What does it mean for Euclidean geometry of the sphere to be inherited from 3-d Space

"The answer is that this Spherical Geometry was merely considered to be inherited from the Euclidean geometry of the 3-dimensional space in which the sphere resides. No thought was given in those times to the sphere's internal 2- dimensional geometry as representing an alternative to the Euclid's plane. Not only did it violate Euclid's fifth axiom, it also violated a much more basic one (Euclid's first axiom) that we can always draw a unique straight line connecting two points , for this fails when two points are antipodal.

Tristan Needham VDG Pg-7, sect.1.2 Spherical Geometry

What does it mean for Euclidean geometry of the sphere to be inherited from 3-d Space? I can't quite understand the language of the above paragraph.

Related

In this setting, when we speak of "geometry", we're talking about a choice of a Riemannian metric on a smooth manifold. I'm sure you can look up the precise definitions yourself, but the idea is that Riemannian metrics are what define the notion of a geometry on a given space; they prescribe at "an infinitesimal level" (i.e at the level of tangent spaces) an inner product, meaning a notion of lengths and angles (once we have an "infinitesimal" description of lengths and angles, we can build up to a general notion of distances, areas, volumes etc, i.e all the basic "geometric" notions).

To say spherical geometry is "inherited" from the Euclidean geometry in $\Bbb{R}^3$ means the following. Let $(x,y,z)$ be the Cartesian coordinates in $\Bbb{R}^3$ (i.e the identity chart) and define $g=dx\otimes dx+dy\otimes dy+dz\otimes dz$, then $g$ is a Riemannian metric on $\Bbb{R}^3$. Now, let $\iota:S^2\to \Bbb{R}^3$ be the canonical inclusion mapping. Then, we can consider the pullback tensor field $\iota^*g$ on $S^2$. Since $\iota$ is an immersion (i.e at every point $p\in S^2$ its tangent linear mapping $T\iota_p:T_pS^2\to T_{\iota(p)}\Bbb{R}^3=T_p\Bbb{R}^3$ is injective) it follows that $\iota^*g$ is indeed a Riemannian metric on $S^2$. This is what we refer to as the Riemannian metric "induced" on $S^2$.

More generally, if you start with a Riemannian manifold $(M,g)$ and you consider an embedded submanifold $S\subset M$, then we can look at the inclusion $\iota:S\to M$, and then look at the pullback $\iota^*g$. This way $(S,\iota^*g)$ is a Riemannian manifold in its own right (i.e it has its own notion of lengths and angles "on an infinitesimal level"), and we call this a Riemannian submanifold of $(M,g)$. We also say that $S$ inherits its geometry from $M$ (more precisely from $(M,g)$).

I'm sure this terminology of tensor fields, smooth manifolds, pullbacks etc has not been introduced by page 7 of the book, but that's what is meant (even if at this stage of the book one does not have access to this terminology). Also, one of course has to define what "straight lines" mean in this general context, but you'll get to all of this as you continue reading.