Extract

Introduction and historical background


   Human beings have been concerned about causality, i.e. what causes what in this world and why, for a long time. With regard to health, for thousands of years, men and women attributed events, such as disease, to a variety of physical and conceptual 'causes', including natural phenomena, e.g. an eclipse, evil spirits or the wrath of the Gods.


   The publication in 1638 of Galileo's book Dascori marks the beginning of the modern scientific era. In his book, Galileo introduced concepts such as description first, explanation second and that description could be carried out using the language of mathematics. Following Galileo, the most notable philosopher who wrote about causation was David Hume, the Scottish 18th century philosopher. In his major work A Treatise of Human Nature, Hume addressed a number of aspects related to causality, and many of the concepts that he proposed remain valid. One of his useful thoughts was that there is no such thing as an impression of a causal relationship. According to Hume, we can perceive by mere observation of a and b that a is above b or to the right of b. He held the view that when we say a causes b, we mean that a and b are constantly conjoined by fact, but not that there is some necessary connection between them. In Hume's view we have no other notion of cause and effect but that of certain objects, which have been always conjoined. Bertrand Russell adds that we cannot penetrate into the reason of the conjunction (Russell on Hume).


   Although practical biomedical sciences tend generally not to be concerned during day-to-day work about causation with such philosophical rigour, the principles of causation must be taken into consideration in all work that examines interactions between events with regard to a possible cause and effect relationship. Francis Goulton, the inventor of finger-printing and cousin of Charles Darwin, measured the length of a person's arm and the size of that person's head and asked to what degree can one of these qualities predict the other. Goulton's experimental work, while technically simple, documented for the first time in the history of science that the correlation between two biological variables can be connected on the possible basis of measurement rather than human judgements, i.e. that the attribute to the correlation to variation of the two organs is partly due to common causes. This led Pearson to state, 30 years later, that his interpretation of Goulton's work was that there was a category broader than causation, namely correlation, of which causation was only the limit. This new concept of correlation brought psychology, anthropology, medicine and sociology in large parts into the field of mathematical treatment. (Pearl, 2001)


   In general, the two fundamental questions about causality are:

1. What experimental evidence is required for legitimate inference of a cause and effect relationship?
   
   
   
2. Given that we are willing to accept causal information about the phenomena, what inferences can we draw from such information and how?