TREVOR PONECH

The Substance of Cinema

Avant-garde filmmaker Hollis Frampton once said
that film art “consists in devising things to put into
our projector. The simplest thing to devise,
although perhaps not the easiest, is nothing at all,
which fits conveniently into the machine.”

1

Frampton’s arch yet earnest recipe for cin-

ema is interesting for the ingredients it
excludes. He invites the thought that cinema is
defined below the level of pictorial imagery:
perhaps the viewers’ perceptual objects need
not be moving pictures in order to be movies.
Perhaps that rectangle of undifferentiated white
light beamed from the projector as it gives its
mechanical performance is all the audience
needs to see in order to have seen a movie.

I begin by discussing two recent analytic-

philosophical attempts to grasp the properties
owing to which something is cinematic. One is
kindred in spirit to Frampton’s unruly remarks;
the other, antithetical. Both try to be much more
exacting about what cinema is. However, in so
doing each obscures the substantive difference
of cinematic from noncinematic representa-
tional systems. Hence the rest of my essay
develops a third alternative. Accepting Framp-
ton’s invitation, I elaborate on the notion that
cinema precedes imagery. The properties by
virtue of which something is cinematic are
narrowly describable structural features and
physical events upon which an image might but
need not supervene. These comprise what I call
a stroboscopic visual display. Instantiation as
this type of substantial thing is the source of a
perceptual object’s cinematic nature.

Several movies, plus a few “paracinematic”

nonmovies, play a heuristic role in my discus-
sion. The works selected are usually considered
highly abstract. Ironically, far from trading
purely in abstractions, the makers of these

works undertake artistic but nonetheless prac-
tical experiments designed to uncover cinema’s
rock-bottom constituent elements. Using
various instruments, materials, and physical
magnitudes, they intuitively explore cinema’s
ontology. These explorations are not works of
philosophy. They are hardly rigorous truth-
seeking procedures providing evidence and
arguments in support of their authors’ conclu-
sions. Rather, they are vehicles for their
makers’ philosophically pertinent ideas about
cinema’s nature. By putting these ideas into
practice and giving them form, the artists sup-
ply us with concrete as well as conceptual
resources that we can mobilize to debate and
refine our ontological hypotheses.

I. ANALYTIC DEFINITIONS AND
CINEMATIC ABSTRACTIONS

Noël Carroll and Gregory Currie, vanguards of
analytic-aesthetic inquiry into cinema, both
adopt causal approaches to defining cinema, but
they identify the relevant causal chains in sig-
nificantly different ways. Carroll indeed does
not associate cinema with one particular etiol-
ogy, opting to identify it with any underlying
processes and materials capable of generating a
certain result. His sense of what is cinematic is
correspondingly permissive. Currie, on the other
hand, accepts just one causal history as defini-
tive, making membership in the class a lot more
exclusive. How these definitional strategies
handle unorthodox cinematic works—some of
which flaunt the question of their ontology—is
an indicator of their overall success.

Currie reserves the term ‘cinema’ for a group

of things constituting “something like a ‘natural

188

Thinking Through Cinema: Film as Philosophy

kind.’”

2

Something belonging therein is “pro-

duced by photographic means and delivered
onto a surface so as to produce, or be capable of
producing, an apparently moving image.”

3

When one uses a movie camera in the standard
way, the mechanisms of photographic produc-
tion along with the physics of optics and light
naturally result in moving, pictorial images of
visible objects before the camera. This much of
his claim is descriptive. He admits that it also
has a stipulative aspect, since he is legislating
that “cinema” be reserved for only that which is
made by photographic means. Ascertaining pre-
cisely the degree of similarity between such
things and other imaginable as well as actual
sources of items that might be called movies is a
job he would rather defer.

4

When pressed,

though, he is willing to pass judgment on non-
pictorial, purely abstract works. Images in
purely abstract works are not simply unrecog-
nizable or unidentifiable as images of some-
thing; they are also produced by means other
than photographing anything, not even a
shimmer of reflected light, in the outside world.
Hence formalist works like Ballet mécanique
(Dudley Murphy and Fernand Léger, 1924) are
not strictly speaking abstract. Stan Brakhage’s
Night Music (1986), made by hand-painting
abstract shapes directly onto filmstrips, paradig-
matically is.

Currie judges that purely abstract film stands

in the same relation to cinema as setting afire a
canvas prepared for painting stands to painting.

5

Such “pyro-painting” might have some artistic
merit but it is not painting per se. Likewise,
nonpictorial movies created entirely by nonphoto-
graphic techniques are not so much movies as
they are “multifarious things you can do with
the cinematic apparatus.”

6

Cinematic art, on Currie’s thesis, diverges

from painting in that it is necessarily imagistic
and ontologically precludes genuine abstrac-
tion, construed along the lines of nondepiction
combined with lack of reference. This seems
more a metaphysical assumption than a descrip-
tion. His thesis also implies that depictive but
nonphotographic animation techniques—from
manually scratching pictures into the film’s
emulsion to digital animation—are uncinematic
for all their pyrotechnics. I wonder if he has
not expelled too much art from cinema, and
cinema from art. I suspect, as does Carroll, that

Currie takes the etiology of one sort of movie as
defining cinema in general. Thus he misses the
ontologically primitive commonalities of super-
ficially different artifacts. Or so I argue below.

Carroll realizes that choice of definitive

causal conditions might say more about norma-
tive presuppositions regarding the nature of
cinema than about the nature of cinema itself.
His solution to this problem is anti-essentialist,
ingenious, and big hearted. But I doubt it is
successful.

Stated roughly, Carroll identifies as cinematic

any two-dimensional image presented in a
detached display, generated by a template tech-
nology, and produced by cinematic means.

7

He

takes pains, though, to dissociate imagery from
depiction. Indeed, he prefers “moving image”
for its wider application to pictorial works
(made photographically or otherwise) as well as
purely abstract ones (including those made by
such heteroclite methods as editing together
strips of clear and opaque leader).

8

Display detachment concerns the way cinematic

and other images perforce involve spatial dis-
location.

9

Just by looking at a movie image of it,

I cannot orient my body in the Grand Canyon’s
direction. Unlike windows, mirrors, and magni-
fying lenses, photographically-produced movie
pictures do not connect us spatially with their
depicta. Say I am watching a movie at a spatial-
temporal location L1. Following Carroll, when
a display, D, at L1 is a photographically (or
nonphotographically) generated picture of some
extra-cinematic referent, S, D is detached from
S’s own space-time location L such that I cannot
orient myself bodily and in time relative to S
merely by looking D-wards at L1. But what if I
were watching closed-circuit, live images of
happenings in the room next door? As Carroll
notes, background knowledge of the imagery’s
connection to that room, not visual experience
of the image, allows one to orient one’s self
spatially-temporally to those happenings.

10

A template is a physical storage, recording,

exhibition, or transmission format by means of
which a cinematic work is disseminated. When
I watch The Player (Robert Altman, 1992),
I could be viewing imagery produced from
a DVD, videocassette, broadcast television
signal, or 35mm film print. None of these items
is identical to the work, The Player, though. The
work consists of story, characterization, irony,

Ponech

The Substance of Cinema

189

actors’ performances, directorial decisions, and
so forth—properties and events no traces of
which are found within the template’s mole-
cules and electrons. Similarly, were my copy of
The Player destroyed or the TV signal
degraded, the work itself would not thereby be
destroyed—only one particular token of its tem-
plate. Moreover, screening the show requires
I view a mechanical performance executed from
this template. At a stage play, visiting a movie
set, or attending a hockey game, I watch people
as they perform their artistic or professional
actions right before my eyes. When I view a
play, film, or game on television or in a movie
theater my proximal perceptual object, the
detached display, results from something a
machine is arationally “doing” in front of me.
This mechanical performance facilitates my
access to peoples’ artistic and sporting actions,
but it is not itself an artistic or sporting perform-
ance.

Finally, Carroll holds that a movie is by

definition something realized by cinematic
means. This claim would be no more interesting
a tautology than “war is war” were it not anchored
in an arresting idea: some movies contain no
movement.

11

Frampton’s Poetic Justice (1971–

1972) is a film of a shooting script sitting on
a tabletop; Michael Snow’s One Second in
Montréal (1969) consists exclusively of still
photos. Yet these movies share with standard
fare the property of having been produced by
cinematic means, that is, using technologies that
make a certain result, moving imagery presented
in a detached display, a possibility.

12

Frampton

and Snow could have employed their filmmak-
ing means to generate moving images, had they
preferred. Whereas it is categorically imposs-
ible for drawings, paintings, or photographic
stills to contain moving imagery, “movement in
a film image is an artistic choice which is
always technically available.”

13

Viewers of paint-

ings know by looking that such works are struc-
turally guaranteed immobile. La Jetée’s (Chris
Marker, 1964) audience confirms that there is
but one fleeting instance of movement in this
work, otherwise consisting of still photographs,
by watching to see if the filmmakers deviate
from their aesthetic policy of immobility.

Carroll renounces any aspiration of discover-

ing cinema’s essence. He does not suppose the
necessary, general features he cites jointly suffi-

cient to define what a movie is; nor does he
wish to define cinema by identifying it with any
actual, particular materials, processes, produc-
tion techniques, causal chains, or media.

14

Still,

his definition serves to group together a broad
range of things by discerning their mutual
difference from theater, painting, sculpture,
Balinese shadow puppet plays, and so forth. At
the same time, it accommodates the fact that
there are many different kinds of movies and
lots of ways to make them. His definition also
supports my intuition that Night Music is as
cinematic as Gone with the Wind (Victor Fleming,
1939). Carroll’s analysis is subtle enough to
allow that motion is a definitive though not
universal cinematic property. On top of that, it
affords an inkling, to which I will return, of a
crucial structural division between display and
template components.

It also turns out that Carroll is ambivalent

about the necessity of a movie being imagistic. I
sympathize with his doubt but it ultimately
makes trouble for him. Carroll imagines a mod-
ernist filmmaker creating an underexposed film
appearing pitch black under ordinary screening
conditions. As this “invisible film” plays, the
artist, on a voice-over soundtrack, talks of the
screen being dark for half the time a movie
plays.

15

Carroll notes that an actual film of this

type exists, Derek Jarman’s Blue (1993), con-
sisting of seventy-six uninterrupted minutes of
35mm film printed blue. He describes Blue as
having “no images,” hence being an invisible
film.

16

Yet he maintains that it is nonetheless

cinematic if it satisfies two conditions: being
made using cinematic means and being “part of
an intelligible, ongoing filmworld conversa-
tion.”

17

I am unconvinced that movies can be image-

less. That claim is only as defensible as one’s
concept of “image.” This term figures in Carroll’s
definition as an unanalyzed primitive. We are
not told what kind of a thing it is nor what
differentiates imagistic from nonimagistic
items. He does allude to an image being an
“intentional visual artifact.”

18

But so are eye-

glasses. Also, if a movie is imageless, does it
therefore lack a detached display? If the answer
is affirmative, the attrition rate in the list of
necessary features doubles owing to the loss of
an arguably integral structural element. Speak-
ing of attrition, that cinematic works have or are

190

Thinking Through Cinema: Film as Philosophy

generated from templates is perhaps only con-
tingently true. In his lecture, Frampton seemed
to be describing a templateless as well as
imageless work. Indeed, during his presenta-
tion, he cast the beam of an empty projector
onto a screen. It would be unsurprising were
Carroll to treat this case as genuinely cinematic,
as it is realized by cinematic means and could
satisfy the psychosocial conditions of cognoscenti
approval.

At least as much nothing fills Carroll’s

description of cinema as Frampton’s. A movie,
he tells us, is a possibly invisible entity result-
ing from an unspecified underlying cause
known (knowable?) to us only by one of its con-
tingent effects. Now that’s cinema as pure
abstraction. Notice the etiological “produced by
cinematic means” condition appears to be the
sole definitive characteristic exclusive to movies,
although Carroll does not say so explicitly.
Note, too, that, intent on defining cinema with-
out invoking one arbitrarily privileged produc-
tion mode, “cinematic means” is untheorized
beyond saying that it is a technological some-
thing that can cause a certain result. This gambit
resembles defining malaria as a living some-
thing potentially causing fever and death. The
structure, operation, and substance of the some-
thing, of cinema, remain as obscure as if
nobody cared what cinema is.

II. PROJECTABLES, VISUAL DISPLAYS,
AND STROBOSCOPY

One method of exploring cinema’s ontology
is to do hands-on experiments designed to
model one’s intuitions. Instead of leaving our
philosophical understanding to itself, let us take
some suggestions and cautions from a few of
these models.

The works I discuss are alike in at least one

respect. Each treats movie making as an occa-
sion not so much to present images of extra-
cinematic depicta as to evoke the constituent
physical elements of images. It is somewhat
like directing one’s gaze at, not through, the car
windshield. More precisely, these works explic-
itly take as their objects three of cinema’s
underlying constituent physical elements: pro-
jectables, light and its characteristic cinematic
behavior, and the luminescent visual display

space. The best way to explain these items and
their connections is to turn directly to some film
experiments.

Brakhage’s Night Music and his earlier

Mothlight (1963) evince unusual artistry in
devising things to put into projectors. Neither
was made with a camera. Both involved meticu-
lously hand-working unorthodox filmmaking
materials. Brakhage created Night Music by
painting hundreds of film frames he mechani-
cally duplicated then edited together. Mothlight
is the product of putting moth wings, grass, and
flowers between two strips of transparent tape,
running the tape through a film-printer, then
editing the resultant print. Exhibition prints of
these movies are themselves worth examin-
ing.

19

Night Music’s individual paintings are

appreciated anew when studied as still frames;
Mothlight, which contains no frames, is seen to
be as much a collage as a film print. However,
making these projectables was but a step in an
artistic process culminating in a cinematic
object itself of focal concern to Brakhage.

Watching any movie, spectators’ eyes are

trained on a bounded space illuminated by a
pulsating light. If, for instance, they are viewing
a (super 8mm, 16mm, 35mm, 70mm) film
print, a projector behind them flashes light
across the screen in front of them about seventy-
two times per second. During that second,
around twenty-four frames pass through a gate
and shutter mechanism immobilizing each one
just long enough to let light penetrate it three
times. The patterns, dyes, opacity, and various
microphysical surface features within the frame
determine what visual information is borne to
the screen by each light pulse. During the milli-
seconds between flashes, no light at all is
projected onto the screen. Hence the audience
attends to a visual display space, the structure
and visible appearance of which derives from
the interaction between the projectable and the
stroboscopically flashing light.

Night Music’s spectators are presented with a

visual display consisting of a seamless succes-
sion of colorful abstract patterns. These stream
before the eyes at a furious pace. But having
double and quadruple-printed frames, Brakhage
keeps individual patterns within the threshold of
visibility: Each lingers onscreen just long
enough to be noticeable; not so long as to give a
sense that the stream of visual change is ever

Ponech

The Substance of Cinema

191

arrested. In this way, the tempo at which one
pattern gives way to the next suggests the
relentless stroboscopy underlying the display’s
visible changes. Likewise, in Mothlight, moth
wings and plants flutter in time with the flash-
ing light that constitutes the visual display
space. The projectable as collage is inert. It is
stroboscopy and the filmstrip’s mechanical
advancing and stopping that reanimate the dis-
jecta membra.

“Projectable” is closely related to “template,”

but they cannot be equated. Both concepts
designate items that are not themselves percep-
tual objects during movie spectatorship. During
Mothlight, we do not look at the print; watching
a DVD of The Player, we do not look at the
Mylar disk or its digital encoding. So both con-
cepts point to a pivotal structural divide that all
movies exhibit, namely: that between the visual
display, which completely captures our atten-
tion as the movie plays, and the exhibition,
transmission, recording, and storage vehicle by
means of which the display is generated at a
given place and time. However, whereas all
templates are projectables, not every projectable
is a template. And whereas one sort of project-
able is an intrinsic property of all cinematic arti-
facts, no template is more than a contingent
feature. The visual display is made out of light.
Thus light is a cinematic primitive, a magnitude
of electromagnetic radiation without which cin-
ema cannot exist. Normally, this intrinsic pro-
jectable carries information from the template
to the display. Frampton’s eliminativist experi-
ment departs from this norm by stripping
cinema to the ontologically primitive. What he
presented to his audience fits nobody’s pretheo-
retical idea of a movie. By turning on an empty
projector, its shutter speed set low to ensure a
noticeable flicker, he nonetheless generated a
stroboscopic visual display. This object has
everything it needs to be an instance of cinema.

III. THE STROBOSCOPIC VISUAL DISPLAY

Visual displays are material items occurring at
particular spatial-temporal locations. Turning
on television, you bring one into being in your
livingroom from time t until t

n

, when you turn

off the TV. At the theater, a display is instanti-
ated on screen, so many meters from your eyes,

until the projector is shut down. That display is
itself a spatial-temporal location, at which cer-
tain items and events arise during t … t

n

. In

either case, the display is a real, substantial indi-
vidual thing. It is a source of ambient energy
triggering a cascade of responses in the percipi-
ent’s visual system.

I agree that such displays are “detached.” My

reasons go a bit beyond Carroll’s, though. First,
D’s ability to carry information about S at L1
does not depend on D being at L2. The display
would carry the same information no matter
where and when it existed. In other words,
D being at L2 in itself does not mean that S
exists or existed at L1. Contrast this situation
with the following: this shadow now cast on this
wall indicates the current position of the sun or
some other light source. Second, the display is
perceptually detached from the observer. D is
located at some distance external to one’s visual
receptor surfaces. The display’s substance,
structure, and dynamics are distinct from those
of the percipient’s internal representational
states; its visible properties are not produced
by or dependent on the percipient’s visual
responses. One relevant upshot here is that a
cinematic item cannot be a “plug-in,” that is, a
film-loaded prosthetic device attaching directly
to your brain and letting you trip out on movie
imagery with your eyes wide shut.

20

I identify

cinema with the visual display. By definition,
the display is an external thing, albeit one that
can be represented by certain of our sensory and
internal states.

Finally, the visual display is detached from

the template. One’s perceptual experience is of
the former, not the latter. Precursors to cinema
like the Phenakistoscope and Zoetrope are non-
cinematic partly because they are structurally
unified, the viewer’s perceptual object and the
exhibition or storage format being one and the
same thing. The Praxinoscope, which projects a
band of cartoon or photographic pictures onto
an overhead screen when somebody manually
cranks them past a steady beam of light, is articu-
lated in the right way, but is as lacking in strobos-
copy as the Phenakistoscope and Zoetrope. On
the other hand, the peep show Kinetoscope
requires looking into a box, through a magnify-
ing lens, at a backlit filmstrip. Fully motorized,
the film is continuously fed between an electric
lamp and a fast shutter mechanism producing a

192

Thinking Through Cinema: Film as Philosophy

flash of light every time a frame moves under
the lens. The viewer’s proximal perceptual
object is the filmstrip itself. The rudimentary
stroboscopic engineering seems to me to make
the Kinetoscope a nonetheless genuinely pre-
cinematic device.

21

Displays are generated from various materi-

ally and technologically different kinds of tem-
plates. Indeed, displays themselves can be made
of all sorts of stuff. The only limitations belong
to human ingenuity and the laws of physics.
Apparently, engineers can now project 2D
video images onto water vapor diffused in
mid-air.

22

Moreover, nothing about the concept

of visual display necessitates that it be two-
dimensional. Carroll defines the detached
display as perforce 2D, to distinguish it from
other standard modes of presenting artworks.

23

However, 3D cinema is imaginable, if not now
feasible. The risk of collapsing or obscuring the
distinction between visual displays and statues,
standard theatrical presentations, music boxes,
and the like dissolves once we recognize the
display’s distinctive primitive features.

I am sitting in a wooden chair crafted by a

small local furniture-making company. Being
made of wood is a trait that chair possesses
intrinsically. It is a physical condition of the
chair itself. Its location in my study to the left of
a bookcase, and currently holding my body, are
among the chair’s broad features. These proper-
ties concern its relations to other things and
parts of the world. Other of my chair’s broad,
relational properties are the historical properties
of having been made by the atelier’s artisans
and demonstrating their virtuoso woodworking
skills.

I take it that the visual display’s narrow prop-

erties rather than its causal or psychohistorical
relations to other things are the most informa-
tive when we try to define cinema. In fact, ‘visual
display’ refers to cinema’s narrow, primitive
properties. Very generally, then, the visual display
is a delimited area of illumination. It is visible
insofar as it is the source of highly organized
packages of photons projecting from a reflect-
ing or light-emitting surface. Pixels and strobo-
scopic motion are two narrow, intrinsic features
it possesses essentially. Patterns are contingent
narrow properties.

‘Pixel’ usually denotes “picture element.” I

use it in a slightly adjusted but related technical

sense. By ‘pixels’ I intend points of light. This
usage converges with descriptions of movie
images as constructed from separate regions
varying independently in spectral distribution.

24

At a basic level of physical description, visual
displays are composed of pixels. More accurately,
I associate the display with a totality, namely, a
field of points of light plus any unilluminated
areas situated spatially-temporally between those
points.

Exhibition of cinematic imagery requires

light be emitted from or projected against a
surface. Interrupting the flow of this light by
turning off the machine is similar to closing a
faucet, as the flow of imagery is thereby
arrested. Contrast this situation with exhibitions
of still photographic prints or painted canvases.
If the salon is plunged into darkness, or light
reflected from these objects is stopped from
reaching our eyes, the works’ visible, artistic-
ally pertinent features are not extinguished, just
made harder to see. The perceptible patterns
constitutive of the paper print’s or canvas’s
image do not themselves supervene on light
reflected or emitted by the image. When we
watch a movie, that which we look at—the
image on display as distinct from the negative
image on the celluloid or the digital encoding
on the DVD—is integral to a certain structure of
light, not extraneous to it. Hence a pixel is any
actual point of light, emitted by or projected
against a screen, and consequently comprising a
visible element of the cinematic image. Rather
than displays being made of images, images are
made out of displays.

A display is more than a field of pixels. It is

also an event. At a nonmicrophysical level of
description, the display and the pixels therein
undergo a distinctive type of motion. Pixels
flash, periods of illumination alternate with
periods of nonillumination, the pixels’ bright-
ness and intensity fluctuates. Turning on a tele-
vision, computer monitor, or movie projector
activates, I have said, a visual display at a given
location, the screen in front of you, from t until
t

n

. There, an uninterrupted stream of changes

occurs to the display’s illuminated condition,
achieved by the regular flashing of the pixels in
a cycle of, for instance, seventy-two times a
second. This is the display’s stroboscopic
motion. This motion is identified with constitu-
tive, ongoing changes in the display’s surface

Ponech

The Substance of Cinema

193

condition: the rapid onset/offset of pixels and
areas of pixels. ‘Stroboscopic motion’ does not
refer to the subjective impression of objective
displacement sometimes induced by the strobo-
scopic flashing of lights or imagery. It signifies
only the actual changes in the visual display, its
transient phases of illumination. It is a property
of the display itself, independent of any percep-
tual impact it might be calibrated to have. To
date, all movies derive from stroboscopic
motion. Once activated, it is by flashing at a
certain frequency and degree of brightness that
the visual display makes the cinematic image
visible to observers.

Carroll asserts that not all movies contain

motion. He is partly right. Movies necessarily
contain stroboscopic motion, but not all afford
impressions of objective displacement, that is, a
visual experience analogous to seeing a bunny
hop across the lawn while watching motion
picture imagery of a bunny hopping across the
lawn. That experience could never arise—no
matter what “cinematic means” produced the
movie—if not for the display’s internal strobo-
scopic motion.

A clock’s minute hand changes position too

slowly for us to see its movement. Stroboscopic
motion is similar with respect to exceeding the
threshold of perceptibility. It consists of transi-
tions between surface states normally occurring
too fast to be seen. Phenomenologically, it does
not seem to observers that the display is flash-
ing. One’s visual experience of the display is
not about its stroboscopic motion. It does not
represent the onset and offset of pixels any
more than visual experience of the minute hand
represents the event of its tiny but nonetheless
real change of position. But there are usually
lots of other things to look at in the display.

I use ‘pattern’ to refer to any of a visual

display’s singletons.

25

Loosely stated, a pattern

is a visibly differentiated structure standing out
from its background. A bit more precisely, a
singleton is a perceptual target, in that it can be
the object of an observer’s perceptual represen-
tation. My notion of pattern includes that area
of La strada’s (Federico Fellini, 1954) image
we are inclined to call “Gelsomina wearing
clown attire.” Embedded in this pattern are
various other singletons, like Gelsomina’s eyes,
the clown nose drawn onto the end of her own
nose, her hat and striped vest, her hands, and so

forth. An undifferentiated, homogenous area of
illumination, such as the effect created by pro-
jecting a completely blue image onto the screen,
might at a global level of description constitute a
pattern. Nor would I be quick to exclude unstable
moiré “patterns” associated with poor television
reception, and even the “snow” associated with a
badly malfunctioning TV. I have no theoretical
need to take any apparently borderline or ques-
tion-begging cases aboard, though, since I do not
maintain that anything and everything appearing
in a display has to be a pattern.

Unlike my psychohistorical pattern of mis-

calculating my capacity to afford luxury goods,
the patterns I am concerned with are not at all
abstract-typical. They belong to the category of
substances. Spiritual salvation, Santa Claus, the
market value of John Currin paintings, and
nothingness are abstract, conceptual entities.
Grasping these takes reason and imagination. A
substance, on the other hand, is a magnitude, a
piece of stuff, or an event that can be the object
of perceptual experience. It is a material, spatio-
temporally localized part of the world’s existing
hardware. As such, it is possible to be in non-
conceptual contact with it. To step on a bug
requires neither a bug concept nor knowledge
that you are stepping on a bug. Likewise, dogs
and infant, prelinguistic children see beetles
without possessing any ideas or words about
beetles.

In (nostalgia) (1971), Frampton shows us, in

separate shots, one after another still photo-
graphs being gradually incinerated on a hot-
plate. While looking at one photo burn, we hear
a narrator’s comments about the next to appear.
Frampton thus provides a visual trope for an
apparent consequence of the visual efferves-
cence of cinema. If you think of movie imagery
as composed of one after another still picture,
then asynchrony and referential confusion are
inevitable. Any given still image flashed onto
the screen is gone before one can form a
thought or express an utterance about it.

This remark raises the problem of individuat-

ing stroboscopically generated patterns. Within
the visual display, patterns are instantiated by
the onset and offset of groups of pixels. Within
fractions of a second, the pixels betokening the
“Gelsomina wearing clown attire” pattern
vanish from the stroboscopic display space;
milliseconds later this pattern is temporarily

194

Thinking Through Cinema: Film as Philosophy

reconstituted in the display by a fresh set of
pixels. And so on, until the pattern is not instan-
tiated with the onset of the next array of pixels.
The Gelsomina pattern, therefore, is a visible
uniformity across pixels.

At this stage my analysis evidently comes

unstuck from a basic truism, which would inter-
pret such uniformities as nonexistent or at best
abstract-typical entities, rather than substantial
ones. Is not one second of Gelsomina imagery
in the display merely a series of twenty-four
discrete, nonidentical, still “Gelsomina” patterns
disjoined by interstimulus periods, each out of
sight before we can form or express and idea
about it?

No. There is also a structural invariant rang-

ing over those patterns. This particular—call it
G—is no less real and substantial an individual
than are you and I and our chairs. Our furniture
and we are individuals by dint of being physical
items existing for finite, continuous time
periods at actual, contiguous spatial locations.
Presumably, each such individual is a unified
whole, inasmuch as it consists of parts and
phases held together over time and space by
various underlying physical connections,
causes, and processes.

26

G is arguably more this

sort of substantial entity than chimera or sheer
theoretical term.

The display, I have said, is a coherent spatial

location. Its parts, states, phases, and changes
constitute a space—an area made of changing
illumination—as unified and contiguous as that
of the table at which I now sit. Therein things
(pixels, patterns) and events (stroboscopic
motion) occur during t … t

n

. If G exists, it

exists at that finite spatiotemporal location, as
part of the display.

Now say that t…t

n

is greater than or equal

to one second, no fewer than twenty-four
Gelsomina patterns, g

1

, g

2

, g

3

, and so forth,

appearing onscreen about three times each.
These are underlying parts of G, which could
not exist in the display without them. Notice
that g

1

… g

n

bear a key relation: they resemble

one another. That is why they are g

1

… g

n

instead of g, e (“elephant”), f (“flower”), and so
forth. To stand in a resemblance relation, they
need not be identical. They only have to be
relatively similar with respect to their grosser
visible surface and structural features. These
features pertain to an object’s shape, size, color,

and texture. They include its edges and bounda-
ries; the thickness, straightness, angularity,
curvature, and junctures of its lines; distances
between points on its lines; its volumetric prop-
erties; its rotation; and its luminescence. Two
singletons, g

1

and g

2

, are similar to the degree

that they are as alike to one another as g

1

(or g

2

)

is alike to itself with respect to grosser visible
surface and structural features. In reality, some
items (g

1

and g

2

) are more similar than others

(g

1

and e), in the current respect. The fact of

g

1

… g

n

’s degree of similarity is reflected in the

display’s appearance.

To the extent that every positively illumi-

nated phase of the display during t…t

n

actually

contains g

1

… g

n

, the display contains the

uniformity G. G is made out of g

1

…g

n

, along

with the pixels that are the display’s essential
narrow properties. But the display has another
underlying, essential internal feature—strobo-
scopic motion. This physical process is central
to G’s existence as an individual. Although real,
G is not the same kind of entity as my table or
my dog. It is held together over time, at its loca-
tion, not by molecular bonds or by cell division
and metabolism, but by the similarity relations
between g

1

… g

n

and by stroboscopy. Strobo-

scopic motion is that which generates and main-
tains this uniformity in the display space. G exists,
continuously and uninterrupted, until another
stroboscopic event, from which it is absent, occurs
in the display. The phase changes and discontinui-
ties uniquely intrinsic to the visual display are
neither extraneous to G’s existence nor disruptive
of it. They are as internal to G as molecular bonds
and cellular division are to other entities, being the
process keeping it on the screen.

The display plus the template-specific

technologies upstream of it are machinery
engineered to generate individuals like G. This
singleton is often that which is on display. Not
only is it presented onscreen, its description
most closely corresponds to that of the content
of the viewer’s perceptual experience—and the
referent of the viewer’s thoughts and utterances.
As a cinematic individual, G is not identical or
reducible to any basal phase, state, or part of the
display. It can have properties missing from
some or all of items g

1

through g

n

. It can get

bigger or smaller, split in half, or metamorphose
into another pattern. It can even undergo objec-
tive displacement. But that is another matter.

27

Ponech

The Substance of Cinema

195

IV. FROM ESSENTIALLY CINEMATIC ARTIFACTS
TO MOVIES

Thanks to historicism’s fashionability, and to
their own inadequacies, essentialist definitions
of cinema are in disrepute.

28

Currie’s natural-

kind definition does not help revive the genre,
since the broad properties it identifies as defini-
tive are merely adventitious. My essentialism
seeks informative truth-conditions describing
narrow, physical features not varying over time
and across social contexts and being almost the
only ones necessary and sufficient to make
something a movie. It has no normative impli-
cations for how cinema should look, how best
to make movies, the proper uses of the medium,
or the criteria for artistic value. It succinctly
indicates the underlying difference between
movies and other kinds of representational sys-
tems. Its analysis is deep and precise enough to
pick out shared primitive traits identifying
Workers Leaving the Factory (Louis and
Auguste Lumière, 1895), Gone with the Wind,
Night Music, Blue, Finding Nemo (Andrew
Stanton and Lee Unkrich, 2003), and fetal ultra-
sound imagery as the same type of thing.

An essentially cinematic artifact consists of a

stroboscopic visual display. It does not matter
what appears in that display—pictures, tetchy
abstract shapes, an undifferentiated monochro-
matic field, words, numbers, or snowy static
interference. It does not matter what template, if
any, generates the display. Nor is the essentially
cinematic identified with a causal chain by
which the display comes into existence and
acquires its visible appearance.

This definition precludes the existence of

literally invisible films. Carroll’s hypothetical
example, the pitch black underexposed film, is
not a cinematic artifact, let alone a movie, if no
light is projected against the screen. Running a
completely opaque filmstrip through the projec-
tor preempts the visual display as surely as
putting a lead plate in front of the lens. In con-
trast, there is nothing invisible about Blue or
any work in which cinema’s intrinsic projecta-
ble, light, behaves stroboscopically within a dis-
play space.

My essentialism also excludes from the cate-

gory of cinema such actual “paracinematic”
pieces as Anthony McCall’s twenty-four-hour
Long Film for Ambient Light (1975), consisting

of an empty, darkened warehouse illuminated
by a bare lightbulb during the evening.

29

Jonathan Walley labels this work paracinematic
because it evokes a dematerialized definition of
cinema, one not limiting or reducing cinema to
the traditional materials, instruments, and proc-
esses of the film medium.

30

Walley finds the

paracinematic conducive to the notion that
cinema is as much a conceptual phenomenon—
an idea temporarily taking certain concrete
forms—as a tangible one.

31

On these grounds

he stakes his claim that Long Film for Ambient
Light ought to be located within the avant-garde
film canon. Granted, McCall abandons depic-
tion to draw attention to light, space, and time—
three properties necessary to cinema. But these
are not nonphysical, immaterial magnitudes;
nor are they exclusive to cinema; nor jointly
sufficient to identify something as cinematic.
McCall’s artwork is about cinema without
either being cinematic or embodying a cogent
definition of cinema. Alas, neither Walley’s
credulity nor the din of filmworld discourse
could turn Long Film for Ambient Light into an
authentic member of any cinematic canon.

McCall’s Line Describing a Cone (1973) is

another of Walley’s examples of paracinema.

32

This work seems without qualification a cine-
matic artifact. Another empty projector gig, it
differs from Frampton’s primarily in that the
light beam widens during the thirty-minute per-
formance and audiences are encouraged to
inspect and interact with the beam itself rather
than stare at the display space on a wall across
the room. Yet neither Frampton’s nor McCall’s
presentation fits the “movie” rubric. A movie is
not only a stroboscopic visual display; it is one
containing an image generated from a template.

No narrow, internal feature makes a visual

display imagistic. Being imagistic involves rela-
tional, psychohistorical constraints. It is a
broad, contingent property of a display standing
in a functional relationship to plans enacted by
its maker(s) or user(s). Namely, it must be
intended to serve some communicative, expres-
sive, or aesthetic purpose. The maker could aim
to invite spectators to contemplate the display’s
purely formal, structural qualities and to explore
freely its shapes, colors, and rhythmic move-
ments. Makers also frequently use cinematic
artifacts to indicate openly their own thoughts
and feelings—although they need have no more

196

Thinking Through Cinema: Film as Philosophy

expectation of audience comprehension than
you do when expressing intimate thoughts to a
baby or dog. Alternatively, image making can
involve attempts to guide viewers to particular
realizations about, for instance, what happens in
a fictional story or what the maker’s ostensible
beliefs are regarding a given topic. In turn, these
realizations might be meant to foster some
appropriate response, like suspense, laughter, or
belief formation. Thus the display is an image
to the extent that it is a vehicle for expression
and/or eliciting in a target audience some antici-
pated response of a kind typically associated
with art appreciation.

Frampton’s presentation encourages compet-

ing intuitions, his display being visually dissim-
ilar to standard movie imagery while having at
least the minimal expressive function of draw-
ing attention to cinema’s intrinsic projectable.
Even were we to grant it image status and agree
to call it, along with Line Describing a Cone,
a kind of cinematic artwork, we would have
no reason to call such artifacts movies. Their
projectors are empty, after all.

I propose my own stipulative use of ‘movie.’

It is no more arbitrary than Currie’s, since it
helps pick out an existing class of ubiquitous
items that really do differ from bare cinematic
artifacts. These items are essentially cinematic
because they consist of stroboscopic displays.
They are movies because they contain images
generated from templates. Here, a template is
further described as something the structure of
which makes an observable difference to the
display’s visible appearance. That is exactly
what film prints, DVDs, videocassettes, and
broadcast television signals do; projectors, tele-
vision sets, satellite transmitters, and video and
DVD players are correspondingly designed to
be structurally coupled with these formats.
During his lecture, Frampton simulated adding
a template by holding a colored gel in front of
the projector’s lens. This gesture is functionally
similar to producing an image from a template.
But since nothing is loaded into the device engi-
neered to couple the template with the display,
it falls short of turning the presentation into a
movie. Had Frampton projected an emulsion-
less, untinted, completely clear reel of film
leader, that would not have been a movie,
either. For nothing about the template would
have made a visible difference to the display’s

appearance or practical difference to its expres-
sive potential.

The notion of imageless movies is oxymoronic.

The idea of imageless displays generated from
informative templates is not. Consider Ernie
Gehr’s History (1970), consisting of a depth-
less, textured black space in which beads of
light flicker on and off in time with the projec-
tor’s rhythm. Gehr made his film without a
camera, exposing raw stock to dim light then
making successive generations of prints. The
resulting movie creates an impression of the
granularity of the medium upon which photo-
graphic images are imprinted. History is one of
Gehr’s attempts to escape representational
film’s subordination to the photo-recorded
event, affirm the material conditions of its exist-
ence, and foreground its “primordial state” as
“patterns of light and darkness.”

33

However,

one can readily imagine a display identical in
every visible respect to History but made with-
out expressive intentions—the result, say, of
improperly storing 16mm film stock. That arti-
fact would be cinematic. It would be produced
by cinematic means and derive from a template.
It would also be imageless. Although some
images are essentially cinematic, there is not
anything essentially imagistic about cinema.

V. CONCLUSION

Someone might criticize my ontology of cinema
for eliding crucial differences. While true, this
charge signals a defect not so much in my ana-
lysis as in certain assumptions about what a
proper definition of cinema ought best to do. I
ignore the differences in how people perceptu-
ally, cognitively, and affectively experience
films on the theater’s big screen versus pro-
grams watched at home on television versus
fetal ultrasound imagery viewed on the obstetri-
cian’s computer monitor. My analysis flouts
major institutional, economic, and cultural dif-
ferences between various modes of production
and spectatorship. It is also oblivious to how
speakers use the word ‘cinema’ and to what
they variously believe it to mean.

Inquiry into cinema’s myriad psychosocial

dimensions is vital, but the immediate job of
defining cinema need not involve a sweeping
anthropological, historical if not historicized,

Ponech

The Substance of Cinema

197

enterprise. As Currie intuits, it can be more
simply and elegantly done by taking advantage
of the fact that there exists “something like a
natural kind” that we can intelligibly use the
word ‘cinema’ to designate. My description of
this natural kind, of course, associates it with a
class of artifacts. Think of it as a concise articu-
lation of cinema’s engineering specifications,
one I hope is empirically grounded in uniformi-
ties across physical properties and processes of
otherwise dissimilar technological artifacts.
Those who say of the fetal ultrasound image,
“That’s not cinematic!” are reacting to the
countless differences from their favored para-
digm cases of cinema but are likely ignoring the
underlying engineering invariants.

While their creations surely have other signi-

ficant facets, Frampton, Brakhage, Gehr, and
others, by modeling their artform’s primitive
parts and processes, provide a resource for
inquiry into cinema’s engineering specifica-
tions, a resource unparalleled by anything
arising from philosophical discourse itself. In
doing so, they remind philosophers and theo-
rists that “the cinema” names not just a welter
of cultural, human phenomena but also a
unified group of tangible things—no less so
than rocks, trees, and airplanes—in the world,
right before our eyes.

34

TREVOR PONECH

Department of English
McGill University
Montreal, Quebec H3A 2T6
Canada

INTERNET

: [e-mail]

1. Hollis Frampton, “Lecture,” in The Avant-Garde Film:

A Reader of Theory and Criticism, ed. P. Adams Sitney
(New York: Anthology Film Archives, 1987), p. 276.

2. Gregory Currie, “Reply to My Critics,” Philosophical

Studies 89 (1998): 357. See also his Image and Mind: Film,
Philosophy and Cognitive Science (New York: Cambridge
University Press, 1995), p. 1.

3. Currie, Image and Mind, p. 4.
4. Ibid.
5. Currie, “Reply to My Critics,” p. 357.
6. Ibid.
7. Noël Carroll, “Defining the Moving Image,” in Theo-

rizing the Moving Image (New York: Cambridge University
Press, 1996), pp. 49–74.

8. Carroll, “Defining the Moving Image,” p. 66.
9. Carroll, “Defining the Moving Image,” p. 63.

10. Ibid. A larger point motivates Carroll’s discussion of

detached displays. He wants to demonstrate that depictive
photographic images are not transparent by dint of their
causal relations to their referents, as Kendall Walton, “Trans-
parent Pictures: On the Nature of Photographic Realism,”
Critical Inquiry 2 (1984): 246–277, contends. Looking at a
photographic picture of the Grand Canyon is fundamentally
unlike looking at it through a window because one would
need supplementary nonperceptual beliefs in order to know
one’s approximate location relative to the Canyon. To see
something, S, is presumably not to need such cognitive addi-
tives in order to relate yourself spatially to S. Hence one does
not see S by seeing a movie image of it.

11. Carroll, “Defining the Moving Image,” pp. 64–65.

See also Carroll, “The Essence of Cinema?” Philosophical
Studies 89 (1998): 323–330.

12. Carroll, “The Essence of Cinema?” p. 329.
13. Carroll, “Defining the Moving Image,” p. 64.
14. Carroll, “Defining the Moving Image,” pp. 49–54.
15. Carroll, “The Essence of Cinema?” p. 329.
16. Carroll, “The Essence of Cinema?” p. 330, n.4.
17. Carroll, “The Essence of Cinema?” p. 329.
18. Carroll, “Defining the Moving Image,” p. 65.
19. At <fredcamper.com/Film/BrakhageL.html> you can

view frame enlargements from Mothlight as well as from sev-
eral of Brakhage’s hand-painted films.

20. I borrow this example from Currie, Image and Mind,

p. 4.

21. Kristin Thompson and David Bordwell, Film

History: An Introduction (New York: McGraw-Hill, 1994),
pp. 3–12, give an overview of the technical developments
and precursors culminating in cinema’s debut. For a
description of the Kinetoscope’s inner workings, see David
Robinson’s From Peep Show to Palace (Columbia Univer-
sity Press, 1996), p. 34.

22. A pair of websites currently herald such “free-space”

displays as the next big thing: <www.FogScreen.com> and
<www.io2technology.com>.

23. Carroll, “Defining the Moving Image,” pp. 70–71.
24. Julian Hochberg, “Representation of Motion and

Space in Video and Cinematic Displays,” in Handbook of
Perception and Human Performance, Volume 1, Sensory
Processes and Perception, ed. Kenneth R. Boff, Lloyd
Kaufman, and James P. Thomas (New York: John Wiley
and Sons, 1986), p. 224.

25. For a discussion of singletons and their role in visual

perception generally, see Steven Yantis, “Attentional
Capture in Vision,” in Converging Operations in the Study
of Visual Selective Attention, ed. Arthur Kramer, Michael
G. H. Coles, and Gordon Logan (Washington: American
Psychological Association, 1996), pp. 45–76.

26. My approach to substance concepts and individuation

draws on the work of Ruth Garrett Millikan, especially her
On Clear and Confused Ideas: An Essay about Substance
Concepts (Cambridge: Cambridge University Press, 2000). My
notion of an individual as a uniformity persisting through some
course of events derives from Jon Barwise and John Perry’s
analysis of this primitive of their situation semantics theory;
see their Situations and Attitudes (MIT Press, 1986), pp. 7–9.

27. My “Cinematic Motion” (in preparation) reopens the

question of whether this phenomenon is real or illusory, and
whether it is a narrow or response-dependent property of
movies.

198

Thinking Through Cinema: Film as Philosophy

28. Carroll discusses medium essentialism’s failings—

especially its tendency toward prescription and its inability
to countenance technological and artistic innovation—in
“Defining the Moving Image.”

29. For a description of this work, see McCall’s “Two

Statements,” in The Avant-Garde Film, p. 252.

30. Jonathan Walley, “The Material of Film and the Idea

of Cinema: Contrasting Practices in Sixties and Seventies
Avant-Garde Film,” October 103 (2003): 20.

31. Walley, “The Material of Film and the Idea of

Cinema,” p. 23.

32. McCall, “Two Statements,” p. 251.
33. Ernie Gehr, “Program Notes for a Film Screening

at the Museum of Modern Art,” in The Avant-Garde Film,
pp. 247–248.

34. My thanks to Murray Smith and Tom Wartenberg for

helpful comments on, and good questions about, the ideas
expressed in this essay.