Though Murder Has No Tongue (36 page)

Sundheim was again probated (given court-ordered psychiatric hospital admission) in February 1938, which falls between discovery of victims nos. 9 and 10. A friend of the doctor completed the official complaint, which provides a fascinating bit of data: “Patient Physician at Massillon State Hospital. In Probate Court for 3rd time as alcoholic. Yesterday became hallucinated
with ideas of fear. Talked about Federal men after him. Police called and taken to jail.” On February 28, 1938, two psychiatrists, West and Stone, examined Sundheim and documented “that at times he has indicated symptoms suggestive of delusions, which however were transient. Also that he had been hallucinated at times when drinking, usually.” Hallucinations and delusions are psychotic symptoms, which represent a loss of touch with reality. Although they can be due to a mental illness, alcohol- or drug-related problems (intoxication, as well as withdrawal) can also result in psychotic symptoms. Considering that Sundheim may, in fact, have been tailed by FBI agents, his drunken ramblings take on a whole new light. In April 1938, when he was again probated, the psychiatrist noted the absence of psychosis.

There is evidence to suggest that Sundheim may have experienced some of the long-term consequences of chronic alcohol consumption. Over time, alcoholics can develop certain vitamin deficiencies that have substantial effects on the nervous system. Peripheral neuropathies (damage to the nerves outside of the spinal cord) result in numbness, weakness, and burning pain of the hands and feet. When Sundheim was evaluated by doctors for his January 1934 inquest of lunacy, he was quoted as saying, “My wife sent me in. I'd lie around and moan all night with these hands. The legs and feet are better but they still burn like a hot pad on them.” It is possible Sundheim had even more devastating neurological problems as a result of his alcohol use. Damage from thiamine deficiency can cause confusion, dizziness, gait abnormalities, and abnormal eye movements. This grouping of symptoms is called Wernicke's encephalopathy. Untreated, Wernicke's can develop into Korsakoff's syndrome, which is a chronic condition affecting memory. Unlike Wernicke's, Korsakoff's can improve with thiamine treatment, but it does not completely resolve with thiamine treatment. An individual with Korsakoff's syndrome has impairments of recent memory and the ability to form new memories (anterograde amnesia). Niacin deficiency can result in alcoholic pellagra encephalopathy. Individuals with this condition can have confusion, clouding of consciousness, and other neurological symptoms.
⁷
Records indicate that Sundheim continued to drink until at least the mid-1950s and possibly until the end of his life, in 1964. Thirty years of heavy drinking certainly makes central nervous system (brain) damage a possibility.

Another possible diagnosis supported by the records is barbiturate dependence. It is unclear whether Sundheim was abusing barbiturates during the time frame of the torso killings. However, he clearly was later in life, as evidenced by correspondence between the director of Dayton's Veterans Administration Center and the secretary of the Ohio State Medical Board. In one letter, the director of the VA Center quotes from the chief of the
Domiciliary Medical Service: Gaylord Sundheim “has been here in the Domiciliary on and off since 1946 . . . because the veteran still has his Ohio Medical license, he freely writes prescriptions for barbiturates. He allegedly prescribes for his aged sister who will only take the medicine he sends her. Instead, the veteran takes the drugs himself and becomes an impossible management problem. The veteran's name is also carried by all the pharmaceutical houses and he receives countless samples of drugs which he is able to become intoxicated on by overdosages.” The director of the Dayton VA Center also wrote: “Although he has been under care at a mental hospital as a psychotic and is still rated as incompetent by the Veterans Administration, we have a continuing problem with him in that he writes prescriptions for barbiturates and tranquilizers in fictitious names for his own use.”

With regard to other diagnostic possibilities, some of the most interesting data comes from Sundheim's own late-in-life writings. I have a few words of caution before I delve into his written communications. By the 1950s, he had been drinking and perhaps also abusing prescription medications for many years, and likely suffering many detrimental effects from doing so. The mental status revealed by his writings from the 1950s and 1960s is not likely to be truly representative of his mental status during the time of the torso killings. It would be inaccurate to conclude that whatever neurologic and psychiatric issues color his correspondence in the 1950s were necessarily also active in the 1930s. In addition, intoxication can result in mood and psychotic symptoms (hallucinations, delusions, disordered thinking). The
DSM-IV
describes substance-induced mood disorder and substance-induced psychotic disorder. Since Sundheim was impaired by chemicals for most or all of his adult life, it is hard to know if he had any mood or psychotic disorders unrelated to drug and alcohol use. To make matters even more complicated, chemical dependence problems frequently coexist with primary mood (major depression, bipolar disorder) and psychotic (schizophrenia) disorders. It is entirely possible that Sundheim had a mood or psychotic disorder and used alcohol as a way to self medicate, although this is impossible to prove now.

We can't ever really know what is in another person's mind. However, a person's writings can give us little glimpses of his or her thought processes. On September 12, 1953, Gaylord Sundheim wrote a letter to FBI director J. Edgar Hoover. It's nearly impossible make any sense out of it even after a third or fourth reading. Was he drunk or high when he wrote it? Is it an example of psychotic thinking, where the connections between the concepts make sense only to the author? Sundheim's February 14, 1954, letter to Eliot Ness and his June 29, 1956, letter to a Cleveland Probate Court judge are similarly nonsensical. They contain examples of neologisms, made up words (for example, “Nessisms”).
There are also several examples of clang associations, words grouped together based on how they sound rather than their meanings or to create a logical statement.

He writes to the Cleveland Probate Court: “Evidently you are protected there by, [illegible] protectives from the seeming ever encroaching Federal Octopus or pusses promoting vicious ends by Haunt Taunt or daunt methods.” My favorite of his clang associations—“truancy in lunacy”—comes from this letter. I detect the slightest bit of paranoia in the writings but, frankly, they are so difficult to comprehend that I may be projecting my own desire to find hidden meaning in them. Whether due to the effects of drugs and alcohol or a primary psychotic disorder (or both), the letters show Sundheim had seriously impaired thinking. His letters from the late 1950s and early 1960s regarding his medical credentials are (mostly) goal oriented and not suggestive of psychosis or severe cognitive impairment.

Given my diagnostic impressions, let's quickly review the diagnoses from professionals who actually personally evaluated Gaylord Sundheim:

Considering advances in science and how our diagnostic criteria have changed over the years, I agree with the doctors from 1963; I think their diagnoses are the best substantiated by the data.

I can't tell you whether or not Gaylord Sundheim was the torso killer. I can, however, examine his known history to cautiously determine whether he was at risk for violence to others. Risk assessment is a routine part of a psychiatrist's work. The goal of a risk assessment is to determine if the individual is low, moderate, or high risk of violence to self (suicide) or others. For a risk assessment to have any value, it has to be thorough. The more extensive the data collection, the more useful the final product will be. Sources of information for a violence risk assessment include a clinical interview with the individual in question; interviews with any of the individuals who witnessed threats or actual violence and with the person's family members, roommates, partner or spouse, and close friends; reviews of any threatening
communications, including recorded phone messages, letters, e-mails, and Internet postings, and of the individual's medical and psychiatric records. The clinical interview is often very lengthy and can cover everything from the individual's taste in video games to his pornography preferences. In the case of Gaylord Sundheim, my sources of information are very few. Therefore, I can only make the most tentative of conclusions regarding risk.

Once a person is identified as being at low, moderate, or high risk for violence, the risk can be further classified. Is he or she at an elevated risk for high-magnitude violence, such as shooting or stabbing? Is the risk elevated for lower magnitude acts, such as hitting or spitting? Is there a specific target for the violence? (It is possible for the person to be at high risk for violence against a specific individual [i.e., a boss or a family member] but low risk as far as the general public is concerned.) Is the risk acute or chronic? Is the violence imminent or contingent on a particular event? To use a well-known example, a man who abuses his wife may be at a chronic elevated risk of violence, but that risk increases acutely at the point in time the wife decides to leave the marriage.

No risk assessment is complete without a risk-reduction plan. If the identified risk is low, the plan may be to simply monitor the individual. If the risk is moderate or high, more extensive steps need to be taken. The first task in constructing a plan is to differentiate between modifiable and unmodifiable factors. A history of past violence is the biggest risk factor for future violence, and obviously unmodifiable. Male gender, lower IQ, and age (late teens to early twenties) are other examples of violence risk factors that cannot be changed through treatment. Modifiable risk factors include drug or alcohol use, certain psychiatric symptoms, access to weapons, unemployment, and residential instability. It is important to note that not all psychiatric symptoms elevate violence risk.

Using the time frame of the torso murders (1934–38), let's now examine Gaylord Sundheim's unmodifiable risk factors. The court documents from his divorce petition allege he had a history of being physically abusive to his children and demonstrating “extreme cruelty” toward his wife. His male gender elevates his risk of violence. His divorce in 1936, ending his marital relationship and separating him from his children, is an important consideration in the equation. Finally, we know that Sundheim was a veteran, which means he received weapon and combat training.

Sundheim also had some modifiable violence risk factors. The most prominent of these was his alcohol use. In his December 1933 Cuyahoga County Probate Court inquest of lunacy, court documents indicate heavy drinking. His 1936 divorce papers discuss “habitual drunkenness” and almost continual
intoxication. In February 1938, he was involuntarily admitted to hospital through Probate Court. In a statement to the court, a friend of Sundheim's said that he “became hallucinated with ideas of fear” and “talked about Federal men after him.” Doctors examined Sundheim and said he “has indicated symptoms suggestive of delusions, which however were transient. Also that he had been hallucinated at times when drinking, usually.” Did Sundheim's drinking make him paranoid, or was he actually referring to Ness and his men? Paranoia, whether delusionally based or due to a real world circumstance, can increase violence risk. When an individual feels endangered, he may use violent means to protect himself. Of course, if Sundheim was paranoid about federal agents, his violence risk would increase with respect to law enforcement, though not necessarily globally. Delusional paranoia is now treatable with antipsychotic medication. In 1938, however, there were no antipsychotic medications. Thorazine did not come along until 1950, and prior to that insulin shock and frontal lobotomies were the most advanced treatment.

Keeping in mind the very limited historical data and my inability to interview Sundheim or review any collateral sources of information, I conclude that Sundheim's risk for physical violence of a low magnitude was mild to moderate during periods of intoxication. This conclusion is based primarily on court documents from the time of his divorce, which alleged that he was habitually intoxicated and had been physically abusive to his family. I have no confirmable evidence to suggest he was violent when sober or that he engaged in high-magnitude violence (choking, stabbing, shooting) at any time in his life. There is nothing to suggest he was assaultive toward non-family members. Even later in Sundheim's life, correspondence from the Dayton VA does not mention violent episodes. With regards to his concern that federal agents were following him (delusional or factual), none of the documentation states that he felt in danger or was contemplating taking action against his pursuers. I chose my words carefully here because I want to make sure that my conclusions are not unsupported assertions. Now, if it could be definitively proved that Gaylord Sundheim committed murder by decapitation, it would alter my risk assessment a great deal. The more information available about the context and motivation for those decapitations, the more complete my risk evaluation would be.

I am at the end of my analysis and no nearer to proving Francis E. Sweeney, M.D., is Gaylord Sundheim is the Mad Butcher of Kingsbury Run. However, I hope I've provided some enlightenment into the factors affecting Sundheim's mental state and risk for violence during the years of the torso killings (1934–38). Most important, I hope to have given you a realistic appreciation for what forensic psychiatry is and is not. Although a part of me still longs
to be Clarice Starling on the hunt for serial killers, I find rewards in the less life-threatening challenges of conducting thorough, carefully reasoned, and well-supported forensic evaluations.