1. Description of the obstacle

Specialization is a most distinctive characteristic of modern science. The trend for the division of intellectual labor can be traced back to the Renaissance or even earlier, but a major impetus has been observed in the 20th century. In physiology, the trend to specialization surprised even the physiologists themselves (Burton 1975). In psychology, the changes were not less dramatic. If, for instance, we look at the first issue of the American Journal of Psychology (at the end of the 19th century), we learn that submissions were accepted in an amazingly wide range of subjects, including mental images, perceptual mechanisms, motivation, neurophysiology, learning, animal instincts, psychological development, psychopathology, hypnotism, etc. (Hall 1887). Inspection of the first issue of the British Journal of Psychology reveals a similarly wide range of interests, including analytical, comparative, genetic, experimental, pathological, individual, and ethnic approaches to psychology (Ward & Rivers 1904). In comparison, one or more journals can be found today for each one of those specialty areas. A few of the thousands of journals available today are: Animal Learning and Behavior, Behavioral Ecology and Sociobiology, Behavioral Neuroscience, Child Development, Cognitive Science, Developmental Psychobiology, Experimental Brain Research, Hormones and Behavior, Journal of Comparative Psychology, Journal of Experimental Social Psychology, Journal of the Experimental Analysis of Behavior, Journal of Feminist Family Therapy, Learning and Motivation, Perception and Psychophysics, Physiology and Behavior, Psychobiology, and Somatosensory and Motor Research.

If specialization is an important part of science, then specificity should be preferred to generality. Indeed, it is not difficult to see that generality (or "epistemological holism") may be an obstacle to the advancement of science. Even a reputable scientist such as Isaac Newton, for instance, made the absurd generalization that the transmission of nervous impulses could be explained by conduction through the same Ether that he had used to explain the action of one planet on another (Livingstone 1962). Generalizations of this type serve no purpose but to constitute epistemological obstacles.

Obviously, the situation is not simple. It is true that, on one hand, science praises specialization and abhors unjustified generalizations. On the other hand, however, generalization is the main goal of any inductive enterprise (which includes science). Indeed, it would be hard to believe that any investigator who uses five rats as experimental subjects is interested solely on those specific five rats. At the least, he/she would like to generalize the findings to all rats from the same colony. It is here that the dialectical nature of scientific investigation can be observed at its best. There are no a priori categories in the dialectical process. The rule is that specificity should be sought wherever it can be found. A concept that is too general today is an obstacle today. But a concept that is specific today may or may not become an obstacle tomorrow. Sometimes an illicit generalization may even result in the eventual creation of a new science (Refinetti 1987), which is obviously positive for the new science but still constitutes an obstacle for the original science. The bottom line is that the good scientist is eternally helpless: whatever he does, he knows he could always do it better. If he is too specific, he misses important generalizations; if he generalizes too fast or too much, he becomes an obstacle to the progress of his science.

2. Examples of the obstacle

An example of hasty generalization is the adoption, by most authors of physiology textbooks, of the neuroanatomical principle that specific ascending tracts carry sensory information from specific sensory modalities. Although there is some empirical basis for this principle (such as the relatively dense afference of pain sensitivity in the spinothalamic tract), the generalization was certainly premature (Norrsell 1980). The consequence of this hasty generalization was the acceptance by the scientific community of a wrong law. It is difficult to think of a stronger epistemological obstacle than a wrong law.

The field of biorhythmicity can provide us with another example of irresponsible generalization. For well over a century, experimental studies in mammals have demonstrated that several biological variables, such as general activity level and body temperature, exhibit a circadian variation (Davy 1845, Ogle 1866, Simpson & Galbraith 1905). Now, based on a single study with one cat and two dogs (and a few specimens of four other species), in which the cat and the dogs did not show any clear circadian rhythm, a group of researchers postulated a separation between animals that are mainly predatory and do not have circadian rhythms (such as cats and dogs), on one side, and animals that are not essentially predatory and have circadian rhythms (such as rats and monkeys), on the other side (Hawking et al 1971). The authors of a major textbook on biological rhythms endorsed the generalization without any further experimental evidence (Moore-Ede et al 1982 [p.235]). Naturally, if empirical research had confirmed this irresponsible guess, it would have been redeemed and I would probably be calling it "lucky intuition." In actuality, research showed that cats have clear circadian rhythms (Johnson & Randall 1985, Kuwabara et al 1986), and I can conclude that the generalization served only to mislead readers of the textbook.

In order not to make unnecessary enemies, let me stress the fact that all of us make generalizations all the time. Small generalizations, and generalizations that turn out to be correct, are usually safe. Obviously, you cannot tell whether the generalization will be correct before it is tested. I myself have suggested, based on data from a single animal, that moderate changes in ambient temperature can mask the circadian rhythm of activity of Armadillidium entrained to a light-dark cycle (Refinetti 1984). Fortunately, no rigorous study of this issue has been published that contradicts my suggestion. On the other hand, Fioretti et al (1974) suggested, based on an experiment conducted with inappropriate equipment, that the circadian rhythm of body temperature of the rat does not persist in the absence of environmental time cues. Unfortunately for them, many studies have since shown that the free-moving rat actually has an endogenously generated rhythm of body temperature (e.g., Eastman & Rechtschaffen 1983, Honma & Hiroshige 1978).

Another interesting example of generality as an obstacle has to do with skin receptors. In 1882, M. Blix noticed that thermal stimulation of some spots in the human skin evoked a cold sensation, whereas stimulation of other spots evoked a warm sensation (see Hensel 1973, Zotterman 1971). A few years later von Frey suggested, based on anatomic studies of the cornea and conjunctiva of the eye, that the cold and warm spots on the skin corresponded to specific anatomic structures, the end bulbs of Krause and the organs of Ruffini, respectively (von Frey 1895). This classification was adopted by most physiology books for over half a century, but empirical research showed that both cold and warm receptors are free nerve endings (Hensel 1974, Lynn 1975). Now, when an erroneous generalization becomes part of the textbooks that will educate the next generation of scientists, there is no doubt that we are dealing with an obstacle to the advancement of science.

Some generalizations are so tempting that people will make them even though the chances of error are evident. Two classes of such generalizations are those related to gender and species. Is it not tempting to assume that, if endotoxins and endogenous pyrogens produce fever in males (e.g., Ford & Klugman 1980, Morimoto et al 1986), they should also produce fever in females? Tempting it is; but it is also wrong, at least in part. Research shows that females develop much smaller fevers than males (Lipton & Ticknor 1979, Murakami & Ono 1987). As for species differences, is it not tempting to assume that, if brown adipose tissue is the major organ responsible for metabolic heat production in response to cold exposure in mammals (Bukowiecki et al 1982, Foster & Frydman 1979), it should also be the major organ in birds? In actuality, most birds do not even possess brown adipose tissue and rely mostly on striated muscle for cold-induced thermogenesis (Barre et al 1987, Saarela & Heldmaier 1987).

Naturally, not all hasty generalizations have already been refuted. For instance, Iversen (1979 [p.121]) says that synapses in the brain are excitatory or inhibitory depending on the type of neurotransmitter that is released presynaptically, whereas Kandel (1979 [p.62]) states that the functional expression of synaptic transmission (i.e., excitatory or inhibitory) is determined by the type of postsynaptic receptor. Because these are conflicting statements, they cannot both be correct. But, to the best of my knowledge, the neuroscience community is not ready yet to decide which of the statements is wrong. It is difficult to see why the authors would make such unnecessary generalizations. To be strictly fair, however, we should probably call the generalizations "heuristic conjectures" rather than "epistemological obstacles" until they are proved to be erroneous.

It is easy to realize that hasty generalizations may constitute epistemological obstacles because they lead to erroneous statements. Less obvious is the fact that generalizations may become obstacles because they provide a false sense of erudition and, consequently, lead to stagnation of research. For example, William James (1931 [v.1, p.8]) defined the concept of mentality in such a general way (i.e., the ability to seek alternative routes to reach a goal when the usual route is blocked) that it could apply to almost any living being and even to inanimate objects (e.g., a river will find an alternative route to the ocean if its bed is obstructed). Such an encompassing concept is so vague that it inhibits rather than stimulates research. Even worse is Wilhelm Reich's concept of the orgone. Starting from the Freudian conception of libidinal energy, Reich conceived the idea of a bioelectric energy (which would encompass the libidinal energy), and ended up with the delusion of a cosmic energy (the orgone) that is absorbed by living organisms through respiration (Reich 1961 [Chap. 9]). Such a concept could explain life, sexuality, mental illness, foreign policy, bad weather, and anything else one might think of. With a theory like this, who needs experimental science? In contrast, a modest generalization such as Hernandez-Peon's conception that the negative effect of lack of attention on perceptual acuity is due to inhibition of receptors and afferent fibers by brain stem mechanisms (Milner 1971 [Chap. 5]) gave rise to a productive line of research on efferent modulation of sensory afference in rat (Hellon & Necker 1976), cat (Kasprzack et al 1970), pigeon (Shortess & Klose 1977), chicken (Miles et al 1972), monkey (Martin et al 1979), and so on.