Autocorrect Was Invented in the U.S. for a Secret Chinese Computer

The strange origin of autocorrect during the development of an American secret Chinese computer.

Every time autocomplete, or its aggravating variant autocorrect, declares war on you, now you’ll know who to thank. Its inventor’s name was Samuel Hawks Caldwell, and it shall live in infamy. Thanks a lot, pal. Caldwell didn’t set out to annoy generations of computer users. His goal was something else completely, and it’s an odd story.

Back in the late 1950s, China was not the technology society it is today. In those early days of computing it was viewed as backwards, and possibly doomed to remain that way due to the difficulties of making their pictographic words understandable to computers.

Moved to action by altruistic ambitions electrical engineer Caldwell began work on a machine called the “Sinotype.” He viewed it as a gift to the Chinese people, writing later, “Many will wonder why this work was ever done or why our military establishment devoted substantial funds and attention to the project,’ he later wrote. ‘The answer to this question seems simple and clear. In selling the idea to the military authorities, the writer had only one real argument … to the effect that a machine for composing Chinese would improve communication among men, and that no improvement of communication ever harmed the cause of peace among men.” His backers were not quite so high-minded: The U.S. Army and Air Force, and the Carnegie Foundation saw the Linotype as a way to disseminate anti-Communist propaganda in Chinese on a scale never before imaginable.

You would think expertise in the Chinese language would be a prerequisite to developing a Chinese computer, but Caldwell didn’t know he language. After all, what he taught at MIT was electrical engineering. So he talked with native Chinese students attending the school and learned that they’d been taught to form words using a specific sequence of brush strokes, much the same way that American children are taught to write “t” by drawing a vertical line and then topping it with a horizontal one. To Caldwell, this meant that “Chinese has a ‘spelling.’

Caldwell enlisted a professor of Far Eastern Languages at Harvard, Lien-Sheng Yang to study the structure of Chinese characters, stroke-by-stroke. It turned out that there were just a handful of common sequences that eventually diverged into specific characters, and that computer analysis could identify the intended character after only five or six strokes, regardless of its final complexity. You see where this is going, right? Autocorrect.

“Autocorrect” in Traditional Chinese.

To Caldwell, this was the breakthrough that would allow him to create his dreamed-of Chinese computer. (To you and me, it’s something else.) He and Yang derived a vocabulary of 22 strokes from which roughly 2,000 Chinese words could be constructed. Each of these was awarded a place on the Sinotype keyboard. (22 was the number of keys on a Western keyboard.)

The Sinotype was nearly announced to the world in the summer of 1959, with the U.S. under the pressure of fears the Chinese would beat them to it. Unsure that it would work as advertised, and not wanting to risk embarrassment internationally, the Eisenhower administration hesitated, and the moment passed.

Caldwell (standing) in 1949 (AP)

Caldwell died in 1960, and his computer was renamed a few times, as the “Chi-coder,” the “Ideographic Encoder,” and the “Sinotype II,” which moved away from a stroke-based keyboard to the popular Pinyin layout.

And what remains of Caldwell’s work? Well, you know it, you probably hate it, and there we have it: autocomplete and its annoying cousin autocorrect.

Related Articles

Major study: Drug overdoses over a 38-year period reveal hidden trends

It's just the current cycle that involves opiates, but methamphetamine, cocaine, and others have caused the trajectory of overdoses to head the same direction

From the study:
Surprising Science
  • It appears that overdoses are increasing exponentially, no matter the drug itself
  • If the study bears out, it means that even reducing opiates will not slow the trajectory.
  • The causes of these trends remain obscure, but near the end of the write-up about the study, a hint might be apparent
Keep reading Show less

Why "nuclear pasta" is the strongest material in the universe

Through computationally intensive computer simulations, researchers have discovered that "nuclear pasta," found in the crusts of neutron stars, is the strongest material in the universe.

Accretion disk surrounding a neutron star. Credit: NASA
Surprising Science
  • The strongest material in the universe may be the whimsically named "nuclear pasta."
  • You can find this substance in the crust of neutron stars.
  • This amazing material is super-dense, and is 10 billion times harder to break than steel.

Superman is known as the "Man of Steel" for his strength and indestructibility. But the discovery of a new material that's 10 billion times harder to break than steel begs the question—is it time for a new superhero known as "Nuclear Pasta"? That's the name of the substance that a team of researchers thinks is the strongest known material in the universe.

Unlike humans, when stars reach a certain age, they do not just wither and die, but they explode, collapsing into a mass of neurons. The resulting space entity, known as a neutron star, is incredibly dense. So much so that previous research showed that the surface of a such a star would feature amazingly strong material. The new research, which involved the largest-ever computer simulations of a neutron star's crust, proposes that "nuclear pasta," the material just under the surface, is actually stronger.

The competition between forces from protons and neutrons inside a neutron star create super-dense shapes that look like long cylinders or flat planes, referred to as "spaghetti" and "lasagna," respectively. That's also where we get the overall name of nuclear pasta.

Caplan & Horowitz/arXiv

Diagrams illustrating the different types of so-called nuclear pasta.

The researchers' computer simulations needed 2 million hours of processor time before completion, which would be, according to a press release from McGill University, "the equivalent of 250 years on a laptop with a single good GPU." Fortunately, the researchers had access to a supercomputer, although it still took a couple of years. The scientists' simulations consisted of stretching and deforming the nuclear pasta to see how it behaved and what it would take to break it.

While they were able to discover just how strong nuclear pasta seems to be, no one is holding their breath that we'll be sending out missions to mine this substance any time soon. Instead, the discovery has other significant applications.

One of the study's co-authors, Matthew Caplan, a postdoctoral research fellow at McGill University, said the neutron stars would be "a hundred trillion times denser than anything on earth." Understanding what's inside them would be valuable for astronomers because now only the outer layer of such starts can be observed.

"A lot of interesting physics is going on here under extreme conditions and so understanding the physical properties of a neutron star is a way for scientists to test their theories and models," Caplan added. "With this result, many problems need to be revisited. How large a mountain can you build on a neutron star before the crust breaks and it collapses? What will it look like? And most importantly, how can astronomers observe it?"

Another possibility worth studying is that, due to its instability, nuclear pasta might generate gravitational waves. It may be possible to observe them at some point here on Earth by utilizing very sensitive equipment.

The team of scientists also included A. S. Schneider from California Institute of Technology and C. J. Horowitz from Indiana University.

Check out the study "The elasticity of nuclear pasta," published in Physical Review Letters.

How a huge, underwater wall could save melting Antarctic glaciers

Scientists think constructing a miles-long wall along an ice shelf in Antarctica could help protect the world's largest glacier from melting.

Image: NASA
Surprising Science
  • Rising ocean levels are a serious threat to coastal regions around the globe.
  • Scientists have proposed large-scale geoengineering projects that would prevent ice shelves from melting.
  • The most successful solution proposed would be a miles-long, incredibly tall underwater wall at the edge of the ice shelves.

The world's oceans will rise significantly over the next century if the massive ice shelves connected to Antarctica begin to fail as a result of global warming.

To prevent or hold off such a catastrophe, a team of scientists recently proposed a radical plan: build underwater walls that would either support the ice or protect it from warm waters.

In a paper published in The Cryosphere, Michael Wolovick and John Moore from Princeton and the Beijing Normal University, respectively, outlined several "targeted geoengineering" solutions that could help prevent the melting of western Antarctica's Florida-sized Thwaites Glacier, whose melting waters are projected to be the largest source of sea-level rise in the foreseeable future.

An "unthinkable" engineering project

"If [glacial geoengineering] works there then we would expect it to work on less challenging glaciers as well," the authors wrote in the study.

One approach involves using sand or gravel to build artificial mounds on the seafloor that would help support the glacier and hopefully allow it to regrow. In another strategy, an underwater wall would be built to prevent warm waters from eating away at the glacier's base.

The most effective design, according to the team's computer simulations, would be a miles-long and very tall wall, or "artificial sill," that serves as a "continuous barrier" across the length of the glacier, providing it both physical support and protection from warm waters. Although the study authors suggested this option is currently beyond any engineering feat humans have attempted, it was shown to be the most effective solution in preventing the glacier from collapsing.

Source: Wolovick et al.

An example of the proposed geoengineering project. By blocking off the warm water that would otherwise eat away at the glacier's base, further sea level rise might be preventable.

But other, more feasible options could also be effective. For example, building a smaller wall that blocks about 50% of warm water from reaching the glacier would have about a 70% chance of preventing a runaway collapse, while constructing a series of isolated, 1,000-foot-tall columns on the seafloor as supports had about a 30% chance of success.

Still, the authors note that the frigid waters of the Antarctica present unprecedently challenging conditions for such an ambitious geoengineering project. They were also sure to caution that their encouraging results shouldn't be seen as reasons to neglect other measures that would cut global emissions or otherwise combat climate change.

"There are dishonest elements of society that will try to use our research to argue against the necessity of emissions' reductions. Our research does not in any way support that interpretation," they wrote.

"The more carbon we emit, the less likely it becomes that the ice sheets will survive in the long term at anything close to their present volume."

A 2015 report from the National Academies of Sciences, Engineering, and Medicine illustrates the potentially devastating effects of ice-shelf melting in western Antarctica.

"As the oceans and atmosphere warm, melting of ice shelves in key areas around the edges of the Antarctic ice sheet could trigger a runaway collapse process known as Marine Ice Sheet Instability. If this were to occur, the collapse of the West Antarctic Ice Sheet (WAIS) could potentially contribute 2 to 4 meters (6.5 to 13 feet) of global sea level rise within just a few centuries."