1996 LECTURE SERIES

Childhood Memories: Here Today, Where Tomorrow?

Dr. Patricia J. Bauer
Institute of Child Development
University of Minnesota
March 13, 1996

When Do Memories Begin?
Think about your childhood. Reflect upon some of your very earliest experiences and try to think about the very first thing that you can remember. Dig way back and think about the very first event that you can remember that happened in your life. For adults, age three and a half is the average age of earliest memory. Over the past 50 to 75 years, there have been a number of descriptions of infantile amnesia and from those descriptions, we've gathered a number of characteristics of early memories.

The first characteristic is, as I've said, the average age of earliest memory is age three and a half. However, there is some interesting individual and group differences in the age of earliest memory and some of those I've listed here. The first is that women tend to have memories from earlier in life than do men, and the average difference is about five months. First borns tend to have memories from earlier in life than do later borns. The average difference between a first and second born is about three months, and the average difference between a first and third or later born is about five months. Adults whose families moved before the age of four have memories from earlier in life than do adults whose families didn't move. And adults who attended pre-school have memories from earlier in life than do adults who didn't attend pre-school and the difference there is about four months. And finally, although there hasn't been very much cross-cultural research on this phenomena, it does seem to be the case that Americans have memories from earlier in life than do Koreans, and the average difference here is about eight months.

Now, in spite of all these interesting individual and group differences, it remains the case that age three and a half is about the average age of earliest memory. One possible reason why adults may not have memories of events that happened before the age of three is because really nothing very interesting happened at that time. While this might be the case and it might be part of the explanation, but we also have absence of memories for events that we know, in fact happened. Researchers Karen Shinegold and Yvette Tenny asked adults to think about an event that they knew in fact had happened and moreover they knew exactly when the event had happened and that event was the birth of a younger sibling. What the researchers did was to identify adults who had a sibling born when they were between one and three years of age, and a sibling born between the ages of three and five, five and seven, seven and nine, and nine and older. They asked these adults a number of questions about the event. For example, they asked, "With whom did you stay when your mom was in the hospital? Where were you when you first met your new sibling? What presents did the baby receive? And what presents did you receive at the birth of the sibling?" What they found is that adults whose sibling had been born when they were between one and three years of age remembered nothing about the event. They weren't able to answer any of those questions. Now with age, there was a steady increase in the amount of information that adults could remember. So if the younger sibling had been born between the ages of three and five, they were able to answer about 10 of the questions and then there was a steady gradual increase in the number of questions that the adults were able to provide with increasing age.

So it really does seem to be the case that as adults, we don't have memories from the first few years of our life. Why is that the case? Why is it that we as adults are unable to remember the events that happened to us earlier in our lives? So as not to disappoint you in the end, I'm not going to be providing you an answer to this question and the reason for that is not because I'm withholding evidence, but because we don't have a complete answer to the question. What I am going to be providing you is the story of what we've learned thus far in our inquiry into this mysterious phenomenon of infantile or childhood amnesia.

Amnesia and Childhood Memories
For just a moment, think about the meaning of the word amnesia. Amnesia implies one of two things. The first is a partial or total loss of memory. By this meaning what we imply is that there once was a memory, that at some later point in time no longer exists. In this sense of the term "amnesia," you at one time knew with whom you stayed when your younger sibling was born, but you've since forgotten that information. The second sense in which we might mean "amnesia" is in terms of the absence of memory. This implies that we never created a memory of with whom we stayed when our younger sibling was born. It's not that we lost something that we once had, but that we never had it to begin with. Now if we're going to unravel the mystery of infantile amnesia, we're going to have to know which of these senses of the term most accurately characterizes the phenomenon. We're going to need to know if young children did have memories that they subsequently lost, or if young children simply don't form memories of the events that happened in their lives. I think the answer ultimately is going to involve both of these senses of amnesia. That is, it involves both an absence of memory and also loss of memory. Importantly, I think that explanation in terms of absence of memory only applies for a relatively brief period in development. I think there is a time when infants simply do not form memories of events that they're able to retain over any long period of time. But importantly, by late in the first year or early in the second year of life, infants and young children can and do remember specific events. This then is going to lead us to the question of why we as adults no longer can get at those early memories. By telling you the bottom line here, I've bypassed a decades' worth of research on infants and young children memory, and what I'd like to do know is to begin a little bit nearer the beginning of the inquiry.

The first thing we had to do in our inquiry into the explanation for infantile amnesia was determine whether infants and young children remember the events of their life. If you think about how we're going to do that, you soon realize that we're going to have a problem here and the problem that we're going to have is how are we going to get at whether this infant or young child is able to remember. If we were interested in memory in older children or adults, the answer would be simple. We'd simply ask. I would ask and you would be able to provide a report of events that happened to you. But when you're dealing with an infant or young child, that's simply not going to work, and the major reason for that is that children of this very young age don't talk. Even by the second year of life when children are making great strides in language development, they still seem to have a considerable amount of difficulty using their language to talk about past events.

This is nicely illustrated through a conversation between a colleague, Judith Hudson at Rudger's University, and her daughter Rachel. Being a good developmental psychologist, Judy had what we lovingly call a longitudinal research unit which is your own child. Judy followed her daughter Rachel's memory development very closely for the first couple of years of life. This is an example of a conversation that the adult Judith and her daughter Rachel had when Rachel was 21 months of age.

Adult: "Do you remember what you did with Aunt Gail and Uncle Tim?"
No response from the child.
Adult: "Did we go eat?"
No response.
Adult: "What did you have to eat?"
Child: "Chair."
Adult: "Rachel sat in a chair. Did Aunt Gail sit next to you?"
Child: "Yes."
Adult: "What did she read to you?"
Child: "Ernie and Bert book."
Adult: "The Ernie and Bert book?"
Child: "Yeah."
Adult: "She read that to you?"
Child: "Yeah."
Adult: "Yeah, did you eat hamburger?"
Child: "Yeah."
Adult: "And chips?"
No response.
Adult: "And did you have ice cream?"
Child: "Yes, ice cream."

It's pretty clear from this example that if you want to find out what if anything the child, rather than the adult remembers, it would be best to use something other than verbal report.

Memory Recollection Via Visuals
To begin to study memory in children at a very young age, about 10 years ago my colleagues and I developed a technique that would not require that the child provide a verbal report. What we did was to develop a technique that would allow children to show us what they remember instead of telling us what they remember. The procedure involves using objects to produce a unique action or a sequence of actions and then allowing the child to imitate our actions. For example, we might give the child two round barrel halves and a block and we would show the child how to make a rattle by putting the block inside one of the barrel halves, putting the barrel halves together and then shaking it to make a rattle. Or you would give them this array of objects and show them how to make a gong by first putting the bar horizontally across the goal post to form a cross piece and hanging the metal disk from that cross piece, and hitting it to make it ring. The way we use this technique is that we first give the objects to the children and just let them manipulate them. We let them explore them and play with them for a period of about a minute. We then take the objects back from the child and show them how to produce a specific event. So with the rattle, we would say, "you know what? I can use this stuff to make a rattle. Watch how I make a rattle." Put the block in the cut, cover it, shake it, and you have to use that intuition when you're working with young children. After this demonstration, we take the objects apart or the product apart and give it to the young child and then we say to them, "now you make the rattle just like I did." Although we tell the child do it just like I did, we probably wouldn't have to. Children really enjoy this procedure and they're quite enthusiastic about producing the activities that we just demonstrated to them. In fact, within a single session, children will learn six to eight of these little events and they'll work with us for about 30 to 40 minutes in order to do so, so they're quite interested in learning how to make rattles.

To determine what children remember, we compare their performance prior to the demonstration to what they did after the demonstration, and any difference in performance we attribute to their memory for the demonstration. To find out what children remember over a longer term, we allow some time to pass and then we invite the children back into the laboratory and we simply put the objects in front of the child once again. Notice that procedurally, that's just what we did when we first exposed them to the particular event. We just put the objects in front and see what they do spontaneously. The difference is that in the meantime, the baby has been shown how to produce the gong or the rattle. Notice that with this technique, we can measure how much the baby remembers and how much the baby forgets. We can measure how much the child remembers by comparing what they did prior to the demonstration to with what they did after the demonstration, and we can determine how much they forgot by comparing what they did immediately after the demonstration to what they did after that time had elapsed. A small difference means that the child forgot very little and a large difference means that the child forgot more. I'm going to be referring to what the children do prior to demonstration as their "spontaneous performance." I'll be referring to what they did immediately after we showed them the event as their "immediate recall performance," and I'll be referring to what they did after the delay as their delayed recall performance. In each of these phases, we look for two things. We look for how many of the individual actions the children produced and also whether they produced the action in the order that we showed them. With the example of the gong, we look to see whether the child first folded the bar across the goal post then put up the metal disk, and then hit the bar. Although making rattles and gongs may not look a whole lot like a verbal report that an adult might produce about a past event, we really think that this non-verbal technique is measuring the same thing as verbal report. In the case of the adult, what she or he is doing is using verbal symbols to tell you about a sequence of actions that cumulated in some action. What the baby is doing is using objects to produce a sequence of actions that cumulates in some outcome. Both of these types of reports make heavy representational demands on the child or the adult. In the case of the story or the action sequence, the outcome of the activity isn't available in the environment. The rattle or the gong doesn't exist in those objects that we give the baby. The only way the child can know what they're supposed to do is to remember it. Moreover, the sequence of actions that we used to make the gong also isn't available in the objects themselves. Again, what the child has to do is to remember how we produced that particular sequence of actions. In fact, the major difference between verbal reports and non-verbal report is in the vehicle that's used. For the adult, the verbal symbols that you use have to come from memory. Whereas for the child, the objects that they used to produce the event actually do exist in the environment for the child to manipulate. Because the order in which the actions are supposed to be produced is not queued in the environment, we're going to be placing heavy emphasis on children's ability to produce action sequences in exactly the order that we demonstrated them.

The Ability to Remember at an Early Age
Now armed with this non-verbal analog to verbal report, my colleagues and I set out to determine what infants and young children are able to remember. I'd like to highlight the fact that when we began our inquiry, we really had no data. There was no information on what infants and young children can remember. Instead, what the field had was a really strong assumption about memory in young children. In fact, if you had stopped virtually any developmental psychologist on the street and said, "What do young children remember?" The likely response would have been this. The basis for this assumption was a strong theoretical model that predicted that there was some qualitative change in mental process that occurred somewhere between the age of two and three. Before the change, it was thought that infants and young children weren't able to think about anything that wasn't right in front of them. It was thought that they could think about the here and now, but that they couldn't think about anything that wasn't in the here and now. It was thought that they lived a real out of sight, out of mind existence. In contrast, after the change, it was thought that children could use verbal symbols to think and to reason and therefore to be able to recall the past. Notice how nicely this assumption of a qualitative change fit with the idea or the phenomenon of infantile amnesia. It made sense that infants and young children wouldn't be able to recall the past. After all, adults didn't have any memories from very early in life. That was consistent with the idea that children simply couldn't form memories. So that was the backdrop against which we began our research on infants and young children's memory for events that happened in the past.

Results of Experiments
Now what do all these data mean? After extensive experimentation, all of our data indicates that by early in the second year of life, children are able to remember specific events over periods of weeks and even months. Over short delays, children are able to remember events that they've experienced only one time and of which they have been reminded. Over longer delays, children's memory is facilitated by multiple experiences and by reminders. In some infants, the ability to remember over long delays already is apparent by late in the first year of life. In other infants it's not. This pattern might be indicative of individual differences in memory ability. What I think is more likely, however, is that it indicates different rates of development. That is, I think that we see these individual differences because at nine months children are in the middle of an important window in the development of the ability to recall over the long-term.

We now can use this information to reflect upon the different senses of amnesia that I mentioned at the beginning of my remarks. One possibility was that infantile amnesia was due to the loss of memories that once were available. Another possibility was that infantile amnesia was due to absence of memories due to an inability to create them. Well it may well be that prior to nine months of age, infants are not able to remember specific past events over any extended period of time. In this case, the culprit in infantile amnesia would be absence of memories. However, once we get into the second year of life, infants and young children can and do create memories of specific events and they're able to retain them over long time periods. That older children and adults aren't able to remember those memories suggests that those memories are lost or later become accessible over the course of development.

Where Do Childhood Memories Go?
Now this leads us back to the question that was posed in the title of this talk Ð Childhood Memories, Here Today, Where Tomorrow? In other words, what is the fate of memories that we know were here today. Where do those memories go tomorrow? Do those memories remain available or do they become inaccessible over the course of development? It's important to note that in imposing this question, what we're really asking is whether these early memories ever are going to be accessible to verbal report? After all, what characterizes infantile amnesia is an inability of adults to talk about early experiences. It's clear that our non-verbal message of testing children's memory is going to remain silent on the issue of whether they ever are able to express these memories verbally. The only way to address that question is to look at children's verbal reports. Catch 22. We've already seen that very young children have difficulty producing verbal reports. In fact, it's not until they are three years of age that they could reliably be counted upon to talk about past experiences. And even then, they only do so with a lot of prompting and support by an adult. Now what's interesting is that children's difficulty in talking about the past contrasts with the ability, their rapidly developing ability to talk about the present. The difference between talking about what you're doing right now and talking about what happened in the past is that right now, it's both physically and temporarily present. In the past, what happened back then is both physically and temporarily distant. The factor that we're interested in is temporal distance, not physical distance. So what do you suppose would happen if we were to eliminate the physical distance requirement? What do you suppose would happen if we put children back in a context in which they'd experienced an event and then asked them to tell us about it?

Testing Children's Recollections
An opportunity to ask that question is afforded by an on-going study in our laboratory. We currently are engaged in a prospecting study of one to two-year-old children's memory for specific past events. Children are enrolled in the study at 13, 16, or 20 months, and at that time we expose them to six events at each of three laboratory visits. Also, at the time that we expose the children to the event, we ask their parents to provide us with some information about the children's language development. We're testing the children's recall after intervals of either one, three, six, nine, or twelve months, and all of those are different groups of children. When we bring the children back into the laboratory after that delay, we test them on the old events that they saw previously and also on three new events.

As has been the case for all the research that I've talked about thus far, we originally were primarily interested in children's non-verbal demonstrations of memory. But what we found is that even though we weren't asking them to do so, the children were talking about the event. What Sandy Whowaka and I did was to characterize children's utterances of their verbalization as either memory-related or not memory-related. Memory-related verbalizations include questions or statements about non-obvious event-related actions.

Regardless of the delay after which we've tested these children they're showing us that they're able to remember the events. They're also telling us. Data shows the number of non-memory related verbalizations is actually a little bit higher on new events than it is on old events and that tells us that the children are perfectly willing to talk about the props or the objects whether they're familiar or not. In contrast, children produce a larger number of memory related verbalizations on the events that they've seen before. Thus, although we're not asking them to, the children are telling us that they remember the past. Thus, verbal evidence of memory for events by children in the one to two year age period is provided. They've also allowed us some insight into what determines whether memory will be verbally expressed.

How Language Development Affects Recollections
Now remember that we asked the parents to provide us some information about the child's language development. One of the things parents provide us is an index of the number of words that their children are producing, the number of words they're actually saying, and we're finding that as measured by parents reports, their productive vocabulary at the time they experience the event is related to their verbal but not to their non-verbal expression of memory. That is, language is not necessary for the children to remember. Even children with very low levels of language development were still remembering the events and they were showing us that they remembered. In contrast, children who had larger productive vocabularies at the time they experienced the event were also able to tell us that they remembered. This isn't just that children who talked a lot when they experienced the event came back and talked a lot at the time memory was tested. It's not just that talkative kids are talkative. Most of the kids when they came back were talking. However, it was only the children who had more language facility the time they experienced the event who were using their language to engage in memory talk. The other kids were just talking. I really want to stress that this doesn't mean that language is determining memory because what we see is that even children with very low levels of language development are able to show us that they remember. However, the availability of language at the time of experience of an event is related to later verbal expression of memory.

We're now finding that verbal expression of memory is related to later verbal expression of memory. The way we're finding that out is that we're inviting these same children to come back into the laboratory when they're between 36 and 40 months of age. And at that time what we're doing is actually eliciting from them verbal reports. We're being very mean to these children. We're getting these props out on the table, the props that they've seen before, but not letting the children interact with them. We're holding them back and we're saying, "Tell me about this. What was this one called? What did we do with this one?" So we're requiring the children to provide us a verbal report about how to produce the event. And what we're finding is the children are able to do so.

At this point, we've only analyzed the data from 12 of these children, all of whom were 20 months of age at the time that they originally experienced the event. What's really intriguing is that what determines children's elicited memory talk at age three years is their spontaneous memory talk from the previous testing session. In effect what we're finding is that greater language capacity at the time of experience of an event predicts early spontaneous memory talk. In turn, early spontaneous memory talk is what is predicting the maintenance of verbal accessibility at the age of three years. It's as if that early verbal expression preserved that memory and preserves it for later verbal expression. What we don't yet know but are very interested in finding out is whether once preserved, these memories would endure beyond the third year. Will it be the case that for the participants in our study some of the earliest verbally expressible memories are of making rattles and of making gongs?

In closing, when we first began our research on infants and young children's event memory almost a decade ago, the field really had very little data on what young children could remember. Rather than by an imperial body, the field was shaped by an influential and tenacious memory ability in these young children. The assumption was that children were not able to remember the events of their lives. That assumption held sway in large part because of a methodological impediment and that is that infants and young children were not able to participate in the task of choice, for most memory researchers that a verbal report. With the development of a non-verbal analog to verbal report, my colleagues and I were able to begin to test that assumption and we found it an unwarranted assumption. We've demonstrated that children in the one to two year age range are able to remember specific past events over periods of weeks and even months. What is more, we now have reason to believe that the ability to remember over the long-term is developing by the end of the first year of life. Not only are children able to show us that they remember, they're also able to tell us that they remember. In supportive context, children are providing verbal reports of memory when we ask them to do so and even when we don't ask them to do so. Early verbal expression of memory is related to parents estimates of children's productive language at the time they experience the events. Later verbal expression of memory is related to earlier verbal expression. Although we've yet to solve the mystery of infantile amnesia, we have gathered some critical clues. We now know that at least by the second year of life the sense of amnesia as loss rather than absence of memory more accurately characterizes the phenomenon. We also know that at least under some circumstances, loss is not inevitable. Some memories seemingly do survive and can be verbally described. Our task now is to establish the circumstances under which memories from early in life survive the transition from infancy to early childhood and become accessible to verbal report by older children and adults.