Is Philosophy Exceptional? A Corpus-Based, Quantitative Study

ABSTRACT

Drawing on the epistemology of logic literature on anti-exceptionalism about logic, we set out to investigate the following metaphilosophical questions empirically: Is philosophy special? Are its methods (dis)continuous with science? More specifically, we test the following metaphilosophical hypotheses empirically: philosophical deductivism, philosophical inductivism, and philosophical abductivism. Using indicator words to classify arguments by type (namely, deductive, inductive, and abductive arguments), we searched through a large corpus of philosophical texts mined from the JSTOR database (N = 435,703) to find patterns of argumentation. The results of our quantitative, corpus-based study suggest that deductive arguments are significantly more common than abductive arguments and inductive arguments in philosophical texts overall, but they are gradually and steadily giving way to non-deductive (i.e. inductive and abductive) arguments in academic philosophy.

KEYWORDS:

• exceptionalism
• indicator words
• philosophical abductivism
• philosophical deductivism
• philosophical inductivism

Acknowledgments

We are grateful to two anonymous reviewers of Social Epistemology for invaluable comments on earlier drafts of this paper.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data Availability Statement

Raw data sets for this study can be found on OSF at https://doi.org/10.1080/02691728.2022.2109529.

Notes

1. Thanks to an anonymous reviewer for this point.

2. It is important to keep in mind the distinction between methods of theory-choice or types of argument, on the one hand, and sources of evidence, on the other hand. For one could accept the empirical evidence pointing to a shift from the traditional methods of academic philosophy toward empirical methods, and yet insist that the sources of evidence that academic philosophers use are still the traditional sources of intuition, introspection, and the like. Similarly, as both Martin (2021) and Martin and Hjortland (2021) observe, one can reject the foundational status of logical knowledge, and thereby embrace an abductivist or a predictivist picture of theory-choice in logic, while at the same time insist that the “data” logicians use are still a priori. This issue is beyond the scope of this paper. For present purposes, our focus is on methods of theory-choice or styles of argument rather than sources of evidence in academic philosophy. But again we are grateful to an anonymous referee for raising this important point.

3. See also Fletcher et al. (2021) for empirical evidence pointing to a turn away from formal, logical methods toward probabilistic methods in academic philosophy.

4. According to McCain and Poston (2017, 1), “Not only are rigorous inferences to the best explanation (IBE) used pervasively in the sciences, explanatory reasoning is virtually ubiquitous in everyday life”.

5. According to Okasha (2016, 19), “Most philosophers think it’s obvious that science relies heavily on induction, indeed so obvious that it hardly needs arguing for”.

6. According to Chakravartty (2017), abduction “seems ubiquitous in science”. See also Douven (2017) on the ubiquity of abduction. On the ubiquity of induction, see Henderson (2020): “We appear to rely on inductive inference ubiquitously in daily life, and it is also generally thought that it is at the very foundation of the scientific method”.

7. Cf. Martin and Hjortland (2021) on logical predictivism.

8. In another empirical study using the methods of data mining and text analysis, Ashton and Mizrahi (2018b) test the “received wisdom” about intuition talk and appeals to intuition in academic philosophy. The results of their empirical study show that, contrary to the “received wisdom”, both intuition talk (as indicated by words such as ‘intuit’ and ‘intuitive’) and appeals to intuition (as indicated by phrases such ‘it seems that’ and ‘it appears that’) go as far back as the 1800s.

9. Ashton and Mizrahi (2018a) use data mined from JSTOR as well. Other philosophers have used JSTOR Data for Research to model topics in philosophy journals. See, for example, Weatherson (2020)’ A History of Philosophy Journals: Volume 1 Evidence from Topic Modeling, 1876–2013 (http://www-personal.umich.edu/ weath/lda/).

10. For an example of the application of corpus-based methods to philosophy of logic, see Mizrahi (2019). Mizrahi (2019, 203) uses corpus-based methods “to test empirically how the idea that ‘logic is obvious’ is reflected in logical and philosophical practice”.

11. According to Salmon (2013), “Expressions such as must, it must be the case that, necessarily, inevitably, certainly, and it can be deduced that frequently indicate that an argument is deductive”, (86), whereas expressions such as “probably, usually, tends to support, likely, very likely, and almost always” typically indicate that an argument is inductive (94).

12. In Hyland (2005) taxonomy of metadiscourse signals, ‘probably’ and ‘likely’ are classified as hedges, whereas in Salager-Meyer (1994) taxonomy they are classified as shields. For a critical assessment of these taxonomies, see Thabet (2018).

13. Thanks to an anonymous reviewer for pressing on this point. As Flake and Fried observe, “When information about the measures in a study is lacking, the information needed to evaluate the validity of the study is also lacking” (Flake and Fried 2020, 459).

14. See Appendix A for details on the text mining methods we used in R.

15. According to Okasha (2016, 19), “science relies heavily on induction”.

16. According to McMullin (1992), abduction is “the inference that makes science”.

17. For a corpus-based, empirical study on whether scientists use mostly deductive terms or inductive terms when they talk about testing hypotheses in scientific publications, see Mizrahi (2020).

18. See also Fletcher et al. (2021) for empirical evidence for a turn away from formal, logical methods toward probabilistic methods in academic philosophy.

19. For an example of a more qualitative approach to corpus analysis, see Sytsma et al. (2019).

Additional information

Notes on contributors

Moti Mizrahi

Moti Mizrahi is Associate Professor of Philosophy at the Florida Institute of Technology. He is the author of The Relativity of Theory: Key Positions and Arguments in the Contemporary Scientific Realism/Antirealism Debate (Springer, 2020), the editor of The Kuhnian Image of Science: Time for a Decisive Transformation? (Rowman & Littlefield, 2018), and the editor of For and Against Scientism: Science, Methodology, and the Future of Philosophy (Rowman & Littlefield, 2022). He has published extensively on the philosophy of science, the scientific realism/anti-realism debate, the epistemology of philosophy, and argumentation. His work has appeared in journals such as Argumentation, British Journal for the Philosophy of Science, Erkenntnis, Philosophical Studies, Studies in History and Philosophy of Science, and Synthese. A major theme in his work has been the application of digital, statistical, and data-driven methods to problems concerning moral, philosophical, and scientific reasoning.

Michael Adam Dickinson

Mike Dickinson is the Planning, Landscape Architecture, & Agriculture Librarian at the University of Illinois. His librarianship has centered on data management. Mike’s research interests include data education, large-scale library reference analysis, and corpus-based textual analysis of philosophical literature.