Can science fully explain what we mean by the self, identity and human values, or are we just the victims of neuro-scientific hubris? Don't worry, there's no need to choose sides, says Julian Baggini

The self: why science is not enough

Can science fully explain what we mean by the self, identity and human values, or are we just the victims of neuro-scientific hubris? Don't worry, there's no need to choose sides, says Julian Baggini

The nature of the self, identity, and human values used to be the preserve of philosophers, but over recent decades psychologists and neuroscientists seem to have thoroughly colonised the territory.

For instance, in the 1960s Roger Sperry and Michael Gazzaniga famously severed the corpus callosum in several people with epilepsy and found that the left hemisphere of the brain could be aware of things the right was not aware of, and vice versa. Then in the 1970s, Benjamin Libet discovered that certain bodily movements were activated in the brain before the person consciously decided to do them, challenging conventional notions of free will.

A decade on, and Michael Persinger brought religious experiences into the domain of neuroscience by inducing them in subjects using transcranial magnetic stimulation. Moving into the 21st century, and neuroscientists such as Todd E. Feinberg and Antonio Damasio continue to use research into the brain to shed light on how our sense of self is created and sustained.

In contrast, my book on self and identity, The Ego Trick, contains just one chapter on the science of the brain. Why? Because I would argue that on the big questions about who we are, recent research has told us a great deal about the physical basis for the emergence of the sense of self, but next to nothing about what a self actually is.

Take the question of what the seat of the self is. For millennia, in the absence of any real understanding of consciousness, it seemed credible to believe that thought required some kind of immaterial soul. The big contribution of empirical science has been to identify the brain, working with the central nervous system and to some extent the whole body, as the main organ responsible for consciousness.

But this very general idea is hardly cutting edge. As long ago as 1664, Thomas Willis published Cerebri Anatome, a detailed attempt to explain how different parts of the brain produced the different "animal spirits" that were believed to power thought and action. More importantly, even before the brain's role in consciousness was fully appreciated, philosophers such as the British empiricists John Locke and David Hume had already worked out that what you are isn't a question of the stuff you are made out of anyway, be it spirit or matter. What makes you the same person over time is, broadly, the continuity of your mental life. The continuity of the same brain in the same body matters only in so far as it makes this possible. As cognitive scientist Douglas Hofstadter memorably put it, "it ain't the meat, it's the motion". We knew this before neuroscience peered into the brain and discovered what makes that motion happen.

But what about the much finer-grained work of recent years, pinpointing exactly which parts of the brain are responsible for the various aspects of consciousness? This is clearly extremely interesting and clinically useful, but philosophically speaking it really only filled in the details and hammered the last nails into the coffins of antiquated views of soul and self. Neuroscientists, for example, agree there is no place in the brain where "it all comes together", no locus of the self in one part of the cerebrum. A sense of self turns out to be something that emerges as the result of most parts of the brain working together.

Using Feinberg's model, the self is a "nested hierarchy". This means that the higher functions of self - self-consciousness, for example - are not independent of the lower functions, like the basic awareness of one's environment, but incorporate and depend on them. So the higher functions of the evolutionarily newest part of the brain, the cerebral cortex, require the more primitive instinctive and emotional functions of the limbic system (amygdala, hippocampus and hypothalamus) and the automatic bodily regulation functions of our "reptilian brains" (the brain stem and cerebellum).

While the detail Feinberg adds to our understanding is scientifically invaluable, in philosophical terms this is no astonishing discovery, just a confirmation of what Hume thought over 200 years ago: there is no thing which is you at all. Each of us is merely a bundle of thoughts, sensations and experiences.

Pretty much the same view was held by the Buddha, who believed that there is no abiding self, just a series of connected conscious experiences. Neuroscience confirms this and explains the mechanics of this centreless self, but it certainly didn't discover it.

What about free will? Surely neuroscience has taught us some serious facts about that? It is a popular view, most recently echoed by Sam Harris in The Moral Landscape, where he cites experiments by Libet and others which show that activity in the brain's motor regions can be detected 350 milliseconds before a person is aware of deciding to move, and that some decisions can be predicted up to 10 seconds before people are aware of having made them. "All our behaviour can be traced to biological events about which we have no conscious knowledge," writes Harris. "This has always suggested that free will is an illusion."

But even here, the philosophers were way ahead of the game. It has been obvious for centuries that if human beings are made entirely of physical stuff, and all physical stuff simply follows the laws of physics, then there is no room in human action for any causal power other than the motion of matter. So to go back to Libet, whether consciousness comes before, after or simultaneously with decision-making, the key point is that thoughts and decisions are produced by no more than a brain working according to the laws of physics.

What else could a decision be but the product of a combination of the present state of being, fashioned by the past, and the environment that a person finds themself in? From at least Hume onwards, many philosophers have understood that the only meaningful sense of free will is action free from coercion or force, not action exempt from the causal necessity of the physical world. To that debate, neuroscience adds nothing.

I don't wish to disparage neuroscience. On the contrary, I am in awe of what is being discovered about the mechanics of mind. But it is simply a philosophical mistake to think that understanding more about the nuts and bolts of the basis of self and identity must add something to our fundamental understanding of what makes us the individuals we are. Some scientists agree. "I don't think the self is ultimately a scientifically tractable question," clinical neuropyschologist Paul Broks told me while I was writing The Ego Trick.

The main reason is that the very notion of a science of the self depends on us identifying its subject - the self - from the perspective of first-person experience. Science can correct false beliefs about what sustains that experience, and it can explain what makes such experience possible, but it cannot change what it means to be a self without erasing the very data it depends on.

Neuroscience may explain the mechanics of the "centreless self" but it didn't discover it

Julian Baggini is a philosopher, with a PhD from University College London on the philosophy of personal identity. Among his books are The Pig that Wants to be Eaten, and The Ego Trick (Granta Books, March) - on which this essay is based. He is co-founder and editor-in-chief of The Philosophers' Magazine.

to go …..

Dreams don't reveal your secrets and desires they are far more important than that. Emma Young reports

The interpretation of nightmares

The interpretation of dreams is the royal road to a knowledge of the unconscious activities of the mind." So wrote Sigmund Freud in his 1900 classic The Interpretation of Dreams. He saw this idea as a "once in a lifetime" insight, and for much of the 20th century the world agreed. Across the globe, and upon countless psychoanalysts' couches, people recounted their dreams in the belief that they contained coded messages about repressed desires. Dreams were no longer supernatural communications or divine interventions - they were windows into the hidden self.

Today we interpret dreams quite differently, and use far more advanced techniques than simply writing down people's recollections. In sleep laboratories, dream researchers hook up volunteers to EEGs and fMRI scanners and awaken them mid-dream to record what they were dreaming. Still tainted by association with psychoanalysis, it is not a field for the faint-hearted. "To say you're going to study dreams is almost academic suicide," says Matt Walker at the University of California, Berkeley. Nevertheless, what researchers are finding will make you see your dreams in a whole new light.

Modern neuroscience has pushed Freud's ideas to the sidelines and has taught us something far more profound about dreaming. We now know that this peculiar form of consciousness is crucial to making us who we are. Dreams help us to consolidate our memories, make sense of our myriad experiences and keep our emotions in check.

Changing patterns of electrical activity tell us that the sleeping brain follows 90-minute cycles, each consisting of five stages - two of light sleep at the start, then two of deep sleep, followed by a stage of REM, or rapid eye movement sleep (see diagram, page 39). There is no characteristic pattern of brain activity corresponding to dreaming, but as far as we know all healthy people do it. And while dreaming is commonly associated with REM sleep, during which it occurs almost all of the time, researchers have known since the late 1960s that it can also occur in non-REM sleep - though these dreams are different. Non-REM dreams tend to be sparse and more thought-like, often without the complexity, length and vivid hallucinatory quality of REM dreams.

Despite their differences, both types of dreams seem to hold a mirror to our waking lives. Dreams often reflect recent learning experiences and this is particularly true at the start of a night's sleep, when non-REM dreaming is very common. Someone who has just been playing a skiing arcade game may dream of skiing, for example (Sleep, vol 33, p 59). The link between waking experience and non-REM sleep has also been observed in brain scanning studies. Pierre Maquet at the University of Liège, Belgium, looked at the later stages of non-REM sleep and found that the brains of volunteers replayed the same patterns of neural activity that had earlier been elicited by waking experiences (Neuron, vol 444, p 535). Many REM-sleep dreams also reflect elements of experiences from the preceding day, but the connection is often more tenuous - so someone who has been playing a skiing game might dream of rushing through a forest or falling down a hill.

Sleep on it

But we do not simply replay events while we dream, we also process them, consolidating memories and integrating information for future use. Robert Stickgold of Harvard Medical School in Boston recently found that people who had non-REM dreams about a problem he had asked them to tackle subsequently performed better on it (Current Biology, vol 20, p 1). Likewise, REM sleep has been linked with improved abilities on video games and visual perception tasks, and in extracting meaning from a mass of information (Neurobiology of Learning and Memory, vol 92, p 237).

"It's clear that the brain does an immense amount of memory processing while we sleep - and it certainly isn't mere coincidence that while our brain is sorting out these memories and how they fit together, we're dreaming," says Stickgold. He suspects that the two types of dream states have different functions for memory, though what these functions are is a matter of debate. Non-REM dreaming might be more important for stabilising and strengthening memories, Stickgold suggests, while REM dreaming reorganises the way a memory is stored in the brain, allowing you to compare and integrate a new experience with older ones.

Jan Born and Susanne Diekelmann at the University of Lübeck in Germany, however, have looked at the same evidence and come to the opposite conclusion - that REM sleep supports the strengthening of a new memory, while non-REM sleep is for higher-level consolidation of memories (Nature Reviews Neuroscience, vol 11, p 114). "I think this means that we're still lost when it comes to understanding the role of different sleep stages in memory," says Stickgold.

Also unclear is how central is the role of dreams in memory formation. During dreaming is certainly not the only time our brains consolidate memories. For example, when we daydream certain areas of the brain, called the default network, become active. We now know this network is involved in memory processing (New Scientist, 8 November 2008, p 28) and many of the same brain regions are active during REM sleep. What's more, daydreaming, like REM dreaming, can improve our ability to extract meaning from information and to have creative insights.

Does this mean we don't actually need dream sleep to process memories? Not necessarily, says Walker, who points out that the way new memories are replayed in the brain is different in daydreaming and dreaming. Rat studies show that the reruns happen in reverse when the animals are awake and forwards when they are sleeping. No one is quite sure what this difference means for memory processing, but Walker believes it shows that daydreaming is not simply a diluted version of sleep dreaming. Maquet agrees. "Different brain states may all have somewhat different functions for memory. Memory consolidation is probably organised in a cascade of cellular events that have to occur serially," he says - some while you are awake, and then some while you are asleep.

Even if dreaming is crucial for memory, Walker for one does not see this as its main function. "I think the evidence is mounting in favour of dream sleep acting as an emotional homeostasis: basically, rebalancing the emotional compass in a good way at the biological level," he says. Everyone knows how a short nap can transform a cantankerous 2-year-old and Walker has shown something similar in adults. He found that a nap that includes REM dreaming mitigates a normal tendency in adults to become more sensitive to angry or fearful faces over the course of a day, and makes people more receptive to happy faces (Cerebral Cortex, vol 21, p 115).

Walker has also found that sleep, and REM sleep in particular, strengthens negative emotional memories (Cerebral Cortex, vol 19, p 1158). This might sound like a bad thing -- but if you don't remember bad experiences you cannot learn from them. In addition, both he and Stickgold think that reliving the upsetting experience in the absence of the hormonal rush that accompanied the actual event helps to strip the emotion from the memory, making it feel less raw as time goes on. So although dreams can be highly emotional, Walker believes they gradually erode the emotional edges of memories. In this way, REM dreams act as a kind of balm for the brain, he says. In people with post-traumatic stress disorder this emotion-stripping process seems to fail for some reason, so that traumatic memories are recalled in all their emotional detail - with crippling psychological results (New Scientist, 21 February 2009, p 34).

As with memory processing, REM and non-REM dreaming may play different psychological roles. Patrick McNamara of Boston University has found that people woken at different sleep stages give different reports of their dreams. REM dreams contain more emotion, more aggression and more unknown characters, he says, while non-REM dreams are more likely to involve friendly encounters (Psychological Science, vol 16, p 130). This has led him to speculate that non-REM dreams help us practise friendly encounters while REM dreams help us to rehearse threats (see "The interpretation of nightmares", left).

So what do they mean?

All this suggests that we couldn't function properly without dreaming, but it doesn't answer the perennially intriguing question: what do dreams actually mean?

For some sleep researchers the answer is simple - and disappointing. Born argues that dreams themselves have no meaning, they are just an epiphenomenon, or side effect, of brain activity going on during sleep, and it is this underlying neuronal activity, rather than the actual dreams, that is important. Walker finds it hard to disagree. "I don't want to believe it. But I don't see large amounts of evidence to support the idea [that dreams themselves are significant]," he says.

Those researchers who refuse to accept the notion that the content of dreams is unimportant point to work by Rosalind Cartwright of Rush University, Chicago. In a long series of studies starting in the 1960s, she followed people who had gone through divorces, separations and bereavements. Those who dreamed most about these events later coped better, suggesting that their dreams had helped. "Cartwright's work provides some of the most solid evidence that dreaming serves a function," says Erin Wamsley at Harvard Medical School. There is no hard data showing that dreaming is not an epiphenomenon, she admits, but the same could be said about waking consciousness.

In fact, Wamsley's own research hints that the form and function of a dream are connected. She worked with Stickgold on the study which found that non-REM dreams boost people's performance on a problem. Their volunteers were given an hour's training on a complex maze then either allowed a 90-minute nap or kept awake. The dreamers subsequently showed bigger improvements, but the biggest gains of all were in people who dreamed about the maze. It didn't seem to matter that the content of these dreams was obtuse. One volunteer, for example, reported dreaming about the maze with people at checkpoints - though there were no people or checkpoints in the real task - and then about bat caves that he had visited a few years earlier. Stickgold didn't expect this to improve the volunteer's ability to navigate the maze, "and yet this person got phenomenally better".

He points out that the dream content is consistent with the idea that during dreaming memories are filed with other past experiences for future reference. "Dreams have to be connected in a meaningful, functional way to improvements in memory - not just be an epiphenomenon," he says. "I say this with fervent emotion, which is what I use when I don't have hard data."

Such evidence may one day be forthcoming, though. In the past there has been no objective way to record what someone is dreaming, but that could change. Yukiyasu Kamitani at the ATR Brain Information Communication Research Laboratory in Kyoto, Japan, and colleagues have used fMRI scans to recreate scenes that volunteers were picturing in their mind while awake (newscientist.com/article/dn16267). The team hasn't yet done this with dreams, but it is theoretically possible, says Kamitani. It wouldn't be like watching a movie, he adds, "but it may be possible to predict what kind of dream a person is experiencing using currently available technology".

Some may think all this peering and prodding at our dream world is taking away its magic but the researchers don't see it that way. While you are dreaming your brain literally reshapes itself by rewiring and strengthening connections between neurons. So although dreams do not reveal the secret you, they do play a key role in making you who you are. "The mystery and the wonder of dreams is untouched by the science," says Stickgold. "It just helps us appreciate better how amazing they really are."

The re-interpretation of dreams

Sigmund Freud was wrong about dreams revealing your secrets and desires. It turns out they are far more important than that

Memory traces from our waking experiences are replayed in our dreams

"While you are dreaming your brain is literally reshaping itself, so dreams play a key role in making you who you are"

A strange form of consciousness we cannot live without

Antii Revonsuo enjoys his nightmares. "At least in hindsight," he qualifies, "as though they were good horror movies where you don't know it's a movie until it's over." But then Revonsuo, at the University of Turku in Finland, thinks that nightmares are the main biological reason for why we dream - they allow us to simulate scary encounters, and so be better prepared for them in our waking life.

"The theory predicts correctly several features of our dream content," says Revonsuo. For example, he and his colleagues have found that about two-thirds of the dreams of healthy adults involve at least one threat. About 40 per cent of these take the form of aggressive encounters - running away from an attacker or getting into a fight. Such encounters are higher among children, accounting for over half of threat dreams in Finnish kids and three-quarters among traumatised Palestinian children.

Revonsuo argues that children's dreams are closer to our evolutionarily original form of dreaming because children haven't yet had a chance to adjust to the modern environment. He has found that between 40 and 50 per cent of children's dreams contain animal characters, often as enemies, which is similar to the instance among adult hunter-gatherers. The figure is just 5 per cent in western adults. "I don't think any other dream theory has made such specific predictions and shown that they hold," he says.

It is a neat idea, but Robert Stickgold at Harvard Medical School in Boston cannot believe that's all there is to dreaming. "I think Revonsuo has made the same mistake as Freud - which is to limit dreaming's functionality. I think dreaming is absolutely about threat rehearsal some of the time. But it's absolutely about other things, too."

Emma Young, KD

Emma Young is a writer based in Sydney, Australia