In trying to analyze the natural world, scientists are seldom aware of the degree to which their ideas are influenced both by their way of perceiving the everyday world and by the constraints that our cognitive development puts on our formulations. At every moment of perception of the world around us, we isolate objects as discrete entities with clear boundaries while we relegate the rest to a background in which the objects exist.
That tendency, as Evelyn Fox Keller’s new book suggests, is one of the most powerful influences on our scientific understanding. As we change our intent, also we identify anew what is object and what is background. When I glance out the window as I write these lines I notice my neighbor’s car, its size, its shape, its color, and I note that it is parked in a snow bank. My interest then changes to the results of the recent storm and it is the snow that becomes my object of attention with the car relegated to the background of shapes embedded in the snow. What is an object as opposed to background is a mental construct and requires the identification of clear boundaries. As one of my children’s favorite songs reminded them:
You gotta have skin.
All you really need is skin.
Skin’s the thing that if you’ve got it outside,
It helps keep your insides in.
Organisms have skin, but their total environments do not. It is by no means clear how to delineate the effective environment of an organism.
One of the complications is that the effective environment is defined by the life activities of the organism itself. “Fish gotta swim and birds gotta fly,” as we are reminded by yet another popular lyric. Thus, as organisms evolve, their environments necessarily evolve with them. Although classic Darwinism is framed by referring to organisms adapting to environments, the actual process of evolution involves the creation of new “ecological niches” as new life forms come into existence. Part of the ecological niche of an earthworm is the tunnel excavated by the worm and part of the ecological niche of a tree is the assemblage of fungi associated with the tree’s root system that provide it with nutrients.
The vulgarization of Darwinism that sees the “struggle for existence” as nothing but the competition for some environmental resource in short supply ignores the large body of evidence about the actual complexity of the relationship between organisms and their resources. First, despite the standard models created by ecologists in which survivorship decreases with increasing population density, the survival of individuals in a population is often greatest not when their “competitors” are at their lowest density but at an intermediate one. That is because organisms are involved not only in the consumption of resources, but in their creation as well. For example, in fruit flies, which live on yeast, the worm-like immature stages of the fly tunnel into rotting fruit, creating more surface on which the yeast can grow, so that, up to a point, the more larvae, the greater the amount of food available. Fruit flies are not only consumers but also farmers.
Second, the presence in close proximity of individual organisms that are genetically different can increase the growth rate of a given type, presumably since they exude growth-promoting substances into the soil. If a rice plant of a particular type is planted so that it is surrounded by rice plants of a different type, it will give a higher yield than if surrounded by its own type. This phenomenon, known for more than a half-century, is the basis of a common practice of mixed-variety rice cultivation in China, and mixed-crop planting has become a method used by practitioners of organic agriculture.
Despite the evidence that organisms do not simply use resources present in the environment but, through their life activities, produce such resources and manufacture their environments, the distinction between organisms and their environments remains deeply embedded in our consciousness. Partly this is due to the inertia of educational institutions and materials. As a coauthor of a widely used college textbook of genetics,1 I have had to engage in a constant struggle with my coauthors over the course of thirty years in order to introduce students to the notion that the relative reproductive fitness of organisms with different genetic makeups may be sensitive to their frequency in the population.
But the problem is deeper than simply intellectual inertia. It goes back, ultimately, to the unconsidered differentiations we make—at every moment when we distinguish among objects—between those in the foreground of our consciousness and the background places in which the objects happen to be situated. Moreover, this distinction creates a hierarchy of objects. We are conscious not only of the skin that encloses and defines the object, but of bits and pieces of that object, each of which must have its own “skin.” That is the problem of anatomization. A car has a motor and brakes and a transmission and an outer body that, at appropriate moments, become separate objects of our consciousness, objects that at least some knowledgeable person recognizes as coherent entities.
It has been an agony of biology to find boundaries between parts of organisms that are appropriate for an understanding of particular questions. We murder to dissect. The realization of the complex functional interactions and feedbacks that occur between different metabolic pathways has been a slow and difficult process. We do not have simply an “endocrine system” and a “nervous system” and a “circulatory system,” but “neurosecretory” and “neurocirculatory” systems that become the objects of inquiry because of strong forces connecting them. We may indeed stir a flower without troubling a star, but we cannot stir up a hornet’s nest without troubling our hormones. One of the ironies of language is that we use the term “organic” to imply a complex functional feedback and interaction of parts characteristic of living “organisms.” But musical organs, from which the word was adopted, have none of the complex feedback interactions that organisms possess. Indeed the most complex musical organ has multiple keyboards, pedal arrays, and a huge array of stops precisely so that different notes with different timbres can be played simultaneously and independently.
Evelyn Fox Keller sees “The Mirage of a Space Between Nature and Nurture” as a consequence of our false division of the world into living objects without sufficient consideration of the external milieu in which they are embedded, since organisms help create effective environments through their own life activities. Fox Keller is one of the most sophisticated and intelligent analysts of the social and psychological forces that operate in intellectual life and, in particular, of the relation of gender in our society both to the creation and acceptance of scientific ideas. The central point of her analysis has been that gender itself (as opposed to sex) is socially constructed, and that construction has influenced the development of science:
If there is a single point on which all feminist scholarship…has converged, it is the importance of recognizing the social construction of gender…. All of my work on gender and science proceeds from this basic recognition. My endeavor has been to call attention to the ways in which the social construction of a binary opposition between “masculine” and “feminine” has influenced the social construction of science.2
Beginning with her consciousness of the role of gender in influencing the construction of scientific ideas, she has, over the last twenty-five years, considered how language, models, and metaphors have had a determinative role in the construction of scientific explanation in biology.
A major critical concern of Fox Keller’s present book is the widespread attempt to partition in some quantitative way the contribution made to human variation by differences in biological inheritance, that is, differences in genes, as opposed to differences in life experience. She wants to make clear a distinction between analyzing the relative strength of the causes of variation among individuals and groups, an analysis that is coherent in principle, and simply assigning the relative contributions of biological and environmental causes to the value of some character in an individual.
It is, for example, all very well to say that genetic variation is responsible for 76 percent of the observed variation in adult height among American women while the remaining 24 percent is a consequence of differences in nutrition. The implication is that if all variation in nutrition were abolished then 24 percent of the observed height variation among individuals in the population in the next generation would disappear. To say, however, that 76 percent of Evelyn Fox Keller’s height was caused by her genes and 24 percent by her nutrition does not make sense. The nonsensical implication of trying to partition the causes of her individual height would be that if she never ate anything she would still be three quarters as tall as she is.
In fact, Keller is too optimistic about the assignment of causes of variation even when considering variation in a population. As she herself notes parenthetically, the assignment of relative proportions of population variation to different causes in a population depends on there being no specific interaction between the causes. She gives as a simple example the sound of two different drummers playing at a distance from us. If each drummer plays each drum for us, we should be able to tell the effect of different drummers as opposed to differences between drums. But she admits that is only true if the drummers themselves do not change their ways of playing when they change drums.
Keller’s rather casual treatment of the interaction between causal factors in the case of the drummers, despite her very great sophistication in analyzing the meaning of variation, is a symptom of a fault that is deeply embedded in the analytic training and thinking of both natural and social scientists. If there are several variable factors influencing some phenomenon, how are we to assign the relative importance to each in determining total variation? Let us take an extreme example. Suppose that we plant seeds of each of two different varieties of corn in two different locations with the following results measured in bushels of corn produced (see Table 1).
There are differences between the varieties in their yield from location to location and there are differences between locations from variety to variety. So, both variety and location matter. But there is no average variation between locations when averaged over varieties or between varieties when averaged over locations. Just by knowing the variation in yield associated with location and variety separately does not tell us which factor is the more important source of variation; nor do the facts of location and variety exhaust the description of that variation.
There is a third source of variation called the “interaction,” the variation that cannot be accounted for simply by the separate average effects of location and variety. There is no difference that appears between the average of different varieties or average of different locations, suggesting that neither location or variety matters to yield. Yet the yields of corn were different when different particular combinations of variety and location are observed. These effects of particular combinations of factors, not accounted for by the average effects of each factor separately, are thrown into an unanalyzed category called “interaction” with no concrete physical model made explicit.
In real life there will be some difference between the varieties when averaged over locations and some variation between locations when averaged over varieties; but there will also be some interaction variation accounting for the failure of the separately identified main effects to add up to the total variation. In an extreme case, as for example our jungle drummers with a common consciousness of what drums should sound like, it may turn out to be all interaction.
The Mirage of a Space Between Nature and Nurture appears in an era when biological—and specifically, genetic—causation is taken as the preferred explanation for all human physical differences. Although the early and mid-twentieth century was a period of immense popularity of genetic explanations for class and race differences in mental ability and temperament, especially among social scientists, such theories have now virtually disappeared from public view, largely as a result of a considerable effort of biologists to explain the errors of those claims.
The genes for IQ have never been found. Ironically, at the same time that genetics has ceased to be a popular explanation for human intellectual and temperamental differences, genetic theories for the causation of virtually every physical disorder have become the mode. “DNA” has replaced “IQ” as the abbreviation of social import. The announcement in February 2001 that two groups of investigators had sequenced the entire human genome was taken as the beginning of a new era in medicine, an era in which all diseases would be treated and cured by the replacement of faulty DNA. William Haseltine, the chairman of the board of the private company Human Genome Sciences, which participated in the genome project, assured us that “death is a series of preventable diseases.” Immortality, it appeared, was around the corner. For nearly ten years announcements of yet more genetic differences between diseased and healthy individuals were a regular occurrence in the pages of The New York Times and in leading general scientific publications like Science and Nature.
Then, on April 15, 2009, there appeared in The New York Times an article by the influential science reporter and fan of DNA research Nicholas Wade, under the headline “Study of Genes and Diseases at an Impasse.” In the same week the journal Science reported that DNA studies of disease causation had a “relatively low impact.” Both of these articles were instigated by several articles in The New England Journal of Medicine, which had come to the conclusion that the search for genes underlying common causes of mortality had so far yielded virtually nothing useful. The failure to find such genes continues and it seems likely that the search for the genes causing most common diseases will go the way of the search for the genes for IQ.
A major problem in understanding what geneticists have found out about the relation between genes and manifest characteristics of organisms is an overly flexible use of language that creates ambiguities of meaning. In particular, their use of the terms “heritable” and “heritability” is so confusing that an attempt at its clarification occupies the last two chapters of The Mirage of a Space Between Nature and Nurture. When a biological characteristic is said to be “heritable,” it means that it is capable of being transmitted from parents to offspring, just as money may be inherited, although neither is inevitable. In contrast, “heritability” is a statistical concept, the proportion of variation of a characteristic in a population that is attributable to genetic variation among individuals. The implication of “heritability” is that some proportion of the next generation will possess it.
The move from “heritable” to “heritability” is a switch from a qualitative property at the level of an individual to a statistical characterization of a population. Of course, to have a nonzero heritability in a population, a trait must be heritable at the individual level. But it is important to note that even a trait that is perfectly heritable at the individual level might have essentially zero heritability at the population level. If I possess a unique genetic variant that enables me with no effort at all to perform a task that many other people have learned to do only after great effort, then that ability is heritable in me and may possibly be passed on to my children, but it may also be of zero heritability in the population.
One of the problems of exploring an intellectual discipline from the outside is that the importance of certain basic methodological considerations is not always apparent to the observer, considerations that mold the entire intellectual structure that characterizes the field. So, in her first chapter, “Nature and Nurture as Alternatives,” Fox Keller writes that “my concern is with the tendency to think of nature and nurture as separable and hence as comparable, as forces to which relative strength can be assigned.” That concern is entirely appropriate for an external critic, and especially one who, like Fox Keller, comes from theoretical physics rather than experimental biology. Experimental geneticists, however, find environmental effects a serious distraction from the study of genetic and molecular mechanisms that are at the center of their interest, so they do their best to work with cases in which environmental effects are at a minimum or in which those effects can be manipulated at will. If the machine model of organisms that underlies our entire approach to the study of biology is to work for us, we must restrict our objects of study to those in which we can observe and manipulate all the gears and levers.
For much of the history of experimental genetics the chief organism of study was the fruit fly, Drosophila melanogaster, in which very large numbers of different gene mutations with visible effects on the form and behavior of the flies had been discovered. The catalog of these mutations described, in addition to genetic information, a description of the way in which mutant flies differed from normal (“wild type”) and assigned each mutation a “Rank” between 1 and 4. Rank 1 mutations were the most reliable for genetic study because every individual with the mutant genetic type could be easily and reliably recognized by the observer, whereas some proportion of individuals carrying mutations of other ranks could be indistinguishable from normal, depending on the environmental conditions in which they developed. Geneticists, if they could, avoided depending on poorer-rank mutations for their experiments. Only about 20 percent of known mutations were of Rank 1.
With the recent shift from the study of classical genes in controlled breeding experiments to the sequencing of DNA as the standard method of genetic study, the situation has gotten much worse. On the one hand, about 99 percent of the DNA in a cell is of completely unknown functional significance and any two unrelated individuals will differ from each other at large numbers of DNA positions. On the other hand, the attempt to assign the causes of particular diseases and metabolic malfunctions in humans to specific mutations has been a failure, with the exception of a few classical cases like sickle-cell anemia. The study of genes for specific diseases has indeed been of limited value. The reason for that limited value is in the very nature of genetics as a way of studying organisms.
Genetics, from its very beginning, has been a “subtractive” science. That is, it is based on the analysis of the difference between natural or “wild-type” organisms and those with some genetic defect that may interfere in some observable way with regular function. But to carry out such comparison it is necessary that the organisms being studied are, to the extent possible, identical in all other respects, and that the comparison is carried out in an environment that does not, itself, generate atypical responses yet allows the possible effect of the genetic perturbation to be observed. We must face the possibility that such a subtractive approach will never be able to reveal the way in which nature and nurture interact in normal circumstances.
An alternative to the standard subtractive method of genetic perturbations would be a synthetic approach in which living systems would be constructed ab initio from their molecular elements. It is now clear that most of the DNA in an organism is not contained in genes in the usual sense. That is, 98–99 percent of the DNA is not a code for a sequence of amino acids that will be assembled into long chains that will fold up to become the proteins that are essential to the formation of organisms; yet that nongenic DNA is transmitted faithfully from generation to generation just like the genic DNA.
It appears that the sequence of this nongenic DNA, which used to be called “junk-DNA,” is concerned with regulating how often, when, and in which cells the DNA of genes is read in order to produce the long strings of amino acids that will be folded into proteins and which of the many alternative possible foldings will occur. As the understanding and possibility of control of the synthesis of the bits and pieces of living cells become more complete, the temptation to create living systems from elementary bits and pieces will become greater and greater. Molecular biologists, already intoxicated with their ability to manipulate life at its molecular roots, are driven by the ambition to create it. The enterprise of “Synthetic Biology” is already in existence.
In May 2010 the consortium originally created by J. Craig Venter to sequence the human genome gave birth to a new organization, Synthetic Genomics, which announced that it had created an organism by implanting a synthetic genome in a bacterial cell whose own original genome had been removed. The cell then proceeded to carry out the functions of a living organism, including reproduction. One may argue that the hardest work, putting together all the rest of the cell from bits and pieces, is still to be done before it can be said that life has been manufactured, but even Victor Frankenstein started with a dead body. We all know what the consequences of that may be.
1 Anthony J.F. Griffiths, Susan R. Wessler, Sean B. Carroll, and Richard C. Lewontin, Introduction to Genetic Analysis, ninth edition (W.H. Freeman, 2008). ↩
2 The Scientist, Vol. 5, No. 1 (January 7, 1991). ↩
What Genes Can’t Tell Us: An Exchange October 13, 2011