Paul Ricoeur

Hermeneutics and the human sciences

Essays on language, action and interpretation
Edited, translated and introduced by
JOHN B. THOMPSON
Cambridge University Press
Cambridge
London New York New Rochelle Melbourne Sydney
Editions de la Maison des Sciences de I'Homme
Paris
144 The theory of interpretation
tution of which the subject would possess the key. In this respect, it
would be more correct to say that the self is constituted by the 'matter'
of the text.
It is undoubtedly necessary to go still further: just as the world of the
text is real only insofar as it is imaginary, so too it must be said that the
subjectivity of the reader comes to itself only insofar as it is placed in
suspense, unrealised, potentialised. In other words, if fiction is a fun-
damental dimension of the reference of the text, it is no less a funda-
mental dimension of the subjectivity of the reader. As reader, I find
myself only by losing myself. Reading introduces me into the imagi-
native variations of the ego. The metamorphosis of the world in play is
also the playful metamorphosis of the ego.
If that is true, then the concept of 'appropriation', to the extent that it
is directed against Verfremdung, demands an internal critique. For the
metamorphosis of the ego, of which we have just spoken, implies a
moment of distanciation in the relation of self to itself; hence under-
standing is as much disappropriation as appropriation. A critique of
the illusions of the subject, in a Marxist or Freudian manner, therefore
can and must be incorporated into self-understanding. The conse-
quence for hermeneutics is important: we can no longer Oppose her-
meneutics and the critique of ideology. The critique of ideology is the
necessary detour which self-understanding must take, if the latter is to
be formed by the matter of the text and not by the prejudices of the
reader.
Thus we must place at the very heart of self-understanding that di-
alectic of objectification and understanding which we first perceived at
the level of the text, its structures, its sense and its reference. At all
these levels of analysis, distanciation is the condition of understanding.
5. What is a text? Explanation
and understanding
This essay will be devoted primarily to the debate between two funda-
mental attitudes which may be adopted in regard to a text. These two
attitudes were summed up, in the period of Wilhelm Dilthey at the end
of the last century, by the two words 'explanation'
For Dilthey, 'explanation' referred to the model of.
rowed from the natural sciences and applied to the hIstOrical dlsclph.nes
by positivist schools; 'interpretation', on the other hand, was a deriva-
tive form of understanding, which Dilthey regarded as the fundamental
attitude of the human sciences and as that which could alone preserve
the fundamental difference between these sciences and of
nature. Here I propose to examine the fate of this opposition m the light
of conflicts between contemporary schools. For the notion of explana-
tion has since been displaced, so that it derives no longer from the nat-
ural sciences but from properly linguistiC models. As regards the
cept of interpretation, it has undergone profound
which distance it from the psychological notion of understandmg, m
Dilthey's sense of the word. It is this new position of the problem, per-
haps less contradictory and more fecund, which I like to explore.
But before unfolding the new concepts of explanation and understand-
ing, I should like to pause at a preliminary fact
inates the whole of our investigation. The question IS thIS: what IS a
text?
I. What is a text?
Let us say that a text is any discourse fixed by writing. to
this definition, fixation by writing is constitutive of the text Itself. But
what is fixed by writing? We have said: any discourse. Is this to say
that discourse had to be pronounced initially in a physical or me.ntal
form? that all writing was initially, at least in a potential way, speakmg?
In short, what is the relation of the text to speech?
145
146 The theory of interpretation
To begin. with, we are tempted to say that all writing is added to
antenor speech. For if by speech [parole] we understand, with
de Saussure, the realisation of language [langue] in an event
of dIscourse, the production of an individual utterance by an individual
speaker, then each text is in the same position as speech with respect
to language. Moreover, writing as an institution is subsequent to
and seems merely to fix in linear script all the articulations
,:hIch have already appeared orally. The attention given almost exclu-
sIvely to phonetic writings seems to confirm that writing adds nothing
to. the phenomenon of speech other than the fixation which enables it
to Whence the conviction that writing is fixed speech,
that be graphics or recording, is inscription of
speech.- an mSCrIptIOn WhICh, thanks to the subsisting character of the
engravmg, guarantees the persistence of speech.
The psychological and sociological priority of speech over writing is
not It may be asked, however, whether the late appearance
of wrItIng has not provoked a radical change in our relation to the very
of discourse. For let us return to our definition: the text
IS a discou.rse fIxed by writing. What is fixed by writing is thus a dis-
course be. said, of course, but which is written precisely
It IS not FIxation by writing takes the very place of speech,
occurnng the SIte where speech could have emerged. This suggests
that a IS really a text when it is not restricted to transcribing
an antenor speech, when mstead it inscribes directly in written letters
what the discourse means.
This of a direct relation between the meaning of the statement
can be supported by reflecting on the function of reading
m relation to writing. Writing calls for-reading in a way which will
enable us shortly to introduce the concept of interpretation. For the mo-
let us say that the reader takes the place of the interlocutor
.as takes the place of speaking and the speaker.
is thus not a particular case of the speak-
relatIon. It is not a relation of interlocution, not an
m.stance of dialogue. It does not suffice to say that reading is a dialogue
WIth author through his work, for the relation of the reader to the
book .IS of a completely different nature. Dialogue is an exchange of
and answers; there is no exchange of this sort between the
WrIter and the reader. The writer does not respond to the reader.
the book divides the act of writing and the act of reading into
two sIdes, between which there is nq communication. The reader is
,
:

J
,
I
I
1
147
What is a text?
absent from the act of writing; the writer is absent from the act of read-
ing. The text thus produces a double eclipse of the reader and the
writer. It thereby replaces the relation of dialogue, which directly con-
nects the voice of one to the hearing of the other.
The substitution of reading for a dialogue which has not occurred is
so manifest that when we happen to encounter an author and to speak
to him (about his book, for example), we experience a profound disrup-
tion of the peculiar relation that we have with the author in and through
his work. Sometimes I like to say that to read a book is to consider its
author as already dead and the book as posthumous. For it is when the
author is dead that the relation to the book becomes complete and, as it
were, intact. The author can no longer respond; it only remains to read
his work.
The difference between the act of reading and the act of dialogue
confirms our hypothesis that writing is a realisation comparable and
parallel to speech, a realisation which takes the place of it as
were, intercepts it. Hence we could say that what comes to wrItmg IS
discourse as intention-to-say and that writing is a direct inscription of
this intention, even if, historically and psychologically, writing began
with the graphic transcription of the signs of speech. This emancipation
of writing, which places the latter at the site of speech, is the birth of
the text.
Now, what happens to the statement itself when it is directly in-
scribed instead of being pronounced? The most striking characteristic
has always been emphasised: writipg preserves discourse and makes it
an archive available for individual and collective memory. It may be
added that the linearisation of symbols permits an analytic and distinc-
tive translation of all the successive and discrete features of language
and thereby increases its efficacy. Is that all? Preservation and increased
efficacy still only characterise the transcription of oral language in
graphic signs. The emancipation of the text from the oral situation en-
tails a veritable upheaval in the relations between language and the
world, as well as in the relation between language and the various sub-
jectivities concerned (that of the author and that of the reader). We
glimpsed something of this second upheaval in distinguishing reading
from dialogue; we shall have to go still further, but this time beginning
from the upheaval which the referential relation of language to the
world undergoes when the text takes the place of speech.
What do we understand by the referential relation or referential func-
tion? In addressing himself to another speaker, the subject of discourse
148 The theory of interpretation
says .about something; that about which he speaks is the
referent of hIS dIscourse. As is well known, this referential function is
by. the sentence, which is the first and the simplest unit of
It IS the sentence which intends to say something true or
real, at least in declarative discourse. The referential func-
so Important that it compensates, as it were, for another charac-
tenstIc of language, namely the separation of signs from things. By
means of the referential function, language 'pours back into the uni-
(according .to an of Gustave Guillaume) those signs
whIch symbohc functIon, at its birth, divorced from things. All dis-
IS, to some extent, thereby reconnected to the world. For if we
dId not speak of the world, of what should we speak?
When the text takes the place of speech, something important occurs
In interlocutors are present not only to one another, but
to the the surroundings and the circumstantial milieu of dis-
course. It m relation to this circumstantial milieu that discourse is
:ully m:anmgful; the return to reality is ultimately a return to this real-
Ity, whIch be indicated 'around' the speakers, 'around', if we may
say.so, the mstance of discourse itself. Language is, moreover, well
eqUIpped to secure this anchorage. Demonstratives, adverbs of time
personal pronouns, verbal tenses, and in general all the
deichc' and' ostensive' indicators serve to anchor discourse in the cir-
reality.which surrounds the instance of discourse. Thus, in
hvmg speech, the Ideal sense of what is said turns towards the real ref-
erence, towards that 'about which' we speak. At the limit, this real ref-
tends to merge an ostensive deSignation where speech re-
the gesture of pomtmg. Sense fades into reference and the latter
mtothe act of showing. .
This is no longer the case when the text takes the place of speech. The
of towards the act of showing is intercepted, at the
same hme as is by the text. I say intercepted and
not suppressed; It IS m thIS respect that I shall distance myself from
may be called henceforth the ideology of the absolute text. On the
baSIS of the sound remarks which we have just made, this ideology
proceeds, by an unwarranted hypostasis, through a course that is ulti-
mately surreptitious. As we shall see, the text is not without reference'
the task of reading, qua interpretation, will be precisely to fulfil
The suspense which defers the reference merely leaves the
tex:, as were, 'in the air', outside or without a world. In virtue of this
obhteratlOn of the relation to the world, each text is free to enter into
1
I
I
I
!
149
What is a text?
relation with all the other texts which come to take the place of the
circumstantial reality referred to by living speech. This relation of text
to text, within the effacement of the world about which we speak, en-
genders the quasi-world of texts or literature.
Such is the upheaval which affects discourse itself, when the move-
ment of reference towards the act of showing is intercepted by the text.
Words cease to efface themselves in front of things; written words be-
come words for themselves.
The eclipse of the circumstantial world by the quasi-world of texts
can be so complete that, in a civilisation of writing, the world itself is
no longer what can be shown in speaking but is to a kind of
'aura' which written works unfold. Thus we speak of the Greek world
or the Byzantine world. This world can be called 'imaginary', in the
sense that it is represented by writing in lieu of the world presented by
speech; but this imaginary world is itself a creation of literature.
The upheaval in the relation between the text and its world is the key
to the other upheaval of which we have already spoken, that which
affects the relation of the text to the subjectivities of the author and the
reader. We think that we know what the author of a text is because we
derive the notion of the author from that of the speaker. The subject of
speech, according to Benveniste, is what designates itself in saying T.
When the text takes the place of speech, there is no longer a speaker, at
least in the sense of an immediate and direct self-designation of the one
who speaks in the instance of discourse. This proximity of the speaking
subject to his own speech is replaced by a complex relation of the author
to the text, a relation which enables us to say that the author is insti-
tuted by the text, that he stands in the space of meaning traced and
inscribed by writing. The text is the very place where the author ap-
pears. But does the author appear otherwise than as first reader? The
distancing of the text from its author is already a phenomenon of the
first reading which, in one move, poses the whole series of problems
that we are now going to confront concerning the relations between
explanation and interpretation. These relations arise at the time of read-
ing.
II. Explanation or understanding?
As we shall see, the two attitudes which we have initially placed under
the double title of explanation and interpretation will confront one an-
other in the act of reading. This duality is first encountered in the work
150
The theory of interpretation
of For him, these distinctions constituted an alternative
:wherein one term necessarily excluded the other: either you 'explain'
In the n:
ann
:
r
of the natural scientist, or you 'interpret' in the manner
of the hIstonan. This exclusive alternative will provide the point of de-
parture for the discussion which follows. I propose to show that the
concept of the text, such as we have formulated it in the first part of this
essay, demands a renewal of the two notions of explanati'on and' t
t ' " In er-
pre and, In of this renewal, a less contradictory conception
of theIr InterrelatIOn. Let us say straightaway that the discussion will
be oriented towards the search for a strict complementarity
and between explanation and interpretation.
opposition in Dilthey's work is not exactly between expla-
and but between explanation and understand-
Ing, InterpretatIon. beIng a particular province of understanding. We
must begIn from the opposition between explanation and un-
ders;andIng. Now if this opposition is exclusive, it is because, in Dil-
they s work, the two terms deSignate two spheres of reality which they
serve to separate. These two spheres are those of the natural sciences
human sciences. Naturp. is the region of objects offered to sci-
entIfIC a region subsumed since Galileo to the enterprise
and since John Stuart Mill to the canons of induc-
lOgIC. MInd IS the region of psychological individualities, into
:whIch each mental life is capable of transposing itself. Understanding
IS such.a transference into another mental life. To ask whether the hu-
can is thus to ask whether a scientific knowledge of
1.S whether this understanding of the Singular can
In ItS own way, whether it is susceptible of universal valid-
Ity. I?l1they affirmatively, because inner life is given in exter-
nal whIch can be perceived and understood as signs of another
mental hfe: 'Understanding', he says in the famous article on 'The de-
of hermeneutiCS', 'is the process by which we come to know
of mental life through the perceptible signs which manifest
It. !hlS IS the understanding of which interpretation is a particular
the signs of another mental life, we have the 'mani-
fixed in a durable way', the 'human testimonies preserved by
, the ',written monuments'. Interpretation is the art of under-
standIng applIed to such manifestations, to such testimonies, to such
of which writing is the distinctive characteristic. Under-
as the knowledge through signs of another mental life thus
prOVIdes the basis in the pair understanding-interpretation; the'latter
What is a text? 151
element supplies the degree of objectification, in virtue of the fixation
and preservation which writing confers upon signs.
Although this distinction between explanation and understanding
seems clear at first, it becomes increasingly obscure as soon as we ask
ourselves about the conditions of scientificity of interpretation. Expla-
nation has been expelled from the field of the human sciences; but the
conflict reappears at the very heart of the concept of interpretation be-
tween, on the one hand, the intuitive and unverifiable character of the
psychologising concept of understanding to which interpretation is
subordinated, and on the other hand the demand for objectivity which
belongs to the very notion of human science. The splitting of herme-
neutics between its psychologising tendency and its search for a logic
of interpretation ultimately calls into question the relation between un-
derstanding and interpretation. Is not interpretation a species of under-
standing which explodes the genre? Is not the specific difference,
namely fixation by writing, more important here than the feature com-
mon to all signs, that of presenting inner life in an external form? What
is more important: the inclusion of hermeneutics in the sphere of un-
derstanding or its difference therefrom? Schleiermacher, before Dil-
they, had witnessed this internal splitting of the hermeneutical project
and had overcome it through a happy marriage of romantic genius and
philological virtuosity. With Dilthey, the epistemological demands are
more pressing. Several generations separate him from the scholar of
Romanticism, several generations well versed in epistemological reflec-
tion; the contradiction now explodes in full daylight. Listen to Dilthey
commenting upon Schleiermacher: 'The ultimate aim of hermeneutics
is to understand the author better than he understands himself.' So
much for the psychology of understanding. Now for the logic of inter-
pretation: 'The function of hermeneutics is to establish
against the constant intrusion of romantic whim and sceptical subJec-
tivism into the domain of history, the universal validity of interpreta-
tion, upon which all certitude in history rests. '2 Thus hermeneutics ful-
fils the aim of understanding only by extl'icating itself from the
immediacy of understanding others - from, let us say, dialogical val-
ues. Understanding seeks to coincide with the inner life of the author,
to liken itself to him (sich gleichsetzen), to reproduce (nachbilden) the
creative processes which engendered the work. But the signs of this
intention, of this creation, are to be found nowhere else than in what
Schleiermacher called the 'exterior' and 'interior form' of the work, or
again, the 'interconnection' (Zusammenhang) which makes it an organ-
152
The theory of interpretation
ised whole. The last writings of Dilthey ('The construction of the his-
torical world in the human sciences') further aggravated the tension.
On the one hand, the objective side of the work was accentuated under
the influence of Husserl's Logical Investigations (for Husserl, as we
'meaning' of a statement constitutes an 'ideality' which exists
m reality nor in psychic reality: it is a pure unity of
meamng wIthout a real localisation). Hermeneutics similarly proceeds
from the objectification of the creative energies of life in works which
come in between the author and us; it is mental life itself its creative
dynamism, which calls for the mediation by 'meanings< 'values' or
'goals'. Th.e demand thus presses towards an ever greater de-
of understanding itself and perhaps
even of mtrospectIon, If It IS true that memory itself follows the thread
,:hich are not themselves mental phenomena. The exter-
lOnsahon of hfe implies a more indirect and mediate characterisation
of the interpretation of self and others. But it is a self and another
in psy.chological tenns, that interpretation pursues;
hon aIms at a reproduction, a Nachbildung, of lived experiences.
ThIS mtolerable tension, which the later DiIthey bears witness to
leads us to raise two questions which guide the follOWing discussion;
Must we abandon once and for all the reference of interpretation to
understandmg and cease to make the interpretation of written monu-
ments a particular case of understanding the external signs of an inner
But if no longer seeks its norm of intelIigi-
blh.ty m understandmg othprs, then does not its relation to explanation,
WhICh we have set aside hitherto, now demand to be reconsidered?
HI. The text and structural explanation
Let us begin again from our analysis of the text and from the autono-
mous status which we have granted it with respect to speech. What we
have called the eclipse of the surrounding world by the quasi-world of
texts engenders two possibilities. We can, as readers, remain in the
su.spense of the text, treating it as a worldlessand authorless object; in
thIS case, we explain the text in tenns of its internal relations, its struc-
ture. On the other hand, we can lift the suspense and fulfil the text in
speech, restoring it to living communication; in this case, we interpret
the two possibilities both belong to reading, and reading is
the dIalectIc of these two attitudes.
Let us consider them separately, before exploring their articulation.
I
I
What is a text? 153
We can undertake a first type of reading which formally records, as it
were, the text's interception of all the relations to a world that can be
pOinted out and to subjectivities that can converse. This transference
into the 'place' - a place which is a non-place - constitutes a special
project with respect to the text, that of prolonging the suspense con-
cerning the referential relation to the world and to the speaking subject.
By means of this special project, the reader decides to situate himself in
the 'place of the text' and in the 'closure' of this place. On the basis of
this choice, the text has no outside but only an inside; it has no tran-
scendent aim, unlike a speech which is addressed to someone about
something.
This project is not only possible but legitimate. For the constitution
of the text as text and of the body of texts as literature justifies the in-
terception of the double transr.er..der..ce of di::;course, towarrlE the world
and towards someone. Thus arises the possibility of an explanatory at-
titude in regard to the text.
In contrast to what DiIthey thought, this explanatory attitude is not
borrowed from a field of knowledge and an epistemological model other
than that of language itself. It is not a naturalistic model subsequently
extended to the human sciences. The nature-mind opposition plays no
role here at all. If there is some fonn of borrowing, it occurs within the
same field, that of signs. For it is possible to treat the text according to
the explanatory rules that linguistics successfully applies to the simple
system of signs which constitute language [langue] as opposed to speech
[parole]. As is well known, the language-speech distinction is the fun-
damental distinction which gives linguistics an homogenous object;
speech belongs to physiology, psychology and sociology, whereas lan-
guage, as rules of the game of which speech is the execution, belongs
only to linguistics. As is equally well known, linguistics considers only
systems of units devoid of proper meaning, each of which is defined
only in terms of its difference from all of the others. These units,
whether they be purely distinctive like those of phonological articula-
tion or significant like those of lexical articulation, are oppositive units.
The interplay of oppositions and their combinations within an inven-
tory of discrete units is what defines the notion of structure in linguis-
tics. This structural model furnishes the type of explanatory attitude
which we are now going to see applied to the text.
Even before embarking upon this enterprise, it may be objected that
the laws which are valid only for language as distinct from speech could
not be applied to the text. Although the text is not speech, is it not, as
154 The theory of interpretation
it were, on the same side as speech in relation to language? Must not
discourse, as a series of statements and ultimately of sentences, be op-
posed in an overall way to language? In comparison to the language-
discourse distinction, is not the speaking-writing distinction second-
ary, such that speaking and writing occur together on the side of
discourse? These remarks are perfectly legitimate and justify us in
thinking that the structural model of explanation does not exhaust the
field of possible attitudes which may be adopted in regard to a text. But
before specifying the limits of this explanatory model, it is necessary to
grasp its fruitfulness. The working hypothesis of any structural analysis
of texts is this: in spite of the fact that writing is on the same side as
speech in relation to language - namely, on the side of discourse - the
specificity of writing in relation to speech is based on structural features
which can be treated as analogues of language in discourse. This work-
ing hypothesis is perfectly legitimate; it amounts to saying that under
certain conditions the larger units of language [langage], that is, the
units of a higher order than the sentence, display organisations com-
parable to those of the smaller units of language, that is, the units which
are of a lower order than the sentence and which belong to the domain
of linguistics.
In Structural Anthropology, Claude Levi-Strauss formulates this work-
ing hypothesis f9r one category of texts, the category of myths:
Like every linguistic entity, myth is made up of constitutive units. These units
imply the presence of those which normally enter into the structure of language,
namely the phonemes, the morphemes and the semantemes. The constituent
units of myth are in the same relation to semantemes as the latter are to mor-
phemes, and as the latter in turn are to phonemes. Each form differs from that
which precedes it by a higher degree of complexity. For this reason, we shall
call the elements which properly pertain to myth (and which are the most com-
plex of all): large constitutive units.
3
By means of this working hypothesis, the large units which are mini-
mally the size of the sentence, and which placed together constitute the
narrative proper to the myth, can be treated according to the same rules
that are applied to the smaller units familiar to linguistics. To indicate
this analogy, Levi-Strauss speaks of 'my themes' in the same way that
one speaks of phonemes, morphemes and semantemes. But in order to
remain within the limits of the analogy between my themes and the
linguistic units of a lower level, the analysis of texts will have to proceed
to the same sort of abstraction as that practised by the phonologist. For
the latter, the phoneme is not a concrete sound, to be taken absolutely
in its sonorous substance; it is a function defined by the commutative
What is a text?
155
method and its oppositive value is determi'led by the relation to all
other phonemes. In this sense it is not, as Saussure would say, a ' s u ~
stance' but a 'form', an interplay of relations. Similarly, a my theme 1S
not one of the sentences of the myth but an oppositive value which is
shared by several particular sentences, constituting, in the language of
Levi-Strauss, a 'bundle of relations'. 'Only in the form of combinations
of such bundles do the constituent units acquire a signifying function.'4
What is called here the 'signifying function' is not at all what the myth
means, its philosophical or existential import, but rather the arrange-
ment or disposition of my themes, in short, the structure of the myth.
I should like to recall briefly the analysis which, according to this
method, Levi-Strauss offers of the Oedipus myth. He divides the sen-
tences of the myth into four columns. In the first column he places all
the sentences which speak of overrated blood relations (for example,
Oedipus marries Jocasta, his mother; Antigone buries Polynices, her
brother, in spite of the order forbidding it). In the second column, we
find the same relation, but modified by the inverse sign: underrated or
devalued blood relations (Oedipus kills his father, Laios; Eteocles kills
his brother, Polynices). The third column concerns monsters and their
destruction; the fourth groups together all those proper names whose
meaning suggests a difficulty in walking straight (lame, clumsy, swol-
len foot). The comparison of the four columns reveals a correlation. Be-
tween the first and second columns we have blood relations overrated
or underrated in turn; between the third and fourth we have an affir-
mation and then a negation of the autochtony of man. 'It follows that
the fourth column is related to the third column as the first is to the
second ... ; the overrating of blood relations is to their underrating as
the attempt to escape from autochtony is to the impossibility of suc-
ceeding in it.' The myth thus appears as a kind of logical instrument
which brings together contradictions in order to overcome them: 'the
impossibility of connecting the groups of relations is overcome (or,
more exactly, replaced) by the assertion that two contradictory relations
are identical, insofar as each is, like the other, self-contradictory'. 5 We
shall return shortly to this conclusion; let us restrict ourselves here to
stating it.
We can indeed say that we have thereby explained the myth, but not
that we have interpreted it. We have brought out, by means of struc-
tural analysis, the logic of the operations which interconnect the pack-
ets of relations; this logic constitutes 'the structural law of the myth
concerned'.6 We shall not fail to notice that this law is, par excellence,
156
The theory of interpretation
the object of reading and not at all of speech, in the sense of a recitation
whereby the power of the myth would be reactivated in a particular
situation. Here the text is only a text and the reading inhabits it only as
such, while its meaning for us remains in suspense, together with any
realisation in present speech.
I have just taken an example from the domain of myths; I could take
another from a nearby domain, that of folklore. This domain has been
explored by the Russian formalists of the school of Propp and by the
French specialists in the structural analysis of narratives, Roland
Barthes and A.J. Greimas. In the work of these authors, we find the
same postulates as those employed by Levi-Strauss: the units above the
sentence have the same composition as the units below the sentence;
the sense of the narrative consists in the very arrangement of the ele-
ments, in the power of the whole to integrate the sub-units; and con-
versely, the sense of an element is its capacity to enter in relation with
other elements and with the whole of the work. TheSf! postulates to-
gether define the closure of the narrative. The task of structural analysis
will be to carry out the segmentation of the work (horizontal aspect),
then to establish the various levels of integration of the parts in the
whole (hierarchical aspect). Thus the units of action isolated by the an-
alyst will not be psychological units capable of being experienced, nor
will they be units of behaviour which could be subsumed to a behav-
iourist psychology. The extremities of these sequences are only the
switching points of the narrative, such that if one element is changed,
all the rest is different. Here we recognise the transposition of the
method of commutation from the phonological level to the level of nar-
rative units. The logic of action thus consists in an interconnected series
of action kernels which together constitute the structural continuity of
the narrative. The application of this technique ends up by 'dechro-
nologising' the narrative, in a way that brings out the logic underlying
narrative time. Ultimately the narrative would be reduced to a combi-
nation [combinatoire] of a few dramatic units (promising, betraying,
hindering, aiding, etc.) which would be the paradigms of action. A
sequence is thus a succession of nodes of action, each closing off an
alternative opened up by the preceding one. Just as the elementary
units are linked together, so too they fit into larger units; for example,
an encounter is comprised of elementary actions like approaching, call-
ing out, greeting, etc. To explain a narrative is to grasp this entangle-
ment, this fleeting structure of interlaced actions.
Corresponding to the nexus of actions are relations of a similar nature
What is a text? 157
between the 'actants' of the narrative. By that we understand, not at all
the characters as psychological subjects endowed with their own exis-
tence, but rather the roles correlated with formalised actions. Actants
are defined entirely by the predicates of action, by the semantic axes of
the sentence and the narrative: the actant is the one by whom, to
whom, with whom, ... the action is done; it is the one who promises,
who receives the promise, the giver, the receiver, etc. Structural analy-
sis thus brings out a hierarchy of actants correlative to the hierarchy of
actions.
The narrative remains to be assembled as a whole and put back into
narrative communication. It is then a discourse which a narrator ad-
dresses to an audience. For structural analysis, however, the two inter-
locutors must be sought only in the text. The narrator is deSignated by
the signs of narrativity, which belong to the very constitution of the
narrative. Beyond the three levels of actions, actants and narration,
there is nothing else that falls within the scope of the science of semiol-
ogy. There is only the world of narrative users, which can eventually be
dealt with by other semiological disciplines (those analysing social,
economic and ideological systems); but these disciplines are no longer
linguistic in nature. This transposition of a linguistic model to the the-
ory of the narrative fully confirms our initial remark: today, explanation
is no longer a. concept borrowed from the natural sciences and trans-
ferred to the alien domain of written artefacts; rather, it stems from the
very sphere of language, by analogical transference from the small units
of language (phonemes andlexemes) to the units larger than the sen-
tence, such as narratives, folklore and myth. Henceforth, interpreta-
tion - if it is still possible to give a sense to this notion - will no longer
be confronted by a model external to the human sciences. It will, in-
stead, be confronted by a model of intelligibility which belongs, from
birth so to speak, to the domain of the human sciences, and indeed to
a leading science in this domain: linguistics. Thus it will be upon the
same terrain, within the same sphere of language [/angage] , that expla-
nation and interpretation will enter into debate.
IV. Towards a new concept of interpretation
Let us consider now the other attitude that can be adopted in regard to
the text, the attitude which we have called interpretation. We can intro-
duce this attitude by initially opposing it to the preceding one, in a
manner still close to that of Dilthey. But as we shall see, it will be nec-
158
The theory of interpretation
to proceed gradually to a more complementary and reciprocal
relation between explanation and interpretation.
Let us begin once again from reading. Two ways of reading, we said,
are .offered to us. By reading we can prolong and reinforce the suspense
affects the text's reference to a surrounding world and to the
audIence .of speaking subjects: that is the explanatory attitude. But we
can also hft the suspense and fulfil the text in present speech. It is this
second attitude which is the real aim of reading. For this attitude re-
veals the true nature of the suspense which intercepts the movement of
the towards The other attitude would not even be possi-
ble If It were not fIrst apparent that the text, as writing, awaits and calls
for a reading. If reading is possible, it is indeed because the text is not
closed in .on itself opens out onto other things. To read is, on any
hypothesIs, to conjOIn a new discourse to the discourse of the text. This
conjunction of discourses reveals, in the very constitution of the text
a.n capacity for renewal which is its open character.
hon IS the concrete outcome of conjunction and renewal.
In the first instance, we shall be led to formulate the concept of inter-
pretation in opposition to that of explanation. This will not distance us
from DUthey's position, except that the opposing concept
of explanatIon has already gained strength by being derived from lin-
guistics and semiology rather than being borrowed from the natural
sciences.
to. this first sense, interpretation retains the feature of ap-
propnatIon whIch was recognised by Schleiermacher, Dilthey and Bult-
In fact, this sense will not be abandoned; it will only be me-
dIated by explanation, instead of being opposed to it in an immediate
and even naive way. By 'appropriation', I understand this: that the
interpretation of a text culminates in the self-interpretation of a subject
who thenceforth understands himself better, understands himself dif-
ferently, or simply begins to understand himself. This culmination of
the understanding of a text in self-understanding is characteristic of the
kind of reflective philosophy which, on various occasions, I have called
'concrete reflection'. Here hermeneutics and reflective philosophy are
correlative and reciprocal. On the one hand, self-understanding passes
through the detour of understanding the cultural signs in which the self
and forms itself. On the other hand, understanding the text
IS not .an end in itse!f; it mediates the relation to himself of a subject
who, In the short CIrcuit of immediate reflection, does not find the
meaning of his own life. Thus it must be said, with equal force, that
What is a text? 159
reflection is nothing the mediation of signs and works, and
that explanation is nothing if it is not incorporated as an intermediary
stage in the process of self-understanding. In short, in hermeneutical
reflection - or in reflective hermeneutics - the constitution of the self is
contemporaneous with the constitution of meaning.
The term 'appropriation' underlines two additional features. One of
the aims of all hermeneutics is to struggle against cultural distance. This
struggle can be understood in purely temporal terms as a struggle
against secular estrangement, or in more genuinely hermeneutical
terms as a struggle against the estrangement from meaning itself, that
is, from the system of values upon which the is based. In this sense,
interpretation 'brings together', 'equalises', renders 'contemporary and
similar', thus genuinely making one's own what was initially alien.
Above all, the characterisation of interpretation as appropriation is
meant to underline the 'present' character of interpretation. Reading is
like the execution of a musical score; it marks the realisation, the enact-
ment, of the semantic possibilities of the text. This final feature is the
most important because it is the condition of the other two (that is, of
overcoming cultural distance and of fusing textual interpretation with
self-interpretation). Indeed, the feature of realisation discloses a deci-
sive aspect of reading, namely that it fulfils the discourse of the text in
a dimension similar to that of speech. What is retained here from the
notion of speech is not the fact that it is uttered but that it is an event,
an instance of discourse, as Benveniste says. The sentences of a text
signify here and now. The 'a<:tualised' text finds a surrounding and an
audience; it resumes the referential movement - intercepted and sus-
pended - towards a world and towards subjects. This world is that of
the reader, this subject is the reader himself. In interpretation, we shall
say, reading becomes like speech. I do not say 'becomes speech', for
reading is never equivalent to a spoken exchange, a dialogue. But read-
ing culminates in a concrete act which is related to the text as speech is
related to discourse, namely as event and instance of discourse. Initially
the text had only a sense, that is, internal relations or a structure; now
it has a meaning, that is, a realisation in the discourse of the reading
subject. By virtue of its sense, the text had only a semiological dimen-
sion; now it has, by virtue of its meaning, a semantic dimension.
Let us pause here. Our discussion has reached a critical point where
interpretation, understood as appropriation, still remains external to
explanation in the sense of structural analysis. We continue to oppose
them as if they were two attitudes between which it is necessary to
160
The theory of interpretation
choose. I should like now to go beyond this antithetical opposition and
to bring out the articulation which would render structural analysis and
hermeneutics complementary. For this it is important to show how each
of the two attitudes which we have juxtaposed refers back, by means of
its own peculiar features, to the other.
Consider again the examples of structural analysis which we have
borrowed from the theory of myth and narrative. We tried to adhere to
a notion of sense which would be strictly equivalent to the arrangement
of the elements of a text, to the integration of the segments of action
and the actants within the narrative treated as a whole closed in upon
itself. In fact, no one stops at so formal a conception of sense. For ex-
ample, what Levi-Strauss calls a 'my theme' - in his eyes, the constitu-
tive unit of myth - is expressed in a sentence which has a specific
meaning: Oedipus kills his father, Oedipus marries his mother, etc.
Can it be said that structural explanation neutralises the specific mean-
ing of sentences, retaining only their position in the myth? But the bun-
dle of relations to which Levi-Strauss reduces the my theme is still of
the order of the sentence; and the interplay of oppositions which is
instituted at this very abstract level is equally of the order of the sen-
tence and of meaning. If one speaks of 'overrated' or 'underrated blood
relations', of the 'autochtony' or 'non-autochtony' of man, these rela-
tions can still be written in the form of a sentence: the blood relation is
the highest of all, or the blood relation is not as high as the social rela-
tion, for example in the prohibition of incest, etc. Finally, the contra-
diction which the myth to resolve, according to
is itself stated in terms of meaningful relations. Levi-Strauss admits
this, in spite of himself, when he writes: 'The reason for these choices
becomes clear if we recognise that mythical thought proceeds from the
consciousness of certain oppositions and tends towards their progres-
sive mediation';7 and again, 'the myth is a kind of logical tool intended
to effect a mediation between life and death'.8 In the background of the
myth there is a question which is highly significant, a question about
life and death: 'Are we born from one or from two?' Even in its formal-
ised version, 'Is the same born from the same Or from the other?', this
question expresses the anguish of origins: whence comes man? Is he
born from the earth or from his parents? There would be no contradic-
tion, nor any attempt to resolve contradiction, if there were not signif-
icant questions, meaningful propositions about the origin and the end
of man. It is this function of myth as a narrative of origins that structural
analysis seeks to place in parentheses. But such analysis does not suc-
What is a text? 161
ceed in eluding this function: it merely postpones it. Myth is not a log-
ical operator between any propositions whatsoever, but involves prop-
ositions which point towards limit situations, towards the origin and
the end, towards death, suffering and sexuality.
Far from dissolving this radical questioning, structural analysis rein-
states it at a more radical level. Would not the function of structural
analysis then be to impugn the surface semantics of the recounted myth
in order to unveil a depth semantics which is, if I may say so, the living
semantics of the myth? If that were not the function of structural anal-
ysis, then it would, in my opinion, be reduced to a sterile game, to a
derisory combination [combinatoire] of elements, and myth would be
deprived of the function which Levi-Strauss himself recognises when
he asserts that mythical thought arises from the awareness of certain
oppositions and tends towards their progressive mediation. This
awareness is a recognition of the aporias of human existence around
which mythical thought gravitates. To eliminate this meaningful inten-
tion would be to reduce the theory of myth to a necrology of the mean-
ingless discourses of mankind. If, on the contrary, we regard structural
analysis as a stage - and a necessary one - between a naive and a critical
interpretation, between a surface and a depth interpretation, then it
seems possible to situate explanation and interpretation along a unique
hermeneutical arc and to integrate the opposed attitudes of explana-
tion and understanding within an overall conception of reading as the
recovery of meaning.
We shall take another step in the direction of this reconciliation be-
tween explanation and interpretation if we now turn towards the sec-
ond term of the initial contradiction. So far we have worked with a
concept of interpretation which remains very subjective. To interpret,
we said, is to appropriate here and now the intention of the text. In
saying that, we remain enclosed within Dilthey's concept of under-
standing. Now what we have just said about the depth semantics un-
veiled by the structural analysis of the text invites us to say that the
intended meaning of the text is not essentially the presumed intention
of the author, the lived experience of the writer, but rather what the
text means for whoever complies with its injunction. The text seeks to
place us in its meaning, that is - according to another acceptation of the
word sens - in the same direction. So if the intention is that of the text,
and if this intention is the direction which it opens up for then
depth semantics must be understood in a fundamentally dynamiC way.
I shall therefore say: to explain is to bring out the structure, that is, the
162 The theory of interpretation
internal relations of dependence which constitute the statics of the text;
to interpret is to follow the path of thought opened up by the text, to
place oneself en route towards the orient of the text. We are invited by
this remark to correct our initial concept of interpretation and to
search - beyond a subjective process of interpretation as an act on the
text - for an objective process of interpretation which would be the act
of the text.
I shall borrow an example from a recent study which I made of the
exegesis of the sacerdotal story of creation in Genesis 1-2, 4a.
9
This
exegesis reveals, in the interior of the text, the interplay of two narra-
tives: a Tatbericht in which creation is expressed as a narrative of action
('God made ... '), and a Wortbericht, that is, a narrative of speech ('God
said, and there was .. .'). The first narrative could be said to play the
role of tradition and the second of interpretation. What is interesting
here is that interpretation, before being the act of the exegete, is the act
of the text. The relation between tradition and interpretation is a rela-
tion internal to the text; for the exegete, to interpret is to place himself
in the meaning indicated by the relation of interpretation which the
text itself supports.
This objective and, as it were, intra-textual concept of interpretation
is by no means unusual. Indeed, it has a long history rivalling that of
the concept of subjective interpretation which is linked, it will be re-
called, to the problem of understanding others through the signs that
others give of their conscious life. I would willingly connect this new
concept of interpretation to that referred to in the title of Aristotle's
treatise On Interpretation. Aristotle's hermenetia, in contrast to the her-
meneutical technique of seers and oracles, is the very action of language
on things. Interpretation, for Aristotle, is not what one does in a second
language with regard to a first; rather, it is what the first language al-
ready does, by mediating through signs our relation to things. Hence
interpretation is, according to the commentary of Boethius, the work of
the vox significativa per se ips am aliquid significans, sive complexa, sive
incomplexa. Thus it is the noun, the verb, discourse in general, which
interprets in the very process of signifying.
It is true that interpretation in Aristotle's sense does not exactly pre-
pare the way for understanding the dynamic relation between several
layers of meaning in the same text. For it presupposes a theory of
speech and not a theory of the text: 'The sounds articulated by the voice
are symbols of states of the soul, and written words are symbols of
words uttered in speech' (On Interpretation, para. 1). Hence interpreta-
What is a text? 163
tion is confused with the semantic dimension of speech: interpretation
is discourse itself, it is any discourse. Nevertheless, I retain from Aris-
totle the idea that interpretation is interpretation by language before
being interpretation of language.
I would look in the work of Charles Sanders Peirce for a concept of
interpretation which is closer to that required by an exegesis which
relates interpretation to tradition in the very interior of a text. Accord-
ing to Peirce, the relation of a 'sign' to an 'object' is such that another
relation, that between 'interpretant' and 'sign', can be grafted onto the
first. What is important for us is that this relation between interpretant
and sign is an open relation, in the sense that there is always another
interpretant capable of mediating the first relation. G.-G. Granger ex-
plains this very well in his Essai d'une philosophie du style:
The interpretant which the sign evokes in the mind could not be the result of a
pure and simple deduction which would extract from the sign something al-
ready contained therein ... The interpretant is a commentary, a definition, a
gloss on the sign in its relation to the object. The interpretant is itself symbolic
expression. The sign-interpretant association, realised by whatever psycholog-
ical processes, is rendered possible only by the community, more or less im-
perfect, of an experience between speaker and hearer ... It is always an expe-
rience which can never be perfectly reduced to the idea or object of the sign of
which, as we said, it is the structure. Whence the indefinite character of Peirce's
series of interpretants.
1O
We must, of course, exercise a great deal of care in applying Peirce's
concept of interpretant to the interpretation of texts. His interpretant is
an interpretant of signs, whereas our interpretant is an interpretant of
statements. But our use of the interpretant, transposed from small to
large units, is neither more nor less analogical than the structuralist
transfer of the laws of organisation from units of levels below the sen-
tence to units of an order above or equal to the sentence. In the case of
structuralism, it is the phonological structure of language which serves
as the coding model of structures of higher articulation. In our case, it
is a feature of lexical units which is transposed onto the plane of state-
ments and texts. So if we are perfectly aware of the analogical character
of the transposition, then we can say that the open series of interpre-
tants, which is grafted onto the relation of a sign to an object, brings to
light a triangular relation of object-sign-interpretant; and that the latter
relation can serve as a model for another triangle which is constituted
at the level of the text. In the new triangle, the object is the text itself;
the sign is the depth semantics disclosed by structural analysis; and the
164 The theory of interpretation
series of interpretants is the chain of interpretations produced by the
interpreting community and incorporated into the dynamics of the text,
as the work of meaning upon itself. Within this chain, the first inter-
pretants serl'e as tradition for the final interpretants, which are the
interpretation in the true sense of the term.
Thus informed by the Aristotelian concept of interpretation and
above all by Peirce's concept, we are in a position to 'depsychologise'
as far as possible our notion of interpretation, and to connect it with
the process which is at work in the text. Henceforth, for the exegete, to
interpret is to place himself within the sense indicated by the relation
of interpretation supported by the text.
The idea of interpretation as appropriation is not, for all that, elimi-
nated; it is simply postponed until the termination of the process. It lies
at the extremity of what we called above the hermeneutical arc: it is the
final brace of the bridge, the anchorage of the arch in the ground of
lived experience. But the entire theory of hermeneutics consists in me-
diating this interpretation-appropriation by the series of interpretants
which belong to the work of the text upon itself. Appropriation loses its
arbitrariness insofar as it is the recovery of that which is at work, in
labour, within the text. What the interpreter says is a re-saying which
reactivates what is said by the text.
At the end of our investigation, it seems that reading is the concrete
act in which the destiny of the text is fulfilled. It is at the very heart of
reading that explanation and interpretation are indefinitely opposed
and reconciled.
6. Metaphor and the central problem
of hermeneutics
It will be assumed here that the central problem of hermeneutics is that
of interpretation. Not interpretation in any sense of the word, but inter-
pretation determined in two ways: the first concerning its field of ap-
plication, the second its epistemological specificity. As regards the first
point, I shall say that there is a problem of interpretation because there
are texts, written texts, the autonomy of which creates specific difficul-
ties. By 'autonomy' I understand the independence of the text with re-
spect to the intention of the author, the situation of the work and the
original reader. The relevant problems are resolved in oral discourse by
the kind of exchange or intercourse which we call dialogue or conver-
sation. With written texts, discourse must speak by itself. Let us say,
therefore, that there are problems of interpretation because the writ-
ing-reading relation is not a particular case of the speaking-hearing
relation which we experience in the dialogical situation. Such is the
most general feature of interpretation as regards its field of application.
Second, the concept of interpretation seems, at the epistemological
level, to be opposed to the concept of explanation. Taken together,
these concepts form a contrasting pair which has given rise to a great
many disputes since the time of Schleiermacher and Dilthey. According
to the tradition to which the latter authors belong, interpretation has
certain subjective connotations, such as the implication of the reader in
the processes of understanding and the reciprocity between interpre-
tation of the text and self-interpretation. This reciprocity is known by
the name of the hermeneutical circle; it entails a sharp opposition to the
sort of objectivity and non-implication which is supposed to character-
ise the scientific explanation of things. Later I shall say to what extent
we may be able to amend, indeed to reconstruct on a new basis, the
opposition between interpretation and explanation. Whatever the out-
come of the subsequent discussion may be, this schematic description
of the concept of interpretation suffices for a provisional circumscrip-
tion of the central problem of hermeneutics: the status of written texts
165