[SOUND] We can think about age of acquisition, right. When essentially learning occurs, early versus late. And interestingly, the notion of age of acquisition of course appears in the bilingual literature. Right. When people talk about second language age of acquisition. Relatively early in my career I discovered this literature on age of acquisition. Now, the interesting part about this was that I was looking for an article, so I'd found an article online, and then I went down to the library to make a photocopy. That was back in the days when they were, all the journals are still lined up in, in a library but that was before PDFs. And what I did is I was looking for this article and I found another article on the age of acquisition. And I said, oh, that's interesting, age of acquisition, you know, it must be about bilingualism. And what I discovered was that in fact it wasn't. It was a study by Morrison and Ellis. And in this study, what they had done is they had looked at age of acquisition, that is, when a word was learned. So we can think about age of acquisition. So for example you can think about words that you might have learned early in your life. Right. And those don't have to be words in two languages, they could be words in a single language. And you might say, well, now, isn't it the case that a word I learned early in life is actually a word that I hear more often, that's called frequency. And at the time, the idea was that frequency, that is how often a word was heard was a primary determinant of how fast someone could read the word. How well they might remember it. How many mistakes they would make when they read it. And the idea was that high frequency words, words that were heard quite often were better or more well represented in the mind than words that were low frequency and not heard as often. Morrison and Ellis suggested that in fact, it wasn't just frequency that was driving the effect. That in fact if we took words that were high frequency, we would find most of those to be words that were learned early in life. So they wanted to take these two things apart. Take how often a word was heard out of when it was heard first. Now, you might say, well, aren't those the same thing? Well, they can be, but they also might not be. So let's take an example of a word. Think about the word dragon. Now, I know, maybe perhaps there recently have been movies about dragons or you might have been watching movies about dragons, but it's not that frequent a word. Right? Most of us adults don't go around in the world talking about dragons every day. But kids do. They do talk about dragons quite a bit and they learn about dragons at an early age. So dragon is an example of a word that's relatively low frequency, but early learned. Right? And so, these two things can dissociate. What Morrison and Ellis did in this study was have people saying the name of a picture, right? That was from a word that was either early or late learned or high or low frequency. Right? Heard often or not often. And what they found was that in fact, when you looked at the effect of early versus late, making them equivalent on frequency, that early was faster than late. But if you did the opposite where you took high and low frequency, but you made them equivalent on when they were learned. So these were matched. These were all roughly the same number of early and late high frequency pictures or words. And roughly the same mix of early and late low frequency words. People were not faster on either one. So their argument was that the primary determinant was early and late and not high and low frequency. It's an interesting result. One that Diane Meschyan, a former graduate student of mine and I took on to look at this, these two effects. So we decided we were going to look at each one, so we created four conditions. Right? Early high frequency, late high frequency, early low frequency, late low frequency. And when we looked at those two conditions, we found in fact, that there was an effect of frequency, how often the word was heard and age of acquisition, early versus late. So both of those played a role. In 2001 I traveled to Leipzig, Germany to visit the Max Planck Institute, at that time, the Max Planck Institute for Cognitive Neuroscience. And there I worked with several researchers. Of course, one of the directors, Angela Friederici also Sonja Kotz and Christian Fiebach. Sonja at the time was a researcher there, Christian was a graduate student. And Christian was very interested in word frequency, he done some work looking at the brain basis of word frequency. So he'd asked people to decide if a word was abstract. Right. So was it something like honesty, or concrete, something like door. And in doing that task, he'd found that there were certain parts of the brain, especially in the frontal lobes, that were more active for the low frequency words. Right? So the brain had to work a little bit harder when deciding if a low frequency word, a word that's not heard often, was an abstract thing or a concrete thing. And when I met Christian, I said, well, what about age of acquisition? And he asked me, what's age of acquisition? Right. I gave him a few articles. He read them. And he said, you know? We could do this. We could actually ask people to rate the age of acquisition. So people are pretty good at telling you whether they learned the word relatively early in their lives or late in their lives. Now, they're not precise. They can't tell you that, you know, when I was three years, five months, two days, right? I learned the word dragon. But they can tell you, you know, I learned that probably around four. And you know, the word fax. I don't know. I learned, you know, I don't know. I could tell you my, around 13, 14, maybe? 15? So I have a pretty good idea, roughly, of what's early and late. I can rate it pretty well. So we asked a bunch of people to rate the age of acquisition. And then we looked at the brain, comparing, right, the frequency of the word, how high and low it was, and when that word was learned. And the results were actually quite interesting. So, the first part which Christian was very happy about was that his word frequency effect still appeared. So you could still find the fact that these low frequency words showed this frontal area. Slightly frontal, roughly in, in language areas. But when we compared the early and late learned words, we found a very different pattern. Right. So for the early learned words, we found activity in roughly the temporal, basically in the auditory cortex. Right. Mostly on the left. And in the area called the precuneus, which is in the parietal lobe. An area involved in integrating information between senses. When we looked at the late learned words, what we found was that in the front of the brain there was bilateral activity not, not in language areas, but actually in front of language areas. Areas involved in what's called effortful semantic retrieval, and having to, to make sense of what this word was. And if, if I had to tell you to the, a more simple type of way to think about it was, what the brains of these subjects seemed to be doing. They were all monolinguals. They were all German speakers, not bilinguals. Is that for early learned words, when they had to make decisions for early learned words, they, essentially, the brain was having to sound out and somehow figure out the auditory footprint of that word. And when people encountered these late-learned words, they essentially kind of related it to something else. So if I had to, if you asked me, you know, dragon. How would I know dragon is a word? Well, I just know it's a word. I don't really have to think about it. Since I learned it early in my life. But if I get a word like fax or a word like abacus, right? Then somehow, I have to relate that to something. So, I might relate an abacus to a calculator. I might relate a fax to a telephone, right? Or to the Internet or to email or something like that. And so the process by which I access information about early and late learned words is different. And that suggests that the way these words are learned is different, right. Learning in childhood is different than learning in adulthood. The brain activity tells us something about the ways in which words are learned. It suggests that words learned early, relatively early in life involve more sensory processing. And words learned later in life involve more effortful processing, which involves connections to other things in the world, to, to other types of things that have similar types of meaning. So this suggests again that there's a difference between early and later types of learning. Even in single words. Even in people who only speak one language. So how do we explain this? And, and how do we conceptualize this? We have a conceptualization within the brain about early and late. But how, how could you think about this in, in a more theoretical type of manner? And psychologists like to think theoretically. So one idea people had at the time was that when you learn a word early in life it's, it's simpler in terms of its phonology. Right. The, the way you learn it auditorily is quite simple and later in life, you kind of have to put together little sound pieces for later learned words. That was one idea at the time. The, the second idea was that in fact, what might happen is that early learned concepts, words form a scaffold, they form a base for later learning. And one analogy I like to use is when somebody arrives at a party. If he arrived relatively early at the party, right? So you get there when there's, maybe you're the first or the second person to arrive at the party. And when you get there, you sort of become a base, if you will. So anyone who arrives after you do will somehow be influenced by the way or the groups that are formed, because of you having arrived early. Right. So essentially, by arriving early, you influence the way that the grouping occurs in a different way, than if you arrive later. If you arrive later to a party, you may actually gravitate towards people that you know or to a group that's already established. And so this was one idea that was used by people who do neural network modeling to think about the difference between early and late learning. Right? That early learning as sort of first arrival influences a system in a more fundamental way. And so that other concepts are built around the early learned concepts or early learned words and that the late concepts also, right, consequently, have to be built around the early ones. And that, that's the difference. And, and our study, Christian Fiebach and I, our study in this respect had some support for both these ideas, right? That, that there's something about the sound that's different about early learned words, but also that late learned words seem to be building around these very early types of meaning that are established.
