In his 1991 book Inference to the Best Explanation, Peter Lipton argued that identification of the best explanation out of a pool of competing ones is not just a matter of identifying which explanation is the likeliest. Apart from likelihood, one also should look for explanation that is lovely – one that provides a sense of understanding.
According to Lipton, epistemic values (e.g. simplicity, fruitfulness, and predictive success) of the explanation are an important source of such understanding. The subtext of Lipton's thesis is that the differences among explanations could be articulated in terms of valuable qualities. For example, some explanations are predictively successful, others greater in their simplicity.
Especially priority disputes provide good grounds for testing this idea. Take the earliest periodic systems, which were developed about 150 years ago. Many associate the Russian chemist Dmitri Ivanovich Mendeleev with the discovery of the system, but before Mendeleev, a number of other chemists had developed systems that are now recognised as the precursors for the modern one.
This backdrop of similarity enables identifying contrasts between the systems. In my PhD research, I argued that especially Mendeleev, Julius Lothar Meyer, and John Newlands emphasised different values when they developed their systems. While no chemist emphasised just one value, Newlands elevated simplicity (“simple relation”), Mendeleev completeness (polnost’), and Meyer carefulness when systematising the elements. (The full story can be read here.)
Investigating how different practitioners emphasised different values gives a helpful way of teasing out differences between competing scientific representations. But is there anything else that values help to uncover? In what follows, I list some additional insights that may be gained if we look into how values guide chemistry.
(1) Values show thematic similarities between distinct design-choices. Often, the output of chemists’ labour includes both pictures and text. When investigating them, one might note how they drew their diagrams in a manner different to their peers, expressed measurements up to a certain degree of precision, or preferred bold conjectures over more modest ones. Initially, the design-features listed might seem like unrelated technical decisions, lacking any apparent connections. But if we seek how value-judgments guided the development of such works, it becomes easier to see connections underlying seemingly distinct design-choices.
Consider the periodic system of Meyer. Authors often bring attention to how his systems were not quite as full as those of the others. In addition, Meyer expressed atomic weights of the elements to a higher degree of precision. Initially, we might not draw a connection between his exclusion of some elements and his choice of expressing some atomic weights precisely. But upon closer inspection, Meyer’s exclusion of elements and high decimals were both signals of the quality of data, as Meyer excluded elements he deemed poorly understood and highlighted accurately measured atomic weights by employing several decimal numbers. The two choices were driven by Meyer’s concern for the quality of data.
(2) Values may give the link between development of scientific representations and their reception. I have argued that valuing of specific attributes played a role in putting together the early periodic systems and explains their differences. This argument related to the processes of developing the systems. But if values help to account for differences in the systems, do they also help in explaining why their reception was different?
Again, the early periodic systems should give us a good example. Why? The systems of different chemists were received in a different manner. Most notably, the Royal Society awarded Mendeleev and Meyer with Davy medal, but initially denied John Newlands an award. Only after Newlands lobbied for his priority, he was awarded one too. In their subsequent explanation of the delay, the president of Royal Society argued Newlands’ system had not seem that complete.
This comment gives an interesting puzzle. Contrary to what the president’s statement states, Newlands’ system had in fact more elements than those of Meyer who was awarded alongside Mendeleev. How is it possible that Newlands’ system, which managed to house so many elements, was still perceived as incomplete?
It might be tempting to dismiss the president’s comment as just an ad hoc reason, but we could also take it seriously. If we inspect Newlands’ system more closely, it becomes clear that there is something missing. Although it had many elements – more than the system of Meyer – all of them were neatly represented in its grid-structure. As can be seen from figure 1, there weren’t enough grids for all the elements, so some of them had to be squeezed into the same grid.
This detail gives us a potential answer to the puzzle: the tight grid of Newlands’ system made it unclear where new, undiscovered elements would be included. This suggests that completeness was not just about accounting for known elements, but also saying something about those that were expected to be discovered.
Figure 1. Newlands' "Law of Octaves" of 1866
In this particular case, looking at values both in reception and in development indicates that values were interpreted differently by different chemists. Although Newlands occasionally stressed the importance of completeness when developing his system, and included many elements, it appears that his understanding of completeness was different to that of the president of Royal Society.
(3) Values allow us to draw connections between someone’s science and other forums of activity. Looking at values guiding chemist’s chemistry also sets the table for examining whether those values also guided some other aspects of their lives. Although I personally did not opt for this approach, there are some illuminating examples of it in the literature. For example, Matthew Stanley examined how values guided the science A. S. Eddington in order to get a better picture of the (dis)connections between his work and his Quakerism. (According to Stanley, examining the connections with the help of broader categories of “science” and “religion” would not have provided as detailed a picture.) Earlier, Kathryn Olesko (1995) has argued that when accuracy emerged as the term describing especially praiseworthy measurements in the 19th century German lands, the term also had gendered import as accuracy carried military connotations. Thus, looking into values in the development of representations allows us to examine connections with values that operated in other forums. So, the framework of values in science can be useful especially when examining contrasts and similarities, and connections and disconnections between scientific representations and different areas of practice.