The expansion of the musical material has been exhausted.
Adorno,
The aging of new music (1954)
Cognitive dissonance was composed between early 2012 and April 2013.
The work is dedicated to Gonzalo Alonso Hernández killed in Brazil for his engagement with environmental defense.
This work was premiered in December 2014 in THE II INTERNATIONAL CONFERENCE SOUND AND AUDIOVISUAL SPACES: SENSORIAL EXPERIMENTATION AND ACTIVE LISTENING. JIEM 2014 – 20th COMPUTER AND ELECTRONIC MUSIC. It was selected in the 2016 call for Works of the New York City Electronic Music Festival. in 2021 was selected for the epicemtrum festival, and in 2022 was selected for the Ars electronica Linz festival.
Introduction
Cognitive dissonance is a term created by the cognitive psychologist Leon Festinger in his work A theory of cognitive dissonance in 1957 [1].
The main idea is that, in the context of conflict, tensions are created. The relaxation of those tensions requires some sort of cognitive effort involving a certain type of change.
In Cognitive-Parametric Music, cognitive dissonance is interpreted in a slightly different way. The idea is to replace the concept of dissonance connected to a property of a musical material, which is defined on a parameter (usually pitch) that can be considered objective (there is a well-defined metric to measure it), with a dissonance not associated with material properties, but the context in which those materials acquire their meaning. Cognitive dissonance occurs when the listener loses the intelligibility of the sound flow. The loss of intelligibility creates a conflict that generates tension that requires rethinking the logic of the work at every moment, producing an additional cognitive effort in listening. Cognitive dissonance depends on pattern recognition, being more relative than the dissonance associated with materials because such recognition is different for each individual. The dissonance depending on the material properties generates tension by the dissonances and their resolution, the cognitive one through moments in which the same regularities are recognized but not in others. The resolution of the cognitive dissonances occurs when there is new intelligibility or sense about what is happening.
Cognitive dissonance research began in the trilogy of The Studies. The idea of the work is based on the superposition of several patterns that move at different speeds, in such a way that it is possible to recognize certain regularities at certain moments but not always. The recognition of these regularities will depend on the relationship between short-term memory and a long-term one. The alternation between moments where the information provided by the memory allows pattern recognition and others that it is impossible generates cognitive tension. Cognitive dissonance is the tension that occurs when the expectations of the speech sound are not accomplished. Cognitive dissonance moves into an ambiguous ground between order and disorder, since there is no cognitive dissonance in a total disorder.
Cognitive dissonance parts from the idea that dissonance is relative and it depends on the context. The dissonance associated with a given parameter is how it has traditionally generated the concept of musical tension. In tonal music, the dissonance is derived from the superposition of certain notes (the explanation of Helmholtz [2] is still the most commonly used, while recent researches improve it, [8]).
Helmholtz’s model sets that the dissonance between a pair of complex tones is produced by the interference between its sinusoidal components. Such interference creates beats that produce a certain roughness. Helmholtz associated this roughness with dissonance. Today the concept of dissonance is better explained by the critical band (to see it in more detail, I recommend: Plomp, Levelt [8]), (for a good explanation you can consult [9]), which although it differs from Helmholtz model, it can be considered as an improvement. Cognitive-Parametric Music is exclusively based on the critical band model.
The capture of the final mixdown of the work with pro tools
Adorno showed us about Berg´s work, (for example the violin concerto) that dissonance is relative. For example, in a non-tonal environment, a series of tonal chords may constitute dissonance. Adorno formulated a concept of dissonance that does not depend on the material, but on the context. Integral serialism and dodecaphony expanded the concept of dissonance associated with a material, with tensions associated with other parameters (such as noise associated with the timber or extreme dynamics associated with the intensity). It also happens simply, generating the concept of cognitive dissonance, although it was not explicitly defined (as quoted from Berg [3]). Enlargement of the dissonance associated with materials led to its end, generating a debate on the topic of what was called the depletion of the dialectics of material (see Adorno the aging of new music [4] and Metzger response [5] The aging of the aging of new music). Serialism extended the concept of tension to other parameters, not only to the pitches, certain dynamic, timbral and rhythmic structures generated a tension derived from the properties of the material, but not directly linked to pitches. Here, is the introduction of the noise as a generator of tension. Thus, in certain serial works, the result of the tension of the work derives from the superposition of several types of tension associated with different material properties. There is a tensioning of the pitches, another rhythmic, etc. Serialism has expanded the type of tension. Little by little, it changed from a tensioning process depending only on a property of materials, to a tensioning process dependent on several properties of materials.
For me, one of the precursors of cognitive dissonance is Luigi Nono. Especially, in his Fragmente-Stille, an Diotima String Quartet generates incredible tension, which cannot be explained by the dissonance of the material used. I.e. The objective quality of material dissonance is lower than the real tension. However, we can perform the opposite operation and check that other works that use very high doses of objectively qualified dissonant materials do not generate even remotely similar tension levels. Here, is the problem of the exhaustion of the dialectics of the material. A new objectively more dissonant material does not warrant a higher tension. There are two metrics for cognitive tension; the correlation (or lack of correlation) between tension and material objective qualities. Modern neuroscience gives us the other one since cognitive dissonance causes the intensive use of the neural circuits that manage errors. (See reference The Cognitive Neuroscience of Music Perezt, Zatorre [6]). While the traditional tension triggers another type of neuronal circuitry. But, in any case, the sense of global tension is always a mixture of both, consciousness cannot differentiate the tensioning. For example, if you have a headache for two reasons at the same time, you will only feel a headache, perhaps more intense, but you will hardly know what pain intensity is caused by what, not even whether it hurts for two different reasons.
Cognitive-Parametric Music is based on the second concept of dissonance and releases the dissonance concerning its dependence on material properties.
In my works, I have used the cognitive dissonance concept in three different ways.
Making of Cognitive Dissonance
At Endgame (1996) I got several overlapping simultaneous processes of different forms of dissonancing a material according to its parametric properties, each one in a different parameter. Generating a type of cognitive context that creates a focus on one of these processes. I got the resolutions to fall in moments of greater dissonance from the material through the use of cognitive models. This would be the embryo of Cognitive-Parametric Music (see [10]).
In the first two studies I used cognitive dissonances to structure formal sections using parametric materials, dissonances are integrated into the concept of parametric cadences, which are nothing more than strategies to provide, to the works based on parametric materials, the possibilities to resolve using cadential structures not depending on the material properties (see [11]).
In Cognitive dissonance and Parametric Formalization Studies (2011), cognitive dissonances are no longer pretended to create a cadential resolution of structural tensions moments, but, through integrational cognitive models, it is intended the association of dissonance sensation with sound pattern recognition tasks, concerning the relationship between long and short-term memories. (See [11]).
The difference between cognitive dissonance and the traditional one is very thin, in fact, it can be said that the dissonance associated with the material with the pitches (defined by Helmholtz as classical dissonance [2]) is a very simple example of cognitive dissonance that uses the cognitive model of Helmholtz [2].
Even more, from traditional classification, The weight of dissonant intervals has a certain relative character. A tritone will be consonant in music in which there are only sevenths and seconds, but dissonant in a context of thirds and fifths.
Example of dissonance associated with the material, about the same synthesis engine. Take care of your ears. Annoying controlled distortion and potentially dangerous.
By contrast, we can understand the concept through the comparison of these two videos. On the same synthesis engine, shown in min 3:30. (very similar to the one shown below), the synthesis engine of Cognitive dissonance is very similar. However, this video works around the concept of objective dissonance associated with material properties. Around 2:17 min, there is a moment of strong dissonant tension generated by a controlled distortion. Feedback with biquad filters inside the object “p live” that manages the entire live electronics. The Gaussian, which can be seen throughout the video – min. 4:22 -, is a stochastic engine, based on Gaussian distribution curves, something widely used in contemporary music, whose pioneer is Xenakis [7]. This video shows the concept of objective dissonance associated with the properties of the material.
Explanation of the work
This work researches the replacement of the concept of dissonance associated with material with the concept of dissonance associated with the cognitive tasks derived from listening. The advantage of cognitive dissonance is that the tension of the work is not based on the objective properties of the materials, but on how the material is processed. Thus, we find a way of saving the possible depletion of the dialectics of materials. Nevertheless, for the tension not to be dependent on material properties associated with a parameter, it is necessary that any sound parameter works as central in the sound stream. To avoid this centrality, I have developed several compositional techniques that avoid parametric centralities. For this purpose, I have created several algorithms in which the sound objects are treated as datum axes. The work uses multiple simultaneous datum axes, none of them determining other ones, nor working as central. For more details consult the composer´s book “Toward an Aesthetics of Cognitive-Parametric Music” in which I speak about Cognitive-Parametric Music as the set of compositional techniques for creating music without a parametric center.
The first works that explore cognitive dissonance are The studies in which the dissonance associated with material properties, and therefore to a particular parameter, is no longer possible.
This work is a superposition of several reference axes in which the parameters of the sound objects are determined by their relative position to the other ones; in such a way that any datum axis cannot be considered central. Each reference axis works as a pattern repeated at a different speed. The position of each pattern relative to each other is measured by the phase. We have a very deterministic system with cross determinations, for example, the timbre of a datum axis is determined by the pitch of another one.
Example of the eighth reference axis in one layer. Each track represents a different reference axis.
In each part of the work, there is a different number of reference axes, as well as different crossed interdeterminations. These techniques have been used for the first time in Parametric Formalization Studies, although in this work I used only two axes. in Cognitive dissonance, I have used up to ten axes and multilayered juxtapositions. In addition, certain characteristic parameters of synthesis, such as the cut-off filter frequency or envelope shape, have been used. The complexity of such formalization has been only possible thanks to its implementation on Max/Msp.
The generation of the work was divided into two distinct phases. The generation of a series of patches of Max, with different axes and different crossed interdeterminations, produced MIDI files (a form of saving events of Datum axes). MIDI events were converted into audio (interpretation of the score, seen from an ancient perspective) in another patch with the synthesis engine, which reproduced only one axis at a time. Finally, I had to do an enormous job of post-production work with Pro Tools to mix all the individual sound files.
The algorithms used to generate the MIDI files (the writing of the score) are very deterministic. They rarely use random variables. The work was intended to operate as a live electronics but the high computing requirement avoids its interpretation in real-time.
As a result, not only the generation of the score but also its interpretation has been made with” Max/Msp “. Only the post-production has been carried out with Pro Tools.
For example, here we see a note generation patch that represents a layer containing 10 simultaneous axes of reference.
The score of the work
Each one of the ten sliders allows settling the speed of pattern repetition (they can be modified at runtime execution of the patch). These sliders feed the object generafase that generates a phase between two different speeds. This phase is passed to the object devuelvexnotas that generate the MIDI events associated with a channel. This object is the only one that includes random variables for the voice conduction, from a given base note. Finally, these MIDI events are recorded as MIDI data, in the standard max object rec, and dumped into standard MIDI files which, subsequently, will be converted into audio using the synthesis engine designed in Msp for this work.
Each fragment of the work uses a different patch, similar to this one, but changes the reference axes that interact to generate the phase, their number, and the way the phase is converted into MIDI events.
Here we have another more complex patch, with a greater level of encapsulation where the wires that join the damidi18t objects between them, represent the cross interdeterminations between the reference axes. The crossed interdeterminations between the axes are matched by groups of 4 instead of two. This is only possible with a higher level of abstraction in the encapsulation of objects. This, for example, is used in the final part of the work. There are approximately 20 patches like this, each one of them with its own idiosyncrasies.
Finally, MIDI files are reproduced in this patch where they were converted into audio files through several objects representing the synthesis engine and introduced alternatively. Each reference ax from the same stratum -or layer- was usually reproduced with a slightly different synthesis engine. Each axis is represented by a midi channel.
The object phasevocodersamplerstereol~ is used in this case, but it changes for each MIDI channel. There are only 5 different types. Each one of them is equivalent to the concept of a traditional instrument. If you perceive more than five timbres is because parametric music induces you to believe so. This patch allowed me to trigger note events during the development of the synthesis engine.
In this example, we can see the skeleton of the synthesis engine, which can be clearly seen as a classic 8-voice synthesizer. The magic of each one of the voices resides in the fftvoicebus2estereo~ objects. In the picture above, you can see how a file is loaded to be analyzed by FFT in real time.
And here is the final mix screen of Pro Tools.
Each object represents the mixdown of all the reference axes of a layer and multiple layers together in premixes. Therefore, it is not possible to make analytical conclusions based on this graph. For example, the second file of the 9-10 tracks is a mixdown of 8 layers, each one containing between 5 and 12 reference axes.
2013
Last revision 3.2 (February 2020)
This work is related to “Three Chants for Computer” and “Homo Homini Lupus” so if you liked it, probably you will like them also.
References:
[1] Festinger. Theory of Cognitive Dissonance (1957).
[2] Helmholtz, H.L.F. On the Sensations of Tone as a Psychological basis for the Theory of Music. New York: Dover. (1877)
[3] Adorno, T. W Alban Berg. The master of the tiny transition.
[4] Adorno, T. W. The aging of New Music. (1954).
[5] Metzger, H. K. Das Altern der Philosophie der Neuen Musik’ (1957).
[6] Peretz, Zatorre R. J. The Cognitive Neuroscience of Music, Oxford, (2003).
[7] Xenakis, Iannis, Formalized Music. Pendragon Press. New York. (1992).
[8] Plomp, r., & Levelt, w. j. M. Tonal consonance and critical bandwidth. Journal of the Acoustical Society of America, 1965, 38, 548-560.
[9] Dowling and Harwood, Music cognition. Academic Press. Orlando Florida. ISBN 0-12-221430-7.(1986)
[10] Busevin, End of game analysis (2000). Unpublished.
[11] Busevin, Towards an Aesthetic of Cognitive-Parametric Music. ISBN 978-84-608-3022-1.(2011)
Other references
Tenney, J. A History of ‘Consonance and Dissonance‘. New York: Excelsior Music Publishing. (1988)
More information
Books
Towards an Aesthetics of Cognitive-Parametric Music
Writings
Fernando Egido 