Indistinguishable Irrational Multiples In (R;+,r)

by Andrew McMorgan 50 views

Hey guys! Today we're diving deep into the fascinating world of model theory, specifically looking at the structure (R;+,r)(\mathbb{R};+,r). This is a follow-up to our previous discussion on definable elements, so if you haven't checked that out, you might want to swing back! We're going to tackle a juicy question: Is every irrational multiple of rr indistinguishable from every other? Let's get into it!

Understanding Indistinguishability

First off, what does it mean for two elements to be indistinguishable in a structure? In model theory, two elements aa and bb in a structure M\mathcal{M} are considered indistinguishable if, for any formula ϕ(x)\phi(x) with a single free variable xx in the language of the structure, M⊨ϕ(a)\mathcal{M} \models \phi(a) if and only if M⊨ϕ(b)\mathcal{M} \models \phi(b). Basically, you can't tell them apart using any of the defined properties or relations within that structure. Think of it like having two identical-looking coins; unless you have some special way to distinguish them (like a hidden mark or a different weight), they're indistinguishable to you. In our case, the structure is (R;+,r)(\mathbb{R};+,r), where R\mathbb{R} is the set of real numbers, '+' is the standard addition operation, and 'r' is a fixed, nonzero real number. We're interested in whether different irrational multiples of this rr behave the same way with respect to all possible formulas in this language.

The Structure (R;+,r)(\mathbb{R};+,r)

Let's break down our structure (R;+,r)(\mathbb{R};+,r) a bit more. We have the familiar set of real numbers, R\mathbb{R}. The operation is standard addition, which is super well-behaved. The 'twist' comes from the constant symbol rr, which represents a specific nonzero real number. This might seem small, but it changes things! The language of this structure consists of:

  1. Variables: Like x,y,zx, y, z, etc.
  2. Constants: The specific number rr (and implicitly, 0, which is definable as x+0=xx+0=x or r−r=0r-r=0).
  3. Functions: The addition operation '+'.
  4. Relations: Here, we don't have any explicit relation symbols like '<' or '=' for arbitrary pairs. The equality symbol '=' is always part of the logical framework itself, but we're interested in definable properties using the given language. So, we're looking at formulas built using variables, the constant rr, addition, and equality.

This means any formula we can write will essentially involve sums and differences of variables and the constant rr. For example, ϕ(x):x+r=5\phi(x): x + r = 5 or ψ(x):x+x=r\psi(x): x + x = r. We are asking if, for any such ϕ(x)\phi(x), if ϕ(a)\phi(a) holds, then ϕ(b)\phi(b) also holds for any two irrational multiples aa and bb of rr. It's like asking if all irrational multiples of rr are interchangeable from the perspective of this structure's rules.

The Core Question: Irrational Multiples

Our focus is on elements of the form kâ‹…rk \cdot r, where kk is an irrational number. We want to know if k1rk_1 r and k2rk_2 r are indistinguishable for any irrationals k1k_1 and k2k_2. Let's consider an element a=kra = k r where kk is irrational. What can we say about aa using only addition and the constant rr?

Any formula Ï•(x)\phi(x) in the language of (R;+,r)(\mathbb{R};+,r) will look something like:

$ \exists y_1, \dots, y_n \forall z_1, \dots, z_m \quad ( (y_i = T_i(x, r, \text{vars})) \land (z_j = U_j(x, r, \text{vars})) \land \dots ) $

where TiT_i and UjU_j are terms built using addition, subtraction, and the constant rr.

For example, if we have a formula like ϕ(x):x+r=2r\phi(x): x + r = 2r. If x=krx = kr, then this becomes kr+r=2rkr + r = 2r, which simplifies to (k+1)r=2r(k+1)r = 2r. Since r≠0r \neq 0, we can divide by rr to get k+1=2k+1 = 2, so k=1k=1. This means that only x=1rx=1r satisfies this formula. If kk is irrational, then x=krx=kr will not satisfy this formula.

Now, suppose we have two irrational numbers k1k_1 and k2k_2, and let a=k1ra = k_1 r and b=k2rb = k_2 r. We are asking if for every formula Ï•(x)\phi(x), Ï•(a)\phi(a) holds if and only if Ï•(b)\phi(b) holds. If this is true, then aa and bb are indistinguishable.

The Key Insight: Rational vs. Irrational Coefficients

Consider a formula Ï•(x)\phi(x) involving addition and the constant rr. Any term in this formula will ultimately be of the form c1x+c2rc_1 x + c_2 r, where c1c_1 and c2c_2 are rational numbers. This is because addition preserves the 'rational coefficient' structure when combined with rr.

Let's write a generic formula ϕ(x)\phi(x). It will involve quantified variables and terms. Each term will be a sum/difference of variables and multiples of rr. For a single variable xx, any term T(x)T(x) in the language of (R;+,r)(\mathbb{R};+,r) can be represented in the form ax+brax + br for some a,b∈Qa, b \in \mathbb{Q}.

So, any formula ϕ(x)\phi(x) is equivalent to a statement about whether ax+brax + br satisfies certain equations or inequalities involving other terms of the form a′x+b′ra'x + b'r.

Let a=k1ra = k_1 r and b=k2rb = k_2 r, where k1,k2∈R∖Qk_1, k_2 \in \mathbb{R} \setminus \mathbb{Q}. Suppose ϕ(x)\phi(x) is a formula in the language (R;+,r)(\mathbb{R};+,r).

Let's consider a very simple type of formula: x+cr=drx + c r = d r, where c,d∈Qc, d \in \mathbb{Q}.

If ϕ(x)\phi(x) is x+cr=drx + c r = d r, then substituting x=krx = k r gives kr+cr=drk r + c r = d r, which means (k+c)r=dr(k+c)r = dr. Since r≠0r \neq 0, this is equivalent to k+c=dk+c = d, or k=d−ck = d-c. Since d−cd-c is a rational number, this formula is satisfied only if kk is rational. But we are considering kk to be irrational. So, for any irrational kk, x=krx=kr will not satisfy such a formula.

What if the formula is more complex? Let Ï•(x)\phi(x) be any formula in the language of (R;+,r)(\mathbb{R};+,r). Any term t(x)t(x) in this formula can be written as q1x+q2rq_1 x + q_2 r for some q1,q2otinQq_1, q_2 otin \mathbb{Q} (this is not entirely correct, let's refine this).

Correction: Let's be more precise. A term T(x)T(x) in the language (R;+,r)(\mathbb{R};+,r) is built from xx, rr, and addition. So, T(x)T(x) will be of the form ax+bra x + b r, where a,b∈Qa, b \in \mathbb{Q}.

For example, if ϕ(x)\phi(x) is x+r=2xx+r = 2x. Substituting x=krx=kr: kr+r=2krkr+r = 2kr. This gives r(k+1)=2rkr(k+1) = 2r k. Since r≠0r \neq 0, k+1=2kk+1 = 2k, so k=1k=1. This formula is only satisfied by x=1rx=1r, which is a rational multiple of rr.

Let's consider a=k1ra = k_1 r and b=k2rb = k_2 r where k1,k2k_1, k_2 are distinct irrational numbers. We want to know if ϕ(a)  ⟺  ϕ(b)\phi(a) \iff \phi(b) for all formulas ϕ(x)\phi(x).

Suppose we have a formula Ï•(x)\phi(x) that can distinguish aa from bb. This means either Ï•(a)\phi(a) is true and Ï•(b)\phi(b) is false, or vice versa.

Consider the structure (R;+,r)(\mathbb{R}; +, r). The set of definable numbers (elements that can be denoted by a term in the language) are precisely the rational multiples of rr. Let SdefS_{def} be this set. If rr is transcendental over Q\mathbb{Q}, then this might be different. But rr is a real number, so we assume standard reals.

Let's analyze the effect of substitution. If ϕ(x)\phi(x) is a formula, and we substitute x=yrx = yr into it, we get a new formula ϕ′(y)\phi'(y) in the language of (R;+,0)(\mathbb{R}; +, 0) with an extra constant rr. The structure (R;+,r)(\mathbb{R}; +, r) is essentially (R;+)(\mathbb{R}; +) with a distinguished element rr.

Consider the simplest possible formulas involving xx: x=qrx = q r for qotinQq otin {Q}. This formula is not definable in the language (R;+,r)(\mathbb{R}; +, r), since qq would need to be part of the language or definable. Only rational multiples of rr are definable as single terms. However, we are talking about arbitrary formulas.

Let a=k1ra = k_1 r and b=k2rb = k_2 r where k1eqk2k_1 eq k_2 and both are irrational. Suppose there exists a formula Ï•(x)\phi(x) such that Ï•(a)\phi(a) is true and Ï•(b)\phi(b) is false.

Any formula ϕ(x)\phi(x) in (R;+,r)(\mathbb{R}; +, r) can be translated. Let x=yrx = yr. The formula ϕ(x)\phi(x) becomes ϕ′(y,r)\phi'(y, r), where ϕ′\phi' is a formula in the language of (R;+,extconstants)(\mathbb{R}; +, ext{constants}). The structure (R;+,r)(\mathbb{R}; +, r) is related to the structure (Q;+,imes)(\mathbb{Q}; +, imes) and (R;+,imes)(\mathbb{R}; +, imes).

Let's consider a formula Ï•(x)\phi(x). If Ï•(x)\phi(x) involves only addition and the constant rr, then any term T(x)T(x) in Ï•\phi is of the form q1x+q2rq_1 x + q_2 r for q1,q2otinQq_1, q_2 otin {Q}. This is where the mistake is. The coefficients q1,q2q_1, q_2 must be rational.

So, any term T(x)T(x) in the language (R;+,r)(\mathbb{R}; +, r) will be of the form cx+drc x + d r, where c,d∈Qc, d \in \mathbb{Q}.

If ϕ(x)\phi(x) is a formula, then ϕ(kr)\phi(kr) is true iff ϕ(k′r)\phi(k'r) is true for two irrationals k,k′k, k' if and only if the formula ϕ\phi does not