UNDER the influence of sociology, some philosophers have argued that science
is culturally determined, and there is no such thing as truth. The psychiatrist
Leslie Brothers wants to have it both ways. She believes in scientific truths,
independent of cultural consent, but argues that our minds are developed and
shaped by our social interactions. So if our perception is culturally
determined, where does that leave objective truth? Moreover, without social
interactions we cannot have a “self” or any feeling of “personhood”—a
hypothesis a little harsh on hermits.

To demonstrate the effect of social interactions on the mind, Brothers
reviews the evidence for innate neural responses to different social phenomena,
such as neurons that “fire” only when detecting a face, or a person walking in a
particular direction, or a specific facial expression. She points out that
damage to certain regions of the brain can interfere with the “self”: those
afflicted may have the delusion that someone else has taken over their self and
imposed their own personality on it.

In Friday’s Footprints Brothers argues that emotions are nothing but
physiological arousal and a disposition to behave to others in certain ways that
are revealed by facial expressions. But some emotions, she points out, have
little to do with other people—frustration, for example, can occur merely
because of the cussedness of inanimate objects. Brothers also queries the
validity of mental terms such as willpower, motives and reasoning. But
scientific advance itself has often depended on inferring the existence of
unseen entities ranging from genes to quarks.

Despite its half-acknowledged return to behaviourism, this is a thoughtful,
if difficult, book. And prophetic: it claims that if 17th-century
neuroscientists had had access to brain scanners, they would have worked on
“godly behaviour and prayer”. Ironically, 20th-century neuroscientists have
just discovered a brain region that underlies religious belief.

O’Connell interprets this scene as more than just “animal behaviour”,
suggesting that the peacekeeper must have a theory of mind—that is, an
understanding of others’ intentions, goals and so on. Perhaps; and again,
perhaps not, for the elderly female’s desire to sleep in peace is enough to
explain her actions.

In her attempt to show that chimpanzees have a theory of mind, O’Connell
relies on too many such doubtful anecdotes. She even assumes that chimpanzees
can learn American sign language, although there is no good evidence for
this.

O’Connell is on sounder ground when she is discussing young children. She
describes the experiments that reveal the age at which they master different
aspects of the mind. It seems that children cannot understand false beliefs
until four or five years old, and cannot recognise a lie until six. It would be
more interesting to know how, rather than when, such mental concepts evolve.

Theory of mind is very much Ă  la mode, even in artificial
intelligence, the only discipline whose workers publish not what they have
achieved but what they hope to achieve. O’Connell piously records the
aspirations of AI devotees to write programs that understand minds.

Despite its naivety, the book is for the most part well written and should
appeal to anyone who thinks infants and chimpanzees are human.

THE road to large-scale computerisation is strewn with banana skins.
According to Tony Collins, 60 per cent of projects fail. He analyses the
reasons, drawing on fascinating case histories.

For the commissioning organisation’s ambition to have the all-singing,
all-dancing system with which to impress the public and steal a march on rivals
can prove disastrous. It was partly for this reason that the Wessex Health
Authority spent about ÂŁ50 million on a system, only a tiny part of which
was ever used. The project was promoted by the then Minister of Health and by
the health authority managers, all of whom wanted a share of the glory. It
became unstoppable, since nobody would report bad news for fear of losing their
jobs. The whole project ran hopelessly out of control until a new information
systems director took charge. She promptly accused some suppliers of fraud: they
had, for example, charged the health authority for a shower for the use of staff
who jogged. Collins, a quietly sarcastic man, writes: “Her hard line against
suppliers so impressed Wessex that it did what governmental bodies tend to do
when confronted with a conscientious, public spirited, apolitical, dynamic
troubleshooter. It made her redundant.” Two senior managers involved in the
project were criticised by the district auditor and subsequently received their
just rewards. They were knighted.

Why has ÂŁ300 billion been spent worldwide on failure? Collins suggests
that much of the blame should rest with the customers, whose requirements may be
too vague—or change during the project, or whose schemes may be just too
complex for the end-user (and for the programmers). He advocates three ways of
dealing with these problems. Before the project design begins, the staff who
will use the terminals must be consulted and persuaded to back the system,
otherwise they may be unable to use it and may end up sabotaging it. Secondly,
staff from the supplier of the computer system should work in the firm so they
can build a realistic picture of what goes on. Finally, the procedures in use by
the organisation should be simplified as much as possible. A sloppy or complex
system can be managed by people but not by computer: people can retrieve a
document filed wrongly, but a computer program cannot recover from a similar
mistake.

And there are other recurring problems: when a system first goes live, it
often becomes overloaded and crashes. This can be the result of end-user error,
as illustrated by the attempt to computerise the dispatch of ambulances in
London. The system was designed to despatch the ambulance nearest to the
patient, so the crew pressed buttons to transmit their location to the computer.
Unfortunately, they often pressed the wrong buttons or took a different
ambulance from the one they had logged into. The computer lost track of where
the ambulances were, and was eventually swamped by the backlog of unattended
calls. The London Ambulance Service reverted to a semi-manual system.

Among Collins’s other recommendations are: don’t use a recondite computer
language or you may be dependent on a single firm for maintenance. Instead, go
for off-the-peg systems wherever possible—they will be well tried. And
never fall for the sunk-costs error—ploughing more money into a project
just because a large sum has already been invested.

The picture Collins paints of British managers is appalling: for the most
part they appear to be lazy, ignorant, gullible and self-seeking. Crash
is essential reading for any organisation thinking of investing in a computer
system. It should also appeal to anyone who enjoys the spectacle of the great
and the good coming to grief.

Those responsible for the more exotic aspects of computing are all intellectuals, and perhaps this is why they believe that intelligence is all that is needed to have a mind and be conscious. They ignore two problems. First, people are not merely intelligent, they experience the world through sensations and feelings. Red, for example, is not just a wavelength, it is a colour, with unique connotations for each of us, perhaps through its associations with fire, flushed cheeks and so on. The sensation of red cannot be conveyed without referring to these associations.

Furthermore, I know perfectly well what the colour red looks like to me, but if others saw red the way I see green and vice versa, they could not communicate the difference (the qualia problem). The quality of sensation is thus ineffable, and could not be captured by any artificial device because we could not put into words or computer language what we are experiencing.

Secondly, computer programs cannot have biological drives. Because they could never need food, sex or love, nor have occasion to be angry, programs could never be conscious in the sense that people are.

Stan Franklin does not attempt to meet these arguments in Artificial Minds, but tentatively affirms that computers could be conscious. He points out that they already perform some intellectual tasks better than human experts – such as playing chess or diagnosing certain kinds of disease. But they work in a very different way – in chess by looking more moves ahead than is humanly possible and in diagnosis by performing the statistics linking symptoms to disease more rapidly and more accurately than physicians.

Artificial Minds tackles three of the outstanding debates in AI, of which the possibility of machines being conscious is only the first. The second concerns the relative merits of connectionism (a novel form of AI) and conventional AI. Until the 1980s, all AI programs were governed by explicit rules and contained symbols that represented concepts (such as objects or relationships) and explicit instructions. The instructions enabled programs to transform the symbols, for example by moving knight to king’s bishop three on a chessboard. Such programs were quite good at solving well defined problems. However, each problem requires its own program. And, because the operations in symbolic AI are serial, lengthy searches are required to locate the information needed at any particular stage.

By contrast, connectionist systems contain no explicit symbols or rules. The commonest kind are known as neural nets, reflecting their structure. They consist of layers of units, each layer connected to the next at random. The first layer is triggered by an input and the last layer represents output. The intermediate layer or layers are called “hidden units”. The connections between layers are triggered in parallel, and their strengths increased or decreased with successive inputs. Input patterns produce changes in different and widely separated connections. You can also teach these systems to give specific outputs to specific inputs.

Remarkable capacities have been found in connectionist systems. If they have learnt to “recognise” a face they will continue to recognise it even if part of it is hidden – the visible parts are able to stimulate many of the same units as the original. And just as the human brain does not break down completely if some nerve cells are damaged, the performance of a connectionist net only deteriorates slightly if some units are destroyed. Like the brain, the system works in parallel and this obviates the need for serial searching. Moreover, the systems are sloppy. Like people, they make mistakes and adding new information may cause partial “forgetting” of older information.

The advantage of the symbolic approach to programming is that it appears to conform more closely to conscious thought, which is serial not parallel. But it turns out that human thought is not entirely serial: people can juggle only six or seven related concepts at any one time, a limit close to that of connectionist AI systems.

So while serial AI cannot juggle ideas, connectionist AI cannot learn to execute tasks that involve rules, such as mathematics. The solution is to introduce some parallel processing into symbol-based systems. But I think there is still a flaw: how could even a combined AI system have a sudden “change of mind” like a human?

Although Franklin comes down on the side of connectionism, he does not meet these objections. It is likely that both types of process are at work in the brain. Moreover, it is certain that our own understanding of the brain must be based on symbols and higher level rules. We could no more understand the working of our 100 billion nerve cells in terms of their connections than we could describe a molecule in terms of quarks.

Franklin introduces a third debate, which I found unintelligible. He asks whether the brain or an intelligent machine must always contain representations of the outside world. His arguments against representation are utterly obscure. Here is one witness he quotes as saying: “parallel activity producers … interface directly to the world through perception and action, rather than interface to each other particularly much.” Note that final qualification. If visual images, smells, sounds and memories are not representations, what are they?

In considering the three debates, Franklin introduces us to other problems, focusing particularly on how to implement common sense. Let’s take a simple situation, in which I experience the thought: “I want a beer.” An AI system might respond by referring to a simple rule written earlier which states: “If you want a beer, order one at the bar.” But, as Franklin points out, behaviour is governed not by rules but by common sense, described in AI terms as “soft constraints”. Off I go to the bar to have a beer. On entering I find the bar is too noisy or too full. There is a corpse on the floor – or my favourite beer has run out. Any of these eventualities and an unquantifiable number of others would be sufficient to make me turn back, breaking the rule: “If you want a beer, order one at the bar.” Franklin agrees that it would be impossible in symbolic AI to search through all these conditions, one of which – that corpse – has never occurred before in my experience. Maybe connectionism could handle the problem, but it has not yet done so.

Following Marvin Minsky, Franklin suggests that the brain consists of separate modules, each of which performs a specialised task. In one version of this idea, the modules are called “demons” and the one that “shouts” the loudest (is most strongly turned on) determines behaviour. Such a specialised architecture might overcome some of the difficulties in implementing common sense, but it is hard to see how it could be learnt, while the problem of interfacing the modules to one another is also formidable. On the other hand, the brain itself is known to have special modules devoted, for instance, to colour vision, aspects of language and even pleasure.

In Artificial Minds Franklin gives a clear and comprehensive survey of the problems confronting AI. He is less good at outlining the proposed solutions. Despite describing programs, he often fails to tell us what they do, and sometimes even leaves us in doubt whether the ideas he presents have actually been implemented. Only rarely does he acknowledge the programs, limitations or use examples to show how they perform a specific task. He concentrates almost entirely on recent work, which despite his laudable enthusiasm, gave me the impression that the subject has not advanced much over the past few years.

Hobson adopts the familiar hypothesis that sleep and wakefulness are governed by the balance of neurotransmitters active in the brain. When the biologically active amines (particularly norephinephrine, a close relative of adrenaline) predominate, we are awake. When their levels and those of acetylcholine are balanced, we fall into deep sleep, but as acetylcholine levels rise, subduing the amines, the sleeper begins to move his eyes rapidly (rapid eye movement sleep or REM) and experiences vivid dreams.

Such dreaming has, according to Hobson, three purposes. First, the activity of acetylcholine, which is believed by some to be involved in memorising, consolidates storage of the previous day’s experiences. Second, it allows people to generalise that experience by connecting it with other events in the flighty and illogical fashion that characterises dreams. Third, since the amines are quiescent, they regenerate, which in turn prevents hallucinations while awake caused by underactivity of the amines.

Although this story cannot be substantiated, it is true that memory is impaired in both rats and people when they are deprived of REM sleep. Moreover, people who have been kept awake for several days (the record is 13) do indeed suffer hallucinations, a phenomenon that occurs, according to Hobson’s account, because their amines are depleted.

On the other hand, it seems unlikely that memory can be reorganised in any useful way through REM sleep. In one of his many distracting autobiographical asides, Hobson argues that he was able to fight off three muggers because he had been practising such fights in his dreams. But knowledge of results is usually needed for learning to be useful. In Hobson’s case, the missing datum was whether his method of fighting would be effective.

Hobson compares the chaotic nature of dreams to delirium or the hallucinations of a schizophrenic, but fails to note that schizophrenia is thought to be caused by the excess activity of one of the amine systems (dopamine) not by underactivity, as his hypothesis demands.

Like all good neuroscientists Hobson feels he must label his model with an acronym. He plumps for AIM, where A is for “Activity” in the cortex (which occurs in wakefulness and dreams), I is for “Information source” (external when awake, internal when asleep), and M is for “Mode” (whether the person is awake or asleep). Unfortunately AIM is too vague to be a model of anything, though it does summarise the author’s basic ideas.

Much of The Chemistry of Conscious States is speculation, made no more convincing by the facetiousness with which it is expressed. Take the following passage: “As for the purpose of this flimflam, we imagine that our aminergic brain cells might benefit from their time off. Perhaps they are able to restore their chemical power by building new molecules and not having to use them. At the same time, our [acetylcholine] neurons are having a field day …”

Neuroscientists have made some important discoveries. They have more or less found out how nerve cells conduct electrical impulses, they have some idea of the incredibly complex method by which one nerve cell fires another at their junction (the synapse) and they have discovered about forty neurotransmitters. But the precise functions of the different neuro-transmitters continue to elude them. And even if they had managed to demonstrate unequivocally that defects in one neuro-transmitter system cause hallucinations, in another depression and in another mania (which they have not) they would hardly have begun to bridge the gap between the action of these molecules and the resulting behaviour and states of mind.

Hobson is perhaps dimly aware of this problem. Instead of talking about the “brain” or the “mind”, he uses the expression “brain-mind” in a vain attempt to conceal the gap. With the arrogance of so many brain scientists, he writes: “Knowing the shared features of all kinds of delirium and their common underlying cause is a huge payoff from the brain-mind paradigm.” But this connection could exist whatever view one takes of the relationship between mind and brain: Hobson’s definition of the two leaves unsolved the way in which too little amine activity produces inchoate thinking, if indeed it does.

Although Hobson’s ideas about the causes of REM sleep and of hallucinations are interesting, if unproven, most readers will find them hard to extract from The Chemistry of Conscious States, which in its disjointed presentation bears some resemblance to a dream. As Bret Harte, the 19th century American writer, remarked: “Do I sleep, do I dream or is visions about?”

Although the paramount importance of genes in shaping personality, talent, intelligence and susceptibility to mental illness is well established, those who back nurture over nature – the environmentalists as opposed to the geneticists – have not changed their spots. Moreover, there are few popular accounts of these findings: to the extent that Out of the Blue presents a clear summary of some of the evidence, it should prove a useful book. Despite its subtitle, it is more about hereditary causation than about depression, though it opens with a good account of what it feels like to be depressed, rightly insisting that depression is a thoroughly degrading experience that ennobles no one.

The predisposition to mental illness is certainly genetic. For example, if one identical twin has manic depression, in 70 per cent of cases the other twin will also be affected. The concordance for schizophrenia is lower, but this does not mean that it is parents who induce this disorder in their children. It is likely that defects in the prenatal environment play a role, for the schizophrenic member of a pair of discordant twins often has physical deformities (particularly in the hands) acquired in the womb and not shared by the other twin. Moreover, in the case of identical twins who are discordant for schizophrenia, the child of the healthy twin is as likely to be schizophrenic as the child of the schizophrenic twin. These findings suggest that parental upbringing has little to do with this disorder.

Recent evidence on IQ and heredity is even more compelling. Identical twins show a correlation of about 0.88 on intelligence tests when they grow up together, and only slightly less when they grow up apart. This figure could not possibly be higher, since if the same twin is tested twice, the correlation between the two tests is also 0.88. Even more striking is the discovery that there is no correlation in intelligence between an adopted child and a biological child raised together, while there is a correlation between biological siblings who are adopted and brought up apart.

It is a pity that Cohen did not stick to the theme of nature versus nurture: he has many digressions, for example, on Popper’s view of science. He also deals at some length with errors in thinking, such as mistaking the cause of an event. This particular digression is an attempt to account for the failure to consider genetic explanations of their data. For instance, the finding that people who have lost a parent when young are predisposed to depression has been interpreted to mean that it is the loss of the parent that is the cause. In reality, the cause could, I think, equally well be heredity – suicidal or depressed parents are more likely to die (or separate) than more normal ones and their children will inherit their genes.

Had Cohen kept more closely to his subject, he could have assembled much more evidence to support his case. He omits, for example, the studies showing that children who are abused or neglected are only slightly more likely to become delinquent than children who are not: once again the difference may be caused by genetic inheritance rather than by the parents’ maltreatment.

Despite his tendency to ramble, Cohen has written an interesting and readable book. It is enlivened by many apt quotations from authors ranging from David Hume to C. S. Lewis. He uses the words of Francis Bacon to characterise his opponents: “For what a man had rather were true he more readily believes. Therefore he rejects difficult things from impatience … things not commonly believed, out of deference to the opinion of the vulgar. Numberless in short are the ways, and sometimes imperceptible, in which the affections colour and infect the understanding” – a sentiment which die-hard environmentalists would do well to reflect.

Despite the massive amount of research that has been done on the subject,
social psychologists know little about personality. They do not agree on
its components, nor on its origins and effects. Most of the suggestions
as to what constitutes personality are based on common sense and, like most
common-sense ideas, they often conflict.

In trying to show how personality influences mental ability, Gary Collier
demonstrates what a quagmire the area is. We can all agree with him when
he says that success in any field requires great effort. Schoolboys not
otherwise remarkable for their intellect may display an amazing ability
at computing, or achieve a knowledge of county cricket that would put Wisden
himself to shame. And nobody can become an expert chess player without having
spent at least 10 000 hours at the board. What is it that drives such people
to practise so assiduously?

The main preoccupation of many social psychologists is the invention
of new personality traits, and it is to these that Collier turns in his
account of motivation. One of them, ‘locus of control’, is the extent to
which people attribute their performance to their own ability or to extraneous
factors such as luck or the difficulty of the task. People who believe they
are not responsible for their own performance will make little effort and
give up at the least sign of difficulty.

It is well established that women tend to attribute their successes
and their failures to external factors far more than do men, but why this
should be so is unclear. Collier suggests that boys are accustomed to overcoming
discouragement because they receive more criticism from their parents than
do girls, but many other explanations – including the level of the hormone
testosterone circulating in the body – are equally plausible.

Among other traits that have been used to account for levels of motivation
are ‘fear of failure’, which prevents people from trying hard enough, and
its stablemate, ‘fear of success’, which has the same effect. The latter
is said to be common among women, possibly because competition and achievement,
particularly in tasks that are traditionally performed by males, may seem
to detract from a woman’s femininity.

But even the validity of the questionnaires used to measure these traits
is doubtful, and the correlations between performance and the level of each
trait are low. Perhaps in consequence, Collier rarely cites the relevant
correlation coefficients, nor – to use that magic word – their ‘significance’.
If, as psychologists often do, you look at enough variables, something,
some time, is certain to turn out to be significant.

Moreover, it is apparent from The Social Origins of Mental Ability that
the study of personality is beset by the difficulty of distinguishing
cause from effect. Do people have a strong need for achievement because
they have done well in the past or does the need for achievement make them
succeed? Is poor reading ability in children the result of watching too
much television, or is the child who cannot read driven to watching television?
And how do we define a term as inexact as ‘too much’? We do not know.

Collier suggests a few educational reforms that might help everyone
to achieve their intellectual potential. It has, for example, been found
that if an older child systematically instructs a younger one, both benefit
but this technique may only work because of its novelty.

Collier also pleads for unspecified social changes to inner cities and
black ghettos. Unfortunately, attempts to change society, however well intentioned,
often have unexpected and undesirable effects. Desegregation of schools
in the US caused a drop in self-esteem in black children because they compared
themselves with their white classmates rather than with one another. The
growth of the social services may have caused the disadvantaged to rely
on outside help and so to feel that they are no longer in control of their
own destiny, a feeling that leads to depression. In the US, even the official
ending of discrimination against black people has had some undesirable consequences.
As Collier points out, in many cities the brighter and better-motivated
black people have moved into white, middle-class suburbs so that there are
far fewer role models in the black community.

All the evidence suggests that making drastic changes to society and
its institutions are a mistake. This lesson is persistently ignored, as
the British government’s recent interference with the National Health Service
and with education show. The reason is that people consider only the benefits
of their ideas, and rarely reflect on the possible drawbacks.

Oddly, we have more knowledge of why people make mistakes in thinking
than of why they get things right, yet research on irrationality is hardly
mentioned by Collier. There is no known method for increasing motivation
nor for producing talent, but one can help people not to make mistakes.
Perhaps Collier should have concentrated on that as a modest but practical
aim.

Stuart Sutherland is director of the Centre for Research on Perception
and Cognition, University of Sussex.

About 40 years ago, the story goes, Eleanor Gibson and her husband,
the renowned J. J. Gibson, visited the Grand Canyon with their infant. Eleanor
believed that perception was to a considerable extent learned. Her husband,
a nativist, thought perception was an innate quality and suggested a simple
but ruthless experiment to resolve the issue: let the child walk around
and see whether it would fall off the edge of the cliff.

Because of her views on perceptual learning, and possibly because she
was a conscientious mother, Eleanor objected to this test. It would appear
from the account she gives in her book that the story has improved considerably
in the telling. It is, however, true that on returning to her laboratory,
she devised a singularly ingenious piece of equipment to provide a way of
testing depth perception in infants without precipitating dangerous falls.

She will be best remembered for this apparatus, and the work she performed
with it. Known as the perceptual cliff, it consists of a sheet of glass,
half of which is supported by a table with a patterned surface. The other
half is suspended over a similarly patterned surface, but a metre or so
lower. On top of the glass, a platform straddles the edge between the ‘shallow’
and ‘deep’ sides. Gibson found that from birth (or as soon as their eyes
opened) many animals, including chickens, goats and rats, moved to the shallow
side rather than the deep side when placed on the central platform. This
provided a convincing demonstration that the perception of depth is innate
in these animals. She also showed that human infants, tested as soon as
they could crawl, also moved to the shallow side.

These experiments were of great importance because they were among the
first to reveal that some high-level perceptual capacities are innate.
They led to a surge of research showing that much more of our behaviour
is innately determined than would have been thought possible 40 years ago.

Among Eleanor’s other important contributions were her demonstrations
(some undertaken with her husband) that motion parallax could be used to
provide information about depth and also her insistence – backed by experiments
– that learning is involved in perception.

An Odyssey in Learning and Perception reprints 33 of her major experimental
and theoretical papers, together with her comments on their origins and
significance. Unfortunately, these comments are distinctly reticent. She
was heavily influenced in her thinking by her husband, but we learn little
of how any disagreements were resolved. I remember J. J. as a kindly man,
but one who could be extremely fierce if his ideas were opposed – so much
so that when I gave talks in the US at Cornell University (where the Gibsons
spent most of their working lives), he would remove his hearing aid and
scowl if I said anything with which he disagreed. It would have been interesting
to know how Eleanor coped with such behaviour, or indeed whether he behaved
in this way with her.

Despite the importance of her experiments, Eleanor’s theorising tended
to be rather vague and sometimes erroneous. For example, she claimed that
perceptual learning consisted of learning to pay attention to the features
that differentiate individual objects. She writes: ‘The development of a
specific response to an item is correlated with the development of specific
responses to the qualities, dimensions, or variables that relate it to other
items . . . for a child to identify an object, he must be able to identify
the differences between it and other objects.’

This statement, however, ignores the obvious fact that few people can
identify the features which enable them to recognise faces. Moreover, an
object is not merely perceived as a set of features – the relationship
between the features is equally important.

It may be that her husband’s influence spoiled her ability to theorise:
she makes much use of the grand but often obscure terms which he was so
good at inventing, such as ‘invariants’ and ‘affordances’, and she believes
that information is ‘picked up’ from the ‘visual array’ and somehow determines
perception without further processing. Like her husband, she denies the
existence of representations (stored models of the environment constructed
in the brain), but neither of them can begin to explain how people can make
judgments about depth or can recognise objects without such representations.
Moreover, Eleanor scorns the work of cognitive scientists, including David
Marr, who have attempted to delineate the complex information processing
that must underlie vision and all other human capacities.

Unfortunately, this is not a book for the general reader. Although
for the most part it is clearly written and easy to follow, the papers of
necessity contain a great deal of technical detail. Eleanor’s comments only
provide snippets of history. Her career started at Smith College, where
J. J. combined sponsoring her MA thesis with marrying her. In 1949 they
both moved to Cornell, but it was not until 20 years later that she received
a faculty appointment there, a fact that she records dispassionately and
without dismay.

I would have liked to hear more about her relationships with fellow
psychologists, about her feelings on publishing her first paper and, above
all, about the seemingly perfect marriage to her husband. When strolling
at night through the beautiful Cornell campus with its lakes and trees,
did they exclaim about the stars or did they indulge in sweet murmurings
about ‘invariants’ and ‘affordances’?

Stuart Sutherland is director of the Laboratory of Experimental Psychology
at the University of Sussex.

Perhaps the least noticed and most disadvantaged minority group is the
left-handed. Their bad press goes back to ancient times. After all, ‘(God)
set the sheep on his right hand, and the goats on the left . . . (saying)
unto them on the left hand ‘Depart from me, ye cursed, into everlasting
fire’.’ Today they suffer from artefacts designed by the right-handed, who
take no account of how well the other 10 per cent of the population can
cope. In the kitchen they have difficulty with tin-openers – although expensive
left-handed versions can be found – but in the workplace they face band
saws and computer keyboards, or attempt to play with hockey sticks, none
of which have left-handed versions.

Stanley Coren traces the causes and consequences of being left-handed.
His account is enthralling and despite its scientific accuracy often reads
like a detective story, for he reveals the thought processes by which he
reached his conclusions. For example, he suspected that on average the left-handed
die younger than others. After much thought and one or two false starts,
he did a quick and dirty study by examining the Encyclopedia of Baseball,
which recorded the deaths of famous players and the hand with which they
pitched or batted. Not satisfied with such an unrepresentative sample of
the population, he perused recent death certificates, and approached the
next of kin to discover the handedness of the dead people. On this basis,
the life expectancy of right-handers appears to be more than seven years
greater than that of left-handers.

This difference is not caused solely by the number of accidents left-handers
incur, although they are certainly accident-prone. Left-handedness is also
accompanied by increased rates of allergies, autism, depression, migraine
and schizophrenia among other disorders.

The condition is also associated with problems at birth: a staggering
54 per cent of premature babies are left-handed. Coren suggests that some
left-handedness is caused by the presence of too much testosterone in the
amniotic fluid surrounding the foetus. Because testosterone retards the
development of the left hemisphere, which controls the right hand, an excess
of it could result in the right hemisphere becoming dominant and hence in
left-handedness. The hypothesis would also explain why left-handers have
an unusually high risk of developing certain illnesses including allergies,
for testosterone slows the growth of the thymus, which produces white blood
cells. In addition, it would explain why more men than women are left-handed,
for they are exposed to more testosterone in the womb.

The Left-Hander Syndrome is a clear, agreeable and consistently intriguing
book. It will not give much comfort to left-handers, though the author is
at pains to point out that many are normal in other ways. Indeed, it is
possible that they go to extremes in terms of talent: although an above
average number suffer from mental disabilities, it should be borne in mind
that Michelangelo, Pablo Picasso, Benjamin Franklin and Charlie Chaplin
were all left-handers. The book deserves to be read for its fascination
alone, but it should also help those who are right-handed both to understand
the problems of their less fortunate fellows and to treat them with more
compassion and tolerance.

Stuart Sutherland is director of the Laboratory of Experimental Psychology
at the University of Sussex.

TODAY as never before, many people devote their lives to postponing
death. They seem to prefer quantity to quality, for they endure ruthless
dieting, margarine and self-punitive jogging. The medical profession aids
and abets them, both when death is a distant prospect and when it is imminent.
Dr Lesley Fallowfield makes a passionate plea for doctors to pay more attention
to the quality of life, rather than attempting merely to keep their patients
alive as long as possible.

One way in which doctors could improve the quality of their patients’
lives is to give more information about the likely course of an illness,
since that makes it easier to bear: in fact, few doctors tell the patient
enough. Fallowfield herself found that patients with breast cancer suffer
less anxiety and depression if told in detail about their prognosis. There
are many other examples: patients undergoing abdominal surgery who were
briefed before the operation about its likely sequelae needed fewer painkillers
and came out of hospital on average four days earlier than patients who
underwent the usual hospital routine. Even terminally-ill patients tend
to be less depressed if they know, rather than suspect, that they are dying.
As Fallowfield points out, it is extremely divisive if the spouse knows
and the patient does not; it is difficult for them to be close to one another
when such dishonesty exists. Fallowfield does not mention the doctors’ standard
reply, namely, that they tell patients if they ask. Unfortunately, this
is little help because many patients hate wasting the doctors’ time and
are too timid to ask enough questions.

Depression often follows many illnesses, such as cancer, and it is not
usually treated because doctors regard depression as a natural consequence.
As Fallowfield remarks, this is like refusing to relieve pain because it,
too, is the normal result of an illness. Even pain does not receive enough
attention: about 70 per cent of terminally ill cancer patients have severe,
unremitting pain, most of which, she claims, could be relieved but is not.

She also stresses the importance of social support in illness. Cancer
remains for many a taboo word and as many as 50 per cent of the friends
of can cer patients avoid them; presum ably far more would refuse to have
contact with a person with AIDS. How we can remove such stigmas is not clear,
but clinical psychologists or psychiatrists could help patients. Research
has shown psychotherapy to be an effective way of dealing with depression
or anxiety resulting from illness. The difficulty here is that there is
a drastic shortage of clinical psychologists in Britain.

The last chapter of The Quality of Life considers how, in deciding on
treatment, it is possible rationally to balance the quality of life for
a patient against the number of additional years of life that a given treatment
is likely to provide. Many treatments, for example, chemotherapy or radiotherapy,
have very unpleasant side effects. Everyone would prefer to live a normal
life for nine years than to live in acute pain for 10, but that is a simple
case.

One technique is to ask people how many years of life with a given disability,
such as an inability to walk, they would trade off against one year’s normal
health. That number of years is then treated as 1 QALY (Quality Adjusted
Life Year – pronounced ‘kwally’) for that disability. Because medical provision
is in practice already rationed, medical workers have tried to assess priorities
by working out the cost per QALY for different treatments. It turns out
that – in these terms – a shoulder-joint replacement is much more effective
than a kidney transplant and that this is more effective than renal dialysis.

At this point Fallowfield flinches, partly through doubts about the
validity of QALYs, partly because their application would deprive some patients
of health care. But some people are already deprived: the only difference
is that they are deprived largely on a haphazard basis in which social class,
wealth, where you live and who you know all play a part.

The Quality of Life is written with clarity and compassion. It is accessible
to the general reader and should be read by all doctors, particularly those
who ‘strive officiously to keep alive’.

Correction: Theories of Visual Perception (Wiley, pp 273, Pounds sterling
29.95 hbk, Pounds sterling 12.95 pbk), reviewed on 28 April, is by Ian E.
Gordon.

Stuart Sutherland is director of the Centre for Research on Perception
and Cognition, University of Sussex.