The conditions for learning musical interpretation in one-to-one piano tuition in higher music education

Research has indicated that one-to-one teaching in higher music education in Western classical music typically favours technical over interpretive aspects of musicianship, and imitation of the teacher’s rather than the student’s explorative interpretation. The aim of the present study is to investigate students’ and teachers’ understandings of how musical interpretation of Western classical music is learned in this context. Semi-structured qualitative interviews with six piano students and four teachers in Sweden were conducted and hermeneutically analysed using haiku poems and poetical condensations. The analysis found that the conditions for learning musical interpretation centred upon students achieving a high level of autonomy, as affected by three key aspects of teaching and learning: (1) the student’s and the teacher’s understandings of what musical interpretation is, (2) the student’s experience of freedom of interpretation as acknowledged by the teacher, and (3) (expectations of) the student’s explorative approach. As none of these aspects were reported as being explicitly addressed during lessons, there might be a need for both teachers and students to verbalise them more clearly to support piano students’ development.

This preludial poem was constructed using the literary practice of found poetry 1 hereafter referred to as poetic condensation (see, e.g., Hølge-Hazelton & Krøjer, 2008;Öhlen, 2003) The practice of found poetry originated in the early 20th century . Found poems are created by combining words, phrases, or entire passages from one or multiple sources and framing the resulting text as poetry in the new context (Perloff, 2012, p. 503). Researchers have used this type of poetry under a variety of names, e.g., research poems (Langer & Furman, 2004) and vox participare (Prendergast, 2009, p. 545).
The conditions for learning musical interpretation 105 denoting an active analytical engagement with empirical material presented through poetry. The term thus functions as a noun (i.e., product) as well as a verb (i.e., practice). The poem above is based on the transcription of an interview with a piano student conducted as part of the pilot study to the research project reported on in this article. The poem incorporates fragments from the Swedish song If you ever reach Samarkand 3 (author's translation), commenting upon and contributing to the poetic condensation. 4 The preludial poem intends to serve a twofold function: first, it sets the scene while at the same time anticipating some of the results of this study, i.e., highlighting the importance of musical interpretation and possible consequences of the (lack of) learning therein. Second, it offers an example of a poetic condensation, a form that-in combination with haiku poems-will be used to present the analysis and results of the interview study, as well as functioning as a transition between sections in this article.
In an earlier article (Holmgren, 2018), I conducted a philosophical poetic inquiry of three aspects of interpretation in music education research in the form of an autoethnodrama. The empirical material for the present article was also used in this earlier inquiry.
However, the aims of the two articles differ, the former centres on autoethnographic aspects, whereas the present one highlights the implications of the findings for music education (research). The text that constitutes the lion's share of this article is not intended to portray itself as an arts-based artefact. Instead, methods for analysis and (re)presentation from the field of arts-based research are integrated into a more traditional type of text. This use is not intended as a critique of (present) arts-based research, nor of more traditional academic writing, nor is it an effort to undo the development within this sphere and start (re)questioning the legitimacy of such research. Instead, I intend to widen the use of artsbased practices to include contexts where this has been more uncommon. In this case, to use practices associated with such research to further the understanding of the learning of musical interpretation.
Essential similarities between understanding viewed from a hermeneutical perspective, musical interpretation, and poetry are the relevance of the temporal aspect the use of poetry can unleash the potential to view the researcher as artist, i.e., as a tool for exploiting complexity to further understanding (and transparency) (Pigrum, 2008, pp. 765-766).
The aim of this article is to investigate students' and teachers' understandings of how musical interpretation is learned. Hence, the focus is on the forming of a musical interpretation, which may include performance as an integrated aspect of such interpretation.
Thus, this stance contrasts to the view that forming of a musical interpretation ends where the performance thereof starts. The study is delimited to the context of one-to-one tuition of piano students in higher music education in the Western classical music tradition in Sweden. The following two research questions were formulated to fulfil this aim: 1. How do students and teachers describe musical interpretation?
2. How do students and teachers describe the conditions for learning musical interpretation?

Background
The term musical interpretation refers to the understanding of a piece of music. Such an understanding is commonly manifested in, but not limited to, musical performance, 5 and musical works are regularly said to be given life through these (cf. Brendel, 2012Brendel, /2013. The notated musical work contains the composer's explicit and implicit instructions to the interpreter (Davies & Sadie, 2001, pp. 497-498). Musical interpretation is the process of selection and application of performance choices on a composition. This selection can be "more or less motivated and coherent" (Lehmann, Sloboda & Woody, 2006, p. 96) and apply both on the micro and macro levels. As the musical score under-determines performances, the interpreter unavoidably has to make countless decisions regarding how the 5 For further context, see, e.g., Levinson (1993) and Neufeld (2012) for discussions of dissimilarities between performative and critical interpretation in music, Davies (2006) for a view of performance interpretation of musical works, and Cook (2001) for a perspective on the relationship between musical performance and analysis.
To develop students' ability to form their own musical interpretations is one of the overarching goals and greatest challenges for higher music education in the Western classical music tradition (Silverman, 2008, p. 249). Higher music education is simultaneously a formally structured education and one that resembles a master-apprentice tuition model (Nielsen, 1999(Nielsen, /2000). In the master-apprentice model, imitation, often without verbal instruction, is a common way of learning, since the master embodies the skills that the student is meant to acquire (Kvale & Nielsen, 1999. Research on higher music education has increased significantly over the past decade. Still, relatively little attention has been directed towards learning in the dominant one-to-one context of instrumental teaching (Gaunt, 2009, pp. 178-179). Nevertheless, there seems to be nearly universal consensus on viewing the one-to-one relation as an "indispensable, intense and intricate" part of higher music education (Gaunt, 2008, p. 230), i.e., what constitutes " 'serious' instrumental and vocal tuition" (Harrison, 2004, p. 206). Hultberg (2010, p. 9) found that the instrumental teachers she studied in higher music education attached great importance to students' playing, and that lessons were largely devoted to mimicry, which the students were highly skilled at. Thus, it was difficult for teachers to evaluate what knowledge the students had internalised and could independently externalise. Consequently, teachers might wrongly form the impression that students had assimilated the teaching content. An evaluation showed that the practical skills (i.e., performance) that students showed in their playing towards the end of a lesson had not been internalised nor could they be successfully externalised (i.e., forming of a musical interpretation) (Hultberg, 2010, p. 9). These results might be due to the (false) premise that if the imitation is similar (enough) to the original, then both must be based on the same representation (Lehmann, 1997, p. 156). It has been found that there is a tendency for the most talented students to mainly engage in imitation of the teacher, rather than develop a sense of responsibility for their own learning (Burwell, 2005). As such, private instruction is described as a comfortable learning environment where students can become dependent on the teacher (Gaunt, 2009, pp. 197 & 203). Thus, students who succeed in imitating the teacher are seen as gifted, and develop quickly (Burwell, 2005, p. 213), at least on the surface level.
A common view among music students is that the development of technique and interpretation are separate, sequential processes. However, many prominent musicians often formulate their interpretive strategies rather early in the practising process, and they usually acquaint themselves thoroughly with the music before they start practising (Lehmann, Sloboda & Woody, 2006, pp. 97-98;Marín, Pérez Echeverría & Hallam, 2012, p. 194). Results show that few first-year students in higher music education observed strategies for developing musical expressivity during their practice sessions (Hultberg, 2008, p. 8). Zhukov (2008, p. 172) found that students in higher music education frequently used professionally recorded versions of works studied as sources for imitation. Although listening to different interpretations, recordings, and performances can increase an awareness of interpretive possibilities markedly, many classical musicians are opposed to this (Lehmann, Sloboda & Woody, 2006, pp. 100-101;Silverman 2008, p. 266). There is, however, no evidence that abstinence from others' performances functions as a beneficial learning strategy (Lehmann, Sloboda & Woody, 2006, p. 101). On the contrary, it has been indicated that active listening and (reflective) imitation can increase the musician's experienced interpretational freedom (Holmgren, 2015). Support is also lacking for approaches where the teacher dictates the interpretation-and the student remains passive-since the student is hindered from forming a personal connection to the score (Silverman, 2007, p. 112).
Instrumental teachers seem to intuitively relate to their own understanding of the music rather than the students' (Lehmann, 1997, p. 157). At the same time, some students do appear to have their own agendas both concerning the content of the lesson and the teacher (Holgersson, 2011, pp. 127-141 & 184). Therefore, students might find it difficult (or choose not) to assimilate the teacher's advice, particularly if they do not understand or cope with the fact that the teacher relates to the musical work (and possibly also to musical interpretation as such) differently. Usually, students continue to pay attention to aspects that are important to them. In cases where students already lack motivation, it might be difficult to bridge the gap between the teacher's and the student's attitudes (Hultberg, 2008, p. 20).
Although students develop many skills in one-to-one tuition, they do not necessarily learn to transfer them to other contexts, such as other pieces within the same genre or category of musical works, e.g., a particular type of movement in Baroque dance suites, or to make independent decisions with regards to how they want to interpret the music (Gaunt, 2009, p. 180;Hultberg, 2000, p. 1;Mills, 2002). Research shows that teaching expression largely consists of teachers presenting their own intuitive thinking, rather than sharing systematised knowledge about expression and its principles (Karlsson & Juslin, 2008, p. 321).
Thus, teachers tend to formulate easy-to-understand rules and basic principles to follow, which may be effective for the teacher, but might not work for the student (Lehmann, Sloboda & Woody, 2006, p. 10).
Engaging students verbally after demonstrating a passage might positively affect their development compared to the teacher only speaking or demonstrating (Lehmann, Sloboda & Woody, 2006, p. 96), as it has been found that students who can describe the interpretation that they heard are better at reproducing it (Lehmann, Sloboda & Woody, 2006, p. 103). In chamber music practice, musicians are expected to make their interpretation explicit verbally, which might be one of the reasons why chamber groups are often obligatory in higher music education (Lehmann, Sloboda & Woody, 2006, p. 103). Moreover, teaching is traditionally seen as an effective way of internalising knowledge (Kvale & Nielsen, 1999Lehmann, Sloboda & Woody, 2006, p. 103). Young, Burwell, and Pickup (2003) found that technique and so-called "commandstyle" teaching was predominant in instrumental lessons in higher music education. Similarly, Zhukov (2008, p. 172) concludes that technique had the highest priority in such lessons. These findings are substantiated by Holgersson (2011, pp. 154 & 185). According to Burwell (2006, p. 345), less of the dialogue during instrumental lessons concerns interpretation than it does technique. This tendency is confirmed by a study of recorded instrumental teaching ranging from municipal, upper secondary, to music conservatory level (Karlsson & Juslin, 2008, p. 314). The results showed that verbal expressions containing the word stem "interpretation" was uttered in total three times by the teachers, while other stems were more frequent, e.g., "good" (153), "play" (274), and "m" (affirmative) (869) (Karlsson & Juslin, 2008, p. 319). Consequently, we may assume that words related to expressive aspects of performance were rarely spoken during lessons. These results do not imply, however, that interpretative teaching does not take place. Earlier studies may have underestimated this aspect due to interpretation not being directly talked about (Karlsson & Juslin, 2008, p. 310). Nonetheless, results from a survey encompassing 51 teachers from music conservatoires in Sweden and England show that these teachers would like to spend less time on teaching technique and more time on teaching expressivity 6 than they do (Laukka, 2004, pp. 46, 47 & 52).
I wonder how often I say the word interpretation it is probably quite seldom 7 6 It must, however, be noted that Laukka (2004, p . 49) emphasises that "play expressively" is not synonymous with "make an interpretation"; the former is about communicating emotions, whereas the latter is about personal expression.

7
This poem is a poetic condensation of transcriptions of the interviews with the four teachers that participated in this study.

Theoretical framework
This section consists of three parts: first, I anchor musical interpretation to approaches within the hermeneutical tradition, constituting the study's theoretical framework; second, I present my views on the relationship between learning and language; and third, I argue for the particular use of poetry in the form of haiku poems and poetic condensations, as employed in this article.
First, the learning of musical interpretation and conditions therein are the focus of the study reported on in this article. Within hermeneutical research traditions, it is important to articulate the assumptions and understandings underpinning a research study, as they will influence how phenomena are viewed, what knowledge is possible to acquire, and how that knowledge is acquired. I subscribe to the view that every musical interpretation is temporary, and to a certain extent, incomplete-as new interpretations are always possible. I do not view the process of musical interpretation as a straightforward process from non-knowledge to objective knowledge. In line with this framework, I hold the view that the interpreter of a musical work is present in a historical context that defines the person's interpretive horizon (Gadamer, 1960(Gadamer, /2013. Accordingly, the researcher faces similar overarching challenges when interpreting empirical material as the musician does when studying a score, even though formal demands may differ.
In order to understand the multifaceted and interpretive nature of the phenomenon studied, a hermeneutical approach based on the philosophy of Gadamer (1960Gadamer ( /2013 and Ricoeur (1986Ricoeur ( /2008) was applied. Even though the points of departure of each of these theorists differ considerably, they are both concerned with language, history, and practical wisdom; thus, it is possible to combine their understandings of hermeneutics (Grondin, 2014, pp. 61-62). In the present study, hermeneutics is seen as an ontological and epistemological base with methodological implications. However, Gadamer's (1960Gadamer's ( /2013) focus on the philosophical content of understanding itself relativises the concept of method without cancelling it out (Dutt & Gadamer, 2001, p. 41 (1986/2008, pp. 151-163) is more concerned about which methods should be followed in interpretation (Grondin, 2014, p. 49). Gadamer (1960Gadamer ( /2013 argues that understanding in the human sciences does not require a methodology; indeed, the scientific ideal of objectivity fails to account for the basic hermeneutic experience of understanding, as understanding appears to be less like stringent rules and more like a fusion of horizons. In the hermeneutical tradition, every human has an individual experience of oneself and one's situation. Thus, the focus lies on humans' interpretations of meaningful material rather than on the constitution of the world itself. For Gadamer (1960Gadamer ( /2013, the idea that the fusion of horizons in understanding is the "achievement of language" is paramount. Therefore, he agrees with Wittgenstein that there exists no private language (Dutt & Gadamer, 2001, p. 56). The concept of a private language is defined by Wittgenstein (1953Wittgenstein ( /2009) as a language, which even in principle is impossible to understand by anybody except a single user of the language. 8 According to Gadamer, language is a we where the individual has no fixed border, and one's place is assigned in relation to each other. Consequently, personal borders and limits of understanding must be crossed in order to understand (Dutt & Gadamer, 2001, p. 56). Gadamer writes that "the text brings a subject matter into language, but this is ultimately the achievement of the interpreter.
Both have a share in it" (1960/2013, p. 406). Thus, as the interpreter intends to understand the text-here viewed in an extended meaning, i.e., including transcriptions of verbal utterances-the "interpreter's own thoughts too have gone into re-awakening the text's meaning" (Gadamer, 1960(Gadamer, /2013. The interpreter's horizon is determinative and seen "as an opinion and a possibility that one brings into play and puts at risk, and that helps one truly to make one's own what the text says" (Gadamer, 1960(Gadamer, /2013. This is what Gadamer describes as the fusion of horizons. He means that the fusion of horizons "takes place in conversation, in which something is expressed that is not only mine or my author's, but common" (Gadamer, 1960(Gadamer, /2013. Second, I define my view of the relationship between learning and language. A teacher might know that the student has associated a compatible meaning with a particular expression or principle if and only if the student's association yields the same (or similar enough) result as the teacher's (Hansen, 2007, p. 50). Thus, learning is neither understood as mere reproduction (Lehmann, 1997, p. 156) nor as associating one particular meaning (i.e., the teacher's) to the expression or principle (Hansen, 2007, p. 50). When the teacher involves the student in describing what the student has heard in a previous demonstration or verbal description, the student is encouraged to work out an ostensive definition (where the pointing out is not necessary as the student has the particular experience fresh in memory) of the subject matter intended to be learned (Hansen, 2007, pp. 40-41). I view learning as including a necessary declarative component: if the student cannot explain an expression or principle in his or her own words, then the learner lacks understanding (Lindström, 2011, p. 4).
Third, I present the reasons for using haiku poems and poetic condensations in this article. The concept of arts-based research emerged during the 1990s (Barone & Eisner, 2012).
Since then, considerable theorisation, as well as the development of domain-specific stylistic conventions, has been carried out in the field. The use of poetry in the context of qualitative research has been motivated by the idea that poems present an essence, and offer an evocative way for (re)presenting results in an often effective (or extreme) condensation, highlighting that qualitative research in itself is a craft (or art), and challenging the (false) dichotomy between fact and fiction (Faulkner, 2007(Faulkner, , 2016Furman & Dill, 2015;Furman, Lietz & Langer, 8 Even though the arguments put forward by Wittgenstein (1953Wittgenstein ( /2009 in support of his no private language-theory has been criticised, the theory has been viewed as correct, see Hansen, 2007Hansen, , p . 50. 2006Leavy, 2015, pp. 63-64;Owton, 2017;Richardson, 1992Richardson, , 2001Öhlen, 2003). Artsbased research, poetic inquiry, as well as haiku and found poetry, have been gaining in popularity over the last few decades; however, they have been used infrequently in (higher) music education research, although there are exceptions (such as Gouzouasis, 2007Gouzouasis, , 2008Gouzouasis, , 2018Gouzouasis & Lee, 2002Gouzouasis & Leggo, 2016;Holmgren, 2018;Lee, 2006;Lee & Gouzouasis, 2017;Prendergast, Gouzouasis, Leggo & Irwin, 2009).
In the present article, poetry is viewed as a form of translation, and as such considered a suitable analytical and (re)presentational tool with which to study musical interpretation (Holmgren, 2018). In addition, the task of translating the poems from Swedish to English was seen as an extra step in the analysis that made it possible to further externalise and elaborate on the interpretative process. The reason for using haiku poems also stems from my prior (amateur) knowledge of the form and the benefits of adhering to a restricted framework, after having written more than a thousand (unpublished) haiku poems myself.
Hence, I have experienced that writing poems induces creativity and forces the words to be selected with the utmost care.

Methodology and material
To align the method with the theoretical perspective and phenomenon studied, qualitative semi-structured interviews (Brinkmann, 2013, pp. 21-25) were used. According to Gadamer (1960Gadamer ( /2013, the fusion of horizons happens in a conversation when something common is expressed. Thus, it follows that studying participants' verbal utterances is a viable way of gaining knowledge about their understanding of the phenomenon. To encourage the generation of rich material based on answers from all participants, interview guides-with slight differences to suit the different perspectives of teachers and students-were created and used. The transcripts of these interviews constitute the empirical material interpreted and analysed.

Participants
The number of participants in this study was ten in total-four teachers and six studentsfrom three institutions of higher music education in Sweden. 9 The number of institutions was guided by the principle of maximum variation sampling (Palys, 2008, p. 697), in order to ensure variety regarding geographical location, tradition, and range of educational programs amongst the chosen institutions. Further, I wished to avoid the risk of a singular view or a potentially polarised discussion.

9
In addition, in the pilot study to the research reported in this article, in total, three interviews were conducted: one each with a piano teacher, a piano student, and a teacher of another instrument, all within higher education of Western classical music in Sweden.
In total, six students were selected: two enrolled in each bachelor, master, and music teacher training program. The selected teachers comprised one full professor and three senior lecturers, all of whom teach piano.

Interviews and transcription
The interviews were conducted during the academic year 2015/2016. The ones with the students lasted about one hour each, whereas the interviews with the teachers lasted closer to an hour and a half (except for one interview which was around one hour long). The interviews, in total approximately 12 hours of digitally recorded audio material, were transcribed verbatim (including stuttering, repetition, and partial words; no pause lengths were noted, nor was emphasis or tone of voice). After that, I revised the language and asked the participants for their approval.

Analytical process
The analytical process consisted of the following overarching eight stages: 1. The interview transcripts were read thoroughly multiples times in an attempt to constantly move focus to and from the parts and the whole, i.e., the concept of the hermeneutic circle (see, e.g., Gadamer, 1960Gadamer, /2013. The transcripts were at all times treated as posthumous, i.e., "complete and intact" , in the sense that "the author no longer can respond" (Ricoeur, 1986(Ricoeur, /2008. Rather than trying to find the "lost intention" behind the text, I aimed to understand what "unfolds, in front of the text" (Ricoeur, 1986(Ricoeur, /2008, i.e., moving from who said what in which situation, to what the text is about (Ricoeur, 1986(Ricoeur, /2008. understanding, implied consequences that were inconsistent with the empirical material (Føllesdal, 2001, p. 375).

A tentative (initial
4. The criterion of correct understanding or a plausible hypothesis (see Føllesdal, 2001, p. 376) was seen as the harmony of all of the details with the whole (Gadamer, 1960(Gadamer, /2013, i.e., how well the temporary hypothesis fitted the empirical material and if other hypotheses fitted at least equally well and were simpler (Føllesdal, 1979, p. 324 (Davidson, 1984, p. 27).
5. During the analysis, four themes were identified, and ten haiku poems crafted according to these themes.
6. Consequently, each of the four themes was presented in descriptive texts that took as their starting points the associated haiku poem(s) in both English and Swedish, the poetic condensations, and my overarching understanding of the empirical material.
7. The preludial poem was created based on the poetic condensation of the transcript from an interview with a piano student conducted as part of a pilot study to the research reported on in this article.
8. The creation and revision processes for one haiku each from the four themes identified were analysed and described to make the process more transparent. In retrospect, the overall creation and revision process consisted of the following three parts elaborated on below: (i) a transposition of the results to a different arena, (ii) the use of different stylistic and analytical devices, and (iii) the reflective translation to English.
i) The transposition of the found characteristics of musical interpretation, learning therein, and persons involved to different settings, mainly (Greek) mythology, classical literature, history of piano playing, and (popular cultural) fiction, isolated the core problems of the learning situation after attachment to the current time, place, and actual persons was removed, i.e., a recontextualisation (Linell, 1998, pp. 154-158).
ii) The devices used were the incorporation of references to other areas of interpretational problems, i.e., translation, highlighting historical aspects and the intertwining of research and art, and the use of allegory and unexpected contrasts in the poems in order to suggest openness for multiple interpretations while emphasising the non-static character of interpretation. Moreover, instead of trying to reduce the potentially contradictory meaning of the poems, the aim was to maintain the focus on the complexity of the studied phenomenon.
iii) The reflective translation from Swedish to English made it possible to analytically interpret the material again, and move it to another (higher) level of abstraction,

Results
In the following section, the piano students' and teachers' descriptions of musical interpretation and the conditions for learning therein are presented. The main results were found to centre on the following four themes:

the student's and the teacher's understandings of what musical interpretation is
2. the freedom of interpretation experienced by the student and acknowledged by the Each theme is introduced below with one or multiple haiku poems (including an indication of which aspects they address), followed by a description of the theme. Finally, I summarise the results as a transition to the discussion. The student's and teacher's understandings of musical interpretation have implications for the goals of teaching musical interpretation, as well as the course of action taken to achieve these goals. The poems above illustrate the importance of considering and trying to understand the composer's intention, suggesting that personal freedom regarding musical interpretation is somewhat limited. This conception also indicates the belief that it is, at least to some extent, possible to get in touch with these intention(s), even in the case of dead composers. Alternatively, it could be understood as a guiding principle, a recommendation that the performer abstains from interpreting too freely. Nonetheless, this understanding implies that the teacher function as mediator between the great master, the composer, and the apprentice's apprentice, the student. Thus, by emphasising both the importance of returning to the composer's original text and the teacher's privileged access to the assumedly correct interpretation thereof, the teacher (and the teacher's private or embodied knowledge) is given unique status, especially if-as indicated in poem two above-the most important information is not always explicit.

Understanding what musical interpretation is
In addition, the participants emphasised that musical interpretation is at least partly a subjective activity, which has implications for the evaluation of student learning. The analysis identified that the freedom experienced by the students seems both to limit their musical interpretation and affect their understanding of what musical interpretation is. Similarly, the freedom that the teacher gives the student with regards to musical interpretation shapes this as a learning object, and deems what is possible for the student to achieve. The understanding verbalised by both students and teachers was that more knowledge opens more possibilities. In other words, they felt that the greatest musicians have greater interpretive freedom than both the participating teachers and their students.

Experienced and acknowledged freedom of musical interpretation
Such a line of thought restricts interpretational freedom and encourages a reproductive stance towards musical works. The teachers considered themselves as having (acquired) good taste, and that they exercised this good taste by avoiding the two extremes of not interpreting enough (lecture about rather than a performance of) or interpreting too much (placing themselves in front of the masterpiece). The ability to navigate appropriately between these two extremes seemed to be an important skill that students might learn during their studies. The teachers also stressed that musical interpretations must make sense to the listener, indicating that they conceptualise a continuum from comprehensible to incomprehensible. Thus, the teachers did not view all interpretational freedom in performance as beneficial to the music or the listener. The students accentuated the importance of playing correctly, particularly those enrolled in the music teacher training program. They explained that before they started to interpret, they tried to play the right notes; that they were afraid of making wrong decisions, and had difficulties forming their own opinions with regards to interpretation. At the same time, the teachers emphasised that it is counterproductive for students to settle on a definitive interpretation too early, and to think that there is only one correct interpretation. Poem seven above expresses the teachers' perception that students, in general, have too little knowledge of Western 12 In this haiku poem, alternative expressions are presented (typographically separated by "/") to emphasise the openness and flexibility of interpretation.
13 Thomas Thorild (1759Thorild ( -1808 was a Swedish poet and philosopher. An aphorism by Thorild is inscribed above the entrance to the main building of Uppsala University in Uppsala, Sweden. The famous aphorism reads: "T änka fritt är stort / men tänka rätt är större" (' T o think freely is great, / but to think rightly is greater') (Thorild, 1794). hjälper dig att säga intention and audience det redan tänkta pressure-tests ideas In the poem eight above, the English haiku takes the analytical interpretation a step further (regarding abstraction and extension) than the Swedish does. The Swedish haiku implies that "the will" ('viljan') to express must be greater than the students' expressive capacity.
The English haiku conveys the stance that interpretations have to transcend what is possible in order to achieve or conceptualise. In the Swedish version, the context is delimited to the educational setting through the word "utbildningsaxiom" (' educational axiom ').
The more general expression " first axiom " in the English version might be seen as relating to either the artistic or educational context, or possibly the overall conditions for human understanding and existence.
Both students and teachers explained that the expectation that students should have an exploratory approach toward musical interpretation shaped their understanding of musical interpretation and the freedom the students were afforded to achieve this. The tendency seemed to be that students in the music teacher training program expected to get direct answers regarding how they should play a specific section or piece, and that they had difficulties forming their own opinions. However, students enrolled in advanced level studies, i.e., master students, seemed to be on more of an equal footing in discussions with the teacher, perhaps due to the students' greater knowledge or the teachers' conceptions thereof, and that they, in general, seem to be better prepared for lessons.
This more equal discussion can be understood as constituting a greater potential for the fusion of horizons in the one-to-one lessons with the master students. There also seemed to be a tendency amongst both teachers and students to verbalise the importance of having the will to accomplish musical interpretations that they could not technically play at present, i.e., the student needed a clear interpretive vision. Such an interpretive vision was understood to guide the student's development in order to realise it. If the student did not have such a vision, there seemed to be a risk that their development would lack direction.

Students' levels of autonomy
Poem 10: Conviction cannot be taught övertygelse you must become me kan jag inte lära ut in order to become you studentens ensak "that is why you fail" The English haiku is more specific than the Swedish version in poem 10 above. The Swedish version expresses that it is impossible to teach the student to acquire "övertygelse" ('conviction'); students have to find it on their own. The English haiku, however, explains that the student first has to assimilate the teacher's way of being in order to realise his or her potential as a musician. This might be understood as an interpretation of what is problematised in the Swedish haiku, and that the problem has found its solution in the English haiku.
A student's level of autonomy was expressed by the participants as a skill or stance reached independently. Students reached this active stance when they were capable of independently forming their own opinions regarding musical interpretation. When the student acted autonomously, it seemed to lead to active decisions regarding interpretational choices, which enabled the teacher to respond to the student's thoughts and thus guide the student further. Consequently, the more developed the conversation with the teacher becomes, the greater the potential for the fusion of horizons. If the student was not able to act autonomously, this more passive stance seemed to imply that the student was and continued to be dependent on the teacher as an external source of answers to the questions at hand. The teachers' understanding seemed to be that it should be the student's responsibility to make decisions regarding interpretational choices, through active development of their own convictions. If the student was not able to accomplish this, the teacher tended to return to the practice of presenting (and demonstrating) how the music should be played, likely hindering both the student's learning of musical interpretation and the development of autonomy.

Summary of results
The analysis indicates that the participants mainly defined musical interpretation as a process resulting in a manifested musical performance in the continuum between noninterpretation and over-interpretation. Consequently, performers are required to do more than merely play what is written, while at the same time they are expected to refrain from imposing themselves upon a masterpiece. The teacher tended to be described as guardian and mediator of the composer's intentions, with a privileged connection to performance history, good taste, and aesthetic judgment. The most significant condition for learning musical interpretation seemed to be the students achieving high levels of autonomy. Low autonomy indicated that the student was dependent on concrete answers from the teacher or other sources, whereas high autonomy indicated that the student took a more active stance towards forming decisions regarding interpretational choices. In sum, the students' levels of autonomy and consequently learning of musical interpretation were primarily dependent on three aspects: (1) both the student's and the teacher's understanding of what musical interpretation is, and how it affects the goals, methods, and content of teaching and learning; (2) the freedom of interpretation experienced by the student and acknowledged by the teacher, and how this affects the boundaries of the student's interpretational continuum (which I propose labelling an interpretational paradigm); and (3) the student's explorative approach and the teacher's expectations thereof, and how this affects the (educational) contract between teacher and student regarding the division of labour as well as responsibility for the student's progress. Further, issues regarding these aspects were not reported as explicitly verbalised or addressed during lessons.

Discussion
In this section, the main results are discussed, drawing on the theoretical framework of the study as well as previous research.

An understanding of musical interpretation
The results suggest that teachers and students rarely discuss what musical interpretation is in-depth, if at all. The lack of aesthetic reflection concerning interpretation is quite remarkable in a context where the interpretation of notated scores is considered to be central.
However, this lack of discussion is in line with earlier research (Karlsson & Juslin, 2008) and also consistent with the view that expressivity is mainly dependent on innate talent and thus not possible to learn (see Sloboda, 1996, for a discussion of this view). The extent of this lack of discussion is difficult to estimate; however, it is reasonable to assume that it affects the conditions for learning musical interpretation in the one-to-one context. If the goals of teaching are not mutually defined and understood, it is difficult to see how teaching can efficiently guide the student toward these goals. Teaching might then not foster a student's independence, but rather bind the student and teacher together in an unhealthy way, and thus hinder the development of student autonomy (Burwell, 2005(Burwell, , 2006Gaunt, 2009 ;Hultberg, 2010;Lehmann, 1997). The complexity of evaluating the student's learning of musical interpretation is expressed in the following poetic condensation based on the transcriptions of the interviews with the students:  (2011) calls indifference. Indifference, in this sense, implies negative consequences for the learning outcome, as students intentionally or unintentionally ignore teachers' suggestions that could have been important for their development (Holgersson 2011, pp. 164-165;Hultberg, 2008, p. 20;Gaunt, 2017, pp. 45 & 58).
Apparently, the focus on the transmission of the how-question (the solution for a particular problem) is prioritised over the why-question (the overarching principles and general thoughts behind the decision). Metaphorically speaking, this can be seen as a limitation that stems from a non-disclosure agreement between the teacher and his or her own former teacher(s). In such a contract, the parties have outlined secret knowledge regarding musical interpretation (i.e., the why) that is shared between them but withheld from third parties. This process firmly inscribes the teacher in the tradition and makes it difficult for the student to inscribe him or herself in the tradition without replicating the teacher's doings. Thus, the process could metaphorically be understood as original sin.
Consequently, musical interpretation (and the learning thereof) becomes an act of reproducing history rather than a device for expressing (the student's) personal feelings (see Leech-Wilkinson, 2019). The persistent appeal to authority, tradition, and prejudgments in teaching and learning musical interpretation in higher music education goes against the principles of Gadamerian hermeneutics (Dutt & Gadamer, 2001, p. 44) and constitutes an important problem that I will return to in the closing part of this discussion.

Experienced and acknowledged freedom of musical interpretation
The students in the music teacher training program seemed to feel restricted in their interpretative domain. This feeling might be due to the students' lack of knowledge or the teachers' conceptions thereof, or to the fact that the teacher prescribes how the music should be performed, thus hindering the students from exploring how they would like to interpret the music (Silverman, 2007, p. 112). Consequently, when students experience that they are allowed limited freedom to engage in musical interpretation, they have little space for musical navigation, and the teacher may feel forced to give steering guidance. On the other hand, if students experience and are afforded a large degree of freedom to engage in musical interpretation, lessons can be devoted to discussions of the quality of the students' interpretation. Then the teacher can focus on helping students to clarify or overcome problems rather than giving definitive answers. Thus, it is important that the student is both aware of and can navigate within the boundaries of freedom of musical interpretation set by the teacher. For students to navigate successfully, they need to have developed sufficient tools, skills, and strategies to be able to cope with the interpretation of the material. Consequently, it is important that teachers focus on making it possible for students to learn these general skills.
That the students in the music teacher training program seemed to be more dependent on the teacher, listened to Western classical music to a lesser extent, were more scared of committing errors, and seemed to be negatively affected by hearing others play better than themselves both within and outside of the educational setting, could be considered pedagogical problems. One explanation for the students' limited acquaintance with Western classical music could be the finding that less than 1% of the time spent in instrumental lessons was used to discuss recordings and the use of libraries (Zhukov, 2008, p. 172). In Zhukov's study (2008, p. 172) she argued that instrumental students, especially at the beginning of their studies, need encouragement to use such resources.
Students in the musical performance programs reported taking a more open and active stance towards making decisions regarding interpretational choices, and this tended to be interpreted by the teachers as a result of natural talent or giftedness rather than something that could be stimulated by the teacher (see Sloboda, 1996). This is in agreement with Bränd ström's (1997, p. 96) results, suggesting that teachers at higher levels in the educational hierarchy are more likely to take an absolute perspective on musicality, i.e., that it is innate, instead of a relativistic perspective, i.e., that it can be learnt. The tendency to view musical skills as innate might also have consequences outside of higher music education.
If future music teachers think that proficiency in musical interpretation is based on giftedness or talent, it will most likely have consequences for their teaching and consequently their students' learning potential.

Expectations of the students' explorative approach
The teachers assumed that students naturally should have acquired an explorative approach to musical interpretation. If it is the case that teachers lack experience in systematically acquiring competences in musical interpretation, and thus lack reflection on the nature of these competencies, these shortcomings may be an explanation for their assumptions about students. Another explanation is that some teachers may tend to depict less successful students as either lazy or ungifted, a view shown in previous studies (Sloboda, 1996).
The teachers in my study tended to state that they thought the master-apprentice model was outdated, yet they did not actively help the students to break free from this convention.
Thus, there seems to be a discrepancy between what is said and what is done, which is in line with earlier research that found that teachers said that students should take responsibility for their learning, while their verbalised approaches to teaching tended to leave little room for that (Gaunt, 2008, pp. 239-240).

Students' levels of autonomy
Students and teachers reported that they rarely or never explicitly shared and discussed their thoughts about what level of autonomy the students were supposed to demonstrate during-and between-lessons. Nevertheless, the teachers seemed to expect the students to be autonomous, but thought that they seldom demonstrated such autonomy in practice.
If the teachers were unsatisfied with a student's level of autonomy, they said they would inform the student. However, the teachers in the study said nothing about how they would help the students develop greater autonomy. Instead, it was left to the student to become autonomous. Consequently, the students were implicitly rendered with the responsibility to approach a problem without the teacher's help, or a problem that was addressed in a way that the student had not understood. This placing of responsibility for learning on the student has a long history going back to before the enlightenment (Jørgensen, 2000, p. 67). Nonetheless, this emphasises the importance of how teachers may handle their own responsibility to encourage or facilitate students' development of autonomy. This question is in line with earlier research which found there was "little understanding of a potential reciprocal relationship between self-confidence and self-responsibility or autonomy in learning" (Gaunt, 2008, p. 239). I agree with Gaunt (2008, p. 240;2009, p. 202) who calls for further research regarding how teaching affects students' autonomy and which teaching methods might positively influence the students' autonomy, i.e., how teachers can help their students avoid being "too closely aligned with them" (Gaunt, 2009, p. 203).

Conclusion
The master-apprentice tradition still seems to be in use, even though some participants in this study stated that they find it old-fashioned and inappropriate in higher education. One of the problems with this model is the non-transparency of the philosophical and practical aspects of musical interpretation, and that knowledge is seen to be embodied by the master, which makes the student dependent on the teacher (Kvale & Nielsen, 1999. Consequently, teachers in higher music education may, at least partly, still need to find a way to develop their practice from that of a "transfer pedagogy" to a "transformative pedagogy" (Carey, Bridgstock, Taylor, McWilliam & Grant, 2013, pp. 5-6 & 9). In addition, one-to-one teaching is expensive, which might become an even bigger problem for higher education in the future if funding is not allocated accordingly. Nevertheless, one-to-one teaching has important strengths that need to be understood and communicated (Carey, Lebler & Gall, 2010, p. 176).
Several philosophical and practical aspects of musical interpretation and the learning thereof still seem to be both non-verbalised and lack analysis of, e.g., the interpretational paradigm (see below) within which the learning takes place, ideas about and expectations for learning outcomes, and what criteria these might be dependent on. The teachers often conclude that learning outcomes depend on the student's given abilities rather than on the teacher's actions or the activities during lessons. Since the present study focuses on the teachers' and students' verbal utterances, further research needs to be conducted concerning how teaching and learning of interpretation in the one-to-one context are carried out in practice. There also seems to be a need for the development of more precise definitions and terminology, e.g., including what the differences and similarities are between interpretation and imitation, interpretation and practice, interpretation and performance, and expression and emotion.
Lastly, more comprehensive philosophical discussion about what interpretation, as a fundamental artistic concept, is supposed to be about, seemed to be lacking. Instead, teachers and students seemed to focus on moulding a particular musical performance. The lack of artistic reflection on interpretation as an action and expression of agency may have consequences for both teaching and learning in the one-to-one context and at a higher level for institutions of higher music education. From a philosophical point of view, both the experienced and acknowledged freedom of musical interpretation can be seen as an interpretational paradigm. 14 If such interpretative paradigms are verbalised, it might be possible to negotiate a shared understanding, i.e., the demarcation of the stage that the musical drama is to be performed in and on. Such negotiations might also facilitate discussions about other relevant topics, including learning outcomes, assessment, and alignment of learning objectives (Pettersen, 2005(Pettersen, /2008Ramsden, 2003). One suggestion is that it would be of interest to try to establish which interpretational paradigms are at work in a particular learning context in order to uncloak the hidden book of rules by which a student should abide. In addition, a fruitful path for further studies could be to philosophically analyse dialogical theories of learning both implicitly and explicitly verbalised about and enacted in the teaching and learning of musical interpretation in higher music education. A good start for a fruitful conversation is formulated as follows by Gadamer: Whoever appeals to authority and tradition will have no authority. Period. The same thing goes for prejudgments. Anyone who simply appeals to prejudices is not someone you can talk with. Indeed, a person who is not ready to put his or her own prejudices in question is also someone to whom there is no point in talking. (Dutt & Gadamer, 2001, p. 44) 14 For further context, see the discussion about the concept Werktreue (Goehr , 2007), the old (original) and the new contract between composer and performer (Kivy, 2007 , pp. 100-103), the argument for the negotiation of the musical work (Butt, 2002, pp . 96-122;Östersjö , 2008), and Challenging performance: Classical music performance norms and how to escape them (Leech-Wilkinson, 2019).

Postlude
Lastly, a reflective, poetic postlude concludes this article, thus giving it a ternary form. 15 However, when the first section returns after the contrasting middle section, it is never the same. Such constant development could metaphorically be understood as akin to the oscillation between the parts and the whole, i.e., one of the fundamental aspects of hermeneutics. He received his master of education in music and master of music from ditto university.
Holmgren's has published in European Journal of Philosophy in Arts Education and presented at Swedish, Nordic, and international conferences. His research interests centre on teaching and learning of musical interpretation in higher education, hermeneutics, languages, translation, and poetry.
15 The poem is based on the study's results, my experiences as a piano student, one relatively recent book about musical interpretation (Carlsen & Holm, 2017), one older text authored by a piano teacher within higher music education in Sweden (Pålsson, 2008), and ends with a commenting fragment from If you ever reach Samarkand also used in the preludial poem.