Argument and Evidence

Critical Analysis for the Social Sciences

Peter J. Phelan, Peter J. Reynolds

• Publisher: Routledge
• Year: 2002
• Print ISBN: 9780415113731
• eBook ISBN: 9781134824007
• Edition: 1
• Topic: Politics & International Relations
• Subtopic: Politics
• Index: Politics & International Relations

Phelan and Reynolds' book is for anyone who needs to evaluate arguments and interpret evidence. It deals with the most fundamental aspects of academic study:
* the ability to reason with ideas and evidence
* to formulate arguments effectively
* to appreciate the interplay between ideas and evidence in academic and media debate
Argument and Evidence presents aspects of informal logic and statistical theory in a comprehensible way, enabling students to acquire skills in critical thinking which will outlast their undergraduate studies. Ideal as a companion for courses on methodology or study skills, Argument and Evidence will also be useful for other disciplines in the social sciences and humanities.

1 Introduction

This book considers and elucidates several ways in which argument and evidence can be used to reinforce or counter a particular stance. But why is this necessary? After all, we have been subjected to arguments from the earliest of ages and surely, by now, we can distinguish between a ‘good’ argument and a ‘bad’ one. As for interpreting evidence, is it not a matter of common sense to see when a particular point of view is supported by the ‘facts’ and when it is not? Further reflection might suggest that there is no such thing as a right argument and no such thing as a wrong one. Yet, if the concepts of right and wrong are inapplicable, how do we choose which arguments to believe? Does it even make sense to refer to the notion of ‘belief when considering arguments? In the following chapters, unambiguous criteria are developed for formulating and evaluating arguments. But what about evidence? What counts as evidence? Can evidence be used to prove an argument? How does one evaluate the strength of a piece of evidence? The answers are by no means clear-cut. If they were, then there would be little scope for controversy to occur.

THE BURDEN OF PROOF

The elusiveness of conclusive evidence is something that students and researchers are familiar with and manifests itself in many ways. It provides one of the main reasons for the existence of most of the legal profession. It is the elusiveness of irrefutable proof that gives rise to the expression ‘beyond reasonable doubt’, which enters into the instruc-tions given to jurors. An aspect of proceedings in British criminal courts is that a person is presumed innocent unless found guilty. The burden of proof falls squarely on the prosecution to present sufficient evidence that the accused person is guilty ‘beyond reasonable doubt’.¹ Yet even this apparently straightforward notion can lead to many difficulties, particularly if extended to applications beyond the confines of the criminal courts. This is illustrated by the following examples, drawn from various walks of life.

Early in 1991, in dawn raids on the homes of families living in the Orkney Islands, representatives of the welfare services snatched a number of children from their beds and took them away from their parents, placing them in the care of the local authorities. There were accusations that the children had been abused whilst under parental care. Rumours of strange rituals abounded and many parents faced court cases. The case against the parents was thrown out and the children were returned to their homes. Public attention then turned on the local officials.² The line of argument implicit in much press coverage was that if the parents were innocent then the officials must, at the very least, be guilty of over-zealously pursuing their public duty. Yet, this does not follow. If the local officials are to be given the same standards of justice accorded to the parents, then surely they too must be assumed innocent unless sufficient evidence is weighed against them. We appear to reach the paradoxical situation of being unable to find either party guilty. The reason is that a sufficient weight of evidence cannot be amassed in support of either case.

Not long ago, a young couple were arrested, following the death of their child. It seemed quite clear that at least one of the parents had a hand in the child’s death. The mother was tried and found ‘not guilty’. The father was tried and found ‘not guilty’. Each was subsequently acquitted. Despite a wealth of evidence to the effect that at least one of the parents was guilty, taken individually it was not possible to amass sufficient evidence against either one.

One of the largest public health scares in recent years occurred in 1990. It was the alleged threat to human beings of contracting bovine spongiform encephalopathy (BSE), as a result of eating infected beef. The Government faced a most tricky dilemma. Should it ban the sale of beef, thereby condemning many farmers and other members of the meat trade to financial ruin or should it discount the threats to public health? A reasonable response would be to look at the evidence. Unfortunately, there was no conclusive evidence then to be found. The Government, through the appropriate Minister, issued public statements to the effect that there was insufficient evidence to establish that BSE could be passed on to human beings.³ At the same time, various measures were taken to reduce the risk of humans eating contaminated meat and the public was reassured. It is a sobering thought that equally accurate statements could have been made to the effect that ‘there was no evidence that BSE cannot be contracted by humans’!⁴ In considering this case—and many like it—a crucial consideration must be where to place the burden of proof. For whatever reason, the Government decided to place the burden on those who alleged that public health was threatened.

THE JUXTAPOSITION OF ARGUMENT AND EVIDENCE

The controversial nature of the cases mentioned above suggests that argument and evidence each have a part to play in settling disputes. It is appropriate to consider how argument and evidence combine in the presentation and evaluation of cases. Yet, to our knowledge, there is no single book which offers a unified treatment both of the structure of argument and of the use of evidence in the evaluation of positions adopted on controversial issues. Much of the academic discussion of the nature and appraisal of arguments occurs in logic texts while serious analytical treatment of evidence is the subject matter of books on statistical analysis. Many textbooks of both kinds are not accessible to the non-specialist. This book does draw on the literature from both disciplines. Equally importantly, it incorporates both a discussion of the requirement that sound arguments allow evidence to weigh in the balance and a consideration of how questions about what counts as evidence impact on the ways in which hypotheses are formulated. This approach aims to be sensitive to the interplay between argument and evidence and their bearing on methods of study and research design.

What follows is not a watered down version of material on formal logic. Our treatment of logical matters applies some of the ideas and approaches to informal logic, developed by philosophers in the USA and more recently in the UK.⁵ Nevertheless, it is proper to respect what experts have to say and so we present the results of the formal logicians’ endeavours so that the reader may acquire skills to recognise patterns of reasoning that can be trusted. In this way, some of the most important contributions in logic can be made available to a wider audience without the necessity for a formal training in some of the more difficult aspects of the discipline.

Similarly, our treatment of statistics should be considered in the context of the aims of this book. For example, most textbooks on statistics address the subject of ‘hypothesis testing’ from the starting point of having already formulated a hypothesis to be tested. From the previous discussion of ‘the burden of proof’, it should be apparent that the formulation of a hypothesis is by no means a straightforward matter. For the issue concerning BSE, should the hypothesis be that BSE does pose a threat to human beings or should it be that it does not? Although, to any reader not trained in elementary statistical analysis, this may seem a matter of simple semantics, the way in which a hypothesis is initially formulated can have a crucial bearing on the outcome of any subsequent test.

THE CONTEXT OF ARGUMENT AND EVIDENCE

In this section we use a topical controversy to explain and illustrate the importance of context to the interpretation of argument and evidence. An issue arose because of suspicions that emissions of radiation from the nuclear plant at Sellafield are causally linked with the incidence of childhood leukaemia in the surrounding area. This attracted much media coverage in the 1980s and continues to be controversial. Sellafield became operational in the 1950s and since then has contributed electricity generated by nuclear reactors to the national grid. It soon acquired a reputation for the reprocessing of nuclear fuel and the management of waste products. By the 1990s, having become a world leader, Sellafield was taking deliveries of radioactive materials for reprocessing from as far afield as Japan. Such a major project, with its inherent dangers for people and the environment, could not fail to raise serious questions about the safety of the undertaking. In 1977, Mr Justice Parker chaired what Stott and Taylor (1980:vii) have called ‘the first major examination in public of the nuclear controversy’, which according to Breach became ‘the arena for fundamentally and intensely differing views of the way the modern industrial state should develop’ (Breach 1978:11). The public interest in this controversy grew so much that it was difficult to imagine that anyone, academic or otherwise, would not have a view on the matter.

On 1 November 1983, Yorkshire Television broadcast a ‘First Tuesday’ documentary entitled Windscale—The Nuclear Laundry. The programme was well publicised and over three million people watched it.⁶ The general thrust of the programme was to draw attention to two apparent facts, namely, that the nuclear reprocessing plant at Windscale, Sellafield, has on various occasions been responsible for discharging substantial quantities of radioactive waste into the environment and that the incidence of childhood leukaemia in the surrounding area is apparently unusually high.⁷ The viewer was steered towards the question, ‘Is Sellafield responsible for these ill-nesses and deaths?’ The impression conveyed was that the company which owned the plant, British Nuclear Fuels Ltd, was to some degree responsible.

The programme and the issue captured the attention of the press. Sufficient public interest and anxiety were generated that the British Government set up a public inquiry. A prominent scientist, Sir DouglasBlack, chaired the deliberations and his report was published in 1984 (Black 1984). The inquiry had specific terms of reference, namely:

To look at the recently published claims of an increased incidence of cancer in the vicinity of the Sellafield site:

1. examine the evidence concerning the alleged cluster of cancer cases in the village of Seascale,⁸
2. consider the need for further research,
3. and to make recommendations.

(Black 1984:7, para. 1.2)

For a public inquiry of the type being considered here, where should the burden of proof lie? The issues were complex. There was much at stake. Macgill points out that it is not possible to answer such a question without reference to the context of the particular inquiry under consideration:

Its outward appearance is that of an authoritative scientific study, investigating the truth of the claim of suggested excess of leukaemia and other cancers in the vicinity of Sellafield, and investigating possible explanations of the cause of this phenomenon. To consider and elevate Black purely as a scientific study, however, is to have too narrow an appreciation of its character.

(Macgill 1987:141)

Macgill goes on to suggest that there are two other aspects of the inquiry, which she calls ‘forensic’ and ‘policy’ aspects, materially affecting the way in which the issues are to be understood and the evidence is to be interpreted. But consider first the context of the inquiry as a purely scientific one.

There is a well tried and generally accepted method available to the scientist to test hypotheses about the relationship between two phenomena. The method requires the scientist to specify a null hypothesis that there is no such relationship and then seek to amass sufficiently convincing evidence that the null hypothesis can be rejected with a high degree of confidence. Only then would the reputable scientist publicly claim that the evidence supported the belief that a relationship exists. Note the parallel with the British legal system.

From the purely scientific perspective, the burden of proof rests with those who claim that the relationship exists. If this approach is applied to the Sellafield cancer-link controversy then the burden of proof rests entirely with those who claim that the link exists. In other words, BNFL is presumed innocent of the cancer-link charge until evidence establishes otherwise. At this point other considerations come into play.

The forensic aspect of the inquiry, mentioned by Macgill, arose because if it were shown that BNFL was in any way responsible for any of the leukaemia cases, then the company or its employees might be liable for damages. In fact, if there were any finding of criminal negligence, then the implications for those concerned could be quite severe. This made the controversy an attractive media topic. The ‘policy’ aspect of the inquiry can be taken to refer to ‘public health and safety’. But, from such a perspective, is it appropriate to assume that something is safe unless it can be shown, beyond reasonable doubt, to be dangerous? When a drug company seeks to market a new product it is required to undertake and document extensive testing to establish, with a high degree of confidence, that the new product is safe and that any side-effects are known. Should an unanticipated side-effect subsequently be identified then the drug company might expect costly litigation. Such was the case with the drug Thalidomide.

From the perspective of public health and safety, if society places the burden of proof with drug companies to demonstrate beyond reasonable doubt that any new product is safe, should it not place an equivalent burden on BNFL to demonstrate that its process for the treatment of nuclear waste does not have any harmful side-effects? Yet Black states that:

We have found no evidence of any general risk to health for children or adults living near Sellafield when compared to the rest of Cumbria, and we can give a qualified assurance to the people who are concerned about a possible health hazard in the neighbourhood of Sellafield.

(Black 1984: para. 6.13)

To ‘have found no evidence of a general risk to health’ is not to establish that such evidence actually does not exist. The possibility of crucial evidence being overlooked or undervalued could not be ruled out no matter how thorough the conduct of the inquiry.⁹

However, it is clear that the Committee themselves were aware of the public health and safety aspect of their inquiry. Since the Committee’stask was to make recommendations for public policy, a rather different and more balanced approach than that actually taken would seem appropriate. With such a ‘balanced’ treatment of argument and evidence there is no necessary presumption of a particular outcome. In these circumstances, the notion of ‘burden of proof is perhaps more appropriately replaced with a ‘balance of probabilities’, more akin to the procedures adopted in UK civil courts rather than those adopted in the criminal courts. This ‘policy’ aspect of the inquiry demonstrates that context has a bearing on the formulation of a hypothesis, on the kind of evidence required and on the ways in which evidence is to be treated and interpreted.

When context affects so much and especially when political considerations and vested interests are not entirely irrelevant, it is not surprising that the inquiry into the Sellafield cancer-link controversy should be chaired by an expert in argument and evidence. In matters of such import, great care is required to preserve neutrality. As authors, our prime concern is to present some ways in which the reader can realistically approach such controversies with some hope of arriving at a conclusion with confidence. Various aspects of the cancer-link debate are used throughout the book, where we subject both the arguments and the evidence to scrutiny.

TO WHOM THIS BOOK IS ADDRESSED

The authors’ intention is that this book should be useful for anyone who encounters serious attempts to evaluate arguments either intrinsically or by recourse to empirical evidence. Also, in conveying one’s own ideas, it is important to present the best case which is defensible in terms of argument and evidence.

In the normal course of events, most of us encounter serious attempts to evaluate lines of reasoning. The controversies mentioned above are typical examples. Media documentaries frequently evaluate arguments on specific issues by reference to what is intrinsically rational or w...

Table of contents

• COVER PAGE
• TITLE PAGE
• COPYRIGHT PAGE
• ILLUSTRATIONS
• PREFACE AND ACKNOWLEDGEMENTS
• 1: INTRODUCTION
• 2: ARGUMENT AND EVIDENCE
• 3: CONTEXT, CONVENTION AND COMMUNICATION
• 4: AN INFORMAL ANALYSIS OF ARGUMENTS
• 5: PATTERNS OF REASONING
• 6: ESTABLISHING VALIDITY
• 7: Critical Analysis In Practice
• 8: ASSUMPTIONS
• 9: EVIDENCE AS GROUND FOR BELIEF
• 10: WHAT COUNTS AS EVIDENCE?
• 11: PRESENTING AND SUMMARISING EVIDENCE
• 12: FURTHERING KNOWLEDGE
• 13: PROBABILITY AND UNCERTAINTY
• 14: PROBABILITY THEORY APPLIED
• 15: ESTIMATION AND RELIABILITY
• 16: TESTING HYPOTHESES
• APPENDIX I: THE SANCTITY OF LIFE
• APPENDIX II: VALID ARGUMENTS IN PREDICATE FORM
• APPENDIX III: STATISTICAL TABLES
• NOTES
• REFERENCES