2.1 The text

1Authorial philology is concerned, as we have seen, with the edition of the original, that is, of the manuscript written directly by the author (an autograph) or written by someone else under the author’s supervision (an idiograph), and of the prints edited by the author. There are many different cases, each of them carrying with it problems that are difficult to summarize in a general overview. To put it simply, the areas in which authorial philology proves most useful can be divided into two fundamental categories: the edition of in fieri texts, which comes under a category more precisely referred to by Cesare Segre as genetic criticism (see Gavazzeni and Martignoni 2009), and the edition of texts in multiple versions, which falls into the so-called criticism of variants.

2.1.1 Edition of in fieri texts

2Let us start with the case of a text attested to by one manuscript, be it an autograph or an idiograph. The manuscript can be clean (as in the case of a fair copy), or it can bear traces of a reworking process. The edition of a clean manuscript is similar in many ways to the edition of a single-witness text in the field of traditional philology (which studies variants introduced through transmission). Of course, an autograph is to be treated differently than an idiograph. An autograph directly exemplifies the author’s writing and phonetic habits. In an idiograph, in contrast, the copyist’s mediation can introduce alterations and/or standardizations, resulting in forms which are alien to the author’s habits. There is a particular problem with regard to errors, which can also be found in autographs. Errors especially occur when authors copy down their own work, that is, when a text is transcribed by an author from a previous document. There is a variety of positions on how errors should be dealt with, ranging from absolute compliance to the text of the autograph to the correction of everything that can be considered no more than a mistake. Whatever the decision, editors have to point out clearly their interventions.

3When there are variants on the single manuscript preserving a text, the philologist will have to establish critically the text and decide which kind of apparatus will better represent the variants found on the manuscript. The question ‘Which version should I choose as copy-text?’ can be answered in two ways.

1. The philologist can decide to transcribe the base-version, that is, the very first version of the text in chronological order. Any further variants will be collected, from the first to the last, in the apparatus (which will be called evolutionary).
2. Alternatively, the philologist can take as a reference the last version as it can be reconstructed from the manuscript. The apparatus (which will be called genetic) will collect the corrections through which the author came to the final version of the text, from the last to the first. As we will see, it is also possible to represent the corrections in a progressive way, that is, from the first to the last, the last being the one chosen as copy-text.

4In the case of unfinished texts, it is necessary to make very difficult choices, especially when the text has not come to a final revision.

2.1.2 Editions of texts in multiple versions

5Of course, the situation is different when a work is preserved in more than one authorial version (manuscript and/or in print). In such a case, philologists must ask themselves whether or not the preserved versions are comparable. If they are, the whole elaboration process can be represented in the apparatus with respect to the writing phase which has been chosen as copy-text. If not, when the versions widely diverge (as in the case of Alessandro Manzoni’s Fermo e Lucia — Seconda minuta — Ventisettana: see section 3.4), or when it is preferable to look at them separately for study purposes (as with Giacomo Leopardi’s Canti in the Gavazzeni edition (Leopardi 2009a): see section 3.5), the philologist can edit each and every compositional phase, presenting the internal variants for each, and create a separate apparatus, establishing a connection between the last stage of each phase and the following stage.

6All such situations can be more or less complex. Authors may have come back at different times on their manuscripts, reused parts of a previous version, or worked simultaneously on different versions (something like this can happen nowadays when an author works on photocopies or on the print-out of a file). What is more, the original manuscript may have been lost, but the version it contained may have been preserved in the ‘indirect’ tradition, that is, in non-authorial ways, or ones not directly depending on the author. In this last case, authorial philology and philology of the copy should work together in order to distinguish any error introduced in the transmission from the actual authorial variants. It is necessary to identify accurately the case under study and make well-motivated choices, taking into account not only the socio-cultural background, but also the appropriateness and effectiveness of the representation (e.g., how easily text and apparatus can be read). As we are starting to see, then, it can be difficult to know which text should be chosen as copy-text, and this choice will always have far-reaching implications, both on the ecdotic and the literary/ cultural levels.

The author’s last will?

7Let us begin with a concrete example. Let us say that a text was republished by an author several times and in different formats. If a publisher appointed us to edit a modern edition of this text, which edition should we prefer? The first, the second, the third, or the last one? The choice is not easy to take. Until the last century, it was standard practice to publish the text which reflected the so-called ‘author’s last will’. This would mean, in our case, that we should opt for the last edition published by the author. However, objections to the concept of the ‘author’s last will’ have been raised in recent years, gradually breaching the communis opinio which elevated this notion almost to the status of a dogma. In order to tackle the various aspects of this problem, it is necessary to distinguish two levels:

• first, we have to take into consideration any element in favour of or against the adoption of the ‘author’s last will’ when establishing which is the reference-text;
• then, when the reference-text has been decided upon, we have to think about the criteria to follow to respect the ‘author’s last will’ for each single reading.

8Our belief is that the two phases regarding the ‘author’s last will’ — that is, the overall textual setting of the text for the publication of the work and the editorial intervention in the case of each chosen reading — should be treated separately (on this, see Italia 2005). The first level concerns the idea of themselves and of their own work which authors may have expressed throughout their life by means of an editorial plan, be it carried out or only envisaged in their mind, as well as the form in which authors may have delivered that idea to readers. The textual choice is indeed a very hard one for the editor precisely because it has a bearing on the new image of an author and of their work which is necessarily established by a new edition. Bruno Bentivogli and Paola Vecchi Galli have commented on the role of the editor in twentieth-century editions: ‘It falls on the philologist to determine the most appropriate editorial strategy for the text and to promote with the publication its most authoritative source: the source may not be identified with the definitive or last version of a book, but perhaps with the most “groundbreaking” and “innovative” one for readers’ (Bentivogli and Vecchi Galli 2002: 163).

9The second level concerns the editorial procedure that we need to apply to the individual readings, and involves linguistic, graphic, and typographic questions. Despite being related to the language and style of every individual author, these matters can be addressed, at least to some extent, in a general way. Twentieth-century critical editions feature a wide range of interventions. Although there are significant differences between one edition and the other, it is possible to identify a few constants, which one day will hopefully lead to the adoption of a universal and standard set of regulation within the academic community. We will now present the problem as it is posed in Stussi (2006), who was the first to examine the elements in favour of and against both choices. The elements in favour of the ‘author’s last will’ (i.e., republishing the last version of the text published when the author was alive) can be traced back to three main motivations:

• the authorial motivation calls for the respect for the personal choices made by the author, a ‘feeling so commonly widespread that it has easily come to dominate the publishing field too’ (Stussi 2006: 191): from this viewpoint, the ‘last will’ seems to provide the readers with a work perceived as ‘more authentic’;
• the historical motivation maintains that the diachronic perspective given by the adoption of the last version of a work allows us to better understand the history of the text and consider it as a historical process: from this viewpoint, the ‘last will’ seems to provide the readers with a ‘more useful’ work in terms of interpreting authors and their work;
• on the grounds that the passage from the first to the last version can be seen as a process through which the work moves towards a more evolved stage, the critical-evolutionary motivation implicitly looks at the last edition more favourably than it does previous ones: from this viewpoint, the ‘last will’ seems to provide the readers with a ‘better’ work.

10However, all the motivations in favour of the ‘author’s last will’ can be overturned:

• authorial motivation: the last will does not always reflect the true intention of the author. Stussi has noted, for instance, that there might be ‘restrictions to the expression of that will in connection with the hereditary succession’, or ‘evident mental disturbances, constraints etc.’ (2006: 191). The textual ‘primacy’ of the first edition exclusively lies in the value attached to it by the author, who links to that edition their own idea of themselves and of their own work;
• historical motivation: the historical perspective can be better appreciated if we consider the process in a diachronic way from the first to the last edition. Only the comparison between the editions can offer historically verifiable information on the language and style of the work, which are otherwise flattened out and oversimplified by a final synchronic image. The first edition also allows us to appreciate the ‘critical reception’ of the work and to acknowledge its ‘tradition’, that is, how the work has had an impact on the literary system;
• critical-evolutionary motivation: the idea that the work evolves from one edition to another towards a better form is a false myth; the last editions are not always the best, and indeed the inherent value of a work can be better appreciated in the first edition, which ‘normally represents the conclusion of the original creative process of a work and is therefore the result of the most intense creative period for the writer’ (Stussi 2006: 192).

11It is easy to see that, depending on the circumstances, these arguments can be applied both in favour of and against the adoption of the ‘author’s last will’ as a criterion to establish which text we should publish. This can lead to what has been called a form of ‘philological Pyrrhonism’, if not to the ‘agnosticism of the self-appointed New Philology’ (Stussi 1994: 292), which can contribute to spreading a sense of annoyance at critical texts and apparatuses, often considered as ‘accessory’ elements of the text. In what follows, we will try to prove that that opinion is wrong. As Paolo Cherchi has noted about the debate sparked off by the New Philology, there are two matters at stake.

12First of all, there is the dialectic between ‘text’ and ‘work’. As Cherchi puts it, ‘The authority of philology has ended up creating so much confusion between “text” and “work” that we feel uncomfortable when we read a work whose text features uncertain readings, although for centuries we have been doing nothing but reading “works”’ (Cherchi 2001: 145). The second issue is the relationship between philology and Italian Studies. The prospect that (Romance) philology might develop in innovative ways into a form of cultural history, and that Italian Studies might grow into a form of comparative literature, has been regarded as an antidote to the crisis of philology. This prospect, however, is not borne out by the facts. On the contrary, such a notion has sometimes contributed to a process of trivialization of the discipline, which can be seen even in very prestigious editions. What is more, this issue has not fostered a general debate on the methods and aims of the philological discipline, especially beyond the field of Italian studies. As a matter of fact, there is no agreement within the scientific community on the terminology to be used and on how corrections and authorial variants should be represented. This has of course resulted in an anti-economic proliferation of signs and abbreviations. Each new critical edition forces the readers to learn a new system of representation (with new symbols, initials and abbreviations), thus complicating the debate even within the same community.

13It is undeniable that reflecting on these issues, as well as on the reasonableness of any ecdotic choice, has far-reaching consequences for the reception of the text and therefore for its interpretation. Let us take as an example the case of the twentieth-century poet Giuseppe Ungaretti.

14Ungaretti’s poems offer an example where an author’s last will only appears gradually, but has ultimately taken form in two editions that contain all the author’s works. These two editions represent or have represented until now the unalterable standard, the ne varietur of Ungaretti’s textual tradition. For Ungaretti, the Mondadori edition of 1942–1945, whose publication he sought and oversaw, concluded the long and tormented variantistic process of the two collections L’Allegria (we can now read this in a critical edition by Cristina Maggi Romano 1982) and Sentimento del tempo (in a critical edition by Rosanna Angelica and Cristina Maggi Romano 1988). The 1942–1945 Mondadori edition features three volumes: the first and the second contain the already mentioned L’Allegria and Sentimento del tempo; the third one is dedicated to the Poesie disperse and was published, at the behest of the author, with a dossier collecting the variants in print as commented on by Giuseppe De Robertis. This represents an exceptional case in which textual tradition and critics are connected by the author himself, who openly directs the entire operation.

15The definitive edition of Vita d’un uomo. Tutte le poesie — edited in 1969 by Leone Piccioni while the author was still alive — gathers the two main collections in their 1942–1945 version (there were very few later authorial changes), together with the last final editions of Il Dolore, La Terra Promessa, Un Grido e Paesaggi, Il Taccuino del vecchio, Dialogo, Nuove, Dernier Jours, and the Poesie disperse, that is, the texts published between 1915–1927 and not included in the definitive editions of Allegria and Sentimento del tempo; another seven texts are grouped together in a separate section of the edition entitled Altre poesie ritrovate. The note to the text of L’Allegria — which was written by Ungaretti himself — significantly declares: ‘As a leopard cannot change its spots, the author, who had defined the abovementioned editions as definitive, could not help introducing at each new time a few small changes of form’ (Ungaretti 1969: 528).

16The choice made by the editors of the two critical editions of 1982 (L’Allegria) and of 1988 (Sentimento del tempo) differs from the one made by Ungaretti in Vita d’un uomo. Cristina Maggi Romano and Rosanna Angelica did not choose the 1942–1945 Mondadori edition (reproduced in the three ‘Meridiani’) as base-text. They selected instead the second 1919 Vallecchi edition for L’Allegria, since this is more representative of the literary pathway of the work than the 1916 princeps, and they chose the initial version of each poem for Sentimento del tempo. In his 1990 ‘N.d.D.’ (‘ [Nota del direttore]’ published in Studi di Filologia Italiana), Domenico De Robertis has explained that the development of Sentimento del tempo essentially took place before the 1933 princeps (after which there would only be textual additions), so that the history of the book can be better understood ‘through the thorough examination of the evolution of the single texts, until its 1933 definition’ (De Robertis 1990: 306). The concept of ‘author’s last will’ has been upset by the reasons put forward in favour of these choices and the objections that even very recently have been made.

17The wide range of proposals which the critical editions have provoked exemplifies how delicate the choice made by the editor is. Let us just look at L’Allegria. One option would be to choose as copy-text the text of V, that is, the 1919 Vallecchi edition: this is the choice taken by Maggi Romano. The alternative option would be to return to U, that is, Il porto sepolto published in Udine in 1916. The adoption of U is recommended by Carlo Ossola (Ungaretti 1990), who suggests making use of the evolutionary variants within the commentary to the text, thus, as Claudio Giunti remarks, ‘putting the critical interpretation before the philological esprit de système’ (Giunta 1997: 174). In a similar vein, Umberto Sereni and Carlo Ossola (1990) called for a critical edition taking into account the transmission in print only: this last solution was adopted in the 1945 Mondadori edition, which nevertheless ‘cannot be called’, as Claudio Vela reminds us, ‘a critical edition’ (Bembo 2001: 1276). A further option would be to stick to the ‘last will’ expressed in M (Mondadori 1942–1945), as proposed by Claudio Giunta (Giunta 1997: 175) on the basis of the ‘historical prestige of the witness’, ‘related, on the one hand, to the exceptional “form” of the 1942 Mondadori print (M), and on the other to the repercussions that that form had on the subsequent work of Ungaretti’; this solution had already been adopted when Ungaretti was alive by the editors of the definitive edition of Vita d’un uomo. Tutte le poesie of 1969, but Giunta (1997: 183–84) proposes to also give ‘the first version published in volume’ of each witness, thus determining a ‘multiplication of the textual items’ of each witness. The existence of such opposite choices is a measure of the liveliness and importance of what is a still-open debate.

18Before we start any critical-interpretative study, it will be necessary, as is now evident, to ask ourselves the following question: ‘What text do we read when we read a text?’.

2.2 The apparatus

2.2.1 Genetic and evolutionary apparatus

19If, as we have seen above, the apparatus is the concrete application of the hypothesis represented by the text, the kind of apparatus to be used in a critical edition will be determined by the choice we will have made about the text. That choice will especially depend on whether or not we stick to the author’s last will; and on how we decide to represent the drafting process, either in a genetic or in an evolutionary way. According to a punctual definition by Dante Isella (2009a: 100), an apparatus can be genetic or evolutionary: what difference is there between one and the other?

1. A genetic apparatus is a graphical way to represent the corrections that have formed over time on a manuscript, or on a print with manuscript corrections, or on a typescript with manuscript corrections in the case of twentieth-century texts. The genetic apparatus is a synthetic and standardized system to represent the genesis of a text, from its first version to the one thought to be its last complete form, that is, the one picked as copy-text. A genetic apparatus should not be considered as a photograph of the text: it is rather a hypothesis made by the scholar on the ways and chronological phases of the writing process.
2. The evolutionary apparatus collects the variants that are subsequent to the stage which we have decided to pick as copy-text: that is, variants which do not belong to a phase in the creation what is yet to become a text, but which rather belong to the evolution of what is already considered a text. Of course, the evolutionary apparatus is not an accurate reproduction of the status of a manuscript: it is an interpretation given by the editor of how the text evolved, from the phase which has been picked as copy-text to the last version which can be reconstructed from the manuscript.

20The fact that an edition is provided with a genetic or an evolutionary apparatus depends exclusively on what the critical editor has decided to choose as copy-text (see Table 1 below). In short, if we decide to pick as copy-text the last version of a text, the apparatus collecting the corrections will be genetic. If, on the contrary, we choose as copy-text the first version of a text, the apparatus will be evolutionary. If we decide to choose an intermediary version as copy-text (e.g., the base-reading of the clean copy of a text immediately before further corrections were made on it), the apparatus will be both genetic and evolutionary. It will be genetic with regard to the corrections which have led to what has been selected as copy-text; and it will be evolutionary as concerns the corrections following the phase represented by the copy-text.

Table 1 Text and apparatus

21When we select as copy-text a version which is not the last one that we can reconstruct from a manuscript provided with further corrections, there are two possible scenarios: the variants persist or do not persist in the complete text. The second scenario (i.e., the variants do not persist) is offered by the eighteenth-century writer Giuseppe Parini’s Il mattino, a work that was analyzed and edited by Isella. Parini’s first and second versions of Il mattino follow a very different compositional logic. Between one and the other, there is an intermediate attempt to correct the first version. This attempt does not follow the logic that will subsequently characterize the second version; it rather belongs to a transitional, experimental phase, one soon abandoned by Parini. In this case, the editor has no choice: it is necessary to distinguish the different writing phases and to avoid any confusion. Thus, the variants concerning the intermediate and provisional phase must be collected in an evolutionary apparatus attached to the first version.

22In general, however, corrections usually lead to some kind of a result, which is at least provisionally stable, and this can be achieved within the same witness (as with the case of an overall revision of the same writing phase), or on a different witness (when the corrections make it necessary for the author to rewrite the text). In this latter case, the choice between a genetic or an evolutionary apparatus is an open one, depending on the editorial criteria. Let us take as an example the manuscript of the most famous poem by Giacomo Leopardi, L’infinito (in the version of the so-called Naples notebook, housed in the Biblioteca Nazionale Vittorio Emanuele III of Naples, C.L.xiii.22). If we look carefully at the text, we see that a few lines have been corrected with a pen which looks different from the pen with which Leopardi wrote the base-version of the text (see Fig. 1). Should we want to publish the text of the manuscript, we would have two options:

1. we can publish the text which would better represent the ‘author’s last will’: in this case, we should select as copy-text the transcription of the last reading which can be reconstructed on the manuscript;
2. we can publish the text in its first draft version: in this case, we should select as copy-text the transcription of the first complete reading which can be reconstructed on the manuscript.

23Let us see the two solutions, and the consequences they have in terms of apparatus.

1. Text corresponding to the ‘author’s last will’

Idillio

L’Infinito

1	Sempre caro mi fu quest’ermo colle,
2	E questa siepe, che da tanta parte
3	De l’ultimo orizzonte il guardo esclude.
4	Ma sedendo e mirando, interminato
5	Spazio di là da quella, e sovrumani
6	Silenzi, e profondissima quiete
7	Io nel pensier mi fingo, ove per poco
8	Il cor non si spaura. E come il vento
9	Odo stormir tra queste piante, io quello
10	Infinito silenzio a questa voce
11	Vo comparando: e mi sovvien l’eterno,
12	E le morte stagioni, e la presente
13	E viva, e ’l suon di lei. Così tra questa
14	Infinità s’annega il pensier mio:
15	E ’l naufragar m’è dolce in questo mare.

24The parts of text concerned with variants are here given in bold (both in the case of immediate and late variants, about which see section 2.3.1). As we have chosen as copy-text the last reading of the text, the apparatus will necessarily be genetic, and it will try to represent the corrections occurring from the first version to the last one.

2. Text corresponding to the base-version (first writing of the text)

Idillio

L’Infinito

1	Sempre caro mi fu quest’ermo colle,
2	E questa siepe, che da tanta parte
3	Del celeste confine il guardo esclude.
4	Ma sedendo e mirando, un infinito
5	Spazio di là da quella, e sovrumani
6	Silenzi, e profondissima quiete
7	Io nel pensier mi fingo, ove per poco
8	Il cor non si spaura. E come il vento
9	Odo stormir fra queste piante, io quello
10	Infinito silenzio a questa voce
11	Vo comparando. E mi sovvien l’eterno,
12	E le morte stagioni, e la presente
13	E viva, e ’l suon di lei. Così fra questa
14	Immensitade il mio pensier s’annega,
15	E ’l naufragar m’è dolce in questo mare.

25As we have chosen as copy-text the first complete reading, the apparatus will necessarily be evolutionary, and it will try to represent the corrections concerning the base-version as far as the last reading which can be reconstructed on the manuscript.

26Of course, the same set of issue applies in the case of different printed versions, or in the case of versions transmitted both by manuscripts and prints. When the versions can be compared with one another, so that we do not have to provide an edition for each version, we have once again two options: either we select as copy-text the last version and represent the preceding writing process in a genetic apparatus, or else we pick as copy-text the first version and collect the subsequent corrections in an evolutionary apparatus. We will later focus on how it is possible to represent variants and corrections in a synthetic way by the means of symbols and/or abbreviations referring to general categories of corrections, without having to provide extended explanations. We will see, in other words, how an editor concretely sets up an apparatus. For now, let us see the general criteria that regulate how variants are represented.

2.2.2 Vertical and horizontal apparatus

27Variants can be represented in two ways: in a vertical apparatus (also called ‘column representation’, Stussi 2006: 187), or in a horizontal apparatus (also called ‘linear representation’, ibid.: 189).

Vertical or column representation

28All the corrections from the first to the last that concern a line are put in columns. Deletions leading to the writing of a new reading and insertions of new readings are identified by using typographic markers such as italics or bold. The reference-text can be the last or the first as reconstructed from the manuscript. Sometimes it is identified with typographic markers such italics or bold; sometimes it is reproduced in full, either in the same page of the text put in columns, or at the beginning of the edition, where it can be read in full either as a starting or a finishing point. In the column representation, temporal succession is set out in vertical form. Lines or textual segments where variants are not found are not repeated, so that is easy to see where and how corrections occur. Clearly, this kind of apparatus can only be used for poetry because the line does not normally exceed a typographic line, so that the variants can be put in columns below it.

29This kind of apparatus was adopted for the first time for Giacomo Leopardi’s Canti in the 1927 edition by Francesco Moroncini. Since then, it has been used in several important editions of poetic texts. Its great advantage is that readers do not have to refer constantly to the apparatus in order to follow the genesis of the text, because text and apparatus are not divided, and they can visually reconstruct the writing process. The disadvantage is that readers cannot read the text in its entirety and free from the corrections, unless one form of the text is reproduced separately, either before the actual critical edition, as Moroncini did with Leopardi’s Canti, or in the upper part of the page, as Emilio Peruzzi did in a subsequent critical edition of Canti (see section 3.5). Less obvious but just as significant is another drawback with this kind of apparatus, namely, that since corrections are grouped together according to the lines, this makes it difficult to see the connections between corrections that relate to bigger syntactic units and thus exceed the length of a verse. This often occurs in poetry, where syntactic and metric units do not necessarily coincide.

30Let us now look at another example from the Naples notebook of Leopardi’s Idilli: La vita solitaria (see Fig. 2).

31Let us focus on lines 64–68 reproduced below. As per the column representation, corrections are given line-by-line. As can easily be seen, this does not help us to understand the syntactic connection underlying the real correction, which occurs between verses 66–67 (see ‘ritorna’ and ‘riede’, here marked in bold). At a first glance, the column representation can be misleading: ‘pensando’ (‘thinking’) seems to be corrected to ‘ritorna’ (‘returns’), while ‘riede’ (‘returns’) seems to be corrected to ‘tosto’ (‘quickly’). In fact, it is ‘ritorna’ that takes the place of ‘riede’ (in other words, ‘ahi ma pensando / Che di lui non si cura anima viva, / Riede al ferreo sopor’ is corrected to ‘ahi ma ritorna / tosto al ferreo sopor’):

64	Odo sonar ne le romite stanze
65	L’arguto canto; a palpitar si move

66	Questo mio cor di sasso: ahi ma pensando
		ritorna

67	Che di lui non si cura anima viva,

Riede al ferreo sopor, chè la più bella
Tosto	ch’è fatto estrano

68	Parte di questa vita il ciel negommi
	Ogni moto soave al petto mio.

Horizontal or linear apparatus

32The horizontal apparatus is based on a clear distinction between text and apparatus, which are separated from one another and graphically distinguished. The text is located in the upper part of the page, while the apparatus is usually located immediately below it. Variants are collected in the apparatus one after the other. Whereas in the column representation temporal succession is set out in vertical form, in the linear representation temporality is represented by the means of horizontality. The part of text concerned with a variant is repeated in the apparatus and is followed by a square bracket. The variant is located immediately after the square bracket. In a horizontal apparatus we can also find numbers and/or letters which refer to various phases of elaboration; abbreviations which indicate the position of the variants; and diacritic signs or different fonts which distinguish the variants from a chronological viewpoint. The end of a series of corrections is marked with the letter T (Text), meaning that the series finishes with what we find in the copy-text (which in turn corresponds, of course, to what we find before the square bracket). We will focus on the meaning of the most commonly used abbreviations and diacritic signs in section 2.5.

33The disadvantage of the horizontal apparatus is that, in order for readers to follow and appreciate the corrections, it is necessary for them to refer constantly to the text placed above. This becomes more problematic when the apparatus is located, due to editorial reasons, at the end of the book — and not, as would be preferable, immediately below the text. If the apparatus is overly complex, it can even be printed apart, in a separate volume. This was the case for the critical edition of Fermo e Lucia, that is, Manzoni’s first draft of the novel I promessi sposi (see section 3.4). Most scholars are more familiar with a representation in which text and apparatus are separated. As well as this familiarity, another benefit of the horizontal apparatus is that it can be used for both poetry and prose. In poetry, the verse number is usually given immediately before the part of text in which the variants appear. In prose, the topographic reference is given with the number of the carta (recto/verso) or page (side of the page); or sometimes with the number of the paragraph when the text is divided into paragraphs made by the editor (this will then need to be explained in the Note to the text).

34Let us see once again the lines 64–68 of Vita solitaria, now represented with a horizontal apparatus:

64	Odo sonar ne le romite stanze
65	L’arguto canto; a palpitar si move
66	Questo mio cor di sasso: ahi ma ritorna
67	Tosto al ferreo sopor, ch’è fatto estrano
68	Ogni moto soave al petto mio.

*

66–68

ritorna… mio.] AN1pensando | Che di lui (see varia lectio) non si cura anima viva, | Riede al ferreo sopor, chè la più bella | Parte di questa vita il ciel negommi. from which T (with pen C)

35Further on, we will see in detail the meaning of numbers, letters, abbreviations and different fonts used in the apparatus.

2.3 Variants

2.3.1 Immediate and late variants

36Let us now return to the manuscript of L’infinito (Fig. 1). If we look carefully at the text, we can easily notice that the corrections are not all of the same kind, since they have been made with different pens, and probably at different times.

• Let us focus for a moment on the correction concerning line 9, where the proposition ‘fra’ is corrected by Leopardi to ‘tra’. It seems that the correction was made with a different pen from both the pen used for the base-text and the pen used for the other corrections: the colour of the ink is more reddish and the stroke of the pen is thinner.
• The text concerned with a variant in line 7 remains the same in both the final versions. Here, the correction was made at the time of the first draft of the text, and it is likely to have been caused by the anticipation of the pronoun ‘mi’ when writing the verse: ‘Io mi’ is corrected to ‘nel pensier mi fingo’.
• Finally, let us examine line 13, where the punctuation mark of the apostrophe mistakenly referred to ‘e’ is corrected with the same pen used for the base-text: ‘e’ l’ is corrected to ‘e ‘l’.

37These examples show that variants cannot be regarded as an undifferentiated unified group. They should be rather understood as series, or layers, of corrections that are chronologically separate from one another. The term ‘layer’ and the geological image of the ‘stratification’ are extremely useful metaphors in order to better understand a manuscript as it appears to the eyes of a scholar: a document that will usually contain two kinds of variants, both immediate and late ones. Immediate variants are made at the time of the writing of the text, and can be recognized as such because they are normally located in the writing line. In the case of an immediate variant, the author has generally deleted a part of text which has just been written and has replaced it with something else; then, the author has kept writing on the same writing line. Coming back to L’infinito, the correction concerning line 7 (‘Io mi’ → ‘nel pensier mi fingo’) was undoubtedly made at the time of the writing of the verse. Had it not been made at that moment, the space between ‘Io’ and ‘mi’ would be difficult to explain.

38Of course, a deletion on the writing line does not necessarily indicate an immediate variant. If a deletion concerns a part of text not necessary for the meaning, then it could have been made at a later time, too. In such a case, the text ‘works’ (that is, ‘it makes sense’), regardless of the deleted part. On the contrary, in the presence of an immediate variant, the text usually makes no sense if we read it with the deleted part. Line 7 of L’infinito, for instance, should be read as follows: ‘Io mi nel pensier mi fingo’. Such a reading would be problematic not only in terms of meaning, but also from a metrical viewpoint, since it would imply that Leopardi wrote a hypermetrical verse, and obviously the metrical aspect is to be carefully considered when the editor works on a manuscript of a poem.

39In order to distinguish between immediate and late variants, it is very important to pay close attention to the way the page is set out. A part of text located in the external margin in place of a part of text deleted in the writing line is very likely to have been added later in time than the base-text. See, for instance, Figure 3.

40A peculiar kind of immediate variant is represented by the implicated variant. This category includes all the corrections implicated in the meaning of what follows and that are above-written, below-written, or aside-written (that is, written beside the base-text, either on the right or on the left of it). The term implication refers to the connection between textual elements: it can be syntactic, as in the case of gender or number agreement, morphological, as in the case of verbal agreement, onomastic, toponymical and so on.

41Let us consider the manuscript of Manzoni’s Fermo e Lucia (Figure 4). Although it is written above, the correction ‘una’ → ‘un’ must be immediate (i.e., it took place at the time of the writing of the base-text), since it is grammatically implicated with the following masculine noun ‘galantuomo’. On the contrary, the correction ‘che’ — which is located in the interline — cannot be considered immediate because the text retains its meaning even without the insertion: ‘che cosa vuol dire parlare’ → ‘che sa che (inserted) cosa vuol dire parlare?’.

42The corrections made after the first draft are called late variants. Strictly speaking, all the corrections not located in the writing line and not implicated with the following text should be regarded as late variants. Of course, a variant can be unmistakably acknowledged as late only in a few cases. One such case is when an author has used two different pens: one for the first writing of the text; the other for the following corrections. Another case is when a text has undergone systematic corrections which have something in common, such as an onomastic or toponymical change, or else when it is possible to identify within the text different graphic or lexical habits belonging to the author. It is therefore clear why it is fundamental for editors to have great familiarity with the language and style of the text on which they are working, and why philology, history of literature, and history of the language are always interdependent in the edition of a text. An author can sometimes come back to the text even shortly after writing a line, making an above-written, below-written, or aside-written correction. In such a case, if seen from a topographical standpoint, the variant could be considered late, while it is in fact immediate when understood in terms of chronology since it takes place at the same time as the writing of the base-text. How is it possible, then, to identify — amongst the many above-written and below-written variants that are not implicated — which variants are truly late, i.e., which variants were truly made in a subsequent moment of revision? How can we identify which variants belong to a layer of corrections later than the first writing?

43In the absence of graphic markers, such as different pens or pencils with different colors, which could indicate different writing phases, it is necessary to take into account a number of factors: the ductus (that is, the stroke of the pen on the paper), the handwriting, the syntactic and lexical connections, the style of the author, as well as the author’s habits in terms of corrections. Of course, none of these factors can give us certainty about whether a correction happened at the same time as the first drafting of the text. However, all of these elements can be taken into examination and contribute — especially if they are all in agreement with one another — to argue in favour of or against a hypothesis about the chronology of a correction. This is the case for the late variants that can be found in the so-called Prima minuta, that is, the manuscript of Fermo e Lucia, Manzoni’s first draft of I promessi sposi (see section 3.4). Some of the variants are to be traced back to a late revision of the text of the Prima minuta. Others are to be traced back to an initial revision of the text of the so-called Seconda minuta, that is, the manuscript of the novel’s revised version entitled Gli sposi promessi. For we now know, in the case of many chapters of the first tome and of a few papers of the fourth one, the Seconda minuta/Gli sposi promessi has been written on the same, thickly corrected papers of the Prima minuta/Fermo e Lucia.

44Although authorial philology — like philology in general — is not an exact science, it works towards the interpretation of the given information with a precise scientific method, arguing for the most plausible hypothesis to explain a problem. When new elements emerge which cannot be explained by a given hypothesis, its validity is suspended.

2.3.2 Horizontal apparatus: Explicit or symbolic

45Since corrections over time are represented by the position of variants vertically in the column, the vertical apparatus does not need abbreviations or symbols. On the contrary, in the horizontal apparatus, the use of markers and symbols keeps the editor from providing verbal and analytical explanations for each and every variant. An effective apparatus must be rational and synthetic, making use of a coherent and consistent system of representation of the same graphic phenomena with appropriate markers. The markers can be provided in an abbreviated form in a symbolic way. Hence, the distinction between:

• explicit apparatus: the apparatus is called explicit when it makes use of abbreviations in order to represent the same graphic phenomena;
• symbolic apparatus: the apparatus is called symbolic when it makes use of symbols for the same purpose.

46In order to represent one of the most common cases in manuscript texts — the correction from one variant to another, Italian editions of authorial philology often make use of the generic abbreviation: corr. in (= corretto in, ‘corrected to’), or of a directional arrow such as →. Different arrows may represent different kinds of variants: a simple arrow such as → may represent, for instance, an immediate variant, while a two-colour arrow such as ➢ may represent a late variant. If a variant is located in the interline, the explicit apparatus can make use of the abbreviation: ins. (= inserito, ‘inserted’), which indicates that the text in the manuscript is inserted in the upper interline (in the rare case of a variant inserted in the lower interline, it is possible to further specify: ins. nell’interl. inf. = inserito nell’interlinea inferiore, ‘inserted in the lower interline’). For this same kind of correction, the symbolic apparatus can make use of special markers such as a slash isolating the inserted word: \word/. Mixed-type apparatuses, explicit and symbolic at the same time, are very frequent: in order to represent in a synthetic and coherent way the corrections, the apparatus makes use of both abbreviations and symbols. Abbreviations are always italicized, so it is easy to distinguish the text of the editor from that of the author. Both abbreviations and symbols are usually explained in a Table, which is normally placed in the edition after the Note to the text.

2.3.3 Photographic apparatus and diachronic apparatus

47We are starting to see that some apparatuses try to account for the dynamics of the corrections, while others try to provide a typographic transcription of the status of the manuscript. This is the fundamental difference which distinguishes the authorial philology practiced in Italy from the French critique génétique and the German Editionswissenschaft, all of which tend to represent the variants as they are found in a manuscript, without distinguishing between text and apparatus. In an attempt to respect the topography of the manuscript, above-written variants are reproduced in the upper interline, while the parts of text inserted in the margin of the manuscript are reproduced in the margin too, and so on.

48Following on from the methods of representation used by Francesco Moroncini for Giacomo Leopardi’s variants, techniques of formalization intended to reproduce the diachronic dynamics of the text have become more and more sophisticated. This has been possible especially thanks to Dante Isella and his students. A transition towards an apparatus understood in a diachronic and systemic way has gradually taken place over the last twenty years in Italy in the field of authorial philology, thus placing the Italian school at the forefront in the European philological context. Let us see a few examples.

49At the outset, the intention of the philologist was — even in the first horizontal apparatuses — to represent typographically the complex phenomenology of the text. The idea was to provide a typographic transcription of the text by means of abbreviations and symbols. A particular effort was made attempting to provide the relevant explanations with appropriate symbols or exponents. In the edition of the Chigiano Codex of Torquato Tasso’s Rime supervised by Franco Gavazzeni (Tasso 1993), for instance, the explanations concerning the variants are given with an alphabetic superscript that refers to where each variant is placed, depending on whether it is above-written (a), below-written (b), written on the right (c), or written on the left (d). This was a highly effective means for a better understanding of a manuscript that is difficult to decipher such as the one considered here (see section 3.3).

50The same set of considerations applies for the most representative apparatuses of twentieth-century works, that is, the ones included in the editions of the works of Carlo Emilio Gadda. The fundamental 1983 critical edition of Racconto italiano di ignoto del Novecento by Isella adapted a triple-filter system in order to represent in a rational way all the textual materials which were not part of the copy-text: the apparatus, the marginalia (metatextual notes) and the alternative variants (see section 3.6). In fact, this edition marked for authorial philology the beginning of a new phase, one that was both more scientific and innovative in terms of methodology.

51In the apparatuses produced from the end of the 1990s onwards, scholars have tried to represent — instead of the topographical location of the variants — the stages in the text’s evolution as connected to one another in chronological terms. The focus has not been on the way in which a variant is graphically realized in relation to the base-text (above-written, below-written, inserted, aside-written, and so on), but rather on the chronological relation which a variant has with the base-text and the other variants too. The main difference — and the main difficulty when it comes to setting up the apparatus — lies in the possibility of comparing the different phases with the final text, as well as in identifying and grouping the variants in relation to a ‘system’: ‘in the apparatus […] the portion of text altered by a variant (i.e., the portion of text that comes before the square bracket) can always be directly compared with the variant, or the variants, that affect it, so that it can be studied directly and autonomously, without having recourse to the copy-text’ (‘Introduzione’ to the critical edition of Canti supervised by Franco Gavazzeni, in Leopardi 2009a: xliv).

52The main advantage of this kind of apparatus is the autonomy it offers to readers in terms of following the genesis of the text, with no regard to its photographic representation. This possibility is even more facilitated when there is the opportunity to have high-definition digital reproductions, which allow us to distinguish — on the base of the ductus and of the hand — the different phases of correcting presented by the text. The apparatuses should not be designed to provide a better interpretation of the autograph; they themselves should be an interpretation of the autograph. Consequently, the reading of the autograph should be intended as a possibility to test and verify — in parallel — the work of philological interpretation and critical analysis carried out by the editor. It is undeniable that, in order to set up an apparatus of this kind, it is necessary to invest much more time than was previously allocated in analyzing the manuscript. It is necessary in fact to understand more deeply the ‘mechanisms’ underlying the correcting process as well as the linguistic construction, whether in prose or in poetry. It is one thing to provide a representation of the variants in a topographic way, but it is quite another to understand the variants in relation to a diachronic system and place them within it.

2.3.4 Horizontal apparatus: progressive or derivative

53Another important difference concerns the distinction between progressive and derivative apparatuses. The part of text concerned with a variant, as we have said several times, is repeated in the apparatus and followed by a square bracket. The corrections following the square bracket can be represented in a progressive or derivative way, according to the order followed in the presentation of the chain of variants. In an evolutionary apparatus the corrections follow a progressive chronological order, from the first to the last. The passages from one correction to another can be explained with an arrow or, as happens in Italian editions, with the abbreviation corr. in (= corretto in). See the following example, where corr. in is translated and abbreviated in English as corr. to (= corrected to):

reading picked as copy-text] A corr. to B corr. to C corr. to D.

54In a genetic apparatus, the corrections follow the exact opposite order: they are reproduced in the apparatus from the last to the first, i.e., in a derivative way. In Italian editions the chain of variants starts with the abbreviation da (translated as from in the example below), followed by the chronologically second to last correction, which means that the reading picked as copy-text is derived from the second to last, and that the second to last is derived from the third to last, and so on, until the oldest reading:

reading picked as copy-text] from D from C from B from A.

55In a few cases, corrections can be represented both in a progressive and derivative way. The correcting phases, for instance, are always represented in a progressive way (e.g.:¹ from which² from which³ etc.), while minor corrections encompassed within the same correcting phase are usually represented in a derivative way (e.g.:¹ from which² (above-written to¹) from which³ (aside-written to¹ and²)).

2.4 Marginalia and alternative variants

56The work of Dante Isella also forms an essential precedent for one of the major methodological innovations introduced over the last two decades in the use of apparatuses. In his aforementioned 1983 edition of Gadda’s Racconto italiano di ignoto del Novecento, Isella successfully rationalized the representation of the different textual levels found in the manuscript with a triple-filter system distinguishing between apparatus, marginalia (metatextual notes) and alternative variants. Let us now examine these elements in their fundamental relation to the text, considered as it were as the fulcrum around which — unlike the methods adopted in French philology — the critical edition should pivot.

2.4.1 The apparatus

57The term ‘apparatus’ refers to a part of text which has a relation of topographical and typographical subordination with the copy-text. The apparatus is usually located in the footer and is in a smaller font size. The apparatus may also be placed at the end of the volume. In this case, the relation between apparatus and copy-text may become an extreme subservient one. In a few cases, the apparatus occupies an entire volume and is in the same font size of the copy-text. One example of this is found in Isella’s edition of Fermo e Lucia (Manzoni 2006).

2.4.2 Marginalia (metatextual notes)

58Isella has given the most exhaustive definition for the term marginalia (Italian, ‘postille’) in the abovementioned edition of Racconto italiano di ignoto del Novecento (Gadda 1983: xxxiv–xxxv):

[Marginalia refers to] the remarks provided by Gadda almost everywhere on the page, commenting on what he has already written or what he is about to write afterwards: the list includes statements of disappointment or satisfaction, as well as words of warning or advice directed towards himself; doubts (sometimes expressed with an interrogation mark), and references to different sections of his text; and sometimes also annotations which can be attributed to a later writing stage and are functional to the rewriting of single passages in a clean version, or to the reuse of single passages outside of the context of the Cahier.

59As they are side annotations to the base-text, the marginalia should be ideally imagined in the margin of the page. In an edition, however, typographical and editorial reasons make it necessary to collect them in a separate section at the end of the text (ibid.):

The interested reader is punctually informed at the occurrence of every marginalia by a conventional cross-reference mark located in the margin of the page in place of the marginalia itself (>): something like a graphic stylization of a hand with a pointed index finger which was frequent in former times.

60The placement of the cross-reference mark in relation to the text is nothing more than a mere typographic and graphic arrangement to indicate the presence of the marginalia. This is all the more appropriate if we consider that the marginalia do not have the same status of the text, but should be considered as metatextual notes (that is, as part of the metatext).

2.4.3 The alternative variants

61According to the definition given by Isella, alternative variants — which are not to be mistaken for genetic or evolutionary variants — are ‘competing readings amongst which the author cannot choose, or amongst which he/she has not made it unequivocally clear whether or not he/she has chosen’ (Gadda 1983: xxxv). In editions, alternative variants are located in the footer (below the copy-text) and are tagged with superscript alphabetic letters, whereas a superscript number is generally used for the notes of the author present in the text. Alternative variants have a relation of parity with the text, both in typographic and graphic terms, since they are in the same font size as the text. The idea underlying this presentational approach is that the editor does not know if the author — in a phase of further revision of the manuscript — would have chosen the alternative variant, or the reading that the editor has selected as copy-text. Consequently, from a theoretical point of view, the alternative variants have the same status and value of the text. The location in the footer and the use of the same font size as the copy-text are, in other words, a way of confirming that they are part of the text and not of the apparatus, i.e., that they have the same status and value of the text as they are potentially part of it. Likewise, marginalia are separated from the actual critical apparatus, as we have seen above, as they do not have the same status and value of the materials therein collected.

62The distinction between text, apparatus and metatext is not only very important in general terms, but it also has remarkable consequences for the editing of single-witness texts which, although they cannot be properly said to be ‘critical’, are nonetheless presented as ‘scholarly’, resembling in every aspect a critical edition with regard to what is found in the copy-text, despite not having an apparatus. These editions — which meet a need for philological precision and accuracy, as well as satisfying material and editorial requirements (the kind of readers they address, the cost of paper, and so on) — attempt to preserve the basic ‘theoretical framework’ we have just seen. They give an account to readers of the alternative variants (located in the footer) and of the marginalia (separately collected in an appendix or in the Note to the text), although they do not offer the genetic and/or evolutionary apparatus; or do not have the space for it. An example of the former case is offered by the works of Pier Paolo Pasolini as edited by Silvia de Laude and Walter Siti for the Mondadori series ‘i Meridiani’.

63This simple, straightforward distinction between different textual levels is the core premise of many critical editions by Isella and his students, editions which contributed to providing specific ecdotic solutions in several complex cases. It is particularly in these editions that took place over time a significant development towards a diachronic and systemic apparatus.

2.5 Diacritic signs and abbreviations

64In order to represent the phenomena found on the manuscript, critical editions in the field of authorial philology have made use of the most diverse range of diacritic marks and systems of abbreviations. Even though the scholars have repeatedly expressed the need for homogenization, there are still no standard criteria. For this reason, readers find themselves forced to familiarize with the various systems adopted by the editor every time they encounter a new edition. Standard symbols are available only for a few corrections, and there is widespread confusion about most of the others. Sometimes, the same symbol can even be used by different editors to represent different, if not opposite, phenomena.

65There is a certain agreement as for what concerns the most frequently used symbol: the square bracket []. It is well known that square brackets are used to mark everything that is due to intervention of the editor, including: restoring letters missing in truncated words (e.g., wor [d]); the explanation of various phenomena; the filling in of lacunae and of punctuation in texts with no print tradition (although it would be preferable to explain how punctuation has been restored in the Note to the text). When we find square brackets, in other words, we should expect an editorial intervention on the text. In a few cases, editors can also make use of the angle brackets < > in order to represent the same phenomena.

66There is some agreement on the use of the series of three dots or ellipsis within square brackets […] or round ones (…) to indicate that part of the text is missing (the ellipsis is regularly used in this way in abbreviated quotations). A closing square bracket] marks in the apparatus the separation between text and variant. What comes before the bracket is the copy-text; what comes after it is the variant (which can be either manuscript or in print), including any symbols and/ or abbreviations used to explain its topography and/or chronology. Inverted angle brackets > < generally refer to a deletion (e.g., >xxxxx<). A deletion can also be noted by the use of the italics (e.g., xxxxx). In several apparatuses, however, the italics can also be used to indicate what does not change (the ‘invariant’), while square brackets are used to signal the deletion. Square brackets are often used, as we have seen, to restore parts of the text, or even when the reading of a word is doubtful: this is of course a very hazardous enterprise, because the use of the same symbols to represent different phenomena, as we are seeing, can cause great confusion.

67Unlike the symbolic apparatus, the explicit apparatus makes use of several abbreviations in order to represent the phenomena found in the manuscripts. To distinguish them from the text (which is in a standard non-italic font), the abbreviations are usually italicized. Here below you may find a list of some of the most frequently used abbreviations found in Italian editions (an English abbreviation is suggested next to the English translation):

Table 2 Abbreviations and their meaning

68In a symbolic apparatus, these abbreviations are replaced with diacritics that are understood in the same way. While the explicit apparatus usually gives us some details on the position of the variant such as whether it is written above, below or in the margin of the text, the symbolic apparatus cannot provide the same information, or must gather it in footnotes attached to the apparatus. This was the approach taken by Isabella Becherucci in her apparatus to Alessandro Manzoni’s Adelchi (see Manzoni 1998).

69In order to represent the diachronic relationship between the variants, the symbolic apparatus makes use of arrows:

• the direct arrow → represents a correction (and thus replaces the abbreviation corr. in ‘corrected to’);
• the inverted arrow ← represents a derivation (and thus replaces the abbreviation da ‘from’).

70In a few cases, in order to represent the chronology of the variants, two kinds of arrows can be used: a simple arrow for an immediate evolutionary variant (→), a two-colour arrow for a late evolutionary variant (➢). In the case of a particularly extensive correcting phase, it can be useful to represent smaller corrections within the same phase.

71Topographic and diachronic details are placed within italicized round brackets and are to be referred to the word that comes immediately before the opening brackets. When details are referred to more than one word, a reference mark is located at the beginning of the part of text concerned with the variant. This mark can have different forms: a black dot, a little star, an asterisk, or half of a square opening brackets (⌈). When on the same line of the apparatus there are more than one different variant, the variants are separated one from another with a fixed blank space (corresponding to the space of four or five characters), a small square figure □, or a tilde ~.

2.6 How to prepare a critical edition

72Let us now see how an editor can prepare the critical edition of a manuscript that features various series of corrections. Let us examine the manuscript of Giacomo Leopardi’s poem to the Moon Alla luna — titled in the manuscript (as we will now see) La Ricordanza — which helps us to understand, because of the number and types of corrections it presents, how an editor should proceed. Like the manuscript of L’infinito that we have discussed above, the manuscript of Alla luna (see Fig. 5) belongs to the so-called Naples notebook, which is housed at the Biblioteca Nazionale Vittorio Emanuele III of Naples (C.L.xiii.22) amongst other papers that belonged to Leopardi. As we have already noted, the manuscript is a clean copy, but still bears traces of interventions of different kinds. The question one might ask when facing such a document is: which textual version should be chosen as copy-text?

73If we decide to document the last textual version of the manuscript, we will have to transcribe the text including in it all the corrections, no matter if they are immediate or come from later phases. We will have to include in the text, in other words, both the corrections made with the base-pen (the pen used for the first writing) and the corrections made at a later time. As we have decided to publish the last textual version, the apparatus will obviously be genetic, giving account of all the correcting phases tagged with superscript numbers. Any deletions made on the writing line will be represented in the apparatus in a smaller font size with a two-point difference from the rest of the apparatus. Further corrections will be indicated with round brackets, which will also be in a smaller font size to help the passage to be read easily and to assist the understanding of the compositional phases.

74Let us see the corrections which interest the title of the poem, changing from¹La Luna to²La Luna o la Ricordanza to³La Ricordanza. The phases are always indicated with a superscript number and separated by a double spacing. If a phase is derived from another with the conspicuous reuse of textual materials, such as one or more letters, that phase is introduced by the abbreviation from which (It. da cui), or by a direct arrow →. How can we understand that the original title was ‘La Luna’ and not ‘La Luna o la Ricordanza’? Even though the ductus and the ink (which is identical in both phases) do not provide enough information, the placement of the text on the page provides useful orientation. If the original title had been ‘La Luna o la Ricordanza’, it is easy to imagine that the author would have placed it right at the center of the page — and not on the right, as the autograph clearly shows. Based on this, we can think that the title ‘La Luna’ was already written, and that, at a later time, Leopardi added on the right of it the second part ‘o la Ricordanza’, changing only at a third further stage the lowercase character ‘l’ (contained in the article ‘la’) to uppercase. This last correction does not seem contemporary to the base-writing, but bears some similarities with a few interlinear corrections in lines 2, 7–8 and 9, which were made at a further stage, and are contemporary to the writing of the texts which follow ‘La Luna’ in the Naples notebook, the so-called ‘second time’ of the Idilli (see Italia 2007b). This important observation gives us the possibility of identifying different levels of corrections in the text, corresponding to different pens, always indicated by the editor with capitalized alphabetical letters (A, B, C, D), which, alongside the superscript numbers, identify each correcting phase. In the case of smaller corrections, such as the one in line 5, the pen used is given within italicized round bracket and in a smaller font size. Further variants can indeed be found in the reconstruction of the same correcting phase, and, where present, they are represented in a smaller font size in a derivative from, such as in the case of line 4:

(su quella selva)^1Asopra quel bosco, from which^2Asopra quel prato, (with prato written over bosco)^3Asu quella selva, (written next to²) from which^BT

75This example gives us the opportunity to go into more detail on the use of the abbreviations from (It. da), from which (It. da cui), and from which T (It. da cui T). The first one from/da is used to represent a correction in which the final text materially reuses one or more letters of a previous reading. The second one from which/da cui, which can be replaced by the arrow →, refers to the reuse, in a subsequent variant, of a major textual portion of a previous variant. When the subsequent variant corresponds to the copy-text, it is possible to use the abbreviation from which T/ da cui T, and this can be replaced with an arrow pointing to T: → T.

76The representation of the variants in lines 7–8 illustrates what we meant when we introduced the notion of diachronic and systemic apparatus. The first definition — diachronic — stems from the fact that the apparatus does not focus on representing the placement of the corrections: for example, it does not indicate where the variants are located in the manuscript, that is, if above, below, or next to the text. It rather focuses on representing the chronology of the corrections, the evolution, that is, of the variants from an earlier to a later form, with each phase identified by a superscript number. Phases are here reconstructed as being four, including the last one, which corresponds to what has been chosen as copy-text. The second definition — systemic — refers to the fact that corrections are not represented individually and linked to the term (or terms) to which they refer from a topographic standpoint; corrections are instead represented in a system, including the variants and the final text too. The variant can obviously exceed the measure of a line and be related to the following line (this is the case of lines 7–8).

77This example shows that it is very difficult, if not almost impossible, for the vertical apparatus — which parcels out every variant and connects it very strictly to the words that are positioned closer to it — to represent effectively the corrections which concern two or more lines (a detailed examination of this is found in section 3.5).

78In the edition of the manuscript, immediately below the band occupied by the genetic apparatus, there is a box collecting the so-called varia lectio, that is, all the variants, quotations, linguistic and metatextual observations which Leopardi would annotate on his own manuscripts. We find them — albeit to a lesser extent than in other Neapolitan autographs — also in the manuscript of Alla luna. Leopardi penned a variant in the right margin of the page, in a position directly opposite to the written text of the poem, probably in a phase of later revision of the manuscript, as indicated by the reddish ink (here identified as pen D). Immediately below the box dedicated to the varia lectio, a further band accommodates the ‘Philological notes’, which are clearly distinguished from the apparatus by being italicized. When notes of this kind are instead included in the ‘Note to the text’, they are generally printed in a standard non-italic font. As these ‘Philological notes’ provide an analytical illustration of the makeup of the manuscript and its correcting dynamics, they serve a very useful function. Such notes also suggest various ways of interpreting the text, and report if and when the reading of one or more words is doubtful. More generally, the ‘Philological notes’ can include everything that editors might like to add in order to justify their choices in terms of apparatus, especially if such explanations cannot fit in the limited space underneath the text. A way to understand whether or not an ecdotic choice has been made judiciously is to reflect on the delicate balance between the need for analytical accuracy and the need for an economical form of representation. An apparatus is successful only when it represents the manuscript and its corrections in the most precise, clear and synthetic way, turning what is at first visual and iconic into a dynamic text.

79Let us now examine how the critical edition of the manuscript would look like. The abbreviations and expressions used in the ‘Philological notes’ have been translated in English as clearly as possible on the model of the abbreviations which would be used in an Italian apparatus (see sections 2.3.2, 2.3.4 and 2.5).

AN p. 1

Idillio

La Ricordanza

1	O graziosa Luna, io mi rammento
2	Che, or volge un anno, io sopra questo poggio
3	Venia carco d’angoscia a rimirarti:
4	E tu pendevi allor su quella selva
5	Siccome or fai, che tutta la rischiari.
6	Ma nebuloso e tremulo dal pianto
7	Che mi sorgea sul ciglio, a le mie luci
8	Il tuo volto apparia; ché travagliosa
9	Era mia vita: ed è, nè cangia stile,
10	O mia diletta Luna. E pur mi giova
11	La ricordanza, e ’l noverar l’etate
12	Del mio dolore. Oh come grato occorre
13	Il sovvenir de le passate cose
14	Ancor che triste, e ancor che il pianto duri!

tit.	La Luna o La Ricordanza]1ALa Luna from which2ALa Luna o la Ricordanza from which3BLa Ricordanza (with L over l)
2	Che, or volge un anno, ]1ACh’or volge un anno, (with an over al) from which2ACh’è presso a un anno, from which3BT sopra] from su (pen A)
4	su quella selva]1Asopra quel bosco, from which2Asopra quel prato, (with prato written over bosco)3Asu quella selva, (written next to2) from whichBT
5	Siccome or] written above Com’ora (pen B)

7–8	a le mie luci | Il tuo volto apparia; chè travagliosa]1Aa le (before al<le>) mie luci | Il tuo viso apparia, perché dolente from which2Ail tuo bel viso | Al mio sguardo apparia, perché dolente3Ba le mie luci | Il tuo volto apparia, che travagliosa from which4DT
9	cangia]1Acangia2Bcambia (written above1) from which3DT
11	ricordanza] from rimembranza (pen A)
12	come] written above quanto (pen B)
14	triste] from tristi (pen B) il] from ‘l (pen C?)

tit.	The text, initially only consisting in the title ‘La Luna’, is corrected to ‘La Luna o la Ricordanza’ with the same pen being used for the base-text (A); subsequently, it is corrected with pen B to ‘La Ricordanza’. The two phases A in the writing of the title are identified thanks to Leopardi’s customary practice of writing the title at the center, immediately below ‘Idillio’ (as in ‘La sera del giorno festivo’ and ‘La vita solitaria’).
2	The first correction (‘Ch’è presso a un anno, ’) is made with pen A; the same applies for corrections in ll. 7–8 and 11.
4–5	The corrections made with thicker and heavier ink belong to phase B.
7–8	As already noted by Domenico De Robertis (in Leopardi 1984: II, 327), the correction of the comma to a semicolon after ‘apparia’ and the accentuation of ‘che’ seem to have been made with the pen with a reddish ink (here called D).
9	The correction of ‘cambia’ in ‘cangia’ was made at a later time (De Robertis in Leopardi 1984: II, 327) with pen D, the same pen which introduces in l. 8 the grave accent on ‘che’ and which Leopardi uses to write the variant in the right margin, directly opposite to the written text ‘ (come sì grato’); in the edition ‘come’ is not in bold because the correction of l. 12 ‘quanto’ → ‘come’, made with pen B (though Lucchesini (in Leopardi 2009: I, 278) thinks that the correction was made with pen A) is thought to have been made before the writing of the varia lectio.
14	The correction over ‘tristi’ obscures with the lower part of the ‘e’ the point of the ‘i’, creating an unusual upward swirl in the formation of this letter. If the correction of ‘’l’ to ‘il’ might belong to phase C, given the serial nature of the intervention, the varia lectio in the margin is closer to phase D, sharing with it the ductus and the reddish color (again De Robertis in Leopardi 1984: II, 327).