Synthese (2015) 192:287303

DOI 10.1007/s11229-014-0568-1

Explanation and the dimensionality of space

Kants argument revisited
Silvia De Bianchi J. D. Wells

Received: 3 March 2014 / Accepted: 27 September 2014 / Published online: 17 October 2014
Springer Science+Business Media Dordrecht 2014

Abstract The question of the dimensionality of space has informed the development
of physics since the beginning of the twentieth century in the quest for a unified
picture of quantum processes and gravitation. Scientists have worked within various
approaches to explain why the universe appears to have a certain number of spatial
dimensions. The question of why space has three dimensions has a genuinely philosophical nature that can be shaped as a problem of justifying a contingent necessity
of the world. In contrast to explanations of three-dimensionality based on anthropic
arguments, we support the search for a theory that provides a justification for the
dimensionality of space based on a combination of deductive and inductive reasoning
applied to science. In doing so, we argue that Kant correctly approached the question in
Thoughts on the true estimation of living forces (1747) by connecting space dimensionality and the inverse square law. In expounding the strategy of Kants argument,
we describe the main features of a general Kantian explanation of the dimensionality of space and discuss them with respect to current accounts of explanation in the
philosophy of science, such as inference to the best explanation and the deductivenomological model.
Keywords Kant Dimensionality of space Causality Explanation Inverse square
law Anthropic argument

S. De Bianchi (B)
cole Normale Suprieure, CNRS-UMR 8547, Rue dUlm 45, 75005 Paris, France
e-mail: silvia.de.bianchi@ens.fr
J. D. Wells
Physics Department, University of Michigan, Ann Arbor, MI 48104, USA
e-mail: jwells@umich.edu

123

288

Synthese (2015) 192:287303

1 Introduction: arguments for the three-dimensionality of space

Explaining the number of spatial dimensions has been a research endeavor for centuries
(see Bolzano 1843; Halsted 1878; Ehrenfest 1920; Whitrow 1955; Tangherlini 1963,
1986; Penney 1965; Barrow 1983, 1988; Jammer 1993, Chapters 45; Tegmark 1997).
The rather counterintuitive hypothesis that extra dimensions exist is the path that many
theoretical physicists have taken, launching their ideas from the ground of the success
of the Standard Model of elementary particles. The way in which the connection
between space and dimensionality is described has philosophical interest and informs
the kind of arguments, even anthropic arguments, that justify the existence of three
dimensions in the universe. The distinction between space dimensions and extension,
for instance, makes it possible to think of branes embedded in a higher dimensional
space as constituting the world in which we live (Durrer et al. 2005). Nevertheless,
problems arise when physicists who are currently looking for possible experimental
signals for the existence of extra dimensions1 appeal to anthropic arguments or orbital
stability arguments. These arguments show weaknesses in their abilities to explain
three-dimensionality, by assuming that a teleological explanation can be taken as a
causal one.2
A teleological argument for the dimensionality of space infers the necessity of three
dimensions from the mere fact that the current state of affairs concerning stability of
planetary orbits or the presence of life on Earth could not be as such if different
dimensionality of space were admitted. These arguments have been developed during
the last three centuries by a tradition started by William Paley, among others, and
they are best-known as anthropic arguments for the dimensionality of space. This
kind of argument is burdened with an inner circularity and a subreption in implicitly
assuming that the explanandum (three-dimensionality) must be part of the ground
of the explanans (the state of affairs as being possible only in a three-dimensional
space). The three-dimensionality of space (jointly with the action of fundamental
forces) may indeed contribute to the stability of planetary orbits or to the development
of life, but these cannot be either the causal explanation for the three-dimensionality
of space nor their immediate direct consequences. The literature is replete with such
explanations for three dimensions that are at their core only found connections of
unique compatibility between a set of seemingly unrelated data and three dimensions.
Two examplary found connections are the celebrated arguments that planetary orbits
are stable (something discerned from collected sets of data) only when there are three
or less dimensions, and that the condition of less than three dimensions creates a
topological barrier to complex life forms (Whitrow 1955). Supporters of this argument
infer three-dimensionality only from data. This line of reasoning is an explanation only
if we admit that anthropic explanations are consistent, which is something that we
doubt. These explanations are guided by an anthropic principle suggesting that some
1 They are: (i) the direct or indirect observation of a Kaluza-Klein tower of states, or (ii) the observation
of deviations in the inverse square law of gravity in short-range experiments. In our paper, we mostly
review arguments provided in favour of the existence of extra-dimensions that appeal to this second kind
of experimental search.
2 Manson (2003) offers a clear overview of teleological arguments in modern science.

123

Synthese (2015) 192:287303

289

characteristics of nature (e.g., three dimensions) must be present not only for intelligent
beings to exist, but also because intelligible beings exist and can produce knowledge
of these characteristics. Our search, on the other hand, thus arises from the profound
dissatisfaction with anthropic principle-based arguments, which in our view offer no
satisfactory explanation of the necessary link between the three-dimensionality of
physical space and the laws of nature.
With regard to arguments developed in the last century, the attribution of three
dimensions to physical space has been viewed as being (1) purely contingent; (2) a
necessary feature of our conception of the world; (3) partly conventional and partly
contingent; (4) partly necessary and partly contingent (Whitrow 1955). In the face of
the question of how many spatial dimensions are there, a simple answer is that at the
energies and spatial distances of ordinary life, it requires three and only three numbers
to specify a location unambiguously. Therefore, we infer that we live in a threedimensional world. Now, it is possible that if we contemplate spatial distances much
smaller than our everyday experience, e.g., at 1020 cm, extra spatial dimensions may
be curled up, or compactified, with tiny radii (Graa 2005). This was first postulated
in the 1920s by Kaluza (1921) and Klein (1926), and later made popular in string
theory (Polchinski 2005),3 theories of topological singularities (Rubakov and Shaposhnikov 1983), and theories of large extra dimensions (Arkani-Hamed et al. 1998;
Randall and Sundrum 1999). Within these theories there are attempts to show that our
(3 + 1)-dimensional space-time is a necessary consequence of a higher dimensional
space when the correct laws of nature are applied.4
In all of these extra-dimensional theories if there was a bug many billions and
billions of times smaller than usa Planck bug with size of 1035 meters for
examplethat bug would see a higher number of dimensions than we see (9 spatial dimensions in the case of string theory). It is for this reason that our reporting of
the number of dimensions is scale dependent, meaning that if we look at large distance
scales such as a millimeter and higher, which we humans probe in our everyday life, we
would report that there are three dimensions, and not the higher number that a Planck
bug would report. Therefore, according to these theories, it is conventional that we
define the number of dimensions as the three that humans experience in ordinary
life. If we assume the meaning of conventionalism in Poincars terms,5 then arguments explaining the three-dimensionality of physical space are conventional if they
assume that the parameter for establishing the number of dimensions is the result of
3 In string theory, which is the most ambitious theory of everything proposed to date, the issue of spatial
dimensions is central. It has been found that in most constructions it is required to have nine space dimensions
and one time dimension for mathematical consistency of the theory. The experimental fact that we perceive
three dimensions is accommodated by proposing various compactification alternatives to the remaining six
dimensions.
4 A recent example of this approach is the following: One of the most fundamental questions concerning
our universe is why we live in a (3+1)-dimensional space-time, and why the universe is expanding. The aim
of this Letter is to provide some evidence that these facts can be derived from a nonperturbative formulation
of superstring theory in (9 + 1) dimensions based on matrix models (Kim et al. 2012).
5 In referring to axioms of geometry, Poincar pointed out that they are conventions and that our choice
among all possible conventions is guided by experimental facts; but it remains free and is limited only by
the necessity of avoiding all contradiction (Poincar 1902, p. 65).

123

290

Synthese (2015) 192:287303

a choice that depends on experimental facts. This implies that several choices may be
possible, but a convention is required to specify this parameter without contradiction.
It remains possible that the true explanation for three-dimensionality implies a contingent link between the number of dimensions and physical space, because nature
could have no further principles in store from which the necessity of three dimensions
could be derived. In other words, three-dimensionality might be nothing more than a
parameter that is to be measured as a property of physical space, and any value is as
good as another from the standpoint of physical law. The search for an explanation
for the origin of three dimensions based on necessity continues unabated today. Arguments for the three-dimensionality of physical space also support their explanations
by deriving them from a priori methods, or deductive reasoning and, more generally, they are guided by the idea of a necessary connection between the parameter of
dimensionality and physical laws. As we argue below, only an approach combining
deductive and inductive reasoning could open the possibility of abandoning anthropic
principles and of identifying necessary grounds and contingent facts that reveal a firm
connection between the number of dimensions and the laws of nature.
2 Kantian explanation for three dimensions: an overview
In what follows we investigate Kants 1747 explanation for three dimensions that
involves an argument that combines deductive and inductive reasoning and attributes
both a necessary and a contingent nature to the connection of spatial dimensions and
physical laws. In the previous section we gave an example whereby the number of
dimensions can have a conventional explanation due to the lack of information of
what is happening to spatial dimensions at extremely small scales, and that there may
indeed be more than three of them. In 1747 Kant accepted the possibility of more than
three spatial dimensions and, in fact, he had to assume this in order to prove that three
dimensionality derives from the inverse square law (ISL).6 However, there is no direct
experimental fact to which he appealed to choose three-dimensionality rather than two
or an arbitrary number of dimensions.7 Thus, we will not consider the conventionalist
account of the dimensionality of physical space in dealing with Kants argument.
The latter is endowed with a hybrid nature. On one hand, Kant used a deductive
approach and even tried to derive a priori the three-dimensionality of space, but this
methodology alone was unable to provide a satisfactory result. On the other hand,
precisely this insufficiency led Kant to combine deductive and inductive reasoning.
Kant did however provide a necessary explanation, namely he suggested that three
dimensionality is necessary by virtue of the ISL. It is somewhat counterintuitive to
todays students and scholars, who learn that the ISL is derived from three dimensions
via Gausss law (Moody and Paik 1993; Callender 2005). Thus, we implicitly hold
6 For an historical account of the inverse square law, see Henry (2011).
7 Other than three dimensional space as derived from the ISL are ruled out from Kants argument by virtue

of the fact that we could not perceive or imagine other than 3D spaces. Euclidean space might be seen as
a convention adopted by Kant, but, as we shall see, his argument aims at finding the physical necessary
ground for the fact that the physical world is experienced to be three-dimensional.

123

Synthese (2015) 192:287303

291

today that three dimensionality is a necessary explanation for the ISL, whereas in
the presence of Kantian argument we proceed in the reverse direction.8 Nevertheless,
the firm mathematical link between 3D and the ISL is now well known. This renders
Kants hypothesis impressive and worth being analyzed. Kants argument also shows
a contingent nature. He recognized that there was an element of arbitrariness to the
number of dimensions due to the arbitrariness of the ISL. There could have been
another law, say the inverse cube law, for instance, and then another dimensionality
of space would have descended from it. Thus, it must be emphasized that even if there
is a necessary implication of three dimensions from the ISL, yet the power law has an
intrinsic contingent nature (God could have chosen another oneto put it in Kants
terms), and thus the explanation for the three dimensionality of physical space relies,
at least in Kants view, on a connection between the law and physical space, which
is both necessary and contingent. This hybrid we will call Kantian explanation. It
suggests that in searching for the explanation of the dimensionality of space, we can
find a necessary proximate cause or causal explanation for it, but also we encounter its
contingent nature, which is embodied by an arbitrary parameter of a law, whether it be
derived from a deeper stage of the present theory of space-time or from another theory
that is still to come, which in turn determines the change of space dimensionality.
3 Investigating Kants 1747 argument
The fundamental connection between spatial dimensionality and the power of the
force law is Kants central claim in 1747.9 Kants argument can be divided into three
parts. The first part is constructed a parte priori (Sect. 9) and the other parts a parte
posteriori (Sects. 1011). At first, Kant tried to explain why space is three-dimensional
by means of an a priori argument appealing to geometrical and mathematical properties
(Sect. 9). Afterwards Kant develops his a posteriori proof (Sects. 1011). He assumes
the existence of mutual interactions among substances and the existence of forces
acting outwards (see Table 1).10 In doing so, Kant admitted a connection between
8 Solutions to the field equations of Einsteins theory of General Relativity reduce to Newtons theory of
gravitation and hence the ISL, F 1/r 2 . Since 3D is a fixed input to General Relativity, just as it is for
Newtonian gravity, the ISL is derived when 3D and other inputs needed to describe the theory are applied.
If space had n dimensions, and the principle of general covariance were maintained, the force law would
reduce to F 1/r n1 . Therefore, one could just as legitimately start with GCTn (General Covariant
Theory in n spatial dimensions) and call ISL an input, and 3D (and General Relativity) would be derived
as a consequence.
9 Kants argument is concerned with a notion of space that is very different from the one adopted later on

in the Critical period, both in the Dissertation (1770) and in the Critique of pure Reason (1781/1787). In
Thoughts on the True Estimation of living forces, Kant is discussing properties of physical space, whereas in
the Critique of pure Reason, specifically in the Transcendental Aesthetics (A19-49/B33-73, pp. 172192)
he considers space insofar as it is an a priori intuition of outer sensibility. As such, this pure intuition
necessarily determines spatiality as the a priori form of outer appearances. For the account of empirical
space as treated in physics, see Kant (1786, AA 4:480481, pp. 1516).
10 Kant discusses a notion of space that is similar to the one presented in the Physical Monadology
(1756). In the latter, Kant distinguished the notion of extension from physical space: substances or physical
monads do possess an extension, because they fill space not by virtue of their mere existence, but by virtue
of their sphere of activity determined by attractive and repulsive forces. In supporting the distinction between

123

292

Synthese (2015) 192:287303

Table 1 The structure of Kants argument

Section 9 (a priori)

Section 10 (a posteriori)

Section 11 (a posteriori)

Critique of Leibniz

Mutual interactions

Possibility of other spaces

Critique of number theory

Deduction of 3D from ISL

Probability of other worlds

metaphysical substances, geometrical (the domain of extension) and physical space

(where bodies act) through the essential actions of forces:
It is easy to show that there would be no space and no extension if substances
had no force to act external to themselves. For without this force there is no
connection, without connection, no order, and, finally, without order, no space.
Yet it is somewhat more difficult to see how the plurality of dimensions in space
derives from the law according to which this force of substances acts externally
(Kant 1747, AA 1:23).
Metaphysics and physics jointly provide sufficient ground for the three-dimensionality of space (Sect. 10), even if the latter is a fact of this world and a contingent
one (Sects. 1011), because God could have chosen an inverse cube law to order
substances, and this would have implied the constitution of another world and another
kind of space. The argument supports the necessary link between dimensionality and
an input law, but also claims that the three-dimensionality of space is a necessary
contingency, or, which is the same, a contingent necessity of this world. This is because
it is probable (wahrscheinlich) that the ISL is the only input law of this world, and thus
the link between the ISL and 3D can be necessary, once not only the ISL is assumed, but
also actual physical interactions respond to forces acting according to this law. The ISL
is thus just a sufficient ground for 3D (Kant 1747, AA:123).11 Kant was led to support
the link between dimensionality of space and force as depending on mathematical
and physical laws, because the necessary ground for the three-dimensionality of space
could not be found in a purely geometrical proof, such as the one Leibniz gave in the
Theodicy:
Because I discern a circular inference in the proof that Herr von Leibniz, somewhere in the Theodicy, takes from the number of lines that can be drawn at
right angles to each other from a point, I have sought to demonstrate the threedimensional character of extension from what can be discerned from the powers
of numbers (Kant 1747, AA 1:23).
Footnote 10 continued
extension and space, Kant combined Cartesian concepts of force, such as De Mairans (Massimi and De
Bianchi 2013), with Leibnizs metaphysics (see Schnfeld 2000; Watkins 2001), in order to modify Newtons
account of space and gravity. In particular, the notion of active force, as presented in 1747, is explicitly
derived from Leibnizs Specimen Dynamicum (1695).
11 As we shall see, Kant arrived at the solution by excluding Leibnizs geometrical argument and proofs
based on the properties of numbers, and (1) by following a metaphysical principle, and (2) by assuming that
the cause of gravity can be known. The latter assumption was compatible with Eulers position on Newtonian physics. According to Euler, indeed, Newtonian mechanics should have been enriched by Cartesian
elements, and physics must have been able to explain phenomena via causal principles (see Henry 2011).

123

Synthese (2015) 192:287303

293

Kant is referring to Leibniz ([1710] 1985, p. 335), who responded to the claim
made by Pierre Bayle (16571706) that the number of spatial dimensions is arbitrary.12 Kant claimed that Leibnizs proof is circular, because it tried to explain
three-dimensionality via properties of geometrical entities that already presuppose
three-dimensional space. Kant also shows the inadequacy of arguments appealing to what we now call analytic geometry and number theory in order to find
the necessary ground for three-dimensionality. In the mid-eighteenth century, analytic geometry was far from being formalized. There were, however, formalizations of algebra and arithmetic applied to geometry as formulated by Descartes
in his 1637 Geometry and later developed by Bernoulli and Euler. And it is not
until 1775 with Lagranges Recherches darithmetique (17731775) that the proper
beginning of the theory of numbers can be identified (Bos 2001; Israel 1998; Panza
2005).
The second part of Kants a priori argument for the necessity of threedimensionality follows from the failure of the Leibnizs attempt of a geometrical
proof. Kant continues then:
I have sought to demonstrate the three-dimensional character of extension from
what can be discerned from the powers of numbers. The first three powers are
entirely simple and cannot be reduced to any other, but the fourth power, as the
square of the square, is nothing but a repetition of the second power. As good
as this property of numbers appeared to me as a means of explaining the threedimensionality of space, it proved to be unsound in its application. For the fourth
power is an impossibility with regard to everything we can represent to ourselves
concerning space by means of the imagination. In geometry one cannot multiply
a square by itself, nor can one multiply the cube by its root; hence, the necessity
of three-dimensionality rests not so much on the fact that, in positing several
dimensions, one does no more than repeat the previous ones (as is the case with
the powers of numbers); rather it rests on a certain other necessity I am not yet
in a position to explain (Kant 1747, AA 1:23).
Kant is here discussing the treatment of solid bodies that appeared in eighteenthcentury arithmetic textbooks.13 It is worth noting that he is not denying the possibility
of deriving a mathematical explanation of three-dimensionality, but he assessed the
insufficiency and incompleteness of analytic geometry in solving the problem. In
particular Kant refers to the insufficiency of the additive exponential system first
developed by Diophantos of Alexandria and diffused in Prussia by Euler (Alten et al.
2003, p. 96).
This second part of the a priori argument contains a metaphysical principle that
also natural science was believed to employ: to same effects correspond the same
12 In Sect. 351 of the Theodicy, Leibniz claims: With the dimensions of matter it is not thus: the ternary

number is determined for it not by the reason of the best, but by a geometrical necessity, because geometricians have been able to prove that there are only three straight lines perpendicular to one another which
can intersect at one and the same point (Leibniz [1710] 1985, p. 335).
13 These topics were discussed in Marchi (1775, p. 189) or Doria (1726, p. 68).

123

294

Synthese (2015) 192:287303

cause(s).14 This is an argument per analogiam based on the principle that the ground
of the effect must be contained in the principle itself or in the cause itself. Kant uses
this argument as a transition to the a posteriori argument (Sects. 1011), where force
is treated as a metaphysical concept,15 namely as a cause that generates effects on
extension (Ausdehnung) and then in space (Raum):
The force by which any substance acts in union with other substances cannot be
conceived without a certain law that manifests itself in its mode of action. Since
the kind of law by which substances act on each other must also determine the
kind of union and composition of many substances, the law according to which
an entire collection of substances (i.e., a space) is measured, or the dimension of
extension, will derive from the laws according to which the substances seek to
unite by virtue of their essential forces. The three-dimensional character seems
to derive from the fact that substances in the existing world act on each other in
such a way that the strength of the action is inversely proportionate to the square
of the distances (Kant 1747, AA 1:24).
Kants argument concludes that the connection between the ISL and threedimensionality does not appeal to pure mathematical arguments, rather it appeals
to physical connections among substances. One of the reasons why he pursued this
strategy is that the reducibility of cube roots was unknown. Therefore, only the ISL
could have received a proper foundation and evidence. On the other hand, the contingency of three-dimensionality is a necessary conclusion if we admit the possibility, at
least, of other kinds of spaces (Kant 1747, AA 1:25) and the possibility of having other
input laws. And precisely from this perspective, the last part of Kants a posteriori
argument is the most fascinating:
Accordingly, I am of the opinion that substances in the existing world, of which
we are a part, have essential forces of such a kind that they propagate their effects
in union with each other according to the inverse-square relation of the distances;
secondly, that the whole to which this gives rise has, by virtue of this law, the
property of being three-dimensional; thirdly, that this law is arbitrary, and that
God could have chosen another, e.g., the inverse-cube, relation; fourthly, and
finally, that an extension with different properties and dimensions would also
have resulted from a different law. A science of all these possible kinds of space
would undoubtedly be the highest geometry that a finite understanding could
undertake. The impossibility we notice in ourselves of representing to ourselves
a space of more than three dimensions seems to me to stem from the circumstance
that our soul likewise receives impressions from without according to the inversesquare relation of distances, and because its nature is itself constituted so as not
only to be thus affected, but also to act external to itself in this way (Kant 1747,
AA 1:2425).
14 The argument is the following: Because everything found among the properties of a thing must be
derivable from what contains within itself the complete ground of the thing itself, the properties of extension,
and hence also its three-dimensionality, must also be based on the properties of the force substances possess
in respect of the things with which they are connected (Kant 1747, AA 1:24).
15 This aspect is borrowed from Leibniz, see Watkins 2001; 2005.

123

Synthese (2015) 192:287303

295

Kant supports his deduction by appealing to the fact that matter impresses upon our
senses according to the ISL and therefore our perception responds to the universal
law that regulates the world. However, Kant introduces here the notion that threedimensionality is a contingent feature of our world, as well as of our perception with
the following consequences:
If it is possible that there are extensions of different dimensions, then it is also
very probable that God has really produced them somewhere. For his works have
all the greatness and diversity that they can possibly contain. Spaces of this kind
could not possibly stand in connection with those of an entirely different nature;
hence such spaces would not belong to our world at all, but would constitute
their own worlds. I showed above that, in a metaphysical sense, more worlds
could exist together, but here is also the condition that, as it seems to me, is the
only condition under which it might also be probable that many worlds really
exist. For if the only possible kind of space is a three-dimensional one, then it
would be possible for the other worlds that I assume to exist apart from the one in
which we exist to be spatially connected with ours, for the spaces are of one and
the same kind. Hence, the question would be why God separated the one world
from the other, since he would certainly have imparted a greater perfection to
his work by linking them; for the more connection there is, the more harmony
and agreement there is in the world, whereas gaps and divisions violate the laws
of order and perfection. It is thus not probable (though it is inherently possible)
that many worlds exist, unless the many types of space that I have just mentioned
are possible (Kant 1747, AA 1:25).
But what was the historical, philosophical and scientific meaning of this fascinating
argument? Callender (2005) suggested that Kants argument highly influenced the
history of philosophy and the history of physics. However, it seems not to be the case
until the twentieth century, because Kants 1747 writing received critical notices and
reviews, such as Lessings16 , in between 1750 and 1752, and it went out of print by
the end of the eighteenth century.17 The revival of Kants argument in its original form
16 According to Lessing, Kant undertook the difficult business of educating the world. He estimated the
living forces, without first estimating his own. The statement appeared in Das Neueste aus dem Reiche des
Witzes, Lessing 1751, p. 32
17 In order to be aware of the hostile environment in which Kants work was received, it is worthwhile
looking at the review that appeared in Nova Acta Eruditorum (Anonymous 1752, pp. 177179), where an
anonymous reviewer focused on two main points of Kants meditations. The first one concerns the notion of
action of substances. According to the anonymous reviewer, Kants notion of a force as being determined
to act outwards is striking (What namely is that force of the same body A, determined to act outwardly?
[Quid enim est vis illa ipsius A, ad extra se agendum determinata? ], p. 178). The sarcastic attitude of the
reviewer is then directed towards Kants conception of three-dimensional space (Anonymous 1752, pp.
178179). The reviewer stated It seems that the author, who in one way or another heard of action in
inverse square of the distances and of three dimensions, unites them as in the same way as we bring together
dreamed-up ideas. [Videas hic hominem, qui aliqua de actione in reciproca duplicata distantiarum, et
de trina dimensione, sando inaudivit, haec ita conjungere, ut somniando ideas conjungimus] (Anonymous
1752, p. 179). The main criticism on this point is that the necessary reciprocal connection among substances
is not perspicuous and the same holds for the mechanics underlying higher-dimensional spaces. Therefore,
this argument is pure speculation and overall, the reviewer claims, Kants works deserve no more of his
paper and time.

123

296

Synthese (2015) 192:287303

is therefore more recent, and it is due to the development of Kantian studies aided by
the diffusion of the 1902- Akademie edition of Kants works and the debate involving
Cassirer on General Relativity and the problem of space (Cassirer [1923] 2004).
Paley (2006) is mentioned by Callender (2005) as one of the supporters of an argument for the three-dimensionality of space similar to Kants. However, as we show,
Kants argument is not an anthropic argument, despite Callenders assertion that it is
(Callender 2005, p. 2), and does not appeal to any teleological explanation, something
that Paley does. Paley does not refer explicitly to Kant (1747) and investigates the
stability of the orbits as a proof of three-dimensionality. Orbital stability arguments
were not known by Kant. It is also claimed by Callender (2005) that Paleys anthropic
approach to the explanation was a development of Kants arguments, which we disagree with. Furthermore, it should be mentioned that Newton was very well aware
of the possible link between the ISL and the three-dimensionality of space, even if
he never published his notes and did not pursue a systematic argument as Kant did.
Newtons arguments for the infinite extension of space and time (duration) are both
conceptual and theological in nature.18 The necessity of three-dimensional space was
linked to the power of gravity acting at a distance. Newton suggested that the inverse
square law arose naturally due to the structure of matter, namely all particles constituting matter interact through forces propagating according to the ISL (Hall and Hall
1962, p. 353). According to propositions LXXII and LXXIII of the Book I, the net
force within a sphere of uniformly dense matter is directly proportional to the radius
from the centre of that sphere, giving an underlying inverse square law force. This
only follows if the net mass within the given radius increases as the cube of the radius,
which combines with the inverse-square law to give the proper net force. Here the
three-dimensionality of space is the crucial ingredient, for in a higher-dimensional
space this result would not follow (Pesic 1998). In the last part of the Principia, The
System of the World, Newton has shown that the ISL can be derived from the motion
of planets, but refused to speculate on the origin of the law. Even assuming that the
ISL can be derived from matter and the power of gravity, as he remarks in the General
Scholium of the Book III, the cause of the power of gravity is unknown, and also
the cause of the properties it determines in the world, e.g. ISL, three-dimensional
space, is still to be discovered (see Hall and Hall 1962, p. 205; Guicciardini 1999,
pp. 7172). Therefore, Paleys (1802) argument should be considered to be related
to a long-standing tradition discussing Newtons theory of gravity (and its possible
cause) and his arguments supporting action at a distance, rather than be linked directly
to Kants proof.19 Kants argument was different and introduced his reflections on
the nature of physical space in the context of his metaphysics when dealing with the
18 See Hall & Hall (1962, pp. 133134).
19 Even if it has been claimed that Kant abandoned the spirit of his 1747 argument and then treated space in

a very different way (Friedman 1992; Watkins 2001; Schnfeld 2000), there are still traces of this argument
in his 1755 cosmology. In Universal Natural History and Theory of the Heavens, Kant claims: Attraction
is without doubt a quality of matter that is just as pervasive as the coexistence that makes space in that it
combines substances by reciprocal dependences, or, to put it more accurately, attraction is precisely that
universal relationship that unites the parts of nature in one space: it therefore extends to the entire expanse
of space into all the reaches of its infinity (Kant 1755, AA 1:308). The order of space depends on attraction,
which in turn regulates reciprocal fundamental interactions. The latter are the constraints that unify space

123

Synthese (2015) 192:287303

297

explanation of the relations between substances and the world they constitute through
their causal connections (Sects. 811). According to Kant, there would be no space
and no extension if substances had no force to act outwardly. For without this force
there is no connection, without connection, no order, and, finally, without order, no
space. Yet it is somewhat more difficult to see how the plurality of dimensions in space
derives from the law according to which this force of substances acts externally.
Having failed in finding an argument for the three-dimensionality of space completely a priori, Kant tries to find a necessary and at least sufficient ground for it via
the a posteriori argument. If the properties of geometrical space and those of numbers
cannot provide the necessary ground, metaphysics and physics must be connected and
provide it. Therefore, Kant proceeded in his a posteriori argument that relies on the
validity of Newtons theory of universal gravitation. From the success of the ISL and
the metaphysical assumption of fundamental interactions among physical substances,
Kant derives the following important consequences: (1) substances in the existing
world, of which we are a part, have essential forces of such a kind that they propagate
their effects in union with each other according to the inverse-square relation of the
distances; (2) the whole to which this gives rise has, by virtue of this law, the property of being three-dimensional; (3) this law is arbitrary, and God could have chosen
another, e.g., the inverse-cube, relation; (4) an extension with different properties and
dimensions would also have resulted from a different law. We have shown here the
features of Kants argument, by considering the role played by historical factors. In the
next section we shall highlight the kinds of explanation that are embodied in Kants
1747 argument and then conclude by recognizing how this is a clear alternative to
anthropic arguments for the dimensionality of space in physics.
4 Causal explanation of physical space
The title of this paper entails the word explanation for a crucial reason. In analyzing
Kants argument for three-dimensionality, it is our purpose to highlight weaknesses
and strengths of current arguments concerning causal explanation and to investigate
the features of a suitable argument for the dimensionality of space without appealing
to anthropic principles. Salmons (1989; 1994) accounts of causal explanation cannot
give guidance to deal with Kantian explanation, because they aim to account for causal
nexus and interactions in the world, rather than assume them, and they also assume that
there must be the conservation of a mark in the spatio-temporal continuum that allows
us to detect the nature of change occurring within a system. This line of argument
is clearly inapplicable here by the fact that in dealing with the question of space
dimensionality we would not be allowed to assume space-time in the explanans, and
we would then be required to have the explanandum as the postulate underlying the
explanans (the ISL), which is contradictory to the method.
into one expanse of space. The counterbalance of attraction is to be found in repulsive forces. Again in 1756,
in the Physical Monadology, Kant has the same problem of balancing the idea of attraction at a micro-level
Footnote 19 continued
and solves it via the concept of repulsive force, which responds to the inverse cube of the distance. This
ratio is also used by Kant in 1786 and later on in the Opus postumum.

123

298

Synthese (2015) 192:287303

Other kinds of scientific explanation appealing to inductive statistical or probabilistic accounts are ruled out as well, because there are no variables to be correlated.
However, Kantian-like arguments show that there are indices (say, the inverse cube
relation rather than the inverse square) that, when modified, can give rise to a different kind of space, and therefore manipulations of the explanans can give rise to
modification of the explanandum.
Reflection on Kants argument suggests an important notion to be investigated here,
namely the notion of generation in relationship to the notion of Law. What does
it mean for a physical law to generate a configuration of space? The fact that space
is three-dimensional is explained by a mode of action, the inverse square law, but
this law in itself does not necessarily imply three-dimensionality only. What we need
is a systematic interaction of substances, through active forces, and matter with its
properties, such as density and so forth. Therefore, at first, the fact that space is threedimensional appears to be partly necessary and partly contingent. Current debates on
causality and explanation are wide, and a long-standing tradition in the philosophy of
science deals with different versions of explanation and notions of causal explanation.
We now highlight how a Kantian argument encompasses at least two approaches
to explanation disclosing a viable path to the formulation of an argument for the
dimensionality of space without appealing to anthropic principles.
4.1 Causal explanations
The idea of causal explanation we need in order to classify Kants argument is multifaceted. We can divide the argument into three parts. The first one discloses a pattern
similar to the inference to the best explanation (Lipton 2004, 2009). The second part
is constituted by essential features of nomological accounts of causation (Sosa 1993,
p. 235). In the last part of the argument counterfactuals play an essential role in shaping
the hypotheses on the possibility of higher-dimensional spaces and on the probability
of other worlds. In what follows, we shall analyze these three parts of Kants 1747
argument and relate them to current positions on explanation.
1. In the first part of the argument (Sect. 9) we can detect the strategy of the inference
to the best explanation (IBE) that Kant used to undermine Leibnizs proof and any
derivation of three-dimensionality from the theory of numbers. IBE descends from
the idea that science is competitive, and this is precisely the image that Kant offers in
1747 when dealing with the possible explanations of the three-dimensional space.
According to IBE, it is only by asking how well various hypotheses would explain
the available evidence that we determine which hypothesis merits acceptance (Lipton 2009, p. 629). In the case of Kants 1747 argument this method is clearly
employed when Kant compares (a) Leibnizs geometrical proof for 3D grounded
on essential properties of geometrical entities and (b) the mathematical argument
that speculates that 3D can be derived from the properties of numbers. According to
Lipton, IBE is a two-stage process, where both stages are filters of potential explanations (Lipton 2004, pp. 5664). Kant first filtered out the implausible explanations
(Leibnizs geometrical explanation, Bayles argument for three-dimensionality as
due to arbitrariness) and then at the second stage investigates the potential explana-

123

Synthese (2015) 192:287303

299

tion based on the theory of numbers, which is plausible, because it is not based on
assumed properties of space, but is ultimately found to be incomplete. The lack of
evidence for the derivation of three-dimensionality from the properties of numbers
undermines the mathematical explanation and strengthens the physical one, which
is the only viable explanation, namely the one that, if correct, provides the most
understanding (Lipton 2004, p. 59) and would best explain the three-dimensionality
of space.
2. The second part of Kants argument (Sect. 10), assumes the evidence of the ISL, and
seems to be simply grounded on the fact that this law implies three-dimensionality,
namely that the law logically implies it as a consequence. According to nomological
explanation (N), an event or state of affairs P (partially) causes (or is a cause or
causal factor of) another Q only if there are actual (initial) conditions I and a law of
nature L such that, by necessity, if P and I and L all obtain then Q must obtain, where
the law L is essential in that P and I alone do not necessitate Q. At a closer look,
the argument goes as follows: an outward force (P) generates the dimensionality
of space (Q), which is three-dimensional, if and only if we assume the conditions
of a mutual interaction among substances (I) and that they interact according to an
inverse square law (L). Only by assuming all these three conditions Q is obtained.
Indeed, the inverse square law in itself could produce something different than
Euclidean three-dimensional spaces (such as spherical surfaces), and certainly the
mutual interaction among substances (I) per se does not dictate a dimensionality:
one needs a force (P) acting outwards to do so. Therefore, the nomological account
of causation is detectable in the use of a metaphysical notion of force that plays a
crucial role in Kants argument.
3. In the third part of the argument (Sect. 11), Kant makes use of counterfactuals to
strengthen his claim (Sect. 10) and makes the hypothesis that God could have chosen
other than the ISL to govern this world, and that if there are other worlds, separate
from our world, their number of dimensions would be unknown to us, because also
the law underlying the unification of our world with others is unknown to us. By
intervening on the explanans (Gods choice of the law, or the law itself) and by
changing the index of the law governing physical interactions, another kind of space
(explanandum) would be not only possible, but also actual in our or another world.
Kants argument shows here the character of contrastive explanation, namely a
version of the causal model of explanation that best supplements the general pattern
of IBE, by attempting to answer the question Why this rather than that? (Lipton
2004, p. 33; Fraassen 1980, pp. 126129). These explanations do not causally
explain an event as such, but rather explain an event together with the absence
of another event, similar in kind. Coming back to Kants argument, the reasoning
applies as follows. In face of the question Why is space in this actual world threedimensional rather than lower/higher-dimensional? Kants reasoning is that it is
probable that other worlds exist, where other than 3D spaces are possible, if the
input law governing the interactions is different from ISL. But we know that in our
world the index of the input law is also constrained by actual physical interactions,
therefore the actual output is a world in which space has three dimensions. In other
words, to explain why the space in our world is three-dimensional, Kant applies
a causal difference (Lipton 2004, p. 71) between higher-dimensionality and our

123

300

Synthese (2015) 192:287303

world, consisting of the cause of the former (an input law different from ISL) and
the absence of a corresponding actual event in the case of the latter.
How do we have to interpret the notion of generation of three-dimensionality in
Kants argument? It is a causal nomological explanation, no doubt. However, and this
is the most relevant part of Kants legacy for us, the causal relation between the ISL
and three-dimensionality is a sufficient but not necessary ground of it. Kant seems
to admit causes that are deterministic and universal but not necessary in themselves.
The ISL is a necessary law of nature, but the relation of three-dimensionality with it
is subject to other factors also. In Kants view, the world could have been different
from how it appears, and we know today that this fascinating statement concerning
the necessary contingency of the world allows us to speculate about extra-dimensions
and higher-dimensional spaces. It is thus to be asked whether Kant is really talking
about physical space when connecting dimensionality and the ISL.
4.2 Causality, other worlds, and the dimensions of physical space
Carnap (1924) investigated the relationship between causality and dimensionality of
space. Even though he discussed Kants transcendental idealism of space and time, he
did not mention Kants 1747 argument. However, Carnaps argument deserves careful
attention for the results that obtain, namely the deep connection of physical causality
and dimensionality of space. For Carnap, in our experience, we have to distinguish
the primary world (die primre Welt) and the secondary world (die sekundre Welt).
The first one consists of sensations that are not yet determined, but in their spatiotemporal order. These acquire spatio-temporal contents in the secondary world. The
concept of the number of dimensions is different in the first and the second world: in
the first we have (2+1) and in the second (3+1) dimensions. According to Carnap, it
is in the secondary world that the physical laws are determined, whereas in the first
world physical laws are still to be determined. Both the three-dimensionality of space
and determinability (or physical causality) are fictions (Fiktionen) in the secondary
world, and they are in a reciprocal logical dependance to each other, namely the threedimensionality of space is derivable from physical causality (Carnap 1924, p. 130).
Carnaps scheme applies to Kants 1747 argument and clarifies the way in which the
probability of many worlds and the possibility of higher-dimensional spaces should
be understood. We claim, indeed, that Kants distinction between extension and space
allows one to think of other than three-dimensional spaces in a way that is reminiscent
of Carnaps notion of first world. We emphasize that precisely the distinction between
a determinable and determined world, namely between the world of possibility and
the world of actuality, enables the compatibility of our three-dimensional space and
other spaces. If this point is conceded, then the assessment of epistemological and
metaphysical aspects related to this approach is a relevant task of current philosophical enquiries. Kant explored the condition of the probability of many worlds, even
possessing extensions with a different number of dimensions. These worlds, however,
would not be in connection with ours and therefore would remain conceivable, but
unknown to us. Kant then believed that this possibility goes against the rational organization of the world, and his explanation is based on the law of continuity of nature
borrowed from Leibniz:

123

Synthese (2015) 192:287303

301

For the more connection there is, the more harmony and agreement there is in
the world, whereas gaps and divisions violate the laws of order and perfection.
It is thus not probable (though it is inherently possible) that many worlds exist,
unless the many types of space that I have just mentioned are possible (Kant
1747, AA 1:25).
Kants words are extremely interesting. He believes that there can be a fundamental
level (that we could identify with the notion of another world) in which space has
more (or less) than three-dimensions. This level falls outside of our perception, but to
eventually prove the possibility of many types of space (be it via a priori arguments or a
posteriori arguments and experiments) would imply not only the possibility in thought,
but also the probability (Wahrscheinlichkeit) that other worlds exist. According to
Kants argument, this procedure would constitute the finding of a possible ground
of their unification, because one would find not only an input law able to determine
the order(s), the extension(s), and therefore the dimensionality of space(s), but also a
unifying physical law of interaction enabling different worlds to communicate.
5 Conclusions
In this paper we have shown and discussed the main features of Kants 1747 argument
for the three-dimensionality of space according to which the latter is portrayed as a
necessary contingency of this world. In expounding the characteristics of the Kantian
argument, we have also shown that different approaches to explanation are combined
in order to obtain a coherent and stronger conclusion. In our view, the combination
of these approaches, which are embodied by the a priori and the a posteriori parts of
the argument, can be related to current views of explanation, such as inference to the
best explanation (IBE) and the DN model. All these approaches are combined in what
we called Kantian explanation for the dimensionality of space. This argument, we
claim, enables us to think of a necessary contingency of the dimensionality of space,
without appealing to theological or teleological arguments. We therefore believe that
in the structure of Kants argument resides a clear alternative to anthropic arguments
that are often reported to justify and explain the fact that space is three-dimensional.
The most important aspect in constructing a Kantian argument is the recognition of a
physical input law that is essential to the causal explanation of the dimensionality of
space.
We claim thus that Kants 1747 argument deserves careful attention and that it can
be the source of inspiration for both scientists and philosophers of science who wish
to reflect upon the combination of different accounts of explanation, to construct solid
arguments for the dimensionality of space, and to set conditions for inferring the possibility of other than 3D spaces. We have shown that it is not only desirable, but also
necessary to ground the dimensionality of space on a physical law. The reliability of
physical laws implying more than 3D spaces should be accompanied with the investigation of the connection between the input law and its necessary consequences, e.g. 3D
brane dynamics and particle propagation into higher-dimensional spaces. However,
without experimental confirmation of such a law and its implications, what is left to us
at present is to recognize only the mere possibility in thought of other than 3D spaces.

123

302

Synthese (2015) 192:287303

Acknowledgments This work has made been possible also thanks to the grant offered by CERN THDivision to Dr De Bianchi in 2012 and 2013. We are grateful to two anonymous referees whose stimulating
comments improved the final version of our paper. We would like to thank Brigitte Falkenburg and Daniele
Cozzoli for fruitful discussions.

References
Alten, H. W., Naini, A. D., Folkerts, M., Schlosser, H., Schlote, K. H., & Wuing, H. (2003). 4000 Jahre
algebra: Geschichte. Kulturen. Menschen. Dordrecht: Springer.
Anonymous (1752). Meditationes de vera aestimatione virium vivarum/Gedancken von der wahren
Schtzung der lebendigen Krfte. Nova Acta Eruditorum, 177179.
Arkani-Hamed, N., Dimopoulos, S., & Dvali, G. (1998). The hierarchy problem and new dimensions at a
millimeter. Physics Letters B, 429(3), 257263.
Barrow, J. D. (1983). Dimensionality. Philosophical Transactions of the Royal Society A, 310(1512), 337
346.
Barrow, J. D., & Tipler, F. J. (1988). The anthropic cosmological principle. Oxford: Oxford University
Press.
Bolzano, B. (1843). Versuch einer objektiven Begrndung der Lehre von den drei Dimensionen des Raumes.
Prague: Gottlieb Haase Shne.
Bos, H. J. (2001). Redefining geometrical exactness: Descartes transformation of the early modern concept
of construction. Dordrecht: Springer.
Callender, C. (2005). Answers in search of a question: Proofs of the tri-dimensionality of space. Studies
in History and Philosophy of Science Part B, 36(1), 113136.
Carnap, R. (1924). Dreidimensionalitt des Raumes und Kausalitt. Annalen der Philosophie under
philosophischen Kritik, 4(1), 105107.
Cassirer, E. ([1923] 2004). Substance and function and Einsteins theory of relativity. New York: Dover.
Descartes, R. ([1637] 1954). Geometry. New York: Dover.
Doria, P. M. (1726). Delle opere matematiche: Considerazione, dissertazioni, ed esercitazioni geometriche,
e la duplicazione del cubo dimostrata per la via generale dEuclide. Con una meccanica de corpi
sensibili, ed insensibili ...Ed. ora accresciuta di nuove considerazioni intorno ai moti, ed alle orbite
de pianeti. Venice.
Durrer, R., Kunz, M., & Sakellariadou, M. (2005). Why do we live in 3 + 1 dimensions? Physical Letters
B, 614(34), 125130.
Ehrenfest, P. (1920). Welche Rolle spielt die Dreidimensionalitt des Raumes in den Grundgesetzen der
Physik? Annalen der Physik, 366(5), 440446.
Friedman, M. (1992). Kant and the exact sciences. Cambridge, MA: Harvard University Press.
Graa, M. (2005). Flux compactifications in string theory: A comprehensive review. Physics Reports, 423(3),
91158.
Guicciardini, N. (1999). Reading the principia: The debate on Newtons mathematical methods for natural
philosophy from 1687 to 1736. Cambridge: Cambridge University Press.
Hall, A. R., & Hall, M. B. (1962). Unpublished scientific papers of Isaac Newton: a selection from the
Portsmouth collection in the University library, Cambridge. Cambridge: Cambridge University Press.
Halsted, G. B. (1878). Bibliography of hyper-space and non-Euclidean geometry. American Journal of
Mathematics, 1(3), 261276.
Henry, J. (2011). Gravity and De gravitatione: The development of Newtons ideas on action at a distance.
Studies in History and Philosophy of Science Part A, 42(1), 1127.
Israel, G. (1998). Des Regulae la Gomtrie/From the Regulae to the Gomtrie. Revue dhistoire des
sciences, 51(2), 183236.
Jammer, M. (1993). Concepts of space: The history of theories of space in physics. New York: Dover.
Kaluza, T. (1921). Zum Unittsproblem in der Physik (pp. 966972). Berlin: Sitzungsberichte Preuische
Akademie der Wissenschaften.
Kant, I. ([1747] 2012). Thoughts on the true estimation of living forces. In E. Watkins (Ed.), Immanuel
Kant: Natural science (pp. 1155). Cambridge: Cambridge University Press.
Kant, I. ([1755] 2012). Universal natural history and theory of the heavens. In E. Watkins (Ed.), Immanuel
Kant: Natural science (pp. 182308). Cambridge: Cambridge University Press.

123

Synthese (2015) 192:287303

303

Kant, I. ([1781/1787] 1998). Critique of pure reason. In P. Guyer & A. W. Wood (Eds.). Cambridge:
Cambridge Univerisity Press.
Kant, I. ([1786] 2004). Metaphysical foundations of natural science. In M. Friedman (Ed.). Cambridge:
Cambridge University Press.
Klein, O. (1926). Quantentheorie und fnfdimensionale Relativittstheorie. Zeitschrift fr Physik A, 37(12),
895906.
Kim, S.-W., Nishimura, J., & Tsuchiya, A. (2012). Expanding (3 + 1)-dimensional universe from a
Lorentzian matrix model for superstring theory in (9+1) dimensions. Physical Review Letters, 108(1),
011601.
Lagrange, J. L. (1775). Recherches darithmtique. Berlin: C. F. Voss.
Leibniz, G. W. ([1710] 1985). Essais de Thodice sur la bont de Dieu, de la libert de lhomme et lorigine
du mal. In E. M. Huggard (Ed.), Theodicy. La Salle Illinois: Open Court.
Lessing, G. E. (1751). Das Neueste aus dem Reiche des Witzes, Monat Julies, 32.
Lipton, P. (2004). Inference to the best explanation (2nd ed.). London: Routledge.
Lipton, P. (2009). Causation and explanation. In H. Beebee, C. Hitchcock, & P. Menzies (Eds.), The Oxford
handbook of causation (pp. 619631). Oxford: Oxford University Press.
Manson, N. A. (Ed.), (2003). God and design: The teleological argument and modern science. London:
Routledge.
Marchi, G. E. (1775). Arithmetica ragionata tratta con somma diligenza da migliori autori. Modena:
Societ tipografica.
Massimi, M., & De Bianchi, S. (2013). Cartesian echoes in Kants philosophy of nature. Studies in History
and Philosophy of Science Part A, 44(3), 481492.
Moody, M. V., & Paik, H. J. (1993). Gausss law test of gravity at short range. Physical Review Letters,
70(9), 11951198.
Paley. ([1802] 2006). Natural theology. Introduction and notes by M.D. Eddy and D.M. Knight. Oxford:
Oxford University Press.
Panza, M. (2005). Newton et les origines de lanalyse, 16641666. Paris: Albert Blanchard.
Penney, R. (1965). On the dimensionality of the real world. Journal of Mathematical Physics, 6, 16071611.
Pesic, P. (1998). Newton and hidden symmetry. European Journal of Physics, 19(2), 151153.
Poincar, H. (1902). Science and hypothesis. In The foundations of science (G. Halsted, Trans.). New York:
The Science Press.
Polchinski, J. (2005). String theory. Cambridge: Cambridge University Press.
Randall, L., & Sundrum, R. (1999). A large mass hierarchy from a small extra dimension. Physical Review
Letters, 83(17), 33703373.
Rubakov, V. A., & Shaposhnikov, M. E. (1983). Do we live inside a domain wall? Physical Letters B,
125(23), 136138.
Salmon, W. C. (1989). 4 decades of scientific explanation. Pittsburgh: University of Pittsburgh Press.
Salmon, W. C. (1994). Causality without counterfactuals. Philosophy of Science, 61(2), 297312.
Schnfeld, M. (2000). The philosophy of the young Kant. Oxford: Oxford University Press.
Sosa, E. (1993). Varieties of causation. In E. Sosa & M. Tooley (Eds.), Causation (pp. 234242). Oxford:
Oxford University Press.
Tangherlini, F. (1963). Schwarzschild field in n dimensions and the dimensionality of space problem. Nuovo
Cimento, 27(3), 636651.
Tangherlini, F. R. (1986). Dimensionality of space and the pulsating universe. Nuovo Cimento B Series 11,
91(2), 209217.
Tegmark, M. (1997). On the dimensionality of spacetime. Classical and Quantum Gravity, 14(4), L6975.
Van Fraassen, B. (1980). The scientific image. Oxford: Oxford University Press.
Watkins, E. (Ed.). (2001). Kant and the sciences. Oxford: Oxford University Press.
Watkins, E. (2005). Kant and the metaphysics of causality. Cambridge: Cambridge University Press.
Whitrow, G. J. (1955). Why physical space has three dimensions? British Journal for the Philosophy of
Science, 6(21), 1331.

123