Blue / Grey hero banner image

Arthur Balfour's Ghosts: An Edwardian Elite and the Riddle of the Cross-Correspondence Automatic Writings, by Trevor Hamilton

Publication Details: 
Imprint Academic ISBN-13: 978-1845409135
Publish date: 
December, 2017

Review by Professor Donald J. West:

Despite the flippant title, this book is a sympathetic, exhaustively researched and well documented study of the famous ‘cross-correspondences’ found in the automatic writings of distinguished SPR members and one professional medium, the American Leonora Piper, during the first third of the twentieth century. With 32 pages of references, it is the most detailed history so far available, revealing the previously unpublished identities, complex inter-relationships and sometime dual roles of the script writers and their investigators, previously concealed by the use of pseudonyms in SPR Proceedings. This book is less preoccupied with evidence for survival than were the discussions in Proceedings or the review by Saltmarsh (1938). The title reflects the fact that so many of the script writers and purported communicators were linked to Arthur Balfour, his siblings and his intimates.

The death in 1901 of F. W. H. Myers, author of Human Personality and its Survival of Bodily Death, stimulated some of his friends to seek contact with his spirit through various techniques of automatic writing. The diverse attitudes, beliefs and practices of these enthusiasts ranged from palm reading, fortune telling, and crystal gazing to traditional spiritualistic séances. Their efforts at ‘automatic’ writing yielded a plethora of short scripts, produced irregularly on impulse, over the next 35 years. The core group of these automatist/mediums was an assembly of distinguished and talented individuals, including Winifred Coombe-Tennant, sister-in-law of Myers, Margaret Verrall, lecturer at Newnham College, Cambridge, Dame Edith Lyttleton, a prominent social reformer and Alice ‘Trix’ Fleming, sister of Rudyard Kipling. Some of the group were already friends.

 The first investigators of the scripts were not disinterested assessors, but a close-knit assembly of intellectuals, including Arthur Balfour’s sister Eleanor, all of whom had been friends of the living Myers. Most of them belonged to the elite group of socially prominent and talented academics who led the SPR in its early years and were generally favourable to the idea of the paranormal. Their integrity and sincerity is unquestioned. Their ‘automatic’ scripts were mostly short statements of a few sentences written in various languages, many being quotations from or references to Greco-Roman classics. Some erudition was required to appreciate the allusions. The investigators decided they could perceive at least two coherent themes in the scripts The ‘Story’ was an attempt by the deceased May Littleton to convince Arthur Balfour of her continuing love for him. The ‘Plan’ was an attempt by the communicating spirits to promote world peace by influencing the birth of specially gifted children.

Cross-correspondences occurred between scripts produced independently by writers on the same or nearby days, working at different locations, but some correspondences appeared also in scripts at later dates, so the field for searching was wide. Correspondences were sometimes easy to spot. For example (p.207) Trix Fleming on 28 February 1906: ‘No not in the Electra. M. will know better’. Margaret Verrall, 28 February1906: ‘Be sorrow spoken, but let good prevail’. This counts as a cross correspondence because Fleming seems to have sensed some quotation from Greek tragedy (actually from Aeschylus’s Agamemnon). Many of the passages identified as cross correspondences were more complex and sometimes difficult to recognise. Correspondences were potentially paranormal only if the ‘automatic’ scripts were produced independently. This was not guaranteed. Some of the contributors of scripts, such as Winifred Coombe-Tennant, were also classed as investigators and given access to all existing scripts and could have been influenced by their knowledge of prior contributions.

The counting of cross correspondences in the total collection of scripts was chiefly the task of J.G. Piddington, who spent years pondering and revising his criteria. What can or cannot be dismissed as coincidence is a matter of subjective and therefore disputable judgement. The book includes enough examples to show the great difficulty of the judgements required, but the investigators insisted on avoiding a box-ticking statistical approach. The credibility of the scripts was affected by the failure of their apparent aim, in particular the failure of the implausible expectation of a messianic child who would promote a golden age. Henry, son of the automatist/medium Winifred Coombe-Tennant, from a liaison with Arthur Balfour’s brother Gerald, was supposed to fill this role. Actually, Henry Coombe-Tennant’s career as a soldier and diplomat proved promising, but he extinguished the hope that he would be the great reformer by becoming a monk! The claim that the accurate references in the scripts to features of the communicators’ earthly lives must be paranormally acquired information is open to doubt now that the opportunities for some of the automatists to mingle with the communicators’ surviving friends is evident. The non-disclosure of relevant information about participants’ identities casts a doubt on the efficacy of the investigators’ intended impartiality.

The majority of Hamilton’s book deals with differing interpretations of the cross correspondence evidence. He discusses at length the pros and cons, but readers hoping for firm conclusions will be disappointed. He poses some basic questions. Did the scripts demonstrate paranormal cognition? Was the assertion of purpose and meaning in the scripts just wishful thinking? He is adept at identifying pros and cons for different answers. To cite one example from Chapter 19, he notes how familiarity with the classics was a badge of upper class status. Inevitably, the authors of ‘scripts’ would be motivated to employ such language, even when it was not their natural habit. The resulting communality of speech would create pseudo correspondences to be mistakenly interpreted as contributing to meaningful messages.  The scripts were a mixed collection produced spontaneously by different people over an extended period. Considering the common interest of writers and investigators in finding correspondences, it is not surprising that many such were identified.

In this reviewer’s opinion, the Hamilton study points to the cross correspondences project having failed to generate any advance in knowledge about psi. Had the investigators been more proactive in guiding their ‘communicators’ to more precise tasks, the results might have been different. One may draw some lessons from this experience. The investigators had an unusually free interchange with seemingly forthcoming communicators, but they were not proactive in trying to steer conversation to verifiable formation on matters as yet unknown to anyone involved with the project. The heritage of the all the endeavours devoted to these cross correspondences is a collection of historic material open to endless and unproductive dispute.


Saltmarsh, H . (1938)  Evidence of personal survival from cross-correspondences. London, Bell. 

A separate review by Guy Lyon Playfair will appear in the Journal of the Society for Psychical Research.