Making sense of what you sense

As we saw in last month's newsletter, A World of Your Own, our view of the world around us is greatly influenced by our expectations. This goes beyond our sensory perceptions, though, and affects our cognitive perceptions or judgment, also.

Note that the terms "sensory perception" and "cognitive perception" are my own. Neuroscientists and cognitive psychologists don't seem to have pinned down the dividing line between sensation and cognition. While there is a lot of published research that there is "top-down influences" from cognitive concepts such as beliefs, memories, and mental constructs that influences our apparent sensation, others have rejected that notion[1]. On the other hand, other research has shown what appears to be memory influences directly in the sensory neurons of mice[2]. Ultimately, I don't think that it's terribly important exactly where the influence happens, and I'm sure that the current state of research does not fully model the human nervous system.

Whether your past experience prevents you from observing what's happening, or merely prevents you from noticing, it does isolate you in A World of Your Own[3].

Recently on Mastodon someone asked "Do you think it's ok to use an LLM for inspiration, e.g. to get started on a book report or research paper, basic structure, points to touch on, etc.?" My response was "If you follow a provided basic structure and list of points to cover, though, then you're ceding more than inspiration to the bot. You're letting it limit you."

They immediately started arguing with me that it was a good starting point and wouldn't limit you. They apparently already had a mind-set that

1. LLMs are trustworthy sources for outlining an area of study, and
2. that they could easily "add creativity and self expression" if they chose.

The ironic thing was that their mind-set had already limited the alternative opinions that they could hear and consider, even though they professed to be curious about other views. They might be able to add some flourishes, but they were apparently unable to consider rejecting the opinion they had already adopted, or setting it aside and thinking about the situation from a different starting point.

They opined that being limited by a mind-set you had adopted was a "human flaw" and not the fault of the tool. At this point I disengaged from the conversation because the mind-set limitations were becoming recursive. I question, though, the value of a tool that doesn't play well with the tendencies of the humans using it. Perhaps you can apply work-arounds to counter those tendencies. Perhaps the most effective work-arounds are those that precede the use of the tool, given what is known about the difficulty of changing one's mental model.

What I'm calling a mindset or mental model, Piaget called a schema, i.e., a mental structure that we use to organize, process, and store information. If you've read the book or done the research, you likely have built a schema organizing the contents of the book, or the research topic. These schemata may, of course, refer to schemata you've previously acquired. It could be, though, that your read the book without really engaging with the ideas presented in it. Or your concept of "research" might be collecting a bunch of URLs to relevant articles. If you haven't mentally processed the book or research topic, you might short-circuit the learning process and uncritically accept the LLM output as The Answer.

According to Piaget, learning something involves assimilation of new bits of knowledge added to our existing schemata. When there's no conflict between the new information and your current mental model, it's pretty easy to enhance the model to hold the new information.

When there is a conflict, it's not so easy. As Leo Festinger noted[4], people strive toward internal consistency, and the dissonance between two beliefs is uncomfortable. When your current mental model and the new information, disagree with each other, it produces cognitive dissonance that you'll try to reduce to achieve consonance.

The easy path, of course, is to think "oh, my previous belief wasn't quite right" and modify your current schema to hold a new understanding. In the case of women doctors and male nurses mentioned in A World of Your Own[5], it was pretty easy to realize that this was a perfectly reasonable situation, even if I'd never experienced it before. I modified my then-current schema of doctors and nurses to remove the gender expectations. This modification of the current mental model is a process that Piaget called accommodation.

Sometimes accommodation goes a bit wrong, however. Perhaps I'm emotionally committed to a strong sense of hierarchy, and my schema of that hierarchy includes that men are higher than women, and doctors are higher than nurses. It's a long step from there to beliefs that individual people may vary a lot, but worth is not an aspect of gender, or that doctors and nurses are complementary roles that can reach a more beneficial outcome working together. To process the information incongruent with the existing schema requires processing capacity, a schema organized such that it is readily modified for this particular new information, and motivation to favor accuracy over maintenance of the current schema. Without these prerequisites, new data is unlikely to change peoples' minds. It's a much shorter step to deny that a woman doctor is "a real doctor" and deny the reality around me.

"Generally, when information is incongruent with the schema, the information may be dismissed as 'bad data,' and the instance stored with the schema will not be revised."[6]

Sad, isn't it, to so easily miss our opportunities to learn more about the world around us while we hold tightly onto what we think we know. Knowing THE answer can blind us so easily.[7]

Have you ever noticed when what you already know has gotten in the way of learning something new? Think how many other times that might have happened and you never noticed.

P.S. Please help me out. I would appreciate it if you could reply with a few words about your reaction to this newsletter issue. Think of it as extreme opt-in, non-persistent and totally voluntary tracking.

And if you'd like to talk further about this topic, you could schedule a Zoom Session with me to talk about it.

Schedule Zoom Session

Or you can also simply reply to this email or send an email to newsletter@idiacomputing.com to continue the conversation. There's a person, not a bot, on this end. I'd really love to hear from you.