The Third Law (Barseghyan-2015)

From Encyclopedia of Scientonomy
Jump to navigation Jump to search

This is an answer to the question Mechanism of Method Employment that states "A method becomes employed only when it is deducible from other employed methods and accepted theories of the time."

The Third Law Barseghyan 2015.png

This version of The Third Law was formulated by Hakob Barseghyan in 2015.1 It is also known as the law of method employment.

Broader History

The basic idea of the third law is not new. A number of philosophers have suggested that our beliefs about the world shape how we engage with the world. Different versions of this idea can be found in the works of Thomas Kuhn, Paul Feyerabend, Dudley Shapere, Larry Laudan, and Ernan McMullin.

Most noteworthy is Larry Laudan’s account of changes in drug trial methods. In his Science and Values, Laudan argued that the discovery of previously unaccounted effects resulted in the formulation of new methods of drug testing.2pp. 38-39 However, while Laudan’s account hints at aspects of the third law, it ultimately conflates methods and methodologies.1pp. 130-131

Ernan McMullin’s accounts of historical methods offer another example of a prototype of the third law. McMullin showed how the hypothetico-deductive method came to replace the Aristotelian Medieval method in the 18th century. In his account, McMullin shows that the employment of the hypothetico-deductivism was a result of accepting that the world is more complex than it appears in our observations.3p. 32-34 These accounts demonstrate how our accepted theories impact our criteria of theory assessment.

There have been many other attempts at explicating the way in which methods change, such as the reconstructions of Plato’s method performed by David Lindberg, or the proposal of synchronous change in paradigm shifts by Thomas Kuhn.

Nevertheless, according to Barseghyan, "what we have had so far is a picture from a bird’s eye perspective. What we lack is the knowledge of the actual mechanism: how exactly can accepted theories shape employed methods?".1p. 133

Scientonomic History

Barseghyan's formulation of the third law was the first attempt to address the problem of method employment in the scientonomic context.

Acceptance Record

Here is the complete acceptance record of this theory:
CommunityAccepted FromAcceptance IndicatorsStill AcceptedAccepted UntilRejection Indicators
Scientonomy1 January 2016The law became de facto accepted by the community at that time together with the whole theory of scientific change.No21 January 2017The law became rejected as a result of the acceptance of Sebastien's new formulation of the Third Law. For details, refer to the modification.

Suggestions To Reject

These are all the modifications where the rejection of this theory has been suggested:

Modification Community Date Suggested Summary Date Assessed Verdict Verdict Rationale
Sciento-2016-0001 Scientonomy 3 September 2016 Accept a new formulation of the third law to make it clear that employed methods do not have to be deducible from all accepted theories and employed methods but only from some. 21 January 2017 Accepted There was a community consensus that "the new formulation of the third law does bring an additional level of precision to our understanding of the mechanism of method change".c1 The community agreed that the new formulation "makes a clarification that, on its own, warrants this modification's acceptance".c2 Importantly, it was also agreed that the modification "solves the paradox of normative propositions".c3

Question Answered

The Third Law (Barseghyan-2015) is an attempt to answer the following question: How do methods become employed by an epistemic agent?

See Mechanism of Method Employment for more details.


The third law explained by Hakob Barseghyan

Barseghyan's formulation of the third law states that a method becomes employed only when it is deducible from other employed methods and accepted theories of the time. "Essentially," Barseghyan writes, "the third law stipulates that our accepted theories shape our employed methods".1p. 132

According to this formulation, a method becomes employed when:

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

In a nutshell, this suggests that accepted theories shape the set of implicit criteria employed in theory assessment.

In practice, the third law states that when a new phenomenon is discovered, this discovery produces an abstract requirement to take that discovery into account when testing relevant theories. This abstract requirement is then specified by a new employed method.

The third law does not stipulate how methods should go about specifying any new abstract requirement. The third law functions as a descriptive account of how methods change, and is not responsible for describing how methods ought to change. As such, it is an effective means of explicating the requirements of other employed methods.

The third law has an important corollary: scientific change is not necessarily a synchronous process, which notably differs from Kuhn's view of scientific change as a wholesale, synchronous process.1p. 151 This corollary is known as the Asynchronism of Method Employment theorem (Barseghyan-2015).

The gist of this theory can be illustrated by the following examples.

Confirmed Novel Predictions

The third law has also proven useful in explicating such requirements as Confirmed Novel Predictions (CNP).1pp. 146-150

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 (~1740), CNP were not a requirement of his professed method, yet they were still provided. This is also true in the cases of Fresnel's wave theory of light (~1820), Einstein's general relativity (~1920), continental drift theory (1960s), and electroweak unification (1970s).1p. 146

On the other hand, Clark’s law of diminishing returns (1900) had no such predictions. They also played no role in the acceptance of Mayer's lunar theory (1760s), Coulomb's inverse square law (early 1800s), the three laws of thermodynamics (1850s), and quantum mechanics (1927).1p. 146

Barseghyan explains that this indicates that is because "we do expect confirmed novel predictions but only in very special circumstances. There was one common characteristic in all those episodes… they all altered our views on the structural elements of the world".1p. 146 For instance, in our key examples, 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.

Barseghyan presents his historical hypothesis that this specific requirement for CNP has been employed in natural science since the 18th century. Assuming he is correct (for the sake of argument), he continues: "The third law stipulates that the requirement of confirmed novel predictions could become employed only if it was a deductive consequence of the accepted theories and other employed methods of the time. So a question arises: what theories and methods does this requirement follow from?".1pp. 147-148

Barseghyan answers the question with two principles. For one, there is a principle, implicit in our contemporary mosaic and accepted since the eighteenth century, that states: "the world is more complex than it appears in observations, that there is more to the world than meets the eye".1p. 148 Thus, observations may not tell the whole story, as what we observe may an effect of an unobservable. Secondly, "it has been accepted since the early eighteenth century that, in principle, any phenomenon can be produced by an infinite number of different underlying mechanisms".1p. 148 "This leads us to the thesis of underdetermination that, in principle, any finite body of evidence can be explained in an infinite number of ways".1p. 148 Therefore:

The abstract requirement that follows from these two principles is that whenever we assess a theory that introduces some new internal mechanisms (new types of sub-stances, particles, forces, fields, interaction, processes etc.) we must take into account that this hypothesized internal mechanism may turn out to be fictitious even if it manages to predict the known phenomena with utmost precision. In other words, we ddo not tolerate "fiddling" with the accepted ontology; if a theory attemptes to modify the accepted ontology, it must show that it is not cooked-up.1p. 148

This abstract requirement can then be implemented in several ways, including through our contemporary requirement of confirmed novel predictions. This is an illustration of the second scenario of method employment.

Thus, in utilizing the third law, one can discover both when certain criteria become an implicit rule and under what conditions they are necessary.

The Double-Blind Trial Method (Two Scenarios for Method Employment)

As Barseghyan explains, the double-blind trial method "is based on our belief that by performing a double-blind trial we forestall the chance of unaccounted effects, placebo effect, and experimenter’s bias".1p. 141

The propositions that this premise is based on in turn derive from theories that are acecpted; for example, "our belief that a trial with two similar groups minimizes the chance of unaccounted effects follows from our knowledge about statistical regularities, i.e. from our belief that two statistically similar groups can be expected to behave similarly ceteris paribus".1p. 142 Similarly, our knowledge of physiology and psychology lead to our understanding that we can void the placebo effect with fake pills.1p. 142 Our knowledge of psychology allows us to understand that researchers can bias patients from their own knowledge of which group is which.1p. 142 Clearly, these premises, although trivial, are currently accepted within our scientific mosaic.1p. 142 Hence, the double-blind trial method, although an implementation of abstract requirements, is still based on our currently accepted theories. This is true in all scenarios of implementation.1p. 142

Thus, methods follow deductively from elements of the mosaic whether they follow strictly from theories and methods or implement abstract requirements. This is an important similarity between the two scenarios for method employment.

Aristotelian-Medieval Method

In Barseghyan’s explication of the Aristotelian-Medieval method, he illustrates how Aristotelian natural philosophy impacted the method of the time.1p. 143 Most notable is the acceptance of teleology – a theory which states that every thing has a nature it seeks to fulfill (e.g. an acorn’s nature is to become an oak tree). The best theories, then, would uncover the nature of a thing. If only the best theories are acceptable, this leads to the abstract requirement that "A theory is acceptable only if it grasps the nature of a thing". It stood to reason that the nature of a thing can only be intuitively grasped by an experienced person. This fundamental belief, combined with the abstract requirement outline above, led to a method which specifies these requirements known as the Aristotelian-Medieval method: "A proposition is acceptable if it grasps the nature of a thing through intuition schooled by experience, or if it is deduced from general intuitive propositions".1p. 145 This is an illustration of how employed methods are deductive consequences of the accepted theories of the time.

Drug Trial Methods

"How exactly can changes in accepted theories trigger changes in employed methods? What is the precise mechanism of method change? How do methods become employed?".1p. 136

Barseghyan presents the example of testing a new drug for alleviating depression to as an example of the third law and in answer to these questions. In summary, 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.1pp. 132-152

Specifically, Barseghyan begins with the question "How can we ensure that the improvement was due to the drug itself and not due to other unaccounted factors?" The question is answered by the implementation of a controlled trial, wherein "we organize a trial with two groups of patients with the same condition – the active group and the control group. Only the patients in the active group receive the drug".1p. 134

What we have here is a transition from one method to another triggered by a new piece of knowledge about the world. The initial method was something along the lines of hypothetico-deductivism: we had a hypothesis “the drug is effective in alleviating depression” and we wanted to confirm it experimentally. Once we learnt that the alleviation may be due to other factors, our initial method was modified to require that a drug’s efficacy must be tested in a controlled trial.1p. 134

Another transition in method occurred when upon the discovery of the placebo effect, or the fact "that the improvement in patients’ condition can be due to the patients’ belief that the treatment will improve their condition".1p. 135 Now,

it was no longer sufficient to have two groups of patients. If only one of the two groups received the drug then the resulting positive effect could be due to the patients’ belief that the drug was really efficient in alleviating their condition. The solution was to organize a blind trial. We take two groups of patients with the condition, but this time we make sure that both groups of patients believe that they undergo treatment. However, only the patients of the active group receive the real drug; to the patients in the control group we give a placebo (fake treatment).1p. 135

Once again, Barseghyan writes, "this is an instance of a method change brought about by a change in accepted theories".1p. 135

But why are we forced to introduce this new requirement to our method of drug testing? Well, because this new requirement follows deductively from two elements of the mosaic – from our knowledge that the results of testing a hypothesis about a drug’s efficacy may be voided by the placebo effect and from a more fundamental requirement that we must accept only the best available hypotheses.1p. 137

Notably, "while the new requirement is abstract (“the possible placebo effect must be taken into account”), the blind trial method is concrete, for it prescribes how exactly the testing should be done. Thus, the blind trial method specifies the new abstract requirement. This is the relation of implementation: a more concrete method implements the requirements of a more abstract method by making them more concrete".1p. 138 That is, the blind trial method is not the only possible implementation of the abstract requirement to take the placebo effect into account.1p. 138 In Barseghyan's words, "the same abstract requirement can have many different implementations".1p. 139

A final change in method occurred when experimenter's bias was discovered:

The researchers that are in contact with patients can give patients conscious or unconscious hints as to which group is which. It is possible that the positive effect of the drug established in a blind trial was due to the fact that the patients in the placebo group knew that they were given a placebo. The method of drug testing was modified yet again to reflect this newly discovered phenomenon. The contemporary approach is to perform a double-blind trial where neither patients nor researchers know which group is which.1pp. 135-136

Double Blind Trial Deduction (Barseghyan-2015-139).png

The double-blind trial method is a further example of the relation of implementation.


No reasons are indicated for this theory.

If a reason supporting this theory is missing, please add it here.

Questions About This Theory

There are no higher-order questions concerning this theory.

If a question about this theory is missing, please add it here.


  1. a b c d e f g h i j k l m n o p q r s t u v w x y z aa ab ac ad ae af ag ah  Barseghyan, Hakob. (2015) The Laws of Scientific Change. Springer.
  2. ^  Laudan, Larry. (1984) Science and Values. University of California Press.
  3. ^  McMullin, Ernan. (1988) The Shaping of Scientific Rationality: Construction and Constraint. In McMullin (Ed.) (1988), 1-47.