What is the Function of Emotions?

About 30 years ago, when I was at university, I had a good friend who was a teaching assistant in computer science, and who was quite interested in artificial intelligence.

Since both of us were night owls, we used to sit up into the wee hours speculating about thinking computers.  I recall once asking him what role emotions might play in a thinking computer.  He answered that emotions were "primitive" thoughts.  That is, they were the half made up, vague, and fuzzy precursors to clear thinking.  And it followed that a thinking computer would not need to have emotions anymore than it would need to have partial or incomplete thoughts.

Thirty years ago, his answer didn't sound quite so outlandish as it might sound to you today.  There was quite a bit about emotions that just was not known back then.  And his notion that emotions were "primitive thoughts" was a respectable guess. 

Yet, if today we know better than to think of emotions as primitive thoughts, then what are they?  Or, in other words, what is their function?

Well, one way to approach that question is to first ask what people would be like if they had absolutely no emotions at all.  As it happens, we actually know a little bit about what people are like when they lack any ability at all to feel emotions.

However, if you think people who cannot feel emotions function like the famous (and fictional) Mr. Spock, you would be wrong.  Mr. Spock functioned quite well without any emotions.  Yet, real people do not function well without emotions.

The reason we know what real people are like when they lack any ability to feel emotions is because -- now and then -- someone has an accident that damages his or her brain in such a way that s/he no longer has any emotions.

One of the few scientific studies of such people that I've heard of discovered they are dysfunctional in a surprising way.  Surprising, because many of us tend to think of emotions as things that typically get in the way of making good decisions.  Yet, if that is true -- if emotions do indeed get in the way of making good decisions -- then we would expect to find that people who have no emotions routinely make better decisions than the rest of us.

In fact, we find they have great difficulty making any decisions at all.

In general, decision making requires us to prioritize things.  That is, in order to make a decision, we must decide what is -- and what isn't -- important to us.  It has been found, however, that people who cannot feel emotions cannot easily prioritize things in such a way that they can decide between one option and another.

Suppose, for instance, that you yourself had no emotions.  One day you are sitting in your office when you hear the news that the stock market is crashing.  You realize that your life savings are in danger.  But it's about noon.  Do you call your stockbroker or go to lunch?

If you have no emotions, it will be very difficult for you to decide which course of action is most important to you.

So, if the scientists are right -- if the above is true -- then at least one of the functions of emotions is to help us prioritize courses of action.  My friend and I didn't know about that 30 years ago when we would sit up late discussing artificial intelligence.  Which makes me wonder what new things will be known about emotions in another 30 years?

1 comment:

  1. It is true that emotion can cloud judgement. And often it does. And sometimes that's a good thing (such as overlooking the faults of someone because you love them).

    That's an interesting study you reference, but if we were better in understanding ourselves, it may not be that surprising. Our vernacular seems to point to emotions in decisions. How often do we say "I feel" or "I like" when making decisions? Back when I worked in an engineering office, we would often half-joke about getting a "warm, fuzzy feeling" about a particular design.

    For humans, I think emotions are the difference between color versus black-and-white, where emotion can be an enriching aspect. From that study, obviously they are necessary too. For AI, I think that there will be a strong drive for programmed emotion, if for no other reason than to drive sales.

    It will be exciting to see what does come in the next 30! :-)


Comments Welcome -- but no flaming. If you wish, you can email me at paul_sunstone@q.com