I was teaching this morning — a small course in grammar — when I was struck with how cool pronouns are because of how much they tell us about the mind. This is one of those things, like the double-slit experiment where the phenomenon can be easily observed, but the underlying mechanisms are actually pretty complex.

So let’s start off with a question, and some observations.

Basic question: why do we we use pronouns in general?
Follow-up question: what happens when we don’t?

Observations:
Let’s take a small invented passage, where, instead of using pronouns, we use the full form of the nouns. (Not actually a quote, but I’m going to set this off in a block-quote.)

Zoe went to the store. And then Zoe tried on some shirts. But Zoe thought Zoe didn’t look that great in the shirts. Zoe thought the shirts were not Zoe’s color. So Zoe got Zoe a cupcake, which was Zoe’s favorite color. Then Zoe sat down and fed Zoe’s self the cupcake. Zoe thought the cupcake was delicious, but Zoe was sad Zoe ate the cupcake in just two bites.

Wow, annoying, isn’t it?

So what’s going on? Why are we so annoyed by this?

To answer this, we have to think about how our brains operate. Our brains are gigantic information processors. They take information from our senses and other sources, and then they process that information into usable forms… into things that we can use to our benefit. This applies to basically any organism with a brain. Consider a small nematode worm. The nematode will have various tactile sensors along the length of its body, may be sensitive to light, perhaps picks up chemicals in the water, and so forth. These sensory observations will help the creature to identify if it’s in contact with things or if it’s moving from light to shadow, for example. But what good is all that? What good is it to be able to sense that you are moving into a shadow? You also need to be able to respond to that and act on that information, and that’s where our brains come in. That shadow could mean something swimming overhead, which might mean a predator, or it might just be the sun going behind a cloud, or a hundred different things. So the first thing that’s important is to be able to identify what these different things in the environment are, and then from there, to form different predictions about how those things act.

This is both fairly simple, and at the same time, fairly complex. Knowing what’s in the environment, at its core, is a form of sampling. We look around to sample the light in our environment. If we were in a dark unknown room, we might reach a hand out to try and feel for and sample where things are. This can give us a quick basic overview of the environment. At the extreme of taking a thorough inventory, we only really need to sample each thing around us once if we have a good memory. On the other hand, being able to make predictions places very different demands on the information you collect. To make predictions, you first need to know: how things act, how consistently they act, and what things act like each other. If you are an organism planning your response, you may also want to know: what behaviors you can take in a situation, and then which actions are most likely to be useful in that situation. Super complex for complex situations. At the core here, making a prediction of an object’s behavior requires you to sample repeatedly from it. At the extreme, being able to predict an object’s behaviors perfectly over all time, you would have to have a perfectly working model of the entire universe.

These two sampling behaviors are at odds, though! One wants us to sample lots of stuff just a little. The other wants us to sample just a few things a whole lot of times! It’s the push and pull that we have to navigate as organisms. In general, we prefer the middle ground — a sort of goldilocks zone (as my good friends and colleagues have described for human learning). Our brains have come up with some internal behaviors to help us find that appropriate middle ground, since too much time at the extremes could be dangerous (too much time sampling everything, we don’t notice threatening behaviors unfolding over time; too much time sampling repeatedly from single things, we may not notice the obvious threat behind us). Two of these constraining internal behaviors are irritation and boredom. How exactly the brain realizes these… that’s beyond me. I won’t pretend to understand the neural basis of boredom. But here’s roughly how it applies behaviorally: if you spend time focusing on complex things without spending time doing more overall sampling, you tend to get antsy and irritable. Think of working on a tough problem, and needing a break to walk around. (That need for a walk to sample your environment and celar your head is a deep, embedded behavior right there!) Or, conversely, if you are in a place where you have looked around and there doesn’t seem to be anything to investigate more deeply, you get bored and antsy. If you were asked to wait in a plain white room, how long would it take before you reached for your phone? And if your phone was dead: oh, terror! Left with your own thoughts? Please, no! There’s even a now classic study that showed humans would rather administer an uncomfortable shock to themselves when bored rather than just sit and do nothing. That’s how strongly we crave new information.

Pronouns are an interesting result of this. The first time new information is presented in a dialog, we tend to use a full descriptor. “Did you know Beth and Jose are going to go see Wonder Woman 1984?” I’m providing you with the whole title of the movie and the names of both of our friends because that gives you specific information about the situation. Doing so allows you to retrieve information about the relevant movie and the relevant people and keep those in mind. (This is in a way a form of deep sampling; I’ll also return to retrieval in just a moment). This has also become a signal in our cultural exchanges for something you do when you bring up new information. If I just use pronouns when I first meet you without any other prior dialogue, “Hey, did you know they are going to see it?”, this can be very irritating because it’s not enough information for you to use to be able to understand who / what I’m talking about. It’s undersampling. Conversely, by continually using the full proper nouns over and over again, it signals that new information is continually being brought into the conversation, when indeed it is not. This is also irritating, because it cues us into deeper sampling about something that we’ve already observed. We are being forced to consider the same information over and over and over, and it very strongly hits our irritation buttons. It may be additionally irritating when someone else does it, because it takes away our power to choose what we sample and investigate — we are being forced to reprocess things against our will. (Note: some kids characters do this. Elmo, I think, is guilty. Many adults also find Elmo extremely annoying.)

Pronouns are neat because they fall into that middle ground of information seeking, but they do so in a really neat way that also tells us about how the mind operates. A pronoun is like a placeholder for information. It can refer to pretty much any type of information. “Look, there’s that rabbit again. It‘s been eating all my lettuce!” (It = that rabbit.) Or “Swimming in the ocean can be dangerous. It is not recommended for beginners.” (It = swimming in the ocean.) There are also other pro- form words, like pro-verbs (not proverbs, mind you), but that’s a bit of a deeper aside. So, we’ve got these placeholders for information. They stand in for information that we’ve collected, but that we can readily pull back up to attach more information to as needed. And, what’s also interesting, we get confused if we use too many of them. That suggests, that whatever’s happening in our minds, we have a limited room in this shortcut / working space. This is one of the main ideas behind working memory (sometimes: short-term memory), and it can be thought of sort of like a modern computer. (Note, I’m mostly sure that modern computers were not designed to operate like human minds, so that they seem to roughly work in the same sort of way is a neat example of how information structure can dictate what the machine that processes it should look like.) Modern computers tend to have some sort of long-term storage or hard-drive. They also have some amount of faster privileged random-access memory (RAM) that is used to storing things in the short term that you are going to access frequently. When programming, programmers have to keep these things directly in mind. You can have lots of resources packaged with a program, but to use and manipulate them, you will want to load them into memory. For efficiency, you also don’t want to load the resources every time you call them, so what people tend to do is to start off a section of program by loading up the resource, and then giving it a short name or a pointer to where that information will be kept in memory.

In a lot of ways, this is like nouns and pronouns. We access information in our long term memories when we hear a noun. If that information is in the topic under discussion, we keep it a privileged place for easy access so that we can refer to it with a short label. But how many things we can keep around in our pronoun space probably tell us about working memory capacity, and as such, the structure of our minds.

So, pronouns — pretty neat! Most of the time, we don’t think about them. But when it come down to it, to understand what’s going on, we need to consider information theory, the function of our own machinery, and we even get to peek a bit behind the curtain into what the structure of the mind might be like. Cool!

One last note:
One of my personal pet-peeves, and something that is directly related to the above, is also touted as good advice among the business crowd. And that’s that you are supposed to repeatedly use someone’s name in order to build up a rapport with them. “Use a person’s name three times when you meet them, and it will help you remember their name and also make them feel a more personal connection to you.” Humbug. There is almost nothing that irritates me more. One, it’s easy to see through it as a cheap trick intended to make me like you more so you can sell me more stuff. But, two, as above, when someone refuses to use pronouns when appropriate, it is deeply irritating because they are strong-arming your own information seeking patterns. Our own names are highly privileged, but they are also used for very specific purposes. Getting our attention is one. Parents scolding their kids seems to be another. Meeting new people for the first time, a third. Repeatedly using my name accesses these “hey, become alert” processes, puts me more on edge, and makes me really, really, want to go buy from basically anyone else. Do we like to be known? Yes. Do we like to be remembered? Yes. But don’t overdo it.