Hi again, everybody!
Jane Curtain & I investigated the Reverse Speech theory in 1997. We found ourselves unable to hear more than a few very short reversals in Oates’ material; even these were short accidental approximate reversals of unconnected FS sequences (as in it’s an honour). In other cases, any resemblance between FS and alleged RS was minimal; for instance, Neil Armstrong’s … small step for man (said on the Moon in 1969) does not reverse to Man will space-walk, as is claimed; the consonants, especially, do not match. However, we found that we were being distracted by Oates’ continual prompting on his video and audio tapes, which induce listeners to hear the alleged RS sequences. Because of this effect and the requirement to subject the RS theory to empirically sounder testing, we turned to experimentation.
Oates’ description of his own experimental methodology is repeatedly obscure and ambiguous, and he is generally reluctant to answer questions seeking clarification (or else repeatedly fails to understand exactly what information is required). The replication of his experiments is thus somewhat uncertain. Nevertheless, we replicated the experiments as best we could; additional variants were introduced where this was thought potentially useful. Forty subjects, divided into four groups of ten, participated in the preliminary experiment. Six short recordings of alleged RS sequences were taken from Oates’ audio tapes and were reproduced in written form. Each group of subjects experienced a different set of procedures:
Group A: The Group A response sheets listed the six written sequences, without any identification of the speakers.
Group B: The Group B response sheets listed six written sequences, which were entirely different from the RS sequences alleged by Oates, but which displayed a) the same number of syllables and b) similar or the same vowel phonemes, within each sequence.
Group C: Group C subjects were not provided with a written list of the alleged RS sequences. They were, however, told that an intelligible sentence was present in each of the six recordings.
Group D: Group D subjects were not provided with a written list of the alleged RS sequences, and they were told NOT told that there WAS an intelligible sentence in each of the six recordings but that there MIGHT be such a sentence.
Groups A and B subjects were asked to tick those sentences which they could clearly hear in the extracts played to them, or to circle any syllables, words or sequences of syllables or words, shorter than the sentences, which they could hear. Groups C and D subjects were asked to record (in normal orthography) any clearly intelligible sequences in English which they could hear, whether these formed the whole of a given extract or only a part of it.
The Group A subjects provided a significantly greater number of ‘correct’ responses than did the other three groups, and the Group B subjects provided a greater number of ‘correct’ sentences than did Groups C and D. In respect of words and syllables, the Group D subjects provided a greater number of ‘correct’ responses than the Group C subjects. It is clear from these results that suggestion (prompting) is a major factor in the hearing of alleged RS sequences. Where the vowel phonemes were the same as those proposed by Oates, 32% of the relevant syllables were ‘correctly’ identified; while, for the same participants, where the vowel phonemes were different from those proposed by Oates, only 18% of the syllables were ‘correctly’ identified. A possible explanation for the fact that the Group D subjects had more success than the Group C subjects involves the idea that they concentrated very hard to hear such sequences, whereas the Group C participants initially believed that the sequences would be obvious and gave up attempting to hear such sequences when they proved difficult to hear.
Further, potentially decisive tests of Oates’ claims suggest themselves, notably tests aimed at determining whether information which (as it seems) could not otherwise be known can be obtained from listening to RS sequences. However, Oates has not been willing to co-operate with linguists and psychologists in arranging such tests.
In summary, it appears (as is agreed by most of the linguists and psychologists who have examined Oates’ theories) that RS is an artefact of the listening process, often encouraged by advance prompting in the material (written and oral) provided by Oates.
There are various other skeptical discussions of RS, from various standpoints. Oates and some RS supporters have attempted rebuttal of some of these criticisms, but very few of these responses manifest anything resembling the level of specifically LINGUISTIC expertise required for dealing with the data, and their authors often misinterpret the criticisms and/or fail to understand what further information is required.
As ever, detailed references on request. Next time, I’ll say a little about Backward Masking and related phenomena.