Jump to ContentJump to Main Navigation
Why Red Doesn’t Sound Like a BellUnderstanding the feel of consciousness$

J. Kevin O'Regan

Print publication date: 2011

Print ISBN-13: 9780199775224

Published to Oxford Scholarship Online: January 2012

DOI: 10.1093/acprof:oso/9780199775224.001.0001

Show Summary Details
Page of

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2018. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a monograph in OSO for personal use (for details see www.oxfordscholarship.com/page/privacy-policy). Subscriber: null; date: 16 January 2019

Toward Consciousness

Toward Consciousness

Chapter:
(p.73) Chapter 6 Toward Consciousness
Source:
Why Red Doesn’t Sound Like a Bell
Author(s):

J. Kevin O’Regan

Publisher:
Oxford University Press
DOI:10.1093/acprof:oso/9780199775224.003.0049

Abstract and Keywords

This chapter looks at the progress that scientists have made so far in making machines that think, perceive, learn, use language, and that have the notion of self. It discusses the development of artificial intelligence, the cognitive self, and the societal self. It argues that to give a scientific account of the self, at least as concerns the cognitive and social aspects, no magical, as yet unknown mechanism is likely to be needed. Making machines with cognitive and social selves is a task that may still take years to achieve, but it seems clear that once we have machines with sufficient cognitive capacities that are embedded in the real world and integrated into a social system, a notion similar to the “I” that humans use to refer to themselves is likely to become so useful that it is “real”.

Keywords:   artificial intelligence, consciousness, cognitive self, societal self, machines

When I was a child, my dream was to build a conscious robot. At that time in the 1960s, computers were just coming into existence, and they were where I hoped to find inspiration. I haunted technical bookstores and libraries and read all about “flipflops” and “core memories.” I memorized complicated diagrams of binary “half-adders” and “full-adders” with “AND” and “OR gates.” At school I astounded the Science Club with a demonstration of how you could build a simple computer that used water instead of electricity, making binary calculations at the same time as spraying jets of water all over the chemistry lab.1

I found an ad in a popular science magazine for a kit to construct a machine called Geniac that solved logical puzzles and played tic tac toe (see Fig. 6.1).2 In the picture the machine looked impressive with complicated dials and lights. To the despair of my parents (who thought I was a bit too young), I insisted on ordering the machine. When it finally arrived from overseas, I spent a feverish week wiring it up after school every day. When I finished, I was disappointed. Admittedly the machine played good tic tac toe. But it was essentially just a sophisticated set of switches. There were hundreds of criss-crossing wires, but it didn’t seem to be intelligent. It couldn’t think. It wasn’t conscious.

I kept wondering what was missing from the machine. Consulting a book in my parents’ bookshelves on the anatomy of the nervous system, I saw diagrams showing complicated connections, leading from different parts of the body to the spinal cord, synapsing wildly with other nerve pathways, going up into the brain and spreading out all over the place. I spent hours poring over that book. How, I wondered, was the brain different from the tic tac toe machine? Was there more than just complicated wiring? What special principle or what additional module would I have to add into the tic tac toe machine so that it would be conscious?

How to build consciousness into a machine is a question that many people have been thinking about for several decades. If we could solve this question, we would make considerable headway in understanding how consciousness arises in humans, which in turn would shed light on the nature of feel.

In this chapter, I shall be looking at the progress that scientists have made so far in making machines that think, perceive, learn, use language, and that have the notion of self. Then, in the next two chapters, I shall discuss two different kinds of consciousness: access and phenomenal consciousness. I shall be arguing that thinking, perceiving, learning, the self, and access consciousness pose no logical problem to science and as such are achievable in these machines—if not now, then at some (p.74)

                      Toward Consciousness

Figure 6.1 Ad for Geniac kit and circuit diagram for tic tac toe from http://www.oldcomputermuseum.com. Used with permission of Old Computer Museum.

point in the future. But we shall see that phenomenal consciousness—or what I shall be calling “raw feel”—is going to need special treatment. The new view of seeing will provide inspiration for how to solve this problem.

Machines That Think, Perceive, Use Language, and Learn

Artificial intelligence (AI) was born in the 1960s when “knowledge-based systems” or “expert systems” were designed in computer laboratories to analyze information (p.75) contained in large, complicated databases. Such systems worked best in very specific domains. For example, they were used to find oil deposits, to give investment advice to banks, to diagnose blood diseases and recommend antibiotics, to optimize the way companies stock warehouses and organize deliveries, to play games like chess, checkers, and “go,” and to solve mathematical problems. After a slow-growing period in the 1980s and 1990s, sometimes called the “AI winter,” such systems have now matured into commercially viable products that are used daily across the world. For example, the spam filter in your e-mail program has been quietly analyzing the mails you have thrown away to fine tune the way it will work in the future. The ads that appear on your Web browser are selected by analyzing the information you have recently been obtaining.

Recent AI systems also use numerical methods based on statistical pattern recognition techniques, some of them modeled on the functioning of networks of neurons in the brain. Such approaches have given birth to active research domains like “connectionism” or “artificial neural networks” and “machine learning,” which are proving successful in applications where pattern recognition is needed. For example, robot vision and control systems guide autonomous vehicles like surveillance drones or driverless cars. Speech recognition systems are used to understand simple telephone commands in airline reservation systems. Systems that recognize car license plate numbers can use “neural network” pattern recognition techniques. Other examples are search engines that ferret out information from the Web, or programs that analyze stock market trends or recognize handwritten addresses on envelopes in mail sorting.

Progress is also being made in language processing.3 Useful devices operate today for specialized purposes like generating multilanguage lexicons and language- independent semantic representations, tagging parts of speech in text corpuses, correcting grammar and orthography, indexing, summarizing text, answering simple questions in help applications, subtitling speeches, and translating computer menus for different countries. One of the earliest automatic translation systems from the 1960s is today a successful commercial company,4 SYSTRAN, which sells the translation technology used by Google, Yahoo!, and Alta Vista, as well as by the European Commission, which circulates documents in more than a hundred languages. There are also frivolous offshoots of early natural language processing research, like ELIZA the automated psychoanalyst, and like the admirable postmodern article generator, a Web site that you can use to generate amusing spoofs of postmodern articles on literary criticism.5

But for all these advances, it remains true that the goal of human-like thought, perception, and language understanding is still far from being attained in machines. To solve the problem, two promising ideas are currently being explored: One is to provide the systems with common-sense knowledge about the world and about human life.6 The other is the idea that in order for machines to attain human cognitive performance their involvement in the world may have to replicate that of humans.7 Humans have bodies and live, move, and interact in particular ways with the objects that they use and with other humans in the real world. Human language and thought are not just raw symbol manipulation: The concepts that are manipulated are constrained by the physical world and the particular way humans interact with it. People live in a shared social environment and have desires, emotions, and (p.76) motivations that play an important role in conditioning thought and communication. Thus, it may be that to think and use language like humans, machines will have to have human-like immersion in the world.

For this reason a domain of robotics has emerged recently called “autonomous robotics,” where the idea is to construct machines that possess sufficient intelligence to control their own behavior, and that move and explore their environments and act of their own accord. Such machines are beginning to be used in situations where communication with a human is impractical, for example, in space or in surveillance. Other examples are domestic robots that clean a swimming pool, mow the lawn, vacuum the living room,8 or entertain grandma and the kids.

A related trend is “developmental robotics” (also called “epigenetic robotics”). This involves building autonomous robots that are initially provided with only basic capacities, but that, by interacting with their environments like a human infant, evolve to master more complex perceptual and cognitive behavior.9 One early example was BabyBot, constructed at the Laboratory for Integrated Advanced Robotics (LIRAlab) in Genoa (see Fig. 6.2). Another baby robot, the icub, is a more recent project financed by the Robotcub consortium of the European Union.10

By constructing a robot that can interact with its environment and with humans and other robots, these projects aim to study how problems such as visual segmentation, tactile recognition, manual manipulation, and word meaning can emerge in a way similar to human infant development. For example, usually machines that visually recognize objects only make use of information extracted through video cameras. BabyBot, on the other hand, recognizes a toy car not only visually but also by poking it around and understanding that depending on which way the robot pushes it, the car will or will not roll. This information allows the robot to link visual information to manual “affordances,” which may then help it to recognize other objects and understand what it can do with them.11

Even if progress has been much slower than some luminaries predicted in the past,12 the AI and robotics communities remain confident that gradually but surely we will have machines that approach or surpass human thought, perception, and language skills. In this venture, it is clear that in order for machines to capture meaning in the same way as humans, engineers will have to take account of the essential involvement that humans have in the real world. Artificial systems may have to be provided with sensors and the ability to move, and to be embedded in the world and possibly even in a social system. With better hardware and closer attention to the way humans are physically and socially embedded, logically nothing should prevent us from understanding and replicating in these systems the mechanisms at work in human thought, perception, and language—even if in practice this may remain impractical for years to come.

But something still seems to be missing.

The Self

We as humans have the distinct impression that there is someone, namely ourselves, “behind the commands.” We are not just automata milling around doing (p.77)

                      Toward Consciousness

Figure 6.2 (Top) Babybot from http://www.lira.dist.unige.it/babybot/robot.htm. and icub from http://www.robotcub.org/index.php/robotcub/platform/images/icub_sitting, and (Bottom) myself conversing with icub. Used with permission of Lira Lab in Geneva and G. Sandini.

intelligent things: There is a pilot in the system, so to speak, and that pilot is “I.” It is I doing the thinking, acting, deciding, and feeling. The self is a central part of human thought, and it also plays a fundamental role in language.

So what is the self? The nature and origin of the notion of the self is an actively discussed subject today.13 Philosophers are trying to decide what precise components the notion of self boils down to. Within psychoanalytical approaches, Freud’s well-known distinction between the id, the ego, and the superego as components of the self, and Heinz Kohut’s self psychology with its own tripartite theory of the self,14 are just two examples within a very rich, diverse, and complex tradition. Very active work is also being done in cognitive psychology. Developmental psychologists and psycholinguists are trying to ascertain how the notion of self develops in the maturing child; cognitive anthropologists look at whether the notion is different in different human societies; cognitive ethologists study which species possess (p.78) a self; and social psychologists investigate how the self is determined by an individual’s social environment.

To get an idea of the complexity of the notion of self, we can begin by looking at how psychologists and philosophers have decomposed and refined the distinctions between different aspects of the self.15 At the end of the 19th century, the philosopher William James16 distinguished the physical self, the mental self, the spiritual self, and the ego. Further refinements were suggested in the 1980s by the cognitive scientist Ulrich Neisser17 with his ecological, interpersonal, extended, private, and conceptual aspects of the self. Yet further concepts have been proposed more recently, such as the primary self, secondary self, cognitive self, conceptual self, and representational, semiotic, contextualized, embodied, social, extended, verbal, dialogic, fictional, narrative,18 private, and minimal19 selves, as well as the notion of self model,20 the essential self, and the transparent self21!

How can we make sense of this plethora of different terms that have been invented to describe the self? In particular, is there anything hidden within the concept of self that somehow prevents it from being built into a robot or approached by the scientific method? To try to bring some order into this discussion, I would like to isolate two important dimensions along which different notions of self are defined22: a cognitive and a social dimension.

The Cognitive Self

The cognitive dimension concerns self-cognizance, namely the amount of knowledge an individual has about itself. We can mark off three tiers in a continuum of self-cognizance, going from “self-distinguishing” to “self-knowledge” to “knowledge of self-knowledge,”23 and ask whether there would be any problem building these tiers into a robot.

Self-distinguishing is a precursor to the cognitive self: It is the fact that an organism acts differently with regard to parts of its own body as compared to the outside world. It is something very basic that doesn’t require a brain or any form of cognitive capacity. For example, an animal’s immune system must distinguish foreign cells from the host body’s own cells. For any animal, moving around in the environment requires implicitly distinguishing itself from that environment. Clearly there is no problem building this into a robot. For example, the Humanoid Robotics Group at MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) has developed Domo, a robot constructed to investigate this stage of the self notion (see Fig. 6.3).

The robot is an upper torso equipped with moveable eyes, head, arms, and grippers. It uses vision-based movement detection algorithms to determine whether something is moving in its visual field. It checks whether by commanding movements of its own body, the movements it sees are correlated with the movements it commands. If such a correlation occurs, it assumes that what it is seeing is part of its own body. In this way it is able to figure out what its own hand looks like, and later, what its own fingers look like.24

Self-knowledge requires a system with cognitive capacities, that is, a system which is more than simply reactive and that can make choices between different courses of action, make judgments, and plan its behavior. Self-knowledge involves (p.79)

                      Toward Consciousness

Figure 6.3 Domo was constructed by Charles Kemp, now at Georgia Tech. Reprinted from Edsinger, A., & Kemp, C. (2006) with permission.

cognitive behavior that makes use of the fact that the system is a whole, with a body that separates it from its environment and other individuals. This allows the system to avoid bumping into other individuals, to find mates, and to avoid predators. It allows it to have the notion of ownership and to distinguish its possessions from the possessions of others. But note that under this definition of “self-knowledge,” the system may not actually know that it has this self-knowledge. Thus, for example, a bird may function perfectly in its environment, choosing courses of action, finding food that it has hidden away, distinguishing itself from others, and separating its territory from other birds’ territory; yet it may not in any sense know that it is an individual. To the extent that computers today can already be programmed to manipulate concepts and make inferences, there is nothing to prevent engineers in the future from programming a robot so that it reasons about itself as an entity that can be distinguished from its surroundings in order to make plans, judgments, and decisions about what to do next.

Knowledge of self-knowledge is the further stage where an individual may have various kinds of meta-knowledge about its self-knowledge. For example, the most basic meta-knowledge is for an individual to know that it exists as an individual. It may also know that other individuals know that they exist as individuals. It may also know that other individuals know that it knows that it exists and that they exist. . . Things can get complicated, leading to finely graded social strategies ranging from selfishness, through misleading other individuals, to cooperating with them. Such meta-knowledge is necessary for an individual to have embarrassment, shame, pride, and contempt. With this level of meta-knowledge of self-knowledge, the individual has a “Theory of Mind”: It can empathize with others, and interpret other individuals’ acts in terms of their beliefs, desires, and motivations. Knowledge of self-knowledge underlies what has been called the “intentional stance” that humans (p.80) adopt in their interactions with other humans: They consider that other humans have intentions and goals just as they themselves do. As concerns robots, knowledge of self-knowledge requires a level of conceptualization in which the robot must make the link between its own plans, judgments, and decisions and those that can be attributed to other agents in a social context. Again, as a matter of concept manipulation, there is in principle no obstacle to building such a capacity into a robot. It would, of course, have to be immersed in a social context, either with other robots or with humans.

Many research teams are currently working on projects in this direction, and although, as compared to human self-knowledge or knowledge of self-knowledge, the robotic self is a pale imitation, progress is being made. Work on these higher notions of self is being done, for example, with the COG platform, also developed at CSAIL (see Fig. 6.4). COG is an upper-torso humanoid robot equipped with visual, auditory, tactile, vestibular, and kinesthetic sensors; it can move its waist, spine, eye(s), head, arms, and primitive hands. COG has been used by a variety of groups at CSAIL to do experiments in object recognition, tactile manipulation, and human–robot interaction. It was actually one of the first humanoid robotic platforms built at MIT, and it is approaching “retirement.” It has provided the inspiration for more recent work.

The ideas being used to study the emergence of the self in COG are based on analyses of what psychologists consider to be the most basic capacities postulated to underlie the self in humans.25 One such basic capacity is the ability to locate and follow an agent’s gaze direction. For this, skin, face, and eye detection algorithms in

                      Toward Consciousness

Figure 6.4 COG, from Scassellati, now at Yale (Scassellati, 2003). Reprinted from Scassellati, B. (2003) with permission from the author.

(p.81) the robot’s visual system allow it to locate eyes and infer where a human is looking. By extending this gaze-following capacity, the researchers hope to implement algorithms for joint attention, that is, algorithms that allow the robot to attend to an object that another agent or human is also attending to.

The COG, Domo, and many other projects are work in progress and at the moment represent only the very first steps toward implementation of different self notions in a robot. But the vitality of this and related research projects suggests that providing robots with a realistic notion of self and accompanying Theory of Mind is an achievable goal. Even if it takes many more years, as concerns the cognitive capacities involved, the problem of building a robot with a self seems in principle solvable.

The Societal Self

The cognitive dimension of the self concerns how an individual’s cognitive capacities allow it to interact with the environment. There is already a social dimension to this, to the extent that the individual interacts with other individuals and can develop a Theory of Mind. But there is another social aspect to the self that concerns how an individual fits into a society and how each individual considers him- or herself within this society.26

Scientists studying this societal aspect of the self agree that even though each of us has the intimate conviction that individually we are just one person, “I” is essentially a melting pot of subselves, each manifesting itself in different conditions, and all having learned to cohabit under the single label: “I.” The 18th-century philosopher David Hume proposed that contrary to the individual’s impression of being a single self, “I” am in fact a composite bundle of different capacities and entities. “I” am like a republic or commonwealth, or like a nation or club whose structure may change over time, but it can nevertheless be usefully described as a single unit.27 Under this view, “I” am a convenient abstraction that “I” use to describe (both to others and to “my” self) how the entity that composes “me” behaves, with “my” thoughts, attitudes, and opinions.28 Daniel Dennett proposes that the self is a “center of narrative gravity,” useful for describing to others and ourselves what we do, want, or believe.29

But how can “I” seem so real to myself, if “I” am merely a cognitive construct? An analogy that helps to understand this is money. Despite just being bits of metal or paper, money is perfectly real in our society. The reason is that money is self- validating: It requires a social consensus to function, but it also creates, through its functioning, a situation where it can continue to function effectively. The self might be a similar self-validating social construct.

Compared to money, however, there is something additional about the self: “I” have the intense conviction of my own self as really existing. Unlike money, which is used by people (and not by money), the self is a notion used by the very selves that the notion defines.30 The concept of self is in that sense not only self-validating like money but also self-referring.

Thus, the reason I experience myself as being so real is that I am telling myself a self-consistent story about myself being real. Furthermore, this story is reinforced by my interactions with other people who have similar stories about themselves and (p.82) about me, and they and I act in ways that cause the stories we are all telling to be true. Said in another way: “I” is part of its own story. The story has evolved so that what we mean by “I” includes being convinced that “I” is real. The story of “I” is precisely the story of a single entity that is the center of its own story. If you are part of the story, you must by definition be convinced that you are real.

Thus, the notion of a self is defined within a social context, from the outside. Yet it is a notion that the self in question must espouse on the inside.31 And although it is only a social construct, it is a self-referring social construct that enables the self to perceive itself as being very real. This rather shocking idea has some interesting consequences. One concerns the question of who writes the story of “us.” The other concerns how to escape from our story.

Social psychologists studying the unconscious influence of cultural prototypes on our behavior show that our everyday actions are more determined than we think by automatic, socially driven influences. We unconsciously espouse images of ourselves as having a certain personality, as belonging to a particular social category, and these cultural prototypes strongly influence the construction of our identity. Indeed, a person’s gait, gestures, speech, taste, and dress are all exquisitely sensitive to their cultural or social context. You can tell a person’s origins from the way he or she moves and talks. I’ve lived in France so long now that my mother (who is American) says she can’t understand me anymore because when I talk, I move my hands like a Frenchman. Even something so primal as the expression of pain32 is influenced by cultural pressures: When I’m among French people and I hurt myself, I exclaim “aïe.” Among Anglo Saxons, I cry “ouch.”33

But even if it is true that the main influence on the construction of our identities is unconscious and dependent on our social environment, do “I” not also play some conscious role in the construction of “I”? This would have to be the exception and not the rule, because if people constantly changed their stories to suit their momentary whims, they would never be the same and the effect would be destabilizing for society. It probably has to be part of “my” story that I should be convinced that I am an entity that cannot just switch to a dramatically different identity. There has to be something like a social taboo preventing us from consciously modifying ourselves.

There are, however, ways in which people can and do escape from their normal stories and change their identity—for example, through alcohol and soft drugs, through hypnosis,34 or through channeling, possession trances, and culturally bound syndromes35 like amok, zar, latah, and dissociative identity disorder (multiple personality disorder). Changes in identity also occur as a result of pathologies like schizophrenia, brain lesions, and traumatic head injury. Take, for example, my friend FP, who knocked his head in a skiing accident and completely lost his long-term memory. I only got to know FP a few months after the accident, but I had noticed that he had a hesitant, inquiring way of acting in social contexts, as though he were searching for approval. As he explained to me years later, he was actually reconstructing his self. He explained how even his slightest action was filled with questioning. He needed to search for his own self in the eyes of others. He would sit down at the table and pick up his knife and fork. But should he pick them up this way or that way? Of course, we all encounter such problems in new social situations where we are not sure how to behave. But FP had a much harder time because not (p.83) only did he have to work out the social constraints, but much more important, he needed to find out what he was like. He would look around and try to judge from the way his friends looked at him whether he was behaving in the way they expected the real FP to behave. Was he the kind of person who sat upright or did he slouch relaxedly? Did he talk loudly or softly? I think this example is interesting because it is an exceptional case where a person has some degree of insight into the social construction process that constructs his own self.

The previous paragraphs only scratch the surface of the active research being done on the cognitive and social aspects of the self.36 However, what they illustrate is that to give a scientific account of the self, at least as concerns the cognitive and social aspects, no magical, as yet unknown mechanism is likely to be needed. Making machines with cognitive and social selves is a task that may still take years to achieve, but it seems clear that once we have machines with sufficient cognitive capacities that are embedded in the real world and integrated into a social system, a notion similar to the “I” that humans use to refer to themselves is likely to become so useful that it is “real”.

Notes

(1.) Doing calculations with fluids was a project in the Amateur Scientist section of Scientific American, May 1966 (p. 128).

(2.) The Geniac machine was conceived by Edmund C. Berkeley, a computer scientist and antinuclear activist, author of the book Giant Brains or Machines That Think (1949). See http://www.oldcomputermuseum.com/geniac.html and especially http://www.computercollector.com/archive/geniac/for more information.

(3.) For machine translation, see the review by Hutchins (2003). Examples of the difficulty encountered in making a computer understand natural language are given in the Supplements to this book on http://whyred.kevin-oregan.net

(4.) In 2010 you could buy the SYSTRAN program for your own PC, starting at only 39 euros, and choose from the Professional Premium, the Professional Standard, the Personal, the Web, the Office Translator, and other versions of their software!

(5.) Weizenbaum’s ELIZA program is on http://courses.cs.vt.edu/∼cs3604/lib/Ethics/eliza.html, the postmodern generator on http://www.elsewhere.org/pomo. The generator was written by Andrew C. Bulhad using the Dada Engine.

(6.) The best known example of an attempt to compile “common sense” is Douglas Lenat’s “CYC” project, which started in the 1980s at Atari corporation and has now become a company, Cycorp. Cyc is an ongoing effort to accumulate human consensual knowledge into a database of more than a million assertions which can be used to understand, for example, newspaper articles or encyclopedia entries. On http://www.cyc.com/ the company says: “Cycorp’s vision is to create the world’s first true artificial intelligence, having both common sense and the ability to reason with it.”

(7.) This point was forcefully made in 1972 by Hubert Dreyfus in his influential book What Computers Can’t Do: A Critique of Artificial Intelligence, updated in 1992 to What Computers Still Can’t Do: A Critique of Artificial Reason (Dreyfus, 1972, 1992).

(8.) The company iRobot, inspired by the work of Rodney Brookes, has a Web site http://www.roombavac.com/. They say they deliver “… innovative robots that are (p.84) making a difference in people’s lives. From cleaning floors to disarming explosives, we constantly strive to find better ways to tackle dull, dirty and dangerous missions—with better results.” They say that if you have iRobot Scooba: “check mopping off your list.” With iRobot Roomba: “check vacuuming off your list.”

(9.) See http://www.epigenetic-robotics.org for ongoing activities in this field, with bibliographic references. For a survey of the related field of developmental robotics, see Lungarella, Metta, Pfeifer, and Sandini (2003).

(10.) See http://www.lira.dist.unige.it/ for BabyBot, created by Giorgio Metta when he was working in the COG project at MIT. It is now part of the larger, European RobotCub project (http://www.robotcub.org/). Videos can be seen on that site as well as on http://www.lira.dist.unige.it/babybotvideos.htm. An overview of some other current efforts in autonomous and developmental robotics can be found on the Web site for the supplements to this book: http://whyred.kevin-oregan.net

(11.) See Fitzpatrick and Metta (2003). The notion of “affordance” is due to the ecological psychologist J. J. Gibson.

(12.) For a very critical article in The Skeptic about the possibilities of AI, see Kassan (2006).

(13.) For a useful bibliography on different approaches to the self, see the Web site maintained by Shaun Gallagher: http://www.philosophy.ucf.edu/pi/ or his excellent review (Gallagher, 2000). See also the special issue of Annals of the New York Academy of Sciences in Vol. 1001, 2003.

(14.) Kohut, 2009.

(15.) I based this quick overview on Gallagher (2000).

(16.) James, 1890.

(17.) Neisser, 1988.

(18.) Dennett, 1991, 1992.

(19.) Gallagher, 2000.

(20.) Metzinger, 2004, 2005.

(21.) Part of this long list is taken from Strawson (1999); see the book by Gallagher and Shear (1999) for a recent interdisciplinary survey of many approaches to the self.

(22.) Here I am going to follow the two dimensions of classification proposed by Vierkant (2003).

(23.) This classification is based on that proposed by Bekoff and Sherman (2004). They use the umbrella term “self-cognizance” to designate the whole spectrum of ways that a system can have knowledge of itself. In their classification they use the three terms “self-referencing,” “self-awareness,” and “self-consciousness” instead of the terms “self-distinguishing,” “self-knowledge,” and “knowledge of self-knowledge.” I prefer my terms because they are more neutral as regards the question of whether any phenomenal quality might be associated with the notion of self, and because they avoid the difficult terms “awareness” and “consciousness.”

(24.) See Edsinger and Kemp (2006).

(25.) As suggested by cognitive psychologists Simon Baron-Cohen (Baron-Cohen, 1997; Leslie, 1994). For robots that imitate humans’ social development, see Breazeal and Scassellati (2002) and Scassellati (2003). See Triesch, Teuscher, Deak, and Carlson (2006) for a discussion of how gaze following might be a learned faculty.

(26.) Because this social self is such an important concept in our society, it is discussed from a myriad of different points of view, in the popular press, in psychological counseling, in psychoanalysis and in academic psychology, sociology, and anthropology. (p.85) In these domains, a plethora of concepts have been defined, including for example self-actualization; self-awareness; the self-concept; self control; self development; self disclosure; self-efficacy; self-esteem; self harm; self help; self-identity; self image; self monitoring; self-perception; attitude change; self-realization; self-regulated learning; self talk.

(27.) The comparison Hume makes actually concerns what he calls the “soul” (Hume, 1793/2003, Book 2.i.vi). Derek Parfit also makes the comparison with a nation or club (Parfit, 1986).

(28.) Why does this abstraction exist and why is it so useful? To understand this, we can appeal to the concept of “meme,” invented by the biologist Richard Dawkins. A meme is the social or cognitive equivalent of a gene. Just as there are genes corresponding to traits replicated in biological reproduction, there are “memes,” which are ideas, skills, or behaviors that stick in the mind, serve a social purpose, and so become part of human society. Memes are “replicators”; they are adapted to human cognitive capacities. They are easy to remember and stable enough to be accurately propagated through social interactions and from one generation to the next. Furthermore, memes may have the self-validating property that they influence the structure of society so as to favor their own preservation and propagation. As suggested by writer Susan Blackmore, the self may be one of the most fundamental memes in human society. The self is a concept which each of us uses in order to understand the goals and desires not only of our own selves but of our fellow humans (Blackmore, 2000; Dawkins, 1976).

(29.) Dennett, 1992.

(30.) Martin Kusch cited by Vierkant (2003) mentions the doubly self-referring and self-validating roles of social constructs like money and selves, though there may be a difference in the way Kusch uses these terms.

(31.) It is hard to believe that this kind of circularity could provide something that seems so real to us. Surely I can sense my existence—I can perceive myself talking and thinking. If I am just a story, then how can this story perceive itself talking and thinking? It’s as though it made sense to say that Sherlock Holmes, who we all know is just a fictional character, really considers himself to have an identity.To make these ideas more plausible, consider another example of a self-referring and self-validating social construct, which is the notion of country or nation. A group of people can proclaim itself as a nation, and from that moment on, the nation takes on a real existence on the international scene. Furthermore there is a real sense in which nations have “selves.” They have national identities, they build alliances, they make pacts and wage wars, they communicate and trade. They can be offended, envious of others’ resources, aggressive, or peaceful. They have national memories and even what might be called split personalities when under stress (e.g., French Resistance under German World War II occupation).Thus, judging from international events and talk in the media, and judging from the dramatic excesses that nations are willing to engage in to preserve and extend themselves, nations are real and really have identities. Because nations generally “express themselves” in different ways than humans, with no single “voice,” the analogy with a human “self” is not perfect, but if we were somehow able to go out and actually interview a nation, and ask it if it really felt it had a self, an identity, and if it really felt it was real, it would reply “yes.” (These ideas are presented admirably by Dennett, 1989.)

(p.86)

(32.) It may be that not just the expression of pain but also pain itself may be influenced by cultural factors, but the evidence is difficult to interpret (see Rollman, 1998).

(33.) Seminal experiments showing unconscious influence of social norms on behavior were done by the social psychologist John Bargh, now at Yale (see Bargh & Chartrand, 1999). In these experiments, participants were told that they were being tested on language proficiency in a sentence unscrambling task. They were given five words (e.g., “They her respect see usually”) and were required to write down a sentence which made use of four of those words. Actually, however, this supposed language proficiency task was used to “prime” participants with different cultural stereotypes. For example, in one experiment, the words used in the sentence unscrambling task were those that were either associated with politeness (e.g., honor, considerate, appreciate, patiently, cordially), or with rudeness (e.g., brazen, impolitely, infringe, obnoxious). It was found that when the participants had been primed for rudeness, on leaving the experimental room they were more likely to interrupt the experimenter who was ostensibly talking to a colleague. In another experiment the sentence unscrambling task used words that primed for an “elderly” prototype (e.g., worried, Florida, old, lonely, gray) versus a neutral prototype (e.g., thirsty, clean, private). It was found that the time it took for the participants to cross the hall on leaving the experiment was longer when they were primed with “elderly” words. A different experiment showed that participants primed for the “elderly” prototype could also not remember as many details about the room in which they participated in the experiment.

(34.) Hypnosis is so easy to induce that any basic text on hypnosis will provide an induction technique that can be used by a complete novice to hypnotize someone else. The Web site http://www.hypnosis.com/ provides a wealth of information, including induction scripts and numerous scripts purported to cure anything from AIDS to constipation to procrastination. Some hypnotists claim to be able to hypnotize people (even people who have not previously been sensitized) in a matter of seconds.The ease with which hypnosis can be induced is consistent with the idea that hypnotic trance is a matter of choosing to play out a role that society has familiarized us with. The suggestion that hypnosis is such a culturally bound phenomenon comes from the very high correlation between hypnotizability and a person’s belief that hypnosis works. Conversely, it may be impossible to hypnotize someone from a different culture, who has never heard of hypnosis. “Playing the part of being hypnotized” would seem to be a culturally accepted way of exploring a different story of “I.” In this story, by simply choosing to do so, one allows one’s “self” to be manipulated by others (or, in the case of self-hypnosis, by one’s own self). A view similar to this as concerns reported involuntariness of hypnotic trance is suggested in Lynn, Rhue, and Weekes (1990). I have read that people who are hypnotized to think that an ice cube burns them will actually develop blisters. Paul (1963) concludes that there may be some effect of suggestion on the formation of blisters, but Benham and Younger (2008) agree with Paul that any effects will generally have causes other than hypnosis itself.This is not to say that the hypnotic state is a pretense. On the contrary, it is a convincing story to the hypnotized subject, just as convincing as the normal story of “I.” Moreover, it produces measurable effects in physiology and on the brain, so much so that clinicians are using it more and more in their practices, for example, in complementing or replacing anesthesia in painful surgical operations. (p.87) For a recent review on hypnosis, including chapters on physiological and brain effects by Barabasz and Barabasz (2008) and by Oakley (2008), see The Oxford Handbook of Hypnosis (Nash & Barnier, 2008). See also the 2009 special issue of Contemporary Hypnosis on hypnotic analgesia introduced by Liossi, Santarcangelo, and Jensen (2009). For a review of the use of hypnosis in the treatment of acute and chronic pain, see Patterson and Jensen (2003). Hardcastle (1999) authoritatively devotes a chapter to expounding the extraordinary effectiveness of hypnosis, which she says surpasses all other methods of pain control. Marie-Elisabeth Faymonville (Faymonville et al., 1995) has performed thousands of heavy surgical procedures under only local anesthetic, using hypnosis. She told me that even though usually it is claimed that only a small percentage of people are sufficiently receptive to hypnosis for them to become insensitive to pain, in her own practice, among about 4,000 patients on whom she had operated, the hypnotic procedure had failed only in isolated cases.As a sampling of mention of hypnosis in the press: In April 2008 a British hypnotist submitted to a painful orthopedic operation where bone was sawed from his hand, without any anesthetic: http://news.bbc.co.uk/2/hi/uk_news/england/sussex/7355523.stm. Referring to the Montgomery et al. (2007) study using hypnosis in breast cancer operations: http://news.bbc.co.uk/2/hi/health/6969298.stm. Dr. Christina Liossi, from University of Wales, Swansea, using hypnosis with local anesthetic in cancer operations in children: http://news.bbc.co.uk/2/hi/health/3642764.stm; also see Hawkins, Liossi, Ewart, Hatira, and Kosmidis (2006).

(35.) These are examples of culturally bound syndromes (Simons & Hughes, 1985; Somera, 2006) where an individual will temporarily change his or her identity. Here is a description of latah, found in Malaysia and Indonesia:

A person exemplifying the Latah syndrome typically responds to a startling stimulus with an exaggerated startle, sometimes throwing or dropping a held object, and often uttering some improper word, usually “Puki!” (“Cunt!”), “Butol!” (“Prick!”), or “Buntut!” (“Ass!”), which may be embedded in a “silly” phrase. Further, he or she may, while flustered, obey the orders or match the orders or movements of persons nearby. Since spectators invariably find this funny, a known Latah may be intentionally startled many times each day for the amusement of others. (Simons, 1980)

Similar syndromes are found in Burma (yaun), Thailand (bah-tsche), Philippines (mali-mali, silok), Siberia (myriachit, ikota, amurakh), South-West Africa, Lapland (Lapp panic), Japan (Imu), and among French Canadians in Maine (jumping).It has been suggested that the phenomenon may be related to the belief that fright or startle can cause people to lose their soul, leading then to the intrusion of evil spirits (Kenny, 1983). This is interesting because it shows how an innate human startle reaction can be profoundly modified to serve cultural purposes. When this happens, people can espouse the syndrome so completely that they claim it is involuntary and that they are powerless to prevent it.There are more dramatic examples where one’s self is involuntarily modified. Often these can be attributed to a person being subjected to strong social or psychological stress. The Malay phenomenon of amok may be such a case. Here, typically, an outsider to a local group is frustrated by his situation or by social pressures and, (p.88) after a period of brooding, suddenly grabs a weapon and wildly attempts to kill everyone around him, afterward falling into a state of exhaustion and completely forgetting what happened. Despite its extreme violence, this behavior may correspond to a historically sanctioned tradition among Malay warriors that people consider is the appropriate response to personally intolerable conditions (Carr, 1978). Similarly indiscriminate homicidal attack behaviors are observed in other societies, in particular recently in the United States, and they may be culturally determined in different ways.Hypnosis, latah, and amok are just a few examples of short-lived changes in the self that can be observed in special circumstances, with other cases being possession trances, ecstasies, channeling provoked in religious cults, oracles, witchcraft, shamanism, or other mystical experiences. More permanent, highly dramatic involuntary modifications of the self can also be provoked by extreme psychological stress or physical abuse, as shown by Lifton (1986, 1989, 1991), for example, in the cases of people subjected to brainwashing by sects, in religious cults, and in war.There is also the case of dissociative identity disorder (formerly called multiple personality disorder). A person with DID may hear the voices of different “alters” and may flip from “being” one or other of these people at any moment. The different alters may or may not know of each other’s existence. The philosopher Ian Hacking, in a discussion of multiple personality disorder, Paris café garçons, and the gay community (Hacking, 1986), suggests that the surprising rise in incidence of DID/MPD over the past decades signals that this disorder is indeed a cultural phenomenon. This may be one explanation why the illness is considered by some to be very suspect. Under the view I am taking here, DID/MPD is a case where an individual resorts to splitting his or her identity in order to cope with extreme psychological stress. Each of these identities is as real as the other and as real as a normal person’s identity—since all are stories. See Hartocollis (1998) for a review of different viewpoints of MPD as a social construction.

(36.) I have also not mentioned the ethical, moral, and legal problems that these ideas pose (see Greene & Cohen, 2004), and I have not discussed the fascinating related idea that what we call free will is a construction similar to that of the self. According to Daniel Wegner (Wegner, 2003a, 2003b), the impression of having willed an action is a cognitive construction that we assemble from evidence that we gather about what occurred.

Notes:

(1.) Doing calculations with fluids was a project in the Amateur Scientist section of Scientific American, May 1966 (p. 128).

(2.) The Geniac machine was conceived by Edmund C. Berkeley, a computer scientist and antinuclear activist, author of the book Giant Brains or Machines That Think (1949). See http://www.oldcomputermuseum.com/geniac.html and especially http://www.computercollector.com/archive/geniac/for more information.

(3.) For machine translation, see the review by Hutchins (2003). Examples of the difficulty encountered in making a computer understand natural language are given in the Supplements to this book on http://whyred.kevin-oregan.net

(4.) In 2010 you could buy the SYSTRAN program for your own PC, starting at only 39 euros, and choose from the Professional Premium, the Professional Standard, the Personal, the Web, the Office Translator, and other versions of their software!

(5.) Weizenbaum’s ELIZA program is on http://courses.cs.vt.edu/∼cs3604/lib/Ethics/eliza.html, the postmodern generator on http://www.elsewhere.org/pomo. The generator was written by Andrew C. Bulhad using the Dada Engine.

(6.) The best known example of an attempt to compile “common sense” is Douglas Lenat’s “CYC” project, which started in the 1980s at Atari corporation and has now become a company, Cycorp. Cyc is an ongoing effort to accumulate human consensual knowledge into a database of more than a million assertions which can be used to understand, for example, newspaper articles or encyclopedia entries. On http://www.cyc.com/ the company says: “Cycorp’s vision is to create the world’s first true artificial intelligence, having both common sense and the ability to reason with it.”

(7.) This point was forcefully made in 1972 by Hubert Dreyfus in his influential book What Computers Can’t Do: A Critique of Artificial Intelligence, updated in 1992 to What Computers Still Can’t Do: A Critique of Artificial Reason (Dreyfus, 1972, 1992).

(8.) The company iRobot, inspired by the work of Rodney Brookes, has a Web site http://www.roombavac.com/. They say they deliver “… innovative robots that are (p.84) making a difference in people’s lives. From cleaning floors to disarming explosives, we constantly strive to find better ways to tackle dull, dirty and dangerous missions—with better results.” They say that if you have iRobot Scooba: “check mopping off your list.” With iRobot Roomba: “check vacuuming off your list.”

(9.) See http://www.epigenetic-robotics.org for ongoing activities in this field, with bibliographic references. For a survey of the related field of developmental robotics, see Lungarella, Metta, Pfeifer, and Sandini (2003).

(10.) See http://www.lira.dist.unige.it/ for BabyBot, created by Giorgio Metta when he was working in the COG project at MIT. It is now part of the larger, European RobotCub project (http://www.robotcub.org/). Videos can be seen on that site as well as on http://www.lira.dist.unige.it/babybotvideos.htm. An overview of some other current efforts in autonomous and developmental robotics can be found on the Web site for the supplements to this book: http://whyred.kevin-oregan.net

(11.) See Fitzpatrick and Metta (2003). The notion of “affordance” is due to the ecological psychologist J. J. Gibson.

(12.) For a very critical article in The Skeptic about the possibilities of AI, see Kassan (2006).

(13.) For a useful bibliography on different approaches to the self, see the Web site maintained by Shaun Gallagher: http://www.philosophy.ucf.edu/pi/ or his excellent review (Gallagher, 2000). See also the special issue of Annals of the New York Academy of Sciences in Vol. 1001, 2003.

(14.) Kohut, 2009.

(15.) I based this quick overview on Gallagher (2000).

(16.) James, 1890.

(17.) Neisser, 1988.

(18.) Dennett, 1991, 1992.

(19.) Gallagher, 2000.

(20.) Metzinger, 2004, 2005.

(21.) Part of this long list is taken from Strawson (1999); see the book by Gallagher and Shear (1999) for a recent interdisciplinary survey of many approaches to the self.

(22.) Here I am going to follow the two dimensions of classification proposed by Vierkant (2003).

(23.) This classification is based on that proposed by Bekoff and Sherman (2004). They use the umbrella term “self-cognizance” to designate the whole spectrum of ways that a system can have knowledge of itself. In their classification they use the three terms “self-referencing,” “self-awareness,” and “self-consciousness” instead of the terms “self-distinguishing,” “self-knowledge,” and “knowledge of self-knowledge.” I prefer my terms because they are more neutral as regards the question of whether any phenomenal quality might be associated with the notion of self, and because they avoid the difficult terms “awareness” and “consciousness.”

(24.) See Edsinger and Kemp (2006).

(25.) As suggested by cognitive psychologists Simon Baron-Cohen (Baron-Cohen, 1997; Leslie, 1994). For robots that imitate humans’ social development, see Breazeal and Scassellati (2002) and Scassellati (2003). See Triesch, Teuscher, Deak, and Carlson (2006) for a discussion of how gaze following might be a learned faculty.

(26.) Because this social self is such an important concept in our society, it is discussed from a myriad of different points of view, in the popular press, in psychological counseling, in psychoanalysis and in academic psychology, sociology, and anthropology. (p.85) In these domains, a plethora of concepts have been defined, including for example self-actualization; self-awareness; the self-concept; self control; self development; self disclosure; self-efficacy; self-esteem; self harm; self help; self-identity; self image; self monitoring; self-perception; attitude change; self-realization; self-regulated learning; self talk.

(27.) The comparison Hume makes actually concerns what he calls the “soul” (Hume, 1793/2003, Book 2.i.vi). Derek Parfit also makes the comparison with a nation or club (Parfit, 1986).

(28.) Why does this abstraction exist and why is it so useful? To understand this, we can appeal to the concept of “meme,” invented by the biologist Richard Dawkins. A meme is the social or cognitive equivalent of a gene. Just as there are genes corresponding to traits replicated in biological reproduction, there are “memes,” which are ideas, skills, or behaviors that stick in the mind, serve a social purpose, and so become part of human society. Memes are “replicators”; they are adapted to human cognitive capacities. They are easy to remember and stable enough to be accurately propagated through social interactions and from one generation to the next. Furthermore, memes may have the self-validating property that they influence the structure of society so as to favor their own preservation and propagation. As suggested by writer Susan Blackmore, the self may be one of the most fundamental memes in human society. The self is a concept which each of us uses in order to understand the goals and desires not only of our own selves but of our fellow humans (Blackmore, 2000; Dawkins, 1976).

(29.) Dennett, 1992.

(30.) Martin Kusch cited by Vierkant (2003) mentions the doubly self-referring and self-validating roles of social constructs like money and selves, though there may be a difference in the way Kusch uses these terms.

(31.) It is hard to believe that this kind of circularity could provide something that seems so real to us. Surely I can sense my existence—I can perceive myself talking and thinking. If I am just a story, then how can this story perceive itself talking and thinking? It’s as though it made sense to say that Sherlock Holmes, who we all know is just a fictional character, really considers himself to have an identity.To make these ideas more plausible, consider another example of a self-referring and self-validating social construct, which is the notion of country or nation. A group of people can proclaim itself as a nation, and from that moment on, the nation takes on a real existence on the international scene. Furthermore there is a real sense in which nations have “selves.” They have national identities, they build alliances, they make pacts and wage wars, they communicate and trade. They can be offended, envious of others’ resources, aggressive, or peaceful. They have national memories and even what might be called split personalities when under stress (e.g., French Resistance under German World War II occupation).Thus, judging from international events and talk in the media, and judging from the dramatic excesses that nations are willing to engage in to preserve and extend themselves, nations are real and really have identities. Because nations generally “express themselves” in different ways than humans, with no single “voice,” the analogy with a human “self” is not perfect, but if we were somehow able to go out and actually interview a nation, and ask it if it really felt it had a self, an identity, and if it really felt it was real, it would reply “yes.” (These ideas are presented admirably by Dennett, 1989.)

(32.) It may be that not just the expression of pain but also pain itself may be influenced by cultural factors, but the evidence is difficult to interpret (see Rollman, 1998).

(33.) Seminal experiments showing unconscious influence of social norms on behavior were done by the social psychologist John Bargh, now at Yale (see Bargh & Chartrand, 1999). In these experiments, participants were told that they were being tested on language proficiency in a sentence unscrambling task. They were given five words (e.g., “They her respect see usually”) and were required to write down a sentence which made use of four of those words. Actually, however, this supposed language proficiency task was used to “prime” participants with different cultural stereotypes. For example, in one experiment, the words used in the sentence unscrambling task were those that were either associated with politeness (e.g., honor, considerate, appreciate, patiently, cordially), or with rudeness (e.g., brazen, impolitely, infringe, obnoxious). It was found that when the participants had been primed for rudeness, on leaving the experimental room they were more likely to interrupt the experimenter who was ostensibly talking to a colleague. In another experiment the sentence unscrambling task used words that primed for an “elderly” prototype (e.g., worried, Florida, old, lonely, gray) versus a neutral prototype (e.g., thirsty, clean, private). It was found that the time it took for the participants to cross the hall on leaving the experiment was longer when they were primed with “elderly” words. A different experiment showed that participants primed for the “elderly” prototype could also not remember as many details about the room in which they participated in the experiment.

(34.) Hypnosis is so easy to induce that any basic text on hypnosis will provide an induction technique that can be used by a complete novice to hypnotize someone else. The Web site http://www.hypnosis.com/ provides a wealth of information, including induction scripts and numerous scripts purported to cure anything from AIDS to constipation to procrastination. Some hypnotists claim to be able to hypnotize people (even people who have not previously been sensitized) in a matter of seconds.The ease with which hypnosis can be induced is consistent with the idea that hypnotic trance is a matter of choosing to play out a role that society has familiarized us with. The suggestion that hypnosis is such a culturally bound phenomenon comes from the very high correlation between hypnotizability and a person’s belief that hypnosis works. Conversely, it may be impossible to hypnotize someone from a different culture, who has never heard of hypnosis. “Playing the part of being hypnotized” would seem to be a culturally accepted way of exploring a different story of “I.” In this story, by simply choosing to do so, one allows one’s “self” to be manipulated by others (or, in the case of self-hypnosis, by one’s own self). A view similar to this as concerns reported involuntariness of hypnotic trance is suggested in Lynn, Rhue, and Weekes (1990). I have read that people who are hypnotized to think that an ice cube burns them will actually develop blisters. Paul (1963) concludes that there may be some effect of suggestion on the formation of blisters, but Benham and Younger (2008) agree with Paul that any effects will generally have causes other than hypnosis itself.This is not to say that the hypnotic state is a pretense. On the contrary, it is a convincing story to the hypnotized subject, just as convincing as the normal story of “I.” Moreover, it produces measurable effects in physiology and on the brain, so much so that clinicians are using it more and more in their practices, for example, in complementing or replacing anesthesia in painful surgical operations. (p.87) For a recent review on hypnosis, including chapters on physiological and brain effects by Barabasz and Barabasz (2008) and by Oakley (2008), see The Oxford Handbook of Hypnosis (Nash & Barnier, 2008). See also the 2009 special issue of Contemporary Hypnosis on hypnotic analgesia introduced by Liossi, Santarcangelo, and Jensen (2009). For a review of the use of hypnosis in the treatment of acute and chronic pain, see Patterson and Jensen (2003). Hardcastle (1999) authoritatively devotes a chapter to expounding the extraordinary effectiveness of hypnosis, which she says surpasses all other methods of pain control. Marie-Elisabeth Faymonville (Faymonville et al., 1995) has performed thousands of heavy surgical procedures under only local anesthetic, using hypnosis. She told me that even though usually it is claimed that only a small percentage of people are sufficiently receptive to hypnosis for them to become insensitive to pain, in her own practice, among about 4,000 patients on whom she had operated, the hypnotic procedure had failed only in isolated cases.As a sampling of mention of hypnosis in the press: In April 2008 a British hypnotist submitted to a painful orthopedic operation where bone was sawed from his hand, without any anesthetic: http://news.bbc.co.uk/2/hi/uk_news/england/sussex/7355523.stm. Referring to the Montgomery et al. (2007) study using hypnosis in breast cancer operations: http://news.bbc.co.uk/2/hi/health/6969298.stm. Dr. Christina Liossi, from University of Wales, Swansea, using hypnosis with local anesthetic in cancer operations in children: http://news.bbc.co.uk/2/hi/health/3642764.stm; also see Hawkins, Liossi, Ewart, Hatira, and Kosmidis (2006).

(35.) These are examples of culturally bound syndromes (Simons & Hughes, 1985; Somera, 2006) where an individual will temporarily change his or her identity. Here is a description of latah, found in Malaysia and Indonesia:

A person exemplifying the Latah syndrome typically responds to a startling stimulus with an exaggerated startle, sometimes throwing or dropping a held object, and often uttering some improper word, usually “Puki!” (“Cunt!”), “Butol!” (“Prick!”), or “Buntut!” (“Ass!”), which may be embedded in a “silly” phrase. Further, he or she may, while flustered, obey the orders or match the orders or movements of persons nearby. Since spectators invariably find this funny, a known Latah may be intentionally startled many times each day for the amusement of others. (Simons, 1980)

Similar syndromes are found in Burma (yaun), Thailand (bah-tsche), Philippines (mali-mali, silok), Siberia (myriachit, ikota, amurakh), South-West Africa, Lapland (Lapp panic), Japan (Imu), and among French Canadians in Maine (jumping).It has been suggested that the phenomenon may be related to the belief that fright or startle can cause people to lose their soul, leading then to the intrusion of evil spirits (Kenny, 1983). This is interesting because it shows how an innate human startle reaction can be profoundly modified to serve cultural purposes. When this happens, people can espouse the syndrome so completely that they claim it is involuntary and that they are powerless to prevent it.There are more dramatic examples where one’s self is involuntarily modified. Often these can be attributed to a person being subjected to strong social or psychological stress. The Malay phenomenon of amok may be such a case. Here, typically, an outsider to a local group is frustrated by his situation or by social pressures and, (p.88) after a period of brooding, suddenly grabs a weapon and wildly attempts to kill everyone around him, afterward falling into a state of exhaustion and completely forgetting what happened. Despite its extreme violence, this behavior may correspond to a historically sanctioned tradition among Malay warriors that people consider is the appropriate response to personally intolerable conditions (Carr, 1978). Similarly indiscriminate homicidal attack behaviors are observed in other societies, in particular recently in the United States, and they may be culturally determined in different ways.Hypnosis, latah, and amok are just a few examples of short-lived changes in the self that can be observed in special circumstances, with other cases being possession trances, ecstasies, channeling provoked in religious cults, oracles, witchcraft, shamanism, or other mystical experiences. More permanent, highly dramatic involuntary modifications of the self can also be provoked by extreme psychological stress or physical abuse, as shown by Lifton (1986, 1989, 1991), for example, in the cases of people subjected to brainwashing by sects, in religious cults, and in war.There is also the case of dissociative identity disorder (formerly called multiple personality disorder). A person with DID may hear the voices of different “alters” and may flip from “being” one or other of these people at any moment. The different alters may or may not know of each other’s existence. The philosopher Ian Hacking, in a discussion of multiple personality disorder, Paris café garçons, and the gay community (Hacking, 1986), suggests that the surprising rise in incidence of DID/MPD over the past decades signals that this disorder is indeed a cultural phenomenon. This may be one explanation why the illness is considered by some to be very suspect. Under the view I am taking here, DID/MPD is a case where an individual resorts to splitting his or her identity in order to cope with extreme psychological stress. Each of these identities is as real as the other and as real as a normal person’s identity—since all are stories. See Hartocollis (1998) for a review of different viewpoints of MPD as a social construction.

(36.) I have also not mentioned the ethical, moral, and legal problems that these ideas pose (see Greene & Cohen, 2004), and I have not discussed the fascinating related idea that what we call free will is a construction similar to that of the self. According to Daniel Wegner (Wegner, 2003a, 2003b), the impression of having willed an action is a cognitive construction that we assemble from evidence that we gather about what occurred.