Lesson Text

Translation

Grammar

46 Introduction to Crimean Gothic

Crimean Gothic (CG) is the name given to the language thought to be thedying throes of the East Germanic branch of languages. All that remains ofthis language is some hundred words copied in a letter of the diplomatOgier Ghislain de Busbecq. The words so transmitted are similar enough tothose of the Biblical Gothic (BG) of Wulfila's translation that scholarsare in general agreement that the language in question is indeed Gothic,but there are some differences which suggest it may not be the latersurviving form of BG itself. It may have formed a separate member of theEast Germanic family, perhaps representing the language of another of themany Gothic tribal factions, but one whose literary records have notsurvived to the present day.

Busbecq himself was a highly educated man of Flanders, born in 1522 in thetown of Comines (Komen). His father was a nobleman and secured for him anexcellent education. He began at the University of Louvain in Brabant whenhe became thirteen, and then continued his education in Paris, Venice,Bologna, and Padua.

Busbecq eventually accepted an appointment in 1554 as an ambassador ofFerdinand I of Austria, who was later to become Holy Roman Emperor(1558-1564). In this appointment, Busbecq was charged with negotiating apeace treaty with the Ottoman Sultan Suleiman I, 'the Magnificent'. Inearly 1555 Busbecq travelled to the Ottoman empire, and returned to Viennathat summer to deliver a letter from Suleiman I to Charles V, then HolyRoman Emperor. He subsequently returned to Constantinople in 1556 andremained in the region for the next seven years as Ferdinand I succeeded tothe throne of the Holy Roman empire. He eventually negotiated a treatywith the Ottoman empire and returned to Vienna in 1562, then continuing toFrankfurt to present the treaty to Ferdinand.

Busbecq was subsequently knighted and continued to serve the imperialfamily in various political roles. In late 1592, then working in France,Busbecq took some time away from his service to visit his homeland. As hetravelled through Normandy, he was seized by a group of soldiers, but setfree the next day. Unfortunately, he quickly took ill and died on the 28thof October, 1592.

Busbecq had the reputation of being a man of many languages. Somebiographers claim he had a native command of Flemish, Latin, Italian,Spanish, French, German, and 'Slavic'. What is meant by the last is notexactly clear. Flemish was of course his native tongue, and his educationevidently provided for a firm grasp of Latin, as the letter shows. Hespent time at universities in Italy, where he was likely to have developeda command of Italian; and it is certain from surviving letters in Frenchthat he had a command of that language as well. Early in his career he wasattaché to Ferdinand's representative Pedro de Castilla, so it islikely that he acquired some proficiency in Spanish. We also know that, whilein Turkey, Busbecq had a habit of copying Greek and Latin inscriptions. Itis therefore likely that he was acquainted with Classical Greek, thoughevidently not with Modern Greek. He also explains some Turkish words inhis letters, so he may have had an acquaintance with this language as well.It is also quite possible that he had learned German while in service inVienna.

Sometime during his second and longer stay in Constantinople, Busbecq hadoccasion to meet with two envoys from the Crimea. One was evidently anative speaker of a Germanic language of the region, the other a nativeGreek speaker who had learned the Germanic language in question. From whatBusbecq says, it appears that the Germanic speaker had lost much of hisability with his native tongue due to constant interaction with Greekspeakers, and the Greek speaker evidently turned out to be the morecompetent in the Germanic language. From the letter it is unfortunatelyunclear as to which one of the gentlemen actually supplied the linguisticdata. Most scholars are of the opinion that it was the nativeGreek speaker, a source of much consternation to present-day linguistsbecause of the fact that his native tongue presumably interfered with hisability in the Germanic language. Neither is it clear what medium wasemployed in the interview: whether Busbecq interrogated the gentlemendirectly (what language would he use for this?), or by means ofinterpreters (presumably via Greek). The letter raises as many questionsas it answers.

Busbecq was certainly no newcomer to linguistic pursuits, and his letterclearly shows that he had an interest in investigating rare linguistic gemsshould occasion arise. Some of the modern scholarly literature has faultedBusbecq for his poor habits as a 'linguistic fieldworker', failing toemploy a phonetically accurate orthography, failing to be consistent inorthographic practice, failing to identify the recorded forms clearly,failing to note the circumstances and method of his linguistic interview --the list continues. But this should not be a surprise, since the data wascollected sometime between 1555 and 1562, centuries before the advent ofmodern linguistic practices. For a man who was actually employed tonegotiate a peace treaty between empires, we are lucky that he took thetime he did to produce what may be the last surviving record of the EastGermanic family of languages.

47 The Corpus of Crimean Gothic

The totality of primary texts concerning Crimean Gothic occupies roughlyfour manuscript pages. Of these, the bulk of the linguistic data properoccupies two pairs of columns, each column containing a word or, rarely,phrase of Crimean Gothic followed immediately by a Latin translation.There also appears a three-line song, orcantilena given space ofits own, but which lacks the Latin gloss one would have hoped for. Thereare a few other words of Crimean Gothic set within the Latin text itself.

The entire corpus is short enough that it is worth reproducing herein full. Below, where there is continuous Latin text, this is placed tothe left of its English translation. Where there are columns ofCrimean Gothic with accompanying Latin, these have been set in threecolumns, the first Crimean Gothic, the next Latin, and finally an Englishtranslation of the Latin. Thecantilena has no accompanying Latintranslation, and so there is no English translation. The Englishtranslation is intended to be close to the Latin, word for word ifpossible, while at the same time making decent sense. No attempt is madeat polished English prose.

In the course of subsequent copying and typesetting, it is possiblethat errors may have crept into the text. For example, some suggestthat in setting one of the columns of the word lists, some letters may havefallen out and been incorrectly replaced. In this way, CGThurnreceived the finaln which properly belonged to CGKor, andCGFisct received the finalt which properly belonged to CGHoef. The proper words should thus be: CG*Thur, *Korn,*Fisc, *Hoeft. Others also suggest that the glossEgo dico'I say' for CGIch malthata is an error forEgo dixi 'Isaid'; similarly the glossVoluntas 'wish' for CGBorrotschis a misprint forVoluptas 'pleasure'. Stearns (1978) proposes thefollowing emendations:

48 Phonology

The phonology of Crimean Gothic is both the primary avenue ofinvestigation into the language and the greatest source ofconsternation. The difficulty stems from the manner in which thelinguistic data was gathered and transmitted. Interpretation of thephonology of the corpus as received suffers from three basic problems:

1. the informant: Whichever of the two gentlemen from the Crimea provided the linguistic data, he was either a Greek by birth, or a native speaker of Crimean Gothic more competent in Greek than in his own native tongue. In either scenario, one may be certain that the pronunciation of Crimean Gothic words was colored by some degree of interference from Greek phonology of the region and period.
2. the fieldworker: Busbecq himself was known to be competent in several languages. Nevertheless, his native language presumably colored his understanding of the sounds he heard. In addition, as is clear from the letter, Busbecq had already formed preconceptions as to the nature of the language and its possible origin. This, coupled with his competence in other languages, perhaps resulted in his amendment of certain words to conform with those of languages he knew and to which he believed the language of the informants to be related.
3. the transmission: Busbecq's original letter appears to have suffered from both scribal emendation and error, introducing a certain degree of uncertainty in some instances as to what it was that Busbecq intially wrote.

Corrections which may be due to the last point have been touched upon inthe preceding section. Below we discuss issues related to the informant'slinguistic skills and Busbecq's own skills and practices.

48.1 Informant Interference

Most scholars interpret Busbecq's letter as suggesting that it was in factthe native Greek speaker who was the primary source of the Crimean Gothicmaterial. If so, it is especially important to understand the sounds ofhis native language, since we expect that he will be more apt to preservethose phonological distinctions of Crimean Gothic which are paralleled inhis own language than those which are not distinguished in his nativetongue. What then was the phonological structure of the Greek which theinformant spoke?

Stearns (1978) suggests the following line of reasoning. We may presumethat, because of the Greek speaker's familiarity with the region, he mayhave been born in or near the Crimea and probably spoke the Greek found inthat region at that time. We may call this languageCrimean Greek ofthe 16th century. Inasmuch as there is scant evidence for the exact natureof this variety of Greek during this period, we must look for some indirectmethod to discover its characteristics.

It turns out that in the late 18th century the Tartars persecuted theChristians of the Crimea, and these Christians requested from Catherine IIpermission to settle in Russian territory. This was granted, and many ofthem, among them a large number of Greek speakers, eventually settled inthe city of Mariupol (Zhdanov) in the Ukraine, beside the sea of Azov. Thedialect of these Greek speakers, which we may termMariupol Greek(MGk), survived into the 20th century. As it ultimately stems from theGreek spoken by natives of the Crimea, we may tentatively equate MariupolGreek with Crimean Greek (CGk). We then further assume that MGk, henceCGk, has undergone little change in the centuries between the gathering ofcurrent linguistic data and the time period of Busbecq's visit to theCrimea. These assumptions are all speculative, but provide at least somestarting point for an assessment of the pertinent phonology.

From studies of Mariupol Greek, we find that this dialect does in factdiffer from standard Modern Greek in some ways. The consonants of MariupolGreek, and thus Crimean Greek, are given in the chart below:

Here the symbolþ is used rather than the more usual theta inkeeping with our focus on Gothic. There does not appear to be anydistinction between long and short consonants in MGk. In addition, MGkpreserves the distinction of voiced and voiceless consonants, andword-final consonants do not undergo devoicing.

The vowels of Mariupol Greek, and so Crimean Greek, are shown in thefollowing chart:

A period immediately following a vowel indicates a tense pronunciation.MGk apparently does not phonemically distinguish vowel length. Several ofthe phonemes had allophonic variants, indicated in square brackets in thecolumn following the phoneme. In particular, the phoneme /e/ hadphonetic variants including palatal on- and off-glides. MGk [je]appears for /e/ in initial position, so that [jéma]corresponds to Modern Greek [éma]; likewise MGk [ej] occursfor /e/ in one-syllable words, so that MGk [psejs] corresponds toModern Greek [pses].

We may now consider a few examples of how our knowledge of MGk phonology,and hence CGk phonology, provides us with a method by which to interpretthe Crimean Gothic corpus.

We are in a position to determine the reflex of PGmc*e in CrimeanGothic. Consider the word Busbecq transcribes asseis. Thediphthong may be Busbecq's rendering of what the Greek-speaking informantpronounced as [sejs]. This mid-front vowel with palatal off-glide,however, is the Greek informant's allophone for /e/ in monosyllabicwords. Hence we expect the informant's [sejs] to correspond to CG/ses/. Thus CG /e/ corresponds to PGmc*e. Compare BGsaíhs, ONsex, OHGsehs.

Consider also the initial the initialie of the words writtenIel, Ies. Some scholars have takenie to represent long[ī]. This is possible, but the allophones of CGk /e/ providean alternate explanation. Since the CGie appears in initialposition, we might expect this to represent a palatal on-glide introducedby the informant. Thus the informant's [jes] may represent CG/es/. Compare RuniceR, OHGer from PGmc*es. Inaddition, since CGk does not distinguish vowel length, we may suspect thatCGk /e/ represents a possible CG /ē/. Then we find in initialposition [je] < CG /ē/, which is perhaps the CG reflex of thePGmc diphthong*ai. Then CGIel = /ēl/, for which compare BGháils and OHGheil 'well', with loss of initialh-.

As regards consonants, we may recall that final voiced consonants in CGkare not devoiced. In this context we note the following correspondences:

The final consonants in the CG reflexes are clearly devoiced, and thisdevoicing cannot be presumed to be a result of interference from theinformant's native language. Thus we find that devoicing of final PGmcconsonants is a feature of CG itself.

The initialp in CGPlut 'blood' poses an interestingchallenge. In this position we expect the voiced plosiveb, as in BGblōþ, since other words give us no indication that an initial PGmc*b should be treated specially. Compare the following: CGBroe 'bread' (cf. OEbrēad, OHGbrōt), CGBruder 'brother' (cf. BGbrōþar), CGBoga 'bow' (cf.ONbogi, OS, OHGbogo), CGBars 'beard' (cf. OEbeard, OHGbart), CGBrunna 'fountain' (cf. BGbrunna). Given thatPlut is the sole exception, it is quitepossible that thep is simply a misprint. If not, we must seekanother explanation. Since the informant's native tongue distinguishesvoiced and voiceless consonants in all positions, we must assume that thisrepresents a true CG [p]. It is then possible that CG had twoallophones for /b/, one strongly voiced [b], and one weaklyvoiced [p]. The latter occurred in initial position when followed by/l/.

A similar explanation may apply to the unexpected occurrence oftinTag 'day' andth inThurn 'door'. Thesepresumably represent the same voiceless plosive [t], where we wouldotherwise expect [d] from PGmc*d, cf. BGdags, OSduri, OHGturi. The PGmc*d in fact occurs as CG [d]in other places:Handa (cf. BGhandus, ONhǫnd, OShand) and CGfyder (cf. BGfidwor, fidur-). As with CG/b/, we may assume that CG /d/ < PGmcd had twoallophones. One was lightly voiced and equivalent to [t], occurringfor initial /d/; the other was voiced, and occurred for /d/between sonorants, i.e. resonants and vowels.

An unfortunate property of the CGk phonemic inventory is that it lacks theglottal fricative /h/, which we would expect to develop at some stagein the history of CG from PGmc*x if the other Germanic languages areany indication. If such a phoneme was in fact present in CG, we wouldexpect the Greek-speaking informant to substitute either [x] orzero. Unfortunately we find the confusing situation in which the CGcorpus showsboth the expectedh in certain words andzero inothers:

This could be an idiosyncracy of the informant, employing [x] in somewords,zero in others. Perhaps more likely, however, is that PGmc*x was in fact lost in CG, and that it was Busbecq's own knowledge ofDutch and German orthography which lead him to write anh in wordswhich were in fact pronounced without [x]. Note that the words withh,Hus, Hoef, Handa, appear in the list of those wordswhich Busbecq believed to be quite similar to his own, whileAnoandIeltsch appear in the list of words which sounded fairlydissimilar to the Germanic languages with which he was familiar.

48.2 Fieldworker Interference

Busbecq was fluent in a number of languages, and familiar with stillothers. Therefore the number of different language sounds with which hewas familar must have been considerable. Some scholars have suggested,however, that Busbecq was not entirely familiar with [þ], hearingthe sound instead as [ts], and therefore transcribingtz.Consider the following examples: CGGoltz 'gold' (cf. BGgulþ),CGStatz 'land' (cf. BG dat. sg.staþa, OSstað, OHGstad,stado 'shore'), and CGTzo 'thou' (cf. BGþu,OEþu, ONþú).

However this same familiarity with a number of languages may have been ahindrance in the area of orthography. It is clear from the letter thatBusbecq already had certain preconceived notions of the relationship thelanguage under investigation had to other Germanic languages. It appearsthat in certain respects he allowed his transcription of the informant'swords to be colored by the orthographic conventions of other languages,particularly Netherlandic and German writing conventions.

Busbecq's transcriptionie may have represented the long vowelī, since this was the practice in Middle Netherlandic and EarlyNew High German. This is likely the intention in the transcriptionMeira 'ant' andSchieten 'shoot', which mimic MiddleNetherlandicmiere andschieten. Given the lack ofphonemic distinction between long and short vowels in CGk, however, theinformant likely pronounced a short vowel. We have also discussed abovehow theie found in CGIes,Iel, andIeltsch likely represents the palatal on-glide of the informant'sinitial [je], and not the monophthong [ī]. If this is true,it shows that Busbecq suffered some inconsistency in employing letters fortranscription.

As discussed in the preceding subsection,h is particularlyproblematic. Its apparent inconsistency on Proto-Germanic grounds islikely the result of Busbecq's own orthographic habits. This letterappeared in the Middle Netherlandic wordshuus, hovet, hant and inEarly New High Germanhaus, haupt, hand(t). What is more, Busbecq wasaccustomed to writing this letter in initial position before a vowel, sincein his native Flemish [h] was lost in just this position, thoughinconsistently retained in the writing. This strengthens the argumentoutlined above, that [h] was lost in initial position in CrimeanGothic, but Busbecq inserted the letter in the writing of those wordsclosest to those of the Germanic languages with which he was familiar.

Scholarly opinion varies on what the transcriptiontz actuallyrepresents. This digraph only occurs in places where one expects PGmc*þ, though*þ also appears ast,th, andd. It could of course be a misprint, but the correspondences with*þ are not chance, so that this is not likely. As mentioned above,Busbecq may have misunderstood [þ] as [ts] and thus writtentz. On the other hand, he may have been attempting to transcribethe voiced fricative [ð]. But since the informant would havedistinguished [ð] and [þ], he may have pronounced [ð]in CGTzo 'thou', but it is less likely inGoltz andStatz. In the end, Busbecq likely heard CG [þ] correctly,but was confounded in how to transcribe it. Perhapstz was hissolution to this dilemma.

49 Morphology

Unfortunately Busbecq has left us with few complete phrases with which toanalyze syntactic relations and the morphology which marks those relations.Instead what we have are for the most past isolated forms devoid ofcontext. This leaves any identification of morphological form tentative atbest.

49.1 Substantives

Scholars generally agree that Crimean Gothic distinguished gender, case,and number, though it is difficult to tell from the data available just towhat degree. If we are correct in interpreting boths andsch as representations of CG /s/, then several words may showthat CG retained the PGmc nominative singular masculine ending-z asCG /s/. Consider the following examples: CGVVintch 'wind'(if a misprint for*VVintsch, cf. BGwinds), CGFers'man' (cf. BGfaírƕus), CGRintsch 'mountain', CHBorrotsch 'desire'. The CG phraseKnauen tag may displaythe ending of the accusative masculine singular. Compare the accusativesingular BGdag, ONdag, OEdæg, OSdag, OHGtag.Only CGOeghene 'eyes' has a plural gloss. This perhaps representsa nominative or accusative plural form, as with BGáugōna, ONaugo, OEēagen, OSōgun, OHGougun.

The phrases preserve some of the CG pronouns. In particular we find CGIch 'I' (cf. BGik, ONek, OEic, OSik, OHGih), as well as CGTzo 'thou' (cf. BGþu, ONþú,OEþū, OSþū, OHGdū). We also have the masculinethird person pronoun CGIes 'he' (cf. BGis, ONer). Theletter may also provide examples of the accusative singular of a thirdperson neuter enclitic pronoun CG-(a)ta, if it is correct tointerpret phrases likeIch malthata as 'I sayit'. Compare BGita, OSit, OHGiz. A different interpretation takes thepronoun to be a demonstrative CGthata, so thatIch malthata would mean 'I saythat'. Compare BGþata,ONþat, OEþæt, OSthat, OHSdaz.

Among adjectives, CGIeltsch 'healthy' may employ the transcriptionsch to reflect a strong nominative singular masculine ending CG/s/, and is thus equivalent to BGháils. The adjectives CGAlt'old' andTelich 'foolish', lacking either transcriptionsorsch, may show strong neuter nominative/accusative singularendings, similar to BGblind. The pronominal alternate ending for thisform, akin to BGblindata, may also be in evidence in the transcriptions-ta and-tha of the following adjectives: CGAtochta'bad' (Lat.Malum), CGVVichtgata 'white' (perhaps misprintedfor*VVitgata; Lat.Album), CGGadeltha 'beautiful'(Lat.Pulchrum). Each of the Latin glosses shows the neuternominative/accusative singular form. The adjective CGKnauen 'good'likely represents the strong masculine accusative singular ending, akin toBGblindan.

49.2 Verbs

Busbecq's data provides us with a number of infinitive verb forms, allending in CG-en: CGSchieten 'to shoot an arrow' (writtenin the manuscript with a lift over thee, rather than thenitself; Lat.Mittere sagittam), CGSchlipen 'to sleep' (Lat.Dormire), CGKommen 'to come' (Lat.Venire), CGSinghen 'to sing' (Lat.Canere), etc.

The two verb forms CGVvarthata andVarthata (both likelyrepresenting*VVarthata) are glossed with the Latin perfectindicative, first and second person singular, respectively. These could bedisyllabic preterite forms, analogous to the plural forms of BG, e.g.nasidēdum 'we saved'; this would then show that such disyllabic formswere extended to the singular in CG. However, such a conclusion might beobviated by reading the final sequence-(a)ta or-thata asneuter accusative pronouns. In this regard, the manuscript glosses thephrase CGIchmalthata as 'I say' (Lat.Ego dico).Since this form appears the same as the preceding, and those are glossed aspreterites, most scholars take Lat.Ego dico 'I say' to be amisprint forEgo dixi 'I said'. However, reading the final twosyllables as a neuter direct object allow the manuscript reading to stand,and perhaps this shows the first person singular, present indicativeactive.

The phrase CGKilemschkop 'drink up your cup' is glossed as a Latinimperative. Reading the final syllable-kop as the CG word for'cup', then that leavesKilemsch as a second person singularimperative. The phrase CGIel vburt 'may it be well' is glossedwith the Latin present subjunctive. The form CGvburt, perhaps for*vvurt, may be a third person singular past subjunctive (optative):cf. OHGwurti fromwerdan.

50 Syntax

There are a meagre few phrases in the Crimean Gothic corpus from which wecan even hope to begin a study of the language's syntax. Most regrettably,thecantilena (song)

does not even contain a gloss. This is a particular blow, because this isthe only possible passage which may have contained multiple clauses, or atleast one extended clause. What we are left with, then, are the followingphrases:

If we take CG-thata to be a pronoun, then we may make thefollowing analysis:

Assuming the informant made no syntactic errors, then it appears that CG istentatively a Subject-Verb-Object (SVO) language. Such an identificationmust be qualified, however, by the structure of the phrase CGKnauentag '(I bid you) good day'. This presumably showsan attributive adjective preceding the noun it modifies, which is a featuregenerally associated with SOV languages. If similar to other Germaniclanguages of a similar time period, Crimean Gothic was likely an SVOlanguage which retained remnants of an earlier period of SOV word order.

If the phrase CGKilemschkop 'drink up your cup' is actually to beanalyzed as an imperativeKilemsch and direct object-kop,then it appears imperative forms occupy first position in an utterance.This order need not hold, however, if the verb is subjunctive: CGIelvburt 'may it be well' appears to be verb final.

previous lesson   |  next lesson