$$\require{cancel}$$

# 19.17 Recognizing a Message

As light travels through space, its intensity decreases by the square of the distance from which it came. Since radio waves are a form of light, this means that when a radio signal is detected over a large distance, it is dramatically weaker than when it was transmitted. This becomes a substantial obstacle when sending or receiving signals over the vast distances of space. In particular, the radio signal becomes so weak that it is difficult to detect the signal against background cosmic noise sources. It is like trying to hear a whispering child in a room full of shouting adults.

When creating a radio message to communicate through space, the message should be sent over a narrow bandwidth and code information through variations in radio frequency. When compared to cosmic sources of radio noise, a narrow bandwidth will provide the best chances of detection. That's because natural sources of cosmic radiation don't have bandwidth as small as a few Hertz. A team at the University of Cambridge discovered a signal with these characteristics in 1967. The discoverers were initially mystified as to the origin of the periodic signals. While trying to find an explanation for these radio pulses, the graduate student who made the initial observation, Jocelyn Bell, jokingly called the signals LGM1, LGM2, etc for "little green men." When the local media picked up on the joke, it was published as a splashy story about aliens. But instead of aliens, the pulsing signals was actually a natural source of radio signals coming from pulsars. They have a periodic signal, which is one of the simplest kinds of variation. It turns out, however, that a periodic signal, like that emitted by pulsars, actually carries very little information, a single frequency or number. In order to create a message with a significant amount of information, the signal must be more complicated.

The optimal method for transmitting a radio message is by pulsing the signal. This way the radio energy is sent in small, concentrated bursts and is easier to pick out from background noise. Pulsed signals correspond directly to the digital way that information is stored and transmitted on FM radio stations, CDs, and computers. Just as the information stored on our computers can be large and extremely complex, an enormous amount of information can be conveyed with a sufficient number of binary elements, or bits. Each bit is a signal which is either 1 or 0, on or off. We might hope that any creature with sense receptors is aware of some version of the distinction between bright and dark, hot and cold, loud and quiet, or 1 and 0.

In order to increase our chances of successfully recognizing an extraterrestrial message, it is helpful to consider what message we would send into space. We might first think of sending an eloquent statement describing our hopes and dreams, but in doing so we risk falling into an anthropocentric trap. To transmit a universal message, the message should be as free from cultural influences as possible. Some have suggested using a language based on mathematics, arguing that it would be able to convey many universal truths. For example, artificial intelligence expert Marvin Minsky has explored the behaviors of all possible artificial processes by creating all possible computers and their programs. This is the computational equivalent of the Miller-Urey experiments, taking the simplest processes and seeing what complexity arises. Minsky created thousands of computational machines, most of which stopped without accomplishing anything. A few got trapped in circles and senselessly repeated the same steps over and over. However, a few performed a counting operation; essentially, they "invented" arithmetic! Dutch mathematician Hans Freudenthal has developed LINCOS, or Lingua Cosmica, a language designed for universal discourse. This language is designed to convey mathematics and physics in a coded form.

Have the messages we have sent so far matched the ideal of a universal language? In 1974, we used the Arecibo radio~telescope to beam a symbolic message in the direction of the globular~star~cluster M 13. The actual message consisted of 1,679 bits of information, or on and off pulses. The logic behind this number of pulses was based on the hope that an (alien) mathematician would notice that 1,679 is only divisible by 23 and 73, two prime numbers. This would suggest creating a two-dimensional pattern with 23 and 73 as either the number of columns and rows. If the alien arranged the bits as 73 columns of 23 bits each, he or she or it would find no discernable pattern. However, if the alien selected the second choice, 23 columns of 73 bits each, a more distinctive pattern is revealed (where the binary numbers are replaced by their equivalents in a visual field, 0 = dark and 1 = bright). The pattern is clearly nonrandom. But most people on Earth, when asked, can't even decode the message! It is a strange visual mixture of crude cartoons and binary representations of the chemical constituents of life. Based on its incomprehensibility to most humans, it is questionable how clear its true meaning would be to an alien intelligence. Regardless, we won't find out for a while since, even traveling at the speed of light, it will take the message 54,000 years to get to M 13!

If other civilizations in the universe have been as forthcoming as ours by sending a message, we may not be able to decipher it. However, we are fairly confident that we will be, at the very least, able to recognize a message of alien origin. For example, even if the alien recipients of our Arecibo message cannot interpret the symbolic images, or have no visual sense at all, they could recognize the intelligent intent of the message. In particular, a pattern of 0's and 1's is highly nonrandom; it has less than a 1 in 105 chance of happening by chance. It is also possible for an expert in mathematics to find a number of patterns in a binary-coded picture such as this. In the case that the message was coded in a language rather than as a picture, a civilization would be able to use pattern recognition techniques that would take advantage of the naturally occurring redundancy of natural languages. For example, the written English language is over 75% redundant, only 1 in 4 letters on average conveys any information at all.