¹ Hölderlin, Sämtliche Werke, ed. Friedrich Beissner (Stuttgart, 1943 ff.). Subsequent reference will be made to volume, part, and page, and in general discussion to the Grosse Stuttgarter Ausgabe as well.

Work toward this investigation was begun in 1961, at the Hölderlin-Archiv der Württembergischen Landesbibliothek, Schloss Bebenhausen b. Tübingen. I wish here to express my thanks to the Ottendorfer Memorial Fellowship Committee, Department of German, New York University, for a fellowship that made direct consultation of sources possible; to the authorities of the Stadtbibliothek, Bad Homburg vor der Höhe, for permission to reproduce photographically from its holdings of Hölderlin MSS; and especially to Maria Kohler, director of the Hölderlin-Archiv, Bebenhausen, for generously supplying me with out-of-the-way materials and helpful information.

This article was published with the help of grants from the American Council of Learned Societies, and the Graduate Research Committee, Dept. of Germanic Langs. and Lits., Univ. of Illinois.

² Manfred Windfuhr, “Die neugermanistische Edition: Zu den Grundsätzen kritischer Gesamtausgaben,” DVLG, xxxi (1957), 425–442; Horváth Károly, “A kritikai kiadások kérdéséről” [“On the Problem of Critical Editions”], Világirodalmi Figyelő [“Observer for World Literature”] [Budapest], iv (1958), 137–146; Hans Zeller, “Zur gegenwärtigen Aufgabe der Editionstechnik: Ein Versuch, komplizierte Handschriften darzustellen,” Euphorion, lii (1959), 356–377; Dietrich Germann, “Zu Fragen der Darbietung von Lesarten in den Ausgaben neuerer Dichter,” Weimarer Beiträge, 1962, H. 1, pp. 168–188.

³ For an effective partial summary see Heinz Otto Burger, “Die Hölderlin-Forschung der Jahre 1940–1955,” DVLG, xxx (1956), 329–331. References to individual reviews will be made below.

⁴ Die Stuttgarter Hölderlin-Ausgabe: Ein Arbeitsbericht, ed. Theophil Frey (Stuttgart, 1942), p. 16.

⁵ “Die Stuttgarter Hölderlin-Ausgabe,” Neue Zürcher Zeitung, Literatur und Kunst, Samstag, 24. November 1951, Bl. 4.

⁶ “Die reine Textgestalt Hölderlinscher Dichtung,” Universitas, xvii (1962), 45.

⁷ DVLG, xxxi, 439.

⁸ Hölderlin, Sämtliche Werke, ed. Norbert v. Hellingrath, Friedrich Seebass, and Ludwig v. Pigenot, 2d ed. (Berlin, 1922–23). Hereafter cited by the abbreviation “Hel.,” volume, and page.

⁹ Friedrich Hölderlin, Sämtliche Werke und Briefe, ed. Franz Zinkernagel (Leipzig, 1914–26). Cited by the abbreviation “Zk.,” volume, and page.

¹⁰ Rev. of Hölderlin, Sämtliche Werke, hg. von Friedrich Beißner, Vols. ii, v, Euphorion, xlviii (1954), 363.

¹¹ i, Part ii, 317–321; also in “Aus der Werkstatt der Stuttgarter Hölderlin-Ausgabe,” in Hölderlin: Reden und Aufsätze (Weimar, 1961), pp. 251–265; “Einige Bemerkungen über den Lesartenapparat zu Werken neuerer Dichter,” Orbis litterarum, Suppl. ii (1958), 15–20. Cf. Horváth, who views Beissner's apparatus with thoughtful approval: “... a variánsok progresszív elemzése (a legrégebbtől a véglegesig) a költői alkotómunka irányát világítja meg, regressziv irányban (a végleges változattól az elsőig) a mondanivaló konkrét értelmét magyarázza” (“ ... the progressive analysis of the variants—from the earliest to the final—illumines the direction of the poet's creative work; in a regressive direction—from the final variant to the first—it interprets the concrete meaning of the subject matter”) (Világirodalmi Figyelő [“Observer for World Literature”], iv, 144).

¹² “Der Hölderlin-Text: Zu Beißners Edition und zum Neudruck der Propyläen-Ausgabe,” DVLG, xxi (1943), Referatenheft, pp. 88–123; “Die Stuttgarter Hölderlin-Ausgabe,” Iduna: Jahrbuch der Hölderlin-Gesellschaft, i (1944), 226; “Zum Fortgang der Stuttgarter Hölderlin-Ausgabe,” Hölderlin-Jahrbuch, vii (1953), 90.

¹³ Hölderlin-Jahrbuch, vii, 91–92, quoting Lachmann, Wort und Wahrheit, vii (1952), 379–381. Cf. the conciliatory tone in Rodolfo Paoli, “La grande edizione delle opere di Hölderlin,” Letterature moderne, vi (1956), 97–104.

¹⁴ Euphorion, lii, 356–377 (cf. n. 2); rev. of Friedrich Beissner (Hsg.), Hölderlin, Sämtliche Werke, Vols. i, ii, v, AfdA, lxix (1956–57), 75–82. Subsequent references to both articles will appear in the text.

¹⁵ In the “Anhang” to Vol. iv of his edition Hellingrath writes: “Endlich habe ich in der Erwägung, dass kein Druck von Hölderlin beaufsichtigt, keiner genaue Wiedergabe der Vorlage ist, für den Text wo irgend möglich mich an Handschriften gehalten. Allzu ängstlich vielleicht; da ich es für einen Hauptzweck unserer Ausgabe ansah—ein ne varietur wird erst viel spätere Zeit schaffen können—, die Forderung treuester Urkundlichkeit aufzustellen, gerade bei einem so leicht missdeutbaren Dichter” (Hel., iv, 270). That there is, and can be, no such thing as a critical edition possessed of an unalterable final validity has been adequately pointed out by Allemann: “Hellingrath hat davon gesprochen, daß in der Entzifferung Hölderlinscher Texte ein ne varietur erst durch eine viel spätere Zeit erreicht werden könne. Beim Beginn des Erscheinens der Stuttgarter Ausgabe mochte es scheinen, diese Zeit sei angebrochen. Heute fragen wir uns, ob ein ne varietur im Bereich der späten Entwürfe überhaupt jemals zu leisten sein wird. Es bleibt das Hauptverdienst der Stuttgarter Ausgabe, gerade in diesem Bereich sehr viele Lesungen im einzelnen richtiggestellt und gesichert zu haben. Aber alles weitere bildet eine unendliche Aufgabe der Hölderlin-Forschung” (AfdA, lxix, 82).

¹⁶ ii, Part i, 165–172; Part ii, 764–784. This is the fair copy version, or the so-called “Widmungsexemplar,” only (described as MS. H ⁵, ii, Part ii, 764).

¹⁷ More recently Walther Killy has examined the relative merits of Beissner's and Zeller's editing techniques (“Der Helian-Komplex in Trakts Nachlaß, mit einem Abdruck der Texte und einigen editorischen Erwägungen,” Euphorion, liii [1959], 380–418). Killy finds Beissner's “staircase” system (“Technik der Kolumnierung,” p. 407) preferable to Zeller's apparatus on the grounds of clarity and an open admission that editorial decisions form a vital part of textcritical work: “Wer auf der Basis der Symbole Zellers ohne Kenntnis des Originals die Handschrift ‘rekonstruiert’, erhält eine schematisierte Topographie der lebendigen Niederschrift. Sie bildet deren Grobstruktur ab und die Lese- und Entscheidungsweise des Herausgebers, aber gewiß nicht das Manuskript. Je schwieriger die Handschrift ist, um so weiter muß sich das symbolische Schema von ihr entfernen” (p. 408). For his Trakl edition Killy has “versucht, aus beiden Methoden zu lernen und an der Kolumnierung zusammengehöriger Textentwicklung festgehalten” (p. 409). Killy's main criticism of Beissner's editing is concentrated on the latter's indiscriminate repetition, in the apparatus, of words and passages not actually repeated in the MS (pp. 407–408). While Killy is right in his insistence on fidelity to the MS, an important lesson taught by Beissner's method is that it is not possible for any apparatus to explore the genetic development of a text without a repetitive presentation of at least some of the variants. For this reason, my readings below will expand Beissner's practice of repetition, but in a systematic manner, making full use of Killy's special double-dotted slashes, /: :/, for purposes of marking all passages repeated in the reading but not in the MS. (p. 411). In keeping with the use of Beissner's apparatus, the sigla notation of the Grosse Stuttgarter Ausgabe (viz. H ², ii, Part ii, 764) will also be consistently followed.

¹⁸ This is done by Ulrich Häussermann, who, in his readings against Beissner, uses the apparatus of the Grosse Stuttgarter Ausgabe throughout (Friedensfeier: Eine Einführung in Hölderlins Christushymnen [München, 1959], pp. 251–260).

¹⁹ Giovanni Pascucci writes: “È còmpito della critica testuale la riconstruzione di uno scritto nella forma più vicina possibile a quella dell'originale perduto, cioè all'autografo dell'autore (constitutio textus)” (I fondamenti della filologia classica [Firenze, 1957], p. 29). Pascucci's definition is preserved here in part only, as the task of text constitution in Hölderlin MSS is not, of course, the recovery of lost autograph texts, but the recovery of lost or obscured intentions. On this latter point cf. Pyritz, Hölderlin-Jahrbuch, vii, 92–94; on the importance of correct textual editing for sound literary scholarship see Fredson Bowers, Textual and Literary Criticism (Cambridge, England, 1959), pp. 1–34.

²⁰ Photographs, approximately actual size (with edges cut, 22.8 cm. wide; actual width, c. 30 cm against MS width 24 cm) (Plates i, iii, vi), over twice actual size (Plate ii) and approximately twice actual size (Plates iv, v, vii, viii) were made on matte paper with a large Aerial Ground camera and Kodak Contrast Process Ortho Film.

²¹ ii, Part ii, 764–766.

²² Described by Beissner, ii, Part ii, 380–381; also in Katalog der Hölderlin-Handschriften, ed. Johanne Autenrieth and Alfred Kelletat (Stuttgart, 1961), pp. 93–94. This MS will hereafter be referred to by “Homburg F” and page.

²³ See the MS of the hymn “Mnemosyne” at the end of the “Folioheft” (Homburg F 90–92), where, besides ink blots, there appear dots (in front of lines of verse) and short vertical separating strokes (between columns of writing) (Homburg F, 91, 92). A facsimile of the MS accompanies Beissner's article, “Hölderlins letzte Hymne” (Hölderlin-Jahrbuch, iii [1948–49], 66–102), following pp. 80 and 96; also in Beissner, Hölderlin: Reden und Aufsätze, Plates iv-vi (as separate enclosure in rear of volume).

²⁴ “Aus der Werkstatt ...,” Hölderlin: Reden und Aufsätze, pp. 256–260; cf. Beissner, “Einige Bemerkungen über den Lesartenapparat zu Werken neuerer Dichter,” Orbis litterarum, Suppl. ii (1958), 15–20.

²⁵ Each of the three decipherments below is presented in the following sequence of steps: (1) the edited text passage as printed in Volume ii, Part I of Beissner's edition, (2) a diplomatic transcription of the MS area under examination, (3) photographic enlargements of the MS area and of individual passages, words, or punctuation marks not observed by Beissner, (4) all available previous readings relevant to the interpretation of the MS area, (5) a set of textual notes explaining the decisions, both paleographic and interpretive, on which I base my reading, (6) my reading of the MS area, following consistently Beissner's “staircase” system of transcription, and (7) a series of discrete texts obtained from my reading and showing the development of the passage in question, from earliest draft (Text i) to fair copy (P₁).

Readings by editors will be abbreviated as follows: H=Hellingrath (Hel., iv, 361–362; cf. above, n. 8); Z=Zinkernagel (Nachlass Franz Zinkernagels. Stuttgart, Württembergische Landesbibliothek, cod. poet. et philol. 4° 195, hereafter referred to by “Nachlass” and page); B = Beissner (ii, Part ii, 773–783); M = my reading. Edited versions of the hymn “Patmos” will be referred to by abbreviations and subscript numbers: P₁=“Patmos, Dem Landgrafen von Homburg” (ii, Part i, 165–172); P₂=“Patmos, Dem Land grafen von Homburg: Vorstufe einer späteren Fassung” (ii, Part i, 173–178). The abbreviation “T” will refer simultaneously to the diplomatic transcription and to the photograph of the corresponding MS area (in decipherments i, ii, and iii below, to Plates i, iii, and vi, respectively). The word “Text” or “Texts” followed by a Roman numeral will in every instance refer to the discrete text solutions following my readings. For ease of reference all transcriptions, readings, and texts have been provided with line numbers. Plain numbers after an abbreviation will in each instance give the appropriate line reference; plain numbers preceding the entry words to a textual note will refer to my reading.

²⁶ Corresponds to Hel., iv, 194, vss. 117–120.

²⁷ Corresponds to Zk., i, 359.

²⁸ For an example, since Beissner, of the practice of anticipating the fair copy text see Allemann, AfdA, lxix, 78–79. Such practice is unsound because it does not take into account the possibility that the poet may have changed his mind regarding the retention or rejection of a variant while preparing the fair copy. See also decipherment iii, Text viii and P₁, below.

²⁹ In the form of my readings, which in all other respects retain Beissner's “staircase” model of transcription (for explanation see above, page 124), I have dispensed with Roman “Ansatz” numerals completely and with Arabic verse numbers partially. Instead I use a technique of multiple solutions wherein one and the same “Ansatz” may receive variant chronological interpretation (as indicated by the italicized word or to the extreme left of an entry so affected). Where the dispensing with Beissner's verse-by-verse presentation has the effect that certain later variants must be introduced by new entry numbers or letters, I continue Beissner's (1), (a), (α) system with my own (i), (i), (A), (a), (l). A + at the end and beginning of two consecutive lines in my reading means that the material so marked constitutes one line, which had to be broken for lack of space.

³⁰ Corresponds to Ed., iv, 196, vss. 178–180.

³¹ This argument is based on an assumption that the relative sizes of individual letters, especially of capitals, are indicative of their relative chronology; the largest writing appears to have been done on maximum available space, i.e., on the blank page. Similarly, the difference in size between the two S's of the two closely located occurrences of the word “Sonne” (T 3, 4, right), or between T of “Triumphgang” (T 3, extreme left) or T of “Tapfern” (insertion above T 5, left), may be taken to have chronological significance.

³² B's entry “(a) Sohn” (4) appears to be misleading; B means not “Sohn des” (T, below 5, 6, right middle) but “Sohn des Höchsten” (below T 5, extreme right), completed by B in entry (1) (b) “frohlokende Sohn des Höchsten” (B 5–6).

³³ Corresponds to Eel., iv, 197, vss. 190–195.

³⁴ Corresponds to Zk., i, 361.

³⁵ This theory may probably be proved wrong by arguing that Hölderlin was in the habit of forgetting to cross out rejected material; cf. in the present decipherment the word “sie” (T 7, extreme right), left uncanceled at the stage of the concluding lines in the final version, “Stillleuchtende Kraft trinkt, mögen / Der Gnade sich freuend sie / Am stillen Blike sich üben” (Text viii, lines 4–6; reflected also in P₁, vss. 194–196). The present argument on the chronology of “nur” is advanced as a possibility only.

³⁶ Orbis litterarum, Suppl. ii (1958), 19.

³⁷ Although we have in Hölderlin's correspondence practically no references to his way of drafting and revising poems, an analysis of the latter, based on MSS of the middle and late odes, has so far been performed by Horst Knospe, “Die Entstehungsweise der Gedichte Hölderlins” (diss. Berlin, Freie Univ., 1956). In summary of the principle of his study Knospe writes: “Hölderlin neigt dazu, das Kernstück eines Gedichtes... oder einer Versgruppe..., die Aussage, die ihm am Herzen liegt, zuerst niederzuschreiben. ... Der Entwurf ist ein Unentwickeltes, ist ganzheitliche, inhaltlich noch nicht durchgebildete Frühgestalt, aber eben schon Gestalt, die Wesen und Gehalt der Endgestalt erkennen läßt” (pp. 28–29).

³⁸ Cf. Jonas Fränkel's view: “... daß es möglich wäre, mit Hilfe eines verzwickten Systems von Zeichen und Klammern eine komplizierte Handschrift wirklich wiederherzustellen, halte ich für Illusion. Im Zeitalter der höchstentwickelten photographischen Kunst erscheint mir die darauf verwendete Mühe als sehr überflüssig. Wenn Ernst Grumach in die Goetheausgabe der Berliner Akademie ein vollständiges Faksimile der Urfaust-Handschrift aufgenommen und dadurch den noch zu erwartenden Apparat zu Faust I merklich entlastet hat, so schlug er meines Erachtens den richtigen Weg ein” (“Zum Problem der Wiedergabe von Lesarten,” Euphorion, liii [1959], 421).