The semiotic model described by C.S. Peirce has been adapted to film and media generally and more recently to sound analysis, but seldom has it been applied to the actual practice of sound design. This paper argues that the model provides not only a comprehensive and powerful tool for the analysis of sound but may also be used as a tool to examine and inform the practice of sound design and production.

By focussing particularly on Peirce’s later model of semiotics, which clarifies and reframes the definitions of some of the principal elements of his system, its utility in application to the analysis of sound is greatly enhanced. It provides an appropriate and flexible conceptual framework which can be applied to the processes involved in creating elements of the soundtrack – the practice of sound design – and a means to understand how the sounds themselves, in conjunction with the image, can be used to create meaning for the audience.

Introduction - Semiotics and Sound

Theo van Leeuwen’s (Van Leeuwen 1999) adaptation of Halliday’s social semiotics is perhaps the preeminent implementation of Saussurean semiotics to sound. In Speech, Music, Sound, the stated aim is to “explore the common ground between speech, music and other sounds” (Van Leeuwen 1999: 1), without using the language of mainstream musicology and linguistics. However, van Leeuwen accepts that using this method to do so also proves difficult:

A semiotics of sound should therefore not take the form of a code book, but the form of an annotated catalogue of the sound treasures Western culture has collected over the years, together with the present experience. A semiotics of sound should describe sound as a semiotics resource offering its users a rich array of semiotics choices, not as a rule book telling you what to do, or how to use sound ‘correctly’. (Van Leeuwen 1999: 6)

Whilst an "annotated catalogue" is useful, it does not take us much farther down the road towards understanding how we understand how sound works in practice or the decisions that practitioners take to create soundtracks. A "code book" need not be prescriptive, but may provide both a language and range of concepts that can be applied to any use of sound. Van Leeuwen’s approach is both original and broad-ranging, but is not sufficiently specific or explicit enough to be easily adapted for an analysis of film sound theory or practice. What is required instead is a system that can be used for the analysis of the soundtrack elements both individually and in combination with other elements, both sound and picture.

A number of film theorists, most notably Christian Metz (Metz 1974a; Metz 1974b), have attempted to apply the linguistic model of semiotics to film, given its apparent similarities with language, to create a study of cinesemiotics. Its linearity, its paradigmatic and syntagmatic structure, and, perhaps most importantly, the ability for film to be reduced to scenes, sequences, shots and individual frames seems to imply a linguistic model and structure. Unfortunately, this reduction of elements into ever smaller subdivisions is not possible with sound. There is no sonic equivalent to a single frame of film. The construction of the soundtrack is largely concerned with creating meaning or the conditions for the creation of meaning by the audience. Whilst the Saussurean model or variations of it have been adopted by a range of theorists, its applicability for an analysis of sound remains elusive. Several film theorists adopted the model of C.S. Peirce in favour of the linguistically-oriented Saussurian model (Deleuze 1986; Deleuze 1989; Silverman 1983), with Wollen (Wollen 1998: 166) being particularly influential in moving away from a Saussurian conception of language to the model proposed by Peirce.

For Silverman, Peirce’s account of realism, a fundamentally important aspect of cinematic representation, is a crucial aspect of the semiotic model. For Peirce, since reality is "provisional", the truth of a representation can be judged by three criteria: its insistence, its recognition by others, or by induction (Silverman 1983). In terms of the soundtrack, Silverman appears not to apply the full breadth of the Peircean model, instead claiming that, “The sound track, exclusive of music, is primarily iconic, simulating the noises of speech, sirens, horns, screams, doors opening and closing, birds, barking dogs, etc... However, because these sounds often alert us to unsuspected or as yet unseen occurrences and objects, they also participate indexically” (Silverman 1983: 23).  Although the concepts of iconicity and indexicality are mentioned specifically, this analysis of sound does not sufficiently take into account the actual practice: the manipulation of sounds, their recording and highlighting, symbolic aspects, fabricated indexicality, and therefore the manufacture of realism.