The Quest for “Authenticity”

Three performances of a Bach’s fugue compared

Tiago Morais Ribeiro de Sousa

1. Bach and the Lute

Paolo Cherici (2001) observes that the lute, having undergone successive changes, first with the addition of supplementary strings on the low register, allowing the widening of the gamut and sonic intensity, and later with a readjustment of the tuning in several simple strings, achieves a new stability in the middle of the 17th century. Within a hundred years the lute was disappearing from the musical scene, without undergoing any significant changes in that process. Meanwhile, the instrument’s centre of popu­larity shifts gradually from Italy to France and then to Germany. Among the foremost composers for the lute in these two countries we find such names as Jean-Baptiste Besard, Denis Gaultier, Jacques Gallot, Charles Mouton, Esaias Reusner and Silvius Leopold Weiss, who came to make friends with J.S. Bach.

Cherici tells us that though it is difficult to accurately determine the dates of Bach’s works for the lute, it is reasonable to think they belong to the Cöthen period (between 1717 and 1723). A considerable portion of Bach’s output in instrumental music springs from this period, since Leopold, prince of Anhalt-Cöthen, for whom Bach worked, fa­vour­­ed profane over sacred music. It is thought that Bach wrote more pieces for the lute than are extant. In support of this hypothesis one should mention from the outset that Bach has certainly played and taught this instrument, as we seemingly have evidence, on the one hand, in the presence of a lute in his collection of instruments, and, on the other hand, in some testimonies from his pupil, Krebs. Second, bear in mind that a good part of his production, mostly that which was bequeathed to his son Wilhelm Friedemann, was lost or destroyed.

The beleaguered story of the “Fuga / del Signore Bach”

Claude Chauvel (2009) helps us to draw the historical path of Bach’s work under analysis. From the second movement of the first Partita in G minor (BWV 1001), from the set of violin pieces titled Sei Solo / a / Violino senza Basso accompagnato / Libro Primo / da Joh. Seb. Bach. Ao 1720, this fugue is, undoubtedly, an arrangement by a lutenist contemporary of Bach, Herr Weyrauch, it is thought to have been written around 1730, based on a version for the lute by Bach himself. There is yet a version of this fugue for the keyboard, catalogued as BWV 539, but nowadays serious doubts are cast upon its provenance. The problem at hand is that of knowing to what extent Weyrauch was faithful to Bach’s musical idea, or whether during this process he felt tempted to add to the work something of his own creativity. We can but critically compare this version with the Sei Solo version, the only one we know to have issued from Bach’s pen. This piece was discovered among the estate of the music critic and organist Carl Ferdinand Becker (1804-1877), with the name Fuga / del Signore Bach, more than one century later.

2.The Question of Historical Authenticity

Before continuing with the study of this fugue, let us make a short speculative digression into an aesthetic and musicological debate directly related to our analysis. By the 1980s, a movement had arisen within the musical and musicological community, advocating a set of suitable modes and principles of performing early musical works – usually works written until the end of the Baroque era. This movement of musicians and music theo­rists claimed that the works written during this period should be performed in ac­cordance with the coeval performance practices of the composer himself and the period in which he composed such works. For what purpose? John Eliot Gardiner, one of the proponents of the “historically informed early-music movement”, explains:

To unlock the codes in the musical language of these Baroque masters, to close the gap between their world and ours, and to release the wellspring of their creative fantasy meant cultivating a radically different sonority (Gardiner (2014): 9).

Bach on the guitar – a widespread practice

Bach's fugue at stake here is included in this type of works. According to this movement the legitimate (or “authentic”) way to perform this fugue is to follow certain historicist guidelines. These guidelines exclude, of course, among a bunch of another modern practices, the use of modern classical guitar – a non-existent instrument at the time of Bach.

One of this work’s purposes is to determine whether, and to what extent, performing Bach on the guitar – a widespread and consecrated practice in the musical world – car­ries implications for the aesthetic or artistic value of what is thus achieved. On the guitar, the problem of “authenticity” sets itself with particular relevance, since Bach did not write any work specifically for this instrument – and yet, he is one of the “composers of reference” in the guitar repertoire. All guitar students are required to work through his lute or cello Suites, or his violin Sonatas and Partitas. The transcription of the Chaconne, from the Partita nº2 in D minor, BWV 1004, for the violin is so well-known in the guitar world that one could almost say that it would not be the least surprising that someone less informed about history came to believe momentarily that the guitar is its “original” instrument. Besides, Bach’s works appear regularly in the recital programs of professional and consecrated guitarists.

The meanings of “authenticity”

Having said that, what are we talking about, precisely, when we talk about “historical authenticity” as this concept applies to musical performances? In what consists the historic­al authenticity of a performance? To better understand this concept, we may contrast it with the notion of “simple authenticity”, with no reference to “historical” authenticity. Now, a performance is authentic in this sense when it is a genuine (well-formed or bona fide) occurrence of a work. Whatever it is that makes a sound event count as an occurrence of the Fugue BWV 1000, those are the features that bestow authenticity upon the performance. This presupposes a fundamental difference that only became obvious with the division of labour between the composer and his per­formers, and also with an enhanced notation system: it is the difference or distinction between a work and its instances (tokens, occurrences or performances). It makes no sense, in this way, to state about the work itself that it is or is not authentic. Only performances can be or fail to be authentic. But being authentic in this simpler sense is merely to be an occurrence of the work, that is, an instance that complies with the score. The problem is to determine what conformity to­­ the score amounts to. If by that we mean only relations of tonal intervals, durations, intensities, etc., that is, only the pure sound structure (with no reference to instrumentation, tone-colour properties, etc. in contrast, e.g. with Jerrold Levinson’s “instrumentalist” view of musical work identity), then nothing prevents a performance of a work by Bach in a modern instrument from being a genuine or authentic occurrence of the work. It suffices for it to comply with the structural properties described in the score and to be aesthetically meritorious. In other words, simple authenticity consists merely in the identity conditions for performances (it coincides with them).

Kinds of “historical authenticity”

What exactly is then the nature of such an additional property, which we qualify as “historical” authenticity? What must we add to “simple” authenticity in order to obtain “historical” authenticity? Peter Kivy (2012, 2007, 1995) lists the following hypotheses:

a) Intentional authenticity: the historical authenticity of a performance consists in realizing the composer’s intentions.

b) Authenticity of practice: the historical authenticity of a performance consists in re­specting the performance conventions of the period in which the work was written, for instance, using only instrumental means available to the author at the moment he wrote the score, etc.

c) Sonic authenticity, which Kivy in his turn divides into two kinds: 1) the historical authenticity of a performance consists in the maximum acoustic resemblance to the way the work was performed in the period in which it was composed, and 2) the historical authenticity consists in reproducing in the minds of listeners today the same musical experience (or the same kind of musical experience) that the listeners contemporary to the work’s author would have had upon listening to it (i.e. with sensible resemblance). The idea is to make the experience “for us” as it was “for them”.

The latter point, the second kind of sonic authenticity – sensible resemblance – seems to collide with an epistemological impossibility. Even if we were to travel in time back to the 18th century in order to listen to a performance of a work by Bach directed by himself, we would not be able to experience such an event as listeners of our own century, informed by the history of music subsequent to the 18th century, thus listening to the music of that period against the background of a wider knowledge of the musical past (thanks to musicological research) and against the colossal background of the musical future (that is, everything which was done from the 18th century onwards), to which listeners in the time of Bach never had access. Kivy illustrates this idea with an example from Mozart:

They heard with eighteenth century sensibilities and ourselves with twenty-first-century ones. They and we would hear the same sounds differently (Kivy (2012): 43).

What should we then say about the composer’s intentions – the intentional au­thenticity? Well, there is a clear sense in which Bach’s works were not meant to be performed with an instrument such as the guitar – this instrument was not available in his time. This was not a choice he could make or that was open to him. Now, one of the intentions we might suppose composers have is that their works be performed as good as possible. Suppose we have reasons to believe that a performance in modern instruments is aesthetically better (alternatively: that it sounds better) than a perform­ance in period instruments, using period techniques. (We need not delve here into the specifics of what “aesthetically better” amounts to, as long as we have a working understanding of it as something which bears certain valuable properties in virtue of which,ceteris paribus, we prefer to experience rather than its alternatives.) Why not use the modern instrument? How shall we prioritize the author’s intentions? Which intention has the prerogative? The intention of using an instrument of a certain type or the intention that performances of his work will sound as good as possible? Such con­siderations serve as well to question the “authenticity of the practices”. Why should we follow the practices in use when the composer was active, rather than the practices we have been developing or the techniques and instrumentation that are now available to us, so as to attain the best possible performance?

About author’s intentions, Kivy (2012: 38) also reminds us that these cannot be reduced to what he could “hear” in his mind when creating the work, nor to what he could “project” about his future performances. It is natural for the creator himself to consider that his work can be performed in diverse ways, instead of being replicated continually in the same way. Having said that, innovative and ingenious performances of the work could never be entirely envisaged in his mind. It is only to be expected, as a matter of fact, that the composer, would be positively surprised with some of these performances.

Suppose historical authenticity is nothing over and above the sheer acoustic re­semblance to what would be a performance contemporary of the author – the first kind of sonic authenticity – with the instruments, technique, tone-colour properties, etc., available at the time, regardless of how different our own experience of such performances is, by contrast with the musical experiences of 18th century listeners.

We now have a characterization of historical authenticity that does not seem riddled with epistemological or phenomenological difficulties. However, why exactly would this be an aesthetically desirable goal? It is in fact odd to want to replicate the physical qualities of sound as it was produced at the time the works were composed, even bearing in mind the impossibility of such a sound being experienced the same way it would have been then.

The “gap” between the score and the performances

Ever since the development of a musical notation and an explicit concept of “musical work”, as contrasted with the concept of “instance” or “performance” of a work, we know there is a gap between the properties prescribed by the score and the totality of properties realized in a performance or performances. Scores do not exhaustively determine performance properties. For instance, decisions regarding ornaments, phras­ing, articulation, as well as the precise meaning of injunctions such as “adagio” or even the way to perform a figured bass, have always been the prerogative of the performer. Furthermore, it is expected from the performer that it be so. A computer rendition of the Fugue BWV 1000 does not satisfy us (in principle) precisely for the absence of (the kind of) properties added by the performer, which do not affect the work’s identity, while giving musical experience its genuine interest. The performer is supposed to contribute with these aspects. So the “gap” is not a defect or imperfection inherent to the notational system, but a key component of value in the western classical musical tradition.

As such, why would it be desirable to eliminate this “gap” between work and per­formance? It is as if the defender of historical authenticity wished to erase the contri­bution one has always expected from performers, in favour not of producing something new, new ways of engaging the learned listener, acquainted with different perform­ances of the same work, but of an archaeological reconstitution, mechanical at worse, of the features of a tradition that is no longer the living musical tradition. We may even argue that such endeavour deprives performance of its authenticity: musicians in the time of Bach did not seek to reconstruct a past tradition but to partake in the production or continuation of a living musical culture, straightforwardly applying their sensitivity and musical taste, rather than try to recreate a taste which was not theirs.

Historically informed vshistorically authentic performance

Kivy (2007: 96-97) expresses his scepticism about the supposed distinction some allege to hold between the concept of “historically authentic performance” and the concept of “historically informed performance”. This change of designation suggests a critical de­velopment of the original historicist project. To be historically “informed” would sup­posedly consist in a more free and flexible approach that would contrast with the dogmatism and normative rigidity that “historical authenticity” would demand. Now, Kivy starts by observing that, strictly speaking, it is obviously absurd to say that a per­formance is historically informed, because only performers can be so. In that sense, the famous conductor Daniel Barenboim, playing the Aria of Bach’s Goldberg Variations on a modern piano with a romantic rubato – because he decided to do so – would consist, auto­matically and definitively, in a performance “historically informed”. But, if we wish to avoid such disconcerting conclusions while referring to the qualities of the per­for­mance itself, we should try to understand what connection is then being said to hold between historical information the performer took care to gather and the character of his performance. If all we are saying is that the performer has the freedom to use that information according to his discernment and aesthetic sensitivity, then we lack an explanation of why one should privilege that historical information rather than some other kind of knowledge. Everything – a poem, painting, a journey, a recital by a renown performer, certain personal experiences – may be an inspiring element. If, on the con­trary, one insists that the historical character of the information is, in itself, a value, criticism about authenticity will do equally well to criticize this (supposedly) renewed movement.

3. Three case studies of consecrated performances

How can this speculative talk about “performance authenticity” work in practice? I attempt to introduce empirical content in the treatment of this issue through a detailed comparison of three distinguished and prestigious performances of this fugue: two on the classical guitar and one on the lute. Through this study it will be possible to observe, with a greater degree of concreteness, the aesthetic and ontological implications of the options taken by each performer.

Three distinguished performers in recent history were chosen for this purpose – Andrés Segovia (1893-1987), Juliam Bream (n. 1933) and Hopkinson Smith (n. 1946). Segovia is an historical reference in the world of the classical guitar, not just for his work as a performer but for his role in bringing this instrument to a wider audience. Segovia enjoyed such recognition that several renown composers created works especially for him, as was the case with Turina or Villa-Lobos. Bream was also recognized as a divulger and performer of the classical guitar. Despite belonging to an older generation of guitar players, his performances are still a landmark for the present generation. Finally, we have a recognized performer of the lute: Hopkinson Smith.

While Segovia and Bream draw on transcriptions, Smith offers us a performance directly based on the score for the lute. We shall try to understand in what measure a more direct connection to the historical sources adds value to the performance. In this case, it is matter of knowing whether Smith’s performance stands out positively from the per­formances on the guitar. However, the question thus phrased raises some problems from the outset, since Segovia takes as a basis of his performance the score for Sei Solo (and not the later transcription for the lute), the only one we know to have been hand­written by Bach himself. So that, if on the one hand we have Smith playing an in­strument which is more “authentic” for that piece, on the other hand we have Segovia per­forming from a score whose credentials of authenticity are more solid. These quest­ions will be approached later in this article. Let us see each of the disputed cases, in sequence.

Segovia’s performance

The performance by Andrés Segovia, of 1928 (Segovia: (2007)), in its first moments, especially on the subject’s third entry, exemplifies something that contemporary guitar­ists in general try to avoid, reserving this technique as an occasional expressive resource: the use of a somewhat “cracked” sound, with a strong deployment of the finger­nail in the attack, in a marked staccato. Segovia’s articulation, for that reason, is not a very delicate one. Yet, despite Segovia’s use of this technique to foreground the subject of the fugue, he seeks throughout his performance to explore, for the subject itself, other kinds of timbre, in different regions of the guitar, with diverse kinds of attack. Concerning agogics, Segovia chooses a generic tempo that seems consensual among performers, with the quarter note around 80 bpm – perhaps Segovia’s tempo is slightly faster. The Andalusian guitarist is the most metronomic of this triad of per­form­ers, not lending his renditions any noteworthy rhythmic variations. We can, of course, observe some degree of agogic freedom, but Segovia chooses, in general, a constant, steady rhythm. His individual choices in dynamic are not prominent, producing some­thing quite homogeneous throughout the whole work. Ornamentation is virtually non-existent, and the only example we are able to listen is that explicitly indicated by Bach on the score. Segovia is the only one of the three using as a main reference the score of Sei Solo, though he does not waive an idiomatic transposition to A minor. Hence we may infer that Segovia was concerned with authenticity, in his unwillingness to use a second-hand text. We could even ask whether it would not be more adequate to define Segovia’s performance not as a performance of work BWV 1000 for the lute – as is specified in the recording – but as a performance of work BWV 1001 for the violin.

Bream’s performance

Julian Bream, in this recording of 1965 (Bream (1993)), begins by distinguishing himself from Segovia, through a timbre which is much closer to what is nowadays regarded as desirable among his peers: a smooth and more full-bodied tone. Notice that Bream presents the subject of the fugue on string number three, instead of string number two or string number one, open – the tone-colour nature of string number three is able to lend a velvetier tone to the melody played on it. Throughout the performance, Bream explores other modes of attack and other regions of the guitar, and of the three per­formers it is he who exhibits the more tone-colour variety. The tempo used in this per­form­ance is approximately the same as Segovia’s and Smith’s, but, together with Smith, the agogic profile is quite wider. As to the dynamics, Bream is the one who seeks to enrich the work with this expressive resource: we hear clear-cut crescendos and dimi­nuendos, distinctively feel the presence of fortes and pianos. For all of this, of the three performers, Bream is seemingly the freest in what concerns taking the work beyond the strictly textual. Finally, this performer, unlike Segovia, chooses the version for the lute, and, as with Segovia, transposes it to the key of A minor, much more natural in the guitar.

Smith’s performance

The performance by Hopkinson Smith, recorded in 1981 (Smith (2009)), most clearly distinguishes itself from the others: the use of the original period instrument is, in fact, something that transports us to another sonic universe. Smith’s timbre is characterized by this dark resonance that seems to echo the darkness of another epoch. The agogics in Smith seems to be the freest – notice how the subject is presented in the first few seconds: a clear accelerando in the first repeated notes drives the theme towards its main melody, thus reconciling it with its just tempo. The dynamics, though present, are restrained. Smith does not impose any dynamic beyond that which emerges from the natural concord with the inherent expressive movement of the music. He allows the music to speak for itself, relinquishing the claim of speaking for it. Again, free ornamentation is virtually non-existent: in this parameter, Smith seems to merely follow what Bach has written down. The chosen key is G minor (the original key). However, Smith, possibly in line with a certain historicist practice, chooses a 415 Hz diapason – supposedly the standard frequency for organs of that period. Today, that frequency entails a tuning almost half step below the usual, and in terms of actual perception the performance is heard as if it was in F sharp minor.

Comparison Synthesis

The following chart schematically summarizes these ideas:




Recording date









Violin, Sei Solo

Tab. Lute

Tab. Lue


A minor

A minor

G minor, 415 Hz






Metronomic, textual

Freer, generous use of rubato

Freer, generous use of rubato



With more variation

Diverse though restrained


Aggressive, staccato

Delicate, more fluid

Delicate, more fluid


Cracked timbre with heavy use of the nail

Full bodied, velvety

Dark, resounding


Strictly as specified in the score

Strictly as specified in the score

Strictly as specified in the score

Chart 1 - Synthesis of characteristics of the three analysed performances

We can see that there are differences between the performances in all aspects, except for the use of free ornamentation. In this regard, all decided to stick to what is textually manifest in the score. It is interesting to note that although Bream uses the same instrument as Segovia, he takes performative choices that seem generally closer to those of the lutenist: the agogics and dynamic freedom, as well as the fluid articulation. Regarding timbre, the organological differences between the two instruments inhibit a parallel approach. Segovia distances himself from their peers in virtually all relevant options: metronomic agogics, uniform dynamics, marked articulation. Even at the tone-colour level, Segovia widened the already significant chasm separating the guitar from the lute, through his use of a sparkling nail attack.

Let us proceed now to a careful discussion of our case studies, examining the aesthe­tic and ontological implications of these options in more depth.

4. A More Detailed Discussion

The value of Smith’s performance

Recalling the above arguments against the historicist project, we would see no con­vincing reasons to attribute a positive value to performances guided by concerns of historical accuracy to one or other aspect linked to the work’s origin in the composer’s intentions and his historical, social and artistic milieu – such as the use of instruments from the period of the work’s creation. However, there is in fact an inclination or propensity to think that Smith’s performance is really the most interesting and engaging. How to explain this? As I said, there are no reasons that, in general and without at­tending the specific case, should incline us to value such historical aspects. We will always have to look at specific cases, considering relevant factors that may enrich the performance. I think the factors involved in the fecundity of Smith’s performance are fundamentally two: 1) the tone-colour profile of the instrument and 2) the expressive options taken by the performer. In fact, the timbre of Smith’s lute endows the music with a sonic character whose richness and profundity are not to be found in the timbre of the guitars in Segovia and Bream. Smith manages to somehow involve us, listeners, in a kind of acoustic obscurity (if such a metaphor is intelligible) which is perfectly adequate to the spiritual depth of Bach’s fugue. The somewhat sparkling character of the two guitarists’ performances gives too much light to a work that, at least in part, takes its nourishment from the darkness – especially in what concerns Segovia’s rather “cracked” attack.

What was just said does not lend support to the general thesis that period instru­ments, for their tone-colour qualities, are in general more adequate to the performance of ancient music. What happens here is that, in this specific case, the timbre of the lute fits better with the peculiar spirit of that specific work by Bach. To see this, it suffices to think that a performance of a Bach’s fugue with a similarly obscure melodic outline would probably gain by being performed on a modern day piano (with all the richness and harmonic diversity that it offers) than on a harpsichord.

As to the second factor – Smith’s performative choices – one can also observe that the lutenist, opting for a very free kind of agogics, was capable of offering us a much more engaging musical experience. The special interest that Smith was able to give his performance is to be adequately conceived more in terms of his freedom and expressive talent than in terms of conforming to a set of baroque conventions. I think Smith’s performance itself is, in a considerable measure, best understood as an artistic creation in its own right and, as such, not adequately conceived by reference to a set of pre-established stylistic rules[1].

Segovia’s option for the Sei Solo version

We should yet take note of the interesting riddle that was adumbrated: given that Segovia’s performance is based on the only score we know to have been handwritten by Bach, couldn’t that bestow it a special status concerning its authenticity? After all, if it is true that the lute transcription made by Weyrauch on the supposedly existent (though never found) lute transcription by Bach was, in some way, modified by the free creativity of the first, we could ask whether what is more essential to the work is its sonic con­figuration (acoustic profile) or the instrument for which it was envisaged. One might argue that if Weyrauch had altered the piece, then Smith’s performance would inherit a kind of adulteration of a significantly more serious nature than what occurs with Segovia’s performance: the altering of the very notes in Bach’s score. The Andalusian gui­tarist sought at least to secure that basic structure. In this sense, we could say that the latter’s performance is the more authentic of the three, for if there are doubts concerning the essential character of the instrumentation, there can be little doubt as to the essential character of the basic sound structure represented in the score.

We have seen that Segovia’s performance is, for the aforementioned reasons, the least interesting of the set. One asks: if decisive historical evidence were forthcoming that Smith’s score was in fact modified, regardless of Bach’s intentions, would that change our judgement of it? Would we rethink the position ascribed to Segovia? This hypothetical situation presents us with the intricate question of the extent to which beliefs seemingly so “external” to musical experience (such as whether the piece was entirely composed by A or B) may affect our fruition and judgement of its object. I think that even granting that such might, to a certain extent, happen, the musical experience generated by Smith’s performance is so intense and piercing that a variation in our beliefs regarding the origin of that version would not suffice to substantially undermine the quality of the experience. In fact, we might come to the conclusion that Weyrauch’s transcription was so skilfully developed that it managed, despite the modifications he decided to introduce, to preserve the original grandeur of Bach’s piece. Anyway, com­paring both versions, the differences between them do not seem, as a matter of fact, all that remarkable. Even if one would argue that, in principle, changing the notes is aes­thetically more harmful than changing the instrument, we must always take into account the degree of such modification. If the shift in timbre implied by the change of instrument is considerable and the changes in the sound structure merely residual, it is not at all obvious which of the resulting performances is the more authentic.

In sum, the categories used in the analysis of all three performances did not decisively withstand considerations about the supposed authorial intentions (which comes close to a true exercise in divination), nor considerations about a greater or lesser conformity to the instrumental means of performance for the period of the work’s inception. The tone-colour nature of the instruments and the performative options explained in terms of expressive freedom, of properly artistic creation and the capacity of each performer to emphasize different aspects of one and the same work sufficed to distinguish the aesthetic value of each.

The modern classical guitar question

Recall that the question with which we started was whether it is appropriate to perform Bach on the guitar. One might say that such is still too vague or ambiguous a way of putting the question, the sense of “appropriate” being rather unclear there, or it being the case that more than one question can be asked using the same words. There are countless possible situations in which we might deem “inappropriate” to play Bach on the guitar, from contexts where purely aesthetic considerations matter to ones where moral reasons are invoked. Now, the debate about historical authenticity in the philosophy of music is focused mainly on two points: an ontological one and an aesthetic one, as well as on the relation between these, including the question whether these are actually two distinct points or just one. Ontologically: is it or is it not the case that a sound event S, intentionally produced in conformity with the score for work W, but ignoring the prescription of instrumental means indicated in that score or in some other record of the composer’s intentions is a bona fide instance or token of work W? In other words, if we play a piece by Bach on the guitar are we still playing Bach? Aes­thetically: is the aesthetic or artistic value of a “historically authentic” performance (one that in the least does not violate the prescriptions of instrumental means given by the composer) superior (ceteris paribus) to that of a performance that ignores or violates the restrictions on historical authenticity? Why?

Moving to a more straightforward answer to that first question thus clarified and applied to our example of the three Bach performances (trying as well to use those three practical examples as an empirical guide to our reflection): despite our having con­sidered that Smith’s performance, precisely in part due to the tone-colour profile of his lute, is the more aesthetically interesting of the set, it does not follow that guitar per­formances of the work are inappropriate or devoid of artistic value. That is, even if it is true that the performance on the lute is aesthetically better, for that piece, than the performance on the guitar, that does not follow from general considerations about historical authenticity but rather from peculiarities of the aural or musical experience of that particular piece.

We have seen that the criteria put forward by defenders of historical authenticity – conformity to performative conventions of the historical period in which the work to be performed was created – do not properly explain the adequacy and value of a musical performance.

If we accept conformity to the score as a minimal criterion of adequacy and this is, in its turn, conceived as adherence to the basic sound structure represented there, then Segovia’s performance will not raise any doubts, since, as was said, this performer was careful enough to follow the score which is known with absolute certainty to have been authored by Bach. The only alteration made to the sound structure consists in a trans­position from the key of G minor to the key of A minor. An idiomatic adjustment that does not relevantly affect that structure (otherwise we would not be able even to transpose a single melody to different keys without ceasing to have that same melody, for its identity would not survive the change).

About Bream’s performance we can say that it followed the score for the lute, which is believed to be a transcription of a version Bach made for that instrument. Now, with­out historical warrant that Bach’s original transcription ever existed, we could, plausibly, doubt the authenticity of the score for the lute. Throwing such suspicion over the score also compromises, a fortiori, the authenticity of Bream’s performance. However, Smith, as it happens, also followed this score. The same suspicion would hold for the lutenist’s case as well as for the guitarist. And yet, when we examine Segovia’s option for the Sei Solo version, we conclude that Smith’s choice of the lute version could not result in anything particularly harmful. So that nothing special in this regard might be attributed to Bream’s performance. Consequently, if we accept conformity to the score as a minimal criterion of adequacy, the adequacy of guitar performances is secured.

The acoustic and technical nature of the guitar

But is this minimal criterion of adequacy a satisfactory one for whoever intends to inquire into the validity and aesthetic merit of these performances? We have seen that to add historicist considerations, following the project of historical authenticity, does not give us anything of special relevance. I think that we should rather inquire to what extent the option for the guitar compromises a musically satisfactory experience of Bach’s work.

I believe two aspects concerning the nature of the guitar should be taken into account:

1) The gamut and technical idiom of the instrument (in what measure the option for the guitar demands the sacrifice of notes in the sound structure, either in virtue of the instrument’s gamut, or of the technical difficulties involved in the use of that gamut).

2) The tone-colour profile of the instrument (to what extent the guitar’s timbre differs considerably from the timbre of the original instrument).

As to the instrument’s gamut, Segovia’s version does not pose any problem, since the guitar encompasses the whole extension of the piece written for violin and Segovia shows the possibility of technical realization of that structure. Bream’s version, being a transcription from an instrument of wider gamut, at least in the low register (the lute) to an instrument with a more restricted gamut in that register, would in principle demand the elimination or transposition of some notes and the structural reorganization of some chords. And about this latter aspect one must say that although the arrangement of the notes is not something that can be ignored, if the main notes secure their function in the harmonic sequence, the effect is not substantively lost. Concerning the problem of note suppression, there are low notes possible in the lute but surpassing the gamut of the guitar. In general, one opts for transposing them one octave up.

Generally, the guitar, even in its lowest tuning with the last string in D, does not include the notes C, B and A we see in the normal tuning of the baroque lute (cf. Cherici (2001): III). These notes normally perform the function of harmonic backup. Despite they’re giving a different weight, density and substance to the music, as long as their harmonic function is quite explicit, the procedure of raising an octave or even suppressing may not relevantly compromise the work’s appreciation.

A closer look to the score

It is however relevant to note that in this specific case there was no need of suppressing any of these notes, since the lowest note in the original version is a D and, given that the guitar version is transposed a whole step up, to A minor, the lowest note in this latter instrument is E, the last open string. Even so, there might have been this need of raising an octave or suppressing, not for reasons related to the instrument’s gamut but for technical reasons: it might have been just too difficult to play all notes of the chord as it appears in the lute version. But not even that is the case. Hence in this aspect there was no sacrifice of the basic sound structure in Bream’s version. Note that Bach’s own violin version does not include some of those notes. That shows that Bach himself considered that whatever is essential to the identity of the piece (its “spirit” we might say), is present in that original “reduced” version (we should perhaps speak of an “expanded” version for the lute). We see that Segovia himself, despite having followed the violin version, has added some of these notes for harmonic backup, just as we see them in the lute version. Hence not even Segovia, who relied on the only entirely trustworthy score, took Bach’s original text quite literally. The question of conformity to the original text carried out by our performers is not, in fact, linear[2].

The beginning of the violin score illustrates some aspects of what was said:

Fig. 1 - Excerpt of the Fugue under examination, in its violin version, BWV 1001b

The circles represent some of the changes. We can see that the third entry of the subject in the violin version takes place earlier than in the lute version, which only oc­curs one measure and a half later. In that one and a half measure of difference it even occurs, in the lute version, a new entry of the subject on the tonic. We can also note the inclusion of harmonic backup notes in the episode after the exposition. This episode con­sists, basically, in a sequence of arpeggios following the circle of fifths. The added notes correspond to the tonic of the chord being played as an arpeggio.

An identical situation occurs in the following sequence:

Fig. 2 - Excerpt of the Fugue, in its violin version BWV 1001b

In this stretch we have a harmonic backup note for each group of 4 semiquavers. Each of those groups contains the 4 notes of a seventh chord (the first group, for in­stance, would correspond to the chord of G dominant seventh). This organization might indicate to us, in fact, that each 4 note group would functionally match one single chord which, in the lute version, would be accompanied by a bass matching the fifth of each chord, that is, each chord would appear in its second inversion. However, it seems to us that a more plausible hypothesis is to organize these 4 note groups into 2 note sub­groups, so that to each subgroup would correspond a distinct chord. In this case, we would again have a chordal progression following the circle of fifths: D minor, G major, C minor, F minor, B flat minor, E flat minor, A flat minor, D minor.

We also note that the lute version is one octave lower relative to the violin version. The guitar version, in A minor, is, then, one whole step higher than the lute version and a minor seventh below the violin version.

As to the second factor, one should take account of the following: if it is true that the first transcription is that of Bach, then the composer himself transcribed this work from the violin to another instrument with very different tone-colour and idiomatic character­istics. Even if the transcription was not made by Bach, it would still be true that it was made by a contemporary of his, who was plainly doing something quite common for the time. This practice of transcribing works for different instruments was frequently done by the composer himself. It suggests us that, in this case as well as in similar ones, instrumentation might not have an essential character. Recall that Bach made transcriptions of this kind for other pieces – e.g., Suites BWV 995 and 1006a for the lute seem to have been rewritten by Bach himself, as well as the Cello Suite BWV 1011 and the Violin Partita BWV 1006. To move from the violin to the guitar even seems some­thing quite less radical than moving from the violin (or the cello) to the lute, or the other way around. Despite all the differences, the general organological attributes of this instrument – its gamut, timbre, articulation, harmonic or polyphonic nature – are in fact much closer to the attributes of the guitar than to those of the violin. Thus, even though we may acknowledge that the question of instrumentation is not an arbitrary one – more extreme cases such as playing Bach on the electric guitar with distortion might, in fact, constitute persuasive counterexamples to the idea that instrumentation is not rele­vant to ascertain the adequacy of the performance – we may consider that the qualities of the guitar fit the framework of what would be expected Bach himself to ac­cept. As was already said, when we speak of the author’s intentions, we cannot limit our­selves to those intentions the author might have had in his musical context and merely with the set of instrumental means available to him. This insight was clearly and most aptly described by Kivy, when he introduced the notion of a “counterfactual in­tention”:

The complication in authenticity of intention is that as the concept of intention is applied in human affairs, allowance is made for what I call ‘‘counterfactual intentions’’. The question of what the Founding Fathers intended in the US Constitution is not only the question of what they intended when they framed it, but what, given those intentions, they would have intended, here and now, given the new circumstances in which we find ourselves, but which they could not possibly have anticipated. And so also with the performing intentions of composers. What Bach’s performing intentions would be today, given the circumstances in which we make music today, are part of what his performing intentions were, when he was alive, even though he could not possibly have anticipated what musical life would be like in the twenty-first century (Kivy (2007): 98).

Perhaps it is more relevant to try and imagine what intentions the composer most likely would have in our own present context. Given the qualities the guitar has acquired today, it is most likely – an insight to be taken with the advisable pinch of salt a speculation of this kind requires – that Bach would have been pleased by the experience of listening to his works for the lute played by Bream or Segovia.

I have argued that the attributes of the guitar are not an impediment to the realization of aesthetically valuable performances of works such as the Fugue BWV 1000, which has served us as an example and practical illustration. Having said that, one should bear in mind that the aesthetic value of a performance of Bach on the guitar will depend not only on the acoustic qualities of this instrument, but also on the way that the transcriber and the performer use those qualities. If, as I have argued, we consider that the guitar has enough qualities to generate an aesthetically valuable performance, to make it possible will depend, to a large extent, on the idiomatic decisions of the transcriber – who must use his knowledge of the sonic and technical possibilities of the guitar, in order to preserve the musical qualities essential to the performed work – and on the performer’s artistic decisions. The performer carries the burden of skilfully applying his talent, aesthetic sensitivity and technical proficiency. To a great extent, authenticity is also a consequence of the degree to which creativity in music is enhanced by the creativity the performer puts into his performance, and here no a priori metaphysical principle will allow us to know beforehand what is good or bad.

5. Final considerations

The genius of Bach finds in the guitar a cluster of conditions appropriate to its artistic expression with all the dignity we owe him. Part of the authenticity we expect from a performance depends on a risk that belongs exclusively to the performer (in this case, the guitarist), who cannot avail him or herself of rigidly pre-established rules, but rather uses his or her talent in such a way as to immerse in the spirit of the work and, from there, in the experiential knowledge of its aesthetic essence, as well as taking visceral decisions about what is really important, to create moments of genuine musical beauty. To deny music lovers the opportunity of listening to the works of this composer through the sound of the classical guitar, that is, to deny this music its singular expression in the countless and ineffable acoustic shades of this instrumental means would be a gross mistake.


Chauvel, C., 2009: Bach: L’oeuvre de luth, Naive, Paris.

Cherici, P., 2001: Opere Complete per Liuto, Suvini Zerboni, Milão.

Gardiner, J.E., 2014: Bach. Music in the Castle of Heaven, Alfred A. Knoff, New York.

Kivy, P., 1995: Authenticities. Philosophical Reflections on Musical Performance, Cornell University Press, Ithaca.

Kivy, P., 2007: On the Historically Informed Performance, in Id., Music, Language, and Cognition. And Other Essays in the Aesthetics of Music, Clarendon Press, Oxford / New York, pp. 91-110.

Kivy, P., 2012: Sounding Off: Eleven Essays in the Philosophy of Music, Oxford University Press, Oxford.

Koonce, F., 2002: Johann Sebastian Bach. The Solo Lute Works, Kjos Music Company,San Diego(2nd ed.).


Bream, J. (1993): Baroque Guitar. CD. Julian Bream Edition. ISBN-13: 009-0266049424.

Segovia, A. (2007): Complete Bach Recordings, 1927-1955. CD. ISBN-13: 067-5754002831.

Smith, H. (2009): Bach: L’oeuvre de luth, Naïve. EAN 822186400036.

[1]Kivy calls this dimension of performance «personal authenticity», contrasting it with the three kinds of historical authenticity mentioned. According to the author, «Musical performers are, I take it, artists – ‘‘performing artists’’ is what they are called. And musical performances, I take it, are works of art. Furthermore, one of the things that artworks are customarily praised for is their originality, one of their demerits is customarily taken to be their derivative nature» (Kivy (2007): 99).

[2] Frank Koonce chooses a transcription that seeks to combine the best of two worlds: «some of the modifications found in the tablature [for the lute] evidently have resulted from technical concessions to the lute and so they must be attributed to the transcriber J.C. Weyrauch. […] For this reason, in the present transcription there are many references to the violin version in questionable places, and notes that seem to have been omitted due to technical considerations relative to the lute were restored in accordance with the violin score» (Koonce (2002): VIII).


Creative Commons License

This work is licensed under a Creative Commons Attribution 4.0 International License (CC-BY- 4.0)

Firenze University Press
Via Cittadella, 7 - 50144 Firenze
Tel. (0039) 055 2757700 Fax (0039) 055 2757712