What is the History of Medieval Optics

Really About?1
A. MARK SMITH
Professor of History, University of Missouri at Columbia

S INCE ITS PUBLICATION IN 1975, David Lindberg’s Theo-

ries of Vision from Al-Kindı̄ to Kepler2 has become the canonical
source for our understanding of medieval optics and its place in
the development of modern optics. Lindberg’s ulterior purpose in writ-
ing this book was to show that, contrary to prevailing opinion, Johan
Kepler’s account of sight, which is based on the casting of point-by-
point images through the lens at the front of the eye onto the retina at
the back, represented a continuation of, rather than a break with, the
medieval optical tradition, whose foundations were laid by the Arab
Ibn al-Haytham, or Alhacen as he came to be known in the Latin
West.3 To be sure, Lindberg concluded, “Kepler attacked the problem
of vision with greater skill than had theretofore been applied to it, but
he did so without departing from the basic aims and criteria of visual
theory established by Alhazen in the eleventh century.”4 So, even if we
credit Kepler with opening the way toward the development of modern
optics, neither the opening itself nor what came out of it during the
seventeenth century marked a fundamental departure from the past.
The transition from medieval to modern optics was evolutionary, not
revolutionary.

1 Read 10 November 2001.

2 Chicago: University of Chicago Press, 1975.
3 Theories of Vision, 205–07. Lindberg’s argument is directed primarily at Steven S.

Straker, “Kepler’s Optics: A Study in the Development of Seventeenth-Century Natural

Philosophy” (Ph.D. diss., Indiana University, 1970) and A. C. Crombie, “Early Concepts of
the Senses and the Mind,” Scientific American 210.5 (1964): 108–16 and “The Mechanistic
Hypothesis and the Scientific Study of Vision: Some Optical Ideas as a Background to the
Invention of the Microscope,” in Historical Aspects of Microscopy, ed. S. Bradbury and
G. L’E. Turner (Cambridge: Heffer, 1967), 3–112. Until recently, it was commonly assumed
that “Alhazen” was the proper Latin form of Ibn al-Haytham’s proper name (“al-Hasan”),
but recent scholarship makes it clear that “Alhacen” was the preferred form during ˙ the
Middle Ages.
4 Theories of Vision, 207.

PROCEEDINGS OF THE AMERICAN PHILOSOPHICAL SOCIETY VOL. 148, NO. 2, JUNE 2004

[ 180 ]
medieval optics 181

Lindberg’s case rests on the fact, which he demonstrated in meticu-

lous fashion, that Alhacen and Kepler based their accounts of vision on
similar theoretical and methodological grounds. Both, that is, used the
same basic model of punctiform light-radiation, and both relied upon
the same ray-analytic procedure in tracing out the logical implications
of that model. That the two reached markedly different conclusions is
beside the point. The path that led to them was the same; Kepler sim-
ply followed it more rigorously. It would therefore seem reasonable to
conclude with Lindberg that, in sharing the same analytic principles,
Alhacen and Kepler operated within the same conceptual framework.
But this conclusion is grounded in a problematic assumption: namely,
that, like Kepler and his seventeenth-century successors, Alhacen and
his Latin medieval followers were primarily concerned with the physics
of light and issues more or less directly relating to it.5 As we shall see in
due course, this assumption applies to the first group, but not to the
second. For, unlike Kepler and his seventeenth-century successors, Alhacen
and his medieval Latin followers were far more concerned with making
sense of sight than with understanding light. Thus, while there is an
undeniable link between Alhacen’s and Kepler’s accounts at the proce-
dural level, the two are worlds apart at the conceptual level.
Before examining Alhacen’s account of vision in detail, let us briefly
set the background. The primary source for that account is the Kitāb
al-Manāzir, or “Book of Optics.” Written in the 1030s, this work com-
˙ books, the first three of which are devoted to a close anal-
prises seven
ysis of visual perception, the second three to an equally close analysis
of reflection, and the last to a study of refraction.6 Alhacen seems to
have intended the Kitāb al-Manāzir as a critical response to Ptolemy’s
Optics, which predated it by nearly ˙ nine centuries. In particular, Alhacen
challenged the central supposition of Ptolemaic optics that the eye
emits visual flux, which passes outward from its centerpoint through
the pupil in a radial bundle that forms a cone, as illustrated in figure 1.

5 See ibid., x–xi. Compared to his predecessors, in fact, Lindberg took a fairly broad view

of pre-Keplerian optics to include not just the physics of light but also the mathematics and
physiology of vision. Yet he drew the line at psychology and epistemology, claiming that “this
limitation is not only expedient but also legitimate, since psychological and epistemological
issues, though raised within the context of visual theory, were never its central concerns.” By
thus precluding psychology and epistemology, Lindberg effectively dismissed the perceptual
aspects of vision to focus on its physical aspects—i.e., how light and color are radiated into
and through the eye.
6 For a critical Arabic edition and English translation of the first three of these seven

books, see A. I. Sabra, ed., Ibn al-Haytham, al-Manāzir I-II-III 5 Kitāb al-Manāzir. Books I-
II-III: On Direct Vision (Kuwait: National Council for ˙ Culture, Arts, and Letters,˙ 1983) and
The Optics of Ibn al-Haytham: Books I-III on Direct Vision, 2 vols. (London: Warburg,
1989).
182 a. mark smith

The rays within this cone are to be

thought of by analogy to fingers that
reach out to external objects and put
us into visual touch with them. The
information garnered from this visual
contact is conveyed back through the
cone of flux in the form of color-
impressions. These color-impressions are
subject to visual scrutiny after reaching
the anterior surface of the eye, and
Figure 1 from them we eventually get an inter-
nal “picture” of the objects they repre-
sent.7 But why posit such visual radiation, Alhacen demanded, when it
is perfectly sufficient to assume that objects are seen by means solely of
what they radiate to the eye? In thus rejecting Ptolemy’s extramission-
ism, Alhacen set himself the task of reformulating visual theory on an
intromissionist basis.
Sometime around 1200 the Kitāb al-Manāzir was rendered from
Arabic into Latin under the title De aspectibus. ˙ 8 All indications are
that this Latin version of the Kitāb al-Manāzir lay fallow for the next
fifty years or so until it began to be studied ˙in earnest by a succession
of Latin thinkers, starting with Roger Bacon. Out of this renewed
interest developed the science of perspectiva, which found its core in
the optical writings of Roger Bacon, Witelo, and John Pecham, all of
which were composed between roughly 1262 and 1280.
The core itself consists of four basic treatises, the earliest of which,
Roger Bacon’s De multiplicatione specierum, was probably composed
in the early 1260s. A few years later Bacon produced his Perspectiva,
which was originally included as part of the Opus majus but was soon
disseminated as an independent work.9 By the mid-1270s the third of

7 For a detailed account of Ptolemy’s model of visual perception, see A. Mark Smith,

Ptolemy’s Theory of Visual Perception, Transactions of the American Philosophical Society,

86.2 (Philadelphia: American Philosophical Society, 1996), 19–35. Cf., however, Gérard
Simon, Archéologie de la vision: L’optique, le corps, la peinture (Paris: Seuil, 2003), 131–64.
8 For a critical edition, English translation, and study of the first three books of the De

aspectibus, see A. Mark Smith, Alhacen’s Theory of Visual Perception, Transactions of the
American Philosophical Society, 91.4 and 91.5 (Philadelphia: American Philosophical
Society, 2001). The subsequent discussion of Alhacen’s visual theory will be based on this
edition and translation.
9 For Bacon’s De multiplicatione specierum , see David C. Lindberg, Roger Bacon’s

Philosophy of Nature (Oxford, 1983); for Bacon’s Perspectiva, see David C. Lindberg,
Roger Bacon and the Origins of Perspectiva in the Middle Ages (Oxford: Clarendon Press,
1996).
medieval optics 183

the core works, Witelo’s Perspectiva, was completed,10 only to be fol-

lowed within five years or so by the last and most popular of the Per-
spectivist treatises, John Pecham’s Perspectiva communis.11 With certain
theoretical elaborations provided by these three key Latin thinkers, the
Alhacenian model of light and vision represented the ne plus ultra in
optics from roughly 1300 to 1600.12
To make proper sense of that model requires, first, understanding
Alhacen’s theory of light and color.13 Start by thinking of light as a real
and inherent quality in any self-luminous or illuminated object. Think
of the object’s surface as a composite of infinite point-sources from
which light propagates in all directions. Think of the propagation itself
in terms of replication, each point-source creating formal representa-
tions of itself in any continuous transparent medium, such as air. Think
of the overall result as a sphere of propagation. And think of each
radius within that sphere as a rectilinear trajectory along which point-
forms of light are transmitted outward from the center.
Transparency, for its part, is what enables light-forms to penetrate
certain objects without hindrance. Opacity balks such penetration and,
in doing so, lets certain objects capture the light-forms reaching them. 14
This capture is manifested on their surfaces as illumination, and, once
illuminated, they become sources of light-radiation in their own right.
But all opaque objects are inherently colored, and light is naturally apt
to mingle with color. So what actually radiates from the surfaces of
visible objects is luminous color, not pure light. Pure light, in fact, is a
mere theoretical abstraction for Alhacen and his Latin disciples. Its
proper function is not to be seen—i.e., it is not really per se visible—
but rather to render color visible.
How we see things by means of the luminous color they radiate to
the eye depends on how the optic system is designed to apprehend it.15
As conceived by Alhacen, whose account derives from Galen, the eye

10 See Sabetai Unguru, Witelonis Perspectivae liber primus, Studia Copernicana 15

(Warsaw: Polish Academy of Sciences, 1977) and Witelonis Perspectivae liber secundus et
liber tertius, Studia Copernicana 28 (Warsaw: Polish Academy of Sciences, 1991). See also
A. Mark Smith, Witelonis Perspectivae liber quintus, Studia Copernicana 23 (Warsaw: Polish
Academy of Sciences, 1983).
11 See David C. Lindberg, John Pecham and the Science of Optics (Madison, Wis.:

University of Wisconsin Press, 1970).

12 Friedrich Risner’s landmark tandem edition of Alhacen’s De aspectibus and Witelo’s

Perspectiva in 1572 under the title Opticae Thesaurus represents the final stage in the
dissemination of this model.
13 For Alhacen on light and color, see Smith, Alhacen’s Theory, liii–liv.
14 For Alhacen on opacity and transparency, see Smith, Alhacen’s Theory, liv–lv.
15 For Alhacen on the optic system, which includes not just the eye but its connection to

the brain via the optic nerve, see Smith, Alhacen’s Theory, lvii–lx.
184 a. mark smith

consists of two nesting, but ec-

centric, spheres. As illustrated
in figure 2, the outer sphere,
centered on A, is contained by
the scleral tunic. The cornea
at the front forms a transpar-
ent continuation of this tunic.
The inner sphere, centered on
B, is enclosed by the uveal tu-
nic. At its front lies the round
Figure 2 aperture that forms the pupil.
The uveal sphere contains three
transparent fluids, or humors, of differing optical density. The rearmost,
and most optically dense, of the three is the vitreous humor. Ahead
of it lies the glacial humor, and ahead of this lies the albugineous
humor. The interface between albugineous and glacial humors forms
a convex spherical section concentric with the eye as a whole and
therefore with the cornea. The interface between glacial and vitreous
humors forms a concave spherical section. Together, these interfaces,
which are defined by a gossamer membrane called the aranea (i.e.,
spider web), mold the glacial humor into the shape of a double con-
vex lens.
Both the scleral and uveal spheres have a small opening at the rear
through which visual axis AB passes. Extending back from their re-
spective openings, the tunics enveloping these spheres form the inner
and outer sheaths of the hollow optic nerve. The optic nerves of both
eyes continue through the eyesockets to join at the optic chiasma, and,
after parting, wend their way to the fore-
front of the brain, as represented in figure
3. There each nerve feeds into the front-
most of three successive cells, or ventricles,
that are replete with animal spirit, which is
a distillate of pneumatized—or “aerated”—
blood that passes from the left ventricle of
the heart through the carotid arteries into
the base of the brain. Serving as the me-
dium for all sensitive, perceptual, and in-
tellectual functions, animal spirit perfuses
the entire optic complex from the front of
the brain to the anterior surface of the gla-
cial humor. There it takes specific form as
visual spirit, which is peculiarly sensitive to
Figure 3 visible stimuli. Thus sensitized, the glacial
medieval optics 185

humor is prepared to accept physical light- and color-impressions in a

visual way.16
Now, let us place a luminous object, such as BCD in figure 4, di-
rectly in front of the eye. According to the Alhacenian account, each
point on the object’s surface will radiate its form through every fac-
ing point on the cornea to the anterior surface of the glacial humor.
Thus, point D on the luminous object’s surface radiates its form through
the cornea to distinct and separate points, such as E, F, and G, on the
anterior surface of the glacial humor. Likewise, every point on the ob-
ject’s surface will radiate its form through the cornea to each point on
the anterior surface of the glacial humor. For instance, points B, C, and
D on the luminous object’s surface all radiate their forms through the
cornea to the same point E on the anterior surface of the glacial humor.
To explain how a coherent visual impression is sifted from this
chaos of overlapping forms, Alhacen appealed to the refractive and
sensitive properties of the glacial humor.17 First, the impinging color-
form must radiate through the refractive interface between albugineous
and glacial humors. Among all such forms, only those that reach this
interface orthogonally, that is, along perpendiculars BE, CA, and DF in
figure 5, will pass unrefracted. The rest, whose radial paths are repre-
sented by DE and CF, will be deflected out of the account. Second, by
analogy to a projectile, forms of luminous color impinge most force-
fully when they strike the interface directly along the orthogonal. The
more obliquely they strike, the more glancing the impingement. Being
most forceful and, therefore, most acutely felt by the glacial humor, the
orthogonal impingements are the only ones that make a sensible
impression on it. The rest are simply ignored. So it is on the basis of its
sensitivity that the glacial humor abstracts a coherent visual represen-
tation, or “image,” B9C9 in figure 6, that is in proper point-to-point
correspondence with its generating object BC. On the basis of its
refractivity, meanwhile, the glacial humor allows each point on the
abstracted image to radiate straight through toward the center of
the eye. The field of effective vision is therefore defined by a cone of visi-
bility with its base in the visible object BC and its vertex at centerpoint
A of the eye. Defining the center of sight, this vertex serves as the

16 This anatomical and physiological model is Galenic in origin, and many of the details,

such as the ventricular structure of the brain and the manufacture of animal spirit, are not
explicitly mentioned by Alhacen. Roger Bacon, however, was explicit in tying the Galenic
model of the brain, in all its details, to the Alhacenian account of vision. See Smith, Alhacen’s
Theory, xxxvii–xlix and lxxxiv–xc.
17 For Alhacen on the refractive and sensitive selection of visual images by the crystalline

lens, which is formed of glacial humor, see Smith, Alhacen’s Theory, lx–lxii.
186 a. mark smith

Figure 4 Figure 5

Figure 6 Figure 7

ultimate reference-point for the geometrical analysis of sight within the

cone of visibility. On that basis, we gauge the distance of objects accord-
ing to how far they lie from the center of sight, whereas we gauge their
size according to a correlation of their perceived distance and the
visual angle they subtend at the center of sight.18
If every point-form on the image abstracted by the glacial humor
passed straight through the eye, then the rays along which they are
transmitted would cross at centerpoint A and continue on in inverted
order. The resulting image would therefore be upside down. But in
meeting the refractive interface between glacial and vitreous humors
before centerpoint A, these rays are deflected outward by being bent
toward the normal as they pass from the optically rarer glacial humor
into the optically denser vitreous humor. In this way, image B0C0 in fig-
ure 7, which is transmitted along these refracted rays, is channeled in
proper, upright order into the hollow optic nerve at the back of the eye.

18 For Alhacen on spatial perception, see Smith, Alhacen’s Theory, lxiv–lxvi.

medieval optics 187

At that point it will have shrunk enough to fit into the narrow opening
so as to be conveyed in proper, upright order through the visual spirit
pervading the nerve. And when it finally reaches the optic chiasma, it
will be fused with its mate from the other eye. Such fused images are
what we normally see in binocular vision, although in abnormal cir-
cumstances, when the two images fail to fuse properly, we see double. 19
Let us go back for a moment to the object represented in this pro-
cess. According to the Alhacenian account, particularly as it was inter-
preted by his Latin followers, all physical objects present themselves to
us in various ways and at various levels. This brings us to the notion of
intentionality.20 Take, for example, Georges Seurat’s well-known Sun-
day Afternoon on the Island of la Grande Jatte. At the very lowest
level, it represents a mere composite of colored dots. As such, it can be
likened to the pointillist representation formed at the anterior surface
of the glacial humor. At a somewhat higher level, though, it represents
a number of bodies in space. At a higher level yet, it represents a group
of men and women distributed in various positions and poses on a
tree-shaded lawn by water’s edge. And at an even higher level, it repre-
sents a social event occurring in Paris during the summer of 1884.
Nevertheless, although we may “see” all of these things portrayed in
the painting, they are not actually there. They are merely implied, or
intended, by the peculiar juxtaposition of colors on the painting’s sur-
face, and they remain implicit until they are subjectively realized through
perception.
Just so, the visual image abstracted by the glacial humor and con-
veyed to the optic chiasma is an intentional, and therefore subjective,
representation of its generating object. As such, the visual image ex-
presses the visible nature of the object it represents. But more to the
point, in expressing the visual realization of that nature, it represents
the very seeing of the object. This representation, in turn, intends a
host of things that are not actually in it, since it consists of nothing
more than colors rendered visible by the sensitized glacial humor. By
Alhacen’s count, there are twenty such ulterior intentions, ranging from
size, shape, and spatial disposition to corporeity, opacity, and even
beauty or ugliness.21 We realize these intentions by submitting the
visual image at the optic chiasma to perceptual scrutiny and judg-
ment. This process is carried out in the brain by the so-called Final Sen-
sor (ultimum sentiens). Syllogistic in nature, the process itself entails

19 For Alhacen on the channeling of visual images into the hollow optic nerve, see Smith,
Alhacen’s Theory, lxi–lxii; for Alhacen on image-fusion, see ibid., lxxiii–lxxvi.
20 On intentionality, see Smith, Alhacen’s Theory, lxxxvii–lxxxix.
21 For Alhacen on the visible intentions, see Smith, Alhacen’s Theory, lxii–lxiii.
188 a. mark smith

discrimination, as well as comparison and correlation of intentional

features, the result being an intentional representation of the object ac-
cording to the full array of its physical properties that can be visually
perceived.22 As has already been mentioned, spatial characteristics,
such as distance and size, are perceptually determined by reference to
the center of sight at the vertex of the cone of visibility. From the per-
ceptual representation formed in this way, we abstract an even higher-
level conceptual representation of the object as a specific or general
type. That, in the end, is how we see something as a horse, or as an an-
imal, rather than as a mere array of colors or as a certain object of a
certain shape and size, lying at a certain distance.
From this account it should be clear that Alhacen and his medieval
followers conceived of vision not as a simple act, but as a complex pro-
cess unfolding in stages, from physical radiation, through brute sen-
sation, to perception and, finally, conception. Each stage is marked by
the formation of a particular intentional representation, which is a vir-
tual likeness in much the same way that a painting is a virtual likeness
of what it represents. This process is illustrated in figure 8. The lumi-
nous color-forms transmitted through air are intentional representa-
tions, or virtual likenesses, of the actual colors on the object’s surface.
Physically radiated to the eye,
these color-forms generate a visual
representation, or image, in the
optic complex. This image is a
virtual likeness of the object at
the level of pure sensation. Trans-
mitted through the spirit pervad-
ing the optic complex, the visual
representation gives rise to a more
abstract perceptual representa-
tion. This intentional represen-
tation is realized in the animal
spirit of the brain. So too is the
conceptual representation arising
from it.
In figure 9 we have an actual
Figure 8 late medieval illustration of the

22 For Alhacen on this process of visual scrutiny and certification, see Smith, Alhacen’s

Theory, lxviii–lxxiii. See also A. I. Sabra, “Sensation and Inference in Alhacen’s Theory of
Visual Perception,” in Studies in Perception, ed. Peter Machamer and Robert Turnbull
(Columbus: Ohio State, 1978), 160–85.
medieval optics 189

perceptual model just outlined.23

The band toward the top of the
head portrays the three ventricles
of the brain in a sort of cutaway
format. The various sense-links to
the first ventricle are specifically
represented by taste (gustus) and
smell (olfactus). The visual and au-
ditory links are drawn in but not
labeled, and the tactile link is miss-
ing altogether. The anterior lobe of
the first ventricle, where all the
sense-links converge, is devoted to
brute sensation, the responsible fac-
ulty being labeled sensus communis.
The sensus communis, or “common
sense,” represents the clearing-house
for all sense-data, including, of
course, visual data. The posterior
lobes of this same ventricle are de- Figure 9
voted to perception, which is car-
ried out by the faculty of perceptual representation (labeled imaginativa,
or “imagination”) and perceptual association (labeled fantasia or
“phantasy”). Connected to the first ventricle by a passageway called
the vermis, the second ventricle contains the conceptual—or cognitive—
faculties, which include the estimativa and cogitativa. This latter fac-
ulty abstracts what I have called conceptual representations from the
perceptual representations transmitted to the second ventricle through
the vermis. And, finally, the posterior ventricle is where the conceptual
representations abstracted by the cogitativa are transmitted for long-
term storage by the memorative faculty, labeled memorativa. Stage by
stage, from beginning to end, the visual process evolves in a continual
succession of replications, each one representing the object at a more
abstract intentional level. The physical, anatomical, and physiological
structure of the visual system is expressly designed to support this suc-
cession. In every respect, therefore, the visual process is naturally dis-
posed to give us a faithful subjective picture of objective reality.24

23 This figure is reproduced from Gregor Reisch, Margarita philosophica (Basel: Schottus

[and M. Furterus], 1508).

24 For a more detailed overview of the Alhacenian model of visual perception within the

broader historical context of Perspectivist optics, see A. Mark Smith, “Getting the Big Picture
in Perspectivist Optics,” Isis 72 (1981): 568–89, and “Picturing the Mind: The Representation
of Thought in the Middle Ages and Renaissance,” Philosophical Topics 20 (1992): 149–70.
190 a. mark smith

All of this is to say that Alhacenian optics is about appearances,

about how things look to the perceiver from a particular point of view—
that is, from the center of sight at the vertex of the cone of visibility.
Ideally, when the light is right, the optic system sound, the object near
enough to be properly viewed, and so on, appearance conforms to real-
ity. But things are not always as they appear. This is especially evident
in the case of reflection and refraction, where objects always appear
dislocated, and often distorted. In plane mirrors things appear to lie
behind the reflecting surface when they actually lie above it. In convex
mirrors they appear behind the mirror and smaller and more bowed
than they really are. In concave mirrors they are subject to a wide variety
of distortions involving apparent place, size, and orientation. In a tub of
water they appear higher and larger than they should. These are all in-
stances of misperception or illusion. As such, they represent anomalies
that demand rectification, and ray-analysis provides the means. It enables
us to reconcile mere appearance, or illusion, with reality on the basis of
key governing principles, such as the rectilinearity of the ray, the equal-
angles law of reflection, the cathetus-rule of image-location, and so forth.
Just how complex the resulting analysis can be is illustrated by a
theorem that occurs toward the end of book 5 of the De aspectibus.
The point of this theorem is to demonstrate that, if A in figure 10 is an

Figure 10
medieval optics 191

object-point placed inside a concave cylindrical mirror, defined by arc

GDEZ, if T is a center of sight within that same cylinder, and if A and
T are placed in an appropriate relationship to one another vis-à-vis
axis HU of the cylinder, then the image of object-point A will reflect to
center of sight T from as many as four points—i.e., M, L, Q, and O—
on the cylindrical segment. Furthermore, each of those points will lie
on a determinate elliptical section on the cylinder’s surface.25 Alhacen’s
ultimate purpose in demonstrating such things is to establish a basis
for determining image-location and, from there, determining precisely
how images appear distorted according to size, shape, location, or orien-
tation, in spherical, cylindrical, and conical convex and concave mirrors.
However “modern” Alhacen’s ray-analysis may appear at a mathemat-
ical level, it is anything but modern at the conceptual level—at the level,
that is, of analytic intent.
From this it should be evident that the ultimate reference-point for
Alhacen’s analysis of reflection is the center of sight—e.g., point T in
the previous figure. And the same holds for his subsequent analysis of
refraction. Furthermore, Alhacen and his Latin followers drew a clear
and crucial distinction between the kind of image formed at the ante-
rior surface of the glacial humor and the kind of image seen in a mir-
ror. The former is invariably referred to by the Latin term forma, or
“form,” which connotes the sort of objective concreteness that an im-
pression made in wax by a signet-ring would have.26 The term ymago,
or “image” properly speaking, is reserved for what is seen in a mirror,
and it is etymologically related to the psychological faculty of “imagi-
nation.”27 In other words, the ymago is a subjective construct, pure
and simple, whereas the forma is both subjective and objective. Or, to
put it anachronistically, the image abstracted by the eye is at least
partly real, whereas the image seen in the mirror is wholly virtual.
Within such an analytic context, then, the science of optics is not about
light-radiation, reflection, or refraction, as we understand them in the
modern, objective sense, but about how we perceive things directly or
by mediation of reflective or refractive surfaces. With these points made,
let us turn, finally, to Kepler.

25 For the text of this theorem, see Risner, Opticae thesaurus, 184–85. The diagram to the

lower left of the figure is reproduced from one of the manuscript-versions of the De
aspectibus (Paris, Bibliothèque Nationale, MS Lat. 7319, f. 233r); the diagram to the upper
right is my reconstruction of the figure to fit the specific parameters of the theorem.
26 For a discussion of the Alhacenian concept of “form,” see A. I. Sabra, “Form in Ibn

al-Haytham’s Theory of Vision,” Zeitschrift für Geschichte der Arabisch-Islamischen

Wissenschaften 5 (1989): 115–40.
27 See Risner, Opticae thesaurus, 125: “Et forma comprehensa in corpore polito

nominatur imago” [And a form perceived in a polished body is called an image].

192 a. mark smith

Kepler presented his account of vision in the fifth chapter of Ad

Vitellionem paralipomena (A Supplement to Witelo), which was pub-
lished in 1604.28 As noted earlier, Witelo was one of Alhacen’s Perspec-
tivist followers, and in this case he clearly stood proxy for him. Like
Witelo, and thus Alhacen, Kepler assumed that light propagates spher-
ically from point-sources. Like Witelo and Alhacen, he assumed that
this propagation occurs within a continuous transparent medium. Like
Witelo and Alhacen, he assumed that light and color intermingle natu-
rally to radiate in tandem from the surfaces of visible objects. And, like
Witelo and Alhacen, he assumed that the transmission of light and
color can be resolved into individual rays within the sphere of propa-
gation. Unlike Witelo and Alhacen, however, Kepler wondered what
would happen if the lens formed by the glacial humor were treated as
nothing more than a refractive body and if all the rays from a given
point-source, not just the orthogonals, went into the formation of the
visual image. Using a glass sphere containing water to represent the lens
and tracing the passage of light through it, he determined that, if the
incoming light is channeled through a small aperture directly in front
of the sphere, all the rays passing through that aperture from a given
point-source A in figure 11 will be focused by the sphere to a particu-
lar spot B on the opposite side. On this basis, he concluded that the
eye, with its pupil and its lens, is designed for no other purpose than to
focus the light emanating from specific points in the visual field to spe-
cific corresponding points on the retinal screen at the back, as is illus-
trated in figure 12. Altogether, these points form an inverted image of
everything within the visual field, an image Kepler pointedly referred to
as a pictura—a painting—to distinguish it from a mere ens rationale,
or “mental entity,” that has only intentional status.29
So far it looks as if, in transposing the visual image from the ante-
rior surface of the lens at the front of the eye to the retinal screen at the
back, Kepler simply followed the imperatives of Alhacenian ray-analysis
to their logical conclusion and, in the process, resolved certain inconsis-
tencies. This is essentially Lindberg’s point. But let us take a closer look.
For one thing, Kepler’s retinal image is upside down. Yet we see things
rightside up, a fact the Alhacenian account takes into full consideration.

28 For a recent English translation of this work, see William H. Donahue, Johannes

Kepler: Optics (Santa Fe: Green Lion, 2000); for chap. 5 see 171–236.
29 As Kepler puts it: “Cum hactenus Imago fuerit Ens rationale, iam figurae rerum verè in

papyro existentes, seu alio pariete, picturae dicantur” [Since hitherto an Image has been a
Being of the reason, now let the figures of objects that really exist on paper or upon another
surface be called pictures]; for the Latin version of this quotation, see Franz Hammer, ed.,
Johannes Keplers gesammelte Werke (München: Verlag C. H. Beck, 1939), 2:174; for the
English, see Donahue, Johannes Kepler: Optics, 210.
medieval optics 193

Figure 11 Figure 12

For another, Kepler’s retinal image is real, not subjective or intentional.

It exists objectively, there for anyone, not just the perceiver behind the
eye, to see. This means that the image itself, not the object it repre-
sents, must be what is actually perceived. If so, then vision can never be
veridical, because it is based on mistaking the physical, objective repre-
sentation for what it represents. For yet another thing, the retinal image
is too large to fit into the optic nerve, much less pass through it to the
brain for perceptual scrutiny. Gone is the peculiar sensitivity of the gla-
cial humor. Gone is the succession of intentional representations passing
through the eye into and through the brain. Gone is the tight linkage
between objective cause and subjective effect so critical to the medieval
account. Gone, as well, is the cone of visibility, and with it the vertex
to serve as a reference-point for spatial perception. All that remains is a
sightless mechanism whose sole function is to focus incoming light in
a specific pattern on the retina. As for what happens to this pattern
afterward, how the visual faculty might actually apprehend it, or how
it might eventually be judged at the subjective level, “This,” Kepler
asserted dismissively, “I leave to the natural philosophers to argue
about. For the arsenal of the optical writers does not extend beyond
[the] opaque wall [formed by the retina].”30
Implicit in this disclaimer is a recognition on Kepler’s part that his
account of retinal imaging is wholly incompatible with the Alhacenian
model of visual perception. And implicit in his evident refusal to see
this as a serious problem is a conception of optics, of its basic aims and
criteria, that is diametrically opposed to that of his medieval predeces-
sors. For them, as we have seen, optics is first and foremost about
sight. Or, to borrow a useful, if somewhat pretentious, buzz-word from

30 Donahue, Johannes Kepler: Optics, 180.

194 a. mark smith

postmodernist criticism, it is oculocentric. The primary goal of medi-

eval optics is therefore to explain how objective reality is subjectively
manifested in the perceptual system through the mediation of light and
color. Every part of that explanation, including the physical account of
light and color, is bent toward the fulfillment of that goal. For Kepler,
on the other hand, optics is first and foremost about light, about how it
acts in its own right apart from any of its sensible or perceptual effects.
In short, Keplerian optics is luminocentric rather than oculocentric. Its
primary goal is therefore to explain light in terms of its objective, phys-
ical manifestations alone. If the physical theory designed to meet this
goal conflicts with the prevailing model of visual perception, then so
much the worse for that model.
Kepler thus posed a dilemma for subsequent thinkers. To adopt his
theory of retinal imaging with all its entailments meant jettisoning the
Alhacenian model of light and sight. In the long run, of course, that is
precisely what happened, and by the end of the seventeenth century the
science of optics had been radically transformed in accordance. Light
was now understood in material rather than qualitative terms, not as a
formal effect propagated through transparent media, but as a contin-
ual succession of tiny bodies shooting, or striving to shoot, through
space and interacting dynamically with other bodies in the way of their
passage. Visual images had been replaced by optical images, some real,
some virtual. The center of sight had been replaced by focal points.
And, perhaps most telling of all, luminosity and color had been trans-
formed from real, objective qualities to mere psychological states bear-
ing no resemblance whatever, intentional or otherwise, to their objective,
physical causes.
At bottom, therefore, Kepler’s account of retinal imaging repre-
sented not a continuation, but a repudiation of the medieval optical
tradition. At issue was the relationship between objective cause, in the
form of light and color, and subjective effect, in the form of perceptual
impressions. Medieval optics was explicitly designed to bind the two as
tightly as possible by means of intentional representations. Keplerian
optics was implicitly designed to sever this bond by interposing the
opaque wall of the retina between the perceiving subject and the per-
ceived object. Out of the resulting disjunction arose not only the modern
science of physical optics but also the mind-body dualism of Descartes
and his philosophical successors.