Scientific American Mind March/April 2011
Living in a Dream World (p 24)
Daydreaming can boost problem-solving and creativity and inspire art and science. But what happens when it becomes an addiction? In this month's Scientific American Mind, Josie Glausiusz investigates what happens when our minds wander—and when daydreaming starts to take over.
Einstein's theory of special relativity and the films of Tim Burton are said to have been inspired by reveries. More generally, daydreaming is thought to have a creative function, perhaps because the waking brain is never at rest. But there is a fine line between mundane musings and incessant, extravagant fantasizing. Michael Kane, a cognitive psychologist at the University of North Carolina at Greensboro, says that daydreaming can be beneficial or harmful, depending on the context and the daydreamer's own goals at the time.
The "default network" is an area of the brain thought to be involved in autobiographical mental imagery and sense of self; it may also be associated with daydreaming. Malfunctions in the default network have been linked to schizophrenia and depression and may impair the ability to daydream. Cordellia Amethyste Rose, who started Wild Minds, an online forum for people who cannot stop daydreaming, describes her compulsive daydreaming as an addiction. "I'm like an alcoholic with an unlimited supply of booze everywhere I go," she says.
Your Avatar, Your Guide (p 58)
Using a dopplegänger, an avatar—a digital doll that vividly resembles you—can have powerful effect on users. Writing in Scientific American Mind, Samantha Murphy demonstrates the potential of avatars beyond gaming, and as vehicles for self-improvement.
Avatars have long been known to influence behavior and thought. Recent brain imaging research shows that people project themselves onto avatars, explaining the potential of digital doppelgängers for treating phobias and anxiety disorders and in helping people lose weight and make better financial choices. Murphy notes that "avatar therapy" is poised to be on the rise in the next few years. Avatar applications may soon become commercial. One avatar-based exercise program is already in early-stage development.
While many uses of avatars are positive, some have sinister undertones. Research suggests that avatars could be used to make us feel more favorable towards a particular politician or influence us into owning or using certain products. Murphy warns that implications of avatars may be far-reaching and unpredictable.
Where Are the Talking Robots? (p 44)
Teaching robots to interpret and use language had proven to be a difficult task. As Joshua K. Hartshorne writes in Scientific American Mind, the problem is not even technological as much as linguistic—understanding the nuances of language is simple for us, but teaching it and programming a machine to learn it, is harder than anyone imagined.
Our ability to readily understand the meaning of ambiguous words in context is the result of millions of years of evolution. "And we accomplish these feats without knowing how we do so," writes Hartshorne, "much less how to teach the skill to an artificial being." To help with bewildering double-meanings and confusing grammar in spoken language, scientists are creating complex statistical programs in which computers break down language into pairs of words, learning which words tend to come before which other words. However, improving this methodology by teaching machines longer phrases is computationally cumbersome. For a language containing a paltry 10,000 words, a machine would need to learn a trillion different three-word combinations. At six words—which is still not long enough to produce natural speech—a computer would need to devour 10 trillion exabytes of information. In 2009, all the digital information on Earth rang in at just 500 exabytes.
One cutting-edge approach to teaching robots to talk taps into global networks. "To help give robots the benefit of real-world experience while bridging the data gap, several recent Web-based projects have sought to enlist the public," says Hartshorne.
About Scientific American
Founded in 1845, Scientific American is the oldest continuously published magazine in the US and the leading authoritative publication for science and technology in the general media. Together with scientificamerican.com and 14 local language editions around the world it reaches more than nine million readers. Other titles include Scientific American Mind and Spektrum der Wissenschaft in Germany. Scientific American is published by Springer Nature, a leading global research, educational and professional publisher, home to an array of respected and trusted brands providing quality content through a range of innovative products and services. Springer Nature was formed in 2015 through the merger of Nature Publishing Group, Palgrave Macmillan, Macmillan Education and Springer Science+Business Media.