Instant Download Oxford Studies in Metaphysics Karen Bennett PDF All Chapter
Instant Download Oxford Studies in Metaphysics Karen Bennett PDF All Chapter
com
OR CLICK BUTTON
DOWLOAD EBOOK
https://ebookmass.com/product/oxford-studies-in-metaphysics-
karen-bennett-editor/
https://ebookmass.com/product/oxford-studies-in-ancient-
philosophy-volume-lvi-caston/
https://ebookmass.com/product/perspectival-realism-michela-
massimi/
https://ebookmass.com/product/oxford-studies-in-
metaethics-13-russ-shafer-landau/
Oxford Studies in Medieval Philosophy Volume 10 Robert
Pasnau
https://ebookmass.com/product/oxford-studies-in-medieval-
philosophy-volume-10-robert-pasnau/
https://ebookmass.com/product/oxford-studies-in-philosophy-of-
law-john-e-gardner/
https://ebookmass.com/product/oxford-studies-in-philosophy-of-
language-volume-3-lepore/
https://ebookmass.com/product/oxford-studies-in-political-
philosophy-volume-7-david-sobel/
https://ebookmass.com/product/oxford-studies-in-political-
philosophy-volume-4-david-sobel/
OXFORD STUDIES IN METAPHYSICS OXFORD STUDIES
Editorial Adviso1y Board IN METAPHYSICS
Elizabeth Barnes (University of Virginia)
Ross Cameron (University of Virginia)
David Chalmers (New York University and Australasian National University) Volume 11
Andrew Cortens (Boise State University)
Tamar Szabo Gendler (Yale University)
Sally Haslanger (Min
John Hawthorne (University of Southern California)
Mark Heller (Syracuse University)
Edited by
Hud Hudson (Western \Vashington University)
Kathrin Koslicki (University of Alberta) Karen Bennett
Kris McDaniel (Syracuse University) and
Trenton Merricks (University of Virginia) Dean W. Zimmerman
Kevin Mulligan (Universite de Geneve)
Laurie Paul (Yale University)
Jonathan Schaffer (Rutgers University)
Theodore Sider (Rutgers University)
Jason Turner (University of Arizona)
Timothy Williamson (Oxford University)
Managing Editor
Christopher Hauser (Rutgers University)
OXFORD
UNIVERSITY PRESS
OXFORD
UNIVERSITY PRESS PREFACE
Grear Clarendon Street, Oxford, OX2 6DP.
United Kingdom
Oxford Unin~rsity Press is a dcpamnem of the University of Oxford. Oxford Studies in Metaphysics is dedicated to the timely publication of new
It furthers the University's objective of excel!ence in research, scholarship. work in metaphysics, broadly construed. The subject is taken to include not
and education by publishing worldwide. Oxford is a registered trade mark of
Oxford University Press in the UK and in certain mher countries
only perennially central topics (e.g. modality, ontology, and mereology) but
© the several conrribmors 2018
also metaphysical questions that emerge within other subfields (e.g. phil-
osophy of mind, philosophy of science, and philosophy of religion). Each
The moral rights of the authors have been asserted
volume also contains an essay by the winner of the Sanders Prize in
First Edition published in 20 l 8
Impression: 1
Metaphysics, an biennial award described within.
All rights reserved. No part of this publication may be reproduced. stored in K. B. & D. W. Z.
a retrieval system, or transmitted, in any form or by any means. without the
prior permission in wriring of Oxford University Press, or as expressly permitted New Brunswick, NJ
by !aw, by licence or under terms agreed with the appropriate reprographics
rights organization. Enquiries concerning reproduction outside the scope of the
above should be sent to the Rights Department. Oxford University Press, at the
address above
' The Marc Sanders Foundation is a non-profit organization dedicated to the revival of
systematic philosophy and traditional metaphysics. Information about the Foundation's
other initiatives may be found at <11ttp://www.marcsandersfoundation.com/>.
x The Sanders Prize in Metaphysics
1. INTRODUCTION
physics, either in certain contexts or in general, can be found in Stein (1970; I 977b);
Malament (1976); Horwich (1978); Friedman (1983, 221-3); Earman (1989); DiSalle
(1994); Leeds (1995); Rynasiewicz (1996; 2000); Dorato (2000; 2008); Belot and
Earman (2001, sec. 10.7); Pooley (2013, sec. 6.1, 7); Curiel (2016); Slowik (2016).
Earman (1989) advocates the need for a te11ium quid.
4 Jill North A New Approach to the Relational-Substantival Debate 5
hole argument, all of which live on in today's discussions) may reasonably At the end, I briefly discuss how the conclusion in favor of substantivalism
suggest a stagnated debate. Each of these aims to show that the opposing side may change witl1 future developments in physics. Yet however the physics
recognizes either too few or too many spatiotemporal facts for the physics; but turns out, the question of relationalism versus substantivalism should be
there are various maneuvers, well hashed-out in the literature, allowing each settled by means of the new type of argument offered here. Hence, if I am
side to escape the charge. Relatedly, given the variety of different understand- right, the substantivity of the debate is secured regardless of future develop-
ings of the dispute, you might think that there is no overarching, well-posed ments in physics, while the conclusion in favor of one view or the other will
question in the vicinity (Curiel, 2016). David Malament is not alone in ultimately be decided by the physics.
wondering whether there is any clear-cut dispute between the two sides:
"Both positions as they are usually characterized ... are terribly obscure. After
they are qualified so as to seem intelligible ;md not too implausible, it is hard to
retain a firm grasp on what divides them" ( 197 6, 317). Certainly all of this hints 2. SPATIOTEMPORAL STRUCTURE AND
at "the fragile health of the substantival-relational debate" (Belot, 1999, 38). THE MATCHING PRINCIPLE
These are reasonable concerns when leveled at traditional conceptions of
the dispute. Nonetheless, I believe that there is a debate that is substantive, I'll begin by arguing that both the relationalist and the substantivalist should
not stagnant, and relevant to physics. The debate that I will present is not posit enough, and not too many, spatiotemporal facts for the physics. As
exactly the traditional one. But it is close enough in spirit that I think it is I will put it, they both should countenance the spatiotemporal structure that
the best way of understanding that dispute, updated to take into account is needed for the physics. (In Section 3, I turn to whether they both can
more recent developments in physics and philosophy. And once we frame do this.) I argue that there is a certain methodological principle we are used
the debate in this way, we unearth a novel argument for substantivalism, to relying on in physics, even if it is not usually mentioned. This principle
given current physics. At the same time, that conclusion could be overrid- guides our inferences from the mathematical formulation of a theoty to the
den by future physics. A seemingly subtle shift yields surprising progress on nature of the world according to the theory. I show by example that we do
a longstanding issue that many people feel has stagnated. generally, and successfully, rely on this principle. The conclusion about
In Section 2, I discuss an idea that will play a central role: structure in spatiotemporal structure will follow from it.
general, and spatiotemporal structure in particular. I will argue that, regard- Consider classical Newtonian mechanics. What does this theory tell us
less of whether you are a relationalist or substantivalist, you should think about the world? Newton thought it tells us that absolute space, a space that
that there are objective, determinate spatiotemporal facts about a world: you persists through time, exists. He argued that phenomena involving inertial
should be a realist about spatiotemporal structure in my sense. This follows (unaccelerated) and non-inertial (accelerated, in particular rotated) motion
from a general principle we rely on in physics. (The traditional debate was reveal this. (Think of his bucket experiment and the spinning globes
about the existence of space and time separately. I discuss the question of example.) Although we nowadays agree that the phenomena indicate a
spacetime, or spatiotemporal structure, updating things to the terms of real distinction between inertial and non-inertial motion, we think that
modern physics.) In Section 3, I will argue that, regardless of whether you Newton was wrong about what's required to account for this distinction.
are a relationalist or a substantivalist, you can be a realist about spatiotem- In today's terms, Newton was arguing for substantivalism about what is
poral structure. I do this by framing the debate in terms of fundamentality ofren called Aristotelian, or Newtonian, spacetime. 2 This spacetime has the
and ground, notions that have gotten lots of press recently in metaphysics. structure to support Newton's idea of absolute space, for it has structure that
I show that this way of putting things captures traditional conceptions of identifies spatial locations over time. But we now know (as Newton did not)
the dispute, while allowing us to formulate the most plausible-if not that Galilean, or neo-Newtonian, spacetime also supports the distinction
entirely traditional-versions of the two main positions on it. (Although between accelerated and unaccelerated motion, without absolute space.
I put things in terms of ground, what's most important is that we make use Spelling this out, Aristotelian spacetime has all the structure of Galilean
of some notion of relative fundamentality.) Finally (Sections 4 and 5), I put spacetime, but it also has absolute space, or an absolute standard of rest or
all the pieces together to show that there is a powerful argument for
substantivalism, or at least a powerful challenge to relationalism, given much 2
Not to be confused with the spacetime that Earman (1989, sec. 2.6) calls 'Aristorel-
of current physics. ian.' I follow Geroch's (1978) use of the 'Aristotelian' and 'Galilean' labels.
6 Jill North A New Approach to the Relational-Substantival Debate 7
preferred rest frame. To remind you of what this means, think of an observer spatiotemporal structure supports a notion or quantity of absolute acceleration
on a platform and another observer on a train moving with constant velocity but not of absolute velocity-"absolute" not in Newton's sense, which assumes
relative to the platform. Each observer feels that he or she is at rest and that the existence of absolute space, but in the sense of being invariant or frame-
the other is moving. Galilean spacetime says that neither one is "correct" or independent. 3
at rest in any absolute, observer-independent sense. Each is simply in All of this suggests that a Galilean spatiotemporal structure is the right
motion relative to the other, and at rest in her own frame of reference. structure for Newton's physics. This is the structure that's required for, or
(Think of a reference frame as a coordinate system attached to an observer, presupposed by, the dynamical laws; the structure that recognizes the spatio-
representing her own point of view.) According to an Aristotelian spatio- temporal facts that the laws do. 4 Newton was wrong to think that a classical
temporal structure, there is an observer- or frame-independent fact, from world must contain absolute space and a concomitant quantity of absolute
among all the observers in constant relative motion, about which one is at velocity: the physics doesn't require it. (If the laws were not invariant under
rest in an absolute, frame-independent sense-namely, the one at rest in changes in inertial frame, then we would infer that extra structure. Such laws
absolute space. For there is a frame-independent fact about whether a given would implicitly refer to a preferred frame.) Notice that we reached this
spatial location is the same location over time, so that an object located there conclusion about the structure needed for the laws independently of the
is at absolute rest. In other words, there is a preferred rest frame: the one relational-substantival debate, an idea that I will return to soon. 5
that's at rest with respect to absolute space. First let me say a bit about "structure." On my understanding (and as it is
Intuitively, an Aristotelian spatiotemporal structure has more structure than often used in physics and mathematical physics), 6 structure has to do with
a Galilean one. It has all the same structure, plus an additional absolute-space, the invariant features or quantities, which are the same in all allowable
or absolute-velocity, structure. It recognizes all the same spatiotemporal facts, reference frames or coordinate systems. Inertial structure, for example, is
but it also says that there are facts about how fast an object is moving with part of a classical spatiotemporal structure: there is an absolute, frame-
respect to absolute space. independent notion of accelerated versus unaccelerated motion. But there
It turns out that these additional facts are not needed for, or recognized is no "absolute-velocity structure." An object's velocity depends on the
by, the physics here. Newton's laws are the same in any inertial frame-they inertial frame we use to describe it. Since Newton's laws are invariant
are invariant under changes in inertial frame-which means that they can be under changes in inertial frame, we infer that the choice of frame is an
formulated without mentioning or presupposing a preferred frame. Since a arbitrary choice in description, and that any quantity depending on that
preferred frame isn't needed in the mathematical formulation of the laws, we choice, like velocity, is merely frame-dependent, not out there in the world
infer that it doesn't correspond to anything physical in the world. An apart from that choice.
absolute standard of rest isn't part of the theo1y's, or world's, spatiotemporal Similarly, we think that a choice of origin is just an arbitrary choice in
structure. The physics does not recognize objective, frame-independent facts description, not corresponding to genuine structure in the world. Choose a
about what velocity an object has. Conclusion: Aristotelian spacetime has coordinate system with a different origin, and the laws always remain the
excess, superfluous structure, as far as Newton's laws are concerned. It same. Since the laws are invariant under changes in origin-they "say the
recognizes more spatiotemporal facts than the laws do.
These laws do recognize facts about objects' accelerations (as Newton
argued). Think of Newton's first law: an object travels with uniform velocity 3 I believe that this sense evades Rynasiewicz' s (2000) arguments against the clarity of
unless acted on by a net external force. This law assumes that there is a any absolute/relative distinction.
4
distinction between accelerated and unaccelerated motion, since it tells Although the inference to a Galilean structure is now relatively standard (Earman
things to behave differently depending on whether they are accelerating or (1970); Stein (1970); Huggett (1999, 194-5); Maudlin (2012, ch. 3)), there is room for
debate. Saunders (2013) and Knox (2014), in different ways, argue that Newtonian
not. In terms of spatiotemporal geometry, the law assumes a distinction physics requires a different structure. I continue as though the above inference is correct. It is
between straight and curved trajectories or paths through spacetime, with in any case agreed that absolute space is not needed, and whatever structure is required, the
the straight ones corresponding to inertial motion, the curved ones to non- example illustrates our reliance on the upcoming principle.
5 A similar point is made by Stein (1970, 271-2), although he goes on to say that, "the
inertial motion. And Galilean spacetime has the structure to support this
question whether ... this structure of space-time also 'really exists', surely seems to be super-
distinction. It has an affine connection, or inertial structure, which provides erogatory" (277). In a way I agree, but I also think that there remains a substantive dispute.
6
a standard of straightness for these trajectories. We might put it like this: this More is in North (2009).
8 Jill North A New Approach to the Relational-Substantival Debate 9
same thing" 7 regardless-we infer that this choice is merely a conventional be srated or formulated without assuming it-they wouldn't make sense
or arbitrary choice in description. There is no preferred-location structure without it. 8
in the world, no coordinate-independent fact about whether a given point Two examples illustrate this. Recall Newton's first law, which tells objects
is "really" the origin. By contrast, the laws of Aristotle's physics are not to behave differently depending on whether they are traveling inertially,
invariant in this way. According to them, there is a preferred-location struc- with uniform velociry, or not. This law would not make sense if there
ture in the world-a location toward which certain elements naturally fall weren't a distinction between uniform and accelerated motion: it presup-
and away from which others naturally rise-and preferred coordinate systems poses it. So the world must be such that there is this distinction. The world's
for describing this structure, namely those with an origin at that location. spatiotemporal structure should distinguish between inertial and non-
We likewise think that different choices of unit of measure are conven- inertial trajectories. Assuming that the laws are about the objective nature
tional or arbitrary choices in description. Change from feet to meters or of the world, there must be objective facts about whether objects are
some other unit for measuring distances, for instance, and the physics always traveling inertially or not. 9
remains the same. Since the physics says the same thing regardless, we infer Consider a different example that I'll return to later. If the laws are not
that there is no "preferred-unit-of-measure structure" in the world. time reversal invariant-if they "look different" when we flip the direction
As I see it, structure corresponds to the intrinsic, genuine, objective of time, swapping past and future-then this suggests a structural, physical
features or quantities, which don't depend on arbitraty or conventional distinction in the world between the two temporal directions. Newton's
choices in description. By contrast, frame-, coordinate-, or unit-dependent laws are symmetric in this sense: any behavior allowed by the theo1y can also
quantities depend to some extent on our arbitrary or conventional choices in happen backward in time. The film of any Newtonian process (a ball
description-arbitrary, since according to the physics any choice is equally thrown in the air, billiard balls colliding) run backward also depicts a
legitimate. Such quantities aren't wholly about the world as it is in itself, but process that evolves with the laws. These laws don't distinguish past versus
are in part about our descriptions of the world, whereas structural features future: they say the same thing regardless of the direction of time. By
are agreed upon by all the allowable descriptions, and so correspond to contrast, the second law of thermodynamics says that entropy increases to
genuine features of the world apart from any of those descriptions. No the future, not the past: gases expand, ice melts, not the reverse. A reverse-
matter which description you use, after all, you get the same result. running film shows something disallowed by the law. Non-time reversal
Spatiotemporal structure in particular concerns the intrinsic, genuine, invariant laws like this mention or presuppose the distinction between past
objective spatiotemporal features of a world, which don't depend on arbi- and future, telling things to behave differently depending on the direction of
trary or conventional choices-that two objects are separated by some time. Such laws would not make sense if there weren't a past-future
amount under a Euclidean metric, say, or that a particle's trajectory is distinction in the world, corresponding to an asymmetric temporal struc-
straight according to a given inertial structure. Notice that this idea of ture, or objective facts about past versus future: they presuppose it. (If you
structure is neutral between substantivalism and relationalism. Both of are worried about this conclusion in the case of the second law, stay tuned:
these views can recognize that there is a distinction between spatiotemporal I return to it later in this chapter.)
facts that are more objective, and those that are frame-, observer-, unit-, or Finally, the principle. The above examples are familiar instances of how
coordinate-relative. we draw certain conclusions about the physical world from the laws that
We are still working up to the general principle. Here's an idea that we govern it. These examples all suggest that we rely on a certain methodo-
have reached so far, which will motivate the principle. As we can see from logical principle, which says to posit in the world the structure that's
the inference to a Galilean structure for Newton's laws, any physical theory presupposed by the laws. We generally posit physical structure in the
will constrain, or help dictate, a world's spatiotemporal structure. We infer world corresponding to the mathematical structure needed to formulate
the structure from the physics in this way. This is because any theory will the laws-such as a Galilean spatiotemporal structure for Newton's laws, an
require or presuppose a certain spatiotemporal structure. In particular, it asymmetric temporal structure for non-time reversal invariant laws, or a
will require the structure needed to support the laws, in that the laws cannot
8
Consider Earman' s statement that "laws of motion cannot be written on thin air
alone but require the support of various space-time structures" (1989, 46).
7 9 Compare Maudlin (2012, 9-12); Pooley (2013, sec. 3).
Brading and Castellani (2007) discuss different ways of spelling out this idea.
10 Jill North A New Approach to the Relational-Substantival Debate 11
preferred-location structure for Aristotle's laws. We infer to the world I have argued that the matching principle is a core methodological
whatever the laws presuppose, whatever there must be in the world for the principle we use to guide our inferences from a physical theo1y to the nature
laws to make sense and be true of it. There should be a match in structure of rhe world according to that theory. Now we can see that this principle
between the laws and the world. Theories obeying what I will call the rells us to posit, or countenance, or somehow be able to talk about,
matching principle are "well-tuned," to borrow a phrase that John Earman spariotemporal structure. For the laws generally talk about, they mention
(1989, ch. 3) uses for a somewhat different idea. 10 (I rake it this is motivated or presuppose, a particular spatiotemporal structure. We should counten-
by a kind of realism. I won't argue for realism here.) ance the particular spatiotemporal structure or facts required for the laws;
As with any guiding methodological principle, this principle won't yield ipso facto, we should countenance spatiotemporal structure or facts in
conclusive inferences, yet it is still a reasonable guide. We cannot be certain general. In other words, the matching principle says that we should be
that there is no absolute space in a Newtonian world, bur it is reasonable to realists about spatiotemporal structure, since the laws presuppose such a
infer that there isn't. Or rake special relativity. The marching principle lies rhing, and we should generally posit in the world the structure that's
behind the thought that there is no preferred simultaneity frame. Since the presupposed by the laws.
laws are invariant under changes in Lorentz frame, we infer that there is Importantly, this conclusion is independent of the relational-substantival
no absolute, frame-independent simultaneity relation. We can't be certain debate. Regardless of your position on that debate, the matching principle
about this, and some people argue that we have other reasons to posit this rells you to believe that there are objective facts about the spatiotemporal
structure (for presentism or for certain theories of quantum mechanics, for structure of a world; to recognize the spatiotemporal facts that are recog-
example). Still, we do generally, and reasonably, rely on this principle. We nized by the laws. You should believe that a Newtonian world has a
take it to be successful. A~ the case of special relativity shows, we need an Galilean spatiotemporal structure, for example (although this claim may
extra reason to disobey it. To put it another way: all other things being be understood differently by the relationalist and substantivalist, as I discuss
equal, we should infer a match in structure between laws and world. Those below). Who would reject the principle? The conventionalist, for one, like
who believe in a mismatch are saying that other things are not equal, and Reichenbach or Poincare, who denies that there is an objective fact about
must argue as much. 11 the "right" spatiotemporal structure of a world: there are no objective
It is sometimes said that the reason to posit a Galilean rather than spatiotemporal facts. Against such a view, the matching principle suggests
Aristotelian structure in a Newtonian world is that the latter would yield that spatiotemporal structure is out there in the world. It is not conventional
in-principle undetectable physical facts. 12 Since Newton's laws are invariant or arbitrarily chosen, as is an inertial frame or origin or unit of measure. 13
under changes in inertial frame, no experiment could ever detect which is This structure exists; it is part of reality. There is an objective, determinate
the preferred frame. Choose any frame in which to run your experiment, fact about what spatioremporal structure a world has, evidenced by its laws.
and the laws always predict the same results. That's right. But I think that The marching principle is nor Quine's criterion for ontological commit-
there is a deeper reason for the inference to a Galilean structure, which is the ment. Quine says that we are ontologically committed to what the variables
match between the mathematical structure of the theory and the physical of our theories must range over in order for those theories to be true. This has
structure of the world. This match is part of our evidence that we have to do with ontology, with what entities exist. The marching principle is about
inferred the correct structure of the world. This is a more fundamental what structure we should posit. It says to align physical structure in the world
reason for the inference than the verificationist-sounding principle to avoid with the mathematical structure required to formulate the laws. This has to do
undetectable physical facts. with what spatiotemporal facts we should recognize, which is not simply a
matter of ontology. To see that these come apart, notice first that a given
spatiotemporal structure, say a Galilean one, can be understood by different
10 Earman suggests that there should be a match between the symmetries of the laws
the laws of quantum mechanics trump special relativity when it comes to inferring this 13
We can agree with Reichenbach and Poincare that those things are arbitrary, since
structure. the laws indicate that different choices are equally legitimate. Spatiotemporal structure is
12 Mentioned, with varying support, in Earman (1989, ch. 3); Ismael and van Fraassen
different. We cannot arbitrarily alter the metric, for instance, and keep the laws the same,
(2003); Roberts (2008); Dasgupta (2009); Maudlin (2012, ch. 3); Pooley (2013, secs. 3-4). not without major compensating changes elsewhere.
12 fill North A New Approach to the Relational-Substantival Debate 13
people as involving different entities: by a certain substantivalist 14 as involving You might wonder: How can the relationalist believe in spatiotemporal
points of spacetime and a relationalist as involving material bodies. (As Tim structure? Isn't this the very sort of thing the relationalist rejects? On the
Maudlin (2015) puts it, to attribute "a mathematical structure to physical other hand, if the relationalist can believe in spatiotemporal structure, you
items" is to say that those items "have some physical features that make them might then wonder what could be left for the two views to disagree about.
amenable to precise mathematical description in some respects". In particular, I'll now suggest that the notion of ground gives the sense in which the
it is not yet to say what the items must be.) Second, two people might agree on relationalist as well as the substantivalist can countenance spatiotemporal
what entities exist-say, points of spacetime-but disagree on the spatiotem- structure, and that this yields a real disagreement that's relevant to physics.
poral structure, for instance on whether the points are arranged in a Galilean The basic idea will be this. Both views can countenance, or believe in the
or Aristotelian way. This will become clearer as we proceed. existence of, spatiotemporal structure. (Whether each one is able to recog-
Question: How should we formulate the laws? It seems as though nize the particular structure needed for the laws is a question that I will be
different formulations can presuppose different structures. If so, then in sidestepping here, for reasons to come.) The views differ on what underlies
order to adhere to the matching principle, we will first need to know how to this structure. Essentially, the substantivalist says that spatiotemporal struc-
formulate the laws, which is a big question. Trust me for now that we can ture is fundamental to the physical world, whereas the relationalist says that
make progress in advance of answering this question. I will return to it at the it arises from the relations between and properties of material bodies.
end of this chapter. Putting this in terms of ground. A grounding relation is an explanatory
Some have argued for a third view, neither substantivalist nor relational- relation that captures the way in which one thing depends on or holds in
ist, called 'structural spacetime realism.' 15 Since that view emphasizes real- virtue of another, without implying that the dependent thing doesn't exist.
ism about spacetime structure, you might think that it is what I am Ground captures a "metaphysical because" in answer to questions about
advocating. I don't have space to address the alternative in detail, 16 but why something exists or some fact holds. (I use the general idea, without
I will note that, despite superficial similarities, it is importantly different entering into debates over its metaphysics. I won't take a stand on whether
from my overall approach. First, I claim that both the relationalist and the ground is properly a relation between facts or objects, but deliberately use
substantivalist should (and can: below) be realists about spatiotemporal both ways of talking. It is generally thought that the grounding relation is
structure, whereas spacetime structural realism aims to be distinct from transitive and irreflexive, and that the grounds metaphysically necessitate
either of those views. Second, I understand the idea of spatiotemporal the grounded. None of these assumptions have gone uncontested, but
structure differently, to encompass any objective, intrinsic spatiotemporal I assume them here. 17)
fact about a world. In particular, countenancing spatiotemporal structure in Using the notion of ground, the relationalist and substantivalist can
my sense does not mean eschewing fundamental physical objects (alterna- each say that spatiotemporal structure exists, that there are objective spatio-
tively, intrinsic properties) altogether, nor the possibility of our knowing temporal facts about a world. They disagree on what the spatiotemporal
about such things, as the structural spacetime realist often seems to do. That structure holds in virtue of; what metaphysically explains the spatiotemporal
said, below we will see one way in which my account mirrors certain claims facts. The relationalist says that a world's spatiotemporal structure is
of the spacetime structural realist. grounded in the features and behaviors of material bodies. All the spatio-
temporal facts are grounded in the facts about material bodies. The sub-
stantivalist says that spatiotemporal structure isn't grounded in anything else
3. A DISAGREEMENT ABOUT GROUND more fundamental to the physical world; in particular, it is not grounded
in material bodies. There are fundamental spatiotemporal facts that are not
In order to say that the relationalist and substantivalist both should counten- grounded in facts about material bodies. Both views can countenance spatio-
ance spatiotemporal structure, I must be able to say that they both can do this. temporal structure or facts; they disagree on what, if anything, grounds this
structure or those facts.
14 See Section 3.3.
15 Different versions are in Dorato (2000; 2008); Slowik (2005); Bain (2006); Esfeld
17
and Lam (2008); Ladyman and Ross (2009). Different accounts are in Fine (2001); Schaffer (2009). Rosen (2010) defends
6
1 See Greaves (2011). the idea.
14 Jill North 11 New Approach to the Relational-Substtmtival Debate 15
I spell out the two views more in a moment. First, a few notes on the use
of ground in this context. Jonathan Schaffer (2009, 363) and Shamik 3.1. Relationalism in terms of ground
Dasgupta (20 l l) also suggest that we can understand this debate in terms
The relationalist says that certain material bodies, and various of their
of ground, but they put things a little differently. They say that the
properties and :-eta~ions, a:·e fundamental, and a world's spatiotemporal
relationalist and substantivalist both believe that spacetime exists, while
structure holds m vtrtue of them. All spatiotemporal structure or facts are
differing on what grounds the existence of spacetime. I say that both (can
grounded in (facts about) material bodies. In saying that "certain material
and should) believe that spatiotemporal strncture exists, while differing on
bodies are fundamental," this means whichever material objects turn out to
w?at grounds the ~xisten~e of that structure. I prefer this way of putting
be most fundamental: certain particles, say. (I assume the fundamental
thmgs because, we II see, it allows us to flesh out the competing views in
relations can include spatiotemporal ones, 20 although the relationalist
different ways, all the while maintaining a genuine dispute that the physics
will weigh in on. might want a different kind of relation to be fundamental, causal ones
being a familiar candidate. I leave this open here. The upcoming argument
It may seem unexciting to exchange a debate about the existence of
spacetime for one about the fundamentality of spatiotemporal strncture. rakes aim at all these versions of relationalism equally. 21 )
So, for example, the fact that a world has a Euclidean spatial structure is
~here has been mucl.1 discussion in metaphysics oflate about doing a similar
kmd of exchange with other existence debates (as in Schaffer (2009)), so grounded in, holds in virtue of, the fact that its particles are, and can be,
that this instance may feel like old hat. There have been some related arranged in various ways, with various distance relations between them.
(I return to this "can be" phrase soon.) The world has a Euclidean structure
thoughts about the spacetime debate in recent philosophy of physics as
well. Thus Carl Hoefer ( 1998) frames the question in terms of fundamen- because (in the metaphysical sense) its particles are, and can be, arranged in
those ways; this is what the spatial structure consists in. Similarly, the fact
tality, ,~s that of how "to understand the basic ontology of the physical
that a Newtonian world has a Galilean spatiotemporal structure is grounded
wo~ld, alt~ough he formulates aspects of the dispute more traditionally,
saymg for mstance that substantivalism is committed to the existence of in the fact that its particles do, and can, behave in various ways, with various
"a substantial, quasi-absolute entity." 18 Gordon Belot (1999; 2000; 2011) spatiotemporal relations between them. The fact that a world has a particu-
says that the relationalist, like the substantivalist, can be a realist in the sense lar spatiotemporal structure is made true by the facts about material bodies.
of''.attribut[ing] to reality a determinate spatial structure," while disagreeing A world has the spatiotemporal structure it does because material bodies
on 'the nat~re oft.he existence of space" (2011, 1). 19 This is close to my own (can) behave in certain ways.
way o.f putting thmgs, although his account is not spelled out in the same Three notes on this use of ground. First, a grounding explanation is
importantly different from a causal explanation. In Kit Fine's words, ground
way (it does not use notions like ground or my conception of spatiotem-
yields "a distinctive kind of metaphysical explanation," in which the objects
poral structure, and it focuses on certain traditional examples), nor does
or facts are connected by "some constitutive form of determination" (2012, 37).
he dra""'. the sam~ conclusions. The more prevalent attitude in philosophy
Particle behaviors don't cause a Euclidean spatial structure. This is rather
of phys:cs, es~ec1ally among those who complain about the substantivity
what the spatial structure consists in or depends on, in a metaphysical
of the dispute, 1s that the debate concerns the existence question. So although
sense. Compare this to more familiar cases, such as the grounding of
my proposed way of understanding of the dispute is not without prece-
facts about the macroscopic world in facts about subatomic particles, or
d~nt, e:en then there are differences, and it is anyway not the prevalent
v1ewpomt. If you disagree with that assessment, though, it will soon be the grounding of mental facts in non-mental facts, or moral facts in non-
moral facts. Ground captures this metaphysical "in virtue of" explanation. 22
clear that novel avenues of argument open up once we are completely
explicit about this shift. As I understand it, when we say that "the fact that x grounds the fact that y,"
. is Hoefer simi!arly argues that this is a substantive dispute, which is likely to remain so
wit.h furu:e ph~sics, and that general relativity supports substantivalism. Yet he puts 2
° Contra Nerlich (1994a, ch. 1).
vanous thmgs differently from how I do, drawing these conclusions for different reasons. 2
i I also assume that t~e objects and relations are equally fundamental, though there may
i 9 Be!ot also says that his formulation, while unorthodox, yields a debate that is be a view with only one rundamental "ontological category" in the sense of Paul (2013).
substannve, relevant to physics, and reminiscent of the traditional dispute. 22 Loewer (2001) discusses the relevant sense of "in virtue of."
16 Jill North A New Approach to the Relational-Substantival Debate 17
this just means that "the fact that y holds in virtue of the fact that x"; i.e. that about material bodies. 24 As we'll see, I think there is an argument for
the holding of the grounded fact consists in nothing more than the holding of subsranrivalism that goes through even if we grant the relationalist the
the grounding fact. ability to ground all the relevant facts in ones she takes to be fundamental.
Second, ground aims to give a "looser" connection between the facts or So for the purposes of that argument, I am going to grant the relationalist
objects involved than that given by a definition. An analogy: I am thinking that ability.
of ground in such a way that it can articulate the view that the biological It is worth mentioning one thing that I do think will be required to
facts are nothing over and above the facts about these systems' particles. ground that structure, which is some version of "modal relationalism."
(You may not hold such a view, but ground can specify what it amounts to.) I suspect that the relationalist will have to countenance facts not only
The history of failed attempts in twentieth-century philosophy of science to about the actual features and behaviors of material bodies, but about their
spell out a "tighter" connection between the reduced and reducing facts by possible ones as well-facts about what spatiotemporal relations can hold, in
means of correspondence rules that define the biological quantities in terms some sense. This is because the actually instantiated relations won't in
of physical ones suggests that this won't work. Yet there is still a way of general suffice to fix the full spatiotemporal structure required for the
capturing the sense in which the biological facts "are nothing but" the physics. (As long as the relationalist can embed the actual relations uniquely
physical facts, which is to say that the biological facts are grounded in the into a certain structure, it seems as though she can talk of the spatiotemporal
physical ones. In an analogous way, the relationalist can say that the facts structure of a world. The problem is that the actual relations may not
about spatiotemporal structure are "nothing but"-are grounded in-the uniquely fix the structure (up to isomorphism) needed for making predic-
facts about material bodies, even if she can't explicitly define the spatiotem- tions about material bodies. 25 ) In order to adhere to the matching principle,
poral structure in terms of the relations between material bodies. A grounding the relationalist will have to go modal. I refer you to Carolyn Brighouse
relation can hold even in the absence of a definitional connection. (This is one (1999) and Belot (2011) for discussion of ways the relationalist might do
reason the notion of ground can help the relationalist, since finding such this and what sort of modality may be involved. 26
explicit definitions is notoriously difficult. Of course, it is not easy to give an (Modal relationalism arguably allows the view to countenance vacuum
account of the grounding of spatiotemporal structure in material bodies worlds, which seem possible according to both classical and relativistic
either, but replacing the definitional requirement with the looser constraints physics. Such worlds contain no material bodies and yet can have a spatio-
of ground can ease some of the burden.) temporal structure. Now, it is open to the relationalist to deny that vacuum
Third, there must be some account of how the facts that the relationalist models correspond to physically possible worlds. Nonetheless, the modal
takes to be fundamental manage to ground all the spatiotemporal facts relationalist should be able to allow for these possibilities. All the facts about
needed for the physics. (For instance, there can't be two worlds with the spatiotemporal structure will still be grounded in facts about material
same fundamental relationalist facts but different spatiotemporal structures, bodies-in facts about how these bodies would behave, if there were any.
since the fundamental facts necessitate the grounded facts.) Simply being a Such a relationalist can arguably even countenance different spatiotemporal
realist about spatiotemporal structure does not guarantee the ability to structures in different vacuum worlds, as general relativity seems to allow
generate the particular structure required by the laws as the matching for. This is not to say exactly how the relationalist can do this, just as I haven't
principle demands. You might be skeptical that the relationalist can do
this. Much of the literature is taken up with this question of how, and
whether, the relationalist's more meager ontology can recognize all the 24 From this perspective, those such as Manders (1982); Mundy (1983; 1992);
spatiotemporal facts we want. 2 3 Huggett (2006); Belot (2011) can be seen as giving accounts of how this grounding
This is a big question, but I won't try to answer it here. I won't try to tell project might go.
25 Examples are in Mundy (1986); Maudlin (1993, 193-4, 199-200); Nerlich
you exactly how the relationalist grounds all the spatiotemporal facts in facts (1994a); Belot (2000; 2011, ch. 2). Field (1984) argues that the modal view is necessary
for the relationalist to solve the problem of quantity. An alternative is conventionalism
(Earman, 1989, sec. 8.6).
26
The view may sound newfungled, bur even Leibniz, according to many, held it: Belot
23 (2011, Appendix D). The liberalized relationalism of Teller (1991) is a precursor to more
A repeated complaint against the varieties of relationalism surveyed by Pooley
(2013) is that the relationalist's resources are too thin to yield predictions of the recent versions. See also Sklar (1974, IILB2); Horwich (1978); Mundy (1986). Objections
phenomena. are in Malamcnt (1976); Field (1984); Earman (1989, sec. 6.12); Nerlich (1994a).
18 Jill North A New Approach to the Relational-Substantiva! Debate 19
said how the relationalist can ground any particular structure in material
bodies. Yet once we grant the (modal) relationalist the ability to ground all 3.2. Substantivalism in terms of ground
the spatiotemporal facts in facts about material bodies, there needn't be a
The substantivalist denies that ,all spatio temporal facts hol~ in virtue of fac.ts
special problem for vacuum worlds.)
Keep in mind that the relationalist might not deny the fundamentality of abou t nnterial
, bodies . A worlds spatiotemporal
, structure Is not grounded . m
c r·es and behaviors of material bodies. fhe fact that a world has a given
any spatiotemporal fact or structure. Depending on the version of the view reatu . ld .
iotemporal structure is a fundamental fact about the physical wor ; m
(see the beginning of this subsection), the fundamental facts may include
spat·cular it is not grounded in facts about material bodies. (Clarifications
ones such as that two particles are separated by some distance, or that one part! ' d
below.) The facts about a world's spatio~emporal structure, .in turn.' groun
particle lies between two others. 27 What's important is that the relationalist
the facts about the spatiotemporal relations b.etween material bod1e:. (The
only allows certain kinds of spatiotemporal facts (if any) to be fundamental,
c mer may only partially ground the latter, smce the grounds may mclude
namely those that essentially involve material bodies and their relations- ror b . .
occupation relations that ma~erial bodie: ear to sp~cet1me pomts or
facts that the substantivalist takes to be nonfundamental. The fact that
·ons depending on the version of the view-see Sect10n 3.3.)
a world has a given spatiotemporal structure is grounded in the facts reg l ' . .
For example, the fact that two particles are some distance apart is
about material bodies, even though these latter facts may include certain
grounded in, made true by, the fact that they are separated b~ th~t ~mount
spatiotemporal ones. More exactly, there is no fundamental spatiotem-
according to the fundamental metric structure (where the metric w11l 1tselfbe
poral fact or structure apart from the structure of, or facts about, material
understood in different ways by different substantivalists-see Section 3.3-
bodies. For ease of exposition, I put this as the claim that all spatiotem-
but will in any case not be grounded in features of material bodies). The fact
poral facts are grounded in facts about material bodies. All spatiotemporal
that a particle is traveling inertially in a Newtonian world is likewise gro~nde.d
structure is grounded in the relations between and properties of material
in facts about the fundamental spatiotemporal structure: the particle is
bodies.
following a straight trajectory because (in the metaphysical sense) its path is
So, using the notion of ground, the relationalist can say that there are
straight according to the world's Galilean structure. (The substantivalist then
facts about a world's spatiotemporal structure, which are distinct from
recognizes nonfundamental spatiotemporal facts or structure of a sort, ab?ut
the facts about material bodies and their relations, but are also nothing
the spatiotemporal relations between material bodies. More exactly, the view
over and above those facts about material bodies-just as one might say
holds that there are fundamental spatiotemporal facts or structure not
that there are real facts about macroscopic systems, which are distinct
grounded in (facts about) material b.odies. Notice t.hat. ce:tain fact~ abo~t
from the facts about their particles, but are also nothing over and above
material bodies, for instance about thell' fundamental 111tr1ns1c properties, wt!!
the facts about the particles.
be fundamental. What's not fundamental are the spatiotemporal facts about
This is a non-standard (if not wholly unprecedented) way of formulating
them.) By contrast, for the relationalist, a world's spatiotemporal structure is
relationalism, which captures traditional thoughts about the view, for
Galilean because the particles behave in certain ways. On that view, the facts
instance that spacetime doesn't "really exist": "spacetime" is nothing but
about material bodies metaphysically explain the fact that a world has the
various features of material bodies; certain material bodies are fundamental,
given structure.
and any spatiotemporal talk or fact is really about them. At the same time,
For the substantivalist, facts about the spatiotemporal relations between
this formulation allows the relationalist to say that spatiotemporal structure
material bodies are nothing over and above facts about how these objects are
exists, that there are objective truths about what spatiotemporal structure a
arranged according to a given spatiotemporal structure. Facts about. a
world has, as the matching principle says we should do. It's just that these
world's spatiotemporal structure, on the other hand, are not grounded 111
things all hold in virtue of what's true about material bodies.
facts about material bodies, and in that way are "over and above" any facts
about material bodies. This captures the traditional conception of the view
as holding that spacetime exists "independently of" material bodies: there is
spatiotemporal structure that is not metaphysically due to material bodies.
2
7 Which of these depends on whether the relationalist thinks that fundamental You may worry that this conception of substantivalism is already discon-
relations can be quantitative. firmed by our current best theory of spacetime. According to general relativity,
20 Jill North A New Approach to the Relational-Substantival Debate 21
the presence of matter affects the local spatiotemporal geometty, which in structure are not fundamental but grounded in facts about God. Yet
turn affects the behavior of matter; whereas on my conception of substan- Newton is still a substantivalist, on my understanding, since the facts
tivalism, there is spatiotemporal structure that is independent of matter. about the spatial structure are more fundamental than the facts about
This worry is evaded by noticing that the interdependence between spa- bodies' spatial relations. 28 To put it another way: the facts about the spatial
tiotemporal structure and material bodies in general relativity is of a structure are fundamental to the physical realm. Analogously, the relationalist
different, causal or nomological, kind from that given by ground. Although will say that all spatiotemporal facts are grounded in facts about material
the substantivalist says that there is spatiotemporal structure that is inde- bodies, regardless of her other metaphysical views, such as whether there is
pendent of material bodies in not being grounded in them-these facts something yet-more-fundamental that lies outside the physical realm. The
about spatiotemporal structure are "metaphysically over and above" the views still disagree over whether spatiotemporal structure apart from mater-
facts about material bodies-she can still allow that the behavior of material ial bodies is fundamental to the physical world. For ease of presentation,
bodies causes a certain spatiotemporal structure in accord with the physical I continue to put the dispute as the question of whether spatiotemporal
laws. Compare: although the dualist says that mental events are not grounded structure is fundamental (to the physical world).
in physical events-mental events are "metaphysically over and above" physical What if there is no fundamental physical level? In that case, the views
ones-she can still allow that physical events cause mental events in accord might still be distinguished by means of the relative fundamentality of the
with the scientific laws. behaviors of material bodies and a world's spatiotemporal structure, depend-
Substantivalism and relationalism, as I understand them, disagree about ing on the details. This may suggest that the debate should be framed in
the fundamental nature ofthe physical world. They both countenance spatio- terms of relative fundamentality. Substantivalism would then be the view
temporal structure or facts, but disagree on whether all such structure or that the facts about a world's spatiotemporal structure are more fundamental
facts hold in virtue of material bodies. Both views can recognize the fact that than the spatiotemporal facts about material bodies, and relationalism
two particles are separated by some distance under a Euclidean metric, for would be the view that the facts about material bodies are more fundamental
instance, or that a world has a Euclidean metric structure. But they will than the facts about spatiotemporal structure. But I don't want to put it this
disagree on whether the metric is itself fundamental or grounded in the way. That way of putting things would imply that either relationalism or
behavior of material bodies. To borrow a phrase that Helen Beebee uses for a substantivalism is bound to be true, regardless of future physics, so long as
different debate, these views "have completely opposite conceptions of what the two kinds of facts are not equally fundamental. Yet intuitively, if nothing
provides the metaphysical basis for what" (2000, 580). The substantivalist like either spatiotemporal structure or material bodies turns out to be funda-
sees a world's spatiotemporal structure as the metaphysical basis for the mental to the physical world, then neither view has been vindicated. You
spatiotemporal relations between material bodies. The relationalist sees could insist that substantivalism would still be correct so long as the facts
material bodies and their relations as the metaphysical basis for a world's about the world's spatiotemporal structure are more fundamental than the
spatiotemporal structure. If we ask, of a Newtonian world, "why (in the spatiotemporal facts about material bodies, and contrariwise for relationalism.
metaphysical sense) does it have a Galilean spatiotemporal structure?" the This strikes me as too far removed from the original views. More generally,
relationalist will answer: "because the particles (can) behave thus and so." I don't think that one of these views must be correct regardless of future
The substantivalist will have no answer (or if there is any answer, it won't physics, and it will depend on the details of that future physics whether one or
reference material bodies: see below). This is a substantive debate about the other, or neither, is correct.
what makes it the case that the spatiotemporal structure needed for the There is another way to put the difference between the views, which
physics holds. I want to be careful with. The substantivalist says that there exists a
fundamental physical space(time); the relationalist denies this. Similarly,
the relationalist denies, whereas the substantivalist accepts, the existence
of spacetime points (or regions) as fundamental physical objects. This way
3.3. Further clarifications
of putting things is familiar and in keeping with traditional conceptions
The substantivalist might not take a world's spatiotemporal structure to be
absolutely fundamental. Newton held that absolute space is a necessary
28
consequence of God's existence, so that the facts about the world's spatial Some argue that Newton wasn't a substantivalist: Stein (1970); DiSalle (2002).
22 fill North A New Approach to the Relationa!-Substantiva! Debate 23
of the dispute. 29 The problem is that it is not entirely clear what it means to seeing it as a fundamental fact about the physical world. This allows us to
say that a physical space-this "peculiar entity" (Belot and Earman, 2001, discuss the dispute, and to evaluate the evidence for either side, while
227)-does, or doesn't, exist; relatedly, whether spacetime points or regions remaining neutral on how the substantivalist wants to understand the
exist as concrete entities. I suspect that this is an underlying reason for the instantiation of that structure or the ontology behind this fact.
unclarity of the debate in many people's minds, especially in the philosophy This dovetails with an idea in spacetime structural realism. Jonathan Bain
of physics community. Some philosophers of physics have worried about (2006) argues that classical field theory (this includes general relativity),
taking spacetime points to be concrete physical entities in particular. As standardly given in terms of a tensor formalism, can be formulated in ways
Malament says, in the context of discussing whether spacetime points are that do not presuppose a differential manifold of points. He describes three
nominalist-friendly, "They certainly are not concrete physical objects in any alternative formalisms one could use (twistor theory, Einstein algebras, and
straight-forward sense. They do not have a mass-energy content .... They do geometric algebra), none of which treat points as fundamental. My under-
not suffer change. It is not even clear in what sense they exist in space and standing leaves it open for the substantivalist to spell out the spatiotemporal
time" (1982, 532). Others have worried more generally that this kind of structure in any of these ways, or even to refuse to choose among them, as
ontological dispute-a dispute that is just about what things exist-is non- Bain himself proposes. (Bain argues that we should be realists about space-
substantive or merely verbai. 30 Howard Stein, in discussing the spacetime time structure and not any particular instantiation of it. He sees this as a
debate, says that, "For me, the word 'ontological' itself presents seriously third view, since according to him the substantivalist is committed to
problematic aspects"; in particular, "Quine's usage [is) not a very useful one spacetime points, but it counts as substantivalist by my lights.)
for philosophy of physics" (1977a, 375). To be explicit, there are four different kinds of view that my conception
As I see it, the debate is about the fundamentality of spatiotemporal of substantivalism is meant to encompass, each of which holds that there are
structure, in particular about whether there is any spatiotemporal structure spatiotemporal facts or structure not grounded in material bodies. First is
(fact) not grounded in the structure of (facts about) material bodies, where what we might call Bainianism, on which one is a realist about spatiotem-
the substantivalist says that there is and the relationalist says that there isn't. poral structure but not about any particular instantiation of it, i.e. not about
Within this framework, there is some flexibility as to how exactly to put the any of the (non-material) objects that could be said to instantiate it. On this
dispute. Neither the matching principle nor my conception of spatiotem- view, the different possible descriptions or formulations or instantiations of
poral structure says how we must construe the nature of spatiotemporal spatiotemporal structure do not really differ from one another: one is an
structure; and I have not taken a stand on whether ground is primarily a anti-realist about those. Second is what we might call uncommitted sub-
relation between objects or facts. As a result, although we can put the stantivalism, on which one is a realist about a particular instantiation of
disagreement as being about whether there exists a fundamental physical spatiotemporal structure-there is a single best way of describing or formu-
spacetime or fundamental spacetime points, we do not have to. Anyone lating the spatiotemporal-structure facts, in terms of a certain kind of non-
squeamish about putting things in ontological terms can still see the debate material object-but one doesn't know what that instantiation or best
as being about the fundamentality of spatiotemporal structure, understand- formulation is; hence we cannot state the view as propounding one or
ing this as being not about whether there exist certain objects (over and another such formulation. Third is what we might call committed substan-
above material bodies), but about whether there are certain facts (over and tivalism, on which one is a realist about a particular instantiation of
above the facts about material bodies): the relationalist says that the fact that spatiotemporal structure, one thinks that there is a best formulation of it,
a world has a certain spatiotemporal structure holds in virtue of the fact and one does claim to know what it is; e.g. it might be the one in terms of
that material bodies behave thus and so; the substantivalist denies this, points (in which case the view approaches traditional substantivalism).
Fourth is the "qualitativist" substantivalism of Dasgupta (2009; 2011), on
which the fundamental spatiotemporal facts are purely qualitative, not
mentioning any entities at all; spacetime is not an entity but a "purely
29 See Field (1980, ch. 4); Mundy (1983); Earman (1989, 12); Brighouse (1994). qualitative structure." One of the things I am claiming is that, when it comes
30 This seems the spirit behind Stein (1970; 1977a); Curiel (2016); perhaps Belot
to the relational-substantival debate, we needn't choose among these ver-
(2011) and some others in note l; in a different way Wallace (2012). There have been
similar thoughts in metaphysics, for example in Hirsch (2011), but it's not clear that this sions of substantivalism. The argument in Section 4 will support each of
is exactly the same idea. them in the same way.
24 Jill North A New Approach to the Relational-Substantival Debate 25
apart from material bodies is nonfundamental; whereas certain material absolute motion" (1998, 460). '
33• Huggett and Hoefer (2009) note other relationalist views denying the relativity of
objects, and certain of their properties and relations, are fundamental.
motion.
At the same time, this is a non-standard, non-traditional take on things 34 T.his ~igns :vith a similar shift away from that question in recent literature,
which allows us to sidestep many of the reasons people feel that the usuai exemplified 1ll Stem (1970; l 977b); Sklar (1974); Friedman (1983); Earman (1989);
dispute has stagnated or become non-substantive. Most importantly, it Belot (1999; 2000; 2011); DiSalle (2006).
3
~ Huggett (1999, ch. 8) discusses the traditional arguments. More recent examples
leaves room for future physics to provide an answer, so that this dispute are 111 Earman and Norton (1987); Belot (2000).
cannot be "merely verbal" or "purely metaphysical." We think that there 36
St~tem~nts intimating that the relationalist cannot believe in spatiotemporal struc-
tur~7 are 1ll Field (1984, 34); Nerlich (1994a); Pooley (2013, 542); Maudlin (2012, 66).
Earman (1989) suggests this at points.
31 38
See the many notions listed in Horwich (1978); Friedman (1983); Earman (1989). Rynasiewicz (1996; 2000) worries about the clarity of all these (and other) distinctions.
26 Jill North A New Approach to the Relational-Substantival Debate 27
This is something that the physical laws give us a handle on, in ways Above I argued that the relationalist should go parrway41 toward adhering
discussed earlier, though I admit that there is more that could be said. to the matching principle by countenancing spatiotemporal structure, and
Perhaps there is nothing else that makes some fact or structure spatiotem- that she can do this by understanding all the facts about spatiotemporal
poral; perhaps there is. 39 Either way, I take the idea to be relatively familiar structure as being grounded in facts about material bodies. I am now
from physics. At least we have some clear cases of spatiotemporal structures, going to argue that really the relationalist can't adhere to this principle,
such as those discussed here. properly understood. The argument differs from the more familiar charge
Third, my conception requires a distinction between material bodies and that the relationalist cannot countenance a particular spatiotemporal fact
other things in the world. Although people have worried about the clarity of or structure.
this distinction, 40 I think that it is clear enough for our purposes. At the Recall that the matching principle says to posit in the world the structure
least, I suggest that we understand the debate in this way, on the assumption presupposed by the laws; that is, to posit physical structure in the world
that we will be able to locate such a distinction. For now I follow Earman, corresponding to the mathematical structure needed to state the laws.
who says that, "It is a delicate and difficult task to separate the object fields Now here is something else about the principle I haven't yet mentioned.
into those that characterize the space-time structure and those that charac- It applies, in the first instance, to the fundamental laws. (By saying "in the
terize its physical contents," while also noting that "the vagaries of this first instance," I mean to indicate that the principle applies at least to the
general problem need not detain us here, since there are clear enough cases fundamental laws, and that this is where we begin constructing our picture
for our purposes" (1989, 155-6). For those wanting argument that the of the world from physics, in that we build a world "from the bottom up."
distinction can generally be made, I refer you to Carl Hoefer (1998) and also I leave it open whether an analogous idea holds for nonfundamental laws.)
David Baker (2005). Given the fundamental laws, we should posit in the world the structure
One will find, in contemporary discussions, the thought that the rela- they presuppose. This is clear from our usual inferences about spatiotem-
tionalist can believe in the existence of spacetime, understanding this as poral structure. Assuming that Newton's laws are fundamental, we infer a
being (somehow) constructed out of material bodies and their features. So it Galilean structure to the world. From different fundamental laws, we infer
may seem like even the traditional dispute (and contemporary versions of it) a different spatiotemporal structure-such as a Minkowskian structure for
was never about the existence of spacetime but its fundamentality, and my special relativity, a preferred-location spatial structure for Aristotle's physics,
own formulation may seem like just a new label for an old dispute. This or a variety of different spatiotemporal structures for general relativity.
however is something of an anachronism. Traditional participants, like The matching principle also tells us to posit, in the fundamental level of
Newton and Leibniz, weren't focused on questions of fundamenta!ity: the physical world, whatever those laws presuppose. The fundamental laws,
they were not thinking explicitly in those terms. Neither, of course, were after all, are about what's fundamental. They don't "care about" or "know
they thinking in spatiotemporal terms. Atthe same time, to the extent that about" or mention the nonfundamental. I take it this is part of what we
we can understand what they were saying in these terms, this shows that my mean when we say that they are fundamental. I also take it that this is a
understanding is, as I claim, an updating of the traditional dispute, using familiar thought. (Michael Townsen Hicks and Jonathan Schaffer (2017)
more recent developments in physics (involving spacetime and its struc- call it orthodoxy. 42 ) For example, it lies behind our dislike of quantum laws
tures) and philosophy (fundamentality and ground). that mention things like "measurement" or "the observer." This isn't to
deny that fundamental laws have consequences for nonfundamental things.
These laws yield predictions for nonfundamental phenomena when we plug
in initial conditions and use various bridge principles. On their own, though,
4. AN ARGUMENT FOR SUBSTANTIVALISM
I now suggest that if we do understand the debate in this way, then there is a
powerful argument for substantivalism, given much of current physics. 41
Partway, since I haven't shown that the relationalist can ground the particular
structure needed.
42
They argue against the idea, concluding that fundamental laws can, and do,
39 Belot (2011) and Brighouse (2014) are two different accounts. mention nonfundamental properties. I agree that an alternative formulation can be useful
40 See especially Rynasiewicz (1996). in practice, bur I think that the best formulation won't mention such things.
28 ]ill North A New Approach to the Relational-Substantival Debate 29
fundamental laws only mention or presuppose or know about things at the I said that if the laws are asymmetric in this way, then we infer an
fundamental level. 43 asymmetric temporal structure in the world. The idea is that such laws
Another way to see this comes from the idea of "the structure presup- presuppose this structure, for they mention or presuppose a distinction
posed by the laws." The sense in which the laws presuppose or require some between past and future, by telling things to behave differently depending
structure is akin to an idea familiar from mathematics. In mathematics, we on the direction of time. But there is more to the stoty. Take the second law
can define different levels of structure by starting with a lowest level, such as of thermodynamics. This law is not time reversal invariant, so it may seem to
a set of points, and then defining other objects that add more structure. indicate an asymmetric temporal structure. However, the second law of
These levels of structure form a hierarchy. The ones "higher up" assume or thermodynamics is not a fundamental law. It doesn't mention a system's
presuppose or constrain levels lower down, in that the higher-level objects particles or other fundamental constituents. It is formulated in terms of
cannot be defined until the lower-level ones have been assumed or defined. higher-level macroscopic quantities like entropy. Whether to infer an
For example, think of adding differential structure to a topological space. objective past-future distinction in the world then really depends on what
This structure indicates, from among the continuous curves specified by fundamental theo1y accounts for the second law, and whether that theory's
the topology, which ones are smooth to varying degrees. In this way the laws are symmetric in time. (It is natural to think that if a past hypothesis
differential structure assumes or presupposes a topology: it cannot be defined, account of thermodynamics is correct, then there is no asymmetric tem-
it doesn't make sense, absent a topology. Higher-level structure is not simi- poral structure; whereas if a non-time reversal invariant theory like GRW
larly constrained by levels lower down-as different metrics, or none at all, quantum mechanics is true (and able to account for thermodynamics) then
can be added to a differential manifold. In other words, a given level of there is. 45 ) The nonfundamental law on its own does not tell us about
structure only "knows"-requires, constrains, presupposes, assumes-things fundamental temporal structure: it is too far removed from the fundamen-
about that level and below. 44 tal level to do that. Only a fundamental law can tell us about this.
Analogously for the structure required by the physical laws. This structure In other words, we posit fundamental structure in the world needed for
is presupposed by the laws in that it must be assumed in order for the laws to the fundamental laws. We recognize as fundamental the facts that are
be formulated or make sense. The laws don't similarly know about- recognized by the fundamental laws. The matching principle applies, in
require, constrain, presuppose, assume-higher-level structure. For funda- the first instance, to the fundamental laws and fundamental level of physical
mental laws, the result is that they only know about fundamental structure. reality. The matching principle as discussed in Section 2 says that the world
Note that the fundamental laws may constrain things higher up in a should "look like" or "fit" its laws. The primary reading of the principle says
different, metaphysical sense: given the fundamental laws and ontology, that the fundamental level of the world should look like or fit its funda-
everything else may be "fixed" in some sense. This is a different sense of mental laws.
constraining from the mathematical notion, which concerns what is needed Now to the argument for substantivalism. First notice that the kinds of
for something to make sense or be defined. The other sense is a metaphysical fundamental laws we are most familiar with are formulated to presuppose
notion that requires additional metaphysical principles concerning the spatiotemporal facts apart from material bodies. These laws mention or
relation between different levels of realiry. presuppose a spatiotemporal structure in addition to material bodies and
An example illustrates and motivates the primary reading of the matching their features. Newton's laws presuppose a Galilean spatiotemporal structure
principle. Recall the discussion of non-time reversal invariant laws. Earlier in addition to the existence of massive particles. These laws assume or
require that the world has this structure, just as the laws of special relativity
assume or require a Minkowskian structure. The laws of Aristotle's physics
43 This is different from Sider' s (2011, ch. 7) purity principle. Purity is a very general
mention a preferred-location spatial structure in addition to the elements
principle about what the fundamental facts or truths can mention. (It says that they
cannot mention nonfundamental concepts.) The above is specific to the physical laws and that move toward their natural places. Similarly for the laws of general
what they presuppose and therefore tell us about the physical world. relativity, even though they allow for different spatiotemporal structures.
44 In mathematics one also talks of a higher-level structure "inducing" a lower-level
one (e.g. "the topology induced by the metric"). This makes it sound as though the
higher-level structure is defined first and it then constrains the lower, but in fact it
45
amounts to the above idea (e.g. once we have defined a metric, there must already be Albert (2000) discusses these two accounts. See North (2008) on why these
implicitly a topology). conclusions about temporal structure are natural.
30 Jill North A New Approach to the Re!ational-Substantival Debate 31
Think of the usual way of understanding the field equations, as saying how above that of material bodies. Fourth premise: the prima1y reading of the
the distribution of matter and energy relates to the spatiotemporal geometty, matching principle. Conclusion: relationalism is incorrect. Substantival-
which in turn affects the behavior of matter. These equations are formulated ism posits the spatiotemporal structure or facts needed for the laws at the
directly in terms of-they mention or talk about-a spatiotemporal struc- fundamental level.
ture apart from material bodies, coded up in the metric tensor, distinct General relativity provides an example. This theory establishes a nomo-
from the stress-energy tensor. (See Hoefer (1996; 1998) for arguments logical connection between material bodies and a spatiotemporal structure
that the metric is most naturally seen as characterizing a spatiotemporal apart from them. On their own, the laws do not say whether material bodies
structure that is not the structure of a material field. This is not uncon- and spatiotemporal structure are at the same level of physical reality, nor
troversial, but is assumed in standard presentations.) The fundamental which is more fundamental if not. Without some further principle, both
laws that we are familiar with make reference to material bodies, but they relationalism and substantivalism seem satisfactory: both recognize facts
also presuppose or make reference to a spatiotemporal structure apart from about material bodies as well as a world's spatiotemporal structure. Enter
those bodies. 46 the matching principle. The substantivalist does, the relationalist does not,
Given that the fundamental laws are typically like this, a problem arises adhere to it.
for the relationalist. The problem is not that the relationalist doesn't You may wonder why the spatiotemporal structure presupposed by the
recognize enough spatiotemporal facts for the physics, a concern lying at laws is apart from material bodies, as premise two claims. After all, the
the root of classic arguments like Newton's, as well as many contempora1y relationalist, in my view, can countenance this structure, but will say that it
ones (see notes 23 and 25). Grant the relationalist enough stuff to ground has to do with the (actual and perhaps possible) spatiotemporal relations
those facts and make the relevant predictions, and there is still a problem. between material bodies. In what way do the laws presuppose a spatiotem-
According to the core of the view, all the facts about spatiotemporal poral structure that is in addition to material bodies? The answer comes
structure are grounded in more fundamental facts about material bodies. from the way that the fundamental laws are usually formulated. (I turn to
The kinds of fundamental laws we are used to, though, presuppose or potential reformulations in Section 5.) These laws are typically formulated
mention spatiotemporal facts apart from material bodies-facts that, for to directly mention material bodies, with a term that directly refers to
the relationalist, are nonfundamental. This violates the principle that the them-such as the mass term of Newton's dynamics, or the mass density
fundamental level of the physical world should contain whatever is needed of some formulations of Newtonian gravitation, or the elements mentioned
for or presupposed by the fundamental laws. in Aristotle's laws, or the stress-energy tensor of general relativity. 47 At the
So the argument is this. First premise: the fundamental laws are about same time, these laws also presuppose that the world has a spatiotemporal
what's fundamental to the physical world; they refer to or presuppose structure apart from those bodies-apart in that it is presupposed by the
things about the fundamental physical level. Second premise: these laws laws in the mathematical sense given above, or else is directly mentioned by
are about, they presuppose or refer to, a spatiotemporal structure, or or coded up in a distinct term.
spatiotemporal facts, apart from material bodies. Third premise: for the Recall that the matching principle tells us to infer that a special relativistic
relationalist, this kind of structure or fact exists at a nonfundamental level, world lacks an absolute simultaneity structure. The laws don't require this
mathematical structure, which suggests that the world doesn't have the
corresponding physical structure. To fail to adhere to the matching principle
is to fail to heed this evidence from the laws about what the world is like.
46 There is a difference between the laws mentioning and presupposing something.
The relationalist fails to adhere to the primary reading of the principle in the
That a law explicitly mentions something implies that the law presupposes it, but not vice same way. The fundamental laws are giving us evidence that spatiotemporal
versa. The laws of general relativity explicitly reference both material bodies and spatio-
temporal structure. The usual Newtonian laws explicitly mention the former yet only structure is fundamental to the physical world, which the relationalist fails
presuppose the latter. (Hence a difference from Quine's prescription (see pp. 11-12): to heed. The relationalist may respond that there are good reasons to
Newton's laws, as usually formulated, presuppose a Galilean spatiotemporal structure;
they don't explicitly mention or quantify over that structure, which the matching
principle tells us to posit.) This difference does not matter here. We use the matching
47
principle to infer structure in the world regardless of whether it is explicitly mentioned or In the context of this debate, both views take certain material objects to exist at the
presupposed. Either way, the laws require it. fundamental level. (Supersubstantivalism would then deny this.)
Another random document with
no related content on Scribd:
Note al Capitolo Undicesimo.
59. Polyb., 35, 4, 4 sgg. I frammenti della nuova Epitome liviana, scoperti tra
i Papiri di Oxyrynchus, hanno portata nuova luce sull’asprezza delle
contese provocate in Roma dalla insurrezione spagnola; cfr. Oxyr. Pap.
IV, pp. 90 sgg., ll. 177; 182-84; 207, 209 — E. Kornemann, Die neue
Livius Epitome aus Oxyrynchus, Leipzig, 1904, pp. 107 sgg.
62. Sulla condizione fatta alla Grecia nel 146, cfr. G. Colin, Rome et la Grèce
de 200 à 146 a. J. C., Paris, 1903, pp. 640 sgg.
CAPITOLO DODICESIMO
I GRACCHI
66. Sui Salassi e sulle miniere aurifere dell’antico Piemonte, cfr. Polyb., 34,
10, 18; Strab., 5, 1, 12; Plin., N. H., 33, 4, 78.
67. Per una illustrazione della legge agraria di Tiberio, cfr. E. De Ruggiero,
Agrariae leges, in Enciclopedia giuridica italiana, 1, 2, pp. 798-808.