Metaphysics of Spacetime: a Report of My

Research

Lu Chen

Koç University

June 2021

1 / 47
My Research during PhD Dissertation (2017-2019)

My Research Question:
What is the structure of continua, such as space or spacetime?

I Is space composed of points?

I Is space gunky ?
I Are there infinitesimal regions?

2 / 47
Motivations

Why spacetime?

I Ubiquitous in science and in experience

“Laws of motion cannot be written on thin air” (Earman 1989)

I All current views face significant problems

I the standard view
I the atomistic view
I the gunky view

3 / 47
Some results

Methodology
Develop new options: philosophically rigorous, mathematically
grounded, empirically constrained.

Dissertation Work
I Two Infinitesimal theories of space based on nonstandard
analysis 12
I A new account of distance for atomistic space in response to
a famous objection3

1
“Do Simple Infinitesimal Parts Solve Zeno’s Paradox of Measure?”Synthese
(2019).
2
“Infinitesimal Gunk,” Journal of Philosophical Logic (2020).
3
“Intrinsic Local Distances: a Mixed Solution to Weyl’s Tile Argument”Synthese
(2020).
4 / 47
Continuation from Dissertation (2019-2020)

I explore a new theory of space with infinitesimal regions based on

smooth infinitesimal analysis, which:4

I regiments scientific reasoning involving infinitesimals.

I gives a new understanding of vectorial quantities.

4
“Smooth Infinitesimals in the Metaphysical Foundation of Spacetime Theories,”
under review in Journal of Philosophical Logic.
5 / 47
Post-Dissertation Work (2020-2021)

1.

6 / 47
Infinitesimals in Geometry

A circle is a regular polygon with infinitesimal sides.

Area= 12 rC . Area= 12 aP.

7 / 47
Infinitesimals in Physics

An example from classical mechanics:

|∆v|/v = |∆r|/r (triangle similarity)

1
v
· | ∆v
∆t
|= 1
r
· | ∆r
∆t
|

a = v 2 /r

8 / 47
A Puzzle About Vector Field

Is the value of an electric field at a point intrinsic to the point?

Figure: pepper in cooking oil forming electric field lines

9 / 47
Smooth Infinitesimal Analysis

Smooth infinitesimal analysis (SIA) is a theory of infinitesimals

that purports to regiment those ideas.
I It can serve as more intuitive foundation for calculus.
I A possible framework for contemporary physics (e.g., Paugam
2012).
I A more attractive framework for understanding vectorial
quantities.

10 / 47
Smooth infinitesimal analysis

Microstraightness (core postulate)

For any smooth curve, and any point on the curve, there is a
straight infinitesimal segment of the curve around the point.

SIA features the smooth line R (instead of the real line).

c.f. nonstandard analysis

But there is a major challenge of interpretation.

11 / 47
Nilpotent Infinitesimals

Nilpotent infinitesimal:
x ∈ R ∧ x 2 = 0.

Microstraightness
entails:
The set of nilpotent
infinitesimals is not {0}.
i.e. Not all nilpotent
infinitesimals are zero.

12 / 47
On the other hand

Theorem
There are no nilpotent infinitesimals that are not zero.

Proof
Suppose there is a nilpotent infinitesimal  6= 0.
 =  · 1 =  · ( · −1 ) = 2 · −1 = 0 · −1 = 0.
Contradiction. The supposition is false!
Note that the proof is also intuitionistically valid.

13 / 47
Intuitionistic Logic

Claim 1. Not all nilpotent infinitesimals are zero.

Claim 2. There are no non-zero nilpotent infinitesimals.
SIA is inconsistent in classical logic!

But Claim 1 and Claim 2 are consistent in intuitionistic logic.

Intuitionistic logic is classical logic without the law of excluded
middle.
Claim 2 intuitionistically implies all nilpotent infinitesimals are not distinct from zero.
But, ¬x 6= 0 does not intuitionistically imply x = 0.

14 / 47
In face of classical inconsistency

Various strategies of reformulating SIA in classical logic have failed.

I e.g. Indeterminacy of identity? (Hellman 2006)

SIA can regiment reasoning involving infinitesimals, but the cost is

classical logic!
Should we pay the cost?
I Give up SIA?
I Give up classical logic?

15 / 47
In face of classical inconsistency

SIA vs Classical Logic

I Hellman (2006): SIA describes an abstract mathematical

structure, but it does not apply to actual space.

I M. Heller, J. Król (2016, 2017): we should weaken classical

logic to intuitionistic logic because of SIA.

16 / 47
My Goal
I will argue that SIA can be reconciled with classical logic.

Strategy
There are classical models for SIA using “sheaf semantics”
(Moerdijk and Reyes 1991)

A geometric interpretation of the models (c.f. Hellman 2006, Bell

2008).

A semantics view of SIA: one of the models is a realistic model of

space according to SIA.

17 / 47
Background: Einstein algebras

Information about manifolds is encoded in “Einstein algebras.”

18 / 47
Background: Einstein algebras

19 / 47
Background: Einstein algebras

(Category : A collection of objects and relations between them.)

20 / 47
Generalizing Einstein algebras

21 / 47
Generalizing manifolds

22 / 47
The Basic Model for SIA

E.g. “Not all nilpotent numbers are zero.” =⇒ There is a nilpotent manifold.

23 / 47
Interpreting the basic model

How does the category of generalized manifolds represent a

theory of our physical space?
I Let one generalized manifold represent physical space.
I Let some embedding maps represent parthood relations.
I Then, some generalized manifolds together with those
embedding maps represent regions of physical space.
The resulting theory: our space has nilpotent regions!

24 / 47
Nilpotent regions

Distinctive features:
I Too short to bend!
All functions on a nilpotent region are linear.
I Straightforward understanding of derivatives and vectors.

25 / 47
Higher-dimensional case

All n-vectors at a point “live on” the n-dimensional nilpotent

region around that point.
I A vector is intrinsic to a nilpotent region!

26 / 47
Nonclassical mereology

I A nilpotent region has a point

as a proper part.
I The nilpotent region has no
other proper parts!

27 / 47
Nonclassical mereology

Violation of Supplementation
Supplementation
If x is a proper part of y , then y also has a proper part z distinct
from x.

The statue and the lump of clay

28 / 47
Conclusion

It is possible to extract a realistic theory of space from SIA that:

I Regiments scientists’ reasoning with infinitesimals;
I Provides a better understanding of vectorial quantities;
I Satisfies classical logic.
I (Generalize Einstein algebras.)

29 / 47
Future work

Which structure is more fundamental?

I Manifolds (spacetime) are more fundamental: spacetime
substantivalism (the geometric approach)
I Rings (matter fields) are more fundamental: spacetime
algebraicism (the dynamic approach)

30 / 47
Future work

Project 1. Beyond Spacetime: the C ∞ -Algebra of all Fields

(in collaboration with T. Fritz (Perimeter))
A novel proposal for fundamental algebraic structures.

Project 2. Can Spacetime be derived from Dynamics?

On Geometric versus Dynamic Approaches to Relativistic
Spacetime.

—End of Talk—

31 / 47
References

Bell, J. (2008) A Primer of Infinitesimal Analysis (2nd Edition). Cambridge

University Press.
Earman, J. (1989) World Enough and Spacetime. MIT Press.
Heller, M., Król, J., (2016) “Synthetic Approach to the Singularity Problem.”
arXiv:1607.08264 [gr-qc]
Heller, M., Król, J., (2017) “Gravity in the Smallest.” arXiv:1706.03541 [gr-qc]
Hellman, G. (2006) “Mathematical Pluralism: The Case of Smooth
Infinitesimal Analysis,” Journal of Philosophical Logic, 35: 621-651.
Moerdijk, I., Reyes, G. (1991) Models for Smooth Infinitesimal Analysis.
Springer-Verlag New York.

Paugam, F. (2014) Towards the Mathematics of Quantum Field Theory.

Springer International Publishing.

32 / 47
First Strategy

I Constructive Mathematics?

Mathematical objects are constructions of human mind and

do not exist independently.
I Claim 1 =⇒ Possible to construct a non-zero nilpotent
number.
I Claim 2 =⇒ Impossible to construct a non-zero nilpotent
number.

33 / 47
Second Strategy

Indeterminacy?

The idea:
“=” means “determinately identical”
“6=” means “determinately not identical”

But, 6= is just a combination of ¬ and =.

34 / 47
The (Pre)Sheaf Model

hW , D, C , v i
W : quotient rings of C ∞ (R).
DR : a presheaf (domain function) that assigns every w ∈ W the
set of its ring elements.
C : homomorphisms (counterpart maps) between members of W .
For any w1 , w2 , if there is a homomorphism from w1 to w2 , then we say w1
“sees” w2 (abbr. w1 Rw2 ).

35 / 47
Examples of Quotient Rings of C ∞ (R)

Quotient Rings of C ∞ (R)

Equivalence classes of members of C ∞ (R) under certain
equivalence relations that preserve the original ring structure

Example 1. C ∞ ([0, 1])

Two smooth functions on R are equivalent iff their values do not
differ on [0, 1]
=⇒ Isomorphic to all smooth functions on [0, 1].

36 / 47
The Ring of Linear Functions

Example 2. The ring of linear functions L

Two smooth functions on R are equivalent iff they have the same
value at 0 and the same derivative at 0.
=⇒ Isomorphic to {f |f = a + bx}.

Nilpotent Elements
There are nilpotent ring elements that are not zero, e.g., f (x) = x.
f 6= 0, but f 2 = 0.

37 / 47
Interpreting SIA

1.¬∀x ∈ R(x 2 = 0 → x = 0)
For any w1 , and for any w2 with w1 Rw2 , it is not the case that, for any w3 with
w2 Rw3 , for all d ∈ DR (w3 ), for any w4 with a map h from w3 to w4 , if
h(d)2 = vw4 (0), then h(d) = vw4 (0).

Put simply: every ring sees some ring that has non-zero nilpotent
elements.
A truth maker: L has non-zero nilpotent elements (e.g., f (x) = x)
and is accessible from every ring.

38 / 47
Interpreting SIA

2.¬∃x ∈ R(x 2 = 0 ∧ x 6= 0)
For any possible world w1 , and for any possible world w2 with w1 Rw2 , it is not
the case that, there is a d ∈ DR (w2 ) such that d 2 = vw2 (0) and for any possible
world w3 with a map h from w2 to w3 , h(d) 6= vw3 (0).

Put simply: in any ring, every nilpotent element has zero as a

counterpart in some ring.
A truth maker: Every nilpotent element in every ring can be
mapped to 0 in R.

39 / 47
A Realistic Interpretation

The presheaf model consists of rings of smooth functions on

standard space.
=⇒ SIA is actually a ring theory??
That wouldn’t help.
We want some similarity between the interpretation of SIA and its
intuitive meaning (a nonstandard theory of space).

40 / 47
Rings Represent Regions

The rings actually represent regions of space, and homomorphisms

between rings are maps and relations between their corresponding
regions.

41 / 47
Einstein Algebras

*Spacetime Algebraicism. Fields exist without an underlying

spacetime. (Geroch 1972, Earman and Norton 1987)

Manifold-Algebra Duality
There is a one-to-one correspondence between manifolds and
smooth algebras such that for any two manifolds M, N, every
smooth map from M to N uniquely corresponds to a
homomorphism from C ∞ (N) to C ∞ (M) and vice versa.
(Rosenstock et al. 2015)

42 / 47
The “Geometric” Condition
A ring is a smooth algebra only if it does not have non-zero
nilpotent elements.

However, without appealing to Manifold-Algebra Duality,

this condition seems arbitrary and unmotivated.

43 / 47
Ring-Locus Duality

Ring-Locus Duality
There is a one-to-one correspondence between quotient rings of
C ∞ (R) (more generally, of C ∞ (Rn )) and loci such that for any
rings A, B, every homomorphism from A to B corresponds to a
unique smooth map from B’s corresponding locus to A’s
corresponding locus.

Examples
C ∞ (R) = the smooth line locus RL
L = the nilpotent locus ∆L .
R = the point locus p.

44 / 47
Loci and SIA

Loci exhibit some desirable features of SIA.

SIA: (∀f : ∆ → R)(∃!a, b ∈ R)(∀x ∈ ∆)f (x) = a + bx.

Loci: All smooth functions on ∆L are linear.

(“Smooth functions” on ∆L are maps from ∆L to RL , which are represented by
homomorphisms from C ∞ (R) to L, which are isomorphic to L.)

45 / 47
Differentiation

The derivative of a smooth function on RL on point p is the slope

of the function restricting to ∆p , the nilpotent region around p.
∆L can be embedded as a nilpotent region of RL around any real number point
through an injective map from ∆L to RL .

46 / 47
Presheaf Semantics

Negation. vw (¬p) = 1 iff for all w 0 such that wRw 0 , vw 0 (p) = 0.

Conjunction. vw (p ∧ q) = 1 iff vw (p) = 1 and vw (q) = 1.
Disjunction. vw (p ∨ q) = 1 iff vw (p) = 1 or vw (q) = 1.
Conditional. vw (p → q) = 1 iff for all w 0 such that wRw 0 , if
vw 0 (p) = 1, then vw 0 (q) = 1.
Universal quantifier. vw (∀xφ(x)) = 1 iff for all w 0 such that
wRw 0 , and for all d ∈ X (w 0 ), d satisfies vw 0 (φ).
Existential quantifier. vw (∃xφ(x)) = 1 iff there is a
d ∈ X (w ), d satisfies vw (φ).

47 / 47