How are Words Connected in Our Minds? Priming

So let’s talk about connections. We can find connections between lots of things, whether it’s evidence connecting people to crimes, or noticing that two items are the same colour or size or shape. Our brains are always seeking out patterns, looking for ways to link things together in every second that goes by. But when we’re seeing or hearing language, what connections do we use? Here’s the quick answer: ALL OF THEM. I’m Moti Lieberman, and this is the Ling Space. When we’re taking in language, we have to cope with a huge rushing intake of new information all the time. In a normal conversation, you’re getting bombarded with probably a couple hundred words a minute, and when you’re reading, it’s even more - the average American adult reads at a brisk 250-350 words per minute. That information stream is so rapid that we’ve got to do something to keep up the pace. If we waited until every sentence ended to decide what to do, we’d be hopelessly lost, way behind the moment in whatever conversation we’re in. You probably know this feeling from trying to have regular-speed conversations in a new language you don’t know well yet. It’s super hard. But in a language you’re fluent in? Just talking normally feels, well, normal! It doesn’t come across as taking much effort. So does that mean we’re all such linguistic geniuses that we can string together clear readings of complex sentences right away, spotting all the right connections on the first try? Not exactly. We actually use a more brute force approach to the matter. Every time you hear a word, your brain doesn’t really have a great idea of what exactly it’s going to need to do with it, or even if it 100% registered the word correctly. So what you do is you pull up all the things that are related to whatever you just heard. Like. All of them. No matter how they’re related. And then as soon as you identify the one you actually need, you just drop all of the rest in a matter of a tenth of a second or two. That’s what your brain does all the time. If you’re listening to me? It’s doing it right now. But how do we know that people do this? Well, linguists have devised a kind of experiment that lets us probe our mental connections, and it all revolves around how long it takes us to recognize whether a certain string of characters qualifies as a word in whatever language we’re testing. In this experiment, you’re seated in front of a computer, and you have to push a button as quick as you can to say whether something like “sheriff” is a word in English. And we compare that to how long it takes you to decide that something like “steriff” isn’t a word. But that kind of word recognition task isn’t enough to tell us about connections. It just forms the base of the main task. What we most want to see here is what happens when we flash you a different word before the one you have to make the decision about. So let’s say that right before you see our target word “sheriff”, you see another word, like “badge.” And in another case, you get a different, probably unrelated word before “sheriff”, like, say, "spoon". If seeing “badge” helps you decide more quickly that “sheriff” is a word than “spoon” does, that implies to us that “badge” activates “sheriff” in your mind. Since it’s already been pulled up out of your mental files, you can tell more quickly that it really is a word. This is known as priming, and so we would say that “badge” primes for “sheriff”, but “spoon” doesn’t. And using this priming task, we’ve found a whole lot about how we connect up words and structures in the brain. The kind that we’ve just been talking about, with “badge” and “sheriff”, is known as semantic priming, because the two concepts are connected by their meanings referencing each other in some way. That’s how we know that, say, “student” will prime for “teacher”, or “parent” for “child”. If you can see a link in meaning, then it’ll probably show this priming effect. But meaning isn’t the only kind of clue that we can get. We also pull up connections for things that sound similar. If you have two words that rhyme, they’ll also prime. Even though they don’t really have any relationship between them, we still find that “cheese” and “please” will prime for each other, because they’re phonologically close. But they don’t have to sound alike – they could also just look alike. The visual form of the word can be enough to prime. “Bomb” and “comb” start differently and have different vowels, and I don’t think that bomb combs are a semantically connected thing, but they’re spelled almost the same, and that’s enough. You’re gonna see some priming there. And that’s not even the end of priming. I mean, one language isn’t even the end of the strings our brain tugs on. For a French-English bilingual, the word “cat” will prime for both “dog” and “chien” - all the versions of dog in their mental dictionary. When bilinguals are thinking in one language, there’s some evidence that we’re pulling up related stuff in another language. Take a look at another French-English priming study. Here, the researchers were asking participants to consider pairs of words where the first was in French and the second in English. Between English and French, there are words that are spelled the same, but are pronounced differently, and don’t mean anything remotely similar to each other. So like, coin in French is corner, but coin in English is a small metal monetary disc. So if you’re accessing French, you wouldn’t really expect that coin would pull up a word like money in English. Corners and money aren’t related. But coins and money are, and the researchers found that even if the French word’s meaning wasn’t related to the English one, as long as the words were semantically linked in English, recognizing the target English word was faster. So we’re pulling cross-linguistic strings – coin does prime for money. But not always - the results for cross-linguistic priming are more mixed, as we’ll talk about back on our website. And anyway, everything we’ve been talking about here is about looking at individual words. Maybe we can’t help but pull up all that information when what we have to focus on is just the words themselves. But it turns out, we do this in the context of sentences, too. And we know this because of experiments that apply the kind of task we’ve been talking about, where you have to make decisions about whether a string of letters is a word or not, at a particular point in a complete sentence. Let’s take a look at one of those experiments. Okay, so here’s our sentence: “The man was not surprised when he found several bugs in the corner of the room.” Now, think about bugs. It’s ambiguous, right? It could mean like crawly insect bugs, or listening device bugs. If when we hear sentences, we prime for related words as we go through, we’d expect both meanings of “bug” to do some work. So that’s what got checked: after “bugs,” people were asked to do that word decision task. And this was on words like “ant,” which related to the insect meaning; “spy,” which related to the secret mic reading; and “sew,” which doesn’t relate to anything. And it turns out, people answer faster about “ant” and “spy”. So we do priming in sentences, too. But what about when it’s really clear what meaning we intend? After all, there's usually going to be some context that’ll help us decide which meaning we want for a sentence. What if the sentence is more like this: “The man was not surprised when he found several spiders, roaches, and other bugs in the corner of the room.” This is almost certainly the ant kind of bug, and so we shouldn’t be surprised that bug primes for ant there. But… we also still prime for spy. You check right after bugs is said, and spy is still there in the listener’s mind. Intriguing! But if we went along priming these unrelated words forever, that’d be pretty challenging for our brains: we’d explode with unrealized possibilities that we kept in our heads. And so the experimenter here tried moving the place where he performed the word decision task down a bit: instead of it being right after bugs, he did it two words later, between "the" and "corner". That’s not a long time ahead – remember, we can read maybe 300 words a minute. And yet, by that point, spy doesn’t do any better than sew for how fast we can identify it as a word. Only ant, the word that was contextually called for, sticks around to get primed. We just drop spy. And so, that’s what our brains are doing all the time. Priming studies have found that we can activate these connections after only 50 milliseconds: that’s a twentieth of a second. That’s not even enough time for people to register they saw a word consciously - did you just see the one that we had flash on the screen? But as soon as it’s clear that we don’t need something – two little words later! – those links we don’t need get dropped. We’re constantly working out all the possible angles that our conversations or our reading material might send us careening around, and then just subconsciously forgetting those links when they turn out not to be useful. Until the next time we hear them, and then all the strings get yanked up once more. That’s the way we deal with our connections when our minds are in prime condition. So we’ve reached the end of the Ling Space for this week. If you let go of the links you didn’t need, you learned that we constantly scan for connections to everything we hear or read; that these connections get activated for when things have similar sounds, meanings, or written forms; that we perform this priming even during sentence processing; and that we activate these bonds quickly, and let most of them go just as quickly. The Ling Space is produced by me, Moti Lieberman. It’s directed by Adèle-Elise Prévost, and it’s written by both of us. Our editor is Georges Coulombe, our production assistant is Stephan Hurtubise, our music is by Shane Turner, and our graphics team is atelierMUSE. We’re down in the comments below, or you can bring the discussion back over to our website, where we'll have some extra material on this topic. Also, try dropping by our new store, where we have a bunch of cool linguistics stuff. Check us out on Tumblr, Twitter and Facebook, and if you want to keep expanding your own personal Ling Space, please subscribe. And we’ll see you next Wednesday. Ābāra dēkhā habē!