The reals as a subset of an ultraproduct of finite fields

Imagine you are a master architect trying to build a perfect, infinite city called The Real Numbers (let's call it R). This city contains every possible number you can think of: whole numbers, fractions, and the messy, never-ending decimals like $\pi$ and $\sqrt{2}$ .

Now, imagine you have a massive, chaotic warehouse filled with millions of tiny, finite boxes. Each box is a Finite Field ( $F_p$ ). Think of these boxes as small, self-contained universes where numbers wrap around like a clock. In a box labeled "13," if you add 1 to 12, you don't get 13; you get 0. These boxes are simple, but they are limited.

The paper by Roee Sinai asks a fascinating question: Can we build a perfect copy of our infinite city (R) inside this chaotic warehouse by combining these tiny boxes?

The Setup: The Ultraproduct Warehouse

The author uses a mathematical tool called an Ultraproduct. Imagine taking an infinite number of these finite boxes and gluing them together according to a specific set of rules (an "ultrafilter"). The result is a giant, non-standard universe called $\tilde{F}$ .

Inside this giant universe, we know we can find a copy of the Real Numbers. However, there's a catch: The copy is "hidden" or "external." It's like trying to find a specific, perfect house in a city where the blueprints don't match the buildings. You can't point to a single internal rule that says, "This is the house."

The Three Ways to Build (The "Construction Methods")

The author explores three different ways to try and "construct" or define this hidden city using the tools available in the warehouse. He calls these methods $\sigma$ -sets, $\delta$ -sets, and Almost Internal sets.

Think of these as different construction techniques:

$\sigma$ -sets (The "Union" Method): You try to build the city by stacking up a countable list of simple, internal rooms.
$\delta$ -sets (The "Intersection" Method): You try to build the city by taking a countable list of large, internal zones and finding the tiny area where they all overlap.
Almost Internal (The "Cut" Method): You use a simple "cut" (like a ruler that stops at a certain point) and a machine (an internal function) to filter out the numbers you want.

The Big Discoveries

1. The Impossible Dream (Theorem 0.5 & 0.6)

The author proves that you cannot build the Real Numbers using any of these three simple methods.

Analogy: It's like trying to build a perfect, infinite library by only stacking a finite number of books, or by only looking at the intersection of a finite number of shelves. No matter how you try, the "Real Numbers" are too complex and "messy" to be captured by these simple, countable rules. They are external—they exist in the warehouse, but they are invisible to the standard blueprints.

2. The "Good Enough" Sub-Cities (Theorems 0.7 & 0.8)

While you can't build the entire Real Number city, the author shows you can build very impressive sub-cities that are almost as good.

The Scenario: If the warehouse contains a copy of the Algebraic Real Numbers (numbers that are solutions to polynomial equations, like $\sqrt{2}$ but not $\pi$ ), we can do something amazing.
The Result: We can construct a sub-city that is Real-Closed (it has all the nice properties of real numbers, like square roots of positive numbers) and is $\aleph_1$ -saturated (it's so rich and full of numbers that it can satisfy any logical pattern you can describe with a countable list of rules).
The Catch: This sub-city is huge. It's so big that it contains $2^{\mathfrak{c}}$ (an unimaginably large number) of different copies of the Real Numbers.
Analogy: You can't build the exact city of New York, but you can build a "Super-Manhattan" that is so perfect and dense that it contains billions of different versions of Manhattan inside it. You just can't pick out one specific version that is "the" Real Numbers.

3. The Role of the Square Root of -1

The paper splits into two cases:

Case A: The warehouse has no $\sqrt{-1}$ (Imaginary numbers). Here, the sub-cities we build are "Real-Closed." They feel very much like the real world.
Case B: The warehouse does have $\sqrt{-1}$ . Here, the sub-cities become Algebraically Closed. This means they are like the "Complex Numbers" (the real world plus the imaginary world). These are even bigger and contain even more copies of the Real Numbers.

The "Cut" and the "Machine"

The author introduces a clever trick to find these sub-cities. He defines a "Cut" (a line in the non-standard numbers) and a "Machine" (an internal function).

Imagine the non-standard numbers are a long, infinite road.
The "Cut" is a barrier placed somewhere on the road.
The "Machine" measures how "complex" a number is.
By filtering numbers that are "simple enough" (below the barrier), the author creates a new field.
The Surprise: If you choose the barrier carefully, the resulting field is a perfect, rich mathematical world that contains the Real Numbers, even though the Real Numbers themselves are too complex to be defined by the barrier alone.

Summary in Plain English

This paper is a mathematical detective story.

The Mystery: Can we find the Real Numbers inside a giant mathematical structure made of finite fields?
The Clue: Yes, they are there, but they are "hidden" (external).
The Failed Attempt: You cannot define them using simple, countable rules (unions, intersections, or simple cuts). They are too complex.
The Success: However, you can build a "Super-Field" using these rules. This Super-Field is so rich and powerful that it contains countless copies of the Real Numbers.
The Conclusion: While you can't isolate the one true Real Number line with simple tools, you can build a universe so vast that the Real Numbers are hiding inside it in a billion different ways.

It's like saying: "You can't point to the exact soul of a human being using a simple checklist, but if you build a universe of perfect clones, the soul is definitely there, hiding in one of them."

Here is a detailed technical summary of the paper "The reals as a subset of an ultraproduct of finite fields" by Roee Sinai.

1. Problem Statement

The paper addresses the structural properties of nonstandard models of arithmetic, specifically ultraproducts of finite fields ( $\tilde{F} = \prod F_p / \mathcal{F}$ ). While it is a known result (via [2], [5], [4], [8]) that the field of real numbers $\mathbb{R}$ can be embedded into such an ultraproduct, these embeddings are "external" sets. The central problem investigated is whether specific copies of $\mathbb{R}$ (or related fields like the algebraic reals) can be constructed using "mostly internal" sets.

The author defines three specific classes of "constructible" external sets to test this hypothesis:

$\sigma$ -sets: Countable unions of internal sets.
$\delta$ -sets: Countable intersections of internal sets.
Almost internal sets: Preimages $f^{-1}[C]$ of a "cut" $C$ (a downwards-closed subset of nonstandard naturals $^*\mathbb{N}$ ) under an internal function $f$ .

The core questions are:

Can a copy of $\mathbb{R}$ be a $\sigma$ -set, a $\delta$ -set, or an almost internal set?
If not, can we construct larger fields (containing $\mathbb{R}$ ) using these methods?
How does the presence or absence of $\sqrt{-1}$ in the ultraproduct affect these constructions?

2. Methodology

The paper employs tools from Model Theory, specifically nonstandard analysis and the theory of ultraproducts.

Nonstandard Model Construction: The author works within a nonstandard model $(V(^*\mathbb{N}), V(^*\mathbb{N}), *)$ generated by an ultraproduct of finite fields. Elements are classified as internal (elements of $^*V(\mathbb{N})$ ) or external.
Height Functions: A crucial technique involves defining internal "height" functions ( $f_Q$ for rationals, $f_{\bar{Q}}$ for algebraic numbers) that map elements of the field $\tilde{F}$ to nonstandard natural numbers $^*\mathbb{N}$ . These functions measure the "complexity" of an element (e.g., the size of the numerator/denominator or the size of matrix entries required to represent the element as an eigenvalue).
Cut-Based Construction: The author constructs subfields by taking the preimage of specific "cuts" (downwards-closed sets) in $^*\mathbb{N}$ under these height functions. By carefully selecting cuts that satisfy specific algebraic closure properties (e.g., ensuring roots of polynomials exist within the preimage), they generate new fields.
Saturation Arguments: The proofs heavily rely on the $\aleph_1$ -saturation of the nonstandard model (guaranteed by the properties of the ultrafilter) to show the existence of elements satisfying specific types (e.g., being larger than all standard integers but smaller than a specific cut).
Algebraic Number Theory: The appendices contain detailed proofs involving resultants, homogeneous polynomials, and ideal theory to establish the existence of algebraic integers with specific properties (Lemma 5.9), which is necessary for embedding $\mathbb{R}$ into certain rings.

3. Key Contributions and Results

A. Impossibility Results for $\mathbb{R}$

The paper proves that no copy of the real numbers $\mathbb{R}$ inside the ultraproduct can be constructed using the defined methods:

Theorem 0.5: No copy of $\mathbb{R}$ $R$ is a $\sigma$ $σ$ -set or a $\delta$ $δ$ -set.
- Reasoning: If $\mathbb{R}$ were a $\sigma$ -set, it would be a countable union of finite sets (since internal subsets of $\mathbb{R}$ in this context are finite), implying $\mathbb{R}$ is countable (contradiction). If it were a $\delta$ -set, saturation arguments allow the construction of an element in $\mathbb{R}$ larger than all standard integers, which is impossible.
Theorem 0.6: No copy of $\mathbb{R}$ $R$ is "almost internal" (i.e., of the form $f^{-1}[C]$ $f^{- 1} [C]$ for an internal $f$ $f$ and cut $C$ $C$ ).
- Reasoning: If $\mathbb{R}$ were almost internal, one could enumerate it using the internal structure of the cut, leading to a contradiction regarding the cardinality or order properties of $\mathbb{R}$ .

B. Constructibility of Larger Fields

While $\mathbb{R}$ itself cannot be constructed this way, the paper shows that fields containing $\mathbb{R}$ can be constructed, and these fields are rich enough to embed $\mathbb{R}$ in $2^{\mathfrak{c}}$ distinct ways.

Case 1: The field does NOT contain $\sqrt{-1}$ (Real-closed case)

Construction: The author defines a specific $\delta$ -cut $C_{\hat{p}}$ based on the solvability of odd-degree polynomials.
Result (Theorem 0.7): The preimage $R^\delta_{\hat{p}} = f_{\bar{Q}}^{-1}[C_{\hat{p}}]$ $R_{\overset{p}{^}}^{δ} = f_{\overset{ˉ}{Q}}^{- 1} [C_{\overset{p}{^}}]$ is an $\aleph_1$ -saturated real-closed field.
- This field contains a copy of $\mathbb{R}$ .
- There are at least $2^{\mathfrak{c}} $distinct embeddings of$ \mathbb{R}$ into this field.
Result (Theorem 0.8): There also exists a $\sigma$ -subfield that is real-closed and contains $\mathbb{R}$ . However, this $\sigma$ -field is not $\aleph_1$ -saturated.

Case 2: The field DOES contain $\sqrt{-1}$ (Algebraically closed case)

Construction: Similar definitions are used, but the cuts are defined to ensure algebraic closure (solvability of all polynomials).
Result (Theorem 0.7 & 0.8):
- There exists an almost internal $\delta$ -subfield that is algebraically closed with cardinality at least $\mathfrak{c}$ (isomorphic to $\mathbb{C}$ if $|\tilde{F}|=\mathfrak{c}$ ).
- There exists an almost internal $\sigma$ -subfield that is algebraically closed.
- In both cases, there are at least $2^{\mathfrak{c}} $ways to embed$ \mathbb{R}$ into these fields.

C. Structural Analysis of Cuts

The paper provides a detailed classification of cuts in $^*\mathbb{N}$ :

Internal cuts are either $\mathbb{N}$ or intervals $[0, \hat{n}]$ .
$\sigma$ -cuts are unions of such intervals; $\delta$ -cuts are intersections.
The paper introduces partition cuts (based on partitions of the index set of the ultrafilter) and proves that under certain ultrafilter conditions (non-P-points), there exist cuts that are neither $\sigma$ nor $\delta$ . This demonstrates the complexity of the hierarchy of external sets.

4. Significance

Refining the Nature of External Sets: The paper clarifies the boundary between "constructible" external sets and the standard real numbers. It proves that while $\mathbb{R}$ is embeddable, it is structurally "too complex" to be formed by simple countable operations on internal sets.
Existence of Saturated Real-Closed Fields: It provides a concrete method to construct $\aleph_1$ -saturated real-closed fields within ultraproducts of finite fields, which are of significant interest in model theory and nonstandard analysis.
Multiplicity of Embeddings: The result that there are $2^{\mathfrak{c}} $distinct copies of$ \mathbb{R}$ within these constructed fields highlights the "wild" nature of nonstandard models; there is no canonical or "simplest" copy of the reals in this context.
Methodological Innovation: The use of "height functions" combined with cuts to define subfields is a novel technique for generating specific algebraic structures within nonstandard models.

5. Conclusion

Roee Sinai's paper establishes that while the field of real numbers $\mathbb{R}$ cannot be directly constructed as a $\sigma$ -set, $\delta$ -set, or almost internal set within an ultraproduct of finite fields, fields containing $\mathbb{R}$ can be constructed using these methods. These constructed fields are either $\aleph_1$ -saturated real-closed fields (in the absence of $\sqrt{-1}$ ) or algebraically closed fields (in the presence of $\sqrt{-1}$ ). Crucially, these fields contain $2^{\mathfrak{c}} $distinct copies of$ \mathbb{R}$, demonstrating that the embedding of the reals in such nonstandard models is highly non-unique and structurally complex.