Învățare Mediată de Funcție într-o Fantomă (Operant Conditioning of a Ghost)

In honor of the Halloween season, here’s a peculiar story from the annals of behavior science that I’ll bet you haven’t heard before.

I learned of it only through a chance encounter while touring the Romanian countryside following a conference in Bucharest.

Count Vasilîi Caculă, nattily dressed as always, circa 1935.

It’s the story of “Count” Alfred Vasilîi Caculă (cha-KOO-luh), who claimed to descend from the noble House of Basarab that established the Principality of Wallachia in the 14th Century. The veracity of this is unclear, as an A.V. Caculă appeared in the records of a Minneapolis General Mills breakfast cereal plant the year before the Count burst onto the scene in roaring-twenties Romania, so it’s possible that the Count was actually born in the States and adopted his Romanian identity in early adulthood. As has been breathlessly reported elsewhere, adding to the mystery are some quixotic potential links between the Count and ongoing General Mills projects, though it’s hard to know how much to make of this.

In any case, whatever Caculă’s origins, his sharp mind and preference for military-style regalia made him a swashbucking figure in Bucharest society during the 1920s and 1930s. In certain quarters, however, he was called “Doctor Blajini” and regarded with suspicion for his hobby of breeding and taming rats. The reason: In Romanian folklore, the blajini are chimeric inhabitants of a sort of mirror earth who have the head of a rat and are the incarnation of children who died before being baptized. Although in the dominant Romanian culture blajini are benevolent, in some rural areas, as the Count would learn, they are seen as morally capricious and capable of harming humans.

Count Caculă was a polymath who was self-taught in numerous languages and complex subjects, but his two great interests were quantum theory and animal behavior, which came together in the story to be related here. Caculă had consumed occult writings by Nikolai Tesla and others, and came to believe that quantum theory held the key to understanding the afterlife, particularly the noncorporeal status of the deceased (ghosts). Because of his linguistic skills he was also able to digest the research of Pavlov, Lloyd Morgan (whose observations of animal behavior at times anticipate operant theory), and Edward Thorndike, from which he derived his ideas about behavior. By 1936, three years before Skinner would publish The behavior of organisms, Caculă had a solid understanding of the distinction between respondent and operant conditioning, which he called “function-mediated” learning (învăţare mediată de funcţii). He saw operant processes, particularly involving social reinforcers, as the primary foundation of human behavior, and apparently published several papers laying out this position in the short-lived Romanian Journal of Genetic Psychology (1933-1940), of which, unfortunately, no known copies survive.

Count Caculă’s two primary interests coalesced in 1938 when a haunting was reported in the remote mountain village of Capu Corbului in northeast Romania. The specific location was Albescu (White) Manor, so named for its white stucco exterior that is quite atypical for Romanian architecture. The building has since fallen into disrepair, but I was able to visit its ruins during my recent trip to Romania (below). According to local legend, the master of Albescu Manor feared his only daughter Yulia would run off with a local scoundrel named Ernesto. To prevent this, from age 11 on he kept her locked in a box or cage with nothing but a cello to occupy her. At 19 she passed away, possibly of a Vitamin D deficiency. Soon afterward the father was gripped by an unexplained seizure and died as well, which locals interpreted as the daughter taking her revenge. No one dared occupy Albescu Manor after that, and it was reported that Yulia’s spirit would roam the building at night.

The author at Albescu Manor, now in ruins with its once distinctive white facade yellowed by age.

Count Caculă believed that quantum theory supported the existence of multiple parallel dimensions, and that death involved a phase shift from ours into a different one. “Hauntings,” he believed, were a sort of Schrödinger’s Cat scenario in which the deceased existed simultaneously in two dimensions. Incredibly, he believed that operant processes were responsible for this. Because death was a dimensional shift rather than a cessation-of-being, Caculă thought that the deceased would retain neural functions and hence be capable of operant learning. The incomplete phase shift was not a haunting but rather a kind of multidimensional behavioral momentum. While alive, The Count reasoned, the deceased’s approach behavior had been reinforced by contact with the living. After death, this difficult-to-extinguish approach behavior is what drew them back across the dimensional divide. Thereafter, Caculă thought, visitations would be intermittently reinforced by encountering the living.

When Count Caculă learned of the situation in Capu Corbului, he saw an opportunity to test his hypothesis. He immediately embarked to the village and set up in an apartment with his beloved pets. He then recruited a small army of local residents to help him conduct a pivotal experiment. Caculă marked a perimeter 10 m from the outside walls of Albescu Manor, and during the hours of 10 PM to 3 AM positioned local residents along it, where they could be easy seen by a spirit wandering within. Density of reinforcement for approach behavior was operationalized in terms of spatial density: The more residents on the perimeter, the more likely they would be seen by Yulia, and the more likely she would be to return.

The experimental design was parametric but simple: 7 two-week experimental conditions, with density of residents on the perimeter ranging from 1 every 6 m to 1 every 232 m. These residents were charged with noting the number of Yulia sightings each night and reporting this back to Caculă. The primary dependent variable was the number of sightings in each two-week period.

As a side note, it’s remarkable that Caculă could pull this off. Capu Corbului has never had more than 1,000 residents and the highest-density reinforcement condition required 256 of them to stand around Albescu Manor for a fortnight. A less charismatic person might not have been able to orchestrate this. However, in a bitter irony, interacting with so many locals on a daily basis may have ultimately been the Count’s undoing. Once the villagers became aware of his rodent flatmates they grew increasingly agitated, and with just three days remaining in the final condition of the study they refused to assume their positions around the perimeter, insisting that they would no longer take orders from a blajini. When the Count, decked out in his habitual military uniform, tried to browbeat the locals into continuing, they returned the favor in a more physical way, bringing about the Count’s untimely end. He is said to be buried in an unmarked grave just steps from where Yulia strolls the halls of Albescu Manor, but his resting place has never been verified.

We know of the “ghost conditioning” experiment only because of a half-Romanian Oxford College student named Marius Bernfela, who was en route to visit Caculă when the tragedy unfolded. Bernfela arrived in time to find Caculă’s records, including a nearly-completed draft of a manuscript, still intact in his rented chambers. He updated the report with the most recent data and dutifully submitted it to the now-lost Romanian Journal. It was published just before the journal was shuttered on the eve of World War II.

I learned of Bernfela entirely by accident while waiting in Piatra Neamț for a bus to Gheorgheni. A rumpled old gentleman sat down next to me and in broken English declared that his name was Bubâiri (there is no direct phonetic translation, but roughly: bū-BAIR-ee) and he was on the same route. We began to chat, and when I mentioned my interest in operant conditioning he became very animated and insisted that we stop over in Capu Corbuliu, where he took me to Albescu Manor and related the story of the Count and Bernfela. Given the total absence of any published record of the events he related, I was of course skeptical. But Bubâiri claimed to be a former employee in the archives of the Romanian Royal Museum. He provided the name of a contact there who, upon my return to Bucharest, gave me entry and, sure enough, buried deep in the archives I found Bernfela’s notebook.

Bernfela’s photograph of the Count’s bizarre data display. Courtesy Romanian Royal Museum.

It revealed that Bernfela had barely gotten out of Romania. As chaos built in the leadup to war, societal institutions began to break down and lawlessness prevailed outside of Romania’s major cities. Bernfela was accosted by bandits who fired on him with small-bore pistols. However, the notebook, secured in his breast pocket, stopped two bullets that otherwise would have pierced his heart. After the war broke out Bernfela was called home and conscripted to fight, which is how his notebook returned to Romania. Sadly he was killed in action.

Aside from the two near-fatal slugs, which are still embedded in its cover (below), the notebook contains a photograph of Caculă’s results (above) plotted in an idiosyncratic graphing technique that the Count devised to compress multi-dimensional relationships into two-dimensional space. As you can see, this display is all but inscrutable. But Bernfela did his best to transcribe the key raw data and, using a log transformation, replotted it in the graph reproduced below. The figure is faded and marred by a 1930s coffee stain, but it shows an apparently linear relationship between reinforcement density and haunting frequency. Bernfela, like the Count before him, took this as support for the ghost reinforcement hypothesis.

I am proud to present these data for independent inspection — likely for the first time since the Romanian Journal went defunct more than eight decades ago. Given the potentially controversial nature of the functional relation that is depicted, you may draw your own conclusions.

Left: Bernfela’s notebook, complete with the two small-bore slugs that could have killed him. Right: Bernfela’s replotting of the Count’s results. Courtesy Romanian Royal Museum.


A recently released graphic novel features a character who’s rumored to have been based partly on the Count.