Anthropological Theory
Copyright © 2002 SAGE Publications
(London,Thousand Oaks, CA
and New Delhi)
Vol 2(4): 433–446
[1463-4996(200212)2:4;433–446;029509]
433
A confusion of tongues
Vincent Descombes
Ecole des Hautes Etudes en Sciences Sociales
Abstract
When we maintain that an anthropologist offers ‘thick descriptions’ of the life of
people, do we mean that it is in any way descriptive? Or, instead, that it is purely
interpretative, like the creative reading of a text by a literary critic? According to
Gilbert Ryle, who invented the term, a ‘thick description’ is just a ‘thin description’
made complex by the addition of adverbial information. According to Clifford Geertz,
the anthropologist has no access to ‘thin descriptions’, he deals with conflicting views
and interpretations given by the participants of ‘social discourses’ and has to address a
permanent ‘confusion of tongues’. Contemporary use of this term appears to rest
upon a hidden conflict between two philosophies of anthropology.
Key Words
fieldwork • Clifford Geertz • hermeneutics • Gilbert Ryle • (thick/thin) description
Numerous anthropologists have rallied to the idea that their fieldwork consisted of doing
what Clifford Geertz calls ‘thick description’ (Geertz, 1973). Their discipline is no more
a science of pure observation than that of history or sociology. If the ethnographer wishes
to convince himself of this, he need only take into consideration his own practical experience.
‘What will I record in my field notebook?’ this ethnographer must ask himself. If I
write, ‘In the middle of the city there is an edifice taller than the others and guarded by
armed men’, I am making use of a rather weak vocabulary, but I am not thereby more
objective because this ‘thin’ description is not what I will be using in order to explain the
lifestyle of the society being studied. In fact, one must record in his notebook a ‘thick’
description of the following genre: ‘In the middle of the capital is found the king’s palace.’
The success of the notion of thick description shows that the anthropological disciplines
which practice inquiry in the field needed a term of this genre to describe their
work. The problem that arises for the investigator in the field is, fundamentally, knowing
how to take into account, in his descriptive work, the difference between nature and
convention, between phusis and nomas. As Aristotle said: fire burns in the same way here
(in Greece) and in Persia, but rights are variable. (Nicomean Ethics, 1134b). As long as
one is content with describing the processes of fire, one has no need to distinguish two
types of description (the thin and the thick). On the other hand, a certain way of
dressing, of walking, of covering one’s head will be regarded here as neutral behavior,
06 ANT 2-4 Descombes (JB/D) 5/11/02 9:51 am Page 433
elsewhere as a punishable offense: the description must be ‘thick’, that is to say, it must
identify the type of conduct from the point of view of the meaning that it possesses in
the context in which the behavior takes place.
However, the fact that the concept of thick description has been well received does
not mean that its status is clear. Is a thick description as descriptive as a thin description?
Is it, like all description, a test of truth in confrontation with reality? In fact, one
generally considered that to adopt this distinction between the thin and the thick was
to accept a revision of the status of the sciences which have to do with convention
(nomos) or meaning. This status would be hermeneutical. In other words, unlike physics
(and all the natural sciences), the anthropological discipline (and all the social sciences)
could not avoid having to pose the ‘hermeneutical problem’.1
THE ‘HERMENEUTICAL’ NOTION OF ANTHROPOLOGICAL INQUIRY
The descriptive science of fire might be satisfied with rendering thin descriptions because
it does not have to take a context into account. The manner in which fire burns, whether
in Greece or in Persia, does not in any way depend on what men here and there think
of it. On the other hand, a knowledge of mores will note that the rights of the Greeks
are not those of the Persians. Law currently in force is a function of the opinion of men
concerning right and wrong. The facts with which the knowledge of mores concerns
itself cannot be detached from what the players think about them. Since these facts
cannot be ‘objectified’, the science of mores is a ‘hermeneutical’ science. The ‘facts of
observation’ do not reveal their significance in and of themselves: the investigator must
make them speak, must interpret them, just as the critic does who decides to read his
text in such and such a context. Anthropology has a hermeneutical status, which simply
means that the cognitive processes asked of a researcher studying a custom or a social
form of life are strictly identical to those of an interpreter, that is to say, not simply a
translator, but rather a literary critic writing an essay on King Oedipus or on an idea such
as A roll of the dice will never abolish chance.
Continuing this reasoning, one will say that all the researcher’s notes, if they have the
slightest anthropological import, are interpretations on his part. But one then sees what
has happened. We started from a distinction between thin descriptions and thick
descriptions depending on whether the vocabulary in which they are given is poor or
rich in cultural terms. However, we finally end up at the idea that there is very little of
description – and perhaps none at all – in the ‘material’ that the investigator will be
seeking in the field.
It is impossible not to wonder here whether the notion of thick description is not
utilized as an equivalent (less openly contradictory) of ‘subjective description’. In the
combination ‘thick description’, the adjective could well contradict the noun, but
without openly admitting to the conflict of the descriptive ideal and the interpretive
means. In fact, the anthropologist will find it difficult to give up the idea that his work
is of a descriptive nature. Why would there be inquiry on site, study in the ‘field’, if not
to provide an empirical basis for the research? Furthermore, the ethnographer prides
himself in practicing a universalist science: he does not go in the field to deliver
judgments or to reform, but precisely to learn how people live and what they say about
themselves. He is driven by all this to introduce himself as the practitioner of a descriptive
discipline.
ANTHROPOLOGICALTHEORY 2(4)
434
06 ANT 2-4 Descombes (JB/D) 5/11/02 9:51 am Page 434
The adjective ‘thick’ comes in to temper a bit this ambition of scientific objectivity.
To say that ethnographers do not practice observation, as do their naturalist colleagues,
but rather thick description, is first of all (at least at first reading) simply a way of recognizing
that anthropological description takes place under special conditions. One
imagines that he has perhaps retained the descriptive status of the inquiry, all the while
doing his part in the interpretive aspect of the process.
But in what way are the conditions of the analysis special? Many anthropologists today
appear ready to say that description is special in that it can never limit itself to facts.
Such is precisely the direction chosen by Geertz. In the final analysis, he explains,
description is necessarily a construction, the fruit of the active imagination of the
researcher (Geertz, 1973: 15). What remains of the descriptive if one grants an epistemological
immunity to the version of facts proposed by the ethnographer? Nothing at all!
The one who says description says factual truth and the possibility (theoretically) of an
empirical truth. On the other hand, a hermeneutical discipline comes under other
criteria of judgment.
THICK DESCRIPTION ANDTHIN DESCRIPTION
In the chapter which serves as a manifesto and which opens his work on The Interpretation
of Cultures, Geertz proceeds with the fusion of two themes borrowed from very
different traditions. The first is that of a hermeneutical status of the humanities (he
returns to the idea, put forward by Ricoeur, according to which the art of reading a
classical text could serve as a model for human knowledge). The second is that of a
contrast between the ‘thin’ description or the purely factual nature of an action and its
‘thick’ description or that enriched by contextual elements (he borrows this distinction
from Gilbert Ryle). Geertz, in this chapter, sets out to defend two theses simultaneously
and to support one by the other: the ethnographer undertakes thick description (Geertz,
1973: 9–10) and never leaves the sphere of interpretation, because what he finds in his
field is formed by interpretation; that which he brings with him to the field are models
or diagrams for the interpretation (Geertz, 1973: 28).
How does the merging of the two themes take place? We will turn first to the way in
which Ryle introduces, in his own words, the idea of thick description. Ryle appeals to
the contrast between two types of description in articles devoted to the activity of
thought (see ‘Thinking and Reflecting’ and ‘The Thinking of Thoughts: What is Le
Penseur Doing?’ Ryle, 1971). How will the author of The Concept of Mind, known for
having taken on the ‘ghost in the machine’, characterize the activity of thinking?
In a style typically Rylian, the articles invite us to consider the character of Rodin’s
Thinker and to ask ourselves how we would describe his activity. The answer will be
sought in a difference between two meanings of the verb ‘to think’: the thought of
someone who is thinking about what he is doing, as, for example, a tennis player who
is concentrating on his game, and the thought of someone who does nothing but reflect
and who, seen from the outside, might seem altogether distracted. So Ryle is in the
process of seeking a form of expression that a phenomenologist would render as follows:
the active man is a thinker in so far as he is present at his task; the man who reflects is
thoughtful, he is detached from immediate circumstances and thus capable of being
present at other tasks or for other problems or beings other than those which give
themselves to him in the form of a close or urgent presence. In order to arrive at the
DESCOMBES A confusion of tongues
435
06 ANT 2-4 Descombes (JB/D) 5/11/02 9:51 am Page 435
distinction he is seeking, Ryle is trying to make clear the idea that, in one case (that of
the tennis player), thinking functions in the way in which a physical activity is performed
(in distracted or attentive fashion), while in the other case (that of Rodin’s Thinker),
thinking is the principal activity (an activity which itself can be described in different
ways: concentrated, intense, prolonged, and so on).
What interests us here is not the attempt to give a non-Cartesian definition of
thinking, but the distinction for which thinking is the occasion: that between ‘thin’
description and ‘thick’ description. As we are going to see, this latter distinction between
two types of description rests on the contrast which has just been made between a subject
whose activity thinking describes (the tennis player) and a subject for whom thinking is
the principal activity (the Thinker). The difference which Ryle seeks to bring out is,
therefore, the metaphysical difference between an action and its mode of execution. Or
again, if one prefers, the grammatical difference between a main verb and its adverbs.
The adjectives ‘thin’ and ‘thick’ may be understood by analogy with the composition
of a sandwich (Ryle, 1971: 482): certain sandwiches, closer to the ‘spread’ of a social
buffet than to a ‘snack’, offer only a fine layer of the foods which they hold together;
others are built by the piling up of various layers of food, construction that can be so
thick that it requires the use of knife and fork for whoever wishes to eat it. Thus it is
that the North American ‘club sandwich’ is composed of three slices of soft bread
between which one places two layers of meat (chicken or turkey), while between the
bread and the meat one slips in lettuce, tomato and mayonnaise.
The difference between the two types of descriptions is introduced by way of the
example of two boys: both wink, but the first one blinks because he is afflicted with a
twitch, whereas the second one blinks in order to wink at his companion. There is, writes
Ryle, a mode of description of these movements, which does not distinguish between
the two: you can tell that the eyelid contracts without being able to tell whether it’s an
involuntary movement or a signal. At this level, disregarding the distinction between the
involuntary and the voluntary makes for a description so ‘thin’ that one cannot tell which
is a twitch and which is sending a signal.
What is the boy doing who is giving a signal? He contracts his eyelid in order to make
a sign addressed to his companion, according to a code established between the two,
hidden from the knowledge of those around them, in order to indicate something, and
so forth. The description of the action has become ‘thick’. Ryle next complicates matters
by adding a third boy who also winks, but only in order to mimic the second one, and
a fourth who parodies the third one. All this aims at bringing out an important distinction:
the description of the boy’s movement who winks intentionally can become
complicated by the addition of new layers in the sandwich description, and yet, in this
ever lengthening description, it is always about one and the same thing. The reality
described is always the same: it’s the act of winking.
Here Ryle thus gives his own version of an idea which is developed in several ways at
that time.2 It is in fact this same idea of descriptive pluralism that one encounters in the
philosophy of action (Elizabeth Anscombe) and in the theory of speech acts (John
Austin).
There are several ways of recounting what the boy is doing, but he is doing only one
thing. In order to give the signal, the boy has to do no more than contract his eyelid:
thus, the description ‘give the signal’ enriches our information, but it establishes nothing
ANTHROPOLOGICALTHEORY 2(4)
436
06 ANT 2-4 Descombes (JB/D) 5/11/02 9:51 am Page 436
more in rebus than that same thing which had been established by the description
‘contract the eyelid’. The entire complexity is in the description: the boy does not have
two tasks to carry out, but only one. He is not, says Ryle, like someone who all at once
welcomes his aunt and pats her dog on the head (which is an example of coordinating
two actions, and not the carrying out of one and the same action).
Having brought this point out, Ryle sketches out a little grammar of verbs of doing
destined to account for the complexity of a thick description. The principle of this
grammar is that certain verbs are verbs of doing in that they signify an autonomous
action (or an activity), whereas others can only describe what the actor does. An example
illustrates this difference of category. An officer cannot order a soldier to obey (without
saying more): in order to give a command to a soldier, he must show him the thing to
be done, for example, to lower his arms. It is impossible to do something which would
be to obey and only to obey. Obedience is not an action that can be ordered independently.
When the soldier lowers his arms in accordance with the order received, he obeys.
Obeying, on the part of the soldier, is certainly not reduced to lowering his arms (for it
is not enough, in order to obey a command, whatever it may be, to lower one’s arms).
But in this case, all the soldier has to do in order to obey the order to lower his arms is
simply to lower his arms.
The difference, from the point of view of logical grammar, will therefore be between
two categories of verbs, that of the main verbs (signifying autonomous actions) and of
‘adverbial verbs’. This difference is understood in reference to the thick structure of the
description: certain verbs are found on the first floor of the descriptive construction
while others can only be found on the upper floors. Thus the verb ‘to obey’ can never
serve to give the thin description of what the soldier does when he carries out an order.
Ryle gathers together in a class of ‘constituently adverbial verbs’ all those verbs which
can never enter into a thin description of the activity: these verbs correspond to
someone’s action on the condition that another verb can specify, in a thin description,
what the agent does. The distribution of the verbs between the two categories is thus a
matter of internal relationships (Ryle, 1971: 483). Ryle explains what he means by
‘internal relationships’ in terms close to those of Wittgenstein. There are lessons that are
not possible to learn if one has not been through previous classes. A child cannot know
what theft is if he has not already learned what property is. An instructor cannot correct
mistakes in calculus if he has not first learned how to calculate. These necessities are not
epistemological, they are logico-philosophical (metaphysical).
FROM DESCRIPTIONTO INTERPRETATION
According to Geertz, Ryle’s brief analysis allows locating the object of ethnography: there
are ‘meaningful structures’ which claim that winks, signals, antics or exercises are
products, perceived as such and interpreted. He writes more precisely: ‘a stratified hierarchy
of meaningful structures’ (Geertz, 1973: 7). According to him, the interest of the
analysis proposed by Ryle is to show how inferences and implications pile up, a factor
which will be able to change a simple gesture into a complicated signal.
In his chapter, Geertz returns to Ryle’s little example and he singles out that difference
between the facial twitch and the wink. The wink, from the point of view of a
simple observer, is indiscernible from a facial twitch. But the ethnographer’s work does
not resemble that of an observer precisely because he would have nothing to conclude
DESCOMBES A confusion of tongues
437
06 ANT 2-4 Descombes (JB/D) 5/11/02 9:51 am Page 437
from a simple movement (like the facial twitch). He is always dealing with meaning (like
the wink) and even winks which point to other signals – a spate of them.3 The corporal
movement, as soon as it becomes pertinent for the ethnographer, becomes part of a thick
description. Between the physical reality of the eyelid which contracts and that which
the ethnographer is going to record in his field notebook, there is an entire thickness of
meanings allowing this physical event to be entered into a cultural category.
Until now, Geertz seems to be inviting us to make use of the analytical tool to structuralist
ends. Studying a culture is to study the ‘codes’ or the ‘structures’ by virtue of
which the events of everyday life are loaded with more or less wide-ranging meanings.
There has not yet been a reference to interpretation in the hermeneutical sense of the
term, that is to say, interpretation of a reading that is based on the radical decision of
the reader. But more precisely, Geertz goes beyond Ryle’s introduction of the thick
description. The latter’s analysis focuses on an artificial example: it is a ‘little story’ of an
Oxford philosopher. We are too far from the ethnographical investigation and still more
so from questions of hermeneutics. Geertz therefore provides an extract from his own
field notebook in Morocco in order to have us understand the pertinence of this idea of
a piling up, one upon the other, of ‘structures of meaning’.
The incident that Geertz reports dates from 1912. It takes place in a mountainous
part of Morocco (the region of Marmusha). At that time, the French forces were in the
process of establishing themselves in Morocco, but were far from controlling the entire
territory. The story presents three groups of players: Jewish merchants, Berbers, and
finally the French Foreign Legion. In the narration put together by Geertz from his
informant, there are four acts.
Act I: The merchant Cohen is the victim of aggression and a theft perpetrated by
prowlers belonging to a neighboring Berber tribe. Two other Jewish merchants are
killed. Cohen manages to escape.
Act II: Cohen goes to the French captain Dumari and declares to him his intention
of asking for compensation due him according to the traditional system of justice.
But the tribe to which the guilty parties belong has not been subjugated. Captain
Dumari does not wish to get involved needlessly and does not authorize Cohen to
have recourse to the justice of the traditional chief of the place (the sheikh of the city
of Marmusha), but he does not prohibit him from pursuing his claim. ‘If you want
to get yourself killed, that’s your business.’
Act III: The traditional chief, once informed by Cohen, puts in place an expedition
in accordance with their customs. The armed group goes into the territory of the
neighboring tribe, takes the shepherd of the tribe prisoner and seizes his flock of
sheep. They’re on the verge of engaging in battle, but decide to negotiate. The wrong
done to Cohen is acknowledged and a fair compensation is granted him. Cohen
chooses the sheep which are due him, in accordance with the traditional rules, by
way of compensation.
Act IV: Cohen returns to Marmusha with the sheep. The French understand nothing
about this and believe him to be in league with the Berber rebels. They confiscate the
sheep and put Cohen in prison. Finally, the unfortunate Cohen manages to get out
of prison but fails to recover his sheep.
ANTHROPOLOGICALTHEORY 2(4)
438
06 ANT 2-4 Descombes (JB/D) 5/11/02 9:51 am Page 438
THE CONFLICT OF INTERPRETATIONS
Geertz offers us this little account as an illustration of the thesis he is in the process of
defending: that the ethnographical description is always thick and that the interpretations
of the players, in addition to facts and observable gestures, always enter into it.
He also talks about ‘constructions’:4 reporting whatever event – for example, the aforementioned
incident – is in reality reporting the way in which someone in particular
(here, Cohen) understood what happened; it is also reporting the interpretation which
someone gives of the way in which the other protagonists have understood or interpreted
the entire incident. If the ethnographer can himself believe, more or less, or make his
reader believe that he is dealing with ‘raw facts’, it is only because he began by adopting,
without saying it and perhaps without realizing it himself, a unique system of interpre-
tation.
But where is the relationship between the Moroccan incident and Ryle’s example?
Geertz himself indicates an important difference: the two boys, in the philosopher’s
example, have a code, which they are using to communicate, whereas the Marmusha
players have instead ‘frames of interpretation’ which enter into conflict. At Oxford, one
is understood without having to spell it out, and even without speaking, by a simple
wink. In that part of Morocco in 1912, things are more complicated: everything must
be explained and, in addition, the explanations themselves are not, in the end, entirely
understood by the actors in the scene. At bottom, Cohen’s misadventure applies to a
‘babelian’ situation, to a ‘confusion of tongues’ (Geertz, 1973: 9). The poor soul does
not know which authority to appeal to. If he lodges his complaint with the French
authorities, they declare themselves incapable of acting. If he turns towards the traditional
Berber authority, the latter enters into action in satisfactory fashion as long as
things take place far from the French fort. But as soon as he returns to the city, the
colonial power reasserts its authority.
Upon replacing the scholarly example of Ryle with the more dramatic adventure of
the merchant Cohen, Geertz has tried to show that the ethnologist, just like the literary
critic, is engaged in a hermeneutical activity. He is not dealing with clearly identifiable
facts but with accounts and interpretations. Everything takes place as if he had an old
palimpsest covered with marks and that he was obliged, by means of interpretation, to
restore its meaning; and of course, to discern what comes from the source, what has been
added or left out by the copyist, what glosses have been added by a biased commentator,
and so on (Geertz, 1973: 10).
Along the way, we have really passed from thick descriptions, in Ryle’s meaning, to
hermeneutical interpretations. Geertz perhaps considers that he has satisfied himself in
adapting Ryle’s analysis to the working conditions of an ethnologist: the thick description,
when practiced on another culture, takes the form of a hermeneutical interpretation.
But, in reality, the change of example had the effect of turning the meaning of
‘thick description’ on its head. With Ryle, thick description assumes thin description:
there could not be upper floors without the ground floor, a ham sandwich without bread,
adverbial verbs without main verbs. With Geertz, concerned about combating the positivist
prejudices regarding the ‘raw fact’ and the ‘data of observation’, ethnography is
only thick description. Thus, one has first passed from a distinction between two types
of description to a distinction between pure description and description mingled with
interpretation, in order to finally arrive at the good old-fashioned opposition at the core
DESCOMBES A confusion of tongues
439
06 ANT 2-4 Descombes (JB/D) 5/11/02 9:51 am Page 439
of hermeneutical philosophy between, on the one hand, the description of facts and, on
the other, the interpretation of meanings.
It is true that Ryle’s example has something narrow and insular about it. He evokes
irresistibly a universe of boarding school and scout games. The entire scene is immersed
in an atmosphere of great cultural homogeneity, not only between the different boys,
but also between the players and the observer who piles up adverbs and thick description.
This is why Ryle does not encounter, at any point, anything that would lead him
to pose the ‘hermeneutical problem’.
For his part, Geertz has taken care to choose as a typical example a babelian scene of
crisis. The merchant Cohen does not know which way to turn, but it is not because he
lacks the resources to understand or to make himself understood; it is rather because the
situation in which he finds himself renders impossible any compensation for the wrong
done to him. The French authorities do not want to concede to the local authorities any
role in the administration of justice, but are not yet capable of exercising total control
themselves. In such a state of things, it is impossible for Cohen to explain what has
happened to him in such a way as to lay claim to his rights.
One cannot therefore say that Geertz has sketched out here the ethnographical
description of an institution or of a life form. It is rather, in this particular case, clearly
a situation where an institution of traditional justice is prevented from functioning by
the intervention of the French, yet without being replaced by another system. On the
other hand, when Geertz explains to us how the traditional system of the penal code
functioned, he does it by referring back to the period when this system was actually in
force: he tells us how things would normally have happened before the arrival of the
French.
The misadventure of the merchant Cohen from Marmusha is due to the confusion
of tongues. But the very notion of a confusion of tongues assumes that there are different
languages, not ‘interpretations’ as numerous as individuals. In order for there to be a
babelian misunderstanding (between Cohen and the Legion captain), it must be true
that, in other circumstances, people normally make themselves understood in their
respective tongues. This is precisely because a person normally makes himself understood
when he expresses himself in his language; thus there is a situation of confusion of
tongues when people perceive that they are not being understood. If the confusion of
tongues were the norm, this would mean that, as a general rule, one is not being understood,
even when one believes oneself to be understood. That would amount to saying
that one is understood just as well (or as poorly) when one is not being understood as
when one believes oneself to be understood, and that the feeling of a particular confusion,
tied to the irruption of foreign groups in the same territory, only manifests a
normal state of overall misunderstanding. Finally, a radical thesis on the babelian character
of communication as such must maintain that there is really no difference between
understanding and misunderstanding.5 In order to escape this contradiction, it is therefore
necessary to maintain that the diagnostic of confusion of tongues must form a sharp
contrast to the normal state of communication. It is not enough that there be several
languages for one to be in a babelian situation: there must be several in number while
one was expecting that there be only one.
ANTHROPOLOGICALTHEORY 2(4)
440
06 ANT 2-4 Descombes (JB/D) 5/11/02 9:51 am Page 440
DESCRIPTIVEVALUE OFTHICK DESCRIPTIONS
Ryle would ask, what distinguishes the upper floors from the description of the ground
floor? One would be mistaken to understand this range of levels in an epistemological
sense. The difference is not that one would find observations in due form on the ground
floor, the upper floors being reserved for personal glosses, for interpretations, for subjective
perspectives. In fact, the difference is that on the ground floor, description is of
the most rudimentary nature: it must consist of a main verb but not of adverbial details.
Adverbs are not qualifications or additions made by the interpreter in order to report on
his reactions or his hypotheses on the process being observed. A description in the form
of verb + adverb is thicker than a description reduced to a main verb, but it is no less
descriptive thereof. Example: ‘The child eats his soup reluctantly’ is no less descriptive
nor more interpretive than ‘The child eats his soup’.
With Ryle everything points to the idea of a system to be elucidated. Among the adverbial
verbs of which he makes a list, one discerns three principal groups, which correspond
to three ways of discovering an intelligible order within an intentional act. The
first two groups correspond to individual actions, the third group comprises actions that
have a social nature, a factor which introduces a particular problem.
1. There is, first of all, the order of a practical syllogism: how is a boy going to behave
in order to communicate information to his companion? He will do it, in conformity
to their code, by winking. ‘Winking’ is the means of the act of communicating
and, therefore, from the point of view of practical reasoning, it is the gerund
clause ‘by winking’ that serves as the adverb of the main verb ‘to show that the
teacher is coming in’. But, from the point of view of the action itself, things are
reversed: that which the boy must do to communicate is the action described at the
thinnest level. (The more one describes the means put into play, the more one
descends on the descriptive scale, since specifying the means employed is to add an
adverbial phrase of means. The more one points to the end, all the while setting
aside, if not the means, then at least their materiality, the more one places oneself
on a higher plane.)
2. Moreover, Ryle proposes several examples that belong to a group of activities which
one can gather under the name of exercises (preparations, training, demonstrations
of putting one’s ability to the test). Here, someone will make a gesture to improve
his performance, or to show that he knows what he is doing or how he must act.
For example, the movements of a tennis player who is practicing repetitive service
movements are exactly (or preferably) those which the player makes during a game,
but there is the added complication that these movements are being carried out with
the idea in mind of acquiring a skill (or attaining a higher level of a skill). The logical
relationship is thus the following: one can carry out an act for its own sake (and not
in order to practice), but one cannot train oneself to succeed in an action without
undertaking that action.
3. Finally, a certain number of these examples remind us that there is a logic to social
acts. In order to obey the command to lower his arms, the soldier has only to do
one thing: lower his arms. The description ‘to obey’ doesn’t enter into competition
with the description ‘to lower the arms’, nor is the former added to the latter (as if
DESCOMBES A confusion of tongues
441
06 ANT 2-4 Descombes (JB/D) 5/11/02 9:51 am Page 441
the soldier had to do both things, lower his arms and obey). The two descriptions
must be arranged on different levels.
This analysis of Ryle’s, it seems to me, is convincing. One could illustrate it by the
following distinction. An officer gives a soldier the order to undertake a certain action
(dig a hole). The soldier does it. He does it by command. Let us suppose that he does
it rapidly. From the logical point of view, we will make the distinction between carrying
out the order of doing A rapidly and rapidly carrying out the order to do A. This distinction
could be expressed with the help of parentheses indicating in which order the
adverbs come in the sentence. In the case where the soldier hurries to carry out the order
received, the description of his action will take the form: RAPIDLY (by command [to
do A]). In the other case, where the order was to do A rapidly, we will have a description
of the form: BY COMMAND (rapidly [do A]).
Thus we clearly have two different descriptions, two different types of behavior, the
distinction being marked by the order in which the adverbs come in the sentence. Or,
if one prefers, by the level that they occupy in the composition of the ‘sandwich’.
The conclusion that emerges from the parallel between Ryle’s analysis and that of
Geertz is now clear. If we retain for the term ‘thick description’ the meaning that it has
for Ryle, then it is undeniable that, under the conditions described by Geertz, there is
no place to speak of thick description, any more than of thin description. This distinction,
with Ryle, introduces the idea of a logical order of descriptions, of an intelligible
organization: I can obey the order to lower my arms only by lowering my arms, not by
doing something which would be and only be to obey. As for Geertz, he is interested in
the complication of his material, but this complication contains nothing of a logical complexity:
no system is able to arrange in order the actions of the Jewish merchant, the
sheikh and the French captain. There is disorder from the historical point of view. The
Jewish merchant wishes to see justice rendered to him. When his complaint is understood
according to one of the rival systems and a fair compensation is given him, he discovers
that this is understood, according to another system, as a breach of law. The
plurality which Geertz draws from his account is a conflict of antagonistic interpretations,
not an order of descriptions linked to one other by internal relationships.
Thus is Geertz poles apart from the philosophy of action which he has cited: the
problem he wishes to pose, and the conceptual tools which he wants to allow himself,
have nothing to do with the distinction between the main verb and the adverb. But this
distinction suffices to render incoherent the idea that the hermeneutical ethnographer
is dealing with description. Since he does not deal with thin description, neither does
he deal with thick description, for the two go hand in hand. Since all meanings are added
by the witnesses or the informants, one never manages to know what the significance of
the gesture was. In Ryle’s example, the significance of the wink is in the wink: there is,
by virtue of the code, a value in the message of the gesture. The significance is not only
an ‘hypothesis’ that the observer would fabricate to ‘interpret’ data. If the observer does
not see that the gesture has meaning, the fact is that he does not see what is going on
before his very eyes; he has been kept in the dark with regard to the two boys’ secret
code.
Is this to say that the anthropologist, if he intends to do descriptive work, could
practice the art of thick description in Ryle’s sense of the term without having to embrace
ANTHROPOLOGICALTHEORY 2(4)
442
06 ANT 2-4 Descombes (JB/D) 5/11/02 9:51 am Page 442
the doctrines of hermeneutics? I believe so, with one small reservation. What the anthropologist
could retain of Ryle is the idea of the complexity of the description as system.
The analysis of the Oxford don should hold the attention of the anthropologists who
call upon Mauss and who have kept the idea according to which the inquiry must
attempt to grasp ‘total social phenomena’. As is proper, the don took for an example a
little scene, a minuscule, isolated episode. However, if one had to describe a complete
social phenomenon (as Geertz did with his famous analysis of the Balinese cockfight),
one would encounter the problem of putting in order the various dimensions of the
phenomenon to be studied. It is then that the idea of an order, following the model of
the introduction of adverbs in a simple sentence, can be of use. Not that an anthropological
description must necessarily adopt purely and simply a logico-grammatical
model. It remains that the problem arises for the investigator to not confine himself to
a multiplication of ‘points of view’ and of ‘aspects’, but to find an articulation, an order,
in this plurality. To speak of thick description here would be to show that description
must not only be multidimensional, but that it must also be organized according to
levels, along the lines of Ryle’s ‘sandwich’.
There is, however, one qualification to be made about the analysis that Ryle proposes
for his example. Geertz was correct in denouncing the following bias: one should, in
order to pile up meanings upon one another, first allow oneself an impeccably positive
description. Now what Ryle suggests is as follows: the thin description appears to be for
him a description which could be accepted by a psychologist of behaviorist persuasion.
Ryle acts as if all ‘intentional’ or ‘semantic’ description were the redescription of an
observable movement. Here he gives in to a well-known temptation, that of considering
that the perception of human behavior begins with raw facts (the behavior) and that
it is then built on this positive basis.
If Ryle remains, in this case, a prisoner of a certain behaviorism, that is no doubt
explained by his concern to exclude dualism. His entire analysis aims at driving home
the point: someone who intentionally undertakes an action does not do two things at
once, on the one hand a sequence of body movements and on the other, a mental act
(that of targeting a goal or of desiring a result). There is only one thing done. Ryle’s
perspective is therefore a critique of the myth of volitions.6 His critique is altogether
effective and there is no reason to return to this point.
This in no way means that we can give ‘thin descriptions’ only on terms acceptable
to a behavioral psychologist. Ryle, among the examples he gives of a thick description
of action, cites the case of someone who obeys an order. The main verb is going to state
what was to be done, the adverbial details will tell in what way it is done (in this instance,
the action is performed in compliance with the order received). The main verb is therefore
going to correspond to the content of the order given by the officer. But it goes
without saying that the officers are not thus limited to ordering exclusively corporal
movements. It is even likely that most of the orders focus on intentional acts. For
example, the soldier is sent ahead to scout around and must return in order to tell if
there is anything to be reported. The description ‘Go and see if there is anything to be
reported’ is thin (it can appear on the ground floor), but it comprises an obviously logical
complexity which prevents seeing in it the reading of a natural movement (as opposed
to an intentional action).
One will note that the example of the wink, by its excessive simplicity, has the
DESCOMBES A confusion of tongues
443
06 ANT 2-4 Descombes (JB/D) 5/11/02 9:51 am Page 443
disadvantage of favoring the illusion of a word-for-word correlation between human
action and observable behavior. The peculiarity of giving the agreed upon signal by
winking is that such an action is very economical: giving such a signal takes very little
time since this is done precisely ‘in a wink of the eye’, the time needed to wink.
The simplicity of the bodily movement of which the action of giving the signal
consists is a characteristic feature of this particular example. As soon as one turns to other
examples, ‘la belle correspondance’ between a simple movement of the body and an
action disappears. This takes nothing away from the physical reality of the action and
thus does not affect in any way what Ryle is attempting to establish (in his antidualist
crusade): to wink voluntarily, there is nothing more to do on the part of the subject than
to wink (at the moment chosen by him). Similarly, in order to go from Paris to Lyon,
the subject has nothing more to accomplish than the movement of going from Paris to
Lyon (by such and such a means). Nevertheless, in this second example, the voluntary
act does not consist of an elementary gesture but of an orderly, complex sequence of
gestures (which are themselves complex). It follows that the ‘thin’ description of a
journey to Lyon will have nothing of behaviorism.
Geertz was therefore wrong to say that ethnography was only thick description. To
eliminate thin description from the palette of which the ethnographer avails himself is
to abolish the contrast between the thin and the thick. This hermeneutical radicalization
renders inapplicable the idea of an organized complexity of levels of description.
Nevertheless, Geertz was right in pointing out that the ethnographer did not have to
build his thick description on a hard rock composed of descriptions wholly decontextualized.
In the notebook where the ethnographer records his ‘field observations’, there will
be no behaviorist readings.
AN IMPROPER SEMIOTIC
Geertz says that cultural anthropology need not claim to aim for a consensus – even
partial or provisional – but that it can do no better than render more ‘refined’ the terms
of the disagreement (Geertz, 1973: 297). It is then a question of a conflict between
anthropologists. But, oddly enough, this ‘hermeneutical’ conception of the discipline
finds its counterpart in the definition of its object. Geertz is not far from giving his characters’
confrontation the appearance of a confusing dialogue between visions of the
world foreign to one another, as if the protagonists which he puts on stage were the representatives
of three traditions who would meet in order to reflect together on the essence
of justice.7 The definition which he assigns to the ethnographical activity changes the
incident that he reports into a sort of philosophical symposium between a Jew, a Berber
and a legionnaire: all three have an idea of criminal justice, but it is not the same.
Thus Geertz, guided perhaps by the example of the boys who send each other signals
back and forth, proceeds to a semiotization of the whole incident which he has reported
in his field notebook. Each group gives a speech in its own tongue, in which it affirms
the validity of its point of view: Cohen considers that he has the right to compensation,
the sheikh wishes to maintain his authority and the French desire to establish theirs.
According to Geertz, the whole constitutes a ‘social discourse’ in several languages
(Geertz, 1973: 18). However, since the confusion of tongues reigns, one could speak
more precisely of a brouhaha8 than of a discourse.
In fact, this incident illustrates rather clearly the absence of a social discourse common
ANTHROPOLOGICALTHEORY 2(4)
444
06 ANT 2-4 Descombes (JB/D) 5/11/02 9:51 am Page 444
to all the protagonists. To persist in speaking of ‘social discourse’ in such a situation is
to leave us with the understanding that it has, in fact, always been thus, that when people
speak to one another in the same tongue, in fact everything happens as if they were
speaking to one another in different tongues.
In order for the object of anthropological inquiry to be presented as the analogue of
a text, we must transform the protagonists of an action into interlocutors of a social discourse.
In order for the thick description to be a hermeneutical activity, its thickness
must not have descriptive value, strictly speaking. These are the two symptoms of one
and the same difficulty, that of linking together the analytical philosophy of action and
hermeneutical phenomenology.
Notes
1 One understands generally by ‘hermeneutical problem’ the point of knowing if we
have reasons to interpret a sign as we do, whereas: (1) there are in reality other
interpretations possible and, among them, other interpretations actually forbidden by
other traditions and other sects; (2) the solution proposed could not claim to be true,
as if there were only one single correct interpretation (that is excluded beforehand by
the ‘universality of the hermeneutical situation’, or if one prefers, by the very constitution
of language, of history and of human existence). Certain hermeneutical
philosophers respond to this problem by sending us back to our radical personal
choice (these are the heirs of existentialism); others tell us that, in practice, we don’t
have a choice, for we must take as a starting point the interpretation which is imposed
upon us by the fact of our belonging to a particular historical tradition (whether we
accept it or contest it).
2 One really could not say that Ryle is borrowing this idea from others, for his entire book
on The Concept of Mind develops the idea according to which certain psychological
verbs, far from signifying extra deeds achieved by the subject, indicate rather the way
in which he does what he does. For example, the candidate who rereads his copy and
who pays attention in order to discover the mistakes that might have slipped in, is not
someone who does two things; he is someone who does one thing – the reading of the
copy – and who does it very carefully. Someone who gives an intelligent answer is not
someone who does two things (answer, be intelligent); it is someone who does one thing
intelligently. Thus do we rediscover the insightful Rylian idea of an adverbial status of
intelligence, of awareness, of attention, and more generally, of verbs of thought.
3 ‘Winks upon winks upon winks’, writes Geertz (1973: 9).
4 In English, the notion of ‘grammatical construction’ is used more liberally than in
French. One speaks not only of constructing the entire sentence according to such
and such a word order (in order to render it word for word), but also of constructing
a word according to a certain syntax, and especially of constructing the meaning of
the sentence, which is the equivalent of giving one such and such an interpretation.
5 Here one cannot help thinking about what Baudelaire wrote: ‘The world works only
by misunderstanding – it’s by universal misunderstanding that the world gets along.
For if, as ill luck would have it, one was understood, one could never be in agreement’
(Baudelaire, 1863, no. 42). But, unlike the ‘post-structuralists’ and the ‘postmodernists’,
in no way does Baudelaire confuse the two senses in which people can
get along between themselves: only understanding each other, or rather agreeing.
DESCOMBES A confusion of tongues
445
06 ANT 2-4 Descombes (JB/D) 5/11/02 9:51 am Page 445
6 One understands by ‘volition’ that mental event which the subject is supposed to live,
or produce, in order to give rise to voluntary movements of his body.
7 In reality, the misunderstanding especially affects the relationship between the
colonial power and the local communities. The account of the incident shows that
the latter had found the paths to coexistence within a like order of justice.
8 It is a matter of no small importance to note that, according to one of the proposed
etymologies, the word ‘brouhaha’ would be the ‘onomatopoeic deformation of the
Hebraic formula baruk habba, “blessed be he who comes” (in the name of the Lord),
Psalm 113’ (Bloch and von Wartburg, 1975).
References
Baudelaire, C. (1863) Mon coeur mis à nu no.42
Bloch, O. and W. von Wartburg (1975) Dictionnaire étymologique de la langue
française. Paris: Presses Universitaires de France.
Geertz, C. (1973) The Interpretation of Cultures. New York: Basic Books.
Ryle, G. (1949) The Concept of Mind. Hutchinson: London. (French translation, 1978
La Notion d’esprit: pour une critique des concepts mentaux. Paris: Payot.)
Ryle, G. (1971) ‘Thinking and Reflecting’ and ‘The Thinking of Thoughts: What is Le
Penseur Doing?’, in G. Ryle Collected Papers. II: Collected Essays 1929–1968.
London: Hutchinson.
VINCENT DESCOMBES is Directeur d’études at the Ecole des Hautes Etudes en Sciences Sociales. Publications
available in English: Modern French Philosophy (Cambridge University Press, 1980); Objects of All Sorts:
A Philosophical Grammar (Blackwell and The Johns Hopkins University Press, 1986); Proust: Philosophy of the
Novel (Stanford University Press, 1992); The Barometer of Modern Reason: On the Philosophy of Current Events
(Oxford University Press, 1992); The Mind’s Provision: A Critique of Cognitivism (Princeton University Press,
2001). Address: Centre de recherches politiques Raymond-Aron, 105 bd Raspail, 75006 Paris. [email:
Vincent.Descombes@ehess.fr]
Translator
Henry M. Smith, Lecturer in French, Department of French & Italian, University of New Hampshire,
Durham, NH 03824. [email: hmsmith@christa.unh.edu]
ANTHROPOLOGICALTHEORY 2(4)
446
06 ANT 2-4 Descombes (JB/D) 5/11/02 9:51 am Page 446