Beruflich Dokumente
Kultur Dokumente
Colin Webster
Colin Webster: Columbia University Classics, 1130 Amsterdam Ave. 617 Hamilton Hall MC
2861, New York, New York 10027, United States, E-Mail: caw2126@columbia.edu
Introduction
When ancient Greek philosophical authors talk about visual perception, they
most often focus their attention on its physical mechanisms. For instance, such
authors as Plato, Aristotle and Theophrastus all examine various arguments for
and against opposing theories, addressing the physical aspects of eyes, light
and sight. They also mention the accounts of Empedocles, Anaxagoras and De-
mocritus, who do the same.1 In stark contrast, the Optics attributed to Euclid is
a fully geometrical treatment of vision, composed of fifty-seven mathematical
proofs.2 The Euclidian author weighs no alternative theories. He makes no phy
1 See especially Pl. Tim. 45b2c4, 67c469a4, Arist. De sens. and Theophr. De sens.
2 Before we begin an examination of the Optics, we need to determine 1) which text we are
discussing; 2) whether Euclid is its author; and 3) when such a text was composed. In regards
to the first question, there are two different versions of the Optics preserved in the manuscript
tradition, now referred to as version A and version B. Since the publication of Heiberg (1882),
edition A has been considered the authentic Euclidian text, while B was thought to have been
a later edition produced by Theon of Alexandria (ca. 335 CEca. 400 CE). Version B was thus
referred to as the Recensio Theonis. Heibergs conclusion was based on the fact that B contains
many difficult and unclear passages and he considered these unworthy of Euclid. In more recent years, however, both Knorr (1991) and Jones (1994) have pushed back against Heibergs
argument and proposed that it is more typical for an editor to add clarifications and additional
proofs, rather than muddy long, clear syntax. Moreover, B employs a highly idiosyncratic lettering scheme, unique to the Optics, while A displays the more standardized alphabetical lettering
of later geometrical texts. Jones and Knorr argue that lettering was more likely standardized
than made more idiosyncratic and therefore A is more likely the recension, while B is an earlier, more authentic version of the text. That being said, both scholars recognize that B still
contains many stylistic oddities and geometric errors uncharacteristic of Euclid in his Elements.
Thus, in regards to our second question above, then, not only do these oddities suggest that
the text contains multiple interpolations, but it also casts doubt on whether Euclid is in fact its
author.
To Knorr and Jones arguments, I will add that Geminus, Astron., Frag. Opt. 22.1415 appears
almost certainly to have had version B in the first century BCE, since he quotes its first lines:
optics establishes that visual rays from the eyes move along straight lines [ ]. This is identical to the first
definition found in B, Let us establish that visual rays from the eyes move along straight
lines [ ], (Eucl. Opt. def.
1, B), while the A edition reads Let us establish that straight lines leading from the eyes move
a distance of great magnitude [
] (Eucl. Opt. def. 1, A). Although Heiberg (1882), p. 134,
seems to acknowledge that Bs first definition is more correct, he nevertheless supplies As
version in both his 1882 and his 1895 editions of the Optics. In light of this evidence and recent
arguments, I will follow Knorr and Jones and use manuscript B as the main text for my investigation.
Lastly, it remains difficult to assess precisely when the Optics was composed or compiled, especially since Euclids own dates are uncertain (although most scholars, contra Schneider (1979),
accept the date of fl. c. 300 BCE). Nevertheless, while the text uses both and to
refer to the visual rays, Netz (1999), pp. 103113, argues that mathematical language tends towards economy, shedding superfluous terms in order to denote each concept by a single moniker (many thanks to my anonymous reader for this observation). To my mind, this suggests an
early provenance, probably sometime in the third century BCE, even if the text nevertheless
contains interpolations added later. Moreover, Bs idiosyncratic lettering system suggests that it
was composed before the practices of the Elements became fully standardized, which further
In its structure, the text displays far greater similarity to contemporary mathematical treatises, such as Euclids Elements, than it does to earlier philosophical
works concerning vision. Yet, while the Optics remains distinctly mathematical
supports this (for the emergence of standard geometrical lettering practices, see Netz (1999),
pp. 1288). An inability to date the text securely, however, has been accounted for in my argument, as well as potential multiple authorship; see below.
3 Netz (1999), pp. 89103, analyzes the practice of starting geometrical texts with definitions of
this type which he points out are actually second-order statements, since they make assertions
about entities rather than providing foundational descriptions. All translations are my own.
in form, it includes many statements that rely on certain assumptions about the
physical process of sight. For instance, even within the above definitions, the
Euclidian author commits himself to the claims that 1) vision results from the
rectilinear propagation of discrete rays [/] proceeding from the
eyes; 2) that these rays fall in the shape of a cone; and 3) that only those points
on which the rays fall are actually visible.4 As a result, objects can actually fall
in between the visual rays (prop. 3), which must therefore move side to side to
fill in the gaps (prop. 1). To be sure, if visual rays can move back and forth in
this manner, they must be real physical entities, rather than simple mathematical abstractions. Still, the text never fully explicates its implicit physical
claims and instead concerns itself solely with using geometry to articulate certain aspects of vision, including how objects of the same size can appear to be
of different magnitudes, how height and movement affect the appearance of
these magnitudes, how various geometrical shapes appear at different angles,
etc.
The sheer multitude of propositions presented in a single treatise has led
scholars to suspect that the Euclidian author must have drawn some of his
proofs from earlier works, especially since, at least a generation prior, Aristotle
mentions as though it were already a firmly established geometrical
science. Even if no extant examples survive, it is likely that optical treatises existed on which the Euclidian author could have relied.5 Scholars have proposed
several potential candidates, including Democritus lost text, the ,6
4 For a discussion of the physical assumptions implicit in Euclids definitions, see Jones
(1994).
5 Arist. Metaph. 997b20, 1077a5, 1078a14; Phys. 194a812; An. Post. 75b16, 76a24, 77b2, 78b37,
79a1120. Plato, by contrast, does not mention optics at all and instead considers the art of
measuring [ ] to be the tool used to decide whether objects that appear to be
smaller when viewed from a distance actually are smaller (cf. Pl. Prot. 356c4e4). Hahm (1978),
p. 62, suggests that the author of the Optics relied on earlier source texts, but makes no claims
as to what they might be.
6 Rudolph (2011), p. 74, suggests that this lost treatise of Democritus was a source text for
Euclid. Yet, despite her claims, as well as those of LeJeune (1957), p. 4, Keuls (1978), p. 66, and
Sider (2004), pp. 1519, it seems unlikely that the was actually an account of
vision or perspective. Thrasyllus (D. L. 9.48 = DK 68A33) lists it as a mathematical work in a
tetralogy otherwise devoted to celestial concerns, which includes Diagrams of the Heavens
[], Diagrams of the Earth [] and Diagrams of the Poles [].
Moreover, in his extant fragments Democritus uses only to refer to rays from the sun
(DK 68A91) and never mentions them in the context of eyesight presumably because (pace
Rudolph) it would conflict with his theory of vision, which involves atoms and impressions in
the air, not rays. Instead, Democritus displays considerable interest in general celestial matters
(cf. DK 68A85, 67B1, 68A87, 68A89, 59A81), in geometrical astronomy (cf. DK 68A86, 59A78,
68A87, 68A88, 59A77, 59A80, 68B14.1) and in explaining eclipses by means of the shadow cast
by the rays of the sun (cf. DK 67A1, 68A89a). Thus, it is more likely that the
provided an account of celestial illumination, not optics.
7 Suida s.v. ; cf. D.L. 3.37. See Tarn (1975). The Suida lists the title of the second
work as , which, if correct, is a hapax legomenon. Later commentators report that Aristotle himself wrote an optical treatise, called the or (Frag. var.
7.43, n. 12). This, however, could be a reference to Mete. 3.372b12377a28, which deals with
the halo and rainbow, books 15 and 31 of the pseudo-Aristotelian Problemata, which deal with
optical problems, or even a later misattributed text, such as the pseudo-Ptolemaic De speculis.
8 For various general accounts of early astronomy, see Heath (1913), (1932); Lloyd (1970), pp.
8098, (1973), pp. 5374, Pedersen (1974), Neugebauer (1975, v. 2), pp. 573776, Evans (1998)
and Graham (2013).
9 Herodot. 1.74 = DK 11A5; cf. Aristoph. Birds 9921009. Heath (1913), pp. 1223 casts doubt on
the idea that Thales knew the mechanism of eclipses and instead insists that he more likely
followed the observational records of the Babylonians. Dicks (1959) takes a skeptical position
towards Thales astronomical teachings in general; cf. Burkert (1972), pp. 413417 and Netz
(2004), p. 246. Most recently, Whrle (2009) has collected all the extant testimony for Thales,
but for a discussion of the prediction of eclipses in particular, see Bowen (2002).
10 Plut. De fac. in orb. lun. = DK 59B18 mentions demonstrating the Anaxagorean proof that
the sun gives its light to the moon [ , ]. Similarly, Plut. Nic. 23 = DK 59A18 states that The first to establish an account of the illuminations and shadow of the moon in an image, most clearly and boldly of all,
was Anaxagoras ( ); cf. 59A5, 59A42 (8), 59A76.
Anaxagoras displays an overriding interest in astronomy, having being condemned for impiety
after calling the sun a red-hot stone; cf. DK 59A1, 59A2, 59A3, 59A11, 59A19, 59A35, 59A72,
59A73. He also addresses other celestial matters to which geometry was generally applied, including rays, orbits and the size of the earth which he considered to be drum-shaped; cf. DK
59A30, 59A47, 59A72, 59A74, 59A80, 59A81, 59A87, 59A88, 59A89.
11 D.L. 2.11 = DK 59A1. Similarly, Clement, Misc. 1.78 = DK 59A36 claims that Anaxagoras was
reportedly the first to publish a book through an image/writing [ ]; cf. Plut. Nic. 23
= DK 59A18 quoted above in n. 10.
12 DK 67A1; cf. DK 68A89a.
13 Archimedes, Quadrature of the Parabola 262.12 claims that although Democritus did not actually prove the proposition, he was the first to discover that the volume of a cone was 1/3 of a
cylinder with the same height and diameter; cf. DK 68B155; Archim. Sphere and Cylinder I 4.9;
Method 430.1. For a discussion of Archimedes reliability, see Cuomo (2001), p. 50. Although
Rudolph (2011), p. 7, suggests that Democritus geometrical discovery may have been the result
of envisioning a cone of visual rays, it is far more likely that it arose from the discussion of
luminous rays and celestial shadows; cf. n. 5 above.
likely that diagrams of some sort were circulating in astronomical texts by this
time even if they did not yet take the axiomatic-deductive form characteristic
of later proofs.14 In any case, by the fourth century BCE both Plato and Aristotle
mention astronomy as a geometrical practice,15 and it flourished during the Hellenistic period in the works of Autolycus, Eudoxus, Euclid, Aristarchus, Hipparchus and others.
Some authors in antiquity seem to have noticed the resemblance that the
Optics bears to this astronomical tradition, since the treatise appears as part of
a compendium called the Little Astronomy, a set of texts that included Euclids Phaenomena, Autolycus On the Moving Sphere and On Risings and Settings, Theodosius Sphaerica, On Days and Nights and On Habitations, Aristarchus of Samos On the Sizes and Distances of the Sun and the Moon and
Hypsicles Risings.16 They must have been circulating together by at least the
beginning of the fourth century CE, since Pappus of Alexandria refers to those
who teach the Treasury of Astronomy [ ],17 and
Heath notes that a third hand in the margin of the oldest manuscript has identified this collection as the Little Astronomy [ ].18
Still, while Heath recognizes the close relationship that several of the texts
within the group bear to one another, he remains confused as to why the Optics was included, suggesting that at most it might have been used to inoculate
students against the Epicurean argument that objects including the sun
must be the size that they appear.19 At first glance, his suggestion seems reasonable, since this text is the only member of the group that does not deal
directly with celestial astronomy. Yet, rather than looking for an explanation
based on the potential utility of this optical text for budding astronomers, we
should first note that there are many parts of the Optics that appear to have
14 Netz (1999), pp. 272275, argues that lettered diagrams and axiomatic-deductive proofs did
not become part of Greek mathematics before 440 BCE. Similarly, Netz (2004) argues that Hippocrates of Chios quadrature of the lunules is one of the earliest such texts.
15 See Pl. Rep. 528d56, 530b6c1, Theaet. 145d12, Phaedr. 274c7d2, Euthyd. 290b7c6, et
al.; cf. Arist. Cael. 291a29b11, 1723, et al. Aristotle uses the term , not ,
although Lloyd (1992), p. 570, discusses the interchangeability of these terms in early Greek
astronomy.
16 Heath (1913), p. 317; cf. Heiberg (1882), p. 152. An Arabic version of the manuscript also
contains the Sphaerica of Menelaus, on which Pappus relies; see Bjrnbo (1902), p. 4, 51 and
55.
17 Pappus, Sunagoge 6.474; see Heath (1913), p. 317.
18 Heath (1913), p. 317; cf. Heiberg (1882), p. 152.
19 Heath (1913), p. 320. For this type of Epicurean argument, see At. 2.21.5; cf. Arist. De anim.
428b24. See also n. 37 below for other visual illusions against which Euclids Optics could
potentially inoculate the reader.
been born within the astronomical tradition itself. First and foremost, the Optics, just like geometrical astronomy, depicts rays interacting with objects. In
this way, it bears both a conceptual and visual similarity to the Little Astronomy as a whole, even if its rays happen to be projecting from eyes and falling
on hypothetical round objects, rather than proceeding from the sun and illuminating particular celestial spheres. Yet, more than simply having a general resemblance, the Optics contains several clues suggesting that it has derived
some of its geometrical content from this parallel astronomical tradition, most
notably the set of propositions 2327, which concern spheres, their magnitudes
and their distances.
Proposition 23 is the first in this particular series. It demonstrates that a
single eye will never see a full hemisphere, since regardless of the spheres size,
a single vantage point will never provide an unimpeded view of any full cross
section:
,
.
In whichever way a sphere is seen by a single eye, it is always the case that less than a
hemisphere is visible, but the part of the sphere that is visible appears circumscribed by
a circle (Eucl. Opt. prop. 23, B).
The Euclidian authors proof is simple and starts by cutting the sphere along its
horizontal plane to give a two-dimensional circle, (fig. 1):
20 Except otherwise indicated, all figures are my own, based on firsthand examination of
manuscripts Parisianus gr. 2107, 2347, 2350 and 2390.
This proof begins by demonstrating that when rays project from a single eye,
they will always run tangent to the circumference of a circle at two points in
front of its full diameter. That is, the rays projecting from a single eye will never
comprehend the circle across its full cross section, since some part of the circle
will always get in the way. In order to extend this conclusion to the three dimensions of a sphere, the Euclidian author simply rotates the proof around
the circles center axis. Since the rays will always run tangent to the sphere at
the same point of the circumference regardless of which circle we take as our
plane, rotating these points will inscribe a vertically oriented circle sitting
slightly in front of the diameter. Thus, we will see a figure that looks like a full
hemisphere, but is in reality slightly smaller.
Proposition 24 seeks to extend this discovery to examine how distance affects
this relationship. It demonstrates that as an eye moves closer to a sphere, it can
see progressively less of it, although the portion that is visible will occupy a progressively larger section of the visual field or, as the Euclidian author states:
The proof that follows simply repeats the geometry of proposition 23, except
that it hypothesizes a second eye, , located slightly closer to the sphere than
the first, (fig. 2).
, , , , ,
, , , . ,
, .
, .
, , , , , .
, .
, . ,
.
, .
, .
Let there be a sphere, whose center is , and from the eye at let the line join at the
center; and let line be extended through at a right angle; and let a circle be drawn
with as its diameter; and let lines , , and be joined. And so, the angles
at points and will be right angles, since they are each on a single hemisphere.
Therefore, the lines and touch the sphere at a single point. Therefore, the rays
from the eye at will fall along the lines and . In turn, let the eye at be moved
to ; and let a circle be drawn with as its diameter; and let lines , , and
be joined. And so lines and touch the sphere at a single point. And the rays proceeding from the eye at will fall along the lines and . Thus, is seen by the
angle at , but is seen by the angle at . But is larger than , although it
appears smaller. For angle is larger than angle , and the things seen by a larger angle
appear larger. Therefore the shape appears larger than , but is smaller (Eucl. Opt.
prop. 24, B).
As Heath notes, although the diagram makes it obvious that the proposition is
correct, the Euclidian author does not actually prove his assertions, since he
does not move through the requisite steps to show how the two visual cones
relate to one another.21 Nevertheless, both theorem 23 and 24 are true and potentially applicable to vision.
Propositions 2527 of the Optics further expand on these conclusions, relying on the previous proofs, but now addressing how binocular vision affects our
perception of spheres. In particular, these theorems address the relationship between a spheres magnitude and the portion of it that remains visible to two
eyes collectively. Proposition 25 states:
,
, .
When a sphere is seen through two eyes, if the diameter of the sphere is equal to the
straight distance between the eyes, a hemisphere will be seen (Eucl. Opt. prop. 25, B).
In one way, insofar as these are the sole theorems in the Optics to consider
binocular vision, they seem to display a unique concern with sight as it actually
takes places in the world. That is, by examining how binocular vision effects
the geometry of spheres, the Euclidian author may at first seem to be demonstrating sensitivity to the fact that none of his previous propositions account for
a very basic aspect of human sight, namely, that we each (generally) have two
eyes. Yet, we may ask whether these particular propositions would conceivably
ever have arisen from a direct meditation on eyesight as it takes place in the
world, especially since it is uncertain whether the Euclidian author would have
22 Cf. Archimedes, Arenarius 1.4-7 (Heiberg, v. 2, p. 244). No trace of this theory appears in
Sizes and Distances, Aristarchus only extant text.
23 At. 1.15.5.
24 Ptolem. Syntax. 3.2.
25 At. 4.13.9 mentions that Aristarchus wrote works concerned with optics, which provides
additional justification for considering the two authors together.
26 For a discussion of how Euclids Optics, Phaenomena and Elements use definitions and hypotheses, see Berggen and Thomas (1996), pp. 910.
27 Aristarchus, Sizes and Distances, hypoth. 1.
The similarity of Aristarchus proof to those in the Optics did not go unnoticed
in antiquity. In his commentary on Ptolemys Almagest, Theon of Alexandria,
editor of Euclids Optics and Elements, claims that both Aristarchus and Euclid
had demonstrated the same facts about celestial illumination.
, . .
On every side, more than a hemisphere [of the moon] is illuminated by the rays of the
sun, since the sun is larger than the moon; and because of this, the light and dark portions of the moon are not divided by its full circumference, but by a smaller one. For
these things have been demonstrated by Aristarchus and Euclid (Theon, Comment. in
Ptolem. Syntax. Math. 957).
In his note on this passage, Heiberg identifies the two proofs to which Theon is
referring as Aristarchus proposition 2 and proposition 27 of the Optics (which is
more or less the inversion of proposition 23 above).29 In most respects, it would
not be surprising that both authors demonstrate a common geometrical fact,
especially since the geometry of spheres and cones remains the same whether
applied to eyes or to the sun. That being said, Theon neither claims that Euclid
has demonstrated a general geometrical truth, nor some particular aspect of vision, but that Euclid has proved an essential feature of celestial illumination. As
it turns out, Theon is at least in some sense correct, since one important
aspect of propositions 2527 betrays the fact that these proofs have been borrowed from an astronomical context: while their geometrical arguments remain
fully applicable and valid for the illumination of celestial spheres, they do not
function for binocular eyesight. Whereas Aristarchus proofs are true, the Euclidian proofs are false.
Optics 2527 all commit the same error, but it is easiest to see in proposition
25 (fig. 4):
, , , , ,
, ,
, .
,
, ,
. , .
For let there be a sphere whose diameter is the line ; and from the points and let
lines and be extended at right angles; and from point let a line be extended
parallel to ; and let one eye be placed at and the other at ; and from the center, ,
let a line be extended parallel to . And so if, while remains stationary, the rectangle is rotated and returns again to the same spot whence it began to be moved, the
shape circumscribed by will be a circle, which runs through the center of the sphere.
Thus, only half of the sphere will be seen by the eyes and (Eucl. Opt. prop. 25, B,
emphasis mine).
At first glance the diagram may intuitively suggest that the proposition is correct: since our eyes are far enough apart, no part of the circle protrudes far
enough to prevent visual rays from striking it across its full diameter. Similarly,
the eyes are also not far enough apart to allow any rays to land behind the
circles diameter either. Thus, the rays projecting straight from our eyes will fall
tangent at the exact diameter the circle. Up to this point, the Euclidian author is
correct: a full semicircle will indeed be visible when the circles diameter is the
same magnitude as the distance between the two eyes. Nevertheless, the proposition does not stop there. Rather, proposition 25 seeks to demonstrate that
these two eyes also comprehend a perfect hemisphere. In order to demonstrate
this, the author suggests that we simply rotate the proof in this case the
rectangle around the circles center axis, . Indeed, it may seem that,
just as with proposition 24 above, the rotation translates the two dimensions of
a circle into the three dimensions of a sphere. Yet, the physical parameters of
vision make this impossible. Put simply, in the physical world, our eyes cannot
rotate; they must remain stationary along a single horizontal plane in our head.
Rotating our eyes around a center axis (by doing a cartwheel?) would at most
produce a whole series of sequential vantage points, and even these would not
allow us to see a full hemisphere all at once. Proposition 25 and for the same
reason, propositions 26 and 27 proves incorrect.
We can also demonstrate that this proof is false by splitting it into two, with
each eye producing its own unique visual cone. This way we can see what each
eye comprehends individually before combining their vantages together. Let us
again consider an overhead view of a sphere with a horizontal circular plane
produced through its middle; and let that circle have a center and a diameter
(fig. 5):
Let a line be extended from to at a right angle to ; and let a line be
extended from to at a right angle to ; and let a line connect and and
be parallel to ; and let one eye be placed at and the other at . We know
from proposition 23 that neither eye will comprehend a full semi-circle on its
own. Thus, let lines representing visual rays be projected from the eye at not
only to its tangent on the circle at point B, but also to its tangent at point A;
and let a line be extended from to the center orthogonal to BA, which it
crosses at point . And so if, while remains stationary, the triangle is
rotated and returns again to the same spot whence it began to be moved, it will
produce a cone with its vertex at the eye and its base circumference resting on
the surface of the sphere (see prop. 23). We can then repeat the same process to
produce a second visual cone extending from the other eye (fig. 6). Let lines
representing visual rays be projected from the eye at not only to its tangent
on the circle at point , but also at point ; and let a line be extended from to
Even this formulation in the Elements, however, seems to have had its origins in
celestial geometry, since the language may have been borrowed from Autolycus
of Pitane (fl. 310 BCE), whose treatise On the Moving Sphere, provides several
elementary geometrical proofs concerning the rotations required for many astronomical calculations. The first proof, dealing with the simple rotation of a
sphere about an axis, contains vocabulary almost identical to the Optics:
30 Euclid uses the same grammatical formulation to define a cone (Eucl. Elem. XI, def. 18) and
a cylinder (Eucl. Elem. XI, def. 21).
31 Aujac (1979), p. 7.
32 Diagram based on Heath (1913), p. 354, fig. 16.
Since the source of the rays in Aristarchus proposition is itself spherical, when
he rotates the parallelogram and the semicircles and to translate
the two-dimensional proof into three dimensions, the rotation inscribes both the
hemisphere of the visible object (the moon) and the edges of the illuminating
object (the sun). Since light rays project from entirety of the spherical sun, Aristarchus geometry actually works. The Euclidian author, by contrast, misapplies
this geometry and thereby produces a false proof. Even though the uncertain
date of the Optics makes it difficult to discern whether the Euclidian author borrows directly from Aristarchus, or whether both authors rely on a common
source (even though Theon mentions no such predecessor), the success of the
proposition within the context of celestial illumination makes it likely that these
proofs about rays and spheres originated out of a concern with the sun, the
moon and the earth, rather than out of a meditation on either eyes or perspective. Even though they have been reformulated, propositions 25-27 still remain
applicable to astronomy, not optics. The Euclidian author (or some predecessor)
has refashioned them based on a simple and apparent resemblance to the geometry of sight.
tators especially those from the discipline of art history investigate almost
exclusively whether the Optics is compatible with modern ideas about perspective; they thus treat the work as though it were fundamentally constructed in
order to explain how to produce two-dimensional images of a three-dimensional
world.34 Others, such as Grard Simon, have pushed back against this reading,
arguing that such an interpretation takes for granted that the text deals with the
propagation of light, rather than, as seems more accurate, the propagation of
sight.35 Others still, such as Thomas Heath and Silvia Berryman, suggest that
the Optics was written to combat false appearances and thereby to solidify visual information as a valid access point to knowledge or, in other words,
these scholars argue that the text is rooted in a set of philosophical concerns.36
All these interpretations, however, implicitly assume that the Euclidian author
possesses a more or less unified goal whether a phenomenon to explain or a
philosophical point to demonstrate that he then uses his mathematical expertise to achieve. Or, to use slightly broader terms, the Euclidian author first has a
set of questions and then seeks to answer them. While propositions 23-27 may
be modest adaptations, they cloud these typical assumptions about the priority
of question (primary) over answer (secondary). Instead, the Euclidian author
would have had a considerable number of proofs at his disposal, which could
act as ready-made tools to be applied to new (physical) questions. Yet, by recognizing the indebtedness that the Optics bears to earlier astronomical texts, we
can ask whether such questions always existed especially since the appearances that the Euclidian author sometimes attempts to explain do not actually
appear in any normal sense of the word. They are phenomena created by the
very explanations that the Euclidian author employs.
The problem of false false appearances is not unique to the propositions
borrowed from celestial geometry. C.D. Brownson and Wilburt Knorr have de
34 Panofsky (1991), p. 66; cf. Andersen (1987). For the most recent discussion in this vein, see
Sinisgalli (2012).
35 Simon (1988).
36 Heath (1913), p. 320. Berryman (1998); cf. Jones (1994), p. 47. This idea appears as early as
Geminus, Frag. Opt. 22, who lists the apparent convergence of columns, the rounding of
towers and unequal things appearing equal as phenomena that optics explains; cf. Procl. in
Euclidis Elementiis 40, who also claims that optics has as its subject the illusions created by
objects at a distance, and he also cites the convergence of parallel lines and the round tower as
examples. Plato presents different examples of the paradigmatic visual illusions, including
straight things appearing bent in water, painting with shadows to make flat surfaces appear
concave or convex (which he calls ) and things of the same magnitude appearing to
be of different sizes at various differences; see Pl. Rep. 602c78; cf. Rep. 523b5c3, Prot. 356c4
e4.
37 Eucl. Opt. prop. 35, B: And if a line is extended not at right angles to the plane (of the
circle), but it is equal to the diameter of that circle, the diameters will appear equal [
, ,
].
38 Brownson (1981); Knorr (1992).
39 Unlike Eucl. Opt. prop. 18, B, the A manuscript includes a very awkward reference to the
eye, claiming that it would be at the very edge of the objects shadow (e.g. on the ground at
point ). This is a completely unnecessary inclusion if one is simply measuring the triangles
(as well as being physically quite difficult). To my mind, this suggests that the inclusion of the
eye has been added in the later edition; see n. 42 below.
40 Even though no theoretical concerns seem to explain the inclusion of these proofs dealing
with similar triangles, the measurement of an object by its shadow becomes crucial for the use
of the gnomon in celestial astronomy and for estimating the height of a distant wall (for instance, when designing siege engines) in dioptrics. Therefore, the propositions could conceivably have been included for their practical use. Yet, in the Optics, the Euclidian author does not
present these propositions as if they have these specific purposes in mind, so arguing that
these are the real reason the proofs were included would be misleading, since it is simply in
the nature of geometry as elemental as this that it can be applied to multiple situations.
The first person verb form seems to allude to the fact that the author himself
discovered the application of this proof perhaps as he wandered around in
the sun with a stick. This suggests that this particular group of propositions
may have been incorporated, not out of a deep concern with articulating vision
per se, but because the Euclidian author had found a way to measure the height
of objects by using their shadows and thought it was relevant or at least related to the subject at hand. In fact, the remaining propositions in this group
seem to be a simple extension of this basic insight, insofar as they utilize the
geometry of similar triangles to describe three comparable examples. At the
same time, however, these solutions are somewhat difficult to enact in practice.
All four rely on already knowing some of the magnitudes involved in the calculation, and it is not always clear how this would be the case in most practical
situations. For instance, Optics proposition 20 suggests that to calculate a given
depth, we could simply stand above it, near the depths edge, and measure the
distances from the edge to our feet and from our feet to our eye. Since this
would produce a right-angled triangle with a known ratio, we could extend the
41 The first person verb does not appear in the A version of the text. To my mind, this further
suggests that B is older, and the proofs have been streamlined and polished in A.
visual ray from our eye along the path that aligns with the edge and mark the
place where it strikes the ground on the floor of the depth. We could then measure back from that point to the place directly under the edge itself, creating
another right-angle triangle with the same ratio of sides. We could then use the
ratio from the first triangle to calculate the missing height of the second. Of
course, this could only work if the depth 1) had a completely vertical side (for
example, a cliff or well); and 2) it was possible to measure from the base of that
wall to the point where our visual rays falls. This would be extremely difficult
to do in almost all circumstances where we could not just as easily measure the
depth itself. The proposition may be true, but its practical application seems
quite minimal. Therefore, we can conclude that the primary justification for including these proofs is that they 1) utilize the geometry of rays and 2) seem to
work mathematically. Any theoretical concern or actual practical application
seems secondary, derivative and perhaps even somewhat irrelevant.
The fact that the Optics relies on earlier astronomical propositions, incorporates erroneous proofs and attempts to measure heights complicates the idea
that the treatise is about explaining light, sight or false appearances in any
straightforward sense. Even though the text has certain physical implications
and relies on certain physical assumptions, vision is not encountered solely or
even primarily as a phenomenon outside the diagrams; it is instead constructed
as a geometrical entity within them. In fact, in several moments within the Optics, the geometry even leaves vision behind. In short, the text reflects a conglomerate of concerns and dependencies, oriented toward a somewhat diffuse
goal: simply to articulate the geometry of vision and rays. In this way, the text
could be said to be as much about diagrams as it is about sight.
The use of the first person verb in the above quotation also helps draw our
attention to a problem that we have not yet addressed, namely, whether we can
treat this mathematical work as though it is the product of a single author.42
The Optics itself survives in multiple editions and, as a survey-type text, likely
drew from multiple sources. As such, we cannot simply assume that the Euclidian author of the Optics was the first to re-fashion the pre-existing proofs from
celestial astronomy to fit an optical context. Rather, the Euclidian author could
42 Netz (2009), pp. 107109 classifies three different types of mathematical works: 1) ludic, 2)
survey; and 3) pedagogical. Ludic works are constructed around a single mathematical purpose
(such as calculating a given value) and will incorporate surprise by delaying the justification
behind intermediary steps in order to dazzle the reader. Pedagogical works privilege lucidity
and lead the reader to an explicit goal (or set of goals). Survey works, like the Optics, attempt
an exhaustive account of a given subject (e.g. Eucl. Elem. X, which provides an account of
irrationals). Netzs discussion of survey works, however, focuses more on theoretical geometry
and does not deal with applied mathematics in particular.
Conclusion
As a consequence of recognizing the various motivations behind the Optics propositions, it is harder to maintain that the text is wholly about addressing a
single philosophical or mathematical goal. It reflects different priorities at different times. Sometimes geometry is used to articulate physical ideas about
sight. Sometimes preexisting mathematical propositions help construct vision
as a geometrical phenomenon. Sometimes geometry is used for more mundane
tricks. Indeed, in recent years, scholars have begun to recognize the diversity of
motivations behind the construction of ancient geometrical treatises. Reviel
Netz has shown that Hellenistic geometers responded to literary as well as
mathematics traditions, demonstrating aesthetic sensibility by incorporating
elements of surprise and delight into their arrangement of proofs.43 Similarly,
Serafina Cuomo has argued that Pappus Sunagoge is designed to engage with a
number of different traditions, including mathematical, scientific and philosophical debates for which reason she refers to the text as a palimpsest.44 In a
comparable way, we can shift the assumption that what Aristotle calls more
physical mathematics [ ]45 is involved in the
pure explanation of natural phenomena by geometrical means. As the above
propositions within the Optics help demonstrate, authors can adopt pre-existing
geometrical proofs, modifying them for subject matter at hand. In many ways,
this remains the fundamental strength of theoretical geometry, since it is precisely by relying on earlier discoveries that knowledge can advance. In the case
of applied geometry, however, this occasionally creates inconsistencies insofar
as repurposed proofs can fail to account for the restrictions imposed by the physical world. In some instances, the phenomena explained by these texts can
even be phantoms of geometry, appearing only within the confines of the diagram. Tracking these inconsistencies allows us insight into how these treatises
were composed and helps us understand the complex relationship between construction and explanation, as well as between mathematical entities and the
objects they supposedly represent.
Works Cited
Andersen, K. Ancient Roots of Linear Perspective, in From Ancient Omens to Statistical Mechanics: Essays on the Exact Sciences Presented to Asger Aaboe, edd. J. L. Berggren and
B. R. Goldstein. Copenhagen: University Library, 1987.
Asper, M. The two cultures of mathematics in ancient Greece, in The Oxford Handbook of
the History of Mathematics, edd. E. Robson and J. Stedall. New York: Oxford University
Press, 2009, 107132.
Aujac, G. Autolycos de Pitane. Paris: Les Belles Lettres, 1979.
Berggen, J. L. and R. S. D. Thomas. Euclids Phaenomena: A Translation and Study of a Hellenistic Treatise in Spherical Astronomy. New York: Garland, 1996.
Berryman, S. Euclid and the Sceptic: A Paper on Vision, Doubt, Geometry, Light and Drunkeness, Phronesis, Vol. 43, No. 2 (1998):176-196.
Bowen, A. C. The Art of the Commander and the Emergence of Predictive Astronomy, in
Science and Mathematics in Ancient Greek Culture, edd. C. J. Tuplin and T. E. Hill. New
York: Oxford University Press, 2002, 76-111.
Bjrnbo, A. A. Studien ber Menelaos Sphrik. Beitrge zur Geschichte der Sphrik und
Trigonometrie der Griechen, in Abhandlungen zur Geschichte der mathematischen Wissenschaften, Heft 14. Leipzig: Teubner, 1902.
Brownson, C. D. Euclids Optics and its Compatibility with Linear Perspective, Archive for
History of Exact Sciences 24, (1981): 165194.
Burkert, W. Lore and Science in Ancient Pythagoreanism. Cambridge, Mass.: Harvard University Press, 1972.
Cuomo, S. Pappus of Alexandria and the Mathematics of Late Antiquity. New York: Cambridge
University Press, 2000.
Cuomo, S. Ancient Mathematics. New York: Routledge, 2001.
Dicks, D. R. Thales, CQ 9, No. 2 (1959): 294309.
Evans, J. The History and Practice of Ancient Astronomy. New York: Oxford University Press,
1998.
Graham, D. W. Science Before Socrates: Parmenides, Anaxagoras and the new Astronomy.
New York: Oxford University Press, 2013.
Hahm, D. E. Early Hellenistic Theories of Vision and Perception of Color, in Studies in Perception: Interrelations in the History of Philosophy and Science, edd. Peter K. Machamer
and Robert G. Turnbull. Columbus: Ohio State University Press, 1978.
Heath, T. Aristarchus of Samos: The Ancient Copernicus. New York: Dover, 1913, repr. 1981.
Heath, T. Greek Astronomy. New York: E. P. Dutton & Co. Inc, 1932.
Heiberg, J. L. Litterargeschichtliche Studien ber Euklid. Leipzig: Teubner, 1882.