Dissertation: The Psychology of Serial Offenders: A Case Study Analysis

THE PSYCHOLOGY OF SERIAL OFFENDERS: A CASE STUDY ANALYSIS

Karl Matthew Jenkins

A dissertation submitted in partial fulfilment of the requirements for the degree of Bachelor of Criminology with Honours

Department of Criminology
University of Adelaide

October 1995

Supervisor: Dr Margaret Conroy

---

Abstract

This dissertation examines the psychological factors that contribute to the development and persistence of serial offending. It proceeds by comparative case study, taking three documented offenders — drawn from England, the United States and Australia, and differing in era, method and apparent motive — and reading them against one another for what they hold in common. The central argument is that serial offending is neither simply inherited nor simply produced by circumstance, but emerges where three strands converge: a developmental substrate of early deprivation and failed attachment; a psychological architecture marked by an absence of ordinary fellow-feeling and a compensating need for control; and an iterative behavioural process by which a method is slowly assembled through fantasy, rehearsal and escalation. A secondary argument follows from the first. Across all three cases the offender is most legible not in his motive, which is frequently obscure even to himself, but in his method, which he repeats. The corresponding failure is institutional: where cases are not connected to one another, a pattern that is in principle readable goes unread. The dissertation does not claim to have discovered a cause. It claims only that the same few features recur with a regularity that is difficult to attribute to chance, and that this recurrence has consequences for how such offenders might be anticipated, identified and, in a small number of instances, prevented.

---

Acknowledgements

I am grateful to my supervisor, Dr Margaret Conroy, for her patience with a project that was slow to find its shape, and for insisting throughout that I say less than I could prove rather than more. I thank the staff of the Department of Criminology, and the staff of the Barr Smith Library, who located a good deal of material that was not easy to find. I thank my parents, who taught me before I understood it that careful work is its own justification. The errors that remain are mine.

---

Contents

1. Introduction
2. Literature Review: Theories of the Serial Offender
3. Methodology
4. Case Study One: Gerald Pryor
5. Case Study Two: Howard Beck
6. Case Study Three: Desmond Frawley
7. Cross-Case Analysis
8. Conclusion

---

Chapter 1: Introduction

1.1 Background and significance

Few categories of crime attract as much public attention, or as little settled understanding, as the conduct of the offender who kills, assaults or abducts repeatedly, over months or years, choosing strangers and returning to the same act again and again. The attention is not difficult to explain. Most violence has a reason that can be located — a quarrel, a debt, a sudden loss of control — and most violence stops once that reason is spent. The serial offender fits neither expectation. He appears to act without a reason of the ordinary kind, and he does not stop of his own accord. It is this — the absence of an ordinary motive, and of an ordinary end — that makes him frightening in a way that a single violent act, however terrible, is not.

Understanding is harder to come by. The literature on serial offending is large but uneven. Much of it is journalistic, concerned with the horror of particular cases and content to leave the offender finally inexplicable, a monster about whom nothing general can be said. A smaller part is clinical or academic, and here the difficulty is of another kind: the cases are rare, the offenders are unwilling or unreliable witnesses to their own conduct, and there is a standing temptation to build large theories on a few well-known examples. Careful comparison — the patient setting of one case beside another — is rarer than the seriousness of the subject would lead one to expect.

This dissertation attempts such comparison, on a deliberately modest scale. It does not propose a new theory of serial offending. It takes three offenders whose conduct is reasonably well documented, sets out each case on its own terms, and then reads the three against one another to establish what, if anything, they share. The method rests on a simple assumption: that a feature which looks, within a single case, like mere biography — the unrepeatable product of one life — may prove to be something more once the same feature is found in cases that have nothing else in common.

1.2 The problem of definition

Any study of serial offending must begin by saying what it means by the term, and here there is no agreement to be had. The most widely cited definition, associated with the Federal Bureau of Investigation and with the work of Ressler and his colleagues (Ressler, Burgess & Douglas 1988), restricts the category to three or more separate events at three or more separate locations, with an interval — a “cooling-off period” — between them. The interval is taken to be the feature that distinguishes the serial offender from the mass or spree offender, who acts in a single sustained episode (Holmes & DeBurger 1988).

The definition is serviceable but plainly arbitrary at the edges. The threshold of three is a convention rather than a finding; an offender caught after his second killing is not a different kind of man from the one caught after his fourth. The “cooling-off period” describes something real but explains nothing, and its language assumes a heat between offences that may not always be there. More importantly, the standard definitions were developed in respect of serial murder, and serial murder is only the most extreme instance of a wider phenomenon — one that includes serial rape, serial arson and serial abduction, in which the same patterned, repetitive, stranger-directed quality appears without a death necessarily occurring (Douglas et al. 1992).

This dissertation adopts a working position rather than a settled definition. By serial offending I mean the repeated commission, by one offender, of the same grave offence against a succession of victims chosen for what they represent rather than for who they are, separated in time, and carried on in the absence of any conventional precipitating motive such as gain or revenge. For reasons of evidence set out in Chapter 3, the three cases examined here are all cases of serial murder. Serial murder is the sub-type for which documentation is fullest, and the one in which the features of interest appear in their clearest form. Where the analysis bears on serial offending more broadly, I say so.

1.3 Aims and research questions

The dissertation has one principal aim: to determine, by close comparison of three documented cases, whether serial offenders share a small set of recurring psychological and developmental features, and, if they do, to describe those features as precisely as the evidence allows.

From this aim, three research questions follow.

First, what does each offender’s history, taken on its own terms, appear to show about the origins of his conduct? This is the biographical question, answered separately for each case in Chapters 4 to 6.

Second, when the three histories are read together, which features recur, and which belong only to the individual case? This is the comparative question, and the central one. A feature found in one offender is biography; a feature found in all three, in the absence of any other connection between them, is a candidate for explanation.

Third, what follows, for the prevention and investigation of such offences, from whatever recurring features are found? This is the applied question, and the one I treat most cautiously. The history of the subject is full of confident profiles that did not hold up against the next case, and I would rather not add to them.

1.4 Scope and delimitation

Three limits on the scope of the study should be stated at the outset.

The first concerns geography. The serious literature on serial offending is overwhelmingly American, and to a lesser extent British. Australian academic treatment of the subject is, at the time of writing, very thin, and the small number of documented Australian cases have drawn more journalistic than scholarly attention. This is itself a finding of a kind, and it carries a warning for the method: frameworks developed in respect of American offenders, in American conditions, cannot simply be lifted and applied elsewhere as though the social setting made no difference. One of the three cases examined here is Australian, included partly so that the imported frameworks can be tested against a local instance rather than merely assumed to fit it.

The second concerns the type of offence. As noted, all three cases are cases of serial murder. This is a limit of evidence, not of interest. I do not assume that the serial murderer and the serial rapist or arsonist are psychologically identical, and where the case material allows a distinction to be drawn, I draw it.

The third concerns the reach of the conclusions. Three cases are three cases. A feature common to all three is established as common to all three, and as worth attention; it is not thereby established as universal, and nothing below should be read as a claim that every serial offender exhibits it. The comparative method can show that something recurs. On this number of cases, it cannot show how often.

1.5 The comparative case-study approach

The decision to proceed by comparative case study, rather than by the statistical analysis of a larger sample, calls for a word of justification, since the larger sample would ordinarily be preferred.

The justification is that the larger sample is not available in any usable form. The population of documented serial offenders is small; the records that survive are uneven in quality, and were compiled for purposes — prosecution, journalism, clinical treatment — other than the one a researcher would wish to put them to; and the variables of real interest, such as early maltreatment or the role of fantasy, are exactly those recorded inconsistently or not at all. A statistical study built on such material would look more rigorous than it actually was.

The case-study approach accepts this constraint instead of disguising it. It does not claim a representativeness it cannot establish. It works by depth: by setting out each case fully enough that the reader can follow the reasoning and disagree with it, and by drawing comparisons only where the material on both sides is sound. Its characteristic danger — that the investigator will find in the cases the pattern he brought to them — can be guarded against in only one way: by stating the expected pattern in advance, in Chapter 2, and by reporting the features of each case that do not fit it as carefully as those that do.

1.6 Structure of the dissertation

The remainder of the dissertation is organised as follows. Chapter 2 reviews the principal existing theories of serial offending — psychiatric, developmental, typological and sociological — and sets out the three-strand framework against which the cases are subsequently read. Chapter 3 describes the method: how the three cases were selected, what sources were used, how they were analysed, and what the limitations of that procedure are. Chapters 4, 5 and 6 present the three cases in turn — Gerald Pryor, an English offender of the late 1960s and early 1970s; Howard Beck, an American offender of the following decade; and Desmond Frawley, an Australian offender of the 1980s whose case remains, in important respects, unresolved. Chapter 7 reads the three cases against one another and forms the analytic core of the dissertation. Chapter 8 states the conclusions, considers their bearing on prevention and investigation, and is candid about what the study has not shown.

---

Chapter 2: Literature Review — Theories of the Serial Offender

2.1 The shape of the field

Explanations of serial offending fall into a small number of families, each developed by a different discipline and each answering a different question. The psychiatric literature asks what kind of person the offender is; the developmental literature asks how he came to be that person; the typological literature asks how such offenders may be sorted into kinds; and the sociological literature asks why they appear when and where they do. The families are not rivals so much as partial answers to different questions, and the mistake most often made in the field is to take an answer to one question as though it answered the others. This chapter reviews each family in turn, takes from each what appears to survive criticism, and sets out at the end the framework — assembled rather than adopted — against which the three cases of Chapters 4 to 6 are read.

2.2 The psychiatric account: psychopathy

The oldest and most durable psychiatric description is that of the psychopath, set out by Cleckley (1941; 1976) in terms that have not been much improved upon. Cleckley’s subject is not the raving madman of the popular imagination but a man who presents, on first and often on long acquaintance, as entirely sane — articulate, plausible, superficially charming — while lacking the ordinary inner life that the presentation implies. Behind what Cleckley called the mask is an absence: no durable attachment, no remorse, no anxiety of the kind that governs ordinary conduct, and a poverty of feeling concealed by a fluent imitation of it. The importance of the description, for the present study, is that it locates the defect not in what the psychopath does but in what he does not feel.

What Cleckley emphasised, and what later writers have sometimes lost, is that the mask is not a disguise consciously assumed but a competence: the psychopath performs normal feeling so habitually, and so well, that the performance is most of what there is. The cluster he described — superficial charm and intelligence, an absence of nervousness, unreliability, untruthfulness without compunction, a want of remorse or shame, poor judgement and a failure to learn from it, a profound egocentricity, and a general poverty of the deeper emotions — has proved remarkably stable across half a century of study. Its diagnostic value lies in the combination; any one feature is unremarkable, but the assembly is not.

Cleckley’s clinical portrait was given a measurable form by Hare (1991), whose Psychopathy Checklist–Revised scores the construct across two broad groups of items: an interpersonal and affective group — glibness, grandiosity, shallow affect, lack of remorse, failure to accept responsibility — and a group describing an unstable and antisocial way of life. Hare (1993) is careful to distinguish psychopathy, so measured, from the antisocial personality disorder of the diagnostic manuals (American Psychiatric Association 1987; 1994), which is defined largely by a history of rule-breaking behaviour. The distinction matters. Antisocial personality disorder, resting on conduct, is common in prison populations; psychopathy, resting on the affective deficit beneath the conduct, is much rarer, and it is the affective deficit, not the rule-breaking, that the serial-offender literature keeps returning to.

Two cautions must be entered against any use of psychopathy as an explanation of serial offending. The first is that the construct is descriptive, not aetiological: the checklist measures the presence of a deficit but says nothing of how the deficit arose, and a measurement is not a cause. The second, and more important, is that the construct is not specific to the offenders studied here. The great majority of psychopaths never kill anyone; psychopathy is found in boardrooms and barracks as readily as in prisons. It cannot therefore be what distinguishes the serial murderer, for it is shared by many who are merely callous. What psychopathy supplies is not a cause but a component — a description of the affective architecture that, in combination with other things, may issue in serial offending. That is the use made of it here, and no more.

2.3 Constitutional and biological accounts

A newer line of work seeks the roots of the affective deficit in the body rather than the history. Studies gathered by Raine (1993) report, in some violent and psychopathic subjects, a lower resting autonomic arousal — a reduced fear response — together with indications of impaired functioning in the frontal regions associated with the regulation of impulse. The work is suggestive, and may in time supply a physical correlate for the coldness that Cleckley could only describe. Two reservations keep it at the margin of the present study. The findings are, at the time of writing, correlational and group-based, and do not establish that any individual offender’s conduct issued from such a cause; and, more decisively for a study of three historical cases, no measurements of this kind exist for the men examined here. The biological account is therefore noted as a possibility and set aside — not because it is implausible, but because the evidence this dissertation works from cannot speak to it.

2.4 Developmental accounts: how it is made

If psychopathy describes the architecture, the developmental literature attempts to describe its construction.

The best-known developmental claim is the so-called triad proposed by MacDonald (1963): a childhood combination of bed-wetting beyond the usual age, the setting of fires, and cruelty to animals, said to foreshadow later violence. The triad has not held up well as a predictor; the evidence for it as a package is weak, and most children who display one of its elements grow into unremarkable adults. Its interest is residual rather than predictive. Each of its three elements is, on its own, a recognised marker of a disturbed and frightened childhood — of a home in which a child is humiliated, neglected or made afraid — and it is the disturbed childhood, not the triad as such, that the literature has reason to take seriously.

That a violent or neglectful childhood raises the risk of later violence is among the better-established findings in the field. Widom (1989), in the most careful study of the question available, followed a large sample of children with documented histories of abuse or neglect and found them significantly more likely than a matched comparison group to be arrested for violent offences in adulthood. The finding must be read with its own qualification, which Widom is the first to state: the majority of abused and neglected children do not become violent adults. Maltreatment raises the probability of later violence; it does not determine it. As an account of serial offending in particular it is therefore necessary but plainly insufficient — a background condition found in many such histories, but found also in many thousands of histories that end in nothing of the kind.

A more specific developmental claim concerns attachment. On the account derived from Bowlby (1969–1980), the capacity for fellow-feeling is not innate but built, in infancy and early childhood, through a reliable bond with a caregiver; where that bond is absent, broken, or coldly inconsistent, the child may fail to develop the ordinary felt connection to others on which conscience depends. The attractiveness of this account, for the present subject, is that it offers a route by which the affective deficit Cleckley described might actually come about: not as something the offender is born with, but as something that fails to form. It is, however, difficult to establish in the individual case, since the relevant events are early, private, and recorded only in their later effects.

It is worth stating plainly what these developmental accounts cannot yet do. They identify conditions that raise the risk of a later affective deficit, but they do not explain why, of two children subjected to comparably brutal beginnings, one develops the deficit and the other a hard-won resilience. The literature gestures at protective factors — a single reliable adult, a temperament less given to withdrawal — without being able to specify them in advance. For the present study this means that the developmental substrate must be treated as a predisposing condition and not a determining one: something found in the histories of these offenders, and plausibly part of what made them, but not a mechanism that could have marked them out beforehand.

A fourth account, sociological in its materials but developmental in its form, is the theory of violentisation advanced by Athens (1992). On the basis of close interviews with violent offenders, Athens described a four-stage process — brutalisation, in which the subject is subjected to and made to witness violence; belligerency, in which he resolves to answer violence with violence; violent performances, in which he does so successfully; and virulency, in which success hardens into a settled violent self-conception. Athens’s account has the merit of describing a process rather than a predisposition, and of grounding it in evidence the subjects themselves supplied. Its limitation, for this study, is that it explains the making of violent men in general and not the particular, patterned, stranger-directed repetition that distinguishes the serial offender from the merely dangerous one.

2.5 Typological accounts: sorting the offenders

A third family of work is concerned not to explain serial offending but to classify it.

The most influential classification is the organised–disorganised distinction developed by the Federal Bureau of Investigation and set out by Ressler, Burgess and Douglas (1988) and in the Crime Classification Manual (Douglas et al. 1992). It is derived from the crime scene. The organised offender plans, brings his own weapon, controls his victim, leaves little behind, and is typically socially competent and employed; the disorganised offender acts on impulse, uses what is to hand, leaves a chaotic scene, and is more often socially isolated and, in some cases, psychotic. As an investigative heuristic — a way of inferring something about an unknown offender from the scene he leaves — the distinction has obvious uses. As a theory of persons it is weaker than its popularity suggests. Many real offenders are “mixed,” displaying organised and disorganised features in the same offence or across a series; the categories describe scenes rather than men, and the inference from scene to person is looser than the confident language of profiling tends to admit. There is, moreover, a circularity to be guarded against: where an offender’s “type” is inferred from the scene and the scene is then explained by the type, nothing has been added but a label. The Crime Classification Manual itself concedes the existence of mixed cases, which is a considerable concession, since a binary that admits a third category for everything it cannot place is not doing the work a classification is meant to do. The distinction is used in this dissertation, but used descriptively, to characterise conduct, and not as an explanation of it.

A second well-known scheme, that of Holmes and DeBurger (1988), classifies serial murderers by motive: the visionary, driven by hallucination or delusion; the mission-oriented, who believes himself to be ridding the world of a despised group; the hedonistic, who kills for gratification, whether sexual, thrill-seeking, or material; and the power-or-control type, for whom dominion over the victim is itself the end. The scheme is useful as a vocabulary, but it inherits the central difficulty of all motive-based classification, which is that motive is the least reliable thing about a serial offender. It is inferred from conduct, or taken from the offender’s own account, and the offender’s own account is, as already noted, the least trustworthy source available. The categories also overlap: the power the control-type seeks and the gratification the hedonist seeks are frequently the same thing under two descriptions. This difficulty is not incidental to the present argument; it is part of it. A classification built on motive is built on the least stable ground in the case, and the dissertation will argue that there is steadier ground to build on.

2.6 Sociological accounts: the social production and its institutions

The sociological literature shifts the question from the individual offender to the society that produces him. Leyton (1986), in the fullest treatment of this kind, argues that the modern multiple murderer is not a timeless aberration but a historically specific figure, and that his killing is intelligible as a kind of violent protest — most often that of a man near the bottom of a social order he cannot rise within, taking his revenge upon those just above or just below him. The strength of the account is that it addresses what the individual theories cannot, namely the distribution of serial murder — why it appears more in some societies and periods than in others. Its weakness is the converse of its strength: in explaining the type it tends to dissolve the individual, treating each offender as a symptom of his society and underplaying the plain fact that the great majority of men in the same position do not kill at all.

A narrower but, for this study, more useful sociological contribution concerns not the making of the offender but the response of the institutions that ought to catch him. Egger (1984) gave the name linkage blindness to the failure of separate police jurisdictions to recognise that a series of crimes is the work of a single offender — a failure of information, co-ordination and assumption that allows a patterned series to be processed as a set of unrelated incidents. To this Egger (1990) added the observation that the victims of serial offenders are disproportionately drawn from groups whose disappearance attracts little notice — the transient, the marginal, those he termed the “less-dead,” whose deaths are under-investigated precisely because the victims were undervalued in life. The two observations belong together. A series escapes detection not only because its author is careful, but because the system that should connect it is fragmented, and because the victims it chooses are those the system is least moved to pursue. This institutional dimension is taken up again in the Australian case of Chapter 6, in which it is central, and in the conclusion.

2.7 Fantasy and the assembling of a method

One further strand of the literature concerns the internal process by which the offence is generated and refined. The clinical observation is old — Brittain (1970), describing the sadistic murderer, placed a rich and compulsive fantasy life at the centre of the picture — and it was given empirical support by Prentky and colleagues (1989), who found detailed, rehearsed fantasy to be markedly more characteristic of serial than of single sexual murderers. On this account the act does not spring fully formed. It is preceded by fantasy, often for years; it is rehearsed in imagination before it is attempted in fact; the first attempts are frequently clumsy, and the method is refined across the series as the offender learns what satisfies the fantasy and what reduces his risk. Two consequences follow that matter to the argument of this dissertation. The first is that the method is not given at the outset but assembled over time, which is why a series has a developmental shape of its own. The second is that the fantasy, being stable, leaves a stable trace: across changing victims and circumstances the offender returns to the same essential act, and it is this returning — the signature beneath the variable method — that makes a series legible as the work of one man.

2.8 Synthesis: a three-strand framework

The foregoing review yields no single theory adequate to serial offending, and this dissertation does not supply one. What it offers instead is a framework that assembles, from the surviving parts of each literature, the conditions whose convergence appears to be required.

Three strands may be distinguished. The first is a developmental substrate: the early deprivation, maltreatment, and failed or broken attachment described, from their different angles, by Widom, by the attachment literature, and by Athens — the conditions under which the ordinary capacity for fellow-feeling fails to form. The second is a psychological architecture: the affective deficit described by Cleckley and measured by Hare, combined with the need for control or mastery that the typological literature, for all its faults, repeatedly identifies. The third is an iterative behavioural process: the fantasy, rehearsal and escalation through which a method is assembled and a series acquires its shape.

The claim of the framework is that none of the three strands is sufficient alone, and that it is their convergence that distinguishes the serial offender. A damaging childhood, without the affective deficit, produces a wounded adult, not a predatory one; the affective deficit, without the iterative process, produces a cold and unpleasant man who never acts; the process, without the substrate and the architecture beneath it, has nothing to drive it. Each strand, taken singly, is common; the convergence of all three is rare, which is consistent with the rarity of the offence.

One further point, drawn chiefly from the typological and behavioural literatures, governs the analysis that follows. The least stable term in any account of a serial offender is his motive — inferred, retrospective, and frequently obscure to the offender himself. The most stable term is his method, and beneath the method the signature: the constant act to which he returns. It follows that the offender is most reliably read not through his reasons but through his repetitions, and that the corresponding institutional failure — the failure Egger calls linkage blindness — is precisely a failure to read repetitions that are, in principle, there to be read.

The framework is offered not as a finding but as an instrument. It states, in advance of the evidence, what the three cases would be expected to show if the convergence it describes is real. Stated in advance, it can be tested rather than merely illustrated; and the chapters that follow are written so that the features of each case which do not answer to the framework are reported as plainly as those which do.

---

Chapter 3: Methodology

3.1 Design

This dissertation is a comparative study of three cases. The case study has a settled place in the social sciences as a method suited to phenomena that are rare, complex, and embedded in a context which cannot be stripped away without distorting them (Yin 1989); serial offending is all three. The comparison of several cases, rather than the close reading of one, is what allows the study to move beyond biography. Its underlying logic is old, and is essentially that of Mill’s method of agreement (Mill 1843): if a number of instances of the phenomenon under study have only one circumstance in common, that circumstance is a candidate for its cause. The logic must be applied here with more modesty than Mill allowed it. Three cases are too few, and the circumstances they share too many, for agreement to prove anything; what agreement can do is to mark a feature as worth explaining, by showing that it persists where almost everything else has changed. The whole design of the study follows from that single idea — to hold the phenomenon constant and vary everything around it, so that whatever remains constant stands out.

3.2 Selection of cases

The three cases were chosen against four criteria.

The first was documentation. Each case had to rest on a substantial public record — a trial and its evidence, an official inquiry, or serious secondary study — rather than on journalism alone, so that the facts relied upon could be traced to something more durable than rumour.

The second was developmental record. Because the framework of Chapter 2 concerns the making of the offender, cases were required in which something is actually known of the offender’s childhood and early life. This is a more demanding criterion than it appears, and it excludes a number of well-known offenders whose early histories are effectively blank.

The third, and the most important to the design, was variation. The three cases were chosen to differ from one another as widely as the first two criteria allowed: in country, in period, in the apparent type of the offender, and in his apparent motive. The reasoning is given above. If three offenders separated by an ocean, a generation, and every obvious circumstance nonetheless share a set of features, those features cannot easily be put down to a common setting, and the case for taking them seriously is correspondingly stronger.

The fourth was practical, and is stated rather than disguised. The study was carried out in Adelaide, from the holdings of a single university library and what could be obtained for it on loan. Cases were necessarily limited to those for which adequate material was available in English. I do not pretend that the three are the most important cases of their kind, only that they are well enough documented, different enough from one another, and serious enough in themselves to bear the weight the argument places on them.

On these criteria the cases selected were: Gerald Pryor, an English offender active in the late 1960s and early 1970s, organised in method and outwardly respectable in life; Howard Beck, an American offender of the later 1970s and early 1980s, disorganised, isolated, and at the disordered edge of the category; and Desmond Frawley, an Australian offender of the 1980s, convicted on three counts and suspected of more, whose case was never fully resolved. The three differ on every axis the design sought to vary. What they are found to share is the subject of Chapter 7.

3.3 Sources and their treatment

The material used falls into three kinds, and was not treated as of equal weight. Adjudicated and official records — the findings of courts and inquiries — were treated as the most reliable, being the product of procedures designed to test evidence. Contemporary reporting of record was used for chronology and circumstance, with allowance made for its purposes. Secondary and scholarly studies were used for analysis, and for facts not otherwise available, weighted according to the sources on which they themselves rested.

One distinction was applied throughout. A good deal of what is “known” about any serial offender comes ultimately from the offender himself — in confession, interview, or boast — and his account is, for the reasons given in Chapter 2, the least trustworthy source in the case. Where the analysis draws on such material, and particularly on an offender’s account of his own fantasies and motives, it is marked as his account, and treated as evidence of what he said rather than as established fact. This is not a small reservation. It bears directly on the strand of the argument concerned with fantasy, where the offender is frequently the only witness.

3.4 Procedure

The cases were prepared to a common template, so that the same questions were put to each. For every offender the study sets out, in the same order: his history and development; the psychological picture, so far as it can be reconstructed; the pattern and method of his offending; and the response of the institutions that investigated him. Imposing one structure on three very different cases costs something in the texture of each, but it is what makes comparison possible: like is set beside like, and the differences that remain are differences of substance and not of arrangement.

With the three cases so prepared, they were read across one another against the framework of Chapter 2, noting for each of its three strands whether it is present, absent, or indeterminate in the case. In keeping with the discipline stated at the outset, equal attention was given to what the framework failed to predict — features common to the cases that it did not anticipate, and features it led one to expect that a case did not supply. A study of this kind is worth little if it records only its successes, and the cross-case chapter is written accordingly.

It should be said plainly that the analysis is the work of one person, with no second reader to check the reading. The judgements that follow are therefore mine, and are open to the disagreement that the full setting-out of each case is intended to allow.

3.5 Limitations and ethical considerations

The limitations of the study are considerable, and are best stated by its author. It is retrospective and second-hand: I never met these men, and work entirely from records made by others, for purposes of their own, often long ago. The developmental material in particular is thin, filtered, and sometimes shaped after the fact by knowledge of how the story ended. The number of cases is small, and no claim is made that they are representative, or that the features they share occur with any particular frequency.

One limitation is graver than the rest, and the argument depends on its being kept in view. The study examines only men who offended. It has no comparison group — no account of the many who suffered comparable childhoods, or who carry the same affective deficit, and who harmed no one. It follows that the study cannot show that the features shared by its three offenders are what set them apart from such men. It can show that the features recur together in offenders; it cannot, on this evidence, show that they are the difference between the offender and the rest. The conclusions are framed with this firmly in mind.

A final word on the handling of the material. The cases concern real deaths and real grief, and the literature on serial offending has a known tendency to forget it — to treat the victims as the furniture of the offender’s story, and to dwell on the details of their suffering for an interest that sits closer to the appetite it studies than its authors would care to admit. This study tries to avoid both faults. Suffering is described only so far as the analysis requires, and no further; the victims are kept in view as persons and not as illustrations; and the offender, while examined as closely as the evidence permits, is not granted the fascination he would no doubt prefer. Understanding such men is necessary. Admiring the thoroughness of their methods, even by an accident of tone, is not.

---

Chapter 4: Case Study One — Gerald Pryor

Gerald Pryor is the first of the three cases, and the clearest example among them of what the literature calls the organised offender. He killed at least five women, and probably more, across the English Midlands and the North between 1967 and 1972; he was an outwardly unremarkable married man with a respectable occupation; and he was caught not through any recognition of the pattern in his crimes, but almost by accident, after which the pattern was assembled in retrospect. He is included here because each of these features — the ordinariness of the surface, the control in the method, and the long failure to connect the crimes — appears in its plainest form in his case, and because the documentation, resting on a full trial and a careful later study (Mercer 1979), is good enough to support the analysis.

4.1 The offender and his history

Pryor was born in 1931 in Walsall, in the West Midlands, the only child of a household the later accounts agree in describing as cold and frightening. His father, a foundry hand, was a heavy drinker given to sudden violence, and seems to have treated the boy with a contempt that did not soften as he grew; his mother, of whom much less is known, was ill for long periods and died when Pryor was eleven. With the outbreak of war he was evacuated, twice, to households in which he was an unwanted addition, and he returned to Walsall in 1944 to a father who had in the meantime taken a second wife and had little use for the child of the first. The record of these years is thin, and much of what survives comes from Pryor’s own later statements, which must be treated with caution; but the outline — early and sustained humiliation, the loss of one parent and the indifference of the other, repeated separation from any stable home — is corroborated well enough by school and welfare records to be relied upon.

He left school at fourteen, undistinguished, and after a period of casual labour did his National Service, of which nothing remarkable is recorded. In 1955 he took the position that would define the outward shape of his life: a travelling salesman for a Midlands wholesale firm, supplying ironmongery and household goods to shops across five counties. The work suited him. It gave him a car, a respectable reason to be anywhere, and long unsupervised hours on the road, and he was good at it — punctual, quietly persuasive, well regarded by the firm for twenty years. In 1958 he married. His wife later described a marriage that was orderly and entirely without warmth; he was controlling in small domestic matters, particular about routine, and absent without explanation for nights at a time, which his work made unremarkable. They had no children. To his neighbours in the Walsall street where he lived until his arrest, he was a dull, polite, unmemorable man, and it was this, above all, that they found difficult to reconcile with what he proved to be.

4.2 The psychological picture

The psychological interest of the case lies precisely in that difficulty. Pryor was not, by any account, mad. He was never found to be psychotic, suffered no delusions, and was assessed before trial as fully responsible for his actions. What the assessments describe instead is the affective emptiness that Cleckley (1976) placed at the centre of the psychopath: a man who could perform the ordinary social emotions — courtesy, sympathy, mild good humour — well enough to pass for normal across decades, while feeling, so far as could be determined, almost none of them. The performance was not a disguise put on for particular occasions; it was the whole of his presented self, and it was seamless enough that his wife of fifteen years claimed to have suspected nothing.

Beneath the performance, the one consistent feature was a need for control. It showed in the pettiness of his domestic regime, in the meticulousness of his work, and, as will be seen, in the method of his killings, which were nothing if not controlled. It is tempting to organise the whole case around this need — to call him, in the language of Holmes and DeBurger (1988), a power-or-control offender — but the temptation should be resisted so far as the label explains nothing it does not first assume. What can be said with more confidence is narrower: that Pryor exhibited the affective deficit and the controlling temperament which the framework of Chapter 2 treats as the psychological architecture of the serial offender, and that he exhibited them in an unusually pure form, undisturbed by intoxication, psychosis, or evident rage.

Of the inner life behind the killings, less can be responsibly said. In a series of interviews after his conviction, Pryor gave an account of long-standing fantasies of domination preceding the offences by many years, and described the killings as their enactment. The account is consistent with the clinical literature on the role of fantasy (Brittain 1970; Prentky et al. 1989), and may well be true; but it is his account, offered when he had every reason to present his crimes as the product of a compulsion he could not help, and it cannot be taken as established. It is recorded here as what he said, and weighed accordingly.

4.3 The pattern and method

Pryor’s victims were young women, most of them between eighteen and their early twenties, whom he encountered in the course of his travels: women waiting for buses on the edges of towns, or walking between villages, to whom a respectable man in a clean car could offer a lift without alarming them. This was the whole of his approach, and its ordinariness was its strength. He did not break into houses or seize women from busy places; he was simply given, by his manner and his circumstances, the few minutes of trust he required.

The killings themselves were consistent to the point of routine. Having driven the victim to an isolated place, he killed her by ligature, quickly, and with little of the frenzy associated with the disorganised offender; the post-mortem records are notable for how little additional injury they describe. He then disposed of the body in open country — on moorland in the Peak, in woodland in the Dales, once in a drainage dyke in the Lincolnshire Wolds — chosen, it appears, for remoteness and for distance from the place the woman had been taken. From each victim he removed a single small possession, a watch or a piece of jewellery, which he kept; a collection of these items, found hidden in his house after his arrest, was among the evidence that convicted him.

Two features of this pattern matter to the larger argument. The first is its constancy. Across five years and a considerable distance, the essential act did not change: the same approach, the same method of killing, the same taking of a token. This is the stable signature described in Chapter 2 — the trace that, in principle, makes a series legible as the work of one man. The second is the deliberate scattering of the crimes across distance and jurisdiction, which had the opposite effect, and to which the investigation’s long failure must largely be attributed.

There is some evidence, though it is not conclusive, that the method was refined over the series. The first identified victim, Carol Beaumont, a nineteen-year-old shop assistant killed in 1967, was found relatively close to where she had last been seen, and the disposal in that case was comparatively careless; the later killings show longer journeys, more remote sites, and greater attention to the removal of traces. If the sequence is read correctly, it is consistent with the literature’s account of a method assembled and improved through repetition rather than possessed from the start.

4.4 The investigation and its failures

The most instructive feature of the Pryor case, for a study concerned with the institutions as much as with the offender, is that for five years no one was looking for a single man at all. The killings fell within the areas of four separate county forces, each of which investigated its own as a local crime. There was, in this period, no national means of comparing cases across forces — the Police National Computer did not begin operation until 1974 — and the comparison that might have revealed the pattern depended on individual officers happening to hear of similar crimes elsewhere, which for the most part they did not. The similarities were striking in retrospect; at the time they were distributed across four sets of files that no one read together. This is linkage blindness (Egger 1984) in its historical form: not a failure of intelligence on the part of any investigator, but a structural failure of a system not built to see across its own divisions.

The failure was compounded by the kind of victim Pryor chose. Several of the women were, in the language of their time, thought to have “gone off” of their own accord — to have left home, or run away — and their disappearances were for a time not treated as suspicious at all. Egger’s (1990) observation about the “less-dead,” the victims whose marginality slows the response to their deaths, applies here in a muted English form: not the destitute, but the young, the working-class, the assumed-flighty, whose absence was easy to explain away.

When the end came, it owed nothing to the recognition of the pattern. In 1972 a constable near Bakewell, noting the vehicles in a lay-by from which a footpath led onto the moor, recorded Pryor’s car among them; when the body of his final victim was found on that moor some weeks later, the note was still in the file, and a routine follow-up of the listed vehicles brought Pryor in for questioning. The collection of tokens in his house did the rest. Only after his arrest, as the four forces at last compared their cases against the man in custody, was the series understood to be a series. The pattern had been there for five years. It was assembled only once there was a name to assemble it around.

Pryor was tried in 1973, convicted of five murders, and sentenced to life imprisonment. He never pleaded guilty and never expressed remorse; the post-conviction interviews on which much of the psychological record rests were given, by his own suggestion, in exchange for small privileges, and have the quality of a performance. He died in prison in 1989.

4.5 The case against the framework

Read against the framework of Chapter 2, the case supports it in part and qualifies it in part. The psychological architecture — the affective deficit and the need for control — is present about as clearly as the available evidence could show it, and is the strongest element of the case. The iterative process is present in the constancy of the signature and, more tentatively, in the apparent refinement of the method across the series. The developmental substrate is present in outline — the violence, the loss, the repeated severing of any stable attachment — but the outline is thin, partly reconstructed after the fact, and dependent in places on Pryor’s own untrustworthy account; it is better described as consistent with the framework than as strong evidence for it. The fantasy element, similarly, rests almost entirely on what Pryor chose to say, and is therefore recorded but not relied upon. The case offers firm evidence for the architecture, fair evidence for the process, and only suggestive evidence for the substrate — an unevenness that will be seen to recur, and which the cross-case analysis of Chapter 7 takes up directly.

---

Chapter 5: Case Study Two — Howard Beck

Howard Beck is included as the deliberate opposite of the first case. Where Pryor was controlled, respectable and mobile, Beck was chaotic, isolated, and rooted to the East Side of Youngstown, Ohio — a district hollowed out in the years after the steel mills closed — where he killed at least three people, and probably many more, between 1978 and 1982. He is, in the language of the typological literature, the disorganised offender to Pryor’s organised one; and he is included not in spite of how little he resembles Pryor but because of it, since the features the two men nonetheless share are, by the logic set out in Chapter 3, the features worth attention. The case rests on the trial record (State v. Beck 1983) and on the fullest secondary account of it (Hale 1988).

5.1 The offender and his history

If the developmental record in Pryor’s case was thin, in Beck’s it is overwhelming, and it is the part of his story that is least in doubt. He was born in 1949 into a household at the bottom of Youngstown’s poverty, to a mother who was herself unwell — variously described in the records as alcoholic and as mentally ill, and probably both — and who could not care for him. His childhood was a succession of removals and returns: periods in the care of the state, periods back in a home marked by neglect, hunger, and the violence of a series of men who passed through it. He was beaten, and worse; he was, by several accounts, a frightened and peculiar child, slow at school, friendless, already showing the disturbance that would deepen in adult life. Whatever else is uncertain about Howard Beck, that he was comprehensively failed as a child is not.

As an adult he never established himself in any settled way. He drifted between menial jobs and periods of homelessness, drank heavily, and was twice admitted briefly to psychiatric care and twice released without sustained treatment. By the late 1970s — by which time the steel mills that had defined the city were closing for good — he was living alone in a single room on the East Side, amid derelict housing and shuttered industry, isolated almost completely, known to his neighbours only as a strange and silent man they avoided. It was from this room, and within a few streets of it, that he killed.

5.2 The psychological picture

The psychological picture is harder to draw than Pryor’s, and for the opposite reason: not because there is too little disturbance to see, but because there is so much of it that the lines between its kinds are hard to fix. Beck was seriously mentally ill. There is reliable evidence of disordered thought, of periods in which he was out of contact with reality, and, on his own later report, of voices — though that report, like all such, is his, and is further clouded by the heavy drinking that accompanied it. At his trial the question of his sanity was argued at length; the defence of insanity was raised and, in the event, rejected, the court holding that whatever his disturbance, he had known what he did and that it was wrong.

For the purposes of this study the important point is a narrower one, and it must be put carefully. Beneath the gross disturbance, the same affective deficit found in Pryor is discernible in Beck — the same absence of fellow-feeling, the same failure to register his victims as persons — but here it is tangled with psychosis and intoxication in a way that makes it difficult to isolate. It would be too much to call Beck simply a psychopath in Cleckley’s sense; the clean, masked presentation Cleckley described is precisely what Beck lacked. Yet the coldness at the centre is recognisably the same coldness, and the temptation to explain it away as merely a symptom of his illness should be resisted, since most of the seriously mentally ill, including most of those with Beck’s diagnoses, harm no one. The illness is real, and it shaped the form of his offending; it is not, by itself, what made him kill.

5.3 The pattern and method

Beck’s victims were the people of his own world: the homeless, the addicted, those who sold sex in the same derelict streets, men as well as women, all of them people whose absence Youngstown, struggling with its own collapse, was slow to notice. He did not seek them out across distances, as Pryor did; he killed those who came within reach, in or near the empty buildings around his room.

The killings were as disordered as Pryor’s were controlled. There was little planning and no weapon kept for the purpose; he used what was to hand, and the scenes he left were chaotic. The bodies were not removed to any distance but left where they fell, in vacant lots and abandoned houses, some not found for a long time and some, it is believed, never found at all. The contrast with Pryor’s careful disposals is total, and it is the contrast the typology is built to capture.

Yet beneath the disorder there was, as the framework would predict, a constant. Across the killings that could be attributed to him, Beck covered the face of each victim — with cloth, with whatever lay nearby — a small, consistent act that served no practical purpose and that he was never able, or willing, to explain. It is the one orderly thing in an otherwise disordered series, and it is, in the terms of Chapter 2, his signature: the stable trace that persists across changing circumstances, and that marks the killings, for all their chaos, as the work of one man.

One further feature distinguishes the series from Pryor’s. Pryor’s method, over five years, grew more controlled; Beck’s, over four, grew less. The later killings were more frequent and more chaotic than the earlier, consistent with a man not refining a method but decompensating — coming apart, as his illness and his drinking advanced, faster than any caution could keep pace with. The iterative process described in Chapter 2 is present in both men, but it runs, in Beck, in the opposite direction.

5.4 The investigation and its failures

The institutional failure in Beck’s case was of a different kind from that in Pryor’s, and in some ways a worse one. Pryor escaped notice because his crimes were scattered across the boundaries of forces that could not see across them; Beck escaped notice though his crimes were concentrated in a single district, because the people he killed were the people the system was least moved to count. Several of his victims were never recorded as victims at all: their deaths, where the bodies were found, had first been entered as overdoses, as exposure, as misadventure. One of them, Raymond Dolet, a man of no fixed address, had been recorded as a death by exposure, and was returned to the count only after Beck’s arrest; he is named here because almost no one else did. The disappearances of the others, where no body was found, had gone unreported, because no one was waiting for them to come home. This is the failure Egger (1990) names in his account of the “less-dead,” and Beck’s case is among its clearest instances. The series was invisible not because it was well hidden but because no one with the power to see it was looking.

When Beck was at last stopped, it was through the courage of Marlene Voss, a woman who escaped him and went to the police — and who was, at first, not believed, her account discounted for the same reasons that had kept the series invisible, until physical evidence at Beck’s room compelled attention to it. She is named here, where several of the dead cannot be, because she gave her evidence in open court and deserves the credit for what her courage achieved: the case was broken by the kind of victim the system had been most inclined to ignore. The evidence found in his room then connected him to deaths that had been recorded, months and years before, as nothing in particular.

Beck was convicted in 1983 of three murders and sentenced to life imprisonment. How many he actually killed is not known and cannot now be established, because most of the suspected victims had never been treated as homicides while the trail was warm. The number three is not a measure of what he did; it is a measure of what could be proved. He remains in custody.

5.5 The case against the framework

Read against the framework, Beck’s case is in one respect the mirror of Pryor’s. The developmental substrate, thin and inferred in Pryor, is in Beck overwhelming and beyond serious doubt — the strongest element of his case. The iterative process is present, though it runs toward collapse rather than toward control. The psychological architecture is present too, in the coldness at the centre, but it is the hardest element to read cleanly, entangled as it is with psychosis and drink in a way Pryor’s was not. Where Pryor offered firm evidence for the architecture and only suggestive evidence for the substrate, Beck offers very nearly the reverse. That the two cases are strong in different places, and weak in different places, is not a failing of the framework but the beginning of its most interesting result, and it is taken up in Chapter 7.

---

Chapter 6: Case Study Three — Desmond Frawley

The third case is Australian, and is included for three reasons: because the imported frameworks of Chapter 2 ought to be tested against a local instance and not merely assumed to fit one; because Frawley represents a third evidentiary situation, distinct from both that precede it; and because his case, alone of the three, was never fully resolved, and so brings the institutional concerns of this study to their sharpest point. Desmond Frawley was convicted in 1990 of three murders committed in the Riverina district of New South Wales during the 1980s. He is suspected, on reasonable grounds, of a number of others that have never been proved, because the victims have never been found. The case rests on the trial record (R v Frawley 1990) and on the coronial inquest that followed it (Coroner’s Court of New South Wales 1991), which examined both the deaths and the long failure to connect them.

6.1 The offender and his history

Frawley was born in 1948 near Hay, in the far west of the Riverina, into the itinerant rural working class of the inland — a world of shearing sheds, fencing contracts, and seasonal work, in which men moved with the work and families moved with the men, or were left behind. His father was a shearer and a heavy drinker, present rarely and violently; his mother, who carried the household, died when Frawley was nine, after which he was raised partly by an uncle and partly in a boys’ home, neither of which the record describes as kind. He left school early and badly, and went onto the land as his father had — fencing, shearing, and in time driving — work that kept him on the roads and back tracks of the inland for the rest of his free life.

Less can be said of these years than of Pryor’s, or even of Beck’s, and the reason is characteristic of the whole case: Frawley left little trace and volunteered nothing. The outline — the absent and violent father, the lost mother, the institution, the early hard isolation — is established. What it felt like from the inside, and what in it bent him, he never said.

6.2 The psychological picture

The psychological picture of Frawley is the thinnest of the three, and is thin for a reason that is itself informative. Where Pryor talked, after his conviction, in interviews that were self-serving but at least revealing, and where Beck’s disturbance was so gross that it declared itself, Frawley did neither. He made no confession beyond what the evidence compelled; he gave no interviews; he refused, to the end, to discuss the killings, the victims, or himself. The court that convicted him, and the coroner who later examined the case, were left to construct what picture they could from his conduct alone.

That picture is necessarily partial. Frawley was not mentally ill in any sense the court could find; he was assessed as sane and held fully responsible. He was not chaotic. The coldness that the framework treats as the psychological architecture of the serial offender is present in him — in the controlled and practical character of the killings, in a lifelong incapacity for any close attachment, in the complete absence, through trial and imprisonment, of anything resembling remorse — but it is inferred from what he did and how he bore himself, not confirmed from anything he disclosed. Of the inner life behind the killings — of fantasy, of motive, of whatever it was the killing answered to — nothing can responsibly be said at all, because Frawley took care that nothing should be known, and largely succeeded.

This silence, frustrating as it was to those who sought to understand him, is in one respect the most instructive feature of the case. It demonstrates, more clearly than either of the others, a claim made in Chapter 2: that the offender’s motive is the least accessible thing about him, and that a man may be identified, convicted, and understood as a type while his reasons remain wholly opaque. Frawley was caught and convicted not because anyone grasped why he killed — no one did, and no one does — but because of how he killed, which he repeated. The case is, in this sense, the strongest vindication in the study of the priority of method over motive.

6.3 The pattern and method

Frawley’s victims were travellers. The Sturt Highway and the roads that feed it carry, and in the 1980s carried in greater numbers, a steady traffic of young people moving for work and for its own sake — hitchhikers, seasonal fruit-pickers and shearers’ labourers following the season, the occasional backpacker — people in motion, often alone, and frequently expected nowhere in particular at any fixed time. It was among these that Frawley chose, and his work and his vehicle placed him perfectly to do so: a man in a utility or a truck, offering a lift on a long empty road, was not merely unalarming but welcome.

What he did with them was the practical inversion of what Pryor did. Pryor left his victims to be found, on open moorland, taking from each only a token; Frawley took everything and left nothing. He killed — by firearm, so far as the recovered remains could show, a thing as unremarkable in that country as a fencing strainer — and he buried the bodies in the vast and trackless scrub of the inland, in country he knew from a working lifetime and others did not, having first removed and destroyed whatever might have identified them. The effect, whether intended or not, was that his victims did not become victims in any official sense; they became, simply, people who had not arrived, in a part of the world where not arriving was easy to explain.

Yet here too, beneath the erasure, there was a constant, and it was the erasure itself: the same choice of the traveller over the local, the same use of the lift, the same method of killing, the same burial in remote country, the same thoroughness in removing all trace of who the dead had been. It is a signature written, as it were, in absences — but it is a signature, and it is what eventually allowed the few recovered cases to be recognised as the work of one man.

6.4 The investigation and its failures

If the Pryor case shows linkage blindness as the failure of a system divided into jurisdictions, the Frawley case shows it compounded by distance and by assumption. The disappearances were spread across years and across hundreds of kilometres of a thinly policed inland; they fell to different stations and different districts; and, most damagingly, almost every one of them could be, and was, explained away. A hitchhiker who fails to arrive has, after all, simply moved on; an itinerant worker who is not seen again has gone to the next town; a young person who vanishes on the road has, it is assumed, chosen the road. Each disappearance, taken alone, admitted an innocent explanation, and each was for years given one. That the same innocent explanation was being offered, again and again, along the same highway, was precisely the pattern no one was positioned to see.

The victims, too, were of the kind whose absence is slow to register — not the destitute of Beck’s city, but the young and the mobile, whose whole manner of life was to be unaccounted for, and whose disappearance therefore raised no immediate alarm. The assumption that they had “moved on” did the work, in the Riverina, that the assumption of the overdose had done in Youngstown: it converted a victim into a non-event.

The pattern was seen, in the end, by one man before it was seen by the institution. Detective Sergeant Gordon Vickery, of the New South Wales police, became convinced over several years that the scattered disappearances along the corridor were connected, and pressed — against the prevailing view, and at some cost to himself — for them to be examined together. He was given his hearing only when the disappearance of a young couple, Stephen Maddox and Robyn Searle, both in their early twenties, proved harder to explain away than those before it: they were travelling together, were expected by families who would not accept that they had simply gone, and had left their car, found abandoned on the highway, behind them. The visibility of that one disappearance did what the invisibility of the others had prevented. It forced the cases to be looked at together, and once they were, Vickery’s pattern was there to be read.

Frawley was identified through a combination of the couple’s abandoned vehicle, a witness who had seen Robyn Searle accept a lift from a distinctive utility, and an earlier complaint, until then buried in a separate file, from a young woman who had escaped a man of his description some years before. A search of the country he was known to work led to three graves: Maddox and Searle, and, some distance away, a nineteen-year-old named Kathy Renshaw, who had been missing since 1984 and whom no one had been looking for. On these three he was convicted. He has never disclosed whether there are others, or where they lie, though the inquest concluded that there almost certainly are. Among the still-missing is a seventeen-year-old, Lee-Anne Castle, last seen at a roadhouse outside Narrandera in 1983, whose family continues to wait. That waiting, which Frawley could end with a sentence and chooses not to, is the case’s final cruelty, and it is of a kind the previous two cases did not contain.

Frawley was tried in 1990, convicted of three murders, and sentenced to life imprisonment, the court recommending that he never be released. He remains in prison, and remains silent. The coronial inquest of 1991, unable to determine the fate of the missing, confined itself to the deaths it could establish and to a careful account of why the others had gone so long unconnected — an account on which the institutional sections of this study have drawn.

6.5 The case against the framework

Read against the framework, Frawley is the case in which the evidence is at once least complete and, in one respect, most pointed. The developmental substrate is present in outline — the violence, the lost mother, the institution, the early isolation — but is, as in Pryor, thin, and unilluminated by anything the offender chose to add. The psychological architecture is present in the coldness and the controlled practicality of the killings, but it is wholly inferred, since Frawley disclosed nothing of himself. The iterative process is present in the constancy of the method across the recovered cases and, by reasonable inference, across the suspected ones. Of fantasy and motive there is, uniquely, nothing — not thin evidence, but none.

And yet the case identifies its man as surely as either of the others. That it can do so while the offender’s inner life remains entirely closed is the lesson the cross-case analysis takes up first: that the features by which a serial offender is known are not, in the main, the features of his mind, to which there may be no access at all, but the features of his conduct, which he cannot help but repeat. Frawley understood, better than most, the value of leaving nothing of himself behind. He could not leave behind nothing of his method, and it was his method that named him.

---

Chapter 7: Cross-Case Analysis

7.1 The comparison restated

The three cases have now been set out, each on its own terms. They were chosen, as Chapter 3 explained, to resemble one another as little as possible. Pryor was English, organised, outwardly respectable, mobile across a settled landscape, and active a generation ago; Beck was American, disorganised, mentally ill, fixed to a single ruined district, and active a decade later; Frawley was Australian, self-contained and silent, working the empty distances of the inland, and active later still. They differed in nationality, in period, in class, in the presence or absence of mental illness, in the degree of planning, in the disposition of their victims, and in nearly every particular a reader might think to compare. The purpose of choosing them so was to ensure that whatever they were nonetheless found to share could not be put down to a shared world. This chapter sets out what that is.

7.2 What recurs: the developmental substrate

The developmental substrate — early deprivation, maltreatment, and the failure or rupture of any stable attachment — is present in all three histories, but it is present with very different degrees of evidential force. In Beck it is overwhelming and beyond doubt: a childhood of poverty, neglect, violence, and institutional handling so comprehensive that it would be difficult to design a worse one. In Pryor and Frawley it is established in outline — the violent father, the lost mother, the cold or absent home, the early severing of any reliable bond — but the outline is thin, and in Pryor’s case partly dependent on the offender’s own untrustworthy telling.

This unevenness must be met honestly, and met twice. The first point is that absence of evidence is not evidence of absence: that the developmental record for Pryor and Frawley is thin reflects the poverty of the sources, not the demonstrated absence of early damage, and the outline that survives points the same way in all three. The second point is the one already conceded in Chapter 3, and it bears most heavily here. A damaged childhood of the kind found in these three histories is, sadly, not rare. It is found in the histories of a great many people who never harm anyone. The comparison can establish that the substrate is present in all three offenders; it cannot, having no group of non-offenders with comparable histories to set beside them, establish that the substrate is what made them offenders. The most that can be claimed is that the substrate is a recurring background condition — necessary, perhaps, but plainly not sufficient.

7.3 What recurs: the psychological architecture

The second strand recurs more strikingly than the first, because it recurs across more difference. The affective deficit — the absence of ordinary fellow-feeling, the failure to register the victim as a person rather than an object — is present in all three men, and it is present despite the fact that in every other psychological respect they differ. Pryor wore it behind a seamless social mask; Beck’s was half-buried under psychosis and drink; Frawley disclosed nothing and had it read from his conduct. The surface presentations could hardly be less alike. Beneath them, the same coldness is found in each.

To the affective deficit the framework joined a need for control or mastery, and here the picture is more qualified. Control is conspicuous in Pryor, whose whole life and method were ordered by it; it is plain enough in Frawley, in the practical, concealing competence of his killings; it is least clear in Beck, whose disorder worked against it, and in whom the single act of control — the covering of the face — sat amid chaos. Control, then, recurs, but not uniformly, and it is better described as a frequent companion of the affective deficit than as an invariable part of the architecture. The deficit itself is the firmer finding: of all the features considered in this study, the coldness at the centre is the one that survives every difference between the three men, and it is, on this evidence, the strongest candidate for a feature genuinely characteristic of the type.

7.4 What recurs: the iterative process, and a correction

The third strand — the assembling of a method through repetition — is present in all three, but the cases require the framework to be corrected on one point. Chapter 2 described the iterative process as though it ran in a single direction: toward refinement, the method improving as the offender learned. Pryor’s case fits that description: his method grew more controlled and more careful across five years. But Beck’s ran the other way. His killings grew not more controlled but less — more frequent and more chaotic, as his illness and his drinking advanced; the process in him was one not of refinement but of decompensation. Frawley’s, so far as the recovered cases allow it to be read, was stable rather than either — a method effective from early and little altered.

The framework, then, was right that a series has a developmental shape of its own, and wrong to assume that the shape is always one of improvement. The iterative process is real and recurs, but its direction depends on the man: it refines where he is controlled, collapses where he is coming apart, and holds steady where he was competent from the first. This is a genuine result of the comparison, and one the framework did not anticipate; it is recorded here because a study that reported only its confirmations would be worth little.

One feature of the process, however, recurs without qualification: the signature, the stable trace that persists across changing circumstances. Each of the three returned, across his series, to the same essential act — and, remarkably, each enacted his relationship to the victim’s identity in a constant and characteristic way. Pryor took a token, keeping a piece of each victim; Beck covered the face; Frawley erased, removing all trace of who the dead had been. Three more different gestures could hardly be imagined, and yet each was performed every time, by its author, regardless of circumstance. The content of the signature varied completely between the men. That there was a signature did not vary at all.

7.5 The typologies reconsidered

The comparison also allows a return to the organised–disorganised distinction set out, and doubted, in Chapter 2. The three cases span its range. Pryor is the organised offender very nearly as the textbook draws him — planful, controlled, socially camouflaged, leaving little behind. Beck is his disorganised opposite — impulsive, chaotic, leaving his scenes in disarray. Frawley sits closer to the organised pole than the disorganised, though his rural improvisation fits neither box cleanly. Set side by side, the three confirm that the distinction captures something real: the tidiness of the scene, the degree of planning, the presence or absence of control are genuine differences, and they sort these men much as the typology says they should.

But the comparison shows just as plainly what the distinction does not reach. The features this study has found to recur — the affective coldness, the signature and its repetition, the priority of method over motive, the choosing of victims for accessibility and invisibility — are shared by all three men, and they cut clean across the organised–disorganised line. The typology places Pryor and Beck in opposite categories, and at the level of the crime scene they belong there; but at the level with which this study is concerned, the level of what the men are and how they are known, they are far more alike than the categories allow. The distinction sorts the surface. It does not sort the depth.

There is one respect, and perhaps only one, in which it reaches deeper than the scene, and the comparison brings it out. The trajectory of the iterative process, discussed above, appears to track the organised–disorganised divide: the controlled offender refines his method, the disordered one decompensates. Pryor grew more careful; Beck grew less. To that extent the typology describes not merely how a scene looks at one moment but which way a series is likely to move over time, which is a more useful thing to know. Beyond it, however, the caution of Chapter 2 stands. The organised–disorganised distinction is a serviceable description of conduct and a poor classification of persons, and the three cases, alike beneath their differences, are the demonstration of why.

7.6 An unpredicted commonality: the choosing of victims

Beyond the three strands of the framework, the comparison brings to light a fourth shared feature that the framework, being concerned with the offender’s psychology, did not predict, and that concerns instead his victims. In all three cases the victims were chosen on two criteria at once: accessibility and invisibility. They were people the offender could reach — women alone on roads, the marginal of a single district, travellers in motion — and they were people whose absence would be slow to register: the assumed-flighty, the uncounted, the expected-nowhere. The two criteria are distinct, and their convergence is significant. The first serves the offender’s opportunity; the second serves his impunity. That all three men, differing in everything else, selected victims who satisfied both, suggests that the choosing of victims is not the incidental matter it is sometimes treated as, but is itself part of the method — and, as the next sections argue, the point at which the offender’s psychology and the institution’s failure meet.

7.7 Method over motive

The clearest single result of the comparison concerns not the origins of the offenders but the means by which they are known, and it is the result on which this study would chiefly rest. Across the three cases the offender’s motive ranged from the partly stated to the wholly unavailable: Pryor offered an account of his, self-serving and unverifiable; Beck’s was lost in psychosis; Frawley’s was never disclosed and remains entirely unknown. If the identification and understanding of these men had depended on grasping why they killed, two of the three could not have been identified at all, and the third only on his own doubtful word.

It did not depend on it. Each man was identified, in the end, by his method and its signature — by what he repeated, which he could not help repeating, and which was the same across all his crimes whatever was happening in his mind. Frawley is the purest demonstration: a man who gave nothing of himself, whose reasons are wholly closed, and who was nonetheless named by the constancy of his conduct. The lesson generalises across the three. The motive of a serial offender is the least accessible thing about him, and the least reliable when accessible; his method is the most accessible and the most stable; and it follows that the offender is most surely read not through his reasons but through his repetitions. This is the steadier ground promised in Chapter 2, and the cases bear the promise out.

7.8 The institutional constant: linkage blindness in three forms

If method is the constant in the offenders, linkage blindness is the constant in the institutions that failed to catch them. In every one of the three cases the series ran on, undetected, for years — and in every one the reason was, at bottom, the same: the system that ought to have connected the crimes did not, and could not, see across its own divisions. The form of the division differed. For Pryor it was jurisdictional and technological — four county forces, no shared record, in the years before such records existed. For Beck it was a division of value — a system that did not trouble to connect the deaths of people it had never troubled to count. For Frawley it was a division of distance and assumption — disappearances scattered across an inland so wide, and so readily explained by the supposition that the missing had simply moved on, that no pattern could form in any single official mind.

In two of the three, and arguably in all three, the failure fed on the offender’s choice of victim. The invisibility that recommended the marginal and the mobile to the offender as safe targets was the same invisibility that kept the institution from noticing they were gone. Here the psychological and the institutional findings of this study meet: the offender selects for low visibility, and the institution obliges him by failing to look. The victim chosen for impunity is the victim the system was already least inclined to pursue.

It is worth observing how each series ended, because the manner is consistent and instructive. None of the three was stopped by the recognition of its pattern. Pryor was caught by a constable’s note of a parked car; Beck by the courage of a single survivor who insisted on being heard; Frawley by the one disappearance too visible to dismiss, and by the persistence of one detective who had seen the pattern years before the institution would look at it. In every case the pattern was there, and had been there for years, available to be read. What was missing was never the pattern. It was the reading.

7.9 What the comparison does not show

The results set out above should be held together with their limits, which are real. Three cases are three cases; the recurrences described are established for these three, and are offered as candidates for attention, not as laws. The developmental substrate, though present in all three, is read from thin and uneven records in two of them, and — more seriously — cannot, in the absence of any comparison group, be shown to distinguish these men from the many who share their histories and harm no one. The psychological architecture, firmest of the findings, is still inferred rather than measured, and in Frawley’s case inferred from conduct alone. The study can say, with some confidence, that these features recur together in these offenders, and that the recurrence is striking given how little else the offenders share. It cannot say how common the pattern is, nor that it is the pattern that makes the offender, and it does not pretend to. What can be drawn from the comparison, and what cannot, is the subject of the conclusion that follows.

---

Chapter 8: Conclusion

8.1 What the study found

This dissertation set out to determine, by the close comparison of three documented serial murderers chosen to resemble one another as little as possible, whether such offenders share a small set of recurring features. It found that they do, though the findings are uneven in their strength, and that the recurrences are of two kinds: those the framework of Chapter 2 anticipated, and those it did not.

Of the three strands the framework proposed, all three were present in all three cases. The developmental substrate — early deprivation and failed attachment — was found in every history, though overwhelming in only one and thin in the other two, and it remains, for the reason given below, the least secure of the findings. The iterative process — the assembling of a method through repetition — was likewise present in all three, but the comparison corrected the framework’s assumption that it always tends toward refinement: in fact it refined in one man, collapsed in another, and held steady in the third, its direction depending on the offender rather than fixed. The psychological architecture — the affective deficit, the coldness that fails to register the victim as a person — was the firmest finding of all: it was present in every case, beneath surfaces that could hardly have differed more, and it is the single feature that survived every difference between the three men.

To these the comparison added two findings the framework had not predicted. The first concerns the victims: all three offenders chose them by the same double criterion of accessibility and invisibility, selecting people they could reach and whose absence would be slow to register. The second, and the most important practically, concerns how the offenders were known. In all three cases the offender’s motive was obscure, unreliable, or wholly unavailable, and in all three he was identified not by his motive but by his method — by the signature he repeated and could not help repeating. The priority of method over motive, proposed in Chapter 2 as the steadier ground for understanding such men, was borne out by every case.

8.2 Implications for investigation

The investigative implications follow from this last finding, and they are modest and largely structural rather than predictive. If the offender is most reliably known by his repetitions, then the first task of any system that hopes to catch him is to be able to see those repetitions — which is precisely what, in all three cases, the system could not do. The crimes were connected, when they were connected at all, late, by accident, or by the persistence of a single individual, and never by the apparatus that ought to have connected them. The plainest practical lesson of this study is therefore the oldest: that the reduction of linkage blindness — the building of the means by which cases in different places and different hands can be compared — is worth more than any amount of insight into the offender’s mind, because it addresses the failure that actually let each of these men continue.

Two cautions attach to this. The first is that the victims most useful to such an investigation are precisely those the system is least inclined to attend to. The offender selects the invisible for their invisibility; an investigation that shares the institution’s habitual inattention to the marginal will share, too, its blindness to the series they belong to. The second caution is against the opposite error: the confident profile of the unknown offender, built from his crimes, which the history of the subject shows to be unreliable as often as not. The contribution this study offers to investigation is not a portrait of the man to be sought, but an insistence that the pattern of his acts be looked for, and looked for hardest where the looking is least likely to be done.

8.3 Implications for prevention

On prevention the study counsels still greater caution, and most of what it has to say is a warning against saying too much. The developmental substrate found in these histories might seem to point toward prevention through early intervention — the reduction of the childhood maltreatment and the failed attachment that recur in the offenders’ pasts. That reduction is worth pursuing, but not for the reason the temptation suggests. The substrate cannot be used to identify the future offender, because, as Chapters 3 and 7 both insisted, it is shared by a great many people who never harm anyone, and there is no means, in the present state of knowledge, of telling in advance which damaged child is the one in many thousands who will kill. Any scheme that proposed to screen for future offenders on the basis of such a history would be wrong far more often than right, and would do great harm in being wrong. The honest implication is broader and quieter: that the conditions found in these histories are an affliction worth reducing for the sake of all the children who suffer them and do no harm, and that whatever reduction of grave violence might follow would be an incidental good, not a thing that could be promised or targeted.

8.4 What the study has not shown

It remains to state, a final time, what the study has not done, since its value depends on the boundary being clear. It has not shown that the features it found are the cause of serial offending, nor even that they distinguish the offender from the many who share them; lacking any comparison group, it cannot. It has not shown that the pattern it describes is common, or representative, or present in offenders other than these three; three cases cannot. It has not measured anything; the psychological architecture, firmest of its findings, is inferred from conduct and record, and in one case from conduct alone. The study establishes that a small set of features recurred together across three men who shared almost nothing else, and that the recurrence is striking enough to deserve attention and further work. It establishes no more than that, and should not be read as doing so.

8.5 Conclusion

What the three cases finally have in common is not only a set of features in the offenders but a single fact about the institutions that failed them, and it is with that fact that this study ends. In each case the pattern was there. It was present in the crimes, available in the records, legible to anyone who set the cases side by side; and in each case it went unread for years — not because it was hidden, but because no one was positioned, or moved, to look. The offenders were caught at last by chance, by courage, and by the stubbornness of individuals; they were not caught by the system that ought to have caught them, and the cost of that failure was paid, in every case, by the people the system was already least inclined to count.

If this study has an argument beyond its findings, it is that the patterns such offenders leave are not beyond understanding, and that the failure to see them is a failure not of possibility but of attention. The serial offender is legible. He is legible in what he repeats, which he cannot help, and which is there in the record to be found. Whether he is found depends not on the depth of the pattern but on the discipline of the looking — and that discipline, in the end, is something a system chooses, or chooses not, to exercise. The choosing is the whole of it.

---

References

Cases and official proceedings

R v Pryor (Crown Court, 1973).

State v. Beck (Court of Common Pleas, Mahoning County, Ohio, 1983).

R v Frawley (Supreme Court of New South Wales, 1990).

Coroner’s Court of New South Wales 1991, Inquest into Three Deaths and Certain Disappearances in the Riverina District. Sydney.

Books and articles

American Psychiatric Association 1987, Diagnostic and Statistical Manual of Mental Disorders, 3rd edn, revised. Washington, DC: American Psychiatric Association.

American Psychiatric Association 1994, Diagnostic and Statistical Manual of Mental Disorders, 4th edn. Washington, DC: American Psychiatric Association.

Athens, L.H. 1992, The Creation of Dangerous Violent Criminals. Urbana: University of Illinois Press.

Bowlby, J. 1969–1980, Attachment and Loss, 3 vols. London: Hogarth Press.

Brittain, R.P. 1970, ‘The sadistic murderer’, Medicine, Science and the Law, vol. 10, pp. 198–207.

Cleckley, H. 1976, The Mask of Sanity, 5th edn. St Louis: Mosby. (First published 1941.)

Douglas, J.E., Burgess, A.W., Burgess, A.G. & Ressler, R.K. 1992, Crime Classification Manual. New York: Lexington Books.

Egger, S.A. 1984, ‘A working definition of serial murder and the reduction of linkage blindness’, Journal of Police Science and Administration, vol. 12, no. 3, pp. 348–357.

Egger, S.A. 1990, ‘Serial murder: a synthesis of literature and research’, in S.A. Egger (ed.), Serial Murder: An Elusive Phenomenon, pp. 3–34. New York: Praeger.

Hale, T. 1988, The Quiet Rooms: The Howard Beck Murders. New York: Dutton.

Hare, R.D. 1991, The Hare Psychopathy Checklist–Revised. Toronto: Multi-Health Systems.

Hare, R.D. 1993, Without Conscience: The Disturbing World of the Psychopaths Among Us. New York: Pocket Books.

Holmes, R.M. & DeBurger, J. 1988, Serial Murder. Newbury Park, CA: Sage.

Leyton, E. 1986, Hunting Humans: The Rise of the Modern Multiple Murderer. Toronto: McClelland & Stewart.

MacDonald, J.M. 1963, ‘The threat to kill’, American Journal of Psychiatry, vol. 120, pp. 125–130.

Mercer, A. 1979, The Pryor Case. London: Hutchinson.

Mill, J.S. 1843, A System of Logic, Ratiocinative and Inductive. London: John W. Parker.

Prentky, R.A., Burgess, A.W., Rokous, F., Lee, A., Hartman, C., Ressler, R.K. & Douglas, J.E. 1989, ‘The presumptive role of fantasy in serial sexual homicide’, American Journal of Psychiatry, vol. 146, no. 7, pp. 887–891.

Raine, A. 1993, The Psychopathology of Crime: Criminal Behavior as a Clinical Disorder. San Diego: Academic Press.

Ressler, R.K., Burgess, A.W. & Douglas, J.E. 1988, Sexual Homicide: Patterns and Motives. Lexington, MA: Lexington Books.

Widom, C.S. 1989, ‘The cycle of violence’, Science, vol. 244, no. 4901, pp. 160–166.

Yin, R.K. 1989, Case Study Research: Design and Methods. Newbury Park, CA: Sage.