Since Theodor W. Adorno’s book
Mahler: Eine Musikalische Physiognomik, many scholars have studied the narrativity of Gustav Mahler’s music. However, insufficient attention has been paid to the analysis of narrativity within the compositional process, particularly in light of the trend in textual studies known as “genetic criticism,” which aims at a hermeneutics derived from the analysis of preparatory materials.  Instead, many literary theorists invoke such an approach in a different way from musicology to explain the interpretive enigmas of modernist literary works, whose narrative fragmentation suggests their aesthetic nature as works in progress.  
This kind of approach seems particularly well suited to Mahler’s music, given that his symphonies can also be seen as “works in progress,” even when completed and published, because of their incessant reworkings, and some authors in the literature on Mahler lead us to believe this. James Zychowicz, for example, notes that Mahler usually made further revisions to the final, even published, versions of his works.
Sometimes, however, Mahler’s feverish reworking can be a specific marker of some programmatic intention, as occurs, according to Stephen Hefling, in the second movement of the Second Symphony.  Considering these aspects, then, the oft-observed modernist internal fragmentation in Mahler’s symphonies suggests their aesthetic nature as works in progress, like the literary works considered in the trend of genetic criticism. A genetic approach thus seems to be one of the most useful ways to understand Mahler’s musical narrativity, particularly in his Tenth Symphony. This is true not because of the work’s incomplete state (Mahler died before finishing it, so it is technically a work in progress, subject to further changes) but because the symphony’s manuscript includes perhaps the greatest number of sketches existing for any of Mahler’s works. In this way, we can reconstruct possible relationships among structure, narrative, and hermeneutics in the compositional process, from the initial sketchy musical ideas to the draft of the last compositional stage. Building on this context, I then hypothesize that this symphony’s compositional process can be compared with that of a narrative, from both a structural and a hermeneutic point of view. 
This article will detail and support that hypothesis.
In section 1, drawing from the literature on musical narrativity, I theorize the concepts of “narrativization” and “narrativity” in music and the narrative-like features through which they appear. In section 2, I explain how these features work in the compositional process of some musical ideas in the first movement of Mahler’s Tenth. Section 3 enlarges this genetic analysis and interprets its data by linking them to a biographical and historical-cultural context of the movement’s compositional process, according to the epistemological perspective presented in section 1, taken from the semiotician Umberto Eco.   Toward Definitions of Musical Narrativization and Musical Narrativity
 Since Adorno’s essay, the discussion on musical narrativity has developed around the key concepts of “voice” and “time,” borrowed from narratology, which occur most frequently in Mahlerian literature.
“Voice” refers to the persona (an external or internal narrator) that is speaking in a given narrative or in its part. Always from a narratological perspective, “time” designates all the temporal relations between the “story,” or what is told in a chronological order of the events of a narrative, and the “discourse,” how it is told in an actual (and not necessarily chronological) unfolding of the events in the text.  
In musicology, the term “voice” has been advanced since Adorno’s book
Mahler: Eine Musikalische Physiognomik, where he was the first to pinpoint an extra-textual level of the act of making music—considered as potentially comparable to a narrative—in a communicative context. “Even where the musical process seems to say ‘I, ’ its correlative, analogous to the latent objective first person of the literary narrative, is divided by the gulf of the aesthetic from the person who wrote the phrase.” After Adorno’s groundbreaking theorization, a vigorous discussion of a possible narrative voice in music has continued upon other conceptual grounds—mainly semiotics and narratology—which are distinct from Adorno’s apparatus. The authors participating in this debate seem to credit the idea that morphological discontinuities and a strong gestural characterization of musical ideas can play a crucial role in their possible imaginative personification by the listener. In such cases, these musical ideas, since they contrast with and stand out from other ones for melodic, harmonic, rhythmic, dynamical, or timbrical aspects, can look like “agents,” dramatic or narrative characters. The adjective “gestural” refers here to the common musicological definition of “gesture” codified by Robert Hatten as “communicative … expressive, energetic shaping through time (including characteristic features of musicality such as beat, rhythm, timing of exchanges, contour, intensity).”  
However, in this process of personification by the listener a further step seems necessary: attributing this “voice” (otherwise impersonal) to someone (for example, a narrator) who performs the action of telling. This passage may doubtless be problematic, given the weakness of referential-denotative meaning in music.
In this trend of musicology, narrative theory is often an essential comparative term, so it is worth considering Gérard Genette’s theory.
Here the voice (or narrator) is defined by its “diegetic narrative level” and by its relationship to the story it is telling. Concerning the narrative level, the voice can be “extradiegetic” or “intradiegetic” if it is external or internal to the story, as happens when the story is told in the third or the first person. Concerning the relationship with the story, the voice can be “heterodiegetic” or “homodiegetic” if it is absent or present in the story itself as its character.  
Within the debate on these issues in the analysis presented in the following two sections, I take ideas from Edward T. Cone’s
The Composer’s Voice and Byron Almén’s  A Theory of Musical Narrative. Cone, in some representative vocal and operatic works identifies musical ideas that can be compared to voices that he connects to a superordinate agency—the “composer’s voice.”  The narrative activity of this “voice” within the piece can be marked, for example, by the orchestral introduction and interludes which comment on the action in operas, or by the piano accompaniment in lieder (he takes as his example Schubert’s  Der Erlkönig). These connective elements then serve as an authorial “set of [expressive] instructions” for the other “voices” of the piece. 
In the field of instrumental music, Cone believes of Berlioz’s
Symphonie fantastique that in the absence of a verbal aid such as a sung text, “the complete musical persona” can be “unitary” (in the case of a solo instrument) or represented by “virtual agents.” Provided that the related musical ideas are highly gestural, these “virtual agents” are groups of instruments that assume roles analogous to those of literary characters. Cone’s search for signs in the musical text of the central role of the composer (“the composer’s voice”), to which the other agents can be traced, relies upon undeniable textual features which refer to a historical-aesthetic context provided by him in his essay. Then Cone considers Berlioz’s  , his notes to this symphony’s first edition, and the composer’s letters: all documents considered by the musicologist for the sake of the composer’s Traité d’instrumentation et orchestration intention and plan. This aspect of Cone’s work is not fully developed, however, requiring a complete reconstruction of the compositional process, which he clearly and rightly calls for.
In continuity with Cone’s theory, Almén identifies musical ideas that can be bearers of “actoriality,” that is, those that acquire “the status of anthropomorphized subjects [and] participate in a narrative trajectory” in the course of the piece.
To demonstrate when these musical ideas can acquire actoriality, Almén quotes Eero Tarasti’s analysis of Chopin’s G-minor  Ballade. Almén points out that “in the actorial analysis … Tarasti tracks two parallel trends, the gradual supplanting of the waltz theme by the quasi-parlando material as the main actor and the gradual emergence of hidden connections between the two themes and recitative-like introduction.”  
 The starting point for reflecting on the “time” dimension in Mahler’s music is always Adorno in his aforementioned monograph on Mahler. In this book, Adorno identifies Mahler’s searching for an alternative temporal order—the novelistic—to that of classical sonata form.
Under Adorno’s influence, this comparison of Mahler’s music with the novel has continued using conceptual grounds—mainly semiotics and narratology—which are distinct from Adorno’s philosophical apparatus. 
Carolyn Abbate was one of the first musicologists to develop this field specifically, especially with regard to Mahler’s music, although she adopts a skeptical position on this issue.
She maintains that music lacks an important narrative feature identified by the philosopher Paul Ricoeur: the capability of “manipulating time, of using tense to achieve a kind of moral distance in recasting the referential object.”  In verbal language, by contrast, this capability is performed through a narrator and a past tense that marks the narrating voice within a denotative-referential mode of signification, but a conjugation is absent in the semiotic system of music.  The evident limit of Abbate’s approach is that she does not consider the possible alternative modes of musical signification that could somewhat compensate for its obvious lack of verb tenses. Thus, her suggestion that music lacks a past tense has triggered a vigorous debate about possible narrative temporality in music. This reflection has led to a search at another semiotic level—that of the connotative—for an explanation, beyond mere referentiality, of the “temporal manipulations” which other musicologists have noticed in Mahler’s and other composers’ music. 
Some authors search for a narrative temporality in self-referentiality (by comparing music to a fiction whose characters can be considered the musical features themselves) and intertextuality (narrative meanings and functions as determined by references to other musical texts through more or less shared codes). Two of these studies, by Robert Samuels
and Vera Micznik,  are especially revealing on the issue of narrative-like “temporal manipulations” in Mahler’s music. 
With this point of view, Samuels, following Adorno, analyzes a passage of Mahler’s Fourth Symphony: the climax at the end of the development. He notes that the border point of this passage (and of the entire developmental section) is a general pause that seems a signal of disquiet, put “not even within the bounds of metre, but simply introduced at the double bar-line.”
The subsequent recapitulation then “begins in the middle of a phrase, with the developmental outcome of first and second subjects.”  Moreover, “the thematic and harmonic scheme have got out of synchronization here … since the first subject begins before the double bar line (disguised by a pause, the voice-leading, and changes of instrumentation and texture).”  As a result of these devices, “formally, and at the level of textual discourse, this moment is the outcome of aporia, presenting the continuity of the formal unit as a fiction.”  In other words, in this passage there is a conflict between codes (those of thematic continuity, motivic development, and formal schemes). These are features of a “quotation-mark music”—Adorno’s definition  —which, I would add, implies narrative-like temporal slippages. 
Following Samuels and Adorno, I suggest that another conflict in this passage, one between all musical ideas which precede this pause milestone, plays a peculiar role in these slippages. On the one hand, there is a ringing sound of sleigh bells in the movement’s introduction, described by Samuels as “an expression of Mahlerian modernism.”
On the other hand, two Haydnesque or Mozartian themes (via Schubert, quoted in the first theme) express a stylized past (termed “archaism” by Samuels).  At the climax, Mahler develops the introduction by distorting it phantasmagorically and dramatically, inserting the trumpet-call motive of the death march subsequently used in his Fifth Symphony. As Samuels notes, “here we have not only music in the past tense, but also future tense too.”  Thus, without the option of a denotative device for a past tense, the composer creates similar temporal manipulations by borrowing past musical styles. 
“Music and Narrative Revisited: Degrees of Narrativity in Beethoven and Mahler,” by Vera Micznik, has made another important contribution in this debate. One of the focuses of her analysis is the notion of the “narrative impulse” that can be triggered in the listener by a piece of music. By responding to Nattiez, who thinks this narrative impulse can be activated only through extra-musical verbal aspects (such as titles and programs), not through the music itself,
Micznik claims that there are also special features of music that make music comparable to narrative. More specifically, if music cannot be narrative with the semantic precision of verbal narrative because of its denotative weakness, it can possess modes of signification that place it somewhere on a range of a greater or lesser degree of narrativity. 
 In other words, according to Micznik, there is not only a listener’s impulse to narrativize music (as Nattiez affirms) but also the composer’s narrative impulse, a strategy whose traces can be narrative-like text features that confer to a musical piece a high degree of narrativity. Specifically, I observe that Mahler’s narrative impulse is ascribable to the breaking of Adorno’s “principle of [motivic-thematic] economy” of classical-romantic music.
As a result of this historical musical turn, Mahler’s symphonies also present, besides this Adornian diseconomy of different musical ideas, that polyphony of characters that literary critic Mikhail Bakhtin identifies in the modern novel.  By drawing from Micznik’s theory, I can identify the following among these narrative-like features of musical pieces having a high degree of narrativity (in comparison with those having a low degree of narrativity):  the emancipation of the little-conventionalized secondary parameters (dynamic level, tempo, texture, register, and so forth), which, in pieces exhibiting a high degree of narrativity, convey meanings more often than the primary parameters (melody, rhythm, harmony), which refer to a more codified (by tonal and formal conventions) self-referential semiotic system;
this happens because secondary parameters refer to general patterns of “processes of accumulation, velocity, dissolution, [and] disorientation,”  so these patterns refer more directly than the primary parameters to an extra-musical world, and thus they are more comparable to the hetero-referentiality of episodes of verbal narrative; so, according to Micznik, in the first movement of Mahler’s Ninth the thematic conflict between the two main themes does not result from their tonal contraposition, as per sonata form conventions; instead, this conflict comes from the gestural contrast determined by secondary parameters and by the articulation of semantic content this contrast can induce in the listeners;   the higher number of “intertextual connotations,” which, according to Micznik,
are musical connotative meanings arising from references to other musical texts of the same or other composers, in terms of quotation or style borrowings; more specifically, these intertextual connotations help make the musical ideas more concrete and referential, although within a primarily connotative semiotic system; so in a piece with a low degree of narrativity—the first movement of Beethoven’s Sixth—Micznik identifies only few “intertextual connotations”; by contrast, in the first movement of Mahler’s Ninth, which has a higher degree of narrativity, she detects many “intertextual connotations”;   the discrepancy between the durations of occurrences during the piece of a given musical idea and an “‘ideal’ temporal … scheme (which could consist of older formal models)”
or the temporal scheme established at the beginning of the piece;  in other words, according to Micznik’s analysis of the first movement of Mahler’s Ninth, the occurrences of its themes do not follow the durational pattern established at the beginning of the movement and the more regular one associated with the older model of classical sonata form.  
Ultimately, these aspects, among others, help to trigger in the listeners a perception of the “discourse” of the piece as a sequence of distinct episodes and tend to attribute to each of them an autonomous meaning. Thus, in Mahler’s symphonies a narrative-like teleology tends to replace the more semantically unitary (and motivic-thematically economical) formal pathway determined by the abstract and self-referential semiotic system, dominated by the primary parameters, of absolute music of Viennese classicism.
The limitation of Micznik’s apparatus, however, is that it does not consider the work-in-progress nature of Mahler’s music. In other words, her theory and analysis work only on a synchronic plane of the final version of the piece, without considering the diachrony of the compositional process. However, Micznik does find significant indications for this perspective. In particular, she reflects on the genesis of a motive whose potential Mahler was likely aware of from the early stage of its conception.
In another passage in her essay, she mentions a “composer’s strategy” which is “highly conventionalized” and “to a large degree predetermined” about tonality in the classical period.  For “a late-Romantic composer like Mahler, it [this “composer’s strategy”] is more likely to present unusual, unexpected discursive ‘narrative’ techniques.”  Micznik seems to be alluding to Mahler’s authorial aptitude, which needs to be theorized, for transforming (by de-structuring) the conventional formal schema across the compositional process. 
To define this aptitude, I use the term “narrativization.” With this word, I indicate a possible composer’s narrative strategy across the composition to narrativize early musical and more traditionally organized musical materials by increasing the degree of narrativity through the addition of the dimensions of “voice” and “time” (the latter including narrative-like aspects 1, 2, and 3) presented above. I then adopt the notion of
narrativity to refer to the possible synchronic traces of the narrativization in each of the versions generated during the compositional process of a given work. My theoretical basis for this concept of narrativization is from historian Hayden White, who investigates this concept by studying text-construction strategies in historical narrative (called “history proper”). In this genre, the historian fills the gaps between historical events presented in a fragmentary manner in annals or chronicles by adding “causal connections,” the voice of the author and other agents, and a temporal plot. 
 Starting from the above discussion, I then place my enterprise in epistemological terms within a perspective of “narrative interpretation” based fully on the consideration of the historical-cultural turn of Adorno’s breaking of the “principle of economy” in Mahler’s music. As with any interpretation, this one is subject to a control horizon that needs to be based on a critical historical perspective suggested by Eco’s theory of interpretation.
That is, in terms of this author, this narrative interpretation works as a hypothesis of an  intentio auctoris (author’s intention), inferable only in a communicational context that, however, is irrecoverable because it is located in the past and happened already in a communicational act. The only way to create this hypothesis is to identify the intentions of the implied author, which the reader can infer from the text ( intentio operis or intention of the work). Then, according to Eco, a validating criterion in this search requires extracting from the text only those specific features that can be linked to a context, as it can be reconstructed through historical criticism, and surrounding the communicational act of a given work. In contrast to everyday conversational communication, however, this linking with the context in artistic communication requires a complex and lengthy creative process, which can be intended as a network of relations, mediated by more or less shared codes, among the author’s social context. In the semiotic theoretical apparatus outlined by Eco, the preparatory materials thus also clearly play a non-secondary role in this process. Narrativization During the Compositional Process
In this section, I will show how the process of narrativization operates during the compositional process through the narrative musical dimensions—“voice” and “time” (the latter including aspects 1, 2, and 3)—presented in the previous section.
I will do this by comparing the narrativity of the last version (an orchestral draft)  of the movement with that of its earlier sketches and drafts, with regard to the musical ideas I call “  a” (see figure 2), “ b” (see figure 3), “ x” (see figure 1), “ y” (see figure 3, red square 2), and “ z” (see figure 3, red square 1). In the light of my following analysis in this and section 3, the initial letters of the alphabet “ a” and “ b” designate the two musical ideas in this labeling system which are comparable to the first and the second themes of sonata form. The other musical ideas of the movement are instead designated with the last letters of the alphabet, “ x,” “ y,” and “z.”  Figure 1: F irst occurrence of x, mm. 1–15  Figure 2: F irst occurrence of a  Figure 2-continued: First occurrences of a and y (in the red square) Figure 3: First occurrences of b, z and y (in red squares 1 and 2) 
In the OD of the movement,
x is an introduction-refrain occurring at mm. 1–15, 39–48, 104–11, 183–93, and 244–52. This idea (see figure 1) is distinct from a and b on account of its more intense chromaticism and tonal ambiguity, its lack of a clear periodic structure, and its secondary parameters of tempo ( Andante), timbre (violas), and register. These features, in contrast to those of a and b (presented below), enable this musical idea to be recognized as semantically distinct from them.
Especially in its first and second occurrences (mm. 16–31 and 49–80) in the OD,
a (see figure 2) is the idea in the movement that more resembles a theme, according to the tonal definition and conventional thematic discursive completeness and fluency of the classical sonata form. If the periodic construction of these two occurrences seems quite regular at first, the first occurrence subtly presents some significant internal breaks in continuity on closer examination. For example, the arrival of y—a gesturally differentiated motive—in mm. 27–31 (see figure 2-continued, in the red square) precludes a continuation of the previously uninterrupted thematic discourse of mm. 16–27. This continuation happens only in the second occurrence of a (mm. 49–80), where the idea reaches more fluent thematic completeness. Here, there is no displacement of y, and the presentation of a comprises two complete periods (mm. 49–57 and 58–80) in an uninterrupted thematic fluency. But again, the durations of the following occurrences (mm. 126–27, 133–34, 141–46, 151–52, 178–83, 213–16, 228–43) of a are irregular and tend to be shorter than the two initial occurrences. From the perspective of the “time” dimension, all these aspects of breaking of continuity (and of the Adornian “principle of economy”) along the occurrences of this theme involve the narrative point 3 of section 1. That is a discrepancy between the durations of occurrences of a and a conventional scheme (which calls for these occurrences to be regular in duration) and a temporal scheme established at the beginning of the piece by the durations of the two lengthy initial occurrences.
 Variability and unpredictability of duration also affect
b, continually opposed by y. Its first presentation, after just two bars, is interrupted suddenly (at m. 34) by two gesturally differentiated motives, z and y (see figure 3, red squares 1 and 2), which “swallow” the possible fluent continuation of the occurrence of this theme. At m. 81, y in the first violins and violas is superimposed on an occurrence of b. But the former does not seem a mere accompaniment; instead, with its intrusiveness, it tends to “hide” b. Then, at m. 87, motive z seems to “help” y to prevail on b at mm. 88–90, where y is the only uncontested idea.
Analogous heterophonic superimpositions of
y happen at the occurrences of b at m. 96 and, in the development, at mm. 116 and 135 (see figure 4). Then the occurrence of b at m. 135 is almost made unrecognizable by the superimposition of y (see figure 4). In the recapitulation, from mm. 153 to 162, b is presented with y, which seems to accompany the former (see figure 5, in the square). Finally, only at the coda in mm. 217–27 does b (see figure 6 and figure 6-continued, in the square) acquire thematic completeness and fluency comparable to that of a in the exposition, in a 10-bar uninterrupted melody. Figure 4: y and a at m. 135  Figure 5: b, mm. 153–62  Figure 6: b, mm. 217–27  Figure 6-continued: b, mm. 217–27 
From the point of view of the “time” dimension, the contrasting features of
x on the one hand and a and b on the other—as if they were two temporal planes of narrative—involve narrative aspect 2 of section 1, “intertextual connotations.” More specifically, in light of Samuels’s view presented in section 1, Mahler seems to borrow the tonal ambiguity and chromaticism of x, rarely found in his music, from composers whose music was often more complex, chromatically speaking, than Mahler’s usual enlarged tonality, well represented by a and b. In this view, Jörg Rothkamm detects in this introduction-refrain a reference to Liszt’s Trauergondel No. 2, and Constantin Floros notes the mood of the prelude from the third act of Wagner’s  Tristan und Isolde. Always by applying Samuels’s ideas, shown in section 1,  x semantically poses a narrative present, signaled by the adoption of more tonally ambiguous and chromatic music to signify the contemporary present, counterposed to what seems a recollection of a musical past in a and b. This recollection seems both nostalgic (the enlarged tonality of a and b) and difficult (the interruptions by y).
From the point of view of the “voice” dimension, Julian Johnson links this musically unusual intro-refrain to Cone’s theory about the “composer’s voice.”
In particular, Johnson interprets the introduction-refrain as a human voice because of its aperiodic, improvisatory character. Moreover, this voice “quickly gets lost and apparently is unable to frame a statement or construct an identity.”  On this basis, then, it seems natural to interpret  x as a “voice” of a “subject” who remembers the past, represented by a and b.
Now that we have looked at these three musical ideas in the last compositional stage, it is time to observe their compositional process. From this perspective, the analysis of the movement’s preparatory materials can confirm, specify, and address further Mahler’s narrativizing compositional strategy, already glimpsed in the OD, across the compositional process. For example, the genesis of the first and second occurrence (mm. 1–14 and 39–48) of
x reveals the “additional” and “intrusive” nature of this musical idea. In fact, the initial version of the initial 78 bars on the earliest pages of ÖNB Mus. Hs. 41000/6  and GvE [missing] tends to refer more strictly to the tonal and formal convention of the sonata form tradition, with its usual presentation of the two themes, a and b, without the two occurrences of x. Only in the following stages is this scheme subjected to a process of narrativization: during the compositional process, the degree of narrativity increases, adding on the later pages in BSB Mus. ms. 22746  and ÖNB Mus. Hs. 41000/6  of the first and second occurrences of x.
Indeed, as above, these additions provide an intertextual connotation semantically distinct from that of
a and b. Thus, always by applying Samuels’s reading key, we can see that by inserting the two occurrences of x Mahler wanted to add a temporal narrative-like plane, absent on the previous pages, to the musical discourse of the traditional sonata. Once x is introduced in the compositional process of the movement, this introduction-refrain changes very little in the following versions of these two occurrences. This lack of major changes between versions also occurs in the genesis of the following occurrences of x (mm. 104–11, 183–93, 244–52, at pages indicated in table 1), conceived at the earliest stage, not as later insertions. Table 1: Compositional chronology of the occurrences of x, mm. 1–15, 39–48, 104–11, 183–93, 244–52
x , mm. 1–15
x , mm. 39–48
x , mm. 104–11
x, mm. 183–93
x , mm. 244–52
BSB Mus. ms. 22746 
BSB Mus. ms. 22746  –
ÖNB Mus. Hs. 41000/6 
ÖNB Mus. Hs. 19646 2364 
ÖNB Mus. Hs. 37817 [1v]
ÖNB Mus. Hs. 41000/6 
ÖNB Mus. Hs. 41000/6 
ÖNB Mus. Hs. 41000/6 
ÖNB Mus. Hs. 41000/6  –PSS110.136BB
ÖNB Mus. Hs. 37817 
BSB Mus. ms. 22744 
BSB Mus. ms. 22744 
BSB Mus. ms. 22744 
BSB Mus. ms. 22744 
BSB Mus. ms. 22744  –
BSB Mus. ms. 22744 
ÖNB Mus. Hs. 41000/1 
ÖNB Mus. Hs. 41000/1 
ÖNB Mus. Hs. 41000/1 
ÖNB Mus. Hs. 41000/1 
ÖNB Mus. Hs. 41000/1 
 From the point of view of the “time” dimension, only genetic analysis reveals that the difficulty of the nostalgic recollection of a past time in the OD of
a and b corresponds to their laborious genesis. For example, the compositional process of a in the exposition includes a complicated swapping of groups of bars among those that in OD will be the first (mm. 16–31), second (mm. 49–57), and third (mm. 58–80) occurrences of this theme. Moreover, this conception of a includes changes in the order of presentation of the first two occurrences. In greater detail, on the preparatory pages BSB Mus. ms. 22744  and BSB Mus. ms. 22744 , the earliest version of the first occurrence (in the OD at mm. 16–31) of a is located after the second occurrence of x (in the OD at mm. 39–48). Then, on page BSB Mus. ms. 22744 , what will in the OD become the second occurrence of a (in the OD at mm. 49–57) is located after the first occurrence of x (in the OD at mm. 1–15). Again, as concerns b, the aforementioned 10-bar melody (mm. 217–27) is conceived as a recovery from a sort of “writing past,” on BSB Mus. ms. 22744  (see figure 7, red square 1), from the earliest sketches ÖNB Mus. Hs. 41000/6  and ÖNB Mus. Hs. 41000/6  (see figure 7, red square 2), after these bars are discarded on their chronologically following (but earlier than BSB Mus. ms. 22744 ) pages ÖNB Mus. Hs. 37817 [1v], ÖNB Mus. Hs. 37817 , ÖNB Mus. Hs. 41000/6 , and ÖNB Mus. Hs. 41000/6 [11v] (see figure 8).  Figure 7: Compositional process of mm. 217–24  Figure 8: Discarding of mm. 217–27, previously conceived in earlier pages  From Narrativity to Meta-Referentiality
The last considerations in the previous section on the compositional process of
a and b make us think that, given the weak denotative resource of music, narrativity in this movement appears not only in the narrative aspects I presented in section 2 but also meta-referentially. In other words, it seems that the composer, through the process of narrativization, wanted to represent within the last version of the movement its act of writing, which appears to be an act of memory. This aspect is evident in the deepening of the genetic analysis of the “voice” and “time” aspects in the movement and in the linking of the outcomes of this examination to a larger biographical and historical-cultural context, according to Eco’s interpretative validating parameters.
The tonal ambiguity and the lack of major refinements during the compositional process and between its occurrences during the movement seem to put
x in the category of Richard Kramer’s “draft-like music,” which uses the artifice to simulate the improvisatory, unrefined writing of the initial stages of the compositional process. In other words, I think it is possible to read this tentative musical idea in light of what Kramer says about Beethoven’s “Tempest” sonata: “the process itself, the act of composing [which] infiltrates the substance of the work.”  As in that sonata, here Mahler’s movement begins with a deliberately draft-like and apparently unrefined introduction that seems “external” to the following  Adagio, since the former is different in tempo and thematic physiognomy from the latter. So, always as in that sonata, later this introduction “enters” (taking the form of a refrain) and becomes integrated into the movement acting as its theme, in turn leading and “composing” the other musical ideas. So also in the OD of the Tenth’s first movement, Kramer’s “subjective figuring of the voice of the composer who composes the work” emerges. 
This role of
x as an authorial “composer’s voice” is strengthened when one considers its other aspects. Steven Coburn’s motivic analysis suggests that this idea contains all four basic cells of the movement, and even of the entire symphony (see figure 9). So the two themes of the movement,  a and b, contain x’s cells 3 and 4 and 1 and 4, respectively, and y has x’s cell 1. Figure 9: Basic cells and motives of the symphony 
Considering these aspects raised by Coburn, and in light of Cone’s reading of the piano accompaniment in Schubert’s
Der Erlk önig, this idea serves occurrence after occurrence as what Cone defines as a “set of instructions”—not, as in that lied, of the expressive character of the story in music but, more deeply, of motivic material of  a and b. In this way, in Almén’s terms, the idea acquires an anthropomorphic actorial status in the narrative trajectory of the movement to represent the voice of the composer who creates other musical ideas from the shapeless motivic material of x.
 From this same perspective of the “voice” dimension, we see
x as an example of what literary theorist Gerard Genette—in his theory, presented in section 1—calls an “extra-diegetic narrator.” In other words, x seems to work as an authorial voice that intrudes in the work and becomes a character-author who, with difficulty, writes the work itself. Specifically, the introduction-refrain during the movement writes the story of a and b. X makes this from its materials, not from nothing. In fact, this author— x—knows the story in advance, as suggested by x’s being the motivic source of a and b. Like the epic storyteller, however, he attempts to remember this story through a fixed formula—the formulaic melodic unfolding of the occurrences of this idea, which lack major changes during their compositional process. So the intrusion of the composer’s voice into his piece is indirect. That is, given the low denotation and referentiality of music, the subject cannot manifest himself by naming himself or using verb tenses but only through the ideographical proxy of representing what he is doing in creating his work as he writes, that is, collecting musical ideas that give shape to the story of the two themes. More specifically, this narrator/author moves his hand to note the uncertain, unfinished melody of x, searching for the right time of the memory of past happenings that he is going to narrate.
a, that time finally arrives, and the story can finally begin, although it is recalled with difficulty. Mahler leaves some enigmatic, isolated traces of this effort in the OD. These traces are Coburn’s cell 4 at mm. 56–57, 59, 60, 61, 62, 63, 67, 75, 77, and 78, whose rhythmic profile suggests the iconic work of this cell as “stuttering memory.” Other such traces are the interruption of fluency of a by y at m. 27 (see figure 2-continued, in the red square) and of b at m. 34 by z and y (see figure 3, red squares 1 and 2). The preparatory materials, however, show more evidence of this difficulty: the “stuttering memory” cells are introduced within a only later in the compositional process (see the red circles in figure 10). Given this later addition, then, these traces seem to be markers of Mahler’s difficulty—seen in the laborious compositional process—in fluently telling this theme by constructing it with x. Figure 10: Addition of Coburn’s cell 4 during the compositional process 
The complicated swapping of bars and occurrences during
a’s compositional process (identified in section 2) represents a difficulty in defining a “before” and an “after” and of musically focusing the right time in the memory. The compositional process of b’s occurrences, compared to that of a’s, appears to be even less fluent, given that its short occurrences are often interrupted by y. As a result, the effort of remembering b is less successful. But an “insight” arrives only at the 10-bar melody, whose “liberating” inspirational nature is revealed by the recovery from the writing past during the compositional process.
Mahler highlights this emancipation of
b through its “theft” of the key of D major at bars 221–24 on BSB Mus. ms. 22744  (via pages ÖNB Mus. Hs. 41000/6  and ÖNB Mus. Hs. 41000/6 ) to x’s occurrence (bars 244–52) on ÖNB Mus. Hs. 37817 [1v] and ÖNB Mus. Hs. 41000/6 , all conceived prior to page BSB Mus. ms. 22744 . In this “theft,” in the long duration and in a non-oppositional y during the occurrence of b at bars 222–25 (see figure 7, red square 3), we can identify meta-referential markers. In other words, these aspects suggest that this recovery is liberating after the asperities of the previous b’s occurrences interrupted by y and after the laborious compositional process of bars 217–27.
These corresponding clues about “voice” and “time” dimensions find surprising—and maybe decisive—support in epistolary documents that tell us something about the compositional process of the symphony in the framework of a dramatic summer of 1910. That’s when Mahler discovered Alma’s extramarital affair with the architect Walter Gropius; a series of epistolary poems written to Alma in August 1910 testifies to his psychical pain in response to this discovery.
One of these epistolary poems, dated by Alma August 17, 1910, suggests that the enigmatic motto-theme I have called
x was probably the musical representation of the Feder (quill) mentioned in that poem: “Die Zeit ist da, die Feder ist zur Hand – / Doch die Gedanken wollen nicht verweilen” (“The time has come, the quill is in my hand— Yet this idea continually eludes me”). According to de La Grange-Weiß, Mahler is saying in these verses that at the moment of beginning to write the work, he “had not yet experienced the moment of intuition akin to the ‘sound of oars playing through the water’ which had inspired the opening of the Seventh Symphony,” in spite of his strong inspirational impulse.  
In another passage of this epistolary poem, we read: “Und was mein Herz auch singt und dringt /
es schweifen all Sinne in die Runde !” (“No matter what my heart desires or sings / My senses rove and wander without aim”). Thus, the uncertain, shapeless, sketchy melody  x seems to be the motoric and ideographic expression of the composer’s wandering without aim of his hand’s movement of writing on paper.
 Ultimately, the stylized uncertainty of this musical idea seems to correspond to Mahler’s difficulty finding inspiration (perhaps writer’s block) at the moment of starting to compose the symphony. So the epistolary poem is not only a heartfelt confession to Alma of his difficulty working at that time but also expresses Mahler’s purpose in representing meta-referentially this problematic writing process in the symphony.
The words of another of these epistolary poems to Alma, dated August 17, 1910, lead to the conclusion of my analysis. In this epistolary document, in a co-mingling of life and art, Mahler speaks directly to his hand, in almost agential terms, as an autonomous, separate entity—“Du süsse Hand, die mich gebunden!” (“Oh sweetest hand that ever held me!”).
Here the bond is that of pervasive and incessant writing which is identifiable with the same life, springing up from his wounds in the final verse: “O Leben—spriesse auf aus meinen Wunden!  ” (“Spring-up, o life-force, from my wounds!”). 
These terms from Mahler’s correspondence suggest that he conceived his creative activity in terms very similar to those that Barthes uses in his essay “The Death of the Author” to describe the literary historical-cultural turn represented by his label “modern scriptor,” whose hand
… cut off from any voice, borne by a pure gesture of inscription (and not of expression), traces a field without origin—or which, at least, has no other origin than language itself, language which ceaselessly calls into question all origins.
What is better described as a “field without origin” than the introduction-refrain: shapeless, endless tonally, and thematically indeterminate?
Barthes’s benchmark for this perspective is the process of writing, and the model for his reflection is Marcel Proust. But the pathway determined by the relationship from
x to a and b during the movement—seeking and (difficult) recovery of memory—brings to mind the work in progress of Proust’s À la Recherche du Temps Perdu. This is evident as we consider the conclusion of this monumental literary work, when the narrator finally finds the inspiration to write the long story that he had recalled during the thousands of previous pages. As in Recherche, in this movement the ongoing action of writing (by recalling) a story assumes the role of the protagonist of the story itself. Then a and b can be considered the gradually regained time in a continuous loss and recovery of the memory of an intermittent, difficult inspiration (or represented as such). This reading is suggested in my analysis by the identification of x as the “author” of the ongoing “story” of a and b, by their difficult thematic fluency in the OD, and—in their genesis—by the composer’s uncertainty regarding a “before” and an “after” for a, as well as by the gradual conquest of fluency from the “writing past” of bars 217–27 for b.
There is a scriptorial unfinishedness in both these works, because the respective literary and musical gestures represent an in-progress writing act, not its definite, finished outcome. Thus, ultimately, the above interpretation suggests giving Mahler the same nickname a critic gave Proust—a name that can also be the label of his narrativizing inspiration in this movement: “the hand that writes.”
 References Theodor W. Adorno,
Mahler: Eine musikalische Physiognomik (Frankfurt: Suhrkamp, 1960); and Mahler: A Musical Physiognomy, trans. Edmund Jephcott (Chicago: University of Chicago Press, 1992). ↑ The phrase “genetic criticism” was coined in France, during the seventies, within the theorization of the
Institut des Textes et Manuscrits Modernes (ITEM), directed by Louis Hay; see Centre d’ histoire et d’analyse des manuscrits modernes, Essais de Critique Génétique (Paris: Flammarion, 1979). For a more recent panorama on this trend of study, see Jed Deppman, Daniel Ferrer, and Michael Groden, eds., Genetic Criticism: Texts and Avant- Textes (Philadelphia: University of Pennsylvania Press, 2004). ↑ Lars Bernaerts and Dirk Van Hulle, “Narrative Across Versions: Narratology Meets Genetic Criticism,”
Poetics Today 34, no. 3 (2013): 281–326. ↑ James L. Zychowicz,
Mahler’s Fourth Symphony, Studies in Musical Genesis and Structure (Oxford: Oxford University Press, 2000), 5. ↑ Stephen E. Hefling, “Mahler’s ‘Todtenfeier’ and the Problem of Program Music,”
19th-Century Music 12, no. 1 (1988): 27–53. ↑ This article originates from my PhD dissertation:
Angelo Pinto, “The Symphony as a Novel: Mahler’s Tenth” (PhD diss., The Open University, 2020). From this thesis also come two articles: Angelo Pinto, “Mahler’s Search for Lost Time: A ‘Genetic’ Perspective on Musical Narrativity,” ; and Gli Spazi della Musica 6, no. 2, (2017): 1–37 “On this Side of the Composing Hut: Narrativity and Compositional Process in the Fifth Movement of Mahler’s Tenth Symphony,” . In them I consider other parts of this symphony than those I focus on here. I wish to thank Frans Bouwman and the publishing house Donemus for their kind permission to use some extracts of his transcription: Gustav Mahler, De Musica 23, no. 1, (2019): 52–125 Symphony No. 10 (Unfinished) : Chronological Transcriptions of the Sketches, Short Scores and Orchestral Drafts, ed. Frans Bouwman, 3 vols. (Den Haag: Donemus, 2017). ↑ Umberto Eco,
I limiti dell’ interpretazione (Milan: Bompiani, 1990), 34, and 103–125; and The Limits of Interpretation (Bloomington: Indiana University Press), 1990. ↑ This section borrows and elaborates text and ideas from my two articles:
Pinto, “Mahler’s Search for Lost Time” and Pinto, “On this Side of the Composing Hut.” ↑ I borrow the terms “story” and “discourse” from Seymour Chatman,
Story and Discourse: Narrative Structure in Fiction and Film (Ithaca: Cornell University Press, 1980). ↑ Adorno,
Mahler: A Musical Physiognomy, 24–25. Original wording: “Aber auch wo der musikalische Verlauf Ich zu sagen scheint, ist sein Bezugspunkt, analog zum latenten objektiven Ich der literarischen Erzählung, durch den Abgrund des Ästhetischen geschieden von der Person, die das Gebilde niederschrieb.” Adorno, Mahler: Eine musikalische Physiognomik, 38. ↑ Robert S. Hatten,
Interpreting Musical Gestures, Topics, and Tropes: Mozart, Beethoven, Schubert, Musical Meaning and Interpretation (Bloomington: Indiana University Press, 2004), 97 and 109. ↑ Gerard Genette,
Narrative Discourse: An Essay in Method, trans. Jane E. Lewin (Ithaca: Cornell University Press, 1980), 212–62. ↑ Ibid.
↑ Edward T. Cone,
The Composer’ s Voice (Berkeley: University of California Press, 1974). ↑ Byron Almén,
A T heory of M usical N arrative, Musical Meaning and Interpretation (Bloomington: Indiana University Press, 2008). ↑ Ibid., 11.
↑ Ibid., 81–114.
M usical N arrative, 229. ↑ Eero Tarasti,
A T heory of M usical Semiotics, Advances in Semiotics (Bloomington: Indiana University Press, 1994). ↑ Almén,
M usical Narrative, 59. ↑ Adorno,
Mahler: A Musical Physiognomy, 63. ↑ Carolyn Abbate,
Unsung Voices: Opera and Musical Narrative in the Nineteenth Century, Princeton Studies in Opera (Princeton, N.J.: Princeton University Press, 1991). ↑ Ibid., 52.
↑ Robert Samuels,
Mahler’ s Sixth Symphony: A Study in Musical Semiotics, Cambridge Studies in Music Theory and Analysis (Cambridge: Cambridge University Press, 1995). ↑ Vera Micznik, “Music and Narrative Revisited: Degrees of Narrativity in Beethoven and Mahler,”
Journal of the Royal Musical Association 126, no. 2 (2001): 193–249. ↑ Samuels,
Mahler’ s Sixth Symphony, 142. ↑ Ibid., 142.
Mahler: A Musical Physiognomy, 96. ↑ Samuels,
Mahler’ s Sixth Symphony, 141. ↑ Ibid.
↑ Ibid., 143.
↑ Jean-Jacques Nattiez, “Can One Speak of Narrativity in Music?”
Journal of the Royal Musical Association 115, no. 2 (1990): 257. ↑ Adorno,
Mahler: A Musical Physiognomy, 87. ↑ Pam Morris, ed.,
The Bakhtin Reader: Selected Writings of Bakhtin, Medvedev and Voloshinov (London: Edward Arnold, 1994), 89. ↑ This aspect can be explained by the fact that the primary parameters are characterized by discrete sign units (notes, duration, chords) which can be assimilable, although in a mainly self-referential system, to the discreteness of language units (words, sentences, and so on, which, however, operate within a denotative hetero-referential semiotic system). Instead, the secondary parameters are not based on discrete sign units, and its semiotic dominion cannot be the one-to-one relation of the denotative system but the polysemic functioning of connotation.
↑ Micznik, “Music and Narrative Revisited,” 226.
↑ Ibid., 227.
↑ Ibid., 211.
↑ Ibid., 206.
↑ Ibid., 235.
↑ Micznik welcomes this aspect from the narratology of Genette,
Narrative Discourse, 33–160. ↑ Micznik, “Music and Narrative Revisited,” 236.
↑ Ibid., 225.
↑ Ibid., 220.
↑ Hayden White, “The Value of Narrativity in the Representation of Reality,”
Critical I nquiry 7, no. 1 (1980): 5–27. ↑ Eco,
I limiti dell’ interpretazione, 34 and 103–125. ↑ From here on, I will reference the manuscript pages with their library signatures, in which the acronym “ÖNB” refers to the Austrian National Library, “BSB” to the Bavarian State Library, and “PSS” to the Paul Sacher Stiftung. The number in square brackets refers to a given page within each source. The siglum “GvE [missing]” refers to a page that belonged to the composer Gottfried von Einem whose original is missing, a copy of which is held in the Gustav Mahler Gesellschaft in Vienna.
↑ The library’s signature for this draft is ÖNB Mus. Hs. 4100. From here on, I designate the last version with the acronym “OD.”
↑ This labeling system is taken from Victor Kofi Agawu, “Tonal Strategy in the First Movement of Mahler’s Tenth Symphony,”
19th-Century Music 9, no. 3 (Spring 1986): 222–33. ↑ Taken, by kind permission of the author and the publisher, from
Gustav Mahler, Symphony No. 10 (Unfinished) : Chronological Transcriptions of the Sketches, Short Scores and Orchestral Drafts , ed. Frans Bouwman (Den Haag: Donemus, 2017), 1:1. ↑ Taken (with modifications), by kind permission, from ibid., 2.
↑ Taken (with modifications), by kind permission, from ibid., 4.
↑ Taken, by kind permission, from ibid., 18.
↑ Taken, by kind permission, from ibid., 21.
↑ Taken, by kind permission, from ibid., 29–30.
↑ Taken, by kind permission, from ibid.
↑ Jörg Rothkamm,
Gustav Mahlers Zehnte Symphonie: Entstehung, Analyse, Rezeption (Frankfurt am Main: Peter Lang, 2003), 102. ↑ Constantin Floros,
Gustav Mahler, vol. 3, Die Symphonien (Wiesbaden: Breitkopf & Härtel, 1985), 102. ↑ Julian Johnson,
Mahler’ s Voices: Expression and Irony in the Songs and Symphonies (New York: Oxford University Press, 2009), 298. ↑ Ibid.
↑ In figure 8, the reader may notice the lack of these bars on these manuscript pages by the written indications (circled in red in figure 8) inserted by the transcriber: “weiter Takt 226” (“further bar 226”) and “weiter Takt 227” (“further bar 227”).
↑ Taken, by kind permission, from Mahler,
Symphony No. 10 (Unfinished), 29. ↑ Taken, by kind permission, from ibid., 28.
↑ Richard Kramer,
Unfinished M usic (New York: Oxford University Press, 2008), 186. ↑ Ibid., 171.
↑ Steven D. Coburn, “Mahler’s Tenth Symphony: Form and Genesis” (PhD diss., New York University, 2002), 87.
↑ Taken, by kind permission of the author, from ibid., 115.
The Composer’ s Voice, 1–19. ↑ Taken, by kind permission, from Mahler,
Symphony No. 10 (Unfinished), 6. ↑ Henry-Louis de La Grange and Günther Weiß,
Gustav Mahler: Letters to His Wife, rev. and transl. Antony Beaumont (Ithaca: Cornell University Press, 1995), 377; my emphasis. ↑ Ibid., 376.
↑ Ibid., 378; my emphasis.
↑ Roland Barthes,
Image, Music, Text, essays selected and transl. Stephen Heath (London: Fontana, 1977), 146. ↑ “
O hne Ausdruck” (“without expression”) is the indication by the composer in a version of Idea x on the manuscript page ÖNB Mus. Hs. 41000/6 . ↑ Maurice Blanchot,
Le l ivre à venir (Paris: Gallimard, 1971), 284; and Maurice Blanchot and Charlotte Mandell, The Book to Come, Meridian: Crossing Aesthetics (Stanford: Stanford University Press, 2003). ↑
Cover picture: Portrait of Gustav and Alma Mahler (1909), photography; by courtesy of Österreichische Nationalbibliothek, Bildarchiv Austria .