Kim Binsted

Kim Binsted

Kim Binsted, Ph.D., received her B.Sc. in Physics at McGill (1991), and her Ph.D. in Artificial Intelligence from the University of Edinburgh (1996).

In 2002, she joined the faculty of the Information and Computer Sciences Department at the University of Hawaii, where she does research on artificial intelligence, human-computer interfaces, and long-duration human space exploration. Dr. Binsted is a co-investigator at the UH-NASA Astrobiology Institute, which formed in late 2003. She was a NASA Summer Faculty Fellow at Ames Research Center in 2003 and 2004, where she worked on sub-vocal speech recognition technology in the Neuroengineering Lab. She was Chief Scientist on the FMARS 2007 Long Duration Mission, a four-month Mars exploration analogue on Devon Island in the Canadian High Arctic. In 2009, she spent her sabbatical as a visiting scientist at the Canadian Space Agency (CSA), working on the CSA's planetary analogues program.

Dr. Binsted is now a co-investigator on a space food study at HI-SEAS (Hawaii Space Exploration Analog and Simulation), and the principal investigator on the next three years of HI-SEAS missions, which will include four, eight and twelve month missions.

Abstract

Tongue-Tied: The Inadequacy of Language for Interstellar Communication

Our goal is to construct a message that another intelligence, completely unfamiliar with humankind, would be able to understand. It is assumed that the recipient is so far away that interaction in any reasonable time frame would be unfeasible, so the message must be entirely self-contained. Furthermore, since we know nothing about our target audience, we would like to keep our assumptions about their scientific knowledge, physiology, cognitive processes, and so on, to a minimum. If they have to be very human-like in order to understand the message, we have failed at our task. All we should demand of our audience is that they have the technical wherewithal to build equipment capable of receiving our message.

Even if the recipients were human, we could not necessarily expect them to be able to decode an arbitrary piece of English text. Historically, human archaeologists have had great difficulty decoding natural language texts written by other humans. In every case analogous to the interstellar decoding problem, to decipher the text they have relied on intermediary encodings (such as the Rosetta Stone, or modern versions of ancient languages) and a familiarity with human nature (e.g., that writings on a tomb are likely to be about the entombed person’s life).

So, the message we send should not be arbitrary. Rather, it should be designed to be understood by an intelligent recipient with minimal knowledge of the sender. It has been argued that a large corpus of the written form of a natural language (English, French, or Chinese, for example) would serve this purpose. The main advantage to this approach is that there are large bodies of existing text that could serve as the bulk of the message. The question is, would it be understood? We believe that a natural-language text message, no matter how large and carefully composed it is, is extremely unlikely to be decodable by a non-human intelligence.