Improvisation is ever-present in both performance and composition and is realized to certain degrees in all musical activity. In this chapter, the different scales of improvisation inherent to different ways of making music will be discussed in detail, drawing a picture of a continuum of activities ranging from those that do not involve very much improvisation to others that are very improvisatory in nature. This discussion is organized around an ordered list containing various degrees of improvisation originally presented in Benson's The Improvisation of Musical Dialogue (2003). The examples he provides will be analyzed and contextualized using music by me as well as others. The goal of this chapter is to dispel the myth that improvisation is “something created on the spur of the moment” and to depict it rather more within a continuum of reworking and recontextualizing.


Making things up as you go along?


- Schwankung am Clavier -

[- undulation at the piano -]

(Schumann, Tagebücher Vol. 1, 411; my translation)


What Schumann calls a Schwankung is typically called an improvisation today. Many other terms can be used, such as extemporization or fantasieren in the German language, and certain colloquial phrases describe the phenomenon as well: “to make things up as you go along,” “on the spot,” or “on the spur of the moment.” Considering that "in music, improvisation marks the simultaneous conception and presentation of art" (Landgraf 2011, 16; author's own emphasis), more modern terms such as "real-time composition" and "instant composing" emphasize the conceptual rather than the presentational aspect of improvisation  consider improvisation scholar Simon Rose and saxophonist-improviser-composer Raymond MacDonald's observation of how "real-time composition leads to a particular focus on listening, through the embodied process of improvisation" (Rose and MacDonald 2012, 196), or the implications behind the name Instant Composers Pool, an Amsterdam-based ensemble working on the verge of composed and improvised music. In this research project, I try to stick to the terms improvisation and extemporization, using the latter occassionally as a synonym for improvisation in order to create some diversity in the vocabulary.


So what is happening when you improvise? Are you really making things up as you go along?


In The Improvisation of Musical Dialogue, Benson claims that improvisation is ubiquitous in musical practice but that no instance of improvisation qualifies as “something created on the spur of the moment out of nothing” (Benson 2003, 30). It is ever-present in that it is always connecting the performer, composer, and audience in the field of musical dialogue in ways that happen differently each time, even though these connections are very much guided by previous practices and experiences (my wording of Benson's general thesis).1 I take his viewpoint earnestly and will base the rest of this discussion of improvisation upon the notion that improvisation is ubiquitous in musical practice. This starting point resonates very well with my personal experience as a practicing musician  from reacting to a loud cough during the quietest passage of a slow movement to responding to a particularly unusual piano or acoustics  there does always seem to be some improvising happening while playing music, albeit sometimes to relatively small degrees (as in the two examples just mentioned). This ubiquity of improvisation in musical practice is then augmented by another notion mentioned in Benson’s book, namely that “each instance [of improvisation] involves a kind of reworking of something that already exists, so the differences concern the ways and the degrees to which this reworking takes place" (Benson 2003, 30). If this is the case, if improvisation is always a reworking of something that already exists, can it ever produce something brand new?


In his book The Field of Musical Dialogue, music philosopher Marcel Cobussen quotes the French philosopher Jacques Derrida who remarks that “improvisation, regarded as a purely spontaneous creation of something absolutely new, is impossible” (Cobussen, 2017, What is Improvisation (Not?) [3]). Derrida's claim indeed resonates with Benson's insistence that every instance of improvisation "involves a kind of reworking of something that already exists," and the Russian philosopher Michael Bakhtin argues that "something created is always created out of something" (Bakhtin 1986, 120; also quoted in Cobussen 2017, What is Improvisation (Not?) [3]). Using Bakhtin and Derrida as well as other philosophers to substantiate his view, Cobussen states that:


improvising can therefore be understood as interacting with and transforming certain historical predispositions or, in a less time-based formulation, engaging with existing structures and rules, accepting, rejecting, appropriating, or building on previous selections and decisions. (ibid.)


Relating to "existing structures and rules" plays a significant role in the improvising presented in this research project because of the attention given to remaining within the context of nineteenth-century piano music. The question may then arise as to whether less idiomatically-oriented improvisation still relates to structures and rules, or "historical predispositions" at all, but Cobussen uses social scientist and philosopher Bruno Latour to argue that "any given interaction seems to overflow with elements which are already in the situation coming from some other time, some other place, and generated by some other agency" (Latour 2007, 166; quoted in Cobussen). Associating the question of the access to newness through improvisation and another concept commonly associated with the term, freedom, music philosopher Eric Landgraf claims that, "far from being the expression of unbridled freedom, improvisation marks a process that acquires a degree of consistency by connecting to  repeating and altering  what has come before" (Landgraf 2011, 16). Within the context of Playing Schumann Again for the First Time, the pursuit of something “absolutely new" would probably transgress the original goal of the project, which namely is to pursue ways of learning to improvise convincingly within the context of Schumann and the nineteenth-century piano repertoire in general. As already mentioned in the Introduction, a practitioner like myself who engages in various musical styles is actually looking for a way to avoid the "new" with this project; by re-engaging with and reworking the old, my intention is to pursue improvisation in the context of Schumann's piano music more according to "the sense implied by the literal meaning of the term derived from Latin, defining improvisation as an unforeseen, unforeseeable, and unplanned activity that is inventive, as any creative ‘doing’ that, voluntarily or involuntarily, unfolds without following a predetermined plan" (Landgraf 2011, 16; author's own emphasis). The lack of a specific predetermined plan, however, should not undermine the fact that "improvisation is linked to the known and familiar in as much as the known and familiar must be recognized as that from which the improvisation has to distinguish itself" (Landgraf 2011, 93).


In 1994, pianist-improviser-composer Frederic Rzewski said that:


Improvisation is a controlled experiment with a limited number of unknown possibilities. It always has rules and a framework. There is no such thing as a 'free' improvisation. [...] Learning to improvise can also be compared with learning to speak in public. Success in both depends upon being able to concentrate on the thing to be done, to somehow ensure that it goes on, and on all of the things that could possibly go wrong; which, although they have to be considered at some point, should not be allowed to interfere with the performance in the moment of doing it. Of course errors must inevitably happen; but part of the interest in watching or listening to a successful improvisation lies in the fact that such errors are human and universal, and therefore a part of everyone's experience. The skillful recovery of a dropped ball is satisfying to observe, because the spectator can at once identify with the player in weakness while admiring the same player's strength. (Rzewski 2007, 104—105)


Rzewski's insistence on rules and a framework in order to improvise coincide well with the analogy of improvisation as "learning to speak in public." In what language is the improviser speaking? In a language that already exists? In a language that the audience can understand? In an essay discussing why performers should study performance, American author and jazz musician José Antonio Bowen uses the language analogy to convince the reader that "the point of learning a new language is that we can eventually speak for ourselves. When we are fluent, we can create expressions never heard before, but still understood" (Bowen 1996, 34).2 Returning briefly to the issue of newness, the reason such a phrase can be understood supposes that it is not “absolutely new” at all, but rather an altered instance of something repeated from the past. Bowen's example also highlights how the rules and framework of a language ordain understanding. Without them, an expression could be understood in multiple ways and might therefore lack meaning due to a lack of context. (During my first trip to South Africa, I could not help but understand the Afrikaans phrase baie danke [thanks a lot] as buy a donkey). Rzewski's admiration of the improvisatory recovery of a dropped ball  equating a dropped ball to an unwanted happening in a musical performance such as a wrong note  also raises the question of degree. What efforts does one have to go through to recover a dropped ball? The size of the court is arguably a lesser factor than the experience and skill that the player has in recovering a ball. When a classically-trained performer misses a note and continues as if nothing happened, was that ball easier to recover than when John Coltrane makes a “mistake” while soloing on the changes from “My Favorite Things” for ten or so minutes? Although the word “mistake” implies something unwanted, an elegant recovery can render the mistake-event as something crucial to the improvisation – this recovery can only be considered elegant if the mistake is surrounded by other behavior that provides coherence to the performance. Improvisation is embedded to various degrees in these two musical “mistakes” as well as in their respective follow-throughs. However, a “mistake” in a higher degree of improvisation (mistaking in which musical tradition one takes part, for example) might have further-reaching consequences.3

In her article “Improvising Impromptu, Or, What to Do With a Broken String,” Lydia Goehr differentiates between the concept of improvisation impromptu – the everyday improvisation of “doing exactly the right or wrong thing in the moment  and the musical improvisation extempore - “when musicians make up music in performance, from this moment forward(Goehr 2016, 459; author’s own emphasis). Rzewski’s examples of a dropped ball and a mistake during a musical performance both constitute instances of improvisation impromptu. Goehr continues by explaining that improvisation impromptu can and does happen during improvisation extempore and that these concepts are differentiated by the fact that “improvisation extempore asks us to attend to what is achieved in the performance as a whole, [while] improvisation impromptu picks out the inspired or exemplary turn in a performance when, on the spot, one does (at best) the right or winning thing” (ibid., 464; author’s own emphasis). A further way to differentiate between the two concepts is in regards to the knowledge that the performer has (or lack thereof) of the improvisatory act before its instantiation:


In improvising impromptu, unlike in improvising extempore, the not-knowing is three fold: we do not know that we will have to improvise at all, or how we will improvise when we do have to improvise, and we do not know how our improvising will turn out, although, if well-trained, we might well feel secure that we will indeed know how to employ our wit to make the right fit. (ibid., 471)


Although I believe that a practitioner also does not know exactly how his or her improvisation will “turn out” when improvising extempore, the core issue of degree that this chapter addresses applies (equally well) to both the improvisational concepts of impromptu and extempore. Improvisation extempore is yielded to various degrees in both performance and composition, and as already described above in the discussion of dropped balls and “mistakes,” improvisation impromptu can emerge as a minor mishap or as a significant hurdle during performance.


The rules and framework themselves can also become parameters for improvisation; or, to continue by speaking of degrees, a large enough degree of improvised musical behavior could even alter the rules and/or framework as well. Discussing Erlend Dehlin's PhD dissertation “The Flesh and Blood of Improvisation,” Cobussen says that a "form of rationality, usually presented as a given set of rules, might become an instrument rather than a restraint jacket, not to be obeyed and followed indiscriminately, but to be used creatively and spontaneously" (Cobussen 2017, Management - Richard Barrett [4]). Dehlin himself articulates that "abstract models such as rules and structures [...] can be used in improvisation as tools" (Dehlin 2008, 94-95). A good example of music that results from breaking rules can be heard in the Eusebius Traum subchapter of the chapter On My Improvisation Methods, where two of the three instances of Eusebius Traum do follow the rules laid out for the musical experiment whereas the third instance does not.4 


"Performers  even when performing music that is strictly notated  do not merely 'perform' but also 'improvise' upon that which they perform" (Benson 2003, 26). Using this perspective as a departure point but then quickly extending it to include composers  composers are also always improvising to some degree  Benson summarizes various degrees of improvisation in both performance and composition in a list, designating a number for each degree to which musical improvisation can take place. The list concludes with the eleventh degree of improvisation (improvisation11), after which he follows with this disclaimer: "by no means is this list meant to be exhaustive. These are only some of the forms and degrees that improvisation may take. I sketch these to show not only how varied improvisation can be but also how ever present it is in both 'composition' and 'performance'" (Benson 2003, 30). Benson's items on the list serve to introduce each following subchapter, which include extended commentary of my own, relating each item on the list to music exposed in this project as well as to examples from pianist-improviser-composer colleagues. I use this detailed list as it helps me to present the various degrees of improvisation apparent in music-making within the specific context of nineteenth-century piano music. By combining Benson's insistence on the ubiquity of improvisation and the degrees to which it can take place along with Rzewski's idea of the necessary existence of rules and a framework, a complete picture of what is happening in Playing Schumann Again for the First Time can start to take form.5




Improvisation1: This sort of improvisation is the most "minimalistic." It consists of "filling-in" certain details that are not notated in the score. Such details include (but are not limited to) tempi, timbre, attack, dynamics, and (to some degree) instrumentation. No matter how detailed the score may be,
 some – and often much  improvisation of this sort is necessary simply in order to perform the piece. Thus no performance is possible without some form and degree of improvisation1. (Benson 2003, 26)

I understand this first degree of improvisation to not yet include either "filling-in" or varying pitch content or rhythm, but that is does include all other musical parameters such as (but not restricted to) those listed by Benson  "tempi, timbre, attack, dynamics, and instrumentation." Disregarding instrumentation because it does not fall within this project’s scope, I also consider articulation, tempo modification, arpeggiation, and hand dissynchronization to be parameters that are involved in improvisation1.6 In the Practicing subchapter in the chapter On My Improvisation Methods, the categories of practicing methods could be divided into those that Benson considers to be improvisation1 and those that extend into a different degree, improvisation2 for example. Methods categorized under experimenting with articulation, dynamics, tempo modification, hand dissynchronization, and character distinction are those that I have adopted to make this first degree of improvisation as fruitful as possible in my piano playing. Already by considering the first degree of improvisation on Benson's list, an important distinction between the goals of this project and Benson's discourse needs to be clarified. Benson is considering the phenomenology of musical practice in general and often (and appropriately, in my opinion) refers to a performance of a score. Returning to the main research question behind this project  how one can learn to improvise convincingly within the context of the nineteenth-century piano repertoire?  working with a score may be a reasonable departure point but certainly is not the final goal if one desires to improvise without the presence of a score (see the discussion of the idea of "circumventing the musical work" in the chapters On Mimesis and Morphosis and On Rethinking the Current Performance Practice of Nineteenth-Century Music). For this reason, “filling in” certain details of a score does not suffice for this project, but “changing” details in a score indeed becomes a useful departure point for taking initial steps towards integrating this degree of improvisation more widely into one's own practice by creating some initial distance from a particular score. In this sense, I deviate from Benson’s improvisation1: whereas he claims that improvisationis taking place all the time and in every performance (regardless of whether or not it is described as such), I agree but also use this as an opportunity to look for ways to systematically practice improvisation1 in order to develop improvisational skills. My deviation from Benson's approach applies to the other degrees of improvisation as well. Thus, more generally stated, I believe that Benson wants to use these scale degrees of improvisation to delineate how improvisation is ubiquitous in both musical performance and composition in his book; and while I agree with this, I also have the goal of pursuing practicing strategies towards learning how to improvise more convincingly. All of the methods presented in the Practicing subchapter thus extend past the idea of "filling-in" and continue with "changing" details in a score in order to find more space for musical results related to improvisationas well as the other degrees of improvisation. This process of “changing” is described as “non-compliance to the musical work” in the chapter On Mimesis and Morphosis. Benson’s improvisation1 serves well to delineate what parameters are involved with improvisation at this degree, and his division of improvisatory activities into a eleven-fold list conveniently helps me to understand which particular elements of musical practice are in play in certain contexts and therefore allow me to conceive of ways to practice various approaches to these elements.

Further considering the
shift from “filling in” to “changing” details in a score can also help to clarify what I mean by practicing in the context of this research project. Understanding the way classical music is generally practiced as preparing in order to realize a score in performance, improvisation1 plays the role of “filling in” details in the score as necessary in order to achieve the ability to perform the score. A parallel mode of practicing emerges when details that may or may not be notated in the score are "changed" in ways that might completely misrepresent the composer's "intentions" for example, but sound interesting nevertheless. I consider both modes ("filling in" and "changing") to be versions of Benson's improvisation1 when they are engaged with tempo, timbre, attack, and/or dynamics (as well as articulation, tempo modification, arpeggiation, and hand dissynchronization), which creates space for improvisation at this particular degree by exploring the availability of options within these parameters, which in turn creates the potential for more significant amounts of improvisation to happen during performance  either in a performance related to a pre-existing musical work or in other types of performance that might not explicitly comply with any particular precomposed work.

As an example of how these two modes are actually already both active in pianistic practice, see Schumann's sketch here to the right and consider what details need to be added to merely realize it in performance, and what other details could be changed in order to turn this sketch into productive practicing. Assuming that it was notated to assist his own practicing and not intended as a musical work, taking a similar approach to changing details in musical scores that do intend to convey a musical work opens up many possibilities towards realizing initial degrees of improvisation in performance.

A final observation on the role of improvisation1 in this project: two performances of Schumann's Fantasy in C Major, Op. 17, which almost exclusively exhibit improvisation1, are presented in the chapter On My Improvisation Methods chapter at the culmination of the spiral entitled Creating layers for improvisation. All of the improvisation that was exhibited in these performances, except the realization of a few trills (Benson’s improvisation2) and the fact that I am “part of a musical tradition” (improvisation11), is a form of improvisation1. This musical presentation that mostly exhibits improvisation1 as the last examples in the spiral attempts to show that all other degrees of improvisation return to influence the ability to engage with improvisation1, just as improvisation1 impacts one's ability to engage with other degrees of improvisation. The discussion of Benson's list that continues here should eventually bring the reader to an understanding of how varied but also how intertwined these various degrees of improvisation can be.



Benson's list ends with improvisation
11  "both composer and performer are part of a musical tradition and they work within that tradition." They oscillate between tradition and innovation, between constraints and potentially unlimited exploration, between fixity and freedom. Benson continues by saying that "by no means is this list meant to be exhaustive. These are only some of the forms and degrees that improvisation may take" (Benson 2003, 30). Upon arriving at the end of Benson's list and considering the musical examples exposed in Playing Schumann Again for the First Time, the question remains as to which degrees of improvisation they exhibit. Many of the items on Benson's list could be taking place all the time while I make music. Improvisation7 ("altering the score by changing the melody line and/or altering the chords"), if understood more generally and extended beyond Benson's references to jazz, can be heard regularly in my music-making. A melody and/or chords are often being altered, regardless of whether or not there is a notated score with which to compare the altered material to any particular original. In extemporized music, the first moments of the performance are almost always altered in some way during the course of performing. Improvisation9 ("using a particular form or style of music as a kind of template") also happens everywhere in the music exposed here, because no instance of my music-making could be considered brand-new in the sense that it does not have some sort of predecessor in terms of genre or template. The list of Czerny's genres (presented in the chapter On My Improvisation Methods and already discussed in this chapter in the subchapter Improvisation9) in which a performer of his time was expected to be able to improvise, and my corresponding list of how the examples in this project could be divided up accordingly, serves to show that – however loosely – all my music-making can be understood in the context of an established genre or musical form. The associating of musical materials that I assign to improvisation5 ("adding or subtracting measures, passages, or even complete sections") is also always present, even when (one could even say especially when) the association of materials happens through embodied knowledge, the kind of knowledge that is guided by the feedback between mimesis and morphosis (as described in the chapter On Mimesis and Morphosis) and that can be acquired and shaped through reflection-in-action (see the chapter Between Listening, Forgetting, and Conceptualizing).

A clearer picture is starting to emerge of how more than one of these
degrees of improvisation is always involved in any given instance of music-making. The summation of the degrees of improvisation that seem to be ever-present here, as well as including more specific degrees that are present due to a particular context, could be described as improvisationx – the layering together of all currently active degrees of improvisation. Improvisationx happens in all music-making, the specific definition of the variable x being the sum of the specific degrees taking place in a particular musical moment. Improvisationx conveniently brings one back to Benson's original thesis – that improvisation is ubiquitous in musical practice, albeit to various degrees. These degrees, the "difference [of which] is far more quantitative than qualitative" (Benson 2003, 30; author's own emphasis) prove to be very useful in understanding how exactly improvisation functions in the music-making exposed in this project.

Concluding by returning to Rzewski's idea of improvisation as a contained experiment with rules and a framework: What are the rules and framework for Playing Schumann Again for the First Time? I propose a short list of rules below, preceded by a definition of the framework, as the guiding principles behind the music presented in this project.


  • The framework: music that can be understood in the context of the nineteenth-century Biedermeier German musical tradition

  • Rule 1: Practice (at the piano, with and without the assistance of notation and a recording device) with the goal of not just realizing a precomposed musical work in performance. Use scores for inspiration to find results that are non-compliant to the score. Use materials such as melodies, harmonies, patterns, elements of timing and rhythm, etc. from Schumann and his milieu, and combine, layer, and alter them.

  • Rule 2: Continue to pursue unforeseen musical results while acknowledging the mimetic and morphosic aspects of the process.

  • Rule 3: Continue to pursue unforeseen results by searching for ways to create music without any particular background text as a guide; for example, systematically create enough distance from a score by Schumann until it is no longer recognizable.

  • Rule 4: Pursue sounding results that fit into the context of the nineteenth-century piano repertoire while simultaneously becoming conscious of your own sonic signature.


Overview of the following chapters

The core of this research project, On My Improvisation Methods, presents over 50 musical examplesmany of them exposing strategies I have learned from performer-improviser-composer Rudolf Lutz. Reflections on the music in this chapter often relate to historical texts (like Czerny's Systematische Anleitung zum Fantasieren auf dem Pianoforte, Op. 200). In the chapter On Sonic Signature, the roles that Schumann and I play in this project will be discussed in more detail by presenting the concept of sonic signature as a way of describing how music bears certain audible qualities that reflect the individual approach of a specific musical practitioner. And in the final chapterentitled Between Listening, Forgetting, and Conceptualizing, the viewpoints of two prominent pianist-improviser-composers, Frederic Rzewski and Robert Levin, as to what is happening during improvisation will be analyzed to present a continuum that exists between the listening, forgetting, and conceptualizing that happen while improvising.


Improvisation11: This is the most subtle form of improvisation. Both composer and performer are part of a musical tradition (perhaps classical, blues, or folk music) and they work within that tradition. But working within a tradition inevitably requires modifying that tradition by augmentation and transformation. One follows the rules of composition and performance; but composers and performers 
 particularly those we consider to be exemplary  also modify those rules and expectations. Therefore, the tradition is itself improvised upon. Any practice or discourse involves such improvisation. (Benson 2003, 2930)

11 hits at the artistic impetus behind Playing Schumann Again for the First Time. As explained in the Introduction, I have engaged in improvisation in many contexts throughout my life, and the goal of this project therefore has been to find and pursue strategies that help the musical results more explicitly remain "part of a musical tradition." By initially remaining very close to Schumann's scores, and then slowly creating distance from his works to find room for ever-increasing degrees of improvisation, I believe this project has helped me (and can inspire to help others) to improvise more effectively within the context of the nineteenth-century musical "tradition." Certainly, as Benson states, "working within a tradition inevitably requires modifying that tradition by augmentation and transformation." The transformation of Schumann's materials through the music-making exposed in this project to creates something outside the scope of a performance of his works, as has been described in the chapters On Mimesis and Morphosis and On Rethinking the Current Performance Practice of Nineteenth-Century Music. Furthermore, the influences that I bring to the music-making that are external to Schumann, and which are then realized in performance with the assistance of my embodied knowledge, ultimately bring certain elements to the music that can only be traced to my practice. I use the term sonic signature to describe this phenomenon in the chapter On Sonic Signature.



Improvisation2: Although this level of improvisation is close to the previous one, it differs in that there is the addition of notes to the score that the performer is expected (by the composer) to supply. Two common forms of improvisation2 are the addition of notes to complete a trill and the "filling-in" required by a score that only supplies figured bass. (Benson 2003, 26)

Benson's two examples, adding notes to complete a trill and realizing a figured-bass score, show how diverse musical practices that exhibit improvisation2 can be; furthermore, some of the skills needed to realize improvisation2 are generally practiced by classically-trained pianists while some are not. All pianists approaching the standard repertoire have certainly found realizations for playing ornaments, either on their own or likely with the help of a teacher, while arguably relatively few pianists who engage with nineteenth-century piano repertoire also practice figured bass. Realizing figured-bass lies at the core of the work I did with Rudolf Lutz to develop the improvisation strategies exposed in the chapter On My Improvisation Methods, and figured-bass scores that I created based on Schumann's scores can be found in the Eusebius Traum and Free Fantasy à la Kreisleriana VI. subchapters. There are many other examples of improvisation2 that occupy the space between completing trills and realizing a figured-bass score; pianist and musicologist Robert Levin's research on performance practice of the music of Mozart, for example, focuses on certain moments in scores notated in the eighteenth century where the performer is expected (by the composer as far as we are able to know) to supply notes. In his lecture Improvising Mozart from 2012, Levin spends time discussing two facets of performing Mozart that I consider to both be examples of improvisation2: embellishing melodies upon their reprise in a piece of music, and extending figurative passage-work in some of Mozart's short-hand notational moments (see for example Levin 2011, 258 for a discussion of extending the preceding virtuoso passage-work into the notated whole-note passage that follows in the Piano Concerto in E-flat Major, K. 482).

A connoisseur of classical pianists' playing styles might think that Robert Levin and Alfred Brendel approach Mozart from two completely different ends of the spectrum; yet Cobussen shows what significant roles improvisation1 and improvisation2 play in Brendel's Mozart performances as well. After summarizing some of Brendel's own sentiments as to how he approaches Mozart, Cobussen states that:

a performer of Mozart's music is required to do much more than simply reproducing what is written down by him (or by editors and pupils). From relatively modest interventions, such as filling in dynamics or ornamentation, to the "quasi-improvisational" elaboration of cadenzas, Mozart's scores require a certain amount of flexibility, creativity, and complicity from the performer in order for justice to be done to the compositions. (Cobussen 2017, Alfred Brendel and W.A.Mozart: The Score as Actant, [2])

Comparing Brendel's and Levin's performances of Mozart makes clear how ever-present improvisation1 and improvisation2 are in the general approach to performing Mozart's music today. Both performers embellish melodies, add ornaments, and apply appropriate passage-work to short-hand notational moments in Mozart's scores like the one described above in K. 482. The interesting difference between Levin and Mozart lies in the presence of improvisation3 (see the following subchapter) in Levin's performances and the lack thereof in Brendel's, in that Levin improvises cadenzas to the Mozart concerti whereas Brendel either plays Mozart's composed cadenzas or writes his own. (The subtle differentiation between improvising a cadenza and writing one's own and performing it will be addressed in the subchapters Improvisation3 and Improvisation9, where composition and performance will be more thoroughly juxtaposed.)

Because Schumann's scores are different than Mozart's, there is less room for improvisation2 to take place when realizing Schumann’s music in performance.7 Therefore, improvisation2 plays a smaller role in this project than other degrees of improvisation in the sense that there are not many moments where Schumann's melodies are embellished or ornaments added; however, Benson's other example of improvisation2, the practice of "'filling-in' required by a score that only supplies figured bass" (Benson 2003, 26) has had a great impact on the work exposed here. Examples of figured-bass scores that I created based on Schumann's scores have already been mentioned above, and realizing figured bass (often also referred to as thorough bass, basso continuo, or Generalbaß in German) in performance has helped me find new practical and embodied knowledge through the process of navigating previously unfamiliar harmonic and contrapuntal scenarios. Describing his realizations of pieces from Johann Mattheson's Grosse Generalbaß-Schule (1731), harpsichordist Jean-Christophe Dijoux says that "[Mattheson's Grosse Generalbaß-Schule] offers a wonderful basis to learn improvisation and even composition by means of continuo playing" (Dijoux 2017, [0:46]). Dijoux claims that performing Mattheson's scores by expanding the bass line and numbered figures into a multi-voice, two-hand musical experience, helps one to develop one’s skills in "improvisation and even composition." I interpret this phrase to mean improvisation in performance and improvisation in composition, as improvisation is present in both activities (in my and Benson's opinions). Dijoux continues by commenting that, "actually, what I found truly exciting is that these pieces have to sound as if they were fully written out, [...] and finally [figured-bass practice, BM] allows you to see and hear the written out music with different eyes and ears" (Dijoux 2017, [5:26] and [5:41]). In other words, the practitioner develops the ability to experience a written-out piece of music as one of many possible realizations of a figured-bass score.



Improvisation10: Here the composer takes a particular piece of music  a common folk tune (as in Aaron Copland's use of the Shaker melody "Tis a Gift to Be Simple") or the composition of another composer (such as Handel's "borrowing" from other composers of his day)  and arranges it or uses it as the basis for a more complex or just simply different work. Whereas it is relatively easy to distinguish between Copland's reworking and the Shaker tune, sometimes the distinction between the material used and the reworked product is not so clear. (Benson 2003, 29)

One of the aims of Benson's book is to make explicit how improvisatory the process of composing is. Commonly understood as only happening in performance, Benson makes clear that improvisation is present in both composi
tion and performance. Improvisation10 indeed happens in composition, and when it also happens in performance it is often described as a "free fantasy" 
 as heard in the example of Happy Birthday Georg!. Czerny's description of certain categories of "real improvisation" – “das wirkliche, selbständige Fantasieren [oder] Improvisieren" (Czerny 1829, 4) – includes a scenario that is very similar to Benson's improvisation10, namely when one engages "in die Durchführung eines einzigen Thema in allen, in der Composition üblichen Formen (in the realization of a single theme in all forms typically found in composition)(ibid.; my translation). In the chapter On My Improvisation Methods, Czerny's categories of improvisation in performance that he considers “freies fantasieren,” as well as his list of forms “typically found in composition” that one should also be able to use for a performed improvisation, are presented.




Improvisation9: Whereas the first eight forms of improvisation are those of the performer, improvisation9 is a compositional form of improvisation. Here the composer uses a particular form or style of music as a kind of template. Thus, Mozart's Cosi fan tutte depends on the opera buffa form, which has relatively strict requirements. How far requirements are followed, though, is subject to improvisation. (Benson 2003, 29)


Improvisation3: The difference between improvisation3 and improvisation2 is purely quantitative. Rather than merely adding selected notes or filling out the chord, the performer adds measures or even whole sections. Examples include Baroque and Classical cadenzas, which the composer (again) expects the performer to supply. Sometimes these cadences are written out by the composer, with the expectation either that the performer follow them to the letter or else as a kind of guide or springboard for the performer's own improvising. In the case of the former, the performer then must decide whether to follow the supplied cadences or create something different (or improvise some admixture of the two). (Benson 2003, 26

Improvisation3 could be described as extemporizing whole sections of music within the context of realizing a precomposed score in performance. The difference that Benson highlights between improvisation3 and improvisation2 is the space where improvisation in performance begins to pose serious challenges. Embellishing a melody as a form of improvisation2 may be challenging enough for some; but adding measures or whole sections of music to a particular score can seem quite daunting at first. Improvisation3 encompasses one of the more practical examples of why a concert pianist would be interested in learning to improvise 
 namely, in order to be able to improvise a cadenza (or multiple cadenzas) within a concerto for piano and orchestra by Mozart, Beethoven, Schumann, Grieg, Tchaikovsky, Haydn, et cetera. For an example of this in Playing Schumann Again for the First Time, see the Beethoven examples presented in the subchapter Haydn and Beethoven in the chapter On My Improvisation Methods. Improvisation3 can be readily heard in Frederic Rzewski's performance of Beethoven's Hammerklavier Sonata (discussed in more detail in the chapter Between Listening, Forgetting, and Conceptualizing), Levin's improvised cadenzas to the classical concerti, and also in the subchapters Waldszenen and Cadenza to the Humoreske.


By calling improvisation9 a compositional form of improvisation, I believe Benson is contrasting performance and composition to draw a distinction between improvisation8 and improvisation9. The question will soon arise as to whether certain types of extemporized performanceare best described as exhibiting improvisation8 or improvisation9. According to Benson, improvisation8 happens when, "using the basic form of the score [...], the performer improvises within those confines," whereas improvisation9 occurs when a "composer uses a particular form or style of music as a kind of template" (ibid.). Returning to the idea of circumventing the musical work (presented in the chapters On Mimesis and Morphosis and On Rethinking the Current Practice of Nineteenth-Century Music), when sufficient distance from any given musical work is created during performance so that there are no longer just explicit references to one particular precomposed work, I believe Benson's wording for improvisation9 more accurately describes what is happening. A performer "uses a particular form or style of music as a kind of template" to guide and structure the improvised performance. These templates are very often a genre delineation, like opera buffa in Benson's example, but could also refer to form (a fugato template, for example). All of the musical examples presented in Playing Schumann Again for the First Time are based on templates to various degrees, and a consideration of which templates correspond to each specific piece can be seen in the chapter On My Improvisation Methods where I correlate specific genres listed in Czerny's Systematische Anleitung zum Fantasieren auf dem Pianoforte to the music presented there.



Improvisation7: This sort of improvisation consists of altering the score (or, perhaps a "chart," a minimal score) by changing the melody line and/or altering the chords. Such improvisation is found in Baroque music and jazz. There are many variations of improvisation7, such as:

7a: the melody line is slightly changed, so that it is still clearly recognizable, as, say, the Gershwin tune "A Foggy Day," but it does not strictly follow the sheet music version

7b: a chord is altered enough to make it a different chord but still remains close enough to the original chord so that the new chord "fits" in the same place as the original chord. 
For instance, jazz musicians routinely alter the chord sequence in the third measure of "I Can't Get Started": instead of two beats each of E7 and A minor, they play one beat each of B minor7, E7, Bbminor7, and Eb7 

7c: the melody line is substantially changed, so that it is no longer completely clear to the listener whether there is any connection to the original melody

improvisation7d: certain chords are changed substantially, although the basic chordal structure of the piece remains

7e: the melody is completely disregarded and an alternative melody (or simply no discernible melody) is put in its place (Benson 2003, 28)

Benson's approach to describing improvisation
7 is mainly considering jazz performance practice. In this project, his description of 7a, 7b, 7c, 7d, and 7e applies to activities that can be seen and heard in the F.A.E. Variations if one considers that the ensuing variations alter certain parameters that are present in the original score  in this example's case, the theme. Because these alterations came about during the process of composing rather than in the moment of performance, this example shows that this degree of improvisation (and arguably all degrees) can be realized in both performance and in composition.




Improvisation4: In this case, a piece is transcribed for either a different instrument or different instruments, for voice (if originally for instruments), or for instruments (if originally for voice or voices). While the notes in the transcription may stay the same as in the original piece, often they are changed to accommodate the new instrument(s) or voice(s). However, usually the note relationships remain the same, meaning that the basic melody line and also the chords are unaltered. Transcriptions generally attempt to render the piece as close to the original as possible, making changes only when necessary (to accommodate different instruments or voices). However, not all transcriptions follow the original score in a strict sense. Depending on just how much a transcription varies from the original, it may become an instance of improvisation5 [BM: does Benson actually mean improvisation6 (arranging) here?]. Transcriptions, of course, can be done by the original composer or someone else. (Benson 2003, 27)

Because Playing Schumann Again for the First Time is a solo instrumental project, transcribing music from one instrument to another is not applicable here; nevertheless, transcription certainly plays a role in the project and serves as a useful concept to begin to discuss other musical activities besides performance that took place during the working process. While this project maintains a clear focus on performance, notating music also played a crucial role in trying to develop my skills as an improviser. How specifically is notation yielded in this project? And can it be present without necessarily engaging in composition? Transcription helps to answer these questions, albeit in a slightly different sense than Benson understands: I often used notation to transcribe audio recordings, attempting to codify a performed improvisation. For an example of this, listen to Jübel für Ruedi to hear a circa two-minute recording of extemporized piano playing which was then later transcribed and emerged as a notated score. My use of transcription, both similar and dissimilar to composition, still reflects Rzewski's thoughts regarding the technique of composition:


The most basic technique of composition is that of transferring information from short-term memory to long-term: remembering an idea long enough so that one can write it down. This process of transference is also one of translation: re-forming an impulse or feeling so that it can be expressed in some kind of symbolic language. (Rzewski 2007, 52)


Transferring and translating are involved in transcribing, and this notion of re-forming an impulse to be expressed in symbolic language is precisely what is happening when one tries to notate what happened in a performed and recorded improvisation. In my view, regardless of whether the transference and/or translation involve notating what happened in a recorded improvisation or converting other impulses into notation, this process constitutes an improvisatory act and could be considered an example of Benson’s improvisation4.8


How else are composing and improvising in performance related? Arnold Schönberg said that "komponieren ist eine Art verlangsamte Improvisation; oft kann man nicht schnell genug schreiben, um mit dem Strom der Gedanken Schritt zu halten (composing is a form of decelerated improvisation; it is often difficult to write quickly enough to keep up with one's own stream of thoughts)" (Schönberg 1976, 140; my translation). Schönberg's comments resonate well with Benson's general thesis that improvisation is inherent in both composition and performance. But why exactly can composition be described as a form of decelerated improvisation? Frederic Rzewski more precisely pinpoints the improvisatory aspect by claiming that "composition is the result of an editing process in which one's impulses are passed through the critical filter of the conscious mind: only the ‘good’ ideas are allowed to pass through" (Rzewski 2007, 52). Schönberg's slow-motion process of translating one's impulses into musical notation can be related to Rzewski's filtering the "good" ideas from the "bad" ones, and his thoughts serve well to further clarify what is happening during this translation:


The most basic technique of composition is that of transferring information from short-term to long-term memory: remembering an idea long enough so that one can write it down. This process of transference is also one of translation: re-forming an impulse or feeling so that it can be expressed in some kind of symbolic language. The most basic technique of improvisation is that of short-circuiting this process of conservation: forgetting – momentarily at least – everything that is not relevant to the objective of expressing an idea immediately in sound. (Rzewski 2007, 52)


Rzewski claims that the process of transferring information from short-term to long-term memory is short-circuited when it needs to be realized immediately in performance. But are the basic techniques of composition and improvisation in performance essentially different? If one's "impulses" when realized during performance result in extemporized music, what might they create when translated into notation instead? And how can subjecting one's impulses to the compositional working process inform the way the same impulses can be translated into sound during performance? This project attempts to provide an answer to these questions with the examples exposed in the chapter On My Improvisation Methods, where the music-making hovers between notation and performance. Mondlicht or Menschenblut, to name two examples, exhibit both improvisation in notation and in performance in that some of the impulses guiding the performance were first notated while others were realized directly during the playing.9



Improvisation8: Using the basic form of the score (such as a typical sixteen-bar blues piece), the performer improvises within those confines. In such a case, there may be no connection to the original melody, or even chords. Whereas the various forms of improvisation7 had at least some connection to the original "piece," here there really is no discernible connection (at least to the listener, although possibly to the performer and the composer of the piece). (Benson 2003, 29)

Maintaining the basic form of the score and improvising within those confines seems to best apply to jazz and
Baroque music, although this practice extended into the Classical and early Romantic musical eras as well. For an example of improvisation8 brought into a trans-stylistic realm, see pianist-improviser Johnandrew Slominski's La Follia Variations (Slominski 2017). Slominski puts the Follia model through many different stylistic filters that range from early Baroque to twentieth-century impressionistic contexts. In Playing Schumann Again for the First Time, the super-impositions (Eusebius Traum, Melancholic Flowers, Kreisleriana V meets Kreisleriana VIII, and R meets R) come closest to exemplifying Benson's improvisation8, although they are different because of the aspect of super-imposing two scores on top of each other. Characteristics from one score (such as melodic or figurative materials) are super-imposed onto the other score’s form and harmonies. The performances are therefore less improvised than what Benson might be imagining because they are dictated by two scores instead of one. The border between improvisation7 and improvisation8 is not always clear in the music presented in this project, partially because pre-determined harmonic structures play a less significant role here than in jazz or Baroque music. Harmonic structures are indeed present - ever-present actually - and consciously utilized, but often in a more transient manner with the tendency to quickly develop from one particular structure to another. The way harmonic structures have been used in this project can perhaps be better understood by considering what are often called "models:" miniature structures that serve as building blocks for creating harmonic music. While working together with Rudolf Lutz, we often discussed certain models using the names assigned to them by Robert Gjerdingen in his book Music in the Galant Style (2007), and further examples of models and their common usage can be found in the Compendium Improvisation: Fantasieren nach historischen Quellen des 17. und 18. Jahrhunderts (2017), edited by Markus Schwenkreis. Lutz and I used the names of these models as a way of communicating and guiding our improvisatory activities together. In Widmung for example, a model that shows characteristics of both Gjerdingen’s Monte and Meyer can be heard, and this particular combination of models was the focal issue around which this improvisation was conceived and performed.



Improvisation5: This level of improvisation goes somewhat further than any of the previous ones, for the performer or conductor or an editor alters the score by adding or subtracting measures, passages, or even complete sections. Depending on the era of the composer, such alterations might be expected (as in Renaissance or Baroque music), allowed (as in some pieces by Beethoven), or explicitly (or implicitly) condemned (say, by a twentieth-century composer). Thus, such changes may or may not be expected by the composer. Those expected or approved could be designated improvisation5a, with those not expected or approved designated as improvisation5b. (Benson 2003, 27)

Adding or subtracting measures, passages, or even complete sections of a score" plays a crucial role in this project in much the same way a lighting technician is involved in a theater production. The lighting illuminates everything that happens on stage, and without it, nothing can be seen. Improvisation5 encompasses the decision-making processes that I went through in order to bring different kinds of musical materials together. The most explicit example of this can be seen and heard in the F.A.E. Variations where excerpts of scores by Schumann and Bach were combined and utilized. Associating materials also played a crucial role in the process of super-imposing a certain attribute from one piece with the harmonic implications of another that resulted in Eusebius Traum, Melancholic Flowers, Kreisleriana V. meets Kreisleriana VIII., and R meets R (this process of super-imposing will be discussed in more detail in the subchapter Improvisation8). Needless to say, combining arbitrary materials “at random” does not necessarily produce interesting results; bringing certain things together that produces interesting results is a creative process that I consider to be an example of improvisation5. Similar to improvising as transcribing, Benson and I have a different understanding of improvisation5: when he describes "adding or subtracting measures, passages, or even complete sections" from a musical work, I imagine he is mostly referring to logistical decisions made for performance purposes (for example, when a conductor decides not to take the repeat of the exposition of the first movement from Beethoven's Eroica Symphony, thus shortening the performance by around three minutes). The process of using materials (not necessarily "adding or subtracting" but rather abducting them and putting them in different contexts altogether) that I am describing as a version of improvisation5 is more related to a creative working process outside of any specific performance context, always with the end goal in mind of arriving at more prolific and diverse improvisation skills. In reference to Benson’s subcategories 5a and 5b, the kind of creative process I describe here falls into the category of improvisation5b because it disregards whether or not the composer might expect or approve of such decisions.


Pianist-improviserBert Mooiman and Karst de Jong situate composing and improvising on a continuum in their article “Historically Inspired Improvisation” by claiming that the difference between the two rests “not so much in the way in which new music is created as in the degree to which the new music is registered” (Mooiman and De Jong 2016). Music can indeed be "registered" through notation or in performance (and if “registration” requires documentation, then recording technology can fulfill this function for performed music). The authors then continue by describing a continuum with completely free improvisation on the one end and notated music with no room for improvisation on the other:

At one end [of the fixation scale] is totally free improvisation, each sound being completely new and in which there is no reference to existing music whatsoever: ultimate spontaneity. The input of the performer is maximal. [...] At the other end of the scale is total fixation: a composed piece of music that has been written down in all conceivable detail and allows no freedom whatsoever to the performer whose input is reduced to zero. (Mooiman and De Jong 2016)

While it is certainly true that some musical scores offer the performer more
room for improvisation than others, it seems to me that the relationship between improvisation, performance, and composition can be better understood by considering the ratio of improvisation allotted to the composer and performer in any particular scenario. Considering that a notated score is the result of translated impulses (see the quote by Rzewski above), a score lying more along the fixed end of Mooiman and De Jong's scale arguably contains more translated improvisation from the composer, in turn allowing for less room for improvisation for the performer. But improvisation itself is ubiquitous and practiced by both parties. In other words, both performer and composer are improvising and the degree to which one engages in improvisation possibly restricts the degree to which the other can improvise. Rather than considering improvisation and composition at opposite ends of a fixation scale, or as guided by different techniques as Rzewski describes above, Benson’s position of considering both performance and composition as activities that each involve improvisation allows for a more complete understanding of how “improvisation in performance” and “improvisation in composition” work.


Cannot all improvisatory activities that Benson organizes according to “types and degrees” (Benson 2003, 26) be exercised either in performance or through notation? Although Benson sometimes uses performance and composition to differentiate between degrees of improvisation, for example between improvisation8 and improvisation9, the way he describes each degree (using “the basic form of the score” and “a particular form or style of music as a kind of template”) could apply to both composition and performance. Returning to the question of which degrees improvisation in performance exhibit when there is no explicit reference to a previously existing musical work, improvisation8 and improvisation9 could both function as suitable categories for such a scenario. In my opinion Benson's description of improvisation9  "us[ing] a particular form or style of music as a kind of template"  more accurately describes what is happening in this case than his description of improvisation8  "using the basic form of [a] score [... and] improvis[ing] within those confines" (ibid., 29).10



Improvisation6: It may 
 in at least certain cases  be difficult to make a clear-cut distinction between improvisation4 and improvisation6. Whereas the former attempts to render the score for other instruments or voices and may make some changes in the process, this sort is characterized by an explicit alteration known as "arranging." Arrangements can be minimal (and so very close to a transcription) or can be quite substantial  so much so that we might even begin to have questions as to whether we are still talking about the same piece of music. Note that transcriptions are often also arrangements (or vice versa), although arrangements are not necessarily transcriptions. One may, for instance, take a piece for piano and rearrange it for piano. (Benson 2003, 2728)

6, in the context of solo playing, engenders the question of the difference between transcription and arrangement; Benson refers to examples for more instruments or voice, but the issue is similar for solo music-making. The two birthday songs exposed here  Happy Birthday Georg! and Lang zal hij leven!  could serve as a useful starting point to consider how transcription and arrangement overlap, and what aspects of these pieces might extend into other degrees of improvisation. Both pieces contain a transcription of the original melody, which is then put through other improvisational processes. Because Lang zal hij leven! remains more faithful to the melody, I consider it to be closer to an arrangement and therefore more comparable to improvisation6 than Happy Birthday Georg!, which uses melodic material that is not just directly related to the original melody  exemplifying improvised activity more closely related to what could be called a "free fantasy" (akin to Benson's improvisation10 or even to what I later describe as improvisationx  the simultaneous presence of many degrees of improvisation, including those on Benson's list as well as other forms of improvisation that may have been overlooked).


Robert Levin's fantasy on themes by Mozart, presented in his lecture Improvising Mozart (which will be discussed in further detail in the chapter Between Listening, Forgetting, and Conceptualizing), serves as a useful example to clarify which aspects of the performance could be considered as arrangement or transcription, and which ones extend to other degrees of improvisation. After an introductory prelude, the passages in his fantasy that refer to the melodies by Mozart suggested by the audience hover between transcription and arrangement. When Levin plays a melody that compliantly relates to one that Mozart notated, this could be considered transcription; when one of his performed melodies references but does not comply exactly with one of Mozart’s notated melodies, this might be better represented by the concept of arrangement. The other passages – an introductory prelude and transition material between the passages that refer to Mozartian melodies – exemplify other degrees of improvisation (among others improvisation3 for example). Comparing Levin's performance, his two lectures Improvising Mozart and Composing Mozart, and Benson's degrees of improvisation can also clarify not just what Levin is doing while he improvises a fantasy but also what forms of improvisation he engages in discussion. During the lecture portion of Improvising Mozart preceding the extemporized performance, he almost exclusively discusses the embellishment of melodic material and extending figurative passage-work (both have been discussed here in the subchapter Improvisation2). In the lecture Composing Mozart, Levin shows how improvisatory the decision-making processes that are involved in composition are, for example by discussing how he came to find the material for the et vitam venturi fugue, part of the Great Mass in C minor, K. 427 that Mozart left unfinished (Levin Composing Mozart 2012, [1:09:251:10:54]). The examples of the birthday songs in Playing Schumann Again for the First Time as well as Levin’s Mozart fantasy confirm Benson’s warning that it is difficult to make a clear-cut distinction between improvisation4 and improvisation6; and, furthermore, that many degrees of improvisation are simultaneously active in any given performance.