ANDREAS, FRIEDRICH CARL, German Iranologist (1846-1930).

i. Life and work.

ii. The Andreas Nachlass.

iii. The Andreas Theory.


i. Life and Work

Andreas was born in Batavia, Java, of Armenian, German, and Malayan descent. After general education in Hamburg and Geneva, he pursued Iranian and other Oriental studies at Erlangen, Göttingen, Halle, and Leipzig universities, and completed his graduate work in Copenhagen and Kiel. In 1887 he married Lou Andreas-Salomé, a well-known German writer and psychoanalyst. Between 1875 and 1881, he conducted field work in India with the Parsees and in southern Iran. His research in Europe focused on the languages and music of Ossetia and the Indo-Afghan borderlands. From 1903 to his death he was professor of western Asiatic philology at Göttingen.

A master of many living languages, Andreas specialized in the history of languages and civilizations, but his interests extended to philosophy and natural history. He excelled in reading difficult Oriental scripts, ancient or modern, and in perceiving the finest nuances of spoken languages, especially their accents. Although reluctantto publish the results of his research, he was generous in making them available to students and friends. Many of his findings are scattered in the publications of other scholars (and not always formulated to his satisfaction); much is still unpublished. Iranologists of several generations, such as Kaj Barr, Arthur Christensen, Bernhard Geiger, Walter Bruno Henning, Paul Horn, Wolfgang Lentz, Herman Lommel, and Oskar Mann, owe him decisive influences on their work. Among his fundamental insights was the recognition that the difference between “Arsacid” (i.e., Parthian) and “Sasanian” (i.e., southwestern) Middle Iranian is essentially one of dialect, rather than time sequence. Working with the Manichean fragments from Turfan, he quickly isolated those texts written in Parthian (which he called the “northern dialect”) and identified another “Pahlavi dialect” as the Sogdian language. Even when his ideas were later modified, they proved important starting-points for research; this is particularly true of the hotly debated “Andreas Theory” (see below). In an equally sharp controversy concerning an alleged Iranian “salvation mystery” Andreas took no personal position, but his translations of the Manichean Turfan fragments provided part of the basis for discussion.



Works: The Book of the Mainyô-i-khard, Also an Old Fragment of the Bundehesh, Kiel, 1882 (partial facsimile of Copenhagen Pahlavi Codex K 43).

With F. Stolze, Persepolis, die achaemenidischen und sasanidischen Denkmäler und Inschriften von Persepolis, Istakhr, Pasargadae, Shāhpūr, 2 vols., Berlin, 1882.

With idem, Die Handelsverhältnisse Persiens, Gotha, 1885.

Pauly-Wissowa, thirty-five entries in vols. I and II. Die Babi’s in Persien, Leipzig, 1896.

“Über einige Fragen der ältesten persischen Geschichte” and “Die Entstehung des Awesta-Alphabetes und sein ursprünglicher Lautwert,” Verhandlungen des XIII. internazionalen Orientalisten-Kongresses, Hamburg September 1902, Leiden, 1903, pp. 93-97, 99-106.

“Mehmed Mesʿud, ein muhammedanischer Katechismus,” Potsdam , Heften zum Christlichen Orient 11/4, 1910.

Epilogue to W. Bang, “Beiträge zur Erklärung des komanischen Marienhymnus,” NGWG, phil.-hist. Kl., 1910, pp. 74-78.

“Vier persische Etymologien,” ibid., 1916, pp. 1-6.

Editions of the Avestan Gathas: Y. 30, ibid., 1909, pp. 42-49; Y. 31 (with J. Wackernagel), ibid., 1911, pp. 1-34; Y. 28, 29, 32 (with idem), ibid., 1913, pp. 363-85; ed., H. Lommel, “Gāthā’s des Zarathustra mit Benützung der Entwürfe von F. C. Andreas” (Y. 43-46, 47-51), ibid., N.S., 1, 1934, and 1935, pp. 67-119, 121-69; ed., H. A. Winkler, “Erklärung der aramäischen Inschrift von Taxila,” ibid., 1932, pp. 6-17.

Ed. A. Christensen, (with K. Barr and W. Henning) Iranische Dialektaufzeichnungen aus dem Nachlass von F. C. Andreas, Berlin, 1939.

The Turfan materials: “Zwei sogdische Exkurse” to Wilhelm Thomsen ; EinBlatt in türkischer Runenschrift,” SPAW, 1910, pp. 307-14; “Bruchstücke einer Pehlewi-Übersetzung der Psalmen aus der Sassanidenzeit,” ibid., 1910, pp. 869-72 (introductory remarks).

The text of the Psalter was edited by Kaj Barr, Bruchstücke einer Pehlevi-Übersetzung der Psalmen, ibid., 1933, pp. 91-152; also ed.

W. R. Henning, Mir. Man. See also: Festschrift Friedrich Carl Andreas, Leipzig, 1916 (with an early bibliography).

Obituaries by W. Lentz, ZII 8, 1931 (with a summary of the last item below and an outline of the materials left by Andreas); G. von Selle, Indogermanisches Jahrbuch 15, 1931, pp. 366f.; E. Littmann, Neue Deutsche Biographie, 1953, I, p. 284.

Biographical outlines, based partly on H. Lommel’s unpublished Erinnerungen an F. C. Andreas und Lou Andreas-Salomé, in the last two chapters of L. Andreas-Salomé, Lebensrückblick, Zurich, 1951, latest rev. ed., Frankfurt, 1974, and H. F. Peters, My Sister, My Spouse, New York, 1962.

An autobiographical sketch by Andreas covering the first half of his life, ed. (with some misreadings) H. Kanus-Credé, Iranistische Mitteilungen 8, 1974.

(W. Lentz)


ii. The Andreas Nachlass

When Andreas died in 1930, his scholarly papers were first gathered into the University Library of Göttingen and sorted by W. Lentz, who published a short descriptive list of them as an appendix to his obituary (“Der handschriftliche Nachlass,” 8 , pp. 14-17). Some parts of the material were destined to go to the Prussian Academy of Sciences or the Prussian State Library in Berlin, but the greater part was to remain in the possession of the University Library at Göttingen.

Although very little of the material was ready for publication, two students of Andreas, using to some extent their own lecture notes and those of their fellows, soon brought out editions of important Middle Iranian texts on which Andreas had been working for years. W. B. Henning began with the Mitteliranische Manichaica aus Chinesisch Turkestan, of which parts I-III appeared in quick succession in the Sitzungsberichte der Preussischen Akademie der Wissenschaften, Berlin, 1932-34. Kaj Barrperformed the same service with the editio princeps of the Bruchstücke einer Pehlevi-Übersetzung der Psalmen, SPAW, 1933. A similar undertaking was the publication of the considerable amount of linguistic material recorded by Andreas himself during his long stay in Iran, 1876-80. This fell into three main parts, relating to the dialects of Fārs, those of central Iran (Sīvand, Yazd and Sō), and some Kurdish dialects. The notes on Kurdish were dealt with exhaustively by Kaj Barr, appearing in the volume IranischeDialektaufzeichnungen,Berlin, 1939, together with Arthur Christensen’s treatment of the Central dialects. The Fārs dialect material however, entrusted to W. B. Henning, fell victim to his many other interests and so lay fallow for many years until, after his death (California, 1967), it was returned to Göttingen. This unpublished material comprises several notebooks, mainly devoted to the “Tāǰekī” dialect of Rīšahr, near Būšehr (Bushire), but also containing notes on the related dialects of Ardakān, Garrāš, Kelāt, Somḡūn, Tangestān, etc., and on some Lorī dialects. After a century, and despite later work by others, this material is still of considerable value. Other material from the same period includes notes on the spoken Persian of Šīrāz, including a short collection of proverbs. Many years later Andreas had an opportunity to widen his knowledge of modern Iranian when, in the First World War, the German authorities set about methodically recording the languages of the more exotic of their prisoners of war (special forms were printed for the purpose). Andreas made extensive notes among Ossetes of the Russian army and Pathans of the British Indian service. This Ossetic material was somehow mislaid during the upheavals of the Second World War—an unfortunate loss, tempered by the fact that A. Christensen had early published his own notes from a similar source (Textesossètes, Copenhagen, 1921). The Paṧtō material, over 1500 pages of notes on the Afrīdī, Mohmand and Yūsofzay dialects, has mostly been returned to Göttingen. Only a very small part of it has so far been exploited, from copies which have found their way to the United States (see W. Skalmowski in ManumentumH. S. Nyberg, Acta Iranica 5, Tehran and Liège, 1975, pp. 243f.) Of living languages, though not of the Iranian family, there remains to be mentioned Brāhūī, copious notes on the grammar of which Andreas had made in 1875 in Gwādar, when en route for Persia.

His Avestan studies resulted in one of his rare publications (in collaboration with J. Wackernagel), an edition and translation of the first five Gathas, Yasna 28-32 (NGWG, Phil.-hist. Kl., 1909, 1911, 1913, and 1931). His draft translations and notes to most of the remaining Gathas, Yasna 43-51, were used shortly after his death by H. Lommel in his publications (NGWG, Phil.-hist. Kl., 1934-37), thus practically exhausting this vein. The remaining notes on Old Iranian and related subjects are rather diverse in nature.

It was to Middle Iranian that Andreas seems to have devoted more paper, if not time. Although there is not much on Book Pahlavi (besides the Psalter, mentioned above), there is a great deal on Pahlavi inscriptions, including Hāǰǰīābād, Naqš-e Raǰab, Persepolis, and particularly Paikuli. E. Herzfeld had provided him with copies of the photos which he later published himself (Paikuli I-II, Berlin, 1924). Andreas has been harshly criticized in this connection (H. Humbach, The Sassanian inscription of Paikuli, I, Wiesbaden, 1978, pp. 14f.). However, any man in his seventies could be proud of the “passiveness” of which he is accused, which produced handcopies of the inscription, a glossary of inscriptional Pahlavi, studies on the ideograms, and an elaborate linguistic commentary. But Andreas’ perfectionism was undoubtedly an enemy to publication. He also left readings and notes on a large collection of coins inscribed in Middle Persian on the basis of impressions also still present.

His studies of the fragments from Turfan were made with the help of a set (incomplete) of photographs supplied to him from Berlin. Some of these photographs are now the only record of fragments which have meanwhile been lost or destroyed (see Boyce, Cat. Man. Script, pp. xxiv, xxxi). He left notes to the pioneering publications of F. W. K. Müller, on the grammar and vocabulary of the texts, and on the characteristic differences between Middle Persian and Parthian, as he had identified the two languages. Some of his transcriptions and translations of the fragments are of texts which have still not been published. The same can be said of his notes on Manichean Sogdian texts, of which he had also begun to compile a glossary. Most of the Christian Sogdian texts on which he had been working, on the other hand, have since been brought out independently by others. His observations on Sogdian phonology and grammar have, of course, never reached a wider circle than his later students. The rest of Andreas’ many notes, which were probably never intended for any publication, covers a wide range of subjects. A random sample of themes includes xwaētwadaθa (kin-marriage), Persian legends, lunar mansions among the Iranians, the Manichean calendar, the ceremonies of the Parsees, and so on. Finally there are his letter-books, five volumes kept from 1887 to 1930, and nine boxes full of letters received from colleagues.

Also among Andreas’ papers there are a number of manuscripts from other hands, though regrettably there is no indication of their provenance. The most interesting of these are all specimens of Kurdish. There are two glossaries of the Mokrī dialect, one labelled 1896, and a short and rather inaccurate grammatical sketch in English, evidently of the same dialect, by one Mesrof Karam, presumably an Armenian Christian. Then there is a translation of St. Matthew’s gospel in a popular transcription and, hitherto unrecognized, a short collection of simple folktales written in Syriac script in the dialect of Gewer, obviously by an Assyrian Christian. The transcriptions of both texts are too inconsistent, however, to give either of these curios value for the dialectologist.

(D. N. MacKenzie)


iii. The Andreas Theory

The theory concerning the transmission of the Avesta as formulated by F. C. Andreas.

History of the theory. The theory was presented in summary form by Andreas at the International Congress of Orientalists held in Hamburg in 1902 and then published in the following year (“Die Entstehung des Awesta-Alphabetes und sein ursprünglicher Lautwert,” Verhandlungen des XII. Internationalen Orientalisten-Kongresses in Hamburg 1902, Leiden, 1903); but no detailed explanation of the theory was ever published. It consists mainly of a list of the letters of the Avestan script with remarks about the alleged paleographic origin of each letter derived from the Book Pahlavi alphabet. How the theory worked in practice is exemplified by an edition of some of the Gathas published by Andrea and J. Wackernagel (NIGWG, 1911 and 1913) which contain the vulgate text, a reconstruction of the Arsacid text, and a phonological interpretation of the reconstructed text.

Andreas imparted more details of his theory to his students. It is reflected in particular in a number of articles by H. Lommel, and F. Altheim provided some modifications and additions in his Awestische Textgeschichte (Halle, 1949). Altheim argued that the idea of vocalising unvocalised texts was due to Creek influence. The transcription of Avestan words by scholars such as A. Meillet, E. Benveniste, and A. Christensen shows the influence of the Andreas Theory in a moderate form.

The theory encountered strong opposition from the very beginning and was eventually given up by all Andreas’s pupils except Lommel. Detailed criticism of the theory was published by G. Morgenstierne (NTS 12, 1942, pp. 30-78) and by W. B. Henning (TPS, 1942, pp., 40-56). The problems involved have been treated in greatest detail by K. Hoffmann (HO IV/1, 1958, pp. 1-19).

The Theory. The Avesta was originally written in a simpler script comparable to that of Book Pahlavi. The Avestan text written in that script was called by Andreas the “Arsacid original,” an impression of which is given by the spelling of Avestan loanwords in Pahlavi texts. Thus, Pahlavi pwlywtkyš can be read as Avestan paoiriiō.ṱkaēša- or as Pahlavi pōryōtkēš, and Pahlavi ʾstw(w)ydʾtʾ/w as Avestan astō.viδātu or as Pahlavi astwihād. Andreas did not base his arguments on that material but it may have played a role in the development of the theory. One would not expect proto-Iranian *-am, Avestan -əm, to have been written in the non-vocalised text as -wm as Andreas maintained, but Avestan fratəma- is written pltwm in Pahlavi script (prtwm in Manichaen script), and this may have influenced Andreas. Ahyā yāsā nəmaŋhā (Y. 28.1) was reconstructed by Andreas as ʾhyʾ yʾsʾ nwmwhʾ, which he interpreted as ohyo yāx nomohāx.

In Sasanian times the Avesta was mechanically transposed into the newly invented Avestan script, which was designed to record all the phonetic distinctions of the oral tradition including full vocalisation. The new alphabet was composed partly of letters taken over from Pahlavi with the same shape and value. Most of the new letters were ligatures of older ones. Thus, e and ē are ligatures of y + y + w. Since Avestan eyo became e in Pahlavi, Andreas assumed that the ligatures for e and ē could derive from yo, , yu and ē/ĕ. In any given case the original reading could be determined only on the basis of linguistic analysis.

According to the theory the traditional pronunciation of the Avestan letters is in some cases wrong. Those responsible for the transcription in the Sasanian period made many mistakes, but they did not usually alter the consonantal frame of the Arsacid original. Thus, a preterite verb form like gərəzaēta “she complained” must be interpreted as *gərəzayata (gurzoyoto) reflecting an earlier writing *gwrzyt (or gwrzyywt with the ligature yyw). The scribes may have been influenced by other cases where Pahlavi ē corresponds to Avestan ōi, aē. Another example of an erroneous Sasanian transcription would be yatāra “which of two.” This was supposed to have been written ytʾr to avoid confusion with ytr for yaθra “where” but really to have represented *yatăra. Avestan oifra (variant reading aoifra)is according to H. S. Nyberg (DieReligionen des alten Iran,Leipzig, 1938, p. 469) not written for *vifra but because of the initial vowel it goes back to ʾwpr, that is *avifra “non-seer” (see H. Lommel, ZDMG 105, 1955, pp. 151 f.).

A dialect difference between Gathic and later Avestan either did not exist or was of little importance. The apparent difference is due to the plane writing used for the Gathas because of their particular holiness. Thus, the gen. sing. masc. ending is spelled -hyʾ in the Gathas and is reflected by -ahyā in the vulgate whereas it is spelled -hy in the later Avesta, which is reflected by -ahe in the vulgate, but both are in reality -ahya (-ohyo)(for criticism of this point see J. Kuryłowicz, Metrik undSprachgeschichte, Warsaw, 1975. pp. 102f.).

The primary task of scholars is thus the reconstruction of the Arsacid original which alone carries authority in linguistic and philological investigations.

Criticism. That the Avesta was originally written in a simpler script comparable to that of Pahlavi and that this Arsacid original was mechanically transposed into the newly-invented Avestan script, which was based in part on the Pahlavi script, was widely accepted. However, in particular one assumption evoked much dissent, namely that Avestan ə, ə̄, o, ō, u, ū all go back to an Arsacid w. Moreover, according to Andreas w was part of the ligatures ą (= ʾwn), β (= wp), v (= wwb), š (= whr), and ž (= wṣ). Andreas was unable to suggest any motivation for the differentiation of w into ō/ǒ, ū/ŭ, and the only possible explanation is that the transcribers wished to record the Sasanian pronunciation precisely. However, although Andreas and Lommel did not expressly deny the influence of the later pronunciation, they did not take it into account in practice. The assumption of w in so many cases and the interpretation of most w’s as o is responsible for the strange appearance of the Avestan text reconstructed by Andreas and Lommel.

Most scholars were unable to take seriously writings based on the Andreas Theory. Nevertheless, support for Andreas’ explanation of v as wwb is Book Pahlavi wb for Avestan v (uu in Hoffmann’s transcription) in loanwords: Avestan aṧavan- (aṧauuan-) appears in Pahlavi as ʾhlwb (= ahlaw). On the other hand, Andreas was mistaken in assuming the existence of older and younger forms of Avestan letters since all the evidence points towards the invention of a phonetic script as a single deliberate act without internal history.

The most unfortunate aspect of the Andreas Theory is the idea of mechanical transposition of the Arsacid text into the Avestan script, which was intended to justify Andreas in mechanically transposing it back again. This led to neglect of textual history and of the influence of later pronunciation on the manuscript tradition. It is difficult to understand the reason for retranscribing Gathic fərā as pwrʾ but young Avestan fra as pr since fərā seems to be merely a narrower phonetic rendering of fra.

The new and more adequate theory of K. Hoffmann combines textual criticism, philological evidence, and linguistic evidence. Moreover, he has also shown the importance of taking into account the pronunciation of scribes and priests, which declined in accuracy throughout the ages. According to Hoffmann the oral tradition of the first transcribers was in general excellent but occasionally showed aberrations for which it is difficult to account. He is accordingly obliged to explain as purely graphic deformations such writings as friiąnmahī instead of expected *frinmahī, huuąnmahī instead of *hunmahī, and kərənūši (Arsacid *-nwšy) instead of *kərənaoši. (The graphic explanation of the gen. plur. -anąm he gave up later.) In these cases Hoffmann argues like Andreas, but whereas Andreas began at once with his retransposition, Hoffmann first evaluates all the attested manuscript evidence and attempts to find a linguistic explanation of the form presumed to have stood in the Sasanian archetype. Only if that is not possible does he resort to graphic interpretation. Thus, if in a given case textual criticism does not enable him to decide between š′, š, and , he decides on the basis of etymology after the fashion of Andreas.

It should be emphasized that the forms of the archetype are not necessarily identical with those of the authors. F. B. J. Kuiper remarked concerning the Avestan manuscripts in general that “more often the text may be expected to reflect the dialectal peculiarities of later scribes” (IIJ 4, 1960, p. 189). Andreas’ aim was in fact to eliminate from the Avesta the influence of the dialect of the scribes with its complicated phonology and to remove the mistakes that had occurred in the course of the presumed mechanical transposition from the Arsacid original into the Sasanian Avesta. However, it is now thought that most of those mistakes took place during the later history of the Sasanian text.



Some Gathas were treated according to the theory by F. C. Andreas: Y. 30, NGWG, 1909, pp., 42-49; by F. C. Andreas and J. Wackernagel: Y. 31, ibid., 1911, pp. 1-34; Y. 28, 29, 32, ibid., 1913, pp. 363-85 (with notes in NGWG, 1931, pp. 304-29).

Some writings based on the theory are: H. Lommel, Festschrift F. C. Andreas, Leipzig, 1916, pp. 97-108; ZII 1, 1922, pp. 185-245; 3, 1925, pp. 171-78; 5, 1927, pp. 1-92; 6, 1928, pp. 126-44; ZDMG 105, 1955, pp. 151-52.

Detailed treatment of the problems involved in K. Hoffmann, Aufsätzezur Indoiranistik, Wiesbaden, I, 1975, pp. 58-76 (= HO IV, 1, 1958, pp. 1-19), 274-87 (= W. B. Henning MemorialVolume, London, 1970, pp. 187-200), 316-26.

(B. Schlerath)

(W. Lentz, D. N. MacKenzie, B. Schlerath)

Originally Published: December 15, 1985

Last Updated: August 3, 2011

This article is available in print.
Vol. II, Fasc. 1, pp. 27-30