14.06.2013 Views

1 The Birth of Science - MSRI

1 The Birth of Science - MSRI

1 The Birth of Science - MSRI

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

348 11. <strong>The</strong> Age-Long Recovery<br />

It seems remarkable to me that the second hypothesis [namely, that<br />

the fourth angle is obtuse] holds if instead <strong>of</strong> plane triangles we take<br />

spherical ones, for in this case not only the angles <strong>of</strong> a triangle add up<br />

to more than 180 degrees, but also the excess is proportional to the<br />

area <strong>of</strong> the triangle. 183<br />

This is a truly remarkable observation: it is (though he was unaware<br />

<strong>of</strong> that) equivalent to the statement that his results consequent upon the<br />

assumption <strong>of</strong> an obtuse fourth angle coincide with the spherical geometry<br />

<strong>of</strong> ancient times. Lambert had in fact redemonstrated some classical<br />

theorems.<br />

<strong>The</strong> Sphaerica <strong>of</strong> Menelaus, from the first century A.D., is the oldest work<br />

in non-Euclidean geometry that has come down to us. It studies the surface<br />

<strong>of</strong> the sphere not as something immersed in three-dimensional space,<br />

but through its intrinsic properties (to use the technical term). 184 Each<br />

theorem, including those on spherical triangles, is proved following the page 426<br />

scheme used in the Elements for plane geometry, but interpreting Euclid’s<br />

straight lines (line segments) as arcs <strong>of</strong> great circles. Naturally, those theorems<br />

<strong>of</strong> plane geometry that depend on the existence <strong>of</strong> a parallel to a<br />

line through a given point are not present, being replaced by different<br />

theorems valid in the spherical case. In particular, the theorem cited by<br />

Lambert on the excess <strong>of</strong> the angle sum <strong>of</strong> a spherical triangle appears as<br />

proposition 11 in Book I <strong>of</strong> the Sphaerica.<br />

To a Hellenistic mathematician, there would be no point even in posing<br />

the question whether one can construct a consistent geometry containing<br />

a theory <strong>of</strong> parallels different from the one in the Elements. This is because<br />

spherical geometry makes it obvious that the answer is yes. 185 It is reasonable<br />

to think, then, that Menelaus’ Sphaerica, containing as it does an<br />

explicit alternative to the geometry <strong>of</strong> the Elements, may have pointed the<br />

way toward the recognition that such alternatives exist. And sure enough,<br />

183<br />

J. H. Lambert, Die <strong>The</strong>orie der Parallellinien, in [Stäckel, Engel], p. 202.<br />

184<br />

However, some theorems <strong>of</strong> intrinsic sphere geometry were already present in the work <strong>of</strong><br />

<strong>The</strong>odosius; see note 71 on page 49.<br />

185<br />

For precision’s sake we ought to point out that in spherical geometry (which has no “parallel”<br />

lines) what must be abandoned is not the fifth postulate in its original form, but an assumption that<br />

Euclid makes implicitly in the pro<strong>of</strong> <strong>of</strong> proposition 16 <strong>of</strong> Book I (see note 38a on page 159). It may<br />

be objected that one must also modify the first postulate, since by interpreting lines as great circles<br />

in spherical geometry, uniqueness fails for lines passing through antipodal points. This fault can<br />

be easily remedied by treating each pair <strong>of</strong> antipodal points as a single point (spherical geometry<br />

on the projective plane).<br />

Revision: 1.11 Date: 2003/01/06 07:48:20

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!