More interesting from our current point of view however is the experience that modern students may very well be able to find some prolongation patterns on basis of the technical exercises they are familiar with since their childhoods – but that their fantasy soon runs dry. We are confronted here with the limited musical value of technical exercises that are only seen as muscular gymnastics, and not as potential music. For whom always practised his scales in the same meter, starting and finishing with a tonic on a strong beat, it is not easy at all to convert this scale on the spot to a different harmonic and rhythmic situation; and we can use a scale to prolongate one chord of the cadence, different types of arpeggios for one or two other ones, but what to do on the remaining harmonies? Czerny comes up with all kinds of imaginative patterns that look vaguely familiar, but which we never actually practised. What the modern player, who wants to improvise preludes in the style of Czerny, needs to develop, is the skill to bring variation to his improvised virtuosic passages. It may be true that a trained player can perform almost all music which he can really think – but this imagination of quick passages needs to be fed with the experience of actually playing them. This is not the same thing as automatizing more different patterns in the sense of being able to play them without thinking; on the contrary, what has to be developed is the quick thinking itself, the versatile but very precise imagination. A way of doing this is by offering a multitude of stylistic possibilities for such patterns – in order to become able to change them; what we need is a catalogue of virtuosic ornaments. The goal will be the development of a creative technique.


4.6 The organization of the catalogue


As the collection was growing, some way of organization became necessary that would be practical for a player, starting to work with this material. Grouping the patterns according to what exactly was embellished – a triad, a dominant seventh chord, a scale – seemed logical, but turned out to be not watertight; for, a chord could in theory be prolongated in many ways, not necessarily always using the members of the chord in the same order. Eventually the choice was made to understand the patterns as diminutions on a melody which on its turn prolongated a chord. This made it easy to discern between two directions: rising and  falling.

Historically Inspired Improvisation


Improvising on basis of 19th-century music making




Bert Mooiman

Royal Conservatoire, The Hague

May 2015


research coaches:

Dr Marcel Cobussen (Leiden University)

Prof. Dr Hans Fidom (VU University Amsterdam)







1. Introduction                                                                                         

2. Premises                                                                                                                                                      

3. Improvisation and fixation: score, context, interpretation and improvisation

           3.1 Introduction                                               

3.2 A student improvisation                                                                                                

3.3 Loci communes                                                                                                                        

3.4 Interpreting a score                                                                                                                            

3.4.1 The first level of meaning: the meaning of notational signs              

3.4.2 The second level of meaning: the meaning of a score                           

4. Ornamental patterns and instrumental technique                                      

4.1 Introduction                                                                                                                                

4.2 Complexity                                                                                                                                   

4.3 Practising technique                                                                                                           

4.4 Czerny as a source of inspiration                                                                         

4.5 A catalogue of patterns                                                                                                   

4.6 The organization of the catalogue                                                                      

4.7 Practising the patterns                                                                                                    

 5. Notes, selected bibliography, musical sources



1. Introduction


This research project does not stand alone. It all started during the academic year 2012-2013 with a small scale research under the flag of ‘KC teachers’ research’. This project, entitled ‘Classical Improvisation’, was intended to be carried on in the form of the Master Research project, presented here. At the same time however, the award of a 4-year grant from NWO, the Dutch governmental organization for the funding of scientific research, made it possible to scale up this research to a PhD project that would focus on the role of improvisation in 19th-century music making. This project officially started in August, 2014, under the supervision of Prof. Frans de Ruiter (Academy of Creative and Performing Arts, Leiden University), Dr Marcel Cobussen (idem) and Prof. Dr Hans Fidom (VU University, Amsterdam). More information can be found on

As a result, the presentation at issue should be seen as a part of the large PhD project, that is still in full swing. In order to show something more or less complete in itself, I chose one topic that, I believe, can be isolated relatively well: the occurrence of ornamental patterns in the work of Carl Czerny and contemporaries, and their relation with instrumental technique. A practical offspring of this topic is the compilation of a ‘Catalogue of Ornamental Patterns’. The accompanying text found its way as Chapter 3 of this essay.

Presenting just this text however would probably give a false impression of my goals. Therefore, the whole is preceded by introductory ‘Premises’ and a chapter that looks into the relation between improvisation and score (‘Improvisation and fixation: score, context, interpretation and improvisation’). The ‘Premises’ aim to sketch the perspective and the starting-point of this research, without discussing in detail the issues mentioned and the choices made. The chapter ‘Improvisation and fixation’ deals with a topic that will most likely be a central one in the final dissertation. Together with this introduction, these three chapters form the main text of this ‘master presentation’. 

Because of the nature of this text – being a part of a whole that is far from finished – no separate ‘research question’ is formulated. The question that overarches the whole project is: which was the role of improvisation in 19th-century music? This question will be approached in three ways, that show a permanent interplay: 1. by looking for and investigating concrete examples of improvisation, that have been described in 19th century; 2. by imitating such forms of improvisation (‘Historically Inspired Improvisation’, see below); 3. by applying these experiences to 19th-century scores.

The title I chose for this essay – Historically Inspired Improvisation - is the successor of the original title ‘Classical Improvisation’. Though the latter is a widely used term to indicate improvisation within the classical music world, it is, in my opinion, unsatisfactory. Calling a piece of music ‘classical’ already has its difficulties – but applying the same term to an improvisation, which by definition happens in the here and now, is in fact really problematic. The change to ‘Improvisation in classical style’ does not help very much: what is a ‘classical style’? And when we choose one: is it really possible to improvise in an ancient style? Does this idea do justice to the innate ‘situationality’ [1] of any improvisation? Analogous to the ‘Historically Informed Performance’ (H.I.P.)  practice, which understands itself more and more as historically inspired, I chose for H.I.I.


2. Premises


1. It is virtually impossible to speak about music in general terms. Especially in the Western world, the variations in contemporary ‘musicking’ are bewildering. Any claim about ‘music’ in general, or about related topics like performing, composing or improvising, is bound to be vitiated by some counterexample. It seems therefore to be important to sketch briefly the perspective of this research. Focusing on the Western music from the 19th century, it strongly concentrates upon art music from this period, and from the surrounding decennia. Folk music, for instance, only is discussed where a connection with such art music can be assumed. In this environment, scored music plays a determining role. It is impossible to imagine the extraordinary development of Western music during the past 1000 years or so without the potentialities of staff notation [2]. This being said, the characteristics of the music at issue, and of the corresponding scores, are quite specific: the importance of harmony for instance, both as a structural factor in music and as an area of music pedagogy, can hardly be overestimated.


2. This being an Artistic Research, the musical personality of the researcher is an intrinsic part of the research topic itself. My background as a classically trained pianist, organist and music theorist determines the way how I approach this issue. In this essay certain terms - ‘free improvisation’, for example – will be used in a way that is relevant for the context of this research, without extensive justification. Some readers might see this as a limitation, since – following the same example - a whole branch of contemporary improvisation seems to be neglected in this way.  However, I think that the opposite might be the case: by making unlimited use of the implicit and explicit knowledge existing in a specific field (in this case: the ‘classical’, score-based music making), it will be possible to show nuances that would remain unnoticed with a more generalist approach. I hope, and believe, that any supposed lack of objectivity will be compensated by the benefit of penetrating more deeply into a country that one already started (and thought) to know.


3. With the word improvisation I indicate the process (or the result) of creating (inventing) new music – or adding new elements to existing music - while performing [3]. The difference with composition lies in the available time: in composition, the creator has time to think things over and to change them before the piece is performed. A composition is usually (but not necessarily) codified in a score, to be performed on a different moment by the composer himself and / or by other performers. In an improvisation, the thinking time is always less than the musical time elapsed: there is no way of smuggling in extra time to ponder things. The definition of improvisation used here is similar to the one offered by Jean-Jacques Rousseau in his Dictionnaire de musique (1768). Rousseau didn’t yet use the term ‘improvisation’ in the modern sense; he used ‘Fantaisie’ instead. Under this heading, Rousseau writes: “FANTAISIE: Pièce de musique instrumentale qu’on exécute en la composant [4] – performing while composing. For Rousseau, composing essentially meant inventing and writing music [5]. Two-and-a-half centuries later, this term has become less easy to define. To evade possible pitfalls, my definition avoids the word ‘composing’.  Improvisation can involve creating a new piece (‘free improvisation’), but there can also be an improvisational momentum in the performance of a previously composed piece, even when not a single note in the score is altered [6]. This definition of improvisation is more restrictive than the one used by Bruce Ellis Benson, who writes that improvisation is “to rework something that already exists (that is, ‘conveniently on hand’), and thus transform it into something that both has connections to what it once was but now has a new identity”[7]. Benson also sees composing itself as a form of improvisation – the crucial element for him being the re-arrangement of already existing material. To me, this last implication seems to be somewhat counter-intuitive [8] - which is why I still prefer to distinguish between improvising and composing.


4. Improvising and composing are two ways of inventing music, that are not essentially different from each other: rather they can be thought of as parts of a continuum, a scale with as extremes the totally free improvisation on one side, and on the other side the completely fixated piece of music – both probably being hypothetical [Cf. the Time-Fixation Diagram]The role of the performer varies theoretically from maximal (in the free improvisation, where the performer is also the inventor) to not existing (in the completely fixated piece, which is so determined that it can be performed without a performer). A listener often cannot decide by ear whether music is being improvised or not; for him, the main difference between the two extremes is that the improvisation after its performance only continues to exist in his memory, whereas the completely fixated piece can be exactly reproduced at any time [9]. In this text, improvisation will mainly be described from the point of view of the performer.

5. The environment of this research is in the colourful area in between those extremes, where an improvisation re-uses already existing (sometimes composed) material, and pieces of music that are meant to last (or to be performed by a group of musicians) are notated in a musical score – with all its imperfections, possibilities of interpretation, and even features that give way to space for improvisation. It is the area of musical styles – a style being a musical ‘idiom’, that can be described as a collection of musical patterns and conventions, shared by more than one piece of music. The extent to which a piece of music is fixated in a score changes from style to style and even within a style. The assertion that improvising and composing are not essentially different processes implies that within styles which favour the fixation of music in musical scores – like 19th-century styles - these scores show patterns and conventions that are relevant for free improvisations [10] in the same style as well.


6. Though a reconstruction of an improvisation from the past seems to be even more out of reach than an ‘authentic’ reconstruction of a historical performance from a score, it can still be the desire of a performer to improvise in an idiom which his peers regard as historical. On basis of the assumptions above, compositions written in the style at issue can serve as a source of information for such ‘Historically Inspired Improvisations’. Elements of the stylistic vocabulary can be analysed and used in a creative way during improvisation. Other possible sources of stylistic information are, for instance, textbooks on improvisation, harmony, composition, basso continuo, etcetera.


7. Even though an improvisation today in a musical style from the past only uses an unavoidably modern view on this very style as a point of departure, it still brings the modern performer into a living contact with that historical style. Actively using stylistic elements for new music, instead of taking them for granted when performing from a score, deepens the understanding of the score (and of the music) itself. It helps the modern performer seeing the score as a link in the chain that forms the process of music making (now as well as then), instead of a code that has to be deciphered. Historically Inspired Improvisation may in this way lead towards an approach that is different from the already traditional H.I.P.: rather than seeking a historically justifiable translation of a score into sound, its goal is in (partly) restoring the original function of the score; rather than focusing on the musical Buchstabe it tries to uncover the Geist of the music.


8. The point of departure for this essay is the conception of a score as a historical text that was written and has functioned within a specific context. This context is many-sided; an improvisational tradition (e.g. a tradition of ornamentation), or a ‘performance practice’ can be part of it. With the time passing, (knowledge about) the original context of a score tends to get lost. Drawing upon Hans-Georg Gadamer’s notion of Horizontverschmelzung [11], interpreting the score is seen as merging our own ‘horizon’ with the original context of the score as much as possible. Not to embark upon this dialogue (Gadamer) with the score (by consciously not using what can be known today about its original context) may be artistically a respectable point of view, but means willingly excluding important parts of its treasures of meaning.




3. Improvisation and fixation: score, context, interpretation and improvisation.


3.1 Introduction


How do improvising and playing from a score relate? How is it possible for a musician to include improvisation in playing from a score, even without changing any note (as stated in the 4th premise)? In this chapter I intend to discuss these questions by examining the character of the musical score, its different levels of meaning, and the connections with improvisation. The starting point however will be a recorded duo-improvisation.


3.2 A student improvisation


In June 2014, a one semester course on ‘classical’ improvisation at the Royal Conservatoire in The Hague was concluded by a session, during which all students presented one or more improvisations in different combinations. The course was an elective, meant for third year Bachelor students. The programme of the course consisted of different types of group improvisations, stylistically ranging from (vaguely) a baroque style to some early 20th-century modal style. Since this was a course for beginners in improvisation, the stylistic focus remained rather open throughout the course: at this stage of the development of students, taking away mental barriers and waking up the musical imagination is often to be preferred to creating style-based restrictions, as often happens. Apart from this, every lesson started with exercises like playing on chord progressions, at first repeating itself (like a ground), later also more unexpected. Also, standard harmonic progressions like cadences, well-known sequences etcetera were used. Next to the harmony-based exercises, building a classical musical sentence was practised.

For the final presentation, I provided the students with a short ‘incipit’, a beginning of a melody. The idea was that the student would start a free improvisation with this incipit, with me accompanying on the piano. That is about all we agreed upon in advance – which is already a lot, as we will see.

The session was recorded with an i-phone. This is an example: an improvisation by an oboe student, using incipit no. 2. 

As can be heard at the beginning of the recording, the player changed my incipit in an interesting way – without doubt intuitively, but analysis in retrospect shows that he adds a motivic connection.

This oboe player had virtually no experience with improvising before he came to my course. We never practised with this theme or with anything similar. Still, the music sounds more or less ‘right’: there are no real clashes, and the form seems to be in balance as well. My experience is that this type of result is not an exception. Many classical musicians seem to think that this must be a matter of telepathy or magic, but I believe that it rather shows something about how music works.


3.4.2 The second level of meaning: the meaning of a score


Where the first level of meaning, the déchiffrage, in fact was about the meaning of the individual signs in the score, the second and ‘higher’ level of meaning is about the musical meaning of the represented sounds. This I would like to term the ‘meaning of a score’. The meaning of a score is not the same thing as the meaning of a piece of music; one could say that the latter is in fact one more level higher in the hierarchy of abstraction – the third level of meaning which belongs to the field of philosophy. This level is certainly worth thinking about (and has indeed been the subject of extensive debate), but is less relevant within the compass of this essay.


Recognizing loci communes

In the hierarchical differentiation between levels of meaning suggested here, we are now getting to the level that can be seen as the one where the identity of a musical composition is paramount – which makes it at the same time the level where a connection with the invention of new music is possible. The meaning of a score is in the relation between the represented sounds and the musical ‘objects’ to which they refer. With a musical object I mean here a Gestalt, in the sense of a pattern, a musical ‘way of putting it’, a musical topos or, as I termed it above: a locus communis. Attaching meaning to a score is recognizing such loci communes in the musical structures that are represented in the signs on the paper. The musician reads the score, imagines the sounding result – which in itself is still meaningless - and at the same time links patterns and gestures in the notated music to loci communes which he already knows by experience or which in another way resonate with what is familiar to him [24]. I call this process: interpretation; interpretation starts with recognition.

These loci communes can be so manifold that an exhausting description would be a futile enterprise. Crucial for this research is that they can function in two directions: they can be the guiding factors for the improviser, and at the same time they can be the ‘patterns’ that make the score meaningful to the interpreter. Very important though is that they guide or narrow down the possibilities which the lack of fixation in a score leaves to the performer. For example, we concluded that the rhythmical notation in most scores is far from mathematically exact compared with the desired sounding result. This implies that the musician, performing from the score, has a certain freedom in timing. But not just anything feels right to the performer, or convinces the listener: the performer’s choices of timing only sound ‘organic' [25] when they make sense in relation to an interpretational locus communis.

This might evoke questions: communis to whom, for example? If a locus communis is a shared structure, a common musical expression, then it must function in a specific situation of music making, in other words: it must be shared by a number of ‘musickers’. This makes it part of the context of a score, or, borrowing a term from Hans-Georg Gadamer: it belongs to its Horizont. When the distance between the reader and the score increases – in time or culturally – the horizon of the reader starts to differ from the horizon of the score. As a consequence, a locus communis can become a locus exclusus, no longer shared by musickers. Or the reverse might happen: the reader recognizes a locus communis that is communis to him and his contemporaries, but would have been unthinkable within the horizon of the score (at the time of composition). Examples of both situations will be discussed briefly in the next paragraph.

It is often very hard to express the character of a locus communis in words – this might be the reason why musicians so often sing passages to each other when they want to make clear how they imagine them performed, or why orchestras usually hate conductors who talk a lot.


Examples of loci communes

Loci communes can for example be musical genres, like Carl Czerny also describes them in the later chapters of the Anleitung zum Fantasieren [26]. When I take a piece of music to be a representative of the genre ‘Song without Words’ (not mentioned by Czerny), this immediately has implications for aspects like tempo, mood, tone colour, phrasing, rubato, balance, etcetera. When my association in a 19th-century composition is ‘Fugue’, I will play a different kind of rubato from when I see the piece as a ‘Phantasy’.

However, there is still a difference between having an idea about the general rhythmic character of a genre and knowing it so well that you could make your own example on the spot. In chapter four of the Anleitung, Czerny describes how the pianist can turn any ‘theme’ (not a theme in the modern sense, but rather an ‘incipit’, or even just a motif) into a beginning of a piece that suits one of the genres that are ‘usual on the pianoforte’ [27]. What follows are a few bars in every genre mentioned. Surely, this information will not always be sufficient for a modern reader.

But loci communes can be much more detailed: I can see a melody in a French baroque composition as an example of a melody which typically needs a specific type of inégale playing; moreover, I can interpret a melody by J.S. Bach likewise, even though I am not certain whether Bach would have done that himself. Not all loci communes are necessarily related to the direct environment of a composition. When Bach’s music is sometimes called ‘motoric’, that might be a later locus communis being connected with his score. And this is certainly the case when 21st-century music lovers think that Bach’s music ‘swings’.

An important group of loci communes are dance movements. Recognizing the Viennese Waltz in a score by Johann Strauss implies stepping into a performing tradition which features a specific rhythmical shift by playing the second beat slightly early. Who plays the three beats in the bar straight, simply doesn’t play a Viennese Waltz. He still performs the score – but misses an important part of its meaning.

In Dvořák’s Dumky Trio we find a passage in the first movement which shows a very specific kind of accompaniment in the piano part (please click to enlarge):

From the same concert with Salonmusik, two examples of improvised preludes:

And finally an example of a performance with hardly any changed notes, but with an improvised timing:


The Urtext-paradigm

Though this may not be the proper place to speculate about the reasons why improvisation gradually disappeared from classical music during the 20th century, two possible causes are connected with what was discussed above, and will therefore be mentioned briefly. One is that the (partial) unpredictability of a performance from a score with improvisational elements is hard to reconcile with the striving for perfection that belongs to a recording. Not only are improvised actions more likely to fail (in terms of containing mistakes) than thoroughly rehearsed and practised ones; also the possibility to cut and paste recordings (with the same perfection as a goal) demands a musical performance that repeats itself as much as possible. And so music recordings can be seen as another way of fixating music (performances) that influenced the musical practice [35]. The other cause for the disappearing of improvisation I see is in the treatment of scores. A heritage from the 19th century is the view of a composer as a genius, whose scores sometimes attain the status of holy texts. With the time passing, however, the original loci communes, still familiar to the contemporaries for whom the musical texts were meant, were forgotten gradually. As a result the scores tended to be taken literally – a phenomenon that can be found in the real holy texts of book-based religions as well, by the way.

Consequently, the 20th and 21st centuries saw the publication of a large number of Urtext-editions of classical music, at first from the 18th, and later also from the 19th century. These editions meant to provide the exact score as intended by the composer, stripped of all later additions by publishers, editors, and others; the corresponding dogma was that the oldest version is usually the best one. Apart from this dogma being questionable, in some cases the construction of such a definitive version was literally impossible because the composer himself produced different equal versions (Bach: St. John Passion; Chopin: Nocturnes). Nonetheless, many of these are beautiful editions, and of course a version of the score that resembles the original one (or something like that) is to be preferred above a version with all kinds of annotations by musicians from a past with which we are no longer connected either [36] – for so old are the compositions that we are still playing!

Where it goes wrong, in my opinion, is not in the idea of an Urtext-edition as such, but in the way how they are often being read. The awe of the score, that many 21st-century musicians have, combined with an ignorance of at least an important part of its meaning, leads to understanding a score as a text that should be taken literally. The heyday of this view on the score can roughly be dated around the 1970’s – 2000’s. Interestingly, also the idea of performance practice (Aufführungspraxis) and HIP was not completely immune against this opinion. I would like to term this attitude: the Urtext-paradigm.

What is wrong with the Urtext-paradigm is that it is based upon a misunderstanding. The score by far does not contain all information because it functions in relation with a context, of which the mentioned levels of meaning form an important part [37]. A score gets its meaning only in relation with this context. Taking the score too literally means taking a simplification for the real thing.   

For first experiences with an improvisational approach of scores, please click on this link for the Brahms concert. 



4. Ornamental patterns and instrumental technique


4.1 Introduction

One of the loci communes in early 19th-century piano music is what I would like to label the ‘ornamental style’: the elegant and dazzling virtuosic style from, roughly, 1810 – 1850, that features ornamental passage work, usually on a standing chord. This style seems to be well-known: even though many of the then famous pianist-composers have been forgotten, it is still very vivid in the works of e.g. Frédéric Chopin. Recognizing this locus communis in a score therefore seems to be unproblematic; however, using it actively and creatively in an improvisation turns out to be a very different matter. Many pianists, trying to follow Carl Czerny’s advices about improvising a simple prelude [38], find out soon enough that actually improvising such passage work is not easy at all.

One would expect that the reason for this is in the fact that it is difficult to improvise fast music. However, this chapter sets off from the experience that it is in principle possible to improvise technically demanding music. This, and the improviser’s desire to expand his skills in this field on one hand, and Carl Czerny assuming such skills in his Anleitung on the other hand, lead to using the first chapter of Czerny’s book as a source of inspiration for the development of a new approach to piano technique. This newness is found in the attempt to turn the hitherto often musically passive character of purely technical exercises into an approach that wants to use technique as a stylistic tool for creating music. By extracting from the works of Czerny and contemporaries a collection of ornamental patterns, an important locus communis of the style of the early nineteenth-century virtuosi is analysed and extrapolated. The study of these patterns not only makes it possible to imitate improvised preludes in the style of Czerny; it also serves a higher goal by compelling the pianist to use his musical imagination actively and to become very flexible in bringing his imagination to the keyboard. The patterns themselves being trans-idiomatic, they are expected to be useful for other instrumentalists and singers as well, though this still has to be researched in practice.

A condition for the effectiveness of the suggested approach seems to be the preliminary acquisition of a sufficient fluency on the keyboard. Therefore, this approach is not meant to replace the first stages of instrumental technique; rather, it should be seen as a follow-up for the accomplished player.



4.2 Complexity


There are similarities between creating music and playing chess. In both ‘games’ different levels of organization play a role: on one hand a general, overarching plan that can roughly indicate the general direction (‘form’, in music) without determining in detail the level of the individual moves; on the other hand there is the ground level of the individual choices that have to be made from moment to moment, and where both in chess and in music similar mechanisms apply: a move prompts a range of possibilities, choosing one of them narrows down this number but at the same time creates new ones, etcetera. If an ordinary chess game, or even more a game of correspondence chess, resembles in this way creating a musical score, leaving time to think things over and to compare different consequences (and, usually, notating the game), a musical improvisation is more like playing fast chess, in which all decisions have to be made on the spot. Indeed, to many players musical improvisation feels like very fast thinking, without any time for reflection upon one’s decisions – let alone for correcting them.

On basis of this phenomenon, one would expect (free) musical improvisations to be less complex than notated compositions, like fast chess games tend to be less balanced and usually contain more mistakes. Indeed, experience seems to confirm that improvised music is often simpler than its notated counterpart. Recordings of Olivier Messiaen’s improvisations definitely show a lesser complexity in the treatment of harmony and modality than his compositions. Now, complexity can obviously be of many kinds: apart from harmony and tonal organization it can relate to form, syntax or texture, but also to instrumental or vocal virtuosity. Given the fact that a performer needs to spend many hours of practising to master technical difficulties in a piece of composed  music, it seems to be obvious that such virtuosic passages cannot be improvised. Interestingly, experience shows that this is not always true. It turns out to be possible to include in an improvisation passage work that, when occurring in a composition that one wants to perform, would need thorough and time consuming practising. An example can be found in my recording of an improvised sonata: in the slow movement, a passage occurs in which the right hand plays quick scales and scale-derived runs across many octaves that would normally certainly be considered a technical difficulty.

The level of technical perfection here is, I believe, satisfactory. How is this actually possible [39]?

It makes one think about what is actually happening when a musician is practising such passages in a composed piece. Is it maybe in reality a process of automation? Is a player in fact ‘waiting’ until the difficult passages ‘play themselves’? The influential Austrian – Dutch violin teacher Oskar Back (1879-1963) is reported to have said that a performer only really masters a musical work when he is able to play it five times in a row without a single mistake [40]. Indeed, being a performer of classical music also means being patient for the mind and the body to get ready for the performance. Or is it true that one can play almost whatever one likes, as long as the music can be thought (imagined) in detail? In that case the process of practising a composition would be essentially a process of appropriation, of making somebody else’s score to one’s own music. This perspective seems to be confirmed by the

well-known phenomenon of mental practising, in which somebody learns a piece or even masters a complex technical difficulty by imagining the music only, without touching the instrument [41]. Interestingly, also here a parallel with the chess game exists in the so-called blindfold chess.


4.3 Practising technique


In this connection, it should be mentioned that musicians usually not only practise the compositions they want to perform, but also separate technical exercises. This habit seems to be connected historically with the upcoming of more elaborate passage work in pieces of music from the Viennese classical period on. Surely it cannot be maintained that virtuosity, in the sense of being able to play very fast passages, didn’t occur in earlier music – on the contrary, startling examples like Sylvestro Ganassi’s La Fontegara (1535) or written out ornamentation in 17th-century instrumental and vocal music can be amplified easily. What changed during the second half of the eighteenth century, however, was a drastic simplification of the harmonic language, which focused much more on the three principal harmonic degrees. This resulted in a much slower harmonic rhythm, leaving space (and also creating the desire) for virtuosic passage work to ‘fill’ the sustained harmonies. Connected with this tendency we see an increase of range for many musical instruments, especially the newly developed pianoforte, of which the keyboard compass grew from five octaves around 1780 to more than seven at the end of the 19th century.

The new aesthetic gave rise to a new kind of virtuosity, based upon regularity and fluency (rather than the baroque ideal of a rhetorical understanding of music); also, this passage work showed a much larger degree of standardization than can be found in earlier music. Soon books with technical exercises started to be published, and practising exercises like scales and arpeggios became a standard issue for the virtuoso. The habit to practise technical exercises next to ‘pieces’ endured until our times. A book like Charles-Louis Hanon’s Le Pianiste Virtuose en 60 Exercises (Boulogne sur Mer, 1873), which is still used today, is based upon the assumption that the fingers should be trained to perfect evenness, which will eventually enable the pianist to play pieces of any difficulty (‘Si les cinq doigts de chacune de nos mains  étaient tout à fait également exercés, ils seraient aptes à exécuter tout ce que l’on a ecrit pour cet instrument, et l’on n’aurait plus devant soi qu’une question de doigté dont on trouverait bien vite la solution’ [42]. The idea is that the pianist, simply by repeating again and again the exercises from this book, will develop the fingers individually to equal strength and independence from each other. (It has to be said that already Chopin strongly opposed against this view [43].) In the 20th century, more developed and subtle (some would say: intelligent) views on the learning process became common property. An important book by Marguerite Long, Le Piano (1959), is organized around specific technical difficulties, and in this way offers a fairly complete overview of the possibilities of traditional classical piano virtuosity up to, say, Ravel. The material presented in her book pretty much corresponds with what most modern pianists think of when they speak about practicing ‘technique’: scales, arpeggios, double notes, octaves.

A brief and unofficial inquiry among conservatoire piano students and teachers showed that opinions about the value of practising separate technical exercises (at least at this level of pianism) differ [44]. Some teachers abandoned them altogether, others expect the students to do such exercises by themselves. The prevailing impression is that, if such exercises are used by more or less developed pianists at all, they are seen as a kind of gymnastics. The goal remains the performance of the masterworks, written by (usually dead) composers.

One striking thing about technical piano exercises is their uniformity. Scales and arpeggios are usually practised in the same rhythmic form, starting and finishing on a tonic, in a square meter, every day again. Surely in this way certain important issues are being developed, e.g. evenness of tone, contact with the instrument, and indeed: gymnastics. One extremely important faculty however will hardly be used in this way of practising: the musical imagination. For other instruments other factors apply as well, like intonation and tone production. For singers the situation is different again, since they develop their instrument itself by means of technical exercises. Despite such differences, a common property is the fact that the exercises serve a goal that is connected with performing a score, not with inventing new music. Already in the performance of classical masterworks the limitations of this approach become obvious: as every pianist knows, one can diligently practice his scales, but in a Beethoven sonata inevitably the same scales will be used just a tiny bit differently – which will make them suddenly twice as hard. The classical musician who practises technique in this traditional way is like a swimmer doing movement exercises on dry land: undoubtedly useful in a certain way, but hardly an equivalent to practising the swimming itself. Just as a swimmer needs water to improve, the musician needs to awake his imaginative faculty.


4.4 Czerny as a source of inspiration


Returning to the experience, that improvised virtuosic passage work is very well possible in the heat of the moment: of course, as an improviser I am curious how this skill can be improved. As argued above, the traditional technical exercises will not be of much help here. Interestingly, some inspiration might be found in Carl Czerny’s Anleitung zum Fantasieren auf dem Pianoforte [45], of which the first three chapters are dedicated to improvising preludes and cadenzas.

The first chapter of the Anleitung deals with the improvisation of simple preludes, about which Czerny writes that they should be played before a composed piece as a matter of course, to warm up, to try the instrument, and to prepare the listeners for what will follow:

‘Es gehört zu den Zierden eines Klavieristen, wenn er, besonders in Privatzirkeln beim Vortrag von Solostücken, nicht gleich mit der Composition selbst anfängt, sondern durch ein passendes Vorspiel die Zuhörer vorzubereiten, zu stimmen, und auch dabei die Eigenschaften des ihm vielleicht fremden Fortepianos auf schickliche Art kennen zu lernen, im Stande ist.ʼ[46]

Only before a very serious composition like Beethoven’s Appassionata sonata, he writes, a prelude would be ill at place [47]. Also when the pianist is playing a solo concerto with orchestra, which has a long orchestral introduction, it would be strange when the pianist would prelude before the orchestra starts to play [48]

Playing a prelude before starting ‘the real thing’ was already an old tradition in Czerny’s days: in fact it is one of the clearest find-spots of improvisation in music. From the earliest notated organ music to the Northern German baroque Praeludia, from the French Préludes non mesurés to Hotteterre’s  advices on how to improvise a prelude on the flute: preludes were everywhere.

The above mentioned Préludes non mesurés served as opening movements to the early French baroque harpsichord suites. In these pieces, composers like Louis Couperin and Jean-Henry d’Anglebert notated only the pitch, not the rhythm. Carl Philipp Emanuel Bach’s Freie Fantasien could be seen as late representatives of this genre. Compared to such preludes, the simple preludes in the first chapter of Czerny’s book are structurally remarkably basic. The starting point is always an authentic cadence – be it just the progression V – I, or a more elaborate form like I – VI - #IVdim7 – I 6/4 – V – I. The following example is taken from the beginning of chapter 1 of the Anleitung:


4.5 A catalogue of patterns


The idea of a catalogue of patterns may ring a bell with musicians who are familiar with ornamentation in early music (and in jazz, of course). Books like the already mentioned La Fontegara by Sylvestro Ganassi or the Tratado de Glosas (1553) by Diego Ortiz offer exactly that: different ways to ornament elementary melodic progressions. Where the perspective in such 16th-century works seems to be primarily melodic, in late baroque publications like the Amsterdam edition of Arcangelo Corelli’s violin sonatas op. 5 (1700) or Georg Philipp Telemann’s Sonate metodiche (1728) the ornaments are clearly more based upon the harmony. Can we assume that Czerny was also thinking in this way? Did he see different virtuosic patterns as ornamental ‘licks’? Did an early 19th-century pianist practice technique primarily with the aim of collecting material for ornamentation?

This question is difficult to answer. An argument in favour of this view is the fact that being a musician in his days must have had a different meaning from today. Composing, improvising, performing: they were all part of  music making, and not separated disciplines like in today’s classical music practice. In that context it is hard to imagine technical exercises being understood purely as a kind of gymnastics for the fingers, meant to support the performance of scores, as  is the modern way of seeing it. On the other hand Czerny nowhere comes up with a comparable list of patterns for improvisation himself. But then, neither did Corelli or Telemann! The only publication by Czerny which seems to come close is the second part of the Vollständige Theoretisch-practische Pianoforte-Schule op. 500, entitled Von dem Fingersatze. In this book he does present a list of pianistic formulas, however with a different goal, namely showing the principles of fingering. Nonetheless this book makes one suspect that there is more to those innumerable patterns than just being exercises for the fingers: after showing many different ways of arpeggiating a major triad, Czerny finally shows the reader a last way, which he calls ‘modern’: and we recognize this type of arpeggiation because it is the very motif on which Chopin’s study op. 10 no. 1 is based [50]! Czerny’s way of putting it suggests that a motif, that for us may be musically empty in itself, did have musical connotations at that time. An early 19th-century pianist could still surprise listeners by introducing new musical effects, as is confirmed by the anecdote about Thalberg who caused a stir with his technique of playing with the thumbs an (undoubtedly popular) melody in the tenor register, surrounded by clouds of arpeggiations in both hands [51].The following example, showing this ‘third hand technique’, is taken from his Fantaisie sur l´opéra «Moïse» de Rossini (1839):

The stylistic framework of the catalogue was in the first instance defined by the work of Carl Czerny: collecting models started with material from the Anleitung itself. Other books by Czerny were included as well, like the Kunst des Preludierens op. 300 and the already mentioned second part of the Pianoforte-Schule. Czerny can be seen as a representative of a group of pianist-composers who frequented the (often newly built) stages of European concert halls [53], and created a furore with their dazzling technique and easily digestible, elegant style, favouring genres like potpourris of famous opera tunes. Comparing compositions of these ‘technicians’ (Die Technischen, as Oskar Bie [54] called them) with the models, acquired from Czerny, shows that the patterns which can be found in Hummel, Clementi, Moscheles, Kalkbrenner and others have a strong similarity with the ones Czerny used. It has to be said that this very style lost much of its shine in our days, though there seems to be something like a Hummel-revival going on – thanks to the rediscovery of the earlier pianofortes. It cannot be denied though that this style had a deep influence on the work of a composer who never lost his popularity until now: Frédéric Chopin. Indeed, also from his work such models can be taken. Many of the ornamental patterns in his music are strikingly similar to Czerny’s; it is the far greater harmonic complexity in Chopin which calls for more chromatic patterns – these patterns were included in the catalogue as well.

A last group of works to be consulted were compositions for other instruments than the piano. Interestingly, the same types of patterns appear in the literature for most other instruments, and even in the vocal repertoire (for instance in the vocal cadenzas that were notated by famous singers like Maria Malibran and Pauline Viardot). Especially those vocal cadenzas look quite astonishing today because of  their sheer virtuosity. The patterns are trans-idiomatic.

This final set-up is also more abstracted than the original literal copies from the actual music. The meter remains undecided, but the notation allows for indicating strong and weak notes by grouping them using eighth note beams, so that the first note of a group will always be relatively strong. This is especially important where leading notes can be used as passing notes or as appoggiaturas, which makes a very different musical effect. Finally, the notation in eighth notes is to be seen as relative: the notes can be of various rhythmic values, as a group as well as in relation to each other. 

The ordering that I chose is of a growing complexity, pretty much like the diminutions in La Fontegara. Some principles become very clear:

1. The ornamental patterns are almost exclusively stressing, embellishing or playing around the members of relatively few chord types: the major and minor triads, the dominant and the diminished seventh chords. Those chords are never blurred; on the contrary, they remain crystal clear. This is because every member of the chord is usually diminuted upon in the same way. This stress on the simple chords echoes the simplification in harmony which belongs to the classical era.

2. The standard way to make a single tone more important is the turn. Especially in Chopin, this is also used to shape an ornamental bridge between two melody tones: the first one (usually the lower one) is stressed by a turn, then the gap is filled with either an arpeggio or a diatonic or chromatic scale, and finally the goal tone is reached elegantly from above. This type of bridge, which matches the counterparts from the belcanto repertoire closely, goes together with a standard dynamic build-up, namely a diminuendo, which is contrary to the ‘natural’ tendency of a rising melody. It may be not exaggerated to assume a stylistic watershed between the pre-1800 and 19th-century musical styles in this respect; a diminuendo on a rising melody in Bach would stylistically not be very likely, a crescendo on the ornamented bridge mentioned above would sound blunt.

3. The simple principles which are at the root of the catalogue make it possible to extrapolate the models that were actually found and insert the ‘missing’ ones. That they do not occur in the relatively small group of compositions consulted, does not mean that they never occur; it is tempting to continue searching the repertoire for those suspected models, but then the list is not meant to give a complete collection, but rather to provide material which could be used in this stylistic environment – and changed accordingly.

4. In the musical literature consulted, virtuosic ornamental patterns are almost exclusively limited to the right hand of the pianist. This is remarkable, because the etudes which are based upon similar patterns – the ones by Czerny himself are a good example – famously train both hands: often first separately, then in combination, a procedure also to be found in Beethoven’s 32 variations WoO 80 (var. 1 – 3). Maybe here the different goals of an elegant prelude or ornamented melody and a didactical practice piece like an etude show: the etude serves to develop the gymnastic aspect of virtuosity, whereas the melodic ornamentations have an aesthetic goal.

4.7 Practising the patterns


Crucial in practising the patterns from the catalogue is not to see them as goals in themselves, in the same way classical scores usually are understood, but to improvise with them: to play with them, to put them into various harmonic contexts, to change them rhythmically. In the first place, this means transposing them to all keys. For the catalogue, the patterns were usually transposed to C major; it is up to the player to not only transpose them to all other major keys, but also to try the possibilities in minor and on diminished chords - it should be remembered that some patterns were found in a minor mode in the composition where they occurred! [For first experiences of students working with the catalogue, please click this link.] 

A systematic way to practise transposing the patterns is by using a modulation scheme. This method may seem to be the contrary of an improvisatory approach, but as it allows for at random combinations of different patterns, it still prepares for the flexibility needed to improvise Czerny’s chapter 1 preludes, because also there the player has to change patterns quickly in between chords. 

The transposing of models already teaches an important lesson about the benefit of practising the catalogue, that goes beyond the mere imitation of a historical style. Unavoidably, the patterns will get into very uncomfortable keys. It is not to be expected that such difficult situations will occur in any composition, let alone that an improviser would use them (unless by accident). The reason why this is uncomfortable is in the fingering: when within a specific pattern in, say, C major, the fingers can move in their natural order (from thumb to little finger) in all positions of the pattern, this most likely doesn’t work in D flat major because the different arrangement of white and black keys makes this physically impossible. The player has to find alternative fingerings, which are much less ‘natural’. Especially in high speed this model suddenly becomes extremely difficult to play. Now, if an accomplished player would practise exactly this difficulty for a couple of weeks, of course he would more or less solve the problem for himself – but only for exactly this situation, including rhythmic and harmonic embedding. The proposed method of ‘improvisatory’ practising however tries to avoid precisely this! By constantly changing the patterns, they will not easily become automatized, and the player will be forced every time to come up with ad hoc solutions. He will have to become able to find effective fingerings on the spur of the moment. In this way, he will train a flexibility which is exactly what he needs when improvising technically difficult passages, also when the situation happens to be slightly different.

It is especially those ‘uncomfortable’ fingerings that are so important for waking up the musical imagination. When the fingering is easy and comfortable, like in the C major pattern in the example, the music almost automatically follows from the physical movement in the hand. Indeed, one could even assume that the physical aspect is – at least partly - the motor of the process: a general idea or imagination of the desired result will be sufficient, the fingers ‘play themselves’. This effect, by the way, is quite well-known in improvisation, for instance when someone improvises, say, an accompanying structure consisting of quick arpeggiations - e.g. in the beginning of the improvised sonata mentioned above:

Surely, here the detailed structure was determined mainly by the physical motor, the imagination providing the more general framework. The phenomenon of the physical aspect influencing the detailed musical content reminds us also of quite a lot of virtuosic music which is in reality less hard than it seems to be for this very reason, like piano compositions by Rachmaninov, and, perhaps in a lesser degree, Liszt. Also the work of Messiaen shows the influence of the properties of the pianist’s or organist’s hand, not only in virtuosic passages but also in vertical chord structures. - How completely different the situation is when the fingering is uncomfortable! When the fingers no longer seem to play themselves, they have to follow the conscious ‘commands’ from the mind. For the improviser, such ‘commands’ have to be rooted in the musical imagination: where else would the music come from? Interestingly, when one practises an ‘uncomfortable’ position like described above, one clearly notices one’s imagination at work, especially when the patterns become more complex melodically (e.g. by incorporating unprepared leading notes or other unexpected leaps). The player has to be very concentrated on performing the pattern he wants, and on quickly finding an appropriate fingering as well. In other words: he has to dive into the musical swimming pool.


[1] Fidom, H., Muziek als installatiekunst. Orgelpark Research Reports / 2; Amsterdam, 2012. (accessed 9 September 2014)

[2] Even when we assume that much music was improvised especially until the 18th century: after all, also Tinctoris’ super librum cantare still needed a book…

[3] Cf. Fidom, H.: ‘Improvisation: the emancipation of an ancient musical skill.’ In: Peeters, P. (ed.): The Haarlem Essays. Bonn, 2014; 252

[4] Rousseau, J.J.: Dictionnaire de musique. Paris, 1768; 218

[5] Rousseau, ibid: 109

[6] But only when this happens ‘on the spot’; a completely predetermined ‘interpretation’, with no room for spontaneity, cannot be called improvisational, even when the score is treated with a lot of freedom.

[7] Benson, B.E.: The Improvisation of Musical Dialogue: A Phenomenology of Music. Cambridge, 2003; 45.

[8] An important point for Benson is the avoidance of the notion of a ‘creatio ex nihilo’ (cf. Benson, B.E.: ‘In The Beginning, There Was Improvisation’. In: Improvisation - Musicological, musical and philosophical aspects. Orgelpark Research Report #3/1; Amsterdam, 2013. (accessed 7 January 2015); § 23-55.  Though this construction put upon the term creation is certainly understandable from the perspective of the Western theological tradition, I still prefer to connect with other meanings of the Latin word creare: to cause, to bring forth, or even to elect – all of them meanings which do not contain the ‘ex nihilo’-element. In French, where the meaning of créer is similar to that of creare, a création can also be the first performance of a piece.

 [9] A recording of an improvisation not so much makes it possible to repeat the performance – rather it helps to recreate the listening experience.

[10] In the context of 19th-century music I call a ‘free improvisation’ an improvisation which does not use any fixation in a score.

[11] Cf. Gadamer, H.-G.: Wahrheit und Methode. (Gesammelte Werke, vol. 1). Tübingen, 1990.

[12] Cf. Caplin, William: Classical Form. New York, 1998.

[13] This does not exclude e.g. Augenmusik: the fact that the score represents sounding music does not prevent it from gaining a certain aesthetic autonomy itself. 

[14] The existence of a melody that has not been notated exactly can be located in the memories of people who know the melody.

[15] An interesting variant of this view is the idea that a composition up to the 18th century, notably for organ, mainly served as an example for improvised music (Cf. Fidom, H.: ‘Improvisation: the emancipation of an ancient musical skill.’ In: Peeters, P. (ed.): The Haarlem Essays. Bonn, 2014; 357 – 358).

[16] It is quite possible to link this strong emphasis on the actual sound to a development in 20th century composition which also focused upon the sound an sich (e.g. Varèse, post-war serialism, Cage, electronic music).

[17] Swenson, Edward E.: The History of Musical Pitch in Tuning the Pianoforte. (Accessed 26 March, 2015)

[18] Cf. the work of Kolja Meeuwsen on the relation between early 20th-century performance practice for string players and the upcoming recording industry.

[19] Even the descriptive value is not beyond doubt, as any musician who ever tried to determine his favourite metronome speed for a specific piece can testify: usually the chosen metronome mark feels less right the next day.

[20] Cf. The Pupils of Clara Schumann. Pearl Gemm CDS 99049

[21] Signs indicating the articulation mainly influence the duration of the tone.

[22] In the practical sense of the reading of a score by a musician.

[23] Like Schumann wrote: “Du musst es so weit bringen, dass du eine Musik auf dem Papier verstehst“. Schumann, R.: Musikalische Haus- und Lebensregeln. (accessed 9 January 2015)

[24] This fragment (3.4.2) makes use of ideas that were formulated earlier in: Mooiman, B: Sporen in het zand. Visie op het theorieonderwijs aan het Koninklijk Conservatorium te Den Haag. Unpublished, 2003.

[25] A term that is often used by musicians themselves.

[26] Czerny, C.: Systematische Anleitung zum Fantasieren auf dem Pianoforte op. 200. Wien, 1829; reprint Wiesbaden, 1993.


[27] Czerny, C.: ibidem, 36 – 37. “Jedes Thema, ohne alle Ausnahme, und wenn es nur aus zwey willkührlichen Tönen bestünde, kann, vermittelst einiger Änderung im Takt und Rhythmus, als Anfang zu allen Gattungen von Compositionen dienen, die in der Musik existieren. Nehmen wir einstweilen folgende auf dem Pianoforte üblichen Gattungen an: a. Allegro (ungefähr, wie das erste Stück einer Sonate); b. Adagio (im ernsthaften Stil); c. Allegretto grazioso (einfach, oder mit Verzierungen in der galanten Schreibart) [sic!]; Scherzo Presto (à Capriccio); e. Rondo vivace; f. Polacca; g. Thema zu Variationen; h. Fuge (oft auch Canon); i. Walzer, Ecossaise, Marsch u. drgl.

[28] Gjerdingen, R.O.: Music in the Galant Style. New York, 2007.

[29] Gjerdingen, 7.

[30] Which also shows that learning a piece for an important part means memorizing it…

[31] An example is a remark Shura Cherkassky once made in an interview, stating that during a performance he sometimes liked to decide on the spot whether he would play a passage very tenderly and quietly, or that he would give the audience a fright by suddenly playing it very loud. This evidence of late romantic capriciousness could be seen as an improvised change of locus communis.

[32] Of course, improvised alterations do offer many more possibilities to react to the other musicians, the audience, the room, the instrument etc. than rehearsed ones.

[33] Peres da Costa, N.: Off the record: Performing Practices in Romantic Piano Playing. New York, 2012.

[34] Scott, A.: Romanticizing Brahms: Early Recordings and the Reconstruction of Brahmsian Identity. (accessed 3-5-2015)

[35] Fidom, H., Muziek als installatiekunst. Orgelpark Research Reports / 2; Amsterdam, 2012. (accessed 9 September 2014); § 22.

[36] Unless one is especially interested in what Czerny had to say about Bach, or Schnabel about Beethoven, of course.

[37] Though obviously, a context encompasses much more than the levels of meaning addressed here.

[38] Czerny, C.: Systematische Anleitung zum Fantasieren auf dem Pianoforte op. 200. Wien, 1829; reprint Wiesbaden, 1993; 5 – 14.

[39] Cf. the notion of Ideo-Kinetics, as developed by L. Bonpensiere in New Pathways to Piano Technique (New York, 1953).

[40] But then, Back appears to have suffered from a severe stage fright, which might have been the real reason for his excessive demands to technical perfection.

[41] This method was widely recommended already by 19th and early 20th century piano pedagogues like Leschetitsky, Leimer, Neuhaus and Kentner. The effectiveness seems to beconfirmed by recent neurological research, though an explanation through mirror-neurons suggests that this process only works when it is based upon a previously acquired skill.

[42] Hanon, Avertissement.

[43] Eigeldinger, J.-J.: Chopin vu par ses élèves. Paris, 2006; 45.

[44] As they did in the past: H. Neuhaus writes that his teacher, L. Godowsky, never practised scales separately. Neuhaus, H.: Die Kunst des Klavierspiels. Köln, 1967; 5.

[45] Czerny, C.: Systematische Anleitung zum Fantasieren auf dem Pianoforte op. 200. Wien, 1829; reprint Wiesbaden, 1993.

[46] Czerny, 5.

[47] Czerny, 15.

[48] Czerny, 5.

[49] Czerny, 7.

[50] Czerny, C.,Vollständige Theoretisch-Practische Pianoforte-Schule op. 500, part 2: Von dem Fingersatze. English edition: On Fingering. London, 1839; 109.

[51] Hildebrandt, D.: Pianoforte. München, 1985; 145-146.

[52] Czerny, 8.

[53] Though Czerny himself was in fact an exception: early in his life, he specialised as a teacher.

[54] Bie, O.: Das Klavier und seine Meister. München, 1898; 161.

3.3 Loci communes


For, an improviser is not as naked on stage as one might think. As soon as he decides to tell something musically, there are all kinds of factors guiding him, that together form a kind of reference system. In this case a couple of hints was made by me as a teacher. First of all, I added a title on the assignment sheet: ‘Lied ohne Worte’. This is a designation of a genre, and it already has important implications: it makes one roughly expect a style (Mendelssohn, or some later romantic salon style); it suggests a certain character and texture (a singing melody with a clearly separate accompaniment); the style makes us expect certain melodic gestures; we expect a certain form (ABA will be likely, with the middle movement in a contrasting key, probably the relative one); we even expect a melody that is generally built in four-bar phrases. On a more detailed level, the suggested opening phrase (my incipit) can easily be seen as the first half of a classical period, implying a next phrase that opens symmetrically and then either closes in the home key or modulates.

Of course, also my piano accompaniment was strongly guiding the oboe player; this is because tonal harmony tends to follow specific patterns, depending on, as William Caplin calls it, the formal function of a passage [12]whether you are at a beginning, middle or an end of a section. But exactly which progressions I chose, and how I arranged them, depended on the possibilities the melody of the oboe player left me. His melodic gestures created expectations as well, resulting in me choosing for one of the harmonic options available at a specific moment. In this way it really was an ongoing process of giving and taking.

Many of these guiding factors (form, texture, melodic and harmonic gestures, etcetera) have a resemblance with concepts that are used in classical music analysis. There is one important difference, though. The process of conceptualization that takes place in analysis is a necessity: it is the only way to speak about music on a level that transcends the mere reflection on the immediate reactions of the listener. However, as soon as concepts are made, a process of reification (Verdinglichung) starts: what once was something that sounded and happened in time (like a specific harmonic progression), now becomes a static object, that has a name and can be labelled – discarding the temporal dimension. This process tends to make the analytical concept less useful for an improviser: he can only use e.g. a specific harmonic progression when it has this temporal quality. For improvisation, the guiding factors should not be seen as analytical objects, as is often happening in theory classes, but on the contrary as musical gestures with a direction, creating musical expectations.

These guiding factors are interesting because they are the connection between this new piece and already existing music, be it in the form of compositions or as elements of a musical style. In fact, they are a clear illustration of Bruce Ellis Benson’s view on improvisation as ‘to rework something that already exists’. I would like to propose the term loci communes (commonplaces) for these factors. Without loci communes, the oboe player and I would not be able to find each other, but more importantly: neither listeners nor performers would understand  music.



3.4 Interpreting a score


In order to examine the connection between improvising and interpreting through loci communes, it will be helpful to do a thought experiment by reversing the process: let’s give our oboe player (or any other classical musician) a score to play. What exactly happens when a musician reads a score? Which processes might be at work? To be sure, there are many ways to approach this question, depending on whether one looks from a practical, pedagogical, philosophical or (neuro)psychological perspective. In what follows, I will choose a direction that makes it possible to connect interpretation with improvisation. Our experiment starts with a musician reading a score.

In the context of this research, a score is a musical text which fixates sounding music by representing it in signs [13]. Since music happens in time, and a score is a static object, a score not only describes music which once sounded (e.g. in the imagination of the composer), but also prescribes future music making when the performer uses the score to perform the music that it fixates. These two functions of a score are at the basis of the following investigations.


3.4.1 The first level of meaning: the meaning of notational signs


When the performer wants to perform from the score, it has to mean something to him. It is not likely that an ‘average’ performer will be concerned too much about the extensive music-philosophical and theoretical debates about meaning in music that have been taking place in the past decennia; nonetheless, his need for a feeling of understanding when confronted with a score is acute. Starting from this basic question, I would like to distinguish between different levels of meaning. On the most basic level, the signs in the score indicate aspects (parameters) of sound. The traditional Western music notation can fixate the pitch, the duration and – less precisely – the dynamics and tone colour. But, as all musicians know, the precision of the notation varies: the neumes of the ninth century, for example, were mnemotic signs, used as an aid to memorize existing [14] melodies. Until the development of the staff notation, these signs could hardly be prescriptive, at least as far as the mentioned parameters are concerned. Or, to take a more contemporary example: a graphic score even doesn’t intend to prescribe exactly the parameters pitch and duration. Tablature notation can be regarded as a variant of staff notation, with the difference that it represents an action instead of a sound. It needs the instrument as a link between sign and sound. For the field of this research however, such variants are not relevant. As for the well-known staff notation, exactly the fact that it seems to be much more precise can cause pitfalls to the interpreter. The different parameters that can be represented in this type of notation will here be dealt with separately.


a. The pitch: the staff notation is based upon the diatonic system, which makes it unsuitable for the notation of micro tones. It is not unthinkable that the development of the staff notation by Guido of Arezzo only allowed to notate approximately the pitch of the already existing Gregorian melodies, which probably contained micro tones: if that is true, the score offered a simplified description of the actual music. When the same score was used later prescriptively in a different musical environment, this simplification tended to be forgotten – with ‘pure’  diatonic new music as a result. The (im)possibilities of musical notation tend to influence the creating of new music.

Within the diatonic system, temperament has varied considerably through the ages. When we consider a score to be describing actually sounding music, this fact can make a rather large difference. When a score describes keyboard music in a situation where for instance the mean tone temperament was prevalent, the prescriptive use of this score can cause quite different effects when the used temperament is not the same. In this way, a chromatic line or a specific chord in a piece by Sweelinck or a contemporary composer can sound very different when played in equal temperament. It has to be said that one could argue about the question whether indeed Sweelinck’s score describes music as it actually sounded in the composer’s environment. An equally possible view is to understand the score as representing some abstract imaginary music, which needs ‘materialization’ in the actual performance [15]. Quite convincingly this could be said about an ‘abstract’ piece like Bach’s Kunst der Fuge, which has no indication of instrumentation. On the other hand, Bach’s masterwork can be considered exceptional in many senses. The approach of H.I.P., I believe, very much sees the score as a representation of music as it actually sounded at the time of composition [16].

For singers and instruments that need intonation, the notated pitch very often is not exactly the same as the sounding one. Also in an environment that uses equal temperament, adjustments of intonation happen constantly and almost automatically, generally speaking mainly guided by the desire for well-tuned triads.

So far only the notation of  pitches in relation with each other was discussed. The absolute pitch however, the use of a diapason, is also an issue. Only in 1859, the standard concert pitch was settled in France at 435 Hz for the A4. It is certain that before 1800, concert pitches could differ a lot between cities. Even in 1850, a piano builder like John Broadwood still used very different diapasons for concert instruments and for pianos at home [17]. The use of a standard pitch of 415 Hz for pre-1750 music, and of 430 Hz for music from the classical and early romantic periods – as usual in H.I.P. - must be seen as a modern invention that is more convenient for today’s musicians than that it is historically justifiable.

Summarizing, the extent to which pitch is fixated in the score leaves much to discuss about. A lot of attention has been paid in the past 60 years or so to different temperaments and the changing of the concert pitch. An almost unknown area in classical music is the expressive use of intonation, like the blue note in jazz. We are so much used to (and proud of) playing ‘in tune’, that this possibility seems to be forgotten altogether. The possibility that the development of playing in tune (very much connected to the upcoming recording industry) probably also meant a certain loss of expressive intonation should not be excluded [18]. Surely such (hypothetical) expressive intonation in performing scores from the past would be suitable for an improvisational approach – which clearly cannot be said about choices regarding temperaments or diapasons.


b. Duration. The notation of rhythm in a musical score is usually even less precise than the notation of pitch. The complex medieval mensural notation system provided a number of different possibilities of subdivision, of which in the now conventional music notation only one – a dual division on all levels – remains. The process of simplification, to be found in the notation of quarter tones in a diatonic system, definitely also applies here: the notational system supposes a regular beat (tactus) that can be subdivided only in a limited number of ways. Any irregularity is hard to notate. A rhythm that is not supported by the notation is apt to disappear when scores become a principle vehicle of spreading music. Only as late as half way through the 20th century more complex and additive rhythms started to be notated – very much to the expense of the readability and the comprehensibility of the scores. The speed of the pulse cannot be shown in the shape of the notes has to be described in additional text.

There are many examples in the history of Western music testifying the experienced lack of precision – or even impossibility - of the rhythmical notation: from the notes inégales in French music and the preface to Frescobaldi’s Toccate to the proper ‘swing feel’ in jazz music. But also on a smaller scale, a performance that would do mathematical justice to the notated rhythm of almost any piece of music would not only sound boring, but even inhuman and dead. The largest differences between different performances of a piece of music from a score are in the interpretation of the notated duration; it can hardly be a coincidence that the attention of an institute like CHARM, analysing early recordings, primarily is directed towards the timing.

Also for this parameter, a discrimination between relative and absolute use can be recognised: so far only the relative tone length was discussed, but since the invention of the metronome around 1820 it is also possible to determine exactly the speed of the musical pulse. Surprisingly, this did not mean the end of discussions about the proper tempo: in fact, composers adding metronome marks to their music (starting with Beethoven) only increased the quarrels. Maybe it can safely be said that the metronome indications in scores describe the tempo the composers intended when they wrote down the music they heard (imagined), but that the prescriptive value is more limited [19]. Many anecdotes about composers toning down the importance of following exactly metronome marks in their scores (up to nowadays) testify this. Apart from this, tempo fluctuation within a piece seems to be an intrinsic part of many styles, even though modern performances often ignore this. Early recordings though, like the recorded performances of music by Schumann and Brahms by former students of Clara Schumann, show that the tempo fluctuations within movements could be very large [20].

Summarizing, the rhythmical notation can in many scores be seen as a simplification. Deviating from (or rather: bringing to life) this abstracted rhythm pre-eminently fits an improvisational approach.


c. Dynamics and tone colour. Neither the dynamics nor the colour of the tone can be expressed in the signs which make the musical score: the musical notes; this information always has to be mentioned separately, and often in words [21]. This makes the indication of the dynamic level or the colour of the tone a relative matter: sotto voce only has a meaning in relation to the sound in other passages, and even dynamics are indicated in relation to the instrument or the room. The (usually verbal) indications used for this parameter in fact rather formulate the impression on a listener than anything that can be measured. So far no score indicated the dynamic level in decibels… Exactly the vagueness of this parameter leaves room for an improvisatory approach.


Summary: the basic level of meaning in a score concerns the ‘déchiffrage[22]: the reading of the score, which is coupling the representation to what is represented, especially the pitch and length of a tone. In this process, the musical imagination plays a vital role. As a reader of a verbal text can imagine the sound of a word before he actually speaks it, so the reader of a musical text (a score) should be able to imagine the sound before performing it [23]. Even for some accomplished musicians this is not an easy task at all. And just as a reader of a verbal text may be able to pronounce the words when he knows how to do this, without necessarily understanding the meaning of them, so a musician may play or sing what is written and still not know what this music wants to tell – which brings us to the second level of meaning in a score.


‘Licks’ or not: this phenomenon is actually quite similar to what happens in several jazz styles. Needless to say that, as soon as a composition becomes part of the ‘museum of musical works’ (Lydia Goehr), this effect becomes harder and harder to appreciate.


Compiling a catalogue of ornamental patterns started with setting up the morphological and stylistic framework. As for the former, only such patterns were selected which could be continued across the whole keyboard, i.e. which could be part of a sequence in the octave. In practice, this meant that the patterns were primarily to be seen as chord prolongations, rather than as embellishments of melodic turns. This discrimination however is sometimes a bit artificial, especially when the musical context becomes more complex than the simpler Czerny preludes. In the compositions of Frédéric Chopin, for instance, it is often hard to decide whether a specific ornament is to be seen as bridging a melodic interval or prolongating the harmony under that melodic turn (often a dominant seventh chord). In such cases patterns which did allow for the mentioned sequence building were understood as chord prolongations. A further consequence of the sequence model is that the patterns are presented like senza tempo models, i.e. regardless of the meter. Musically this typically means cadenza style passages, where the musical pace pauses for a while. However, this is not the entire truth: almost all of the patterns (maybe the scale derived patterns excluded) still show a difference between strong and weak tones. What in fact happens in such senza tempo passages, is that inside the sustained or pausing pace of the music, the ornamental structure still shows minor ‘sub-beats’, that have no connection with the main tempo of the piece. Seen this way, senza tempo only means that the main tempo is suspended. This can clearly be seen in the following example [52] from the Anleitung:

Improvised prelude connecting the end of the Destiny Waltz (Sydney Baynes) and the beginning of El Choclo (A. Villoldo)

Which locus communis does the piano accompaniment refer to? On basis of the standard way of playing the 2/4 metre, the bass notes will be stressed (because they are on relatively strong beats), and the right hand chords will be plaid lighter. This way of playing may seem to connect with a locus communis that is very familiar in the popular music of the 20th century. However, this would be an interpretational choice that hardly matches the other loci communes of Dvořák’s piece – and can be called anachronistic. More convincing is to connect with other comparable accompaniments in 19th-century dance music, especially music that is based upon gypsy music. A typical performance within that locus communis pays special attention to those right hand chords on the weak beats by playing them rather sharp and energetically, resulting in an exciting counter-rhythm. This way of playing also dictates the ‘right’ tempo quite compellingly. 

An interesting kind of (hypothetical) locus communis can be found in Robert Gjerdingen’s influential book Music in the Galant Style [28]. In Gjerdingen’s theory, so-called schemata (specific harmony / voice leading situations) play a role, comparable to the above mentioned loci communes; in his book, the function of the schemata is compared with the figures of an ice-skater, or the standard roles in the Commedia dell’Arte. What makes his ideas so interesting in view of the topic in question is in the first place the fact that the suggested schemata / loci communes are hard to appreciate for a modern interpreter. In the words of Gjerdingen: ‘Whereas casual observers of ice-skating competitions may see only a variety of glides, spins, and jumps, a connoisseur sees salchows, axels, lutzes, and camels.' [29] This may indeed be a problem with loci communes from the past: often we are even not aware of their existence! A second point of interest is that Gjerdingen writes about schemata / loci communes from the point of view of the creator of music; they are elements of a galant musical etiquette, and the musician was to be judged after his ability and inventiveness to handle them. I will return to the important implications of this in a following paragraph.

To conclude this elaboration of the idea of loci communes, I would like to ponder briefly the possibility that the interpreter sometimes may be unable to connect the score to a locus communis. Though this may happen not very often, I consider it to be demonstrating the plausibility of the view on interpretation presented above. When a musician doesn’t recognize any familiar pattern in the score, he can’t make sense of it, and feels like not understanding the music. Interestingly, in such cases it turns out to be also very hard to memorize the music, or even to be able to play it at all (when the music is complex)[30].


Improvisation within the score

Is there any improvisation possible on this second level of meaning, the meaning of a score? Surely, this is not the level where we can speak of degrees of fixation easily. Rather it is in the mind of the interpreter that meaning is attributed to a score, that loci communes are connected with notated sounds. An improvisational element in this process would imply changing the meaning during the performance, which might sometimes be possible [31], but is in general not very likely. Mostly, it will not be the meaning itself which is improvised during the performance from a score; rather the meaning provides the playing ground for the game of improvisation. In this respect, it is telling that the schemata of Gjerdingen play a part that resembles that of the loci communes. The same loci communes that give meaning to the musical text, can serve as a starting point for the creation of music. The ‘gypsy accompaniment’ type that Dvořák used in his Dumky Trio makes the performer understand the score, but at the same time inspired the composer and just as well can be the basis of a new free improvisation – or of an improvisational approach of Dvořák’s score. On a detailed level, many equivalent ‘solutions’ are possible: the chords could maybe have different positions, even the choice of chords could sometimes be a bit different; or the pianist might add some ornamental music. But also without changing anything in the score, there are many possibilities for improvisation, e.g. by playing the syncopated ‘counter-rhythm’ less or more clearly, or by playing with the dynamics. As long as the ‘idea’ of the locus communis remains intact, the meaning doesn’t change by such alterations to the score (improvised or not [32]).. The amount of ‘fiddling’ with the score that is regarded as acceptable depends on many circumstances. In this case, I believe we can trust that Dvořák (and his contemporaries) expected the musicians not to change the chords, though he might not have been bothered by some changed octave doublings in the bass register. Dynamically a lot seems to have been possible. As many early recordings, and especially research like Neal Peres da Costa’s [33] and Anna Scott’s [34] show, late 19th-century performances most likely were a lot more adventurous than many modern musicians think. Those recordings give us an impression of freedom, of a sovereign attitude of the musician towards the score. Still, not just anything was possible. Where, then, is the border between gripping freedom and whimsical nonsense? For me, the answer can be found in the loci communes. They provide meaning, they make music sound like a language, they set the frame within which the game can be played. The ‘gypsy accompaniment’-locus communis allows for many different harmonic and dynamic approaches – but the specific movement with its rhythmic energy is essential.

Also Czerny provides loci communes as bases for improvisation, e.g. in chapter 4 where he shows different transformations of elementary thematic material according to different genres (cf. note 27). Apparently, Czerny could assume that his readers knew the loci communes he mentioned; they didn’t need further explanation, just a short example was enough. For us, the situation is clearly different.  

To conclude, an example of a score-based performance with many improvisational elements: the well-known tango El Choclo, performed by a light orchestra. The piano part, played by the author, contains many improvised additions which, I believe, do not change the meaning of the score. Despite the difference in style, this technique is essentially the same as improvising a realization of a figured bass in a baroque piece.

This harmonic structure clearly links the prelude to the cadenzas which started to embellish the solo concertos and arias in the course of the eighteenth century. Czerny does not explain how to develop these harmonic patterns into a prelude – he merely gives (many) examples. Invariably every chord is ‘prolongated’ by virtuosic passage work, each chord with a new pattern. The character of these runs forms another important difference with the baroque keyboard prelude: they resemble the type of runs which started to come up together with the slower harmonic rhythm during the Viennese Classical period. One thing Czerny’s examples of simple preludes share with their baroque counterparts however is the fact that a clear meter is often absent. This means that the patterns, that resemble the runs which we find in so many classical sonatas, here often appear in a senza tempo context.

In other words: in his simple preludes Czerny uses material that matches the type of passage work that came up in classical keyboard music, developed into a characteristic of the highly ornamented style of the early 19th-century keyboard virtuosi like Hummel and Clementi, and forms the basis for the innumerable piano etudes which Czerny composed himself. We see the classical piano technique being applied in those preludes! This makes us expect that the modern pianist, whose daily gymnastic exercises are based upon this type of technique, should be able relatively easily to follow Czerny’s advices and improvise such preludes himself.

Experiences with conservatoire students tell us that this is not the case. One reason is that Czerny’s examples are not like musical recipes: he does not really explain how to do it. Even for such simple music as the preludes of Chapter One in the Anleitung there are aspects which for a modern student need explanation. In the fifth example [49] for instance, the harmonic structure is embellished with secondary dominants and other ornamental harmony in a way that a modern piano student tends to find hard to come up with by himself: 

Improvised sonata, fragment of slow movement. The whole sonata is accessible through

Vocalise, S.Rachmaninov (improvisation with only minor changes to the score)

Improvised prelude connecting the end of the Pizzicato Polka (J. Strauss) and the beginning of the 5th Hungarian Dance (J. Brahms)