METAPHILOSOPHY Vol. 31, Nos. 1/2, January 2000 

REPRESENTATION AS THE REPRESENTATION OF EXPERIENCE
F. R. ANKERSMIT

ABSTRACT: This essay deals, mainly, with the notion of representation. Representation is associated with texts and, as such, is contrasted to the true singular statement. It is argued that the relationship between the text and what the text represents can never be modeled on the relationship between a true singular statement and what the statement is true of, and, furthermore, that the former relationship is aesthetic while the latter is epistemological in nature. This aesthetic relationship between the represented and its representation is investigated from the perspective of historical writing. This offers an interesting perspective because historical texts are representations, but they also aim at giving us “the truth” about the past. Historical representation, therefore, is the kind of representation coming closest to the cognitive claims traditionally investigated by epistemology. The notion of aesthetic and historical experience is explored in order to define the relationship between historical representation and what it represents.

Introduction
In the last few decades the notion of representation has acquired a prominent role in contemporary intellectual debate.
The explanation is a certain imbalance in twentieth-century investigations of the relationship between language and reality. These investigations – as best exemplified by analytical philosophy as it developed in the postwar period – focused mainly on notions such as reference, meaning, truth, sense data, and the epistemological problems surrounding these notions. Useful as these investigations have undoubtedly been, they ordinarily took their point of departure in the simplest utterances – “the cat is on the mat” being the paradigmatic example – because of the assumption, at first sight not unreasonable, that the kind of problem occasioned by the specimens of a more complex use of language could be adequately dealt with only after a satisfactory solution had been found for those simpler epistemological technicalities.
However, contrary to expectations, these epistemological problems turned out not to permit solutions that were acceptable to all participants in the discussion. And the result was that philosophical debate got stuck in the debate on these elementary uses of language, while the problems occasioned by the more complex uses of language were never addressed.

149
Because of that, hermeneuticists, literary theorists, semioticians, historical theorists, theorists of poetry, rhetoric, or the pictorial representation of reality, and so on, were all left very much to their own devices when attempting to deal, as well as they could, with the question of how the relationship between complex chunks of language and reality should be defined for their respective disciplines. Two observations are in order here. In the first place, the theorists of these more complicated uses of language all tended to use th term “representation” for referring purposes. But, second, though the term “representation” was almost universally adopted, everybody tended to see his own discipline as presenting us with the paradigmatic examples of representation.
For obvious reasons this has seriously hampered the debate on representation. If an interdisciplinary discussion of representation took place at all, this discussion rarely was more than a dialogue des sourds. And though postmodernism must be praised for its effort to develop an interdisciplinary discussion, its well-attested lack of interest in and respect for philosophical rigor made things worse rather than better. Hence, though the notion of representation was frequently used and
much discussed in recent decades, a thorough and well-considered analysis was never given by any theorist: the term functioned rather like a coin that was passed from hand to hand without anyone ever closely scrutinizing it. Therefore, in trying to bring some light to the darkness in which the notion of representation is still hidden, this essay will not begin with a historical survey of the debate on representation up till now. For this strategy could only result in a prolongation of the present impasse. Instead, I propose to begin again with the idea of a true singular statement.
However, my effort will not be to explain representation in terms of that idea, but to explain just where representation goes beyond the singular statement and all that one could say about the relationship between it and the world. Put differently, my point of departure will be the conviction that an important step toward a correct understanding of the notion of representation
will have been made as soon as one realizes that the relationship between a representation and what it represents is essentially different from the relationship between a true statement and what it is true of.
The Singular Statement, the Text, and Representation As the etymology of the word suggests, a representation makes present
what is absent. And it might certainly seem that a true singular statement should, if only because of this, be considered to be the paradigmatic case of representation. For in asserting true statements, we ordinarily use language to compensate for an absent reality. For example, you may say “the cat is on the mat” to somebody who is for some reason or other not in a position to observe this part of reality for himself – perhaps because he is in another room, or because he is asking you about the cat over the

150
telephone. And then the true statement will be as good as reality itself for this person and will rightly seem to him to be a “representation” of reality in the proper sense of the word.
This trivial and unpretentious observation may already deepen our insight into representation and suggest how representation differs from a true statement. I said just now that the true statement about S may, in a way, be as good as S itself for us, if we are unable for some reason or other to see S ourselves. But this phrase “for some reason or other” should make us think. For this qualification is less innocent than it may seem at first sight. Its very casualness suggests that this contingent inability to see S for ourselves can always be remedied “in one way or another.” More specifically, the phrase suggests that subjects of knowledge are in principle interchangeable: “for some reason or other” you happen to be elsewhere and that’s why I am telling you about the cat. But if you were sitting where I am sitting now, you could have seen for yourself that the cat is on the mat.
This seems to be an essential condition of the statement’s being true. For truth is no private matter – hence the perennial attraction exercised by the notion of the transcendentalist, or cognitive intersubjective, ego as soon as the notion of truth is under discussion.
However, suppose a student asks you about a book he had not read and you tell him what you know about it. Surely, nothing seems wrong, at first sight, with this phrase “telling somebody what you know about a book.” But it will then not be easy to indicate “where” the student should place himself in order to see for himself the truths you mentioned (as in saying “the cat is on the mat”). Reading the book is obviously not the right answer to this “where” problem, since people interpret books differently.
I am sure that any other answers one might be tempted to try would be even less plausible. Hence, in this case at least, there is a problem with the interchangeability principle. We should perhaps agree with Gadamer’s view that there is no room for this “interchangeability of knowing subjects” – so essential to the notion of truth – in the context of reading and interpreting books. Gadamer is undoubtedly correct in inferring from this that hermeneutic interpretation can never be reduced to truth. What
you tell the student about the book, your “representation,” surely “compensates” for the student’s ignorance to a certain extent, but never in the way that a true statement may plausibly be said to do so.
Things may get worse. A book is at least something that you can take into your hand; you can read it, and everybody who does so will see in it exactly the same chapters, sentences, words, and letters as any other reader. Even Derrida could not deny this – though Derrida, being Derrida, would undoubtedly devise some sophistry attempting to do so. But think now of the kinds of things that historians represent in their writings, for example, the French Revolution. What are, so to speak, the “chapters,”
“sentences,” “words,” and “letters” of the French Revolution that will look exactly alike to all its students? Obviously this is a question impos-

151
sible to answer – not because we happen to lack the means to do so
(though we may indeed have an idea of what would be a good answer to
such a question), but simply because it is a nonsensical question. Here
nothing could be said to be given, objectively or intersubjectively.
Hence, not only will we agree with Gadamer that all readers will come
up with a different reading of books offering representations of the French
Revolution; it will also be impossible to affirm unequivocally what the
French Revolution is, that is, of what is represented in these books. The
French Revolution is not something like the state of affairs of the cat’s
lying on the mat or like a text (seen as a complex of words, sentences,
chapters, etc.) – not simply because of its immense complexity, but, even
more, because the French Revolution is part of the past and therefore
unobservable. This is far from being a contingent fact about historical
representation. For Arthur Danto was undoubtedly right in claiming that
historical representations essentially compensate for the past’s absence
(Danto 1968, 95). Since we have no access to the past as we have to
present states of affairs about our cat (or even texts viewed as long rows
of letters, words, sentences, etc.), historians write their histories of the past
to compensate for this absent past. It will now begin to dawn on us that
both from the perspective of the subject and from that of the object the
relationship between the representation and what it represents is essentially
unlike that between a true statement and what the statement is true
of.
Let me offer a more pointed formulation. We must realize that the relationship
between a true singular statement and reality can be investigated
epistemologically, although this does not hold for representation.
Epistemology investigates the relationship between reality, meaning,
reference, and truth and presents us with a philosophical matrix within
which such notions are usually analyzed. It is hardly necessary to explore
these matters here: to have a bearing upon representation it should be
sufficient to point out that we can be certain a priori that representation, as
such, can never be adapted to the parameters of epistemological debate.
For representation admits of a certain liberty regarding the relationship
between language (the representation) and reality (the represented) that
would be utterly unthinkable even in the most liberal of epistemologies. A
circle may “represent” a city on a map, the letter o, the sun, the earth, a
human face, and so on (Ankersmit 1994, Chapter 4), and what it actually
represents is to a large extent dependent on context and tradition or even
simple convention. It hardly needs to be said that such a degree of freedom
would make nonsense of the whole enterprise of epistemology.
One might conceive of an extreme conventionalist epistemology, such as
that advocated by Hegel in his Phenomenology, and argue that epistemology
must itself be historicized. In the writings of authors such as Kripke or
Davidson, we may find attempts to develop such an epistemology. We may
be sure, then, that such a historicist and conventionalist epistemology will

152
be able to deal with conventions such as circles representing the sun, or a
city on the map, by giving a justification (probably causal) of these
conventions. And then we would have made the world safe for epistemology,
by disposing of the threat posed by representation. However, this
strategy would reduce epistemology to toothless irrelevance. For what
could possibly be the use of an epistemology justifying all our incompatible
and purely contingent conventions? Our reaction would probably be
that under such circumstances, we would be better served by a history
(and thus a representation) of the different ways in which words and
things have been tied together in the past than with such a conventionalist
epistemology. In other words, the only epistemology that can be reconciled
with representation is an epistemology that presupposes (historical)
representation instead of conferring such standing to any would-be bearer.
It is decisive that epistemology (whether historicized or not) is, by its
very nature, incapable of dealing adequately with representation. For if we
admit a represented and its representation, both will give rise to exactly
the same epistemological problems. For example, the categories of the
understanding in Kantian epistemology do not and could not distinguish
between a representation and what is represented without properly fulfilling
the tasks assigned to them within the framework of Kantian epistemology.
The Kantian categories of the understanding do not care whether
they are applied to a landscape or to a picture of that landscape – if all is
well in epistemology, they will be just as effective in both cases.
Epistemology does not enable us to articulate philosophically what differentiates
the represented from its representation, just as, for example, the
purely physical properties of two books are useless if we want to articulate
their difference in content. In other words, by its very nature epistemology
is indifferent to the problems of representation and therefore is of
no help in solving the problems of representation. Furthermore, as soon as
epistemology succeeds in distinguishing categorically between the represented
and its representation, something will seem amiss, since its sui
generis work transcends that distinction. Hence, the fact that a true singular
statement is epistemology’s favourite research object automatically
prevents the statement from serving as a paradigm of representation.
And, indeed, all attempts in the last few decades to achieve an epistemological
reduction of the representation of reality to the problems of true
statements have utterly failed. The following may explain the failure: logically,
a historical representation is a proposal – that is, a proposal to see
part of the past from a certain (metaphorical) point of view. Put differently,
the historian’s representation is a proposal for the organization of knowledge
(i.e., the knowledge expressed by the singular statements that constitute
the representation) and therefore is not knowledge itself.1 Similarly,
__________________________
1 Our age, with its overwhelming mass of information, has every reason to be interested in how knowledge is organized by the historian. But that is what historical writing has

153
the alphabetical organization of the telephone directory does not reflect
some order existing among the people listed in the directory (nor even
among their telephone numbers). The alphabetical organization is, instead,
a device for making knowledge (about their telephone numbers) accessible.
And so it is with historical representation: it gives us an organization
of knowledge about the past – making the past as accessible as possible –
though other historians may disagree and argue that another organizational
representation will be better.
I hasten to add this note. The fact that representations are proposals
about how to organize our knowledge (of the past) certainly does not
imply (as is often argued by structuralist and poststructuralist theorists of
history) that we cannot rationally assess the merits and the demerits of
competing representations. Fortunately reason has more strings to its bow
than merely showing us how to discover truth. We must realize that
although we cannot possibly characterize a proposal (even the wisest or
the most stupid) as either “true” or “untrue” – because it would be a category
mistake – that hardly prevents us from being able to say (un)reasonable
and (un)tenable things about them. The proposal to light the fire if we
feel too warm is not a “false” but a “stupid” proposal. Someone might
object that we could rephrase our proposal into a set of epistemologically
analyzable statements (e.g., “lighting a fire will increase temperature” or
“if one feels too warm, temperature must be lowered”), but that would be
of no help to the epistemologist. For the representation suggested by the
historian is a proposal regarding what (is thought to be) the best representation
of the past; and the notion of the best representation, in its turn,
necessarily (if my account is correct) implies making a proposal that
would affect any rephrasing of the notion of representation in terms that
the epistemologist might ask us to consider. We would otherwise lack any
criterion bearing on how to reformulate the proposal in the idiom to which
the epistemologist would like to restrict himself. Similarly, it is possible to
paraphrase statements about feelings, like being in love, as statements
about hormones and their effects – but the notion of being in love is of
course the indispensable basis for advancing pertinent statements about
physiology. Without the first we could never tell which statements of the
second sort would yield a correct paraphrase.
To conclude: it is not a true singular statement but the set of such
always been about. The notion of metaphor can help us to understand how this organization
of knowledge is achieved in and by the historian’s text: for example, the metaphor “the
earth is a spaceship” organizes knowledge of our earth into a coherent whole not only by
furthering our understanding of the earth’s ecosystem, but also by suggesting how best to
deal with it (Ankersmit 1983, Chapter 7; 1994, Chapter 2). One might say that metaphor
defines, organizes, and structures a context. The significance of context, thus understood, is
nicely brought out by Saffo: “It is not content, but context that will matter most a decade
from now. . . . In a world of hyperabundant content, point of view [as always defined by
metaphor] will become the scarcest of resources” (Saffo 1994, 74, 75).

154
statements that constitutes a historical narrative and that constitutes the
paradigmatic representation. It is, therefore, to the notion of such a set of
statements that we shall soon turn.

Narrative and the Work of Art as Representations
An obvious candidate for representation is, of course, the work of art. Hence, a comparison between historical narrative and work of art may deepen our insight into the nature of representation. Flint Schier has successfully argued for a crucial logical difference between a true singular statement and the representation of reality afforded by a work of art (Schier 1986, 115ff.). Suppose we compare a picture (or even only a photograph) of Marlon Brando in which Brando looks surly with the statement
“Brando is surly.” In the case of the statement, we distinguish clearly and unambiguously between the subject-term and the predicate-term and, hence, between what the statement refers to and the property being attributed to it. No such distinction is possible in the case of the picture or photograph: we cannot divide the picture into two parts, of which the former corresponds to Brando and the other, to the surliness attributed to him. Referring and attribution always go together in the picture (Ankersmit 1995).
Much the same holds true of the historian’s representation of the past.
The historical text consists of individual true statements, while their totality is the representation proposed by the historian; hence, from the level of the true statement to the narrative, we move from the epistemological level to that of representation. Elsewhere I have argued that this shift in level from epistemology to aesthetics becomes possible thanks to the presence in historical narrative of a logical entity whose existence and function have not been noted before. This entity, which I call the “narrative substance,” can be defined as follows. All the individual true statements in a narrative that determine its character must be considered to be the parts of the narrative substance that is proposed in the historical representation in question. On the basis of this definition we can formulate statements such as “N1 is F,” “N1 is G,” and so on, where ‘N1’ is the name of the
narrative substance presented in the historical text, F and G are predicates in p, q, and so on, and p, q, and so on, are the true statements contained in the text.
In the first place, the notion of the narrative substance thus defined will enable us to clarify and to explain a number of properties of (historical) representation, of the nature of historical debate, and the (nonepistemological) criteria by which to decide the acceptability of individual historical representations of the past (Ankersmit 1983, Chapters 5, 8). In the second place, this may contribute to a better understanding of the asymmetry between individual statement and (narrative) representation. For statements like “N1 is F” and “N1 is G” are all analytically true, since they

155
can analytically be derived from the notion N1. And that means that we
cannot distinguish in the case of these statements between a subject that is
referred to and the properties that are attributed to it – the predicate is
already contained within the notion of the subject and, therefore, finds
itself, so to speak, on both sides of the copula. As already argued, the
meaning of the true singular statement in a historical narrative is therefore
twofold: the statement asserts that something is the case in reality (p) and,
at the same time, the statement contributes to the definition of a certain
representation of the past (“N1 is F”). And we can never isolate these two
functions of the statement from each other, just as we cannot say of an
individual brushstroke in a portrait that the brushstroke refers to the sitter
or attributes a certain property to him or her.
There is a difference between the historical text and the painting,
however: we could not say of the individual brushstroke what we can say
of the narrative’s individual statements, namely, that they can be either
true or false.2 But apart from this, there is a close symmetry between the
historical and the artistic representation of reality; they cannot both be
reduced to the matrix of epistemology, and thus they require a new and
open investigation of the relation between the representation and the
represented.

Historical Representation
If, as we observed a moment ago, the relationship between the represented and its representation is, in the case of the painting, still obscured by the problems occasioned by the individual brushstroke, it becomes clear that historical representation is ideally suited to aiding us in investigating the relationship between the representation and what it represents – hence, to investigating the notion of representation as well. The problem of historical representation exceeds the interests of historical writing and of historical theory, in that the analysis of historical representation helps us penetrate the secrets of representation and of what lies beyond the framework of epistemology.
Where, in historical writing, do we actually find the representation and
____________________________________
2 This places us in the dilemma of either conceiving of the individual brushstroke as a representation, consisting of components that can be said to be either true or false (as in the case of the statement), or of conceiving of the brushstroke as being already such a component. But whatever way we would decide, the question would arise what could be meant by the notion of the truth of falsity of these components. However, Goodman’s exemplification might prove helpful here: if the component correctly exemplifies the corresponding part(s) of the represented, the component could be said to be a “true” rendering of it. Obviously, the application of Goodmanian exemplification to the components of the work of art does not imply that under the aegis of exemplification the notion of truth can be introduced in aesthetics in order to sustain the notion of the truth of the painting as a whole. We need only recall here that the truth of the individual statements of a historical representation has no implications for its quality. And so it is in painting.

156
the represented? Perhaps an introduction and synopsis by the historian
would help the reader of a historical text identify those statements that
constitute its narrative substance, but, often, that is doubtful. More often
than not, it will not be clear whether a statement actually serves this function
or merely serves the cognitive purposes of the historian. The double
aims of all historical writing – to tell the truth about the past and to
achieve a specific narrative representation of the past – inhere indissolubly
in the text as it is presented to the reader. With the exceptions just
mentioned, the reader has no adequate clue about how strictly or unambiguously
the two levels of meaning may be distinguished from each
other.
However, this does indicate what means the reader does have at his
disposal. For if the historian wishes to present a specific representation,
something can be said in a negative way, that is, by contrasting this representation
with other representations of (roughly) the same part of the past.
For example, suppose that in a representation of the Enlightenment – say,
Foucault’s – statements can be found about how the ideology of the
Enlightenment effected a certain regimentation of the human individual,
whereas such statements are rare or absent in more conventional representations
(focused, say, on the liberation of humanity from its selbstverschuldeten
Unmündigkeit). We should then be able to identify, at least
partly, those sets of statements that determine the nature of the narrative
substance separately proposed in the two accounts. In any case, without
such contrasts, we cannot hope to succeed in determining the substance of
narrative representations. Narrative representations are recognizable only
by their contrasts with other representations.
Four consequences follow. First, we must associate representation with
“negativity” rather than with “positivity.” For the specific image of the
past that is offered acquires its contours only in the (implicit) denial, and
not in the confirmation, conveyed by other representations. Here you find
the essentially “dialectical” character of representation – the very site, in
fact, where the logic of representation agreed with the Negativität so much
valued by the young Hegel. Indeed, Hegelian dialectics achieves its march
through history thanks only to denial and negation; historical insight into
the past comes into being in a similar way. Second, we can now grasp the
notion of “intertextuality” so much emphasized by the deconstructivists.
Representation is “intertextual” in the sense that it is relational; the conditions
for its existence are not entirely intrinsic to itself by depend on the
existence of other representations. From that perspective, one might as
well say that the text, or representation, is what it is not and that the generation
of representational meaning is the result of a process of “dissemination”
via other texts. Third, it will be clear by now that competing texts are
part of a Wirkungsgeschichte having its origin in a specific historical problem
and that clarity with regard to a historical representation can be
achieved only by viewing it in the context of such a Wirkungsgeschichte.

157
This is how Hegelian dialectics, Derridian deconstruction, and
Gadamerian Wirkungsgeschichte are connected in representation.
Nevertheless, within the framework of this essay, a fourth conclusion is
of even greater importance. Although Hegel’s dialectic promises the ultimate
triumph of Absolute Mind, no such guarantee can be given for historical
representation. Representation requires debate, as deconstruction
suggests; and representations cannot be created ex nihilo, according to the
Gadamerian Wirkungsgeschichte. And yet, debate and tradition cannot
ensure the optimal representation of the past. On the contrary, if the statement
can be tested for its truth and falsity by observation (and I wish to
distinguish clearly here between observation and experience), the logic of
representation suggests that the actual development of representations is
determined most decisively by interpretive traditions (and all kinds of
circumstantial factors influencing such traditions), not by the intrinsic
nature of what is represented. Hence, representational meaning has a
tendency to withdraw within “the prisonhouse of language” – the “prisonhouse”
of representations – to use Jameson’s expression.
This finding is not only generally agreed on, but also regretted, which
is perhaps why people speak of a “representation crisis” in contemporary
modernism/postmodernism disputes. Current theories of representation,
from Hegel to Gadamer to Derrida, confirm that a direct contact with the
represented is impossible; contact is always mediated by other representations
and by a representational history. In this sense, the representations
may be said to repress the represented.
We begin to see here the point of the priority of representations over
the represented – observed by such theorists as Gombrich and Danto and
by such postmodernist authors as Baudrillard.3 Already in “The Decay of
Lying,” Oscar Wilde had pursued the paradox that life imitates art, that
an acquaintance with Balzac’s novels might well reduce our friends to
mere shadows of Balzac’s characters. Great art is not a copy of the world;
it is, rather, the world that imitates art. The truth is that we ordinarily
know the represented only through its representation. Here, we may agree
with Baudrillard when he describes the representation as a “simulacrum,”
a “hyperreality” that is more “real” than represented reality itself. This
priority of the representation over the represented is already evident from
the fact that such relatively weak and aimless factors as tradition and
convention are ordinarily more powerful in determining the nature of
representation than the represented itself. (Think of the history of art.)
Hence, we need not restrict ourselves to art, nor even to how the media
represent the events of our contemporary world – in the newspapers and
on television – in order to appreciate the powers of representation:
_____________________________
3 Needless to say, the priority of the representation over the represented is an entirely different matter from my claims discussed at the beginning of this essay regarding the irreducibility of representation to epistemology.

158
representation forms a great deal of our quotidian reality. In this sense,
aesthetics is an integral part of our daily life. Without representation we
would have no represented – and that would mean the loss of a great part
of our world and our grasp of it. (This is no postmodernist version of
idealism.) Similarly, although we would not say that merely using the
name “United States of America” created a certain part of the globe’s
surface, we may still say, reasonably, that a representation gives us access
to something that was not “there” before we had the relevant representation.
Thus, the United States did not exist in a certain sense before people
used the name to designate it, even though using the name certainly did
not create rivers or fields or mountains. Idealism is a crude philosophical
notion distinctly unsuited to the realist import of representations.
In a sense, then, without representation we would lack what is represented
(in a sense contrary to that in which physical states of affairs
obtain, even in the absence of statements describing them). If it were
objected, for example, that in the case of a portrait the represented surely
precedes its representation, it may be remarked that if we were presented
with different paintings of the same sitter, his so-called objective features
are certainly not, as such, decisive in determining “the best representation.”
It is not photographic accuracy that makes us prefer Titian’s portrait
of Charles V to the one painted by Van Orley. It would be wrong-headed
to regret the priority of the representation over the represented as if there
might be some lamentable imperfection or incurable defect of representation
– hence, proof of its sad incapacity to satisfy the demanding standards
of true statements. That representations do not satisfy such standards is
precisely why we have representations: they contribute something more
than does a true statement, something that conveys a sense of order in
daily, not scientific, reality, that makes it possible for us to live life and
deal with our world. A “true representation” of reality would be just as
useless to us as the facsimile of a text handed to us after we had asked how
best to interpret it.

The Representation Crisis and Historical Experience
Still, though it is only representation that makes our world a livable world, “our” world is enclosed within itself by tradition, by language, by our representational habits and conventions. And this raises the question of whether we might ever break through the magic circle of representation and the traditions governing it. That an affirmative answer must be given was already suggested at the end of the previous section. For it is certainly difficult to indicate exactly what guides and constrains our traditions of artistic representation. Surely Goodman and Gombrich are right to criticize Ruskin’s “myth of the innocent eye”; surely the way the world appears to us and all artistic representations of the world are mediated by tradition. But that does not prevent the specific history of art from being

159
richer and more variegated than other histories; the constraints of context
and tradition may be stronger in the history of representation than elsewhere.
But it is precisely this that makes artists eager to escape from their
constraints. Where tradition is strongest, the will to break with it may be
strong as well.
Once again, historical writing affords the best departure for investigating
the chances of escaping “the prisonhouse of representation.” Although
artistic representation has, from ancient times, favored the precept docere,
placere, et movere, art has tended to emphasize placere et movere (though
there is indeed a tradition spanning Vasari and Goodman that features the
cognitive aspects of art). By contrast, historical writing has viewed as its
primary (if not its only) task the presentation of the past wie es eigentlich
gewesen. The disciplinary goal of historical writing has always been
mainly cognitive. Hence, if we associate the cognitive dimension of
historical writing with the urge to escape from the “subjectivity” of tradition
and context, historical representation will, perhaps better than artistic
representation, bring us to the strange no-man’s-land between the domain
of the true and the domain of the beautiful that is so rarely visited by
philosophers.
If, then, we ask ourselves how historians sometimes succeed in breaking
out of “the prisonhouse of representation,” we should be well advised
to consider what has been called either “historical sensation” or “historical
experience.” Many historians and poets since Herder and Goethe have
recorded how, in a moment of supreme historical grace, they underwent a
direct and immediate contact with the past.4 The most important features
of “historical experience” have been summed up by Huizinga (Huizinga
1950a, 564ff.; 1950b, 71ff.). According to Huizinga, historical experience
is typically effected by relatively trivial objects or events, such as an
antique print, an old song, or entering a building that has not changed for
centuries. The explanation is, first of all, that, say, a painting by Titian or
Raphael is, for us, so much the paradigm of the development of the history
of art that we find ourselves unable to view it as the immediate expression
of a certain historical reality. Second, historical experience is “an ebriety
of the moment,” as Huizinga puts it; it is something the historian undergoes
rather than constructs or conveys at will. And, third, in historical
experience, the historian has the conviction of being in a direct and
completely authentic contact with the past. In this connection, Huizinga
tellingly links historical experience to the sense of touch rather than to
sight or hearing. The eye and the ear are our most “educated” senses: what
they present to the mind is always part of a complex history of seeing and
hearing and hence is in need of decoding. The sense of touch, however,
adapts itself to the most direct and immediate contact with reality. The
_______________________________
4 For a more detailed “phenomenology” of historical experience, see Ankersmit 1993 and the last chapter of Ankersmit 1994.

160
sense of touch does not require a medium in the way visual and auditory
perceptions do. By the way, this is also why the Aristotle of De anima and
De sensu elevated the sense of touch over the other senses and why we
may find in his conception additional insight into the nature of historical
experience.
Two questions need still to be answered. First, can the claim be justified
that historical experience really transcends a representational tradition
(or Wirkungsgeschichte)? And, second, if so, how can the content of
historical experience find its way to a representation of the past?5 If we
start with the first question, we may observe that almost all of twentiethcentury
philosophy opposes the idea of the possibility of a direct and
immediate contact with reality that historical experience suggests. We
have all become Kantians in one way or another: we all dismiss out of
hand a contact with reality that is not mediated by language, narrative,
scientific theories, Kantian categories of the understanding, or so forth.
It is not my intention to attack this nearly universal consensus. Indeed,
all experience, even historical experience, is irrevocably contextual.
Surely, when Huizinga had his historical experience of the late Middle
Ages during his visit to the van Eyck exhibition in the summer of 1902,
from which his The Waning of the Middle Ages originated, he could only
have acquired his special susceptibility to those paintings through what he
already knew of the period. Even so – and this is crucial – it does not
follow that we should also accept the contemporary dogma of the contextbound
character of the content of experience. My thesis is precisely that
although it is undeniable that the occurrence of a historical experience is
context bound, the same need not be true of the content of the experience
occasioned.
A metaphor may clarify my intentions. Suppose one looks down from
an airplane to the ground beneath. Often clouds will prevent us from
seeing the ground, but when there is an opening in the clouds, we will
have an unobstructed view. This is how it is with historical experience.
Most often the clouds of tradition and context prevent us from seeing the
past itself, but that does not preclude the possibility of a direct view of
the past in the momentary passing of these “contextualist” clouds. We are
surely justified in arguing that it is the clouds that determine whether this
may happen or not – and in that sense historical experience is indeed
context bound. However, though the clouds may determine the fact that
we see the ground (or the past) at all, they cannot determine what we will
actually see under favorable circumstances. This unwarranted shift from
the “that” to the “what” is the non sequitur that we may discern in many
contemporary arguments in favor of the context-bound character of all
experience. Furthermore, the metaphor suggests that what historical
_______________________
5 I shall not deal with this question here. For an attempt to answer it, see Ankersmit 1997.

161
experience offers the historian will not have the character of a deeper penetration of or extrapolation from knowledge (of the past) already possessed (i.e., what is given by context), but a sudden awareness of an “openness” in the contextual clouds themselves – an awareness that, in a
sense, requires and is determined by the nature of those contextual clouds.
Thanks to this, something of the past that was hitherto obscured by context
and tradition becomes visible for the first time. It is at such moments that
the historian succeeds in breaking out of “the prisonhouse of language and
of representation” and that a fresh contact is made with the past that makes
possible new avenues of historical research not dictated by prevailing
context and tradition. Speaking more generally, insofar as the development
of historical writing is determined not by a logic of its own
(Ankersmit 1994, Chapter 5), but by what the past itself is actually like,
historical experience realizes this direct and immediate contact.
The metaphor of the airplane suggests a further determination of the
conditions of the possibility of historical experience. We can see the
ground from the airplane only when the airplane (and we ourselves) are
located above an opening in the clouds. Similarly, the past can be experienced
only when there is a certain congruity between the relevant part of
the past and the subject of experience. Hence, all preoccupation with these
contextualist and traditionalist clouds (favored in Gadamer and Derrida) is
at odds with the desire to escape from the representation crisis. More
specifically, we may anticipate that that preoccupation necessarily leads to
an attack on experience itself. Both Derrida and Gadamer themselves
were well aware of this: Derrida’s statement “il n’y a pas dehors texte”
eliminates experience together with its potential object. If there is nothing
outside the text, there is nothing to be experienced “out there.” Gadamer
is even more explicit: “Wir wissen, was für die Bewältigung jeder
Erfahrung ihre sprachliche Erfassung leistet. Es ist, also ihre drohende und
erschlagende Unmittelbarkeit in die Ferne gerückt, in Proportionen
gebracht, mitteilbar und damit gebannt würde” (Gadamer 1960, 429).
Language as the embodiment of context and tradition, as “das Haus des
Seins das verstanden werden kann,” destroys, as Gadamer himself explicitly
recognizes, the structure of experience and places itself between us
and the world in the way that the Kantian categories of the understanding
did for the first time two hundred years ago.
However, this momentary overcoming of context, tradition, language,
narrative, and so on, is possible only if the harmony between the historical
subject and the relevant part of the past is actually realized. This can
be clarified with the help of Dewey’s theory of aesthetic experience,
developed in his Art as Experience (Dewey 1987, 287 ff.). We should
note, in this connection, that historical experience can be seen as a variant
of aesthetic experience. For not only is historical experience most often
effected by works of art from the past; but, more important, the submission
of our perceptual apparatus to the object of experience is the defining

162
feature of both historical and aesthetic experience. Turning, then, to
Dewey’s account of aesthetic experience, we find that Dewey distinguishes
between two extremes in our experience of reality. In one, we can
undergo reality passively in the way a stone rolls down a hill without
“being aware” of what happens to it. The other extreme presents itself
when we react to reality with what Dewey calls a “mechanical efficiency,”
that is, when the impulses sent out from reality to us are merely “read” as
the symbols or signs of the existence of certain states of affairs in reality.
In the latter case, reality disappears, so to speak, behind the “screen” of
symbol or sign. In the former, the subject of experience is reduced to irrelevance;
in the other, the object, the reality, becomes an irrelevance, a mere
occasio (to use Malebranche’s terminology) of subjective processes.
Dewey emphasizes (in a way closely similar to the Aristotelian notion of
mesotès) that aesthetic experience comes into being only when the right
balance has been struck between the extremes. (One is reminded that
communication between equals is far more complex, far richer, than that
between master and slave.) Both aesthetic experience and historical experience
come into being only when neither reality nor subject is the
absolute master in the relationship.
If this requirement between subjects and objects of experience is not met, if either the subject or the object is too strong a partner, we shall find a loss in the authenticity of experience itself. We can now fathom what motivates contemporary resistance to the very idea of a direct experience of reality. Indeed, when experience is exclusively related to its simplest components, to the most elementary perceptions, to “sense data,” as Western epistemology has been in the habit of claiming since Descartes’s
method of resolving disputes, it need not surprise us that experience completely collapses beneath the heavy load of tradition and context that hardens into fixed theory and prejudice. Indeed, the “indivisible atoms” of elementary Protokollsätze, “sense data,” immediately lose their inner structure (if they have such a structure) when they have to be reduced to already existing categories, theories, narratives, languages, and so on. However, if the content of experience has the character of a “surface” or
a “volume” rather than of a “point,” it will be able to resist successfully the powers of language, tradition, and so on.
Hence, the direct contact with the past afforded by historical experience is possible thanks only to the complexity of the content of experience itself. Paradoxically, the direct experience of reality is possible solely on the occasion of a complex contact (that seems initially to favor context and tradition), whereas, in truth, a mere “pointlike” experience yields unproblematically to the pressures of context and tradition. The modernist’s reduced and simple analytic posits are not the natural enemies of context and tradition but, in fact, its best friends. This explains another feature of historical experience. The two requirements we have arrived at – the complexity of experience and the Deweyan and

163
Aristotelian equality of subject and object – imply that what reality offers the subject should, in a certain sense, have its counterpart in the subject.
There must be an appropriate “harmony,” just as the person looking down from the airplane has to do so at precisely that moment when the clouds have momentarily lifted. This is why historical experience is always so personal and even tied to a specific moment in a person’s existence, why it cannot be repeated at will, and, more important, why it shares a strong element of self-awareness with the experience of the sublime, as defined by Kant.6

Once Again, Statement and Representation
Let us return for a moment to the singular statement and representation. I want to argue that, in accord with our foregoing discussion, the singular statement – that traditional paradigm of knowledge, the prime object of epistemological investigation in Western philosophy – does not deserve the name, since it is essentially “imperfect,” or unable to be analyzed completely as it stands, in the sense that it always needs further “perfection” through connection with either a lawlike statement or a narrative
representation. The lawlike statement perfects the singular statement because it expresses a certain relationship between predicates and, in this way, eliminates the indeterminacy that always exists in the relationship between the subject-term and the object-term of a true statement. For we may make as many true statements about the subject of a statement as we like, but there is no analogue in a lawlike statement. Representation, in turn, perfects the true singular statement because, as we have seen, it binds all statements contained in the representation by a relationship of analyticity. This places the epistemologist’s enterprise in a peculiar light. For if singular statements are by no means the perfect expressions of knowledge one used to prize, would that not imply that most of the epistemologist’s efforts have always been misguided? For if a singular statement has its
perfection only in relation to some lawlike statement or representation, it follows that the statement can be fruitfully investigated only after its relation to a lawlike statement and/or a representation has been adequately defined. In sum, epistemology has the bad luck fastening on a hybrid piece of language that it took to be the key to all the secrets surrounding the relationship between language and reality. For a true singular statement is, in fact, a complicated mix of the “purer” uses of language that we find in lawlike statements and narratives. Contrary to appearances, these two uses of language are more elementary than what we find at the level of a true singular statement.
_____________________________
6 For a further explanation of this self-awareness and for a correction of Kant’s argument to the effect that this self-awareness involved in the experience of the sublime is a recognition of our moral destination, see Vall 1994.

164
It may be objected that if a representation consists of statements (as I have been arguing), such statements are presupposed by representation and not the reverse. However, our notion of the (types of) things that we discern in reality and to which the subjects of statements typically refer is itself the result of a typification process of representations (Ankersmit 1983, 155–69). To mention an example: suppose a specific narrative substance regarding the Enlightenment were accepted by everybody, not
just historians. Then the notion of “Enlightenment,” which, as a representation, initially served only to organize our knowledge of that period, would change into the name for a certain object. From then on, sentences like “the Enlightenment is G” are no longer analytical statements, as is always the case in representations, but statements with the same general logical structure as “this house is white.” In this way representation can turn into a description of (that is, supplies categories of) objects, and
narrative proposals can become the names of things (Ankersmit 1994, 88ff.). If someone were to object that this account still has singular statements at its deepest level – hence, that representation presupposes true statements – I should answer in the following way. First, we should conceive of daily reality (and the objects contained in it) as a hierarchy.
That is to say, if we have statements regarding things, for example, like historical individuals, books, paintings, and so on, these statements can be used within representations, and if these representations get codified by context or tradition, new sets of things come into being in conformity with the process described in the previous section. But at the deepest level we will not find statements, only experience; for, both logically and temporally, experience precedes statements. In the most primary contact with reality, reality is still not codified in the way that daily reality always is, but then of course true statements will not be able to be formulated. As understood here, experience denies (or perhaps defies) convention and tradition and the way these determine the categories of things that are met with in reality, while, at the same time, it possesses a complexity that could be construed in accord with the notions of objects and their properties as structured in the statements of a later phase. In the beginning was not the object (as modernist philosophers like the Strawson of Individuals believe), or the word or language (as postmodernists and hermeneuticists believe). In the beginning was experience.
If we recall that convention, tradition, and context (not epistemological criteria) determine the relationship between representation and the represented, it follows that the (categories of) objects that we discern in reality are not the result of some quasi-divine decision but the result of convention and tradition. As we know, the Foucault of Les mots et les choses provides some striking examples of the culturally determined and even arbitrary way in which we cut up the universe into (categories of) individual things. Accordingly, the following picture obtains: first, we have the reality that is investigated by the sciences (and that corresponds to

165
lawlike statements), and, second, we have our daily reality, which is the result of a culturally determined codification process of our representations of the world. One cannot say that one of these two “realities” is more basic or more fundamental than the other. It would be misguided to conceive of the reality of the scientist as having its foundations in daily reality (as we often claim). First, the very solidity and objectivity of daily reality, which seem indeed to anticipate and reflect the rigidity and severity
of scientific argument, are, as we saw a moment ago, the product of tradition, context, and convention and typically lack the solid foundations on which all scientific research is built. Second, in many ways, scientific reality is the result not of an analysis of the nature of daily reality, but of an alienation from it. For the codification of representations (that gives us our daily reality) is nullified by science, by dissolving the things of that world into their lawlike properties and by discovering deeper patterns in
those properties that do not have a counterpart in daily reality.
Here science and representation move into opposite directions: the lawlike statement leaves daily reality behind, whereas representation, in an ever expanding and changing system of semantic layers, constructs ever new codifications of daily reality upon older codifications, thus making our daily reality more and more complex. The reality of the physicist may appear to be fundamentally “unreal” to us, but it consists of, in Leibniz’s apt phrase, phenomena bene fundata. By contrast, daily reality, in spite of being so much more “real” to us than the physicist’s reality, is poorly grounded because its shaky foundations rest in vaguely recurring patterns of how we represent the world (or, more precisely, how we experience it). Here we may discern the deceitful authenticity of daily reality and the reason we should associate authenticity not with the experience of daily reality, but with the kind of aesthetic experience I mentioned earlier.
Indeed, much of this repeats itself on the level of experience. I must (to gain the point) invoke a bit of intellectual history. Ever since Dilthey, Lukasiewicz, and Mates, it has been pointed out that most of Western epistemology and metaphysics after Descartes may be viewed as a reprise of classical Stoicism. We must think here primarily of the Stoic speculation about the so-called logoi spermatikoi believed to inhere both in reality and in rational thought about reality – thus guaranteeing the possibility of reliable knowledge. It may be argued that most of Western philosophy since seventeenth-century rationalism has been a series of attempts to develop an epistemological and metaphysical definition of these Stoic logoi. Alternatively, philosophers were always looking for the tertia comparationis that would enable us to argue from reality to knowledge, and vice versa. Needless to say, the search made sense only on the assumption that there were indeed such things as logoi or tertia of the kind invoked. It should not surprise us where these tertia were sought. Their roots might be situated in the objects of the world: they were supposed to

166
provide patterns repeated on the level of experience and knowledge. This is how the empiricist and positivist systems actually originated. The logoi or tertia supplied the forgotten metaphysical assumptions behind such systems. Hence, the world of things cannot form the unshakable basis of knowledge the positivists and empiricists supposed: we now have every reason to agree with Hegel’s criticism (in the Phenomenology) leveled against empiricism. Again, the world of daily reality proves to be an
abstract world. All these approaches, so popular nowadays and so widespread among such different branches of contemporary philosophy as analytical philosophy, philosophy of science, and deconstruction, are obviously descendants of Kant’s Copernican revolution.
We now understand why the kind of experience discussed in this essay can play no more than a subordinate role in contemporary thought. For it is obvious that these tertia tended to take the place of experience – as intermediaries between subjects and objects. Stoicism and its postmodernist variants succeed to the extent that they eliminate experience. Empiricist systems tended to see reality as the stronger partner in the “transmission line” from reality to knowledge, whereas rationalist systems tended to see the subject as the stronger partner. But both systems required tertia in arguing from one to the other. For related reasons, postmodernism is no more than a continuation of modernism, not the radical break it has always claimed to be. Because of postmodernism’s sovereign disdain for reality, tertia tended to become useless parts of the metaphysical machine
or, perhaps, wheels that invariably produced twentieth-century textualist variants of nineteenth-century idealism. All that postmodernism has contributed to contemporary thought originated in the linguistic definition of the tertia it proposed, but it soon turned into a profoundly counterintuitive linguistic appropriation of all reality.
A break with both modernism and postmodernism can therefore be achieved only if we are prepared to abandon the tertia proposed. Only then can the equilibrium that Dewey attributed to aesthetic experience be achieved. The tertia will always and inevitably destroy this equilibrium. This is where postmodernism, for all its alleged interest in aesthetics, has so sadly and conspicuously failed. It need not surprise us, therefore, that Stoicism never attempted to develop even a rudimentary aesthetics.
Despite their own aestheticism and their aestheticist rhetorics, postmodernists have never really been aware of the nature and force of the challenge posed by aesthetics. This is why, in the end, they have remained too close to (modernist) Stoicism to be able to offer a viable alternative. For what the postmodernists insufficiently appreciated is that what interests us in the work of art, and why we claim to discover its never-ending “newness,” is that it challenges each limited content that we might ascribe
to the tertia. It is precisely the absence of the tertia, or, more specifically, our awareness of their absence, that prevents the domination of subjects over objects in the sense postmodernists advocate. For by eliminating the

167
tertia, we break the link by which the subject and the object might dominate one another.
We see, therefore, what aesthetic experience and historical experience have in common, how the former contributes to a better understanding of the latter, and why (in both) the complexity of the content of experience ensures the directness and immediacy that Huizinga associated with historical experience.

Conclusion
Modernists believe that the true singular statements holds the key to the most important philosophical secrets. Postmodernists, with their interest in texts and the representation of reality mediated through texts, demonstrate what is wrong with the underlying intuition of twentieth-century philosophy of language. Hence, postmodernism is a correction and displacement of modernism.
Nevertheless, we must not forget that postmodernism uncritically accepts one of the major tenets of modernism: namely, the idea that our theories, our narratives, our language – in short, all those variants of the Kantian categories of the understanding that have been invented in our resourceful century – determine our knowledge and experience of the world. Here, the hermeneuticists, the deconstructivists, and other theorists of context and tradition can be said to have been even more radical than their modernist predecessors.
We reach a stage beyond postmodernism only if we consistently and relentlessly detranscendentalize both modernism and postmodernism. We can achieve this goal if we take aesthetics as our point of departure in reflecting on how we ourselves relate to the world, if we ruthlessly eliminate all tertia from our conceptual inventory, and if we recognize that the notion of experience is best suited to carrying out these new philosophical tasks. We should then have moved from modernist truth, via postmodernist representation, to post-postmodernist experience.

Department of History
University of Groningen
Oude Kijk in ’t Jatstaat 26
P.O. Box 716
9700 AS Groningen
The Netherlands

References
Ankersmit, F. R. (1983). Narrative Logic: A Semantic Analysis of the Historian’s Language. The Hague: Martinus Nijhoff.
——. (1993). De historische ervaring. Groningen, the Netherlands: Historische Uitgeverij.

168
——. (1994). History and Topology: The Rise and Fall of Metaphor. Berkeley: University of California Press.
——. (1995). “Statements, Texts and Pictures.” In The New Philosophy of History, edited by F. R. Ankersmit and H. Kellnerr. London: Reaktion Books.
——. (1997). “Sprache und historische Erfahrung.” In Historische Sinnbildung, edited by J. Rüsen. Reinbek.
Danto, A. C. (1968). Analytical Philosophy of History. Cambridge: Cambridge University Press.
Dewey, J. (1987) Art as Experience. New York: Berkley Publishing Group.
Gadamer, H. G. (1960). Wahrheit und Methode. Tübingen, Germany: J. C. B. Mohr (Paul Siebeck).
Huizinga, J. (1950a). Verzamelde Werken II. Haarlem, the Netherlands: Tjeenk Willink.
——. (1950b). Verzamelde Werken VII. Haarlem, the Netherlands: Tjeenk Willink.
Saffo, P. (1994). “It’s the Context, Stupid.” Wired (March).
Schier, F. (1986). Deeper into Pictures. Cambridge: Cambridge University Press.
Vall, R. van de. (1994). Een subliem gevoel van plaats: Een filosofische interpretatie van het werk van Barnett Newman. Groningen, the Netherlands: Historische Uitgeverij.