by Rupert Sheldrake
In Carroll's comments he criticises me for making ad hominem remarks, as for example when I say 'Carroll is a committed sceptic who is strongly motivated to try and discredit the positively and statistically significant results of these tests'. The fact he is a committed sceptic is very relevant because it strongly affects his approach to data that goes against his beliefs. But it is entirely hypocritical of Carroll to object to any element of ad hominem argument, given his own style. In the article about morphic resonance on his Skeptics Dictionary, much of his discussion is a personal attack on me and very little is about morphic resonance. For example, he says, "He prefers Goethe and 19th century vitalism to biochemistry." He has simply made this up. Nowhere do I say I prefer Goethe and 19th century vitalism to modern science; this is a typical example of Carroll's rhetorical style in which he creates false dichotomies. In a similar vein, he goes on to say, "He prefers a romantic vision of the past to a bleak picture of the world run by technocrats who want to control nature... He prefers metaphysics to science, though he seems to think he can do the former but call it the latter".
He also criticises me for what he calls "poisoning the well". But Carroll is again being hypocritical. He himself poisons wells to an extraordinary degree. For example, instead of discussing the empirical evidence for morphic resonance, he says, "It is no more empirical than L.Ron Hubbard's 'engram', the alleged source of all mental and physical illness." Morphic resonance has nothing to do with L Ron Hubbard or with illness, and Carroll is simply trying to create guilt by association.
He then comes to criticising details of the analysis of the data from the N'kisi experiments and in particular the omission of trials in which the parrot said nothing. We omitted these trials because the parrot could neither be right nor wrong if it said nothing, and this is a standard practice in analysing experiments with animals and children. But, as Carroll knows, when one of the reviewers raised this same point and asked me for the data, his analysis of the data including all the trials in which the parrot said nothing made almost no difference. As he said, and as I quoted in my response to Carroll, "I did a permutation test on the entire dataset and found a p-value that differed only trivially from the one stated in the article. Although the authors have done an analysis that I would not have done (by omitting data), it makes no difference to the results." So Carroll's criticism is one that has already been made and answered, and makes no difference to our conclusions.
As Carroll points out, one of the target words, "flower', was repeated very frequently by N'kisi. Again, one of the reviewers already made this point, and we therefore analysed the data omitting all the cases where the parrot said "flower," to see whether the result was still significant when this word was omitted. Using a randomised permutation analysis with 20,000 random permutations, the results excluding "flower" was still strikingly significant (p= 0.006).
It is difficult to do experiments with animals, who cannot be instructed about the exact rules of the tests they are undergoing. It would have made things much easier if this parrot had behaved like a human subject and followed instructions. But that's not the way animals behave, and any research on them has to take into account their actual behaviour, not an idealised behaviour that an experimenter might wish for. In this series of tests, by using independent evaluators of the transcripts and an independent statistician to analyse the results, we ensured that the data were as reliable as possible. The results were highly significant statistically. Carroll's criticism adds nothing to the understanding of animal behaviour, and adds nothing to the peer review process which this paper had already undergone. The only purposes it serves are to muddy the waters, to indulge his passion for polemics, and to justify his prejudices.