Professional Documents
Culture Documents
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
http://www.jstor.org/page/info/about/policies/terms.jsp
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 support@jstor.org.
The University of Chicago Press is collaborating with JSTOR to digitize, preserve and extend access to Critical Inquiry.
http://www.jstor.org
This content downloaded from 205.207.178.2 on Tue, 31 Mar 2015 15:44:15 UTC
All use subject to JSTOR Terms and Conditions
David Summers
372
This content downloaded from 205.207.178.2 on Tue, 31 Mar 2015 15:44:15 UTC
All use subject to JSTOR Terms and Conditions
Critical Inquiry
Winter 1989
373
art historical problems. At the same time, formal analysis remains the
way in which "the work of art itself" is talked about, and if it is simply
abandoned, then the history of art is placed in the paradoxical position
of being unable to speak in significant ways about the objects of its
peculiar concern, which is not even to mention the problems of fashioning
histories of these objects. An art historian's art historian disavows the
speculations of the "methodologians" and prides himself or herself on
going about what needs to be done in the way demanded by the matter
at hand. It is not possible to avoid methodological questions, however,
when they begin to touch the most basic procedural and pedagogical
matters. Moreover, it is precisely because of the specific character of its
own theoretical past that the history of art now finds itself in a situation
in which the discussion of works of art "themselves" is a problem. It is
that problem I will begin to address in this essay. I wish briefly to examine
the notions "form" and "formal analysis," to consider the relation of
these notions to more general issues of idealism and materialism, and to
raise the question of how kinds of description might be devised that
sidestep old alternatives but still make it possible to talk about art as
distinct from other human endeavors in a historically significant way. I
will not, of course, in a single study succeed in completely overhauling
ideas that so thoroughly permeate the conceptual and institutional structure
of the history of art. Still I believe all these problems and questions must
begin to be addressed before it will be possible to consider higher problems
of art historical interpretation or the further, perhaps more elegant problems of the ways in which art historical interpretation relates to interpretation in fields not immediately concerned with saying what we see.
I shall begin with what Michael Podro calls "interpretative vision."'
This is the art theoretical version of the broadly Kantian notion that
consciousness constitutes its world. It is an idea absolutely central to the
modern intellectual discipline of the history of art. According to this idea,
which is now so familiar as to seem truistic (or axiomatic) in many fields
in the humanities, the artist is not just a recorder of appearances but a
shaper and interpreter of them. Heinrich W61fflin begins his Principles
of Art History with the story of four painters who set out to paint exactly
the same thing just as it was and came up with four quite different
This content downloaded from 205.207.178.2 on Tue, 31 Mar 2015 15:44:15 UTC
All use subject to JSTOR Terms and Conditions
374
David Summers
This content downloaded from 205.207.178.2 on Tue, 31 Mar 2015 15:44:15 UTC
All use subject to JSTOR Terms and Conditions
CriticalInquiry
Winter 1989
375
This content downloaded from 205.207.178.2 on Tue, 31 Mar 2015 15:44:15 UTC
All use subject to JSTOR Terms and Conditions
376
David Summers
This content downloaded from 205.207.178.2 on Tue, 31 Mar 2015 15:44:15 UTC
All use subject to JSTOR Terms and Conditions
Critical Inquiry
Winter 1989
377
This content downloaded from 205.207.178.2 on Tue, 31 Mar 2015 15:44:15 UTC
All use subject to JSTOR Terms and Conditions
378
David Summers
Against the backdrop of the opposition of form and content, the ideal
of the unity of form and content also provides familiar alternatives.
Iconographers might argue that meaning has priority in works of art
because form serves meaning, and there is a corresponding position that
meaning must be given form. Sometimes efforts are made to combine
these possibilities, as when generalizations about the formal styles of
periods or places are worked into other kinds of history, or, more commonly, when the works of an artist are worked into his or her biography.
At this point I would like to introduce some more general questions
of interpretation that will be seen to be closely related to the idea that
art is essentiallyformal in nature. Rosalind Krauss has characterized what
might be called usual interpretation as based on the "notion that there
is a work, x, behind which there stands a group of meanings, a, b, or c,
which the hermeneutic task of the critic unpacks, reveals, by breaking
through, peeling back the literal surface of the work."5 In such a view
the understanding of how images mean underlying usual interpretation
is very much like the traditional notion of allegory. In fact the metaphorical
language of "surfaces" and "rinds" is the same language of integumentum
and involucrum used to explain allegory in the late Middle Ages, when
allegory provided a major justification for the blandishments of poetry
(and painting). Poetry, as the formula went, is truth (or at least higher
meaning) set out in a fair and fitting garment of fiction. The lies of the
poet lead us by an appeal to sense to meaning higher than the appeal
itself. Botticelli's Primavera really sets forth the realm of the Natural
Venus,6 and if that is not what it sets forth, it shows something like that,
some idea of a similar level of abstractness, unity, and simplicity in comparison to the complexity of the surfaces and appearances of the painting.
The demonstration of such significance at the heart of works of art is
never easy, and the example chosen, Botticelli's Primavera, has provoked
a series of scholarly interpretations. These interpretations do not differ
in the relation assumed to hold between the painting and its own meaning,
however. This explanation of higher, unitary meaning is very deeply
lodged in our expectations (and explanations) of images, and has been
for a long time. A Renaissance portrait was thought to show not just
someone's appearance but the soul through the appearance, and a similar
opposition of higher and lower was also argued to hold for emblems, in
which the image was the "body," the text the "soul."7 Allegory, like
naturalism itself, may be said to depend on a kind of transparency of
5. Rosalind E. Krauss, "Poststructuralism and the Paraliterary," The Originalityof the
Avant-Gardeand OtherModernistMyths (Cambridge, Mass., 1985), p. 293; my emphasis.
6. See Erwin Panofsky, Renaissance and Renascences in WesternArt (New York and
Evanston, Ill., 1969), pp. 188-200.
7. E. H. Gombrich, "IconesSymbolicae:Philosophies of Symbolism and their Bearing
on Art," SymbolicImages: Studies in the Art of the Renaissance, II (London, 1972), pp. 12391.
This content downloaded from 205.207.178.2 on Tue, 31 Mar 2015 15:44:15 UTC
All use subject to JSTOR Terms and Conditions
Critical Inquiry
Winter 1989
379
This content downloaded from 205.207.178.2 on Tue, 31 Mar 2015 15:44:15 UTC
All use subject to JSTOR Terms and Conditions
380
David Summers
This content downloaded from 205.207.178.2 on Tue, 31 Mar 2015 15:44:15 UTC
All use subject to JSTOR Terms and Conditions
Critical Inquiry
Winter 1989
381
This content downloaded from 205.207.178.2 on Tue, 31 Mar 2015 15:44:15 UTC
All use subject to JSTOR Terms and Conditions
382
David Summers
This content downloaded from 205.207.178.2 on Tue, 31 Mar 2015 15:44:15 UTC
All use subject to JSTOR Terms and Conditions
Critical Inquiry
Winter 1989
383
circularity and essentialism on the one hand, and to avoid being party
to the fabrication of some of the most murderous myths of modern times
on the other, we must avoid such historical inferences altogether. And
if Gombrich's critique of form and expression is taken seriously, then
some very basic art critical and art historical habits must be called into
question. If Gombrich is right and we cannot see through form (or
through all form) to some kind of meaning, then new ways must be
devised for talking about works of art and for making historical inferences
from them.
Gombrich's critique of form and expression is closely linked to his
critique of historicism. Again following Popper, Gombrich has argued
repeatedly that the sort of "spirit of peoples" thinking that has justified
racism and genocide in our time is a brand of essentialism, or pseudoessentialism (if such a term is possible), that seems to make the ends
of history visible, thus to justify the liquidation of groups seen not to
have a place in the scheme of history. The programmatic ideological
skepticism animating Gombrich's arguments unequivocally rejects totalizing
historical schemes of all kinds and dismisses the Hegelian tradition of
historical interpretation, idealist and materialist alike.
The arguments to this point might be summarized to say that art
historical interpretation is often conducted as if directed toward some
essence, formal or thematic or both. What has been called strong formalism
pursues a visual essence, weak formalism pursues a thematic essence,
and the "unity of form and content" really means that essential form
agrees with animating subject matter. As we have also seen, the modern
notion of form has always implied development. This is to say that the
history of art in being formalist is also historicist. In this essay I shall
consider three meanings of the term historicism. The first, just briefly
encountered in Gombrich's arguments against it, arguments he derives
from Popper, is that laws of history are formulatable and that in general
the outcome of history is predictable. The second, which I shall discuss
in the example of Walter Benjamin, is the idea that history is a universal
matrix prior to events, which are simply placed in order within that
matrix by the historian. The third is the idea that a thing is meaningful
insofar as it is a part of a development. The history of art is heavily
involved in all of these arguments. We need only consider the teaching
of history of art surveys, in which periods, styles, and, at least implicitly,
whole cultures "turn into" something else, and in which this "turning
into" is shown to be a more or less necessary development. This assumption
of continuity contains the further assumption that there is one thing that
changes (art) that is always essentially the same (formal); it also makes
series of events or sequences of period styles into "developments" and
makes art historians look for the ways in which one series of works,
regarded either morphologically or physiognomically, changes into another
one. Such radical diachronicity, of course, has the great advantage of
This content downloaded from 205.207.178.2 on Tue, 31 Mar 2015 15:44:15 UTC
All use subject to JSTOR Terms and Conditions
384
David Summers
This content downloaded from 205.207.178.2 on Tue, 31 Mar 2015 15:44:15 UTC
All use subject to JSTOR Terms and Conditions
CriticalInquiry
Winter 1989
385
assumption that art itself has or is some kind of history. The opposite
of formalism is "contextualism," a set of methods based on the idea that
art is in a very real sense made by historical situations. One setting out
to do formalist history will be concerned with works of art "in themselves";
one setting out to do contextualist art history will be concerned with the
circumstances in which art was made. If these options were to be carried
into practice as they stand, it would mean that formalists would not be
interested in circumstances and contextualists would not be interested
in works of art themselves. Very few art historians really occupy either
of these extremes, which are probably impracticable. The possibility of
the logical purity of either extreme clearly exerts a pull over art historical
writing, however, and this pull is reinforced by the relation of these ideas
to other intellectual issues.
The term "context"is of course a broad one, and art may be interpreted
in relation to a number of external factors. One of the most powerful
contextualist methods has been Erwin Panofsky's iconology. The context
in question here is predominantly cultural and intellectual historical,
however, and for that reason such investigations may be said to belong
to the idealist tradition of art history. (Iconography, the technique preliminary to iconological interpretation according to Panofsky, may also
be used for the purposes of political interpretation, which may point to
a very different general definition of context.) Even though it is not
formalist, such art history is consistent with the formalist and idealist
version of art history as the history of the spirit. This leads to what I
think is the real force of the category of "idealist art history," which is
only understandable if "idealist" means not materialist. "Materialist art
history" points to another definition of context, not as cultural historical
but as social, political, and economic historical. And to return to the story
with which I began, it is now clear why it was necessary to apologize for
a "formalist"paper topic, to justify it by indicating its ideological dimension.
Although ideology is also a complex term-especially in relation to the
question of historical idealism or materialism-its use has at least the
rhetorical effect of transforming an apparently idealist topic into a materialist one.
In the next section I am going to examine materialist art history,
argue that it also is ultimately unsatisfactory as a kind of essentialism,
and close by offering some alternatives outside the bounds of these nineteenth-century metaphysical categories.
I wish to begin by examining certain arguments from Benjamin's
brief "Theses on the Philosophy of History," which seem to me to provide a paradigmatic and influential treatment of what Benjamin himself
calls "materialistichistoriography."'5 His arguments are set in terms closely
related in one or another aspect to those we have been considering.
15. See Walter Benjamin, "Theses on the Philosophy of History," Illuminations, ed.
Hannah Arendt, trans. Harry Zohn (New York, 1969), pp. 253-64.
This content downloaded from 205.207.178.2 on Tue, 31 Mar 2015 15:44:15 UTC
All use subject to JSTOR Terms and Conditions
386
David Summers
Benjamin also wrote about "historicism," but used the term in a special
sense, as we have seen. Benjamin understood the word to refer to the
presumption of a kind of Newtonian continuous time into which all
events can be placed. The method of historicist history is "additive,"
which is to say that what we call "contributions" can be assumed all to
have a place in the same great but abstract narrative. The historicist
historian need not worry about more than setting events in their proper
order, and so, beyond the presumption of underlying time, what Benjamin
calls "universal"(that is, historicist) history "has no theoretical armature."16
Benjamin opposes all of this to his "materialistic historiography." He
considers it imperative that historicist history be rejected because historicism, by making the present seem to be the cumulative progressive
consequence of what has gone before, can be seen to justify the status
quo. Human history, Benjamin wants to say, is neither neutral nor is it
positive progress; it is instead endless carnage and suffering, Hegel's
slaughterbench. The very assumption of the absolute continuity of history
is acquiescence in oppression and murder.
"Materialistic historiography," in short, sees the terrible confluence
of history with momentary victory, and Benjamin wants to eliminate the
political and moral anesthesia of historicism once and for all by denying
its assumption of underlying continuity. Events are not continuous with
one another, they are disjunctive; and materialistichistoriography embraces
this disjunctiveness, making it a part of the historian's own procedure,
which cannot simply be justified by what needs to be done in some
subscientific sense. The past may inevitably be at hand, but if it is alive
it is made alive in the present. The rejection of continuity thus implies
a willfully violent hermeneutics in which the past is appropriated to the
purposes of the present. The historian, Benjamin says, seizes the past
with virile force in a revolutionary transformative act. I linger over these
ideas because such ideas are widely diffused; but to keep to Benjamin's
arguments, they are based on an extreme dichotomy corresponding to
extreme circumstances. Benjamin was trying to cut away the intellectual
underpinnings of fascism and to do so rejected the entire tradition of
what he called universal history. To do that he juxtaposed absolute continuity and absolute discontinuity.
Both Benjamin and Gombrich formulated their arguments under
the immediate pressure of the cataclysmic threat of the rise of Nazism
and both rejected what they called "historicism." However different their
understanding of this word may have been, for both writers the rejection
of historicism meant the rejection of negative principles of continuity
that had to be given up not least because of their horrible moral and
political consequences. The solutions to the problem of discontinuity
16. Ibid., p. 262.
This content downloaded from 205.207.178.2 on Tue, 31 Mar 2015 15:44:15 UTC
All use subject to JSTOR Terms and Conditions
CriticalInquiry
Winter 1989
387
offered by the two writers point in two quite different directions, raising
an alternative that will guide much of the remainder of this essay.
If discontinuity is materialist, then continuity is (at least by implication)
idealist. If Benjamin does not in fact say this, it is a tack taken by other
"materialist" writers and has a place among the issues of this discussion.
Historicism as Benjamin understood it is essentialist if not idealist because
it implies that time itself is a progressive principle of change. As such it
misplaces the locus of significant historical transformation from class
conflict to a metaphysical (or physical) principle.
Since it is based on the rejection of continuity, practitioners of materialistic historiography might be expected to favor synchronic over
diachronic explanation. This again gives a more specific definition to
"context," which is not only economic and political, but is structurally
rather than causally related to whatever is to be interpreted. Art is not
to be explained primarily in relation to previous art, and this position
may easily be translated into terms of the earlier argument. "Form" was
said to be a principle of continuity, expressive of the culture to which it
belonged. Form thus had a kind of built-in historical cogency, but at the
same time it made it difficult (if not impossible) to explain anything but
evolutionary change and relations to broader "spiritual" factors. This
difficulty was acknowledged in the conclusion of one of the greatest and
best-known formalist essays in the literature of the history of art. At the
end of his Principles of Art History W61fflin acknowledged that his formal
principles, which had been used to describe a development from Renaissance through baroque art, could not explain why the same development
should end and begin again. He regarded this recommencement (that
is, the return to a linear style in art around 1800) as "unnatural" and
attributed it to "profound changes in the spiritual world," to "a revaluation
of being in all spheres" (P, p. 233). Meyer Schapiro defined a similar
problem in more general terms in his essay on style, this time suggesting
social rather than spiritual historical context as a solution.
The principles by which are explained the broad similarities in
development are of a different order from those by which the
singular facts are explained. The normal motion and the motion
due to supposedly perturbing factors belong to different worlds;
the first is inherent in the morphology of styles, the second has a
psychological or social origin. It is as if mechanics had two different
sets of laws.17
This paradoxical state of affairs has by no means ended; rather the
pendulum has swung now to one, now to the other side. The availability
17. Meyer Schapiro, "Style," in AestheticsToday, ed. Morris Philipson (Cleveland and
New York, 1961), p. 97.
This content downloaded from 205.207.178.2 on Tue, 31 Mar 2015 15:44:15 UTC
All use subject to JSTOR Terms and Conditions
388
David Summers
This content downloaded from 205.207.178.2 on Tue, 31 Mar 2015 15:44:15 UTC
All use subject to JSTOR Terms and Conditions
Critical Inquiry
Winter 1989
389
historical complexity of the kind Clark describes to the actual characterization of the paintings that make up the backbone of his history is more
subtle and perhaps problematical. Despite his disavowal of such things,
Clark does seem to draw analogies between the formal/visual structure
of the work and its meaning. The Burial at Ornans is simple and straightforward in "structure"; there is "no single focus of attention, no climax
towards which the forms and faces turn." In general, the image is impossible
to characterize as a sum of its parts. Clark seems to wish to say that these
characteristics are not formally expressive so much as semiotic; it is a
"lack of open, declared significance which offended most of all" (IP, p.
83). The painting is, so to speak, objectively evasive; the contradictory
responses of the critics are therefore appropriate, and the reasons for
them are visible to us in the painting. In another painting, the Peasants
of Flagey, "colour is at war with form," and in terms of form, Courbet
has "broken the whole surface into a mosaic of distinct and clashing
shapes" (IP, p. 84). This was "perfectly deliberate," and the painting is
disjunctive and powerfully ununified.
Taken altogether, these arguments combine observations about the
intrinsic expressiveness of disjunctive pictorial forms with an insistence
on semiotic disjunction or ambivalence. It is not clear to me that these
two things are really compatible, a question to which I shall return shortly.
For the present, however, it is enough to observe that if the latter semiotic
alternative were generalized into a program for the history of art and
extended beyond a few years in Courbet's career as a painter (that is,
beyond the reach of basically synchronic explanation), then the history
of art might be written not as a history of form but as a complex history of codes. Such a method would depend on the historical reconstruction
of the likely meaning of any instance of the many codes discernible in
a work of art, and all higher interpretation would have to be based on
this reconstruction. If, for example, composition itself involved hierarchy
(as many have argued that it does), then it might be argued to have
ideological significance, and it might further be argued that a painter's
avoidance of such visual order within that tradition of meaning was
therefore also ideologically significant. Such a historical argument could
be made without recourse to intuitive inference from the expressive
nature of form. But to keep to the matter at hand, if a history were to
be written in which formal and/or semiotic discontinuities evident in
works of art were connected to contradictions in society, it is not clear
what kind of visual continuities might be accounted for. As we have seen,
idealist art history had the advantage of built-in formal/visual continuity,
and the outright rejection of formalism is a rejection of that advantage.
But if there are evident continuities in the history of art, why should
they be ignored as a problem? And how are they to be explained?
Clark addresses this question in his more recent book, The Painting
of Modern Life. In the first place, he accepts the idea that there is a
continuity, which he might not have done, and he further accepts the
This content downloaded from 205.207.178.2 on Tue, 31 Mar 2015 15:44:15 UTC
All use subject to JSTOR Terms and Conditions
390
David Summers
This content downloaded from 205.207.178.2 on Tue, 31 Mar 2015 15:44:15 UTC
All use subject to JSTOR Terms and Conditions
CriticalInquiry
Winter 1989
391
This content downloaded from 205.207.178.2 on Tue, 31 Mar 2015 15:44:15 UTC
All use subject to JSTOR Terms and Conditions
392
David Summers
the mediation of these metaphors ("CG,"p. 152). Here again, the argument
is evidently intended to make a jump between particular instances (why
a stock speculator as a stock speculator might have been drawn to the
paintings of Seurat) to a level of generality covered by the formalist idea
of modernism (why flatness should be a continuous issue in the dominant
tradition of modern art making). It is evident that to concede autonomy
to the issue of medium is also to concede the possibility of a history of
art not reducible to fairly direct historical parallels of the kind made at
the microeconomic level of patronage. In order to avoid this concession,
Clark identifies all evidence of medium-that
is to say, "flatness," the
of
characteristic
modernist
identifying
painting-as "negation and esThis
him
to
trangement."
permits
say relatively few things about modern
art (since its essence in all its instances has already been defined), but it
allows him to say things about modern society he wishes to say. By
denying the "ordinary consistency" of medium in favor of all kinds of
gaps and silences, modernist art becomes an expression of the "negativity"
of modern life ("CG," pp. 152, 154). Modernist art is "the black square,
the hardly differentiated field of sound, the infinitely flimsy skein of
spectral colour, speech stuttering and petering out into etceteras or excuses." It is an art "in which ambiguity becomes infinite, . . . a mere
mysticism of sight" ("CG," p. 154). What concerns us here is once again
the nature of inference from works of art to historical generalization,
the way in which flatness/surface/medium becomes the relation of the
modernist artist to historical circumstances. It is once again very difficult
to separate a statement such as "negation is the sign inside art of this
wider decomposition; it is an attempt to capture the lack of consistent
and repeatable meanings in the culture-to capture the lack and make
it over into form" ("CG," p. 154) from the old idealist statements based
on the assumption that we may see through formal essences to historical
essences. Such formulations to my mind run exactly counter to Clark's
admirable insistence on real historical connection and on the role of art
as a possible (and perhaps inevitable) agent of social change.
Why do such contradictory patterns of explanation occur? I believe
it is because both "idealism"and "materialism"as a priori bases of historical
investigation demand the suppression of one or another kind of historical
evidence. Idealism and materialism are alternative principles of the highest
generality. This is not to say that they are simply different ways of describing
the same thing since each involves the relative deepness or priority of
one or another principle, that is, the generality means that one kind of
thing is always able to be explained in terms of the other ("mind is the
highest form of matter"; "matter is something about mind"). The whole
question thus revolves around the point of which principle is explanatory
relative to the other, and if we turn these distinctions to history, it means
that some kinds of evidence are always explanatory relative to others. I
observed above that a most basic task of the history of art is the explanation
This content downloaded from 205.207.178.2 on Tue, 31 Mar 2015 15:44:15 UTC
All use subject to JSTOR Terms and Conditions
CriticalInquiry
Winter 1989
393
of why works of art look the way they look. This is not a trivial statement;
it provides a criterion according to which both idealist and materialist
art history are of limited explanatory usefulness. Either alternative must
exclude kinds of evidence that bear on the explanation of the appearance
and change in the appearance of series of works of art. Idealist art
historians tend to be unconcerned with the patronage, use, and reception
of images, which are manifestly part of their legitimate history, and
materialist art historians tend to avoid any reference to cultural history
or the technical history of art traditions themselves because they are
assumed to bear merely a "superstructural" relationship to a deeper,
"truer" historical principle.
How is this dilemma to be resolved? We might consider the following.
When Marx inverted Hegel's scheme of history, he retained one essential
thing, namely its absolute totality. This totality followed inevitably from
the continued reduction of all historical process to a single metaphysical
principle, but it also retained a vision of something like overarching
providential purpose in history and in human action. In the terms of
this argument, however, totality is based on a most general kind of essentialism, which is finally disenabling for historical interpretation.
At this point it is necessary to distinguish carefully between idealism
If idealism is essentialism then materialism is its opposite;
essentialism.
and
if
but both idealism and materialism are kinds of essentialism, as I have
argued they are, then both may have another opposite, which I shall call
functionalist. A functionalist history of art is not a history of how art has
changed, or even in a simple sense of what art "means." It is rather a
history of what art has done, or to put two goals together, it is a history
that explains why works of art look the way they look in terms of what
art has been meant to do. Such an enterprise might embrace and continue
earlier forms of investigation and interpretation, but it must also give
them a new orientation and encourage study in new directions.
I am aware that the term "functional" is not unproblematical, and
I wish to use the term in a way related to what I understand Gombrich
to have meant by it in the essay on Raphael's Madonna della Sedia I
considered earlier ("M,"p. 76). Function must be given a special definition
in order to be a solution to the problem it purports to solve. The slogan
"form follows function," for example, simply makes a telos of function
and demands that interpretation proceed in the "usual" allegorical mode
discussed above; that is, we see through the forms to the function (or
we should be able to do that) just as there should be a unity of form and
content. This difficulty is only displaced if we define art in terms of its
social function, since this also implies that we have adequately understood
art when we have done that. Functionalist art history would not proceed
from the work "essentially" understood in any way to context; rather it
would be based on the assumption that works of art are radically cultural
or historical and that they are therefore always meaningful in the cir-
This content downloaded from 205.207.178.2 on Tue, 31 Mar 2015 15:44:15 UTC
All use subject to JSTOR Terms and Conditions
394
David Summers
cumstances in which they are made and that they continue to be meaningful
in new circumstances into which they survive. This definition and the
definition of the interrelations of these multiform meanings are the beginning (but not the end) of art historical understanding.
Before returning to Gombrich's essay, I wish quickly to establish a
few general principles. The problem with which I am concerned is art
historical description, and the first part of the paper was concerned with
an argument to the effect that "formal analysis" is a method of description
that raises more problems than it solves, that by its very definition does
not allow us to ask or answer many questions we might like to ask or
answer, and that therefore puts an undesirable stop to both explanation
and interpretation.
I am going to begin by following a lead offered by Michael Baxandall
and develop the idea that the language of art history is basicallyostensive,21
that when we make art historical arguments involving works of art (as
opposed, say, to documentary arguments about them) we in effect indicate
their characteristics to our audience as we write or talk about them, just
as we usually lecture with some sort of a pointer in one hand.
Ostensivity is a principle of basic aesthetic significance. When Alexander Baumgarten began to stake out the field of modern aesthetics, he
distinguished between what he called "intensive"language, which is general
and analytic, and "extensive" language, which is particularizing, sensate,
and "poetic."22 Beginning from this division, language is used by the
philosopher and the poet in different ways, the latter favoring the metaphors and figures shunned by the former. The poet on such a view
fashions an artificial reality, and does this precisely by avoiding analytic,
logical language, which would at once catapult the matter into the nonpoetic
condition of the general. So in a perhaps more modest way does the art
historian, with the difference that the reality to which the language refers
is not the imagined or constructed particular but the work of art itself.
Some descriptions of the work may be straightforward- size, support,
medium, condition, for example-which are always specific to the work
under discussion. But there are also more complex modes of particularity
related to more complex kinds of interpretation. Baxandall, taking as his
text a paragraph by W61fflin, has argued that, in addition to personal
reactions to works (ugly, chilling), we use a variety of metaphorical language,
some of it formal ("a thicket of lines"), and also language from which
agency can be inferred. This last category points to what Baxandall calls
"inferential criticism," which I shall discuss in detail later on. It is clear
21. See Michael Baxandall, "The Language of Art History," New LiteraryHistory 10
(Spring 1979): 453-65; hereafter abbreviated "LAH."
22. Alexander Gottlieb Baumgarten, Reflectionson Poetry,ed. Karl Aschenbrenner and
W. B. Holtner (Berkeley and Los Angeles, 1954), p. 43. Extensive and intensive language
correspond to confused and distinct ideas; more extensive language is clearer and more
poetic.
This content downloaded from 205.207.178.2 on Tue, 31 Mar 2015 15:44:15 UTC
All use subject to JSTOR Terms and Conditions
CriticalInquiry
Winter 1989
395
This content downloaded from 205.207.178.2 on Tue, 31 Mar 2015 15:44:15 UTC
All use subject to JSTOR Terms and Conditions
396
David Summers
"work" (not a neutral term in this context) in line with that assumption,
then asking why such a thing might have been done. In the case of
Donatello, we might try to explain the general conclusion of the description,
that his relief is based on the "principlesof chiaroscuro," by trying to find
evidence outside the work for such an implicit course of action. We would
quickly ascertain that Donatello really only left his sculpture to speak for
itself, but that focus on light and dark as the elements of representation
is to be found in early fifteenth-century Florentine writing about painting,
and that Donatello's "pictorialrelief" (as every textbook calls it) considerably
antedated both the discussion and the realization of similar optical effects
in painting. We might, in short, come to a provisional but defensible
understanding of why Donatello's relief might have been given the appearance it has. And we might go a step further. If we compare Donatello's
sculpture to that of others, then we begin to understand how novel his
relief style was, and if we pursue the question of why and how, for what
purposes and for whom it came to be desirable in the late Middle Ages
and the Renaissance to make such optical art, then we may become
involved in very complex and basic historical questions indeed. The explanation of its evident optical character might prove to be consistent
with the values given to it by its novelty and by its manifest virtuosity,
values that would continue to inform the art of the modern era at the
beginning of which Donatello stood.
Inferential criticism points partly in the direction of the sort of proximity to the work of art familiar in connoisseurship and conservation.
The same characterization of material, facture, and quality of facture
used to establish the historical relations of the work might also be used
to attribute it to Donatello. The technical studies necessary to conservation
may show the making of the work of art-its own internal historywith a kind of stratigraphic clarity not otherwise evident. Such matters
are of course not cut and dried, but still works of art are in many cases
as distinguishable technically as they are "stylistically" (which is to say
that certain techniques and operations underlie certain appearances;
Riegl's opticality develops together with the ever more prevalent use of
the sculptor's drill). The pigments used in a painting are in themselves
linked to local geology, botany, or patterns of commerce; they are products
of extracting, gathering, burning, and precipitation. Materials themselves
are also of relative status and meaning, as are jade, gold, ultramarine,
and bronze.
Before turning to issues of material and facture, I would like to
examine further my major theme, the notion of form. In Western philosophy the distinction between "form" and "matter" has always been a
deeply and simply gendered one, and this durable and pervasive opposition,
rooted in equally durable and pervasive patriarchal social institutions,
has had the deepest formative consequences for our notions of artistic
making, of imagination ("conception"), and, in its latter-day form, of
This content downloaded from 205.207.178.2 on Tue, 31 Mar 2015 15:44:15 UTC
All use subject to JSTOR Terms and Conditions
CriticalInquiry
Winter 1989
397
"creation." According to this view, form has always been an active, "male"
principle opposed to passive, "female" matter, which was of course considered necessary (or a necessary evil) but was also considered lower than
form and relatively nonexistent.24 The prime example of the relation
between form and matter was always art, usually sculpture. Bronze (matter)
may be poured into any shape, and the shape (form) is determined by
the imagination of the craftsman. The bronze may become any form
whatever, as the craftsman wishes.25 In real historical terms it is hard to
imagine what might correspond to such passive "matter," and yet the
history of art as the history of imagination tends to be written as the
history of imagined forms corresponding to this infinitely malleable stuff.
If we think of works of art in terms of their facture as I have begun
to suggest, that is, if we think of them as indexes of all the purposeful
processes of their making, most of them were clearly collaborative efforts.26
If we think of a statue of the pharaoh simply as an expression of "the
genius of the individual artist" or of the "Egyptian imagination," then
we may be little concerned with the stone "formed" in making it, since
that form (except perhaps as a "block" in its own right) is incidental in
the terms of the argument. But the stone was quarried, that is, cut
squared from the earth, and moved with great expense of labor before
being made into an image. Both the power to command stone to be
quarried and moved and the power to command skilled craftsmen to
fashion it in the pharaoh's image as pharaoh are visible in the work, not
in the sense that the forms of the work express meanings to us, but in
the sense that the work was squared, is of hard stone, is closely similar
to the images of other pharaohs, and those features can be indicated
and explained. They are explained by a process of "contextualization,"
that is, by the explanation of why those people then and there did what
they evidently did do. This example is sufficient to suggest that fairly
obvious characteristics of works of art may-one might rather say mustlead to more or less specific social historical understanding.
24. Aristotle, in Physics 192a, called matter a "mother" of what comes to be, "a joint
cause, with the form.... The truth is that what desires the form is matter, as the female
desires the male" (The Basic Worksof Aristotle, ed. Richard McKeon [New York, 1941], p.
235). Gerda Lerner, The Creation of Patriarchy (New York, 1986), pp. 205-11, provides a
discussion of the pervasive form-matter distinction in Aristotle's philosophy.
25. Aristotle, Metaphysics,1032a- 1038b.
26. On such "actor-orientation," see "LAH," p. 463, where Baxandall cites Clifford
Geertz, "Thick Description: Toward an Interpretive Theory of Culture," The Interpretation
of Cultures:SelectedEssays (New York, 1973), pp. 3-30. Baxandall's ideas seem to me to be
an adaptation of Geertz's (and finally Gilbert Ryle's) notion of "thick description," in which
case other kinds of description, the physical state of the work, personal and metaphorical
language, would drop into the category of the "thin." "The thing to ask ... is not what
... ontological status is. It is the same as that of rocks on the one hand and dreams on
the other-they are things of this world. The thing to ask is what their import is: what it
is, ridicule or challenge, irony or anger, snobbery or pride, that, in their occurrence and
through their agency, is getting said" (Geertz, The Interpretationof Cultures, p. 10).
This content downloaded from 205.207.178.2 on Tue, 31 Mar 2015 15:44:15 UTC
All use subject to JSTOR Terms and Conditions
398
David Summers
This content downloaded from 205.207.178.2 on Tue, 31 Mar 2015 15:44:15 UTC
All use subject to JSTOR Terms and Conditions
Critical Inquiry
Winter 1989
399
This content downloaded from 205.207.178.2 on Tue, 31 Mar 2015 15:44:15 UTC
All use subject to JSTOR Terms and Conditions
400
David Summers
This content downloaded from 205.207.178.2 on Tue, 31 Mar 2015 15:44:15 UTC
All use subject to JSTOR Terms and Conditions
CriticalInquiry
Winter 1989
401
This content downloaded from 205.207.178.2 on Tue, 31 Mar 2015 15:44:15 UTC
All use subject to JSTOR Terms and Conditions
402
David Summers
necessary for "pre-iconographicaldescription"(our recognition that an Annunciation, before we know it is an Annunciation,is a winged man alighting
in the room of a reading woman) has to be supplemented by the "controlling
principle" of the "history of style," which is "insight into the manner in
which, under varying historical conditions, objectsand eventswere expressed
byforms."3'Obviously this opens up a whole system of possibilities covered
neither by iconography nor by the definition of form as the neutral
substratum of recognizable images. What it suggests is that, like themes,
the means of representation (as these elements of the "history of style"
might be called) are themselves members of series; they are not just
formal series but are culturally specific configurations changing at rates
different from those of iconographic motifs. This in turn suggests some
of Panofsky's own earlier essays, "The History of the Theory of Human
Proportions as a Reflection of the History of Styles," or "Perspective as
Symbolic Form," where basic organizing devices, rather than being consigned to the realm of noncontent, are treated as historical subjects in
their own right, as subjects for historical explanation."2
Let us consider the further example of modeling. Modeling is something that falls into the category of form, that is, into the category of the
arbitrary, the configurations of which are determined by the style and
judgment of the artist. But as Gombrich has pointed out in several places,
modeling does not appear in the history of art with the kind of randomness
such an explanation might lead us to expect. The overwhelming majority
of examples of modeling are in Western art, and in certain periods of
Western art, classical antiquity and the naturalistic and neoclassical styles
begun in the late Middle Ages and the Renaissance.33 If we consider
what modeling is and does, then it becomes much more than one of the
ways in which content happens to get presented. Modeling may be defined
as the systematic transition from light to dark or from one hue to another
(usually the former) within a shape. When it is modeled, a shape becomes
an apparent three-dimensional form, and the area surrounding the shape
becomes a virtual space. Modeling is thus an integral part of the Western
project of the imitation of appearances of physical things, but it is a part
the examination of which adds a rich dimension of meaning to the more
general notion of mimesis. Modeling not only puts a form in space, it
places it in light and in implicit relation to a viewer. Moreover, it puts
the language of painting in relation to the psychology of perception and
to the science of optics, both of which began at about the same time and
in the same place as modeling began in painting. (Such connections only
31. Panofsky, Meaning in the Visual Arts, pp. 40-41, and Studies in Iconology,p. 15.
32. See Panofsky, "The History of the Theory of Human Proportions as a Reflection
of the History of Styles," Meaning in the Visual Arts, pp. 55-107, and "Die Perspektive als
'Symbolische Form,' " V6rtrageder BibliothekWarburg4 (1924-25): 258-330.
33. See the introduction to Summers, The Judgment of Sense: Renaissance Naturalism
and the Rise of Aesthetics(Cambridge, 1987), pp. 1-31, for a fuller treatment of this issue.
This content downloaded from 205.207.178.2 on Tue, 31 Mar 2015 15:44:15 UTC
All use subject to JSTOR Terms and Conditions
Critical Inquiry
Winter 1989
403
This content downloaded from 205.207.178.2 on Tue, 31 Mar 2015 15:44:15 UTC
All use subject to JSTOR Terms and Conditions
404
David Summers
might be given to them. The first of these questions concerns the relation
of the analysis and reconstruction I have described to other kinds of
interpretation. The position I am trying to set out might be summarized
to say that it is necessary to understand as best we can what was meant
in order to try to figure out what was not meant in any specific instance,
to understand the lines in order to read between them. I have been
concerned with the problem of description and its relation to the historical
reconstruction of what was meant, or what might have been meant, but
the further question of what more is revealed in the actual combination
of meaningful elements in any work of art points to other interpretive
measures.
The works from which we begin are particular, and we begin to try
to understand them by considering them aesthetically in the sense of
assessing them precisely in their particularity. When we analyze them by
placing their constituent parts in series, that is, by trying to establish
what they might have meant, we necessarily treat them both partially
and in general. Although such seriation may be highly significant in itself
(we might write the history of altarpieces, perspective, the theme of
Judith and Holofernes, the trade in paintings), none of these is sufficient
for the historical interpretation of the works from which we began. As
mentioned above, we must go beyond analysis and make synthetic judgments. Thesejudgments are double in the sense that they are ourjudgments
and that they are also judgments about the character of the particular
synthesis of elements in the work. They are, in other words, judgments
about the historicalparticularity of the work that elucidate our experience
of the work as it stands before us, usually in spaces (museum spaces)
which, as machines for formalist looking, are fundamentally incompatible
with the kind of historical imagination I am trying to describe. Analysis
and reconstruction, which bring the work into view as historical, that is,
which contextualize it in any number of possible dimensions, thus heighten
both the past and present particularity of the work. Since works of art
are almost always decontextualized, either because they have been removed
from their original spaces of use or because we who are writing their
history do not belong to the spaces and times for which they were made,
this historical particularity is always more or less imaginary (thoseother
spaces of use, times, purposes, audiences). This imaginary particularity
informs our understanding of the work, that is, it changes our judgment
of it in the present. This brings us back to the work as it stands before
us. The synthesis consequent to analysis is the historian's judgment,
shaped by the initial judgment of the work, which the new judgment
then modifies. These judgments should not be regarded merely as supplements to the inherent limitations of analysis and reconstruction; rather
they are integral with the particularity of the work to be explained.
contribute
take that example-may
Psychoanalytic interpretation-to
significantly to the explanation of the particular character of a work of
This content downloaded from 205.207.178.2 on Tue, 31 Mar 2015 15:44:15 UTC
All use subject to JSTOR Terms and Conditions
Critical Inquiry
Winter 1989
405
art as it stands and to the understanding of specific ways in which meaningful elements were treated or related when it was made. Such interpretation, like the synthesis of the historically particular work in imagination, on which it also depends, is informed by reconstruction but is
finally rooted in the specific relation between interpreter and work.
As for social and economic interpretation, it seems to me that much
might be accomplished simply by taking seriously material and facture
and the various contexts of production and use to which they point,
rather than trying to construe "style" in a way that links it to social or
economic circumstances, which is to my mind clearly an old idealist habit
turned to new purposes.
The second issue I wish to raise briefly is that of uniformities among
works of art in unrelated cultures. Formalism involved interpretive procedures that assumed uniformities and then explained them. But if formalism is rejected and all art is assumed to be bound to different contexts,
how are formal uniformities to be explained? How, for example, are we
to account for the practical universality of the appearance of hierarchical
images based on bilateral symmetry and other kinds of planar order? I
believe such uniformities arise because images are always embodied and
share real space with those who see and use them. This real space (as
opposed to the virtual space we see in the kinds of images to which we
are most accustomed) may be constructed in ways that exploit the significant
values of our own spatiality. The meaning of some images is stated more
in these terms, others less. This real spatiality means that what we call
art possesses kinds of significance irreducibly different from the significance
of texts. Works of art must be understood in terms of the real spatial
conditions of presentation that assume any number of conventional forms,
that is, any number of culturally specific forms.34
Art historians seldom examine the implications of their own habits
of interpretation, preferring instead to limit themselves to the application
of these professional habits to individual problems or to borrow "approaches" from other more self-critical fields. In this essay I have tried,
by examining some of the major categories of the history of art as it has
been discussed in this century, not only to revise old patterns of description
and inference but also to argue against all kinds of interpretation I have
called allegorical and against both idealism and materialism as sufficient
principles of art historical explanation. Such nonmetaphysical art history,
everywhere intersecting with other kinds of history but reducible to none
of them, implies that art itself, rather than being an inclusive (or exclusive)
category is defined by the historyof art, by all that which we call art has
done. It also implies a future potential history of art that is all the things
34. See Summers, "The 'Visual Arts' and the Problem of Art Historical Description,"
ArtJournal 42 (Winter 1982): 301-10; Summers, "This Is Not a Sign: Some Remarks on
Art and Semiotics," Art Criticism3 (1986): 30-45; and Schapiro, "On Some Problems in
the Semiotics of Visual Art: Field and Vehicle in Image-Signs," Semiotica 1 (1969): 236.
This content downloaded from 205.207.178.2 on Tue, 31 Mar 2015 15:44:15 UTC
All use subject to JSTOR Terms and Conditions
406
David Summers
that what might be called art might do. I have also tried to argue for a
kind of historical discourse based on the assumption that any work of
art was made by people in real circumstances and that in those circimstances
it was a meaningful thing to have done. Such an assumption is consistent
with the history of the changing uses to which works of art are put after
they are first made. Both the meanings art had and the meanings it has
subsequent to its making are analyzable and reconstructable (which is
not to say recoverable) by the same means. Beneath the level of aesthetic
characterization of the whole work a high degree of unanimity about
aspects of works that may seem unimportant is possible and at this same
level much fundamentally significant history is possible. At the same
time, these arguments mean that art history will always be at least explicitly
critical because both analysis and synthesis involve judgments about the
work at hand. But it is perhaps most important to insist on the possibility
and the necessity of preserving the ostensive and consensual discourse,
the reconstructive discourse, by means of which richer contextualization
and therefore richer historical criticism are possible.
This content downloaded from 205.207.178.2 on Tue, 31 Mar 2015 15:44:15 UTC
All use subject to JSTOR Terms and Conditions