Is human history explicable in terms of general principles or laws? This question has been debated extensively. Some scholars insist that history is based largely on a few major laws, playing out against a background of far less important noise. Others argue, instead, that history is so full of contingencies (or accidents) that general or universal laws are blown to bits. I am not competent to review this debate but find myself most sympathetic to the intermediate position taken by Hempel (1942) in his call for a nomological-deductive approach to history. Basically, Hempel argued that historical events can be explained only by reference to various general (deterministic or probabilistic) laws that causally link preceding events or conditions to the event being explained. For example, an account of why an automotive radiator cracked during a frost would involve both historical contingencies and general laws relating temperature to pressure
(Hempel, 1942). Similarly, events in human history can be explained by ''showing that the event in question was not 'a matter of chance', but was to be expected in view of certain antecedent or simultaneous conditions'' (Hempel, 1942) and the operation of several, often implicitly assumed, general laws. This nomological-deductive methodology waxes and wanes in popularity (Kincaid, 1996; Mclntyre, 1996), but it seems logical in principle. Naturally, one may debate whether human behavior is predictable enough to yield the kind of laws that are needed for nomological-deductive explanations (Beed and Beed, 2000).
Evolutionary biologists have likewise debated the role of general laws in explaining the past, which in their realm is phylogeny. Some have argued that natural selection is a universal law that can be used to explain the emergence of many, if not most, biological features. Others have countered that natural selection is a mathematical truth, rather than an empirically determined law (Sober, 2000). More importantly, many biologists have pointed out that the results of natural selection are not highly predictable. Gould (1989) made this argument when he declared that rewinding the tape of life on earth and playing it again would not lead to a repeat performance. Biological history is full of accidents, of happenstance. Therefore, Gould argued, evolutionary explanations must be crafted one event at a time, without recourse to general laws. On the other hand, Gould did grant that evolution is constrained by diverse physical principles, by rules of construction and good design, and by some scaling rules (Gould, 1986, 1989). In his view, ''the question of questions boils down to the placement of the boundary between predictability under invariant law and the multifarious possibilities of historical contingency'' (Gould, 1989, p. 290). Gould placed this boundary ''so high that almost every interesting event of life's history falls into the realm of contingency'' (Gould, 1989, p. 290). This appears to be an extreme position, for many other evolutionary biologists place that same boundary lower. They tend to be far more impressed than Gould by the degree of convergent evolution in the history of life (Carroll, 2001; Willmer, 2003). They look, for example, at the convergent similarities of eyes in vertebrates and octopi and conclude that some design rules for eyes exist. In sum, disagreements persist about the placement of Gould's boundary between predictability and contingency, but most biologists accept that evolutionary explanations must involve at least some causal laws (Bock, 1999).
Given this context, it is not surprising that neu-roscientists are conflicted about the importance of general laws for explaining the evolutionary history of brains. Marsh (1886) had proposed that brains consistently increase in size over evolutionary time, but later authors vehemently disagreed (see Jerison, 1973; Buchholtz and Seyfarth, 1999). Personally, I think that Marsh did have a point, for brain and body size have both increased, at least on average, in several vertebrate lineages (see Striedter, 2005). Still, Marsh's laws were merely descriptions of phylogenetic trends, not causal laws. The first explicitly causal law of brain evolution was Ariens Kappers' (1921) law of neurobiotaxis, which states that cell groups in evolution tend to move toward their principal inputs. Unfortunately for Ariens Kappers, later studies showed that cell groups do not move quite so predictably and called into question some of the mechanisms that supposedly produced neurobio-taxis. The next major putative law of brain evolution was Ebbesson's (1980) parcellation principle, which states that brains become more complex by the division of ancestrally uniform cell groups into daughter aggregates that selectively lose some of their ancestral connections. This principle was strenuously criticized by most comparative neuroanatomists, mainly because its empirical foundation was shaky (see Ebbesson, 1984). Although a weak version of Ebbesson's theory, stating merely that brains become less densely connected as they increase in size, is probably defensible (Deacon, 1990a; Striedter, 2005), the strong version of Ebbesson's original idea has failed the test of time: plenty of data now show that brains evolve not only by the loss of connections, but also by creating novel projections.
Confronted with this abundance of failed brain evolution laws, most evolutionary neuroscientists have emphasized only a single, undisputed regularity of brain evolution, namely, that numerous aspects of brain structure and function are highly conserved across species. Specifically, they focused, a la Geoffroy St. Hilaire, on the existence of common plans of construction and highlighted molecular homologies between invertebrates and vertebrates (see above). This has been productive. It is important to note, however, that the principle of phylogenetic conservation predicts stability and does not deal explicitly with change. Is brain phylo-geny subject to just a single law, which states that brains change little over time? Or are there also laws of evolutionary change in brains? I affirmed the second possibility (Striedter, 2005), but laws of evolutionary change in brains are no doubt difficult to find. C. J. Herrick, a founding father of evolutionary neuroscience, put it well:
Most scientific research has been directed to the discovery of the uniformities of nature and the codification of these in a system of generalizations. This must be done before the changes can be interpreted. The time has come to devote more attention to the processes and mechanisms of these changes... but it is much more difficult to find and describe the mechanisms of... [the] apparently miraculous production of novelties than it is to discover the mechanical principles of those repetitive processes that yield uniform products (Herrick, 1956, p. 43).
The last few years have seen an uptick in the number of studies that address evolutionary change and novelty in brains (Aboitiz, 1995; Catania et al., 1999; Rosa and Tweedale, 2005), and modern research on brain scaling and developmental constraints (see above) has advanced our understanding of the regularities that lurk within brain variability. In addition, a rapidly increasing number of studies is beginning to reveal genomic changes that are probably linked to changes in brain size and/or structure (e.g., Dorus et al., 2004; Mekel-Bobrov et al., 2005). Therefore, the time Herrick discussed, when evolutionary change becomes a focus of analysis (see also Gans, 1969), is probably at hand.
Thus, I envision a future in which most evolutionary neuroscientists will embrace many different laws, some dealing with constancy and some with change. A few philosophers of science (e.g., Beatty, 1995) might decry such a vision, because they think that any natural law deserving of its name must apply universally, in all contexts and without room for other, countervailing laws. I have no training in philosophy, but think that all scientific laws apply only in specified domains and given assumptions (Striedter, 2005). In the real world, particularly in the complex world of biological systems, most laws or principles are sometimes excepted. This does not make them useless but, instead, prompts us to ask what causes the observed exceptional cases (West and Brown, 2005). If we understand the causal basis of our laws, then the exceptions should, with further work, become explicable. In other words, I think that evolutionary neuroscientists can fruitfully avail themselves of Hempel's nomological-deductive approach to history. To some extent, they always have.
Was this article helpful?