to think about. At the same time I was struck by how similarly researchers in psychology and linguistics approach issues. Hopefully I can continue to learn from my colleagues in psychology about how human behavior and language behavior might be shaped by various motivations and needs.