The next major stage in the history of attempts to create a Chorasmian dictionary, and the second to be interrupted by death, is that of David N. MacKenzie (1926-2001). Initially a specialist in Pashto and Kurdish, MacKenzie moved into Old and Middle Iranian as a student of Henning at SOAS in the 1950s. As a faculty member of SOAS, by the mid-1960s, he had evidently become interested in Chorasmian (despite his own admission (1971:1) that he “never studied any appreciable amount of Khwarezmian” with Henning) and had begun studying the available material in earnest. Already in 1968 he had prepared a preliminary edition of the Qunyat al-Munya and its Chorasmian phrases, but this project was paused for several decades until he was able to obtain access to better manuscripts of the text from Soviet archives.

MacKenzie’s work towards a Chorasmian dictionary also seems to have begun in the late 1960s, even before he received the fragment of Henning’s dictionary which he edited and published in 1971. Indeed, it was his intimate familiarity with Chorasmian and its structure that enabled him to publish his extensive, critical reviews of Benzing’s work between 1970 and 1972. Although MacKenzie seems to have produced several kinds of preparatory material for a dictionary over the decades, he—like Henning—never published any of it during his lifetime, despite devoting much time to the project after his retirement in 1994. A glossary to his edition of the Qunya (1990) is thus his only published lexical material.

Once MacKenzie passed away and his Nachlass was made available to scholars interested in finishing some of his unfinished projects, some insight could be obtained into the—fairly significant—progress he had actually made on Chorasmian. As Desmond Durkin-Meisterernst first reported in English and German papers on Chorasmian lexicology (both 2005), MacKenzie had, sometime in 2001, printed out a close-to-final draft of his finished entries. According to my count, these total 1888 completed entries covering the letters from alif to most of ghayn (breaking down as follows: ʾalif 950 ʿayn 47 B/b 320 β 72 C/c 113 č 81 D/d 59, δ 90, F/f 65 G 1 Γ/γ 90). MacKenzie also had other kinds of materials in various states, including basic lists of all attested words, concordances to the texts in which they occur, handwritten/typed drafts of entries up to the letter n, and his own card catalog of the complete lexicon. It is clear as well that all his work on a dictionary was based on his own editions of all the existing Chorasmian material in Arabic script—besides the Qunya (1990), he also published corrected decipherments of the Yatīmat al-dahr family of manuscripts (1996).

It seems that MacKenzie envisaged his dictionary to have two main parts: the complete glossary with definitions and attestations, and a part with etymological discussion. The latter part, which he separated out from the former, probably in order to finish it more quickly, does not seem to be preserved (or perhaps was never really drafted). It is also unknown how much longer MacKenzie thought he needed to draft the remaining entries; it probably would have required a few more years. However, it is obvious from the existing near-final draft that he had essentially all of the Chorasmian material gathered and organized, and only needed to write it up into the format he chose for the lemmas. MacKenzie therefore got much further than did Henning, and his material was even more thorough—only to also unfortunately be bested by time.

MacKenzie’s dictionary format largely follows that created by Henning decades previous, although with a slightly more thorough transliteration system, more extensive attestations of course, and a rather complex system of abbreviations which were probably designed to shorten the overall length of the work in print.

From the sample page above, one can see just how much more extensive the full dictionary is than previous published materials such as Henning’s Fragment or the glossary to Mackenzie’s Qunya. These are the materials—in the first place, the printed draft of the completed entries—which form the basis for the Chorasmian dictionary project. The existing materials as well as the project’s methods will be surveyed in the next blogpost in this series.

Approximately concurrent with Henning’s late work on the Chorasmian lexicon in the 1960s, another effort was underway back in Germany. The German Turkologist (and Nazi cryptanalyst) Johannes Benzing (1913-2001) had begun editing the Chorasmian material preserved in the Muqaddimat al-Adab, primarily based on the manuscript Yusuf Aǧa 5010 held in Konya and published in facsimile by Togan in 1951. Benzing was not a scholar of Iranian languages, but was primarily interested in Chorasmian as a possible substrate language to the Turkic variety that became current in the region from the 13th century on (known as Khwarezmian Turkic or pre-Chagatay). His collection of this material published in 1968 (Das Chwaresmische Sprachmaterial einer Handschrift der “Muqaddimat al-Adab” von Zamaxšari) therefore ended up beset by problems.

The work is first of all something of an eclectic edition. Benzing combines his idiosyncratic transliteration of the Chorasmian glosses in the Konya manuscript with the Arabic and Persian equivalents and Latin translations as presented in the 1843 edition of the Muqaddima by Johannes G. Wetzstein (1815-1905). Benzing leans in the direction of a text edition, listing each word or form as it appears in order in the manuscript, rather than gathering them into an alphabetized grouping. As a guide to the Chorasmian lexicon, this means that the edition is rather hard to use. Moreover, as a guide to a single manuscript, its eclectic arrangement with ms pages mixed with the Wetzstein entry numbers means that it is confusing to peruse.

But the main problem was that Benzing made little effort to interpret the spellings of Chorasmian words, many of which were partially or totally unpointed. He used a complicated transliteration scheme to try to indicate what letters should have been pointed a certain way and which remained unclear. Lacking knowledge of Iranian languages and philology, he wasn’t able to resolve these problems through comparison and etymology, and ended up mis-presenting numerous words, many of which would have been decipherable, behind their unpointed Arabic skeletons, to someone with knowledge of Old or Middle Iranian. This is precisely what MacKenzie pointed out in a series of five (rather brutal) review articles on the book that he published from 1970 to 1972. His appraisal of the work can more or less be summed up by his statement that “Benzing can be faulted for many misreadings and, worse, wrong generalizations from them” (1970:541). These reviews are so extensive that reference to Chorasmian words preserved in the Muqaddima is simply not possible without them.

Credit for efforts to create a comprehensive, or at least extensive, glossary of the Chorasmian language are due, in the first place, to those medieval scribes who thought it useful to insert Chorasmian translations into copies of the Arabic Muqaddimat al-Adab (more on that, and on them, in due course!).

In modern times, the idea of compiling all the attested Chorasmian materials into a dictionary goes back to Walter B. Henning, the pioneer, together with Zeki Velidi Togan, of Chorasmian studies. Having begun working on Chorasmian in the 1930s, Henning had lost interest for a while due to the plagiarism of A.A. Freiman and did not publish anything on the language until the 1950s (Henning 1955:423).

It is unclear when Henning began returning to Chorasmian matters, but by the mid-1950s point, he had organized the words and phrases attested in the known manuscripts of the Muqaddima and Qunya (all discovered by Togan) into a card catalogue. At a 1954 conference, he announced “I have now compiled nearly a complete glossary which I hope to publish in the near future” (Henning 1956:43). Other obligations and projects intervened to delay the conversion of this card index into publishable dictionary form, however, and it was only in the year or so before his death in Jan. 1967 that Henning began to draft full entries.

At the time of his passing, approximately 260 entries had been completed, by hand as was his norm, filling 108 pages. These entries were based mainly on two manuscripts of the Muqaddima: the substantial glosses in the Yusuf Aǧa MS 5010 ms. (Togan 1951), and Hacı Beşir Aǧa MS 648 (at that time unpublished), as well as on some then-unknown manuscripts of the Qunya available to him.

Owing to the nature of the language, in which hundreds of words begin with a vowel written by means of the letter alif ʾ, these 260 entries only take us from ʾ- up to ʾkw-. Substantial as it is, this amounts to barely halfway through the first letter of the alphabet! The pages were sent after Henning’s death to David MacKenzie, who by then had established himself as the major authority on the language. The latter apparently had no access to other materials from Henning. MacKenzie published the entries with essentially no modification, adding only about 140 brief additional entries where warranted by cross-references in the completed part, as A Fragment of a Khwarezmian Dictionary in 1971. The handwritten pages were eventually returned to Henning’s Nachlass.

Besides the initial work in collecting, deciphering, and etymologizing Chorasmian words—though much of this was expanded and improved upon by MacKenzie later—one important thing about the Fragment is that Henning, perhaps inadvertently, established the lemma format which MacKenzie’s subsequent work on the Chorasmian dictionary would adhere to exactly. We’ll explain what we think are the downsides of this format in another blogpost; suffice it to say for now that it is not very comprehensible to someone not very familiar with Chorasmian literature.

Here ends Henning’s part in the story of a Chorasmian dictionary. Although his card catalogue is still extant, and may have some interesting notes and etymologies, it does not seem to have been used by any subsequent scholars. Moreover, the correct understanding of Chorasmian grammar and historical linguistics was developed much, much further in the following decades by Henning’s student David MacKenzie. It was he who took up the project of finishing a Chorasmian dictionary, laboring over it for the next three decades.