Entia non multiplicandum sunt praeter necessitatem (William of Occam)
For every complex problem there is an answer that is clear, simple and wrong (H.R. Mencken)
In this website I aim to provide argument and evidence to support my suggestion that the clearest understanding of the development of chronic disease in individuals and populations can be obtained by a multifactorial approach in which it is acknowledged that all the risk factors associated with disease and that most forms of disease can arise through several mechanistic pathways. I draw many of my examples from the field of smoking and health not only because of my familiarity with it but also because it includes a variety of human behaviours and a range of important chronic diseases. In doing so, I impose a difficult task upon myself. Smoking and health is a diverse field including the interpretation and understanding of epidemiological and biological data, together with some knowledge of tobacco and its production, the design and construction of cigarettes, the composition and toxicology of tobacco smoke and various aspects of human behaviour.
There are no experts in smoking and health. The field is too broad. If I include other fields relevant to my objective, the breadth increases. I certainly cannot claim technical expertise in all, or perhaps any, of its components. I have therefore inflicted upon myself the task of drawing and interpreting information from a range of fields of which I have varying levels of knowledge. There are two principal difficulties in this. Firstly, how do I know whether or not a given study was adequate technically. I am not competent to assess the relative merits of different types of gene array in studies of gene expression. Neither am I sufficiently familiar with the social and demographic composition of all the countries from which epidemiological studies have been published to determine whether or not the population studied was adequate to support the authors' interpretations. Guidance is available however. The peer review process which generally precedes the publication of a paper is supposed to ensure that work described sound in all verifiable respects. This is not infallible however. I have read papers which have passed through this process but which refer to experiments which were not physically possible with the apparatus described. Guidance may also be obtained from subsequent papers and correspondence referring to a publication of interest. Silence is also informative. If no further work on a subject appears or if it is not cited by others, it is possible that it was technically inadequate.
Secondly, how can I assemble information of many types from many sources to produce a coherent whole? Again some guidance is at hand. At the top of this page quotations from two philosophers, ancient and modern, provide conflicting advice on how to approach the interpretation of complex information. The words of the good friar, described by some as Occam's razor, can be translated broadly to suggest that, unless we have good reason to do so, we should prefer the simple explanation over the complex. Mencken, on the other hand and with the wisdom of the cynic, warns us to be suspicious of the simple explanation.
Both of these sages provide good advice, but how do we reconcile one with the other? The answer is, of course, that we cannot. Instead we must keep both in mind and balance one against the other as we reflect upon the schemes which we erect painfully to make some sense of large quantities of diverse and sometimes contradictory information. Is this interpretation too complex or too simple? The answer is seldom absolute, but Occam's razor may enable us to pare away unnecessary layers of elaboration while Mencken may cause us to re-examine an explanation which is just too neat and simple and ignores some embarrassing inconsistencies in the data. This is the application of judgement, a most unreliable tool. Just as few of us admit to being bad drivers, most of us feel we have exquisite judgment. Judgement, or rather the need for it, was inimical to the early philosophers who saw it as an extension of reason which, in turn, as the application of the individual mind, was the enemy of logic. Logic is the application of a set of universal rules of thought which, through their very universality, are proof against the individual's quirks and subjectivity. Where these rules came from, if not the human mind, is a question which is better returned to the philosophers.
Does logic help us in our quest of making sense of the mass of information on chronic disease which biologists and epidemiologists have provided? I once heard a senior medical official pronounce that as nearly all cases of lung cancer arise in smokers, it was logical to conclude that smoking caused lung cancer. He was wrong. Logic tells us that smoking is neither a necessary nor a sufficient cause for lung cancer. It is not a necessary cause as some cases arise in non-smokers. The observation that the majority of smokers do not contract lung cancer shows that smoking is not a sufficient cause of the disease. At this level, therefore, logic, as an assistant to our deliberations, abandons us. Nevertheless, I find some of its elements helpful. In a quest such as this, two of the most useful words in the English language, as are doubtless their counterparts in other tongues, are 'if' and 'then'. Faced with a new idea or a fresh combination of the facts, we can ask "if [this] is the case, then [something else] must follow". That is, every proposition has its consequences which can be used either to test its validity or to explore it further or both. We can even add that if a proposition does not lead to such enquiry, it is no advance upon its predecessors. This brings us to Karl Popper (Popper, 1982) who maintained that a hypothesis had value only if it were capable of falsification. In other words, an idea which cannot be tested is no idea at all. This provides essential guidance in a work such as this which explores a mass of complex data which could so easily conclude that the matter was too complex for resolution or generate nebulous suggestions remote from practical application. I am also mindful of Thomas Kuhn (Kuhn, 1970). In his 'Structure of scientific revolutions' he offended some scientists by suggesting that scientific theories do not arise as a result of the careful assembly and integration of facts. Rather, a single discovery may stimulate an idea. This stimulates further effort which may support and extend the idea. As research continues inconsistencies and anomalies are found which have to be accommodated within the theory and in some cases the theory eventually collapses under their weight. I thus pay heed to Popper and Kuhn and try to ensure that my theorising neither ventures into untestable territory nor struggles to accommodate too many awkward inconsistencies. My approach in this matter is therefore to question everything, both observations and conclusions, at every stage in the argument and to apply such interrogation not only to the material which I survey but to the tentative conclusions which I draw.
Like many others I was a child who, from time to time, presented my father with the collection of cogs, screws and wires which was all that remained of some device dismantled in an attempt to understand its workings. I had a capable and kindly father who knew more than I did, and who would restore the mess of parts to full and integrity and provide an instructive commentary to the process. The moral of this commonplace tale is that one learns something by taking things apart but even more by putting them together again. Analysis and synthesis are necessary companions. This website is largely an attempt at synthesis. I study others' analyses of risk factors and molecular pathways and try to construct a whole which accounts for most of what I have learned. There may be several ways of achieving this, but Occam's razor guides me to that which uses most of parts and discards the least. In doing so, I do not assume a superior position. I could do nothing without the painstaking work of the analysts who worry at problems until they have an indivisible unit of information. They too may fit this into a larger scheme and if this differs from mine it may be that their aims and directions differ too.
When I read a paper, I will skim the introduction to find where the authors are heading and then fall upon the tables and figures and try to determine what they mean and what story they bear. Only then will I read the rest of the paper and compare my thoughts with those of the author. Sometimes I agree and sometimes I do not and if the latter, it may well be because, again, our aims and directions differ. Sometimes I will find something in the data which is of great interest to me but which was incidental to the authors' main conclusions. Sometimes I cannot reconcile the authors' conclusions with my own. Usually, but not always, this is because I have misunderstood something. I pay little heed to authors' names and affiliations. I read dutifully and carefully the latest offering from someone I would cross the street to avoid and sometimes I am pleasantly surprised.
This synthesis is therefore based more upon other peoples' results than on their conclusions. This is not arrogance but simply the application of my context to the work of others. This is almost essential when revisiting the older literature. Much of this is relevant today, but often in a context quite different from that of the time. I should add at this point that in the case of smoking and health, reference to older work is essential. From the epidemiological perspective, most of the work upon which the whole edifice of conventional wisdom is founded was conducted before 1970. To discard this would be to discard the essential framework of smoking and health. It is nevertheless necessary to revisit and examine what it actually found and to determine whether it is really capable of supporting the elaborations which have been built upon it subsequently.
Finally I must deal with the philosophical difficulties associated with objectivity, belief and truth. Objectivity requires that our approach to a subject gives proper weight to all the information available and is not coloured by extraneous factors particular to ourselves. This, even with the best will in the world, is difficult to achieve. Most of us have some opinions about the world about us and what is right and wrong or just and unjust and it is both tempting and easy to colour our scientific opinions accordingly. Such bias may be overt or subconscious. I attempt to deal with this through awareness of the possibility and by testing the basis of my opinion against alternative possibilities which may oppose it. This is certainly not an infallible safeguard against partiality, but I do try.
Perhaps more insidious, because it can be less obvious or seemingly more justifiable, is to favour one viewpoint over another because of belief or conviction. Philosophically, this carries us across the boundary between the scientific and the metaphysical. We like to think that the scientific method involves the careful assessment and balancing of observations which are available to and testable by everyone. Belief, on the other hand, comes from beyond this realm. It is something which descends upon us from outside which we can accept or reject but cannot analyse. Let me illustrate this with a quotation from the literature which I provide with my assurance that I make no imputations about the character or integrity of the lady in question.
Most investigators who have worked with tobacco-specific nitrosamines (TSNA) are aware that these compounds are potent carcinogens that occur at high levels in various tobacco products. Therefore, most are convinced that TSNA cause at least a sizeable proportion of tobacco-related cancers in man. There are, however, a number of problems with trying to prove this. (Preston-Martin, 1991)
This can be taken in at least two ways. Firstly, and this is probably the case, the author was using rather unfortunate language to express objectives for research. There is, she is saying, sufficient evidence to support a view that tobacco specific nitrosamines may be involved in the development of some of the cancers to which smokers seem to be more prone than non-smokers to support further research in this area. Alternatively, the statement could be interpreted as saying that, despite a lack of supporting evidence, a number of people sustain the irrational belief that tobacco specific nitrosamines contribute to the toxicological properties of tobacco smoke. The author has made a subtle transition from the logical to the metaphysical. In choosing between these alternatives, one's objectivity is strained by the authors use of unqualified terms like 'potent carcinogen', 'high levels' and 'sizeable proportion' all of which could be questioned in the absence of the bases for comparison on which they were made.
This is why I tend to set aside authors' opinions and examine such supporting evidence as is presented.
Finally, and fundamentally, we have to accept that, as far as the biological sciences are concerned, the meaningful use of such standards of judgement as 'proof' and 'truth' is inaccessible to us. In some disciplines, one can establish whether a proposition is true or false. The statement a = b is true where a = 1 and b = 1 and false where a = 1 and b = 2. In any given case, the truth of the statement can be proved or disproved. The statement (tobacco specific nitrosamines) = (cancer) is not amenable to such a test. Part of the problem is that the proposition is too broad and any test requires more precise statement of dose, species and site of cancer. But even in the best defined case, such as a carefully controlled test in animals, the answer will not be true or false. Some stubborn brute in any adequate test will not develop cancer when the others do, or vice versa, so that the best answer we can get is not true or false, but usually or seldom. Put in other terms, biological observations are not discrete or absolute but are measures of probability. We cannot, therefore, establish truth or proof but must settle for a balance of probabilities and any decision on this balance depends upon whose weighing machine is used. Regrettably, it is all a matter of opinion.
Like everyone else, therefore, I cannot in this website claim to have proved my case but have offer an opinion as to the best interpretation of the information which I have considered. It may be that I have omitted important observations from my consideration. If this is so, it is not deliberate and my excuse is that the total available is too large to take every item into account.
My aim is not to convert others to my point of view but to provide a basis for them to reach an opinion of their own.