Microsoft Word - LAUDISA, Laws are not descriptions


1 

 

Laws are not descriptions 

 

 

Federico Laudisa 

Department of Human Sciences, University of Milan Bicocca 

Piazza dell’Ateneo Nuovo 1, 20126 Milan, Italy 

 

 

 

 

 

Abstract 

 

The view that takes laws of nature to be essentially nothing more than descriptions of 

facts is still rather popular. The present paper, 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 what makes a scientific theory of natural phenomena a 

system rather than a list. Finally, the implications that nomic primitivism might have on 

the issue of the status of the wave function in that particular formulation of quantum 

mechanics known as Bohmian mechanics are considered. 

 

 

 

 

 

 

 

 

 



2 

 

1  Introduction 

About ten years ago I spoke with Einstein about the 
astonishing fact that so many ministers of various 
denominations are strongly interested in the theory of 
relativity. Einstein said that according to his estimation 
there are more clergymen interested in relativity than 
physicists. A little puzzled I asked him how he would 
explain this strange fact. He answered, a little smiling, 
"Because clergymen are interested in the general laws of nature 
and physicists, very often, are not. " 

 
Frank 1949, p. 349 

 

According to a common sense intuition, the very idea of natural law entails that 

the law governs the phenomena falling under its domain of application. According 

to this intuition, the true role of a natural law is exactly that of providing for the 

relevant phenomena such ‘governing’ things as order, structure, evolution over 

time, and the like. Bas van Fraassen, just to mention one out of many possible 

examples, echoes such a widely felt intuition when in his Laws and Symmetry he 

recalls that one of the main tasks traditionally attributed to natural science is 

exactly “to state the laws which the things in the universe obey” (van Fraassen 

1989, p. 18, emphasis added)1. Although it does not prove anything per se and 

needs to be supplemented by robust arguments in order to be convincing, this 

intuition seems to have an undeniable appeal and van Fraassen’s wording is 

especially apt in our case: when there is someone or something that governs, there 

is always also someone or something that obeys and, in the case of natural 

phenomena, for them to obey means that they are constrained by some form of 

necessity or ‘nomicity’ (be it deterministic or indeterministic): we tend to think 

that a freely falling stone cannot choose whether its motion will be uniform or 

accelerated!  

As we know from the history of philosophy, however, it is highly 

controversial whether we can legitimately project such necessity or nomicity 

onto the world or we should limit ourselves to locate them within the knowing 

subject. For instance according to a reductionist approach, we should resist the 

governing view of laws, since it does not seem inevitable to assume such a modal 

                                                           
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 after laws. 



3 

 

feature for the notion of law. In such an framework, it is not a conceptual truth 

that laws govern, so that it is perfectly possible to entertain a non-governing 

view of laws without being plainly inconsistent: in this framework “laws are 

purely descriptive of the particular matters of fact” (Beebee 2000, in Carroll 

2004, 256). 

For a philosophical position, however, to be logically consistent is not 

equivalent to be convincing. In the present paper I will argue against this 

particular descriptive emphasis of the non-governing view of laws, by showing 

further motivations – in addition to the most familiar ones – why such view of 

laws turns out to be especially defective. Although it is not plainly inconsistent 

to deny that the modal character is intrinsic to the notion of law, I will argue 

that the purely descriptive view has such counterintuitive implications that it 

fails to countenance features that any decent notion of law cannot, after all, 

afford to miss. In fact, the only real motivation for defending a descriptive view 

of laws appears to be a quest for ontological parsimony – the whole reality 

conceived as nothing but an astronomical amount of unconnected local facts – a 

motivation that requires a significant price to pay in philosophical terms. On the 

basis of the claim that laws should be rather viewed as grounding natural 

processes in a non-purely descriptive fashion, I will then elaborate on one of the 

possible options that are alternative to the regularity view, namely the so-called 

primitivist approach to laws: a special attention will be then devoted, in the last 

section, to the implications that this form of nomic primitivism might have on 

the issue of the status of the wave function in that particular formulation of 

quantum mechanics known as Bohmian mechanics.  

 

 

2 The descriptive view of laws 

 

The debate on the nature of lawhood is a complex, intertwined network of 

issues. Starting from a very general dichotomy, according to which either laws 

are somehow a part of nature or they pertain only to scientific theories (hence – 

lastly – to us as knowing subjects), the controversy proceeds toward further, 



4 

 

more specific issues: regularity vs. necessity, the role of Humean supervenience, 

the relation with the use of laws in scientific practice, the governing vs. the non-

governing status of laws, and so on. Ned Hall claims, for instance: 

 

By far the most central and important question about laws of nature is this: are they 

mere patterns in the phenomena (patterns that are in some way salient, to be sure – but 

still, nothing more than patterns)? Or are they something more, something that 

somehow governs or constrains those phenomena? Disagreement over this issue 

constitutes the Schism in contemporary philosophical work on laws of nature. (Hall 

2012, p. 1)2 

 

This sort of disagreement is in non-linear relation with different, possible 

positions over the issue whether we should think that there really are items like 

‘laws’ and the search for such ‘things’ is the essence of science: anti-laws 

philosophers like Bas van Fraassen or Ron Giere have vehemently denied it: 

 

The claims of some philosophers, that scientists seek to discover laws of nature, cannot 

be taken as a simple description of scientific practice, but must be recognized as part of 

an interpretation of that practice. The situation is complicated, of course, by the fact 

that, since the seventeenth century, scientists have themselves used the expression «law 

of nature» in characterizing their own practice. (Giere 1999, p. 84) 

 

But if one accepts as a working hypothesis the existence of ‘law-like’ items in the 

inventory of the world, a major issue is of course what are the terms in which 

should we conceive the relation between such items and the particular 

phenomena that in some way or another ‘fall under’ them. It is at this stage that 

the controversy ‘governing vs. non-governing’ arises. The “Governing answer 

[…] insists that there are genuine laws of nature and furthermore that these 

laws govern or even produce the events of the world” whereas “the Non-

Governing answer […] has it that there are genuine laws of nature, but that 

they do not govern or produce the events of the world. The mosaic of events 

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



5 

 

displays certain patterns, and it is in the features of some of these patterns that 

we find laws.” (Cohen, Callender 2009, p. 2). In a Humean, reductionist 

perspective, we should be careful in admitting suspect versions of necessity in 

nature so that, as a consequence, we should not require from laws any 

‘governing’ role as a constitutive feature. In the global metaphysical view of the 

world that is on the background of (any version of) views like this, any modal, 

governing feature in mentioning laws must be traced back to us as subjects, not 

to the world, that in itself is nothing but a collection of sparse entities, usually 

conceived as discrete. Hall captures the Lewisian terse way to characterize this 

view: “Drawing inspiration from Hume, [Lewis] took the fundamental 

ontological structure to consist in the pattern of instantiation by space-time 

points of perfectly natural monadic properties, together with the facts about the 

spatiotemporal relations among those points.” (Hall 2012, p. 5) 

In a regularist brand of a reductionist perspective, what is admitted is just the 

existence of a(n astronomical) number of matters of fact and the existence of a 

certain number of regularities connecting them (Mumford 2004, 32): although it 

is far from trivial specifying what a regularity exactly is in non-modal terms, a 

(naïve) regularity view denies the existence of necessary connections or laws 

whose role in some way or another would be to ground the regularities. The 

more sophisticated versions of a regularity view, summarized in the so-called 

Mill-Ramsey-Lewis best-system view, turn out to be an elaboration of how we as 

subjects should organize regularities, if we want them to play the role that non-

Humean views usually associate to the notion of law. The now classic 

formulation of David Lewis – according to which the ‘laws of nature’ are the 

universal statements that belong to all the true deductive systems with a best 

combination of simplicity and strength (Lewis 1973, 73) – makes explicit the 

disregard for what is often interpreted as the instrinsically modal character of 

natural laws. According to Peter Menzies, for instance, this is exactly the main 

defect of all regularity, best-system approaches to laws: 

 

The defect is that they fail to capture the modal character of laws of nature. […] It is in 

virtue of this modal feature of laws that they fail to conform to Humean supervenience: 



6 

 

two worlds agreeing on actual occurrent facts may differ in their modal structure. 

Lewis’s theory fails because all such regularity must fail; for they try to fashion modal 

facts from the thin actualistic resources of Humean empiricism. But to do this is to 

commit what might be called the actualist fallacy of trying to derive conclusions about 

the possible and the necessary from premises about the actual. No amount of 

sophisticated talk of ideal systems that combine simplicity and information content will 

make this fallacious inference virtuous. (Menzies 1993, pp. 200-201) 

 

In the wake of the position echoed by Menzies, there might be then a tempting 

move for the anti-regularist: by assuming the reliability of a strong modal 

intuition on laws, the anti-regularist might count a governing view of laws as an 

instance of this modal intuition and define laws as abstract governing entities, 

with the effect of dismissing the non-governing view of laws on simple 

conceptual grounds.  

Now, no matter what the position on the irreducibility of modal features of 

laws we hold, objections of the sort argued for by Menzies look like a petitio 

principii, since whether laws of nature should be conceived modally or not is 

exactly the point under discussion. Moreover, a regularist may well resist the 

claim that the notion of law is to be defined in governing terms: she will try to 

claim that such move is far from unavoidable and that, as a consequence, a non-

governing conception of laws is perfectly consistent. This is what Helen Beebee, 

for example, tries to accomplish by claiming that one can define law as 

governing only by assuming an analogy between laws in nature and laws in 

other domains such as theology, politics, ethics and so on: but if this assumption 

is rejected (and it can be rejected), the governing feature of laws fails to be a 

conceptual truth (Beebee 2000).  

According to Beebee the governing intuition about laws might be justified 

only if we assume a deep similarity between the way in which phenomena are 

constrained by natural laws on one side and the way in which moral individuals as 

believers, citizens and so on are subject to prescriptive laws on the other. This 

similarity fails, however, for at least two reasons. First, in the human, 

prescriptive domain it is possible for an individual to violate laws, something 



7 

 

which is not easily the case with natural laws: indeed, it seems constitutive of 

moral and political laws that they can be broken, since it is from this possibility 

that their prescriptive role derives its meaning. Second, moral and political laws 

contemplate forms of punishment for their violation, something which hardly 

makes sense if extended to natural laws. What does all this show? Well, it is 

very plausible to think that if common sense does take it to be part of the concept 

of law of nature that those laws govern, then it does so only because of a tacit 

assumption that laws of nature operate in a way that is analogous to the way that 

other laws – laws which really do govern – operate. But that assumption cannot 

be maintained, since the alleged governing nature of natural laws would have to 

be entirely unlike the prescriptive nature of moral and other laws. (Beebee 2000, 

in Carroll 2004, p. 260) 

Now even if, as a working hypothesis, we assume the plausibility of these 

arguments3 and concede that the analogy between natural laws and other sorts 

of law is not entirely plausible, or implausible at all, the question remains: what 

exactly does the regularist view implies about the very notion of law? First of all, 

let us focus on a linguistic stipulation. If by ‘law’ we agree to mean 

‘regularity+something-grounding-the-regularity’, it follows that in strictly naïve 

regularist terms there are no laws at all: according to the metaphysical view on 

the background of the regularist stance, the world is a lawless place. On the 

other hand, we may simply decide to equate ‘law’ and ‘regularity’, without 

charging the term ‘law’ with the responsibility of embodying some principles – 

the principles ‘in virtue of which’ the regularity would be supposed to hold4 - 

over and above the regularities themselves. The latter is the option adopted by 

Beebee 2000 who characterizes the divide between Humean and anti-Humean 

stances in these terms: 

 

For the anti-Humean, laws (unlike accidentally true generalizations) do something – 

they govern what goes on in the universe – and they therefore require some sort of 

                                                           
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. 



8 

 

ontological basis […] that gives them this ability. Humeans, on the other hand, do not 

require laws to “do” anything: like accidentally true generalizations, laws are at bottom 

merely true descriptions of what goes on. Thus for the Humean there is no need for any 

ontological distinction between laws and accidents. (Beebee 2000, in Carroll 2004, pp. 

258-259) 

 

Under this linguistic stipulation concerning laws, then, let us focus on what in a 

Humean framework a purely descriptive view of laws might entail. Humeans 

deny the need for a search after something that might ‘ground’ a regularity: they 

“do not require laws to «do» anything: like accidentally true generalizations, 

laws are at bottom merely true descriptions of what goes on”. The descriptive 

view seems therefore to support the idea that a law is literally nothing but a sum 

of facts: “for the Humean, since the laws are descriptive, what the laws are 

depends on what the facts are” (ibidem). What I would like to argue for in what 

follows is that such view, although it cannot be ruled out in terms of logical 

consistency, entails features that are seriously contrived in order for the 

descriptive view of laws to be plausible. I will take into account four points: (1) 

determinism, (2) the implications of a possible set-theoretic reading of a law-as-

sum, (3) the problem of the cardinality of the set of facts, (4) the problem of 

uniformity.  

(1) The case of determinism  

Let us consider then first the case of deterministic evolution over time. 

According to Beebee, determinism is useful to express coincisely the non-

governing relation that by a Humean standpoint is supposed to hold between the 

law and the facts that intuitively are accounted for by the law:  

 

We can characterize determinism in the following rough and ready way: the state of the 

universe at any given time together with the laws of nature determines what the state of 

the universe will be at any future time. But what does “determines” mean here? For the 

Humean, the laws and current facts determine the future facts in a purely logical way: 

you can deduce facts from current facts plus the laws. And this is just because laws are, in 



9 

 

part, facts about the future” (Beebee 2000, in Carroll 2004, p. 257, emphasis in the 

original text).  

 

Two minor points, to begin with, that are far from clear. What would the notion 

of fact about the future exactly mean by a strictly Humean, descriptivist 

viewpoint? It seems that, in the metaphysics this viewpoint presupposes, there is 

simply – at every given instant of time – a given collection of facts, none of 

which can be ‘about’ something that does not exist yet. A fact is what it is just 

because there is nothing beyond it and this is just the role that it is supposed to 

play in this lawless image of the world: to be nothing but what it is, full stop. 

Moreover, no matter whether the notion of fact about the future is consistent with 

a Humean metaphysics, I do not see how a Humean can be legitimated to say 

that laws are facts but only ‘in part’: What is the remaining part? Does it need to 

be ‘added’ to the purely factual part of laws? Was not the core of the Humean 

view of laws that these are literally nothing but facts? 

 But let me come to the main point. This Humean understanding of 

determinism overlooks that its ‘determining’ capability is grounded on a specific 

mathematical formulation of a dynamical law. Take the Newton 2nd law. The 

general form of this dynamical law states a proportionality between force and 

acceleration: when a specific formulation for a kind of force is inserted into the 

2nd law, we obtain a mathematical equation that, under non-trivial conditions 

turns out to be integrable. This is what justifies us to assume that the knowledge 

of a given (initial) state and of what is the force (if any) acting on the system 

makes it possible to determine future states of the system (and also past states, if 

the evolution satisfies time reversal invariance). So the crucial point is that the 

determination is possible due to the functional relations among physical 

quantities, relations that are encoded into the mathematical formulation of 

Newton’s law: this encoding can hardly be reduced to any ‘sum’ of facts and 

cannot be easily accounted for in a purely descriptive view of laws. In addition to 

that, let us consider the relation between determinism and predictability. There 

are well-known physical situations (deterministic chaos, three-body problem, …) 

in which, apparently, the phenomena are governed by deterministic laws but we 



10 

 

are not able to predict what the future states will be, namely we are not able to 

‘produce facts’ – in the Humean-like jargon – on the basis of the dynamical law at 

hand. What is left of the Humean-kind-of-law if we simply cannot deduce the 

future facts on the basis of the pair <initial conditions, law> even if the 

phenomena are governed by a deterministic law? 

(2) Laws-as sums: a set-theoretical reading?  

The idea that laws are ‘sums’ of facts lends itself to receive a set-theoretic 

reading. So let us suppose that we have the following collections of states S at 

their respective times t 

Fn = {..........., S(tn-1), S(tn)},   

Fn+m = {..........., S(tn), ......, S(tn+m-1), S(tn+m)} 

If we assume that any of these states is a collection of values of a set of relevant 

physical quantities, each state may well represent a ‘fact’ in the regularist vein, 

since each state works as a sort of snapshot at its time t(*) of the physical situation 

at stake. Moreover, we assume that the states both in Fn and Fn+m are obtained as 

the computational output of the algorithm implicit in one and the same 

deterministic dynamical law.  

Now, the intuition tells us that we have here the same law accounting for the 

evolution of our system at two different times. But, since in a Humean reading 

there is nothing to a law except a ‘sum’ of facts, in the present case the Humean 

reading seems to have no resource to avoid the paradoxical consequence that 

here we deal with two sort-of-laws, since we have two different sums! Even 

worse: at any successive instant of time a new law-as-a-sum is generated. But, 

then, being parsimonious on any alleged modal features of the world seems 

paradoxically to imply a wild and uncontrollable generation of laws-as-sums 

over time, a phenomenon in strong tension with an aspiration to a metaphysical 

economy. On the other hand, should there be a unifying principle according to 

which Fn and Fn+m might be shown to be just two instances of one and the same 



11 

 

law, this principle could not simply supervene on the states and because of this 

non-supervenience this account would immediately become a non-Humean one.  

(3) The problem of cardinality 

An additional problem (but perhaps for another paper) for the regularist view 

might concern the cardinality of the set of all possible facts. Is that set finite or 

infinite? Since it seems hard to devise a robust argument to argue that facts are 

absolutely in finite number, let us turn to the infinity horn. In this case, our 

intuition is likely to suggest that the set of all possible facts should be conceived 

as a countable set, but to a large extent what is intuitive or not is a matter of 

convention: there seems to be no unique way of cutting reality into facts so as to 

ensure that there are exactly as facts as natural numbers, so who are we to deny 

that the set of all the facts might have the cardinality of the reals? In this case, 

however, we might have a hard time simply in discriminating one fact from 

another, so that even the most basic statement of a regularist view – there are 

just facts – seems to be under threat. 

(4) The problem of uniformity 

Sometimes the above Humean view of a lawless world in terms of ‘mosaic’ of 

individual, local matters of fact is formulated with the aid of what has been called 

recombination (Hall 2015) or redistribution (Hildebrand 2013) principle. Namely, 

given a mosaic of individual, local matters of fact, there is no privileged way of 

setting up the mosaic, since this alleged privileged way would amount to a sort 

of ‘ordering’ principle over and above the mosaic itself, a consequence that a die-

hard Humean could not admit: hence any possible combination or distribution of 

such matters of fact is equally possible. In this case, however, a ban on any 

constraints over the possible arrangements of individual, local matters of fact 

seems to have the undesirable consequence that it fails to explain why our world 

is in fact highly uniform, since the absence of constraints makes non-uniform 

arrangements as likely as the uniform ones. The recourse to the idea of pattern is 

of no help here: if we have sufficient reasons to ascribe one or more ‘patterns’ to a 



12 

 

given arrangement, those patterns embody de facto a kind of uniformity that the 

no-constraints ban is unable to account for. 

 

 

3 Primitivism about laws 

 

As we have seen, there are a number of reasons why one should be dissatisfied 

with the anti-necessitarian viewpoint. In particular, from that viewpoint it turns 

out to be difficult to cope with that seemingly irreducibly modal aspect that 

informs our explicit and implicit way of employing or referring to laws: for 

instance, the demand according to which laws must cover ‘possibilities’, and not 

just actualities, and the demand according to which, when we ask a law to explain 

facts, we search after what is in virtue of which facts obtain. Moreover, a look at 

the actual scientific practice when analyzing how laws work does nothing but 

increase the dissatisfaction: one of the most perplexing points of the above 

discussed features of a Humean stance on laws of nature is that of implying a 

truly structureless world, an implication that seems hard to reconcile with a 

scientific image of the natural world, even broadly construed. 

It is from this perspective that we will be concerned with an alternative to the 

descriptive view of laws, namely the so-called primitivist approach5. Basically, 

primitivism about laws can come in two varieties: in its metaphysical dimension, 

laws belong to the fundamental inventory of the world whereas in its conceptual 

dimension, laws are not to be reduced to more primitive notions:  

 

My analysis of laws is no analysis at all. Rather, I suggest we accept laws as 

fundamental entities in our ontology. Or, speaking at the conceptual level, the notion of 

a law cannot be reduced to other more primitive notions” (Maudlin 2007, p. 18).  

 

                                                           
5 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 paper we will focus on the Maudlin version. 



13 

 

When a given notion Ni is chosen as primitive within a set {N1,…, Ni,…,Nn} of 

possible alternatives, the general question arises immediately of what would be 

the deeper motivation for ascribing to Ni rather than to Ni+1 or Ni-1 the status of 

‘primitive’, a question the Maudlin’s reply to which reads as follows: 

 

Taking laws as primitives may appear to be simple surrender in the face of a 

philosophical puzzle. But every account must have primitives. The account must be 

judged on the clarity of the inferences that the primitives warrant and on the degree of 

systematization they reveal among our pre-analytic inferences. (Maudlin 2007, p. 15)6 

 

On the background of the obvious remark, then, according to which there must 

be some entity or notion that is selected as primitive anyway, the choice of laws 

as primitive may have at least two significant strong points. First, the primitive 

status of laws allows one to have a more effective and stimulating confrontation 

with the role of laws within specific scientific theories: the above reference to actual 

scientific practice should be read in this sense. Second, the primitive status of 

laws promises to plausibly accommodate interrelated notions – causation, 

explanation, counterfactuals – that, together with lawhood, appear to form a true 

conceptual network. In fact, laws as primitives can be reasonably seen as able to 

translate causal relations into nomic ones, as grounding counterfactuals if 

similarity of possible worlds is formulated in terms of compatibility with given 

laws and, finally, they can preserve their role in explanation. Let us briefly focus 

on the two last points. 

The connection between laws and counterfactuals is obviously deep, due to 

the issue of whether modality is intrinsic to the very notion of law or not. As we 

recalled earlier, laws are intuitively supposed to account not only for the actual 

phenomena but also for possible ones, and are supposed to account for the 

uniformity of the natural world in terms of non-contingent, law-governed 

processes.  In the Maudlin version of primitivism, laws as primitives can handle 

                                                           
6 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, i.e. Fundamental Law 
Of Temporal Evolution (Maudlin 2007, p. 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. 



14 

 

this modal core of the lawhood intuition through their capacity of generating 

(classes of) models. A given law gives rise to possible worlds to the extent that it 

sets boundaries to the occurrence of phenomena. All that according to precise 

prescriptions fixed by the law remains within the boundaries is ‘possible’: “the 

possible worlds consistent with a set of laws are described by the models of a 

theory that formulates those laws” (Maudlin 2007, p. 18). If setting the 

boundaries for the validity of a law is what allows for possibility in the 

primitivist framework, necessity is obtained at a very low cost: since it is the 

very compatibility with laws that generates a set of models, the laws themselves 

must hold in all models of the set and therefore display a nomic necessity (in the 

usual, possible-world language, Maudlin 2007, p. 21)7. As to the connection with 

explanation, the primitive status of laws allows us to select any account of 

explanation we like in which laws play a sufficiently crucial role, without 

worrying whether the plausibility of the model of explanation we selected is 

threatened or not by some more fundamental notion in terms of which the notion 

of law is reduced: we have explanation of an event f whenever we have a nomic 

subsumption of f under the relevant set of laws L, namely whenever we may 

show that, given L, the event f is what we should expect (Maudlin 2007, pp. 34 

ff). 

 Since no wide-ranging philosophical view concerning such a deep issue as the 

issue of laws on nature can go unchallenged, let us take into consideration some 

possible objections to the primitivist account. A first point concerns the status of 

laws in terms of their alleged ‘fundamentality’. At a given stage of development 

of a scientific theory, we may have reasons to think that a given law is  

fundamental, a circumstance that seems to go along well with the claim that laws 

are primitive endowments of the natural world’s ontology. History of science, 

however, has taught us that laws that were supposed to be fundamental turned 

out to be only special cases of more general laws and still history of science, 

                                                           
7 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 say – 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.   



15 

 

jointly with philosophy of science, suggests that there are no reasons to think 

that there is a foreseeable end to this process of ever-increasing generality. How 

are we to cope with this problem? It hardly looks reasonable to assume that the 

whole network of laws is primitive from the start: would not it be awkward to 

suppose that, for instance, Kepler’s laws are as primitive as Newton’s laws, and 

these in turn as primitive as the Einstein’s field equations of general relativity? 

Although clearly primitive and fundamental are not equivalent concepts, one 

might think that the conventionality inherent in selecting which laws are 

supposed to be fundamental and which derivative might be at least disturbing for 

primitivism. On the other hand, in different areas of science results have been 

obtained that prove in principle our inability to grasp some kind of knowledge – 

from the undecidability theorems to the black hole information loss theorems –  

so that the idea that we might be unable to access part or the totality of the really 

fundamental laws need not contradict the possibly primitive ontological status of 

some of these laws. 

 A second point has been raised concerning the alleged failure of the 

primitivist account in its explanatory power. According to Tyler Hildebrand, for 

instance, the primitivist approach fails in this respect exactly like Humeanism 

(Hildebrand 2013). As we recalled earlier Humeanism – in accepting any 

distribution of properties as possible (according to the recombination principle) – 

has no resources to explain the circumstance that there seems to be a certain 

uniformity among distributions and that not all distributions seem to be on a par 

in this respect. Hildebrand argues that primitivism about laws suffers from an 

essentially identical problem: according to Hildebrand, taking laws as primitives 

has the two following, significant implications: 

 

First, PGL [what Hildebrand calls the “Primitive Governing Laws” view] allows no 

explanation of why we have the laws we do as opposed to other possible laws. If laws are 

wholly primitive, on what basis can we explain why we have one rather that another? 

[…] Second, PGL places no restrictions on the content of laws, so laws could give rise 

to every logically consistent distribution of natural properties. (Hildebrand 2013, p. 5) 

 



16 

 

A preliminary, meta-philosophical kind of remark is that a principled theory 

concerning natural laws is supposed to address a de iure problem – namely, a 

problem like “what is a law of nature?” or “how does a Humean or a non-Humean 

view contribute to a deeper understanding of laws of nature?” – and not a de facto 

problem, like the contingent problem of why we have laws that set constraints 

for certain systems, quantities or processes and not for others. For this reason, 

the latter kind of problem is something for which primitivism need not provide 

an independent explanation. 

In a more substantial vein, Humeanism and primitivism are far from having 

the same problem. While Humeanism – due to the anti-necessitarian flavor of the 

adopted recombination principle – is forced not to have resources to privilege 

certain distributions over others, primitivism is not: in assuming laws as 

belonging to the metaphysical, fundamental inventory of the world, I hold that 

primitivism has the freedom to choose the particular way in which laws are 

supposed to be primitive. In this sense, I propose to include in the irreducible 

property of being primitive for a law the very fact that certain distributions of 

natural properties are privileged over others: otherwise, what does a natural law 

stand for, if not for dictating a certain nomological structure for (a portion of) 

the natural world? According to this proposal, taking laws to be primitive in the 

above sense might recall the Strawsonian sort of solution to the induction 

problem, according to which there cannot be any question as to the rationality of 

the employment of induction, since by being rational we mean, amongst other 

things, using induction. The primitive status of laws would be exactly what 

gives us the right to consider laws as governing entities, since imposing a 

structure over the whole set of possible distributions of natural properties is one 

of the features by which we conceive nomicity for the natural world8. In other 

words, nothing in the formulation of PGL prevents that, if we allow PGL itself 

its Strawsonian role, we might see it somehow negotiating with natural reality 

so as to incline toward certain arrangement of natural properties instead of 

others. 
                                                           
8 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 

 

 

4 Nomic primitivism and the interpretation of the wave 

function in Bohmian mechanics 

 

In addition to its interest by a strictly philosophical viewpoint, the primitivism 

on laws of nature may have a significant connection with a recent option 

discussed in the area of the foundations of contemporary physics, and in 

particular within the issue of the status of the wave function (usually denoted by 

the Greek letter ψ) in quantum theory, in fact, one of the most controversial 

issues in that area (Albert, Ney 2013). As far as the usual meaning attached to 

the traditional ψ in standard quantum mechanics is concerned, a ‘wave function’ 

ψS is, according to textbook presentations, a formal tool that can be assigned to a 

single microsystem S and whose main role is to enable an observer to calculate 

the probability of obtaining one of the permissible outcomes in a well-defined 

experimental setting. There was a time in which some of the founding fathers of 

quantum theory – from Schrödinger to De Broglie, from the early Born to 

Einstein – entertained the idea that the wave function could be some sort of 

physical entity in its own right and not just a convenient computational tool. As 

a matter of fact, it soon became clear that the peculiar status of the mathematical 

space in which the wave function lives prevented this idea to hold in naïvely 

realistic terms and the wave function acquired rapidly the purely informational 

role that the orthodox formulation of quantum mechanics attributed to it for 

decades. In more recent years, however, the obsolescence of the Copenhagen 

orthodoxy and the flowering of interpretational analysis of quantum mechanics 

led to a growing dissatisfaction with a merely epistemic view of the wave 

function, so as to push the pendulum toward radical views in the opposite 

direction: the most notable example is the wave function ontology or configuration 

space realism (Albert 1996), according to which “the fundamental space in which 

entities evolve is not three-dimensional, but instead 3N-dimensional, where N is 

the total number of particles standardly thought to exist in the three-

dimensional universe” (Monton 2002, p. 265).  



18 

 

 In a particular formulation of quantum mechanics, namely Bohmian 

mechanics, an interpretation has been proposed according to which the wave 

function can be conceived as a nomological entity, namely an entity “more in the 

nature of a law than a concrete physical reality” (Goldstein, Zanghì 2013, p. 96). 

This interpretation proposes somehow a shift from the question of the nature of 

the wave function to the question of its role: namely the issue is not what kind of 

object the wave function is but rather what the wave function does for grounding 

the ontology in a nomological sense (Dürr, Goldstein, Zanghì 2013, Goldstein, 

Zanghì 2013)9. It is this nomological interpretation that naturally raises the 

question of how and to what extent this foundational debate concerning 

quantum theory may connect with the above considerations on the status of 

natural laws  In particular, two basic questions arise:  

1) How exactly should we understand the meaning of ‘nomological’?  

2) Under the hypothesis that such meaning can be made sufficiently clear, can 

such an entity as a ‘nomological’ one be really thought of as part of the physical 

world and in what sense?  

Standard Bohmian mechanics is an observer-free formulation of (non-

relativistic) quantum mechanics, according to which the latter describes quantum 

particles and their trajectories in physical space and time: in doing this, Bohmian 

mechanics is said to provide a space-time ontology of non-relativistic quantum 

mechanics, namely a class of well-specified kind of objects and properties 

displayed in space-time that quantum mechanics is supposed to be about: being 

the primary target of the theory, this space-time ontology is called primitive 

ontology 10. In this framework, the wave function plays a crucial role: in addition 

to satisfying the Schrödinger’s equation, the wave function determines the 

particles’ motion via the especially Bohmian addition to the ordinary structure of 

                                                           
9 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 BM, see Laudisa 2012. 
10 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, Zanghì 2013. 



19 

 

quantum mechanics, namely the so-called guiding equation. On the basis both of 

this central role of the wave function in its basic formulation and its generically 

‘realistic’ flavor – the theory is ‘observer-free’ to the extent to which all 

observation-related notions are derivative and not primary in the theory – a 

decisive question that deals with the very foundations of Bohmian mechanics is 

then how to make sense of the wave function, without incurring either into the 

Schylla of a purely instrumental view of it on one side or into the Charybdis of a 

heavy ‘object-like’ view of it on the other.  

 As far as the usual meaning attached to the traditional ψ in standard quantum 

mechanics is concerned, a ‘wave function’ ψS is, according to textbook 

presentations, a formal tool that can be assigned to a single microsystem S and 

whose main role is to enable an observer to calculate the probability of obtaining 

one of the permissible outcomes in a well-defined experimental setting.  If we 

start thinking differently about the ψS, however, as is the case with the Everett 

interpretation or the Bohmian mechanics, the only system S to which a ψS can be 

genuinely attached is the whole universe, a circumstance that yields deep 

implications. In Bohmian mechanics the wave function of the universe – let us 

denote it with Ψ – plays a crucial role at different levels. It is Ψ that the basic 

equations of the theory govern in a fundamental sense and when we discuss the 

status of the wave function as a nomological entity, it is Ψ that we are primarily 

talking about.  

Within this framework, conceiving the Ψ in nomological terms presents some 

remarkable advantages. First, the nomological interpretation allows one to 

justify the circumstance according to which the particles’ motion is affected by 

the wave function without affecting it back: when we think of a ‘law’ determining 

in some sense the phenomena falling under its application domain, we do not 

expect any reaction behavior by the law-governed phenomena on the law itself! 

Second, the nomological interpretation seems able to accommodate more 

naturally the celebrated problem of time in quantum cosmology. If the Ψ as a law 

is supposed to aptly govern the particles’ motion and the related properties, only 

the latter are supposed to change over time and not the wave function: if it is a 



20 

 

law that governs the way in which things in the world change, we expect the 

law-governed things to depend on time, not the law itself: “from a Bohmian 

perspective, the timelessness of Ψ is not a problem. Rather, it is just what the 

doctor ordered.” (Goldstein, Zanghì 2013, p. 100). Third, the nomological 

interpretation seems especially suitable for a further feature of Bohmian 

mechanics, namely the possibility to recover from Ψ the wave functions ψS′, 

ψS′′,…,  that can be assigned to well-defined subsystems S′, S′′, … of the 

universe in a suitable sense. This possibility obtains in Bohmian mechanics 

through the definition of the so-called conditional wave function: I argue that this 

definition, in addition to allowing a well-defined notion of a subsystem of the 

universe, is a further instance of what it might mean for the Ψ to be mainly a 

nomological entity. Let Ψt  be the wave function of the universe at time t and let 

S be a specific system under scrutiny (clearly, S is also a subsystem of the 

universe whose wave function is Ψ). Let us suppose further to decompose the 

configuration of the universe Q = (X, Y) into the configuration X of the system S 

and the configuration Y of the rest of the universe, something that we may call 

the ‘environment’ of S. Then we define the conditional wave function of S at time t 

by the expression 

ψt (x) =  Ψt (x, Y) 

A remarkable implication of this definition is that, under suitable conditions, the 

conditional wave function of S does indeed satisfy the Schrödinger’s equation for 

S, namely, given the wave function of the universe, each of the wave functions 

ψS′, ψS′′,… of the subsystems S′, S′′, … of the universe behaves just like ordinary 

quantum mechanics prescribes they should phenomenologically behave11. 

But in addition to the ‘conservation’ of the state phenomenology, which is an 

important property anyway, there is a further point that is directly relevant to 

our present philosophical discussion:  the joint action of the Ψ and the primitive 

ontology in fixing the state of the ordinary systems that we deal with. Suppose 

an observer O who is operating with a specific system S, on which she is about to 

                                                           
11 A rigorous justification that the conditional wave function is phenomenologically ‘well-
behaved’ can be found in Dürr, Goldstein, Zanghì 1992. 



21 

 

perform a measurement. By assuming a primitive ontology of actual 

configurations that are independent from preparation and measurement 

procedures, BM allows O to concentrate on S by putting to work the whole 

galaxy of the space-time locations of all the subsystems that are ontologically 

fixed anyway but that are not directly relevant to the measurement on S: in a 

sense, the whole ontological web of the (per se existing) space-time locations 

works as a fixed parameter, except for the x that figures as argument of the 

conditional wave function ψ (x). But if we agree that the specification of states for 

subsystems of the universe is a fundamental nomological task of a theory with a 

wide scope such as quantum mechanics is usually assumed to have, the 

nomological role that the wave function takes on in BM is apparent here. The 

Ψ does not belong to the basic endowment of primitive ontology, but it is in its 

‘interplay’ with primitive ontology that it expresses its role of state 

determination: no direct entity-realism concerning the wave function is assumed, 

but rather a primitive-ontology-based theory in which a precise story is told on 

how the wave function – although not belonging itself sic et simpliciter to the 

inventory of primitive ontology – implements an observer-free description of 

quantum systems in terms of such primitive ontology.  

On the background of our preceding discussion on nomic primitivism, at least 

two general points are in order. First, the general view of laws as primitive 

entities may turn out to be well-suited to the role that the wave function as-a-

law occupies in Bohmian mechanics, since such role appears not reducible to 

anything more fundamental in the overall structure of the theory. Second, the 

specific nomic role that the wave function plays in Bohmian mechanics may 

exemplify the idea that a nomic entity like a wave function in Bohmian 

mechanics can be ‘part of’ an ontology without being necessarily conceived as an 

ordinary, stuff-like object. As we have seen, the wave function is not in itself a 

piece of primitive ontology, but it plays the role of generating and making 

possible the ‘displacement’ of primitive ontology and its features in space-time. 

The picture that emerges is, so to say, two-dimensional: the entities included in 

the primitive ontology belong to the dimension of what fundamentally exists in 

space-time, but the wave function belongs to the dimension of what tells 



22 

 

primitive ontology how to behave in space-time. Therefore, a sensible way to do 

justice in philosophical terms to this role of the wave function is to take the wave 

function itself as primitive, although in a nomic sense. 

 But what is this ‘two-dimensional’ talk I am referring to? Basically, it is a way 

to contemplate two different fashions in which something can be part of physical 

reality according to a Bohmian perspective: the first pertaining to the primitive 

ontology, the second to the wave function as-a-law. Unless we allow ourselves to 

distinguish at least these two different ways of being part of physical reality, we 

can hardly avoid the conclusion that the wave function, if it is to be something 

more than a computational tool, is real to the same extent as the entities of the 

primitive ontology, something that in Bohmian mechanics makes no sense. It 

might be correct to claim, as Esfeld et al. 2014 do, 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, p. 780). But no sharp distinction does not imply no distinction 

at all. We may think to have a sort of plane, spanned by two dimensions: that of 

the primitive ontology entities XPO and that of the nomological entities YN: each 

point in the XPO−YN plane represents an implementation of a possible behavior 

of PO in terms of a nomological entity (like the wave function). It is plausible to 

hold that what the theory is about is not just PO but PO plus the nomological 

entities in terms of which PO behaves, and nomological entities such as the wave 

function are essential in order for the behavior of PO to be described by the 

theory: as Esfeld et al. themselves say “we should be open-minded enough to 

concede that the ontology of a physical theory may contain other kinds of objects 

than just particles and fields” (Esfeld et al. 2014, p. 778)12. 

 

 

                                                           
12 In their recent paper, Esfeld, Lazarovici, Hubert and Dürr discuss two different options about 
the lawhood intuition that is supposed to ground the nomological interpretation of the wave 
function in Bohmian mechanics: “the Humean one that regards the law as a contingent regularity 
and the one that anchors the law in a disposition of motion of the particles” (Esfeld et al. 2013). In 
my opinion, neither is convincing and more persuasive than the primitivist option, but a detailed 
analysis and support for this claim are outside the scope of the present paper.  



23 

 

5 Conclusions 

 

On the 6th of April, 1922 Albert Einstein was invited to join at the Société 

Française de Philosophie. Asked of his opinion about the relation of his relativity 

theory with the epistemology of Ernst Mach – an epistemology which is 

remarkably similar to the regularist framework in important respects – Einstein 

replied: 

 

There does not appear to be a great relation from the logical point of view between the 

theory of relativity and Mach’s theory. For Mach, there are two points to distinguish: on 

one hand there are the immediate data of experience, things we cannot touch; on the 

other there are concepts which we can modify. Mach’s system studies the existing 

relations between data of experience; for Mach, science is the totality of these relations. 

That point of view is wrong and, in fact, what Mach has done is to make a catalogue, not a 

system. (Einstein 1923, p. 253, emphasis added) 

 

What in memory of the Einsteinian claim we can also call the catalogue view of 

laws proposed by the regularists turns out to be implausible for analogous 

reasons. This view is logically consistent but implausible in that it misses what 

makes a scientific theory of natural phenomena a system rather than a list. It is 

not true that to assume that there is an ordered structure for phenomena – that 

order being a consequence of the operations of laws – is a requirement defensible 

only in analogy with the use of the notion of law in other domains: it is simply 

that nature turns out to be incomprehensible without such a structure. 

 

 

 

 

 

 

 



24 

 

References 

 

Albert D. 1996, “Elementary Quantum Metaphysics”, in J. Cushing, A. Fine, S. 

Goldstein (eds.), Bohmian Mechanics and Quantum Theory: An Appraisal, 

Dordrecht, Kluwer, pp. 277-284. 

Albert D., Ney A. (eds.) 2013, The Wave Function. Essays on the Metaphysics of 

Quantum Mechanics, Oxford University Press, Oxford. 

Beebee H. 2000, “The Non-Governing Conception of Laws of Nature”, Philosophy 

and Phenomenological Research 61 (3), pp. 571-594 (reprinted in J. Carroll (ed.), 

Readings on Laws of Nature, University of Pittsburgh Press, Pittsburgh 2004, pp. 

250-276, page references are to the reprinted edition). 

Carroll J. 1994, Laws of Nature, Cambridge University Press, Cambridge. 

Carroll J. 2008, “Nailed to Hume's Cross?”, in J. Hawthorne, T. Sider and D. 

Zimmerman, (eds.), Contemporary Debates in Metaphysics, Basil Blackwell, Oxford. 

Cohen J., Callender C. 2009, “A Better Best System Account of Lawhood”, 

Philosophical Studies, 145, pp. 1-34. 

Dürr D., Goldstein S., Zanghì N. 1992,  “Quantum Equilibrium and the Origin of 

Absolute Uncertainty”, Journal of Statistical Physics 67, pp. 843-907. 

Dürr D., Goldstein S., Zanghì N. 2013, Quantum Physics without Quantum 

Philosophy, Berlin, Springer. 

Esfeld M., Lazarovici D., Hubert M., Dürr D. 2014, “The Ontology of Bohmian 

Mechanics”, British Journal for the Philosophy of Science 65, pp. 773-796. 

Fine A. 1996, The Shaky Game: Einstein, Realism and the Quantum Theory, Chicago, 

University of Chicago Press. 

Frank P. 1949, “Einstein’s Philosophy of Science”, Review of Modern Physics 21, 

pp. 349-355. 

Giere R. 1999, Science Without Laws, University of Chicago Press, Chicago. 

Hall N. 2012, “Humean Reductionism About Laws of Nature”, Unpublished 

manuscript. 

Goldstein S., Zanghì N. 2013, “Reality and the Role of the Wave Function in 

Quantum Theory”, in Albert, Ney 2013, pp. 91-109. 



25 

 

Hall N. 2015, “Humean Reductionism About Laws of Nature”, in B. Loewer, J. 

Schaffer (eds.), A Companion to David Lewis, Wiley & Sons, pp. 262-277. 

Hildebrand T. 2013, “Can Primitive Laws Explain?”, Philosophers’ Imprint, 13, pp. 

1-15. 

Lange M. 2000, Natural Laws in Scientific Practice, Oxford University Press, 

Oxford.  

Lange M., 2009, Laws and Lawmakers, Oxford University Press, New York. 

Laudisa F. 2012, “The Uninvited Guest: 'Local Realism' and the Bell Theorem”, 

in De Regt H., Hartmann S., Okasha S. (eds.), EPSA Philosophy of Science: 

Amsterdam 2009, Springer, Berlin, pp. 137-150.  

Lewis D.K. 1973, Counterfactuals, Harvard University Press, Cambridge Mass. 

Maudlin T. 2007, Metaphysics within Physics, Oxford University Press, New York. 

Menzies P. 1993, “Laws, Modality, and Humean Supervenience”, in J. Bacon, K. 

Campbell & L. Reinhardt (eds.), Ontology, Causality and Mind: Essays in Honour of 

D. M. Armstrong, Cambridge University Press, Cambridge, pp. 195-224. 

Monton B. 2002, “Wave Function Ontology”, Synthese 130, pp. 256-277. 

Mumford S. 2004, Laws in Nature, Routledge, London. 

Van Fraassen B. 1989, Laws and Symmetry, Clarendon Press, Oxford.