Laws Are Not Descriptions

Abstract

The view that takes laws of nature to be essentially nothing more than descriptions of facts is still rather popular. The present article, on the contrary, defends the claim that the only real motivation for defending a descriptive view of laws—the quest for ontological parsimony—entails too high a price to pay in philosophical terms. It is argued that nomic primitivism, namely the alternative option that takes laws to be primitive fundamental entities in our ontology, is decisively more appealing, since it is the crucial role assigned to laws that makes a scientific theory of natural phenomena a system rather than a list. Finally, the implications that nomic primitivism might have for the issue of the status of the wave function in that particular formulation of quantum mechanics known as Bohmian mechanics are considered.

Acknowledgements

I presented the main ideas of the article at an early stage at the international conference Philosophy of Science in the 21st Century—Challenges and Tasks, Lisbon, 4–6 December 2013, and at the annual meeting of the Deutsche Physikalische Gesellschaft, Philosophy of Physics Section, Berlin, 17–21 March 2014. I wish to thank the audiences for their stimulating questions. I have also discussed parts of the article at different stages with Mauro Dorato, Aldo Filomeno, Shelly Goldstein, Tim Maudlin, and Nino Zanghì. Finally, I wish to thank the referees of this journal: their remarks and advice contributed to improve significantly the final outcome.

Notes

[1] As is well known, however, van Fraassen himself strongly opposes the view that one of the main tasks, if not the main, of natural sciences is the search for laws.

[2] Hall (2012) is an unpublished manuscript, which is a much more comprehensive version of Hall (2015).

[3] The issue of possible violations of laws of nature, for instance, is not an idle one.

[4] In our opinion the first strategy is definitely less confusing but, provided uniformity is guaranteed, no serious conceptual harm is caused by the second strategy.

[5] For instance, a best-system implementation of the descriptive view can manage the set-theoretic objections, exactly because the ‘best’ choice in singling out the axioms that are suitable candidates for performing as laws can be done over different sets, that appear as merely different descriptions among which the optimal combination of simplicity and strength is recovered.

[6] As the authors emphasize, the term ‘Humean supervenience’ is ‘a name shared by many different theses, differing from one another in subtle ways, though they are all intended to capture the same general view of the world’ (Earman and Roberts 2005a, 2).

[7] This point recalls a critical point of the Lewisian best-systems analysis: when we say that the laws of nature belong to all the true deductive systems with a best combination of simplicity and strength, what do we mean exactly by ‘best’? (See for instance Roberts 2008, 10.)

[8] In addition to non-locality, which threatens the original Lewisian intuition of what a Humean base is supposed to be (Maudlin 2007, 50–64).

[9] Even worse: according to some versions of the Everettian quantum mechanics, the only fundamental entity is the ‘big’ wave function, namely the wave function of the whole universe, and the wave functions attached to single systems are derivative, perspectival and in some sense endowed with a second-order reality (Vaidman 2014).

[10] Metaphorically we might say that the ‘fact-like’ law seems to simulate a ‘governing’ law.

[11] Under the assumption according to which a descriptive view of laws is a reductionist view in a serious sense, the primitivist approach is an anti-reductionist view which includes similar, but mutually non-equivalent positions such as Carroll (1994, 2008), Lange (2000, 2009), and Maudlin (2007): in the present article we will focus on the Maudlin version.

[12] In his proposal, Maudlin selects a specific form of fundamental law as particularly apt to cover many law instances in the special sciences, namely what he calls FLOTE, that is, Fundamental Law Of Temporal Evolution (Maudlin 2007, 12). This option raises immediately the question (that we will not touch here) of whether this choice compels one to assume also time—entering the FLOTE-kind of laws—as a primitive entity or not.

[13] An additional advantage seems to be that, if we assume laws as primitives, the principle according to which we generate models only in terms of their compatibility with laws requires—so to speak—a ‘minimum’ of modality: possibility is exactly law-compatibility, whereas necessity is obtained simply from the fact that the validity of the law is a sort of fixed point in any law-generated model.

[14] Here Kepler’s and Newton’s laws are just examples of what might happen to laws that at a certain stage are taken to be fundamental and that, at a later stage, change this status. The charge that some might raise—Kepler’s and Newton’s laws are not plausible candidates for lawhood anyway because they are false—is irrelevant here: the focus here is on how we should assess the level of fundamentality of certain laws with respect to others, not on their actual truth or falsity.

[15] I concede that the point about conventionality is an epistemological one, whereas the primitivism about laws is mainly an ontological claim (remark by a referee). My reply is that this overlap between epistemology and ontology is innocuous and, at any rate, this sort of overlap is often inevitable in debates like that on laws of nature.

[16] I said one of the features and not the feature, otherwise I would have turned back to the above mentioned claim according to which we should define laws in that way, a claim that—as I have recalled earlier—Humeans can consistently resist.

[17] It is in this sense that Bohmian mechanics might be interpreted as ‘en-theorizing’ a form of nomological realism for the wave function, similarly to what several years ago Arthur Fine proposed in order to assess the very nature of the Einsteinian use of such notions as ‘realism’, ‘causality’, and ‘determinism’ (Fine 1986). For some remarks on a more general en-theorizing strategy concerning Bohmian mechanics, see Laudisa (2012).

[18] There is now an extended literature on the details of such an ontology and its connections with many of the general, interpretive issues in the foundations of non-relativistic quantum mechanics, and this is not the place to provide a review. For a recent presentation and re-assessment, see Dürr, Goldstein, and Zanghì (2013).

[19] A rigorous justification that the conditional wave function is phenomenologically ‘well-behaved’ can be found in Dürr, Goldstein, and Zanghì (1992).

[20] Esfeld et al. (2014) mention the primitivist option but, as matter of fact, they rapidly dismiss it, assimilating its fate to the fate of ontological interpretations of the wave function, that they reject: for they argue that, if we ground the nomological interpretation of the wave function onto nomic primitivism ‘there is no sharp distinction between a nomological and an ontological interpretation of the wave function, since laws belong to the stock of physical reality as well’ (Esfeld et al. 2014, 780). But, as we attempted to show above with the argument of the two-dimensional X_PO–Y_N plane, no sharp distinction does not imply no distinction at all.

[21] For further references on the debate over the status of the wave function in quantum mechanics, see for instance Solé (2013), Dorato and Laudisa (2014), and a recent special issue of Synthese devoted to the issue of the status of the wave function (Hoefer and Solé 2015).