The Kuhnexus CaseBase: Would it Help to Usher in a New Paradigm?

 Introduction

The website of the Society for Psychical Research (SPR) has a page, dated 20 November 2009, entitled ‘Kuhnexus: database of the best cases of paradigmchallenging phenomena’ (retrieved 14 March 2013).  This is listed as a ‘member project’, and states that it is intended to be a:

 “Collection and classification of the best-in-class paradigm-challenging cases, to form a basis and reference for research and theory building.  Includes development of a suitable classification system and case quality criteria.  The project is led by David Rousseau, Chairman of the [SPR’s] Research Activities Committee and Director of the [independent] Centre for Fundamental & Anomalies Research (C-FAR).”

The idea is to take the huge amount of research that has been done on anomalies, or “special puzzle pieces” as they are termed, and make the whole range accessible to as wide an audience as possible.  More, the special puzzle pieces can then be used to form a framework which will enable interconnections to be seen between areas that are currently considered as separate entities.  This framework will enable a deeper understanding of the phenomena to be achieved.  This task will be facilitated by modern technology, which can greatly assist with storage and keywords, assisting in the search for correlations:

“For example, one could easily locate the best apport cases, whether associated with séances, poltergeists, jots [just-one-of-those-things], metal-bending children or UFOs. Databases can be accessed over the internet, which means they can be widely used. And because they are easy to update, they remain current and relevant.”

The aim, then, is to create a database with:

“the most important evidence for each kind of anomaly that seems to occur. We will concentrate on those anomalies that seem to pose a significant challenge to orthodox explanations. Since these are the anomalies that have the potential to trigger a paradigm shift, we dub them Kuhnia, in honour of Thomas Kuhn, and the database will be called The Kuhnexus. Significant challenges to orthodoxy include more that just psi phenomena, and we will include such Kuhnia as ball lighting, acupuncture, brain plasticity, Penrose crystals, etc.”

How do you define an important case?  Well,

“There are many things that can make a case important. The obvious one is its evidentiality, which is useful for people asking how confident we are that a specific type of anomaly occurs. Beyond that, though, there are cases that are important for their implications, or for what they tell us about the variability or characteristics of a phenomenon. There is also significance in features that are common to different types of cases. By creating a database of cases, which can be sorted or filtered in different ways, different readers will be able to focus on cases that are most relevant to their specific interests.”

Challenging the Paradigm Systematically

The project has been a long time in gestation.  David Rousseau outlined the rationale behind it in an article in the SPR’s Journal in 2002, ‘Challenging the Paradigm Systematically: A New and Generic Approach to Classifying Anomalous Phenomena’.  I have had to simplify a fairly complex paper, and it is worth reading in its entirety.  Essentially, the idea is twofold: use a systems approach to develop a classification which would improve the organisation and accessibility of data on anomalous phenomena; and make the information more useful in developing a framework to challenge “orthodox theories.”  It is a structured method for organising data, a programme looking at anomalies research in the context of knowledge management.

Because 'anomalous phenomena’ are here defined broadly to encompass “reliably established phenomena that significantly challenge our world-view,” they extend beyond what are generally conceived to include the paranormal and psi phenomena.  Rousseau gives as examples that he would include but which fall outside psychical research/parapsychology: “cryptozoology, UFOs, cold fusion, dark matter, the placebo effect, homeopathy, ancient technology, ball lightning, etc.”  He contrasts these with puzzles that can be expected to be solved within the orthodox framework, such as “schizophrenia, genetic defects and tumours.”  The system would build an “anomalies catalogue” which would define and classify these anomalies without presuming explanatory models.

Within psychical research, Rousseau posits a link between spontaneous and laboratory findings.  Traditionally the two have been viewed as generally separate domains, but Rousseau sees spontaneous phenomena as comprising, in aggregate, those phenomena that are studied in the laboratory.  The distinction, he argues, is obscured by the way in which the term ‘anomalous phenomena’ is used to cover both levels, the molecular and aggregated.  Spontaneous phenomena can be regarded as “synergistic systems” that can be broken down into their components; if you can understand these elements, you should be able to understand the range of larger-scale phenomena that they constitute.  This would not only establish links between field and laboratory phenomena, but possibly also links between seemingly disparate field phenomena (Rousseau gives the example of dowsing and mental mediumship).  The implication is that we could determine the shape of a phenomenon by mixing the sub-elements together in the correct proportions, rather like the ingredients in a recipe. Rousseau’s hope is that his approach will reveal relationships between anomalous data, and suggest further lines of research.

He notes that unfortunately a problem is that at present we know as little about how laboratory phenomena operate as we do field ones, necessitating a phenomenological approach, describing effects “in order to avoid embedding presumptions about mechanisms into our terminology.”  There is the need for a classification system that does not favour any particular conceptual framework:  “An important feature of this system is that in both the definition and the classification ... we refer only to observed or inferred effects, and not to (presumed) causes or mechanisms.  This makes the classification system robust against the theoretical development of the paradigm models...”  But an obvious objection to this method is that if you are only considering effects, how can you make a judgement on aetiology?  There may be surface similarities, but that does not mean that causes are related.

Also, if you are trying to identify components at the molecular level, one cannot know whether they are the same across phenomena at the macro level, or only superficially similar.  There is a parallel with convergent evolution – simply because characteristics appear alike, it does not necessarily mean that they are linked, such as erroneously assuming a close relationship between bat and bird because they can both fly. Rousseau proposes a classification system paralleling that used for diseases, i.e. described in terms of symptoms.  He talks of “benchmarks” for normal behaviour.   However, if I have a temperature, it is easily established by reference to my typical body heat output, but benchmarks for allegedly paranormal occurrences are not so easily established.

Rousseau’s object is to “include all reliably reported phenomena that would, if true, embarrass our leading theories of how the world works.”  But he does not tackle the complex issue of how reliability can be determined, a significant weakness when assessing the strength of cases.  Examples of apparently strong cases that turn out to be weak when re-analysed are many, including the Chaffin Will Case, examined by Mary Roach in Six Feet Over, S. G. Soal, whose work was held up to be of a high standard until SPR member Betty Markwick examined the data (‘The Soal-Goldney Experiments with Basil Shackleton; New Evidence of Data Manipulation’, in SPR Proceedings, 1978), even the Creery Sisters, who fooled the early SPR investigators, about which Edmund Gurney had to issue a statement in the SPR’s Proceedings shortly before his death in 1888 (‘Note Relating to Some of the Published Experiments in Thought-Transference’).

What is more usual however, is a messy situation in which researchers produce findings that are challenged by critics; free, increasingly acerbic, debate ensues; and finally peters out without a clear resolution as the debate moves on.  Examples would be the 2011 presentiment research by Daryl J. Bem (‘Feeling the Future: Experimental Evidence for Anomalous Retroactive Influences on Cognition and Affect’), and, from the other side, the debate surrounding Richard Wiseman’s 1992 speculations in the SPR’s Journal about how Eusapia Palladino might have cheated during the 1908 Naples sittings (‘The Feilding Report: A Reconsideration’).  Rousseau argues that “there will be differences of opinion between researchers as to what constitutes ‘reliably established’ and ‘significantly challenges’ [sic] but these are not by themselves impediments to this definition’s useful application in research.”  That is optimistic.  For this to work, there would need to be an arbitration mechanism to adjudicate between possibly widely diverging estimates by researchers who might or might not have an axe to grind.

Even this would not be definitive, though.  There can be no final arbiter of reliability, of what presents a challenge for our paradigm, and therefore should be included.  Defenders of a subject and their critics will always argue over validity, and where you draw the line.  Many would argue that homeopathy, one of Rousseau’s examples, does not actually challenge our world view at all because it can be adequately explained by the placebo effect.  There are harder examples:  How would complex cases like Enfield or Scole be evaluated?  What sort of consensus could be reached on their veridicality?  In order to not pre-judge, such a system would need to be set with very liberal criteria, risking the introduction of a lot of noise.

Both experimental and anecdotal research can be disputed, as the examples above indicate, but it is particularly easy for the critic to dismiss the latter as mere anecdotes.  The plural of these may or may not be data, but either way, they possess lower evidential value compared to controlled experiments, and assessing their value on a scale is far from an exact science.  Having read a large selection of ‘ghost’ literature, I can attest that the stories exhibit wide variation in degrees of plausibility.  Witnesses are not pure sources of information, they have motivations, secrets, agendas, varying levels of fantasy proneness, leaving aside the basic difficulties of eyewitness testimony.  Investigators ask leading questions, perhaps jazz things up.  These factors can all be assessed for reliability in the eyes of the evaluator, but we have to be aware of the dangers of unknown levels of distortion.

Even if the reliability issue is somehow overcome, there are problems in building a hierarchy of elements which can be aggregated.  The assumption is that these elements are additive and can be combined to form a larger anomaly.  As an example of this additive view of phenomena, Rousseau argues that a poltergeist might be seen in terms of a range of more “fundamental” anomalies, such as “PK [psychokinesis], apports, thermodynamic anomalies, acoustic anomalies...” etc.  However, this does not take into account emergent properties; the whole might be greater, and rather different, than the sum of its parts.

The reference to acoustic anomalies is an interesting one because much has been made in recent years of an analysis that Barrie G. Colvin made of a range of recordings of raps collected at alleged poltergeist sites, reported in the SPR Journal (‘The Acoustic Properties of Unexplained Rapping Sounds’, 2010), but the effects have been replicated by C. J. Romer using clearly non-paranormal means, namely banging around his house, as detailed in his Polterwotsit blog.  Romer adds that this is not necessarily to say that Colvin’s raps were not paranormal in origin, only that they can be replicated by non-paranormal means.

Colvin’s recordings are open to interpretation as the raps themselves were ephemeral.  Rousseau distinguishes between transient and non-transient effects.  The former leave no lasting trace, such as seeing a ghost.  A non-transient effect would be one you can examine at leisure.  Rousseau’s examples of this are “thoughtography and metal objects deformed by PK.”  These are definitely non-transient, in the sense that one can examine an artefact, but there is still an obvious problem for Rousseau’s taxonomy, working out what is a paranormal object and what is not, if you cannot link the surface effects (symptoms) to unambiguous causes.  It may be non-transient, but you don’t know if it has a paranormal origin.  Perhaps it is unfair to expect any such a scheme to be able to make this determination, but then Rousseau seems to be making claims on which his approach cannot deliver.

The April 2002 SPR Study Day

To coincide with the Journal article, Rousseau elaborated on these ideas at an SPR study day entitled ‘Making the Evidence Count’, chaired by Julie Rousseau.  A report on the event was written by Nicola Holt (Paranormal Review, No. 26, April 2003).  David Rousseau spoke of a scoring system which would, in Holt’s words, “enumerate each case study, capturing information about the significance of each case, an indication of its value and quality.”  Rousseau presented ten criteria, each weighted from 1 to 10, giving a total score out of 100.  As far as I am aware Rousseau has not issued his own list so it is worth quoting the ten in full:

1          Is this case a clear challenge to orthodoxy – does this case reflect an anomaly?

2          Witness or researcher credibility – is their report reliable (e.g., is there a vested interest in a particular outcome)?

3          Depth of evidence, e.g. rich and detailed reporting, immediacy and accuracy of reporting, good research and ‘documentation’.

4          Corroboration – multiple witnesses on multiple occasions.

5          Corroboration – repetition of physical anomalies or physical records/evidence.

6          Consistency – shares features with other cases of the same category.

7          Test of time – the case has stood debunking.

8          Test of time – the case has attracted no criticism due to its strengths.

9          Correlation – the case correlates with an independent objective variable, e.g. ganzfeld success and local sidereal time.

10        Predictability/verifiability – repeatability criteria.

Other speakers at the study day were asked to choose strong cases to discuss, in order to see how the criteria would work.  I was present in the audience, and remember wondering how they hoped to get agreement between researchers on scoring these criteria, which in any case are not orthogonal.  Clearly, according to the Paranormal Review report, others also raised concerns about the practicalities, and problems finding consensus when there were “widely different viewpoints in the field”, as Holt worded it (an understatement; put two psychical researchers together and you will more than likely get three opinions).  Holt’s conclusion was that “While evidence of any kind may never lead to irrefutable ‘proof’ of anything, the anomalies catalogue would at least enable researchers to form opinions based on carefully classified and easily available information, thereby assisting in the construction of theories and models ... based on the ‘best available evidence’.”

Giving researchers the tools to construct theories is praiseworthy, but as an indication of the difficulties presented by the proposed scoring system, take the example Mary Rose Barrington gave at the study day:  This was a case (P. 384) from 1882 reported in the SPR’s Proceedings, Volume 5, 1888, in a paper by Eleanor Sidgwick, ‘On the Evidence for Premonitions’.  As Holt recounts it,

“The case she [Barrington] finds most compelling is a spontaneous case from the Sidgwick collection of 1882, with a high 'randomness factor', called the 'Schweitzer case' or 'the Henry Irving case'. On the 18th July 1882, Mrs Schweitzer dreamt that her son, Fred, died, falling over a cliff. In the dream he was in the company of another man, whose name she enquired, he replied 'Henry Irving'. In waking life, Mrs Schweitzer anxiously sent a message about the dream to her son, who was on a business trip in the Midlands, asking him to come home as soon as possible. Fred replied that he was fine. However, rather than going home immediately, he went to the coastal town of Scarborough. He was keen to go riding in the Forge valley and was accompanied by a new acquaintance he met there. Unfortunately, Fred fell from his horse, hit his head on a rock and died instantly. Later, Mrs Schweitzer met the man who had been with her son; she asked if his name was Henry, indeed it was, Henry Deverell. She told him about her dream, and he said that when he did amateur acting he was known as Henry Irving, Junior. Mary Rose argued that this is a very strong case (awarding it 87 points). It was well documented, with the original notes available and good first-hand evidence from all the parties, in addition the 'Henry Irving precognition' is exceptionally 'random' or 'improbable'.”

That is a score of 87%, which is impressive.  But another scorer may find it less so, and award a lower mark based on precisely the same evidence.  Mrs Schweizer (not Schweitzer) had her dream, which did not correspond in detail to the eventual accident, on 18 July 1882; the accident took place on 26 July and was reported in the York Herald, 28 July; but Mrs Schweizer only wrote to the SPR on 28 October 1882, and the SPR did not follow the case up until April 1888.  Interestingly, Holt not only misspells Mrs Schweizer’s name, but in the original letter, the acquaintance tells Mrs Schweizer that he is introduced at private theatricals as Henry Irvin, not Irving.  Fred did not die instantly, but three hours after his accident.  It is so easy to get details wrong, and these were easily checkable; how much more problematic is eyewitness testimony, which cannot be checked.

Yes, the case is strengthened by Mrs Schweizer meeting “Henry Irvin, jun.”, which apparently accorded with her dream.  However, she says that she only recalled her dream on meeting Mr Deverell, and she could easily have heard the name Henry as the name of her son’s companion when she was told the details of Fred’s accident, and forgotten it in her grief.  Even more, surely the case is weakened by the length of time, well over three months, between having the dream and recording it, with an intensely traumatic event in between, during which many details could have been invented, omitted or rearranged in her memory.  Mrs Schweizer stated that Mr Deverell and her son had signed an account “substantially the same” as that in her letter, but we are not told when this happened, what the differences with the letter were, or whether anybody at the SPR actually saw it.  Sadly Mr Deverell could not be asked for his recollections in 1888 as he had drowned in 1883, though he could not have vouched for the accuracy of the dream as he was only told about it after Fred’s death.  Mrs Schweizer’s sons were unlikely to contradict their mother.

Given these issues, eighty-seven seems a rather high score, and might be seen to reflect the way in which the scorer is predisposed to view the evidence.  Another scorer might consider this not to be a strong case (I tried scoring it myself and gave it 51%, and I thought I was being generous).  The idea of scoring anomalies sounds fine in theory, but in practice tells us as much about the scorer as it does about the case being scored because scoring cannot be calibrated.  Nor are we told what counts as a significance level.  Presumably a case would have to reach some threshold in order to count as a strong case, but there is no indication what it would be, or how exactly how set.  Further, once you have done your scoring, how exactly does that convince a sceptic?

A Database of White Crows

Despite the mixed feedback at the 2002 study day, in 2005 Rousseau “launched the Kuhnexus CaseBase Project, a collaborative project to establish a database of the best cases of paradigm-challenging phenomena.” (The reference is on the website of another organisation founded by Rousseau, the Centre for Systems Philosophy, of which he is Director, retrieved 14 March 2013.)  At the same time, he published an article in Paranormal Review (no. 34, April 2005), ‘Collecting Cases for a Database of White Crows’.  The 2009 SPR website page quoted above is an abbreviated version of the start of this article, but here Rousseau goes into more detail about what a case is, what constitutes an anomaly, what kind of cases the project team are looking for, and how the reader might help.  Cases are defined as “a historical narrative that rests on records that authenticate it to some degree...”, and he requests readers to submit those cases they consider to be most significant.

The criteria have grown to fifteen, but they are not listed, and there is no evidence that fifteen would provide any more objectivity than ten would, nor that they would allow a white crow to emerge.  The article’s title is drawn from William James: essentially, you only need one white crow to disprove the statement that “all crows are black.”  James’s white crow was mental medium Mrs Piper, but while she has had many champions, most recently Michael Tymn, who has written Resurrecting Leonora Piper: How Science Discovered the Afterlife (2013), it is unlikely that even an entire database full of examples as good as Mrs Piper would have much impact in terms of paradigm change on mainstream science.  It is perhaps instructive that Rousseau, when referring to best cases, puts “best” in inverted commas, as if himself unsure of their status.  He does exhort the SPR membership to get involved supplying cases so that the team can select the best ones, but curiously the project website shown at the end of his article is that of C-Far, indicating a blurring of ownership between the SPR and Rousseau’s own organisations.

Kuhnexus CaseBase and Nigel Buckmaster

The patient reader may be wondering why this project, ambitious in scope but unlikely to succeed, is still of interest.  The reason is because it is now potentially of much greater scope than suggested in Rousseau’s articles.  There have been a number of references in previous SPR Annual Reports to Kuhnexus CaseBase, in David Rousseau’s Research Activities Committee (RAC) reports (Rousseau chairs the RAC).  In the latest Annual Report (covering October 2011 to September 2012, released in March 2013), he has much more on the subject, taking up almost half his RAC space: 

“Kuhnexus CaseBase: We [i.e. the RAC] continue to support this project, which is aimed at collating the best cases in every class of anomalous phenomena relevant to psychical research. The project team ... continue to develop the project framework and collect relevant cases, and anticipate that the collection will one day be made available as an on-line resource...

“As reported last year, a generous bequest towards this project has been made by the late Mr. Nigel Buckmaster ... The funded project will be a significant undertaking, but given the preparatory work already done is it is likely to gain momentum quickly ... It is anticipated that the research resources established under this project will bring new momentum to psychical research and establish psychical research’s significance to a wide range of important open issues in the orthodox worldview.”

David Rousseau’s statement on the Financial Position in the 2011-12 Accounts (he is also the SPR’s Hon. Treasurer) mentions a total of £600,000 in the Buckmaster bequest (p.13), so it can be expected that a generous portion of the legacy might be devoted to the CaseBase project; indeed, no other purpose for the disposal of the money is mentioned.  If a substantial sum is going to be spent on such an idea, there needs to be a clear understanding beforehand about how the project might work in practice, and whether Mr Buckmaster’s money will be effectively spent (it is worth pointing out that the same Annual Report states that during 2011-12 the SPR awarded a mere five grants for research, totalling £12,670).  So Rousseau’s various papers are of more than academic interest, they have practical implications.

While there are issues with some of his ideas, others could easily be implemented.  A database of Kuhnia would be an improvement, in terms of accessibility and searchability, on an enterprise like Bill Corliss’s Sourcebook Project, though it would be a long time before it was anything like as comprehensive.  The online library run by Lexscien (an entity created by C-Far) has a range of publications, the SPR’s Proceedings and Journal, its magazines Psi Researcher and Paranormal Review, as well as publications issued by other organisations.  This could easily be expanded to make hitherto obscure material accessible (I have myself suggested adding the run of the SPR’s Myers Memorial Lectures, issued as pamphlets, to the database), plus perhaps a selection of unpublished cases in the archives and any unpublished supporting documentation for those that have been published.  Rousseau’s 2005 Paranormal Review article states that most cases assessed will already have been published, so the novelty would be ease of access, plus the scoring system and classification to find commonalities, rather than the publication of new cases.

The Lexscien search facilities could be improved to assist searches across a broad range of data, and perhaps cases tagged in order to facilitate the hunt for the type of cross-phenomenon feature discussed in Rousseau’s 2002 paper.  This could be a project to involve the SPR membership in a Wikipedia-style exercise, and as Wikipedia allows discussion of its content by contributors, so could the Lexscien pages.  The ten or fifteen criteria could be used as a way of filtering out the clearly weak cases which have no supporting evidence, and new cases could be invited on the same basis, increasing the SPR’s archive in the process.  Such an approach would cost only a miniscule fraction of £600,000, leaving the bulk available to support actual research.

Anything more elaborate, such as the scoring exercise, or finding similarities that may or may not be spurious, runs the risk of using resources on something no more reliable than any previous case collection, and with as much chance of changing the current paradigm.  A paradigm shift happens, Kuhn tells us, when the current one produces anomalies that can no longer be ignored.  Even if a superb quality-controlled collection is gathered together, is this really going to change anyone’s view, or will those who are not convinced carry on ignoring the damned data?  This may seem unduly pessimistic, but we have good evidence gathered over the last 130 years that it will take a lot more than a collection of Kuhnia, however extensive and efficiently collated, to trigger a paradigm shift.