Mechanism of Scientific Change

What is the mechanism of scientific change? How do epistemic agents take stances towards towards epistemic elements? How do changes in a scientific mosaic take place?

Along with the question of the ontology of scientific change, this question is one of the two most general questions of scientonomy. It is safe to say that any general theory that attempts to explain changes in a certain domain normally has two major ingredients – an ontology of the entities and relations that populate the domain as well as some dynamics of how these entities and relations change over time. Scientonomy is no exception as it attempts to understand what sort of elements, agents, and stances, play part in the process of scientific change (ontology) and what the mechanism of these changes is (dynamics). Answering both of these questions is crucial for our understanding of the process of scientific change.

In the scientonomic context, this question was first formulated by Hakob Barseghyan in 2015. The question is currently accepted as a legitimate topic for discussion by Scientonomy community.

In Scientonomy, the accepted answers to the question can be summarized as follows:

• An element of the mosaic remains in the mosaic unless replaced by other elements.
• If a theory satisfies the acceptance criteria of the method employed at the time, it becomes accepted into the mosaic; if it does not, it remains unaccepted; if assessment is inconclusive, the theory can be accepted or not accepted.
• A method becomes employed only when it is deducible from some subset of other employed methods and accepted theories of the time.
• If a pair of elements satisfies the compatibility criteria employed at the time, it becomes compatible within the mosaic; if it does not, it is deemed incompatible; and if assessment is inconclusive, the pair can become compatible, incompatible, or its status may be unknown.
• A theory becomes rejected only when other theories that are incompatible with the theory become accepted.
• A method ceases to be employed only when other methods that are incompatible with the method become employed.
• A methodology can shape employed methods, but only if its requirements implement abstract requirements of some other employed method.
• If an accepted theory is taken as the final truth, it will always remain accepted; no new theory on the subject can ever be accepted.
• Sociocultural factors can affect the process of theory acceptance insofar as it is permitted by the method employed at the time.
• When two mutually incompatible theories satisfy the requirements of the current method, the mosaic necessarily splits in two. When a theory assessment outcome is inconclusive, a mosaic split is possible. When a mosaic split is a result of the acceptance of only one theory, it can only be a result of inconclusive theory assessment.
• Transitions from one state of the mosaic to another are not necessarily deterministic. Scientific change is not a strictly deterministic process. The process of method change is not necessarily deterministic: employed methods are by no means the only possible implementations of abstract requirements. The process of theory change is not necessarily deterministic: there may be cases when both a theory's acceptance and its unacceptance are equally possible.

Broader History

Ludwik Fleck

Ludwik Fleck, an epidemiologist, made one of the earliest attempts to understand scientific change as a social process, and to develop a conceptual framework for understanding how scientific communities function.1 His most comprehensive work was Genesis and Development of a Scientific Fact published in 1935.2 For Fleck, cognition was necessarily a collective social activity, since it depends on prior knowledge obtained from other people. New ideas arise within thought collectives, which are groups of people who participate in the mutual exchange of ideas. As an emergent consequence of mutual understandings and misunderstandings within such a group a particular thought style arises. An established thought style carves the social world into an esoteric circle of expert members of the thought collective, and an exoteric circle who are outside the thought collective. How individual members of a thought collective think and perceive within the relevant domain is determined by the thought style. Scientific facts are socially constructed by thought collectives. Reality in itself cannot be known, but the thought style can be compared with reality through observation and experiment, and may be revised or abandoned on the basis of such interactions.2 1 The thought style of a particular collective can, at most, be only partially understood by members of other collectives, and may be completely incommensurable with the thinking of some other collectives.

Thomas Kuhn

Drawing partially on Fleck’s ideas, physicist and historian of science Thomas Kuhn published his ideas about scientific change as The Structure of Scientific Revolutions in 1962.3 Kuhn spoke of scientific paradigms, which are constellations of theoretical and metaphysical beliefs, values, methods, and instrumental techniques shared by a scientific discipline. A paradigm determines which questions are asked of the natural world by observation and experiment. Adherents to a paradigm engage in normal science, which solves the puzzles needed to expand the range of natural phenomena that can be explained using the paradigm. Eventually, anomalies may be unearthed. These are phenomena that recalcitrantly resist explanation in terms of the paradigm. If anomalies persist and grow in number, practitioners seek fundamentally new approaches. If a new approach is successful at resolving salient anomalies and is deemed to hold promise for solving new puzzles, a scientific revolution may result, in which a new paradigm replaces the old. Because paradigms are holistic networks of theories, methods, and values, they are incommensurable with one another, meaning that the terms and categories of the old paradigm cannot be translated into those of the new. Adoption of a new paradigm thus appeared, especially to Kuhn’s critics, to involve a kind of non-rational leap of faith.43

Paul Feyerabend

In his Against Method, published in 1975, philosopher Paul Feyerabend, an epistemic anarchist, launched a much more radical attack on the idea of a fixed scientific method, and on the rationality of science.5 On his account, science does not possess the regularities that would make a science of science and a theory of scientific change possible. Social constructivists likewise favored an historically contingent, relativist, and particularist view of science, which they supposed was incompatible with a coherent theory of scientific change.

Irme Lakatos

Philosopher Irme Lakatos, a proponent of the rationality of science and of a fixed scientific method launched a new account of scientific change with his Methodology of Scientific Research Programs in 1970.6 Lakatos sought to challenge both Kuhn and Feyerabend. He saw interrelated scientific theories as constituting research programs. Unlike Kuhn, he believed that scientific fields typically host multiple competing research programs and rejected the idea of coherent unitary paradigms. Not all theoretical constituents of a research program were assigned equal importance. The hard core of a research program consisted of those theoretical claims that were indispensable to it. Adherents to a research program attempt to explain an increasingly wide range of natural phenomena in terms of the core claims. This is the positive heuristic of the research program. The protective belt consists of those theoretical assumptions that allowed the application of the hard core to an increasing range of cases. Scientists used their ingenuity to protect the hard core by making alterations to the protective belt so as to protect the core from falsification. The protection of the hard core is a research program's negative heuristic. A progressive research program is one that makes successful novel predictions. A degenerating research program is one whose predictions repeatedly fail, and whose protective belt must be altered in an arbitrary, ad hoc fashion to protect the hard core from falsification. Lakatos rejected Kuhn’s distinction between normal and revolutionary science, and supposed that a revolution occurs when scientists simply switch allegiance from a degenerating research program to a progressive one.

Larry Laudan

In his 1984 Science and Values philosopher Larry Laudan accepted growing empirical evidence that the methods of science had changed with time.78 "Our views about the proper procedures for investigating the world", he wrote, "have been significantly affected by our shifting beliefs about how the world works".8p. 39 However he did not accept Feyerabend’s anarchism, or his view that a coherent theory of scientific change was impossible. Laudan proposed a reticulated model of scientific rationality in which other theories, methods, and research aims all interact in the assessment of a theory, with all three subject to alteration or replacement in the light of the others. Like Lakatos, he supposed that scientific theories were linked into logically related groups which he called research traditions, and rejected the radical holism of Kuhnian paradigms. Laudan distinguished between the acceptance of a theory by a scientific community as the best available and pursuit of a theory as holding potential. Similar ideas were adopted as part of the Barseghyan theory of scientific change.9

Scientonomic History

Acceptance Record

Here is the complete acceptance record of this question (it includes all the instances when the question was accepted as a legitimate topic for discussion by a community):

Community	Accepted From	Acceptance Indicators	Still Accepted	Accepted Until	Rejection Indicators
Scientonomy	1 January 2016	This is when the community accepted its first answer to this question, The Theory of Scientific Change, which indicates that the question itself is legitimate.	Yes

All Theories

According to our records, no theory has attempted to answer this question.

If an answer to this question is missing, please click here to add it.

Accepted Theories

According to our records, no theory on this topic has ever been accepted.

Suggested Modifications

According to our records, there have been no suggested modifications on this topic.

Current View

In Scientonomy, the accepted answers to the question are The First Law (Barseghyan-2015), The Second Law (Patton-Overgaard-Barseghyan-2017), The Third Law (Sebastien-2016), The Law of Compatibility (Fraser-Sarwar-2018), Theory Rejection theorem (Barseghyan-2015), Method Rejection theorem (Barseghyan-2015), Methodology Can Shape Method theorem (Barseghyan-2015), Dogmatism No Theory Change theorem (Barseghyan-2015), Sociocultural Factors in Theory Acceptance theorem (Barseghyan-2015), Necessary Mosaic Split theorem (Barseghyan-2015), Possible Mosaic Split theorem (Barseghyan-2015), Split Due to Inconclusiveness theorem (Barseghyan-2015), Underdetermined Method Change theorem (Barseghyan-2015), Underdetermined Theory Change theorem (Barseghyan-2015) and Scientific Underdeterminism theorem (Barseghyan-2015).

Mechanism of Scientific Inertia for Epistemic Elements

The First Law (Barseghyan-2015) states: "An element of the mosaic remains in the mosaic unless replaced by other elements."

The following passage from The Laws of Scientific Change summarizes the gist of the law:

According to the first law, any element of the mosaic of accepted theories and employed methods remains in the mosaic except insofar as it is overthrown by another element or elements. Basically, the law assumes that there is certain inertia in the scientific mosaic: once in the mosaic, elements remain in the mosaic until they get replaced by other elements. It is reasonable therefore to call it the law of scientific inertia.9p. 123

The First Law for Theories

An accepted theory is not rejected unless there is a suitable replacement, even though sometimes that replacement may simply be the negation of the theory. For example, Issac Newton's theory of universal gravitation produced small errors in predicting the movements of the planet Mercury.9p. 125 Throughout the eighteenth and early nineteenth century, it was noted that predictions of the time when the disk of Mercury would appear in transit across the sun's disk were off, sometimes by hours, or even as much as a day. These anomalies caught the attention of the French mathematician Urbain Jean Joseph Leverrier, who proposed an explanation consistent with Newton's theory in 1859. Mercury, he supposed, was being perturbed by the gravitational pull of an unknown planet orbiting closer to the sun. The hypothetical planet, named Vulcan, was searched for, but never found.10 Newton's theory had other predictive failures as well, but these did not lead to the rejection of the theory. It was not rejected until after 1915, when Albert Einstein showed that Mercury's movements could be explained by his new theory of gravity, the general theory of relativity.11

The First Law for Methods

Formulated for methods, the first law states that the implicit expectations employed in theory assessment will continue to be employed until they are replaced by some alternate expectations.

Mechanism of Theory Acceptance

The Second Law (Patton-Overgaard-Barseghyan-2017) states: "If a theory satisfies the acceptance criteria of the method employed at the time, it becomes accepted into the mosaic; if it does not, it remains unaccepted; if assessment is inconclusive, the theory can be accepted or not accepted."

According to this formulation of the second law, if a theory satisfies the acceptance criteria of the method actually employed at the time, then it becomes accepted into the mosaic; if it does not, it remains unaccepted; if it is inconclusive whether the theory satisfies the method, the theory can be accepted or not accepted.

Unlike the previous formulation of the second law, this formulation makes the causal connection between theory assessment outcomes and cases of theory acceptance/unacceptance explicit. In particular, it specifies what happens to a theory in terms of its acceptance/unacceptance when a certain assessment outcome obtains.

In addition, this new formulation is clearly not a tautology because it forbids certain logically possible scenarios, such as a theory satisfying the method of the time yet remaining unaccepted.

Mechanism of Norm Employment

The Third Law (Sebastien-2016) states: "A method becomes employed only when it is deducible from some subset of other employed methods and accepted theories of the time."

The initial formulation of the law, proposed by Barseghyan in The Laws of Scientific Change, stated that a method becomes employed only when it is deducible from other employed methods and accepted theories of the time.9p.132 In that formulation, it wasn't clear whether employed methods follow from all or only some of the accepted theories and employed methods of the time. This led to a logical paradox which this reformulation attempts to solve.12

This reformulation of the law makes explicit that an employed method need not necessarily follow from all other employed methods and accepted theories but only from some of them. This made it possible for an employed method to be logically inconsistent and yet compatible with openly accepted methodological dicta.

In all other respects, this formulation preserves the gist of Barseghyan's original formulation. According to the third law, a method becomes employed when:

1. it strictly follows from some subset of other employed methods and accepted theories, or
2. it implements some abstract requirements of other employed methods.

This restates Barseghyan's original suggestion that accepted theories shape the set of implicit criteria employed in theory assessment. When a new theory is accepted, this often leads to the employment of an abstract requirement to take that new theory into account when testing relevant contender theories. This abstract requirement is then specified by a new employed method.

The evolution of the drug trial methods is an example of the third law in action. For example, the discovery of the placebo effect in drug testing demonstrates that fake treatment can cause improvement in patient symptoms. As a result of its discovery the abstract requirement of “when assessing a drug’s efficacy, the possible placebo effect must be taken into account” was generated. This abstract requirement is, by definition, an accepted theory which stipulates that, if ignored, substantial doubt would be cast on any trial. As a result of this new theory, the Single-Blind Trial method was devised. The currently employed method in drug testing is the Double-Blind Trial, a method which specifies all of the abstract requirements of its predecessors. It is an apt illustration of how new methods are generated through the acceptance of new theories, as well as how new methods employ the abstract requirements of their predecessors.9pp. 132-152

In Barseghyan’s explication of the Aristotelian-Medieval method, he illustrates how Aristotelian natural philosophy impacted the method of the time. One of the key features of the Aristotelian-scholastic method was the requirement of intuition schooled by experience, i.e. that a proposition is acceptable if it grasps the nature of a thing though intuition schooled by experience. The requirement itself was a deductive consequence of several assumptions accepted at the time. One of the assumptions underlying this requirement was the idea that every natural thing has a nature, a substantial quality that makes a thing what it is (e.g. a human's nature is their capacity of reason). Another assumption underlying the requirement was the idea that nature of a thing can be grasped intuitively by those who are most experienced with the things of that type. The requirements of the intuitive truth followed from these assumptions. The scholastic-Aristotelians scholars wouldn’t require intuitive truths grasped by an experienced person if they didn’t believe that things have natures that could be grasped intuitively by experts.

The third law has also proven useful in explicating such requirements as Confirmed Novel Predictions (CNP). According to the hypothetico-deductive method, a theory which challenges our accepted ontology must provide CNP in order to become accepted. However, the history of CNP has been a point of confusion for some time. By the Third Law, one can show that the requirement of CNP has not always been expected of new theories. When Newton published his Principia, CNP were not a requirement of his professed method, yet they were still provided. On the other hand, Clark’s law of diminishing returns had no such predictions. This is because Newton’s proposal of unobservable entities, such as gravity and absolute space, challenged the accepted ontology of the time, while Clark’s simply accounted for the data already available. Thus, in utilizing the Third Law, one can discover both when certain criteria become an implicit rule and under what conditions they are necessary.

Mechanism of Compatibility

The Law of Compatibility (Fraser-Sarwar-2018) states: "If a pair of elements satisfies the compatibility criteria employed at the time, it becomes compatible within the mosaic; if it does not, it is deemed incompatible; and if assessment is inconclusive, the pair can become compatible, incompatible, or its status may be unknown."

The law of compatibility links the compatibility criteria with various assessment outcomes. If compatibility is defined as the ability of a pair of elements to co-exist in the same mosaic, then the assessment for compatibility is essentially the process by which the epistemic agent decides whether any given pair of elements (i.e. theories, questions, methods) can be simultaneously part of their mosaic. Such an assessment can yield three possible outcomes - satisfied, not satisfied, and inconclusive.13p. 73 Accordingly, the law of compatibility states that if a pair of elements does satisfy the compatibility criteria of the time, then it is deemed as compatible. If, however, an element is taken to be incompatible with the other one, then the pair is deemed as incompatible. Finally, the assessment of compatibility may be inconclusive. In this case, the pair may be deemed compatible, incompatible, or its status may remain unknown. The diagram below summarizes the relation between assessment outcomes and their effects:

Mechanism of Theory Rejection

Theory Rejection theorem (Barseghyan-2015) states: "A theory becomes rejected only when other theories that are incompatible with the theory become accepted."

According to the theory rejection theorem, a theory becomes rejected only when other theories that are incompatible with the theory become accepted.

Implicit in the theorem is the idea that each theory is assessed on an "individual basis by its compatibility with the propositions of the newly accepted theory".9p. 168 If it turns out that a previously accepted theory is compatible with the newly accepted theory, it remain in the agent's mosaic.

Barseghyan notes that, although we normally expect a theory to be replaced by another theory in the same "field" of inquiry, this is not necessarily the case. For example, he writes, "HSC knows several cases where an accepted theory became rejected simply because it wasn’t compatible with new accepted theories of some other fields".9p. 171

Barseghyan summarizes the theory rejection theorem as such:

In short, when the axioms of a theory are replaced by another theory, some of the theorems may nevertheless manage to stay in the mosaic, provided that they are compatible with the newly accepted theory. This is essentially what the theory rejection theorem tells us. Thus, if someday our currently accepted general relativity gets replaced by some new theory, the theories that followed from general relativity, such as the theory of black holes, may nevertheless manage to remain in the mosaic. 9p. 171

Mechanism of Method Rejection

Method Rejection theorem (Barseghyan-2015) states: "A method ceases to be employed only when other methods that are incompatible with the method become employed."

According to the method rejection theorem, a method ceases to be employed only when other methods that are incompatible with it become employed.

Role of Methodology in Scientific Change

Methodology Can Shape Method theorem (Barseghyan-2015) states: "A methodology can shape employed methods, but only if its requirements implement abstract requirements of some other employed method."

A methodology can affect an employed method when it implements one or more abstract requirements of another employed method. Thus, the role normative methodology plays in the process of scientific change is a creative role, in which methods are changed through the implementation of other abstract requirements from some other employed method.

This theorem follows from former description of the third law, which states that a method becomes employed only when it is deducible from other employed methods and accepted theories of the time.

This description of the third law leaves room for methodologies’ to play an active role in scientific change in cases when a concrete method fulfills the requirements of an employed abstract method. The same abstract requirements can usually be implemented in a wide range of different ways. For instance, if there is a whole array of concrete cell counting methods all implementing the same abstract requirement that when counting the number of cells, the resulting value is acceptable only if it is obtained with an "aided" eye.9pp. 151-152 In such cases, methodology can play a decisive role in method employment; what later becomes the requirements of the employed method can be first suggested as a methodology. Thus, the double-blind trial method was first devised as a methodology, as a set of explicitly stated rules, and only after that did it become actually employed as a method of drug testing.9pp. 240-243

Sebastien (2016)'s new definition of methodology offers an alternate means for methodologies to shape methods, although this is not stated in the existing formulation of the theory. Because methodology is understood as a subkind of Normative Theory, it should be possible for the Third Law to deduce an abstract method from a set of theories including some of the normative methodologies a community holds about their method. In this way, it would be necessary for this method to take into account how the community believes its method works, in any concrete implementation of said method, just as a community takes into account descriptive theories (e.g. the placebo effect and experimenter's bias) when employing a new method.

Changeability of the Scientific Mosaic

Dogmatism No Theory Change theorem (Barseghyan-2015) states: "If an accepted theory is taken as the final truth, it will always remain accepted; no new theory on the subject can ever be accepted."

No theory acceptance may take place in a genuinely dogmatic community. "Namely," as is noted in Barseghyan (2015), Barseghyan notes, when introducing the theory rejection theorem in Barseghyan (2015), "theory change is impossible in cases where a currently accepted theory is considered as revealing the final and absolute truth".9p. 165

Role of Sociocultural Factors in Scientific Change

Sociocultural Factors in Theory Acceptance theorem (Barseghyan-2015) states: "Sociocultural factors can affect the process of theory acceptance insofar as it is permitted by the method employed at the time."

Sociocultural factors can impact the process of a theory's acceptance when the employed method of the community allows for such factors to affect the process. This is derived by the Second Law alone. For example, a community which ascribes infallible power to a leader or a group of leaders is in a position to accept a theory in virtue of the leaders. Furthermore, such factors can guide a scientific community to reject a theory based on the acceptance of another social theory with which it is at odds.

Barseghyan’s Laws of Scientific Change break from the traditional language used in philosophy of science, of internal versus external factors in the mosaic. External factors, a term that has traditionally referred to the influences of societal trends, politics, religion, and so on, if defined as “elements not included in the mosaic” then we must accept that these do not affect the mosaic at the time by the the very definition. This is the result of the fact that the 2nd law introduces new theories in the context of the accepted methods at the time. As a result, the language of “external” factors is problematic.9

Socio-cultural factors ought to be defined more explicitly. The question is, instead, whether factors such as economics, politics, and religion can influence the theories accepted in the mosaic. It follows from the Second Law that theories are assessed by the method in the mosaic at the time. Therefore, if the method at the time mandates economic, political, religious, or other social requirements to be met by a theory before it is accept, only then do socio-cultural factors influence theory acceptance.

Barseghyan provides the example of a hypothetical religious community, with an accepted belief (i.e. theory) that holds that the religion’s High Priest always grasps the true essence of things. By the Third Law, a method may be employed the mosaic that states that any proposition is acceptable, given that the High Priest utters it. In this case, it would appear as though socio-cultural factors are influencing, if not dictating, the course of scientific change in the community. This should not be confused with a case where a High Priest or other elite enforces their beliefs unscientifically, through threats, bribery, or otherwise. Should this happen, the change would be unscientific, as it would violate either the method employed at the time (and thereby the Second Law), or it would be creating a method in the mosaic which does not follow from the accepted theories at the time (and thereby the Third Law).9

Mechanism of Mosaic Split

Necessary Mosaic Split theorem (Barseghyan-2015) states: "When two mutually incompatible theories satisfy the requirements of the current method, the mosaic necessarily splits in two."

Necessary mosaic split is a form of mosaic split that must happen if it is ever the case that two incompatible theories both become accepted under the employed method of the time. Since the theories are incompatible, under the zeroth law, they cannot be accepted into the same mosaic, and a mosaic split must then occur, as a matter of logical necessity.9pp. 204-207

As shown in the figure above, the necessary mosaic split theorem follows as a deductive consequence of the second law and the zeroth law. Per the zeroth law, two incompatible elements cannot simultaneously remain in a mosaic, and per the second law any theory that satisfies the method of the time (and the assessment of the theory by the method is not inconclusive) is accepted into the mosaic. This creates the apparently contradictory situation where either of the two theories A) must be accepted because it satisfies the employed method and B) must not be accepted because it in not compatible with another accepted theory.

The necessary mosaic split theorem is thus required to escape the contradiction entailed by the acceptance of two or more incompatible theories. In a situation where this sort of contradiction obtains the mosaic is split and distinct communities are formed each of which bears its own mosaic, and each mosaic will include exactly one of the theories being assessed. By the third law, each mosaic will also have a distinct method that precludes the acceptance of the other contender theory.

Two examples are helpful for demonstrating mosaic split, one formal example and one historical example. Suppose we have some community C' with mosaic M' and that this community assesses two theories, T₁ and T₂, both of which satisfy M'. Let us further suppose that T₁ and T₂ both describe the same object and are incompatible with one another. According to the second law both T₁ and T₂ will be accepted because they both satisfy M', but both cannot simultaneously be accepted by C' due to the zeroth law. The necessary mosaic split theorem says that the result will be a new community C₁ which accepts T₁ and M₁, which precludes their accepting T₂. Simultaneously a new community C₂ will emerge which accepts T₂ and the resulting theory M₂, which precludes their accepting T₁.

Barseghyan illustrates the necessary mosaic split theorem with the example of the French and English physics communities circa 1730, at which time the French accepted the Cartesian physics and the English accepted the Newtonian physics.9p. 203 These communities would both initially accepted the Aristotelian-medieval physics due to their mutual acceptance of the Aristotelian-medieval mosaic until the start of the eighteenth century9p. 210 but clearly had different mosaics within a few decades. According to the second law both the Cartesian and Newtonian physics must have satisfied the methods of the Aristotelian-medieval mosaic in order to have been accepted, but since both shared the same object and posited radically different ontologies they were incompatible with one another and could not both be accepted, per the second law. The necessary result was that the unified Aristotelian-medieval community split and the resulting French and English communities emerged, each with a distinct mosaic.

Possible Mosaic Split theorem (Barseghyan-2015) states: "When a theory assessment outcome is inconclusive, a mosaic split is possible."

Possible mosaic split is a form of mosaic split that can happen if it is ever the case that theory assessment reaches an inconclusive result. In this case, a mosaic split can, but need not necessarily, result.9pp. 208-213

As pictured, the possible mosaic split theorem follows as a deductive consequence of the second and zeroth laws, given a situation a situation where the assessment of two theories obtains an inconclusive result. This will happen when it is unclear whether or not a theory satisfies the employed method of the community. We can easily imagine such a scenario: suppose we have a method for assessing theories about the efficacy of new pharmaceuticals that says "accept that the pharmaceutical is effective only if a clinically significant result is obtained in a sufficient number of randomized controlled trials." The wording of the method is such that it requires a significant degree of judgement on the part of the community - what constitutes 'clinical significance' and a 'sufficient number' of trials will vary from person to person and by context. This introduces the possibility of mosaic split when it is unclear if two contender theories satisfy this requirement.

Carrying on the above example, suppose two drugs are being tested for some condition C: drugs A and B. We'll call T₁ the theory that A is more effective than B at treating condition C and T₂ the theory that B is more effective than A at treating condition C. These two theories are not compatible, and so cannot both be elements of the mosaic according to the zeroth law. Suppose further that both are assessed by the method of the time, meaning that both are subject to double blind trials. In these trials drug A is clearly superior to drug B at inducing clinical remission, but drug B has fewer side effects and is still more effective than a placebo and has had more studies conducted. Even if we accept T₁ we may have reason to suspect that T₂ better satisfies the method. We can interpret this in two ways: by supposing that our assessment shows that we should accept T₁ and that our assessment is inconclusive about T₂ or by taking both assessments to be inconclusive. In the first case it is permissible according to the second law to accept T₁ and to either accept or reject T₂, and in the second case both may be accepted or rejected.

Because any time an assessment outcome is inconclusive we may either accept or reject the theory being assessed we always face the possibility that one subsection of the community will reject the theory and another subsection will accept it. In these cases the two communities now bear distinct mosaics and a mosaic split has occurred. However it is important to note that the ambiguity inherent in inconclusive assessments means that it is never entailed that there will be competing subsections of the community. A community may, in the face of an inconclusive assessment, collectively agree to accept or reject the theory being assessed. Thus, in cases with an inconclusive assessment mosaic split is possible but never necessarily entailed by the circumstances.

Split Due to Inconclusiveness theorem (Barseghyan-2015) states: "When a mosaic split is a result of the acceptance of only one theory, it can only be a result of inconclusive theory assessment."

Split due to inconclusiveness can occur when two mutually incompatible theories are accepted simultaneously by the same community.

Determinism vs. Underdeterminism in Scientific Change

Underdetermined Method Change theorem (Barseghyan-2015) states: "The process of method change is not necessarily deterministic: employed methods are by no means the only possible implementations of abstract requirements."

The third law allows for two distinct scenarios of method employment. A method may become employed because it follows strictly from accepted theories or employed methods, or it may the abstract requirements of some other employed method. This second scenario allows for creative ingenuity and depends on the technology of the times, therefore it may be fulfilled in many ways and allows underdeterminism 9p. 198.

Underdetermined Theory Change theorem (Barseghyan-2015) states: "The process of theory change is not necessarily deterministic: there may be cases when both a theory's acceptance and its unacceptance are equally possible."

The process of theory assessment under the TSC is underdetermined for two reasons. First, only theories that are constructed are available for assessment. Whether or not a theory is ever constructed is, at least partly a matter of creativity, and is therefore outside the scope of the TSC. Second, it is at least theoretically possible that a process of theory assessment will be inconclusive. This might be because the requirements of the method employed at the time might be vague (e.g. Aristotelian requirements of "intuition schooled by experience").9p. 199-200

Scientific Underdeterminism theorem (Barseghyan-2015) states: "Transitions from one state of the mosaic to another are not necessarily deterministic. Scientific change is not a strictly deterministic process."

Scientific underdetermination is the thesis that the process of scientific change is not deterministic, and science could have evolved differently than it did. Hypothetically, two scientific communities developing separately could experience an entirely different sequence of successive states of their respective scientific mosaics. Even without the TSC, the implausibility of scientific determinism can be seen by considering the process of theory construction, which is outside the present scope of the TSC. Theory construction requires creative imagination, and the formulation of a given theory is therefore not inevitable. Still, underdetermination can also be inferred as a theorem from the axioms of the TSC.9pp. 196-198

This deductive inference occurs as follows. The theorem is the consequence of two related theses: that of underdetermined method change, and that of underdetermined theory change. Since method change and theory change are exactly the transitions in the scientific mosaic, showing that neither method change nor theory change is deterministic is sufficient to imply the Scientific Underdetermination theorem (SUT).

The underdetermination of method change follows from the 3rd Law: “A method becomes employed only when it is deducible from some subset of other employed methods and accepted theories of the time.” As a result of this law, a method can be employed in two ways: either it strictly follows from other accepted theories and employed methods, in which case the change is, in fact, deterministic, or it implements the abstract requirements of some other employed method. In the latter case, the change is underdetermined since abstract requirements can give rise to many different implementations. As a concrete example, we have an accepted theory that states, “when counting the number of living cells, the resulting value is acceptable only if it is obtained with an ‘aided’ eye.” A number of different methods can implement this abstract requirement, like the plating method or the counting chamber method.9p. 198 Thus the method is underdetermined by the abstract requirements, so the process of method change implementing these requirements is not deterministic, which is exactly the statement of the underdetermination of method change.

The underdetermination of theory change comes from the 2nd law and the possibility of inconclusive theory assessment. The 2nd law states that “In order to become accepted into the mosaic, a theory is assessed by the method actually employed at the time”. This assessment can result in conclusive acceptance, conclusive rejection, or it can be inconclusive. In both conclusive cases, the theory change is deterministic, but if the theory assessment is inconclusive, then the theory can be accepted or rejected. So the process of theory change is not necessarily deterministic, since it is, in fact, possible for assessment to be inconclusive. The question of whether there actually exist cases of inconclusive theory assessment is a task for the history of science, and is irrelevant to our discussion.

These two theses combine to form the SUT, since changes in theories and methods are all the transitions that occur in the scientific mosaic, and we have seen that the underdetermination of theory and method change follow deductively from the 2nd law, the 3rd law and the possibility of inconclusive theory assessment.

Related Topics

It has the following sub-topic(s):

• Changeability of the Scientific Mosaic
• Determinism vs. Underdeterminism in Scientific Change
• Mechanism of Compatibility
• Mechanism of Discipline Acceptance
• Mechanism of Discipline Rejection
• Mechanism of Method Rejection
• Mechanism of Mosaic Split
• Mechanism of Norm Employment
• Mechanism of Question Acceptance
• Mechanism of Question Rejection
• Mechanism of Scientific Inertia for Epistemic Elements
• Mechanism of Theory Acceptance
• Mechanism of Theory Demarcation
• Mechanism of Theory Pursuit
• Mechanism of Theory Rejection
• Methods Shaping Theory Construction
• Role of Ethics in Scientific Change
• Role of Methodology in Scientific Change
• Role of Non-Social and Environmental Factors in Scientific Change
• Role of Practical Considerations in Scientific Change
• Role of Sociocultural Factors in Scientific Change
• The Status of Holism and Ripple Effect
• Theories Shaping Core Questions

This topic is also related to the following topic(s):

• Ontology of Scientific Change