P. Mendès France Univ., Grenoble II, Business Management School E.S.A.




Ph. D. in Management Sciences





For the language analysis of the organization





volume 3/3 - Summary, Auto-language analysis,

final presentation


Presented and defended in public

the 23rd November 1995 by :         Richard NOLASQUE




Members of the jury :


Director of the Ph.D. :     Mr. Jean MOSCAROLA

      Professor at the University of Savoie


Official witnessess :   Mr René THIEBLEMONT

      Lecturer at the University of Savoie


      Mr Claude JAMEUX

      Professor at the University of Aix-Marseille II


Other members :   Mr Humbert LESCA

      Professor at the P.Mendès France Grenoble II Univ.


      Mr Jacques TRAHAND

      Professor at the P.Mendès France Grenoble II Univ.



French Summary :


Pour un audit langagier de l'organisation :

novembre 1995


Cette thèse défend le champ des potentialités d’une approche langagière de l’entreprise en vue de mieux gérer ses ressources humaines et cognitives. Par la gestion d'actèmes (= couple acte / sens, association entre un acte et un pourquoi), de langage partagé, et de cohésion organisationnelle. Techniquement, on invente 3 manières de déterminer le langage partagé.




Summary in English:


For “the language” analysis of the organization

November 1995


This conception of the day-to-day management is related to human and cognitive resources of the organization. The "actem", which means "couple-action", forms the general framework of the practical management : language as the tool to manage the members according to the goal of the organization. An example of application to the management of cohesiveness is developed, using 3 relatively-new lexical analytical ways to identify the "shared language".


For “the language” analysis of the organization



Summary of the Ph.D.


P. Mendès France Univ., Grenoble II, Business Management School E.S.A.-C.E.R.A.G.

Author : Richard NOLASQUE     Director of Research : Jean MOSCAROLA




This research explores the possibilities of a language conception of the management of the organization. This is linked with the fields of the Communication, of the Human Resources Management and of the Management Information Systems. In the perspective to establish a method to evaluate the problems of the organization based on the speech analysis, we bring a certain distance to the true way of managing the enterprise.

Indeed, we are not always conscious of the language dimension in an organization. It is invisible to the user (according to Aktouf86) and to the receptor. It is difficult to scientifically analyze why some of the researchers, who approach the commercial world, do not utilize this filter or net ? However, the development of some textual software statistical analysis permits a rigorous analysis of the same.

The outcomes are interesting for the Management Sciences. It concerns analyzing how it works, what is the identity and what is the internal communication. More specifically, we are going to emphasize that, creation of group-feeling, which is the spirit of a common goal, the involvement of all its members, could be achieved through the proper management of the language. According to Martinet, “to manage”, is to manipulate signs and symbols.  These signs and symbols are “words” that are shared between the employees. Managing the language dimension in an organization is a new means to gather all the odds to one side, thereby facing a competitive environment.

The research philosophy takes its roots in two paradigms : the positivism and the constructivism, stating on the pre-existing or not of the reality. Indeed, the language allows not only managing the items of the world but, moreover, it allows the building of realities. This is why this research can be qualified of constructivist, systemic and of engineering type. Otherwise, we will also call the socio-linguistics.

To support the idea of a true management of an organization through the language, we will commence by exposing the theoretical side, which is the primary objective. Secondly, we will technically present, the textual statistics, permitting the merging of real actions along with the previously exposed ideas. Finally, in the third part Applications, we will apply the ideas and the technique on the real organization.

 A framework : The language as a tool of management

In the first chapter, we will bring out the obvious interest of the language approach for the management sciences. In the second chapter, we will detail the language encountered in the two paradigms (positivism, constructivism). In the third chapter, we will conclude towards the generic thesis of a language management of the organization.

The interest of a language approach

To prove the interest of a language approach of the organization, we are firstly going to define the contrast between the management model and its practice.

Despite of the theories laid down, the prevalence of some of the problems such as malfunctioning, distortion in information, loss of motivation, unexplained counter-performances, rumors, cultural differences etc.... cannot be ruled out. In management, there is a gap between the practitioners and the theorists. In practice, communication problems that exist in the enterprises seem to be rather the rule than the exception. For instance, at times the employees are not very well aware of the identity of their enterprise, some of them do not know the existence of the other, some even ignore the needs of the others (as explained by Maslow, about the need of truth and of social order) and so on.

The enormous environmental pressure such as the market of competition amplifies the emergency of an action (according to Etchegoyen90) and the cult of acquiring a successful performance (according to Erhenberg91). "The decisional approach prefers the action rather than the understanding" (according to Crozier77 p.265). The power (Mintzberg) disturbs the necessary occurrence of the actions in the ideal process of communication (according to Habermas) because one considers the secret as a protection, an asset. The outcome paralysis the thought, the distance and one understands the whole difficulty of the scientific approach of management.

In reality, the way from the management model to the practice is always done through the verb, the language, which filters and twists the intention issued by the model. The way to apply the theory differs, depending on the language that one uses.

So, taking a closer view of this idea : the language, this concept is able to take into account the information, the communication, the culture and the systemic. Parallel to its ability to take some distance compared to the contents, its intrinsic operational use is included in the fact that it manages words with the physical reality, observable by all the actors in the enterprise. At the same time, flexible and easy to handle, it is also stable and structured. To summarize, it constitutes a true dimension, being everywhere and parallel to each action in the enterprise.

After the references to the theories of the linguistics (Saussure, Benvéniste, Chomsky) we emphasize on the gathering side of this concept, concerning the management phenomenon. In fact, the “language” allows us to take a distance from the day-to-day management.

It is an informational tool, the basis able to send some data and some orders. It can reduce the uncertainty, for instance make conscious verbally the cognitive filters. It is also the ultimate tool of communication, the relational medium shared by every person, in most of the interactions. The language is also a cultural tool (social linguistic identity) : it can clearly identify and thinly re-deal some "strokes" (cf. Transactional Analysis), some answers to the psychological needs, an ideology, and the intimate beliefs of the enterprise toward the actors. It is one of the roles of internal communication, as seen achieved when the Manager officially speaks authoritatively.  The language also allows taking into account the ecology of the enterprise (words more or less disordering or reordering).

To summarize, one can consider the language as a mean to manage the emotional pressure of the day-to-day management, to understand the employees and to answer to their psychological needs.

Two paradigms in the heart of the management

Language and positivism

According to this view, the management is to handle some measures of the reality (management control, exact result figures...) and to insure/secure the "making of order" of the various businesses of the enterprise. The language, rational and analytical, in this positivist paradigm, would be a tool to stabilize the organization.

Firstly, one must talk about the famous information theory of Shannon. The information is there a statistical measure? The more unlikely it is, the more it brings information. Some messages thus "contain" an informational value more important than some others. However, this message is disturbed by some "noises", some filters (including the language) which twist and alter the original intention/will. The cybernetics adds the idea of retro-action between the sender and the receiver.

The whole conception takes its root in the principle, that the meaning pre-exist before any exchange. It is placed in the sentences or in the context or in the intention.

Lastly, in summarizing the role of the language in this paradigm, it would be a tool "factor for gathering stability". By improvising the right movement of information, utilizing more precise and practical words and adopting the language of the receiver, one could reduce the disorder and the uncertainty and thus stabilize the organization.

But, this conception is based on unsure statements : The mechanical, non-human, and mono-meaning status of the inter-personal communication, or the fact that the medium is the center and even the cybernetics, allowing the message to have two ways, stays in this paradigm.

In fact, we arrive at some reasoning about the system inertia. We face a resisting behaviour of the system against the act to make some order, as if the disorder was naturally increasing (entropy) and as if there existed a law of the required variety for a system to work. In conclusion, we arrive at the necessity of a more informal and human language.

Language and constructivism

According to this view, the management is to permanently build some adapted shapes and to live in a reality with a minimum complexity. The language, in this paradigm constructivism and systemic, would be a tool to change the organization.

Firstly, we develop the theory of communication more expressively and relational to the idea of information. In the heart of the day-to-day management, one can say that to communicate is not only to exchange and to share (Etchegoyen90) but, also to act on other people.

This whole conception takes its root in the principle that the meaning is built, by the interaction between several persons by some specific world models to each individual or by the speech act itself.

Lastly, in summarizing the role of the language in this paradigm, the language would be a tool factor of "dis-ordering" change. In managing the contradictions, the rumors, we could increase the disorder and the uncertainty and thereby promote the change in the organization.

Thus, we logically arrive in the idea of equilibrium between these two sides.


To manage the stability and the change

Every ecological system, alive or not, must balance its development between these two ideas. The language allows potentially managing this paradox in the heart of the organizational sciences.

It is here that we advance a thesis. There would systematically be, in every organizational entity (from the actor/ real person up to the enterprises) some "gathering and ordering" speeches and some "dispatching and disordering" speeches. The first type shapes or trains a spirit of group-feeling, useful to increase the efficiency of the collective action and vice versa for the second creating disorder. Furthermore, towards this idea, there would be a shared language (factor of cohesion) facing some language zones (factor of dispatching). What interests us, is the more general idea of “the language” as a tool of management.  Therefore, we commence exploring the same.

The language as a management tool

Management through the language uses the semantic and pragmatic functions of the language.

The language as a tool of meaning

The semantic linguistics puts light on this function of the language. The language is an intermediate (filter) between the reality and the organizational entity. It permits the approach to reality (pre-existing or built) by the means of the mental representation and thus to change the perception of the reality (reframing of the meaning). It is also possible to evaluate the organizational entity.

Moreover, the language is a tool of culture because it allows the holding of the cohesion and of the internal coordination on one side and the definition of the relationship of the organization with its environment on the other side (Laroche91). It comes through the beliefs, which reduce the disorder and play on the identification of each to its organization. Through proper management of the language, the shared values can or cannot increase the cohesion.

The language as a tool of action

The linguistic pragmatics (Austin) makes the word a tool of power. More precisely, the specificity of the action through the language is well acknowledged by the phenomenon of cognitive dissonance (Festinger) where the (verbal) information creates the reality (Watzlawick). The action through the language must be accompanied by facts; this is the congruency between "to do" - "to say". If not, it makes things to be disordered or unbelievable. Otherwise, to find words can sometimes be enough to trigger an action. Lastly, the main active principle is summarized by the idea of spreading or dealing words amidst persons.

The facts involved by an action through the language are the material basis to make the persons more sensitive. One can even imagine a group of persons, with a preference for some particular kinds of words, to be sent toward some departments or parts of the organization, just to deal its words.

To mean and to act

The language is at the same time a passive representation and an active device in the organization. It has a role of an intermediate, which can at will, amplify or reduce the strength of the meaning and the action. To manage through the language would consists in handling some couples (or associations, or links) of a meaning and an action, which can be called "actems". For example, in giving the right meaning to an action of an organizational change. We conclude on the perspective of a language analysis of the organization.

A technique : the textual statistics as a means of investigation

We are now going to summarize the technical facts permitting the application of the previous ideas into an organizational entity, for example in an enterprise. To do so, we will firstly examine the ability to scan or not to scan the dimension of the language. Further, towards the second and the third chapter, we will examine the techniques of lexical analysis to make the language zones obvious amidst the actors.

The observation of the language in the organization

To screen the organizational entity, we will observe the most obvious components of the source of the language, which are its members.  The members produce mainly conversation, communication and some exchange of information. Now, in order to gather and collect these words, the place of the researcher in the organization, must be well understood.  She/He must disturb its observation the least possible.

The data (pre-existing or built) are written or spoken. We go into further detailing the interview because it is achievable even in small structures which have very few writings of internal communication. It consists in "stimulating" the world model of the person interviewed with some special words on the entrance of the model.

The researched information is uncommon because we only search the organizational speech, in mass, whatever are the expressed opinions and contents. We define this kind of speech by a high proportion of words related to the management of the organization. This speech is generated in the mouth of the interviewed persons with the help of an interview guide, structured (predefined questions) or not (chatting more or less lead by the interviewer).

The computerized components by lexical analysis

Lexical analysis will be utilized to deal with these texts. Now, by avoiding the tool words, for instance, "of, to, by", we will count the other words which are in repetition. The conception underlying this technique is related to the speech acts theory (cf. Austin). It commences with the principle that each word in a speech is the result of a micro-decision, a speech act, from the author. It is this choice, more or less consciously, which shapes the statistical unit permitting to approach the mental representations of the actor. The more frequent a word is spoken, the more it is linked with a meaning in the mental model of the individual. 

We put the hypothesis that a lexicon takes into account a speech which takes into account a person. So, one approximates a person and its lexicon. To cope with a lexical analysis, one firstly extracts the lexicon of the texts (one per actor) previously gathered and one counts the number of times that a same word is repeated (its frequency). This manner to conceive the language is analytic and reduced. We are conscious to loose the syntax and rhetoric dimensions of the analyzed speech, without evoking the various meaning of a single word.

The analyzed components are : the actors, real (working persons) or virtual (computed average person, summarized executive lexical profile...). The words, "nature" or gathered into lexical fields, for example the words expressing the theme of a performance or of a war. We can also measure the grammatical dimension : which is the speech structure that is not often perceived, like the negation, the percentage of adjectives, the length of the sentences, the number of "I"...

Encoding the data consists mainly in putting the words in rows and the actors in columns.  It allows a classical Statistic Analysis of Data (Factorial, Typology...).

The shared language and the language zones

In placing the words (or fields) in rows and the actors (real or virtual) in columns, one shapes a chart which shows the number of sentences of the actor where one finds the word. A factorial analysis permits easily to notice which actems are noticeable. One identifies the specific words of an actor through the detail of a Chi² test : the statistically different words of an actor define technically its language zone. This is how, that some groups of actors like those qualified of "recent members" and "old members" can have specific fields that they share equally and others which oppose them. The first ones belong to the shared language, and the second ones belong to the language zones. The words that are statistically shared and spread with homogeneity between all the actors are the shared language. The words especially attracted by a specific organizational sub-group are a language zone, the one of this group of actors.

To compute the shared and unshared language, we create a lexical differential, a technical procedure dealing with the differences of the relative frequencies for the same words between the given actors. We also create the crossed sorts method, which deals with a sum of differences of the percentage of part Chi² on total Chi², the sum which must be approximating to zero, to identify the shared words.

We conclude by some considerations detailing the crossing of the investigation to the action.



Some applications : case studies

In the third part of the book, we cope with some practical cases through a language approach in the goal of evoking the real main possibilities of analysis, check-up and action in an organizational entity. Thus, we should imagine some language analysis of bigger organizations with a true statistical value. We detail mainly the check-up step, basis of further language actions to be decided with the leaders of the organization.

Case #1 : Exploration through the organizational speech

To experiment the language approach, we begin with a small survey (10 interviews) launched in 1992 in a small printing factory, composing of 18 employees. We gather some socio-organizational variables (type of management activity, status, sex, number of years of membership) and conduct some interviews, under the shape of an informal talk.  Half-directed, we search to stimulate verbally the interviewed person with words like "customer, product" in the goal to achieve a general organizational speech. These long texts, recorded on tapes, become a material that can be analyzed through counting the repeated words.

Thus, we identify that the Technical Manager uses much less visual vocabulary than the persons interviewed. This can explain a part of his difficulty in communicating. The General Manager could so support the Technical Manager in working and speaking more on the visual dimension (boards, charts...).

By looking at the frequency chart, we define the shared language of the enterprise ("to do, to say, can, machines, customers, problem, time, price") which is best represented by the actor #7, said the average actor. The action of gathering the members, to be done by the General Manager, could so consist in filling its speeches and "full of meaning" actions of the theme of "machines, customers", and publicly to value the actor #7 which is the language heart of the enterprise.

Case #2 : Analysis through the organizational malfunctions

Instead of a general interview, now, it will be rather focused on internal communication and information malfunctions.

Case #2a

In a small enterprise, we identify a lack of feeling of responsibility and a problem with time. We identify and propose a sub-group of members to be publicly valued, thus enabling them to spread into the enterprise some needed words to improve the situation.

We bring out obviously the sex, the number of years of membership and the status, which are the main points to define the language zones and the shared language.

Case #2b

In nearly ten small identical enterprises, we analyze the responses to the queries concerning the malfunctions and their suggestions in improving the situation. We detect some adjectives, a specific verbal tense and verbally shape several beliefs.

We can summarize the average actor speech references by "information, work, people, enterprise, meetings, workshop" and the relating words, for instance, "to lack, to do, to know, to say, always". This last word is generally rare to this high frequency level and it can prove that habits and strong beliefs play an important role in these 10 enterprises.

We now conclude the more or less shared language by utilizing the crossed sorting method. The words "problems, General Manager, technical, to do, to lack" define some strong language zones. We now determine the shared language, by detailing it through typical sentences including every shared word. These sentences accompanied with actions and verbally linked, would have a gathering effect on the members of the organization. We can also think that it would be useful to deal with some speeches by putting light on the place and role of everybody. We can also use more adjectives, present and future tenses, and a less extreme expression.

Case #3 : external communication analysis

We focus now on writings sent by 4 business parks of the region "Rhône-Alpes". We put the hypothesis that a sociolinguistic identity could be a success factor.

The technique of the lexical differential allows detecting "strong" and "place" as shared language, on the opposite of "international" not at all shared. A speech about the international subject could so perhaps lengthen the "distance" and give a kind of disorder between the various business parks.



To conclude in general, we can say that this thesis considers the organization like a words system interlinked with a problems system. To change one can involve a change in the other. This approach allows gaining some new "information" about the organization and its members. The language approach puts the light on a knowledge hole about the day-to-day management, and opens the way to invisible and soft methods of action.

However, the language analysis has also some bad points. The syntax is not measured, there is some confusion about the meaning of the words, and the non-verbal dimension is totally ignored. It is a narrow approach, very unstable and, worst of all for the chief manager, it is not at all a materialistic approach. Moreover, the analysis is still slow. At last, like in every pioneer exploration, the method is still lacking experience.

But on the other side, some advantages are noticeable. The added value of the textual analysis consists in the scientific micro-statistics, with safety, automation, and rigorous procedures (Ghiglione91 p.51). Its results allow confirming or not some other types of analysis. The secret of opinion expressed is kept because all the sentences are broken into anonymous words in packs. The method is also easily reproduced in other organizations. Moreover, the verbal language is still the only mean to access to the mind representations and values of the persons working in the organization. Its advice to improve the internal communication in the organization has the advantage to be real and practical (to say) and based on precise words easy to reproduce by anybody. This sensitivity to the words allows a different approach from the classical ones. Some new ideas can appear to better understand and act in the organization.

At last, there are some perspectives about the voice directly grasped by the computer, some other ways to gather words from a group, some standard grids of lexical fields, some specific beliefs appraisal procedures, to study 100 enterprises at the same time, to measure how much is shared the general politics amidst the employees, to measure the ability to absorb new words as an indicator of the change ability, to teach students about the words used by the best managers, best sales persons, best enterprises...







Pour un audit langagier

de l'organisation





Thèse pour l'obtention du doctorat nouveau régime

en Sciences de Gestion


Présentée et soutenue publiquement

le 23 novembre 1995 par :           Richard NOLASQUE








membres du jury :

directeur de thèse :    Monsieur Jean MOSCAROLA, professeur à l'université de Savoie,

rapporteur :            Monsieur René THIEBLEMONT, maître de conférence à l'université de Savoie,

rapporteur :            Monsieur Claude JAMEUX, professeur à l'université d' Aix-Marseille II,

membre du jury :  Monsieur Humbert LESCA, professeur à l'université P. Mendès France Grenoble II,

membre du jury :  Monsieur Jacques TRAHAND, professeur à l'université P. Mendès France Grenoble II






- Des problèmes de gestion


- La communication par le langage


- De nouveaux outils d'analyse





Une question de recherche :


Quelles sont les possibilités ouvertes par une approche langagière de la gestion de l'organisation ?











1. Un concept :

le langage comme outil de gestion





2. Une technique :

la statistique textuelle






TROIS études de cas


1.1. Entre Positivisme et Constructivisme :







              théorie de l'                         théorie de la



sens pré-existant                sens construit

transmission                           génération    

réduction de l'entropie    entretien de la diversité

instructions et ordres           management participatif

questionnaires, collectes        discussions interactives










                        Le LANGAGE



1.2. Entre Signifier et Agir : le langage



Acte        Langage           Sens






L'actème en situation de gestion :



Donner du sens à un acte


ACTE                                      SENS 

"signer une vente             c'est appartenir au groupe"  

"négocier un emprunt          pour  gagner un combat"


Donner un acte à du sens


            SENS                          ACTE 

"plus de pouvoir              donc  licencier un salarié"  

"plus de performance          ce n'est pas livrer en retard"     




ACTES                               SENS

COORDONNES                    PARTAGE



1.3. un exemple : la gestion de la cohésion



Langage partagé

et zones de langage




Diagnostiquer la cohésion en observant le langage



Agir sur la cohésion par le langage












2.1. Considérer le texte comme un matériau





Ecrit pré-existant et oral stimulé par l'observateur


L'acte de langage, le nb d'occurrences représentatif des acteurs


Le champ lexical :

- l'énoncé

- l'énonciation

exemple : MOI.DIC = {ego, j, je, m, ma, me, mes, moi, mon}




L'analyse lexicale :

- avantages : rigoureux, analytique

- inconvénients : syntaxe ignorée, polysémie, non-verbal ignoré, biais dus à l'observateur...


2.2. Déterminer le langage partagé


méthode de l'écart-type/moyenne :


champ Pierre      Paul  Jacques     ...   ...   éct/moy    

moi   209   198   127   60    107   183   79    69    35%  

on    198   207   132   139   165   180   176   164   13%  


                  en nombre d'occurrences relatives


méthode des tris croisés :


champ hommes      femmes      somme des écarts 

problèmes   +5%   -10%        15   

dit         -3%         +7%         10   


                  en écart de pourcentage "chi² partiel / chi² total"


méthode du différentiel lexical : (fréq(A)/100 - fréq(C)/100)


fréquence abs.    A           fréquence abs.    C          

activités   58                activités   40         

entreprises 80                entreprises 50         


fréquence rel.    A           fréquence rel.    C          

activités   3,00%       activités   1,00%      

entreprises 1,25%       entreprises 1,72%      


diff. A - C :                

activités   2,00%      

entreprises -0,48%           


                  en différence de fréquence relative



3.1. Le contexte :


      cas         n°1         cas n°2           cas n°3    

deg. d'explo.     fort        moyen       faible     

niveau :    1 PMI 10 s. 10 PMI 100s 4 Z.A.E.   

thème :     D.O. gén.   dysfonct.   dév. éco.  

analyste :  observateur non-obs.    non-obs.   

données :   oral              semi-oral   écrit

communic. : interne     interne           externe    

anal. lex. :      éctyp/moy   tris croisés      diff. lex. 


3.2. Des exemples de résultats :


      cas n°1     cas n°2     cas n°3    

Lang.Par.   "clients, machines"     "réunions"  "lieu, force"    

ex. ZDL :   "boulot"                "PDG" "international"                    

diagnostic  les employés ne se mettent pas en position d'agir     des croyances sont sources de dysfonctionnements   l'argument géographique est insuffisant pour attirer les entreprises     

action :    prononcer leurs prénoms dans les discours contester ces croyances trouver d'autres raisons 


Les possibilités ouvertes par l'audit langagier :




            - GERER LA CULTURE


Les atouts de l'audit langagier :

niveau d'approche organisationnelle rarement utilisé

collecte facile et rigoureuse

action langagière souple et informelle

un outil pour gérer l'immatériel

Les limites de l'audit langagier :

procédure et résultats peu validés

saisie et traitement encore lent

réduction de la réalité à des discours

Des perspectives :

- Techniques :

Combiner écrit et oral ; Etudier la dynamique

- Pistes de recherche :

Dans quelle mesure le langage est-il un facteur de succés ?

Quelle est la proportion idéale de langage partagé / ZDL ?