So, in general how good are we at predicting how long things will take? In a review of research on prospective memory published in 2010,15 Roger Buehler from Wilfrid Laurier University and his colleagues in Canada looked at research asking individuals to estimate how long particular activities would take them. They found that people were generally optimistic in their estimates, tending to discount past failures to complete things on time, and generally underestimating how long tasks actually took to complete. In other words, we seem to believe that our future selves are going to be superheroes at doing things quickly – new you excels at doing things quickly, even if old you was slow. New you is efficient, old you was lazy.
We can probably all remember times when we’ve done this ourselves, thinking we will get up early to go for a run, followed by an early breakfast, then finishing off that work by noon, having an effective lunch meeting, catching up on all our emails, visiting the dentist, going to yoga, cooking a five-course dinner, cleaning all the things, going out for drinks with friends, and then heading home for some incredible sex before having a restful night’s sleep, all in the same day. When has that kind of day ever happened? Yet, how many times do we sit there at night thinking ‘Yes, tomorrow is going to be that kind of day.’
Another reason we may be impossibly optimistic in this kind of time estimation is because we may remember how long each of these tasks takes when considered in isolation, but we forget how long it takes to task-switch and move from one task to another. Also, we may forget that generally after a period of exertion our cognitive resources are depleted and need to recharge before we can effectively begin another task. In other words we remember certain elements of how long things take, and discount others.
Further, according to Roger’s own research from 1994,16 and the review of the literature by him and his colleagues in 2010, we only think of ourselves as future-time superheroes. When estimating the efficiency of others we are actually comparatively pessimistic, overestimating how long it will take them to complete tasks, and often predicting that they will run into problems that will delay their completion of the task. Researchers in this area have found that this effect applies across different kinds of task estimation, suggesting that our prospective memory abilities as they relate to other people are equally bad in work or personal contexts – we overestimate how much time our friends are going to need to get work done, and how long it will take them to meet us for coffee.
In terms of our autobiographical memories, this means that every time you encode a memory of an event – which will invariably have time-related elements to do with duration and chronology as part of it – you are encoding it through the filter of how you feel, how much is happening that day, and various other biasing factors. Time is not objective, so it is open to the same subjective biases as everything else. And these initial biases, like many of the other perceptual biases covered in this chapter, colour our memories.
Through the telescope
Having an appreciation of time perception is also critical for understanding what is referred to as retrospective timing. This has to do with our sense of the duration of events after we have experienced them: estimating how long you were playing a video game, for example.
‘Intuitively (without thinking or counting), I have the impression that this game session lasted ___ minutes and ___ seconds.’ This was one of the questions that Simon Tobin and his research team from Laval University in Canada asked over 100 participants who were regular gamers, in research they published in 2010.17 They asked these gamers to come in and play a gaming session in the lab; they wanted to see how good these people were at estimating how long a session of gaming took. They found that when participants were asked shortly after playing, a 12-minute session was generally perceived as proportionally longer than a 35- or 58-minute session. That is to say, the participants were likely to overestimate how long the short session took by a ratio of 1.4:1, estimating that the 12-minute session actually took closer to 17 minutes. They were more accurate with the longer sessions – for 35- and 58-minute sessions many players got the timing almost exactly right. So our retrospective timing of shorter events is considerably less reliable.
Beyond our ability to remember how much time we wasted playing a computer game, our retrospective memory also helps us to remember our entire life timeline. It is useful for timing the duration of individual events, and the duration between events. It allows us to date an event autobiographically, so that we know the relative recency of an event in relation to today, be it earlier today or ten years ago. Retrospective memory allows mental time travel, enabling us to peek into our personal past.
One type of cue we use to help us date memories is known as ‘landmarks’. Landmarks are important events such as the assassination of JFK in 1963, 9/11 in 2001, or when Russia invaded the Ukraine in 2014, which people can use to orient themselves along their time continuum – it may be easier for us to chronologically place a personal event if we remember it in relation to these landmark events. For example, I went on vacation to Cuba after 9/11, but before the Ukraine invasion. This narrows down the time frame somewhat when we are talking about a lifetime of events. These landmarks can also be personal, using events such as a high-school graduation or a wedding as markers. So, instead of pinpointing in relation to historical events, you may remember you went to Cuba after graduation, but before you got married, for example.
Of course, our estimation of landmark events themselves is also fascinating, since we can make mistakes dating them as well. Because the timing of commonly held landmark events is far easier for researchers to verify than personal landmarks, they are a useful phenomenon for studying time estimation across different people. At least that is what a team led by George Gaskell from the London School of Economics thought.
In 200018 they published one of the largest investigations of its kind, involving over 2,000 participants who had taken part in a census survey in 1992 which had looked at how good the British general public was at estimating when historical landmark events occurred. They examined two events: Margaret Thatcher’s resignation as prime minister of Britain, which had occurred 19 months before the study, and the Hillsborough football disaster, which had occurred 37 months before the study. They wanted to see whether the public could date the events accurately, which they defined as dating the events within one month of their actual occurrence.
The results were fascinating. Overall, only 15 per cent accurately dated the resignation of Margaret Thatcher, and even fewer, around 10 per cent, accurately dated the football disaster. Performance was equally poor across all ages, so it did not seem to matter how old the participants were at the time the landmark events occurred. Instead of accurately dating the events, the overwhelming majority of participants engaged in what is referred to as temporal displacement, or ‘telescoping’, which means moving things around in time. We have a tendency to do this. In particular, we often remember things that happened more recently as having happened longer ago than they actually did. Conversely, we often remember things from long ago ‘as if it were yesterday’.
The telescoping study by Gaskell and his colleagues found that most people reported the event as happening more recently than it actually did (forward telescoping), so closer in time to today. A significant number, however, reported it as being earlier than it actually was (backward telescoping), so being longer ago. For the more recent event, Margaret Thatcher’s resignation, which had only happened 1.5 years before the interview, 40 per cent of participants forward-telescoped, while 31 per cent backward-telescoped. The opposite was found for the event that happened over 3 years before – with 29 per cent forward-telescoping and 42 per cent backward-telescoping. Although these effects have been shown in different configurations by other researchers, what they all agree on is that in the time estimation of our memory we have a tendency to push some events even further into the past than they really were, while others we draw closer to today.
Indeed, the bia
s seems to enter when a memory hits about three years. Things that happened less than three years ago we generally assume were less recent than they actually were, while things that happened more than three years ago feel more recent. While telescoping is due to a complex interplay of memory biases, one reason we might particularly think that things happened more recently than they actually did (forward telescoping) is because landmark memories are often very accessible. We can recall these important life events easily and with a lot of detail, just like memories of things that happened much more recently. We thus interpret this easy access and high vividness of the memory as indicating that it must have happened fairly recently.
Landmarks are an important mainstay of our memory timelines, yet they are riddled with predictable errors. They help us navigate our personal chronology, but it’s easy to see how they might prove somewhat unreliable anchors. So, not only are we bad at estimating how long things are going to take, and how long tasks just took, but our recollections can play tricks on us that move even important events up and down along the timelines of our lives.
The good old days
Let us think about that timeline of our lives a bit more. Time-travelling through our memories, we may find that some events stick out more than others. If we think about the characteristics that these memories have in common, we may notice that the most vivid are the most emotional, most important, most beautiful or most unexpected events of our lives. We may also notice that our memories cluster. And they often seem to cluster around particular periods in our lives.
This is a phenomenon called the reminiscence bump, and may help explain ‘the good old days’ and the ‘when I was your age’ comments. The reminiscence bump means that we do not remember all ages in our life equally. In 2005 a study involving 2,000 participants between the ages of 11 and 70 from the US and the Netherlands was conducted by a team lead by Steve Janssen at the University of Amsterdam.19 They wanted to answer the question ‘What remains in a lifetime of memories?’
Apparently what remains most are memories from between the ages of 10 and 30. The findings of the study supported what others had shown before them – that before the age of five most people report almost no memories. Then, between five and ten, the number of memories begins to increase, hitting a peak for both genders in the late teens. This period of increased reported memories stays quite high until the early twenties, when it begins to drop and then stabilise for the remaining decades. So we seem to retain the most memories of our teens and twenties.
This effect appears to be completely cross-cultural. A study conducted in 2005 by Martin Conway from Durham University in the UK20 found that it does not seem to matter whether you are from Japan, China, Bangladesh, England or the US – everyone has the same reminiscence bump.
But while the density of memories we recall from this critical time may be the same for all of us, there may be cultural differences in the nature of the memories we have for this time frame. Chinese participants showed a preference for memories that involved groups and events focused on social situations, while US participants generally remembered more self-centred events. Chinese participants reported more memories that focused on themes such as childbirth, interacting with neighbours and colleagues, and intimate relationships, while Americans reported more personal themes such as individual success, frustration, fear, or nightmares. So, our reminiscence bump is the same around the world but with slight variations in the content of our most memorable life experiences according to the culture we live in.
One explanation for the reminiscence bump may be that it is related to our emergence of a real sense of self, which seems to be a largely universal phenomenon. At what age did you form a stable identity? The chances are that if you are a woman, your you-ness first really shone through when you were between 13 and 14, and if you are a man, you probably settled into yourself a bit later, between the ages of 15 and 18. This age range also happens to be the peak of the bump, at least according Steve Janssen’s research team.
These are the memories that defined us. They are the memories that made us who we are. And, whether or not they have been tainted by perceptual and memory biases, they are the memories we seem to cherish and remember most vividly.
Yet, despite defining us and being so important for our identities, it seems clear that our memories can have inbuilt flaws as a result of the ways our perceptions can be fooled – by visual illusions, our level of arousal, and even from having a poor grasp on the seemingly intuitive ability to sense time. And this is just the tip of the iceberg. Every single one of our perceptual abilities is imperfect. Our vision, our hearing, our taste, our sense of heat, our tactile senses, our vestibular sense of balance, our proprioception of our body in space – every one of these can be fooled.
As the philosopher George Berkeley said, ‘esse est percipi’ – ‘to be is to be perceived’. Only our perception of reality matters. It means that our misperceptions of reality can be placed into our memory systems to be later recalled in spite of never accurately representing more objective reality. While they may generally be close enough to be functionally useful, the truth is that probably every single one of our memories – even the clearest – contains perceptual flaws and inaccuracies, right from the outset.
3. DANCING WITH BEES
Roofies, sea slugs, and laser beams
Why brain physiology can lead our memories astray
SO YOU WANT to read a book on memory but don’t want to hear too much about brain biology? You are not alone. Head on past this chapter then, please. I hereby give you a pass if you don’t want to find yourself knee-deep in animal studies, biochemistry, and the history of memory theory. You will still be able to understand the following chapters without any understanding of these topics. However, if you do like all of these science-y topics, which really demonstrate to you what memory is, keep reading. And if you have chosen to keep reading, let me introduce you to Kathryn Hunt.
Kathryn Hunt is covered from head to toe in a beekeeper’s suit. She is slowly making her way to a hive that is absolutely brimming with Bombus terrestris – bumblebees. There are hundreds of them, and she is starting to wonder why she signed up for this. They look cute, being one of the fluffier and rounder varieties of bees, and they are generally not aggressive, but walking towards hundreds of them one cannot help but experience a tinge of anxiety. This is a heart-stopping exercise. Is it even going to work?
Hunt need not have worried – the results of her study were groundbreaking. When she conducted it, she was an eager young researcher at Queen Mary University in London, looking at behavioural decision-making in animals, focusing on whether memory changes in a predictable way as it relates to certain behaviours, rather than just degrading steadily over time. She came from a background of researching how animals are affected by, and interact with, their environments. In this field, known as cognitive ecology, she had previously looked into the food preferences of guppies, and at foraging behaviour in leafcutter ants. It was not until she began to collaborate with Lars Chittka, a leading contemporary expert on bees, that her attention turned towards bumblebees.
Hunt and Chittka wanted to see if they could create false memories in bees. Their research, published in 20151 in the journal Current Biology, was intended as a demonstration of some of the fundamental physiological processes that are thought to underlie false memory. Bees have a highly evolved social life, a complex communication system, and exceptional ability to learn new information. Their memory seems to work in a broadly similar way to our own, making a study of it also relevant to the science of human memory.
The particular type of bumblebee they used, the aforementioned Bombus terrestris, was already known to have a highly developed memory for colours and patterns, including an excellent memory for specific flowers. It had also been shown to have a particularly good capacity for remembering multiple things at once. In short, it has exceptional bee-memory. The researchers wanted to see whether this exceptional memory co
uld be a tricked into becoming a disadvantage for the bees rather than an advantage.
The researchers presented the bees in the study with two unique flowers, one at a time, both of which contained delicious nectar – a flower with black and white rings, and a colourful yellow flower. The bees would therefore have learned to associate these flowers with the nectar they were looking for. During a subsequent test, the bees were given a choice of three flowers: the one with black and white rings, the yellow one, and a new flower that was yellow with black rings. When the bees were given this test minutes after they had encountered the original two flowers they rarely went for the new combined flower, instead correctly picking the flower that had most recently provided them with nectar. They displayed accurate short-term memory.
However, when tested a day or three days later, some of the bees developed a preference for the new ‘merged’ flower – even though they had never actually seen it during their training, only during the test, and it had never given them nectar. By the end of the study about half of the bees preferred the merged flower over the correct one.
At this point some people might have concluded that the errors in performance meant that the bees had failed to learn which flowers reliably contained nectar, or perhaps had learned and then forgotten. But if the bees had simply forgotten which flowers gave nectar and which did not, they should have selected each of the three flowers presented during the test equally often, rather than showing a marked preference. Instead, Hunt and Chittka interpreted the bees’ behaviour as indicative of a false memory having developed – the bees had mixed up their memory of one flower with their memory of the other, leading them to ultimately prefer the mixed flower because it featured both the features they had come to associate with nectar, rather than just one. They had created a merged false memory, and this muddled memory had real behavioural consequences.
The Memory Illusion Page 6
