20th WCP: Philosophy of Science and the Theory of Natural Selection

In the 1980s, the hitherto-dominant normative-prescriptive conception of philosophy of science became the subject of a debate which continues to the present time. Some philosophers of science suggested that the proper aim of the discipline is the description of scientific evaluative practice.

There is a modest version and a robust version of descriptive philosophy of science. The aim of the modest version is the historical reconstruction of actual evaluative practice. Given that scientists preferred one theory (explanation, research strategy...) to a second, the modest descriptivist seeks to uncover the evaluative standards whose application led to this preference. For instance, the modest descriptivist may seek to uncover the standards implicit within such evaluative decisions as Aristotle's rejection of pangenesis, Newton's rejection of Cartesian Vortex Theory, or Einstein's insistence that the Copenhagen Interpretation of quantum mechanics is incomplete. Pursuit of a modest descriptive philosophy of science may require a certain amount of detective work, particularly for episodes in which the pronouncements of scientists and their actual practice do not coincide.

The conclusions reached by modest descriptive philosophy of science are subject to appraisal by reference to standards applicable to historical reconstruction in general. There is no distinctively philosophical task of appraisal. The modest descriptivist is a historian with a particular interest in evaluative practice.

The robust version of descriptive philosophy of science derives from, or superimposes upon, the conclusions of modest descriptivism, a theory about evaluative practice. The theory is put forward as a contribution to our understanding of science. It purports to explain why science is as it is. A robust descriptive philosophy of science typically includes the claim that scientific evaluative practices exhibit

certain patterns or conform to certain principles. Of course, not every historical instance will exhibit a pattern exactly or conform precisely to the requirements of a principle. But a successful robust descriptive theory must help us to understand at least some important episodes from the history of science.

A number of philosophers of science have sought to account for the growth of science by reference to the Theory of Organic Evolution. Michael Ruse has observed that "evolutionary philosophies of science" may be subdivided into two types. (1)

1) The "Evolutionary-Analogy" View develops an analogy between the growth of science and the operation of natural selection upon a pool of variants subject to environmental pressures. On this view, competing concepts, theories and methods of inquiry engage in a competitive struggle from which the "best adapted" emerge victorious.

2) The "Evolutionary-Origins" View attributes the growth of science to the application of epigenetic rules that have proved adaptive within the course of evolutionary history. On this view, science develops as it does because certain methodological rules and evaluative principles have become encoded in our genes. Acting on these rules and principles presumably proved adaptive for our proto-human ancestors.

The "Evolutionary-Analogy" View and the "Evolutionary-Origins" View are, in the first instance, robust descriptive philosophies of science. In addition, some evolutionary theorists have drawn normative-prescriptive recommendations from their theories about scientific progress.

The Evolutionary-Analogy View

Stephen Toulmin, David Hull, Donald Campbell and Karl Popper have developed versions of the Evolutionary-Analogy View. Toulmin maintained that the development of scientific concepts is an evolutionary process in which the fittest conceptual variants survive. (2) In a period of "Kuhnian Revolutionary Science", for instance, the victorious "paradigm" (set of concepts) is the one that best resolves the "anomalies" (disciplinary pressures) that gave rise to the adaptive crisis.

Toulmin held, moreover, that the identity-through-change of a scientific discipline is analogous to the identity-through-change of a biological species. Since we believe that we understand the interrelationship of the concepts 'variant', 'environmental pressure', 'adaptation' and 'fitness' within the theory of organic evolution, successful analogical extension of this interrelationship to the history of scientific evaluative practice presumably conveys understanding of "how science works".

David Hull sought to show that organic evolution and conceptual change are specific instances of a "General Theory of Selection Processes". (3) The General Theory specifies a mechanism by which entities pass on structure over time. The mechanism involves an interrelation among "interactors" and "replicators". Interactors compete with one another in response to environmental pressures. The resultant competitive differential adaptation of interactors causes differential success rates among replicators.

Replicators are entities which give rise to copies of themselves. In the organic realm, "replication occurs primarily at the level of the genetic material". (4) Interactors are entities subject to competition within some specific environment. In the organic realm, "interactors" include not only living organisms, but also genes, chromosomes, cells and kinship groups. (5)

Within the history of science, replicators are concepts and beliefs, and interactors are individual scientists and individual research groups. "Concept-replicators" are individual entities, but they exist in various contexts of interrelatedness. Beliefs also are individual entities, and include commitments to methodological principles and standards of appraisal.

Hull maintained that the history of science, like the history of organic forms, is the result of selective pressure operating upon a set of variants. The history of science is a history of "lineages", theories that change over time while retaining self-identity. Theories, like species, are entities determined by phylogeny, and not by possession of a common structure or set of properties. What counts is descent and not sameness of content.

Hull noted that, on a phylogenetic understanding of theory-life, "unappreciated precursors do not count." (6) Thus, Patrick Matthew's unnoticed formulation of the principle of natural selection (1831) is not part of the lineage of natural-selection theory. He noted also that a "phylogenetic" reconstruction of Darwinian Evolutionary Theory reveals a tree of descent whose branches include "Darwin's Darwinism, late nineteenth-century Darwinism, neo-Darwinian Darwinism, the new synthesis Darwinism, and so on". (7)

Hull's descriptive philosophy of science is robust. He maintained that we may increase our understanding of evaluative practice within science by reference to the General Theory of Selection Processes. Indeed, he claimed that otherwise puzzling aspects of evaluative practice are readily understood when viewed from the standpoint of the General Theory of Selection Processes.

One such puzzle is the success of science in policing the activities of its members. (8) The great majority of professional organizations are ineffective at imposing discipline upon their errant members. Science is a striking exception to the general rule. On Hull's model, an individual scientist is an interactor subject to the pressure of selection. Her "fitness" is established by publishing results that subsequently are acknowledged and utilized by other scientists. It is not in the long-term self-interest of a scientist to falsify or fabricate data, or to distort the conclusions reached by other scientists. Consequently, scientists collectively are quick to condemn and punish those occasional miscreants who undermine the process by which lineages are created.

A second puzzle is the vehemence of priority disputes among scientists. If the goal of science is the formulation of increasingly more powerful theories, what does it matter who receives credit for them? On the General Theory of Selection Processes it matters greatly. Theories are phylogenetic entitities. An interactor's fitness is measured by her contribution to such lineages. Any challenge to an interactor's role in the creation of a lineage is a serious threat indeed. Given Hull's Evolutionary Model, it is not surprising that priority disputes often are heated controversies.

Unfortunately for the Evolutionary-Analogy Program, there are important disanalogies between organic evolution and the growth of science. L. J. Cohen pointed out two such disanalogies.

In the first place, the process by which variants are produced within a breeding population takes place independently of the process by which the "better adapted" individuals succeed in the struggle to survive and reproduce. Mutation is a spontaneous, random process. As Cohen put it

the gamete has no clairvoyant capacity to mutate preferentially in directions preadapted to the novel ecological demands which the resulting organisms are going to encounter at some later time. (9)

The situation is otherwise in science. Variant scientific concepts, methodological rules and evaluative standards are consciously created in order to overcome recognized deficiencies in older concepts, rules and standards. Thus there is an important relationship between the formation of scientific concepts and the subsequent fortunes of the theories within which they occur. "Variation" and "selection" are not uncoupled processes within science.

In the second place, biological species are not analogues of scientific disciplines. A biological species is a population of similar individuals each of which is a representative of that species. The same is not the case for a scientific discipline. A scientific discipline includes concepts, invariant and/or statistical relations among concepts, theories about underlying mechanisms, procedural rules and evaluative standards. These diverse ingredients are interrelated in complex ways. Even if we restrict attention to scientific concepts, it is clear that a concept such as 'force' does not instantiate physics in the way that 'Cigar' instantiates the species Equus.

Cohen emphasized that the identity-through-change of a biological species is markedly dissimilar to the identity-through-change of a scientific discipline. A biological species retains its identity provided that a set of individuals with similar characteristics at time t2 resembles in relevant respects another set of individuals with similar characteristics at time t1. But the identity-through-change of a scientific discipline is not of this type. In order to solve conceptual problems within a discipline we need a set of interrelated concepts, not a population of concepts with similar characteristics. Hence changes within a discipline involve a restructuring of an 'evolving' concept's relations to other concepts and not just a replacement of concepts similar to c1 by concepts similar to c2.

The adequacy of an Evolutionary-Analogy theory of science depends on the importance of the above-mentioned disanalogies. Cohen maintained that the independence of variation-generation and selection is an essential feature of the theory of natural selection. He concluded that the analogy to the growth of science fails. Toulmin and Hull, by contrast, conceded that this disanalogy exists, but insisted that the Evolutionary Analogy nevertheless provides a useful theory of science.

Donald Campbell sought to reinstate the evolutionary analogy by shifting attention from "random mutations" to "blind trials". He acknowledged that scientific beliefs, unlike biological variants, are not produced randomly. The scientist has in mind a problem to be solved and a history of prior attempts to find a solution. However, if scientific progress results from the selective retention of blind trials, then the core of the evolutionary analogy may be retained. Every random trial is a blind trial, but a trial may be blind without being random.

In Campbell's usage, a "blind trial" is a trial that satisfies three conditions: 1) the trial is independent of environmental conditions; 2) a successful trial is no more likely to occur at one point in a series of trials than at any other point in the series, and; 3) no trial in a sequence of trials is put forward as a "correction" of a prior trial. Campbell claimed that

a blind-variation-and-selective-retention process is fundamental to all inductive achievements, to all genuine increases in knowledge, to all increases in fit of system to environment. (10)

Campbell thus maintained that the goal-directed decisions of scientists to entertain specific hypotheses are "blind" forays into the unknown, and that those hypotheses that prove "nonadaptive" under testing are eliminated.

The "Selective-Retention-of-Blind-Variants" View is a program for further research. In order for Campbell's program to succeed as a robust descriptive philosophy of science, two conditions must be fulfilled. The first condition is that the Evolutionary Analogy--amended to require selective retention of "blind" variants--must fit important episodes from the history of science. The second condition is that the "fit" has explanatory force.

Kepler's work on the orbit of Mars is a promising candidate for Campbellian reconstruction. Kepler hypothesized a number of ovoid orbits for Mars ("blind trials") before hitting upon an ellipse. The "blindtrials" picture also may be superimposed upon various programs to "save the appearances". The list of such programs includes Babylonian astronomy (the use of linear zigzag functions to calculate the day on which the next new moon appears), Ptolemy's mathematical models for calculating the zodiacal positions of the planets, the nineteenth-century algebraic chemistrys of Benjamin Brodie and Josiah Cooke, and contemporary econometric modelling of macroeconomic forces.

The modified Evolutionary Analogy is less promising as an account of theories about underlying mechanisms. It remains to be shown that Descartes' Vortex Theory, the Kinetic Theory of Gases, Molecular-Orbital Theory and Plate Tectonics Theory are results of the selective retention of blind trials.

Ron Amundson has insisted that the adequacy of "selection explanations" depends on the degree to which certain "central conditions" are met. These central conditions place restrictions on variation and sorting. Variations must be spontaneous, abundant and heritable modifications that are "nondirected with respect to the environmental needs of the organism". (11) And given a pool of such variants, there must be a

preferential persistence of those variations which happen to be suited to the environmental needs of the organism or species- and (most importantly) this sorting mechanism is itself nonpurposive. (12)

Amundson maintained that the above conditions are necessary conditions of explanatory success for analogical applications of the theory of Natural Selection. He noted that challenges to proposed selection explanations often take the form of a denial that the requisite central conditions are met.

Campbell's "blind-variation-and-selective-retention" theory of scientific change would appear to be subject to two principal challenges. In the first place, the generation of variant hypotheses is neither random nor blind. Rather, hypotheses are put forward in response to recognized inadequacies within the scientific environment. In the second place, the sorting process that results in a decision to reject a high-level hypothesis involves judgments about background knowledge, auxiliary assumptions, experimental procedures, and sometimes even metaphysical principles. Amundson declared that

the process of rejecting a 'falsified' hypothesis is often more like the literary critic's negative assessment of a poem than like the cold wind's freezing of the baldest polar bear. (13)

Campbell sought to blunt this type of criticism by acknowledging the existence of "shortcuts" for the blind-variations-and-selection-process. The formulation of hypotheses designed to address perceived deficiencies within a scientific domain may serve as a shortcut for the method of blind trials. Campbell insisted that such shortcuts themselves are an inductive achievement. He declared that we have acquired a

wisdom about the environment achieved originally by blind variation and selective retention. (14)

Purposefully generated (nonblind) hypotheses may be part of a "phylogenetic lineage" the earlier members of which were generated by the blind-variation-and-selective-retention process.

Campbell's introduction of "shortcuts" raises anew the question about the explanatory force of his Modified Evolutionary Analogy. Does it have explanatory force to argue that scientists entertain goal-directed hypotheses today because at sometime in the past our ancestors engaged in blind trials (upon which selection operated)?

Karl Popper endorsed Campbell's version of the Evolutionary Analogy. He held that the conjectures of scientists are analogous to variations and that refutations are analogous to the selective retention of variants. On this view, the "phylogeny" of science is the lineage of theories that survives the rigors of severe testing.

Popper's opposition to inductivism is well known. He repeatedly insisted that there can be no successful algorithm for theory-formation. Popper likened the position of the theorist to the

situation of a blind man who searches in a dark room for a black hat which is--perhaps--not there. (15)

The theorist, like the blind man, proceeds by trial-and-error, coming to learn where the hat is not, without ever reaching a certainty immune from rejection in the force of further experience.

Popper is correct to emphasize the role of creative imagination in the formulation of scientific hypotheses. The problem-situation does not dictate a solution to the theorist. However, neither are hypotheses formulated independently of the problem-situation. Popper's "black-hat image" is quite misleading. Scientific conjectures are "blind" only in the sense that the outcome of subsequent testing is unknown. They are not "blind" in Campbell's sense of being "independent of the environmental conditions of the occasion of their occurrence".

There is a further difficulty in Popper's particular use of an evolutionary analogy. Popper insisted that scientists ought formulate bold, content-increasing conjectures that run a high risk of falsification. But the Darwinian picture of descent-with-modification is a gradual accretion of small adaptations. Popper claimed explanatory value for an "Evolutionary Analogy" that includes pious references to "Darwinian Theory". But he also introduced "Lamarckian" emphases and "Saltation" effects that are inconsistent with that theory.

Of course, there have been disputes over the specific content of Darwinian Theory. However, if Amundson's "central conditions" express the core of this theory, Michael Ruse is correct to conclude that

Popper has been no more successful than others in making traditional evolutionary epistemology plausible. The growth of science is not genuinely Darwinian. (16)

The Evolutionary-Origins View

The Evolutionary-Origins View, like the Evolutionary-Analogy View, may be defended as a purely descriptive theory about science. The Evolutionary-Analogy View is that there is competition leading to differential reproductive success within both organic evolution and science. The Evolutionary-Origins View is that scientific inquiry is directed by the application of epigenetic rules that have become encoded in homo sapiens in the course of evolutionary adaptation. We have certain capacities and dispositions because it was advantageous for our ancestors to have them.

Michael Ruse called attention to several epigenetic rules that appear to inform human evolution: 1) the partitioning of the (continuous) spectrum into discrete colors, a partitioning that takes place in diverse human cultures, presumably because it confers adaptive advantage in the struggle for existence; 2) the "deep structure" of language uncovered by Chomsky and others, and; 3) the prohibition of incest. Ruse suggested that there exist additional epigenetic rules that govern the creation of science: 1) formulate theories that are internally consistent; 2) seek "severe tests" of theories (Popper); 3) develop theories that are "consilient" (Whewell), and; 4) utilize the principles of logic and mathematics in the formulation and evaluation of theories. (17)

Critics of the Evolutionary-Origins View have pointed out that there is evidence that human beings often make decisions that are inconsistent with these supposedly "genetically-hard-wired" rules. Human subjects affirm the consequent with impunity, succumb to the "gambler's fallacy", and erroneously conclude that the probability of (A & B) is higher than the probability of A alone. Ruse acknowledged that this is evidence against the Evolutionary-Origins View, but insisted that it is

better surely to suppose that much of the time we do not think particularly carefully or logically simply because it is not really necessary to do so, but when pressed we can do so and for very good reasons, namely, that those who could not tended not to survive and reproduce. (18)

This is unconvincing. If certain dispositions are acquired in the evolutionary process because of their adaptive value, then these dispositions ought be uniformly actualized. Ruse is forced to subdivide human actions into those that conform to the epigenetic rules (performed by scientists) and those that do not conform to those rules (performed by nonscientists in cases where it is "not really necessary" to conform). Ruse does not argue that those who fail to apply the epigenetic rules are likely to succumb to evolutionary pressures. Instead he introduces the ad hoc hypothesis that nonconformity occurs in cases in which conformity is not necessary. It is debatable whether such a move is consistent with the empirical method required by the position of evolutionary naturalism. (19)

Normative-prescriptive content may be added to the Evolutionary-Origins View by endorsing the move from

(1) The application of methodological rule R and evaluative standard S were adaptive responses to former ecological pressures P.

(2) R and S ought be applied by scientists today.

Ruse usually resisted the temptation to make this move. However, he sometimes teased the reader with suggestive adaptive scenarios. For example, he contrasted the responses of two hominids to evidence of the presence of tigers. Hominid #1 takes the existence of feathers, blood, paw marks in the mud, and growls from the bushes to establish a consilience of inductions, and flees. Hominid #2 views the same evidence but fails to see the importance of consilience. Ruse then asked "which of these two hominids was your ancestor?" (20) Ruse has led the reader to the point where it is natural to conclude, not only that the disposition to apply a standard of consilience has had adaptive value, but also that this standard ought be applied within science today.

However, Ruse did not issue explicit normative-prescriptive claims on behalf of epigenetic rules. (21) (22) Instead, he compared epigenetic rules to David Hume's "dispositions". Hume had observed that we organize our lives by reading "necessary connections" into Nature. We act on the expectation that correlations experienced in the past will continue to hold in the future. Ruse accepted Hume's account of the dispositions involved in our commerce with the world and appended to the Humean account a theory about the origin of these dispositions.

Hume denied that a rational justification can be provided for our expectations of regularity. Past uniformity does not entail the future continuation of that uniformity. Ruse accepted this Humean claim as well. He suggested that the only "justification" for implementing epigenetic rules is that these rules did arise during the course of human evolution. Ruse acknowledged that to provide a theory about the origin of a rule is to fall short of providing a justification for continuing to implement the rule.

There is a further difficulty for normative-prescriptive versions of evolutionary naturalism. In biological evolution "fitness" is a balance between a successful adaptation to present environmental conditions and the retention of the capacity to respond creatively to future changes in those conditions. In a particular case, successful adaptation may be achieved at the expense of a loss of adaptability. That this has occurred becomes evident only with the passage of time.

The normative-prescriptivist evolutionary naturalist stipulates that a particular evaluative decision is correct provided that it promotes "fitness" in the long run. But how can one know at the time a decision is made that it will do so? One may appeal to the fact that similar decisions in the past have proved to have survival value. However, it always is possible that the ecologically unique present situation requires a different response.

The most adequate appraisals are those rendered long after the fact. Survival is the best indicator of fitness. It is survival that establishes a continuing retention of adaptive capacity in the face of changing conditions. Judgments about contemporary evaluative options are much less secure.

The perceptive normative theorist will accept this limitation on prescriptive adequacy. She will acknowledge the possibihity that what appears at the time to be a "fit" evaluative response may turn out subsequently not to be such. But this need not be a recipe for scepticism. It is no more reasonable to expect certainty in the philosophy of science than it is to expect certainty in science.