The semiotic properties of music – a perspective from functional-cognitive semantics

1 Introduction

The issue of what meaning music conveys – if any – is an ancient topic, which continues to raise its head not only in the research literature, but also in concert presentations, reviews and teaching, illustrating that it has not yet received a consensual clarification.

This article presents a new theoretical perspective on the issue, inspired by advances in cognitive semiotics on the one hand and in functional-cognitive linguistics (my own field of expertise) on the other.

The basic question is this: What precisely is it about music that makes it meaningful to listen to?

There are a number of answers to this question, a common denominator being that music in some sense opens the door to deep feelings, or – more broadly – to vital and intensely meaningful experience. However true this is, it does not put the finger on how music manages to elicit these responses in us. Also, such answers are broad and generic; they do not tell us what the difference is between the potential of a Mozart opera and a Bruckner symphony to move us. What is it that makes us feel the rage of the Queen of the Night as a property of the music itself, not just of the stage persona, or makes us feel that the piety and awe of a Bruckner symphony is not just in the composer, but also in his music?

This is both an everyday and an academic question. In an everyday perspective, the question arises when trying to share experience after a concert performance. Typically, there is no straightforward way to achieve this successfully. There is one obstacle that is immediately apparent, i.e. the fact that there is no direct translation between linguistic and musical meaning – a feature known as “the ineffability of music”, cf. Raffman (1993) – but there is more to the problem than that.

Happily, in most everyday situations, implicit and intuitive understanding may be fully adequate for sharing meaningful experiences. But if mutual intuitive understanding fails to appear, it would be nice to be able to put your finger on discrepancies and seek to elucidate them by reference to a consensual understanding of the properties of musical meaning.

This gap between the precise and formal properties of music and its potential for eliciting deep human experience is the point of departure for the suggestions made below. The gap has produced a division that appears in many forms in the literature on music, between on the one hand authors who emphasize the importance of the formal properties and are reluctant to let the fuller panoply of music’s evocative potential into their analysis, and on the other hand authors who freely analyse pieces of music in terms of the structures of feeling and experience they evoke.

It should be emphasized at this point that the division is not as clear cut as it may appear. The existence of both sides of the issue is generally taken for granted. Also, even the most formally oriented authors typically include experiential or ‘affective’ aspects of music in their account (cf. Jackendoff 2009; Lerdahl and Jackendoff 1983). The problem is precisely how to link the two sides up. Taking my point at the formal end, I try to offer some suggestions, inspired by the type of approach to language I practice, for how we can get closer to the experiential potential.

From cognitive semiotics it draws on the semiotic hierarchy (cf. Giacosa 2023; Zlatev 2018), which offers a richer and more comprehensive approach to meaning than found in the linguistic tradition.

From cognitive-functional linguistics (cf., e.g., Engberg-Pedersen et al. 2019; Harder 2016), it draws on an ontology of meaning that links up formal, sociocultural and individual aspects of the process of conveying meaning and in doing so offers suggestions about how they hang together. Two interconnected shifts in perspective are central:

The first involves a differentiation into three distinct, but interdependent locations of meaning. In other words, meaning is not a monolithic property, but a complex formation that is inherently dependent on factors in three different locations. These are (1) actual events, (2) mental architecture (here called ‘competency’, with a ‘-y’) and (3) the social community, manifested in the set of contextual options (‘affordances’) that are available for the understanding of meaningful input. All entrenched social activities, including language and music, share this tripartite existence.

The most basic ontological site is events. Without actual events of linguistic communication, language could not exist. Without events of musical performance-and-appreciation, music could not exist. To avoid misunderstanding: the term events as used in this connection does not refer to ‘events in general’ - it refers specifically to events consisting of actual occurrence of linguistic communication or musical performance-and-appreciation.

But such events do not exist in a vacuum. They can only occur if human beings with appropriate cognitive prerequisites (‘mental architecture’) are available, cf. (2) above.

Further, while ‘raw’ events of communication and appreciation of sound can occur without pre-existing competencies or options, human language and music cannot be understood without reference to pre-existing cultural patterns, cf. (3) above. A language like Danish exists in the collective culture, not as an anatomical feature of an individual brain. These patterns have grown out of actual events, but once they have sedimented into established meaning-conveying options, they have a form of existence that is not reducible to individual minds or events.

The second point involves a principled differentiation of the conveyance of meaning into two stages of a semiotic event: the input stage, consisting of the properties of the signs, and the output stage, consisting of the full interpretation. The input-stage properties are those derived from the pre-existing options, while the full interpretation arises in actual events of communication. In language, the difference is between the meanings associated with the linguistic string, and the meaning conveyed to the addressee. These two cannot be fused into one, and languages like Danish have separate words for them (betydning and mening).

The point of this article is an attempt to show how the concept of ‘music in itself’ can be analysed in a way that understands it as carrying meaning. Importantly, this does not imply that meaning in music is the same thing as meaning in language (more on this in Section 3). The semiotic hierarchy provides a framework for showing ways in which music ‘in itself’ (inherently) shares meaningful properties with basic human experience. These shared properties are proposed as the basis for going beyond musical form in analysing ‘music-in-itself’.

However, the argument goes on to claim that the meaning that is inherent in ‘music-in-itself’ does not capture the full richness of musical meaning. This is due to a complexity in the ontology of meaning, which – in language as well as in music – inherently presupposes that ‘input-level meaning’ is enriched and brought to completion by the participants in the actual event.

With an example: Just as a description of the lexical entry for the word daddy is incomplete as an analysis of the meaning of an actual utterance Daddy!, the abstract musical properties of a C major chord are incomplete as a description of its contribution to an actual musical experience. And just as actual situated understanding is the point of having a language, actual musical experience is the point of having music.

The interplay between analyses focusing on formal structure and analyses focusing on conveyed content (e.g., emotional aspects) is a central theme in musical semiotics (cf. Cook 1996; Tiits 1995), The split between focusing on form and meaning is familiar also in the linguistic tradition. The linguistic framework on which this paper is based, however, integrates meaning and structure, seeing structures as meaning-bearing.

The argument is structured as follows: In Sections 2 and 3, I introduce the features from functional-cognitive linguistics that I hope to show have revealing implications for the analysis of musical meaning. In Section 4, I introduce the features of the semiotic hierarchy that are essential for my purposes. In Section 5 I give a minimal catalogue of types of musical meaning, divided into four subsections. Section 6 sums up the conclusions.

2 The complex ontology of language and music: events, competencies and cultural options

In functional-cognitive linguistics the basic ontology of the domain of language takes a different form than in mainstream linguistics. The dichotomy that pervades the linguistic tradition distinguishes between a level of actual events of language use and a level of underlying structure. Instead, the ontology proposed here (cf. Harder 2010: 174, 2016)) is based on a division between three closely related but distinct objects of description: actual events, cognitive competencies, and the system of options (all of which are structured).

I regard this division as sort of obvious, at least as far as events and competencies are concerned.

The system of options requires a little more justification. A description of it answers the question: What linguistic signs, including mechanisms for sign combination, are available in the community? Languages as sets of options may offer (for instance) choices between more or less polite forms of address, between past and present tense (indicating whether you are talking about what is the case now or what was the case then), between written and spoken utterances (etc., etc.). This mode of being is shared with other social institutions, such as education systems or legal systems.

This tripartite ontology, I hope to show, can throw light also on some of the complexities of musical meaning. As in language, actual ‘usage’ events (=performances) are fundamental: actual experiences of music are the basis on which all else rests. In language as well as in music, such events depend on the cognitive competencies of participating individuals to experience such sequences as meaningful events (musical events could not exist if human beings were devoid of hearing). And finally, what distinguishes music from other meaningful experiences of sound sequences (such as the stream of sounds one may experience while walking through a rainforest or a city) is the existence of a system of musical options with non-random relations between them, out of which the composer’s choices in individual music events constitute a selection. I base the account on classical tonal music, where the presence of a pre-existing set of options may be more obvious than for other types of music.

In music, however, it has a different basis than in language, where the options depend entirely on conventions in the community. Convention has a role in classical music, but it is not constitutive of musical meaning, as it is in the case of linguistic meaning. The system consists of the set of musical options in the community, with both cultural and natural elements: tones, rhythmic patterns, harmonic intervals, the major and the minor key, alternative choices of instrumentation, etc., (cf. also Lerdahl and Jackendoff 1983).

3 ‘Input’ and meaning construction in music and language

An important observation about music is that its meaning is ‘unconsummated’, in the phrase of Susanne Langer (1967): Rather than being fully specifiable, the meaning of a piece of music is open to interpretive options.^[1]

Langer contrasts this with the ability of linguistic meaning to convey ‘consummated’ propositional meaning. But in fact, linguistic meaning is not as complete as Langer assumes. As generally assumed in the functional-cognitive tradition, conventional meaning does not take the addressee all the way to a precise, propositional message content. Rather, linguistic meaning as specified at the level of the pre-existing pattern must be understood as ‘input’ to the construction of precise communicated meaning. Meaning in the pre-existing system of options is in principle underspecified in relation to meaning as part of actual communication. I have used the metaphor of ‘recipe’ about encoded meaning (Harder 1996: 214): analogously to a cook, the addressee needs to carry out the instructions encoded in the linguistic ‘recipe’ in order to arrive at the intended output. The whole point is in the output, not in the recipe: in the dish that is served and in the full situational understanding.

Importantly, this emphasis on interpretation by an addressee is not an observation limited to art, as in literary interpretation. It is a quintessential aspect of the whole mode of being of meaning in language – a presupposed aspect of that semiotic whole of which abstracted input meaning is also one part. To illustrate the degree to which meaning construction is inherent rather than external to meaning in language, consider the function of using language to refer to objects in the world.

This function has been at the centre of the understanding of meaning in language from classical antiquity onwards. For that reason, it has typically been assumed that reference to the world was built into language itself, in contrast to contextual, pragmatic factors. However, from a functional-cognitive point of view it is clear that the mechanism by which language comes to stand for things we refer to crucially depends on a process of ‘meaning construction’ by the addressee, taking the step from linguistically encoded meaning (‘betydning’) to message meaning (‘mening’). When a speaker asks, e.g., can I borrow the book?, the request cannot be understood unless the addressee makes the connection between the linguistic phrase the book and the actual book in question. For obvious reasons, language as a set of pre-existing options cannot accomplish this task on its own. Between conventional linguistic meaning and ‘consummated’ message meanings, a process of ‘meaning construction’ is a built-in prerequisite in order for linguistic communication, also in this heartland of traditional referential semantics, to be possible at all (cf. Harder 2009).

This two-way ontology is applicable to music, too. It is true that musical meaning is unconsummated – in precisely the same sense that linguistic meaning is unconsummated. The ‘recipe’ metaphor allows us to agree with Langer and Stravinsky that music – in abstraction from actual events of performance-and-listening - does not have ‘consummated’, specified meanings such as definite feelings or attitudes. But at the same time this places the input-level properties as merely the initial stage of the whole story of musical meaning. For exactly the same reason as in language, music would be meaningless if the abstract, input-level meaning was left hanging, without being brought to ‘consummation’ by the audience.

It may appear that this account presents meaning in language and music as too similar – especially when it comes to the distinction between meaning viewed as part of a pre-existing system as opposed to the ‘meaningfulness’ of situated musical experience. While it is arguably plausible to see music as a meaningful experience, it is less obvious to postulate a system of musical meanings that are analogous to meanings in a linguistic system. Let me therefore make clear that musical meaning differs in fundamental ways from the conventional, symbolic-conceptual meaning that is the prototype of linguistic meaning. Musical meanings are not arbitrary, they do not refer to aspects of extra-musical reality, and do not form a system that is separable from the expression side of music. Nevertheless, I argue that it makes sense to postulate a meaning potential for music that can be tapped by composers in advance of actual performances. What a piece of music conveys to an audience depends in systematic ways on the choices made by the composer – even though this is not all there is to say.

It may be helpful to point to properties of language that have some of the same features as music, intonation being an example: Intonation does not convey symbolic-conceptual categories, has features such as pitch and rhythm, and yet contributes essential features to the messages conveyed by spoken utterances. It would take us too far to pursue the parallel here; the idea is just to suggest that the possibility of operating with a potential that exists in advance of actual usage or performance events is not limited to the classic prototype area of linguistic semantics. Below, I try to provide an outline of what this potential consists of for music.

4 The semiotic hierarchy

So what is the nature of the input-level meaning that an audience has to base its understanding on?

In a semiotic perspective, this question refers to the nature of the sign meanings that feed into the process of musical appreciation. The semiotic hierarchy is helpful because it broadens the analysis of meaning in mainstream cognitive-linguistic semantics in two ways:

Instead of being based on the level of cognition and conceptualization, it places the analysis of meaning more broadly in the whole process of life.

Instead of predicating the analysis on a unified view of meaning, it offers a multilayered analysis.

The hierarchical aspect provides a directionality from a basic layer of ‘life’ to successively more differentiated and sophisticated levels of meaning. The multilayered approach allows the analyst to differentiate between contributions to the total meaning belonging at different layers beyond the basic level.

The Zlatev-Giacosa version includes five layers: Life, subjectivity, intersubjectivity, sign function and language. The relationship between the layers is conceived in terms of Fundierung such that “the lower level both provides the ground for the higher and is ‘sublimated’ by it” (Zlatev 2018: 5, quoted by Giacosa 2023: 19). Fundierung has parallels with the cognitive-linguistic notion of ‘grounding’ (Harder 2010: 79).

The semiotic hierarchy embodies the insight that signs stand on the shoulders of phenomena that are not signs: Signs presuppose intersubjectivity (without awareness of an addressee, signs would be meaningless); intersubjectivity presupposes subjectivity (without subjective awareness, there would be nowhere to put awareness of an addressee), which again presupposes life (apart from science fiction).

It may appear surprising that it is only at the second highest level of the semiotic hierarchy that we find the sign. This may appear to contradict the status of lower levels as part of an overall semiotic whole: How can they be semiotic, if they fail to reach the level of the sign?

The answer is in the Fundierung relationship between the layers. Since the hierarchy is understood as a stepwise build-up of meaning towards a complex process involving semiosis, meaning has to be available before it can be conveyed by signs. (In this, it contrasts with a purely structural-semantic analysis, where the content plane is defined only in relation to the expression plane). The role of a specific layer of the sign is to indicate the stage at which expressions arise as independent vehicles of meaning, on top of the previous layers of meaning that are ‘submerged’ in human experience.

The primordial step towards semiosis can be captured through an analytic level that is more basic than the rise of the division into index, icon and symbol, the three basic Peircean sign types. The level is constituted by what, with a symbolic homage to Peircean terminology, may be called an ‘identisign’.

An identisign is one in which a ‘brute’ phenomenon achieves the superimposed function of conveying information about itself. As an example, when I wake up in the morning, blue skies may make me aware of blue skies: a ‘brute’ object functioning as a signans with itself as signatum. This, I argue, is the basis for the semiotic status of musical meaning. The step towards ‘sign’ status is implied in the idea of a performance before an audience, where music is cast as the conveyor (a ‘vehicle’) of a meaningful experience, not just as a sequence of sounds.

The point of the ‘identisign’ concept is that it offers a perspective on the perennial issue of whether music means anything, i.e. whether it has sign status – or music just ‘is’ music. Stravinsky (1935: 53) is perhaps the most strident voice in favour of the latter position:

… music is, by its very nature, essentially powerless to express anything at all, whether a feeling, an attitude of mind, a psychological mood, a phenomenon of nature, etc….Expression has never been an inherent property of music. That is by no means the purpose of its existence.

In 1959, Stravinsky elaborated on his earlier dictum: Given a chance to repeat himself, he said, he would rephrase the remark; it was not so much that music is ‘powerless to express anything’, he explained, but that ‘music expresses itself’.

In terms of the argument of this paper for musical meaning, the claim is that we can accommodate what I see as the core of Stravinsky’s position into an account of the expressive powers of music by understanding music as an ‘identisign’. Music can be said to convey meaning, even on Stravinsky’s terms – to the extent that there is something inherent in music that is meaningful ‘in itself’. This is not all there is to say about musical meaning (cf. Section 5.3 below), but I see it as the foundation on which all else rests.

The identity between signans and signatum in music has been expressed in different ways in the music literature. The analysis proposed here differs from one implication that is sometimes drawn, and which is also a potential interpretation of Stravinsky as quoted above: that this reflects an inward-pointing ‘art-for arts’-sake’ aspect. The point here is the opposite: There are properties of music which make themselves felt by listeners as vital aspects of human experience, precisely because music makes them directly available.

Thus there is a clear difference between the ‘identisign’ analysis proposed here and an analysis of music as ‘self-referential’ – cf., e.g., the discussion in Vuust and Roepstorff (2008: 136), which also points to the limits to self-referentiality in music. One reason it is mistaken to suggest that music refers to itself is that musical signs do not refer at all.^[2] Rather, the input-level meanings of musical signs consist in drawing on aspects of human experience, not in order to refer, but in order to use them as constituents of new and composite experiences which ideally have never been heard before (and thus cannot be referred to).

5 The meanings of music: a minimal catalogue

6 Conclusions

The brief catalogue of musical meanings discussed above will hardly surprise a musical audience; the point in this article is to view it in the light of the ontology proposed above, especially the distinction between input meaning and full musical experience. The main features of the analysis are the following.

The key point of the ‘identisign’ analysis was that music inherently shares features with human life. Those features can be captured in the basic layers of the semiotic hierarchy. A distinction was made between features that are fully inherited from basic processes of life (e.g., the heart rhythm), and features that take processes of lived experience as their starting point and ‘sublimate’ them (cf. (Zlatev 2018: 5) by imposing new structures upon them.

The analogy with language is predicated on the illustration example of reference: on the one hand, abstract linguistic meaning cannot inherently link up with a referent – on the other hand, the whole point of a referential expression is to convey precisely such a link. What I suggest, in other words, is that the identities between life and music described above inherently serve as invitations (or ‘instructions’) for a construction process that generates a full musical meaning – which is thus a presupposed, constitutive part of music.

Ultimately, the input-level options invoked by the composer are in two ways ontologically embedded in full musical experience: full musical experiences are the background from which they are selected for use by the composer – and they are prompts for the audience to produce a new musical experience. So full musical experience is truly alpha and omega in the process.

This also throws light on the status of emotions in musical meaning. As argued in Langer (1967), to say that a particular sequence encodes an emotion such as ‘love’ or ‘anger’ runs counter to the nature of musical meaning. On the other hand, it is hardly conceivable to mobilize an adequate experiential response to a piece of music with input properties such as (dis)harmony, intensity, descent or ascent (etc.) without also invoking emotions. Properties such as intensity, anticipation and climax are dependent on there being something that is intense, is anticipated or reaches a climax – and this has to be supplied by the audience.

A natural question for the reader to ask at this point would be: How exactly does this complicated theory of the nature of musical meaning differ from what we already knew? The traditional, well-established view I take to be the one expressed by Rasmussen (2011: 41), who after quoting Celibidache on emotions in Mahler’s music, interrupts himself and goes on: But I apologize! Music can resemble emotions, it can conjure up emotions, but music does not “speak” in emotions, it speaks in tones. As pointed out by Jørgen I. Jensen (2024: 248f), a similar dilemma can be pointed out in relation to Carl Nielsen, who on the one hand clearly distanced himself from attributing ’concrete or positive thought content’ to music but on the other hand in his titles invoked meanings that might be understood as programmatic.

The attempt to say something new in this article can also be described as a contribution to the understanding of how music can convey more than strictly formal properties, without having to jump across the chasm to the full experiential side. It may also be understood as an attempt to expand the understanding of what music is.

In the quote from Rasmussen above, music refers to the input only, while the emotional response is in the audience alone and external to music. I have suggested that a fuller understanding requires us to see the tonal input as a constituent of a larger process, something that cannot be rightly understood on its own: music is ontologically dependent on the response side as well as on the input side.

The crux of the inspiration from language is in the role of a linguistic utterance as something that can only be understood as carried out by a speaker and a hearer acting together. In Austin’s words, without uptake no speech act can be performed (Austin 1975: 116–117).

For music, this involves a claim that the music must be understood as having a range of meaning-bearing properties, already at the level of the strictly musical ‘artifact’ – but that these properties have the constitutive role of inviting the audience to carry out their necessary part in creating the full musical experience that is the whole point. Put differently: just like language, music exists in the relation between an input and an audience.