How to Make Analogies in a Digital Age

Author(s): Whitney Davis

Source: October , Summer, 2006, Vol. 117 (Summer, 2006), pp. 71-98
Published by: The MIT Press

Stable URL: https://www.jstor.org/stable/40368434

REFERENCES
Linked references are available on JSTOR for this article:
https://www.jstor.org/stable/40368434?seq=1&cid=pdf-
reference#references_tab_contents
You may need to log in to JSTOR to access the linked references.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms

The MIT Press is collaborating with JSTOR to digitize, preserve and extend access to October

This content downloaded from

132.174.250.101 on Thu, 12 Oct 2023 19:52:19 +00:00
All use subject to https://about.jstor.org/terms
 How to Make Analogies
in a Digital Age*

WHITNEY DAVIS

I will start with a truism even though, like all truisms, it reduces and c
flates distinctions and nuances to which I will need to revert: painting - in
sense of "painting as an art" - has been a classic analog mode of representa
whether or not every painter in the history of art has used it this way. It cannot
a surprise, then, that it might be specifically in painting - in its specifica
painterly and pictorial achievements and reversions - that we might best be
to gauge the depth and direction of digital technologies of image productio
well as notions of phenomena, appearance, and representation associated
them in their theoretical foundation and recent history.

The best logical analysis of the analog property of painting as an art - for
him, denning depiction as a symbol system - remains Nelson Goodman's Languages
of Art, published in 1968. It will be useful to recall Goodman's terms, not least
because "analog" and "digital" do not figure prominently among them despite
their increasingly widespread circulation at the time his book was written.1
Indeed, Goodman seems to have intended Languages of Art partly as a founda-
tional explication (and to some extent a critique) of contemporary vocabularies
of information, computation, and representation - including a growing tendency
to suppose that their domains of "analog" and "digital" exhaustively parse the

* This essay originated in my response to a wide-ranging exhibition at the Kunstmuseum

Wolfsburg in 2003, from which I have taken my examples of contemporary painting; see Gijs van Tuyl,
ed., Painting Pictures: Painting and Media in the Digital Age, exh. cat. (Bielefeld: Kerber Verlag, 2003). A
preliminary version was presented at a conference on "Art Since 1989" organized by Alex Alberro at
the University of Florida, Gainesville, in March 2006. 1 thank Alberro and Hal Foster, Anthony Grudin,
Branden Joseph, James Meyer, Terry Smith, Julian Stallabrass, and Malcolm Turvey for incisive com-
ments. And thanks to my colleagues in the Center for New Media at the University of California at
Berkeley for many invigorating discussions.
1. Nelson Goodman, Languages of Art: An Approach to a Theory of Symbols (Indianapolis: Hackett,
1968).

OCTOBER 117, Summer 2006, pp. 71-98. © 2006 October Magazine, Ltd. and Massachusetts Institute of Technology.

This content downloaded from

132.174.250.101 on Thu, 12 Oct 2023 19:52:19 +00:00
All use subject to https://about.jstor.org/terms
 72 OCTOBER

entire field o
ingly basic p
masks a host o
symbol system
Thus it does n
"inscription")
which I will r
and precise ch
equally impor
struggled to o
wrote, "the di
illustrate than to define."3

Especially in the historical context of the rise of electronic automation,

advocates had promoted the supposed precision of digital computing relative to
the supposed approximateness of analog instruments, including human vision as
understood at the time. Languages of Art addressed this prejudice: it proposed rig-
orous terms within which the "sensitivity and flexibility" of analog symbol systems
(including symbol systems identified with vision, natural language, and the arts)
could be compared systematically - and often favorably - with the "definiteness
and repeatability" attainable with digital computing machines. As Goodman
urged, the key issue - it is phenomenal, technical, logical, social, cultural, and eth-
ical - is to settle "the maximum required fineness of discrimination" to be relayed
in a representation or computed by a machine.4 Goodman affiliated his ecumeni-
cal approach to this question with the perspective advocated by John von
Neumann's "The General and Logical Theory of Automata" in 1948.5 Although
2. Languages of Art was also, of course, an extension of the distinctive constructivist and nominalist epis-
temology developed in Goodman's doctoral dissertation of 1940, published a decade later as The Structure of
Appearance (Cambridge, Mass.: Harvard University Press, 1951); it was continued in his Ways of Worldmaking
(Indianapolis: Hackett, 1978), where he summed up his view as "radical relativism under rigorous restraints
that eventuates in something akin to irrealism" (ibid., p. x). Indebted to Rudolf Carnap's Der bgische Aufbau
der Welt of 1928, revised in 1961 ( The Logical Structure of the World and Pseudoprvblems in Philosophy, trans. R. A.
George [Berkeley: University of California Press, 1967]), Goodman shared Carnap's break with the idealist
(Cassirerean) and the Husserlian and Heideggerean phenomenological traditions (see Michael Friedman,
A Parting of the Ways: Carnap, Cassirer, and Heidegger [La Salle, 111.: Open Court, 2000]) - though in his own
way he pursued Cassirer's notion that symbolic forms constitute multiple world versions (see Ways of
Worldmaking, chap. 1-2). Many expositions and extensions of Languages of Art (including severe and stinging
criticisms) have been published. I single out Richard Wollheim, "On Nelson Goodman's Languages of Art," in
On Art and the Mind (Cambridge, Mass.: Harvard University Press, 1974), pp. 290-314, and Catherine Z.
Elgin, With Reference to Reference (Indianapolis: Hackett, 1983). Unfortunately, though Wollheim deals exten-
sively with both pictorial representation and painting, he does not really address analogicity - while Elgin,
in addressing analogicity in great detail, says little about painted pictures.
3. Goodman, Languages of Art, p. 160. For lucid expositions and critiques of the distinction
between analog and digital, see David Lewis, "Analog and Digital," Nous 5 (1971), pp. 321-27; John
Haugeland, "Analog and Analog," Philosophical Topics 12 (1981), pp. 213-26; and William Demopoulos,
"On Some Fundamental Distinctions of Computationalism," Synthese 70 (1987), pp. 76-96.
4. Goodman, Languages of Art, p. 161.
5. John von Neumann, "The General and Logical Theory of Automata," in Cerebral Mechanisms in
Behavior, ed. Lloyd A. Jeffress (New York: John Wiley & Sons, 1951), pp. 1-41.

This content downloaded from

132.174.250.101 on Thu, 12 Oct 2023 19:52:19 +00:00
All use subject to https://about.jstor.org/terms
How to Make Analogies in a Digital Age 73

Goodman did not cite this comment, von Neumann tacitly a

itality might be embedded in analogicity (and perhaps vice
recursion-regress?): "The decisive property of a switching or
always found in one or the other of its two extreme discr
only very little time transiently in the intermediate states t
ing continuum."6 It might be, I think, that the "fineness"
symbol systems requires recognition of these "intermediate
tion will be impossible if the mechanism only represents th
But paintings can always do both.
One of Goodman's most radical conclusions about ana
simple: "many [symbol] systems are of neither type," includ
But this insight has been almost entirely forgotten in current
I will pursue my own variant of it here. In an age in which
mediation of information has become ubiquitous, many sym
said to belong to "neither type" because they conjoin both
take to be a new development - cannot be discerned to bel
be sure, and to modify Goodman's statement slightly, in th
of such mongrel type seldom survive [d] long in computer
measure to "the length of the message" - a parameter that typi
power and speed of the computing machine.8 By the same
powers of digital computers should enable an immense proli
cation of "mongrel" analogicities and digitalities. Has this
If not, why?

II

To illustrate what he described as the highly "autographic" character of ana-

log drawings, paintings, and related pictorializations grouped under the general
rubric of "the sketch" - especially works that exploited the immanent aesthetic or
artistic possibilities of these media - Goodman liked the example of Katsushika
Hokusai's prints of Mount Fuji, produced in the 1830s. Several of these, including
Mount Fuji in Clear Weather, might be compared with Paul Klee's Ad Parnassum
(1932) in order to remind us that the logical and morphological properties that
concerned Goodman do not seem to be the exclusive province of any single cul-
tural tradition of depiction (such as Japanese ukiyo-e) and its investigation and
inflection by artists (such as Hokusai) working within its definitions of intelligibil-
ity. In depicting the profile of Mount Fuji, each and every discernible or

6. Ibid., p. 20.
7. Goodman, Languages of Art, p. 162.
8. Ibid., pp. 162, 164. In the late 1960s, when Languages of Art was written, "length of the message"
was a dominant constraint on the deployment of digital computers in the replication of virtually all
symbol systems. Now, it has faded away in the replication of most of them - including some that were at
one time unprocessable.

This content downloaded from

132.174.250.101 on Thu, 12 Oct 2023 19:52:19 +00:00
All use subject to https://about.jstor.org/terms
Top: Katsushika Hokusai. Mount Fuji in Clear Weather, ca. 1834.
Bottom: Paul Klee. Ad Parnassum. 1 932. © Artists Rights Society (AKS),
New York/VG Bild-Kunst, Bonn.

This content downloaded from

132.174.250.101 on Thu, 12 Oct 2023 19:52:19 +00:00
All use subject to https://about.jstor.org/terms
How to Make Analogies in a Digital Age 75

discriminable variation or perturbation of Hokusai's line - o

ous line - seems to us to have value. This wealth of seem
particularity constitutes both the very possibility and - in it
mate identity of the pictorialization of Mount Fuji by Hok
autographic projection of the famous site, despite the ap
blance of the image to such anonymous, diagrammatic
electrocardiogram (a notional point of comparison cited by
ple every modulation of the line in Hokusai's print might
formation in the topography of Mount Fuji itself, perhaps
formations all at once. (Of course, they might not do so; th
skipping of Hokusai's brush.) In the electrocardiogram, by co
nates and abscissas of the points through which the line pas
segments of the line - and all of its modulations - cannot r
arrhythmias in the patient's heart; they certainly are not
arrhythmia and the temperature in the clinic; and - at all t
nates and abscissas - they must denote the patient's heartrat
Hokusai's mode of mark making functions as what Good
scheme. It is to be distinguished from other symbol schem
script and a musical score, or what Goodman calls notation
tions - as in the example of the cardiogram - can look ver
to) the picture. In their non-no tationality (what would typi
ity), pictorial schemes afford directions and domains
particular of reference to putative objective correlates (e.g.,
critics and historians have often wished to describe as rich a
Goodman's underlying theory of symbol systems - the burd
Art - characterizes this richness and ambiguity in terms of

9. Katsushika Hokusai, Fugaku Hyakkei [One Hundred Views of Mount Fu

1849), vols. 1, 4. For the example, which I have fleshed out somewhat, see Go
p. 229; Goodman does not specify the Hokusai that he describes, and I ha
the One Hundred Views because they were printed in monochrome. In Ways
describes how a "single charged line" (p. 14) functions in a drawing by art
whom he collaborated (with choreographer Martha Gray and composer J
projects in the 1970s (see Nelson Goodman, Of Mind and Other Matters [
University Press, 1984], pp. 69-71). For Goodman's characterization of autogr
painting - as distinguished from the so-called allographic art instantiated in
of a printed musical score - see Languages of Art, pp. 113-22. In the chap
Goodman addressed long-standing problems of copying and forgery, the cor
relayed in notations for it, and the like. All of these problems of replica
grounded in a comprehensive theory of notation and its typology of mod
marks or inscriptions on the one hand and characters or correlates on the ot
Languages of Art (especially pp. 127-224). For my purposes, Goodman's
taken analytically to encompass modes of representation such as the con
painting as an art - even though these modes are non-notational in his t
ments on autographic and allographic modes of representation in contemp
drawing and painting - with obvious relevance to the algorithmic organi
digital imaging used in it or to replicate it - see Kirk Pillow, "Did Goo
LeWitt?," Journal of Aesthetics and Art Criticism 61 (2003), pp. 365-80.

This content downloaded from

132.174.250.101 on Thu, 12 Oct 2023 19:52:19 +00:00
All use subject to https://about.jstor.org/terms
 76 OCTOBER

density and r
ial scheme is
differentiatio
morphologica
order of notat
no mark or in
"a" belongs to
must be theo
belong to one
character A it
(articulation).
atively dense
not hold. In a
one and the s
flank of Mou
depicts just on
flank of Moun
and if it is not
Strictly spea
infinitely ma
Looking ahead
representation
if we were to
salient modul
able pixels p
cannot discov
inexchangeabl
in principle t
and incarnati
This possibili
zons for the
pictorial sche
stitutive in
density - wo
might try to
ture.12 As Go
the line, its c

10. Goodman, La
11. Ibid., p. 136.
12. Ibid., pp. 229-30. I have added this emphasis on the possible morphological indiscernibility of
the "analog" and "digital" representations - the dense/replete and the disjoint/articulate symbol
schemes respectively - for reasons that will become apparent. It is not, however, a point pursued by
Goodman, who tended to stress the palpable difference - however fine or subtle - between the phe-
nomenal morphology of analog and digital schemes.

This content downloaded from

132.174.250.101 on Thu, 12 Oct 2023 19:52:19 +00:00
All use subject to https://about.jstor.org/terms
How to Make Analogies in a Digital Age 77

the paper - none of these is ruled out, none can be igno

among the pictorial properties of a sketch can be dismis
the sake of simplicity, we can define density as conjoint
jointness and (2) nonarticulation at the syntactic level, r
Goodman puts it, to conjoint conditions of (3) nondisjoin
lation at the semantic level (i.e., repleteness).15
To be sure, this formulation has drawbacks when we extend Goodman's
account of pure notations all the way to non-notational symbol schemes like
depiction. A mere parallel between syntactic and semantic levels of representa-
tion - relatively dense and replete, respectively - is not, we might think, what
depictions essentially involve: indeed, an essential difficulty in discriminating the
syntax from the semantics seems to characterize depiction as it functions in drawing
or painting - especially in drawing or painting treated as an art.16 I cannot pursue
this problem in detail here. But Goodman actually states five requirements for nota-
tion: in addition to parallel disjointness ([1] + [3]) and parallel articulation ([2] +
[4]), a condition of (5) uniform unambiguousness must apply both to marks or
inscriptions and to the characters or objective correlates with which they comply.17
For most purposes it suffices to say that the density ([1]+ [2]) and the repleteness
([3] + [4]) of depiction constitutes its (5) ambiguity: in depiction, identifying both
the pictorial mark itself and its objective correlate (what it putatively "complies
with" in a relation of pictorial reference) becomes difficult because the pictorial
scheme proffers a relatively dense and replete syntactic and semantic field. In a sim-
plified statement of Goodman's model as it applies to depiction, richness - the
conjoint conditions of density and repleteness - creates ambiguity, and ambiguity -
our uncertainty about the mutual distinction and boundedness of inscription and
correlate - creates richness.
Still, there can be rich symbol schemes (relatively dense and replete) that do
not produce the peculiar ambiguities of depiction - most important, the constitu-
tive ambiguity of depiction at the levels of both mark and character (compliance
class or objective correlate) relative to one another. We might call this the ambiguity

13. Ibid., p. 229.

14. Ibid., p. 192.
15. For "parallel," see ibid., p. 149. A full statement of the syntactic requirements for notation, and
by implication a definition of the relatively dense syntax of depiction, can be found in Languages of Art,
pp. 130-48; for the semantic requirements, and by implication a definition of the relatively replete
semantics of depiction, see Languages of Art, pp. 148-54.
16. In his investigation of depiction, Goodman did not concern himself systematically with drawing
or painting (kinds of "sketch") as an art, though he did provisionally identify characteristic "symptoms
of the aesthetic." Not surprisingly, these include density and repleteness, as well as another property,
exemplification, not specifically addressed in the theory of notation {Languages of Art, pp. 252-55).
Still, Goodman's theory of notation provides certain nominalist-constructivist grounds for widely
accepted phenomenological definitions of painting as an art - for example, in terms of representational
richness and ambiguity or (as Wollheim has argued) in terms of material identity between the type and
the token (of the object) of art (see Art and Its Objects [Cambridge: Cambridge University Press, 1980],
and Painting as an Art [Princeton, N.J.: Princeton University Press, 1987]).
17. Goodman, Languages of Art, pp. 148-49.

This content downloaded from

132.174.250.101 on Thu, 12 Oct 2023 19:52:19 +00:00
All use subject to https://about.jstor.org/terms
 78 OCTOBER

of the syntac
Richard Woll
ment that we
to the project
painting.19
In this regar
symbol system
expression to
torial art wor
not possible t
Andre Bazin,
photograph o
knew them -
tationality, i
depiction.20
In a telling experiment, for example, Fry made a copy of a drawing by Paul
Klee: he used a ruler to trace its each and every line. But the copy, a kind of scan,
cannot fully capture the slight wobble of Klee's line - a wobble that helps depict the
little man's tense liveliness. Fry offered this juxtaposition in part as a visual proof of
his kind of formalism - its unformalizable attentiveness to the configurative inter-
minability of pictorial art works and other complex manmade forms.21

18. Other ambiguous symbol schemes do not have this peculiarity. In the written text of a poem
(because it is a notation), the ambiguity of marks and their compliants (letters that spell words) must be
modest - even if the ambiguity of the poetic image (the imaginative picture it creates) can be extreme. A
painted pictorial representation (/that poetic image can transfer - it must transfer - the ambiguity into
both the definition of the mark (the painting) and its discriminability from the correlate (the figured poet-
ic image). In this respect, the ambiguity of depictions and the ambiguity of descriptions must be distin-
guished. Because they do not notate, pictures do not describe. The implications for computationalism -
and the ubiquitous "computer metaphor" for the human mind - are obvious; as Ned Block has put it neat-
ly, "the computer metaphor goes naturally with descriptional representations, but it is not at all clear how
it can work when the representations are nondescriptional" ("Mental Pictures and Cognitive Science,"
Philosophical Review 92 [1983], p. 535; by "descriptional," Block means digital - i.e., Goodman's disjoint
and articulate). For my own crude efforts to describe the awa/og-computational (density-preserving)
image-construction program of depiction, see Whitney Davis, "The Origins of Image Making," Current
Anthropology 27 (1986), pp. 193-215, and "Replication and Depiction in Paleolithic Art," Representations 19
(1987), pp. 111-47 (both reprinted in Replications: Archaeology, Art History, Psychoanalysis [University Park:
Pennsylvania State University Press, 1996], pp. 46-94).
19. See Richard Wollheim, Two Kinds of Formalism (Barcelona: Fondation Tapies, 1994).
20. See especially Last Lectures by Roger Fry, ed. Kenneth Clark (Cambridge: Cambridge University Press,
1939); Andre Bazin, What Is Cinema?, trans. Hugh Gray, 2 vols. (Berkeley: University of California Press,
1967-71); and Wollheim, Painting as an Art.
21. Last Lectures by Roger Fry, pp. 22-23, 57-58. It is perhaps no surprise that the first illustration in
Goodman's Languages of Art (p. 2) reproduces a perspectival drawing of a floor and/or a facade from
Klee's 1925 Pedagogical Sketchbook. Following Klee, Goodman mobilized the drawing to help launch his
argument that marks can only be interpreted in terms of the symbol system in which they function - not in
terms of what they resemble or in terms of a world whose visual configuration they supposedly reproduce.
But we could readily cite it as a perspicuous instance of pictorial richness and ambiguity in the terms con-
sidered here: the drawing is both floor and facade - and dense and replete in either case/both cases.

This content downloaded from

132.174.250.101 on Thu, 12 Oct 2023 19:52:19 +00:00
All use subject to https://about.jstor.org/terms
How to Make Analogies in a Digital Age

Roger Fry. Tracing of a drawing by Paul Klee,

from Last Lectures by Roger Fry. 1939.

On the face of it, Fry's experiment would pose a challenge - albeit a chal-
lenge somewhat avant la lettre - to the analytical diagramming, and especially to
the computational digitizing, of pictorial art works that must characterize another
kind of formalism. In these operations, a mechanism tries to "see," compute,
transform, and replicate a scene (it is often called "the target scene" by computer-
vision specialists) using technology and algorithms like Fry's copy-scan in which a
straight line is placed wherever one of Klee's original lines (straight or not) is to
be found. In digital imaging - and taking Fry's analytic diagramming to one tech-
nical and logical endpoint - the replicative mechanism breaks the scene down
into discrete units (in Goodman's terms, disjoint and articulate marks putatively
compliant with the drawing, disjoint and articulate or not) in order to "process"
(or, in Goodman's terms, to notate) both its inscriptional order and its objective
correlate, that is, how the scene depicts what it does. Throughout, and because of
its conversion of dense and replete pictoriality relayed in the original representa-
tion, the scan must overlook variations or modulations that have not been
programmed into the digital replications - dense, replete variations and modula-
tions that cannot be programmed when the original image (the mark and its
correlate together in ambiguous parallel, or [l] + [2] + [3] + [4] + [5]) cannot be ren-
dered in discrete, differentiated units, especially in the highly abstracted form of
identical, interchangeable units.

This content downloaded from

132.174.250.101 on Thu, 12 Oct 2023 19:52:19 +00:00
All use subject to https://about.jstor.org/terms
80 OCTOBER

To be sure,
ing can certa
a digital tech
mark - or th
when the pa
notationally
replete picto
marking tha
possibly did
mark making
had a notable
Stiegler's Tech
For my pur
crete inscrip
the disjoint
Paleolithic p
torial, which
to disjoint, a
tion, or dat
marking, Gr
detain us her
tizing constr
image-scan w

22. A vigorous
Walter Seitter, "
erally his Physik
trades on two s
crete (if not dis
makes the case
sentational syst
pictorialization
Goodman has b
ple, Antony Rai
[1972], pp. 1-21
computer-assiste
23. Andre Leroi
Press, 1993); Be
Beardsworth an
points out, disc
required in nota
many symbol sy
we have seen, th
the mark "A" is
not be discrete e
acters A, B, and
terms of the p
(Dordrecht: D. R
here I cite the d

This content downloaded from

132.174.250.101 on Thu, 12 Oct 2023 19:52:19 +00:00
All use subject to https://about.jstor.org/terms
How to Make Analogies in a Digital Age 81

rendered or not, disjoint and differentiated or not - in

When digital replication tries to capture the disjoint and
notations and the dense and replete properties of non-n
like depiction, it must do so indifferently in a discrete g
differentiation and discontinuity. This digitization need
Fry's diagram of Klee's drawing. Indeed, Fry's diagram -
theoretical exercises that it was meant to parody and r
digitally configured image scanning. But in this particu
and part-incorrect at its threshold of resolution - it is i
richness of the image it relays.

Ill

I want to emphasize one primary property of analog representation, what

has been called "continuous correlation."24 In a traditional clockface, the sweep of
the hour and minute hands uninterruptedly correlates with the passage of time in
hours and minutes - but not only in hours and minutes. In most clock faces, we
can also see the sweep between two hour numerals as five seconds; some clocks
count this out digitally as five ticks. In principle the sweep also shows milliseconds
between each second - and nanoseconds between each millisecond. The continu-
ous correlation goes all the way down, whether or not we can see it. To be sure,
the clock is not an example of dense, replete pictorial representation in
Goodman's preferred sense. It is already rather "digital" - disjoint and articu-
late - because it parcels time in metrically equivalent units marked by discrete,
discontinuous notches or ticks. (Interestingly enough, the dense repleteness of
the "sweep" of time tracked by the hands of the clock is like the dense repleteness
of the "smoothness" of Mount Fuji depicted in the continuous correlation of
Hokusai's line; there is a palpable analogy between these analog images of time
and place.) A fully digital clock - as technology and as representation - is differ-
ent from its analog alter ego, an unmarked clockface, and from the analog/ digital
hybrid (marking or ticking time as well as "sweeping" it). Discrete, discontinuous
units of equivalence are everything: in representing the passage of time, one num-
ber simply replaces another, or, more exactly, one number-combination (e.g.,
"12:00") replaces another number-combination ("11:59"). As Walter Seitter has
written, "the technology of digital representation does not truly come to the fore
until the clock . . . show[s] only figures that can change rapidly while remaining in
the same position so that the quantity of the measurement parameter can be read

24. This useful term is Kent Bach's ("Part of What a Picture Is," British Journal of Aesthetics 10 [1970],
pp. 119-37). Bach explicitly developed his account as a refinement and (re)specification of
Goodman's model of density and repleteness, but it is not necessary to engage the intricacies here. For
the overall approach, see also Michael Polanyi, "What Is a Painting?," British Journal of Aesthetics 10
(1970), pp. 225-36. For technical perspective, see Clare D. McGillem and George R. Cooper,
Continuous and Discrete Signal and System Analysis (New York: Holt, Rinehart, and Winston, 1974).

This content downloaded from

132.174.250.101 on Thu, 12 Oct 2023 19:52:19 +00:00
All use subject to https://about.jstor.org/terms
 82 OCTOBER

immediately.
granulation b
let alone any
might coin a p
tinuous corre
correlation c
lively human
units of their
But then ag
reason, I think
ence between
importance
diversity of r
systems. If a
they might h
historical and
many new m
image - typic
digital mode
recursions of the two modes.

The de-differentiation of analog and digital modes in contemporary new

media partly derives from exponential increases in the power and speed with
which electronic computation can handle the "combinatorial explosion" - an
increase of orders of magnitude in the length of the message - confronted by so-
called computer vision when it scans a photograph or other target scene in order to
generate a digitized copy-picture. Equally important, it derives from ever-more-canny
programs based on early image-recognition protocols written by L. G. Roberts,
Adolfo Guzman, M. B. Clowes, Gilbert Falk, and others in the mid-1960s and early
1970s. These programs and their vastly enhanced progeny can algorithmically rec-
ognize ever more content in a digitally defined and digitally analyzed image such
as a prototype drawing, photograph, or film - or, of course, video and television
input, whether delayed or in real time. The early programs only recognized simple
25. See Seitter, "Painting Has Always Been a Digital Affair," p. 30. For Goodman's discussion of
clocks, see Languages of Art, pp. 157-59. As his example of the contrast between analog and digital
computers, Goodman cites an ordinary pressure gauge on the one hand (pure analog) and a coin-
counter displaying numerals on the other hand (pure digital); as he points out, "an ordinary watch,
read in the most usual way, combines analog and digital computers" in its sweep and notches or ticks
(Languages of Art, p. 160, my emphasis). For a technical perspective, see D. E. Hyndman, Analog and
Hybrid Computing (Oxford: Pergamon, 1970). "Analog" is often used synonymously with "continuous"
by computer scientists - and "digital" with "discrete." But the term "discontinuous" will be more helpful
to me than the term "discrete" because it suggests disjointness and articulateness in Goodman's sense -
and maintains the parallel with analog density and repleteness. Moreover, as Malcolm Turvey has pointed
out to me, in dealing with works of art - qua physical objects - the term "discrete" introduces misleading
connotations; all art works - whether singular or multiple, analog or digital, tokens identical with their
type or types that have only one token - must involve discreteness in the sense specified by Goodman
(i.e., they are not overlapped as "part of another).

This content downloaded from

132.174.250.101 on Thu, 12 Oct 2023 19:52:19 +00:00
All use subject to https://about.jstor.org/terms
How to Make Analogies in a Digital Age 83

Left: L. G. Roberts s line finder program. 1 965.

Right: Adolf o Guzman, see. 1969.

uniformly lighted polygons - as in Roberts's line finde

obscene, and Falk's interpret.26 Now three-dimensional
(and can computationally generate and manipulate) curv
ing, colored, shadowed, reflecting, and moving objects o
of brightness and gloom and chiaroscuro - though "pol
Manovich notes, remains the program of choice.27 T
and replicate an increasingly great proportion of what
purview of powerful analog modes of imaging. In fact,
replicate all of it.

26. For Roberts's line finder program, see "Machine Processing of

James T. Tippett, et al., eds., Optical and Electro-Optical Information Process
1965), pp. 159-98 (the image above left adapts Roberts's figure
"Decomposition of a Visual Field into Three-Dimensional Bodies," in A
Interpretation and Classification of Images (New York: Academic Press,
right adapts Guzman's figure p. 167); for Clowes's obscene, see "On Seeing
(1971), pp. 79-116; and for Falk's interpret, see Gilbert Falk, "Interpreta
Three-Dimensional Scene," Artificial Intelligence 3 (1972), pp. 101-44. For
Winston, ed., The Psychology of Computer Vision (New York: McGraw-Hil
Digital Image Processing (Englewood Cliffs, N.J.: Prentice-Hall, 1979).
Margaret A. Boden's detailed exposition and lucid critique of the program
Natural Man (New York: Basic Books, 1977), pp. 179-244, in particular he
rial explosion" (pp. 221-29) identified by Hubert Dreyfus and others
(see Drefyus, What Computers Can't Do: A Critique of Artificial Reason [New
27. I quote from Manovich's important chapter "Navigable Space,"
pp. 254-55. Manovich maintains that "navigable space" based on po
world created with this technique is a vacuum containing separate o
aries" - is fundamentally different from the "systematic space" proj
drawn or painted) linear-perspective pictorializations. It is true that
programs, as opposed to analog-perspective pictorial representatio
(e.g., the "distance" and "scale" of objects relative to one another and
though some of them do see or read "background," and they mus
("behindness"). Nonetheless, and as Manovich acknowledges, "the Cart

This content downloaded from

132.174.250.101 on Thu, 12 Oct 2023 19:52:19 +00:00
All use subject to https://about.jstor.org/terms
 84 OCTOBER

As I want t
increasingly
resentation
duplicate wi
tions that can be rendered in any analog mode. The analog and digital
replications are often indiscernible from each other and frequently they are func-
tionally interchangeable. That is to say, analog and digital modes are now
themselves wholly continuous. They are continuously correlated with each other
despite internal discontinuity on the one (digital) side and continuity with the
external correlates on the other (analog) side.28

IV

At this point we should turn to the artistic "co-mingling" of analog and digital
modes, to use the term applied by Ludwig Seyfarth to certain new media art works.29
It is possible, I suppose, that art works are essentially indifferent - constitutionally
impervious or constitutionally resistant - to any categorical distinction of their ana-
log and digital faces. I will not try to characterize the ontology of art in these
suggestive terms. But this interdigitation - creating a continuity of the discontinu-
ous that is parallel to a discontinuity of the continuous - appears to be one
constitutive concern in the construction, installation, and display of many contem-
porary pictorial arts, including paintings or highly painterly works. And if the real
indiscernibility and interchangeability of analog and digital has de-differentiated
these modes, then the representational value of their distinction (if any remains)
can only be generated figuratively in analogies to this condition - a range of possi-
bilities that must depend on what the image-producer (and the replication
mechanisms he or she used) has taken to be dense and replete (or not) in the
inscriptions and referential correlates secured in various media.

into computer graphics software and often into the hardware itself." This was clearly true in the tight
constraints that were placed on - and the spatial or perspectival "knowledge" that had to be provided
programmatically to - the early line-finder and other polygonal-modeling programs of the mid-1960s
and early 1970s (see Boden, "Adding the Third Dimension," Artificial Intelligence and Natural Man, pp.
180-90). By the mid-1980s, the application of Benoit Mandelbrot's mathematics of fractals enabled
emergent computer graphics programs to recognize and construct such "natural" entities as trees, coast-
lines, mountains, and (at last) human bodies (Benoit B. Mandelbrot, The Fractal Geometry of Nature [New
York: Freeman, 1982]; for naturalistic modeling, see R. F. Voss, "Random Fractal Forgeries," in R. A.
Earnshaw, ed., Fundamental Algorithms for Computer Graphics [Berlin: Springer, 1985], pp. 805-35).
28. I have tried to heed Manovich's remarks on "the myth of the digital" (The Language of New
Media, pp. 52-55); as Manovich suggests, since the 1980s, at least, the logical distinction between ana-
log and digital has subsisted disjunctively in relation to the technical capacities of analog and digital
replication relative to one another.
29. Ludwig Seyfarth, "The Best Destinations Are Just Down the Street: Where Painters Travel to
Today," in Painting Pictures, p. 43. Seyfarth refers specifically to the work of the painter Ben Edwards, to
which I will turn momentarily.

This content downloaded from

132.174.250.101 on Thu, 12 Oct 2023 19:52:19 +00:00
All use subject to https://about.jstor.org/terms
How to Make Analogies in a Digital Age 85

In Jeff Wall's well-known photographs - large tran

light boxes - painterly properties and concerns palpabl
staging, and presentation of the photographic image. A
have remarked, Wall's light boxes seek in part to analog
new medium, to recuperate or to rival - the glowing o
master paintings. But another obvious analogy for the
from master painting - is the art-historical, museologica
of arraying transparencies and slides of art works on li
And this practices, or its echo in the work, introduces o
Indeed, Wall's works construct a dense and replete tape
between - collations and possible interconversions of - a
Despite the canonically painterly configuration of t
boxes are never wholly analog - or never seem to be wh
definitely determined to be wholly analog - because the
ous in some respects and, for all we can tell, digitally di
box has a window-like size and format; Untangling (199
2.2 meters wide. Due to its material projection from the
entirely believe that we are looking through an opening
sion (in the case of Untangling, into an industrial ware
same time, the large size of Wall's transparency and a
likely distance from real observers of the light box tend

Jeff Wall.
Untangling. 1994.
Courtesy the artist
and Marian

Goodman Gallery,
New York.

This content downloaded from

132.174.250.101 on Thu, 12 Oct 2023 19:52:19 +00:00
All use subject to https://about.jstor.org/terms
86 OCTOBER

that we are l
fidelity by t
also in index
component in
trompe-Uoeil
late. This disc
a sign - and
gized but als
devices of "s
paintings and
of images the
play (Wall's l
they functio
ones - of "pa
afforded by e
then, the ima
of art - in th
sion of "pain
and airport s
photograph
digital-multi
In fact, the
photograph -
projected on
affixed. And
instead of fa
television scr
Not a photog
no longer in
relay of the a
Resolution o
seems to be v
relative weig
configuratio
material and
the chiaroscu
generated in
have been re
figuratively u
brings the an
or discontinu
tography in

This content downloaded from

132.174.250.101 on Thu, 12 Oct 2023 19:52:19 +00:00
All use subject to https://about.jstor.org/terms
How to Make Analogies in a Digital Age

this effect: at this scale, and lighted from behind, many digital p
despite high resolution - -just might reveal their discontinuities in corre
the illuminated scene, which could no longer be seen as everywher
with the uniform radiation emanating within the light box. For tha
Wall's analogy would fail; its delicate recursions would break down.
then, digital technology cannot really be used to make the work. The
cally figures its possibility. By the same token, however, the figu
depends essentially on emergent technical identities and phenomen
bilities of analog and digital.
Wall is not a new media artist. By contrast, painter Ingrid Cala
within the world of technique now defined by software - the digi
and printing of pictures and paintings - like Photoshop, Paintbox, a
In one series, the clarity and visible computability of the intricate su
derives from a sequence and layering of transfers and duplications.
image must be distinguished from a mixture, a soak, of dripped or
and spots that cannot be untangled or separated - that is, differenti
registrations. Calame begins by tracing the contours of spots, stain
imprints in the streets or on walls in Los Angeles. In making and o
traces, she partly obeys - she actively literalizes and digitizes - recom
painters proffered by Leonardo da Vinci and other proponents of t
imagination. These configurations have been transferred at scal
enlarged in square-by-square transcriptions, assigned color values, a
hand with enamels on square sheets of aluminum - nearly two meter
case of Ckckhckhckg/ruh
(2002). One could do this
entirely on a computer - the
painting analogizes a possi-
ble, if possibly mad, print
out - if it were not for the
fact, of course, that one can-
not do it entirely on a
computer. Printing in non-
computational media hovers
nearby: the painting analo-
gizes color lithography or
silkscreen as if routed

through a disjunctive printer

- perhaps a printer that
wants to be a painter or a
painter who actually is a
computer.

Ingrid Calame. Ckckhckhckg/ruh. 2002. Courtesy the artist and James Cohan Gall

This content downloaded from

132.174.250.101 on Thu, 12 Oct 2023 19:52:19 +00:00
All use subject to https://about.jstor.org/terms
88 OCTOBER

In fact, Cala
But this circu
could be gene
reproduced b
tion and the
analog proced
tion of visibl
symbol schem
analogizes tec
in a continuo
continuous t
painting that
in digital mo
analog paintin
Analogy - lik
thus analogy
purposes, an
respects with
secured and r
in what seem
image looks l
quality" mon
of a real pla
garage and a
can seem to d
finest discrim
rative lookin
light-box/ph
television br
then, the ana
ative to a cor
respects. The
son, measure
"two by twos
figure, analog
correlates. W

30. I am indebte
Down the Street
and deeply mult
and glazes" in tr
signature" - unl
selves and which
dure not unlike
pursued in the p

This content downloaded from

132.174.250.101 on Thu, 12 Oct 2023 19:52:19 +00:00
All use subject to https://about.jstor.org/terms
How to Make Analogies in a Digital Age 89

not that they are stains in the street or even that they are p
street. (We also say that they look like drip paintings by
like the raveled ropes in Wall's Untangling, sort of; and s
image relays its analogization: the fact that Calame's stain
hand-painted, neither quite like real stains nor quite
images, is figuratively crucial.
Both Wall and Calame seem to remain on the "old media" side - a side

weighted toward analog modes of representation - in their construction of

tion, despite its analogical engagement with the possibilities of digital rep
of the figure itself. This would seem to be only commensurate with the po
of a richly analogical figuration - constituted in limited continuities betwe
figures and ambiguous sort-of correlates. Still, analogical figuration does n
to be essentially tied to analog mediation.

With this in mind, we can formulate a question for new media art or, more
broadly, for any artists working in a culture of digital media and digitized images.
What happens when artistic analogy - striving for the richness and ambiguity of
analogy and art - arrives at the indiscernibility and interchangeability of analog
and digital? - when it is indifferent to analog and digital modes of representation
in securing apparently identical figurative interactions? Indeed, what happens
when digital modes become more popular or preferred - literally required by the
digital filtering and compression of images, analog as they might be in an
unprocessed state, when they are uploaded to and downloaded from the Web or
local networks? In the visual and spatial arts, analog modes of representation his-
torically did the heavy lifting of artistic analogy, despite crucial digital mediations.
Now, though probably not for the first time, the emphasis has changed. In fact,
the equation might appear to be reversed. In visual/spatial configurations, digital
modes constitute analogies.
In contemporary art we find the emergence of what might be called the
New Analogy: of palpably or at least discernibly discontinuous correlation as the
ground of the figure. In principle this need not be inimical to the construction
of analogical figuration as such. Analogy emerges in limited continuities with
part-correlates, a relation that conceivably can be handled just as well in the dis-
joint and differentiated terms (or notations) of the digital mode as in the dense
and replete terms (or images) of the analog mode. But not surprisingly - and
precisely because of the new indiscernibility of analog and digital modes and the
new indifference of image making in using them - the new analogies, digitally
relayed, look quite like the old, analogically generated. Indeed, they must look
quite like the old if they are to do any deep or rich work of analogization, at least
for the time being.

This content downloaded from

132.174.250.101 on Thu, 12 Oct 2023 19:52:19 +00:00
All use subject to https://about.jstor.org/terms
90 OCTOBER

In Massimo
C-prints beh
beach at Ricc
and from tw
facts of tem
because the
it seems to b
Moreover, ev
tion, and dis
a big Deardo
structed platf
A close insp
have related
them: they a
ers - most o
giant podium
anyone looki
digitally in b
does seem a
whether by
series of Ital
Berlusconi in
to observe m
standing of
Italian norm
bances of th
leisure, delud
the social cor
of the beach
twinned ana
images: we se
In contrast
tographs - o
techniques -
removing (th
of the repr
synthesizing

31. Quoted in A
tion about Vital
of their own an
belie this asserti
naturalistic pain
manipulated pict

This content downloaded from

132.174.250.101 on Thu, 12 Oct 2023 19:52:19 +00:00
All use subject to https://about.jstor.org/terms
How to Make Analogies in a Digital Age 91

Massimo Vitali. Riccione Diptych. 7997

photo-views.32 Hong Kong Stock Exchange (1994), a well-

five meters wide, incorporates such digital alteration - m
nical power and perfection that we must take the slight
in the pictures to function in the figurative analogy. A
analog index - "you are really there, and this really hap
detail of objects and movements" - motivates the digital
Vitali and Gursky are less rather than more "new me
near, that is, the analog, side of the New Analogy. In the
discontinuous correlations, interchangeabilities, and
digital imaging, though visible in the replication, have
into the putative identity of the objective correlate - na
tures under the conditions of consumer and market ca
much digitization the works do or do not deploy, the p
ously correlated - thus as analog in the analogizations -
of the objective correlate. Among other things, this corr
ficity and individuality of the depicted social units - r
beach at Riccione or working at the Hong Kong stock e
ble, even if they could be synthetic: they are not an
interchangeable. According to the artistic analogy, digita
tures and in the social world - can and probably does m
less indifferently.

32. The characterization of Gursky 's editing methods as "aggressive

"Andreas Gursky," in Painting Pictures, p. 201; for information on Gurs
Andreas Gursky, exh. cat. (New York: Museum of Modern Art, 2001).

This content downloaded from

132.174.250.101 on Thu, 12 Oct 2023 19:52:19 +00:00
All use subject to https://about.jstor.org/terms
92 OCTOBER

VI

Compared to
shore of new
a Bourgeois S
The Monumen
landscaping f
mages to Lui
photographs
and metric pr
porate parks
already digitiz
have been com
photographi
coordinate gr
of Edwards's c
the superimp
looks like det
As Seyfarth h
figurative an
marble, or st
"corporate ar
course, contin
tionality, ex
correlates. Non
tions of digiti
so on) remain
actual digitizi
and between
digital pictor
Both painted
ric space of p
representatio
in it. (Indeed
early-sixteen
tions of per
technology in
ends it has su

33. Ludwig Seyfa

34. See Painting
Historical Possibi
trans. John Goo
35. Quoted in Se

This content downloaded from

132.174.250.101 on Thu, 12 Oct 2023 19:52:19 +00:00
All use subject to https://about.jstor.org/terms
 Benjamin Edwards. Monuments of Passaic. 2002.
Courtesy the artist and Greenberg Van Doren Gallery, New York.

This content downloaded from

132.174.250.101 on Thu, 12 Oct 2023 19:52:19 +00:00
All use subject to https://about.jstor.org/terms
94 OCTOBER

Presumably
potential ma
mercialized s
use digitized
heterogeneit
structed wor
digital units
sented. Anal
might well b
conclusion in
correlate pse
To be fair,
that would b
painting in a
tions, it se
proportional
of an averag
unit that ap
each side (or
digital! We c
of a real, dig
regimes of s

Edwards. The
Courtesy the art

This content downloaded from

132.174.250.101 on Thu, 12 Oct 2023 19:52:19 +00:00
All use subject to https://about.jstor.org/terms
How to Make Analogies in a Digital Age 95

no further than Vitali's or Gursky's, and it lacks their

about spatial and social reality outside the picture - even if
that in Edwards's figurations a digital mode - as oppose
analog mode - analogizes the digitized world. Argua
Edwards to figure that world any more rigorously. As is
new media New Analogies, perhaps it simply captures him
Perhaps it is premature to venture categorical critical
logical, political, and cultural environment that responds pe
the unprecedented pace and proliferation - the explosion
tized society and its digital self-processing. These include th
the art world, uncontested - rise of psychologies of digitali
as - the very retinal-optical basis and neural-cortical iden
imagistic knowledge. David Marr's computationalist Visio
appeared a mere fourteen years after Goodman's Langu
Goodman's constructivism - despite its reiteration of E
model of knowledge relative to symbolic form, that is, natu
art - surely helped enable Marr's immensely influential
both "visual" and "mental," is not only computational b
visual-imagistic knowing had generally been understoo
replete analog computer known to, possessed by, and provid
In relation to the new digital model of vision - overlaid
modes of representation and digital technologies of repl
decisively whether Edwards's contribution represents buyin
ping out. The New Analogy contains elements of all thre
Edwards's implied equation of digitized world, digital m
too simple, even though he hopes to find its disjunctions.

36. David Marr, Vision: A Computational Investigation into the Human Repre
Information (New York: W. H. Freeman, 1982). In turn, Marr's work was
research of David Hubel and Torsten N. Wiesel in the early 1960s on the v
On its appearance, Marr's book generated great excitement; it motivated
doctoral dissertation in art history in the early 1980s - that the "sociolo
combined with what I called a cognitive history and algorithmic ana
Davis, The Canonical Tradition in Ancient Egyptian Art [New York: Camb
xvii, 6; for comments on Cassirerean epistemology as it constituted th
gy in art history - its movement between "visual" and "mental" images
and Pictoriality," Res 46 [2004], pp. 9-31). For representative applicat
biology of vision to an array of problems in human imaging, see Ro
Visual Arts (Cambridge, Mass.: MIT Press, 1994); Shimon Ullman, High
and Visual Cognition (Cambridge, Mass.: MIT Press, 1996); Donald D. H
We Create What We See (New York: Norton, 1998); and Zenon W. Pylyshy
What You Think (Cambridge, Mass.: MIT Press, 2003). Despite the pub
tertexts - I might cite Michael Podro's Depiction (New Haven: Yale Un
Maynard's two superb books The Engine of Visualization: Thinking Th
Cornell University Press, 1997) and Drawing Distinctions: The Varieties o
Cornell University Press, 2005) - this debate has yet to be joined, eve
most important nexus for art history and criticism to address at the adv

This content downloaded from

132.174.250.101 on Thu, 12 Oct 2023 19:52:19 +00:00
All use subject to https://about.jstor.org/terms
96 OCTOBER

In the recen
using digital
digital proce
in the image
world. The s
haps "new,"
drawings ph
basis for fre
and acute ang
the picture -
constructed
seems to em
GH), and "fr
transversal o
tric line of
artistic analo

Above left: Alber

Above right: Per
explicate Circli
Torben Giehler.
2002. Courtesy t
Koenig, Inc., Ner

This content downloaded from

132.174.250.101 on Thu, 12 Oct 2023 19:52:19 +00:00
All use subject to https://about.jstor.org/terms
How to Make Analogies in a Digital Age 97

Giehler's paintings address digital constructi

(2002), it is the "polygonal modeling" at the orig
1960s, and hence of today's computer-assisted de
As before, however, the figuration cleaves c
since achieved in painterly mode - albeit mo
Mondrian's, that were close to modern digitalit
some of the historical conditions for it. In more
ing, Ed Ruscha's Metro Plots series (begun in
Ruscha has said that for him these paintings we
on the highway - no pun intended. A lot of [m
Every Building on the Sunset Strip of 1966] was str
involved architecture, perspective, and, someti
works came out of raising myself above the sub
on it in an oblique manner."38 Technically we c
perspective, with its raised standpoint, from Gi
seem to be continuously relocated around the ed
jection plane; its spatial correlate seems to be
Ruscha's Metro Plots, Giehler's Circling Overlan
customary perspective projection (ABCD) as arc
ing a digital net around the analog image. Adm
Ruscha's "plotting" and Giehler's "circling" is fin
tional continuity of analog and digital. Perhaps
seen the world recognized and represented in G
different worlds have been probed in "raising [o
cling overland." To revert to my introduction, Gieh
gives us the intermediate states that are only im

37. For characteristically incisive comments, see Ludwig

Pictures, p. 200; for Giehler's technique, see Max Henry, Torbe
38. "Interview with Ed Ruscha" (www.eyestorm.com/artist/

Giehler. Mont Blanc. 2002.

Courtesy the artist and Leo
Koenig, Inc., New York.

This content downloaded from

132.174.250.101 on Thu, 12 Oct 2023 19:52:19 +00:00
All use subject to https://about.jstor.org/terms
 98 OCTOBER

materially fo
alized there.
morphing st
ogy of movem
My point h
New Analogy
tion of the
continuous c
extra-pictor
unless and u
new media d
current new
versely, anal
closed) new (
kinds. But in
digital analog
nitions of an
possibility o
cance - even
mediations.
How is this to be done? Bite before bit, we might say: to make analogies in a
digital age, analog your digitizations! What's between your fingers but that world
into which you have thrust your hand?

This content downloaded from

132.174.250.101 on Thu, 12 Oct 2023 19:52:19 +00:00
All use subject to https://about.jstor.org/terms