Pluralism and the Problems of Demarcation

Pluralism and the Problems of Demarcation

In 1983, Larry Laudan proclaimed the “Demise of the Demarcation Problem” (Laudan 1983). I will argue that the ‘simple demarcation’ problem should indeed be abandoned. However, abandoning the quest for simple demarcation does not entail that demarcation problems are intractable. Though the dream of a single criterion to demarcate ‘science’ from ‘non-science’ is shattered, we can still approach demarcation problems by adopting pluralism. I argue that there are a number of distinct demarcation questions, each making different considerations appropriate, and requiring different solutions. Pluralist demarcation accepts that no single criterion is appropriate to every demarcation question. Moreover, contextual factors are vital in determining the relevant considerations which compose a successful demarcation criterion for a particular demarcation problem.

I develop and defend pluralist demarcation. The argument has three stages, demonstrating the plurality of demarcation criteria (and considerations), questions, and contexts. Each stage is presented in opposition to monistic demarcation strategies—simple demarcation, foundational demarcation and complex monism, respectively. Having ruled out these monistic rivals, I consider rival pluralistic positions. I consider low-key pluralism, network demarcation, and pragmatic demarcation, arguing that none are as appealing as pluralist demarcation.

 

1: Pluralist Demarcation

 

1.1: The Failure of Simple Demarcation

 

When Laudan (1983) declared the demise of the demarcation problem, he meant what Reisch terms ‘simple demarcation’ (1998). The simple demarcation problem demands a criterion by which we can distinguish science from non-science. A great many attempts at simple demarcation populate the philosophy of science literature, from Popper’s (1963) falsifiability criterion, to verifiability (e.g. Ayer’s (1946)), Lakatos’ progressiveness of research programmes (1980a, 1980b), and multi-part criteria such as Thagard’s (1988). Each putative demarcation criterion has its failings. For instance, falsifiability struggles with the Duhem-Quine thesis (Duhem 1906; Lakatos 1968), historical criteria such as progressiveness are inapplicable to new fields, and most of Thagard’s criteria prove individually unnecessary and jointly insufficient for scientific status (see e.g. Derkson 1993). The details of the failure of simple demarcation need not concern us here. Laudan does not so much argue that simple demarcation has failed as state it. In support, we should recapitulate and develop two observations:

  • There is much variation amongst even the paradigmatic examples of science (compare quantum physics to evolutionary biology to medical trials), let alone between typical and atypical examples of science.
  • There is perhaps greater variety still in non-scientific practices.

In support of (2), consider first the distinction between pseudo- and non-science. Intuitively, pseudosciences are non-sciences with scientific pretensions (Hansson 2008). So, we might claim that astrology, Intelligent Design Theory and homeopathy are pseudoscientific non-sciences while, say, philately, literary theory and modern art are simply non-sciences (cf. Glymour 1982). Of course, there is a great deal of variation here—some non-sciences have more scientific features than many pseudosciences; compare the analytic sophistication of literary or musical theory with pseudoscientific practices such as chiropractic (Dupré 1993, 242-3). Demarcation of pseudosciences from sciences may have to be very different from demarcations of non-science from science.[1]

Laudan concludes that demarcation problems are intractable. We should stop attempting to distinguish science from non-science, and instead discuss the confirmation, corroboration and predictive power of our theories. However, there is another option—pluralist demarcation.

1.2: Plurality of Demarcation Criteria

 

The first step in establishing that pluralism is viable is to demonstrate the plurality of demarcation criteria. Here, I quickly summarise the range of considerations which are used to form demarcation criteria. I use a helpful framework from Resnik (2000), who distinguishes four branches of components of demarcation criteria—historical, psychological, sociological and epistemological. To be clear, I do not intend these to be the four demarcation criteria which are used in pluralist demarcation (rather, they are classes of potential criteria and considerations), nor an exhaustive survey. Nor do I intend that every demarcation criteria fits into any one of these classes, or that the groupings do not overlap—many criteria could be placed in multiple classes. The grouping is intended to ease comprehension only.

Historical criteria judge the scientific status of theories and research programmes according to the way they have developed. For instance, Lakatos’ progressiveness criterion sorts research programmes into sciences and non-sciences depending upon whether their development is progressive (accounting for additional empirical content at each problem-shift) or degenerative (non-progressive) (1980a, 1980b). Thagard (1978, 1988) developed Lakatos’ criterion, emphasising relationships to competitor programmes—research programmes are pseudoscientific when they are degenerative and there is a progressive competitor. Such criteria are very important in appraising the scientific status of research programmes and theories, and in understanding historical change in scientific status (Resnik 2000, 254). For instance, historical criteria make sense of how astrology and alchemy could have scientific origins, but subsequently became non-science (Thagard 1978). However, they are incapable of answering every demarcation question—the analysis is muted, for example, when faced with new fields and theories. Moreover, while these criteria are quite effective in demarcating established research programmes and theories, they are usually inappropriate considerations for demarcating scientists from non-scientists or scientific work from non-scientific work. As Lakatos emphasised, it is neither irrational nor unscientific to work within a degenerating research programme in the hope of turning it progressive (1980a, 59-64).

Sociological criteria apply much more clearly to classifying researchers and their work. Sociological criteria demarcate science from non-science according to the values and attitudes of the communities engaged in the work, including shared behaviours and norms (Merton 1973; Collins 1982; Gieryn 1983). A prominent example is Merton’s criterion—he identifies the distinctive norms of scientific communities as universalism, communism, disinterestedness and scepticism (1973). Problematically, these criteria seem unable to distinguish non-scientific individuals within scientific communities (such as frauds) and vice versa. They also may strike us as inappropriate considerations for demarcating research programmes and theories in many contexts—when we ask whether a theory is scientific, we sometimes do not mean ‘Is the theory being tested/used by a scientific community?’ Theories can be scientific, we might argue, even if they are never employed by a scientific community (unconceived or underappreciated theories, for instance). Moreover, scientists can clearly work on or with non-scientific theories.[2]

Psychological criteria and considerations are useful for distinguishing scientists from non-scientists. Criteria such as Giere’s (1988) focus upon the belief-forming processes of individuals, demarcating scientific from non-scientific processes. While this would be a useful way of distinguishing individuals, and even communities, it again seems inappropriate to judge the scientific status of a theory according to the thought-processes by which it came to be believed. Non-scientific theories can be believed (and certainly are often rejected) for scientific reasons (e.g. Laplace’s rejection of the God ‘hypothesis’), while scientific theories are sometimes believed or ignored for non-scientific reasons. Moreover, the analysis is again mute when a theory is not being actively considered.

Epistemological criteria include the methodological principles and epistemic values shared by scientists. Accounts of value in science developed by Kuhn (1977) and McMullin (1982) have been used in this way, especially by Lugg (1987, 1995). Scientists and scientific communities share these values, and scientific theories exemplify these virtues. Attempts have been made to reduce the question of demarcation to a question of scientific methodology[3] (e.g. Lugg 1987, 1995). Theories tested by scientific methods are scientific theories, and precisely this testing is science. However, again, in some contexts methodological criteria are inappropriate. Non-scientific theories can be tested scientifically (for instance, clinical trials of alternative medicines and statistical studies of astrological predictions). Scientific theories are sometimes asserted dogmatically. Moreover, there may be many scientific methods, some of which are employed in non-science (Dupré 1993, 242-3). There may be sciences which do not use paradigmatically scientific methods. In summary, methodology generally fails to delineate the domain of science. A focus upon values of science can be useful. James (1982), for instance, distinguishes the values of rigour, accuracy, consistency and transparency as hallmarks of science. But again, many non-sciences and non-scientists are systematic, consistent, transparent and rigorous. Moreover, for many purposes we still consider bad science to be science. James’ criterion might exclude poor applications of scientific methodology, such as a student who fails to grasp the principles of an experiment, from science.[4] Finally, epistemological factors include theory confirmation. Laudan (1983, 124) emphasises ‘well-testedness’ and reliability as hallmarks of science. However, including confirmation in a criterion changes the boundaries of science again; it excludes putative scientific theories which are not (yet) well-tested. Moreover, this analysis unappealingly (for most purposes) judges past scientific theories which have subsequently been disconfirmed as non-scientific.

There are a range of considerations which can be appropriate to demarcation. Each criterion considered above could be a demarcation criterion in its own right, or could be affixed to others, within or beyond its class, to form compound demarcation criteria. Compound criteria, usually in the form of lists, have been widely defended. Thagard (1978) argues for a three-part criterion, highlighting the theory’s epistemic qualities, the sociological motivations and behaviour of the community, and the historical context with respect to competitor theories. Derkson (1993) presents a seven-component criterion for pseudoscience, identifying epistemological factors (confirmation and reliability), psychological factors[5], and sociological factors such as pretension and community isolation.

Pluralist demarcation holds that we should formulate and use the criterion which is most appropriate to a particular demarcation question, in context. We have already seen that the range of existing criteria struggle with the breadth of subject-matter they are supposed to demarcate—theories, individuals, communities, thought-processes, etc. Different considerations were appropriate to different parts of this subject-matter. The second stage of my defense of pluralist demarcation is to demonstrate the range of distinct demarcation questions, and to suggest that some criteria and considerations are appropriate to some questions, but not to others, while no single criteria is appropriate to all demarcation questions.

 

1.3: Plurality of Demarcation Questions

 

Perhaps the biggest stumbling-block for simple demarcation is the range of subject-matter about which demarcation questions can be, and are, asked. In Section 1.2, we encountered some demarcation questions—what distinguishes scientific theories from non-scientific theories? What makes a research programme scientific? Which individuals are scientists? Which communities are scientific? Which are the scientific thought-processes, methodologies, and values? What is the domain of science? I do not pretend to present an exhaustive list of demarcation questions here, but I do defend the claim that these are all genuine demarcation questions. We can and do ask for demarcations of: fields, works, individuals, communities, theories, research programmes, institutions, practices, events, methods, methodologies, thought-processes, attitudes and values, evidence, and content/domain.

A common strategy to try to avoid engaging with this range of questions, which I will call ‘foundational demarcation’, takes one demarcation question as fundamental, attempts to answer that question, and then uses this answer to address the other questions. Usually only one or two other questions are explicitly addressed, and the details of even these are unexamined. Disputes about demarcation criteria are often really about which demarcation question is foundational—for instance, Derkson and Lugg debated whether the question of who is a scientist (Derkson 1993, 2001) or what is a scientific methodology (Lugg 1987, 1995) is foundational. Usually, opponents of foundational demarcation positions concentrate on attacking the adequacy of the specific demarcation criterion provided.[6] I, on the other hand, want to address the general tactic. There are two ways to attack foundational demarcation as a strategy: we could show that the considerations appropriate to certain demarcation questions are inappropriate to others, and/or we could problematise the crossover from one question to another.

The former strategy was demonstrated in Section 1.2, above. For instance, we saw that some criteria are useful for demarcating scientific individuals (e.g. psychology, sociology, values) while others are less appropriate (e.g. historical criteria, confirmation). For instance, the degree of confirmation or the history of the theories an individual or community work with does not determine the scientific status of that individual or community. The same example functions in reverse—the psychology of the individuals and sociology of the community advocating a theory is often a poor guide to the scientific status of that theory. Similar problems are readily seen when considering scientific methodology (again, epistemological but not historical criteria are appropriate, for instance). Demarcating between scientific and non-scientific works may make considerations such as stylistic conventions appropriate parts of a demarcation criterion—but these considerations are inappropriate for the other demarcation problems. Clearly, different criteria are appropriate to different demarcation problems; using the criteria from one to solve another will fail.

The second strategy hammers home this point. In most cases, defining scientific concepts in terms of other concepts is highly problematic. Suppose we begin with scientific methodology as foundational. We have already considered that not necessarily every scientist uses scientific methodology, and that not everyone using a scientific method is a scientist. This becomes particularly acute when considering practitioners and technicians. Practitioners such as healthcare workers may be considered scientists, perhaps because they practice based upon scientific theories rather than implement scientific method. Meanwhile, technicians may follow scientific methodology, but not be considered scientists—certainly, not in many contexts.[7]

Similarly, not every theory which can be or has been tested scientifically is a scientific theory (in every context), nor is every method used to test a scientific theory a scientific method.[8] Similarly, not every field comprising scientific theories is a scientific field, nor do scientific fields and research programmes contain exclusively scientific theories or employ exclusively scientific methods.[9]

Not every scientific work is necessarily produced by scientists, or according to a scientific methodology, or within a scientific community, according to scientific values, etc. Scientific values may not cohere precisely with scientific methodology (especially if we consider actual practice rather than ideals), nor with the values of scientists or scientific communities[10]. Scientific communities may include non-scientists, while scientists may operate outside any scientific community, or within a non-scientific community. Consider frauds operating within scientific communities, or scientists such as Michel Gauquelin who tests astrological theories through statistical-scientific methods (e.g. Gauquelin 1975), and Edzard Ernst and colleagues (e.g. Ernst, Pittler, and Wider 2006) who systematically tests alternative medicines through epidemiological methods.

There may be non-scientific theories which make claims within the domain of science. For instance, in the USA, Intelligent Design Theory has been repeatedly ruled non-scientific (McLean v. Arkansas Board of Education 1981; Kitzmiller v. Dover Area School District 2005), but clearly makes claims about biology, within the scientific domain (see Kitcher 1982). Similarly, the scientific domain may not be limited to those claims which can be tested by scientific methodologies. Often, claims beyond this domain are made based on scientific values. The principle of parsimony, or Ockham’s Razor, is often used by scientifically-minded people to reject claims about the existence of unobservable entities—God, angels, faeries, etc. (see e.g. Worrall 2004) The claim that God does not exist is beyond the scope of scientific methods, but may not be beyond the domain of science, or of scientific values. Moreover, not every claim which could be tested scientifically is within the domain of science—metaphorical and poetic claims, for instance, though often empirically testable, arguably do not fall within a scientific domain. Similarly, scientists can be and are interested in claims beyond the scope of science or scientific method, even within their roles as scientists. For example, scientists and scientific communities regularly confront and consider ethical, legal and policy questions within their capacity as scientists (Douglas 2009; Daubert v. Merrell Dow Pharmaceuticals 1993).

Finally, not everyone who upholds scientific values and has scientific psychological traits is a scientist. I identify with most scientific values and psychological traits, but I am no scientist. Nor does every scientist exhibit precisely the values and psychological processes identified as scientific. They may fall short of these ideals, or uphold other values, or use different processes without being necessarily non-scientists. Not every work, theory or methodology exhibits scientific values, or is propounded or produced by those who do. There may be multiple clusters of scientific values and traits, and different ways of thinking about and doing science.

I hope I have succeeded in undermining at least some of the regularly presumed links between scientific concepts, and thus between demarcation questions. I do not wish to argue that there are no such links—we may be able to find some inter-definable concepts, or demarcation questions which can be adequately solved with reference to demarcation criterion from a number of other questions. All I wish to demonstrate at this stage is that foundational demarcation makes very strong assumptions which have not been defended, and which we should treat sceptically. Foundational demarcation is an implausible programme, and has met its “demise” alongside simple demarcation. With that in mind, pluralist demarcation looks appealing: we can address each demarcation problem individually, using the considerations which are appropriate to that problem to construct adequate criteria. These criteria need not be applicable to other demarcation problems beyond the ones they were crafted to address (though some may well be).

 

1.4: Plurality of Contexts

 

Even given a plurality of criteria and a plurality of questions, there is still a strong rival position to pluralist demarcation. A complex monistic position could be developed, offering a single demarcation criterion in response to each demarcation question. This gives a plurality of criteria, but each applicable to specific subject-matter, without overlap between the criteria. In effect, we have a complex non-overlapping monism; several questions with singular answers, combined. Pluralist demarcation is distinct from this complex monism. Pluralism holds that not only are there a plurality of demarcation questions with distinct answers, but there are a number of different demarcation criteria which can be appropriate to the same demarcation questions, depending on context. To show this further thesis, we must go beyond a plurality of demarcation questions, and consider the variation in contexts of demarcation questions, and how this affects the considerations which are appropriate to a demarcation criterion.

To demonstrate that the plurality of contexts for demarcation questions supports pluralism, we must show (a) there are a plurality of contexts, and (b) that different considerations are appropriate in forming demarcation criteria depending upon context. It will suffice to show a range of different contexts, rather than to attempt an exhaustive account of all the many contexts in which demarcation questions can be asked.

Clearly, demarcation questions may be asked in abstract, philosophical contexts, or by practicing scientists wishing to clarify their thought or self-image. The abstract philosophical context may make quite idealised considerations appropriate—we look for ideal scientific methods, qualities and values. We are aware that science in practice falls short of this ideal, but this would make a misguided criticism of such idealised demarcation projects. By contrast, the practicing scientist may have more pragmatic concerns in forming demarcation criteria—they want to know how to tell the scientists from the non-scientists they meet, to separate the scientific works they must give their professional attention to from the non-scientific which they can set aside. Ideals will be comparatively inappropriate in this context[11]—social norms, behaviours, conventions and attitudes, as well as epistemological criteria, are more appropriate considerations.

Demarcation also has great importance beyond the academic setting. In the courts, demarcation is essential because of the privileged position of scientific evidence and witnesses. The Daubert standard (Daubert v. Merrell Dow Pharmaceuticals 1993) presents a composite demarcation criterion for demarcating scientific from non-scientific evidence in the USA. Daubert equates scientific evidence with evidence gained by scientific methods, presenting a primarily epistemological criterion. However, Daubert also adopts, from the previous Frye standard (Frye v. United States 1923), a sociological component of community consensus—Frye relies on ‘general acceptance’ within the scientific community as the test of scientific status. Community acceptance is salient and appropriate in a legal context because of the need for consensus and absence of reasonable doubt. Moreover, in the legal context, ‘science’ usually means well-established, confirmed theory and method. In contrast to the academic context, in which science also includes formulating and working with novel, untested theories and methods, the legal context views ‘science’ as the scientific canon—a body of knowledge confirmed through established methods.[12]

A similar demarcation is made in the policymaking context. When policymakers require scientific advice, or implement recommendations based on scientific theories and work, they usually intend ‘science’ as well-established and well-confirmed bodies of theory and work. Their concern is largely one of reliability. In policy, ‘scientific method’ is often equated with ‘reliable method’ (see Douglas 2009). So, epistemic values are appropriate components of demarcation criteria in this context. Community consensus, by contrast, is less significant—dissenting voices are still considered scientific if backed by well-confirmed reliable work.[13] However, the context is very different when appointing a scientific advisor—an advisor’s own work does not need to be highly-confirmed to qualify her for the role, rather socio-psychological factors are most appropriate.

By contrast, the context of research funding and policy-setting for resource distribution does not demarcate science as a well-confirmed body of theories. The demarcation problem here is not which projects deserve funding, but which qualify for consideration for resources earmarked for ‘science’ at all. Funding earmarked for scientific research is distributed to both well-confirmed and relatively-unconfirmed theories. Clearly, satisfying some epistemological criteria is necessary; if the research does not utilise a scientific methodology it cannot qualify (e.g. literary analysis or textual exegesis). Degree of confirmation, however, is an inappropriate consideration.

One particularly lively demarcation debate surrounds educational policy and curriculum-setting. In the U.S.A., creation-science advocates have repeatedly argued that creation-science should be taught in biology classes alongside evolutionary theory, or alternatively that evolutionary theory is non-scientific, and must be excluded from classrooms. Precisely which factors are appropriate to demarcation in this case is a hot issue (see e.g. Kitcher 1982; Ruse 1984). Following Popper (1976), some creationists have argued that only falsifiability is appropriate (Kitcher 1982, 39-44). Degree of confirmation and epistemic values are irrelevant. Even if creation-science is rejected in the scientific community, exhibits none of the epistemic virtues of scientific theories and has a number of empirically-discoverable unscientific failings[14], the fact that its subject-matter is empirical entails that it is scientific, and so should be taught (Reisch 1998, 342). Realistically, this just acts as a practical reductio of the claim that falsifiability is the appropriate criterion here. Furthermore, the limited curriculum space and the non-specialisation of the audience suggests a particular demarcation; we want to select those theories and methods which exhibit particular epistemological criteria – simplicity, scope, fecundity and empirical adequacy within the domains school-children are likely to encounter. Furthermore, a high degree of confirmation and historical progressiveness are contextually important (Dupré 1993, 242-4).

Curriculum-setting raises a specific kind of demarcation question—questions that in fact address membership of specific fields. Magnus (2009, 6-7) argues that the creation-science debates in educational policy are better understood as whether creationism is part of biology, rather than whether it is scientific. There are specific constraints of subject-matter, methodology and community-norms which identify biological science within the sciences. Biology qua ‘subject-matter’ seems to include creation-science (and any theory about organisms), while biology qua ‘scientific discipline’ includes only theories which satisfy further epistemological and sociological criteria. The relevant sense of ‘biology’ in educational policy is the latter.

Similar conflicts arise in medicine, where there are debates surrounding funding for Complementary and Alternative Medicine (CAM). The CAM debate has two forms; whether homeopathy or acupuncture are scientific and whether they are part of medicine. The former question makes subject-matter and methodological criteria relevant, while the latter invokes confirmation, epistemic virtues and specific community aims. It may be sufficient, for some senses of ‘science’, to make empirical claims and test them scientifically, as some homeopathic researchers do. However, the context of funding debates in medical science makes epistemological criteria—confirmation, coherence with the scientific canon, and plausibility—extremely important.

We could discuss further contexts in which demarcation questions are raised, and draw further contextual distinctions within the contexts discussed above[15]. However, the point is clear: different considerations are appropriate as components of demarcation criteria to different extents depending upon the context of the demarcation question.

 

We have seen three strands converge in a favour of pluralist demarcation. There are a range of criteria, each appropriate to some demarcation questions and inappropriate to others. There are a range of demarcation questions, each of which may require different responses to others. Moreover, there are a range of contexts in which these demarcation questions are asked, influencing whether specific components are appropriate to the construction of an adequate demarcation criterion. Pluralist demarcation is the claim that there are many demarcation criteria which are appropriate to answering demarcation questions, and the appropriate criterion must be determined with reference to the question being asked, and the context of that question.

 

2: Alternatives to Pluralist Demarcation

 

I have outlined the arguments against simple, foundational and complex monist demarcation. However, there are other positions available. Pluralist demarcation is not the only position which remains plausible despite the variation presented in Section 1. In this section, I discuss these contenders and argue that each is less appealing than pluralist demarcation.

 

2.1: ‘Low-Key’ Pluralism

 

Godfrey-Smith (2010) coined the term ‘Low-Key Pluralism’ in the causal pluralism debate, to describe positions in which we simply identify existing concepts/criteria and use these in the correct context.[16] Low-key pluralism in the demarcation debate would hold that we already have adequate demarcation criteria for each demarcation question and context, and so once the plurality of questions and contexts is recognised, our work is done. Pluralism is sometimes accused of being a “cop-out” position—low-key pluralism is the position such criticisms address.

Besides the accusation of a cop-out, low-key pluralism is implausible in the demarcation debate. One need only consider the Daubert standard (Daubert v. Merrell Dow Pharmaceuticals 1993) to see that the demarcation criteria used in practice are lacking and would benefit from philosophical attention[17]—they may use factors which are contextually inappropriate, misunderstood, or vague. According to pluralist demarcation, the philosophical work of responding to demarcation problems is only beginning. We face a new horde of specific problems, with real-world implications, which require philosophical scrutiny to ensure coherence, appropriateness and specificity.[18]

There is a similar position to low-key pluralism, which we could call ‘intuitive demarcation’, in which we simply use our existing intuitions, informed by the context and the specific demarcation question, to separate science from non-science. That is, science in a specific context is whatever we conventionally call science in that context. This approach conflates the descriptive and normative questions of demarcation, eliding what we do call science with what we should call science. Demarcation is not an idle definitional matter—status, legal admissibility, educational standards, policy decisions and resource allocation all depend sensitively upon our answers. Moreover, as controversies such as creation science and CAM show, there is a lack of consensus. Finally, to rely upon scientific consensus to define science will beg the question against those traditionally excluded from scientific communities (Feyerabend 1999). Whenever scientific status conveys epistemic weight, authority and privilege, the question is non-trivial and a principled, well-analysed response to each demarcation question is desirable.

 

2.2: Network Demarcation

 

Reisch’s Network Demarcation (1998) offers a more nuanced version of the low-key account. Per Reisch, science consists of a network of theories and fields, and the unity of that network is paramount. Demarcation problems, for Reisch, are problems of integration into the network. If a field, theory, work, etc., cannot be integrated without disrupting the network and damaging its problem-solving abilities, it is unscientific. For Reisch, we know which network is the scientific one, and the disputes discussed above are merely disputes about the practicalities of integration into the network, not our understanding of what science is.

First, we should note that a network approach, while a reasonable way of demarcating fields and theories, seems to fail to demarcate scientists from non-scientists, or provide answers to questions about attitudes, values, methods and methodologies. In other words, Network Demarcation is at best part of a solution to demarcation problems. Second, Reisch makes a strong assumption about the unity of science, which may be unsustainable given an appreciation of the disunity of scientific communities and the many tensions and incompatibilities both between and within scientific disciplines. There may rather be many scientific networks. If that is the case, Network Demarcation will fail unless it can provide specific criteria by which whole networks are judged scientific or non-scientific—otherwise, Reisch’s cause-celebre, Intelligent Design, could set up a separate creation science network, and would be classed as scientific. So, Network Demarcation offers us a new demarcation question (“What is a scientific network?”) rather than a general solution or a defence of intuitive demarcation.

Finally, we should consider the relationship between pluralist demarcation and the unity of science. From the above remarks, it may seem that pluralist demarcation slips in disunity of science via the backdoor. In response, we should first distinguish between unity and unification. Unification, the creation of links between fields, is compatible with disunited science. While unity proposes that the goal of science should be to create a single consistent whole, unification aims to forge links between fields to answer complex scientific questions. Integrative pluralism, such as Brigandt’s (2010), argues against unity of science as an aim, but agrees that unification, in the sense of creating links between scientific fields, is indispensible for sophisticated problem-solving. Following Mitchell (2002), he argues that multiple causal factors are relevant to the analysis of complex phenomena, and so linking analyses from multiple disciplines, across ‘levels of analysis’ is needed. Explanatory unification, at least in Kitcher’s sense (1981), is compatible with disunity of science (Dupré 1993, 227-8), and can be used as part of a demarcation criterion in pluralist demarcation. Moreover, nothing in pluralist demarcation says that science cannot create a single grand theory of everything, or fulfill other such unified goals—rather, it claims that (for instance) whether that grand theory, and other theories, are classed as scientific is dependent upon the context in which the question is asked. Even the grand theory of everything, in its infancy and before resounding confirmation, would be unlikely to count as scientific in a courtroom. But that should not upset or surprise anyone searching for unity—the courtroom is not the context which interests them.

 

2.3: Pragmatic Demarcation

 

Perhaps the most similar position to pluralist demarcation in the existing demarcation literature is Resnik’s “Pragmatic demarcation” (2000). He defends a core of scientific values drawn from a range of concerns. However, these are always insufficient for demarcation because the practical interests and concerns in any practical demarcation must be considered. Following Rudner (1953), Resnik argues that these practical considerations are the consequences of categorising something as science or non-science (2000, 261-4).

Although pragmatic demarcation makes demarcation vary by context, Resnik retains core “themes” of science, appropriate in every context, suggesting: “testability, empirical support, progressiveness, problem-solving ability” (2000, 264). We have already seen that each of these considerations is inappropriate in some contexts. Moreover, these considerations often fail to cross over to other demarcation questions—surely, for instance, to be a scientist, one need not work on empirically supported, progressive theories with problem-solving ability. In reality, much genuinely scientific research is ultimately fruitless, many theories are never confirmed, and many research programmes fail to progress. Pluralist demarcation rejects the core themes, although happily acknowledges that these will often be vital parts of appropriate demarcation criteria for many demarcation questions in many contexts, and that the consequences of demarcation are an important determinant of the appropriateness of considerations in formulating demarcation criteria.

 

 3: Conclusion

 

I have offered a three-part defence of pluralist demarcation: there are a plurality of demarcation criteria available, these criteria (or rather, the considerations comprising these criteria) are appropriate to different demarcation questions to different extents, and contextual factors further affect the appropriateness of different considerations. Pluralist demarcation holds that many demarcation criteria, namely those which are most appropriate to each demarcation question we wish to answer in the context in which it is asked, are needed to adequately respond to demarcation problems. Simple and foundational demarcation strategies fail. Pluralist demarcation is not a cop-out, and leaves considerable philosophical work to be done and challenges to be met. The most appropriate demarcation criterion for each demarcation question, in context, must be formulated—and the appropriateness of different considerations needs clarification and rigorous assessment. As such, pluralism represents an alternative way of addressing the (many) problems of demarcation, not a solution in itself.

 

References:

Ayer, A.J. 1946. Language, Truth, and Logic. New York: Dover.

Brigandt, I. 2010. “Beyond Reduction and Pluralism: Towards an Epistemology of Explanatory Integration in Biology.” Erkenntnis no. 73:295-311.

Collins, H. 1982. The Sociology of Scientific Knowledge. Bath: Bath UP.

Daubert v. Merrell Dow Pharmaceuticals, Inc. 1993. 509 U.S. 579.

Derkson, A.A. 1993. “The Seven Sins of Pseudoscience.” Journal for General Philosophy of Science no. 32:329-50.

———. 2001. “The Seven Strategies of the Sophisticated Pseudoscientist.” Journal for General Philosophy of Science no. 32:329-50.

Douglas, H. 2009. Science, Policy and the Value-Free Ideal. Pittsburgh: Pittsburgh UP.

Duhem, P. 1906. The Aim and Structure of Physical Theory. Princeton: Princeton UP.

Dupré, John. 1993. The disorder of things: metaphysical foundations of the disunity of science. Cambridge, MA: Harvard University Press.

Ernst, E., M. H. Pittler, and B. Wider. 2006. The Desktop Guide to Complementary and Alternative Medicine: An Evidence-Based Approach. Philadelphia: Mosby: Elsevier, Ltd.

Feyerabend, P.K. 1999. “How to be a good empiricist: a plea for tolerance in matters epistemological.” In Paul K. Feyerabend: Knowledge, Science and Relativism. Philosophical Papers, Vol. 3, edited by Preston, J, 78-103. Cambridge: CUP.

Frye v. United States. 1923. 293 F. 1013 D.C. Circ.

Gauquelin, M. 1975. “Spheres of Influence.” Psychology Today no. 7:22-7.

Giere, R. 1988. Explaining Science. Chicago: Chicago UP.

Gieryn, T.F. 1983. “Boundary Work and the Demarcation of Science from Non-Science: Strains and Interests in Professional Ideologies of Scientists.” American Sociological Review no. 48:781-95.

Glymour, C. & Stalker, D.,. 1982. “Winning through Pseudoscience.” In Philosophy of Science and the Occult, edited by Grim, P, 75-86. Albany: State University of New York Press.

Godfrey-Smith, P. 2010. “Causal Pluralism.” In Oxford Handbook of Causation, edited by Beebee, H, Hitchcock, C and Menzies, P, 326-37. Oxford: OUP.

Hansson, S.O. 2008. “Science and Pseudo-science.” Stanford Encyclopedia of Philosophy.

James, E.W. 1982. “On Dismissing Astrology and Other Irrationalities.” In Philosophy of Science and the Occult, edited by Grim, P. Albany: State University of New York Press.

Kitcher, P. 1981. “Explanatory Unification.” Philosophy of Science no. 48 (4):507-31.

———. 1982. Abusing Science: the Case Against Creationism. Boston: MIT Press.

Kitzmiller v. Dover Area School District. 2005. U.S. 400 F. Supp. 2d 707.

Kuhn, T. 1977. “Objectivity, Value Judgement and Theory Choice.” In The Essential Tension, 320-39.

Kumho Tire Company Ltd. et al. v. Patrick Carmichael et al. 1999. 526 U.S. 137.

Lakatos, I. 1968. “Criticism and the Methodology of Scientific Research Programmes.” Proceedings of the Aristotelian Society no. 69:149-86.

———. 1980a. “Falsification and the Methodology of Scientific Research Programmes.” In The Methodology of Scientific Research Programmes: Volume I: Philosophical Papers, edited by Worrall, J. Cambridge: CUP.

———. 1980b. “Science and Pseudoscience.” In The Methodology of Scientific Research Programmes: Volume I: Philosophical Papers, edited by Worrall, J. Cambridge: CUP.

Laudan, Larry. 1983. “The demise of the demarcation problem.” In Physics, philosophy and psychoanalysis, edited by R S Cohen and Larry Laudan, 111-127. Springer.

Lugg, A. 1987. “Bunkum, Flim-Flam and Quackery: Pseudoscience as a Philosophical Problem.” Dialectica no. 41:221-30.

———. 1995. “Pseudoscience as structurally flawed practice: a reply to A.A. Derkson.” Journal for General Philosophy of Science no. 26:323-6.

Magnus, P.D. 2009. “What can Species tell us about Theories?” draft, retrieved from http://www.fecundity.com/job/species-analogy.pdf (04/01/11).

McLean v. Arkansas Board of Education. 1981. 529 F. Supp. 1255 1258-1264.

McMullin, E. 1982. “Values in Science.” Proceedings of the 1982 Biennial Meeting of the Philosophy of Science Association no. 2:3-28.

Merton, R.K. 1973. “The Normative Structure of Science.” In The Sociology of Science: Theoretical and Empirical Investigations, 267-78. Chicago: Chicago UP.

Mitchell, S. 2002. “Integrative Pluralism.” Biology and Philosophy no. 17:55-70.

Popper, K.R. 1963. Conjectures and refutations : the growth of scientific knowledge. London: Routledge & Kegan Paul.

Popper, Karl R. 1976. “Darwinism as a Metaphysical Research Program.” In But is it Science?, edited by M Ruse, 144-55. Prometheus Books.

Reisch, G.R. 1998. “Pluralism, Logical Empiricism and the Problem of Pseudoscience.” Philosophy of Science no. 65 (2):333-48.

Resnik, D.B. 2000. “A Pragmatic Approach to the Demarcation Problem.” Studies in the History and Philosophy of Science no. 31 (2):249-67.

Rudner, R. 1953. “The Scientist Qua Scientist Makes Value Judgments.” Philosophy of Science no. 20:1-6.

Ruse, M. 1984. But is it Science?: Prometheus Books.

Thagard, P. 1978. “Why Astrology is a Pseudoscience.” In Proceedings of the Philosophy of Science Association, edited by Asquith, P and Hacking, I, 223-34. East Lansing: Philosophy of Science Association.

———. 1988. Computational Philosophy of Science. Boston: MIT Press.

Worrall, J. & Ratzsch, D.,. 2004. “Does Science Discredit Religion?” In Contemporary Debates in Philosophy of Religion, edited by Peterson, M L and Vanarragon, R J, 59-94. Oxford: Blackwell.

 

Footnotes:

[1] Moreover, the terms have a very different loading. Pseudoscience is almost always pejorative, while ‘non-science’ is, in some contexts, neutral or positive.

[2] Consider, for instance, clinical epidemiologists testing the claims of homeopathy in randomised controlled trials (see e.g. Ernst, Pittler, and Wider 2006).

[3] Such a project is an example of ‘Foundational demarcation’, for which see Section 1.3.

[4] Sometimes, this would be an appropriate judgment, but in other circumstances not—see Section 1.4.

[5] Derkson (2001) considers thought-processes which overvalue co-incidence to be distinctively non-scientific.

[6] For example, Derkson and Lugg concentrate upon showing that the other’s criterion makes certain miscategorisations of sciences as non-sciences and vice versa.

[7] If this seems unfair to technicians, consider the legal context; in the U.S.A., for instance, the Carmichael ruling distinguishes between scientists and technicians in the context of expert scientific testimony (Kumho Tire Company Ltd. et al. v. Patrick Carmichael et al. 1999).

[8] The court of public opinion is one non-scientific “test” of a scientific theory. Another is historical precedent. Some methods might count as scientific even if they do not or cannot be used to test scientific theories, for instance due to impracticality. Not every scientific method can be used to test every scientific theory, many can be used to test non-scientific theories, and scientific theories are often subjected to non-scientific tests.

[9] Of course, the same arguments can be made from the opposite direction. Equally, we could argue that not every non-scientific field or research programme is devoid of scientific theories, uses exclusively non-scientific methods, etc.

[10] This latter point is particularly clear when we contrast Merton’s (1973) values of scientific communities with the epistemic values listed by Kuhn (1977) and McMullin (1982). The values used in a community’s work are not necessarily the same as the values of the working community.

[11] Although we may try to demarcate scientists in practice by their aspiration towards such ideals, this runs the risk of picking out only idealistic, uncynical scientists (and well-intentioned pseudoscientists).

[12] Furthermore, although omitted from Daubert, replicability is an important component of scientific status in the legal context, but not necessarily in other contexts.

[13] A similar demarcation criterion would probably be appropriate to a less controversial demarcation—that of the subject-matter of scientific/non-scientific journalism.

[14] For instance, the consistent misunderstanding of the 2nd Law of Thermodynamics (Kitcher 1982, 89-96).

[15] Consider, for instance, the variation in what counts as ‘science’ for curriculum-setting according to the age and level of the students.

[16] In causal pluralism, for example, low-key pluralism would hold that we already have several concepts of causation; we just need to recognise this plurality and use them.

[17] If Daubert still seems reasonable, consider the multi-condition criterion presented by Judge Overton in the McLean v. Arkansas “creation science” trial (McLean v. Arkansas Board of Education 1981).

[18] Anyone who accuses pluralist demarcation of copping out of the ‘big question’ of simple demarcation would need to provide a reasonable alternative position which takes on that big question. Pluralist demarcation holds the position that the question is unanswerable because insufficient information is provided.