Encryption algorithms in survival evidence

The parapsychological literature contains a variety of phenomena suggesting, with various degrees of persuasiveness, that some aspect of the human psyche survives the death of the brain. Among them are apparitions, cases suggestive of reincarnation, dreams, hauntings, mediumship, NDEs (near-death experiences), and OBEs (out-of-body experiences). The very idea of life after death is not free of potential philosophical (as distinct from empirical) problems: Flew (1987), (Geach (1987), and Penelhum (1987) argue that transfer of memories or personality traits are not sufficient to identify some incorporeal entity with some once-living person. Nevertheless, these arguments are far from conclusive, and experiments that shed light on this issue are interesting because of their implications.

One proposed method works as follows. A person picks some phrase (ideally, one that is difficult to guess), makes a record of it in a secure location, and, once he is deceased, attempts (supposing of course that "he" has "survived") to communicate this phrase to someone who is living. Living persons who receive this message are able to compare it to the record and, if it matches (a "hit"), have evidence for some sort of survival of death. A well-known related (but not a sealed-message) attempt was made by Harry Houdini, who tried to get a message to his deceased wife. This method has at least two potential pitfalls: cheating (i.e., the living person discovers the phrase through normal means), or "super-ESP" (clairvoyance of the time when the person was alive and knew the phrase).

Proving the existence of super-ESP is as important, or almost as important, as proving the existence of life-after-death; so the elimination of cheating becomes an important goal in these experiments. Potential solutions involve increasing security around the record of the message (putting it in a safe at a well-known law firm, and so forth), but with the complexity of the scheme, the potential for cheating somewhere along the line increases. Another fatal problem is that once the package is opened to check someone's attempt at the message, the experiment is finished. Several variants of this "sealed message" test have been used.

Thouless (1946-1949) proposed a hand-encryption scheme whereby the original message would be encoded and destroyed. This scheme has the advantage that (1) many attempts at "guessing" the message can be made, and (2) there is no permanent record of the original message itself to serve as a target for ESP by living persons. Stevenson (1968, 1976) attempted to improve on this method by eliminating the need for encrypting the message in favor of a combination padlock scheme.

Computer Cryptography as an Improved Method

I propose an idea that will eliminate the problem of cheating from this sort of experiment and will also improve on several other aspects of the previous two schemes. It is based on the similarity between this problem and the issue of identity authentication in computer science. The key is never to store the message in plain text form, thus avoiding the possibility of cheating. The method depends on the existence of trap-door or one-way algorithms that can be used to encode a message but cannot be reversed. The method works as follows: The original person uses such an algorithm on a computer to encode his phrase into a string (sequence) of symbols, never leaving a record of the original phrase. Then he publishes the encrypted form, along with the algorithm used to encrypt it. Once he is deceased, the only way to obtain a "hit" is to have the message communicated by paranormal means. When someone thinks he has received the message, he simply runs it through the published algorithm, and the result is compared to the published encoded form. If it matches, the message is a 100% hit. Because the original phrase is not written down anywhere, it cannot be used to cheat. The added advantages over Thouless's (1946-1949) method (which likewise avoids the cheating problem) are:

1. Given current computer capabilities and the state of the science of cryptanalysis, cryptographically strong encodings are so complex that performing them by hand is no longer feasible.

2. The computer thus naturally provides an easy way for anyone, especially people not skilled at mathematics or cryptography, to participate in the experiment, using a very strong form of encryption.

3. No special locks or encoding knowledge are needed to participate in this experiment. By using a commonly available technology (the home computer), the number of participants can be drastically increased. Because the factors affecting whether a physically dead entity can communicate are unknown, increasing the number of such experiments is important; it increases the chances of finding one who, for whatever reason, is a good post mortem communicator.

4. Finally, a person who is encrypting a message by hand (i.e., on paper, because few can accomplish a serious encoding in their minds) creates an extended spatiotemporal event. This event, which includes the paper where the data are written, the subject's mental concentration on the text, and so on, renders the experiment more susceptible to the effects of later retrotelepathy and retroclairvoyance. The computer method would seem to cut down on this possibility because the original message would exist in the memory of the computer only for milliseconds. It is plausible to assume that it is more difficult to retrieve a message that existed for milliseconds in a computer's memory than to retrieve one that was written down on paper for some part of an hour and vigorously thought about (as happens when the person goes through the steps of coding it).

5. Of course, the proposed method shares the desirable properties of Thouless's method: It can be performed many times without appreciably increasing the odds of false positives, and no record of the clear text is available for potential cheating.

The nature of the encoding algorithm is such that it is impossible (given physically plausible computational abilities) either to reverse the coding (decode the encrypted string into its text form) or to guess at the message by an exhaustive search of all possible strings. This is also an improvement on Stevenson's (1968, 1976) method because the combination lock gives only a 1-in-125,000 chance that a random guess will be correct, whereas the chance of accidentally finding the solution to most algorithms easily handled by a computer is astronomical.

The other problem with the padlock test is discussed but dismissed by Stevenson (1968): that the padlock itself can serve as a target for clairvoyance by the living, who may use a paranormal examination of its workings to discover the appropriate sequence of numbers. Stevenson dismisses this possibility because of the apparent difficulty of translating the positions of various pins and levers into their respective numbers. However, much research (for example, Jahn & Dunne, 1987, and the references therein, and much parapsychological research with nonhuman animals as subjects) has amply demonstrated that many types of ESP work equally well when the subject has no understanding of the inner workings of the mechanism he is trying to affect/perceive. It does not seem implausible to suppose that a gifted clairvoyant could get the numbers from the padlock itself and then make up a message that is based on those numbers.

My suggestion is that as many people as possible should participate in this type of experiment; because the algorithms can be performed on any home computer, it is very cheap, and potentially very profitable. The nature of the algorithm is such that it is able to support a practically unlimited number of attempts without substantially increasing the odds of a random guess's being correct. This is important because of the apparent low percentage of the deceased who are able or willing to convey testable messages to the living. It is reasonable to assume that not everyone (provided he survives) enters exactly the same state after death. The ease of use and ready availability of the method proposed in this paper, combined with its ability to support a large amount of tries without compromise of security, makes it probable that a larger number of people participating in such an experiment will lead to the participation of a deceased individual who is for some reason able and willing to communicate his message. The way to increase the odds of a success is to get as many people as possible to participate.

A Specific Example

There are several candidates for appropriate algorithms. One of the most common is RSA. It uses the fact that it is easy to find large prime numbers but that it is computationally very difficult to factor large integers into prime factors. There is also DES and the Blum-Blum-Shub cryptosystem. Good

descriptions of these algorithms can be found in Denning (1982), Gaines (1944), Merckle (1982), Pfleeger (1989), Price and Davies (1984), and Sinkov (1968).

One of the best choices is PGP. This package was created by Philip Zimmerman and can be freely obtained by anyone by using anonymous FTP on the Internet. The software is on the node src.doc.ic.ac.uk, in directory/COMPUTING/SECURITY/SOFTWARE/PGP. It will run on MS-DOS (i.e., IBM PC and clones), UNIX and VMS workstations, and many other types of common computers.

PGP is an advanced package, with many options (spelled out in the included documentation). One simple example of its use is as follows (text to be typed into the computer is boldface):

1. Create a file with a message in it. The message should be of appreciable length (e.g., several paragraphs). For this example, suppose this file is called "message.dat" and contains the following text: "I am Mr. Jones, and I have survived bodily death."

2. Invoke PGP: type pgp -c -a message.dat

3. PGP will now ask for a pass phrase. This phrase should be at least one sentence long, something that is not easy to guess and is likely to be recalled for years into the future. As the pass phrase is entered, it does not echo on the screen.

4. When this is done, the file "message.dat.asc" is created. The data in this file should be made public, and now appears as the following:

pgAAAE7ViEOfHst76gs8/NLcazh00Iw5gmJ30kHpREd03E4qz3i0vAtEFlup3 I39tjXu6q6lmp382A04+0teUV4NbGMMk4RtKZeE7r8naGTEK4o= =1ZXn5

5. When, at an appropriate time, someone thinks he has paranormally received the pass phrase, he can attempt to decode by typing the following: pgp message.dat.asc. PGP then asks for the pass phrase. If an incorrect pass phrase is entered, the program informs the user of this fact and does nothing. The correct pass phrase restores the original text, which can then count as a direct hit.

Possible Problems with This Method

There are a couple of potential pitfalls in this technique, however. The first is that the original person may leave a record of the message while claiming not to have done so, and someone may discover it. There is not much that can be done about this, except that the most impressive evidence would come if a well-known skeptic was the one to be the "sender." Also, because the sender is by definition deceased when the message is checked, there is no financial (or other) incentive for him to cheat in this way.

Another pitfall involves the apparent nature of ESP transmission. Most of the parapsychological evidence suggests that messages are transmitted by content, not by exact wording. This means that even a strong hit may not appear as such because if the message is off by even one symbol (or a word replaced by a synonym, and so forth), the encoded form will not match that of the published sample. This means that this test is too stringent, that it will have a potentially high percentage of false negatives; however, the most important property is that it will eliminate false positives. This, of course, is the most important issue when attempting to prove the existence of paranormal phenomena.

The super-ESP hypothesis is another potential problem. If pure clairvoyance exists (and this is an open question, for it is difficult to separate pure clairvoyance from telepathy in experiments), then there is no way to determine whether the message was obtained from an existing discarnate entity or by paranormal knowledge of the past. This methodology leaves super-ESP as a formal possibility but cuts down on the plausibility of such an occurrence (compared to the other two proposed methods) because of the shift to a computer instead of a person doing pencil-and-paper coding. That this method does not rule out super-ESP is bad from the standpoint of proving the survival hypothesis, but, of course, establishing the existence of such an ability would be of paramount importance to science.

The final potential pitfall of this technique is that someday, when

mathematical knowledge and computer power grow beyond a certain capability, the algorithm may be reversed, so that the message can be decoded; also, an all-out search among all possible strings may become feasible. Some of these algorithms have been mathematically proven to be irreversible. In addition, most common cryptographic algorithms are being continually subjected to cracking tests; any compromise of an algorithm quickly becomes public knowledge among the computer security and cryptography community. If a "hit" is encountered, it is easy to check with the relevant professional group and to ignore it if they have utilized a method that has been reversed. All of these issues notwithstanding, encryption algorithms provide a useful tool for survival experiments.

REFERENCES

DENNING, D. (1982). Cryptography and data security. Reading, MA: Addison-Wesley.

FLEW, A. (1987). Is there a case for bodied survival? In A. Flew (Ed.), Readings in the philosophical problems of parapsychology, (pp. 347-361). Buffalo, NY: Prometheus Books.

GAINES, H. (1944). Cryptanalysis. New York: Dover Publications.

GEACH, P. (1987). Reincarnation. In A. Flew (Ed.), Readings in the philosophical problems of parapsychology, (pp. 327-337). Buffalo, NY: Prometheus Books.

JAHN, R. G., & DUNNE, B. J. (1987). Margins of reality. New York: Harvest/HBJ. MERCKLE, R. C. (1982). Secrecy, authentication, and public key systems. Ann Arbor, MI: Research Press.

PENELHUM, T. (1987). Survival and disembodied existence. In A. Flew (Ed.), Readings in the philosophical problems of parapsychology, (pp. 338-346). Buffalo, NY: Prometheus Books.

PFLEEGER, C. (1989). Security in computing. New York: Prentice Hall.

PRICE, W. & DAVIES, D. (1984). Security for computer networks. New York: Wiley.

SINKOV, A. (1968). Elementary cryptanalysis: a mathematical approach. Math Association of America. New York: Random House.

STEVENSON, I. (1968). The combination lock test for survival. Journal of the American Society for Psychical Research, 62, 246-254.

STEVENSON, I. (1976). Further observations on the combination lock test for survival. The Journal of the American Society for Psychical Research, 70, 219-229.

THOULESS, R. H. (1946-1949). A test of survival. Proceedings of the Society for Psychical Research, 48, 253-263.

COPYRIGHT 1994 Parapsychology Press
COPYRIGHT 2004 Gale Group