an international and interdisciplinary journal of postmodern cultural sound, text and image
Volume 7, March-April 2010, ISSN 1552-5112
Towards a Philosophical Account of Explanation in Behavioral Genetics
On one level of generality the topic of behavioral genetics seems relatively uncontroversial; but at another level behavioral genetics is much more contentious. If we ask whether genes are central to the development of basic capacities for behavior, then it will widely be agreed that of course genes are very important; change or delete certain key genes early on and certain behavioral capacities simply will not develop. But, if we ask about the role of specific genes in determining the appearance of specific behavioral traits, then things are much less clear. In spite of significant media hype over the past several decades regarding the alleged discovery of particular genetic structures for very particular behaviors, to judge by the meager list on certain government run web-sites, we are not in a position to have much confidence in progress towards accumulating significant information about simple direct connections between the presence of specific genes and the appearance of specific behaviors. This, in turn, has led to significant pessimism about the whole behavioral genetics project. It has prompted others to suggest that a more nuanced approach to behavioral genetics is needed, and to propose certain revisions. I fall into the latter category. If the behavioral genetics project is to continue to be viewed as a viable scientific project, then it is in need of clarification and modification as well as support. In this presentation I will offer a philosophical case for a revised behavioral genetics, building on important work of others.
For an exploration of a revised version of behavioral genetics to be worthwhile, we must first make sure that the basic elements of such a view do not face serious problems. But, the very concepts that constitute any form of behavioral genetics, namely those of genes and behavior, are far from enjoying unproblematic status. The continued use of a gene-centered approach to scientific investigation has been severely critiqued; while the gene-centered approach has been ably defended methodologically, there are other important arguments that demand the attention of defenders of behavioral genetics.
In this paper I begin by examining a particularly disturbing eliminativist argument from Evelyn Fox Keller against the continued use of the very concept of the gene. If Fox Keller’s argument were to work, then any attempt to continue with or attempt to revise behavioral genetics would be doomed. In the course of replying to Fox Keller’s argument a revised, functional concept of the gene is presented and defending. Using this revised conception of the gene I then consider how appeal to a functional approach to the gene can itself lead to a more general functionalist revision of the basic behavioral genetics project. In the third part of the paper I then turn to examining the advantages with respect to scientific explanation that such a functionalist account can provide. And, I end by considering how such an account might provide some help in dealing with additional ethical worries, including additional ethical arguments from Fox Keller against the continued use of the concept of the gene as well as ethical concerns that have been raised regarding behaviorally designed babies.
II. Keller’s Herculean Burden Argument
Evelyn Fox Keller comments in her 2000 book, The Century of the Gene as follows:
…Johannsen’s little word [gene], so innocently conceived in the early days of this century, had had to bear a load that was veritably Herculean. One single entity was taken to be the guarantor of intergenerational stability, the factor responsible for individual traits, and, at the same time, the agent directing the organism’s development. Indeed, one might say that no load seemed too great—as long, that is, as the gene was seen as a quasi-mythical entity. But by the middle part of the century, the gene had come to be recognized as a real physical molecule…just a bit of DNA …and here…the history of genetics takes its most surprising turn. Both the excitement and the triumph of the new science of molecular biology came from the remarkable evidence it provided suggesting that, incredibly enough, the gene, now understood as a self-replicating molecule of DNA, was a structure equal to its task. Yet, with the maturation of molecular biology, the impracticality…that load has become steadily easier to discern.
New kinds of data gathered over the last few decades have dramatically fleshed out our understanding of the parts played by genes in cellular and organismic processes, and in doing so they have made it increasingly apparent how far the weight of such a load exceeds what any one single entity can reasonably be expected to bear, and hence, how appropriate that it be distributed among many different players in the game of life. 
From this extended comment we can derive an interesting eliminativist argument that I will refer to as Fox Keller’s Herculean Burden Argument:
1. Genes are required to [a] be the sole guarantor of intergenerational stability, [b] the sole factor responsible for all individual traits, and [c] the agent that directs the organisms development.
2. Genes are identical with sequences of DNA.
3. Research results establish that mere sequences of DNA cannot by themselves perform the three required tasks of maintaining intergenerational stability, trait fixity and organismic developmental.
4. Thus, mere DNA sequences cannot perform the tasks required of DNA.
5. Thus, genes cannot perform the tasks required to be genes.
6. Thus, talk of genes should be eliminated in favor of different concepts.
I will assume that there are some here who will find Fox Keller’s argument controversial. My suggestion is that friends of genes, when faced with Keller’s claims, should not only reject the second premise, arguing that genes were something more than mere DNA sequences; they should also reject the first premise as being too strong a requirement for genes. While some purists may be horrified, I will argue that as long as the elements discussed are significant or major determining elements in development, composed of DNA and combined with perhaps something else, relatively stable and significantly involved in trait fixity, then, they could still count as genes. The specific objections I propose to Fox Keller’s argument are the following:
Objections to Fox Keller’s Herculean Burden Argument
1. First objection: Genes need not be the sole determining elements in development
2. Second objection: Genes need not be solely responsible for trait fixity
3. Third objection: Genes need not be completely stable
4. Fourth objection: Genes could be something over and above [or other than mere] DNA
Before specifically considering these objections I would like to motivate my response to Fox Keller’s argument by considering an analogy between genes and another important concept from the history of science, namely, the concept of ‘atom’. When we compare what we currently say about atoms with what was once said about atoms we find significant change. Originally conceived as smallest unit of matter, of radically different kinds, as not transformable into another kind … all of these claims about atoms have gone away over time. Still, we continue to talk about atoms. Atoms are a terribly small unit of matter, and the smallest of an important way of distinguishing an important set of kinds of matter from one another, namely elements, they represent kinds that have radically different properties, even if they are composed of very similar parts, and they are transformable into another kind, but the method of transformation is theoretically well understood, and poses no serious conceptual problem for the current conception of atoms and their parts. Thus, from the case of atoms we see that requirements originally associated with a scientific term can be revised as greater theoretical understanding is gained, as major changes in theory occur and are incorporated, and that the term can remain, and can continue to play a useful if somewhat changed--but still importantly related--role. This is the alternative message we should take from Fox Keller’s discussion concerning the concept of genes. Why so? Let us now consider each of her specific complaints in turn.
First, Fox Keller’s developmental requirement seems too strong. While genes are typically thought of as important in determining how an organism develops, one need not assign all developmental responsibility for development to genes. Genes are now thought of as working in normal biological contexts; these contexts are now recognized as providing much that is needed to enable the organism to develop. Further, one might decide to think of genes as being more passive than in the common lay interpretation. But, partially passive genes would still seem to count as genes. One need not think of genes as little omni-causal agents telling other parts of the cell what to do, where to go, what to make, etc.
While casual discussion of genes does often refer to causal activities that genes initiate, as in, “his alcoholism gene switched on and caused him to start hitting the bottle pretty hard”, it is not clear at all that gene talk must take causal agency as an absolutely essential, necessary condition of gene-hood. Whether genes have agency or not, their being background conditions that help serve to lead to particular courses of development, suffices for their having causal influence. Further, there is no clear contradiction in claiming that certain genes do not causally initiate any activities, but take part in causal activities initiated elsewhere. What is crucial is that genes continue to figure causally in the biological explanation of inheritance. To construct a serious objection to genes based on causality, one would need evidence that putative “genes” play no discernable causal role in adequate discussions of inheritance, a more difficulty task.
With respect to stability of gene structure, Fox Keller makes the following claim:
The stability of gene structure thus appears not as a starting point but as an end-product-as the result of a highly orchestrated dynamic process requiring the participation of a large number of enzymes organized into complex metabolic networks that regulate and ensure both the stability of the DNA molecule and its fidelity in replication….Moreover, not only are the mechanisms controlling stability and mutability held in a delicate balance, but that very balance is under cellular regular, and it shifts in response to the particular environment in which the cell finds itself. All this is a far cry from the traditional view of DNA as an inherently stable molecule subject to occasional random errors, and it suggests an even further departure from the traditional view of evolution as a process of cumulative selection of those exceedingly rare mutations that happen to result in increased fitness…. [31, 34-35]
Fox Keller makes the point that the stability of DNA, which had previously been taken for granted by most, is not something DNA is responsible for but is something it owes to cell activities. DNA cannot by itself provide the stability required for genetic transmission. Since genes are sequences of DNA, genes cannot by themselves provide the stability required for genetic transmission. But, we can grant this point and still continue to talk sensibly of genes. If we can accept passive genes, it should not be too far fetched to think of genetic stability as a product of other forces, not something that genes are able to achieve for themselves on their own.
The important point concerning stability is not really who or what is important for genetic stability, but what genetic stability is for. There are good reasons for maintaining genetic stability that concern with how species are able to survive in particular niches for which they have evolved specific adaptations. How the stability of factors responsible for these adaptations is maintained does not seem to be of importance in determining whether these factors are genes.
The most serious worry Fox Keller presents concerns genetic fixity: how to determine what a gene actually is. Problems arose when it was discovered that gene splicing could produce a variety of products. Why does this worry demand our attention? When the structure of DNA was discovered and the putative identification of genes with sequences of DNA advanced, the alleged payoff was that structural identity conditions of genes could be given. As philosophers know, one way of securing an item’s existence is to provide clear identity conditions; structural chemical identity conditions seemed an excellent way to solidify this worry about gene identity. But, as Fox Keller’s comments indicate, mere structure is not necessarily enough in all cases. Since the same structure can apparently sometimes give rise to different elements which produce a variety of different proteins, it is no longer clear how to account for the identity conditions of genes. Without identity conditions Quineans among us urge that no entity will be found, either.
Fox Keller rejects the attempt to try to found identity conditions for genes functionally:
Fifteen years ago Richard Burian observed…’There is a fact of the matter about the structure of DNA, but there is no single fact of the matter about what the gene is.’ In the interim, things have only gotten worse….The complications brought by the new data are vast…Taken together, they threaten to throw the very concept of “the gene”—either as a unit of structure or as a unit of function—into blatant disarray.
Techniques and data from sequence analysis have lead to the identification not only of split genes but also of repeated genes, overlapping genes, cryptic DNA, antisense transcription, nested genes, and multiple promoters (allowing transcription to be initiated at alternative sites and according to variable criteria). All of these variations immeasurably confound the task of defining the gene as a structural unit.
Similarly, discovery of the extensive editorial process to which the primary transcript is subject, of regulatory mechanisms operating on the level of protein synthesis, and others operating even on the level of protein function confound our efforts to give a clear-cut functional definition of the gene.” [66-67]
Since neither functional nor structural criteria will work, and since these two criteria exhaust the sources for identity conditions, Fox Keller concludes that the gene is a concept in trouble.
While we should admit that this is an interesting worry, rather than wring our hands and hold out for future discovery of a hidden novel purely structural feature to solve the identity conditions problem, there is a more promising present strategy. While waiting we can suggest, instead, that a combination of structural and functional elements might work. That is, suppose we propose to identify a gene as a particular DNA sequence of DNA typically assembled in a particular way in certain typical contexts and typically produces specific protein/s.
That is, let us seriously consider a functional role identity for the gene:
The functional role approach to gene identity:
Identify a gene as a specific DNA sequence, S, that plays a typical functional role, R, of typically producing protein P in context C in an organism O.
This approach has several virtues. First, it can accommodate the aberrant findings to which Fox Keller refers above. Second, it is consistent with what we should have been expecting regarding the notion of the gene. After all, the gene was originally conceived teleologically, as an item with certain functions. While some may have assumed that the teleological stage was eliminable upon the discovery of DNA, as Fox Keller has admirably demonstrated, this assumption has been short lived. If we re-introduce teleology into our account of the gene, we are really just returning to our honest biological roots, and are no longer pretending that chemistry all by itself can completely explain living phenomena. In reply, one may, of course, pose skeptical worries about how likely we are to find functional roles for DNA sequences, fears of multiple sequences with the same role, qualms about cases lacking characteristic functional roles, etc. But, none of this shows that no such functional roles are to be found.
The sketch of what constitutes a gene will only work if characteristic activities and characteristic outputs can be shown to account for regularities of inheritance that must be explained. If these regularities can not be tied to similar internal regularities of the sort mentioned above, where there is sufficient differentiation between genes in terms of constituents, relationships and outputs, there would no longer be any need for talk of individual genes but of the general, undifferentiated mechanism of inheritance. If any blob of DNA whatsoever could produce any particular feature you choose, then there would be no need to postulate specific individual elements of difference-makers and similarity-makers, which is an important part of what genes have always done.
III. The Causal Role Account of the Path from Gene to Behavior
In the previous section I proposed a rough functional role notion of identity for the gene as a way to avoid certain conceptual worries. While this proposal does not make explicit mention of behavior, once we have agreed to consider implementing the functional role approach with respect to gene identity, we are then in an interesting position with respect to asking about the biological explanation of behavior. For the functional role approach to genes naturally invites consideration of relevance of the nature of functional roles that we have inherited from work in the last century in the philosophy of mind on the mind-body problem. But this raises a number of questions: How is such an approach to be construed? What elements should it contain? How are the elements to be related? What can we learn from this approach?
Let us briefly recall some basic Post-Identity-Theory philosophy of mind. Mind-Body Functionalism, the currently reigning orthodoxy on the nature of the relation of the mind to the body for at least the past two decades, is the view that mental states are to be defined, not in terms of specific physical substances or specific physical properties, but in terms of the causal role played by an item, such as a neural state, in terms of responding to specific typical inputs and in terms of generating specific typical outputs, and, importantly, also in terms of relations to other such states. Specific cases of mental states are items, of whatever sort, that instance the right sort of causal role. The earliest versions of the view were concerned to account for the possibility of artificial intelligence, and hence made reference to machine states as definitive. But, while finding machine states for machines was obvious, it has always been difficult to imagine clearly how a machine state characterization of human mental states would suffice to distinguish mental states from other, non-mental activities. Some early causal role theorists seemed to think that reflective common sense was sufficient to determine the causal roles of such states. Other theorists held out for the contrasting view that it will be future science that will reveal to us the specific elements that make up the causal roles that characterize specific mental states. While this latter approach might strike some of the scientific pluralists here present as being rooted too firmly in scientific monist commitments, I think one can interpret scientific functionalism so as not to require commitment either to monism or pluralism in science. This said, it should also be noted that some functionalists explicitly utilize specifically biological notions, such as function or teleology, to provide guidance for determining the nature of specific causal roles. And, I would suggest, given that specifically teleological biological concepts, such as hearts, livers, etc, are defined specifically in terms of function and not structure, this last approach is most appropriate for understanding a biological concept, such as the gene, which was hypothesized initially as a functional item before any inkling of its structure was known. The key notion that all of these versions of functionalism have in common is a rejection of a simplistic type-type identity approach to the relation between mental and physical state and its replacement with a functional-role/realization model of this relation.
While the notion of function is not sufficient to show how a revised behavioral genetics should go and how one can derive explanatory power from it once so conceived, as it is not plausible to attribute enough individual functions to all genes, it is a good start. We attribute heart status to diverse items, earthworm hearts, Jarvik-7’s, etc. It is not the material used to construct the item or the exact design, but what the item could be reasonably said to do in the right circumstances, the rest of the body behaving appropriately, the environment cooperating, etc. This broader characterization helps us see how appropriately to relate genes to behavior.
It should be noted in passing is that an appeal to functional characterization should also make it easy for us to see the connection between traditional Mendelian genes and molecular genes. Mendelian genes are defined in terms of the typical phenotypic results whereas molecular genes are defined in terms of their chemical structures and their biochemical products. Given the functional characterization of genes we see that phenotypic results are simply a later stage in the causal role that includes, as an earlier stage, the chemical production of appropriate proteins. Thus, the molecular genes may be seen as more specific “realizations” of items playing general causal roles that also include their Mendelian counterparts. Given my commitment above to genes as functionally defined in a broad enough sense to include both the Mendelian and molecular varieties, there are additional consequences that also follow when we attempt to impose a functionalist account on genetic activity.
First, functional causal role accounts can include typical causes as well as typical effects. When it comes to genes, there seem to be relatively limited or standard typical causes, including the basic inheritance facts of meiosis which, in turn, requires a sufficient number of parents behaving appropriately, and, more unusually, environmental mutational forces. In addition, as Fox Keller notes above, there are important cellular sustaining forces that come into play.
Second, functional role accounts typically recognize the need to accommodate interconnections between functionally characterized items. Standard mind-body functionalists recognize important interconnections in their causal roles between mental states such as beliefs and desires; it is certainly also the case that we need to account for the polygenic contributing causes of complex behavioral in addition to most other complex traits. Thus, the same gene could operate with a potentially wide variety of other sets of genes to produce an equal variety of different items. This, in turn, would demonstrate that one also needs to think of diverse functional roles for diverse sets of genes of genes!
Third, functional accounts importantly emphasize typical effects. And, as we note that other items defined in terms of causal roles, such as hearts and livers, it is the typical effects that are the most prominent. It is with trying to enumerate typical effects involving genes, however, that we find a more elaborate series of items that will help carry us toward behaviors.
The first effect, with the discovery of the molecular phase of gene activity we are now aware of the production of proteins as being a crucial activity for genes to perform. It is via protein production that we can hope to fill in the otherwise strange and apparently magical gap from postulated special item in the body to phenotypic element observed in the progeny. Protein production is relatively unimportant unless it can be shown to lead, under normal circumstances, to typical effects in the body. But, a second effect factor, the body itself, must be characterized in such a way as contributing significantly to the particular effects in question. In addition, as a third factor, it is widely agreed to be equally of importance that the organism’s environment provide significant items for the production of states of the body that would make it possible for behaviors to be manifested when appropriately stimulated. A fourth, and for the purposes of this session, the importantly relevant effect, is the manifestations of typical behaviors, which again depend on either internal and external stimuli or both. It is with the addition of this last factor, combined importantly with the other three factors that we can see what an improved and revised approach to behavioral genetics would look like, though I shall consider below whether it is still appropriate to treat the view that I shall propose as a version of behavioral genetics.
A rough picture of this causal role involving genes as well as other factors may be represented as follows:
Rough Schematic Model of a Causal Role account of Genes leading to Behavior
ES…. EF ES
↓ ↓ ↓
RC → G → MP → IS → DB → B →….
↑ ↑ ↑ ↑
IF ISC…. IF IS
ES = environmental support DB = Dispositions to behave
RC = reproductive causes IF = internal forces
ISC = sustaining causes EF = external environmental forces
G = genes ES = external stimuli
MP = molecular products IS = internal stimuli
IS = internal states MB = manifestations of behavior
[…. = continued!]
According to this schema, environmental support and internal forces lead to reproductive causes; these reproductive causes lead to genes, which depend (for stability, etc.) on sustaining causes and continued environmental support; genes then lead, with continued support, to molecular products; these molecular products, again with continued support, lead to internal states; these internal states lead, in turn with continued support, to dispositions to behave; these dispositions to behave are triggered by a combination of internal and external stimuli, resulting in behavior.
As my perspective is that of a speculative philosopher and not a bench biologist, I would be the first to grant that the above model is very rough indeed. There are, undoubtedly, missing elements, such as additional feed back loops, that either I have forgotten, of which I personally am ignorant, or perhaps remain to be discovered. But, in one sense concern about exact detail here is unimportant. What I want to urge is that there is a model to be found that will do what I am proposing can be done here, namely to provide a quasi-functionalist, causal role linkage, however complicated, between genes and behavior that reserves and important role for genes, such that were the genetic input significantly to change, there would also be a significant change in behavior.
It needs to be emphasized that I am proposing a quasi-functionalist approach which emphasizes the complex causal role from genes to behavior. I am not proposing the apparently obviously unreasonable view (based on our present state of knowledge) that there are distinct and specific biological functions corresponding to each instance of the above schemata. And, I am not proposing that all stages of the causal role sketched in the above rubric must, themselves be subject to a functionalist definition, as philosophers of mind who embrace functionalism urge for all mental concepts. While some are helpfully construed functionally, such as genes, certain mental states tied to certain behaviors, maybe even the concept of “the body”, the same does not seem obviously to be true of certain environmental aspects of the rubric, for example.
We need to emphasize that the rough schematic quasi-functionalist approach, as borrowed and modified from the philosophy of mind, stands in contrast with a standard functional analysis found typically in biology. Functional analysis in biology ends with identifying particular functions for biological items in specific contexts with respect to particular goals. In the absence of a relevant biological goal or end there cannot be a function. The case with genes, however, seems importantly different. While it is reasonable to assign functions to certain activities of certain genes in terms of bringing about certain states which would be needed to accomplish certain biological goals, it seems most unlikely that every single human gene and every active combination thereof equally has a function in this sense.
Thus, one can speak of the “functional role” of a set of genes or of a “functionalist account” of some specific genetic activity without thereby being committed to finding a corresponding distinct biological function which that activity carries out. For example, it seems likely that there is a set of genes which do have the function of enabling speech, as speech is clearly an important element for normal human functioning, and, given our evolutionary history, it has been important for humans to have it. But, although there also seems undoubtedly to be a genetic contribution, it also seems, so far as we now know, unlikely that the correspondingly involved set of genes would have the function of creating a voice of a specific quality, such as a first tenor voice. There need not be a specific function for every distinct functional role.
This quasi-functionalist proposal is offered as an account for showing how to amplify and incorporate current practices in behavioral genetics within a larger philosophical framework. Let us briefly note how this proposal differs from the four main competitors. The Quasi-Functionalist approach differs from Classical Behavioral Genetics, not in ignoring it, but in terms of attempting to incorporate it. The same holds true for Modern Molecular Genetics. Again, the same holds for the Environmental Approach; it occupies a central role. The same can be said for the DST approach, emphasizing, however, the need for a more prominent role for genes.
A further question: What is intended by the use of the arrow, “→” in the above schema? Nothing as strong as implication or entailment, or nomic connection can be meant here. The relation needs to be something much looser, something like “typically/ideally leads to”, where this is to be understood as involving a causal contribution. The notion of a causal role seems to be just this loose. There are many different things that can break the chain of events in a causal role. Yet, causal roles, including those which involve genes, still seem to be that out of which what is distinctly biological is created and what ultimately needs to be studied if one is concerned to study biological phenomena biologically.
A final objection will end this section. Why should one consider the present quasi-functionalist proposal as an example of Behavioral Genetics when other significant elements are involved besides genetics and behavior? In my defense I will note that my proposal involves a process that begins importantly with genes and ends with behavior. It does not, however, claim that all behavior is exclusively genetically determined. So, it is not genetic determinism. But, not all forms of behavioral genetics need to be thought of as forms of genetic determinism. Nor does my proposal claim that all behavior is determined by external forces. So, it is neither cultural nor environmental determinism, either. Still, it is conceived as a multi-factored version of behavioral determinism, one which reaches back for explicit inspiration at least to the Enlightenment. If the term, “Behavioral Genetics” is not sufficiently broad/flexible to include the view I propose, I welcome an alternative that would do a better job of characterizing the present proposal, one that includes a nod to classical behavioral genetics, which investigates the significance of genetic input across the wide spectrum of human behaviors, as well as recognizing that environmental and non-genetic internal factors are also importantly involved in specific behavior production.
IV. Philosophy of Science
Let us now consider how a broad functionalist perspective might deal with various additional concerns from the philosophy of science, including methodology and explanation. First, let us turn to the topic of explanation. When we ask about explanation in general we are asking, at base, for an insightful fitting together of various phenomena within a broader context. We want to know why we to expect particular phenomena occur instead of alternative phenomena. While for several decades in the last century it may have been widely accepted that all truly scientific explanation had to follow a single pattern, this is not longer the case. It is now widely recognized that there are a variety of different sorts of explanation that can occur in serious science. In what senses can the proposed rubric help determine the sorts of explanation that are possible in broad behavioral genetics appeals? There are two answers.
First, there are many opportunities for different kinds of explanation to occur throughout the [gene → behavior] path; few if any of the major forms of explanation are not, eventually, going to find a place somewhere. Beginning with the classical DN model and its causal variant, we can find a place for deductive nomological explanation in the existence of the various laws of nature that seem to govern certain well-understood genetic combinations at the beginning of the rubric. At this point in our understanding, however, given the great complexities of relationships in the rubric proposed, it seems unreasonable to suppose that there will be interesting Deductive Nomological [DN] or direct causal connections, even with generous ceteris paribus clauses, between specific genetic states and specific behaviors. But, opportunity of possible appeal to Inductive Statistical [IS] explanation seems much more open. Many of the connections between states of the rubric seem to be ones where the possibility of finding a probabilistic connection seems rather significant. But, unlike, for example, causal appeals, many will find probabilistic connections to be incomplete and inconclusive. Something more closely fitting the phenomena together, more closely establishing natural connections seems wanted. I have already indicated that functional explanation plays a role here, both in terms establishing the ground of all behavior and, perhaps, for a few very special behaviors. But, I have also claimed that the explanatory insight one can achieve by functional analysis will be limited. So, we end our discussion of the first point relating to explanation by asking: what more helpful explanatory account can we give that is more specific to the situation at hand?
This brings me to the second answer I want to propose regarding explanation. How does one typically explain the attribution of a particular trait to a member of a particular group? While biology is one important domain where occurs, we need to consider that something similar also happens in a wide variety of domains, including the other natural sciences, social sciences and studies, religious explanations, etc. How might we account for the explanation within the context of attributing a particular feature to an item across this wide category of items?
As a start let us consider the following outline of an explanatory framework:
Outline of Trait Acquisition Explanation
Explaining the acquisition of a trait by an individual item typically involves the following
1. Identifying the trait to be acquired
2. Determining the origin of the process
3. Distinguishing the stages in the acquisition process
4. Discovering the mechanisms for trait acquisition
5. Ascertaining the relative effectiveness of the mechanisms
6. Establishing how the trait acquisition process moves from one stage to another
7. Recognizing the internal environmental factors affecting acquisition
8. Finding out the external environmental factors affecting acquisition
9. Detecting how the various factors interact in particular contexts
10. Assessing the measure of likelihood of the process working in a particular case
The complex form of explanation that is at work here is not, in spite of what one might first imagine, an ad hoc invention that is limited only to cases of behavior genetics. Instead we can view trait acquisition in much broader framework. We should note that this explanatory framework, while nicely fitting the quasi-functionalist model proposed above, also can be seen to apply to a wide variety of alternative explanatory approaches. Such views as Special Creation Theory, Developmental Systems Theory, Classical Lamarkianism, Evolutionary Explanations, Narrative History Accounts, Standard Marxist Interpretations, Classical Behaviorist Explanations, even certain fundamentalist religious accounts of trait acquisition, would all seem to fall nicely within this 10-stage theory.
This form of explanation, however, has several limitations. First, it is “mechanistic”. But the notion of mechanism here is no more conceptually complicated than its original early Greek sense of that which brings something about in a rather regular way. As such, it seems consistent with what we typically require of a scientific explanation. Second, it is also a stage based theory, which implies a development, either of groups or individuals over time. So, it is limited in applicability to explaining items that occur historically. Still, most traits possessed by items would seem to fall into this category, ensuring the approach wide applicability. And, there might also be two different applications of the approach to account for the same trait. This might occur, for example, if we could produce both (a) a developmental account of the occurrence of a particular behavior in an organism as well as (b) an “Evo-Devo” account of the same behavior evolutionarily. And, third, it is aimed at attempting to account for complex phenomena, involving a variety of different influences. This is, at the same time, a strength of the approach, as there are numerous complex phenomena the explanation of which seems to invite its use, as well as its most serious challenge, as we will require, in a particular case, some way of sensibly bringing the various diverse influences sensibly into consideration one with another.
I want the causal role view here proposed to be consistent with scientific pluralism. But this interpretation is likely to lead to an initial objection: why should scientific pluralists not oppose the present quasi-functionalist approach, which is itself a single, comprehensive account?
Here is the answer I propose. If the “big picture” presented above is roughly on track, then among the questions we need to ask include, most prominently, how can we proceed to study the various aspects of this proposal so as to make investigative “progress”? Attempting to relate all of the various aspects of all of the various stages all at once seems like a recipe for disaster. But, the plan of limiting our study to attempting to determine the nature of the relations between neighboring stages or between stages related by a gap while simultaneously holding other stages as constant as we can makes some sense. And, this seems to be just what the various traditional approaches to behavioral genetics do.
The problem is that without a “big picture”, even one as rough and full of flaws and limitations as the above is, one is likely to regard the efforts of those engaged in different explanatory processes as being hopelessly naïve or terribly misguided, instead of seeing how their various efforts might contribute to the overall project of explaining how the process of acquiring behavior in the largest biological sense might work.
Now, one should grant defenders of scientific pluralism that there are serious dangers in being a slave to all big theories. The history of science provides many good examples of this danger. But, it would be a mistake to regard all appeals to “big-picture” thinking as either necessarily dangerous, or totally empty of positive value. It depends upon the specific context of investigation and upon the shared assumptions that all researchers are willing to make. In biology, for example, there is a very big theory the filling in of whose many details we are all regularly involved—and this, of course, is Darwinian Evolution. But, biological researchers are, standardly, conscious Darwinians. If we grant that Darwinian thinking is not dangerous thinking from a scientific perspective, which we all should, then it is not mere bigness that is bad. What distinguishes Darwinian from other large scale thinking may be the fact that Darwinian thinking is not necessarily tied to specific mechanisms, or a fixed time-table, or a single simple way of explaining things. So, suitably flexible big pictures are permissible, especially if they encourage a variety of approaches for engaging in scientific investigation. The same, I suggest, holds with regard to the quasi-functionalist approach to connecting genetics and behavior.
There is a second likely objection from scientific pluralists that needs to be considered. This is the objection to a simple addition of the various diverse features from different approaches. Longino, for example, claims that one cannot regard the four different standard approaches to the addressing the appearance of behaviors as being ones that can be added.  For Longino these different approaches are distinctly in conflict with each other; thus, it is inappropriate to assume that it is meaningful to combine conflictual elements.
In response, I would point to the important scientific pluralist claim that scientific pluralism is not reducible to an “anything-goes” form of scientific relativism. But, in order to avoid the charge of “anything-goes” relativism there needs be a common framework within which the players agree to participate. Once the common framework is articulated and in place, however, it is open to serious researchers to appeal to a variety of approaches appealing to element within the common framework to investigate the phenomena. While it is important to note that the correct ways of currently construing certain approaches would make any current attempts to combine data from them into single model incoherent, this by itself does not show that future efforts, by changing particular aspects in different approaches, could not do just that. What would be needed, here, would be a general argument against the coherence of any such efforts at rapprochement. And, I am doubtful that such an argument can be given.
So, with regard to the question, “Must pluralists oppose the combining of approaches?” I think the proper answer is, “No’; the pluralist needs only hold that such a combination must, itself, be viewed as yet another, separate approach. This combination approach, moreover, should instead be viewed as conflicting with the other current approaches. Further, we may note that a combined approach may be likely to give rise to further alternatives with which it will conflict.
I want to end this section on methodology by briefly noting several additional objections that might be raised against the quasi-functionalist approach here presented, objections based on concerns with functionalism itself, which would threaten the methodological utility of the proposed quasi-functionalist approach to behavioral genetics. First, some have worried about a certain ‘emptiness’ in the functionalist approach. This charge is worth taking seriously so long as we have not figured out anything about what the relevant elements in the causal role might be and how they might be related. Second, the reliance on functionalism in the philosophy of mind, in spite of several decades of popularity, has not actually produced significant, detailed definitions of mental states that can withstand scrutiny. It is natural to worry whether one should have higher expectations for behavioral genetics. Third, there is a standing worry to the extent that qualia are implicated in any causal role instance from genes to behavior. And, fourth, there is a related worry, namely the problem of distinguishing standard vs. non-standard causal roles that almost all varieties of functionalism inherit. Thus, while the functional role approach to connecting genetics and behavior is suggestive, certain inherited worries also need to be faced.
V. An Ethical Conclusion?
I would like to conclude with consideration of two different ethical concerns regarding behavioral genetics. First, there are two ethical arguments against genes from Fox Keller to consider. She concludes her 2000 discussion as follows:
….Where my own sympathies lie should now be apparent. Genes have had a glorious run in the twentieth century, and they have inspired incomparable and astonishing advances in our understanding of living systems….But these very advances will necessitate the introduction of other concepts, other terms, and other ways of thinking about biological organization, thereby inevitably loosening the grip that genes have had on the imagination of life scientists these many decades. My hope is that such new concepts and new ways of thinking will soon work to loosen the even more powerful grip that genes have recently come to have on the popular imagination. For if…the term gene may in fact have become a hindrance to the understanding of biologists, it has perhaps become even more of a hindrance to the understanding of lay readers, misleading as often as it informs. As a consequence, it shapes popular hopes and anxieties, in ways that are often off target and in fact counter-productive to effective discussion of public policy even where the issues are real and urgent. At my most optimistic, I even imagine the possibility that new concepts can open innovative ground where scientists and lay persons can think and act together to develop policy that is both politically and scientifically realistic. [pp. 147-148]
I read this final comment from Fox Keller as proposing two additional ethical arguments against continued gene talk. First, she is worried that gene talk misleads scientists and, especially, that it causes the lay public to holding erroneous and anxiety-producing ideas. In reply, it should be noted that Fox Keller’s worries would be worth taking seriously if there is no change in the current use of the concept of gene and if the public buys into a simplistic one-gene-one-behavior view. But, if the approach I am suggesting is accepted as a replacement for the earlier, simplistic view, it is hard to see how gene talk embedded in a functionalist approach such as that suggested above could have the effects Fox Keller fears.
Second, Fox Keller also seems to be claiming that current gene talk is counter-productive to effective public policy discussion and that there is a need for concepts which would further better public policy discussion. Again, I would agree that a simplistic understanding of behavioral genetics would not be useful for meaningful public policy discussion. The acceptance of the more nuanced account of how genes lead to behavior presented here would have a similar salutary effect on public policy discussion without requiring massive linguistic dislocation.
As the case of the atom considered early on demonstrates, it is probably just too late for us to eliminate gene talk. Current evidence indicates that the brief against the gene is just not strong enough for us to want to abandon the term altogether. A better strategy is to improve our use of the term, make it fit better what we know about how inheritance works and how it is effective. In this way, we can achieve the results Fox Keller wants without having to engage in future linguistic revision. I am betting that the concept of the gene will still be around for quite a while.
Second, given that the concept of the gene is going to be with us for a while, there are certain aspects of the issue of behavioral genetic engineering to which the proposals of this discussion are relevant. Let us consider that certain genetic manipulations that are likely to lead to particular behavioral modifications are proposed. What sorts of changes should cause us ethical concern? If one can establish that certain genetic changes are likely to have certain negative behavioral results in likely circumstances of development, this would seem to count against allowing such manipulation. Suppose, for example, that one of the behavioral effects of creating a super-strong human through genetic manipulation would also cause the individual also to have, given likely common environmental factors that cannot be eliminated, dangerous aggressive tendencies. This would create a serious ethical concern about such manipulation. Changes that risk making life worse, either for developing individuals or for society as a whole, will correctly incur public moral disapproval. (Compare with individuals whose drug or alcohol consuming behavior during pregnancy negatively affects the developing fetus.) But, on the other hand, suppose one can establish that certain desired behavioral modifications that can be tied effectively to genetic manipulation within the range of “the normal”, that is, are features that are had by healthy human beings. And, further suppose that there are no likely negative behavioral side effects that could be expected in those circumstances. In such a case much negative reaction to these “normal” behavioral modifications will also be neutralized. (Compare with individuals who maintain a healthy diet, take special vitamins, or listen to Mozart during pregnancy with the goal of bettering fetal development.) Both of these cases indicate how the enhanced behavioral genetics model can be used to help us get a better ethical grip on different kinds of genetic manipulation for the purpose of behavioral enhancement. Both rely on trying to figure out what the normal parameters of development are. And, as I have argued above, this is precisely what a revision of behavioral genetics should have as its primary goal. Further, one can note that once we have a good sense for the parameters of the system, then we will also be in a better position to evaluate and plan against environmental threats.
Let me now summarize. In this discussion I have presented and replied to an attempt by Evelyn Fox Keller to eliminate gene talk. The strategy I adopted to avoid her complaint led to a functional account of genes. This functional account, in turn, led to my proposing a functionalist framework for doing behavioral genetics. My attempt to account for the specific kind of scientific explanation provided by the framework subsequently led to my articulating a general explanatory framework for trait acquisition, which has broad application, both inside and outside of science. I then considered two sorts of methodological complaints: first, concerns from scientific pluralists were addressed in an attempt to keep my functionalist proposal for behavioral genetics within the pluralist fold; and, second, various possibly applicable traditional worries for functionalisms that might also apply to a functionalized behavioral genetics were noted. I concluded by showing some ways in which this approach also helps deal with ethical arguments against continued gene talk and with ethical worries about genetic manipulation. Thus, we have seen in this discussion that if we rethink the behavioral genetics project, we can use it as a way to deal with an impressive variety of problems.
an international and interdisciplinary journal of postmodern cultural sound, text and image
Volume 7, March-April 2010, ISSN 1552-5112
D. M. 1973 Belief, Truth and Knowledge,
Armstrong, Armstrong, D. M 1968 A Materialist Theory of the Mind, Routledge & Kegan Paul,
Carroll, Sean 2005 Endless Forms Most Beautiful: The New Science of Evo-Devo,
Robert M. 2007 Babies
by Design: The Ethics of Genetic Choice,
John 2004 The Philosophy of Mind, 2nd
Christopher 1991 Sensations,
Longino, Helen 2006 “Theoretical Pluralism and the Scientific Study of Behavior,” in
Kellert, et. al., pp. 102-131.
Keller, Evelyn Fox 2002 Making Sense of Life: Explaining Biological Development with
Metaphors, and Machines,
Helen Longino, and
C.K. Walters, eds. 2006 “Introduction: The Pluralist Stance,” in Kellert, et. al., Scientific
Helen Longino, and
Waters, eds. 2006 Scientific
Jaegwon 2005 Physicalism, or Something Near Enough,
Kitcher, Philip 2001 “Carl G. Hempel,” in D. Sosa and A. Marinich, eds. A Companion
to Analytic Philosophy, Blackwell,
Philip 2007 Living with
David 1980 “Mad Pain and Martian Pain” in Ned Block,
of Mind, Vol. I,
MA, pp. 216-222.
Lycan, William 1981 “Form, Function, and Feel,” Journal of Philosophy, Vol. LXXVIII
(1981), pp. 24-50
Millikan, R. G. 1984 Language, Thought, and Other Biological Categories:
New Foundations for Realism, The MIT Press,
Paul Griffiths, and
Russell Gray, 2001 “Introduction: What is Developmental Systems Theory?” in Cycles
Contingency, ed. Oyama,
Parens, Eric 2004 “Genetic Differences and Human Identities: On Why Talking
about Behavioral Genetics is Important and Difficult,”
N. Press, eds. 2006 Wrestling with Behavioral Genetics: Science, Ethics and Public
Press, Nancy 2006 “Social Construction and Medicalization: Behavioral Genetics in
Context,” in Parens, et al., pp. 131-149.
Putnam, Hilary 1967 “The Mental Life of Some Machines,” in Intentionality, Minds
Perception, ed. H-N Castañeda,
Quine, W.V. Quine, W.V.O. 1969 "Speaking of Objects", in Ontological Relativity and Other Essays,
Salmon, W. Salmon, W. C. 1984 Scientific Explanation and the Causal Structure of the World,
Shoemaker, Shoemaker, Sydney 1984 “Some Varieties of Functionalism,” Philosophical Topics, Vol 12,
Turkheimer, Eric 2006 “Mobiles: a Gloomy View Research into Complex Human Traits,”
in Parens, et. al., pp. 100-107.
Waters, C. K Waters, C. Kenneth 2006 “A Pluralist Interpretation of Gene-Centered Biology,” in
Kellert, et. al., pp. 190-214.
Wimsatt, William C. 1972 “Teleology and the Logical Structure of Function Statements”
Studies in the History and Philosophy of Science, 3: 1-80.
Wright, Larry 1973 “Functions,” The Philosophical Review, 82: 139-68.
 See Carroll, 2005, passim.
 NIH website on Behavioral Genetics.
 See Press, 2006.
 See, for example, Waters, 2006.
 Including several authors at this conference.
 See Waters, 2006.
 Fox Keller’s Herculean Burden Argument is eliminativist in several ways. First, she seems to hold that the existence of genes is unlikely given the empirical evidence she considers. Second, she seems to be claiming that certain key features of genehood can be seen not to apply to any existing biological items, so, strictly speaking, it is incorrect to continue to use the term ‘gene’. Third, Fox Keller sometimes claims that we be able to do genetics perfectly well without using the term ‘gene’. And, fourth, she seems to hold that the term ‘gene’ is tied to a defective scientific hypothesis that needs to be replaced with a better hypothesis.
 It is worth noting that, while Fox Keller proposes the elimination of talk of genes, she does not also offer and argue for a specific alternative. This, of course, does not show the Herculean Burden argument to be misguided. But, the inability to come up with a ready replacement for gene-talk might be taken as a sign that there remains an important role for genes to play.
 See also Carroll, passim.
 Whether certain items are full blown causes, partial causes, background conditions, standing conditions, INUS conditions, or whatever, as long as they are a part of the causal story of inheritance that will suffice for those items being considered to be genes.
 Quine, 1969, coined the dictum “No entity without identity.”
 This is very close to Armstrong’s 1968 central-state characterization of mental states.
 See Kitcher, 2007, p. 24.
 To abandon the search based on what Fox Keller presents seems premature. Keller cites a few cases in which stressed cells can cause the same sequence to produce a protein other than what it normally produces; but, this poses no problems for functional analysis.
 There could be information that really would spell the death knell for genes. Imagine two different cases. Suppose, first, we were to discover that two genetically identical individuals, exactly the same genome down to the last amino acid, who were placed in extremely similar environments, with extremely similar forces acting upon them, yet who developed in radically different ways in almost all cases. One was short and fat, the other tall and lean; one was smart, witty, musical and artistic, the other was below average, unfunny, tone deaf and artistically challenged, etc. Second, suppose we were to discover two genetically very diverse individuals, with radically different genomes, who were placed in extremely different environments, with extremely different forces acting upon them, yet who develop in radically similar ways in almost all cases. Both individuals were often mistaken, one for the other. Both were the same height, weight, intelligence, and of the same sort of artistic bent. Both had the same major/minor in college, were in the same profession, etc. If there were regular examples to be found of the above two sorts, reliance on “genes” to explain traits acquisition would appear compromised.
 See Heil, 2004.
 See Putnam, 1967.
 This is a plausible initial interpretation, for example, of Armstrong (1968). But, Armstrong later (1985) allows for future science to have the last word.
 See Shoemaker, 1981.
 See Lycan, 1981, and Millikan, 1984.
 There are, of course, interesting partial cases. Hill (1991) still holds out for the possibility of type-type identity in some cases. Others, such as Kim (2005), only embrace functionalism partially, urging that some phenomena, qualia, cannot be functionally explicated.
 See Nissen, 1997, for a comprehensive discussion and critique of work on functions. Nissen mostly considers naturalistic interpretations of function-talk. See also Plantinga, 1992, for an explicit defense of a super-naturalistic interpretation. The account of Wright, 1973, still enjoys wide support. See Nissen for a critique of Wright.
 Note that it all begins with behavior!
 See Carroll, 2005.
 My indebtedness to Turkheimer, 2006, should be evident.
 Following Armstrong, I take dispositions to be internal states. See, Armstrong, 1973.
 It should also be noted that some of the stages in this process are themselves likely to be mental states, especially if one is considering commonly discussed behaviors as schizophrenia, aggression and intelligence. Given that mind-body functionalism is a most attractive candidate for providing the analysis of such traits for philosophers of mind, it makes some sense to attempt the extension of the functionalist strategy to the whole of the behavioral genetics process.
 One reason is that functionalists typically do not view themselves as being required to produce specifics to support their approach. (See Shoemaker, 1981.) They think it sufficient to indicate the general form of the view whose specifics will be filled in by further appropriate research.
 We can easily imagine, however, a change of environmental circumstances which led us to attribute a particular function in this latter case, were it to be found out, for example, that 1st tenors had better luck than 2nd tenors, baritones or basses in attracting fertile altos and sopranos.
 See Longino, 2006.
 It is worth considering which complications that supporters of DST push in terms of feedback loops between all of the different aspects of their view should be included in the proposed model.
 It is also interesting to note a certain historical irony in my proposal. I am proposing that one look to current philosophy of mind for guidance for a model of how to do behavioral genetics. But, if one looks back to the seminal work of Armstrong in motivating an early version of what became the functionalist account of mind, one finds Armstrong making reference to the sort of identity that exists between the gene and DNA. I find it somewhat ironic that a certain perhaps overly simplistic conception of the relation between the Gene and DNA example helped to motivate what developed into functionalism in the philosophy of mind. And now, I am proposing that we look to these very developments in the philosophy of mind to help clarify how we think about genes. Given the complexities in our knowledge of DNA molecules since 1968 when Armstrong first made this claim, it is appropriate that we re-think how we view this.
Armstrong is clear that his original claim was a type-type claim, but he came to admit that a multiply instantiated version of the view might be acceptable. (Armstrong 1985)
 See Kitcher, 2001.
 See Salmon, 1984. Fans of Salmon’s approach will note that Salmon’s “mark” criterion for causality may fare better in connection with explanation of genetic influences than in other areas.
 We might look to classical Mendelian genetics for cases such as the following, AA + aa → Aa, a simple monozygotic cross, as well as more complex cases for examples that would work for both Hempel and Salmon’s accounts of scientific explanation.
 As for pattern-based explanation, the situation is basically up in the air. While a Kitcher-style appeal to unifying pattern cannot be ruled out here [Kitcher], what yet seems missing is any clear sense of what such a unifying pattern might look like, given the great complexities involved. Still, it is not unreasonable to think that the functional role account linking genes and behavior, were it to develop, would very likely be characterized precisely by common unifying patterns.
 See, e.g., Plantinga’s explanation of theist vs. revealed Christian belief. (Plantinga, 2000.)
 Those accounts which limit the mechanisms involved, or internal or external influence, will not take all of the steps into consideration, and, there may be additional steps that should be added. It should also be noted, as a confirmation of the standard critique of the view, that the defender of Intelligent Design would be hard pressed to say how the explanatory account might work. (Compare Kitcher, 2007, passim.)
 It would not work, for example, to explain the presence of traits that have always been possessed by some item.
 See Carroll, op. cit.
 Compare Longino, 2006.
 But, see Kitcher 2007, for an attempt to explain its apparent danger for other reasons.
 Longino, 2006.
 One can support this charge by noting how it works in other contexts. Consider a religious controversy in which three different items are proposed to explain a particular trait. One cannot assume that it makes any sense whatsoever to add these three approaches together. Suppose, for example, a Manichean, a Calvinist, and a Catholic are trying to explain aggressive behavior. The appeal to a bad god’s interference, to bad effects of ‘The Fall’, and to inappropriate use of free will might be the various elements in play. But, a combination of these three elements would make no sense at all to any of these three discussants!
 Waters, op. cit.
 A super-naturalist appeal to “The Fall” to explain antisocial aggressivity or homosexuality must be ruled out of court, for example.[compare Plantinga, 2000)
 To pursue the example from footnote 45, this would involve proposing a fourth religious perspective that recognized multiple gods, some bad, in addition to free will and The Fall.
 This suggests a further question: must scientific pluralists hold that there will never be a general agreement as to how to do research in a particular area? The history of science, while full of disagreements, is also not totally devoid of examples of agreement. [Were this not the case, then Kuhn could never have proposed his views as a critique of accretionism!] I would suggest that the answer should be, “No”, only that such agreement is to be viewed as another approach, which may be temporary.
 Armstrong, for example, before he admitted to being a functionalist, once referred [in discussion] to mind-body functionalism as ‘fig-leaf materialism’. The ‘emptiness’ charge faces all new approaches to a particular scientific problem, and is only worth taking seriously when the approach has received significant effort and continues not to produce ‘the good’.
 See D. Lewis, 1981, for an attempt to avoid this worry. Super-naturalist versions may avoid the problem; see Plantinga, 1993. Further, some may argue that developments in Evo-Devo might provide a suitable standard to avoid the worry.
 The evidence on inheritance from ordinary experience, the natural sciences, and the Twin Studies is just too intermingled with gene talk for us to expect that a different set of concepts would provide better insight and be easily substitutable for current talk of genes.
 For a comprehensive recent defense of genetic engineering, see Green, 2007.
 As I write this paper [late May, 2008], a study has recently been announced in the popular press commenting on the continuing link between lead paint and a decline in basic intelligence.
 I am deeply indebted to many insightful discussions with L. Russert Kraemer on many of these topics over many wonderful years.