Show Less
Open access

Linguistic Construction of Ethnic Borders

Edited By Peter Rosenberg, Konstanze Jungbluth and Dagna Zinkhahn Rhobodes

This volume focuses on the linguistic constructs involved in ethnic borders. Ethnic borders have proven themselves to be surprisingly long-lived: in nearly all European countries and beyond, border demarcation, exclusion of foreigners, and minority conflicts are some of the most persistent challenges for nations and societies. Which linguistic factors play a role in the formation of these borders, especially those drawn along ethnic lines? Which linguistic constructs contribute to the negotiation, establishment and maintenance of ethnic groups and identities? Under which conditions can processes of linguistic convergence, hybrids, or transcultural identities be observed?
Show Summary Details
Open access

Boundary, a metalinguistic concept at the core of language deformability

Helena Topa Valentim


Boundary, a metalinguistic concept at the core of language deformability

Abstract: In diesem Artikel wird das Konzept der Grenze als methodologisches Instrument zur Erklärung sprachlicher Phänomene behandelt. Der Fokus liegt hierbei auf der Betrachtung der Art und Weise, in der dieses metalinguistische Konzept eine Bezeichnungsdomäne integriert, die ein kompositorisches Konstrukt ist und deren Definition durch ein relationales Netz der Repräsentationen auf unterschiedlichen Ebenen determiniert wird. Ferner wird ihre Operationalität in der Beschreibung der linguistischen Alterität verifiziert, die direkt von der metalinguistischen Differenzierung abhängt.

Der reflexive Charakter der metalinguistischen Aktivität verlangt den Bezug auf das Repräsentationssystem, welches eine Art von symbolischer Übersetzung bilden kann. Laut der Theorie Predicative and Enunciative Operations (PEO), welche von dem französischen Sprachwissenschaftler Antoine Culioli vorgeschlagen wurde, bezieht sich die metalinguistische Repräsentation hauptsächlich auf die Konstruktion der Grenzen, um das Funktionieren von Sprache zu erklären. Dieses Model zeigt die Form einer topologischen Struktur (cf. Culioli / Declés 1981), die durch Mathematik inspiriert wird. Die Anwendung der Topologie und ihrer Terminologie zeigt einerseits ihren Status unter unterschiedlichen Methoden der metalinguistischen Repräsentation insbesondere auf kognitiver Ebene und andererseits die linguistischen Produktions- und Reproduktionsprozesse, die den Sprachgebrauch ausmachen.

Schlagworte: Aussageform, Grenze, Bezeichnungsdomäne

Keywords: Enunciation, Boundary, Notional Domain, Language Deformability


When examining the way languages work, we are led to believe that language activity is generally understood as appealing to metalinguistic representational parameters. At this abstract, cognitive level, and in an attempt to discover phenomena and to explain them, there is primarily a boundary between what is immaterial – “what we mean” – and what has materiality (verbal materiality in this case) – “what we say”. As we speak and communicate with each other, we try to match those two things, but we do so by using multiple modulations, and often with many mistakes and even with misunderstandings. ← 249 | 250 →

I start by considering the boundary as a primitive cognitive concept which shapes the way we represent the world. This way of representing the world naturally results from our perception of otherness; more specifically, it results from the phenomenon of transition, which is identifiable first of all in space-time (after Einstein, no more categories dissociated). Therefore, we have an abstract sense of boundary, taken here as a topological primitive notion corresponding to an operational concept in the framework of Predicative and Enunciative Operations (PEO), proposed by the French linguist Antoine Culioli. Through this theoretical framework, Culioli proposes a metalinguistic representation in order to explain the global functioning of language.

“Enunciate” / Utterance and language deformability

The understanding of how language works allows us to identify the concept of “enunciate”, a term with which we can designate linguistic sequences. As Culioli states (2002, p. 27), the choice of the term “enunciate” instead of the English term “utterance” can be justified. “Enunciate” is an old term. Seneca translates the term lekton (what the Stoics called to something incorporeal, i.e., something without materiality) with the word enuntiativum. After the Middle Ages, there was the relationship between the dicibile (the speakable) and the dictum (the said). The Middle Ages scholars conceived a dictum and a modus, that is, a sort of thought content or propositional content. Therefore, we have the conception of an immaterial representation, an abstraction. The term enuntiativum, employed by Seneca to explain what lekton is, has enuntiare as the verb, with the meaning of “making out” or “to bring up”. That’s why “enunciate” assumes the significance of a transition from something “speakable” to something “said”. The concept of “enunciate” includes this sense corresponding to something which is not a priori; on the contrary, it is constructed by an “enunciator” and reconstructed by a “coenunciator”. In English the available word is “utterance” (which, for simplicity, we will use from now on). Utter means “to externalize” (etymologically, ut = out), but with an emphasis on the agent. In contrast, “to enunciate” puts the emphasis on the enunciating act, on the construction or the production of something.

Consequently, the enunciator responsible for the utterance’s production is not identified with the speaker or the sender. The enunciator is the subjective origin, which results necessarily as an intersubjective entity. Thus, underlying any statement, there is always an enunciator and a coenunciator as theoretical entities – not as flesh and blood entities. In contrast, pragmatically speaking, the speaker and the sender (in the sense of Jackobson’s proposal) are individuals and flesh and blood entities – in the specific case of the sender, with a focus on coding. ← 250 | 251 →

We can therefore look in this manner at language activity: conceiving the relationship between something abstract and inaccessible because of its immateriality, and something accessible and materialized. In other words, language operates between something cognitive or mental, and an utterance. In linguistic activity, we move from the “speakable” to the “said”; in order to understand the way it functions, we raise at least two questions: about the way one operates this transition, and about the way we can manage the boundary between these two dimensions.

Thus, the enunciator is the agent, the responsible entity for this transition, and the one who constructs meaning. However, as the constructed meaning is also reconstructed, it would be interesting to realize the way this non-symmetric mechanism works between both enunciator and coenunciator – the fact that, very often, there is no perfect understanding or transparency on this issue is due to this asymmetry. Benveniste (1966) theorizes about the “formal apparatus of enunciation”, but the question on how a subjective activity can be intersubjective remains without a conclusive answer.

When speaking about intersubjectivity, we speak of the otherness inscribed at the core of linguistic activity, which lies at the boundary between entities in their radical individuality. As many authors have stated, the explanation for the constant modulation that characterizes any language lies in the concept of intersubjectivity. An utterance is always modulated. These modulations acquire the form of a game of grammatical categories, which obtain different values of referential determination. Thus, we have a subjective game, a mode game, a time game, and even an intonational game.

Look for instance at this interactional sequence of utterances (adapted from Culioli 2002, pp. 219–220):


A-  In gyms people don’t really practice sports. They run, they stretch, they jump... People who really practice any sport make their exercises in a judicious and targeted manner.

B   Yes, but why not say that running, stretching and jumping the way people do in gyms is already practicing sports? After all, it is to stir. Don’t you think it is better to stir a little bit than not doing anything?

The construction of the different values happens through a number of modulations. This is a simple example of how we are always facing representational deformability, discontinuities, and heterogeneous phenomena, and this sequence illustrates exactly what enunciation is: namely, something other than strict syntax, semantics or pragmatics. As a construct resulting from intersubjective modulation and adjustment, enunciation crosses all these dimensions, enunciation is transcategorical: it ← 251 | 252 → is simultaneously prosodic, syntactic, semantic and pragmatic. In this sense, the heterogeneity of the intersubjective modulation or adjustment phenomena cause the barriers (boundaries) of the traditional linguistics sub-disciplines to fall.

The focus of our interest should fall on the effects of non-coincidence – on the necessary adjustments between speakers – because through linguistic activity we operate a transition from a non-material representation to a material representation. In other words, through linguistic forms, we operate a transition from an abstract and cognitive representation to the visibility and tangibility of linguistic forms.

The explanation of how this “embodiment” allows an adjustment between the speakers requires an anthropological linguist’s perspective. It demands that we consider all interactional phenomena, not excluding the prosodic and pragmatic phenomena. Thus, on the basis of transcategoriality, we can examine another boundary: the one often established between what is linguistic and what is not linguistic – or extra-linguistic. Considering that referring is one of the language activity dimensions, the question is to know whether reference is external to linguistics. Any answer has to consider the fact that reference to the situation is not foreign to language activity, as it will contribute to the production and the recognition of utterances. In order to prove this, we only have to look at how the big questions of reference are internal to language activity – for instance, something exists or doesn’t exist, something is set or isn’t set, etc.

Culioli, in a published interview (2002, p. 73), compared reference to what happens when you show children a movie and ask them to tell you what they saw. On the one hand, children know that what they’re telling you didn’t actually happen: it’s a film. On the other hand, there is always “the illusion of the film”: they narrate it as if the events actually happened. We can recognise it in the temporal values attached to the linguistic forms they construct. In fact, in terms of what language allows us to refer: both, something that happened (a trip that I made to Paris, for instance) and something that did not happen (a film or a dream that I report) are referential. From the linguistic point of view, both are real events because they refer to representations that don’t have the status of reality. Therefore, we always speak about the visible and the invisible, taking into account that everything is relevant to language. The linguistic forms are by definition constructed on those issues that we tend to consider not being linguistically relevant, such as inference or, as Culioli (ibid.) expresses it, what we designate as “common and shared knowledge”.

Moreover, when we consider the transcategorical nature of the study of language, we study the complex relationship between what is inside the text (the intratextual relationships) and the situational context. The fact that language is not a homogene ← 252 | 253 → ous concept proves this complexity. As a representational activity, language allows us to build representations. As a referential activity, language allows us to construct reference (i.e., it refers to what “we speak about”). On the other hand, language is a regulatory activity; thus, in addition to the regulations that are internal to language (meaning that all languages are characterized as both stable and deformable), there are regulations between speakers. We as speakers always want to lead the other to the representations that we aim at, and in doing so incur the risk of failure, and for that reason we resort to any necessary adjustments in order to avoid this failure.

Linguistic forms play a key role in this intersubjective game of building-rebuilding representations and reference. Linguistic forms are markers of cognitive and linguistic operations simultaneously, and they underlie the utterances and referential values. For instance, the use of the indicative or the subjunctive mood in propositional complements of subjective verbs (as to believe or to think) is a case of this cognitive and linguistic operations marking. In Portuguese, as in other Romance languages, the opposition indicative / subjunctive marks intersubjective values, described as modality values.

In the following utterance, the use of the indicative mood (telefonou) in the subordinate clause indicates the enunciator’s belief. The enunciative distance naturally implicated by any belief (which is not a certainty) is marked by the verb (acredito) in the main clause.

(2)  Acredito que a Ana telefonou indicativo (I believe that Ana called.)

In the utterance (3), the use of the subjunctive mood in the subordinate clause (tenha telefonado) indicates that the enunciator drifts away from his belief.

(3)  Não acredito que Ana tenha telefonado conjuntivo (I don’t believe that Ana called.)

In a negative interrogative sentence like (4), the use of the indicative mood in the subordinate clause is oriented to what the enunciator believes, and simultaneously to what he believes his interlocutor should believe in. Consequently, it expresses the enunciator’s belief, presupposing “I believe”. That’s exactly the reason why this sentence has a controversial interpretation.

(4)  Não acreditas que Ana telefonou indicativo? (Don’t you believe that Ana called?)

It is very common that a small linguistic mark – even a simple intonation mark – is enough to trigger various types of operations and, consequently, to introduce a relevant change.

(5)  Foi um belo dia. (It was a wonderful day.)

(6)  *Se foi um belo dia. (literally: If it was a wonderful day.) ← 253 | 254 →

(7)  Ah! Se foi um belo dia! (literally: Oh! If it was a wonderful day! – which in English would be “Oh! It was really a wonderful day!”)

We observe that the first utterance (6) isn’t acceptable, whereas the last utterance (7) is acceptable. Through the ascending intonation mark, (7) corresponds to the determination of a high degree of value for the property of being “um belo dia” (a wonderful day). Consequently, it acquires an appreciative modal value. This same linguistic construction, without this specific intonation characterization (as in 6), wouldn’t have this modal value: it would be an assertion and wouldn’t be acceptable.

Notional Domain: a metalinguistic representation

Heterogeneity is inherent to any language, and represents a challenge to developing theories about it. If heterogeneity of phenomena arises primarily, any theory must address the problem of how this transition from homogeneity to heterogeneity operates. Through this process, you are on the way to finding some form of homogeneity. Therefore, theorization shows nothing more than a “deformable consistency” (Culioli 2002, p. 88). When we emphasize this deformability, we choose a specific epistemological perspective consisting of observation of the data, identification of the relevant problems, and consequently, conception of a descriptive and explanatory reasoning. The linguistic facts reveal that there is always a “subjective grammar”, a set of lexical and adjustment operations. Therefore, it is essential to adopt a scientific paradigm based upon the classical logic. Actually, linguistic phenomena cannot be reduced to a binary perspective or to an exclusivity of values, as conceived by the trait theory of structural grammar. But the boundary between those apparent exclusive values is not something dimensionless. So it is evident that there is a need for devising a gradient which enables “a more or a less value”. It enables, more specifically, the existence of some values which are neither one thing nor the other: that is to say, some values which do not correspond to a certain property any longer, but also do not correspond yet to another property. This gradient becomes an appropriate metalinguistic feature to account for the deformability of cognitive representations.

The next utterances show empirically how this proposal of a dimensionless boundary is absolutely operative in a metalinguistic reasoning.

(8)  John is tall.

(9)  John is very tall.

(10)  John is really tall!

(11)  John is tall, tall!

(12)  John is a tower! ← 254 | 255 →

(13)  John is not really tall.

(14)  John is kind of tall.

(15)  John is not tall.

(16)  You can’t say that John is tall, but you can’t say that he is small.

(17)  This is not a house, this is a hole.

(18)  This one, yes, is a real house / one of those houses!

(19)  This is not a house; this is the house!

(20)  This is not really a house even if it has everything to be it.

In both groups of utterances – from (8) to (16) and from (17) to (20) – we’re dealing with the cognitive representations of what we lexicalize as “tall” in the first group, and as “house” in the second. The modulations and the consequent deformability are introduced through small linguistic marks. The stability in all these utterances is definable in terms of each of these notions: tall and house correspond both to representations constructed culturally, and simultaneously dependent upon the enunciator’s individual experience. But we can recognize above all a relevant modulation, explainable by operations, both of language and cognitive order.

Many of these empirical facts cause complex problems in the study of languages. Structuralism worked on the “all or nothing”, hence with a clear defined boundary with no extension. However, as a matter of fact, and as these empirical data show, we don’t have a two-value system. In these utterances, we have the “really tall” value (10), but we also have the “not really tall” (13) and the “kind of tall” (14) values.

We can understand this different configuration of boundaries through empirical data from texts produced in an interaction on cooking, as proposed by Culioli (2002, pp. 217–218), recalling Claude Lévi-Strauss (Le cru et le cuit 1964 / The raw and the cooked, 1969).

Culioli begins by referring to the fact that cooking is a cultural issue, a question of representation cognitively inscribed in our minds. As a representational construction, cooking involves two states: the “unbaked” and the “baked”. A potato, for instance, is unbaked before baking, but from the moment it begins to bake, it is no longer strictly unbaked, but it is not yet baked. During baking, the potato becomes more and more baked. So, when a potato is baking, the more baked it is and the less unbaked it becomes. Thus, there is a moment in which it is still not fully baked, even if it is no longer truly unbaked. After a while, there is a moment when it’s already baked. We consider the process to be unidirectional because, according to our knowledge, it is no longer unbaked at all, and once it is baked, it is definitively baked. According to our cultural knowledge, baked is an irreversible state: we don’t have something like “*debaked”, as we have, for instance, “defrosted”. ← 255 | 256 →

The relevant piece from this empirical example pointed in Culioli (2002, p. 218) is that, as a consequence of this process, we have (exactly when baking hasn’t started) “unbaked” at the beginning, and then at the end of baking we say “it’s baked”. Thus, there is one last moment when it is still unbaked, even in to a minimal degree, the same way there is a moment when we declare “it’s baked”. So we build a state at one pole (unbaked), and on the other pole we build another state (baked). The condition of being baked results from a process that includes an intermediate state with a first point in which it isn’t already totally unbaked, and one last point in which it is still not truly baked.

Using this empirical situation, Culioli demonstrates that we cognitively build a first point of “baking” and a last point of “unbaking”, but in between these two values we construct a boundary which integrates a gradient, and which includes both attractors (unbaked / baked) and the values of less and less unbaked and more and more baked.

In this case, the complementary arises in a completely different way than in the case of systems having two values, where each value is necessarily the opposite of the other. Within this dispositive, the complementary of “unbaked” can be “baked”, but can also be the boundary of “baked”. In other words, the complementary of “unbaked” is defined as the time from when it is no longer unbaked (in the strict sense), and when it is also not yet baked.

Notion of UNBAKED (adapted from Culioli 2002, p. 118)


We could take into account the case of the complementarity between unbaked and baked, and incurring the recognition of binary opposition just as the structuralism did. However, we can consider the complementary in a way that it corresponds to “unbaked” and the boundary to the other value. In some cases, we can even take into account the boundary isolating it as we wish in relation to “unbaked and / or baked”. For instance, “The spaghetti is al dente” (al dente being the point of baking pasta (“pasta al dente”) when it is baked “perfectly”, offering a slight resistance to the bite).

As Culioli acknowledges, we realize that in order to describe these phenomena, a classification isn’t enough. It is necessary to make a calculation. These are phenomena that have led Culioli, with the help of Jean Blaize Grize (in logical-mathematical aspect) and François Bresson (cognitive psychology), to introduce ← 256 | 257 → a topological structure which corresponds to a primitive cognitive devise. All notions (of unbaked, baked as well as any other notion like home, coming, eating, etc.) are representable by a domain – a Notional Domain – with topological properties. In addition to the dimensioned Boundary, the Notional Domain consists of an Interior and an Exterior. As the notion is /P/ (here as a variable), the Notional Domain is (p, p ‘)1.

The Notional Domain allows one to distinguish topologically what is internal (what belongs to it), what is external (what does not belong to it – that is, the alterity), and what is on the boundary of this domain. As a topological structure, the Notional Domain is representable in space coordinates as having an Interior, an Exterior and a Boundary.

The Interior zone (I) is where what is truly /P/ (i.e., with all the properties /P/) is situated.

In both utterances, (8) and (9), the enunciator builds a linguistic occurrence of the notion /TALL/ located within the notional domain.

(8)  John is tall.

(9)  John is very tall.

The Interior defines an open area that is organized around an Organizing Centre which functions as the Attracting Centre, where the notion’s high degree properties are located. Thus, the Organizing Centre is where you find the Type Occurrence of a notion.

In the following utterances (10), (11) and (12), the enunciator makes use of various linguistic resources to build linguistic occurrences of the same notion, /TALL/, located precisely in relation to the Attracting Centre. Consequently, these linguistic occurrences have the value of high degree. Producing these utterances, the enunciator says that “John” (the enunciate subject) has all the properties inherent to “being tall”.

(10)  John is really tall!

(11)  John is tall, tall!

(12)  John is a tower!

The same thing happens in utterances (18) and (19), with the difference of being related to another notion, /HOUSE/ also presenting different linguistic resources in order to construct the high degree. For instance, in utterance (19), the definite ← 257 | 258 → article is the mark of this identification operation between the linguistic occurrence and the notional Attracting Centre.

(18)  This one, yes, is a real house / one of those houses!

(19)  This is not a house; this is the house!

In the Interior of the Notional Domain, all the notion’s occurrences are indiscernible, i.e. identified one with the other through the relation they all have with the Type Occurrence. So, it is not possible to establish a final occurrence. This metalinguistic reasoning is marked by the indefinite article in (19) – “a house”.

Turning back to the example of cooking, it is interesting to show that, in order to mean “unbaked” before even the slightest change, we say something like:

(21)  It’s unbaked, really unbaked.

The possibility of an utterance like (21) means that we can construct a value through a self-localization, that is to say, referring to “what is unbaked, justifiably referred to as unbaked”. This leads to an inaccessible representation, which can only be said and cannot be shown which corresponds precisely to the construction of the high-degree.

Continuing the topological explanation of the different values that we can associate to a linguistic construction, we can define a Gradient in the Interior of the notional domain, through which the degree of approximation or detachment of the linguistic occurrence in relation to the Organizing Centre is regulated. This situation is illustrated in utterances (13) and (14):

(13)  John is not really tall.

(14)  John is a bit tall.

However, the Notional Domain also includes the Exterior zone which contains what is not truly /P/ concerning the construction of a linguistic complementary, i.e. everything that doesn’t show the inherent properties of the notion. Therefore, the established relations between the values of a notion in its Interior, and the values in the Exterior, are complementary.

(15)  John is not tall.

(17)  This is not a house, this is a hole.

In (15), the fact of excluding the properties of the notion /TALL/ implicates the construction of its complementary notion: if “he is not tall”, then it is presupposed that “he is small”. In (19), the enunciator constructs the complementary of something “not being a house” as being “a hole”. The complementary is, in these cases, linguistically constructed. ← 258 | 259 →

Utterances (16) and (20) illustrate the construction of an occurrence in the Boundary zone of the Notional Domain, the zone containing what is no longer /P/ and yet is not really /not-P/.

(16)  You can’t say that he is tall, but you can’t say that he is small.

(20)  This is not really a house even if it has everything to be it.

From a topological perspective, this issue is also applicable to the metalinguistic representation of the linguistic occurrences which are constructed in relation to a time axis (cf. Dufaye 2009, pp. 109–110).

(22)  I have lived here for 5 years.

(23)  I have lived here since my son was born.

In utterances like (22) and (23), how can we represent the boundary between the linguistic event (“have lived here”) and the time interval specified by the adverbial? We would say that, once again, it depends on the point of view: in the first case, the boundary belongs to the linguistic event “have lived here”, whereas in the second case it belongs to the time interval specified by the adverbial “since my son was born.”

Final remarks

One of the most crucial characteristics of this metalinguistic device lies exactly in the design of the boundary, not as a dimensionless threshold, but as having a structure itself and therefore also having a dimension. The result is a more complex system that mainly helps to explain the diversity of specific configurations in the utterances’ determination, thus providing some metalinguistic representations that help to build the “story” of linguistic representations. Therefore, it doesn’t deal with objects as isolated entities with clear-cut boundaries. In this sense, it is a procedure of abstraction: the objects are abstractly constructed, acquiring a theoretical status and showing how language refers to itself and not to an extra-linguistic object. The utility of this approach also lies in the fact that it reveals language to be more complex than a compositional analysis proposes. One starts from the principle that, in order to build utterances, we are all provisioned with the same device of construction-deconstruction, which allows us to identify linguistic units as traces of operations. It also reveals that we all have a stock of common operations at our disposal, and that is why we can re-construct what has been produced by others.

Therefore, there is a mental activity to which we can give a kind of crude representation, but we’ll always face difficulties in realizing it. As Culioli suggests, ← 259 | 260 → we know that what happens in and through language is not linear; it can only be represented in a very complicated geometric organization. Through linguistic activity, we’re always building a certain type of object that is provided with formal properties. This object will be anchored in a reference space and be subjected to a permanent intersubjective adjustment.

This perspective also allows us to deal with another type of boundary: the diversity of languages. Investigating the diversity of languages, and trying simultaneously to build a theory of language, involves addressing the problem of the comparability between languages, or always looking for what is common among their diversity. Addressing the problem of comparability between languages necessitates a higher level of abstraction. Each language is unique even if we can translate it. Thus, because of its comparability, it is possible to find a set of operations that exists in all languages. On the one hand, we have the lexical and grammatical notions based on operations; on the other hand, we have the characteristic markers of a given language. What varies from language to language is the correspondence between markers and operations which are specific to a given language.

Therefore, a language is neither a code nor a nomenclature. We deal with procedures, categorization processes, representation and adjustment resources, all defining a deformable system. This deformability means that there are no finished utterances. Every utterance is produced by an enunciator, such that the coenunciator will by himself reconstruct representations through the markers. The coenunciator never avoids that the phenomena such as polysemy, ambivalence, vagueness, and therefore an eventual misunderstanding can occur.

Such a view on the functioning of language leads us to some interesting questions, particularly about how language relates to reality. According to this reasoning, the question of the outside and the inside, and thus the concept of the Boundary is put another way, and as with any other categories, it cannot be forcibly put into the perspective of “all or nothing”. This kind of approach leads to an increase in significance, as it is contrary to the ancient linguistic rationality that doesn’t address the issue of the unsayable, because it doesn’t consider the relevance of intersubjectivity and adjustment.

In this sense, and also due to this idea of Boundary as a metalinguistic concept with dimension, language is unlimited. Because of the deformability principle, we can theoretically say anything, even the unsayable, and there is never a final word. It is this operation is that allows for the existence poetry, and that reveals the system’s complexity which results from constant adjustments. ← 260 | 261 →


Benveniste, Émile: Problèmes de linguistique générale. Volumes 1 and 2. Gallimard: Paris 1966.

Culioli, Antoine: Variations sur la linguistique. Entretiens avec Frédéric Fau. Klincksieck: Paris 2002.

Culioli, Antoine: Pour une linguistique de l’énonciation. Volume II. Collection L’Homme dans la Langue, Ophrys: Paris 1990.

Culioli, Antoine / Desclés Jean-Pierre: Systèmes de représentation linguistiques et métalinguistiques. Coll. ERA 642, Université Paris VII, DRL, 1981.

Culioli, Antoine et al.: La théorie d’Antoine Culioli. Ouvertures et incidences. Éditions Ophrys: Paris 1992.

Declès, Jean-Pierre: «Quelques opérations de localisation». Projet interdisciplinaire de traitement formel et automatique des langues et du langage (PITFALL) 15 and 16, Université Paris 7: DRL, 1975.

De Vogüe, Sarah: «La langue entre discours et cognition». In: Actes du colloque international pour les 10 ans de la MSHS de Poitiers. Université Paris X – Paris Ouest Narterre La Défense / CNRS: UMR 7114, 2011.

De Vogüe, Sarah: «Construction d’une valeur référentielle: entités, qualités, figures». La référence 2, Travaux linguistiques du CERLICO 12, 1999, pp. 77–106.

Dufaye, Lionel: Théorie des opérations énonciatives et modélisation. Ophrys: Paris 2009.

Franckel, Jean-Jacques: «Situation, contexte et valeur référentielle». Pratiques 129/130, 2006, pp. 51–70.

Valentim, Helena T. / Lejeune, Pierre: „Contexto ou condição de interpretação do enunciado“. Cadernos WGT. Centro de Linguística da Universidade Nova de Lisboa – Faculdade de Ciências Sociais e Humanas: Lisboa 2009, pp. 3–11.

1  In (p, p’), p represents what is /P/ (the validation zone); p’ represents what is not /P/ or is /non-P/ (the complementary or non-validation zone).