Text Linguistics and Classical Studies Dressler and de Beaugrande's

UNIPA Springer Series
Mauro Giuffrè
Text Linguistics
and Classical
Studies
Dressler and de Beaugrande’s
Procedural Approach
Editor-in-chief
Carlo Amenta, Department of Economics, Management and Statistics Sciences,
University of Palermo, Italy
Series editors
Sebastiano Bavetta, Department of Economics, University of Palermo, Italy
Calogero Caruso, Department of Pathobiology, University of Palermo, Italy
Gioacchino Lavanco, Department of Psychology, University of Palermo, Italy
Bruno Maresca, Department of Pharmaceutical Sciences, University of Salerno,
Italy
Andreas Öchsner, Department of Engineering and Information Technology, Griffith
University, Australia
Mariacristina Piva, Department of Economic and Social Sciences, Catholic
University of the Sacred Heart, Italy
Roberto Pozzi Mucelli, Department of Diagnostics and Public Health, University of
Verona, Italy
Antonio Restivo, Department of Mathematics and Computer Science, University of
Palermo, Italy
Norbert M. Seel, Department of Education, University of Freiburg, Germany
Gaspare Viviani, Department of Engineering, University of Palermo, Italy
More information about this series at http://www.springer.com/series/13175
Mauro Giuffrè
Text Linguistics
and Classical Studies
Dressler and de Beaugrande’s Procedural
Approach
123
Mauro Giuffrè
Department of Human Sciences
University of Palermo
Palermo
Italy
ISSN 2366-7516 ISSN 2366-7524 (electronic)

ISBN 978-3-319-47930-9 ISBN 978-3-319-47931-6 (eBook)
DOI 10.1007/978-3-319-47931-6
Library of Congress Control Number: 2016955076
© Springer International Publishing AG 2017

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part
of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission
or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar
methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are exempt from
the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authors or the editors give a warranty, express or implied, with respect to the material contained herein or
for any errors or omissions that may have been made.
Printed on acid-free paper
This Springer imprint is published by Springer Nature

The registered company is Springer International Publishing AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Contents
1 Alle Wege Führen Zum Text . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

The Text in the Beginning: Structural and Generative Linguistics . . . . . 2
Structural Linguistics. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Generative Linguistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
The Text from a Logical Point of View: the Linguistic Turn . . . . . . . . . 15
Generative Semantics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
Model-Theoretic Semantics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
The Text in Text Linguistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
Text Linguistics and Its Domain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
The Development of Text Linguistics . . . . . . . . . . . . . . . . . . . . . . . . . 26
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
2 The Procedural Approach to Texts . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
The Point of View of Dressler and de Beaugrande . . . . . . . . . . . . . . . . . 36
The Scope of Cognitive Science . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
Cognitive Psychology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
Neuroscience and Artificial Intelligence Research . . . . . . . . . . . . . . . . 40
Linguistic and Philosophical Research . . . . . . . . . . . . . . . . . . . . . . . . 41
Procedural Semantics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
The Procedural Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
Text Production and Reception . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
Communicative Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
Regulative Principles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
The Concept of Textuality in the Procedural Approach:
Seven Criteria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
Cohesion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
Coherence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
Intentionality and Acceptability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
Informativity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
v
vi Contents
Situationality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
Intertextuality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
3 The Procedural Approach to a Text . . . . . . . . . . . . . . . . . . . . . . . . . . 75
Rhetoric, Cognitive Psychology and Ancient Literary Genres . . . . . . . . . 76
Titus Livius and the Gallic Sack of Rome . . . . . . . . . . . . . . . . . . . . . . . 79
Cohesion and Coherence in Livius, Ab Urbe Condita V, 44 . . . . . . . . . . 89
Cohesion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
Coherence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
Conclusions of This Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
Chapter 1
Alle Wege Führen Zum Text
Towards an empirically motivated

grammatical theory of verbal texts.
János Sándor Petőfi, Studies in Text Grammar
In a naive sort of way, the concept of text and its definition are already rather clear.
In fact, everyone has an intuitive concept of what a text is. However, like other
linguistic concepts that have the same sort of spontaneity (e.g., sentence or utter-
ance) the concept appears to be less than precise. In common usage, as in the
non-specialized scientific disciplines, the term is mostly used to refer to linguistic
utterances. In this restrictive sense, a text is the physical manifestation (written or
oral) of a message sent by an issuer to one or more recipients so that it can be
subjected to interpretation and then understood. However, the complexity of the
definition increases significantly in a specialized dictionary, such as A Dictionary of
Linguistics and Phonetics:
A pre-theoretical term used in LINGUISTICS and PHONETICS to refer to a stretch of
language recorded for the purpose of analysis and description. What is important to note is
that texts may refer to collections of written or spoken material (the latter having been
transcribed in some way), e.g. conversation, monologues, rituals and so on. The term
textual meaning is sometimes used in semantics as part of a classification of types of
MEANING, referring to those factors affecting the interpretation of a SENTENCE which
derive from the rest of text in which the sentence occurs - as when, at a particular point in a
play or novel, a sentence or word appears whose significance can only be appreciated in the
light of what has gone before. The study of texts has become a defining feature of a branch
of linguistics referred to (especially in Europe) as textlinguistics, and ‘text’ here has central
theoretical status. Texts are seen as language units which have a definable communicative
function, characterized by such principles as COHESION, COHERENCE and informa-
tiveness, which can be used to provide a FORMAL definition of what constitutes their
identifying textuality or texture. On the basis of these principles, texts are classified into
text types, or genres, such as road signs, news reports, poems, conversations, etc. The
approach overlaps considerably with that practised under the name of DISCOURSE
analysis, and some linguists see very little difference between them. But usage varies
greatly. Some linguists make a distinction between the notions of ‘text’, viewed as a
physical ‘product’, and ‘discourse’, viewed as a dynamic process of expression and
interpretation, whose function and mode of operation can be investigated using
PSYCHOLINGUISTIC and SOCIOLINGUISTIC, as well as linguistic, techniques.
© Springer International Publishing AG 2017 1

M. Giuffrè, Text Linguistics and Classical Studies,
UNIPA Springer Series, DOI 10.1007/978-3-319-47931-6_1
2 1 Alle Wege Führen Zum Text
A similar distinction sees ‘text’ as a notion which applies to SURFACE STRUCTURE,

whereas ‘discourse’ applies to DEEP STRUCTURE. From the opposite viewpoint, some
linguists have defined ‘text’ as an abstract notion, ‘discourse’ being its REALIZATION.
Apart from these theoretical distinctions, there is also a tendency for texts to be thought of
as monologues, usually written, and often very short (e.g. no through road), whereas
discourses are often thought of as dialogues, usually spoken and of greater length. (Crystal
2006, dictionary entry TEXT)
When one tries to define the ‘text’ object in a scientific manner, the complexity
of the definition increases; it also varies according to the discipline to which the
definition refers. Although an integrated framework is still lacking that allows the
complex object called ‘text’ to be investigated in a firmly established multidisci-
plinary perspective, linguistics is the discipline that has taken on the task of pro-
viding this definition because, since they are linguistic artefacts, texts belong to the
field of analysis of this discipline. For many years, which scientific discipline
should deal with texts was not established, and only in relatively recent times has
linguistics taken up this task. Indeed, at first it did not recognize the text as a subject
of analysis.
Modern linguistics, which developed in Europe and the United States, consid-
ered the sentence as the maximum level of analysis. Progressively, the domain of
linguistic research expanded to include the text on an ongoing basis, and in the
twentieth century in Europe and the United States, text linguistics was configured as
a separate discipline.
The Text in the Beginning: Structural and Generative

Linguistics
Scientific linguistics did not establish the text as its object of study at the outset; it
concerned itself with sentences as the highest level of analysis. Therefore, in
accordance to the maximum level of analysis that it was pursuing, it could be called
sentence linguistics. As regards the methods used for analysis, these can be divided
into structural and generative linguistics.
Structural linguistics analyses the expressions of a natural language according to
different levels of minimum units. The levels are the phonological, morphological,
syntactic and semantic, which pertain, respectively, to the following minimum
units: “phoneme”, “morpheme”, “syntagmeme” and “sememe”. Generative lin-
guistics on the other hand aims to create well-formed expressions in a natural
language thanks to a system of rules of re-writing that work recursively. In a
narrower sense, a generative grammar is a formal mechanism that operates from a
four-part base (“alphabet of terminal symbols”, “alphabet of nonterminal symbols”,
“axioms of grammar” and “production rules”) to produce phrases.
The first type of linguistics has its celebrated πρῶτος εὑρετής in Ferdinand de
Saussure, and finds its greatest exponents in scholars such as Benveniste and
The Text in the Beginning: Structural and Generative Linguistics 3
Hjelmslev. Outside of Europe, structural linguistics developed in the course of the

twentieth century in the United States, too, where it gave birth to generative
linguistics.
Structural Linguistics
In Europe, the founder of what came to be called structural linguistics was

Ferdinand de Saussure. Although he was a refined comparativist, as is clear from
his fundamental work Mémoire (1879), and a tenacious student of anagrams,
Ferdinand de Saussure owes his fame to Cours de linguistique générale (1916),
published by his students Charles Bally and Albert Sechehaye. Besides a few
interpretation problems that remain open today, all the textbooks and manuals of
history of linguistics present this author as the most representative, but, like all
fathers of the various disciplines, de Saussure is universally seen today as an
inimitable model and as the source of all the limits of linguistic research.
The role of pater linguisticae that is unanimously attributed to him stems from his
idea of language, which we would like to explain by using a metaphor taken from a
game: he himself compared language to a game of chess (a game which de Saussure
loved). In this game, the various pieces do not have an absolute value, but rather a
value based on the fact that each is different from another and performs and functions
differently within the game. Some pieces are interchangeable (for example, a red
piece may be replaced by a blue one), but the game continues as long as the pieces
constitute a valid configuration within the game. Therefore, the game continues as
long as the rules are not broken; conversely, the game ends instantly.
Metaphors aside, the game is language, the pieces are the minimum units (on the
various linguistic levels) and the rules are the grammar of a natural language. The
great merit of de Saussure is to have conceived of a concrete idea concerning the
study of languages and to have maintained that language could be studied as a
system. To sum up, his conception can be called structural linguistics.
The core of the Genevan linguist’s thought is summarized in well-known text-
books and manuals through a few famous dichotomies, signifiant versus signifié,
langue versus parole, syntagmatic versus associative, synchronic versus diachronic.
The first dichotomy is internal to the linguistic sign. The linguistic sign is
constituted by the union of “concept” and “acoustic image”; the relationship
between these two parts is arbitrary (and not necessary). In fact, there is no reason
why the sound image of the word tree should refer to the concept ‘tree’ better than
the sound image of arbor. It should be noted that all segnic models elaborated later
are dependent on some form of “Saussurism”.1
1
For example, both the model of signs developed by Hjelmslev and that of Petőfi have this
characteristic in common. However, only Petőfi asserts that the communicative situation is
determined both by the signifiant and the signifié.
The second dichotomy is perhaps the most controversial. Indeed, it has been
explained as both the opposition between abstract and concrete, and between code
and speech act, and as between social and individual dimensions. The last of these
ways of understanding the dichotomy appears particularly interesting because in
Saussure’s vision, langue falls fully into the realm of linguistic research, while the
parole is excluded.2
The third dichotomy opposes “syntagmatic” to “associative” and refers to the
fact that the various linguistic elements can relate to each other in different ways
within a communication; this ability has a dual nature. An initial relationship, stated
in praesentia, regards the terms that cannot be present simultaneously within the
same word phrase, because a given substance cannot be qualified simultaneously by
both elements (for example, the adjectives “black” and “white” cannot be used
simultaneously in relation to the same substance, because skin is either white or
black). The two adjectives given as example may, however, be unrelated in
absentia, because white and black evoke each other (one being the opposite of the
other).3 This dichotomy can be extensively applied to the various levels of the
language system (phonetic and morphological as well as syntactic and lexical), but
it remains difficult to fit it completely within the conceptual system developed by
Saussure.4
The fourth and final dichotomy opposes synchrony and diachrony. This dis-
tinction intends to highlight the distance between the study of language as a system
and the study of the evolution of a natural language over time. With this dichotomy
de Saussure gave linguistic science a revolutionary formulation for his time, though
it now appears “classic”. Indeed, linguistics had been placed until then in the
domain of the “human sciences”, but now became part of the domain of the “hard
sciences”. It should be remembered that at the time of Saussure’s teachings, the
only form of linguistics that could consider itself scientific was historical and
comparative linguistics (to which Saussure himself gave a fundamental contribu-
tion, especially in his Mémoire). In contrast to the mainstream, Saussure’s ideas
brought the synchronic perspective to the fore. This approach clearly demonstrates
Saussure’s desire to place language within the domain of the Exact Sciences
(Naturwissenschaften) rather than define it as part of the Humanities
(Geisteswissenschaften). This has been consided a reflection of both the linguist’s
2
Chomsky, as we shall see later, founded his own reflection on the dichotomy competence versus
performance, which often has been associated with the opposition of the Genevan linguist.
Competence in Chomsky’s sense, however, is universal and innate in man; langue in Saussure’s
sense, on the other hand, is social and historically determined.
3
The term “associative” was replaced by Hjelmslev with the term “paradigmatic” in an attempt to
eliminate the halo of psychology inherent in Saussure’s term.
4
The extensive application of this Saussurian dichotomy allowed Roman Jakobson to classify a
number of phenomena in the categories provided by de Saussure: figures of speech
(metonymy = syntagmatic, metaphor = paradigmatic), the literary genres (prose = syntagmatic;
poetry = paradigmatic), aphasic disorders (linear order of words = syntagmatic; choice of
words = paradigmatic).
family tradition, which contained noteworthy exponents in the fields of chemistry

and physics, and his early studies in these fields (Hjelmslev’s father, too, as we shall
see below, was devoted to geometry). Beyond the question of the origin of this
perspective, the historiography of linguistics conserves the fact that de Saussure
promoted this perspective in the interest of making the study of language scientific.
This perspective is in fact a starting point for the diachronic study perspective,
because a study of the evolution of language can be made only by having a notion
of the functional categories that organize a language into a system.
The evaluation of this dichotomy is particularly dependent on the “position” that
scholars assume today with respect to linguistics. First, it is important to point out
that one cannot avoid taking a position, whatever it is, with respect to this issue.
Even scholars of the school of Prague consider the “synchrony versus diachrony”
opposition too rigid. In our view, the synchronic point of view seems intrinsically
linked to the diachronic perspective, so much so that one is inseparable from the
other. It should certainly be kept in mind that historiographical paradigms of
twentieth-century linguistics can be drawn only by giving a value to this opposition
and only in the light of the position taken with respect to this dichotomy.
Deeply convinced of the value of the exact sciences, Ferdinand de Saussure
suggested that the scientific study of language should resemble the study of
chemistry and physics more than that of history. As long as this is the prevailing
perspective in linguistics, linguists will continue to identify themselves as heirs to
the point of view of the Genevan linguist. This is evidenced by the existence of the
category referred to as “post-Saussurian” in the various textbooks and manuals of
history of linguistics and by the fact that “Saussurism” is the lowest common
denominator of theoretical linguistic perspectives of the twentieth century: struc-
tural linguistics, generative linguistics and text linguistics.
Structural linguistics, born thanks to a Swiss scholar, grew and developed on
French soil. Émile Benveniste was a French linguist, and in a certain sense a
pioneer in the field of text linguistics. Benveniste’s production is vast, so it is
difficult to give a quick overview without doing the author an injustice. In a very
general way, one can say that he divided his commitment equally between historical
linguistic research and general linguistics. The first area, his historical research,
culminated in the publication of Le Vocabulaire des institutions indo-européennes
(1969), while the second area is well represented by an important collection of
articles entitled Problèmes de linguistique générale (vol. I 1966, vol. II 1974). This
anthology contains contributions regarding the examination of modern languages,
the status of linguistics as an autonomous discipline and the relationship between
structural linguistics, analytical logic and psychoanalysis.
Benveniste’s position is not that of an orthodox structuralist, unlike Hjelmslev,
whom we shall discuss later. This aspect is evident already in the way his work is
set up. First, the parallelism, usually recognized in the literature, between the
dichotomy langue versus parole in Saussure and the dichotomy langue versus
discours in Benveniste should be reconsidered. Secondly, one can not overlook
Benveniste’s criticism of Saussure’s concept of ‘arbitrariness of the sign’.5 Finally,

it is above all the tendency of Benveniste to move in the direction of a study of the
complexity of the ‘text’ that marks the difference with an orthodox structuralist
position.
He strongly emphasizes that reducing language to the arrangement and solidarity
of its elements (regardless of whether the type of structuralism is French or
American) is a deep misunderstanding of the linguistic datum. This occurs because
«the linguistic “datum” is not, considered in this way, a primary datum with which
all one has to do is break it down into its component parts; it is actually a complex,
some of whose values come from particular qualities of the elements, others from
the conditions of their arrangement, still others from the objective situation».6
Benveniste refuses to reduce human communication to a theory of information
and the various languages to communication tools. Given that relationships with
others and with the world are the constitutive moments of speech, the structure of
utterances can only be that of dialogue. In this sense, dialogue consists of texts, and,
on the whole, is itself a text. At the same time, the analysis of the nature of the sign
and the reference requires constant attention to the extra-linguistic dimension of
language, which must necessarily correspond to a sort of “anthropological project”.
Benveniste’s concept of ‘culture’ renders indissoluble the act of communication
and the development of values constituting human groups and cultures.7 Having
renounced a reductive conception in which natural language is the result of the
combination of formal data, he, therefore, also rejected the reconstruction of lan-
guage on a causal basis, through the history of etymology, and instead gives full
importance to the meaning borne by language and everything that in it has his-
torically been deposited and is modified daily. Language, by reproducing
extra-linguistic reality, is the mirror of reality itself: inter-subjective relationships,
relationships with the world, the birth, transformation and maintenance of
institutions.8
The relationship with analytical logic that he identifies as a corollary of this
position is worth mentioning. Generally the concept of ‘discourse’ in Benveniste is
recognized as a precursor of the concept of ‘performative’. In Analytic Philosophy
and Language, Benveniste again makes the distinction between performative and
constative utterances. This distinction had been proposed by John Langshaw
Austin, a philosopher of language whom we will discuss later. Austin’s criticism
was made to mitigate their differences and is based on what in every constative
5
This point was specifically developed in Lepschy (1979).
6
The English version of Benveniste (1966) is Problems in General Linguistics (1973), p. 11. The
essay from which we quote is titled Recent Trends in General Linguistics.
7
From this concept was born Le Vocabulaire des institutions indo-européennes, in which
Benveniste goes beyond traditional institutions of law, power and religion to investigate less
apparent forces that act upon techniques, ways of life, social relations, processes of speech and
thought.
8
Thinking of Benveniste, Dessons (1993) suggests the notion of language should be thought of
anthropologically, in terms of culture.
utterance is extralinguistic, which is not necessarily always clear. Although

Benveniste recognized the impossibility of reducing language to its merely formal
aspect, he nevertheless argued that this hypothesis could be plausible only if one
took account of the deep knowledge of the semantic aspects of a language. In
addition, if the major concern of analytical logic is to described how an utterance
serves to perform an action (so that formulating an utterance corresponds to per-
forming the action), Benveniste emphasizes the fact that an utterance is an indi-
vidual act of appropriation of the language. And as such, it is a constituent of the
subject. This is possible only because an utterance implies the presence of another
person and a reference to the reality proper to every speech.
In communicative interaction, in fact, for references to be possible, the sharing of
experience is necessary, which is expressed in the alternation of roles of “issuer”
and “receiver”, and in coherence, acquired during communication and based on the
assumption that the extra-linguistic reference is constant.9 For these reasons,
Benveniste appears to be the first scholar of structural linguistics to be highly
interested in the ‘text’ dimension as the ultimate level of linguistic analysis.
Outside of France, the teaching of Saussure was well received in Denmark by
Louis Trolle Hjelmslev. He published many works and is generally recognized as
the most orthodox exponent of structural linguistics of the twentieth century in
Europe. His first important book was Principes de grammaire générale (1928),
which forms the base of his whole conception; then he published La catégorie des
cas. Étude de grammaire générale (1935), which analyses in detail the general
category of grammatical case, providing a broad set of empirical data to support his
hypotheses. Hjelmslev’s work represents a constantly-evolving theory on the
problem of epistemology in linguistics.
In Omkring sprogteoriens grundlæggelse (1943), translated in English as
Prolegomena to a Theory of Language (1969), he criticized the prevailing methods
of the day in linguistics, proposing a model that was increasingly less descriptive
and more scientific. In his view, making linguistics scientific meant developing
Saussure’s ideas and extending them to their extreme consequences. In the
mid-Thirties he aimed for ever greater formalization and kept linguistics separate
from other sciences.
This attitude usually aroused two different reactions: in the first case, there was a
reaction in the form of rejection of the theoretical perspective, an explicit call to
distrust and resist any hybridization of the field; in the second case, which regarded
a minority, from an historiographical perspective, the secondary literature focused
on developing the point of view of the Danish linguist, trying to highlight the most
fertile aspects of his thinking, which, precisely because it was radical, made no
concessions to academic fashions and aimed directly at understanding the structure
of a system of signs.
According to Hjelmslev, a sign is a function of two forms, the “content-form”
and the “expression-form”, and this is the starting point of linguistic analysis.
9
Benveniste (1966), pp. 96–106. This essay is entitled The formal apparatus of enunciation.
However, each segnic function also originates from two substances: the substance
of content and the substance of expression. The substance of the content is the
psychological and conceptual manifestation of the sign. The substance of the
expression is the material substance in which a sign is manifested.
The attempt to use formal methods in linguistics is evident in his Résumé of a
Theory of Language (1975). It is the structure of the text that provides the first
indication of this attempt. On the one hand, the system seems blatantly
logical-mathematical. The book opens with the introduction of some undefined
terms and the exposition of five principles, which, like the five Euclidean postulates,
indicate the way forward by deduction (it is worth remembering that he was the son
of a geometry scholar). The Résumé seems to aspire to building a mechanism, a
super-algorithm that can determine whether a given object of communication is
semiotic or symbolic and whether it is language or not. Each term is formalized
with a corresponding symbol so that it can be inserted in calculation procedures. In
another area, however, things seem to progress in a different way. Hjelmslev, in
fact, believed that language is not a formal logical system and that linguistics is not
a field of research linked to logic; indeed, the latter must remain clearly distinct
from linguistics.
Hjelmslev’s objectives become clearer if we compare them with his epistemo-
logical goals. The Danish linguist had two objectives, which he had declared
already in the period of publication of Prolegomena: the first was to identify the
object linguistics was to study; the second was to obtain for the human sciences the
formal rigour typical of the exact sciences.10 When he began to talk about types of
dependence, Hjelmslev distinguished between two types of correlation (disjunctive
functions) that can exist between language units: in the first, exclusion, the corre-
lates (the two terms of the opposition) have no common variants; in the second,
participation, the correlates have common variants. The participations are correla-
tions, but correlations that are contradictory. He also identified the possibility that a
language may contain «extreme participations», in which «the participants have the
largest possible number of common variants» (Hjelmslev 1975, p. 62).
For Hjelmslev there are contradictory structures that are functional in languages:
already in La catégorie des cas. Étude de grammaire générale he had reiterated that
classical logic is only one of the possible realizations of a forma mentis, the human
10
The biggest problem is the kind of approach that he has in mind. At first glance, the first principle
set out in the Fundamentals and re-proposed in the Résumé appears almost obvious: «The
description must be free of contradictions, exhaustive and as simple as possible. The requirement
of absence of contradictions takes precedence over that of an exhaustive description. The
requirement of exhaustive description takes precedence over that of simplicity» (Hjelmslev 1975,
p. 45). The hierarchy between the different principles, however, is perplexing. Why, for example,
subordinate the completeness of the theory to its coherence? We must keep in mind that Hjelmslev
is explicit about coherence, which must be strong, that is, one must totally exclude contradictions
from the theory and not merely limit them, pending subsequent solutions.
mind, which is a “sub-logical system” in which the principle of non-contradiction is

not automatically present. So language is not a consistent object in all respects.11
As a result of this position, he was portrayed as a proponent of a scientific
method of analysis, almost a new form of semiotics, but, although later scholars
interpreted his position as that of a semiotic researcher (for example, this is the
opinion expressed by Algirdas Julien Greimas in his essay of 1966), it is plain that
Hjelmslev has argued that no sign can be interpreted unless it is contextualized (i.e.,
one cannot treat its functives—expression and content—as general connotative
mechanisms). For this reason, in his view, although the substance of the content is
important, the object of analysis is the form. Hjelmslev thus never modified the
linguist’s point of view: for him, meaning is given by the form of the content.
Outside of France, where Saussure disseminated his teachings, and Denmark,
where Hjelmslev studied and developed them further, the principles of structural
linguistics developed in other places, too, for example in the United States, and the
ideas circulating there around the 1930s were very similar to those in Europe.
Leonard Bloomfield, a contemporary of Hjelmslev and founder of structural
linguistics in America, was the author of Language (1933). This work was con-
sidered a landmark by American scholars in the 1940s and 1950s. As was the case
in Europe, the discipline did not consider the text as the level of analysis;
Bloomfield, in fact, considered the sentence as the natural subject of linguistic
research, and he defined it as «the largest unit with an inherent structure»
(Bloomfield 1933, p. 70), or «an independent linguistic form, not included by virtue
of any grammatical construction in any larger linguistic form» (Bloomfield 1933,
p. 170).
Bloomfield believed that language, understood as a complex social phenomenon,
should be studied by different disciplines and that, in the framework of these
different disciplines, linguistics should describe the structure of speech acts without
addressing the psychological and physiological conditions present in the use of
natural language. In addition, he also excluded the study of meaning from linguistic
science. This attitude of Bloomfield, in fact, isolated linguistics from disciplines
such as psychology, for which he had a kind of rejection; it also cut the study of
semantics out of linguistic research because the study of meaning had been
excluded as well. These studies, however, were made both by linguists and, even
more, by logicians, as we will see later.12
To sum up, he expressed a neopositivistic attitude, which was the basis for the
belief that «every statement must be verifiable by checking certain physical,
objectively measurable conditions, and scientific discourse must avoid referring to
any phenomenon (psychological, mental, etc.) that was not directly connected to
11
The Italian word uomo can refer to both the entire category (the genus homo) and a part thereof
(the human male); one can suppose that the word “thing” constitutes the lexical equivalent of
extreme participation because it is a term in which the largest possible number of common variants
is obtained.
12
On this aspect, see Fries (1975).
data that could be controlled by an outside observer» (my translation of Lepschy

1990, p. 447).
Bloomfield’s attitude is truly quite distant from the perspective of those who
argue that linguistics should have strong connections with the disciplines that deal
with human language and should attempt to integrate the research accomplished in
the various branches.13 The responsibility for having isolated, like watertight
compartments, the disciplines that study natural language can be traced back to
Bloomfield, and, since his work was influential for about twenty years, the
development of structural linguistics in the United States necessarily suffered the
consequences of this approach. To avoid any possible misunderstandings, however,
it should also be noted that Bloomfield’s attitude is much more scientific than that
of his European colleagues at that time, especially with regard to the adoption of a
rigorous terminology.
Generative Linguistics
In the United States, generative linguistics was the offspring of structural linguis-
tics. The main student of Bloomfield was Zellig Sabbetai Harris. Following the
footsteps of his teacher, he excluded the study of meaning from theoretical lin-
guistics; however, unlike Bloomfield, who ignored the contemporary research being
done on syntax and logical semantics, Harris took a position with respect to this
issue. On the basis of the position of Rudolf Carnap, expressed in The Logical
Syntax of language (1937),14 he argued that the interests of linguists were different
from those of logicians.15
Using simple formal tools, Harris pursued the analysis of natural language with
an obstinate empiricism, and in so doing went beyond the limits imposed on
linguistics: while considering the sentence as the largest unit of the domain of
grammar, he carried out his analyses on texts and therefore tested the limits of
methods of analysis and classification, and brought these limits to light in Methods
in Structural Linguistics (1951) and in Discourse Analysis (1952).
13
The attitude of contemporary scholars with respect to Bloomfield is exemplified in a particularly
illuminating way by the following words: «Today, these positions are generally discredited and it
is not clear how it was possible that, for nearly three decades, linguists seriously thought they
could say something interesting about language, and about psychological phenomena inextricably
relevant to it, on the basis of a model that one could hardly imagine more incongruous and
unsuitable to the object under study» (my translation of Lepschy 1990, p. 447).
14
The excessive pessimism of Carnap is easily seen, for example, in the following statement: «In
consequence of the unsystematic and logically imperfect structure of the natural word-languages
(such as German or Latin), the statement of their formal rules of formation and transformation
would be so complicated that it would hardly be feasible in practice» (Carnap 2001, p. 2).
15
See Harris (1963), p. 69.
The name of Harris is linked to the distributional method. It transcends the limits
of the sentence to «representing the order of successive occurrences of members of
a class» (Harris 1952, p. 8) within phrastic chains, i.e. texts. This method analysed
the distribution of morphemes by means of “equivalences”, which create «chains in
which the elements are the same or had the same environment» (de Beaugrande and
Dressler 1981, p. 17). To increase the number of equivalences and thus to make
analysis more exhaustive, Harris applied the notion of “transformation”. Thanks to
transformation he could manipulate expressions in order to obtain, transformation
after transformation, a “transform” with the maximum number of already realized
equivalences, beyond which it was no longer possible to obtain any new ones.
Since equivalences say nothing in terms of meaning, but manifest at most «what
criteria a new sentence must satisfy to be formally identical with the sentences of
the text» (Harris 1952, p. 493), it is unclear «what Harris’ method is supposed to
discover» (de Beaugrande and Dressler 1981, p. 17), and for this reason, perhaps, it
did not have much effect in the long run.16
All things considered, Harris can be attributed with two merits: first, to have
introduced the concept of transformation, which was then taken up and modified by
Chomsky; second, to have conducted research not only at the sentence level, but
also at the level of text. This research, however, did not consider the ‘text’ object
from a theoretical point of view, but only from a practical point of view. The
research did not address textuality theoretically, but was only empirically
motivated.
A pupil of Harris, Avram Noam Chomsky, attempted to solve all the problems
that did not find a sufficient explanation in the framework of the distributional
method. Chomsky’s contribution can claim the merit of having two fundamental
characteristics.
First, Chomsky overturned the perspective of the distributional approach: rather
than proceeding from hard data to the identification of the theoretical unit, he started
with theoretical formulations and sought confirmation or invalidation from the
empirical data; in other words, he transformed the inductive process of the distri-
butional method into a deductive procedure. The result was that Chomsky’s the-
ories could prove valid upon empirical verification; however, empirical verification
is subject to the data on which it is conducted. Chomsky no longer selected data
having a pre-established number of statements belonging to a text (which was the
typical corpus employed by the distributionalists), but rather the infinite number of
all possible statements of a natural language. In his opinion, this was the only way
to evidence the creativity of the speaker, that is, the ability to produce—and
understand—a potentially infinite number of sentences. This creativity is part of
competence, i.e. the system of (innate) principles that allows an adult human being
to be constantly able to build and receive new communications, in the sense that
they have never been spoken or heard before, to recognize spoken or perceived
sentences as acceptable or unacceptable, to interpret a sentence by eliminating the
16
See Prince (1978).
ambiguities that it contains and to operate transformations on a single sentence,

such as shifts from the active to the passive form and from the affirmative to the
negative (and vice versa).
Second, Chomsky transformed linguistics into the grammar of competence,
given that he took on the task of discovering and describing the system of com-
petence that exists in the mind of the speaker. He however omitted from the field of
his research its implementation, that is, the empirically observable variety of per-
formances in which the linguistic competence of a speaker is manifested.17
Working further down the path opened by his teacher, Chomsky considered the
use of formal methods in linguistics valid and then began to explain and describe
competence and linguistic concepts by means of formal tools of mathematics, logic
and (newly born, at that time) computer science: he decided to use the theory of
automata and the theory of recursive functions. With the introduction of these
methods in linguistics, Chomsky thought he could address the question of the
infinite number of expressions in a natural language, the absence of limits to the
length of an expression in a natural language, the distinction between grammatical
and non-grammatical utterances, as well as semantic and structural ambiguities. In
other words, he tried to overcome the problems which constituted the limits of the
distributional approach.
The grammatical theory of Chomsky does not deal with the creation of com-
petence, that is, it is not concerned with single performances that evidence the
linguistic behaviour of a speaker; his theory deals with predicting the generation18
of all possible utterances of a natural language, to which a structural description is
assigned.19
To transformational grammar, «many criticisms were addressed, above all that of
being too abstract, of referring to a non-existent ideal speaker-listener in a perfectly
homogeneous linguistic community, of ignoring real linguistic material, of under-
estimating the differences between actually attested languages, aiming at a universal
grammar whose structures end up being curiously similar to those of English» (my
translation of Lepschy 1990, p. 453); all these criticisms could be supported by
some very convincing arguments.
17
Chomsky’s concepts of performance and competence are often associated in the linguistic
literature with the concepts of parole and langue introduced by de Saussure, which are probably
more familiar to the reader. While it can be said that there is a certain correspondence between
performance and parole, the same cannot be affirmed for competence and langue. Chomsky’s
concept, in fact, concerns the system of mental rules inherent in speaking, while Saussure’s
concept regards the system of rules established by a socialized collective.
18
We must understand this «generation» as a mathematical concept; for example, given a linear
equation x = 3y, different values of x are obtained by assigning different values to y.
19
The structural description consists of a “deep structure”, which would be the abstract syntactical
organization. This is the basis of a sentence and is responsible for its semantic interpretation, that
is, the assignment of meaning. The name surface structure is given to the syntactic organization
evidenced by the sentence in relation to its linear manifestation.
With the publication of Syntactic Structures (1957),20 Chomsky introduced new

concepts to the field of linguistics, new terminology and a methodology that used
logical, mathematical and computer models in the formulation of linguistic theories.
His formal methods, considered as a means to make observations of natural lan-
guages in an explicit way controllable by others, were not accepted by all linguists,
but even so, they were used. We can therefore say that Chomsky, by putting
mathematical theories at the service of studies of syntax, indicated a new direction
for linguistic studies that can be called the “algebraic line” of research, oriented to
syntax.21
To better understand the methodological breakthrough of Chomsky, we should
remember that between the Fifties and Sixties the coveted goal of several linguists
was to reach one of these two conditions22: either to construct a linguistic theory in
a formal way, that is, in the language of mathematics or logic, or formulate it
20
In this exposition we will not refer to the proposals of Chomsky (1957) but to the more widely
disseminated version contained in Aspects of the Theory of Syntax (1965). In fact, the history of
generative grammar is now more than forty years old and, therefore, in a full exposition we should
take into account all the revisions of the theoretical framework, some of which are substantial,
though it has always remained focused on syntax. The theory of Generative Grammar has in fact
evolved from the so-called Standard Theory (in Chomsky 1968) to the following later versions:
Extended Standard Theory (in Chomsky 1972); Revised Extended Standard Theory (in Chomsky
1975); Government and Binding Theory (in Chomsky 1981); Minimalist Program (in Chomsky
1993).
21
The dichotomy between the algebraic approach and the logical approach regards the use of
formal methods in linguistic research. «Formal methods used in today’s linguistics rest upon two
lines of research. The first line goes back to such workers as Frege, Carnap, Reichenbach,
Ajdukiewicz and others in the field of philosophy of language. We want to call it the logical line as
opposed to the algebraic line, initiated by the works of Chomsky and others in formal language
theory. These two lines differ greatly as to their orientation on syntax or semantics. The logical line
can justly be called semantically oriented, the algebraic line syntactically oriented. We want to
point but that the notions of ‘syntax’ and of ‘semantics’ refer to formal syntax and semantics in
artificial systems. The algebraic line in formal language theory is not to be confused with gen-
erative grammar in the line of Chomsky. Generative grammar in this form makes use of algebraic
language systems partly studied by Chomsky himself» (Eikmeyer 1980, p. 2).
22
If we choose to borrow the words of Andrea Bonomi, we must recognize that «it is certainly no
coincidence that the use of formal logical instruments found a first, natural field of application in
the area of syntax, that is, the most “abstract” of the fields of research according to the boundaries
of Carnap, in the sense that among the data which it is supposed to account for there are no
elements that require an extra-linguistic reference (as in semantics), nor those of a contextual
nature (as in pragmatics). If this conceptual paradigm, which sees in syntax the culmination of a
gradual theoretical selection (and impoverishment) of the data to be analysed, is correct, then it is
not difficult to explain why, symmetrically, syntax was the historic starting point of the systematic
use in linguistic of methodologies developed within the logic of contemporary mathematics» (my
translation of Bonomi 1983, pp. 148–149).
explicitly, that is, in a regulated language that uses specific terms in the most precise
way possible. Both types of formulation would have allowed a linguistic theory to
be automated, that is, rendered applicable on a computer.23
Thus the influence of research on artificial intelligence, which in those years was
the centre of scientific debate in America, drove Chomsky to focus on the for-
mulation of a formal linguistic theory, the theory that currently makes him the most
celebrated linguist of the second half of the twentieth century.
Two limits, however, are attributed to Chomsky: first, he anchored, perhaps
more than before, linguistic research to the sentence, that is, he neglected some of
the stimuli known to Harris, which could have been taken further into account.
Indeed Harris, while following an essentially structural method, devoted himself
constantly to the analysis of texts, and not sentences, in the application of his
models. These stimuli, as we shall discuss in depth below, can be considered
historical precedents of the genesis of text linguistics. Secondly, though he con-
sidered the study of meaning, he artificially discriminated between logical
semantics and linguistic semantics (of which the apparatus developed by Katz and
Fodor is a clear manifestation);24 in this way he confined semantics to an inter-
pretative dimension.
From the point of view of research on the text, both structural and generative
linguistics are, essentially, two aspects of the same type of study, i.e. linguistics
aimed at the study of the sentence.
23
From a technical point of view, it was hypothesized that a linguistic theory formulated in the
language of mathematics or logics could be automatized in order to be run on a computer. If the
formal system had the complexity of at most a Turing machine, it could be automatized by suitable
algorithms. Yet though these algorithms may be suitable for processing the system from a theo-
retical point of view, they are not necessarily effective relative to a given computer installation.
Indeed the implemented algorithm may prove too complex and take too much place and time, and
therefore would not be justifiable. If complete automation was impossible, an attempt at partial
automation would have to be made in a sort of man-machine cooperation, similar to the model
developed in the field of HCI (= Human-Computer Interaction), see Eikmeyer (1980), pp. 5–6.
24
To illustrate the functioning of Katz and Fodor’ apparatus, we present the semantic represen-
tation of the English noun ‘chase’ (a hunt), taken from Katz (1967, p. 169): (((Activity of X)
(Nature:(Physical)) ((Motion)
(Rate:(Fast)) (Character:(Following Y))

(Intention:(Trying to catch ((Y) (Motion))).
The incoming lexeme (called lexical entry) is provided by a syntactic indicator, one or more
semantic indicators, one or more differentiators and one or more selective restrictions. In this way,
each lexical entry is composed of a series of elements that allow different combinations and form
the so-called componential tree. Going through the entire componential tree, i.e. choosing between
alternative routes, one arrives at the end of the path and thus at a certain reading.
The Text from a Logical Point of View: the Linguistic Turn 15
The Text from a Logical Point of View:

the Linguistic Turn25
European and American linguists saw no relationship between their semantics and
logical semantics until the mid-Sixties. They moved only in the path traced by
tradition, following Saussure and Hjelmslev, who denied there were contacts
between linguistics and “psychological sciences”; and following Bloomfield, who
ignored the investigations into syntax and logical semantics going on around him;
and following Harris, who had argued that there was a distinction to be made
between the fields of linguistic and logic. However, some changes began to occur:
although the distributional approach of Harris did not have lasting consequences,
one publication did dare to defy its primacy in the years in which it was dominant.
Reichenbach published a manual titled Elements of Symbolic Logic (1948),
which contained a chapter on the analysis of conversational language.26 Reversing
the position of Rudolph Carnap (perhaps overly pessimistic?),27 the logician
severely criticized the grammatical tradition for the way in which expressions of
natural language are ascribed to grammatical categories. He saw the points of
contact between grammatical tradition and the logical systems employed by Frege
and Pierce. These systems had fuelled a misconception of the logical structure of
natural language. Therefore, he made proposals concerning the description of
deictic expressions, adjectives, adverbs, tenses and modes.
With generative grammar, formal methods were established in linguistics, first
through syntax and then through semantics. At first, as we said, it was a mathe-
matical type of formalism, and for this reason it has been defined an algebraic
approach; then, in correspondence to the growing orientation towards semantic
problems, the formalism employed began to be taken from different types of logic.
The formulation of a model for semantics brought intense criticism down on
Katz and Fodor, both from the side of logicians and from that of linguists.28
Linguists criticized the semantics of Katz and Fodor especially for the assumptions
on which the analysis is based, namely the fact that the meaning of an expression
25
The present section is an extended version of Giuffrè (2011b), pp. 11–14.
26
For a general overview of Reichenbach, see McMahon (1976).
27
The statement to which Reichenbach refers is contained in Carnap (2001), p. 2 (see note 14).
28
The best known locus vexatus of Katz and Fodor is the following: «Since a complete theory of
setting selection must represent as part of an utterance any and every feature of the world which
speakers need in order to determine the preferred reading of that utterance, and since, as we have
just seen, practically any item of information about the world is essential to some disambiguation,
two conclusions follow. First, such a theory cannot in principle distinguish between the speaker’s
knowledge of his language and his knowledge of the world, because, according to such a theory,
part of characterization of a linguistic ability is representation of virtually all knowledge of the
world that speakers share, and since a theory of the kind we have been discussing requires such a
systematization, it is ipso facto not a serious model for semantics» (Katz and Fodor 1963, p. 179).
This step testifies to an undue restriction of the concept of competence that makes the whole
system too rigid.
can be broken down into basic elements and can be represented by the association
of the parts. If this so-called componential analysis could prove valid for individual
lexemes, it did not give adequate guarantee of validity for phrases, and still less for
sentences; that is to say, in general, for the meaning assigned to the lexical input.29
Among logicians, however, the most rigorous critic was Yehoshua Bar-Hillel,
who accused Katz and Fodor of identifying semantics with lexicology on the
grounds that they had neglected all the elements of logical semantics.30 He argued
that, as part of their semantics, relations such as ‘implication’ could not be derived
and that it was impossible to reconstruct semantic fields, while he reiterated that the
task of linguistic semantics was to arrive at a definition of ‘derivability’.31
The strength of Bar-Hillel’s criticism of Katz and Fodor and the gradual increase
in logical research into syntax, aided by the fact that Chomsky himself had
employed this route, meant that within the Chomsky school, scholars began moving
29
The distinction between ‘meaning’ and ‘reference’, quite common in logic-oriented linguistics, is
homologous to that between ‘intension’ and ‘extension’. The distinction dates back to Carnap. The
explanation—taken from (Lyons 1977, p. 171)—is that «the intension of a term is the set of
essential properties that determine the applicability of the term» while «with the extension of a
term is meant the class of things to which it is applied». So the first term refers to specific content,
to its individual properties, the qualities necessarily possessed by the term to be defined
(definiendum); the second to all objects denoted by the same sign. It follows that the extension of a
lexeme is the set of individuals to which the lexeme can be referred as an attribute and that the
extension of a proposition is the set of cases in which it is true. So it is clear that «to give the sense
of a language expression Eo in the natural language being described, i.e. the object language, is to
translate it into a language expression Em in the metalanguage—the language of the semantic
representation, which may be the same as the object language; thus dog means ‘canine quad-
ruped’» (Allan 1992, p. 395) and it should not be forgotten that «a dictionary gives the decon-
textualized sense of a word, abstracted from innumerable usages of it. Dictionary users must
puzzle out its denotation, i.e. what a speaker or writer uses the word to mean in the world evoked
by a text in which the word appears» (Allan 1992, p. 394).
30
Echoing the position inaugurated by Bar-Hillel, Bonomi and Usberti used this criticism in Italy.
Both note that Katz and Fodor, by rejecting extra-linguistic knowledge, dissolve the relationship
between sense and denotation in a relationship between signs, thus proving to be more lexicog-
raphers than semanticists; they argue that «to the extent that one wants to effectively build a
semantic theory, one is […] forced to make assumptions, though implied and inadequate ones,
about what the meaning of the words is» (my translation of Bonomi and Usberti 1971, p. 99) and
therefore, in direct opposition to Katz and Fodor, they believe it is possible «to assign sets of
appropriate traits to lexical items only to the extent that we use these traits as words of the
language. So not only does the meta-language fail to give us an explanation of what the meaning
is, but the “meta-linguistic competence” that we could invoke [viz. invoked by Katz and Fodor]
turns out to be precisely linguistic competence» (my translation of Bonomi and Usberti 1971,
p. 103). As regards meta-language, both explained that «if the explanation of the lexical meaning
of a word occurs by means of other words, a circular process of interdefinition begins whereby, in
order to define the meaning of a term, you use terms that must themselves be defined. In the best
case one produces a chain of definitions resulting from synonymy, which, after going through the
entire dictionary, take you back to square one, still undefined» (my translation of Bonomi and
Usberti 1971, p. 103).
31
See Bar-Hillel (1970).
away from the “orthodox” position. Thus, a current opposed to the semantic
interpretation of Katz and Fodor arose within generative grammar, and it was called
generative semantics.
Generative Semantics
These scholars objected that the theory did not take semantics and problems related
to it into due consideration and they began to extend the domain of linguistics,
arguing that grammar should concern itself not only with the generation of gram-
matical sentences, but also with the structuring of semantic representations. The
post-Chomskyan scholars were gradually attracted to reflections on meta-language
because, due to the opposition that had arisen against interpretative semantics, in
the late Sixties semantics had become the field of study of natural language where
linguists and logicians would most often clash.
The generative semanticists, who had substantially incorporated the views of
Bar-Hillel in their research, became the natural ally of logicians and philosophers of
language against Chomskyan orthodoxy;32 generative semantics was the crowbar
with which to open linguistics to formal methods once and for all. Indeed the
convergence between the interests of logicians and those of generative semanticists
caused the algebraic approach that had characterized Chomsky, Katz and Fodor to
regress, and the strengthening of the logical approach in linguistic research.
The central question of semantics, which is a typical frontier discipline and
closely related to logic and semiotics, concerns semantic representations. For
scholars, semantic representations constituted (and still constitute) a continuous
battlefield, because they are built through the use of a metalanguage.33
The continuous discussion about the kind of metalanguage to be used in
semantic representations brought linguistics to proceed in the direction of logic and
the philosophy of language. Since the days of Aristotle, logic has been used in the
representation of meaning. Standard logic is able to define truth conditions only in
propositions connected by particular uses of and, or and if/then, while non-standard
logic operates with postulates of meaning;34 but a valid metalanguage for the
32
Among the philosophers of language that maintained close relations with the generative
semanticists, Maxwell John Cresswell stands out. He frequently manifested his intention to further
his studies of the proposals of generative semantics; it could also be argued with some degree of
support that philosophers like Donald Davidson and David Kellogg Lewis filled the gap between
linguistics and logic, stabilizing contacts between the two disciplines.
33
On meta-language or metalanguage, the necessary reference is Lyons (1977), pp. 5–13.
34
From the logical point of view, providing truth conditions is a way of saying what the propo-
sitions mean. The concept can be clarified by saying that «[…] it seems uncontroversial that if a
sentence α is true and another β is false, then α and β do not mean the same. Truth-conditional
semantics assumes that, in essence, the meaning of a sentence is equivalent to the conditions under
which that sentence is true. The task then is to formulate an adequate and precise account of truth
semantics of natural languages must be at least as comprehensible as a natural

language and of the same class of notation.35 No logical language currently
achieves this goal; however, for a complete formalization of natural languages, this
is the goal to aim for.
The generative semanticists argued that semantics (and not syntax) generated the
deep structure of sentences. Among them we should give particular mention to
James McCawley and George Lakoff. Both were confronted with the problems
described above and tried to overcome the difficulties of Katz and Fodor’s model by
proposing a form of semantics that went outside the domain of meaning and
included the domain of extension as well.
The first argued that, like meta-language for semantic representations, it was
necessary to choose a language of symbolic logic and change it based on linguistic
motivations36; the second, however, went even further and evidenced the need to go
beyond the context of intension, using research on semantics conducted in the field
of modal logic. He proposed that sentences received truth conditions through a
formalism that was more elaborate than that of symbolic logic.37
Apart from the individual points of view and the various technical aspects, both
scholars occupy an important place the present work because they made several
attempts to employ logical systems, especially in regard to compositional rules, to
explain phenomena of natural language. Generative semantics was a healthy off-
spring of generative grammar, but it proved matricidal: it maintains that an inte-
grated system of several logical languages can serve as a language to a linguistic
theory.
Perhaps generative semantics is blamed for the defect of not sufficiently dis-
tinguishing, in semantic representations, between the rules of formation and the
rules of transformation, and not assigning correlates to semantic representations;
that is to say, the generative semanticists are not able to say what kind of entities
meanings are and how the meaning of an expression in natural language depends on
the meaning of its components.38
One might observe that «it is well known that many linguists have viewed with
particular suspicion the notion of logical form, as it is seen as unrelated to the syntax
and semantics of natural languages, and because through it, an additional level of
linguistic representation would be introduced, one with no independent motive to
exist and thus not required. One could even say that, in a certain sense, the rejection
of this notion has often been the mainstay of various arguments that question the
legitimacy of the application of certain logical tools in linguistic inquiry» (my
(Footnote 34 continued)
conditions in a way which makes it plausible to suppose that knowing the meaning of a sentence is
knowing its truth conditions» (Cresswell 1992, p. 404).
35
See McCawley (1981).
36
See McCawley (1972).
37
See Lakoff (1970).
38
See Hall-Partee (1976).
translation of Bonomi 1983, p. 150), but the real question to ask is: «can we avoid
representing somewhere (in the syntax, the semantics or independently) that type of
information that logicians and philosophers have traditionally linked to the (often
vague) concept of logical form? An affirmative answer to this question would mean
that linguists must cease trying to account for an essential aspect of the competence
of a speaker, that is, the ability to link different sets of logical consequences to the
grammatical structure, or, to say it more precisely, to assign different truth conditions
to the same sentence» (my translation of Bonomi 1983, p. 150).
Therefore, the inability to respond affirmatively to the above question obliged
logicians and linguistic experts to converge in search of a solution. This conver-
gence, however, raised a fundamental problem: because each logical system studies
only a limited section of the expressions of natural language (propositional logic
studies connectives, predicate logic studies quantification, modal logic studies
phrasal modifiers, etc.) the need arose, in order to formulate a linguistic theory, for a
comprehensive framework that could combine the individual fragmented and
specialized logical systems into a single logical system that was uniform and
general enough to describe all the constructions that occur simultaneously in the
texts of a natural language.
Two consequences ensued: first, the most extensive level of analysis of linguistic
research was no longer the sentence, but the text. It would serve as a testing ground
for an empirical verification of the integrated theoretical framework formulated—it
was on this path, therefore, that linguistics pressed for the creation of text lin-
guistics. Second, a degree of explicitness and an ability to integrate the studies of
different areas was achieved that had been never been seen previously.
Model-Theoretic Semantics
Logicians tried to build a semantic system of natural languages based on extension.

Richard Montague took up the challenge launched by Reichenbach and proposed
an alternative vision: he considered it appropriate to employ logical and mathe-
matical methods to develop a system of syntax and semantics useful to the study of
natural languages, so his objections to Chomsky were similar to those raised by
generative semanticists, but he possessed greater argumentative rigour and firmly
stressed the need for a coplanar development of syntax and semantics.39
Following the evolution of thought in the entire tradition of analytical logic of
language, linked to the names of Frege, Ajdukiewicz, Tarski, Carnap, Quine,
Kripke and Lewis, he argued that natural language expressions could not be
translated into corresponding logical forms by traditional logic. It intuitively made
39
The theory of Montague is presented in Stegmüller (1975), pp. 35–64, where much attention is
paid to philosophical and epistemological aspects. Further introductions in Dowty (1979) and
Dowty et al. (1981). More general is Chierchia (1992).
too many transformations when it tried to overcome the problems that arose when
the boundaries of individual logical systems were crossed. For this reason he
conceived of an ambitious project to construct a consistent logical system, one that
was sufficiently general, and which, as we mentioned before, could treat all
expressions of natural language. He called that project natural logic.
Montague’s theory seems extremely complicated: first, he refused to subject the
expressions of natural language to the intuitive changes of analytical logic, and to
vary only the procedure for applying the intuitive changes (because it would have
meant describing them through an already existing logical apparatus!); second, he
refused to modify the logical tools at its disposal to avoid the risk of rendering them
less accurate.40
Despite the high degree of complexity, the theory and work of Montague rep-
resented a first attempt to deal in an integrated way with syntax and semantics by
adopting formal methods, and gave impetus to the research of the philosophy of
language and the development and foundation of philosophical logic. In addition,
the activities of Montague marked the definitive prevalence of the logical approach
in linguistic research: logic was victorious over mathematics in the controversy over
which methodology to use in the formalization of linguistics. This meant the
prevalence of logic over algebra. Natural logic contains in nuce the determination
that expressions of natural language are to be treated on a level higher than that of
the sentence; it is an attempt to build a logical apparatus capable of regimenting
texts, rather than statements. It therefore represents a decisive step towards the
genesis of text linguistics.
The stage of development of the formal tools did not immediately allow a valid
linguistic theory to be proposed for statements; in fact, the semantic systems studied
by Montague, although they were subsequently revised, expanded and deepened,41
did not provide sufficient explanations of linguistic meaning.42
These studies should be given the merit of having decreed that the way to solve
the semantic problems is to combine different types of logical systems and to give
primary importance to model-theoretic semantics. Model-theoretic semantics
comprises three parts. First comes a syntactic description of the linguistic items-words,
phrases, sentences, etc. of the language in question; this is without reference to meaning.
Second comes a description of a class of (usually abstract) language-independent entities
40
The reasoning is flawed by the prejudice that historical languages are imperfect and the tools of
logic are eternal creations of the human mind. This assumption is shared by many logicians,
including Montague. On this subject see Bonomi (1973). Other scholars such as Eric Havelock and
Walter Ong, the theorists of so-called “Greek Orality and Literacy”, propose another point of view:
Havelock challenges analytic philosophers’ assumption that logical procedures have always been
rooted in human nature; Ong affirms that the illusion that logic is a closed system possessing
perfect coherence has been encouraged by the written word and even more by the press (see La
Matina 1994, pp. 99–143).
41
Cresswell was interested in continuing the work of Montague in Cresswell (1973). For the
development of natural logic in the context of generative grammar, see Hall-Partee (1976).
42
See Suppes (1973).
which are the meanings of expressions in different categories. Finally comes a function
which assigns a meaning to each linguistic item in such a way that whole sentences will be
correlated with entities which adequately represent their meaning in the language being
described. The most common such theories are those developed within the possible world
semantics […]. A possible world is a complete and total way that the world could be. One
of these possible worlds is the actual world: this is the way things are. In possible-worlds
semantics, the truth conditions of a sentence are the ways the world has to be if the sentence
is true, and so the assignment function must end up by giving to each sentence, at each
contextual index, a class of possible worlds. The meaning of a sentence is in one sense
conventional, but in another not. The meanings of the separate words are conventional; but
once given, they determine, in conjunction with the syntax of the sentence, the meanings of
complex items. The most explicit treatments of natural language in a possible-worlds
framework are those inspired by the work of Montague […] (Cresswell 1992, p. 404).
Therefore, the solution to the most important problems of semantics might be

found thanks to the possible-worlds theory. This is also what Keith Allan suggests:
truth-conditional and model-theoretic semantics, developed by Montague and by Cresswell,
exploit the notion of ‘possible worlds’ to distinguish sense from denotation, and have rules
for assigning denotations (extensions) to senses (intentions). They seem to offer the best
hope for representing the semantic effect of one surface-syntactic category upon another in
the same structure. Such systems may one day offer a satisfactory semantic metalanguage
for natural language (Allan 1992, p. 397).
Thus, formalization took hold of linguistics, first in the studies of syntax and
then in those of semantics. When these theoretical methods became known in
Europe, where conditions were favourable to the birth of text linguistics, linguistics
came to a turning point, the so-called linguistic turn.
We can define as a turn the path that began with the rejection of formal methods
and which subsequently led to their use in linguistics, first in a prevalently algebraic
approach (Harris and Chomsky) and then in a logical approach (generative
semantics and Montague). Moreover, linguistics, which was initially isolated from
the neighbouring disciplines as if they were watertight compartments, has pro-
gressively extended its domain to stably include semantics, a discipline in continual
contact with logic, the philosophy of language and semiotics.
Text linguistics, because it became a separate discipline in Europe only after the
gradual formalization that sentence-based linguistics underwent in the twentieth
century in the United States, cultivated the aspiration of reaching the formal rigour
typical of disciplines such as mathematics, logic and computer science.
The Text in Text Linguistics43
Text linguistics was born as an independent discipline using the coincidence of two
factors: the impact of the dispute between interpretative semantics and generative
semantics on European scholars, and the development of studies on texts
43
This section is an extended version of Giuffrè (2011b), pp. 15–17.
(especially literary texts) in the context of structuralism and semiotics. As we

anticipated in the title of this chapter by using the title of a work by Peter Hartmann,
from a post-eventum perspective it would indeed seem that alle Wege führen zum
Text.
A homeland and a precise date of birth can be attributed to the genesis of text
linguistics. It was born in the Federal Republic of Germany and the German
Democratic Republic by way of three fundamental publications in the mid-Sixties:
Text, Texte, Klassen von Texten (1964) by Peter Hartmann, Tempus. Besprochene
und erzählte Welt (1964) by Harald Weinrich and Review of Z. S. Harris’ Discourse
Analysis Reprints (1965) by Manfred Bierwisch.
Peter Hartmann, as a theorist, dealt in terms of general methodology with the
need to establish a level of linguistic analysis that was higher than that of the
sentence, arguing that language does not exist and that it cannot be analysed unless
it is in the form of a ‘text’. Harald Weinrich oriented his studies to the concept of
‘text’ in such an explicit way that the subtitle, in Italian, was translated Le funzioni
dei tempi nel testo (The functions of tenses in the text). Harald Weinrich’s work
analysed a very rich corpus of narrative texts and proposed new categories suitable
for stylistic and literary studies. His work also outlined the possibility of defining
literary genres on the basis of a verbal typology. Manfred Bierwisch, in his review
of Harris, showed that an entirely implausible text can satisfy the equivalence
criteria of Harris and consequently he became interested, and was among the first,
in the problem of coherence (his criticisms of Katz and Fodor’s semantics are also
interesting). He sought a solution to this problem by attempting to define the
concept of a ‘text in L’; he maintained that, just as a theory of the sentence has the
task of defining a sentence, a textual theory must provide a definition of text.
The publication of these three works testifies to the fruitful interaction of
methods derived from research in the United States and currents of thought in
Europe. But the theoretical reflection on the text in Europe had been strong even
before these three studies of the mid-Sixties. For example, both of the following
authors could be considered precursors of text linguistics: Tudor Vianu, author of
Arta prozatorilor români (1941) and Eugeniu Coșeriu, author of Determinación y
entorno. Dos problemas de una lingüística del hablar (1955–1956).44
The first was lacking in a rigorous use of scientific terminology; the second
started from a general reflection on language and detected three levels in the
manifestation of the faculty of language: a universal level of speaking, a historical
level of individual natural languages and the individual level of the text. In con-
nection with these levels, there are three different types of linguistics: of speech, of
the language and of the text. The essay deals mainly with the linguistics of speech
(“del hablar”), which was supposed to solve the remaining problems of linguistics
of language (both structural linguistics and, on a lower level, generative). The
perspective of Coșeriu, however, is quite peculiar.
44
See Dragos (1986), p. 180.
The Text in Text Linguistics 23
As far as the objective of establishing the text as the highest level of linguistic
analysis, beyond the ideas of these two precursors, which are in a certain sense
peculiar, the most prominent reference points for the generally accepted view were
undoubtedly Zellig Harris and Louis Hjelmslev.
In relation to Harris, it must be remembered that his suggestions about the text
contained in Discourse Analysis (1952)45 had been largely ignored (for that matter,
even his student Chomsky had ignored them) and it was noted that following the
publication of Review of Z. S. Harris’ Discourse Analysis Reprints (1965) by
Manfred Bierwisch they received more attention, at least in European circles. It can
be said that «Harris’ extension to the text of methods used in connection with
sentence grammar assumes that between text and sentence the same relationship
exists as the one between phrase and morpheme, and between morpheme and
phoneme. But this assumption of Harris’ is incorrect: in fact, sentences do not
constitute a text in the same way in which morphemes constitute a phrase, or in
which a morpheme is composed by phonemes. It is not a quantitative increase, it is
not just one additional level» (my translation of Conte 1977, p. 27). Harris there-
fore, although this premise of his was incorrect, attributed a theoretical value to the
empirical evidence of the ‘text’, which was the subject of his tenacious analysis.
As Bierwisch noted, other structuralists became interested in the text in addition
to Harris. For example, Hjelmslev, too, examined the text from the empirical point
of view, although he included this examination within a theoretical reflection on the
concept of ‘text’ that now appears unsuitable; as illustrated precisely by Conte,
«[…] according to Hjelmslev, every act of language in a given language is a text;
indeed, a living language is itself a text, an unlimited text […]. The text of
Hjelmslev is […] not a linguistic unit, but rather a mere form of existence of a
language. […] Hjelmslev’s text is not a concept on which it is possible to found text
linguistics. […] Hjelmslev’s is, in short, […] not a theory of the text, but through
the text» (my translation of Conte 1977, pp. 24–25).
By identifying Hjelmslev and Harris as the most reliable landmarks, once again
text linguistics seems having been generated by the convergence of theoretical
issues arrived in Europe from United States and by the progress of the European
Structuralist tradition studies. To capture a substantial difference between “upwards
extension to the sentence” and “text”, it was in fact necessary to recognize a
“qualitative (and not merely quantitative) difference between sentence and text”
(my translation of Conte 1977, p. 18). Scholars noted, though not immediately, that
«texts indeed appear to be something more than mere sequences of sentences, and
the comprehension and formation of texts seem to be governed by a specific
competence of the speaker: textual competence. (This competence is distinct from
sentence competence by which the ideal speaker-listener formulates and compre-
hends grammatically correct sentences)» (my translation of Conte 1977, p. 18).
The initial stage can be dated to the mid-Sixties, and, in addition to the three
founding works already mentioned, the studies of Harald Weinrich should be
45
See Conte (1977), pp. 482–483.
mentioned as well, because he was definitely the one who introduced the notion of
Textlinguistik in the meaning still employed today.46 At that time the discipline
went beyond «[…] the limits of the single sentence, for [it considers] sequences of
sentences, but […] does not yet reach an independent discussion of the text. But it is
only by thematizing the hierarchical structure of a text, the overall semantic
coherence, that the step from the sentence to the text can be taken» (my translation
of Conte 1977, p. 17).
Text Linguistics and Its Domain
The birth of text linguistics came about through an interweaving of internal and
external causes. The internal causes were constituted by the need to find a theo-
retical framework suitable for treating and attempting to solve all the problems that
weighed upon sentence linguistics, both in European and American structural and
generative linguistic tradition. The main problem—among the many to be
addressed—may be illustrated by an example by Maria-Elisabeth Conte:
Carla goes to the pool every morning.
In winter, Claudia is never home on Sundays because she goes skiing.
And Giancarlo has even won a silver medal in Montreal.
All my kids are athletes (my translation of Conte 1977, p. 16).
which she comments by noting that «although […] there are neither any repetitions
nor uses of pronouns, nevertheless […], if it were a mother speaking of her three
children, [it] is a coherent text» (my translation of Conte 1977, p. 16).
Coherence is the main problem that text linguistics set out to solve, and the
problem is exclusively linguistic. If the vision of linguistics as a series of watertight
compartments is discarded and one adopts the views of those who argue that
linguistics must have strong connections with all the disciplines that deal with
human language, and, moreover, that it possesses the ability to benefit from
research from various sectors, coherence takes on fundamental importance.
In the genesis of text linguistics, the external causes are represented by the
pressures exerted by various disciplines in the direction of building an integrative
paradigm for the study of texts. Dressler (1972) identified some of these disciplines,
and de Beaugrande and Dressler (1981) subsequently elaborated and further eval-
uated them. They include rhetoric, stylistics, literary studies, anthropology, soci-
ology and conversational analysis, philology and semiotics.
Among the disciplines mentioned, rhetoric is without doubt the one with the
longest tradition and the most authority. Its subject is the text, in the sense that «the
text is the result of rhetoric and is built by the speaker for […] acts of persuasion; in
46
This use of the concept was ratified in Weinrich (1967).
the various phases of that activity the text is configured both from a structural and
communicative standpoint, because rhetoric provides the tools required to achieve
both this total linguistic unity and its transmission, in which the discourse is con-
served as a whole» (my translation of Albaladejo 1989, p. 12).
According to de Beaugrande and Dressler, text linguistics and rhetoric have an
affinity that can be summarized in the following five points: «(a) accessing
arranging of ideas is open to systematic control; (b) the transition between ideas and
expressions can be subjected to conscious training; (c) among the various texts
which express a given configuration of ideas, some are of higher quality than
others; (d) judgements of texts can be made in terms of their effects upon the
audience of receivers; (e) texts are vehicles of purposeful interaction» (de
Beaugrande and Dressler 1981, p. 13).
The two scholars assert that classical rhetoric was intensely committed to finding
an answer to questions such as: «How are discoverable structures built through
operations of decision and selection, and what are the implications of those oper-
ations for communicative interaction?» (de Beaugrande and Dressler 1981, p. 14).
Bringing the thought back to the instrument by which it is expressed, the linguistic
turn had indeed a new and extremely interested look at rhetoric that, for centuries,
had dealt with the text as its object of study, studying its effects and proposing
classifications and definitions for its various functions. For this reason, this ancient
discipline became the centre of great theoretical interest in the second half of the
twentieth century.47
Stylistics deserves to be considered on an equal level with rhetoric. It studies «all
relationships above the level of the sentence» (my translation of Dressler 1972,
p. 15). From the moment the most extensive level of analysis of linguistic research
became the text and no longer the sentence, the fields of study of linguistics
coincided with those of stylistics; and «despite the diversity of approaches, nearly
all work reflects the conviction that style results from the characteristic selection of
options for producing a text or set of texts. Hence, we might look into the style of a
single text; of all texts by one author; of a group of texts by similar authors; of
representative texts for an entire historical period; and even of texts typical of an
overall culture and its prevailing language» (de Beaugrande and Dressler 1981,
p. 14).
In regard to literary studies, anthropology, sociology and conversational anal-
ysis, we limit ourselves to saying that all three disciplines deal with texts, the first
eminently regards texts that can be qualified as “literary”, the second regards the
text as an artefact of culture, the third regards texts from the perspective of their
exchange in relations between human subjects. Especially valuable for the
47
In addition to this, rhetoric owes its rebirth in that period to other causes as well. The art of
persuasion taught how to convince the general public, which was not reached easily persuaded by
complex scientific evidence. In mass society, these needs become progressively stronger, because
political and civil rights are extended to the entire population and, from an economic standpoint,
consumerism extends to the masses. On the subject of advertising, Barthes (1970), p. 21 begins his
discussion with Aristotelian rhetoric. See Chap. 3.
newly-formed science of textual study is the contribution provided by philology,

which studies texts from a historical perspective and from the point of view of their
material transmission.48
Last but not least, we must not overlook the role of semiotics, which provides
«the most comprehensive framework for the analysis of the aspects of language»
(my translation of Petőfi 1977, p. 199).
In the words of Petőfi: «the study of language (or semiosis or semiotics) was
divided by Morris into three branches: syntax, semantics, pragmatics. They can be
characterized approximately as follows: syntax regards only the relations between
linguistic expressions; semantics regards relations between expressions and the
objects to which they refer; pragmatics regards relations between expressions, the
objects to which they refer and the users or contexts of use of the expressions» (my
translation of Petőfi 1977, p. 199). Text linguistics appeared as a response to the
need «to give linguistic contributions to disciplines that deal with complete texts
[…], to explain—using a consistent text—a series of phenomena that cannot be
studied adequately in isolated sentences, […] to integrate semantic and pragmatic
data with grammatical data, […] to follow general scientific guidelines, including
integrated studies, thus avoiding the reductionism of other linguistic trends» (my
translation of Bernárdez 1982, p. 31). The disciplines that we have discussed,
however, are not «simple antecedents discovered a posteriori for the sake of text
linguistics. Indeed, they have a direct impact on its birth and its development, and
even today they are precursors of some of the most prominent trends» (my trans-
lation of Bernárdez 1982, p. 21).
The Development of Text Linguistics
Text linguistics, although it was just born, immediately became the centre of tur-
bulent development. The main direction of this development was represented by the
perspective of Weinrich, which had a large following; a second direction, a
minority, was the perspective of Coșeriu, and today it remains less followed. For
ease of exposition, we shall now examine this minority perspective briefly. In
Tübingen, where he was Professor of General Linguistics, Coșeriu developed his
perspective on the basis of the system of the mid-Fifties, and his work culminated in
the publication of Textlinguistik. Eine Einführung (1980).
In this work the Moldovan linguist proposes a form of text linguistics that is
articulated in three different directions: one form presumes to be linguistics per se
and argues that every analysis must be conducted on a text, since it is the vehicle of
all observable phenomena. According Coșeriu, this form of text linguistics is an
“extended” form (or, to employ once again the terminology we used before,
48
For the contribution of the philological perspective to an integrated science of textual study, see
La Matina (1994), pp. 21–49; La Matina (2001), pp. 21–92; and La Matina (2004), pp. 283–388.
sentence linguistics with an expansion of the domain). A second form of text

linguistics (that corresponds to the current that enjoyed the most success in the
Seventies, which will be discussed immediately below49) developed “transphrastic
grammars” designed to examine all the purely linguistic phenomena that undermine
traditional grammar’s explanatory power (e.g., topicalization, the use of pro-forms,
ellipses, replacement). According to the author, these phenomena fully justify the
existence of a discipline that studies them, but one configured only as a study of
“supersentences”, because the text is seen as a level of grammatical structure of a
language, while for Coșeriu it is a level of idiolinguistic structure, not rationally
necessary, but empirically very frequent.
The last form of Textlinguistik identified, and which truly deserves that name, is
the part Coșeriu deals with that «coincides with philology correctly understood and
also with hermeneutics» (my translation of Coșeriu 1980, p. 62). It is based on the
autonomy of the text and on the specificity of textual functions; for this, critics50
associated him with the semantic criticism of Antonino Pagliaro, a theory of textual
interpretation in which Coșeriu is referred to explicitly. For Coșeriu, this discipline
is thus a linguistics of meaning that analyses literary and poetic texts in particular.
In this way, the particular role of hermeneutics appears clearly, and is based not
only on language, but also on culture.
Despite Coșeriu’s affirmations, the discipline developed in the direction of
transphrastic grammars, and the main perspective of Textlinguistik was modified in
two successive steps: the formulation of textual grammars and of textual theories.51
The first step is the formulation of textual grammars (Textgrammatiken) to
describe textual competence (Textkompetenz).
What does textual competence consist of? The speaker of a language is able to distinguish
an incoherent agglomeration of utterances from a coherent text. […] The speaker may also
paraphrase a text, summarize it, condense it, i.e. grasp its overall meaning. The speaker is
able to notice if a text is interrupted or unfinished. He can give a title to a text, or vice versa,
match a text to a title […]. The speaker knows how to classify texts according to various
types or sorts of texts (Textsorten). […] All of the particularities of every normal speaker
justify – indeed, they postulate – the construction of a textual grammar that explains the
textual competence of the speaker (my translation of Conte 1977, p. 18).
49
Among the leading scholars listed in Coșeriu (1980) are Hartmann, Kummer, Koch, Dressler,
Schmidt, Weinrich and Petőfi.
50
First, by Donatella Di Cesare (2000) in her introduction to Italian translation of Coșeriu (1980)
and by Tullio de Mauro (2007) in his preface to Coșeriu.
51
These steps are identified in Conte (1977) but, in truth, the scholar declares that her distinction
«is not chronological, but typological. They are three types of a theoretical development, and not
necessarily three stages of a temporal sequence» (my translation of Conte 1977, p. 14). Later she
specifies that the types she identifies «[…] are, rather than types of text linguistics, ideal types,
Idealtypen (in the sense of Max Weber) for a possible history of text linguistics» (my translation of
Conte 1977, p. 22). We believe that the author’s thought is not excessively stressed if we claim
that they are not only typological, but today—quasi forty years after being written—they appear
chronological as well.
The largest amount of work on the concept of textual grammar was carried out
by a research team that had formed around Peter Hartmann.
The group of scholars was financed by the Deutsche Forschungsgemeinschaft
(DFG) and met in Constance, where Hartmann taught. The purpose of the Projekt
Textlinguistik was to generate Mr. K.’s favorite animal by Bertolt Brecht from a
system of abstract rules (a textual grammar) and an appropriate vocabulary. The
experiment failed, as foreseen directly by Werner Kummer in his Kummer (1972a)
and indirectly by János Sándor Petőfi in his Petőfi (1971),52 who participated in the
project.
Kummer triggered a controversy with the team of researchers (Jens Ihwe, János
Sándor Petőfi, Hannes Rieser and Martin Rüttenauer), grouped together by Peter
Hartmann in Constance. Arguing that the generation of Brecht’s text was an
assumption (insufficiently motivated) of the researchers and not a result of their
grammar, Kummer launched the first stone (1972a), which was followed by the
researchers’ response (Ihwe and Rieser 1972) and his counter-response (Kummer
1972b).
In spite of the failure of the project, two fortunate consequences followed: the
first was the founding, in 1972, of the series Papiere zur Textlinguistik/Papers in
Textlinguistics, the first volume of which was dedicated to the Constance project
and written by the Konstanz Projektgruppe Textlinguistik (PTK) in the 1974.53 The
second was two theoretical proposals that arose within that group of scholars, that
are very well known in the field of text linguistics, one by Teun Adrianus van Dijk
and one by Petőfi.
The model proposed by van Dijk, in general terms, works like this: a text begins
with a general idea; progressively, this idea is developed into specific meanings and
these then constitute in the form of sentences the individual passages of the text.
Conversely, given a text, it will be possible by means of special operations
(deletion, generalization and construction) to go back to the basic idea expressed by
the text. These operations, performed on the macro-structure, are carried out within
a procedure-based model, which uses (expanded) predicate logic to treat the deep
semantic structure of texts, which, according to van Dijk, is isomorphic with respect
to the deep semantic structure of the sentences.54
52
In Petőfi (1971) the difficulty of applying a transformational grammar to the generation of a text
is underlined.
53
The series, published by Helmut Buske in Hamburg, is a very valuable tool for the study of text
linguistics. It is accompanied by the equally prestigious Research in Text Theory/Untersuchungen
zur Texttheorie, founded in 1977 and published by Walter de Gruyter & Co. The two series are the
main tools with which to begin studying text linguistics.
54
This postulated isomorphism is very interesting from the perspective of psycholinguistics; the
macro-structures would thus possess a psychological reality, that is, they would be the global plans
underlying the production and reception of texts.
The theoretical work of Petőfi is divided into the three different periods that have
been identified in his more than forty years of study. In the first phase he formulated
a generative textual grammar within the research team working in Constance and
attempted to reconcile the positions of the two main currents in generative gram-
mar, acting as mediator between interpretative semantics (Chomsky, Katz and
Fodor) and generative semantics (McCawley and Lakoff). His textual grammar with
a non-linearly fixed base (Textgrammatik mit nicht-linear festgelegter Textbasis)
was conceived within a generative framework and was able to make use of con-
ceptual and operational logic. This grammar was generative because it was able to
foresee a deep structure of the text (textual base) which, through certain rules,
generated the linear manifestation of the text. The textual base was non-linearly
fixed, that is, it was composed of semantic representations that were
non-determinate with respect to the linear manifestation of the text. In the second
period, when he published Vers une théorie partielle du texte (1975), he dealt with
the problem of semantic representations according to the concept outlined by
Montague, and to the textual grammar formulated in the early period he added a
component dedicated to metric, rhythmic and phonic issues, and thus internal to the
text: this component was called co-textual (kotextuell). Semantic interpretation, in
terms of models and possible worlds, was entrusted to a contextual component
(kontextuell). Both components belonged to a partial theory of text (une théorie
partielle du text), referred to by the acronym TeSWeST (Text-Struktur
Welt-Struktur Theorie) and also had subsequent designations TeSReST
(= Text-Struktur Relatum-Struktur Theorie) and VeSReST (= Vehiculum-Struktur
Relatum-StrukturTheorie). The final result of Petőfian thought, which emerged in
the last period, is the full inclusion of semantics in the strict sense and the birth of a
new discipline that was given the name Semiotic Textology.55
The second phase was the one in which text linguistics became the field in which
textual theories were created. This, more than a phase, was a real breakthrough.
And this came about due to the fact that within the disciplinary statute of text
linguistics the role of semiotics began to change. Semiotics, in the conception of
Charles William Morris that we mentioned earlier, is divided into three branches:
syntax, semantics, pragmatics. The breakthrough consisted in the fact that text
linguistics fully incorporated the study of pragmatics.
The inclusion of pragmatics in the domain of text linguistics meant replacing the
goal of formulating a textual grammar suitable to describing textual competence
with a new goal, a more ambitious and methodologically problematic one.
This new goal was to formulate a text theory that could describe the speech act
within the context of the communicative situation. Taking the pragmatic aspects on
board changed the fundamentals of text linguistics: the role of semiotics increased
and, as the connection with analytical logic grew weaker, contact grew stronger
55
A comprehensive presentation of Semiotic Textology in Petőfi (2004), which is written in Italian
language. For English audience, see Giuffrè (2011a).
with so-called linguistic phenomenology,56 another current of research on natural

language.
It has already been pointed out that to produce semantic representations, truth
conditions must be established for an expression. Truth conditions can be estab-
lished based on a possible world, as well as on a communication context in which it
is possible to assign a relatum to an expression to be interpreted. In other words,
semantic interpretation always requires pragmatic information. Analytical logic, we
have seen above, had an impact on the history of sentence linguistics. It made and
makes use of symbolism and rigorous formalism. Different from and in opposition
to analytical logic, linguistic phenomenology—linked to the names of the “later
Wittgenstein” (i.e., in the so-called “second stage” of his studies), of John
Langshaw Austin, John Searle and Herbert Paul Grice—did not use the language,
techniques and symbolism of mathematical logic.
Instead, phenomenology employed ordinary language in a refined and regulated
way, both as object and as a metalanguage. The phenomenologists remain far from
the opinion that we can know or guess the meaning of expressions in natural
language without defining the intentions of the speaker, the communicative con-
texts and the effects that producers of text intend to bring about on the receiver. In
short, in direct opposition to the analytical logicians, they considered pragmatic
information a requirement for linguistic analysis. The so-called speech acts theory
began to develop.57
Very briefly, they argued that in linguistic communication, direct or indirect
actions are performed regardless of whether they are intended to be made. The text
is the result of an act: to produce a text is to act on the world. The acts that can be
made during a series of utterances are either direct acts (which take place in the very
act of enunciation: illocutionary acts) or indirect acts (a secondary and unpre-
dictable effect of utterance: perlocutionary acts). The inclusion of this set of
56
Linguistic phenomenology is the term used to refer to the research conducted by Austin in the
field of the philosophy of language. He argues that «our common words have many more nuances
in their potential use and are capable of more distinctions than what philosophers realise; and facts
related to perception (discovered by psychologists, for example, but also observed by ordinary
people) are very different from and far more complicated than what has been thought. It is essential
here, as elsewhere, to abandon […] the deeply rooted cult of dichotomies that appear to be neat
and tidy» (Austin 1979, p. 19). Austin (as well as his students Grice and Searle) accepted the basic
principles of the “use of language movement” (i.e. the philosophy of the later Wittgenstein) and
stated that «we need more of a framework in which to discuss these uses of language. […] I think
we should not disperse too easily and talk, as people are apt to do, about the infinite uses of
language. Philosophers will do this when they have listed as many, let us say, as seventeen; but
even if there were something like ten thousand uses of language, surely we could list them all in
time. This, after all, is no larger than the number of species of beetle that entomologists have taken
the pains to list» (Austin 1979, p. 234). As you can see, there is considerable distance between
Austin and linguistic phenomenology in general on one hand, and the thinkers of the “verification
movement” (the tradition of logical positivism that Carnap founded).
57
On the speech acts theory, see also Conte (1983).
problems in the domain of text linguistics meant scholars were forced to differ-
entiate their positions significantly on the basis of their orientation to the issues
themselves.
References
Modern Studies
Albaladejo, Tomaso. 1989. Retórica. Madrid: Editorial Sintesis.

Allan, Keith. 1992. Semantics. An overview. In International Encyclopaedia of Linguistics, ed.
William Bright, 394–398. New York-Oxford: Oxford University Press.
Austin, John Langshaw. 1979. Philosophical Papers, ed. James Opie Urmson and Geoffrey James
Warnock. New York-Oxford: Oxford University Press.
Bar-Hillel, Yehoshua. 1970. Language and Information. Selected Essays on their Theory and
Application, ed. Yehoshua Bar-Hillel. Reading, Massachusetts: Addison-Wesley.
Barthes, Roland. 1970. L’ancienne rhétorique. Communications 16: 172–229.
Benveniste, Émile. 1966. Problèmes de linguistique générale I. Paris: Gallimard.
Benveniste, Émile. 1969. Le Vocabulaire des institutions indo-européennes. Paris: Minuit.
Benveniste, Émile. 1973. Problems in General Linguistics. Miami: University of Miami Press.
Benveniste, Émile. 1974. Problèmes de linguistique générale II. Paris: Gallimard.
Bernárdez, Enrique. 1982. Introducción a la lingüística del texto. Madrid: Espasa Calpe.
Bierwisch, Manfred. 1965. Review of Z. S. Harris’ Discourse Analysis Reprints. Linguistics 13(1):
61–73.
Bloomfield, Leonard. 1933. Language. New York: Holt.
Bonomi, Andrea. 1973. La struttura logica del linguaggio. Milano: Bompiani.
Bonomi, Andrea. 1983. Linguistica e logica. In Intorno alla linguistica, ed. Cesare Segre, 148–
171. Milano: Feltrinelli.
Bonomi, Andrea, and Giorgio Usberti. 1971. Sintassi e semantica nella grammatica trasfor-
mazionale. Milano: Il Saggiatore.
Carnap, Rudolf. 1937. Logical syntax of language. New York: Humanities [Republished as
Carnap, Rudolf. 2001. Logical Syntax of Language. Abingdon (UK) - New York: Routledge].
Chierchia, Gennaro. 1992. Logica e linguistica. Il contributo di Montague. In Introduzione alla
filosofia analitica del linguaggio, ed. Marco Santambrogio, 287–365. Roma-Bari: Laterza.
Chomsky, Noam Avram. 1965. Aspects of the Theory of Syntax. Cambridge, Massachusetts: MIT
Press.
Chomsky, Noam Avram. 1957. Syntactic Structures. The Hague: Mouton.
Chomsky, Noam Avram. 1968. Language and Mind. New York: Harcourt Brace & World Inc.
Chomsky, Noam Avram. 1972. Studies on Semantics in Generative Grammar. The Hague:
Mouton.
Chomsky, Noam Avram. 1975. Reflections on Language. New York: Pantheon Books.
Chomsky, Noam Avram. 1981. Lectures on Government and Binding : The Pisa Lectures.
Dordrecht: Foris Publications.
Chomsky, Noam Avram. 1993. A minimalist program for linguistic theory. In The View from
Building 20, ed. Ken Hale and Samuel Jay Keyser, 1–52. Cambridge, Massachusetts: MIT
Press.
Conte, Maria-Elisabeth (ed.). 1977. La linguistica testuale. Milano: Feltrinelli.
Conte, Maria-Elisabeth. 1983. La pragmatica linguistica. In Intorno alla linguistica, ed. Cesare
Segre, 94–128. Milano: Feltrinelli.
de Saussure, Ferdinand. 1879. Mémoire sur le système primitif des voyelles dans les langues
indo-européennes. Leipzig: Teubner.
Coşeriu, Eugeniu. 1955–1956. Determinación y entorno. Dos problemas de una lingüística del
hablar. Romanistiches Jahrbuch 7 (1): 24-54.
Coşeriu, Eugeniu. 1980. Textlinguistik. Eine Einführung. Tübingen: Narr.
Cresswell, Maxwell John. 1973. Logics and Languages. London: Methuen.
Cresswell, Maxwell John. 1992. Truth-conditional and model-theoretic semantics, Vol. III. In
International Encyclopaedia of Linguistics, ed. William Bright, 404–405. New York, Oxford:
Oxford University Press.
Crystal, David. 2006. A Dictionary of Linguistics and Phonetics. Oxford: Blackwell.
de Beaugrande, Robert-Alain, and Wolfgang Ulrich Dressler. 1981. Introduction to Text
Linguistics. London-New York: Longman.
De Mauro, Tullio. 2007. Prefazione. In Eugenio Coseriu. Il linguaggio e l’uomo attuale. Saggi di
filosofia del linguaggio, ed. Cristian Bota and Massimo Schiavi, 9–16. Verona: Fondazione
Centro Studi Campostrini.
de Saussure, Ferdinand. 1916. Cours de linguistique générale, ed. Charles Bally and Albert
Sechehaye. Lausanne-Paris: Payot.
Dessons, Gérard. 1993. Émile Benveniste. Paris: Bertrand-Lacoste.
Di Cesare, Donatella. 2000. Introduzione. In Linguistica del testo. Introduzione all’ermeneutica
del senso, edited by Donatella Di Cesare, 11–19. Roma: Carocci.
Dowty, David. 1979. Word Meaning and Montague Grammer: The Semantics of Verbs and Times
in Generative Semantics and in Montague’s PTQ. Dordrecht-Boston-London: Reidel.
Dowty, David, Robert Wall, and Peters Stanley. 1981. Introduction to Montague Semantics.
Dordrecht-Boston-London: Reidel.
Dragos, Emil. 1986. Cohesion and coherence in Romanian linguistics (A historical survey). In
Research in Text Connexity and Text Coherence, ed. Michel Charolles, János Sándor Petöfi
and Emil Sozer, 179–197. Hamburg: Buske.
Dressler, Wolfgang Ulrich. 1972. Einführung in die Textlinguistik. Tübingen: Niemeyer.
Eikmeyer, Hans Jürgen. 1980. Formal methods in text semantics. In Some Aspects of Formal
Foundation in Text Semantics, ed. Hans Jürgen Eikmeyer, Wolfgang Heydrich and János
Sándor Petöfi, 1–18. Bielefeld: Projekt Formale Textsemantik (= Materialien des
Universitätsschwerpunktes Mathematisierung der Einzelwissenschaften, Heft XXVI).
Fries, Udo. 1975. Studien zur Textlinguistik. Braumüller: Frage- und Antwortsätze. Eine Analyse
an neuenglischen Dramentexten. Wien-Stuttgart.
Giuffrè, Mauro (ed.). 2011a. Studies in Semiotic Textology in Honour of János S. Petöfi
(= Sprachtheorie und Germanistische Linguistik Supplement 1). Münster: Nodus
Publikationen.
Giuffrè, Mauro. 2011b. How did text research lead to semiotic textology? In Studies in Semiotic
Textology in Honour of János S. Petőfi (= Sprachtheorie und germanistische Linguistik
Supplement 1), ed. Mauro Giuffrè, 9–26. Münster: Nodus Publikationen.
Greimas, Algirdas Julien. 1966. Eléments pour une théorie de l’interprétation du récit mythique.
Communications 8(1): 28–59.
Hall-Partee, Barbara. 1976. Montague Grammar. New York San Francisco London: Academic
Press.
Harris, Zellig Sabbetai. 1951. Methods in Structural Linguistics. Chicago: University of Chicago
Press.
Harris, Zellig Sabbetai. 1952. Discourse Analysis. Language 28: 474–494.
Harris, Zellig Sabbetai. 1963. Discourse Analysis Reprints. The Hague: Mouton & Co., Publishers.
Hartmann, Peter. 1964. Text, Texte, Klassen von Texten. Bogawus. Zeitschrift für Literatur, Kunst
und Philosophie 2: 15–25.
Hjelmslev, Louis Trolle. 1928. Principes de grammaire générale. Copenhague: AF Høst.
Hjelmslev, Louis Trolle. 1935. La catégorie des cas. Étude de grammaire générale. première
partie. Vol. Acta Jutlandica VII 1. Aarhus: Universitetsforlaget.
Modern Studies 33
Hjelmslev, Louis Trolle. 1943. Omkring sprogteoriens grundlæggelse. København: Akademisk

Forlag [English version published as Hjelmslev, Louis Trolle. 1969. Prolegomena to a Theory
of Language. Trans. F.J. Whitfield.Madison (Milwaukee) and London: University of
Wisconsin Press].
Hjelmslev, Louis Trolle. 1975. Résumé of a Theory of Language, vol. XVI. København: Nordisk
Sprog- og Kulturforlag.
Ihwe, Jens, and Hannes Rieser. 1972. Versuch einer Exploration des „Versuchs einer Exploration
der neuentdeckten Formelwälder von der Insel Mainau“ von Werner Kummer. Linguistische
Berichte 18: 56–58.
Katz, Jerrold Jacob. 1967. Recent issues in semantic theory. Foundations of Language 3: 124–194.
Katz, Jerrold Jacob, and Jerry Alan Fodor. 1963. The structure of a semantic theory. Language 39
(2): 170–210.
Kummer, Werner. 1972a. Versuch einer Exploration der neuentdeckten Formelwälder von der
Insel Mainau. Linguistische Berichte 18: 44–53.
Kummer, Werner. 1972b. Zum „Versuch einer Exploration des, Versuchs einer Exploration der
neuentdeckten Formelwälder von der Insel Mainau’ von Werner Kummer“. Linguistische
Berichte 19: 78–79.
La Matina, Marcello. 1994. Il testo antico. Palermo: L’Epos.
Lakoff, George. 1970. Linguistics and natural logic. Synthese 22(1): 1–17.
Lepschy, Giulio Cesare. 1979. L’arbitrarietà del segno. Stampatori Editore: Intorno a Saussure.
Torino.
Lepschy, Giulio Cesare. 1990. Storia della linguistica. Bologna: Il Mulino.
Lyons, John. 1977. Semantics, 2 vols. Cambridge: Cambridge University Press.
McCawley, James. 1972. Syntactic and Logical Arguments for Semantic Structures. Bloomington:
Indiana University Linguistics Club.
McCawley, James. 1981. Everything that Linguists Have Always Wanted to Know about Logic
(but Were Ashamed to Ask). Oxford: Blackwell.
McMahon, William. 1976. Hans Reichenbach’s Philosophy of Grammar. The Hague: Mouton.
Petőfi, János Sándor. 1971. Transformationsgrammatiken und eine ko-textuelle Texttheorie.
Athenäum: Grundfragen und Konzeptionen. Frankfurt.
Petőfi, János Sándor. 1973. Towards an empirically motivated grammatical theory of verbal texts.
In Studies in Text Grammar, ed. János Sándor Petöfi and Hannes Rieser, 205–275. Dordrecht,
Boston, London: Reidel.
Petőfi, János Sándor. 1975. Vers une théorie partielle du texte. Hamburg: Buske.
Petőfi, János Sándor. 1977. Semantica, Pragmatica, Teoria del testo. In La linguistica testuale, ed.
Maria-Elisabeth Conte, 195–223. Milano: Feltrinelli.
Petőfi, János Sándor. 2004. Scrittura e interpretazione. Introduzione alla Testologia Semiotica dei
testi verbali. Roma: Carocci.
Prince, Ellen. 1978. Discorse analysis in the framework of Z. S. Harris. In Current Trends in Text
Linguistics, ed. Wolfgang Ulrich Dressler, 191–211. Berlin: de Gruyter.
Reichenbach, Hans. 1948. Elements of Symbolic Logic. New York: Macmillan.
Stegmüller, Wolfgang. 1975. Hauptströmungen der Gegenwartsphilosophie. Eine kritische
Einführung. Stuttgart: Kröner.
Suppes, Patrick. 1973. Logic, methodology and philosophy of science IV. In Proceedings of the
4th International Congress for Logic, Methodology and Philosophy of Science, Bucharest,
1971, ed. Patrick Suppes, L. Henkin, Gr. C. Moisil and A. Joja. Studies in Logic, Methodology
and Philosophy of Science IV. Amsterdam: North-Holland.
Vianu, Tudor. 1941. Arta prozatorilor români. Bucureşti: Eminescu.
Weinrich, Harald. 1964. Tempus. Besprochene und erzählte Welt. Stuttgart: Kohlhammer.
Chapter 2
The Procedural Approach to Texts
There is no such thing as the “philosophy” of language. There

is only linguistics; and grammar is an essential part thereof.
Louis Hjelmslev, Principes de grammaire générale
In Chap. 1 we have shown that there is no simple scientific definition of the concept
‘text’ because although texts are linguistic artefacts, for a long time they were not
considered the highest and most important level of linguistic analysis. Instead,
general linguistics, which developed in Europe and the United States, has always
considered the sentence as the highest level of analysis.
During the twentieth century, in Europe and in the United States, the domain of
linguistics gradually expanded to stably encompass links with some neighbouring
disciplines, such as semantics. As a result, linguistics took on the task of recog-
nizing the text as a valid subject of analysis, considering it as a sort of maximum
extension. In the second half of the twentieth century, many linguists worked from
this perspective and ended up creating a new autonomous discipline, text linguis-
tics, which was able to integrate the research results achieved by semantics. In this
way, and in order to deal with the most controversial issues of semantics, scholars
also admitted pragmatics into the domain of text linguistics.
This circumstance created scientific debate between various scholars’ points of
view, specifically the positions taken by each with regard to specific problems and
the methods used to deal with them. Some of them employed the formalism of
analytical logic that had developed in the tradition of Frege-Lewis (see Chap. 1,
19–21); other scholars leaned toward the tradition of linguistic phenomenology,
cultivated by the “later Wittgenstein”-Grice current (see Chap. 1, 29, 30). For this
reason, the most outstanding scholars of text linguistics work in different per-
spectives according to the current they belong to.
Now, I shall attempt to illustrate the orientation of the following specific authors,
seeking to highlight the similarities and the differences between their points of view
and the other textualists’ ones. In order to do this, I wish to present a leading book,
the Introduction to Text Linguistics by Robert-Alain de Beaugrande and Wolfgang
Ulrich Dressler.

36 2 The Procedural Approach to Texts
This book has two versions: the English and the German one. Both of them were
simultaneously published in 1981 by the two authors, who had prepared a common
work, that originated two manuscripts, each of them in his author’s mother tongue.
In Italy the translation appeared for the first time in 1984 and a second renovated
edition was published in 1994. My work refers to the English version of 1981 and
to the second Italian edition of 1994.
This choice is a tribute to Robert-Alain de Beaugrande, a passionate scholar,
who loved the truth and the unfettered, no-holds-barred scholarly debate. Defying
the academic conventions, he put all of his research works on his web-pages for
free. I thought that the best way to remind him and his unconventional ideas was
using the page numbers of his free printable online English version of the
Introduction (1981) in my citations. It was available until 10 August 2012, date of
its archiving, due to copyright issues. His defiance to scholar conservative enter-
prise lives again in this work through this device.1
The Point of View of Dressler and de Beaugrande
On the back cover of the second Italian version of the volume (1994), you can read
as follows:
The science of the text does not pretend to reach the formal rigour typical of mathematics
and logic, and, to some extent, even of linguistics, since its unit of measurement – the text –
does not lend itself to absolute categorizations and regulatory paradigms. In the light of the
results achieved so far, the science of linguistic structures above the sentence level is
characterized by qualitative rather than quantitative predictions, and its main task is to
design models to represent the complex processes of production and reception of texts,
processes that are not observable in the laboratory, but reconstructed in the conceptual
framework of cognitive science and communication theory.
R.-A. de Beaugrande and W. U. Dressler, leading experts in the field together with
Schmidt, van Dijk and Petőfi, in this systematic and updated introduction intend to make
known to the general reader not only the methods and models of the discipline, but also the
close relationship it has with psychology, anthropology, literary criticism and semiotics (my
translation of de Beaugrande and Dressler 1994, p. IV).
Reading this page prompts two interesting questions that can help us identify the
different perspectives of the various textualists:
1. Why should text linguistics aspire to the formal rigour typical of mathematics
and logic, and, within certain limits, of linguistics [viz. sentence linguistics]?
2. Who sought to impose intricate mechanisms and a rigorous formalized approach
on the text, which is the unit of measurement of text linguistics?
To answer the first question, we must refer to the history of linguistics outlined
in Chap. 1. As stated before, in Europe and in the United States sentence linguistics
1
See Lemke and van Helden (2009).
The Point of View of Dressler and de Beaugrande 37
had come a long way: initially, formal methods were rejected. Later, linguistics
recuperated them, first in the prevailing algebraic approach (Harris and Chomsky),
and then in a logical approach (generative semantics and Montague). First through
studies of syntax and then through studies of semantics, formalization took hold of
linguistics, and though it had initially been isolated from neighbouring disciplines,
it had gradually extended its domain to include semantics. Therefore, the answer to
the first question is that text linguistics, born from the gradual formalization of
sentence linguistics during the twentieth century, was configured as a separate
discipline based on the formal rigour typical of disciplines such as mathematics and
logic.
To answer the second question, we must necessarily elaborate a broader anal-
ysis. The text that prompts the question derives from the point of view of Dressler
and de Beaugrande; therefore, we must first understand the perspective adopted by
the authors. To do this, let us look again at a part of the page just quoted:
the […] processes of production and reception of texts […] are not observable in the
laboratory, but reconstructed in the conceptual framework of general cognitive science and
communication theory.
This statement brings the status of text linguistics closer to the cognitive science.
Opposing the use of logical and mathematical formalism, the two scholars carry out
an expansion of the domain of linguistics. Considering semiotics as having been
divided into the three branches (syntax, semantics, pragmatics) identified by Morris
(see Chap. 1, p. 26) in their vision pragmatics assumes a specific weight. In fact,
they define pragmatics as «the domain of plans and goals, and questions of use are
freely treated in syntax and semantics as well» (de Beaugrande and Dressler 1981a,
b, p. 22). With regard to the actual use of language, they argue that «if human
language users are in fact demonstrably unable to make such a distinction [viz.
between sentences and non-sentences] consistently […] grammaticality of sen-
tences is only a default in a theory of language as human activity, that is, something
assumed in absence of contrary specification» (de Beaugrande and Dressler 1981a,
b, p. 24). This position—in contrast with generative grammar, which has its centre
in the distinction sentence/non-sentence—places the focus on pragmatics.
Going even further, both argue that «as the distinctions of sentence/non-sentence
and text/non-text lose importance, the gradations of efficiency, effectiveness, and
appropriateness gain […]. Those factors control what people say at least as much as
do the abstract rules of grammar and logic» (de Beaugrande and Dressler 1981a, b,
p. 24). This last position, however, critical of the use of logical and mathematical
formalism, which regiment texts rigidly, gave pragmatics a priority role because it
is able to process flexible principles.
In light of the above considerations, let us re-read the second question:
2. Who sought to impose intricate mechanisms and a rigorous formalized approach
on the text, which is the unit of measurement of text linguistics?
The “intricate mechanisms” and the “rigorous formalized approach” to which the
authors refer were employed by textualists belonging to a different tradition than
that of de Beaugrande and Dressler and who operated according to the logical and
mathematical formalism developed by the Frege-Lewis current. They are Teun
Adrianus van Dijk and János Sándor Petőfi, presented in Chap. 1, p. 28.
On the same page quoted above, they are mentioned as the leading experts in the
field of text linguistics, without specifying, however, that they belong to a tradition
different than that of the authors of the Introduction. Both theoretical proposals, that
of van Dijk and that of Petőfi, by integrating the perspectives of analytic philosophy
without ignoring the issues raised by the phenomenological approach, delve deeply
into the various problems; but it is clear that in the attempt to mediate between the
demands of the two fronts, they are vulnerable to criticism from both areas of
research. Indeed, they end up being too casual for the analytical philosophers and
too strict for the phenomenologists, and thus leave themselves open to criticism.2
The position taken by Dressler and de Beaugrande in the Introduction is a veiled
criticism: they accuse the two scholars, especially the Hungarian one, of excessive
rigour. This is evident by reading the passage in which the two scholars mention the
theories of Petőfi. They argue that:
Setting aside the technical details of Petőfi’s evolving model, we can view it as illustrative
of the issues which logic-based text theories will have to face. Either established logics are
employed, so that much of the texts’ nature is lost from view; or the logics are modified to
capture texts more adequately […]. Petőfi foresees intricate mechanisms to mediate
between real texts and logically adequate versions of texts. Whether this undertaking will
succeed, and whether it will then clarify the interesting properties of texts, remains to be
seen. Perhaps a less rigorous, formalized approach would do more justice to the approxi-
mative way humans use texts in everyday communication (de Beaugrande and Dressler
1981a, b, p. 19).
The approach to text studies suggested by Dressler and de Beaugrande avoids

the use of logical bases, and the point of view they propose is close to that of the
cognitive science. So, it is time to take into consideration the cognitive science.
The Scope of Cognitive Science
There is no single definition of “cognitive science”.3 First of all, scholars do not

even agree if we should speak of cognitive sciences or of cognitive science. In the
first case it would be clear that an attempt is being made to establish a field that
brings together different disciplines, which share the same object of study but have
different methods and knowledge. In the second case, however, the idea of a
unifying model is suggested, or at least of a synergy between disciplines that
2
On this point, see Chap. 1, p. 15.
3
For Italian audience useful introductions to the cognitive science are Tabossi (1998) or Legrenzi (2002).
The Scope of Cognitive Science 39
operate in the light of shared results. The cognitive science is relatively recent, if we
consider the year of their birth as 1978, when a Cognitive Science conference was
organized at the University of California in La Jolla, outside San Diego, California.
Despite the problem of knowledge that has always run through philosophical
thought, the conference attempted to photograph the state of the art of the cultural,
scientific and technological changes which since the Seventies had attracted great
attention to issues related to the human mind. Scholars, each according to his view,
include a number of different disciplines in the great family of the cognitive science.
This, however, does not prevent us from defining a broad group of disciplines that
constitute the backbone of this field. Since the subject that the cognitive science
examine is basically “knowledge”, they are faced with the task of answering
questions like: how does the cognitive system of a person work? How does it
reason, store information, perceive and “know” the world? In general, what do
intelligent systems consist of and how do they work? Can intelligent machines be
built, and how? To answer these questions, a group of different disciplines are
called upon to interact in search of an integrated perspective constituting the
framework of cognitive science.
The list of disciplines can be compiled in various ways; according to the needs of
this discussion, we shall take into account a small group of disciplines. For this
reason, in addition to cognitive psychology, which is one of the undisputed and
unanimously recognized pillars of the cognitive science, we shall also consider the
contributions of neuroscience, artificial intelligence research, as well as the lin-
guistic and philosophical research (considered simplistically as a single block).
Cognitive Psychology
In the first half of the twentieth century, behaviourism dominated the field of psy-
chology. The intent of this line of research was to make the discipline scientific by
adopting the experimental method and forgoing assertions not based on direct
observation. The behaviourists had to base their studies only on observable beha-
viour, excluding any mental process from theories that had been or were being
formulated. In other words, the mind had to be considered a sort of “black box”,
something unobservable and whose functioning, therefore, could only be the subject
of unprovable hypotheses. For this reason, experimental psychology for decades was
based on the stimulus-response mechanism: the subject (human or animal), sub-
jected to certain stimuli, produced responses that the researcher observed and
attempted to classify. This gave rise to a great mass of experimental data, which
highlighted the existence of certain mechanisms; these mechanisms, however, were
anchored to situations that were simplified, limited and unrepeatable.
In the Fifties, interest in the mental mechanisms on which human behaviour
depended became more widespread. Internal processes, as such not directly
observable, regulate these behaviours; nevertheless, a number of more or less
general hypotheses were progressively elaborated about the way these processes
worked. Although still unaware of itself, this was the birth of cognitive and
behavioural psychology.
In addition to its interdisciplinary configuration, cognitive psychology had other
significant characteristics. First, it was interested in the cognitive processes (per-
ception, attention, memory, language, thinking, creativity), and these were recog-
nized as possessing both structural autonomy and reciprocal interrelationships and
interdependence. In the second place, the mind was conceived as an information
processor, with a predetermined organization of a sequential type and a limited
processing capacity along its transmission channels.
Neuroscience and Artificial Intelligence Research
Cognitive psychology relied on an analogy between mind and computer that was
based on notions of information, channels, sequences of transmission and pro-
cessing of information, the input and output of data to and from the processor and
memory structures. To explain this structural and functional organization, flow-
charts were used, formed by units (boxes) having specific tasks (perception,
attention, etc.) and by communication paths. In early models, information pro-
cessing was conceived as a horizontal multi-stage process; when the operations of
one stage had been completed, the next stage began, and so on.
In the Seventies, new models were presented that evidenced both the possibility
of sending feedback from one stage of the process to previous ones and the pos-
sibility for operations from a later stage to be activated even if the former stages had
not already processed the information.
The contribution of neuroscience refers to the study of the functioning of the
human nervous system, especially the fields of neuropsychology and neurophysi-
ology, through analogy with computers.4 With the passage of time, more and more
knowledge was acquired and today the influence of neuroscience on cognitive
science is significant. The hypothesis of so-called neural networks derives directly
from neuroscientific research.5
Artificial intelligence research comprehends all studies aimed at building intel-
ligent machines (or, in other terms, machines that are able to simulate intelligent
behaviour). Today, disappointed by the results of artificial intelligence, one prefers
to speak about expert systems, i.e. systems that can solve specific problems.
A prime example could be that of medical diagnosis; a second example, however,
could be represented by HCI.6 It was pointed out that a user who is not familiar with
a given system still manages to interact with it, recalling experiences and models
4
Of particular interest is Tabossi (1998).
5
See in particular Changeux (1983).
6
On HCI (= Human-Computer Interaction), see Chap. 1, footnote 23.
acquired in similar situations.7 In general, the process through which many tech-
nological innovations are metabolised seems largely metaphorical. Often, to be
accepted by potential users, a new technology must present a familiar look. In other
words, it needs to present itself in such a way that it expresses continuity with the
past and parallelism with other existing technologies (same appearance, same
functions, etc.). This way the user will tend to use models of behaviour and user
experience that he or she had developed in other circumstances. The same thing
happens with the spread of a new medium,8 and with human language, which takes
advantage of the global models that we discussed earlier.
Linguistic and Philosophical Research
A large contribution to the cognitive science had been made by scholars studying
semantic issues in the linguistic and philosophical research area. The father of the
idea of the ‘modular mind’, one of the most popular in this field, is Jerry Fodor,
whose semantic model, developed together with Katz, we outlined previously.
Another important linguistic contribution is the theory of the metaphorical nature of
the conceptual system by George Lakoff and Mark Johnson.
The two scholars assume that there is no difference between a semantic system
and a conceptual system. A conceptual system is the basis of knowledge of the
world that individuals have, and therefore of the way they interact with it. The
concepts are instruments of categorization of reality, which allow us to pigeonhole
individual everyday communicative interaction in larger (and known) groups.
However, there are also abstract concepts, with which we build our reasoning.
These concepts, and the reasoning in which they are used, are the way individuals
represent reality; we rely on them to build assumptions and create behaviours that
we believe contribute to our survival. It thus follows that the conceptual system is
fundamental both for thought and action. This conceptual system is manifested
mainly through verbal language. Simplistically, we could say that nouns express the
categories in which individuals classify worldly objects (concrete and abstract), and
some conjunctions represent the types of logical connections used in reasoning.
Other elements (visual, audio, etc.), however, also belong to the conceptual system.
Lakoff and Johnson stress the high frequency of metaphors in speech, not for
their aesthetic function, but as evidence of the structure of the conceptual system.
The idea of Lakoff and Johnson is to study this conceptual system (the way the
7
See Mantovani (1995), pp. 132–138.
8
Bolter and Grusin (1999), inspired by McLuhan (1962), use the concept of ‘remediation’, i.e. the
process by which every new medium of communication conserves certain characteristics of its
predecessors.
concepts are structured in and among themselves) through its manifestation in

verbal language. They note that: «But our conceptual system is not something we
are normally aware of. In most of the little things we do every day, we simply think
and act more or less automatically along certain lines. Just what these lines are is by
no means obvious. One way to find out is by looking at language. Since com-
munication is based on the same conceptual system that we use in thinking and
acting, language is an important source of evidence for what that system is like»
(Lakoff and Johnson 2003, p. 4).
According to the two researchers, the concepts themselves have a metaphorical
organization and to prove this they take as an example the concept ‘argument’ and
the metaphor of “war”. This principal metaphor (that is, the idea that in some
respects an argument is like a war) has different realizations in language:
ARGUMENT IS WAR
Your claims are indefensible.
He attacked every weak point in my argument.
His criticisms were right on target.
I demolished his argument.
I never won an argument with him.
You disagree? Okay, shoot!
If you use this strategy, he’ll wipe you out.
He shot down all of my arguments.
[…]
It is important to see that we don’t just talk about arguments in terms of war. We can
actually win or lose arguments. We see the person we are arguing with as an opponent. We
attack his positions and we defend our own. We gain and lose ground. We plan and use
strategies. If we find a position indefensible, we can abandon it and take a new line of
attack. Many of the things we do in arguing are partially structured by the concept of war.
Though there is no physical battle, there is a verbal battle, and the structure of an argument
– attack, defense, counterattack, etc. – reflects this. It is in this sense that the
ARGUMENT IS WAR metaphor is one that we live by in this culture; it structures the
actions we perform in arguing.
Try to imagine a culture where arguments are not viewed in terms of war, where no one
wins or loses, where there is no sense of attacking or defending, gaining or losing ground.
Imagine a culture where an argument is viewed as a dance, the participants are seen as
performers, and the goal is to perform in a balanced and aesthetically pleasing way. In such
a culture, people would view arguments differently, experience them differently, carry them
out differently, and talk about them differently. But we would probably not view them as
arguing at all: they would simply be doing something different. It would seem strange even
to call what they were doing “arguing.” Perhaps the most neutral way of describing this
difference between their culture and ours would be to say that we have a discourse form
structured in terms of battle and they have one structured in terms of dance. This is an
example of what it means for a metaphorical concept, namely, ARGUMENT IS WAR, to
structure (at least in part) what we do and how we understand what we are doing when we
argue. The essence of metaphor is understanding and experiencing one kind of thing in
terms of another. It is not that arguments are a subspecies of war. Arguments and wars are
different kinds of things – verbal discourse and armed conflict – and the actions performed
are different kinds of actions. But ARGUMENT is partially structured, understood, per-
formed, and talked about in terms of WAR. The concept is metaphorically structured, the
activity is metaphorically structured, and, consequently, the language is metaphorically
structured.
Moreover, this is the ordinary way of having an argument and talking about one. The
normal way for us to talk about attacking a position is to use the words “attack a position.”
Our conventional ways of talking about arguments pre-suppose a metaphor we are hardly
ever conscious of. The metaphor is not merely in the words we use - it is in our very
concept of an argument. The language of argument is not poetic, fanciful, or rhetorical; it is
literal. We talk about arguments that way because we conceive of them that way – and we
act according to the way we conceive of things (Lakoff and Johnson 2003, pp. 4–6).
According to their hypothesis, the conceptual system is organized mainly in a

metaphorical way. For example, this occurs when some of the concepts—the ones
that are particularly abstract or distant from the world of experience—would
otherwise be unclear. Metaphorization lets us see certain elements in terms of
another better known and more understandable concept; it, in turn, may have been
metaphorized previously.
A typical example would be abstract concepts. Probably, the first units of
meaning with which individuals confront each other are those related to the per-
ception of the world, which is, then, the basis of knowledge and of the entire
conceptual system. On this basis individuals build the remaining part of their
conceptual system, often thanks to subsequent metaphorical extensions.
In this regard, the two scholars observed the following human behaviour: «we
typically conceptualize the nonphysical in terms of the physical—that is, we con-
ceptualize the less clearly delineated in terms of the more clearly delineated»
(Lakoff and Johnson 2003, p. 59).9 As we saw in the case of “argument”, which can
be metaphorized through war or dance,10 different metaphors will produce different
concepts or different aspects of the concepts.
Metaphors do not only structure individual concepts, but also sets of concepts.
Some initial concepts can thus be used to arrange the relations between other
concepts amongst themselves. Of great importance are the so-called metaphors of
9
In general, Lakoff and Johnson define as structural metaphors those that structure a concept in
terms of another concept. They insist on the idea of structuring because, as we have already seen,
conceptual metaphors never exist as an isolated case. In other words, if, for example, I am
considering the metaphor “discussion is a war”, I will not stop at this initial consideration, but I
shall articulate the metaphor by perceiving a number of similarities between the situation “dis-
cussion” and the situation “war”. The two concepts, however, never overlap completely (otherwise
they would represent the same concept). This means that the metaphor necessarily highlights
certain aspects of the metaphorized concept (those that exhibit a similarity with the other concept),
while it neglects or hides the others.
10
For a proposal of interpreting ‘argument’ as a metaphor of ‘dance’, on the basis of Hegel’s
thought, see Giuffrè (2013).
orientation, so defined because many of them relate to spatial orientation: up-down,

in-out, front-back, etc.11
This broader conception of metaphor, understood as a cognitive mechanism, is
the basis of the global models in the procedural approach of Dressler and de
Beaugrande, which addresses the problem of identifying the way in which these
models are used by people during verbal interaction.
Procedural Semantics
The Introduction ascribes a central role to semantics in the procedural approach.

The authors argue that in order to ascertain what expressions mean, one must build
a system of operational semantics that operates on the basis of human thought
processes, unlike the logical semantics of the structuralists, the semantic interpre-
tation of Katz and Fodor and the model-theoretic semantics adopted by the van Dijk
and Petőfi. According to Dressler and de Beaugrande, procedural semantics is the
heart of the approach to texts.
Since procedural semantics was born as part of the cognitive science, it has not
been universally defined, as it is not part of a single perspective. Indeed, three
qualifying aspects can be identified: first, the relationship between semantics and
understanding; second, semantics lacks autonomy in the proper sense; finally, the
relationship between meaning and concepts.12
The procedural approach makes semantics a kind of theory of understanding,
which undertakes the task of describing the ways individuals understand linguistic
expressions. On this level, classification schemes are developed, to which we have
referred above in relation to global models. Because understanding is a mental
activity, meanings become “cognitive content” and for this reason, the mental
processes that construct meanings and make them part of the overall conceptual
framework are of capital importance.
Procedural semantics is non-independent, at least from two points of view: if it is
isolated from all cognitive abilities, the study of meaning cannot be addressed; in
the procedural perspective, linguistic knowledge cannot be separated from
extra-linguistic knowledge, which is related to factual reality. This is due to the
interaction and reciprocal influence of cognitive models and linguistic models.
11
These spatial orientations derive from the human body itself and how it functions in the physical
environment. The metaphors of orientation give the concept a spatial orientation. For example, in
expressions such as “Today I am in high spirits”, the expression for “happy” has a high position.
The fact that the concept it contains is oriented upwards determines a spatial location for that
mood. These metaphorical orientations are not arbitrary, as they have a basis in physical and
sensory experience. But, although the opposition up-down, in-out, etc. are physical in nature, the
metaphors of orientation based on them can vary from culture to culture (Lakoff and Johnson
2003, p. 33).
12
See Violi (1997).
Procedural Semantics 45
The third key point is the relationship between meanings and concepts.
Compared to Saussure’s view, which decreed the priority of the linguistic system
over forms of thought and therefore discriminated clearly between linguistic
meaning and conceptual meaning, the procedural perspective affirms there is a
direct connection between the two elements. According to the theoretical positions
of the various authors, this connection has different configurations: for some,13
language retains a clearer specificity and has greater variations with respect to a
more defined conceptual framework; for others,14 semantic and conceptual cate-
gories are more overlapping; the latter is the case of Dressler and de Beaugrande.
In this second perspective, some have even hypothesized a single level of mental
representation, which is considered as a conceptual framework containing infor-
mation related to language, perception, motor skills, etc. with probable corre-
spondences and possible compatibilities. On this level procedural semantics
investigates the relationship between language and perception, in an attempt to find
parallels between the linguistic system and the spatial-perceptual system.
Procedural semantics, given its mentalist assumptions, does not accept the idea that
meanings are linked to states of reality (while the perspective of logical semantics
does), but assumes that language and the perceptual system interact.15
Dressler and de Beaugrande explicitly restate their positions towards the prob-
lem of semantics as a whole. The two authors believe that framing the texts and
knowledge about the situation in which communication takes place within a logical
system is probably absurd.16 Rather, it appears more convenient for them to act in
the opposite direction: models must be built that are acceptable from the cognitive,
perceptual and sociological point of view, and only later should we search for types
of logic to use as a formal basis. In other words, they refute the perspective of
logical-philosophical semantics Chap. 1, p. 30.
In the aforementioned line that includes Frege, Russell, the “later” Wittgenstein,
Austin, Grice and Searle, the philosophy of language developed by focusing on the
philosophical understanding of the essence of language and how it works, thanks to
Willard van Orman Quine, Donald Davidson, Nelson Goodman, Saul Kripke and
Hilary Putnam. Their studies made it possible to go beyond two paradigms of
linguistic philosophy: the first is the one that identifies philosophical problems with
13
This is the position of Fillmore (1968).
14
Jackendoff also holds the same position and has spoken out against the “syntax-centric” nature of
Generative Grammar. In the various versions of Generative Grammar listed in Chap. 1 footnote 20,
Chomsky argues that syntax is the only generative component within a language. Jackendoff
instead believes that phonology, syntax and semantics are generative, and that they are connected
to each other reciprocally by means of special components having their own rules. Rejecting the
mainstream of syntax-centric Generative Grammar, the contribution of cognitive semantics—
which Jackendoff helped develop—principally concerns meaning. He disputes that it is syntax that
determines semantics, and not, however, the contrary. Syntax requires a relationship with
semantics to be able to produce eutactic outputs; see Jackendoff (1996, 2002).
15
On the whole issue see Traini (2006).
16
See de Beaugrande and Dressler (1981a, b), pp. 55–56.
linguistic problems, whereby language itself would be the cause (this first model,
which we mentioned earlier, may be an indication of a sort of “dissolution ten-
dency”); the second paradigm, however, is the one that identifies philosophical
problems as problems concerning the meaning of words. From this perspective,
philosophy should help solve the problems by precisely verifying the meaning of
words (“resolution tendency”).17 Both trends share a lack of interest in the psy-
chological component of meaning (anti-psychologism) and a strong interest in the
relationship between natural languages and the real world (referentialism).
In concrete terms, this relationship was theorized in two ways: without any
mediation between linguistic sign and extra-linguistic referent, or with mediation,
that is, through categories that connect signs and referents. On the one hand, the
relationship without mediation is supported by Quine and Kripke, for example, who
consider nouns and other word classes connected to the referents on the basis of an
act of designation that is valid for all the possible worlds in which the element
exists. On the other hand, the relationship with mediation is typical of the proce-
dural semantics of de Beaugrande and Dressler. According to them, logic is in no
way able to explain a large number of complex sequences of reasoning that humans,
however, are clearly capable of performing.
In expressing this idea, they are much closer to the orthodox tradition of
structural linguistics, represented by the position of Hjelmslev (from whom we took
the epigraph of this chapter, provocatively), than to van Dijk and Petőfi. This gives
an idea of the distance between their perspective and that adopted by the other
textualists.
In the perspective of the procedural approach, knowledge and meaning are
sensitive to the situations in which they are used. In principle, the connection
between a concept and the relationships that it activates in the mind of an individual
can be considered a case of problem-solving. It thus follows that it is possible to
deal with this situation in the manner described previously. When individuals who
use a text come upon vague portions of content, they build more or less plausible
hypotheses on what the text means.
Processes such as making inferences are also performed by the participants in a
conversation in complete harmony with the conditions under which they receive the
text, that is to say, with the communicative situation. So the focus of research
becomes how to classify and systematize the ways in which these events occur, and
not to prove that all individuals repeat the same process. Even in this respect,
therefore, the authors of the Introduction distance themselves from logic-
philosophical semantics.
Dressler and de Beaugrande’s procedural approach to semantics is also a sig-
nificant change compared to Montague’s model-theoretic semantics. While the
philosopher tried to change the instruments of formal logic in order to regiment
natural languages, the procedural approach overcomes the problem at the
foundation.
17
See Marconi (1999).
Since natural languages obey global cognitive models that are flexible according
to the situations in which they are used, it is impossible to find formal means able to
represent all possible elements: in fact, a part of these elements is oriented by
context; however, the other elements are conditioned by aspects that could even be
precognitive and justified only on the basis of processes of perception or cognition.
In the procedural approach, language is no longer an infallible mechanism, as the
logic-oriented tradition of linguistics of the twentieth century had considered it.
The procedural approach does not require increasingly precise semantic markers
to be set in order to constitute the logical apparatus of a textual theory; it indicates
the need to engage in the search for types of global cognitive models on the basis of
which textual production and reception functions. The point of arrival is not
knowing how language works in general (the scope of general linguistics), but
finding out how the human mind operates to achieve its goals (the scope of cog-
nitive science), especially when it plans to produce texts as outcomes of the human
faculty of language (the scope of text linguistics).
For Dressler and de Beaugrande, the centre of the epistemological framework of
linguistics is not logic, but the realm of cognitive science. Linguistics is the study of
the human faculty of language, one of the faculties the human mind employs to
achieve its goals; but language is expressed concretely only through texts.
Therefore, the only linguistics possible is text linguistics.
It is useless to ask how a language works; instead, it is much more important to
set out clear principles on which the cognitive ability of men and women operates
and see how, in each case, that faculty has to bend itself to conform to useful
models within communicative interaction in order to fulfil various purposes. And
these models are not the same for everyone (and thus are not general and universal,
as competence is), but the result of the interweaving of a common physiological
basis and individual experiences of perception and cognition.
A comprehensive semantic theory should integrate the three dimensions exam-
ined above. The first is the intralinguistic dimension, i.e. the relationship between
the elements of a linguistic system. This dimension, although we have repeatedly
noted that it is unable by itself to give an explanation of meaning, is nevertheless
important, because it puts the focus on the social and cultural aspects of meaning
and gives an account of the differences between languages (linguistic relativism).
The second dimension is the cognitive dimension, which refers to the relationship
between the lexical structure and the conceptual structure. We must consider
whether it is appropriate to maintain a fixed distinction between lexical units and
deep conceptual schemas. The third dimension is the extra-linguistic dimension,
that is, the relationship between language and the world. The key issue here is the
role of experience, to which language refers. The problem of reference is formu-
lated in the perspective of semiotics, whereby an objective reference to an external
reality is no longer of primary importance, but rather a reference to experiential
content. Certain lexical aspects seem to strongly confirm the importance of per-
ceptual experience in the constitution of meaning; and since a portion of perceptual
experience is irrepressibly individual, there will be an irreducibly subjective area in

language, too. Therefore, it will be necessary to give up the tradition of
logical-philosophical semantics and also the path followed by van Dijk and Petőfi,
based on Montague’s model-theoretic semantics.
The irreducibly subjective element had already been glimpsed by Hjelmslev,
who had argued that in language there is an irreducible pre-logical element. This
consideration had prompted him to say that «there is no such thing as the “phi-
losophy” of language. There is only linguistics; and grammar is an essential part
thereof» (Hjelmslev 1928, p. 27), as I indicated in the epigraph.
Part of the literature18 argues that in human communication there are at least two
kinds of knowledge: the first, called “declarative knowledge”, concerns beliefs
about facts in the sense of the organization of events and real world situations; the
second, called “procedural knowledge”, refers to facts or beliefs adapted to certain
types of uses and/or operations.19 Meaning, as a property of language inserted in a
text, is configured as a special case of acquisition, memorization and use of
knowledge in the context of all the activities carried out by individuals. Since
usages of language are highly differentiated, they are normalized according to social
agreements.
Linguistic expressions used during a communicative interaction activate rela-
tions with corresponding concepts that are deposited in a mental workspace of
memory.20 This workspace, from a functional standpoint, is larger when the rela-
tionship between the text reception and production (for example, in the case of
participation in a dialogue) activates integrated and memorized simple models.
These models may have different aspects according to the processing demands of
the moment.
In procedural semantics, the memorization and use of knowledge are regulated
by two different principles.21 The first principle, defined as episodic, depends on the
personal experiences of the individual; the second principle, inherent, reflects the
models governing the organization of knowledge, i.e. the structures of events and
situations. Episodic knowledge must be reconnected to the situation of the specific
circumstances in which the individual experiences occur; and so it is burdened with
many accidental traits. Intrinsic knowledge, however, is related to common traits.
The importance of global patterns is evident in the production and reception of a
text: a topic, in fact, may be articulated through the repeated usages of modules.
According to the principle of economy, global patterns reduce the complexity of
processing. Therefore, language should not be separated from people’s experience;
rather, it should be placed within models whose use can be explained on the same
18
Miller and Johnson-Laird (1976), Winograd (1976), Levesque (1977), Schneider (1978),
Levesque and Mylopoulos (1979).
19
Goldstein and Papert (1977) and Winston (1977).
20
This issue was dealt with in detail by Miller and Johnson-Laird (1976).
21
The point is dealt with in further detail by Tulving and Donaldson (1972).
level as other perceptual and cognitive processes.22 If text linguistics must perform
this task, it must necessarily move away from logic and the philosophy of language
and draw closer to the cognitive science.
The Procedural Approach
According to the procedural approach, the object being analysed—a text—is

considered to be the outcome of a process. And if a text is a material object
resulting from the completion of a process, linguists have the task of describing the
various levels of language according to how they are used. The Introduction, as we
have seen, recognizes that even different traditions of linguistics agree that language
has a “systematic nature”, by virtue of which each of the elements belonging to the
different levels of analysis has a function that contributes to the overall operation.
Despite this, research has produced different outcomes because after identifying the
levels of analysis, linguistics has been primarily concerned with the systems of the
minimal units.23
The systematization of an object being analysed depends on the scientific pur-
pose to be achieved. To describe an object, it is sufficient to identify the set of
properties that make it classifiable in a reliable manner; to explain an object,
however, one must identify the principles by which the object has certain proper-
ties, and verify the cases observed in empirical reality. Descriptions of a language
are without doubt possible and often exclude the considerations necessary to
explain it. Otherwise, to explain language, the concept of interaction should be
employed rather than that of modularity.24
In modularity, the factors of a process are viewed as independent of each other;
in interaction, however, the components are seen to interlock and control each
other. Modular systems are more easily realized; this is why the linguistic models
considered until now are modular, that is, the theoretical paradigms within which
structuralists and generativists carried out their work. Ultimately, they do not offer
effective operating systems for the comprehensive analysis of how language is used.
Actual communicative behaviour can be explained only by the use of interactive
systems.25 In the real time of human communication, a linguistic model that
maintains an autonomous syntax (such as Generative Grammar) finds itself facing a
sort of combinatorial explosion: a calculation of the structures and the alternative
22
In this direction Minsky (1975), Miller and Johnson-Laird (1976), Rumelhart (1977).
23
See de Saussure (1916), Hartmann (1963), Chomsky (1965), and Coșeriu (1975).
24
25
Walker (1978) reports a series of tests that support the idea that abstract distinctions are not
recognizable in the verbal processes but can only be derived from the communicative situation.
readings of a given communication, even a simple one, would require astronomi-

cally vast amounts of time to analyze.26
Basing their work on this idea, Dressler and de Beaugrande affirm the need to
study texts in human communication using the procedural approach. It, ultimately,
does not bother to identify structural units; rather, it focuses on the operations that
regulate the units when language systems are implemented. From a procedural
perspective, the text is defined as a result of the operations regulating the units in
the use of a system. Both argue that by employing the procedural approach, one can
empirically evaluate under realistic conditions the accuracy with which a person
produces a text, placing the morphological and syntactic material in sequential
order and then using it in actual practice.27
According to them, the validity of the models used in the procedural approach is
to be verified in relation to natural human activities; therefore, scales of different
priorities must be established on the basis of the object being studied. Furthermore,
it is essential to develop a new observational model for some of the “traditional
issues” of traditional linguistics.
The concept of ‘distinctive trait’, for example, should include the features
detected by experimental phonetics and psycholinguistics, which should comple-
ment the information derived from descriptive linguistics. In addition, the structural
analysis of utterances to which different meanings can be assigned—a recurring
issue in generative linguistics—could be oriented in the direction of the processing
strategies of psycholinguistics, which deal with the ways in which individuals
exclude or eliminate a posteriori the ambiguities in everyday language. Even the
concept of ‘assumption’ should be modified in its application to the single utter-
ance, developing a macro-theory that deals with the general application of “world
knowledge” in human communication, which would fall partially in the domain of
behavioural psychology. Finally, the interest of logic in the ‘existence of objects’
and the ‘truth conditions’ of sentences should be replaced with a search for the
minimum cognitive thresholds that act as a dividing line, in order to understand the
extent to which participants in a communication are able to recognize objects and
believe what is said.
While Dressler and de Beaugrande do not propose to set aside the distinction
between sentence and non-sentence, they believe that we must assume that speakers
are normally able to recognize the grammaticality of sentences without doubt and in
a consistent manner. That said, it must be admitted that in the absence of expressly
contrary indications, the standard case of human communication is valid utterance.
Only those realizations that absolutely cannot be tolerated in human communication
are non-texts because they clearly contravene the criteria of textuality: this occurs
when participants in a conversation are totally unable to recognize cohesion and
coherence in the language produced.
26
In the early models of computerized language processing, this factor was particularly evident, as
pointed out by Woods (1970).
27
See de Beaugrande and Dressler (1981a, b), p. 25.
The Procedural Approach 51
Text Production and Reception
Communication functions as a constant upheaval and reestablishment of stability

through the interruption and restoration of the continuity of a text. By attributing
value to the continuity of relationships and connections, the processing of a text
presents itself as a formal problem-solving process that takes place in successive
stages dominated by various principles. The problem is considered solved when
there is a connection without interruptions between the original state and the end
state.28
The two states can be connected by employing three different research strategies
to solve the problem: in the first case, the problem-solver tries to find a text’s
purpose, proceeding by paying little attention to alternatives; when he encounters
an obstacle, he retreats just far enough to resume progress towards the finish line
(depth-first search). As a second alternative, the problem-solver considers only a
partial and more attainable goal and evaluates alternative paths to reaching it; he
then repeats this process until he reaches the primary purpose (breadth-first search).
The last research strategy is the means-end analysis, through which the solver
identifies the key differences between initial and goal state and then gradual reduces
them one by one. Each state encountered on his path becomes a control centre from
which he can move in any and all directions.
The Introduction defines the model of text production and reception as a
non-rigorously sequential arrangement of phases of processing dominance.
Dominance is invoked because it presents two advantages: to begin with, it is not
necessary to assume that the processes of a given phase suspend those of other
phases, and moreover, it lets us go beyond the opposition between modularity and
interaction discussed earlier.29
In text processing, the first phase is the planning phase. A person who produces a
text has a purpose.30 From this point of view, text production is only an interme-
diate step in achieving this purpose. The producer, through a means-end analysis,
attempts to evaluate which text among the many that could be produced will help
him the most in reaching his goal in a given situation.
Once the text type has been selected, the second phase begins: ideation. The
authors propose an explicit correspondence between the concept of ‘ideation’ and
the inventio of classical rhetoric, the search for ideas. The inventio is perhaps the
most important part of rhetoric for Aristotle. It is, as Barthes (1970) says, more than
invention, it is discovery. It is the inventio that must guide our search for the proof,
that is, the building blocks of our discourse. Aristotle’s main concern, contrary to
that of Cicero and Quintilian a few centuries later, is to construct solid and con-
vincing reasoning. The proof Aristotle speaks of can essentially be of two types:
technical and non-technical. Non-technical proof is not derived from our reasoning
28
29
30
Petőfi (2004) and van Dijk (1972) define this the ‘dominant intention’.
and our speech, but comes in some manner from the extralinguistic universe. It is
what in a modern courtroom trial we would call the evidence. For Aristotle,
technical proof is much more important because unlike non-technical proof, it is not
ready-to-use and must be processed first. It is, in fact, the fruit of reasoning, which
can occur through induction or through deduction (as in dialectics). Falling within
the category of inventio are also the so-called topica, i.e. the set of arguments (or, to
be more precise, more or less abstract argumentative models) that are available for
use in any discourse and which a good speaker must always be ready to retrieve at
the appropriate time. A prime example is political oratory, in which the producer
must often conceal the true purpose of his text. An idea is a conglomerate of
conceptual contents that offers several control centres for the creative and sensible
text production. The transfer of a design structure onto a idea is complicated to say
the least; the degree of difficulty increases when it is not useful to talk openly about
one’s purpose.
The third phase is the development, whose function is to expand, develop and
connect the ideas that have been selected. The development is a sort of search for
the various arrangements of content stored in memory. The development oscillates
between two extremes: recalling combinations without altering their contents and
connecting contents in a new way to create original combinations.
Through expression, the fourth phase, the contents that have been gathered thus
far are arranged and links between the various levels are organized so as to create an
organized overall structure. If the contents have already been expressed in the past,
we can speak of preferences in the choice of expression.
The search for proper expression can be seen as partially corresponding to the
concept of dispositio of classical rhetoric. The so-called dispositio is the second part
of rhetoric. Its function is to arrange material from the inventio in an effective
manner within a discourse. There are various opinions regarding the dispositio. For
example, according to Aristotle, only two parts of speech are important: the
statement (which introduces the main idea) and the proof or confirmatio (where the
idea is confirmed by the evidence). The statement is often called the narration
(narratio). Barthes (1970), for example, associates the two; Aristotle, however,
observes that it is only the forensic speech which requires a regular narrative
(διήγησις), a full and detailed statement of what has happened before. So, he prefers
to speak of statement, since the statement of the case is necessary (otherwise, what
is to be demonstrated?) while a discourse does not always include a narration of
facts.31 According to other authors, however, the statement is only a part (albeit the
central part) of the narration. The dispositio was concerned with not only the parts
of a speech, but also the order in which facts and arguments had to be arranged.32
31
«For narrative only belongs in a manner to forensic speech, but in epideictic or deliberative
speech how is it possible that there should be narrative as it is defined, or a refutation?» (Aristotle,
Rhetoric 1414, 3a–b), in the English translation by John Henry Freese, fellow of St. John’s
College, in the Loeb Classical Library (1947), book III, XIII.3, p. 425.
32
The classical tradition required facts to follow the natural order, that is their true chronological
order. Later, especially in the Middle Ages, however, this rule was often violated in order to create
The last phase is parsing. Expressions are arranged in the surface text and
inserted into grammatical dependencies. It should be remembered that there is a
certain asymmetry between the repertoire of conceptual relations and that of
grammatical dependencies.33 The process that produces the surface text expresses
certain preferences (for example, the sequential arrangement of grammatically
interdependent elements).
The Introduction warns scholars not to conceive of these phases as a linear
sequence because it is conceivable that they affect each other simultaneously in an
alternation of dominances. It is not to be excluded that there is a principle according
to which, during production, the materials demonstrate specific organizational
trends and impose them on the producer. In practice, the continuous production of
texts generates interpenetration between the individual phases of text production.
According to the two authors, in principle there is no limit that induces us to
consider the production of a text completed; instead, there is a threshold-term at
which point the producer considers the product satisfactory in relation to its pur-
pose. Similarly, the qualitative assessment of the recipient determines the amount of
processing potential that the recipient is willing to employ. But we cannot speak of
an absolute conclusion in the reception of a text either, and we must speak again of
a threshold-term at which point the recipient considers his understanding of the
processed material satisfactory.
Two other possibilities should not be excluded: first, another individual can
change the surface text, possibly even improving it; second, another individual can
parse a text even more deeply than the recipient to whom the text was addressed.34
The Tartu school of semiotics and the semiology of Roland Barthes argue that a
literary text is inherently polysemic. From the perspective of the procedural
approach, these prescriptions should be applicable to any type of text without ruling
out a priori their belonging to one literary genre or another.
If the final product becomes the documentation of the decisions made within
processes of selection and combination, the surface text becomes significant due to
specific narrative effects (analepsis, prolepsis, etc.). In these cases, we are dealing with an artificial
order. As for the arguments, however, classical tradition mentions three different methods:
(1) ascending order: from weak to strong arguments; the advantage is that the last arguments used
are more easily remembered; (2) descending order: from strong to weak arguments; the advantage
is the strength of the first argument’s impact; (3) Nestorian or Homeric order: the strongest
arguments are placed at the beginning and at the end (so named because in the Iliad Δ, Nestor
places less reliable troops at the centre of his formation).
33
It has been observed by Indo-Europeanists that modern European languages such as German,
English and Spanish have a far less extensive repertoire of grammatical dependencies than Finnish,
Hungarian or Caucasian languages, which are equipped with many grammatical cases capable of
indicating conceptual relations. Hjelmslev (1935) has already reflected on the correspondence
between semantic relations and grammatical cases, assigning a kind of supremacy to the
Finno-Ugric languages.
34
the existence of other versions that exist virtually and might have appeared if they
had been chosen. The different ways in which different subjects can receive the
same text must also be examined.
Communicative Functions
The procedural approach considers a text as a system made up of a vast set of

functions. Whereas a language is the virtual system of available choices that can be
made but which have not yet been selected, the text is a specific organization that
has already been realized: an actualized relationship between elements in which
certain possible selections have been made and implemented. «This utilization is
carried out via procedures of actualization» (de Beaugrande and Dressler 1981a, b,
p. 25).35 Consequently, the text is defined as a «cybernetic system which contin-
ually regulates the functions of its constituent occurrences» (de Beaugrande and
Dressler 1981a, b, p. 25).
During a communicative interaction, the participants possess knowledge systems
concerning the language and the content of their messages and intentions; when
these knowledge systems diverge, the stability of the textual system is disturbed. It
must therefore be re-composed thanks to an integration that regulates the divergent
case. If insoluble discrepancies persist, use of the text is blocked; but under standard
conditions participants in communication tend to safeguard the stability of the
system, maintaining a continuity between meaningful occurrences and relevant
context. The participants in a conversation are more likely to infer that there is
unspoken content than to abandon the idea that the speech is coherent and contains
information.
Although they have several options to identify relations between the elements of
the text, some are more satisfactory or more likely than others. To the extent that
knowledge is distributed by a community in a preferential order, the outcome of
text processing will be similar for almost all the community’s members. The
individuals in question will be induced to pay more and more attention to their
common preferences.
According to the Introduction, the awareness that preferential knowledge exists
certainly does not require the members of a community to comply with it; indeed,
the opposite occurs: a text that conforms fully to established knowledge will pro-
vide zero information. All that is known fully—or in cybernetic terms, is totally
stable—is devoid of interest for the human mind.
A language that transmits contents that are already known is easy to produce and
receive because the sender and the recipient are moving along a routine path. For
35
In this passage, the concept of ‘execution’ expressed by de Beaugrande would be an actual-
ization of competence during execution; there is thus a huge difference between this idea and the
analogous concept of ‘performance’ utilized by Chomsky.
this reason, however, the contents of this language are not incisive and end up being
boring. On the contrary, a creative language whose content is unusual and which is
expressed from unique perspectives has a significant influence and a bewitching
charm; it also proves enormously difficult to process. It offers the recipient a
challenge and the pleasure of interpretation.
A good speaker or a proficient writer gives the impression that he produces texts
in a totally relaxed manner; the suspicion does arise though that the limited time
required for the expression of these texts, written or verbal, may be counterbalanced
by a proportionately much longer period of preparation. In all likelihood, these
individuals invest a great deal of consciousness in order to obtain such results. This
may explain why an individual with experience in textual production can correct
other people’s texts without having participated in their thought processes. The
producer of a text may often find it more difficult to correct himself since he is
already familiar with the concepts in question and is thus unable to detect inefficient
expressions or, even worse, errors.
Regulative Principles
In the procedural approach, a text can be recognized due to the fact that it is realized
on the basis of three regulative principles—efficiency, effectiveness and appropri-
ateness—that determine certain pragmatic characteristics. If a text is to be regarded
as a tool, it is obvious that it can be used with greater or lesser skill. A cleverly
constructed text will prove to be functional, that is, it will allow the person who
produced it to reach his goals; these goals, however, can be reached with greater or
lesser degrees of efficiency, effectiveness, and appropriateness.36
A text is efficient if it is able to achieve in an economical manner the goals for
which it was realized, that is to say in a manner that requires the recipient to make
the least interpretative effort possible. An efficient text will consist of a well-known
code, will have features of linguistic explicitness, will be full of elements that
contribute to its linguistic and thematic unity, and will not introduce too much new
information.
An effective text is one that completes its task forcefully and energetically. For
example, an informative text may ensure its effectiveness through graphical devices
and structural or expressive elements and content (e.g. the use of special characters
or the use of a transgressive style) that make it especially easy to remember the
information it conveys. Normally, very effective texts tend to be inefficient; vice
versa, efficient texts are not very effective. This is not surprising because the
effectiveness and efficiency of a text are functions of its predictability. It is easy to
understand that documents that do not require inferences to be made or a large
amount of information to be accessed are relatively predictable for the recipient and
36
for this reason they are efficient; on the contrary, those which have the opposite
characteristics are effective. Appropriateness, finally, is the characteristic of texts
whose general approach is appropriate to the content that they aim to convey.
In the Introduction, the two researchers warn scholars that these elements have
not yet been thoroughly investigated in the scientific literature because the com-
plexity of the operations surpasses the descriptive ability of the models developed
heretofore. Despite the great quantity of operations possible, the number of types of
operation that deserve to be investigated is, according to them, relatively limited.
These operations direct the content of the communications in a way that can be
compared to the rules of grammar and logic; moreover, by reflecting on the average
level of abstract knowledge of grammar possessed by native speakers of their
mother tongue, it could be argued that in everyday texts, criteria of textuality count
more than grammatical rules.
Before directly analysing a specific case in the next chapter, we must illustrate
the criteria by which textuality is identified in the model of Dressler and de
Beaugrande. In fact, most of their Introduction is devoted to the discussion of the
criteria that fully determine the properties that, in a scientifically fixed definition,
the ‘text’ object must possess.
The Concept of Textuality in the Procedural Approach:

Seven Criteria
Cohesion
The first criterion is cohesion.37 The authors use this term (from the Latin verb
cohaerere = to be joined, attached) to indicate the function of syntax, which is to
impose an organizational model on the surface text in the context of the commu-
nication. In their view, syntax manifests a lesser number of classes and structures
than that of conceptual relationships; this is seen by the observation that individuals
retain the surface structures of a text in “working memory”, but store the conceptual
content in long-term memory.38
The two authors find full correspondence between syntactic functions and
cognitive factors. The principal units of syntax are highly distinct models of
dependencies: the phrase, the clause, the sentence. These tools contribute to sta-
bility. Recurrence is defined as the repetition of these elements and patterns.
Cohesion is more evident within a phrase, a clause, or a sentence than in two or
more such units. However, they believe that it is not easy to explain how these units
become structured when used in language.
37
See de Beaugrande and Dressler 1981a, b, pp. 32–53.
38
For support they indicate Wright (1968).
The Concept of Textuality in the Procedural Approach: Seven Criteria 57
Cohesion is achieved through various means: in units such as phrases, clauses,

and statements, cohesion is ensured by the inclusion of various elements in the
grammatical dependencies; in longer texts, however, the main operation is the
enumeration of elements and models already in use, which can be reused, modified,
or combined together.
The first mechanism is that of repetition.39 The direct repetition of elements is
called recurrence, because the original item appears again—it reoccurs. Recurrence
takes place at different levels, but of particular interest is lexical recurrence, that is,
the repetition of words or expressions. Recurrence is mainly used to reinforce one’s
opinion, but it can also be used to deny. Partial recurrence entails using the same
basic word components but shifting them to a different word class. Recurrence has
the disadvantage of reducing the level of information transmitted. For this reason
certain techniques such as parallelism are used, which involve the repetition of
structures. This also occurs in the case of paraphrase or synonymy. Repetition also
concerns the use of pro-forms, which do not always co-refer to elements of the
same grammatical class (e.g. a pronoun with a noun). These correspondences are
preferences that re-use grammatical frameworks that have already been parsed. The
primary motivation for using pro-forms, in general, is efficiency. There is a point
where a trade-off is established between compactness and clarity. Pro-forms save
processing effort because they are shorter than the expressions they replace.
However, if those expressions are difficult to find, the savings are lost again on
search operations.
Another cohesive device is ellipsis.40 It is defined as a perceptible discontinuity
in the surface text that is detected when the text is processed. In general, ellipses are
based on the partial commonality of certain structural components. The typical case
is that of the anaphoric ellipsis. The ellipsis of the subject or other unnecessary
elements indicates how complex the interaction between cognition and syntactic
conventions is. The procedural approach attempts to clarify why ellipses are more
present in certain conditions. If the function of syntax is to produce a surface
structure that determines which hypotheses can be formulated in regard to the
organization of the underlying concepts and relationships of a text, then an
underdeveloped syntax would constitute a substantial processing strain. Both the
use of pro-forms and ellipses are based on a trade-off between compactness and
clarity. The use of texts without ellipses is more time-consuming, but it can be
noted that very large ellipses require very intense research.
Another expedient of cohesion is junction, which makes it easier to identify
relationships among events and situations. The authors believe that there are at least
four basic types of junction: conjunction, disjunction, contrajunction, and subor-
dination.41 Junction is very intricate. The use of junction is required infrequently
given that speakers can recognize certain relationships by referring to their general
39
Repetition had already been studied by Weinrich (1972) and van Dijk (1972).
40
It had been analysed by Dressler (1970) and Halliday and Hasan (1976).
41
In particular, Lakoff (1971), Halliday and Hasan (1976), van Dijk (1977).
knowledge. Through the use of junctions, the producer of a text influences the way
the recipient rebuilds the relations. In this perspective, junctions underline the fact
that the participants in a communication are involved in an interactive communi-
cation. The various types of junction may appear simple, but they are only
apparently simple; in fact, it is significant that in impromptu speech they are
omitted.
Thanks to the computer and via the so-called transition network it is possible to
construct a representation using the concept of nodes linked by connecting bran-
ches. In a transition network, structures of phrases and clauses are used as means to
construct (and then verify) hypotheses regarding the reliability of various elements.
These networks contain the users’ expectations and strategies, and express gram-
matical rules in the form of procedures for the use of the rules.
These operations can also be examined in another perspective. The processor of
the text could put each occurrence on a waiting list of partial results until the
macro-status was completed, and then could sort the various results into a gram-
matical dependency network. In this way a network within each single phrase
would be constructed.
Repeating this procedure for all the phrases, the sentence is not parsed as a linear
sequence, but rather as a labelled transition network: the nodes represent the
grammatical states and the branches represent grammatical dependencies. The role
of such a network would be to organize the surface structure according to the most
direct access, so that the linear text could be read off it during production, or traced
back to it during reception.
Coherence
The authors’ second criterion is coherence.42 They mean the ability that a linguistic
expression has to transmit knowledge (any possible meaning, thus a virtual ability).
The two linguists define actual meaning or sense as the knowledge actually
transmitted in a text. A text makes sense because there is a certain continuity in the
knowledge activated by the expressions used. This sense of continuity is the
foundation of coherence and it represents mutual accessibility within a combination
of concepts and relationships.
The two scholars are surprised by the fact that in traditional debates the meaning
of individual expressions or the content of isolated concepts is the centre of so
much controversy. They explain that in the procedural approach, virtual meaning
(= meaning) and the actual meaning (= sense) are procedures used to apply one’s
knowledge to a wide range of tasks and activities deriving from texts. They also
insist on the consideration that in the textual world, both expressions and contents
appear fairly stable and delimited.
42
See de Beaugrande and Dressler 1981a, b, pp. 53–72.
The authors thus propose that instead of deconstructing meanings and concepts,
we should examine the attribution of conceptual meaning to linguistic expressions.
Indeed, the construction of textual worlds is well-documented in human commu-
nication. The two authors empirically attempt to progressively decrease the insta-
bility of concepts, providing as much precise information about communicative
situations as possible. In this framework, they define the content of a concept as an
ordered set of hypotheses regarding the access of cognitive elements within a
current pattern.
According to the procedural approach, meaning—as a property of language—is
only a special case of acquisition of knowledge. During the processing of texts,
surface expressions are taken as cues to activate the combination of concepts and
relations. The knowledge on which the use of texts is based should be formalized in
global models that are to be re-utilized according to each specific representation.
These global models may appear in different forms depending on the processing
needs of the moment.43 The recipients then verify their hypotheses about the main
subject and the organization of the textual world.
They believe that some types of global patterns are probably stored as integral
“chunks” because they are used so frequently. “Frames” are global patterns that
contain common-sense knowledge about certain basic concepts. These frames
indicate, in general terms, what the connections are but not the order in which the
related concepts must be expressed. In addition to frames, “schemas” need to be
identified, that is, global patterns of events and states in ordered sequences. “Plans”
are global patterns of events and states that refer to an intended goal. Finally,
“scripts” are stabilized plans called up to define the roles of participants and their
expected actions.
According to Dressler and de Beaugrande, during text production and reception
all the global models regarding the development of the arguments, the sequence of
events, and the characters in the various situations ineluctably become important.
Coherence is the result of the concepts and relationships that are unified in a
network, whose core is made up of main topics and knowledge spaces. The con-
cepts function as steps for building a continuity of sense and the extent of pro-
cessing will vary according to the task. The points that are strategically important
for processing are called control centres, which correspond to the primary concepts
and sub-concepts.
According to their model, the textual world is organized through a comparison
of beliefs about the real world and a person’s own knowledge. In this way, lin-
guistic expressions activate knowledge and text users define a textual world that
does not appear correspondent to standard knowledge as “fictional”.
In addition to the connections, the authors define the state of the links that are
made by different operators, which include:
43
Global models are treated in a plethora of studies; among others, see Petőfi (1976), Schank and
Abelson (1977), van Dijk and Kintsch (1978) and Allen (1979).
determinateness operator (δ)

typicalness operator (τ)
proximity operator (π)
projection operator (ρ).
The processes of activation and realization of inferences are mechanisms that
expand, develop, and integrate the content of a text. Inferences are established to
deal with discontinuities. Activation is expanded when certain points of a stored
model are stimulated.
Text linguistics (see chap. 1, p. 24) was concerned with cohesion and coherence
even before the advent of the procedural approach. In fact, the cohesion of a surface
text and the coherence of textual worlds, even though they are the most obvious
criteria of textuality because they contribute to the production of the meaning of the
text, are not in themselves sufficient to discriminate between a text and a non-text.
This stems from the observation that in human communication, even structures that
are not entirely cohesive and coherent can function as texts. Therefore, the
Introduction indicates five other criteria of textuality.
Intentionality and Acceptability
Intentionality and acceptability, rather than two separate policies, seem more like a
separated pair of the same property, part of which is oriented to the issuer of a text
and another to the recipient. As a criterion of textuality, intentionality is the
counterpart of acceptability. The first regards the issuer and demonstrates that his
product is intended to be cohesive and coherent; the second, however, regards the
receiver in that he must accept the text produced by the issuer as a cohesive and
coherent construct.44
In order to participate in a communicative interaction, a linguistic structure must
be validated, that is recognized and accepted by issuer and receiver. Therefore, the
attitude of text users must also be taken into account by the criteria of textuality.
Attitudes relate to the tolerance towards any disturbances in cohesion or coherence
that do not undermine the purpose of the communication. The attitudes of those
who produce a text are indicated by the authors as intention, while the attitudes of
those who receive the text are indicated, correspondingly, as acceptability.
Intentionality may be considered in both a narrow and in a wider sense. Strictly
speaking, the issuer always considers his text product as cohesive and coherent;
however, limitations of time and processing resources do not always allow the text
to fully realize this intention in the presentation phase. The mutual dependence
between cohesion and coherence and communicative intention sometimes produces
44
On intentionality and acceptability see de Beaugrande and Dressler (1981a, b), pp. 72–86.
somewhat complex situations. For example, the issuer may lapse into faults in
coherence because he wishes to conceal some of his knowledge or actions (we
could think of a police interrogation in which inconsistencies are used to evaluate
the strength of an alibi).
Beyond this strict sense, intentionality also has a broader sense. It refers to the
means the issuer uses to pursue and achieve his goals. From this point of view, the
authors point out that studies relating to intentions have been conducted in several
disciplines (sociology, psychology, philosophy and artificial intelligence research)
and that linguistics has been particularly influenced by philosophy.45
The speech acts theory of Searle, developed following Austin’s footsteps within
the phenomenological current that we mentioned in Chap. 1, p. 30 made significant
contributions to linguistic pragmatics. Among the most successful results of phe-
nomenological studies are the so-called conversational maxims of Grice.46 In the
tradition of behaviourism, human language has been investigated as a reactive
response to an external stimulus in the environment.
Both phenomenology and behaviourism have overlooked the human ability to
envision alternative future states and to work toward a particularly desired one. This
faculty, typically human, is to make plans. Of course men and women are not
all-knowing or all-powerful and are influenced by the environment, but people also
have a threshold of plan activation. This is the degree of awareness of the steps
required to start developing a plan. When one of the steps looks uncertain, the
planner has a problem. Hence, planning is an elaborated, comprehensive type of
problem-solving applied to advancing the planner’s own state toward a goal in an
evolving situation. The action of producing texts is driven by a specific plan
whenever the issuer intends to steer the situation to his own end.
In the procedural approach, communication is a case of active planning and
participants submit texts as acts of speech.47 For example, a plan might require
inducing beliefs in participants if they are useful to achieving the goal. Such a
project can be problematic if the belief collides with empirical evidence.
The correspondences between intentionality and acceptability are extremely
complex. Under stress or time pressure, people often produce utterances which they
might feel disinclined to accept under normal circumstances; conversely, they
accept utterances from other people which they would be reluctant to produce.
Some studies try to show that people may not be aware of their own speaking
styles, or those of their social group.48 And people may shift between styles of text
production in order to obtain desired social roles in different social contents, taking
45
See also Schank and Abelson (1977), Cohen (1978), Allen (1979).
46
The maxims listed by Grice (1975) concern cooperation, quantity, quality, relevance and
manner, and make it possible to monitor the logical implicatures in conversations.
47
See Bruce and Newman (1978).
48
See Blom and Gumperz (1972).
advantage of the different qualities of certain options. In view of all these consid-
erations, the conclusion that language can scarcely be described or explained except
in terms of texts in real settings again seems inescapable.
Like intentionality, acceptability also has a meaning in a narrow and in a broad
sense. In a strict sense, acceptability presupposes a certain degree of tolerance for
discontinuities or discrepancies, provided that the overall context can be perceived.
Thanks to the progress made in research aimed at demonstrating that the set of all
allowed utterances is a grammar, more and more scholars subscribe to the opinion
that acceptability must be considered what is actually accepted in a human com-
munication, while grammaticality is what is determined by abstract criterion.49
The two scholars, however, are uncertain about the correlation between the two
concepts and assume that the crucial difference lies between the virtual system and
the process of actualization. In fact, when there is adequate justification, actual-
ization is necessary, even if it means bypassing the organization of the virtual
system: this principle is the fundamental distinction between the subject of study in
the Exact Sciences and in the Humanities.
According to scholars, the rigid application to a text of concepts derived from the
exact sciences seems to remove the subject of investigation from its proper domain
—that of an imperfect human artefact; a formal configuration ends up becoming a
diagram, in the semiotic sense, but not an explanation. From this point of view, they
find that logical and mathematical formalisms are inadequate tools and we can
therefore understand the criticism that the authors make of the framework of Teun
Adrianus van Dijk and of János Sándor Petőfi.
A certain prospect of validly linking the abstract and the concrete level might be
offered by probabilistic considerations.50 Grammar would thus consist of a series of
vague instructions, among which is syntactic well-formedness. Communicative
situations would however bring the communicating parties to recognize gram-
maticality. The concept would thus define a factor which, through interaction with
others, constitutes acceptability. If the terms of the statements manage to evoke
mental images, we are more likely to accept the statements themselves.
Beyond this strict sense, acceptability also has a broader sense. Acceptance is an
action in its own right and entails entering into discourse interaction, with all
attendant consequences. Refusing acceptance is conventionally accomplished by
explicit signals. Participation in discourse would, as a default, be assumed to imply
acceptance. If acceptance is denied, the textuality is compromised. It is also pos-
sible to block acceptance voluntarily by not maintaining coherence and thereby
discouraging the interlocutor.
49
The gradual progress has been highlighted by the studies of McCawley (1972) and Lakoff
(1973).
50
The observations in Greenbaum (1973) are made in this direction.
Informativity
The fifth criterion identified is informativity.51 To illustrate this principle, the two
authors begin with the now classical information theory of Claude Shannon and
Warren Weaver.52 It is based on statistical probability and argues that the value of
information increases in proportion to the number of possible alternatives. In the
view of the authors, it is fairly well agreed that this model of statistical probability is
not applicable as such to natural language communication. To count up all the
sequences of a language like English is out of the question. Even if it were not, the
occurrence of most elements depends on factors other than the occurrence of the
preceding element.53 Nevertheless, according to de Beaugrande and Dressler the
notion of probability cannot be eliminated from a textual theory and a model for the
use of texts. Statistical probability should be replaced with contextual probability.
The procedural approach uses transition networks to represent cohesion and
coherence, and these networks are based on contextual probability. In fact, what
really matters are the classes of occurrences that are more or less likely to be linked
to systematic combinations. The degree of probability would vary in the different
systems. A sequence might be syntactically probable, but conceptually improbable.
Contextual probability is a complex amalgam of factors because there is a pro-
gression of steadily more specialized expectations applying in various degrees
during communication.54
The first source of the receiver’s expectations comes from the real world. It is
constituted, from the point of view of the receiver, by the social model that dom-
inates the situation in which he lives and by its contours. In this world, certain
propositions are held to be true; these are facts which a person or group considers to
be generally applicable to some “real” or recoverable situation or event. As a whole,
they form the system of beliefs. Some facts are so firmly entrenched in our manner
of thinking that they act as defaults. Should any such facts be violated in a textual
world, there must be explicit, unmistakable signals. Humans seem to apply con-
sistent strategies of apperceiving and arranging the real world, and integrate their
sensations into a model of the world via a highly skilled act of attention.55
Whatever knowledge is acquired is continually used as a bridge to annex further
knowledge. For instance, frames, schemas, plans, and scripts are used for matching,
integrating, and controlling large amounts of current material.
The second source of expectations derives from the definitiveness of the lan-
guage used in the text. In the principal modern European languages, many com-
binations derive from arbitrary conventions.
51
52
The theory is presented in Shannon and Weaver (1949).
53
This is maintained by Sprung (1964).
54
The whole discussion of these issues is based on Miller and Johnson-Laird (1976).
55
See Miller and Johnson-Laird (1976).
A third source of expectations is the technical arrangement of the elements with

respect to their informativity. For example, highly informative elements tend to
appear toward the end of a clause and to receive a high key. It might be concluded
that the first source of expectations about the real world and its facts would be
independent of language altogether, whereas the second source (formal conven-
tions) and the third (informativity signalling) would vary from one language to
another. For the authors of the Introduction this issue is, however, in great dispute.
The diversity of formal conventions among languages is uncontested; but there is
little agreement about whether this diversity also impels the users of language to
organize the world in different ways.
The fourth source of expectations is text type. Text types are global frameworks
controlling the range of options likely to be utilized and affect the use of phonetic,
syntactic and informative modules.
The fifth and final source of expectations is the immediate context where the text
occurs and is utilized. If, as we claim, actualization can override the conventional
organization of virtual systems, this source might modify the expectations drawn
from the other four sources.
Some receivers can expect some sorts of occurrences to be more dominant and
frequent than others; this corresponds to the notion of style. Informativity can be
increased on occasion by breaking out of one’s own established style. Style is of
particular importance in literary or poetic texts, with the result that producers must
expend considerable care and attention upon selection procedures and the receiver
expects a higher than normal focus to be placed on it; in fact, the choices and
production options vary more in poetic texts than in the standard organization of the
language. As a yardstick to evaluate the novelty of the contents of a text, informa-
tivity plays a key role in choosing and arranging the various options within a text.
The Introduction proposes a scale consisting of three orders of informativity:
first-order or upper degree, second-order or lower degree and third-order or
apparently outside the set altogether. The first order corresponds to a rather trivial
level informativity in that it contains elements that are well integrated into the
system. The first order could be considered the standard case, that is, where
selections are obvious in the absence of contrary indicators. When defaults or
preferences are overridden, i.e., when occurrences are below the upper range of
probability, we obtain second-order informativity. To get to the second degree, that
of greater informativity, a mechanism of upgrading-downgrading can be used.
Third-order informativity, composed of occurrences that are outside the set of more
or less likely options, is comparatively infrequent; on the other hand, it is more
interesting, because the text receiver must do a motivation search—a special case of
problem-solving to find out what these occurrences signify and how they can be
integrated back into the continuity of the context. A successful search will show
that the occurrence in question was within the range of options after all, though
accessible only via some mediation. Accordingly, the third-order occurrence will be
downgraded to the second order and will thus become new knowledge. The same
mechanism also applies to upgrading and it should be noted that it may be appli-
cable to human relationships in general.
The presence of at least some second-order occurrences would be the normal

standard for textual communication, since texts purely on the first order would be
difficult to construct and extremely uninteresting. Through upgrading and down-
grading, first- and third-order occurrences are transformed into second-order
occurrences within a system constantly seeking to maintain the continuity of sense.
In addition, a producer can design a text so that interest is kept at a high level
through expectations and thus satisfy his intention. Informativity is of primary
importance in providing a logical explanation of why certain options are used in
any given context.
Situationality
The sixth criterion of textuality that the scholars have identified refers to the
communicative situation and is called situationality.56 It refers to the factors which
make a text relevant in a communicative situation. Usually, the effects of a situation
are perceived through a form of mediation and are proportionate to how much a
participant’s own beliefs and goals are used in the communicative interaction.
There are significant correlations between texts, speech actions and commu-
nicative situations which do not constitute simple reactions to the perceptible evi-
dence within a communicative situation. The standard case, however, is that in
which the content of a text is distant from the evidence of empirical reality, due to
the mediation based on the opinion, the beliefs and the purpose of the issuer. The
acceptability of a text is more dependent on the credibility and the relevance the
participants in the communication assign to attitude than the correctness of the
references to reality. Speech actions can be considered as realizations of general
strategies aimed at monitoring and managing the different types of situation in
which people interact.
When, during discourse, texts are used that can steer communicative interaction
in the direction of the aims of the participants, we are dealing with situation
managing. The use of monitoring is typical of situations that do not reflect the
expectations of the issuer, who, for his part, aims to overcome the discontinuity or
to strengthen his expectations. On the other hand, managing calls into question
higher purposes that require considerable mediation. The boundary between situ-
ation monitoring and managing is not clearly marked and varies according to the
assessments of the individual participants. It seems that speakers are somewhat
complacent when they disguise managing as forms of monitoring in order to give
the impression that the situation is evolving spontaneously towards the desired goal.
From the theoretical point of view, the authors envisage a distinction between
monitoring and managing in terms of dominances. A variant of monitoring could be
56
the case in which texts are reactions to external inputs. To describe a material object
in the form of the text is a case where normal classification strategies are used. At
times these are disturbed by the presence of objects or events that the speaker
considers highly unlikely. A fundamental case of improbability occurs because of
an imbalance in frequency. If someone repeats something more frequently than
normal, monitoring will be called into play; this can also occur in attempts to
explain and thereby downgrade a surprisingly high frequency of repetitions.
Monitoring can be a sign of lack of continuity and require downgrading when the
actions do not appear sufficiently motivated. The monitoring of a communicative
situation within the terms described is similar to problem-solving. The issuer of a
text takes an object or an event and treats it as the main topic of his text.
There are two possible outcomes: either the receiver realizes this and does not
express any reaction, or else he finds a way to downgrade it without making it seem
a betrayal of expectations. By integrating the event, the problem becomes more
apparent but arouses the willingness of participants to reaffirm their own criteria and
to seek confirmation in the criteria of others. When they seem to be unsuccessful,
expectations are reinforced in a contingent situation.57 However, departures from
the evidence of the situation are allowed in some text types, especially in dramatic
texts.58 As a subclass of literary texts, they have a different organization from
standard objects and events and require the user to mediate intensively.
Situation monitoring may encourage the use of pro-forms instead of abstract and
conceptual definitions to describe objects and events. For example, when
57
Edmondson (1980) considers such a process similar to negotiations. When several participants
have different views on what is happening or is going in front of them, it is likely that the situation
will be subjected to monitoring.
58
The following is an exemplificative case in Plato: «SOCRATES: Of course! So that, in
Epicharmus’s phrase, “what two men spake erewhile” I may prove I can manage single-handed.
And indeed it looks as though it must of sheer necessity be so. Still, if we are to do this, for my part
I think we ought all to vie with each other in attempting a knowledge of what is true and what
false, in the matter of our argument; for it is a benefit to all alike that it be revealed. Now I am
going to pursue the argument as my view of it may suggest; but if any of you think the admission I
am making to myself are not the truth, you must seize upon them and refute me. For I assure you I
myself do not say what I say as knowing it, but as joining in the search with you; so that if anyone
who disputes my statements is found to be on the right track, I shall be the first to agree with him.
This, however, I say on the assumption that you think the argument should be carried through to a
conclusion; but if you would rather it were not, let us have done with it now and go our ways.
GORGIAS: Well, my opinion is, Socrates, that we ought not to go away yet, but that you should
go through with the argument; and I fancy the rest of them think the same. For I myself, in fact,
desire to hear you going through the remainder by yourself. SOCRATES: Why, to be sure,
Gorgias, I myself should have liked to continue discussing with Callicles here until I had paid him
an Amphion’s speech in return for his of Zetus. But since you, Callicles, are unwilling to join me
in finishing off the argument, you must at any rate pull me up, as you listen, if it seems to you that
my statements are wrong. And if you refute me, I shall not be vexed with you as you were with
me; you will only be recorded in my mind as my greatest benefactor. CALLICLES: Proceed,
good sir, by yourself, and finish it off» [Plato, Gorgias 505 e–506 c, with the English translation by
Walter R.M. Lamb, M.A. sometime fellow of Trinity College, Cambridge, in the Loeb Classical
Library (1946), volume 5, pp. 462–465].
monitoring occurs that is completely divergent from the situation or the events
themselves, managing will entail. Situation managing can be profitably explored in
terms of plan theory. Stabilized plans are developed only for situations whose
managing is routinely demanded in a given society. In other situations, participants
must adapt to a range of variable factors and protect their goals as best they can.
They can scan texts from other participants to recognize the latter’s goals; or they
can simply postulate default goals by assuming that most other people will have the
same desires as they do themselves. If resources are too limited for fulfilling every
participant’s goals, conflict can be expected to result. Conflicting goals lead to
conflicts in how the same event or situation is monitored.
Monitoring must incorporate a set of methods to obtain the approval of others and
favour their cooperation. It is called goal negotiation.59 This negotiation occurs
because many goals are not obtainable through the actions of one agent. You can
simply ask someone to do or say something, or invoke a speech act or a piece of
information about something; you might inform the person of reason why they
should be co-operative or invoke that reason; you could bargain to do them a favour
in return, or you could bargain to give them some object they would desire. If all these
discourse actions fail, you could threaten people, overpower them, or steal what you
want. When a planner moves down this list toward steadily more extreme actions, we
can use the term planbox escalation. Planbox escalation entails a trade-off.
The individual creates a plan in which there is a balance between efficiency,
effectiveness and appropriateness. Asking, invoking, and informing are easy and
demand no expenditure except of the processing resources needed to produce the
text. Bargaining commits you to an expenditure of material resources, but it provides
a greater incentive in many cases (the authors point out that close friends might be
offended by the suggestion that you won’t help out without reward). Threatening,
overpowering, and stealing commit you to an expenditure of physical resources, but
they suppress further negotiation; their real disadvantage is that they render the goal
unstable, because people will often try to avenge themselves or recover their prop-
erty. Most societies have institutional measures for discouraging the extreme plan-
boxes of overpowering and stealing. Threatening is easier to carry out and conceal,
but also highly problematic. If threatened people don’t believe in your ability to carry
out the threat, it matters little whether you can or not: your goal will not be reached.
Planbox escalation is therefore a normal response to continued failure.
Intertextuality
According to the authors of the Introduction, the last criterion of textuality is

intertextuality.60 It means that there are interdependencies between the production
59
Schank and Abelson (1977) take into account the goal negotiation.
60
See Chap. IX (de Beaugrande and Dressler 1981a, b, pp. 110–127).
and reception of a given text, depending upon the participants’ knowledge of other
texts. Through mediation this knowledge is applied proportionally to the expanse of
time and to the processing activity between the use of current text and the use of
prior texts. Generally, mediation occurs at different degrees: a deeper mediation is
found when people develop and use text types, namely classes of texts expected to
have certain characteristics for specific purposes. On the contrary, a smaller
mediation is found when people refer to well-known texts, quoting or alluding to
other texts, as it happens for literary works or famous speeches. A standard case of
mediation is found in ordinary conversation, when people report, summarize, or just
reply to other texts. Here, mediation is extremely slight.
Linguistic typology deals with systemization and classification of different text
types. In the past, linguistic studies have compared typologically utterances coming
from different modern European languages. This typology applies to virtual sys-
tems, i.e. the potential structures of languages. A text typology has to deal with
actual systems in which selections and decisions are to be made. Many actualised
instances do not illustrate the exact characteristics of an ideal type, neither
according to their completeness, nor according to their accuracy. Expectations and
demands related to a text type can be modified by the requirements of the com-
municative situation when text exchanging can occur; if this is the case, between
actual occurrences and ideal linguistic types there are immanent discrepancies. In
order to establish a textual typology, linguistic typology faces another challenge: to
take into account text types per se.
According to Dressler and de Beaugrande, a text typology has to be correlated
with typologies of discourse actions and situations. Unless the appropriateness of a
text type to its setting of occurrence is judged, the participants cannot determine the
means and extent of upholding the criteria of textuality. For example, the demands
for cohesion and coherence are less strict in conversation, while they are elaborately
upheld in scientific texts. In poetic texts, cohesion can be organized according to
unconventional principles. Some text types can be defined considering their con-
tributions to human interactions, thus might also recognize some dominances.
From their point of view, the authors affirm that knowledge spaces may be
enriched. There are different kinds of enrichment: descriptive texts, whose control
centres could be objects or situations, would enrich knowledge spaces proposing a
frequency of conceptual relations for attributes, states, instances and specifications.
Narrative texts, whose control centres could be actions and events, would enrich
knowledge spaces proposing a frequency of conceptual relations such as cause,
reason, facilitation and time proximity. Argumentative texts, whose control centres
could be the acceptance of certain beliefs (as true vs. false), or the evaluation of
certain ideas (as positive vs. negative), would enrich knowledge spaces proposing a
frequency of conceptual relations such as reason, significance, volition, value and
opposition. Although the descriptive, narrative and argumentative functions are
theoretically separated, in actual texts they may be found at the same time; there-
fore, many texts often develop into a mixture of these functions.
Without doubt the argumentative function is dominant. It is not only the surface
text to cause the assignment of a text to a text type, but its function in human
communication. Literary texts also contain these functions. ‘Literariness’ is a

characteristic recognized in a text where “text-world” stands in a principled rela-
tionship of alternativity compared to the shared version of the “real world”. The
socially accepted model of the “real world” often contains some discrepancies with
regard to text-worlds of literary texts. Through the comprehension of such dis-
crepancies in literary texts, we sharpen our awareness of the organization of the
“real world”, which is not an objective datum. On the contrary, it is the result of the
growing of social cognition, interaction and negotiation. What motivates insights
into such an organization is alternativity. When ‘literariness’ is recognized as a
property of the whole class of literary texts, poetic texts become a definite subclass.
Taking into account the alternativity as a criterion, poetic texts show the highest
level of alternativity; on the contrary, scientific texts alternativity show a standard
level; moreover, didactic texts show the lowest level. In fact, according to the poetic
function, the first class of text is intended to motivate insights into the organization
of expression as negotiable; the second class serves to clarify knowledge referring
to a special domain of facts; the third class requires the presentation of solid
background knowledge, in order to disseminate widespread knowledge.
The question of text types goes beyond conventional linguistic methods and
melts into the conditions of text usages in communication. A text type is a set of
strategies in order to produce, predict and process textual occurrences, and so it acts
as a main determiner of efficiency, effectiveness and appropriateness. However, it
can hardly take into account the task to provide absolute boundaries between
members and non-members of a class.
Another issue concerning intertextuality is that of text allusion, i.e. the ways
people refer to previously known texts. Although a text producer can refer to any
previous text from a theoretic point of view, in practice he uses only well-known
texts, in order to facilitate accessibility to the receiver. Between the time of the
production of the first text (the original one) and the time of the production of the
following text (the genesis of the actual text) there might be a considerable distance.
It is in conversation that occurs the least mediation concerning intertextuality.
During a conversation, intentionality and situationality regulate its organization,
although both of them cannot offer a full account. In the same conversation, a text
cannot just be relevant to producers’ intentions and to the circumstance, but it
would be relevant to other previously used texts and in order to generate the
subsequent ones. «Topics must be selected, developed and shifted» (de Beaugrande
and Dressler 1981a, b, p. 114). An ordinary conversation may have one or different
topics; generally, a single topic is prevalent only in a limited period of time. In order
to individuate a topic of a conversation, you should investigate the density of
concepts and relations within several textual worlds presented in the texts, that
compound the whole conversation and are connected to each other. A single text
has topics potentially susceptible to further developments. To decide what to say
about a topic, participants in a conversation take in account the informativity of
potential contributions. Usually the most problematic and variable aspects of a topic
to be developed are those not yet established, because they are still subject to
changes.
During a conversation, either participants may violate some social conventions

or shared principles, or their intentions and beliefs may appear to be discrepant or
unmotivated. If this is the case, a participant may resort to monitoring, which can be
addressed to the style of the presentation and indirectly to the ability of the issuer;
moreover, a participant’s state may be monitored as inadequate for the discourse.
The state of a participant may also become the object of monitoring, based on the
evidence of his/her texts. During a conversation, the emergence of a problem will
cause participants to monitor the behaviour, the manner of speaking, the state, the
intention and the knowledge possessed by other participants. If a participant in a
conversation makes some unpredicted object or event into a text topic, he integrates
it into the “commonly shared real world”. The unexpected elements arise in the
external setting of the discourse, when situation monitoring occurs. A text receiver
can request to know the motivation behind the unexpected occurrence; as a con-
sequence, both the producer and the receiver can interact in solving the divergence.
When the conversation requires negotiating standards for belief and behaviour,
other problems arise: the conversation coherence is not univocally recognized in its
component texts. Typical human activities and viewpoints have to be distinguished
from specific events and situations that participants want to talk about.
In such a way, the notion of ‘text-world model’ might well be expanded to that
of a ‘discourse-world model’. The so called ‘discourse-world model’ would be an
integrated configuration of concepts and relations lying beneath all the texts that
compound a discourse. In order to minimize conflicts among the different
‘discourse-world models’, the situation monitoring would help for possible
disagreements.61
It is a matter of fact that people’s implicit knowledge is difficult to observe and
study, because it only emerges when it leads to some disturbance or discrepancy.
Obviously, even though it is necessary for making sense, a vast amount of infor-
mation goes unnoticed. A substantial amount of implicit knowledge could be
brought to light through experiments; when people are required to recall the content
of a text, there is a systematic pattern of additions, omissions and changes, which
may follow rules that could be systematized because of their analogy with cognitive
processes.
The third and final issue of intertextuality is extremely important and concerns
the making of reports and summaries of texts that have been read. The reception
process of a text is guided by the receiver’s placement of that text within a schema.
A balance is always set between prior knowledge and knowledge presented by the
text. All conflicts between a schema and a presentation should always be resolved
in favour of the schema; and schema-related materials should always be recalled
better than others. Dressler and de Beaugrande underline that these are trends and
are certainly not infallible.
There is evidently a gradation in which trace abstraction, construction and
reconstruction all participate to some extent, and are the results of processes of
61
This proposal is also supported by Reichman (1978), Rubin (1978) and Webber (1978).
inferencing and spreading activation. These constructive or reconstructive processes

act upon the content of the text, and expand, update, develop or complement it. The
first process occurs in the case of specific discontinuities, while the second is
deployed for activating one point in a knowledge pattern. The type of linkage
followed, evidently corresponds to the topic material and the frames that apply.
Considering text as a cybernetic system in which processing is devoted to a
maintenance of continuity of sense, relation and access among elements within a
level or on different levels are the centres around which the criteria of textuality are
fixed. In understanding and recalling the content of a text, a receiver assumes as
main priority the maintenance of a continuous pattern. In order to accomplish this
task, the human mind draws a paradigm. If the actual traces of the presentation
appear to be discontinuous, the receiver always uses previously stored knowledge.
Under normal conditions, a receiver has no clear motivation to create an exact
trace of the text content of a precise text. Firstly, he tries to maintain the continuity
of sense; when it is not possible, if typical world knowledge matches quite well the
content of text-presented knowledge, he can prove a partial trace of the text content
(i.e. trace abstraction).
It should be determined whether the interaction between stored world-
knowledge and text-presented knowledge follows systematic tendencies. On the
one hand, it is believed that strategies and trends must be similar in order to ensure
reliable communication between different individuals. On the other, it is almost
impossible to make predictions because of the amplitude of variations; there are,
indeed, results of various experimental tests demonstrating that receivers perceive
interference between text-motivated and text-unmotivated inferences.62
Afore mentioned systematic tendencies could concern text-presented knowledge
and patterns of previously stored knowledge: (1) text-presented knowledge is
privileged in understanding and recall if it matches patterns of stored knowledge;
(2) text-presented knowledge is privileged if it is attachable to the main entries of an
applied global pattern, such as a frame, schema, plan or script; (3) text-presented
knowledge is altered to produce a better match with patterns of stored knowledge;
(4) distinct elements of text-presented knowledge become conflated or confused
with each other if they are closely associated in stored knowledge;
(5) text-presented knowledge decays and becomes unrecoverable if it is designated
accidental or variable in world knowledge; (6) additions, modifications and changes
performed via spreading activation or inferencing become indistinguishable from
text-presented knowledge.63
In consideration of the above, one has to wonder how two people can understand
each other, since the understanding of each depends on the knowledge already
possessed by the other.
62
The main studies in support of this are Johnson (1977), Meyer (1977) and Rumelhart (1977).
63
These systematic tendencies are fully explained in de Beaugrande and Dressler (1981a, b),
pp. 123–124.
References
Classics
Aristotle. Rhetorica. In Aristotelis opera. Edidit August Immanuel Bekker, Academia Regia
Borussica. Berlin: Reimer, 1831–1870 [rist. De Gruyter, 1960]. In the English translation by
John Henry Freese, Fellow of St. John’s College, in the Loeb Classical Library (1947).
Plato. Gorgias. In Platonis opera. Edidit Iannes Burnet, Oxford Classical Texts Library. Oxford:
Oxford University Press, 1903. In the English translation by Walter R. M. Lamb, fellow of
Trinity College (Cambridge), in the Loeb Classical Library (1946).
Modern Studies
Allen, James Frederick. 1979. A Plan-Based Approach to Speech Act Recognition. Toronto:
University of Toronto.
Blom, Jan-Petter, and John Gumperz. 1972. Social meaning in linguistic structures: Code
switching in Norway. In Directions in Sociolinguistics: The Ethnography of Communication,
ed. John J. Gumperz, and Dell Hymes, 407–434. New York: Holt.
Bolter, Jay David, and Richard Grusin. 1999. Remediation. Understanding New Media.
Cambridge: MIT Press.
Bruce, Bertram, and Denis Newman. 1978. Interacting plans. Cognitive Psychology 2(3):
195–234.
Changeux, Jean-Pierre. 1983. L’homme neuronal. Paris: Fayard.
Chomsky, Noam Avram. 1965. Aspects of the Theory of Syntax. Cambridge: MIT Press.
Cohen, Philip Raymond. 1978. On Knowing What to Say: Planning Speech Acts. Toronto:
University of Toronto.
Coşeriu, Eugeniu. 1975. Sprachtheorie und allgemeine Sprachwissenschaft. 5 Studien. München:
Fink.
de Beaugrande, Robert-Alain, and Wolfgang Ulrich Dressler. 1981a. Eine Einführung in die
Textlinguistik. Tübingen: Niemeyer.
de Beaugrande, Robert-Alain, and Wolfgang Ulrich Dressler. 1981b. Introduction to Text
Linguistics. London-New York: Longman.
de Beaugrande, Robert-Alain, and Wolfgang Ulrich Dressler. 1994. Introduzione alla linguistica
testuale (trans: Muscas, S.). Bologna: Il Mulino.
de Saussure, Ferdinand. 1916. Cours de linguistique générale, ed. Charles Bally, and Albert
Sechehaye. Lausanne-Paris: Payot.
Dressler, Wolfgang Ulrich. 1970. Textsyntax und Übersetzung. In Sprachwissenschaft und
Übersetzen, ed. Peter Hartmann, and Henri Vernay, 64–71. München: Hüber.
Edmondson, Willis. 1980. On negoziation in discourse: Contras and counters in exchange
structure. Grazer Linguistische Studien 11: 28–44.
Fillmore, Charles John. 1968. The Case for cases. In Universals in Linguistic Theory, ed. Emmon
Bach, and Robert T. Harms, 1–88. New York: Holt, Rinehart & Winston.
Giuffrè, Silvia. 2013. Il pensiero in movimento (Ipotesi per una filosofia della danza). In Creatori
di senso. Identità, pratiche e confronti nella danza contemporanea italiana, ed. Massimo
Schiavoni, 49–62. Roma: Aracne. doi:10.4399/97888548662255
Goldstein, Ira, and Seymour Papert. 1977. Artificial intelligence, language and the study of
knowledge. Cognitive Science 1(1): 84–123.
Greenbaum, Sidney. 1973. Informant elicitation of data on syntactic variation. Lingua 31:
201–212.
Modern Studies 73
Grice, Herbert Paul. 1975. Logic and conversation. In Syntax and Semantics 3: Speech Acts, ed.
Peter Cole, and Jerry L. Morgan, 41–58. New York: Academic Press.
Halliday, Michael, and Ruqaiya Hasan. 1976. Cohesion in English. London-New York: Longman.
Hartmann, Peter. 1963. Theorie der Grammatik: Allgemeinste strukturgesetz in Sprache und
grammatik. The Hague: Mouton.
Hjelmslev, Louis Trolle. 1928. Principes de grammaire générale. Copenhague: AF Høst.
Hjelmslev, Louis Trolle. 1935. La catégorie des cas. Étude de grammaire générale. première
partie. Vol. Acta Jutlandica VII 1. Aarhus: Universitetsforlaget.
Jackendoff, Ray. 1996. Semantics and cognition. In The Handbook of Contemporary Semantic
Theory, ed. Shalom Lappin, 539–559. Oxford: Blackwell.
Jackendoff, Ray. 2002. Foundations of Language. Brain, Meaning, Grammar, Evolution.
London-Oxford-New York: Oxford University Press.
Johnson, David. 1977. On relational constraints on grammar. In Syntax and Semantics VIII:
Grammatical Relations, ed. Peter Cole, and Jerrold Sadock, 151–178. New York: Academic
Press.
Lakoff, George. 1971. On generative semantics. In Semantics: An Interdisciplinary Reader in
Philosophy, Linguistics and Psychology, ed. Danny D. Steinberg, and Leon A. Jakobovits,
232–296. Cambridge: Cambridge University Press.
Lakoff, George. 1973. Fuzzy grammar and the performance/competence terminology game, ed.
Claudia Corum, Thomas Cedric Smith-Stark, and Ann Weiser. Papers from the Ninth Regional
Meeting, Chicago Linguistics Society, April 13–15, 1973 (CLS, 9). Chicago: Chicago
Linguistics Society. 271–291.
Lakoff, George, and Mark Johnson. 2003. Metaphores We Live By. New Edition with a new
Afterword. Chicago: University of Chicago Press.
Legrenzi, Paolo. 2002. Prima lezione di scienze cognitive. Roma-Bari: Laterza.
Lemke, Jay L., and Caspar van Helden. 2009. A tribute to Robert de Beaugrande. Functions of
Language 16(1): 1–3.
Levesque, Hector. 1977. A procedural approach to semantics networks. University of Toronto,
Dept. of Computer Science, M.Sc. thesis, Toronto.
Levesque, Hector, and John Mylopoulos. 1979. A procedural semantics for semantic networks.
In Associative Networks: Representation and Use of Knowledge by Computers, ed. Nicholas V.
Findler, 93–120. New York-San Francisco-London: Academic Press.
Mantovani, Giuseppe. 1995. L’interazione uomo-computer. Bologna: Il Mulino.
Marconi, Diego. 1999. La filosofia del linguaggio. Da Frege ai giorni nostri. Torino: UTET.
McCawley, James. 1972. Syntactic and Logical Arguments for Semantic Structures. Bloomington:
Indiana University Linguistics Club.
McLuhan, Marshall. 1962. The Gutenberg Galaxy. Toronto: University of Toronto Press.
Meyer, Bonnie. 1977. What is remembered from prose: A function of passage structure. In
Discourse Production and Comprehension. Discourse Processes: Advances in Research and
Theory, ed. Roy O. Freedle, 307–336. Norwood: Ablex.
Miller, Geraldine, and Philip Johnson-Laird. 1976. Language and Perception. Cambridge:
Harvard University Press.
Minsky, Marvin. 1975. A framework for representing knowledge. In The Psychology of Computer
Vision, ed. Patrick Henry Winston, and Berthold Horn, 211–277. New York: McGraw-Hill.
Petöfi, János Sándor. 1976. Some Remarks on the Grammatical Component of an Integrated
Semiotic Theory of Texts. Bielefeld: University of Bielefeld.
Petőfi, János Sándor. 2004. Scrittura e interpretazione. Introduzione alla Testologia Semiotica dei
testi verbali. Roma: Carocci.
Reichman, Rachel. 1978. Conversational coherency. Cognitive Science 2: 283–327.
Rubin, Andee. 1978. A Theoretical Taxonomy of the Differences between Oral and Written
Languange. Cambridge: Bolt, Beranek & Newman.
Rumelhart, David Everett. 1977. Introduction to Human Information Processing. New York:
Wiley.
Schank, Roger, and Robert Abelson. 1977. Scripts, Plans, Goals and Understanding. Hillsdale:
Erlbaum.
Schneider, Peter. 1978. Organization of Knowledge in a Procedural Semantic Network
Formalism. Toronto: University of Toronto.
Shannon, Claude, and Werner Weaver. 1949. The Mathematical Theory of Communication.
Urbana: University of Illinois Press.
Sprung, Lothar. 1964. Zur Psychologie des Gedächtnisses. Zeitschrift für Psychologie 1: 35–51.
Tabossi, Patrizia. 1998. Intelligenza naturale e intelligenza artificiale. Bologna: Il Mulino.
Traini, Stefano. 2006. Le due vie della semiotica. Teorie strutturali ed interpretative. Milano:
Bompiani.
Tulving, Endel, and Wayne Donaldson. 1972. The Organization of Memory. New York-San
Francisco-London: Academic Press.
van Dijk, Teun Adrianus. 1972. Some Aspects of Text Grammars: A Study in Theoretical
Linguistics and Poetics. The Hague: Mouton.
van Dijk, Teun Adrianus. 1977. Text and Context. London-New York: Longman.
van Dijk, Teun Adrianus, and Walter Kintsch. 1978. Cognitive psychology and discourse:
Recalling and summarizing stories. In Current Trend in Text Linguistics, ed. Wolfgang Ulrich
Dressler, 61–80. Berlin: de Gruyter.
Violi, Patrizia. 1997. Significato ed esperienza. Milano: Bompiani.
Walker, David E. (ed.). 1978. Understanding Spoken Language. New York-Amsterdam: North
Holland.
Webber, Bonnie. 1978. A Formal Approach to Discourse Anaphora. Cambridge: Bolt, Beranek &
Newman.
Weinrich, Harald. 1972. Thesen zur Textsortenlinguistik. In Textsorten. Differenzierungskriterien
aus linguistischer Sicht, ed. Wolfgang Raible, and Elisabeth Gülich, 161–169. Frankfurt am
Main: Athenäum.
Winograd, Terry. 1976. Towards a procedural analysis of semantics. Stanford: Stanford
University.
Winston, Patrick Henry. 1977. Artificial Intelligence. Reading: Addison-Wesley.
Woods, William. 1970. Transition network grammars for natural language analysis.
Communication of Association for Computer Machinery (ACM) 13(10): 591–606.
Wright, Patricia. 1968. Sentence retention and transformation theory. Quarterly Journal of
Experimental Psychology 20(3): 265–272.
Chapter 3
The Procedural Approach to a Text
Haec nostro more nos diximus, Epicurii dicunt suo;

sed quae dicant videamus, quo modo neglegamus.
Cicero, Tusculanae Disputationes
If human communication is viewed from the perspective of the procedural

approach, conceptual content becomes more important than grammar rules. In fact,
even if human communication sometimes does not obey the rules of grammar, it
continues to flow incessantly nonetheless. However, if it enters a domain where
grammaticality is missing, this domain will be characterized more by the absence of
conceptual support than by a lack of respect for grammatical rules. The distinction
between grammatical and non-grammatical communication also depends on
extra-textual factors such as the tolerance of the interlocutor, his prior knowledge,
and the textual typology. If the distinction between text and non-text decreases, the
importance of other criteria will doubtlessly increase in order to mark the textuality
of a linguistic artefact. These criteria are precisely identified by de Beaugrande and
Dressler, who devote most of the Introduction to them.
The concept (quae dicant), therefore, and not the grammatical correctness of an
expression (quo modo), is the centre of human communication in the perspective of
cognitive science. Rhetoric has long tried to teach us how to persuade a large
audience, one which is not easily moved by an argumentative discourse full of
complex content. The art of persuasion has today become the focus of great the-
oretical interest. This broader conception of metaphor—understood as a cognitive
mechanism—is the basis for defining global models within procedural approach.
For this reason, the problem arises of how to identify the way individuals use such
models during verbal interaction.
In the second half of the twentieth century, when it was understood that the
faculty of language is connected directly to thought and is the instrument by which
thought is expressed, the linguistic turn prompted us to look again at rhetoric with
renewed interest. This ancient discipline had dealt with language—the same object
of study of text linguistics—for centuries, studying the effects and proposing
classifications and definitions for various functions.

76 3 The Procedural Approach to a Text
If people confront each other in a relatively non-violent way through language,

then language becomes the embodiment of a faculty by which individuals con-
stantly negotiate the possible achievement of their goals. To successfully use the
faculty of language, people need to develop a technique in order to win the “war”,
i.e. a handy and flexible “weapon” that can connect concepts and be used in
communicative interaction as a tool in their daily struggle for success in their
negotiations (that is, to achieve the goals to which de Beaugrande and Dressler
refer). This weapon is rhetoric.
Rhetoric, Cognitive Psychology and Ancient Literary

Genres
In the mid-twentieth century, due to the broad expansion of political and civil rights
and the rise of large-scale consumption, a so-called “mass society” was born in the
Western world. As a result of this new socio-economic situation, it became
increasingly necessary to learn how to convince the general public, in order to guide
its consumption, but it was not easily persuaded by complex scientific evidence.
Therefore, that was exactly what the art of persuasion taught.
Barthes found a strong connection between Aristotelian rhetoric and advertising.
He argued that: «the rhetoric of Aristotle is mainly a rhetoric of proof, of reasoning,
of approximate syllogism (enthymeme = ἐνθύμημα); it is a logic that has been
voluntarily degraded, suited to the level of the “public”, i.e. to common sense and to
general current opinion. Extended to literary productions (which was not its original
purpose), it would implicate primarily the public’s aesthetics rather than the aes-
thetics of any given work. It is the reason why, mutatis mutandis and taking into
consideration all (historical) proportions, this rhetoric would be well suited to the
products of our so-called “mass” culture, which is the reign of Aristotelian
“verisimilitude”, and this is specifically “what the public believes possible”» (my
translation of Barthes 1970, p. 21).
In the Fifties rhetoric began to be the subject of important new studies. In the
United States, Yale University was prominent at the forefront, and its chief expo-
nent was the psychologist Carl Iver Hovland.1 He, along with his group, led
numerous experiments on attitude change and persuasion, focusing particularly on
the factors used by a receiver to evaluate the credibility of a source, that is on the
elements that make a message compelling and their optimum arrangement inside of
a text. In Europe rhetoric was also rediscovered in the same period, but from a
different point of view. In Exercices de style (1947), Raymond Queneau composed
a brief text with 98 variations in which the text is rewritten using different meta-
phors, or different styles, or a different arrangement of its parts. Chaim Perelman,
1
A complete bibliography of Hovland’s works, a summary of the Yale school and other psy-
chological research aimed at the study of rhetoric is found in de Montmollin (1969).
Rhetoric, Cognitive Psychology and Ancient Literary Genres 77
another French scholar, published Traité de l’argumentation. La nouvelle

rhétorique (1958) with the intent of continuing Aristotle’s work on rhetoric, with a
particular focus on the study of persuasion. In Italy, too, several works recognized
the importance and merits of rhetoric in mass society.2
Umberto Eco clearly expressed his position as follows: «The reduction of phi-
losophy and other forms of argument, though once considered indisputable, to
rhetoric is an achievement of, if not reason, at least of reasonableness, which has
become cautious towards any fanatical and intolerant faith. In this sense, rhetoric,
initially the art of persuasion—understood almost as subtle deception—is increas-
ingly recognized as a technique of human reasoning, controlled by doubt, subject to
all the historical, psychological and biological influences weighing on every human
act. But there are varying degrees of persuasive speech. And between these degrees
a continuous series of shades can be detected, ranging from honest and cautious
persuasion to persuasion as deception. Let us say, from a philosophical discourse to
techniques of propaganda and mass persuasion» (my translation of Eco (1968),
pp. 84–85).
However it was not in mass society that rhetoric was born.3 Ars rhetorica
originated almost certainly in the Greek culture of Sicily—generally called Siceliot
culture—in the fifth century BC. Greek and Latin sources were aware of this lineage
and gave two different versions of the genesis of this discipline.
According to the first, Thrasydaeus, tyrant of Akragas (now Agrigento) and
Himera, was defeated by Hiero I, tyrant of Syracuse, in 472 BC and was forced to
flee. Akragas and Himera, now free from tyranny, adopted first an oligarchic and
then a democratic political system. Empedocles of Akragas, the philosopher, was
one of the leading proponents of democracy, which he passionately defended in
public assemblies. In 466 BC, Thrasybulus, the brother of Hiero I, came to power in
Syracuse. The Syracusans, however, unable to bear any new tyranny, overthrew the
system with a democratic revolution. Corax, who has been called the first true
rhetorician, was a public figure close to Hiero I and was also the teacher of Tisias.
Through his political activities, he tried to exert a conservative influence on the
citizens’ opinions. Corax and Tisias were among the first to teach this art. It is not
clear what their exact role was, but their names have been handed down by all
ancient authors. We cannot be sure of the nature of their early teachings and
writings about rhetoric. According to some theories, the first texts on rhetoric were
collections of speeches that were chosen to serve as exempla: by reading, studying
and imitating speeches that had already been written, one was expected to gradually
learn the art of speaking and discussing issues in public. It is likely however that in
addition to anthologies of useful models, there were also veritable manuals (called
τέχναι) that circulated as well, in which the fundamentals of rhetoric were explicitly
explained, and also discussed from a theoretical perspective. Tisias, since he was
the author of a τέχνη, was strictly speaking the first rhetorician.
2
Among the best known scholars, Eco must be remembered; see Eco (1968).
3
On the origin of ancient rhetoric, see Montanari (1998, pp. 403–408).
In the Latin version, the origin of rhetoric follows roughly the same contours,
though it is presented in a different way. The tyrants of Syracuse, after conducting
extensive expropriations, distributed the fertile farmland to loyal mercenaries. After
the fall of tyranny, the citizens then filed a series of lawsuits in order to regain
possession of their farms. This intense forensic activity led many to begin won-
dering about how best to convince a jury.4
It should be noted that in both versions, the birth of rhetoric is linked to the
revival of public life after the overthrow of tyranny; and on this aspect all ancient
sources agree. Although the two traditional versions do not provide a clear
understanding of what the nature of rhetoric was at that time (political rhetoric
aimed at convincing citizens during public debates or judicial rhetoric arising from
claims of land ownership), one can nevertheless delineate from the very beginning
the existence of two of the three kinds of rhetorical speech: genus deliberativum
(i.e. aimed at convincing citizens during public debates) and genus iudiciale (i.e.
judicial rhetoric). The last of the three, genus demonstrativum, seems to have
developed later.
Rhetorical speech is divided into three main kinds, distinguished by the nature of
the recipient. This distinction is made by Aristotle in the following passage: «The
kinds of Rhetoric are three in number, corresponding to the three kinds of hearers.
For every speech is composed of three parts: the speaker, the subject of which he
treats, and the person to whom it is addressed, I mean the hearer, to whom the end
or object of the speech refers. Now the hearer must necessarily be either a mere
4
This information comes from Cicero (in particular paragraph 46): «[45] This age therefore first
produced at Athens an orator all but perfect. For the ambition to speak well does not arise when
men are engaged in establishing government, nor occupied with the conduct of war, nor shackled
and chained by the authority of kings. Upon peace and tranquillity eloquence attends as their ally,
it is, one may say, the offspring of well-established civic order. [46] Thus Aristotle say that in
Sicily, after expulsion of tyrants, when after a long interval restitution of private property was
sought by legal means, Corax and Tisias the Sicilians, with the acuteness and controversial habit of
their people, first put together some theoretical precepts; that before them, while many had taken
pains to speak with care and with orderly arrangement, no one had followed a definite method or
art. He says further that Protagoras wrote out and furnished discussions of certain large general
subjects such as we now call commonplaces; [47] that Gorgias did the same, writing particularly in
praise or in censure of given things, since he held that it was the particular function of oratory to
magnify a thing by praise, or again by disparagement to belittle it; that Antiphon of Rhamnus
produced some similar writings, concerning whom we have the trustworthy assurance of
Thucydides that no man ever pleaded his case better, when in his hearing Antiphon defended
himself on a capital charge; [48] that as to Lysias, it was only in the beginning of his career that he
professed the art of rhetoric, but afterwards, seeing that Theodorus was a more skilful theorist and
teacher, though dry as a speaker, he began to compose speeches for others and abandoned the
profession of teacher. He tells also how Isocrates with similar alternation at first denied that there
was an art of speaking, while at the same time he was writing speeches for others to use in court;
but when it happened repeatedly that he was summoned as having violated a law like ours
“providing against circumvention or chicanery by judicial process”, he ceased to write speeches
for others and devoted himself wholly to the composition of theory and models of oratory».
Cicero, Brutus 45–48, in the English translation by George Lincoln Hendrickson, fellow of
Branford College, in the Loeb Classical Library (1952), volume 3, book V, XLIV, pp. 149–151.
Rhetoric, Cognitive Psychology and Ancient Literary Genres 79
spectator or a judge, and a judge either of things past or of things to come. For
instance, a member of the general assembly is a judge of things to come; the dicast,
of things past; the mere spectator, of the ability of the speaker. Therefore there are
necessarily three kinds of rhetorical speeches, deliberative, forensic, and
epideictic».5
This partition will be maintained in the following centuries, and each of these
types of discourse will have a different purpose: the deliberative discourse served to
decide what is useful or harmful to the community, the judicial discourse what is
right or wrong, and the epideictic discourse what is beautiful or ugly.
The text which we intend to analyze according to the criteria of procedural
approach is of the deliberative type. It is a speech delivered to the elders in the local
Senate of Ardea by General Marcus Furius Camillus, and its purpose was to arouse
support for helping Rome, which had been militarily occupied by the Gauls in 390
BC during their invasion of Lazio. The historical event is known as the “Gallic sack
of Rome” and the oration is contained in a passage of the work of the Roman
historian Titus Livius.
Titus Livius and the Gallic Sack of Rome
Titus Livius (whose cognomen is unknown) was a Roman historian who lived
between 59 BC and 17 AD. Born into a wealthy family, he devoted his entire life to
literary pursuits and was the author of a monumental history of Rome, traditionally
known by the title Ab Urbe Condita libri CXLII. This work described Roman
history from the traditional date of its founding (April 21, 753 BC) to the death of
Augustus (14 AD).
Though Livius composed most of his work under Augustus, scholars6 have
repeatedly stressed the attachment to republican values that emerges from his
writings and his veiled desire for a Republican restoration. His work however never
indulges in an overt celebration aimed at the princeps.7 Due to the lack of books
relating to the Republican period of crisis and the rise of Augustus, none of his
political beliefs can be expressed with certainty; however, it can be affirmed with
certainty that Livius was critical of some of the values embodied by the constitu-
tional law promoted by Augustus. On the other hand, this attitude did not bother the
5
Aristotle, Rhetoric 1358b, in the English translation by John Henry Freese, fellow of St. John’s
College, in the Loeb Classical Library (1947), book I, II. 22. 2–3, p. 33.
6
The most prominent Italian Latin scholars interested in aspects of Livian ideology are Mazza and
Pianezzola, and we refer specifically to Mazza (1966) and Pianezzola (1969).
7
Augustus called Livius by the nickname “Pompeian” due to his pro-Republican views. This fact
is reported by Tacitus (Annales IV, 34), which in turn relies on information taken from the lost
historian Cremutius Cordus.
princeps, who entrusted him with the education of his nephew Claudius, who later
became emperor.8
The Ab Urbe Condita libri project included a total of 142 books grouped into
groups of ten, called decades for this reason. Only 35 books have survived, i.e. the
first decade (books I–X) and books XXI–XLV (two decades and a pentad). Only
short fragments of the others are known;9 nevertheless, the contents of the rest of
the work are known through summaries of the individual books called periochae
(περιοχαί). Due to the large size of the work, several compendia, called epitomi
(ἐπιτομαί), written by various authors, were already in circulation in antiquity;
among these we must give special mention to the careful work done by Eutropius.10
One peculiarity of Livius’ work is his method of using historical sources: he did
not make use of original documents, even if these existed, but rather used mainly
literary sources.11 For this reason it is believed that the real talent in the work of Livius
is not seen in his reliability as a historian, but in the beauty of his literary work.
The literary and artistic value of Livius’ work can be appreciated on two different
levels: the precepts and the enjoyable quality of the narrative. In regard to the first, he
criticized the customs of his era and considered them decadent; on the contrary, he
exalted the values that gave eternal fame to the Urbe (above all, virtus and frugalitas)
and thereby contributed to the construction of the myth of Roma Aeterna. In regard to
the second, his style of exposition established a clear departure from the style of prior
historians (in particular, from Polybius and Sallust) and he favoured a general char-
acter of greater eloquence, proceeding in the tradition of Herodotus and Isocrates. For
this purpose, he adopted a narrative prose, with the frequent use of dramatization.12
8
It is believed that the influence of Livius on Claudius became particularly evident in the final
period of Claudius’ reign, when his oratory grew full of quotations from Roman history, drawn
faithfully from Livius’ work.
9
Among which the fragment of book CXX is especially significant, as it contains the account of
the death of Cicero.
10
Eutropius was a master of rhetoric and was probably of Italian origin. At the request of the
Emperor Valens, to whom he was magister memoriae and secretary, he wrote Breviarium ab Urbe
condita, a work in ten books that summarized Roman history from the founding of the city until
the death of the emperor Jovian (364 AD). In fact, Eutropius employed many sources, but among
them the role of Livius is preeminent.
11
For example, Livius gives two versions of the disappearance of Romulus: the first is a mythical
version, and speaks of his ascension among the gods; the second is a secular version, according to
which Romulus was killed. Livius does not favour one of the two versions, but leaves the decision
about which to give credit to to the discretion of the reader. In particular, Livius argued that the
lack of trustworthy sources prior to the sack of Rome in 390 BC made the task of reconstructing
the archaic period very difficult.
12
The narrative intent was, nevertheless, framed within an annalistic schema: Livius used a style in
which historical chronology and narration alternate, and he interrupts his narrations to announce
the election of new consuls, since this was the system used by the Romans to mark the passing
year and the one traditionally adopted by historians.
Titus Livius and the Gallic Sack of Rome 81
The narrative style of Livius exhibits its sense of drama in two main areas: the
first is the role played by the characters, which serve as exempla. Livius portrays his
characters as absolute protagonists, as if they were paradigms of human passions.
The second area of dramatization is the speeches “pronounced directly” of his
characters—most are in the form of direct speech, but Livius also makes use of free
indirect speech. These speeches are composed more to stimulate the imagination of
the reader and to support the myth of Rome Aeterna than to represent historical
truth, unless done so in a merely abstract way; in other words, Livian speeches
respond to narrative and psychological needs and are structured according to typical
forms of ancient rhetoric.13
One of the best known characters used as an exemplum of the military value of the
Romans and as the main protagonist of the first decade is General Marcus Furius
Camillus, saviour of the Urbe during the so-called “Gallic sack of Rome” in 390 BC.
He is the protagonist of one of the most traumatic episodes in the Republican period
of the city: the conquest by the Senon Gauls led by Brennus, known by the name of
Clades Gallica (Gallic defeat). This episode, about whose historicity there is no
doubt, is narrated not only by Livius, but by other ancient historical sources as
well: Polybius Histories II, 18, 1–3,14 Diodorus of Sicily Book XIV, 113–117,15
13
The sense of drama can also seen in the use of the typical Greek techinque known as Deus ex
Machina to save a narrative situation that is about to fail.
14
«18. On their first invasion they not only conquered this country but reduced to subjection many
of the neighbouring peoples, striking terror into them by their audacity. Not long afterwards they
defeated the Romans and their allies in a pitched battle, and pursuing the fugitives, occupied, three
days after the battle, the whole of Rome with the exception of the Capitol, being diverted by an
invasion of their own country by the Veneti, they made on this occasion a treaty with the Romans,
and evacuating the city, returned home.» Polybius, Histories II, 18, 1–3 in the English translation
by William Roger Paton, Ph.D. of Stanford University, in the Loeb Classical Library in 6 volumes
(1923), volume 1, p. 285.
15
«117. While the Romans were in a weakened condition because of the misfortune we have
described, the Volscians went to war against them. Accordingly the Roman military tribunes
enrolled soldiers, took the field with their army, and pitched camp on the Campus Martius, as it is
called, two hundred stades distant from Rome. Since the Volscians lay over against them with a
larger force and were assaulting the camp, the citizens in Rome, fearing for the safety of those in
the encampment, appointed Marcus Furius dictator. … These armed all the men of military age
and marched out during the night. At day-break they caught the Volscians as they were assaulting
the camp, and appearing on their rear easily put them to flight. When the troops in the camp then
sallied forth, the Volscians were caught in the middle and cut down almost to a man. Thus a people
that passed for powerful in former days was by this disaster reduced to the weakest among the
neighbouring tribes. After the battle the dictator, on hearing that Bola was being besieged by the
Aeculani, who are now called Aequicoli, led forth his troops and slew most of besieging army.
From here he marched to the territory of Sutrium, a Roman colony, which the Tyrrhenians had
forcibly occupied. Falling unexpectedly upon the Tyrrhenians, he slew many of them and
recovered the city for the people of Sutrium. The Gauls on their way from Rome laid siege to the
city of Veascium which was an ally of the Romans. The dictator attached them, slew the larger
number of them, and got possession of all their baggage, included in which was the gold which
they had received for Rome and practically all the booty which they had gathered in the seizure of
the city. Despite the accomplishment of such great deeds, envy on the part of the tribunes
prevented his celebrating a triumph. There are some, however, who state that he celebrated a
and in more detail by Plutarch, who wrote a Life of Camillus, a biography of the
General.16
Livius’ version of the facts is found in Ab Urbe condita V, 37–55. He narrates
that after the Gauls departed from Senigallia (from “Sena [River] Gallica [of the
Gauls]”)17 for a raid in southern Etruria, they first defeated the Etruscans of Chiusi
easily, and then invaded Lazio. The Romans sent them ambassadors of peace, but
no treaty was signed and the Gauls proceeded with their invasion. There was then a
Roman attempt to stop the Gauls about eleven miles from the City, where the Allia
river flows into the Tiber,18 but this ended in a major defeat with heavy casualties
among the troops. That day, July 18, was given the name dies Alliensis, in memory
of the bitter defeat, and it was recorded in calendars as a dies nefastus (day on
which human activities were illicit), as it had become synonymous with disaster.
Pressed by the Gauls, the surviving soldiers withdrew haphazardly inside the walls
of Rome, forgetting to close the doors of the city. The aggressors were thus able to
enter and pillage the city.19
The Gauls then broke into the Senate. The senators, who had been informed of
what had happened and had chosen to remain seated in their pews in a hieratic
manner, let themselves be massacred to the last man. The rest of the Romans took
refuge within the fortifications of the Capitol, which was then placed under siege.
During this siege, the Gauls climbed the hill in an attempt to enter, but the
quacking of geese consecrated to Juno and preserved in the temple dedicated to her
triumph for his victory over the Tuscans in a chariot drawn by four white horses, for which the
people two years later fined him a large sum of money. But we shall recur to this in the appropriate
period of time. Those Celts who had passed into Iapygia turned back through the territory of the
Romans; but soon thereafter the Cerii made a crafty attack on them by night and cut all of them to
pieces in the Trausian Plain». Diodorus of Sicily, Book XIV, 117 in the English translation by
Charles Henry Oldfather, professor of ancient history and languages at the University of Nebraska,
in the Loeb Classical Library in 12 volumes (1923), volume VI, pp. 315–319.
16
More details concerning the entire historical period of the battle against the Gauls are found in
Plutarch, Life of Camillus, 15–32 (Plutarch’s Lives, with the English translation by Bernadotte
Perrin, Professor at Yale University, in the Loeb Classical Library (1968), vol. II, pp. 127–177).
17
See the debate in Chilese (2013).
18
A stream located at the 18th kilometer of the Via Salaria, whose area of confluence with the
Tiber is now called “Fosso della Bettina”.
19
During the devastation of the city, the State Archives, which contained many documents dating
back to the founding of Rome, were lost. All the events preceding the battle are thus largely
legendary and difficult to reconstruct historically, as can be deduced from Livius’ preface: “As for
the events relating to the founding of Rome or earlier, I seek neither to affirm they are true nor to
deny them: their appeal is due more to poets’ imagination than to the reliability of the informa-
tion.” Nevertheless, many modern historians believe that at that time there could not have been
many pre-existing chronicles or documents.
on the Capitoline Hill warned the defenders of the besiegers’ presence and the
attempts failed. The Romans then asked the Gauls to end the siege in exchange for
a ransom of pure gold, and they accepted. In this context, Brennus, the leader of
the Gallic expedition, ostentatiously cheated and weighed his sword as well, in
order to obtain even more gold, screaming in contempt: “Vae victis!” (Livius, V,
48).20
At this point of the story, Livius, with the clear intention of saving the honour of
Rome, uses the typical Greek Deus ex Machina technique to insert into the story the
providential arrival of General Marcus Furius Camillus, a character who embodies
all Roman virtues, an exemplum.21 Historically he was of noble family and his
cursus honorum began with his appointment as censor in 403 BC, and later
included four victories, five dictatorships and the honorary title of Pater Patriae
(Second Founder of Rome).22 His appointment as dictator came about during the
20
Vae Victis is a phrase that literally means “Woe to the vanquished”: while the Romans were
weighing the gold they had to pay the Gaul as a war tribute, one of them protested because the
weights were rigged. Brennus then drew his heavy sword and added it to the plate of the weights to
be matched in gold, thus making the calculation even more unfair; he simultaneously exclaimed
“Vae Victis” indicating that the terms of surrender are unquestionably dictated by the winners,
based on the right of the strongest. Since this story clearly tends to discredit the image of the Gauls,
the general opinion of historians is that the historical veracity of the episode is non-existent. The
phrase has become proverbial due to the spread of this account and it is frequently used as a bitter
comment on exceptionally cruel oppression or on excessive zeal against a defenceless opponent. It
has been used as a synonym for the terms of surrender of Italy after its defeat in World War II, with
obvious ulterior meanings.
21
Eutropius presents General Marcus Furius Camillus summarizing Livius’ version as follows:
«Twenty years afterwards, the people of Veii resumed hostilities. Furius Camillus was sent as
dictator against them, who first defeated them in battle, and then, after a long siege, took their city,
the oldest and richest in Italy. He next took Falisci, a city of no less note. But popular odium was
excited against him, on the ground that he had made an unfair division of the booty, and he was
condemned on that charge and banished. Soon after the Galli Senones marched towards Rome;
and, pursuing the Romans, whom they defeated at the river Allia, eleven miles from the city,
possessed themselves of the city itself, no part of which could be defended against them, except the
Capitol. After they had besieged it a long time, and the Romans were suffering from famine,
Camillus, who was in exile in a neighbouring city, attacked the Gauls unexpectedly, and gave them
a severe defeat. Afterwards, on receiving a sum in gold, to desist from the siege of the Capitol, they
retreated; Camillus, however, pursued them, and routed them with such a slaughter, that he
recovered both the gold which had been given to them, and all the military standards which they
had taken. Thus he entered the city for the third time in triumph, and received the appellation of a
second Romulus, as if he also had been a founder of the city». Eutropius, Breviarium ab Urbe
Condita, I, 20 in the English translation by John Selby Watson, Rev. of London, in the Henry G.
Bohn edition (1853). The translation is online avalaible.
22
The dictatura was one of the extraordinary Roman honores dating back to the origins of the
republic, and it remained active until the Second Punic War. «The dictator was created through a
solemn dictio, which was the ritual appointment that was possible only in agro Romano. The
appointment was made when one of the consuls, by arrangement with his colleague and the
Senate, entrusted the post of dictator to someone for a period as long as necessary to respond to the
needs (usually military, but not only) which had prompted its creation, and in any case not more
than six months. The relation between the dictatorship and the power of the consul, moreover,
would seem to have been such that the dictator would in any case expire with the expiration of the
war against the Etruscan city of Veii, in which he defeated the inhabitants and, to
avoid further wars, destroyed the Etruscan city completely. However, after being
accused of distributing the spoils of war unjustly, he chose voluntary exile at Ardea
in Lazio, demonstrating his patrician pride.23 From there, the general urged the
elders in the local Senate to take up arms to defend the city, along with the rest of
Lazio, in a vehement deliberative speech, which is our case study below.
The Gauls, in fact, after sacking Rome, turned against Ardea and threatened to
besiege it; the Ardeati, after entrusting their leadership to General Marcus Furius
Camillus, managed to repel the siege; meanwhile, the exiles who had been cast out
of Rome re-appointed him dictator and, after gathering his troops, he returned to the
Capital with the cry «The honour of the fatherland is not defended with gold, but
with iron weapons!» (ferro non auro reciperare patriam) and routed the Gauls.
While the Gauls attempted to retreat to their territories, the General pursued them,
defeated them, recovered the coat of arms and the spoils they had obtained and
returned to Rome, where he received the honorary title of Pater Patriae (Second
Founder of Rome).
Livius’ version is openly aimed at highlighting the values of Roman virtus and at
exalting the myth of Roma Aeterna: the dictator, although he has been removed
from the homeland by his fellow citizens, puts aside his personal grudges during the
siege and rushes to save the country in danger; secondly, the peoples of the Latin
League, though ill-tolerating the evil yoke that Rome forced on them and having
consuls’ term of office, during which the dictio had been made. The imperium of the dictator was
not susceptible to the intercessio of the Plebeian tribunes; it was summum and was therefore
imposed on all the other magistrates. The constitutional function of this office was to overcome any
political impasse resulting from the dynamics of ordinary constitutional balances (based on a
delicate system of checks and balances) that could block the res publica in times of severe tension
and political uncertainty. With the appointment of the dictator (and the consequent suspension of
all other magistrates and tribunes), the consuls and the Senate took the initiative and invested a
person with supreme powers, someone able to realize, once every type of veto had failed, the tasks
that had been or could be hindered or obstructed. But in so doing, they also bore the political
responsibility for the initiative, exposing themselves to the consequences of public opinion that
was based on the success or failure of the actions of the dictator. On taking up office (which
occurred extemplo, i.e. immediately), the dictator had to appoint in turn a magister equitum. The
exercise of military power was, for both, subject to the lex curiata de imperio» (my translation of
Cerami et al. 2006, p. 57).
23
Probably, the real causes of the defamatory accusations were the excessive power and authority
of the patrician; moreover, the archaic Roman mentality considered his triumphant entry into
Rome on a chariot drawn by white horses an excessive personalization, which may have fuelled
the irritation that exploded in senatorial spheres against him.
tried several times to regain their lost independence, in the face of barbarism enter
the battle to repair the damaged prestige of the Capitol.24
The episode is framed very differently from Livius in modern historiography: the
Gallic sack of Rome was the official start of the Roman-Celtic conflict, which was
prolonged by a series of conflicts and lasted until the Po Valley was finally con-
quered by the Romans in the second century BC, becoming a province of Cisalpine
Gaul25 after having been progressively Romanized over the previous 150 years.
Historiography however is unanimous in the belief that the first Celtic threat
against Rome was not an episodic expedition in order simply to raid, but part of
large migratory movements originating in north-eastern Gaul and central Europe.26
These movements had as their final result the Gallic occupation of many areas of
northern Italy and the central and northern lands bordering the Adriatic that Rome
was interested in; instead, the Celtization of much of Europe had taken place
without Rome’s knowledge, whose interest in security in the fourth century BC was
above all local. In fact, until then, Rome had had to deal primarily with Aequians,
24
In the context of events regarding the Latin League, the Ardeati were alternately allies or
enemies of Rome (under Tarquinius Superbus, Rome launched an initial attack against Ardea,
which was not successful; however, in 509 BC, under the first treaty between Rome and Carthage,
the city was among the Romans’ allies). Ardea was the ancient capital of the Rutulians, one of the
oldest peoples of Latium Vetus, which included the towns of Antium, Satricum and Lavinium. The
city, which has extremely ancient origins, is located in the western part of the Alban Hills. Its
origins were described by various versions of the myth, all linked to the story of the landing of
Aeneas on the coast of Lazio, and therefore also to the birth of Rome (Dionysius of Halicarnassus,
Ovid, Virgil). The city was one of the most important centres of southern Lazio (from the 8th to the
9th century BC) and it possessed skilled craftsmanship and wealthy commerce. It was renowned
for the production of weapons and ornaments and was favoured by its position at the crossroads of
Latin, Volscian and Etruscan populations and by its port at the mouth of the river Incastro (Inui
Castrum). During the fifth century BC the city fought against the Volscians.
25
The main centres of Cisalpine Gaul were Mediolanum (Milano), inhabited by the Insubrians,
Verona, a Cenomanian settlement, and Bononia (Bologna), occupied by the Boii, which was
known to the Etruscans as Felsina. The Senones, in particular, settled in a strategic area of the
Marche, between the present cities of Pesaro, Macerata and Ancona: their position allowed them
easy control over access to the Tiber Valley and to the Adriatic roads towards Puglia and
Campania; see Eluère (1992, p. 68).
26
Written texts and archaeological evidence agree on the dual origin of the Celts: on one hand
north-eastern Gaul (Champagne), whence came the Senones who settled in the Marche and the
hinterland of the Adriatic coast; on the other Central Europe (Bohemia), whence came the Boii.
There is significant archaeological evidence of ties already in the fifth century BC between these
two regions and the Italic-Etruscan and Italiot area, and of the persistent links between the peoples
who settled in Italy and their lands of origin. The artistic productions of the Remi of Champagne,
whose quality is clearly indebted to Greek-Etruscan influences, provide some of the evidence; the
continuity of relations is evidenced by the development in Bohemia of a new and original local art
form, though it was influenced by the cultures of the Italian peninsula through the mediation of the
Celtic-Italic environment; see Kruta (2004, pp. 38–42).
Volscians, Sabines and Etruscans—neighbouring peoples and threats to its

security.27
In fact, only three powers dominated the equilibrium of the Italian peninsula in
the fifth century BC: Etruscans, Carthaginians and the Greeks of Magna Graecia
(Italiots and Siceliots). The Celtic invasion of Italy in the fourth century BC and the
conflicts that followed projected Roma into a broader dimension, but the ancients
were well aware that the echo of the Gallic sack of Rome reverberated well beyond
regional or peninsular limits.28 The causes that drove the Celts to migrate are not
known. Archaeological evidence testifies to a planned migration and agrees with
the consensus of modern historians on the Celtic occupations in Italy and the
consequent battles. Far from being spontaneous and improvised movements, Celtic
migrations appear to be carefully planned operations and strategic elements in the
play of interests and alliances on the Italian peninsula.
Most likely a part of this scheme was the military and strategic plan of
Dionysius I, tyrant of Syracuse, one of the most sophisticated and ambitious players
in the political arena of that time. In the context of the traditional Greek-Punic
conflict, the tyrant had long been opposed to Carthage, which, in turn, had woven
alliances with the ancient Etruscan world to limit the expansion of the Greeks of
Magna Graecia (Italiots and Siceliots) in Italy. Dionysius I competed against the
poleis of the Italiot League and manifested his desire to replace Etruscan control of
the commercial marine traffic in the Tyrrhenian Sea with that of Syracuse and
27
The diplomatic relations of Rome had elevated to an international level only in some cases,
which led to direct contact with Carthage. However, these relations were framed in the wider
context of the regulation of Etruscan and Punic interests. Aristotle mentions such relationships
(Aristotle, Politeia 1280a) and the bilingual inscription (Phoenician and Etruscan) contained in the
gold foils of Pyrgi (Corpus Inscriptionum Etruscarum, 6314–6316) is considered a confirmation of
relations with the Phoenician—if not directly Punic—environment. The broad scope of this ancient
convergence had culminated in the second half of the sixth century BC in the anti-Phocean alliance
of the Battle of Alalia, which was decisive in defining the balance of power in the Tyrrhenian just
before the fifth century BC (see Acquaro 1988). Relations with Carthage had continued even after
the end of the Roman kingdom when Rome, at the very beginning of its Republican experience,
signed a treaty with Carthage—the first of a series—to which we find reference in Polybius
(Histories III, 22), dating back to the early transition stage from monarchy to republic, but which
obviously should be considered as part of the normal relations between the Etruscan area and
Carthage. Prior to the fourth century BC, other wide-ranging relations were maintained with
Etruscan capitals such as Clusium, with poleis such as Cuma, and with Sicily—relations arising
mainly in regard to the commerce of wheat (cf. Ogilvie 1976, pp. 78–159).
28
The events of the Gallic sack of Rome had a profound impact not only on Roman history, but
also on Etruscan history and that of other Italic peoples. The echo of this event also reverberated in
the tradition of Greek historiography: Heraclides of Pontus, a philosopher living just after these
events, incidentally gave an account in his book Περὶ ψυχῆς (On the soul) of a «rumour coming
from the West that an army, originating in Hyperborea, had conquered a Greek city called Rome,
located in an unspecified place of the Great Sea.» The news was also reported by Aristotle,
Theopompus and other authors of the fourth century BC; see Ogilvie (1976, p. 166).
expand his power along both sides of the Adriatic, seconding his advisor Philistus
and exploiting the new balance of power after the peace of Antalcidas.29
On the basis of various sources and archaeological evidence, historians today are
generally in agreement about the Celtic occupation of Italy and the related military
battles: the Celtic penetration into Italy was not a spontaneous event, but rather the
result of the careful planning of Dionysius I of Syracuse. The invasion was sup-
ported, perhaps even co-directed, by one of the key players in the political balance
of the Italian peninsula: this ambitious tyrant who was fighting with Carthage for
supremacy and was in open conflict with the πόλεις of the Italiot League. In this
context he looked with favour upon the entrance of a new player in the geopolitical
arena, because he could reap many benefits from it: he aimed at broadening his
influence in Sicily and on the Tyrrhenian and Adriatic shores of Italy, and dreamed
of influencing even the Illyrian and Adriatic geopolitical area, too; for this reason he
was the first to openly contrast the Etruscans, whose alliance with Carthage had
long created a partition of the Tyrrhenian Sea between the two peoples.
At the beginning of the fourth century BC, Rome had just made a qualitative
leap forward in its history, both in regard to territorial expansion and increased
military organization. After the victorious conclusion of the war against Veii, the
political balances of the Etruscan capitals and their traditional internal tensions
underwent some changes: in fact, the hostility against Veii was somewhat con-
cealed thanks to the neutrality expressed by the other eleven Etruscan cities grav-
itating around Fanum Voltumnae;30 even Caere had offered its alliance to Rome.
The alliance between Rome and Caere during the attacks against Veii could well
be placed in the context of expansionist interests and hegemonic aims of the Siceliot
tyrant. Dionysius I, either as counsellor or commander, favoured the descent of a
new and belligerent subject into central and northern Italy, because the Celts had the
potential to trigger disturbing effects on the balance of the region, especially against
the Etruscans.31 The story of the siege of Reggio also fits well in this same
29
The Peace of Antalcidas or the King’s Peace (386 BC) was signed between Antalcidas, com-
mander of the Spartan navy, and Artaxerxes II of Persia to impose peace on the enemies of Sparta,
against Athens. The Peace of Antalcidas is the first example of a peace treaty guaranteed by
sanctions, ratified by all the Greek states, and having no time limit; the treaty safeguarded Spartan
hegemony over Greece, but at the cost of recognizing Persian hegemony over Asia Minor and
Cyprus. The terms of the deal, announced at Sardi in the winter of 387–386 BC, were formally
accepted in Sparta in 386 BC.
30
The fanum Voltumnae (sanctuary of Vertumna) was the Etruscan federal sanctuary, mentioned
by ancient sources, but of uncertain origin. It was dedicated to the god Vertumnus (probably an
aspect of the god Tinia). Every year in the spring, it was the meeting place of the heads of the
twelve Etruscan cities, who came to elect the supreme head of the federation. Religious festivals
were also held there and decisions were made in regard to domestic and foreign policy; see Torelli
(2000, pp. 275 and 282).
31
The attribution of this role to the Syracusan tyrant appears to be consistent with the subsequent
payment of Celtic mercenaries, who had been found on a labour market in which the Doric
emporium of Ancona probably had a central role; its port was close to the territory of the Senones;
see Sordi (2004).
framework: while favouring the entrance in Italy of the Celts, Dionysius I ruthlessly
repressed the Greek city-state at the southernmost tip of the Italian peninsula.32
The Gallic sack of Rome transferred the entire area into the sphere of the
political interests of the Greeks of Magna Graecia (Italiots and Siceliots):33
Polybius incorporated the news into traditional Greek historiography and dated it
chronologically. Since the date was linked to well-known events of Greek history, it
represents a placement of the event—specifically 387–386 BC—that is more reli-
able than the year 390 BC referred to in the Roman tradition.34
Although the vision of contemporary historiography is widely documented and
sufficiently substantiated from a scientific point of view, from a literary perspective
the General’s speech, contained in Ab Urbe condita V, 44, undoubtedly reaches
both the aims that Titus Livius proposed in the drafting of his historical work:
historiography should be both pleasant to read—after being built according to
appropriate forms of rhetoric—and encourage the recipients of the work to believe
in the great power of Roma Aeterna and its values. The oration of the General,
which, as will be explained in detail later, achieves both goals with great technical
ability, in English reads:35
V, 44
Men of Ardea, my ancient friends, and of late my fellow citizens, - since your
goodness would have it so and my own fortune has made it necessary, - let none of
you suppose me to have come forward in forgetfulness of my condition; but cir-
cumstances and our common peril oblige every man at this crisis to contribute what
he can to the general defense. And when shall I show gratitude for your great
kindness to me, if I am backward now? Or when shall you have need of me, if not in
war? ‘Twas by this art that I stood secure in my native City: unbeaten in war, I was
32
The repression of Reggio cost the tyrant both the admiration of Plato and the indignation of the
public opinion of the entire Greek world at the Olympics held in 388 BC. Dionysius received
various manifestations of hostility, for example that expressed by Lysias in his Olympic Oration by
comparing him to Artaxerxes; see Sordi (2004).
33
In relation to the Celts who had descended into Italy, Polybius speaks of ἑταιρεῖαι, i.e. military
“societies of friends” (Polybius, Histories II, 17). Archaeological research, though unable to
subtract the question from the domain of hypothesis, suggests that the members of this martial élite
had a common characteristic: that of land ownership. It may have been a small aristocracy,
uninterested in flaunting obvious differences in status (judging from the funeral accessories) and
assembled in military intertribal organizations; see Kruta (2003, pp. 241–250).
34
The traditional Roman chronology is not very accurate because it was based on a calendar that
was both lunar and solar, in use before the Julian calendar was adopted. That chronology is marred
by a fundamental error, which produces a systematic deviation of 3 to 4 years; it is based on the
time sequence of the annual honores and a sequence of dates fixed in a conventional manner at a
later date, probably following the intervention of Varro; see Ogilvie (1976, p. 166).
35
The English translation is in the Loeb Classical Library (1960), volume 3, book V, XLIV,
pp. 149–151, by Benjamin Oliver Foster, Ph.D. of Stanford University.
driven out in time of peace by the thankless citizens. But you, men of Ardea, have
now an opportunity of requiting the Roman People for such great benefits as you
yourselves are mindful of, - nor need I cast up to you things which you remember; -
and your city has an opportunity to win from our common enemy great renown in
war. That people now drawing near in loose array has been endowed by nature with
bodily size and courage, great indeed but vacillating; which is the reason that to
every conflict they bring more terror than strength. This may be seen in their defeat
of the Romans. They captured the City, which lay wide open; but a handful of men
in the Citadel and the Capitol are holding them at bay; already, oppressed by the
tedium of the siege, they are departing and roaming aimlessly through the
country-side. They greedily gorge themselves with food and wine, and when night
approaches they erect no rampart, and without pickets or sentries, throw themselves
down anywhere beside a stream, in the manner of the beasts. Just now success has
rendered them even more careless than they are wont to be. If you have a mind to
protect your city and not to suffer all this country to become Gaul, arm yourselves in
the first watch, and follow me in the force, not to a battle but a massacre. If I do not
deliver them up to you fast asleep, to be butchered like cattle, I am ready to submit at
Ardea to the same fate that I endured at Rome.
Cohesion and Coherence in Livius, Ab Urbe Condita V, 44
In this part I present a case study in order to make de Beaugrande and Dressler’s
procedural approach applicative.36 The selected text offers a sample of the per-
suasive force exerted on a political assembly that has powers of decision in matters
of war. This is the original Latin text of General Marcus Furius Camillus’ speech to
the inhabitants of Ardea:
1. “Ardeates” [inquit] “ueteres amici, noui etiam ciues mei, quando et uestrum
beneficium ita tulit et fortuna hoc eguit mea, nemo uestrum condicionis meae
oblitum me huc processisse putet; sed res ac periculum commune cogit quod
quisque possit in re trepida praesidiis in medium conferre.
2. Et quando ego uobis pro tantis uestris in me meritis gratiam referam, si nunc
cessauero? Aut ubi usus erit mei uobis, si in bello non fuerit? Hac arte in patria
steti et inuictus bello, in pace ab ingratis ciuibus pulsus sum.
3. Vobis autem, Ardeates, fortuna oblata est et pro tantis populi Romani pristinis
beneficiis quanta ipsi meministis—nec enim exprobranda ea apud memores sunt
—gratiae referendae et huic urbi decus ingens belli ex hoste communi pariendi.
36
For other case studis in classical texts, see Giuffrè (2011a, 2012).
4. Qui effuso agmine aduentant gens est cui natura corpora animosque magna
magis quam firma dederit; eo in certamen omne plus terroris quam uirium
ferunt.
5. Argumento sit clades Romana. Patentem cepere urbem: ex arce Capitolioque iis
exigua resistitur manu: iam obsidionis taedio uicti abscedunt uagique per agros
palantur.
6. Cibo uinoque raptim hausto repleti, ubi nox adpetit, prope riuos aquarum sine
munimento, sine stationibus ac custodiis passim ferarum ritu sternuntur, nunc ab
secundis rebus magis etiam solito incauti.
7. Si uobis in animo est tueri moenia uestra nec pati haec omnia Galliam fieri,
prima uigilia capite arma frequentes, me sequimini ad caedem, non ad pugnam.
Nisi uinctos somno uelut pecudes trucidandos tradidero, non recuso eundem
Ardeae rerum mearum exitum quem Romae habui.”
Text processing requires us to analyze the surface text as an organization of
grammatical dependencies; in fact, the linear sequence of a text (often) presents
ambiguities that can be removed by dealing with the textual elements as part of a
“modifier-to-head” dependence structure. Identifying this structure is therefore
fundamental and must be accomplished in order to avoid treating the individual
clauses and consequently the texts as linear sequences, treating them instead in the
perspective of labelled transition networks, which the procedural approach proposes
as model of formal representation. In these networks, the nodes represent gram-
matical states and the branches of conjunction represent the dependencies. These
dependencies make it possible to identify already used syntactical patterns and
elements that combine together and thereby modify each other reciprocally. Once
these operations have been carried out, it is possible to provide a representation of
the cohesion of the text according to the procedural approach.
After the nodes and branches of the networks have been determined and the
reused syntactic and expressive modules have been recognized, the expressions of
the surface text can be considered. These expressions are thus inputs that can
activate concepts. This should not be understood as a kind of search for terms in a
“mental dictionary”, but as the recognition in the surface text of those concepts that
contribute in a fundamental way to determining the unity and continuity of
meaning. These concepts are defined as control centres, because they represent
strategically important points where text production and reception are realized.
Therefore, the so-called control centres of transition networks are of the greatest
strategic importance for the use of texts in communications and in text processing.
They may also be called Primary Concepts and in the Introduction are sorted into a
typology. This typology is necessary to represent coherence, as it is found in our
case study.
Cohesion and Coherence in Livius, Ab Urbe Condita V, 44 91
Cohesion
Before providing a representation of the cohesion of this text, the manner in which
it will be read should be explained. The cohesion of this text was analyzed by
performing the first of two phases for each paragraph of the text. To determine the
“modifier-to-head” structure, the various paragraphs of the text under examination
were divided into clauses as indicated below. The representation of the individual
clauses identifies the “modifier-to-head” structure and it appears in Figs. 3.1, 3.2,
3.3, 3.4, 3.5 (first paragraph); 3.7, 3.8, 3.9 (second paragraph); 3.11, 3.12, 3.13,
3.14 (third paragraph); 3.16, 3.17, 3.18 (fourth paragraph); 3.20, 3.21, 3.22, 3.23,
3.24 (fifth paragraph); 3.26, 3.27, 3.28, 3.29 (sixth paragraph); 3.31, 3.32, 3.33,
3.34, 3.35, 3.36, 3.37 (seventh paragraph).37 They show, with some simplification
made for ease of reading, the labelled transition networks for each clause. Within
the network, the various elements of the text appear in different geometric shapes:
the rectangle indicates the “head” and the predicate, the ellipse indicates the
“modifier,” and the rhombus indicates the adverb. Latin prepositions appear on each
connector.
The representations of the individual clauses allow the general representation of
each paragraph to be made and all the elements necessary for decoding are listed in
the legend. These representations are in Figs. 3.6 (first paragraph); 3.10 (second
paragraph); 3.15 (third paragraph); 3.19 (fourth paragraph); 3.25 (fifth paragraph);
3.30 (sixth paragraph); 3.38 (seventh paragraph—Section “Rhetoric, Cognitive
Psychology and Ancient Literary Genres”), 3.39 (seventh paragraph—
Section “Titus Livius and the Gallic Sack of Rome”), 3.40 (general representation
of paragraph 7).38
The representations of each paragraph, in turn, allow a representation of cohe-
sion through a labelled transition network single of the overall text to be made. It
appears in Fig. 3.41.39
The clauses identified in the first paragraph are:
1. [inquit] Ardeates ueteres amici, noui etiam ciues mei, quando et uestrum ben-
eficium ita tulit et fortuna hoc eguit mea, nemo uestrum condicionis meae
oblitum me huc processisse putet; sed res ac periculum commune cogit quod
quisque possit in re trepida praesidiis in medium conferre.
• Ardeates ueteres amici etiam noui ciues mei
• quando et uestrum beneficium tulit ita et mea fortuna eguit hoc
• nemo uestrum putet me oblitum condicionis meae processisse huc
• sed res ac periculum commune cogit conferre in medium in re trepida
• quod praesidii possit quisque.
37
If you are reading this book in print, then you will see the figures in greyscale.
38
If you are reading this book in print, then you will see the figures in greyscale.
39
If you are reading this book in print, then you will see the figure in greyscale.
Fig. 3.1 Clause 1: Representation in a labelled transition network

Fig. 3.6 General representation of paragraph 1

The clauses identified in the second paragraph are:

2. Et quando ego uobis pro tantis uestris in me meritis gratiam referam, si nunc
cessauero? Aut ubi usus erit mei uobis, si in bello non fuerit? Hac arte in patria
steti et inuictus bello, in pace ab ingratis ciuibus pulsus sum.
• si nunc cessauero, et quando ego referam gratiam uobis pro uestris tantis
meritis in me?
• Aut si in bello non fuerit, ubi usus mei erit vobis?
• steti in patria hac arte et pulsus sum in pace ab ingratis ciuibus, inuictus
bello.

The clauses identified in the third paragraph are:

3. Vobis autem, Ardeates, fortuna oblata est et pro tantis populi Romani pristinis
beneficiis quanta ipsi meministis—nec enim exprobranda ea apud memores
sunt—gratiae referendae et huic urbi decus ingens belli ex hoste communi
pariendi.
• Ardeates autem fortuna est oblata vobis
• et pariendi belli decus ingens huic urbi ex hoste communi
• et referendae gratiae pro pristinis tantis beneficiis populi Romani
• quanta ipsi meministis (enim nec sunt exprobranda ea apud memores).
The clauses identified in the fourth paragraph are:
4. Qui effuso agmine aduentant gens est cui natura corpora animosque magna
magis quam firma dederit; eo in certamen omne plus terroris quam uirium
ferunt.
• Qui aduentant effuso agmine
• gens est cui natura dederit animosque corpora magna magis quam firma
• ferunt in certamen omne plus terroris quam uirium eo.
The clauses identified in the fifth paragraph are:

5. Argumento sit clades Romana. Patentem cepere urbem: ex arce Capitolioque iis
exigua resistitur manu: iam obsidionis taedio uicti abscedunt uagique per agros
palantur.
• clades Romana sit argumento
• cepere urbem patentem
Fig. 3.18 Clause 3:

Representation in a labelled
transition network
• iis resistitur exigua manu ex arceque Capitolio

• iam uicti taedio obsidionis abscedunt
• uagique per agros palantur.
The clauses identified in the sixth paragraph are:
6. Cibo uinoque raptim hausto repleti, ubi nox adpetit, prope riuos aquarum sine
munimento, sine stationibus ac custodiis passim ferarum ritu sternuntur, nunc ab
secundis rebus magis etiam solito incauti.
• ubi nox adpetit
• repleti ciboque uino hausto raptim
• sternuntur passim prope riuos aquarum ritu ferarum
• ab secundis rebus incauti nunc magis etiam solito sine munimento, sine
stationibus ac custodiis.
The clauses identified in the seventh paragraph are:
7. Si uobis in animo est tueri moenia uestra nec pati haec omnia Galliam fieri,
prima uigilia capite arma frequentes, me sequimini ad caedem, non ad pugnam.
Nisi uinctos somno uelut pecudes trucidandos tradidero, non recuso eundem
Ardeae rerum mearum exitum quem Romae habui.
Fig. 3.21 Clause 2:

transition network
• Si est uobis in animo tueri moenia uestra

• nec pati haec omnia fieri Galliam
• capite arma prima uigilia
• sequimini frequentes me ad caedem, non ad pugnam.
• Nisi tradidero trucidandos uinctos somno uelut pecudes
• non recuso eundem exitum rerum mearum Ardeae
• quem habui Romae.
The second step of our examination of cohesion is the identification of any
elements and patterns that have already been employed and which combine together
and modify each other reciprocally here in this text. As the reader can see by
observing Figs. 3.1, 3.2, 3.3, 3.4, 3.5 (first paragraph); 3.7, 3.8, 3.9 (second para-
graph); 3.11, 3.12, 3.13, 3.14 (third paragraph); 3.16, 3.17, 3.18 (fourth paragraph);
3.20, 3.21, 3.22, 3.23, 3.24 (fifth paragraph); 3.26, 3.27, 3.28, 3.29 (sixth para-
graph); 3.31, 3.32, 3.33, 3.34, 3.35, 3.36, 3.37 (seventh paragraph), many syntactic
and lexical parallels are present, which are typical patterns of ancient oratory.
Next, we provide in Fig. 3.41 the general representation of cohesion in the text
considered, which indicates all of the grammatical relationships present within the
surface text.


Fig. 3.35 Clause 5:

transition network
Fig. 3.38 Representation of paragraph 7 in a labelled transition network (Section “Rhetoric,

Cognitive Psychology and Ancient Literary Genres”)
Coherence
The grammatical relationships present in the surface text are not sufficient to
illustrate the full complexity of the meaning conveyed by the text, which is con-
structed by means of the “concepts” that are activated by it. These concepts
determine the unity and the continuity of meaning within a text, which is the
principal property that users seek in a text in order to use it. Since these concepts
(which the procedural approach defines as the control centres of the text) are sorted
in the following typology,40 we must refer to it in order to represent the coherence
in the present case study.
(A) OBJECTS: conceptual entities with a stable identity and constitution;
(B) SITUATIONS: configurations of mutually present objects in their current
states;
(C) EVENTS: occurrences that change a situation or a state within a situation;
(D) ACTIONS: events produced intentionally by an agent.
The concepts that do not belong to the group of primary concepts are placed in a
group of Secondary Concepts, which was fine tuned in the Introduction. It was
originally presented in de Beaugrande (1980), where a detailed explanation is
provided, while here we have listed only the technical aspects and a precise
explanation of each in the footnote.41 Please bear in mind that for ease of reading,
only the primary concepts are shown in the present case study.
40
See de Beaugrande and Dressler (1981), p. 59.
41
(A) STATE: the temporary, rather than characteristic, condition of an entity; (B) AGENT: the
force-possessing entity that performs an action and thus changes a situation; (C) AFFECTED
ENTITY: the entity whose situation is changed by an event or action in which it figures as neither
agent nor instrument; (D) RELATION: a residual category for incidental, detailed relationships
like ‘mother-child’, ‘boss-employee’, etc.; (E) ATTRIBUTE: the characteristic condition of an
entity (cf. “state”); (F) LOCATION: spatial location of an entity; (G) TIME: location of a situation
or an event in time; (H) MOTION: change of location; (I) INSTRUMENT: a non-intentional
object providing the means for an event; (J) FORM: shape, contours, and the like; (K) PART: a
component or segment of an entity; (L) SUBSTANCE: materials from which an entity is com-
posed; (M) CONTAINMENT: the location of one entity inside another but not as a part or
substance; (N) CAUSE: a set of relations that shows the way a situation or event affects the
conditions for other events or situations; (O) ENABLEMENT: sufficient but not necessary con-
dition for the realization of an event; (P) REASON: relationship in which a human action happens
as a rational response to some previous event; (Q) PURPOSE: situation or event which is planned
to become possible via a previous situation or event; (R) APPERCEPTION: operations of sen-
sorially endowed entities during which knowledge is integrated via sensory organs;
(S) COGNITION: storage, organization and use of knowledge by a sensorially endowed entity;
(T) EMOTION: an experientially or evaluatively non-neutral state of a sensorially endowed entity;
(U) VOLITION: activity of will or desire of a sensorially endowed entity; (V) RECOGNITION:
successful match between apperception and prior cognition; (W) COMMUNICATION: activity of
expressing and transmitting cognitions by a sensorially endowed entity; (X) POSSESSION:
relationship in which a sensorially endowed entity is believed (or believes itself) to own and
control an entity; (Y) INSTANCE: a member of a class inheriting all non-cancelled traits of the
Fig. 3.39 Representation of paragraph 7 in a labelled transition network (Section “Titus Livius
and the Gallic Sack of Rome”)
Before proceeding to the representation of the coherence of the text, it would be

best to indicate the best way to read it. The coherence of our text was analyzed by
realizing the conceptual networks appearing in Figs. 3.42 (paragraphs 1–2); 3.43
(paragraphs 3–4); 3.44 (paragraphs 5–6) and 3.45 (paragraph 7). It is to be noted
that the conceptual networks appearing in Figs. 3.42, 3.43, 3.44 and 3.45 are dif-
ferent from the transition networks appearing in Figs. 3.1, 3.2, 3.3, 3.4, 3.5, 3.7,
3.8, 3.9, 3.11, 3.12, 3.13, 3.14, 3.16, 3.17, 3.18, 3.20, 3.21, 3.22, 3.23, 3.24, 3.26,
3.27, 3.28, 3.29, 3.31, 3.32, 3.33, 3.34, 3.35, 3.36 and 3.37. The difference is that
the conceptual networks do not represent the terms contained in the surface text, but
concepts, even if they still contain Latin words. It would be better a clear differ-
entiation between transition and conceptual networks; however, it has not happened
class; (Z) SPECIFICATION: relationship between a superclass and a subclass, with a statement of
the narrower traits of the latter; (Aa) QUANTITY: a concept of number, extent, scale, or measure;
(Bb) MODALITY: a concept of necessity, probability, possibility, permissibility, obligation, or of
their opposites; (Cc) SIGNIFICANCE: a symbolic meaning assigned to an entity; (Dd) VALUE:
assignment of the worth of an entity in terms of other entities; (Ee) EQUIVALENCE: equality,
sameness, correspondence, and the like; (Ff) OPPOSITION: the converse of equivalence;
(Gg) CO-REFERENCE: relationship where different expressions activate the same text-world
entity (or configuration of entities); (Hh) RECURRENCE: the relation where the same expression
reactivates a concept, but not necessarily with the same reference to an entity, or with the same
sense (de Beaugrande and Dressler (1981), pp. 59–60).
Fig. 3.40 General representation of paragraph 7 in a labelled transition network
yet for two reasons. First of all, scholars have not a common point of view con-
cerning the best way to represent the textual meaning. Secondarily, it would be
better for the text processing to keep the formal analogies while accomplishing the
analysis at several linguistic levels.
After identifying the “modifier-to-head” structure of the various paragraphs of
the text, the various elements of the text that present a direct logical relationship
(although they may not have a direct grammatical relationship) were linked. The
representation of the coherence of the individual clauses is thus extended and
108
Fig. 3.41 Representation of cohesion in Livius, Ab urbe condita V, 44 through a labelled transition network
3 The Procedural Approach to a Text
becomes the extended transition network that meets the criteria required by the
procedural approach.
Figures 3.42 (paragraphs 1–2), 3.43 (paragraphs 3–4), 3.44 (paragraphs 5–6)
and 3.45 (paragraph 7), representing extended transition networks, show, with a bit
of simplification, a pair of labelled transition networks joined to arrive at a single
extended transition network. Each single one also shows the logical connections
between the elements of the text. The various elements of the text, divided by
category, are shown in Figs. 3.42, 3.43, 3.44 and 3.45. The figures present a
rectangle used for the “heads” and the predicates, an ellipse for the “modifiers” and
a rhombus for the adverbs.
The representations of the coherence in each paragraphs (contained in
Figs. 3.42, 3.43, 3.44 and 3.45) allow a general representation of the coherence
through an extended transition network of the full text to be made. It appears in
Fig. 3.46. By following the main connectors, the argumentative flow of the case
study can be seen. This flow adheres to the criteria of ancient genus deliberativum,
both Greek and Latin.
We provide the individual representations in the Figs. 3.42, 3.43, 3.44 and 3.45,
which after determining the logical relationships between the various textual ele-
ments allow a global representation of the coherence of the text to be realized
through an extended transition network. This representation appears in Fig. 3.46.
The representation of the coherence of the text of our case study through an
extended transition network appearing in Fig. 3.46 allows us to follow the topic
connexions of the entire speech along the path of the main connectors.
Our case study was constructed by using the expressive patterns of classical
rhetoric. A typical deliberative oration contains three partitions:
• incipit cum benevolentiae captatione;
• narratio;
• explicit cum peroratione.
It is not possible to deal with these partitions extensively in this chapter; how-
ever, it is possible to underline that the first part shows the speaker’s attempt to
empathise with the audience. The second part of the speech reveals the speaker’s
persuasive skills; he uses different topics according to the subject of the oration. The
final part contains a closed end with the prayer, reiterating and summarizing the
topics covered to the public and appealing to its benevolence.
In the speech of the General, the three partitions are as follows:
• incipit cum benevolentiae captatione (paragraphs 1–2);
• narratio (paragraphs 3–6);
• explicit cum peroratione (paragraph 7).
The criteria governing the organization of an ancient deliberative oration can be
identified in Fig. 3.47, following the geometric shapes, starting from lower left and
proceeding clockwise direction. Moreover, you have to note that, in accordance
with the compositional patterns of persuasive technique, the density of connections
Fig. 3.42 Paragraphs 1–2 representation in an extended transition network



Fig. 3.45 Paragraph 7 representation in an extended transition network
highlights the main actors of the persuasive strategy actualized by the orator. They
are namely the members of the local Senate of Ardea (the audience of the speech),
the Gauls (the enemies) and the General himself (the text producer). Each actor is
connected to specific modifiers and heads, characterizing his role.
114
Fig. 3.46 Representation of coherence in Livius, Ab urbe condita V, 44 through an extended transition network
3 The Procedural Approach to a Text
Fig. 3.47 Representation of topic connexions in Livius, Ab urbe condita V, 44 through a

conceptual network
Looking at the representation of the topic connexions in Fig. 3.47, the three
partitions can be grouped in the following order:
• incipit cum benevolentiae captatione (paragraphs 1–2);
• explicit cum peroratione (paragraph 7);
• narratio (paragraphs 3–6).
In accordance with the expressive form of classical rhetoric, the three-part
structure of the speech requires the first and last blocks to contain certain fixed
components: in the first block the speaker illustrates his situation to the audience; in
the last block he offers a perspective for the future development of current events,
proposing a specific event to be realized in the near future, according to the deci-
sions made by the deliberative assembly (this is exhortation in the strict sense). In
the central block the speaker tries to establish a certain empathy from the emotional
point of view, in order bring the audience over to his own positions (movēre). From
this point of view, the entire oration seems a typical expressive canvas of ancient
historiographical narrative: in the literature there are many speeches made by
generals and politicians exiled from the motherland who defend the motivations of
their ungrateful countrymen in front of a deliberating public assembly. This
interpretation is confirmed by the fact that within each of the three blocks that make
up the text, one of the lexemes—representing a docking station of the commu-
nicative situation in which the performance of the oral text takes place—is repeated:
it is the word “Ardea/Ardeates” (which appear respectively in paragraphs 1, 7 and
3). In this perspective, this lemma acts as an actualization of a text belonging to a
fixed expressive module. The text can thus be inserted and used within interesting
intertextual comparisons with other orations belonging to the same literary
sub-genre as General’s speech, that is, the deliberative speeches pronounced by
politicians or military figures exiled from the motherland. Following this line, the
first two blocks are to be considered fixed, while the central block of the narratio
must be readjusted each time to the present circumstance; this makes the text
attributable to the specific repertoire of the experienced politician, who when
necessary must make speeches belonging to this literary subgenre.
Conclusions of This Work
Before re-examining the key points of our reasoning, I should briefly discuss the
background of this paper. My entire work grew from a specific initial assumption.
Because we want the work to appear theoretically well-founded, I believe it is
appropriate to fully state and explain this assumption.
In artificial intelligence research, the term “Natural Text Processing” (NTP) was
coined. This terms, in opposition to the type of processing done by machines,
identifies an operating module used by humans to describe, analyze, explain and
evaluate texts. Since these objectives involve several disciplines, among which
Linguistics, Philology and Literary criticism appear most prominent, the module
must necessarily be interdisciplinary and the disciplines must produce results that
are mutually accessible and verifiable.
The description and analysis of a text fall in the domain of Linguistics, the
explanation of a text in that of Philology and the (artistic) evaluation of a text in that
of Literary Criticism. Thus, Natural Text Processing is an interdisciplinary module
involving Linguistics, Philology and Literary Criticism. These disciplines deal with
texts, and specifically those texts in which, according to both tradition and the
consensus of readers, the quality of ‘literariness’ can be recognised. This per-
spective leads us to face the problem of the definition of two concepts: ‘text’ and
‘literariness’. The first element, the text, is an object of which only an intuitive
Conclusions of This Work 117
concept is sufficiently clear. Attempts at providing a scientific definition for it

intersect different areas of research, including the Exact Sciences and the
Humanities, and therefore it is difficult to arrive at an unambiguous definition, one
that is able to satisfy the interdisciplinary interest it has aroused. As regards ‘lit-
erariness’, the second element, the difficulty is the same: how can you define the
dividing line between literature and non-literature? Scholars of Aesthetics and of
Literary Theory, especially, have laboured long and hard on this age-old problem.
Even in this case, while not wishing here to delve any deeper into the problem, it
must be said that the discipline of Literary Criticism is applicable only when the
problems of how to describe and analyze a text have been overcome.
In other words, the philologist and literary critic can operate in the moment in
which the datum, that is the starting point, is a text that possesses the quality of
literariness. But, in general, should a philologist or a literary critic need a reason or
motive to be familiar with textual linguistics and its paradigms in order to process
natural texts? Literary Criticism, which is the discipline for the evaluation of a text,
is (chronologically) the last of the disciplines of the interdisciplinary module to
participate in the process. From this point of view, the discipline that first begins to
make the module function is the one that describes the text, and the description of
the text is the purpose of textual linguistics itself.42
Having stated the issue in these terms, it becomes clear that Text Linguistics
serves as a reference discipline for Literary Criticism. Therefore, it is a sort of
Grundlagenwissenschaft of Literary Theory. The basic framework of the present
work was based on these considerations, which are also supported by many other
scholars.43 For example, the following passage from Pier Marco Bertinetto supports
this position well:
Because no one objects to the idea that the art critic deals with colours and shapes, or the
fact that the music critic has insight into notes and instruments, we should not be surprised
if a literary critic turns his attention to the substance texts are composed of: namely, to
language and grammar. But the problem, put in these brutally simple terms, is likely to
appear overly trivial. In fact, the attitude of the art critic is one thing, while that of the
decorator is something else, though he, too, is professionally involved with shapes and
colours. Similarly, the cognitive efforts of the music critic are not to be confused with those
of a man running a store selling CDs and instruments. The difference lies in the particular
task of criticism; it is an exercise whose purpose is to emit value judgements, and by its
nature it cannot be satisfied with mere descriptions (no matter how detailed) of the object
analyzed, or simple acknowledgements of existence motivated by merely declarative goals.
It follows that in literary criticism there must be an unbridgeable gap between the study of
the verbal structure of the text and the aesthetic evaluation that can be made of it. But even
this manner of approaching the problem is unsatisfactory, because in fact, between the
intellectual activity of the critic and the verbal material to which he devotes his attention
stands – whether he likes it or not – a field of independent research (linguistics), with its
own statutes and even its own methodological canons. It might be objected that musical
criticism, too, has its ancillary disciplines, such as counterpoint and harmony, which study
sound from an absolutely technical standpoint; but, in fact, it is commonly believed that
42
See also Giuffrè and Scibetta (2014).
43
See, among others, Segre (1983), p. 331.
these components are a normal part of the critic’s body of knowledge. Nobody, I think,
would be willing to consider anyone who could not read a score as a music critic. In the
case of literary criticism, however, it seems that such an assumption does not exist:
common opinion does not ever consider that the literary critic must have a knowledge of
linguistics. […] the separation between linguistics and criticism originated in the very
moment these two disciplines began to exist in their modern sense. Without this premise,
there would be no reason for such a fierce debate in recent decades about the legitimacy of
the union between these two fields of research, that contend (in part) the same objects of
analysis. […] It would be enough to surpass the barrier of romantic thought to be immersed
in a cultural climate that takes into consideration, even in classroom practice, a close
intermingling of evaluation of the work and a thorough investigation of its material char-
acteristics and this is also what the centuries-old tradition of rhetoric meant. […] the very
existence of this debate [i.e. the one between critics and linguists] (and still under way,
despite certain prejudiced stances) goes to show that many maintain faith in a recompo-
sition of critical knowledge and linguistic knowledge (or at least a significant portion of the
latter) – a recomposition that recognises the verbal materiality of the text as an absolute
priority (my translation of Bertinetto (1983), pp. 246–248).
In the representation provided by Bertinetto, many believe and trust that

Linguistic and Literary Criticism will be able to find a plane on which to confront
each other and produce mutually observable results, contributing to the definition of
a common domain for studies centred on human communication. No doubt, my
position is consistent with Bertinetto’s; indeed, if possible, it is even more radical.
I believe, in fact, that the artistic evaluation of a mental artefact (i.e. ‘text’) must
first perform a process of analytical linguistic recognition. Philology is based on
that recognition; however, it is conditioned by the artistic evaluation that is then
made of the various texts. It thus follows that the disciplines that comprise the NLP
module (Linguistics, Philology, Literary Criticism) are strongly interrelated and
provide a mutually interchangeable set of results, produced through their practical
application. Again in Bertinetto, we read that:
Of course, there was a time when the appearance of a form of “linguistic” criticism might
have alarmed some, due to its intrusiveness; but the positions were then largely reversed,
and the most glaring manifestations of intolerance now come clearly from the other side. It
will be useful to remember that the most substantial and lasting results have been achieved
by those who were able to put intolerance aside and (while not diminishing the rigour and
coherence of the method) have maintained a constant and productive dialogue with the
opposite shore (my translation of Bertinetto (1983), pp. 248–249).
We can only hope that the scientific community keeps the debate and dialogue
open between linguists, philologists and literary critics, and that their communi-
cation is ongoing and productive.
From our point of view, a reconsideration of these disciplines is necessary, and
they must be considered in the light of an integrated study of the text. In particular,
this is because a given “text object” should be considered in light of its material
transmission during a certain historical development, and its artistic evaluation, too,
cannot be carried out without a synchronic and diachronic perspective. For these
reasons, if the philologist and literary critic share the goal of building an integrated
paradigm for the scientific study of texts, they must take possession of a wealth of
linguistic knowledge that Bertinetto hoped could be re-composed.
Moreover, it should be noted that the search for an integrated perspective of this
kind was also the goal of the eminent linguist Louis Hjelmslev.
This is not surprising at all, because, as we mentioned in Chap. 1, among
European linguists he was one of the most open to text studies. Above all, he
complained that philologists were to blame if there were obstacles to the estab-
lishment of such an integrated framework:
The search for such an aggregating and integrating constancy is sure to be opposed by a
certain humanist tradition which, in various dress, has till now predominant in linguistic
science. In its typical form this humanistic tradition denies a priori the existence of con-
stancy and the legitimacy of seeking it. According to this view, humanistic, as opposed to
natural, phenomena are non-recurrent and for that very reason cannot, like natural phe-
nomena, be subjected to exact and generalizing treatment (Hjelmslev 1969, p. 11).
This work is an attempt to benefit those who are calling for a re-composition of
linguistic knowledge with philological and critical knowledge, and the inclusion of
Philology and Literary Criticism in the framework of a comprehensive integrated
study of the text, which can be carried out by eliminating the dichotomy between
synchrony and diachrony.
Now, one last observation to fully expose my point of view. The reader might
say, ok, let’s integrate Philology and Literary Criticism through the use of linguistic
models. But which models should we use in an interdisciplinary operating model
such as NLP?
Since Literary Criticism is performed on texts and not on speech, it cannot use
linguistic models per se, but must take its own models from Text Linguistics. For
this reason, in Chap. 1 I explained the genesis and development of Text Linguistics
through a discussion sub specie historiae and I described the domain of this field of
study. In Chap. 2, however, I illustrated a specific approach to text processing: the
theoretical paradigm developed by Robert Alain de Beaugrande and Wolfgang
Ulrich Dressler. It was presented both from a historical and from a technical point
of view. I argued that this theoretical proposal—known as the procedural approach
—has vast analytical capabilities. Finally, in this chapter, I applied the procedural
approach to a case study, the speech of General Marcus Furius Camillu, presented
in Titus Livius, Ab urbe condita V, 44. My goal was to illustrate the key stages of
the procedural approach for text processing.
Among the various analytical approaches developed by Text Linguistics, the
procedural approach is the best paradigm because more than others it facilitates the
direct integration of the descriptive and analytical processes operated by Text
Linguistics with the other processes of disciplines of the interdisciplinary NLP
module. Other approaches such as those of other textualists like Petőfi and van
Dijk, although they are more complex and sophisticated, may grind to a halt when
faced with the difficulties arising from their formalization. These difficulties,
especially those regarding Semiotic Textology, may appear disproportionate to the
objective pursued.44
44
I have dealt with his position in Giuffrè (2011b).
Instead, the procedural approach has the advantage of being more easily oriented
to operations and activities performed by humans. Regardless of the ways that these
units could be cognitively configured (and there may be a large number of possible
configurations), the cognitive units could be permanently stored as data.
Consequently, the various configurations should be tightly woven together or
should be reassembled on a case by case basis. In this way the textual world may be
arranged according to its cognitive content, which can then be compared to the real
world in a manner that is undoubtedly complex, but often only approximate, and in
most cases certainly far from the rigour with which textualist approaches based on
logical grounds would prefer to treat the text.
Cognitivism therefore has the advantage of focusing on the knowledge conveyed
by texts, rather than on references to extra-linguistic reality. In Text Linguistics,
rather than focusing on the statement “expressions refer to entities”, the procedural
perspective is based on the assertion “a linguistic expression activates knowledge”.
Transition networks only serve as a tool for representing what this activated
knowledge actually is.
Our book would ideally elicit the reaction of both philologists and literary critics,
because while the first show a certain technical-applicative propensity in textual
research, they rarely debate theoretical issues; the second, for their part, debate
theoretical issues while sometimes lowering their attention during the applicative
phase. Paolo Orvieto highlights the opinion of philologists with these words:
[…] if criticism was born with a parasitic (and didactic) physiognomy, as a simple tool for
the restoration, analysis and evaluation of the literary object and/or the subject-author, then,
in its libertarian frenzy, it appears to be an art itself, almost a hysteron proteron that can be
practised even in the absence of the texts themselves. Besides, those years are still fairly
recent – and I do not know if they have passed completely – when you would read the
various critics-writers (like Barthes) or the great theorists and “mythoclasts” who were not
always French (consider Eco), or the methodological-inventive ones (like Agosti) in and for
themselves, rather than as the vestibule and the initiation to the literary text. The eman-
cipation from one’s fathers – whether they be positivists, Croceans, academics and scholars,
historicists and Marxists, structural-semeiologists, etc. – was one of critics-children’s
greatest moments, and was also and above all expressed by the iconoclasm of each new
criticism: this in fact is its function, both in a social and cultural as well as an anti-Oedipal
sense. On the one hand, therefore, we find a concatenated system of methods and
counter-methods, of positions successively abandoned – which are then, at times,
re-proposed years later; on the other hand, a type of criticism that has evolved in a straight
and monochrome line, almost without history: it is philological (in a general sense) and
erudite criticism, perhaps because it is thinks it possesses the stigma of ideological neu-
trality, perhaps because it is linked to the academic world, from the Alexandrians to the
Humanists, to the erudite eighteenth-nineteenth century scholars, to the present. But, in the
end, even academics (in the sense imposed by Lanson) – some more and some less – are
philosophically (and politically) biased: they are positivists, but their ideology is frozen,
and no longer evolves: they are never à la page and, therefore, do not undergo crises due to
obsolescence. Maybe it’s true: when criticism follows itineraries of obscurantism and
sceptical relativism, then one clings to the model, ne varietur, of philology (my translation
of Orvieto (2003), p. XVI).
These words reveal a dual attitude towards Philology: it is praiseworthy because

it manages to limit the effects of obscurantism and sceptical relativism to the field of
criticism, but it is also the legacy of a frozen and neo-positivist ideology.45 With
regard to the opinion of literary critics, it is again Orvieto who highlights some
interesting aspects:
Criticism is also part of the book market: is the reader interested in – when he is not a
student – following archaeological excavations in cultures which for him are at first distant
and foreign, becoming familiar only after great effort; or is he interested in immediately
finding in criticism specular reflections of the reality in which he lives and breathes, that is,
portions of his own life? A critical edition – sometimes with hundreds of pages of notes –
has an audience of a mere handful of insiders; but other books, for example on post-colonial
criticism, have become best-sellers, and not only in Anglo-Saxon countries. Indeed, the
decision to update a text to the present time, though it may thereby satisfy the palate of a
wider audience (and the market), does have its risks, which surpass any economic
self-interest: it leads to relativism, scepticism. Relativism that Italian criticism tends […] to
limit: the reader actualizes, though with a wide range of options, structures and meanings
that are virtually inscribed in the text. It must be said that Italian criticism, viscerally
concrete and basically historical in nature, has always had for the most part a holy terror of
critics “free of bonds” (my translation of Orvieto (2003), p. XVIII).
The risk identified by Orvieto is that some philologists—there aren’t many, in

truth—in order to avoid being accused of neo-positivism, might aspire to be à la
page, and thereby fall into relativism, which easily becomes an ideological reading
of texts. By reading texts with the aid of an alchemy comprised of scalpels and
strainers, they might separate what they like from what they do not like, according
to their personal idiosyncrasies; and in doing so, true criticism would not be the
result of their work, but rather—according to Orvieto—criticism “free of bonds”.
References
Classics
Aristotle. Politeia. In Aristotelis opera. Edidit August Immanuel Bekker, Academia Regia
Harris Rackham, Fellow of Christ’s College (Cambridge), in the Loeb Classical Library
(1932).
Aristotle. Rhetorica. In Aristotelis opera. Edidit August Immanuel Bekker, Academia Regia
John Henry Freese, Fellow of St. John’s College, in the Loeb Classical Library (1947).
Cicero. Brutus. Recognovit Henrica Malcovati, Bibliotheca Scriptorum Graecorum et Romanorum
Teubneriana. Leipzig: Teubner, 1965. In the English translation by George Lincoln
Hendrickson, Fellow of Brandford College, in the Loeb Classical Library (1952).
Cicero. Tusculanae disputationes. Recognovit Max Pohlenz, Bibliotheca Scriptorum Graecorum et
Romanorum Teubneriana. Leipzig: Teubner, 1918. In the English translation by John Edward
King, Fellow of Lincoln College (Oxford), in the Loeb Classical Library (1927).
Corpus Inscriptionum Etruscarum II. 1, Fasc. 4 (Tit. 5607–6324), ed. Cristofani M. Two volumes;
I Tituli, II Tabulae et Indices. Florence: L’Impronta di Scandicci, 1970.
45
On Philology and its fundamentals, see Giuffrè (2016).
Diodorus. Bibliotheca historica. Ediderunt August Immanuel Bekker, Ludwig August Dindorf,
Friedrich Vogel, Curt Theodor Fischer, Bibliotheca Scriptorum Graecorum et Romanorum
Teubneriana. Leipzig: Teubner, 1888–1906 [rist. 1985-1991]. In the English translation by
Charles Henry Oldfather, Professor at the University of Nebraska, in the Loeb Classical Library
(1923).
Eutropius. Breviarium ab urbe condita cum versionibus Graecis et Pauli Landolfique addita
mentis. Recensuit et adnotavit Hans Droysen, Monumenta Germaniae historica, München:
Weidmann, 1978. In the English translation by John Selby Watson, Rev. of London, in the
Henry G. Bohn edition (1853).
Livius. Ab Urbe Condita Libri CXX. Recognovit Wilhelm Weissenborn, Bibliotheca Scriptorum
Graecorum et Romanorum Teubneriana. Leipzig: Teubner, 1915. In the English translation by
Benjamin Oliver Foster, Ph.D. of Stanford University, in the Loeb Classical Library (1960).
Plutarchus. Vitae parallelae. Ediderunt Claes Lindskog, Konrat Ziegler, Bibliotheca Scriptorum
Graecorum et Romanorum Teubneriana. Leipzig: Teubner, 1914. In the English translation by
Bernadotte Perrin, Professor at Yale University, in the Loeb Classical Library (1968).
Polybius. Historiae. Edidit Theodor Büttner-Wobst, Bibliotheca Scriptorum Graecorum et
Romanorum Teubneriana. Leipzig: Teubner, 1889–1905. In the English translation by
William Roger Paton, Ph.D. of Stanford University, in the Loeb Classical Library (1923).
Tacitus. Ab Excessu Divi Augusti. Edidit Erich Koestermann, Bibliotheca Scriptorum Graecorum
et Romanorum Teubneriana. Leipzig: Teubner, 1952. In the English translation by John
Jackson, translator, in the Loeb Classical Library (1937).
Modern Studies
Acquaro, Enrico. 1988. “Fenici ed Etruschi.” In I Fenici. Mostra a Palazzo Grassi di Venezia, ed.
Sabatino Moscati, 532–537. Milano: Bompiani.
Bertinetto, Pier Marco. 1983. I modelli linguistici nella critica. In Intorno alla linguistica, ed.
Cesare Segre, 246–266. Milano: Feltrinelli.
Cerami, Pietro, Alessandro Corbino, Antonio Metro, and Gianfranco Purpura. 2006. Ordinamento
costituzionale e produzione del diritto in Roma antica. Napoli: Jovene.
Chilese, Luciano. 2013. Origine del toponimo Senigallia. Senigallia: Biblioteca Comunale
“Antonelliana” di Senigallia.
de Beaugrande, Robert-Alain (ed.). 1980. Text, Discourse and Process: Toward a
Multidisciplinary Science of Texts. Norwood (New Jersey): Ablex Publishing.
de Beaugrande, Robert-Alain, and Wolfgang Ulrich Dressler. 1981. Introduction to Text
Linguistics. London, New York: Longman.
de Montmollin, Germaine. 1969. Traité de psychologie expérimentale, ed. Germaine De
Montmollin, Roger Lambert, Robert Pagès, Claude Flament and Jean Maisonneuve. Vol. IX.
Paris: Presses Universitaires de France.
Eco, Umberto. 1968. La struttura assente. Milano: Bompiani.
Éluère, Christiane. 1992. L’Europe des Celtes. Paris: Gallimard.
Giuffrè̇, Mauro. 2011a. “Demosthenes and the Greek Oral Culture in the framework of Semiotic
Textology. An application of János Sándor Petőfi’s theory to Classical Greek Literature.” In
Studies in Semiotic Textology in Honour of János S. Petőfi (=Sprachtheorie und germanistiche
Linguistik Supplement 1), ed. Mauro Giuffrè, 65–97. Münster: Nodus Publikationen.
Giuffrè, Mauro (ed.). 2011b. Studies in Semiotic Textology in Honour of János S. Petőfi
(=Sprachtheorie und Germanistische Linguistik Supplement 1). Münster: Nodus
Publikationen.
Giuffrè, Mauro. 2012. Theognis of Megara and the Divine Creating Power in the Framework of
Semiotic Textology: An Application of János Sándor Petőfi’s Theory to Archaic Greek
Modern Studies 123
Literature. Journal of Logic, Language and Information 21(3): 325–346. doi:10.1007/s10849-

011-9155-8.
Giuffrè, Mauro, and Salvatore Alessandro Scibetta. 2014. Semiotic Textology as the
“Grundlagenwissenschaft” for Philology. János Sándor Petőfi’s Scientific Inheritance.
Sprachtheorie und germanistiche Linguistik (Nodus Publikationen) 24(2): 183–210.
Giuffrè, Mauro. 2016. “Philology Within the Framework of Semiotic Textology. János S. Petőfi’s
Scientific Legacy II.” Edited by Margarita Borreguero Zuloaga and Luciano Vitacolonna.
International Workshop János S. Petőfi In Memoriam (April 23rd, 2015 Universidad
Complutense de Madrid). Cambridge Scholars Publishing. [in press].
Hjelmslev, Louis Trolle. 1969. Prolegomena to a Theory of Language. Trans. F.J. Whitfield.
Madison (Milwaukee) and London: University of Wisconsin Press.
Kruta, Venceslas. 2003. La grande storia dei Celti: la nascita, l’affermazione e la decadenza.
Roma: Newton & Compton.
Kruta, Venceslas. 2004. I Celti e il Mediterraneo. Milano: Jaca Book.
Mazza, Mario. 1966. Storia e ideologia in Livio. Roma: Bonanno.
Montanari, Franco. 1998. Storia della letteratura greca. Roma-Bari: Laterza.
Ogilvie, Robert Maxwell. 1976. Early Rome and the Etruscans. Atlantic Highlands (New Jersey):
Humanities Press.
Orvieto, Paolo. 2003. “La critica e il problema dell’interpretazione.” In Storia della letteratura
italiana, XI. La critica letteraria dal Due al Novecento, ed. Paolo Orvieto, XIII–XX. Roma:
Salerno.
Perelman, Chaim, and Lucie Olbrechts-Tyteca. 1958. Traité de l’argumentation. La nouvelle
rhétorique. Paris: Presses Universitaires de France.
Pianezzola, Emilio. 1969. Tradizione e ideologia. Livio interprete di Polibio. Patron: Bologna.
Queneau, Raymond. 1947. Exercices de style. Paris: Gallimard.
Segre, Cesare (ed.). 1983. Intorno alla linguistica. Milano: Feltrinelli.
Sordi, Marta. 2004. Il mondo greco dall’età arcaica ad Alessandro. Milano: Jaca Book.
Torelli, Mario. 2000. Gli Etruschi. Milano: Bompiani.

Text Linguistics and Classical Studies Dressler and de Beaugrande's

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Text Linguistics and Classical Studies Dressler and de Beaugrande's

Uploaded by

Copyright:

Available Formats

UNIPA Springer Series

ISSN 2366-7516 ISSN 2366-7524 (electronic)

© Springer International Publishing AG 2017

Printed on acid-free paper

This Springer imprint is published by Springer Nature

1 Alle Wege Führen Zum Text . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

Towards an empirically motivated

© Springer International Publishing AG 2017 1

A similar distinction sees ‘text’ as a notion which applies to SURFACE STRUCTURE,

The Text in the Beginning: Structural and Generative

Hjelmslev. Outside of Europe, structural linguistics developed in the course of the

In Europe, the founder of what came to be called structural linguistics was

family tradition, which contained noteworthy exponents in the ﬁelds of chemistry

Benveniste’s criticism of Saussure’s concept of ‘arbitrariness of the sign’.5 Finally,

utterance is extralinguistic, which is not necessarily always clear. Although

mind, which is a “sub-logical system” in which the principle of non-contradiction is

data that could be controlled by an outside observer» (my translation of Lepschy

ambiguities that it contains and to operate transformations on a single sentence,

With the publication of Syntactic Structures (1957),20 Chomsky introduced new

(Rate:(Fast)) (Character:(Following Y))

The Text from a Logical Point of View:

semantics of natural languages must be at least as comprehensible as a natural

Logicians tried to build a semantic system of natural languages based on extension.

Therefore, the solution to the most important problems of semantics might be

The Text in Text Linguistics43

(especially literary texts) in the context of structuralism and semiotics. As we

Text Linguistics and Its Domain

newly-formed science of textual study is the contribution provided by philology,

The Development of Text Linguistics

sentence linguistics with an expansion of the domain). A second form of text

with so-called linguistic phenomenology,56 another current of research on natural

Albaladejo, Tomaso. 1989. Retórica. Madrid: Editorial Sintesis.

Hjelmslev, Louis Trolle. 1943. Omkring sprogteoriens grundlæggelse. København: Akademisk

There is no such thing as the “philosophy” of language. There

© Springer International Publishing AG 2017 35

The Point of View of Dressler and de Beaugrande

The approach to text studies suggested by Dressler and de Beaugrande avoids

The Scope of Cognitive Science

There is no single deﬁnition of “cognitive science”.3 First of all, scholars do not

Neuroscience and Artiﬁcial Intelligence Research

Linguistic and Philosophical Research

concepts are structured in and among themselves) through its manifestation in

According to their hypothesis, the conceptual system is organized mainly in a

orientation, so deﬁned because many of them relate to spatial orientation: up-down,

The Introduction ascribes a central role to semantics in the procedural approach.

experience is irrepressibly individual, there will be an irreducibly subjective area in

The Procedural Approach

According to the procedural approach, the object being analysed—a text—is

readings of a given communication, even a simple one, would require astronomi-

Text Production and Reception

Communication functions as a constant upheaval and reestablishment of stability

The procedural approach considers a text as a system made up of a vast set of

The Concept of Textuality in the Procedural Approach:

Cohesion is achieved through various means: in units such as phrases, clauses,

determinateness operator (δ)

Intentionality and Acceptability

A third source of expectations is the technical arrangement of the elements with

The presence of at least some second-order occurrences would be the normal

According to the authors of the Introduction, the last criterion of textuality is

communication. Literary texts also contain these functions. ‘Literariness’ is a

During a conversation, either participants may violate some social conventions

inferencing and spreading activation. These constructive or reconstructive processes

Haec nostro more nos diximus, Epicurii dicunt suo;