Cinematic Discourse Notes 5:
Semiotic & Discoursal Connections

A) Saussure's influence

Ferdinand de Saussure regarded language as a master semiotic system. Among the concepts used in semiotic analysis influenced by Saussure are that of langue and parole: langue refers to the language system, whereas parole is the actual use of language. In the same way, film production involves a system of possibilities – the langue – whereas the actual production of a film refers to the parole: the actual realisation of the film from the system of possibilities. Thus language has a system, from which actual speech may be realised, and this is true of cinema as well.

Another set of concepts that has been influential is histoire and discours: histoire looks at the historical development of a linguistic entity, whereas discours looks at the system at a particular point or short span of time. Related terms to histoire and discours are diachronic and synchronic respectively. The kind of linguistic philological research predominant during Saussure's time looked at language from a historical, and hence diachronic angle, but Saussure emphasised that synchronic research on language should also be done, so that the language system could be more clearly understood.

Saussure's influence extends beyond linguistics. He has also been influential in literary, film, cultural studies, and semiotics in general, as can be seen in the following:

B) Peirce's influence

Symbol, icon & index

One question that should be asked in relation to semiotics in general and the use of semiotics in cinema in particular, is whether it should be based on linguistics. Another important influence on semiotics, both in general and in relation to cinematic semiotics, is the American philosopher, Charles Sanders Peirce. Peirce's approach to semiotics does not depend on language as its basis. According to him, there are three types of signs:

  1. symbol

  2. icon

  3. index

Peirce's signs have to do with the relationship between the signified (what the sign refers to), and the signifier (the representation being used in language or other semiotic systems to refer to the signified). Or, it has to do with the relationship between a signifier and another signifier.

The symbol and icon are defined by their relationship between the signified and the signifier: for a symbol, the relationship is arbitrary; for an icon, the signifier is motivated by its signified. In the case of the index, a signifier either has an inherent non-ambiguous relationship with its signified or it leads to the assumption that another signifier will follow or exists simultaneously with the earlier signifier:

  1. symbol: arbitrary relationship between signifier and signified

  2. icon: the signifier is motivated by the signified

  3. index: either an inherent relationship between the signified and signifier or a close relationship between signifiers.

The symbol

Most cultural signs are symbolic: the relationship between cultural signifiers and what they signify is usually arbitrary. Hence people belonging to different cultures may interpret a particular cultural sign differently.


Other definitions of the word ‘icon’ need to be noted here, before we go further, especially in the study of cinema, as they are different from how the word is understood in semiotics, even if there are some resemblances. A cinematic icon, for example, is a film star, or a representative or charismatic figure in the cinema who is particularly well-known, or who has a prototypical association with a genre, a period or films of a certain language (such as the icons of Hong Kong cinema, Tamil cinema etc.). Also, a religious icon is not straightforwardly iconic in the semiotic sense. The Christian cross, for example, is an iconic representation, according to the Christian faith, of the cross on which Jesus was crucified. But to many Christians, it also has a wider religious significance which extends beyond this – such as the representation of God's love and of Christ's sacrifice – which thus makes the cross symbolic as well.

As cinema is visual and its signs, arguably, are not always arbitrary, but represent what we see, it may be argued that cinema is generally iconic. For example, a woman in a film, represents, well, a woman, and the cinematic woman as a whole certainly resembles a real-life woman. The position that all cinematic signs are iconic or indexical is taken by Peter Wollen, and in fact, looking backwards, Charles W. Morris went one step further by claiming that all aesthetic signs were iconic (see my brief discussion in NT 12.5). However, although cinema is obviously more iconic than a novel, for example, the claim that all cinematic signs are iconic is not so straightforward, and needs to be qualified. The cinematic sign is more complicated, and a sign which appears to be an icon on the surface, may have features of the symbol and index as well. Moreover, ‘persons’ in cinema are complex entities, and cannot be straightforwardly classified in relation to one type of sign.


As noted in the previous sub-section, Wollen believes that cinematic signs are both iconic and indexical. It has been argued that the indexicality of many cinematic signs has a close relationship to cinematic realism. This realism is based on the assumption that the same logic that operates in the world at large, also operates in cinematic discourse. The difference between an index and icon lies in the suggestiveness of the former. So an inferred meaning or consequence, which must not be purely interpretive, is activated by a sign. Dark clouds in a film suggests, for example, that rainfall is likely, or long shadows indicate that an event in a film occurs either during sunrise or sunset. In this light, the index has a part to play in the coherence of a film, as some things are not mentioned in the film, but it is understood, as something should be present or should occur if a sign is given.

Peirce's influence on other scholars

Peirce has been a highly influential figure in general and in film studies in particular. He has influenced the thinking of scholars such as Umberto Eco and Julia Kristeva. In cinema studies, in addition to Wollen, mentioned above, the semiotics of Peirce was certainly more important than Saussure's in the thinking of the philosopher Gilles Deleuze on film.

C) Discourse

The definitions of discourse and discourse analysis have been broached at the beginning of notes no. 1. It should be clear by now that one of the important questions is whether the concept of discourse can be applied to the cinema. What we have here is not only a linguistic conception of discourse, but a non-linguistic conception of discourse as well, or discourse that does not restrict itself to language. This conception of discourse is best put by Adam Jaworski and Nikolas Coupland in their introductory chapter, “Introduction: Perspectives on Discourse Analysis,” to their Discourse Reader (p. 7):

It is worth emphasising that discourse reaches out further than language itself. When we think of discourse in the wider context of communication, we can extend its analysis to include non-linguistic semiotic systems (systems for signalling meaning), those of non-verbal and non-vocal communication which accompany or replace speech or writing (see Hodge and Kress 1991 for an overview of social semiotics). Discourse practices include the ‘embodied’ or more obviously physical systems of representation, for example performance art, sign language or, more generally, what Pierre Bourdieu has called the ‘bodily hexis’...). Other non-verbal discourse modes include painting, sculpture, photography, design, music and film...

If we look at some current definitions of discourse in the social sciences, many of them implicitly or explicitly regard the concept as trans-linguistic, and not merely confined to language. This is the case in Foucault's concept of discourse formation, which was discussed in Notes no. 2. Although one of the manifestations of a discourse formation is the statements within it that are encoded in language, this does not define it completely. Foucault's conceptions of the discourse formation and the statements that are found within it can be compared to the linguistic text and the sentences that it contains, but the former set of conceptions are not equivalent to the latter, as Foucault explains in his Archaeology of Knowledge (1969/2002: 130–131):

A statement belongs to a discursive formation as a sentence belongs to a text, and a proposition to a deductive whole. But whereas the regularity of a sentence is defined by the laws of language (langue), and that of a proposition by the laws of logic, the regularity of statements is defined by the discurisve formation itself. The fact of its belonging to a discursive formation and the laws that govern it are one and the same thing...

What is the difference between a linguistic approach and a discoursal approach to the text? Foucault's observation, made during the sixties (1969/2002: 30), when the discipline of discourse analysis had hardly established itself, was prescient, and one wishes that more linguists and discourse analysts, even today, would take note of one of the differences between the linguistic and discoursal approaches:

The question posed by language analysis of some discursive fact or other is always: according to what rules has a particular statement been made, and consequently according to what rules could other similar statements be made? The description of the events of discourse poses a quite different question: how is it that one particular statement appeared rather than another?

That linguistic analysis is based on the discovery, formulation or analysis of rules, is well recognised today, and is not only a feature of Chomsky's transformational-generative grammar and the grammars put forward by his followers. Indeed, it has an ancient history, and stretches back to the Indian linguist P??ini (circa 520 – 460 BC), who formulated 4,000 rules for the grammar of Sanskrit. Rule-following however, does not seem to be central to discourse, where many of the so-called ‘rules’ are less rigid and are prone to variation and the unpredictable influence of external factors. In fact, a different paradigm is needed in discourse analysis, and this has been well highlighted by James Paul Gee in his An Introduction to Discourse Analysis: Theory and Method. To Gee (2001: 48, 52), the conception of the mind that is needed for the analysis of discourse should view it as ‘a pattern-recognizer and builder’ and not as a rule-follower.

Coming back to the basic fact that discourse analysis must not adopt a strictly linguistic approach, or an approach that concentrates on just the language of the text, even if the object of analysis deceptively consists entirely of language, we can go back again to Foucault, who saw, more than half a century ago, the limitations of this approach. Here is his rather lyrical advice on this matter in his The Archaeology of Knowledge (1969/2002: 54):

Of course, discourses are composed of signs; but what they do is more than use these signs to designate things. It is this more that renders them irreducible to the language (langue) and to speech. it is this ‘more’ that we must reveal and describe.

Looking at or analysing this ‘more’ of the language of discourse comes naturally in the analysis of cinematic discourse. What we must bear in mind here is that even if we concentrate on the language of a particular cinematic passage, more than just language or a linguistic analysis must be brought in, in order to do a satisfactory analysis.

D) Some conceptions of discourse

Whether discourse is viewed linguistically or non-linguistically, or, more realistically, if we are to mix the two inherent features of discourse together, there are some common conceptions about what it should be. Here are some definitions of, and approaches to, discourse.

i) Intra-connected and holistic

The term discourse analysis began with Zellig Harris who perceived the importance of the analysis of connected discourse, and not just individual elements, such as words and morphemes, viewed in isolation. There must thus be some intra-connectivity between the parts of a text before it could be analysed as discourse. For the text to be analysed as discourse, it must also be viewed as a coherent whole, and not in bits and pieces, which is a tendency found in the analysis of the lower levels of language in linguistics. In a discoursal approach, even when a part of a text is selected, it is usually viewed in the context of the text as a whole, either informally, or more formally, as a constituent of the text or as a discourse move, within the context of the text.

ii) Interactive

Another attempt to look at discourse views it from an interactional perspective. Viewing discourse as essentially interactive does not mean that discourse should be restricted to the overt forms of interactive discourse, such as argument and debate. This formulation is well put by Leon James and Barbara Gordon (1978) in their document, “Applied Psycholinguistics in Social Psychology”:

...all discourse is interactional discourse, i.e., discourse is produced by more than one individual. This is of course most obvious in the study of conversation where it appears that participants take turns at talk: it is clear that the discourse visible in a transcript is interactionally produced. But this is equally true in both writing and in interior dialog where there appears to be only one person producing the discourse. However, it is a matter of common observation that writers change their discourse in response to the intended or imagined audience, showing that it too is a form of interactionally produced discourse. In self-talk (or interior dialog), the person acts as if there is an audience: reports of interior dialog produce transcript-like segments in which the person addresses himself or herself using the pronouns [I, you, we] along with the appropriate verb form, thus indicating that the discourse produced in self-talk is also a derivative form of interactional discourse.

The all-encompassing formulation here is that it is not only spoken discourse which is interactional, but written discourse as well. In fact, viewing the text as a coherent whole (the previous sub-section) usually necessitates viewing it from an interactional perspective. The written text does not exist as an autonomous entity, and the same can be said about the quasi-linguistic cinematic text. Thus the text, whether it is linguistic or non-linguistic, must be interactive, or there must be some interactional elements involved when it is viewed in its appropriate contexts, before it could be considered as discourse. But is interactivity sufficient for discourse to take place? Must communication be added to interactivity before we can view the interaction as discourse?

iii) Communicative

The idea of communication, and whether it is a necessary conception for discourse, has been briefly, but critically, discussed in notes no. 2. It can be argued, in response to the question at the end of the previous section, that interactivity alone is not sufficient, and that some kind of communication must occur before discourse can be said to have taken place. But before going to this claim, let us take a look at another contention: that communication is necessary for language. If this is true, then we can say that communication occurs in the cinematic text whenever there is language. But is this the case?

It has been argued that communication is one of the essential criteria for the definition of language, and hence it is necessary if we are considering linguistic discourse. But as Noam Chomsky has convincingly argued, in his Reflections on Language (p 69):

I can be using language in the strictest sense with no intention of communicating. Though my utterances have a definite meaning, their normal meaning, nevertheless my intentions with regard to an audience may shed no light on their meaning. But communication is only one function of language, and by no means an essential one.

Chomsky has stated elsewhere, equally convincingly, that if we say that communication is necessary for the definition of language, then we have to say ‘that bacteria have language skills. They communicate, after all.’ He elaborates:

There is a virtually religious dogma among many philosophers that the “function” of language is communication, that language evolved as a communication system, and that language would be of no value apart from communication. We know that none of that is true. Probably 99% of your use of language is internal. Is that of no value? As for evolution, there’s quite good reason to believe that communication was an ancillary function of language, matters understood quite well by leading evolutionary biologists, for which we now have substantial empirical evidence.

Let us look at the problem of communication more generally, leaving aside the question of whether communication is necessary for language. The contention that communication is needed for discourse to occur can be seen in one of the dictionary definitions of discourse, which is rooted in the etymology of the word: ‘Discourse is communication that goes back and forth (from the Latin, discursus, "running to and from")’.

However, although going back and forth may be interactive, it does not mean that communication has occurred. A game of tennis, for example is interactive, but no communication occurs between the players (unless of course they talk to each other while playing).  Communication involves the imparting of information, and we can say that broadly and minimally, there needs to be some sort of communication before discourse can arise. I am in general agreement with Robyn Penman here (in “Facework & Politeness: Multiple Goals in Courtroom Discourse,” 1990: 15):

All discourse analysis is concerned with making sense of a fundamental human phenomenon — communication. At the heart of this sense-making process is the assignment of meaning.

It needs to be added that for discourse to take place, communication need not be viewed in the narrow sense, and need not be viewed comprehensively. In other words, if X sends message 'A' to Y, the message need not be received in its complete form and understood fully, before we can analyse the text from a discoursal perspective. (For some critical views on the approach to cinematic discourse in terms of communication, see NT, chapter 11, section 11.9).

iv) Ideological

Teun van Dijk defines ideologies as ‘systems that are at the basis of the socio-political cognitions of groups’ (“Ideological Discourse Analysis”, 1995: 138). Thus the importance placed on the cognitive approach in these notes means that the conceptual prerequisites for ideological analysis have been laid down.

But do ideologies have an inherent relationship with discourse? Van Dijk reminds us elsewhere that the relationship is not exclusive: ‘ideologies are typically, though not exclusively, expressed and reproduced in discourse and communication, including non-verbal semiotic messages, such as pictures, photographs and movies’ (“Discourse Analysis as Ideology Analysis”, 1995: 17). That ideologies are also present in non-verbal discourse indicates that an ideological analysis should not restrict itself to language, or portions of a text that are encoded in language, and this is somethng to bear in mind in the analysis of cinematic discourse.

Although ideologies are not exclusively expressed in discourse, van Dijk further notes that discourse is usually the best place to present and to find ideologies: ‘among the many forms of reproduction and interaction, discourse plays a prominent role as the preferential site for the explicit, verbal formulation and the persuasive communication of ideological propositions’. Thus ideologies are omni-present in discourse, and it would be odd to simply ignore them in discourse analysis in general, and the analysis of cinematic discourse in particular, in spite of some intrinsic difficulties in the analysis of ideology in the cinematic text (see NT, chapter 11, section 11.10).

v) Rhetorical

At the onset, we can say that everything that is rhetorical is also discoursal. Thus, determining that a text is rhetorical is a surefire way of determining that it is also discoursal. But this formulation doesn't work the other way round. In other words, the discoursal need not necessarily be rhetorical. In this connection, it has been pointed out that ‘not all discourse is rhetorical’, because ‘[o]ne of the components [of rhetoric] could be missing—no audience, no exigence that can be modified by discourse, no audience that can act. Scientific and poetic discourse are not rhetorical.’ Thus if we want to discuss cinematic rhetoric (NT, chapter 11, section 11.11), it can be assumed that certain assumptions of cinematic discourse would also be involved in the discussion.

But it needs to be pointed out as well that not all aspects of cinematic rhetoric are linguistic; some aspects are visual. Non-linguistic rhetoric is not something new that came into being with cinema. As the seminal advocate of rhetorical concepts in the analysis of literature, Wayne Booth (2004: 181), has reminded us, ‘non-linguistic rhetoric long predated language’. Thus, the analysis of cinematic rhetoric does not mean that we are restricted to the analysis of the language found in film.

vi) Pragmatic

According to an oft-cited quotation by Gillian Brown and George Yule in their book, Discourse Analysis (1983: xiii),

...the analysis of discourse, is necessarily, the analysis of language in use. As such, it cannot be restricted to the descriptions of linguistic forms independent of the purposes or functions which these forms are designed to serve in human affairs

The study of language use is often taken as the definition of pragmatics. But pragmatics is not necessarily restricted to how language is used, and may function non-linguistically. This conception is also relevant to cinema, which does not consist of only language.

The pragmatics of cinema concerns the contextual or situational analysis of the cinematic text as a whole, and not only its linguistic components. This is how the pragmatic approach to cinema is described by Francesco Casetti in his Theories of Cinema, 1945-1995 (1999: 253):

Film is connected with the social environment where it appears; with the needs, habits, expectations, attitudes that characterize it; with the time and place where it is produced and projected; with the action of its promoters and beneficiaries, with their respective orientations, intentions, abilities; with the complex of the texts that accompany it, sometimes materially (as complements to the program: fliers, short films, etc.). In a word, the (filmic) text is related to its context, that is, the “surroundings” where it operates, or at least intends to.

vii) Multimodal

Discourse per se is not mono-modal, but is essentially multimodal. As Philip LeVine and Ronald Scollon (2004: 2–3) have noted (in their introductory article to Discourse and Technology: Multimodal Discourse Analysis):

...all discourse is multimodal. That is, language in use, whether this is in the form of spoken language or text, is always and inevitably constructed across multiple modes of communication, including speech and gesture not just in spoken language but through such ”contextual“ phenomena as the use of the physical spaces in which we carry out our discursive actions or the design, papers, and typography of the documents within which our texts are presented.

Thus the analysis of discourse essentially involves a multimodal analysis. Even when concentrating on just the language of a particular text, non-linguistic factors intrude, and they must be acknowledged. In this regard, Norman Fairclough has emphasised that ‘It would be quite artificial to conceive of discourse in exclusively verbal terms’ (Language and Power, 1989: 27).  In a way, the analysis of cinematic discourse is advantageous to discourse analysis, because it brings this essential feature of discourse up front; we are not fooled into thinking that we can do discourse analysis without looking at non-linguistic factors, as these factors are not only important, but they may, at times, be more important than the linguistic ones.

viii) Visual

That discourse can be non-linguistic has been implied or stated in much of the discussion on the definitions of discourse and approaches to it in the previous sub-sections. Of the non-linguistic modes of discourse, the most important is probably the visual. In this connection, Eduardo de Gregorio Godeo (2005) has observed , ‘within contemporary socio-cultural theory,... discourses often cease being practices of linguistic nature to acquire a non-linguistic nature - more often than not visual’. Visual discourse is defined by Zhou and Feiner as ‘a series of connected visual displays’. This falls in line with the definition in sub-section i above of discourse as intra-connected and holistic. Following this definition, it can be said that the cinematic text, when compared to individual photographs or single-frame paintings, is inherently discoursal. Within cinematic theory itself, there is Metz's conception of image discourse (see NT, 11.9, last paragraph), which puts an emphasis on the fact that we must go beyond language in order to analyse cinematic discourse satisfactorily.


As shown above, there is a plethora of definitions of, and approaches to, discourse. Not all possibilities have been given. Although the term discourse has been primarily used in language studies to refer to the extended inter-connected (or intra-connected) use of language, we notice that it is difficult to restrict its analysis to just the analysis of language, even if we want to confine ourselves to language. All discourse analysis is ultimately multimodal, even if the object of analysis consists entirely of language. As noted, this is evident in the analysis of cinematic discourse, which serves an important function in highlighting the importance of multimodal factors in the analysis of discourse in general.


EL3222 Home.

© 2010-2018. Ismail S Talib. Last revised: 24 September 2018.