Throughout, π denotes the map described in claim 1, and coordinates of points of Rm+n are compared using the index decomposition recorded in the statement: every k∈[m+n] lies in [m] or equals m+j for exactly one j∈[n], and not both.
Claim 1. Let w∈Rm+n and let π(w)=(ξ,η) with ξk=wk for k∈[m] and ηj=wm+j for j∈[n]. Then ι(π(w)) and w have the same coordinate at every index of [m+n], by the two defining clauses of ι, so ι(π(w))=w. Conversely, for (ξ,η)∈Rm×Rn the first entry of π(ι(ξ,η)) has kth coordinate ι(ξ,η)k=ξk for k∈[m], and its second entry has jth coordinate ι(ξ,η)m+j=ηj for j∈[n]; hence π(ι(ξ,η))=(ξ,η). So ι and π are mutually inverse, and ι is a bijection.
Claim 2. Two points of Rm+n are equal exactly when they agree at every index of [m+n], so it suffices to compare coordinates at k∈[m] and at m+j with j∈[n].
By Sum of Points of Rn the kth coordinate of ι(ξ,η)+ι(ξ′,η′) is ι(ξ,η)k+ι(ξ′,η′)k. For k∈[m] this is ξk+ξk′, which is the kth coordinate of ξ+ξ′ and hence of ι(ξ+ξ′,η+η′); for the index m+j with j∈[n] it is ηj+ηj′, which is the jth coordinate of η+η′ and hence the (m+j)th coordinate of ι(ξ+ξ′,η+η′). This proves the first identity. The second is identical, using Scalar Multiple of a Point of Rn and that the kth coordinate of λz is λzk; the third is identical, using the coordinatewise description of the difference in Difference, Dot Product, and Orthogonality in Rn.
Claim 3. Let c:[m+n]→R be the family ck=ι(ξ,η)kι(ξ′,η′)k, so that by Difference, Dot Product, and Orthogonality in Rn
ι(ξ,η)⋅ι(ξ′,η′)=k=1∑m+nck.
By Splitting a Finite Sum at an Index this equals ∑k=1mck′+∑j=1ncj′′, where c′ is the restriction of c to [m] and cj′′=cm+j. For k∈[m] we have ck′=ξkξk′, so ∑k=1mck′=ξ⋅ξ′; and for j∈[n] we have cj′′=ηjηj′, so ∑j=1ncj′′=η⋅η′. This proves the first identity of claim 3.
Taking ξ′=ξ and η′=η and using claim 1 of Elementary Properties of the Euclidean Norm on Rn, which gives ∥z∥2=z⋅z in each Euclidean space, yields ∥ι(ξ,η)∥2=∥ξ∥2+∥η∥2.
Claim 4. By claim 2 of Elementary Properties of the Euclidean Norm on Rn, dE(u,u′)=∥u−u′∥ in each Euclidean space. By claim 2 of the present lemma, ι(z)−ι(z′)=ι(ξ−ξ′,η−η′), so by claim 3,
dE(ι(z),ι(z′))2=∥ι(ξ−ξ′,η−η′)∥2=∥ξ−ξ′∥2+∥η−η′∥2=dE(ξ,ξ′)2+dE(η,η′)2,
which is the first assertion.
Write α=dE(ξ,ξ′), β=dE(η,η′), γ=dE(ι(z),ι(z′)) and μ=d×(z,z′), all nonnegative because a metric takes nonnegative values. By Product Metric on the Cartesian Product of Two Metric Spaces, μ is the maximum of α and β, so α≤μ and β≤μ by claim 1 of Elementary Properties of the Maximum of Two Elements, and μ equals α or β by claim 2 of that lemma.
Since μ is one of α,β, we get μ2≤α2+β2=γ2, using that the other square is nonnegative by Nonnegativity of Squares in an Ordered Field and claim 3 of Elementary Arithmetic in an Ordered Field. As μ and γ are nonnegative, claim 2 of Monotonicity of Squaring on the Nonnegative Elements of an Ordered Field gives μ≤γ.
For the upper bound, α≤μ and β≤μ give α2≤μ2 and β2≤μ2 by claim 2 of Monotonicity of Squaring on the Nonnegative Elements of an Ordered Field, so
γ2=α2+β2≤μ2+μ2≤μ2+(μμ+μμ)+μ2=(μ+μ)2,
the middle inequality holding because μμ is nonnegative, and the last equality being field arithmetic. Since γ and μ+μ are nonnegative, claim 2 of Monotonicity of Squaring on the Nonnegative Elements of an Ordered Field gives γ≤μ+μ.