Motivated signs and multimodal analysis in Gunther Kress’s semiotics

3.1 Context and motivated signs

I begin analysing the scene he sketched in his opening sentences:

Kress’s text is entirely in print mode, introducing a book about multimodality. That is not a problem for multimodal theory but the contrary. Kress’s description is in words, which I have yet to show are motivated signs, but their meaning, in the picture he paints, comes from motivated signs. This route to the meaning of his text can be called ‘translation’ (p. 10), a simple, powerful resource for multimodal semiotics.

This story conveys some features of the sign-maker who is creating the meanings that are carried by the book, who is at this time reading the sign he is about to analyse, a sign directing customers to Morrison’s supermarket. By focussing on this figure in this interaction (himself) he provides motivated signs for a key meaning in his theory, the role of ‘socially formed individuals’, which he elaborates later in the text (p. 10).

I make two main points from this text. Firstly, it shows the importance of context for multimodal analysis. Secondly, all details he identifies as meaningful are partly motivated signs, e.g., he goes to work by bus (ordinary member of the work force, commuting like many of his readers, an ‘everyday’ semiotician – though also a semiotician). All these signs are motivated, and also supported by convention, to use Kress’s term. That is, his readers are expected to understand his meanings because of background knowledge. Some are constrained by this conventional knowledge but others arise out of it.

He reports some signs which tell him that the makers of this sign did not create it for him or bus commuters. For instance it passes by too quickly and is at the wrong height. His interpretation draws on these signs, and on his general knowledge (‘conventional’) about the immediate environment. It has a clear semiosic effect, controlling who receives the message and who will not. In Kress’s report, these signs in effect negate this message for one class of users. It invisibly directs the semiotic traffic without most members of the public being aware of what is happening.

All these meanings refer to what has been called ‘context’. ‘Context’ is an important term in linguistics, sociology and semiotics, prominent in Halliday’s work, used freely though not technically in Kress’s work, significantly absent in Chomsky’s. I use it here as a consensus term to describe the generic relation between ‘text’, and what accompanies it, con, ‘with’.

Kress is not alone in analysing context. On the contrary, his work relies on the importance of ‘context’ established by other theorists in other disciplines. This raises complex issues of ‘originality’ and ‘interdisciplinarity’. ‘Pragmatics’, a well-recognised discipline in the Text and Talk toolkit, also does so. Kress discusses this positively (see, e.g., 2010: 56–58), but distinguishes pragmatics from multimodal semiotics (p. 58). I would not emphasise such distinctions. Pragmatics is a complex, reactive discipline (see, e.g., Jucker 2012 for a history) – a sister discipline which could change readily to incorporate analysis of motivated signs.

The potential contribution of Multimodality to this set of disciplines to show how ‘context’ as a well-recognised object for analysis is mediated by a large array of signs, many motivated to some degree. Multimodal analysis including motivated signs can enable, extend and refine a complementary analysis of context.

3.2 Composition as source of signs

The joint work of Kress and Theo Van Leeuwen on layout and composition was a breakthrough in social semiotics, which has hardly been recognised outside the field. That gives urgency to my task here, to explain how this incorporated a radically new theory of signs in general, including motivated signs. In ‘Reading Images’ Kress and Van Leeuwen (1996) combined ideas from art and the media with ideas from Michael Halliday’s linguistics and from Critical Linguistics. This work crystallized in a magisterial article (Kress and Van Leeuwen 1998), in which they analysed composition in newspapers. This project continued in the latest edition (Kress and Van Leeuwen 2021).

There is always continuity with earlier stages of thought, as was the case here with social semiotics and in Kress’s acknowledged debt to Halliday. But I argue that Kress and Van Leeuwen essentially discovered a new kind of sign which had been effectively invisible before, a new theoretical entity, in which motivated signs played a crucial role.

I illustrate the scale of this discovery by examining signs that Kress produced in Multimodality. I use Figure 1 as my text. In the beginning I attend to a set of signs associated with composition and typography, signs too humble to be included in most accounts of the meaning of print texts.

Figure 1:

First page of Kress’s Multimodality.

I begin with the first sentence. ‘1. Meaning is the issue’ is a cluster of motivated signs with many features. It is the largest in size, bolded, higher in the page, and surrounded by space. Order is also a signifier. This is the first sentence. Order is also signified by another modality. It has the number 1 attached to it, a different mode but with the same message. A further motivated sign of importance is the repetition of this sentence as a running header on all other right-hand pages of the chapter.

That is, this sentence is a site where at least six signs converge, all motivated, for most people too small and trivial to justify the term ‘sign’. Each has one similar meaning, something like ‘important’. Most of these signs are also used significantly in the immediately following text. The ‘sub-heading’ Multimodality: simple really is in bold, therefore important, but in a smaller font than the main heading, higher up than the following text, surrounded by space but less than for the main heading: less important than the main heading, more important than the main text.

Where the six signs seemed to reinforce each other in the first instance, here they produce a more nuanced meaning out of a more complex relationship between the signs. The format of the main text carries meanings by the absence of these signs, signifying it is more everyday, less important than either heading.

In addition to messages from these motivated signs, this ordinary text carries other meanings from motivated signs arising from what it is. For instance, it has the regularity of print produced by a printing process, on quality paper. This says that this is a high status text, worthy of serious attention.

These are all motivated signs. They are all also guided by conventions, which in turn shape the making of meaning, for Kress and his readers. These two kinds of sign, motivated and conventional, co-exist; they are not opposed. It is not the case that highly conventional signs – e.g., firstness signifying importance – are not also well-motivated, and the converse.

As signs, these features do not map directly on to strictly linguistic signs, understood as words. They apply to a whole sentence in the first case, to an aberrant sentence in the second, and to many sentences and paragraphs in the third. Many apply to large parts or the whole of the text. They are clearly part of the semiotic code determining meanings in the print mode, but they do not use word meanings to do so.

They all shape the text and its meanings, rather than provide what could be called ‘content’, what would be understood as its ‘meaning’, in this case Kress’s understanding of meaning. In my analysis they all have a common meaning, which could be paraphrased in the verbal code as ‘this is important/not’. Since there are at least six of them we could see this as a massive redundancy, at least six signs with one meaning. But if we look at any one of them, firstness or order for instance, we can see each of these has many meanings.

I call this the principle of multiplicity. I put this formally: In semiotic systems, every signified or meaning may be expressed through more than one signifier, and every signifier may contribute to more than one signified.

This system and its component signs, all motivated and all supported by convention, do several things. They organise the meanings of the text, saying that one thing is more important than another. They frame the meaning of the text, overall and piece by piece, fulfilling a function that Kress (2010: 10) sees as fundamental for semiotics: ‘there is no meaning without framing’. They also contribute, directly and indirectly, to truth claims for this piece of text and for the text as a whole.

I repeat my claim here that Kress and Van Leeuwen in effect discovered a class of signs which had not hitherto been recognised, combined with functions that had been marginally but inadequately recognised in other contexts, not integrated into a full semiotic or linguistic theory. These signs are clearly motivated, but also conventional. They seem to frame meanings without being meaningful themselves.

As an early step in my project to bring out the importance of motivated signs in the production of social meaning I make a preliminary claim: all semiotic modes, including all modes of verbal language, are supported by sign systems which include motivated signs at their centre.

3.3 Meanings of modality

In the previous section I claimed that Kress’s analysis with Van Leeuwen of compositional and typographic signs was a breakthrough for Multimodality. In this section I focus on the functions and meanings carried by those signs, which were also novel and not well understood.

The immediate antecedent for exploring these meanings and functions comes from ‘modality,’ a key term for Kress in his critical linguistics phase. This concept was inspired by Halliday’s (1970) seminal work in grammar about a group of words and morphemes collectively called ‘modal auxiliaries,’ such as ‘may’ and ‘can’. Halliday took the term from traditional linguistics, whose long, complex history is well described by Van Der Auwera and Aguilar (2016).

It derived from Latin modus, way or manner, a vague word referring to many different aspects of language and logic. In linguistics and logic it was the ancestor of the English word modality, which unfortunately had as many meanings as modus. Under the influence of the great German philosopher Kant’s version, modalität, the term came to refer to the major system dealing with basic categories of reality, necessity and possibility. Halliday’s term reflects this line of thought, incorporating an important dimension into a linguistics that sorely needed it, and still does.

Halliday recognised how widespread this function was, pervading very many different aspects of language. He did not label it as such, but this illustrated the Multiplicity principle, where one meaning (reality value) is carried through many signs and modes, and the same sign could often have more than one meaning.

Kress and I found this analysis brilliant and inspiring, but became disappointed that Halliday and his followers did not seem to appreciate its potential. In his 1970 article he had examined two systems, ‘modality,’ and ‘modulation’. He worried whether modality could be both, and decided not. From then on this rich ambiguity disappeared from his work (e.g., Halliday 1985; Halliday and Mathiessen 2014), and ‘modality’ largely remained in its important but restricted use.

We developed it as a rich, complex system, connected by motivated signs, though at the time we did not note that fact. We saw it as concerned with a major semiotic function, checking the truth of every text against reality, but also always serving other functions. We noted another intriguing quirk, explicable in terms of motivated signs. As Shakespeare showed with his Lady Macbeth (‘Methinks the lady doth protest too much’), over-statements of certainty can be interpreted as the opposite, motivated signs of self-doubt.

Social semiotics (Hodge and Kress 1988) applied the concept to other semiotic modes, discovering that a process that had hardly been recognised in verbal language was everywhere in different semiotic modes. However, ‘modality’ was officially retired in Kress’s latest work, replaced by ‘validity’ (Kress and Van Leeuwen 2021). Kress had many motives. One was the potential confusion with ‘Multimodality’, which used ‘modality’ in a different sense. Another was the lack of transparency of the original term, which was not an accessible term for this key idea.

In the latest edition of Reading Images (2021), Kress and Van Leeuwen explained their choice of ‘validity’. This term usefully put the crucial issue of the ongoing battle over reality back in the centre of social semiotics and multimodality. It contested the doctrine of relativism and the dismissal of ‘reality’ from semiotics, associated with Berger and Luckmann's famous slogan 'the social construction of reality' (1971) and poststructuralists such as Derrida (1988), who used Saussure’s authority to claim that ‘there is nothing outside text’.

This battle is so important that it needs clear terms, and ‘validity’ serves that function. But behind the concept that Kress earlier called ‘modality’ lay a more complex idea that should not be lost because of problems with the word. Here I invoke another theorist, Gregory Bateson, from another discipline (psychology, anthropology), known to both Halliday and Kress. Bateson (1972) proposed a model of communication which was semiotics under another name. He identified a crucial cybernetic mechanism he called a ‘metacommunication loop’, which moderated main messages along three parameters: on reality or truth; on interpersonal relationships; and on the coherence of messages.

Until his death Kress revered Halliday’s work, but from early on he departed from it in large and small ways. This complicating ambivalence creates difficulties for pinpointing Kress’s originality.

In the section that follows I explicitly go beyond what Kress actually said or wrote, or even what he would probably have accepted. Halliday proposed a tripartite structure, with three macro-functions: the ideational function, concerned with the way signs contribute to ideas about reality, the ‘interpersonal function’, the way they act on social relationships, and the textual function’, which refers to signs which organise texts into coherent wholes in their context.

Halliday assigned ‘modality’ to the ideational function. Kress believed it belonged to the interpersonal function as well. I propose that the terms of this tripartite structure can be reframed to correspond to Bateson’s model, and applied to all networks that emerged from analysis of ‘modality’. I offer this complex principle: all semiotic systems contain a set of signs which characteristically act on other signs to control three key aspects of meaning acting through them; their relation to reality, the social relations they are embedded in, and their role in constructing functioning texts.

3.4 Multimodal analysis of ‘doing theory’

The article so far has implicitly used multimodal analysis of Kress ‘doing theory’. In this section I focus more explicitly on that task. The concepts ‘meaning and ‘modality’ are key terms in Kress’s Theory as this would commonly be understood.

The normal assumption I contest is that Multimodality has nothing relevant to say about such objects and the texts that carry them. According to this assumption, Theory is produced and contested only in a monomodal verbal mode. My analysis needs to show the opposite: that all texts are irreducibly multimodal, and motivated signs play a significant role in interpreting even what theorists mean using the verbal code.

I reproduce text from Figure 1, eight words plus selected non-verbal elements.

The first phrase initially may seem a familiar object for ‘normal’ linguistics. Its meaning seems unproblematic, and can be paraphrased as ‘meaning is important’. The second phrase disturbs the apparent simplicity of meaning. It can be paraphrased as ‘multimodality is simple’, but that simplicity is complicated by the adverb ‘really’.

On the face of it, this adverb seems like a validity/modality marker which reinforces the main proposition, adding nothing to its content. However, in this context it has the effect of almost negating the adjective. Kress claims that multimodality is simple while implying that he knows his readers will not find it so. It works like an intentional version of the Lady Macbeth effect (‘Kress doth protest too much’). This interpretation relies on motivated signs. The phrase is recognised as a translation from spoken to written form of oral forms, interpreted as Kress’s response to an initial response of doubt from student readers: ‘Really?’ ‘Yes, really’.

The comma after ‘simple’ plays a slight but significant role in how the phrase is interpreted and how it relates to reality, its ‘validity/modality’. The comma is a sign in the written code, preserved in the print code, to mean a pause in speech. Both motivated signs, comma and pause, signify a break in the flow of communication, which signifies a range of things.

Linguists see two kinds of meaning here. Vocabulary is usually divided into two categories, content words like ‘meaning’ which carry meaning, and other words, like ‘the,’ often called ‘function words’. Of the five words in the first sentence, only two are ‘content words’. The other three have a modality/validity function. They have either no meaning or a kind of meaning that is hard to fix.

For instance, ‘where’ seems to refer to a location, which is not specified. But if we try to tease out a meaning about place it becomes clear it is not about an actual place. It is like a logical category, limiting the scope of the main sentence, equivalent to saying: ‘This proposition only refers to some instances …’ This logical proposition is based on/motivated by a spatial metaphor used to represent logical relationships.

‘The’ is a typical and important kind of sign within the set of modality/validity signs. I ask a probe-question: What is a the? What does it look like? The lack of an answer shows that it does not have a content-meaning. However, for Peirce (1956 [1931]: 58) it is an indexical sign, pointing to reality which is its content, while the act of pointing is itself a motivated sign, in this case signifying Kress’s urgency.

However meaningless it may seem, ‘the’ is clearly doing something. On this page alone there are 25 instances, as against 3 instances of the main content word, ‘meaning’. That ratio is typical for written English, where words from this category usually are among the 10 most frequent words in a text.

But what does this number mean? This question falls directly on a common split affecting semiotics, between ‘quantitative’ approaches, supposedly about counting things without any interest in meanings, and ‘qualitative’ approaches, which normally investigate meanings without counting things. But as Peirce (1956 [1931]: 58; 2: 298) insisted, mathematics is full of motivated signs.

I illustrate the point. ‘Meaning’ occurs 3 times in this page, more than any other content word. This fact may seem unsurprising and not especially meaningful for readers. There are not enough numbers to back that up. However, this book provides an index. This was probably not compiled by Kress, so it may not express his meanings, but the numbers communicate someone’s meanings.

‘Meaning’ has 37 entries, compared to 21 for ‘multimodality’. ‘Communication’ has 22, slightly more than ‘multimodality’, but less than ‘meaning’. Only ‘mode’ has more entries, 54, than ‘meaning’. But ‘meaning’ is listed for the first time on page 51, and there is no definition. In fact, Kress does not define ‘meaning’ or ‘multimodality’ anywhere in a book which purports to be about the second and in which the first is ‘the’ issue.

These numerically-generated meanings are all at least partly motivated signs. All perform a validity/modality function as parts of a complex evaluation of core meanings – ‘meaning’ and ‘multimodality’. I translate this meaning into words, as can always be done in multimodal analysis, but this does not detract from the motivatedness of the signs that constitute them.

In this multimodal reading ‘meaning’ is as important as the validity signs of composition and the count of items in the index says. The lack of a definition may signify that Kress found definitions difficult here. But his frequent use of the term signifies a contrary sense, that it is easily understood and used.

In this case multimodal analysis may not contribute obviously to the content, but the motivated signs provide vital evidence for what this theorist thought about his topic. Meaning and multimodality are both simple and not simple, and both are connected, to an issue that is of over-riding importance but defies an easy solution. As meta-signs do, it shows Kress evaluating his own thinking, using motivated signs that throw light on how he ‘does Theory’.

4.1 Kress on Saussure

This idea helps explain what otherwise might seem like Kress’s confusion. When Kress specifically analyses Saussure, he does not reject his core idea out of hand. On the contrary, he says, of the idea that the relation “between sound shape and meaning is an arbitrary one … that (it) seems plausible enough” (2010: 63). ‘Seems + plausible + enough’ are all validity/modality markers, motivated signs of being reasonable.

In fact he always had a complex, nuanced view of Saussure. In 1985 he distinguished the official dogmatic Saussure from a complex thinker for whom ‘there is a constant tension between these contradictory tendencies’ (1985: 86). For me, Kress and Saussure are both more admirable because of the tensions they recognised.

He also gives a complex counter-example from Wittgenstein (1935), who used an invented example of a game of chess, in which players lose one piece and agree to replace it with a supposedly arbitrary button. He agreed that Wittgenstein’s appeal to ‘convention’ was plausible, too. But he pushed back, playing Wittgenstein’s own fictional game, adding further details in which the players could have two choices – one arbitrary, the other motivated. He suggests, plausibly, that they would probably choose motivated signs, e.g., black buttons for black pieces.

Both times, Kress’s motivated signs signify compromise. He accepts with Wittgenstein that signs initially generated as motivated signs (e.g., large crowned figures as king and queen), also had conventional meaning in the game. Kress’s counter-suggestion in effect agrees with this interpretation while also insisting that the movement from motivated to conventional signs would be complemented by an impulse towards motivated signs. This is a picture in which two forces, motivation and conventionalization, are always in play.

Kress then analysed Saussure’s core argument. I revisit his argument.

1. The history of language change as meaningful is evidence against Saussure’s account of it as meaningless. Kress refers to Raymond Williams’ Keywords (1976) as example. Ironically, Williams deployed Saussure’s speciality, histories of words, on his own core vocabulary, taking for granted that these accessible histories provided insight into these contemporary motivated meanings. This practical tool would be useful for Text & Talk readers.

2. Saussure does not have an adequate theory of levels. Saussure’s mono-scalar vision of language can be seen as lacking a fundamental dimension, like a ‘flat-earth’ theory. For instance, he does not distinguish between sounds, features and extended landscapes of sound. I would argue for a more extended multiscalar analysis (Hodge 2017), than Saussure’s inadequate theory.

3. Saussure does not recognise the role of social individuals, as against a single speaking mass. This defines out of existence the role of individual agency. Kress wishes to emphasize individual agency, not deny social agency. The point of his critique of Saussure is not to propose the opposite to Saussure, but to make all positions along each continuum thinkable, from relatively free to highly constrained, as shown empirically not applied dogmatically.

4.2 A multimodal reading of Saussure

I now turn to Saussure’s own theoretical text, to see what a multimodal reading can add to complement other kinds of analysis. My corpus comes from his famous work.

Saussure’s text is in French and that is what I analyse. However, I give an English translation to indicate the point that the two languages are not incommensurably different. More important, in context typographic conventions are even more similar, and help frame the meaning of the two texts in comparable ways.

From apparently anecdotal evidence we can say that Saussure’s text was multimodal, a fusion of written and spoken modes. We can infer that Saussure started with a written text, delivered as spoken discourse, transcribed, written up and published by students. For Multimodal theory this data is not meaningless or irrelevant. It motivates a serious search to find traces of this process, even if the nature of the data resists full reconstruction.

There are some small traces of this multimodality in the French. (5) a was probably written, probably on a blackboard. (5) b is a translation into standard written French. (5) c is more expansive, still written but closer to speech.

Differences in mode have subtly different validities. The heading captures the idea in an authoritative form which does not invite discussion. The first expansion is still authoritative, but it is a full sentence which could be disagreed with. The third is still writing, but closer to Saussure thinking aloud and talking. These represent three different kinds of validity. (5) a is the official position. The third may be closer to ‘what Saussure really thought’, even if it diverges from the official position.

In this passage Saussure introduces two key terms, ‘arbitraire’ and ‘inmotivé’, and elsewhere mentions a third, ‘la convention’, which Kress prefers. I use Williams’ strategy (1985) to drill down to interrogate the meanings and histories of words. In English ‘arbitrary’ is a negative word, with a range of meanings from ‘capricious’ through ‘relative’ to ‘despotic’. Both the English and French go back to classical Latin arbiter, a judge, via 16th century Late Latin arbitrarius, relating to court cases.

Saussure the great scholar would have known this history well. It has one feature that Williams often found, an apparent sharp change in meaning, explicable in terms of, motived by, context. In mediaeval France and England judicial processes were notoriously corrupt and inefficient, twisted to serve the interests of the powerful and wealthy. ‘Arbitrary’ absorbed this ideology and its context, which Saussure picks up and transmits with his use of the word.

‘Convention’ has a less complex history, deriving from Latin con ‘with’, plus venir ‘to come’ (come together). ‘Motivated’ raises different issues. It is not presented directly as a word. It comes in the form of its negation, immotivé. Moreover, this negation is not presented directly through ‘not’, but incorporated into, blurred with, the verb.

These are all examples of transformations, a process that has played a prominent role in Kress’s theories since his early days (see Kress and Hodge 1979). It is also a major class of motivated signs, alongside negation, so important that it is a fundamental premise for all his semiotic theories.

In this early work the term was taken from Chomsky, who defined transformation as a purely grammatical operation with no effects on meaning, and changed into a semiotic tool which exposed hidden motives and meanings, especially of the powerful. Transformations are a major class of motivated sign, where each change signifies motives and forces that lead to it. In the present case, transformational analysis reveals evidence in the form of motivated signs for changes that bear on judgements about the validity of Saussure’s core claims about ‘motivation’.

There is another powerful operation at work in this passage. Saussure assembles two meanings of the linguistic sign, arbitrary and unmotivated, then declares them to be both the same. That is, he negates these differences. Again, ‘negation’ is a motivated sign.

Since good judges such as Eco (1976) claim the two words are not the same we can say that the assertion is ‘arbitrary’, in the everyday sense. However, it is not conventional, since it appears that no one before Saussure claimed the two words mean the same. Nor is it unmotivated, since Saussure makes the choice to coin the link before our eyes. Since this term normally applies to individuals, Saussure’s restriction to the action of an impersonal social mass is a personal innovation. This innovation as motivated sign contradicts his stated doctrine of the arbitrary nature of the sign, but Saussure here as elsewhere freely innovated, inventing new meanings while claiming that no one can do that.

This critique removes what has been presented as a problem for Kress, the apparent contradiction in proposing motivated and arbitrary and conventional signs. This problem is an artefact of Saussure’s arbitrary binary premises. A more open set of choices is plausible and more congenial for Kress and for multimodal semiotics. All signs are motivated, to some degree. They are also conventional, to some degree. They are also possibly arbitrary, depending on what that word means.

4.3 Motivated signs in Saussure

My multimodal reading picks up another intriguing use of motivated signs in the same argument, a page earlier (Figure 2).

Figure 2:

Saussure on the sign.

Commentators do not usually comment on the multimodality of this text. It is primarily a diagram, an icon in Peirce’s terms, and also contains words: i.e., Saussure thinks in motivated spatial signs about words. In fact Saussure has many diagrams, more than Peirce, inventor of iconic signs, more than Kress or Derrida. Saussure thought importantly in images, in motivated signs, yet his theory when expressed in words overvalued verbal signs and had an inadequate theory of these nonverbal signs.

I provisionally decode motivated signs in this image to produce a meaning. Ovals are conventional/motivated signs for self-contained wholes, bounded unities. Solidity of lines is a motivated sign for strong boundaries, including the internal dividing line. Arrows signify forces acting directionally. In this case they are equal and opposite, representing two halves acting and reacting with each other.

I call this model a ‘Saussurean dyad’. Like all readings of motivated signs this is open to different readings, but this one is surprising because his theory expressed in words seemed to propose no relationships within linguistic signs. In this diagram the only possible signifier of non-relationships seems to be the blank space between the two circles. How do we explain such contradiction?

In the accompanying text, Saussure (1983 [1917]: 78) explains that the image represents only ‘a two-sided psychological entity’, with two elements intimately linked’. In this context ‘psychological’ acts effectively as a negative. It says that this image of a highly motivated sign does not apply to actual linguistic signs. They are cut off from reality by an unstated Cartesian knife.

Variations of this image are repeated seven times in the book, signifying its importance, while Saussure has no diagram of the arbitrary sign as in the official doctrine. Did his ‘motivated’ diagram-producing left hand not know what his ‘arbitrary’ verbal hand was saying?

I turn to the words embedded in the image, beginning with the basic theoretical pair, ‘concept’ and ‘image acoustique’. Concept is motivated, derived from the Latin con, with, plus cept from Latin capere, to grasp. It is also conventional, well recognised by many generations of speakers. But image acoustique combines two known words into a single new concept. This is the same paradox we have already seen. Saussure claims that no individual can innovate in a language, but he here shows that capacity, as a socially formed agent and creative maker of language, as in Kress’s theory.

His creativity did not stop there. He proposed a new pair, signifié, ‘signified’, and signifiant, present participle, ‘signifying’, both derived from signifier ‘to signify.’ This pair of terms was invented by Saussure. The three words have a strongly motivated relationship to each other. These motivated terms have become the standard words in semiotics, not the idiosyncratic, less well motivated pair Saussure originally invented. Saussure’s example shows how motivated signs work, his own included.

In this discussion I do not criticize Saussure’s contradiction. On the contrary, my point is to show how complex, problematic and interesting his meanings are, though they have usually been taken to be parts of a considered, authoritative narrow position. My aim is to build a stronger, more applicable account of motivated signs, using as resource Kress’s battle with Saussure and Saussure’s work refracted by multimodal analysis.