1. Introduction

1.1. Musicology and performances

¶1 The idea of studying performances, in the plural, may seem strange. Studying music has generally meant either studying performance, at a conservatory for example, or studying musical works and ideas about music, as typically happens in universities. But why would we study performances? Are they worth studying? Are they not too ephemeral, too unpredictable, too whimsical to teach us anything useful about music? Or, in the case of studio recordings, too artificial? And anyway, how would we do it? Would it involve taking that analytical and historical approach to music that universities have fostered and applying it to performance? Or is it something that conservatory students would do by listening in order to refine their own playing or singing? And why would it be interesting? Surely it’s the work itself that deserves our attention, not what some performer has done with it.

¶2 If you’re a performer I hope you found that last sentence as infuriating as I do. Why do academics and (on occasion) critics have such a low opinion of performers compared to the almost limitless respect they seem willing to afford composers? If you’re a scholar you may feel that there’s an important truth in the long-standing belief that works endure while performances pass by. But in that case I hope I’ll be able to persuade you that performances are much more the work than we have traditionally supposed, that performance traditions influence the ways we think about works over long periods of time, and that performers have things to teach us about pieces of music that are every bit as interesting and true as the most subtle analyses and commentaries.

¶3 It’s certainly the case, though, that up till now musicology has not seen performance this way. On the contrary, musicology, when it’s dealt with performance at all, has seen its own role as educative, telling performers how they ought to play a piece, either in order to accord with an analyst’s view of the structure of a work, or in order to reproduce a historical performance practice ‘appropriate’ to it. The idea that performers might have something to teach musicologists has, until quite recently, been inconceivable. 1 One could probably argue, with some historical support, that this wasn’t always so. As we shall see in the next section, the rise of musicology, from about 1800, coincides (and it can hardly be coincidence) with a relatively rapid change in attitudes to music. Music of the past became interesting for the first time just when the focus for thought about music changed from concern with its effect in performance towards concern with a piece’s inherent nature. A piece became less a sequence of sounds that pleased the ear and more a creative statement of a gifted mind. It became a work: the outstandingly gifted composer became a genius. 2 Musicology developed partly in order to enable us to understand how music had reached this exalted state (music history), and partly in order to explain the work of genius (music analysis).

¶4 This shift of attention away from performance towards the composer and his work brought a number of advantages. Some were commercial: it was possible to exploit and market works much more widely than performances. Some were intellectual: concentrating on the work gave one a text to study, something one could see and write about in such a way that one’s readers could check one’s insights against a score and learn to share them. Some were spiritual: by studying composers’ works it became possible to conceive music as a procession and at the same time as a progression in which each composer and each work took its proper place; an immensely reassuring and optimistic view of one’s place at the pinnacle of musical achievement so far.

¶5 Of all these, the most significant from our perspective is the view of music as founded in a notated text. From the later Middle Ages onwards, notation contributed increasingly to the transmission of music from person to person and from place to place, so that by the time music came to be identified with musical works notation was able to provide a quite detailed representation of those aspects of works thought to be the most essential. Not surprisingly, then, works and scores became increasingly synonymous. What the composer wrote down gradually came to matter rather a lot, and musicologists increasingly (especially from the later nineteenth century onwards) saw one of their most important functions as ensuring that published scores presented precisely the notes put down by the composer. As a consequence, performers in the twentieth century were increasingly expected to follow that notation strictly and without deviation, and analysts increasingly believed that by studying the written notes they could reach an understanding of the essence of the work. The piece became its notation to the extent that in English ‘the music’ came to mean ‘the score’, as in ‘I’m just going to look at the music’. It takes a moment’s thought, now that this way of thinking is so ingrained, to see why that phrase is oxymoronic.

¶6 Towards the end of the last century a backlash against the text-driven view of music began to strike. Partly it was a reaction against the idea that the accurate reproduction of a composer’s text produced an accurate representation of the composer’s intentions—so that the notion of historically informed performance came under attack. Partly it was reaction against the idea that analysis could explain the meaning of musical works—so that music began to be seen as a representation of a culture rather than as a self-contained autonomous art form. Both these trends have brought welcome doses of realism into the world of academic music, the one allowing once again the possibility that performers might have something to bring to scores; the other making it easier, after a century of rampant pseudo-science, for musicologists to admit that what music means depends not just on what it consists of but also, and to a very large extent, on who we are. But even so, musicology is still, for the most part, operating with the notion that works are one category of music and performances another.

¶7 Something else that has made performances hard to think of studying has been the problem of talking about music as sound. A performance is not simply a neutral representation of the pitches and durations notated in a score. Most details of a sequence of sounds cannot be specified in existing notations. If a notation capable of specifying every detail were developed, no human performer would be able to realise it. And yet those details do play an essential role in how music is perceived, which in turn feeds back, as we shall see, into how it is performed. A musicology that took account of more of this information would have to have a means of observing and considering it in detail. But sound is extremely hard to discuss in non-scientific terms, much harder than the sensations through which we perceive the other arts. One can talk about how a painting looks, without having to invoke measurement, in far more detail than one can talk about how a piece of music sounds: light is not just described as bright or dark, there is a rich vocabulary of words we can use to discuss colour and shape. Even touch can be described in several ways, hard, soft, rough, smooth. We talk about sound much more than touch, but the only words we use that are specific to hearing are loud and quiet; every other word we use is borrowed from another sense (bright, dark, hard, soft, sweet, silvery) or identifies the sound producer (percussive, brassy) or a style of behaviour (bold, timid) or a kind of motion (liquid, halting). In the same way as with scent (something smells of raspberries, or is lemony) we can talk about musical sound most easily by imagining what it is like. We use metaphor and simile, in other words. (There are good reasons why, and we shall come to them.) But that is not a lot of help if one is trying to explain precisely how sounds work in relation to one another, which is something musicology has become very good at by restricting its attention to the things specified in scores.

¶8 It’s also true to say that a lack of vocabulary both results in and is symptomatic of a lack of perception. We don’t actually notice sound very much when we listen to music. Those of us trained so thoroughly that we can write knowledgeably about music tend to hear it, instead, in terms of the categories in which we’ve been taught to think of music—melody, harmony, variation, development, and so on—all things that can be seen relatively easily in a score. The precise ways in which those categories are presented in a performance, and the quality of sound at any moment, are generally things to which we pay less attention. Consequently, although it may seem self-evident if one stops to think about it that this is what music is, musicology tends not to think that a piece of music is the sensual experience of hearing it.

¶9 Yet sound is an ever-present factor in our response to music. What music does to us, it does not just through the interaction of pitch material organised into form, but also by changes or irregularities (often minute) in loudness and timing, changes that, by a process we shall study in later chapters, produce changes in the way we feel. Music is not so much a thing ‘out there’ as an interaction between sound and a hearer (who may also be a performer); it happens in the hearer, not in the score. But to acknowledge that, one must also admit that music involves performance. Without that admission, an appreciation of the quality of sound, its constantly varying details, and its role in generating emotional responses to music, is unlikely. Put together the difficulty of considering sound with the belief that the composer deposits music in a score, and one arrives at a mentality in which performance must inevitably seem peripheral.

¶10 By contrast, in most cultures of the world, and in most musical subcultures in the West, music only happens in performance. Apart from those who study western classical music, what people everywhere mean by ‘music’ is something that one does and listens to. That’s not to say that musicology has been wrong to use notation as a medium through which to study music, or to use the notion of a work as a heuristic, making possible certain kinds of thoughts that are useful but that would otherwise be extremely hard to have. Musicology has achieved extraordinary things, and has had a profound (and, since it has begun to train composers, an ever-increasing) influence on composition, which to an extent has made its approach increasingly appropriate. But musicology has achieved much of this at the expense of the otherwise universally-shared sense that how music sounds while one hears it is fundamental to what it is, and at the cost of isolating itself from the vast audience of music lovers who seek to understand what they experience.

¶11 Perhaps this may help us understand why musicology is peripheral to the interests of most lovers of music. Nicholas Cook, in his book Music, Imagination, and Culture, distinguishes between ‘musical’ and ‘musicological’ listening, the former being what most people do (including, Cook suggests, musicologists when they are listening for pleasure), while musicological listening focuses on hearing the kinds of things that musicology studies. What seems to be happening is that, while all listening to music uses general principles of auditory organisation—grouping sounds, for example, in order to identify relationships between them, or interpreting them in order to assess the information they signal—expert listeners will tend to group sounds by their understanding of the grouping of musical materials in composition—for example hearing relationships between themes and between tonal areas—and to assess meaning in terms of the function of those musical materials in a musical structure, while non-experts will tend to group sounds by their proximity in time, and to interpret their meaning in terms of their inherent beauty and the kinds of memories and associations they seem to evoke. It seems very likely that musicians hear the things non-musicians hear but ignore them to a considerable degree when listening musicologically. Much less certain is that non-musicians hear the kinds of groupings musicians do, but recognise them only subconsciously, although this is what music analysis has tended to suggest in order to substantiate its claim to be revealing truths about how music works. A sceptical view of the claims of musicological listeners might lead one, as Cook suggests, to ‘conclude from all this that the conventional theory of music, in which sonata forms, tonal structures, and thematic relationships play so large a part, is no more than a theory of unheard forms, imaginary structures, and fictitious relationships’. 3 Cook quotes the musical phenomenologist, Thomas Clifton, to the same effect: ‘…for the listener, musical grammar and syntax amount to no more than wax in his ears.’ 4

¶12 It seems hard to deny that musicology has made relatively little impression on most listeners to western classical music, despite popularised presentation of its findings in books, programme notes and broadcasts. The difficulty is not just one of vocabulary. Most people have enough idea of what melody, harmony, repetition and variation are to be able to follow a reasonably analytical discussion of music with which they are familiar. The problem is rather that such a discussion has no obvious correlative in the way they hear the music as they listen. People who are not musically trained simply don’t listen to form defined by thematic material and harmonic progression. They listen to each moment as it happens and respond to it through feeling, not through conscious analysis. Cook may well be right to suggest that musicologists do this too, while at the same time following those more technical processes that they are trained to hear, even though it is only the latter that they use in order to explain what they feel. An interesting study by Watt and Ash proposes, on the basis of experimental evidence, that ‘when music is perceived, it is assigned attributions that would normally be assigned to a person.’ 5 My suggestion here is that for musicologists this person appears to be someone rather like themselves, who engages in a complex internal debate, presenting contrasting hypotheses (represented by themes) in intellectual (harmonic) contexts and working out the ideas they suggest. For everyone else the person represented seems more like an actor, responding emotionally to situations that develop around him, situations that may be calming or enlivening, beautiful, disturbing, frightening, or thrilling. It’s not hard to see that the experiences of this listener depend to a very considerable extent on the way performers shape the musical structure notated in the score. 6 What may be less immediately obvious is that the musicological listener receives a great deal of information by the same route.

¶13 Another challenge for musicology, then, and one that needs to be faced alongside finding ways of studying music as sound and in performance, is to reintegrate into its view of what music is, and of what musicology’s business embraces, the evident fact that music works with our emotions. At one time this was taken for granted. In the seventeenth and eighteenth centuries musical ‘affects’ were a leading topic for intellectual discussion of music. But as we’ve seen, during the nineteenth and twentieth centuries the study of historical and musico-linguistic facts came to be valued much more highly. Indeed, it became a matter of principle to keep emotional response to the material out of musicological work. 7 More recently the situation has entirely changed. At the end of the last century, ‘new musicology’ made a point of validating subjective response, bringing in its wake benefits for all of us. While, until now, new musicology has not paid very much attention to the idea that performance, rather than simply the mind’s response to the score, might be seen as the crucial trigger for subjective response, 8 what it has achieved in talking about scores makes it far easier now to see how studying music as a mechanism generating directed emotional response might be professionally achievable.

¶14 At the same time, musicology has begun to notice what’s been going on in science for the past three-quarters of a century and increasingly since the 1980s. With the development of experimental psychology and neurology it has become increasingly evident that the brain responds to sounds in predictable ways. Subjectivity only seems riotous; many of our responses to music are shared, and those that aren’t are still the result of coherent and potentially understandable processes. There are good reasons, then, why a particular note in a phrase is stretched, and they are not just music-analytical; 9 there are good reasons why when he sings that note Pavarotti sounds anguished, and they are not just theatrical. Behind these layers of reasoning are physiological processes that determine not just what performers do but also how we feel about it. And the full meaning of those moments, which musicology exists to explain, depends on a great many factors that are currently the business of other disciplines. Thinking about what the performance of music does to us, therefore, has the potential—indeed, it makes it necessary—first to open up musicology and ultimately to integrate it with scientific disciplines, because it’s here, and not in history, that the aims of science and musicology meet. We all want to know what makes music meaningful.

¶15 Of course, this is to look ahead rather a long way. But enough has been done already in affective science and the psychology of music perception to give us some pointers towards the processes by which music engages the emotions when we perform and listen to it. 10 Much of this work has powerful explanatory value for musicians, and we shall have to take account of it below. At the same time, what musicians do has itself much to teach us about how the brain processes sound. Musicians in fact have more to offer than many realise. Performances provide the data. So there is a great deal more at stake in this new field than simply studying performance as a branch of current musicology; just how much will, I believe, gradually become apparent over the coming decades. Here we shall do no more than touch on a few of the possibilities. But it is certainly one of the aims of this book, while also offering practical approaches to studying details of performance from recordings, approaches with historical, analytical and performance-orientated applications, to suggest areas in which we might now start to look for answers to much wider questions.

1.2. Musicology and recordings

¶16 I hope that what I’ve said so far helps to explain why I think there’s a case for studying music as performance and from performances. In particular, we can hope to discover things about how music works in sound and about what performers do to make it work. What I’m aiming to do in this book, therefore, is to take a step towards a musicology that studies how music works in us through performance.

¶17 How might we do this? We’ll need a way of studying sounds, and what performers do to make them, and we’ll need to start, at least, to think about how we perceive them. As I’ve suggested in talking about musicology and science, it’s immediately clear that even a limited programme of work, focussing on what performers do, is going to involve bringing together several areas normally kept separate, including performing as an activity, acoustics, sound communication, perception and cognition, the emotions. But before we can look at any of those things we’re going to have to have some performances to study. And that means looking rather closely at recordings.

¶18 There were a few academic studies of recorded performances before modern times. Louis P. Lochner quotes perhaps the first analysis of a performer’s style from recordings, made in 1916 by Eugene Riviere Redervill who examined Fritz Kreisler’s vibrato. 11 Several studies by Carl Seashore’s research group at the University of Iowa in the 1920s took records as their sources of data, partly on the grounds that it enabled others to repeat their experiments exactly. 12 But musical academia has on the whole, at any rate until recently, taken a fairly dim view of recordings. 13 Robert Philip tells how, when he ‘wrote to the professor of music at Oxford, Sir Jack Westrup, to ask whether I might be able to do [research on recorded performance] at Oxford, he replied, “I feel bound to tell you that I could not recommend this topic to the Faculty as appropriate for research.”’ 14 Partly, this is because performance has traditionally, among musicologists, been looked down upon. Musicologists, as I have suggested, were thought to understand, performers merely to do. And that snobbery is still very obvious in the relative importance in the academic mind of learning about music, of playing it and of listening to it on disc. In universities playing is often a bonus, a way of getting good marks for an extra skill if you have it. Listening to recordings is considered to be a kind of shortcut to study, saving students the bother of reading scores. Understanding scores and knowing some history—and the self-delusions involved in that phrase would make another book—is what really gets you the credit. It was surprisingly recently that British university music departments started to make a serious effort to use recordings. Even when I started as a lecturer, in the mid-1980s, most staff still made little use of recorded examples in their lectures but read orchestral scores at the piano in order to illustrate a point. It’s a hugely impressive skill, of course, and a musical one. But what mattered was to illustrate a point about the musical argument, the argument in the score, which was thus seen as a record of a kind of intellectual process in which notes conversed or argued with one another until the composer managed to get them all to agree at the end. This had nothing to do with any particular manner of performance; that was simply a matter of fashion, entirely separable from the essential nature of the music. Music existed in its purest and most ideal form on an abstract plane, and it was there that one must strive to encounter it, not in the unreliable and emotionally charged domain of any particular performer’s take on the piece. Consequently, one bought lots of scores for the library but many fewer recordings. And the increasing importance of record collections was seen as a thoroughly undesirable consequence of incoming students’ increasing inability to read scores. One bought recordings because the poor dears weren’t competent to read music any more.

¶19 I don’t think anyone could honestly maintain that music students are as ready to read music from scores as they once were. But to argue that their understanding of music was richer when they worked only from scores one has to have what now looks like a rather narrow view of what ‘music’ is. For anyone who believes, as I’ve begun to argue above, that music’s meaning depends on far more than just that argument between the notes, it seems quite obvious that the visceral impact of sound, and the variation between performances of scores, need to be among the factors available for study when we work on a piece. How our bodies respond to music, and how differently they respond to different performances, are matters of fundamental importance if we want to understand what music is.

¶20 The main danger of using recordings is that we may use too few of them, getting used to just one view of a piece and losing sight of music’s essential ‘unfixedness’. This is a subject we’ll come back to later on. For now it’s enough that we begin to entertain the thought that the impossibility of pinning down the identity and full meaning of a piece of music in one ideal performance might be not a problem but rather a defining feature of music, one of the main sources of its power. 15

¶21 There is, of course, a contradiction between celebrating the unfixedness of music and arguing for the value of recordings, which fix a performance in such an unnatural way. But the contradiction is one that we must simply accept as a necessary compromise in the interests of having something to study. The problem can be alleviated by taking a lot of recordings of the same piece: although each is fixed, the variability between them gives a good sense of the apparently endless variety of approaches that can be taken to turning scores into sound. If we take enough samples, as it were, we can come close to a sense of the openness of music to performance—performance in the sense of something made without full knowledge of how it will be made until it happens. Ideally one would study this process in real time. There is valuable work to be done gathering responses to music as it happens. Indeed it’s probably the most valuable work that can be done with music, and in time people will get a lot better at it. But it’s at an early stage, and at the moment there is probably more to be learned from seeing recordings as performances to which we can return, and from studying details of them closely. It’s also worth bearing in mind that, for reasons that are not yet well understood but probably have to do with pre-conscious triggering of responses to sound stimuli, 16 and thus significant construction of musical meaning in real time or at most in short-term memory, we tend to hear repeated performances on record to some extent as if for the first time: we don’t necessarily (although eventually we may) get bored or find ourselves unsurprised or unmoved by hearing the same details again and again. 17 There will be more to say about the relationship between recordings and performances below.

1.2.1. Expertise beyond musicology

¶22 Academics, in fact, are coming very late to the study of recorded performance which, as we can see, has been going on outside academia for the better part of a century; and we have an awful lot of catching up to do. Supreme in their hold on facts are the specialist collectors, who may be interested primarily in one kind of repertoire, even in just one performer, or in the output of a single record label, but who accumulate in that field unrivalled knowledge from which academia has much to learn. 18 There is a great deal of listening and talking to be done with experts who have far more experience of working with recordings than any of us. Most are now middle-aged or older, and much of their knowledge is undocumented. I mention this because it should be a priority for academics in this field to seek out collectors, discographers, and enthusiasts wherever they can find them, to learn as much as possible, and ideally to offer university space for their collections, whether documentary or on disc. Many lifetimes’ work will disappear into skips over the next decade or so unless we do something now.

¶23 What sorts of knowledge will be lost? Knowledge about how records were made, for a start: much of this is lost already; it is now very difficult to answer quite obvious questions about the processes used by different companies and engineers in the making of 78rpm discs and cylinders; questions whose answers would make a significant difference to the way we understand what we hear when we play them. 19 Almost all the really detailed information about the recording of dynamics and tempo on piano rolls has gone. 20 Partly this is because much was kept secret for commercial purposes. But a huge amount could have been preserved if only anyone had thought it worthwhile.

¶24 Equally, a great deal has been recovered by enthusiasts in modern times, above all, discographies—lists of recordings made, of what is on them, by whom and where and when they were made, where (if anywhere) they were published. If you’ve not yet tried to assemble this kind of information you won’t realise just how difficult it is. 21 There is no universal discography comparable to bibliographies such as RISM, RILM, or The Music Index. Most published discographies are woefully inadequate for our purposes: even modern discographies habitually don’t bother with an index of titles, or if they do then they don’t bother with an index of artists. Modern discographies are much better about dates of recording, issue and deletion, but older publications (and some of the largest discographies are fifty or more years old) tend not to offer them. The most reliable way to find out what was recorded—and even this, supposing one had time for it, would not be very reliable—would be to read through every annual catalogue produced by every record label over the past 100 years. 22 These are some of the reasons why the work of enthusiasts is so valuable, and why it’s so important to preserve it. The alternative is for a generation of academics to do it all over again. And to be brutally frank, it is not the sort of work that academics do any more: we are all too busy having good ideas. There are not many universities, now, where you could get a PhD for a discography, though few things in this field could be more valuable.

1.2.2. Record reviewing

¶25 It is easy, too, to underestimate the significance of eighty years of record reviewing. The first specialist journal of the gramophone, 23 the Phonographische Zeitschrift, began publication as early as 1900, and was concerned with the applications of recording: what could it be used for and how? 24 It started to review recordings in 1905/06. Record reviews appeared regularly in newspapers in the UK much later, in the early 1920s, but at a significant moment. As an article in The Times reported on 22 July 1921,

A member of a large gramophone manufacturing firm said yesterday that the so-called ‘popular’ records were not nearly so popular as they used to be. Far better business was being done with records of good pieces of music and of good songs by good singers, although the latter were more expensive than the ‘popular’ records. Records of comic songs and ragtime music combined quantity with cheapness; but what the gramophone user required at the moment was quality. He preferred fewer records, so long as those few were good.

The change in taste had been unexpectedly rapid. During the war the ‘popular’ record was ‘all the rage’, but after the war came industrial depression and now not many records were being bought for amusement pure and simple. People simply could not afford to do so, and the result was that the buying of records of any kind was now largely confined to the so-called ‘cultured’ classes. More of those were buying records than ever before, and they had created such a demand that the supply of good records had been steadily growing.

My informant added that Wagner, Stravinsky, and the Russian composers seemed to be most in favour at present, and good vocal records had a steady sale. 25

¶26 In this context it’s easier to see why the newspapers of the ‘cultured classes’ chose this moment to begin regular reviews. Musical Times had already started a monthly column in January 1921 and the idea quickly spread to newspapers. The Daily Telegraph, after occasional pieces in 1921, began regular record reviews on 4 March 1922. The Times likewise included occasional items during 1921, but then heralded regular reviews, which began on 9 March 1922, with an essay on the gramophone and close listening on 23 February 1922 and a review of Indian language records the following week. The Scotsman began reviewing records on 18 May the same year. All this suggests a widespread perception by then that recordings were of substantial artistic interest, worth reviewing alongside books.

¶27 It’s at this point, therefore, that writing about recorded performance began in earnest. So when the first periodical devoted to record reviewing, The Gramophone (later just Gramophone), was founded by the novelist and recent convert to music on record, Compton Mackenzie, in 1923 it was able to recruit reviewers with some newspaper experience. 26 The most remarkable of Mackenzie’s catches was Hermann Klein, born in 1856, a student of Manuel Garcia, the younger, previously music critic of The Sunday Times and a man able to remember meeting and hearing many of the most famous composers and performers of the second half of the nineteenth century. 27 It was Gramophone that set, and for decades (until the appearance of illustrated radio record reviews after the Second War) maintained, the standard for record reviewing. Its reviewers were remarkably loyal, 28 and over the decades one can map both their responses to very gradual changes in performance and changes in their own taste. 29 It’s worth bearing in mind, also, that reviews represent an enormous depth of knowledge about musical performance, and although it can rarely be plumbed in the standard rather brief reviews, the thematic surveys that became increasingly common towards the end of the last century often have much to teach us about how performances may usefully be described and compared in words alone.

¶28 Musicologists tend, privately at least, to scorn the comparison of performances that one finds in the pages of magazines such as Gramophone, Fanfare, and Diapason, or in radio slots such as the BBC’s ‘Building a Library’ which compares multiple recordings of one piece and recommends one as the best choice: the flaws in that view of performance need no underlining, but actually the sorts of details to which reviewers are attending are those that we shall find ourselves studying, albeit more closely, when we try to understand what makes a performance interesting. The best reviewers have very good musical ears, and are able to notice and understand the significance of features in performances that most listeners only sense rather vaguely, if at all.

1.2.3. Metaphor

¶29 Just as important, and just as interesting from our point of view, is reviewers’ highly developed ability to find analogues in words for the features they hear. We shall see later that metaphor is very much more than just a regrettable convenience for talking about music in performance, necessary to fill the huge vocabulary gap between our emotional responses to music and the technical language available for describing sounds. Far from being a symptom of perceptual and intellectual failure, metaphor is fundamental to human perception, an index of the ways our brains connect up stimuli to generate knowledge and ideas. 30 A skilful record reviewer is telling us things about performances that are true. If we only understood the mechanism that connects the metaphor with the sound on the one hand and the sensation on the other we would understand one of the most important aspects of the way that music works in us.

¶30 What kinds of metaphors are most commonly used? I’ve made a fairly random selection from roundups of ‘recent reissues’ in two record review magazines that happen to be within reach as I write, but I find them quite typical. 31

Suggesting manner or deportment

¶31 sluggish, ponderous, tired, tense, vital, rigid, heavy, clumsy, undernourished, swashbuckling, urgency, sultry, animated, steady, correctness, low-risk, confident, patient, efficient, unselfconscious, radiance, wit, rapture, spontaneity, sympathetic, dogged, poetic, profound, tentative, eloquent, affectionate, tenderness, drama, alluring, matronly, graceful, poise, unostentatious, ardent, impassioned, demonic, unfussy, seriously, energetic, vibrant, drive, vivid, seedy, dodgy, scruffy, fecund, rugged, indomitable, burlesque, bluff, chipper, stern, conscientious, imploring, unearthly, brazen


¶32 bright, dull, monochrome, polished, sparkling, ripe, fresh, burnished, spacious, dense, sunbursts, beam, sparky, monolithic, grandeur, sinewy


¶33 soft, hard, cold, warmth, incisive, coarseness, roughness, bite, crisply, chilling, cold-blooded, stunning, punchy, flexible


¶34 exciting, thrillingly, captivating, delightful, elation, touching, goose-bumps, surge of emotion, hair-raising, chill to the marrow, stirs the blood


¶35 snap, lyricism, rasping, beery clangour

¶36 It’s striking how few there are from the vocabulary of sound. By far the majority describe human behaviour, whether expressed through features that can be seen or features that can be deduced; fewer are derived from the characteristics of things that we can sense through sight or touch (there are none of smell). A number describe feelings we have, but not as many as one might expect. By contrast, a great many have implications for motion. They can be used as descriptions or explanations of particular manners of human movement; they imply something about the speed or ease or regularity or predictability with which one thing follows another, or the kind of impact that is felt when two things move into contact, or when we first catch sight of something. They take us back, in fact, to some of the most essential observations that living creatures need to make in order to know what kind of situation they are in and how they should respond to it. Fundamentally they are all to do with assessing what might happen next. The extent to which our response to music depends on assessments as practical and, in an evolutionary sense, as basic as these will become clearer later on when we look more closely at the way humans respond to sound in life and in music. And when we come to look at the ways in which our minds make connections across domains—for example, between sound and manner, or between sound and touch—we’ll begin to understand the kind of work that these metaphors do and the essential role metaphor plays in our assigning of meaning to music. For now it is enough to recognise that the ability to find the right metaphor for a moment in a musical performance is a very high-level musical and intellectual skill, something that the best record reviewers have to an admirable extent and from which students of performance on record have much to learn.

¶37 So musicologists of performance are very far from being the first to study performance on record. There is a lot of expertise on which we would be foolish not to draw. In the case of the collectors and enthusiasts, we need to make an effort to make contact and to persuade them that we can be trusted with the knowledge they have accumulated, often over a lifetime. In the case of record reviewers their work is there to study, but guided, hands-on experience would not come amiss. Both these points have implications for the teaching and development of our field.

1.3. Why now?

¶38 As we’ve seen, it’s taken a very long time for musicology to get around to studying performances. Why does it seem to make sense now, when for so long it did not? First, the idea that music exists independently of performance, although a staple of musicology and the philosophy of music, is beginning to look distinctly shaky. For most musical cultures in the world (including western popular music) it is nonsensical. For western classical music it’s conceivable only because of notation. The more we believe that notation encodes the work, rather than simply providing sketchy performance instructions using which a performer can make the work, the more inclined we are to believe that works exist in some abstract yet ideal form independent of any performance. It is more than anything else the modern availability in CD reissue of 100 years of recorded performances that has upset those beliefs. Now we can hear that works we thought were fixed (or fixable, given a good enough editor) have actually changed radically in performance over the past century, it’s much easier to think of works as open—fields of possibilities, underdetermined by notation, that seem to change all the time.

¶39 Secondly, there has been the coincidental undermining of the notion of authenticity which gradually, since the mid-1980s, 32 has killed off the belief that there is an ideal reading, even if no one is capable of achieving it, in the shape of the composer’s ideal performance. 33 If I may be autobiographical for a moment, it was listening to recordings that made me realise, sometime in 1983, that the claims of the early instrument authenticists were unsustainable. 34 And I can see now much more clearly than I could then how recorded-sound evidence of performance practice puts a bomb under the notion that one can reconstruct performance styles using only written materials.

¶40 Thirdly, and like it or not, the relativism theorised by post-modernism, but operating in practice in so many aspects of contemporary western culture, has made it seem far more reasonable than it used to seem to accept different instances of the work as equally valid. Clearly one can take that too far; there has to be discrimination on grounds of competence, at least, but one has to recognise that much is a matter of taste. People play in ways that reflect their own understandings of music, and above a certain (rather high) level of technical ability it is probably not sensible to object solely on the grounds that one’s own view is different. Performances that vary as much as those we can now so easily compare are powerfully symbolic of the range of cultures and orientations that we nowadays (in the West) accept as equally valuable.

¶41 Fourthly, and closely related to this, is what one might call post-historicism, the notion, which is only just beginning to appear in musical studies, that history isn’t what it claims to be and cannot be known except from a radically distorted modernist viewpoint. Recordings have a provocative relationship with both that and with the traditional view of music history. On the one hand they tell us—allowing for the shortcomings in the recordings—how it sounded; on the other they tell us that our idea of what is ideally musical is radically unlike ‘theirs’ and cannot encompass it. As a result we cannot know what it meant to hear those recordings ‘then’, let alone the performances behind them. And for that reason early recordings are, in the end, more fully grist to our mill than they are windows on the past.

¶42 For all these reasons, and no doubt many others, the time to study performances through recordings has come. For the first time, to a significant body of musicologists it seems sensible and desirable. 35 What makes it possible, though, has to do less with changing ideas than with developments in commerce and technology. First is the fact that after 100 years of music recording, and thanks to its huge commercial success, we have accumulated enough data to enable us to see how much of musical performance changes over time. Secondly, we at last have sufficiently flexible, powerful and affordable technology to enable comparisons of recordings, on both small and large scales, to be made easily. Digital encoding and personal computing open up fields of study that would have been impractical, and perhaps impossible before. 36

1.4. What will this book do?

¶43 My aim in this book is to set out some ideas about why music might be worth studying from performances rather than from scores, about how we might do it, and about how our view of music will change as a result. I must emphasise that I am writing as a musicologist, and so I am working within some fairly conventional limitations, confining myself to western classical music, focussing on canonical nineteenth-century pieces, and assuming some general knowledge of classical harmony and form. At the same time, what I want to do is to encourage a new generation of music students, performers, and music lovers too, to think about this music in a rather different way. I shall try to persuade you that how the music feels while we are listening to it matters more than how it was composed. Understanding the way it feels is very much musicology’s business, and always has been, albeit often at several removes. But there may be a case now for moving it closer to centre stage. With so much data easily to hand, now seems the time to develop some techniques for making use of it. By proposing some, and by looking at the issues that surround them, I’ll try to suggest some of the directions musicology might take in order to encompass music as performance within its borders.

¶44 I think we have room for this now. We probably know most of what we shall ever know about music history, at least up to, say, the middle of the last century (there is still a lot to be established about more recent events). The study of scores has produced wonderful insights into the relationships that form between notes when composers set them down on paper. But the study of what performers do, and of what that means to listeners, has barely begun within musicology. In the meantime, though, a lot has been happening that is relevant in science, especially within psychology and neuroscience. And it seems very clear that science is best-equipped to answer many of the questions that we might want to ask. But musicians have a contribution to make too, and over the next few years the extent of that may perhaps become clearer. In what follows I shall be making many references to studies that are provocative for ideas we might have about how music works; we shall see that many things that musicians sense are being confirmed by the findings of science. What I hope to persuade you is that a musicology of performance that learns from science, and that teaches it as well, could be hugely worthwhile, and that as time goes on the boundaries between musicology and science, at any rate in the field of performance studies, may become rather harder to discern.

¶45 The things I don’t do in this book, but might have done, or should have done, are legion. I don’t deal with ensemble performance, except for a few passing comments on the relationship of singer and accompanist in Lieder. There are some good reasons for this: I want to study details and to understand who is responsible for them and why, and that is much harder to do when there are a number of players; I also want to be able to use spectrum visualisations, and for them to be useful at all it has to be possible to say which elements map each performer and each note. In due course better techniques will be developed for studying string quartet or orchestral playing, for example, but we’ll make more progress now with simpler materials.

¶46 A much more serious omission is any detailed discussion of the relationship between technique and sound. I talk a lot about the sounds performers make, but hardly at all about what they do with hands and voices to make those sounds. Yet technique is perhaps the most important element in the whole process, since it makes everything else possible. I have a good reason for not dealing with it and also a poor excuse. The reason is that it would have made the book twice as long and several times as complicated; the excuse is simply that I am competent in only one of the instruments on which I focus (the piano), and while I have rather more technical skill in others of the same family I don’t regard them as adequately transferable (and I’m not sure harpsichord playing offers the best focus for this particular book). Nevertheless, technique is an essential part of the subject. Its interaction with sound is hugely complex, but, as Rebecca Plack has shown in a fine study of early recorded singing, 37 when one is able to examine that interaction authoritatively one’s comments on performances can be much more detailed. Being able to distinguish between technical convenience (or even necessity) and expressive intention, often to show their mutual dependence, greatly enriches one’s understanding. Of course there is a danger that having a very firm view of what makes effective expression, derived from the way one does it oneself, can blind one to the virtues of other kinds of musicianship. But the advantages of being able to say what a musician is doing with their fingers or their breath far outweigh that in most cases.

¶47 The book is entirely concerned with moment-to-moment details. It doesn’t cover the statistical investigation of large data sets which, as we have seen from recent work by Cook and Sapp on Chopin mazurkas, 38 and by Spiro, Gold and Rink using the same data, 39 can offer a powerful approach to understanding style. Again, this is partly because I lack the statistical skills, partly because we all worked within a larger research project— CHARM —and the variety of approaches was intentional, but partly too because for me it’s those moments that make music endlessly fascinating, exciting and moving. To understand how music feels that way, and why, is what drives me as a musicologist.

¶48 What will this book do for you? If you’re someone who loves music, perhaps performs for fun, and reads about it in order to learn more, you may feel that so much of the book is obvious that you can’t believe musicology needs a special case to be made for studying it. Of course performances are the music, of course the things that performers do from moment to moment have a profound impact on what the music means to us, of course we need to pay attention to the musical surface and not just to deep structure or to historical and cultural context. But you’ll also understand, I hope, why this is so difficult for musicologists. If you’re a performer, I very much hope that the variety of performance styles we’ll examine may open your ears to a range of possibilities far wider than is currently practised, though you’ll also recognise that to reproduce many of them would be commercially suicidal. But conceptual blending is, as we shall see, one of the things humans are quite good at, and I hope it may be difficult to come away from this book without a sense that, if not anything goes, at any rate a lot more can go today than one might think. If you’re a university music student, especially one who is thinking about a career there, I hope I might be able to persuade you that musicology is full of potential for some thinking about music that has more in common with listening to it and performing it. For the closer those three activities are, the closer we are to understanding what music is.


For the emerging counterview see Joel Lester, ‘Performance and analysis: interaction and interpretation’, in ed. John Rink, The Practice of Performance: studies in musical interpretation (Cambridge University Press, 1995), 197-216, esp. 198; Peter Johnson, ‘Performance and the listening experience: Bach’s “Erbarme Dich”’, in Nicholas Cook, Peter Johnson & Hans Zender, Theory into Practice: composition, performance and the listening experience (Leuven University Press, 1999), 55-101, esp. 56-7; Roy Howat, ‘Debussy’s piano music: sources and performance’ in ed. Richard Langham Smith, Debussy Studies (Cambridge, 1997), 78-107. Back to context...
Lydia Goehr, The Imaginary Museum of Musical Works: an essay in the philosophy of music (Oxford University Press, 1992). Back to context...
Nicholas Cook, Music, Imagination, and Culture (Oxford University Press, 1990), 3. Back to context...
Thomas Clifton, Music as heard: a study in applied phenomenology (New Haven: Yale University Press, 1983), 71. Back to context...
Roger J. Watt & Roisín L. Ash, ‘A psychological investigation of meaning in music’, Musicae Scientiae 2 (1998) 33-53, at 49. There is much discussion of metaphorical mapping between musical and person in Charles O. Nussbaum, The Musical Representation: meaning, ontology, and emotion (Cambridge, Mass.: MIT Press, 2007), esp. chapter 3. Back to context...
For a thought-provoking treatment of musical meaning in listening see Eric Clarke, Ways of Listening: an ecological approach to the perception of musical meaning (Oxford University Press, 2005). On musicology’s ignorance of the listener’s perspective see esp. p. 122-4. Back to context...
Aaron Ridley is particularly good on the consequences of this attitude in the philosophy of music. The Philosophy of Music: theme and variations (Edinburgh University Press, 2004), esp. 7-15. Back to context...
This remains true even of the very recent book by Lawrence Kramer, whose wonderfully evocative descriptions of the effects of compositions seem oblivious to (or take too much for granted) the fact that they refer not to the work per se but to the effects of a performance of it (Kramer (2007), 23-4). Back to context...
Ian Cross, ‘Music analysis and music perception’, Music Analysis 17 (1998) 3-20. Back to context...
A very fine overview is John Sloboda and Patrik Juslin, ‘Psychological perspectives on music and emotion’, in ed. Patrik N. Juslin and John A. Sloboda, Music and Emotion: theory and research (Oxford University Press, 2001) 71-104. Another is Alf Gabrielsson and Patrik N Juslin, ‘Emotional expression in music’ in ed. Richard J. Davidson, Klaus R. Scherer & H. Hill Goldsmith, Handbook of Affective Sciences (Oxford University Press, 2003), 503-34. Back to context...
Louis P. Lochner, Fritz Kreisler (New York: Macmillan, 1950), 272-3. Back to context...
Ed. Carl E. Seashore, The Vibrato, Studies in the Psychology of Music 1 (Iowa City: University of Iowa, 1932), see esp. p. 111; Carl E. Seashore, Psychology of the Vibrato in Voice and Instrument, Studies in the Psychology of Music 3 (Iowa City: The University Press, 1936); ed. Carl E. Seashore, Objective Analysis of Musical Performance, Studies in the Psychology of Music 4 (Iowa City: The University Press, 1936). For an early study of Chopin playing on record see Józef Kański, ‘Über die Aufführungsstile der Werke Chopins: einige allgemeine Probleme der Aufführung auf Grund von Schallplattenaufnahmen’, in Zofia Lissa (ed.), The Book of the First International Musicological Congress devoted to the Works of Chopin: Warszawa 16th–22nd February 1960 (Warsaw: PWN Polish Scientific Publishers, 1963), 444-54. Back to context...
During the first half of the twentieth century this was true of the musical profession as well. See chapter 4 of Timothy Day, A Century of Recorded Music: listening to musical history (New Haven: Yale University Press, 2000), 199-256. Back to context...
Robert Philip, Performing Music in the Age of Recording (Cambridge University Press, 2004), 1. Back to context...
For a philosophical justification see Michael Krausz, ‘Rightness and reasons in musical interpretation’, in ed. Michael Krausz, The Interpretation of Music: philosophical essays (Oxford, 1993), 75-87. Back to context...
Sloboda and Juslin (2001), 92-3. Back to context...
For some discussion of this see Stephen Davies, Musical Works and Performances: a philosophical exploration (Oxford University Press, 2001), 305; and, with some theoretical grounding in studies of perception, Nussbaum (2007) esp. 216-17. Back to context...
See outstandingly the contents of ed. Larry Lustig, The Record Collector: a quarterly journal of recorded vocal art, with meticulously researched articles and discographies. Also Classic Record Collector (previously International Classical Record Collector), edited until recently by Tully Potter. Back to context...
For fascinating work reconstructing early recording processes, however, see especially the publications of George Brock-Nannestad, and Roger Beardsley’s online Guide to 78s . Back to context...
For examples of the furthest we can now get see the contents of The Pianola Journal . Back to context...
For a valuable introduction to discography see Simon Trezise, ‘The recorded document: interpretation and discography’, in ed. Nicholas Cook, Eric Clarke, Daniel Leech-Wilkinson, and John Rink, The Cambridge Companion to Recorded Music (Cambridge University Press, 2009). For useful advice on presenting discographies see the Association of Recorded Sound Collections guidelines. Back to context...
There are fine collections on the open shelves in the British Library (London), in the Humanities 2 reading area. For the Gramophone Co. (HMV, later EMI) there is a full run at The EMI Archive, accessible by appointment. The King's Sound Archive's collection is listed at http://charm.kcl.ac.uk/discography/disco_catalogues.html. Substantial runs are available in most sound archives, though they are not always visible in online catalogues. Back to context...
I exclude trade publications, such as The Phonogram (1891). Back to context...
The first two issues are discussed in Daniel Barolsky, ‘Romantic Piano Performance as Creation’, PhD dissertation, University of Chicago, 2005, 42. I am extremely grateful to Barolsky for checking issues for record reviews. He finds that ‘from July 5, 1906 (Jahrg. 6, 27 [possibly before, as issues from 1905 and early 1906 were not available]), there is a regular section called “Phonokritik” written by Max Chop which reviews recent releases.  It’s organized by record company (i.e. Odeon, Polyphon, Beka, Favorite, etc.) and reviews everything from the sound quality to the composition and performance.’ (Personal communication.) Back to context...
A Correspondent, ‘The Gramophone Habit: “Popular” Records Now Unpopular’, The Times, no. 42779, 22 July 1921, p.7 col. F. I am extremely grateful to Nick Morgan for drawing this and other articles to my attention and for his feedback on an earlier draft of this section. Back to context...
On the history of Gramophone see Anthony Pollard’s fascinating Gramophone: the first 75 years (Harrow: Gramophone Publications, 1998). For Gramophone reviewing as a cultural phenomenon see Colin Symes, Setting the Record Straight: a material history of classical recording (Middletown, CT: Wesleyan University Press, 2004), esp. chapters 7 & 8, pp. 182-211. Back to context...
A fascinating collection of Klein’s reviews is to be found in William R Moran, Herman Klein and the Gramophone, being a series of essays on the Bel canto (1923), the Gramophone and the Singer (1924-1934), and reviews of new classical vocal recordings (1925-1934), and other writings from the Gramophone (Portland, Or.: Amadeus, 1990). Back to context...
Over the (so far) 85 years of Gramophone there have been, for example, only three principal reviewers of song, Hermann Klein, Alec Robertson (from 1934) and Alan Blyth (1972-2007). Back to context...
Barolsky (2005) offers a valuable survey of the early years of Gramophone (40-63) and briefly discusses (63-4) the German equivalent, Fono Forum, which began in 1956. The French Diapason was founded in 1955, and the American Fanfare in 1977. Record reviews appeared in numerous specialist music and hi-fi magazines also, either side of the mid-century. The best way to access record reviews is still Kurtz Myers, Record ratings; the Music Library Association's index of record reviews (New York: Crown, 1956), and Index to Record Reviews (Boston: Hall, 1978, 1983 & 1989), continued in Mark Palkovic & Paul Cauthen, Index to Record Reviews 1987-1997 (New York: Hall, 1998). An admirably full set of the relevant magazines is to be found on the open shelves in the British Library (Humanities 2). Back to context...
George Lakoff & Mark Johnson, Metaphors We Live By (University of Chicago Press, 1980, rev. ed. 2003); Zoltán Kövecses, Metaphor and Emotion: language, culture, and body in human feeling (Cambridge University Press, 2000). Back to context...
Andrew Aschenbach in Gramophone 82 (August 2004) 36-8; Nigel Simeone in International Record Review 5/2 (July/August 2004) 24-5, plus a few that were in my preparatory notes for this section. For a much more colourful example, also by Achenbach, try ‘Disquiet below the surface’ (on Vaughan Williams’ Pastoral Symphony), Gramophone July 2002, 36-40. Back to context...
The classic study here is Richard Taruskin, ‘The pastness of the present and the presence of the past’, originally published in ed. Nicholas Kenyon, Authenticity and Early Music (Oxford University Press, 1988) 137-210, but usefully reprinted with other work on this theme in Taruskin, Text and Act: essays on music and performance (Oxford University Press, 1995), 90-154. See also John Butt, Playing with History: the historical approach to musical performance (Cambridge University Press, 2002). For an attempt to restore the historical status of historically informed performance see Peter Walls, History, Imagination and the Performance of Music (Woodbridge: Boydell, 2003). Back to context...
Still very much alive even in Adorno, Towards a Theory of Musical Reproduction: notes, a draft and two schemata (Cambridge: Polity, 2006). Back to context...
In preparing a lecture in the Music Faculty at the invitation of Peter Le Huray, written up as ‘The limits of authenticity’, Early Music xii (1984) 13-16. Back to context...
To assist musicology to work with performance on record was the purpose of the AHRC Research Centre for the History and Analysis of Recorded Music: see www.charm.kcl.ac.uk for details of its work. Back to context...
The publication of Robert Cogan’s New Images of Musical Sound in 1984 alerted me to the possibility of studying performances in real detail using spectrograms, and together with my colleague Phil Gaffney at Churchill College, Cambridge, I looked into the practicality of a PC-based spectrographic analysis system, using as the display a television on its side (refreshing left-to-right as the music moved across the screen). We quickly abandoned the idea. Cogan had been using specialist equipment at IBM, and it was clear that at that time nothing less would do. The continuing development of the personal computer in the interim, of course, has transformed the situation, and what was then a wild idea is now a routine exercise for audio engineering students. Cogan’s work was pioneering, but although he is often assumed to have been the first to work on recordings using spectrograms, in fact a previous form of the technique was developed by Metfessel as early as 1926 and was used with precision and sophistication by the Iowa psychology lab. See Milton Metfessel, ‘The vibrato in artistic voices’, in ed. Seashore (1932), 14-117 at 15 & 379. The equipment available to Seashore’s team at the height of its productivity is outlined in Harold Seashore, ‘An objective analysis of artistic singing’, in ed. Seashore (1936), 12-171 at 14-6. Back to context...
Rebecca Plack, ‘The Substance of Style: How singing creates sound in Lieder recordings, 1902-1939’, PhD dissertation, Cornell University, 2008 Back to context...
Nicholas Cook, ‘Performance analysis and Chopin’s mazurkas’, Musicae Scientiae 11 (2007), 183-207 Back to context...
Neta Spiro, Nicolas Gold and John Rink, ‘Plus ça change? In how many different ways can Chopin’s mazurka Op. 24 no. 2 be performed and why?’ CHARM Newsletter 2008 (pdf file), 4-9 Back to context...