How many H2O molecules are wet?
Pseudo categorical mistakes and the scope of empirical research
If you ask an educated person, how many water molecules it takes such that they become wet the most likely answer you will get is this:
(a) H2O molecules are a conceptual construction within the framework of chemistry for the purpose of describing facts and processes about the transformation of matter. Chemical formulas enable us to calculate how much hydrogen and oxygen is needed to synthesize water, and vice versa, i.e., how much of the gases you get when you use energy to split the water molecule.
(b) Wetness is a subjectively experienced phenomenon, the quality of which is derived from immediate experience. Things are wet when we experience water on our skin and other forms of wetness, such as, for example, wet streets are derived from this experience. We would have no idea what wetness is if we did not have the subjective experience of the quality of wetness.
(c) From (a) and (b) it follows that the question we asked at the outset mixes up two different realms of being which should not be mixed up: when we talk about the number of molecules, we talk about conceptual entities within the framework of chemistry, whereas when we talk about wetness we talk about subjective experience.
(d) The answer to the above question therefore can only be that the question itself is badly posed and contains a categorical mistake. Just like the question, which number is sweeter, 3 or 5, the question of how many water molecules it takes to implement wetness, does not make any sense.
This argument sounds reasonable and thoughtful. However it is wrong. The answer to the question, in fact, is very simply 6, as Gregory and co-workers have found out by calculations and carefully constructed experiments (1).
This example clearly shows how easy it is to misjudge the scope of empirical research. It also shows that our thinking of problems as either conceptual or empirical is too simple minded and that in reality things are more complex. Our semantic memory contains conceptual and empirical knowledge and it is not clear in every case whether a certain question belongs to the empirical or conceptual realm. Hence one should be careful when talking about categorical mistakes. They may turn out to be pseudo categorical mistakes.
Of course, if somebody were to ask, whether there are married bachelors, whether there is a color of red with a slightly green tint, or whether there are personality disorders lasting for only two weeks, it is easy to see that these questions are not resolved by any additional fact but rather by clarifying the concepts of bachelors, colors, and personality disorders. However, what about questions such as whether there is successful sex between different species or whether there are correct delusions? The quick answers to these questions — a species is defined by successful sex between members of one species, from which it follows, that there is no successful sex between members of different species; delusions are defined as fixed false beliefs from which it follows that there are no correct delusions — are wrong. We know that sex between species happens in 10 to 20 % of all species and is much more common than it was thought. We also know that, for example, the delusion of jealousy may be actually true but nontheless is a delusion.
Why are these ideas particularly relevant to psychiatry? It may be argued, that every scientist should think of the methodological and conceptual underpinnings of his subject, and of science in general, every once in a while. Within a scientific specialty that changes rapidly as is the case in present day psychiatry, such reflection on methods and presuppositions is no luxury for Sundays and celebration speeches, but rather is urgently needed and happens to be a major and basic task of the involved and responsible professional. The reason is that major break-throughs within a scientific specialty always shake the bearings and foundations of that very specialty. Just think of the different answers that biologists would have given you during the last century to the question of what life really is. Because psychiatry is undergoing rapid change brought about by discoveries in the neurosciences, its conceptual framework is, in a way, constantly battered by this process.
There are examples of this from my own work: In 1987 I published a monograph on hallucinations summarizing the approximately 1500 papers on the subject that had appeared in the major journals over the past decades. I was convinced that even the most recent neurophysiological methods are of little use when it comes to the study of the purely mental phenomenon of an hallucination. I wrote back then that "not all questions in which it appears may be answered by experience, can in fact be answered by experience" (Spitzer, 1988) and I quoted examples of that. Thirteen years later the developments in neuroscience have taught me a different lesson. The question, for example, of whether hallucinations are similar to perceptions or dissimilar to perceptions, cannot only be answered by precise descriptions of the patients´ experience but also by functional neuroimaging. The same thing is true as regards the question of whether it is possible to realize for oneself whether one is hallucinating or not, or how we could se tup criteria for the identity of hallucinations (i.e., criteria for answering the question of how to reasonably talk about one, two or three hallucinations). Of course, there is no way in which a scanner can replace critical and analytical thinking on conceptual foundations. However, we should be careful with bold statements about the usefulness of empirical research in psychiatry. A case study done by my colleagues and myself clearly demonstrates this point: a female patient who told us, that she heard voices was stimulated by repetitive transcranial magnetic stimulation in cortical areas responsible for language. However, her hallucinations did not change at all. They occurred every four days and were accompanied by other symptoms with a general appearance similar to hysterical symptoms. In the end, it turned out, that the woman patient did not hear voices at all, but rather produced what people have called "hallucinatory behavior" and only said that she had heard voices.
By the very fact that modern functional brain imaging allows us to relate the workings of the brain to the functions of the mind we have, in addition to purely subjective experience, a second way to refer and relate to mental states. This objective reference to mental states allows us to answer questions empirically, which used to be thought of as being purely within the realm of conceptual analysis. Therefore, the framework in which questions, that used to be called "purely philosophical" ones can be discussed, has changed. We therefore have to ask many of the so-called "old and big questions" again. We have to get rid of old prejudices and dogmas, and start having a fresh look at them.
Original title: Spitzer M (2000) Wie viele H2O-Moleküle sind naß?: Über vermeintliche Kategorienfehler und die Reichweite empirischer Forschung. (Editorial). Nervenheilkunde 19(3):104-105