An old prussian grammar: the phonology and morphology of the three catechisms 9780271011714, 0271011718

1,233 78 12MB

English Pages [184] Year 1975

Report DMCA / Copyright

DOWNLOAD FILE

Polecaj historie

An old prussian grammar: the phonology and morphology of the three catechisms
 9780271011714, 0271011718

Table of contents :
INTRODUCTION
PHONOLOGY
THE NOUN
THE ADJECTIVE
THE NUMERAL
THE ADVERB
THE PRONOUN
THE PREFIX
THE CONJUNCTION
THE NONFINITE FORMS OF THE VERB
WORD INDEX
FOOTNOTES
REFERENCES

Citation preview

AN OLD PRUSSIAN GRAMMAR: The Phonology and Morphology of the Three Catechisms The tentative transcription of the Old Prussian text is: Kayle rekyse thoneaw labonache thewelyse, Eg koyte poyte nykoyte pen(n)ega doyte. The tentative translation (by Valdis Zeps) is: Hello, Sir! It isn''t the good fellow (or: You aren't the good fellow) if you want to drink (and) don't want to pay money. The Old Prussian passage appears on folio 63ra of MS Basel, Oeffentliche Bibliothek der UniversitM.t Basel, F.V.2. The Old Prussian text immediately follows the Questiones Super Quattuor Libras Methororum of Nicole Oresme, dated 1369, and immediately precedes--and was clearly written prior to--the undated Registrum g_uartium l!] librorum Methororum, which lists all the questions treated by Oresme. The text was discovered by Stephen C. McCluskey.

William R. Schmalstieg

The Pennsyh:ania State University Press University Park and London

To My Family Emily, Linda, and Roxanne And in Memory of My Parents Dorothy R. Schmalstieg And

John W. Schmalstieg Copyright © 1974 The Pennsylvania State University All rights reserved Printed in the United States of America

.. I

Library of Congress Cataloging in Publication Data Schmalstieg, William R. An Old Prussian Grammar. Includes bibliographical references. 1. Prussian language--Grammar. I. Title 74-8736 491 1 .91 PG8204.S3 ISBN 0-271-01171-8

TABLE OF CONTENTS PREFACE .....................................ix INTRODUCTION ......•..........................1 MAJOR DOCUMENTS ..............................4 PHONOLOGY Orthography................................8 Vo,.tTels. � ..................................16 Stress....................................22 Consonants................................26 THE NOUN ....................................29 THE ADJECTIVE ...............................82 THE NUMERAL Cardinal nmnerals........................106 Ordinal nmnerals.........................108 THE ADVERB.................................112 THE PRONOUN ................................124 Third person personal pronoun............125 Possessive pronouns......................126 Reflexive pronoun ........................130 Demonstrative pronouns ...................132 Interrogative-relative pronoun...........135 Demonstrative pronoun adjective ..........136 Non-gendered personal pronouns...........137 Encliti c pronoun.........................140 THE PREFIX.................................141 THE PREPOSITION............................144 THE CONJUNCTION ............................ 147 THE VERB The personal endings ..................... 148 The future tense ........................•153 The imperative...........................153 The preterit tense .......................155 The thematic vowel.......................164 The present tense........................165 THE NON-FINITE FORMS' OF THE VERB The infinitive ...........................222 The supine..................•............222 The present participles ......•....•......223 The past participles.....................224 THE PERIPHRASTIC TENSES .................... 227 WORD INDEX•.•.•..•....... , ........•.. , .....229 FOOTNOTE·s •.•.....•.... , .•..................304 REFERENCES. � ...............................346

PREFACE The assumption behind this work is that there was indeed a system to the Old Prussian language, i.e. there was a limited number of distinctive phonological units, a definite and limited number of morphological elements and a syntax which was not much different from the syntax of the other Baltic languages. With some exceptions I have established the forms of Old Prussian merely by means of a comparison with the extant Lithuanian and Latvian. The orthography of Old Prussian is so confusing that there are doubtless numerous errors in my guesses. On the other hand it seems preferable to interpret the evidence rather than to accept it at face value. In many cases I am frankly guessing, but I would defend an interpreted guess which relates a feature of Old Prussian to an existing Baltic language to an uninterpreted guess which relates a feature of Old Prussian to a feature of Tokharian, Sanskrit or Greek. 1 Prof. Anthony Salys recounts the story told to him by Prof. Georg Gerullis. The latter lived in Latvia for some time and had occasion to ask the Germans the Latvian name of the city Frauenburg. The Germans consistently told Prof. Gerullis that the name of the city was Salde in Latvian, even though the real Latvian name was Saldus. It is interesting to note that the German rendering was completely consistent, but consistently wrong� Here it is my pleasant duty to thank the following scholars who were kind enough to take time from their busy personal schedules to spend many hours discussing in detail with me many of the problems of Old Prussian. Although many others have also contributed, I should like to thank particularly Prof. A. Girdenis (Vilnius), Prof. B. Jegers (Northern Illinois); Prof. Terje Mathiassen (Oslo), Prof. Vytautas Maziulis (Vilnius), Prof. A. Nepokupnyj (Kiev), Prof. Christian S. Stang (Oslo), Prof. V. N. Toporov (Moscow) and Prof. Pavel Trost (Prague).

I should also like to express my gratitude to the .American Council of Learned Societies, the .American Philosophical Society· and The Pennsylvania State University (Central Fund for Research). All of the latter organizations contributed greatly to my support during the preparation of this grammar and enabled me to travel to Europe to consult with colleagues and work in European libraries.

I I

I I

,I

INTRODUCTION

The Baltic languages form a group within the Inda-European language family. The extant Baltic languages, Latvian (also known as Lettish), and Lithuanian are classified as East Baltic languages whereas the extinct Old Prussian is classified as a West Baltic language. In earlier times the speakers of Baltic languages were referred to as Aistians, although it is not known whether this term denoted all the Balts or just the Old Prussians. The Aistian peoples (Aestiorum gentes) first appear on the historical scene in chapter XLV of Cornelius Tacitus' Germania. Tacitus wrote, ''Passing then to the east along the shore of the Suebic (Baltic - WRS) sea, we find the tribes of the Aestii, who have the same observances and general appearance as the Suebi, while their language is more like the British tongue. They worship the Mother of the Gods. As the symbol of their religion they carry figures of boars. They believe that, without weapons or protection of any other kind, this charm preserves a devotee of the goddess from harm even among his enemies. They rarely use iron weapons, far more frequently clubs. They labour at the cultivation of crops and fruit trees with a perseverance which is in contrast with the usual indolence of the Germans. They also scour the sea, and are the only people who gather amber. They themselves call it glesum and they find i� in the shallow water or actually on the shore. Like barbarians they have never discovered or inquired by what natural process it is produced." (Translation from Fyfe, 1908, 117.) If these Aistians are Balts or even Old Prussians, then, of course, the comparison with the British language (a Celtic tongue) is not correct except inasmuch as both Celtic and Baltic are Inda-European languages. Several decades after Tacitus the Greek scholar, Claudius Ptolomaeus, in his Geographia (book III, chapter V) mentions two tribes, the Galindai and the Soudinoi, which are commonly

assumed to be the Old Prussian tribes known otherwise as the Galindians and the Sudovians respectively. Cassiodorus, the secretary of the Ostrogothic emperor Theodoric(ca. 454-526), wrote a letter of thanks(some time between 523 and 526) to the Hestis(dat. pl. 'Aistians'?) in which the latter are thanked for gifts of amber which had been sent to the emperor. In the middle of the sixth century the Gothic historian Jordanes relates that the Aistians (Aestii) live near the Baltic sea not far from the mouth of the Vistula. Einhard(ca. 770-840), the Frankish biographer of Charlemagne, mentions Vita Karoli briefly the Aistians(Aistii) in his �� a.d. Magni written around 830 In the introduction to the Universal History of Orosius(which Alfred the Great caused to be translated into Anglo-Saxon) we read of the account of the traveler Wulfstan who apparently' mentions the Aistians in such phrases as 'to Est-um, mid Est-um, East-land, of East-lande.' From other place names which are also mentioned, = e.g. Est-mere( = Aismares), the Ilfing( Elbing) river, Wislemuda( = mouth of the Vistula), it appears that Wulfstar1may have meant the Old Prussians by the name Aistians. Since this is true for Wulfstan it is sometimes assumed that in general the name Aistians, when used by Tacitus, Cassiodorus and Einhard may have referred to the Old Prussians(see Mafiulis, 1966, 13). The word Prussian is apparently found for the first time in the form Bruzi in a geographical description dating from the ninth ·century. Later u in about 956 the name B rus is mentioned in an ler Ibrahim ibn Ja'kub trave account by the Spanish (BUrus) have their ians Pruss who states, "...the tic sea). They (Bal settlements near the ocean not know the do and have their own language ors." languages of their neighb From the Monumenta Poloniae historica ed. by August Bielowski, I, Warsaw, 1960, Ma�iulis (1966, 14) quotes the forms Pruzze, Pruze, Pruzzorum, Pruzziae, Pruzis, Pruzorum, Pruzos, Prucorum,

3 Pruciam, etc. The word Pruzzi is mentioned by Adam of Bremen(eleventh century) and the Russian Nestor chronicle mentions the Prousi. After the tenth century the name became common in many forms including the variants Borussus(= Prussian) and Borussia(= Prussia). The Old Russian chronicles mention a tribe by the name of Goljad' in the twelfth century in the Smolensk area. It is commonly thought that these are Galindians, a tribe whose name is probably _ derived from Baltic galas, cf. Lith. galas 'end.' The Sudovian tribe is frequently equated with the Jatvingians in early texts and we find the latter mentioned in the Hypatian chr�nicle: Ide �olodi�i�n na Jatvjagy i_ vzja zemlju ixn Volodimir went against(attacked) the Jatvingians and took their land." According to the Cronicon terre Prussie (Chronicle of the Prussian land) by Peter of Dusburg(fourteenth century) the Prussian land is divided into eleven parts, the first part of which is Colmensis and Lubovia which was "...almost destroyed already before the arrival of the brothers of the Teutonic order" (ante introitum fratrum domus Theutonice). The other parts are Pomesania, Pogesania, Warmia, Nattangia, Sarabia, Nadrowia, Scalowia, Sudowia, Galindia, Bartha (also Plicka Bartha). Under German pressure the Old Prussians either died out or were Germanized. There were probably only a few speakers or none at all by the year 1700. On the title page of the II Catechism which had been located in the St. Petersburg Public_ Library there was the statement: Diese alte Preusnische Sprache ist nuhnmehr gantz und gar vergangen worden. Anno 1677 ein einziger alter Mann auf der Curis� Nahrung wonend, der sie noch geko� gestorben, doch sollen noch solche daselbst sein "This Old Prussian language has completely disappeared. In 1677 a single old man, who lived in the Curonian Neringa and who still knew the language, died, but there are still said to be some there." (See Mafiulis, 1966, 25.)

MAJOR DOCUMENTS

In addition to the fragments, place names and personal names there are five major documents in Old Prussian, viz. the Elbing vocabulary, Simon Grunau's vocabulary, the I, II, and III Catechisms. The III Catechism is also cormnonly known as the Enchiridion.2 The oldest Old Prussian manuscript still extant is the Elbing vocabulary which forms a part of the so-called Codex Newnannianus. This codex itself dates from around 1400 and is apparently a copy of the original, which was probably composed at the beginning of the fourteenth or the end of the thirteenth century. (See Trautmann, 1910, XXIV.) According to Marchand (1970, 112), "The Elbing vocabulary presents the usual kind of conceptual dictionary found in medieval Latin and German manuscripts. Its only unusual features are those which bespeak the Prussian condition, e.g\, words for 'sled,' 'fire-hole,' etc. It is most certainly not, as Berneker and, following him, Trautmann would have us believe, drawn up for legal purposes." The Elbing vocabulary contains 802 words in German in the left hand colwnn with the Old Prussian equivalents to the right. Between the years 1517-1526 the monk Simon Grunau wrote his Prussian chronicle in which he included approximately 100 Old Prussian words. The words are in no particular order and among the Old Prussian words are apparently some Polish and Lithuanian forms. Many word.s appear without an ending or with a very distorted ending. (See Endzelins, 1944, 13.) Previously seven copies of this manuscript were known, but on the 28th of January, 1970 an eighth copy was found in Helsinki. (See Valentin Kiparsky, 1970, 219.) Since I shall devote separate studies to both the Elbing vocabulary and Simon Grunau's vocabulary I shall say no more about them here. The most important documents in Old Prussian are the three catechisms which are the subject of this study. All three catechisms are printed and

5 in all make up 66 pages of Old Prussian text not including those pages which are a parallel version of the text in German. It appears from the introductory material in the catechisms that there had existed other religious texts in Old Prussian ' but the whereabouts of these is unknown. It is assumed that the catechisms are written in the Samlandian dialect of Old Prussian since we find a statement to that effect in the introduction to the I and II Catechisms, to wit: "Derwegen auch dieser alte und gemeyn Catechismus ist jnn undeUdscher Prettssnischer sprach wie die uff Samland sonderlich am rechten prettssnischen orth und strich gebreUchlich aus F. D. unsers gnedigsten herrn beuelch jn druck verordnet." (See Trautmann, 1910, 1 and 8; Ma�iulis, 1966, 82 and 98.) But those features which are taken as dialect features (Trautm�n, 1910, XXI), such as the supposed passage of e to T and o to u ( after labials) in the catechis�s may well-refl;ct either vagaries of German orthography or phonological changes in statu nascendi. One may asswne that Old Prussian, like all other languages, had dialects, but the orthographic evidence of the Old Prussian monuments does not really give as good evidence as is commonly supposed.3 The I Catechism was published in the press of Hans Weinreich in Konigsberg (now Kaliningrad) in 1545 in an edition of approximately 197 copies. This edition, published in smaller and larger Schwabach script, has 16 unnumbered pages. In addition to the title page and two pages of introduction (all in German), there are 12 pages of parallel German and Old Prussian texts. The German text which served as a basis for this translation is Luther's Smaller Catechism published in 1531 in Wittenberg. The II Catechism also was published in Konigsberg in 1545 in the press of Hans Weinreich. A total of 192 copies were printed. The introduct�on to the II Catechism claims that it is a corrected version, presumably a corrected

6

version of the I Catechism. (See Mafiulis, 1966, 37, 99.) Thus, for example, in the II Catechism the word ryeky 'kingdom' is used instead of laeims (used in the I Catechism), an apparent mistranslation of German Reich. The word laeims probably meant 'reich' in the sense of English 'rich,' cf. Lith. laime 'luck, good fortune.' In the II Catechism we find the word reddi 'false' instead of the German word falsch which is used in the Old Prussian text of the I Catechism. On the other hand the II Catechism is not always an improvement on the I Catechism. Thus in place of the I Catechism word menentwey 'to mention' used in the second commandment to translate German fiiren we find in the II Catechism the word westwey which probably means 'to lead,' cf. Lith. vesti 'to lead.' Thus the thought of the German: Du salt den namen Gottes nicht unniitzlich fiiren 'Thou shalt not take the name of God in vain' is better translated by menentwey 'to mention' than by westwey 'to lead.' The III Catechism or Enchiridion was translated by Abel Will, who used a German original based on the Small Catechism or Enchiridion published in 1543. Abel Will was from the parish which is known in Lithuanian as Pabe6iai and is rendered in German variously as Pobeten, Pobethen, Pubeten and in the Enchiridion itself as Bobeten. (See Mafiulis, 1966, 136.) Abel Will's predecessor in the same parish was Michael Will and it is known that in 1527 the local mill was assigned to a man named Matthew Will. All this is used as evidence for th� assumption that Abel Will did indeed come from· that parish. Abel Will's friend and protector at the court of Duke Albert (Albrecht) of Prussia was Johann Funck (1518-1566), the son-in-law of Andreas Osiander (1498-1552), who taught that justification by faith was not an imputation, but rather an infusion of the essential righteousness or divine nature of Christ. Funck, the leader of the heretical Osiandrist movement, was put to death in 1566 and Abel Will was imprisoned.

7 Having gone blind Abel Will took up residence at one of the Konigsberg hospitals in 1575 where he died sh�rtly afterwards, perhaps in the same year. (See Maziulis, 1966, 39.) Abel Will apparently did not know enough Old Prussian to translate the catechism by himself. When an official from Grlinenhof took Will's translator, Paul Megott, for statute labor, Will wrote a letter to Funck (dated 26 July 1554) complaining that he would be unable to continue with the translation without the help of a translator. Leskien (1876, 60) says that this letter is proof that Will either knew very little Old Prussian or none at all. Trautmann (1910, XXX-XXXI), defending Will, says that the phonetic rendering of the text is excellent. Trautmann points to the fact that Will was apparently sometimes able to distinguish the intonations s�mething which August Schleicher, who spent his life working on linguistics, was unable to do for Lithuanian. I would counter Trautmann however by pointing out that most non-native s;eakers ' require a certain length of time before they begin to hear distinctive intonations in languages they may be learning. One may compare the very short length of time during which Schleicher was in contact with spoken Lithuanian with the much more extensive length of time during which Will had the opportunity to hear Old Prussian spoken. Whether Will ever got his translator back or not we do not kriow. In any case the III Catechism or the Enchiridion was published in 1561 in the press of Johann Daubman, although the number of published copies remains unknown. It should be mentioned here that David Robinson, 1972, has pointed out the existence of one copy each of the first two editions of the Old Prussian catechism in the Niedersachsisches Staatsarchiv and one copy of the first catechism in the British Museum.

PHONOLOGY Orthography Quite possibly one of the most serious mistakes in the analysis of the Old Prussian documents is the assumption of a high degree of phonetic (or phonemic) accuracy in the transcription. This assumption of phonetic or phonemic accuracy seems to need very little refutation. Even highly trained phoneticians can make serious mistakes in transcription. And the German scribes copying down Old Prussian weren't even firnt year students of phonetics at a university. It seems likely then that either the palatalization was not perceived by the German ear or else that it was perceived as a consonant plus a following /j/. This undoubtedly explains such spelling vacillations which occur either with or without a following' i. Thus, for example, the word for 'waits' is -;pelled either gieidi (where the i following the _g_ denotes palatalization) or geide (where there is no i following the initial£). Likewise we encounter the word for 'fi�h' spelled penckts (palatalization unmarked) vs. pyienkts (palatalization marked). I propose that unpalatalized consonants in Old Prussian were labialized by following non­ front vowels. Following EndzelTns (1943, 18) I have supposed the existence of palatalized consonants in Old Prussian. I assume also that this palatalization was phonemic as in Lithuanian and Russian. An interesting feature of languages which have phonemic palatalization is that such languages frequently have non­ phonemic labialization also. One can compare, for example, the Russian situation where labial consonants followed by certain vowels are labialized, cf. Russ. � 'to wash', most 'bridge,' etc. This phonetic labialization carries no phonemic burden, but it is nevertheless noted by non-native speakers.

Thus Andre Martinet says (1955, 356)," ...un Fran�ais qui entend les mots byl et most pourrait @tre tente de transcrire bwil et mwost." This phonetic labialization of consonants was misinterpreted by various German-speaking scribes who tried to mark it in a rather haphazard fashion by writing -u-, -ua- -wa­ or -£_- where one might not exp;ct thes; lett;rs . otherwise (Schmalstieg, 1968 c, 189-190). 5 The follow�ng v cillations in orthography � are to be explained in this way: (The words given below �re in the Enchiridion unless they are marked with I or II which denote the I and II Catechisms respectively.) Labialization unmarked

Labialization marked

Norn. sg. masc. kawijds (interrogative and relative pronoun)

kuwijds

Norn. sg. fem. aucktimmisikai 'authority'

aucktimmisk;

A�c. sg. fem. mergan 'maiden' Ace. sg. fem. prabutskan 'eternal' Ace. sg. fem. crixtianiskan 'Christian' Ac:,c. sg. fem. peroniskan 'community, parish' (in this example the word-final syllables are to be compared·' the .£ may denote a labialized

-

/a/)

mergwan (I and II) prabitscun (I) prabusquan (II) crixtianiskun krichstianisquan perronisquan (II) perroniscon (I)

9

11

10 Ace. sg. fem. sallaubiskan 'marriage'

salobisquan (II)

Ace. sg. masc. malnijkikan 'child'

malnijkikun

Gen. pl. masc. [?] grikan '[of] sins'

grecon (I) grekun (I) griquan (II)

Adverbs deineniskai 'daily' laimiskai 'richly' perarwiskai 'certainly' etnrwingiskai 'mercifully'

deinenisku 6 laimisku perarwisku etnrwingisku

Prefix papagauts 'conceived' pakunst 'to watch over'

pogauts pokunst

Verbs asmai 'am' asmu, asmau polinka 'remains' polijnku (The form polijnku is not a 'conjunctive'; it is doubtful that Old Prussian even had a 'conjunctive.') Sometimes the testimony of related forms in Old Prussian or other languages suggests that we are encountering an attempt to record the labialization of a velar or a labial. Thus we find a nom. pl. masc. malnijkiku 'children' but also a nom. pl. masc. malnijkai 'id.' The voe. pl. masc. waikai 'servants' probably reflects the same form as the nom. pl. masc. waikui 'id.' Old Prussian kurwan 'dem Ochsen' which occurs once in the Enchiridion (beside curwis 'ochze' in the Elbing vocabulary) is apparently cognate with Lith. karve 'cow. I The -u- in the Old Pr�ssian words may be an attempt to render the /a/ af'ter a velar consonant. Among the various Germanic dialects known

in the East European area was certainly Middle Low German, which was used in the Baltic provinces. (See Lasch, 1914, 1.) Thus it would hardly be surprising if some of the Middle Low German orthographic practices found their way into the Old Prussian texts, although Lasch (19) says that the contemporary Old Prussian provinces had little significance for the Middle Low German standard language (Schriftsprache), since the influence of the Middle German language of the orders was strong there. Of 37 grand masters of known origin only two were of Low German origin and only one was of importance for the literature of the order, viz. Luder von Braunschweig, although his family was in the High German literary tradition. (See Helm and Ziesemer, 1951, 36.) The origin of the other colonists was, however, more varied. And Helm and Ziesemer say (39) that even those works which had their origin in the land of the order do not have a linguistic unity. Those active in literature were from various regions and they did not immediately lose their local speech habits. Thus Helm and Ziesemer ( 39) say: "Das lautliche Bild der einzelnen DenkmHler ist oft recht differenziert, so dass die jeweils gebrauchte Sprachform aufzunehmen ist und mit gewissen Beschr!l.nkungen aus ihr Schlilsse auf die Herkunft des Verfassers gezogen werden kBnnen." As Trautmann (1910, XVI) shows, there was, however, at least some influence of Middle Low German vocabulary in Old Prussian. He quotes the words h8fftmannin 'captain' from Middle Low German (MLG) hBvetman, instran 'fat, grease' from MLG inster, penningans 'money' from MLG peninge, predickerins 'pastors' from MLG prediker, reisan 'time' from MLG reise, reckenausnan 'account' from MLG rekenen, clines 'bran, pollard' from MLG klfen, broakay 'shorts' from MLG brdk S. Th. Hoffheinz (1872, 459-460) concludes that the contemporary East Prussian dialect of High German was also influenced by Low German

12

and Lithuanian and Old Prussian. The Germanic ol (= Gothic /o/, Old High German /uo/) was f;equently written as� in Middle Low German according to Lasch (1914, 96). She goes on to say: "Durch den dialekt gewghrleistet ist uo... im brandenburgischen und Zerbstischen. Hier wird in der ganzen mnd. periode konsequent� geschrieben. Sonst kommen u im anfang auf dem ganzen gebiete vor, am stirksten elbostf�lisch und nordnds., besonders ostelbisch." According to Lasch (112): "Die diphthongierung von .9..•••, die heute auf einem weiten gebiet zu beobachten ist, setzt schon in frlliier mnd. zeit ein. Mangel an geeigneten zeichen zur darstellung, sowie die traditionelle orthographie (die durch die nordalbingisch-Llibeckischen verh�ltnisse gestlitzt war, s.u.) beschr�ken die be�eichnung , der diphthonge in �lterer zeit auf wenige beispiele, die aber doch genligen, den vorgang schon flir diese periode zu sichern." One can then imagine that on the basis of the Middle Low German orthography the letter� may have been used to denote an Old Prussian vowel which appeare� to the German scribes to be /u�/ or /ua/. Now if we accept Burwell's proposal (see below) that Baltic */o/ and */a/ had merged as */o/ which later passed to /ua/, it would not be surprising to find an Old Prussian orthographic� in places wheE_e we would posit an etymological Baltic */a/ or */o/, cf. Old Prussian muti 'mother' which could be the German way of rendering [muate] which stands for an earlier */mate/. Possibly the orthographic� occurred only a�er labials and velars because the stronger labialization was noted there. Af'ter a dental this labialization might not have been noticed so easily and this fact may explain the numerous form� with orthographic§:., cf. Old Prussian datwei 'to give,' etc. Gerullis (1922, 270) gives alternate forms of certain Old Prussian place names as follows:

(1) Wogenis, Ugeyne (today Uggehnen) (2) Uppin, Woppe (today Oppen) (3) Wundithen, Wondithen (today Wonditten) (4) Wutterkaym, Woterkeim (today Wotterkeim) (5) Warmediten, Wormedith, Wurmdit (today Wormditt) (6) Warkaym, Workaym, Wurkaym (today Workeim) (p.

13

208)

(7) Worelauke, Worlauks, Worlavken, Wurlauken

(8) Wormen, Warmen (9) Worwayn, Wurwaynen (10) Worennye, Wrenie, Vuoronnye, Vuoreine, Worenyge, Werennye, Worennie, Vorennie, cf. Lith. Varen�. Under the heading for the name Worit Gerul lis compares Lith. Voryte, the name of a brook, and proposes that the Old Prussian word derives from a root cognate with Lith. voras 'old' (with the addition of the suffix -rt-). In the Enchiridion we encounter the Ol d� Prussian word urs 'ol d,' which is also thought to be cognate with Lith. voras ( *varas. The Enchiridion word urminan 'red' is usually considered to be related to the Elbing vocabulary word wormyan 'id.' and Simon Grunau's warmun'id.' Likewise the Enchiridion words gallu, gallu 'head' are surely cognate with Lith. galva 'id.' I suspect then that Old Prussian orthographic u may have stood also for /va/ or /va/. Thus Old Prussian urs should perhaps be phonemicized as /var[a]s/, urminan perh_§J)s as /varminan/ (cf. wormyan and warmun), · gallu, gallu perhaps as /gal va/. This orthographic peculiarity may then also be explained by the tendency of some Middle Low German texts to use u to denote /uo/. (See Lasch, 1914, 96-97.)- Perhaps Baltic /va/ and /va/ 3ounded like German uo to the German� speaking scribes. A common method of denoting a long vowel in Middle Low German was to write a double vowel, viz. aa, ee, ii or sometimes to write a second vowel letter which by itself would denote a different quality. (See Lasch, 1914,

14

24.) Thus, apparently, ae, ai or §}I_ could

denote /a/, ei, � could denote /e/, ou, �' oi, 9J£_ could denote /o/, and �' ui, � could denote /u/. According to Lasch (1914, 25) there seems to be no difference in the use of orthographic � or i:_ as a second component. She quotes the following examples: raid, jair, hues, uit, altaer, oest, oestzyde, oisten, oistside. I would suggest that in Old Prussian the word daeczt (= /dat[a]s; from the II Catechism) furnishes an example in which orthographic ae denotes (See Maziulis, 1966, 45.) Th;­ orthographic sequence ai also sometimes seems to represent /a/. An �ample is furnished by Old Prussian mensai 'flesh, meat' which is also written as mensa. Likewise in Old Prussian aucktimmisikai 'authority' (also written aucktimmisku) word-final -ai probably denoted /a/. Compare also the forms crixtisnai, crixtisna and crixtisna. It might be pointed out here also that probably word-final orthographic sequences such as -ai and -ei did not necessarily denote a long vowel in Old Prussian, but just what the German scribe felt had to be written as a long vowel (e.g., possibly just a stressed vowel). Since German was apparently losing word-final contrastive entities, the German scribes may well frequently have perceived as long a vowel which from the Old Prussian point of view was short. It should be noted, however, that words ending in -ai, -ei or -i may merely show the addition of�he particl; /-ai/, a particle similar to the cognate particle in Lithuanian. Zinkevicius (1966, 431) says that in Lithuanian dialects this particle may be added to pronouns, numerals, adjectives and certain verbal forms. For example, from 'I' we find asai, from du 'two' we find dujai-,-from b�ltas 'white' we have the definite adjective b�ltasai and from scl{a 'turns' we have suk.ai. Vacillations between o and a are well attested in Middle Low German texts. Thts is undoubtedly the result of the fact that the

/a/.

i'I

I

as

15

latter language was undergoing very complex vocalic changes. Thus the phoneme /a/ probably had Q-coloring already in Middle Low German. (See Lasch, 1914, 64.) Note the vacillation between �' vun, � (Lasch, 33) and dan, den, don 'dann' (Lasch, 34), etc. Occasionally in Old Prussian we find the representation of a voiced consonant where on etymological grounds we would expect an unvoiced consonant. Some examples are (1) bokl.::_smans /paklusmans/ 'obedient', cf. poklusmai 'id.', (2) agins /akins/ 'eyes ' cf. ' . L i"th · � k' I id. I , Old Prussian accodis /akutis/ I vent' exhaust hole' I cf. Lith. akt.te 'bud I (etymologically a diminutive of akls).7 One further remark about the system of tran�cription is that we do not find spelling consistency. Such consistency is certainly a laudable cultural goal but it is clearly ethnocentrism on the part of linguists to assu me that this is a universally mainuained cultural value for all periods of human history. One need merely examine the writing of a semi literate in any language in order to note­ numerous vacillations in spelling. We must therefore not be surprised if Abel Will or some other German scribe writes the same word in several different ways. Such vacillations in spelling are found in other dead languages such as Old Church Slavic and Hittite, etc. In this book the digraph te has been rendered by� and the digraphtl has for the most part been rendered as ue. -

Vowels

1:.

The Proto-Inda-European vocalic system is usually reconstructed as a five-vowel system, each vowel of which had both a short and long counterpart. In addition several shwas (one of which may be the reflex of a laryngal consonant) are usually posited, but these can be disregarded in any study of Old Prussian. In all of Baltic the short */o/ merged with short */a/ (as in Slavic and Germ anic), but only in West Baltic (i.e. Old Prussian) c an we posit a merger of */o/ with */a/. In East Baltic apparently under most circumstances */o/ and */a/ were kept separate, cf. Lith. duoti 'to give' (< *do-, cf. Greek didomi 'I give') as opposed to Lith. m6teris 'woman' ([2] /i/ (semme g �eme 'id.'], sidans 'havin 'earth' [cf. Lith.� �Td• 1 0 • down '] , � sit 'to sat' [cf. Lith. sesti 'meal' [cf. Lith. �dis 'food']); [1] /o/>[2] /u/ (noseilis 'spirit' beside nuseilin). It is

19 difficult to know, however, whether this apparent innovating system really reflects the Old Prussian situation, or whether these alternate spellings just reflect the confusion of the German scribes. One notes, however, that apparently Abel Will could distinguish the stress on the second element of the diphthongs which derive from long vowels, cf. Old Prussian sounan 'son' (= Lith. acc. sg. S�1t, 'id.'), bout 'to be' (= Lith. btti 'id.'), etc. This could perhaps serve as support for the belief that the diphthongal writings here really did denote diphthongs. It still seems likely to me that the contrast between /e/ and /a/ was just barely phonemic and that in position a�er a /j/ or palatalized consonant there was no contrast at all. (In other words the situation may have been similar to that of modern Lithuanian where /e/ and /a/ contrast only in word initial position. The Lithuanian orthography is morphophonemic so that in words like joja 'rides,' vercia 'turns' the phonemic sequences /je/ and /ce/ are written -� and -�ia respectively although there could be no contrastive */ja/ and */ga/.) In my transcription of the Old Prussian words I follow the example of Lithuanian orthography by writing /ja/ and /Ca/ even though the sequences in question may have reflected phonetfc [je] and [Ce]. Thus a word like trinie 'threatens' may have been phonetically [trine], but I would write it morphophonemically /trina/. The diphthongs in Old Prussian are /ai/, as in /laika/ laiku 'holds to,' /au/ as in /ausautenikamans/ auschautenrkamans 'debtors' and /ei/ as in /eit/ eit .r goes.' The sequence [eu] may have existed in Old Prussian, but it was merely one phonetic realization of the underlying /au/. In Balta-Slavic the sequence */eu/ passed to */(j)au/ so that if there remains any phonemic distinction between etymological */eu/ and */au/ it is expressed by

20

the /j/ or the palatalization of the consonant preceding that /au/ which derives from earlier * I eu/. Compare Lith. li audis 'p.eople' ' Latvian iaudis 'id.', Old Church Slavic ljudbje, Old High German liut, all apparently deriving from a source *leud-. In the Elbing vocabulary we find the words skewre 'sow,', bleusky 'sedge,' geauris 'cormorant,' gleuptene 'smoothing board,' keutaris 'ring-dove,' keuto 'skin; hide,' peuse 'pine tree.' Even supposing that the orthographic sequence eu did indeed represent phonetic [eu], one would assume that the preceding consonant was palatalized so that phonetic [e] in this sequence stands for underlying /a/, the fronting being determined by the preceding palatalized consonant. Examples of the diphthong /ui/ are more difficult to find and perhaps occur only in borrowings such as the word /kuilis/ cuylis 'stud-boar' (found in the Elbing vocabulary),' cf. Lith. kuilys 'id.', Latvian kuilis, all probably from Slavic *kylb. The word muisieson 'greater,' which I originally thought ( 1 964, 215) to embody an example of this diphthong is to be phonemicized /mai6esan/, the letter u merely being the representation of /a/ a�;r /m/. The correct phonemic interpretation of this word, by the way, does away with some of the traditional difficulties in connecting the word with Gothic mais 'more, rather.' There is no good reason for positing long diphthongs in Old Prussian, all of these having been shortened in the Balta-Slavic, or at the very latest, the common Baltic period. Indeed, if Old Prussian had retained a contrast between short and long diphthongs, it would be the only Baltic or Slavic language to have maintained this contrast as a contrast of length into the historical period.1 1 It is possible that the distinction of long vs. short diphthongs was retained as a difference in type of intonation. In other words the original short diphthongs may show a circumflex intonation, whereas the original long diphthongs may show an acute

21 intonation, cf.,e.g. (with an etymologically short diphthong) Lith. varnas 'raven,' Serbo­ Croatian vran, Russian v6ron, etc. vs. Lith. varna 'crow,' Serbo-Croatian vrina ' Russian vorona, etc. But there -is no evidence that vocalic length had a phonemic function in the first part of either pure or 'mixed' diphthongs. It can be noted then that if the orthographic sequence seyr 'heart' (found in the Elbing vocabulary), really represents /ser/, it is the only word in the entire Baltic (or Slavic) language family which would retain a long vowel plus sonant in word-final position. Thus the �honemicization /ser/ seems highly dubious. 1 Sometimes the existence of the orthographic sequence -on in certain cases which are assumed to be�enitives plural is used as justification for the belief that long diphthongs were shortened separately in word­ final position in Old Prussian. Thus we find the expression (7, 1) Et werpsannan grecon Vorgebung der sUnden 'forgiveness of sins,' although farther on in the same I Catechism the same expression is rendered (7, 29) as att werpsannan grekun - vergebung der sunde� The variation between grecon and grekun is merely a chance vacillation in orthography. The final written -on is quite meaningless, cf. the acc. sg. masc. deickton 'something' or the numerous pa�t passive participles ending in -.9.£, e.g. Popeisaton 'written,' ensadinton 'established,' podaton 'given' for which some kind of nominative case would seem to be required by the syntax. In conclusion, then, in the morphophonemic transcription of Old Prussian in the following sections I will assume a four-vowel vocalic system which presupposes the merger of both long and short */o/ with /a/. All of the 'diphthongs' will be considered to be sequences of two short vowels. Vocalic length is also phonemic. /

Stress In the foreword to the Enchiridion Abel Will wrote: "Damit aber der leser solche sprach nach jrer Natuerlichen art verstendiglich lesen koenne: und es die zuhoerer auch verstehen ' ist dieses fleissig zu mercken, das die Fuenff Vocales gemeiniglich durch eine lange Pronunciation aussgesprochen werden, Derwegen solche Buchstaben jhre sondere zeychen haben muessen. Wo nun diese nachfolgende verzeychnus an einem solchen buchstaben im wort erfunden, muss derselbige mit seinem gewohnlichen accent pronuncij rt werden. � � I Q � ij_." (See Trautmann, 191 0, 1 84.) In many cases the macron is used over vowels which correspond to long stressed (either with the acute or circumflex intonat�on) vowels of Lithuanian. Examples: Old Prussian Lithuanian dat 'to give' duoti I id. I salin 'grass' Zoll I id. I idis 'meal' tdis 'food' postat 'to become' pastoti 'to become pregnant' This can be used as evidence that there was a contrast between long and short vowels in Old Prussian. In diphthongs the macron may stand on the initial or the second element. The macron may stand on the second element, however, only if this is i or u. The macron never stands on one of the sonants, i.e. �' E_, !_, 1., although it seems reasonable on analogy with the pattern of Lithuanian to assume that /m, n, r, 1/ were stressable as· the second element of 'mixed' diphthongs. In principle the macron on the first element in the Old Prusssian word corresponds to the circumflex in Lithuanian, whereas the macron on the second element in the Old Prussian word corresponds to the acute in Lithuanian.

Examples: Old Prussian eit I goes I laikU 'holds to I kaulinS 'bones I laiskas 'booklet'

23

Lithuanian et'ti 'to go' laiko 'holds' · k�ulas 'bone' lafskaS I letter I (but more commonly laiskas) The macron here certainly does not reflect phonemic length in the Old Prussian words, even though the German scribes may have heard the vowel as long. From.the Old Prussian point of view it may have been amplitude or relative pitch stress which was marked.13 In this regard it is ins.tructive to compare the remarks of A. Laigonaite (1958), who argues that relative amplitude, not relative pitch furnishes the phonemic basis for the difference between the Lithuanian acute and circumflex intonations. In those syllables traditionally marked with the acute there is relatively greater amplitude in the first part of the syllable. In those syllables traditionally marked with the circumflex there is relatively greater amplitude in the second part of the syllable. In Lithuanian diphthongs the stressed element is relatively longer than the unstressed element and I assume the same to be true for Old Prussian. One may suppose, then, that Abel Will, in whose native German vocalic length was phonemic, noticed an allophonic feature of the Old Prussian stressed diphthongs, i.e. the fact that sometimes the first element, sometimes the second element of the diphthong was long. Since his German ear did not dispose him to hunt for this phenomenon when /1, m, n, r/ functioned as the stressed second element he never marked this. There is, of course, also the good possibility that the printer would not have had the signs even if Will had noticed the phenomenon. (See Trautmann, 1910, 194.) One may note, however, that sometimes the macron is written on the first elem�nt of one of the 'mixed' diphthongs (i.e. a diphthong with the second element

24

/1, m, n, r/) when the corresponding diphthong in Lithuanian has a circumflex intonation. Examples: Lithuanian Old Prussian (ace. sg.) algq, 'salary' algas 'wage' martin 'bride' (ace. sg.) marci1t 'id.' ains antran 'each (ace. sg.) antr� other' 'second' kumpas 'crooked' etkillllps 'again' rankan 'arm' (ace. sg.) rankfl_ 'id.' As I have mentioned earlier (p. 19) stress on the second element was marked even in those diphthongs which have their origin in an etymologically long vowel. The stress on the second element in Old Prussian is equivalent to a Lithuanian acute. Examples: Lithuanian .,, Old Prussian sounan 'son' (ace. sg.) sun1t, 'id. ' geiwan 'alive' �vas 'id.' buti 'id.' bout 'to be' Perhaps this evidence supports the assumption that the Old Prussian circumflex and acute were distinguished in long vowels, although it was impossible to mark this distinction in the case of a single vowel. Since it seems to be a fairly safe assumption that in the diphthongs the vowel which Abel Will marked with a macron was never long, we may then also suppose that elsewhere vowels marked with a macron may merely have seemed to be long to the German ear, but in fact may not have been long at all. The problem is particularly evident when we ask about the meaning of the macron in non-initial syllables. One can assume that in principle the stress should fall on the initial syllable for the German, so that when he heard an accented vowel in some syllable other than the initial, he may have assumed this vowel to be long. Thus, for example, Old Prussian nom. sg. fem. antra 'second' (cf. Lith. antra 'id.') and pienc�fif'th' (cf. Lith. penkta 'id.) then might well have a long stressed vowel in the

25 final syllable, but there is no way of knowing for sure. I tend to doubt that in Old Prussian perwedda 'lead' or wedde the final vowel is really long. It is possibie that in the word tans 'he' (which seems to derive from */tanas/) the macron marks stress rather than length, although perhaps it marks both length and stress. 14 According to Trautmann (1910, 185) short accented vowels are marked by doubling the following consonant: Old Prussian buttan (cf. Lith. butas 'apartment'), Old Prussian kittan 'other' (cf. Lith. acc. sg. klt� 'id.'), Old Prussian dessimton 'ten' (cf:-i:;rth. desimt 'id.'). EndzelT"ns (19 43, 19), however, doubts that the doubling of the consonant always means that the preceding vowel is stressed and cites several examples (ku:mpinna 'stops, impedes,' skijstinnons 'purified') in which an unstressed short vowel precedes a doubled consonant. It seems that Abel Will used the device of doubling consonants to mark a short vowel on the model of German. I conclude then that vocalic length is phonemic: vowels are either long or short. Short vowels count one mora; long vowels count two morae. In diphthongs (including tauto­ syllabic sequences with /1, m, n, r/ as the second element) both the first and the second element count one mora each. Diphthongs therefore count two morae. Amplitude or pitch stress is phonemic. In each lexical item there may be only one peak (either of pitch or amplitude). This peak may occur on a short vowel or on the first or second part of a long vowel or a diphthong.

Consonants The Old Prussian consonant phonemes exist in plain and palatalized variants except for /j/ which is pure palatalization and perhaps for /s/ and /z/[?] which are discussed below. The palatalization of consonants before front vowels is not marked since it is automatic. The labials are /p, b, m, v/. Examples: Plain variants Palatalized variants 15 /pa-paika/ popaika /kneipa/ knieipe 'deceives' 'obtains' /pa-ba[j]int/ pobaiint /bi[j]a/ bia 'fears' 'to punish' (with automatic palatalization of the /b/ by the following front vowel) /maita/ maita /mentama/ mentimai 'we 'nourishes' lie' (with automatic palatalization of the initial /m/ by the following front vowel) /varga/ wargel6 /vest/ west 'to lead' 'bothers' (with automatic palatalization of the /v/ by the following front vowel) The dentals are /t, d, n/. Examples: Plain variants Palatalized variants /turet/ turTt 'to he.ve' /mentama/ mentimai 'we lie' /gei.da/ gej de, gieidi /dat/ dat 'to give' 'wait' /sadina/ saddinna, /trina/ trinie sedinna 'puts' 'threatens' The dental spirants are /s, z/. Examples: Plain variants Palatalized variants /sadina/ saddinna /septmas/ septmas, sedinna 'puts' sepmas 'seventh' (with automatic palatalization of the

27 /s/ by the following front vowel) /zalen/ salin 'grass, /zemen/ semmin, semmien, weed' semien, semman 'earth' (with automatic palatalization of the /z/ by the following front vowe 1) Ordinarily German orthographic practices do not make any allowance in the Catechisms for a distinction between /s/ and /z/. The word zuit 'enough' is written with an initial�' but the word seems to be borrowed from Polish� 'full.' There may have been a palato-alveolar spirant (and /z/?) derived from */sj/ (and */zj/?). This is highly tentative, but I give below several possible examples. Plain variant Palatalized variant /au��audesnan/ /tavi�an/ tawischen 'neighbor' auschaudisnan 'faith, hope' I have assumed that Proto-Baltic */ii and */t/ have passed to /s/ and /z/ as in Latvian. The velar stops are /k, g/. Examples: Plain variants Palatalized variants /is-rankTt/ isrankit /en-kaitetai/ ankaitTtai 'to save' (with 'tempted' automatic palatalization of /k/ by the following front vowel) /gimtun/ gemton 'to be /pa-gadint/ pogadint 'to destroy' born' (with automatic palatalization of /g/ by the following front vowel) The liquid /1/ functions as a consonant in the following examples. Plain variant Palatalized variant /pa-laipsan/ /pra-leitan/ pralieiton pallaipsan 'shed' (with 'commandment' automatic palatalization of the /1/ by the

/s/

28

I, i

following front vowel) The trill(?) /r/ functions as a consonant in the following examples. 17 Plain variant Palatalized variant /is-rankit/ isrankit /kafausnan/ kariausnan 'struggle, fight' 'to save' The consonant /j/ may be found in the following example: /juren/ jurin 'sea' The letters f and h are found in such Old Prussian words as falsch 'false' (Trautmann, 191 0, 334) and h3f�mannin 'captain'(Trautmann, 1910, 345), but probably neither represents a phoneme native to Old Prussian. The h found in the words hest 'is' and haese 'from' is of no significan�

THE NOUN In the Baltic languages the noun vacillates between stem classes more than in other Indo­ European languages. (See EndzelTns, 1957, 63.) Thus, for example, we find Lith. bradas 'ford' beside brada, Latv. klavs 'maple tree' beside klava and malks 'gulp, swallow' beside malka. The situation is even more confusing with regard to the i-stems, JQ-stems and �-stems which have become quite mixed up, usually as a result of the passage of the old i-stem noun to the JQor �stem category. Th�s, for example, in Old Prussian the situation could well have been similar to the Lithuanian situation where the i-stem dative singular avei '[to the] sheep' ;ould have been the same as the e-stem dative singular katei '[to the] cat.' (See Kazlauskas, 1968, 200-206, where we also find many examples of the passage of the old i-stem nouns to the J.£_-stem category.) In contemporary Lithuanian we frequently find the same word represented in � a number of different stem classes, e.g. kande 'moth,' also kandls (i-stem), kandZia; usnis 'thistle' (i-steml, be;ide usnt; obells 'apple tree' beside obele; giria 'forest' beside dialect gire; dialect medzias 'tree, wood' beside standard medis; dialect kelias 'knee' beside standard kelis. (See also Senn, 1966, 105.) Sometimes the Lithuanian iu-stems are confused with the J.s?..-stems, cf. Lith. a.mzius 'age' beside dialect a.mzis. (See Senn, 1966, 1 05.) As a result, then, both of the vagaries of Old Prussian orthography and the possibility of shi� of stem class on the part of the Baltic noun it is extremely unwise to state unequivocally to which stem class each noun belongs. Thus I assure the reader that the proposed reconstructions are quite tentative and at best seem only probable to the author of these lines. The o-stem declension is reconstructed as follows. The nominative singular ending is reconstructed as -as, usually reduced to -�,

30

cf. Lith. laiks 'time' beside laikas. Thus we find the Old Prussian nom. sg. masc. deiwas 'God' one time beside the more u�ual deiws for this case. I asswne also that the form labbas 'property' which Trautmann (1910, 366) labels a genitive singular may well be a nominative singular since the syntax would seem to require a nominative case in the context. A rather complex sequence of consonants (cf., e.g. Lith. bbkstas 'tower') could favor the retention of the -as in the nominative singular, cf. Old Prussian deicktas 'something. ,19 Leskien, 1876, 31-32, argues that the Old Prussian o-stem genitive singular ending -as could not-really be explained as deriving from an earlier*-asja, because the genitive singular demonstrative pronoun steisei, steise, steisi, stesse, stessei, steisai seems to presuppose the ending -sja. If this latter explanation of the pronoun is correct, then it� iS impossible to suppose that deiwas derives from an earlier*deivasja. In Old Prussian, differently from Slavic, there is a clear distinction between the nominal and pronominal endings in question and both endings have cognates in related languages. Leskien suggests that rather than to asswne a separate ending in *-as which would have existed parallel to an ending*-a (reconstructed on the basis of the Lithuanian and Latvian*o-stem genitive singular) it is most probable that the ending *-as was taken over from the a-stem nouns. Kazlauskas (1968, 173-174) suggests that indeed both the pronominal and nominal paradigms could have had a word-final*-sja (cf. Sanskrit tasya ' [of] that' and v0kasya '[of the] wolf'). This word-final*-sja p assed first to*-sje and then to*-�. Kazlauskas then says that the ending -� could have been retained longer in the pronominal paradigm because in the latter paradigm there are more cases having a dissyllabic form, cf., e.g., the dat. sg. masc. stesmu '[to] the.' In the nominal paradigm, on the other hand, there was

31 a tendency for dissyllabic endings (preswnably such as*-� from earlier*-asja) to become monosyllabic and the word-final short vowel -e was lost. Kazlauskas also points out that attempts have been made to derive the Old Prussian ending -as from Proto-Inda-European *-oso, although the latter ending is indeed itself disputable, since in recent times people are inclined to derive the Germanic ending from *-osjo or*-es,jo and the Slavic pronominal form ceso I of what I could be a late innovation. He also mentions the Gothic o-stem genitive singular in -is (cf. Gothic wulfis '[of the] wolf') and th;-Hittite form antub.sas 'man' which serves both as a nominative and a genitive singular. I might suggest that in Hittite there may have been a distinction between the nominative and genitive forms which the orthography does not reveal, but on the basis of the evidence it �8uld be impossible to prove such an asswnption. The suggestion which appeals to me most is that of Andr� Vaillant (1958, II, 30) who writes: " ...la desinence -as du vieux prussien represente la superpositio;;:-�*-� de la finale -s du genitif dans les autres types de flexion, r";m. -as, -is (lit. -os, -es), etc., sans doute par l'intermediaire des masculins en -is, gen. -is, qui confondaient les deux types en -";£2_(gen. lit. -io) _et en -i- (gen. lit. -ies)." If Leskien (1876, 34-35) is right in equating the Old Prussian phrase kas arrientlaku with Lith. kas aria ant lauko �which plows on the field,' the final -u of laku may re12_resent the original genitive si-;gular ending*-..§:_ or indeed Old Prussian -a.21 The o-stem dative singular presents a problem since it seems to show both the ending -ai from the a-stems and the ending -u from the *a-stems. Th; forms ending in -ai which could b; considered dative singular ar;-as follows. The word bitai 'in the evening' which occurs twice is, �ccording to Endzelins (1944, 85) not a dative singular, but a locative singular. In

32

the expression schieison malnijkikai 'of this child' EndzelTns suggests a correction of the word-final -i to -g_. In his opinion the same is to be done in the expression enstesmu wirdai 'in this word,' particularly in view of the fact that the expression stesmu wirdan, apparently a dative object of the word druwe 'believe,' occurs elsewhere in the Enchiridion. Likewise Endzelins proposes that in the expression stesmu kermeneniskan istai bhe pouton 'the bodily food and drink,' which would seem to be governed by the preposition sirsdau which requires the dative case, the word I"stai 'food' is to be corrected to I"stan if it isn't a contamination of Tst and Y-dai. According to Endzel'fns (1944, 84): "Auf -u enden grThu, maln'fku, piru, siru, waldniku, wo -u... nebst zem. -ou resp. -iI auf urbalt. -o (� lat. -o) zurlickgeht das d;rch Einbusse eines -i aus id;. -oi (> gr. -Q und li. -ui) entstanden is=r:-'' Maliulis (1970, 116) derives the Lithuanian dialect *£-Stem dative singular (viik)-� directly from an uncontracted *-£ in Proto­ Indo-European. He derives the u-stem form (sun)-uo directly from Proto-Baltic *-au. The standard Lithuanian dative singular ending (vili)-ui and dialect (vilk)-� are, in his opinion, borrowed from the u-stem endings (sun)-ui and dialect (sun)-� respectively. The �-stem ending -� comes directly from one type of Proto-Inda-European dative singular form which is essentially nothing but the pure stem. The second u-stem dative singular ending *-au (from anoth;r type of �-stem dec}ension) would correspond exactly to a Slavic ending -�- Thus Matiulis (1970, 118) quotes a personal letter (31 January 1960) from V.N. Toporov who wrote: "I am convinced that the ending -� in the dative singular of the a-stems in Slavic vl�k-u has nothing in comma:; with *oi (as Meillet, and after him many others, thought) and I am ready to think that it is really from the �-stems." While not wishing to commit myself on all Q..



-

-

33 the details of Maziulis' arguments, I would accept the clear relationship between the 2,.and u-stem dative singular endings and that in Baltic at least there was borrowing from one stem class into the other. Likewise I am inclined to accept Toporov's conclusion that the Slavic a-stem dative singular ending -u has nothing-to do with a posited *-oi and� least the possibility that it was taken from the u-stem nouns, although I might suggest ;-lternatively that the Slavic form derives from a pronominal ending *-am which passed to *-uN and then *-i (with sporadic loss of the nasalization). (See Schmalstieg, 1971, 145.) Matiulis (1970, 268) says further that in West Baltic the a-stem dative singular *-o is represented in Old Prussian (sir)-� '[to the] heart' and that this ending -� must have merged with the u-stem dative singular in -u, cf. Old Prussian Tpeck)-�. It turns out, ho;ever, that there are no u-stem datives singular attested in Old Prussian �less one accepts Ma¥iulis' interpretation of the word pecku in the expression kas stesmu pecku swaian perdin dast Der dem Vihe sein Futter gibet 'who gives to the cattle his (apparently with wrong reference in Old Prussian) food.' If Matiulis is right and this is indeed a dative singular, one may then assume that this �-stem dative singular ending was transferred directly to the 2,.-stems. On the other hand there is good evidence to show that orthographic -u after velars is one way of rendering /ai/. One-can compare the adverbs deineniskai 'daily,' perarwiskai 'certainly' which are also spelled deinenisku, perarwisku, etc. More examples are found on p.

10.22

Therefore the orthographic -� occurring after labials and velars is not very good evidence for a phoneme /u/. On the other hand word-final orthographic -� does occur after r in two words, viz. siru 'heart' and piru 'community, parish, congregation.' The first word occurs in the expression kai ious

34

sturnawingisku prei siru immati - das jhr ernstlich zu Hertzen nehmen 'that you take earnestly to heart.' These two words are the only two attested Old Prussian words in which an assumed dative singular ending -u occurs a�er a consonant other than a labi-;i or a velar. In fact they are apparently the only two words in Old Prussian with a word-final -u preceded by !.·17 I could imagine no particular phonetic reason why one should asswne that the labialization of a phoneme /r/ would be more noticeable. We do find such words as tickromien 'right [hand],' tickromiskan 'right,' peroni 'parish, community' in which the orthographic o must denote a short or long /a/ if there is n-;;- short or long /o/ in the phonemic system. One can also point to the fact that in the Elbing vocabulary German ros 'steed' is rendered by Old Prussian russis and that in the Enchiridion we find the apparent gen. sg. masc. russas '[of the] horse.' Similarly in the Elbing vocabulary we find the Old ,.., Prussian word roaban as a cognate to Lith. raibas 'variegated.' Here the Old Prussian phonemic sequence /ai/ is rendered by oa just as it is in the Elbing vocabulary word-;poayno 'scum on fermenting beer'(cf. Lith. spaine 'bubbles of foam on water). So elsewhere we do find phonemic /ri/ rendered by both ro and ru. Since it would be possible to imagine a dative singular *a-stem ending /ai/ as well as /u/ I leave the choice open. I do not believe that the evidence is sufficiently clear to favor clearly one alternative more than the other. The *2.-stem accusative singular ending is just what one would expect from the comparative point of view, viz. -�. It has been commonly asswned in the past that the *2.-stem neuter nominative-accusative sipgular should be the same as the *a-stem masculine accusative singular. On the basis of the similarity of the endings in Greek (-on),

35

Latin(-wn) and Sanskrit(-�) an original identity in Inda-European was taken for granted. When Hittite was discovered, however, it was found that the adjectival inanimate gender nominative-accusative endings -a and -an were apparently optional. (See Friedrich, 1940, 16.) On the basis of the evidence of the Lithuanian neuter adjective, e.g.' gera 'good,' sveika 'healthy,' etc. and Slavic dobro 'good,' nova 'new,' it might be just as well to asswne_a_ Bal to-Slavic neuter nominative-accusative ending *-o(or *-a) rather than *-om (or *-am). One go-;;-d reaso; for assuming a Balta-Slavic *-o (rather than *-s�!!!) is that we would expect-*-om to develop into Slavic -�(cf. the *o-stem acc. sg. gradi 'city' [the fi�al back jer-of which seems to derive from *-om]). A Balta-Slavic or Proto-Inda-European *-om should be represented by -8c,. in Lithuanian, cf. the *2.-stem ace. sg. vyr-�. 'lhe Germanic evidence regarding the neuter *2.-stem ending is quite ambiguous. Thus Gothic juk 'yoke' is usually said to derive from *�-om, but there seems to be no reason why the form in question shouldn't derive just as well from *�-.£· Thus in the entire North European(i.e. Baltic, Slavic and Germanic) language group there would be only Old Prussian which would give us evidence for an ending *-om(> -an) rather than *-2.·( -�).23 Since we have no very good assurance that the authors of any of the Old Prussian docwnents were very clear about when to use the nominative and when to use the accusative case, we must be very circwnspect about the use of the Old Prussian evidence. Therefore many of the forms which I have labeled as neuters may indeed be merely masculines or feminines in the accusative case. Thus the assignment of any form to the neuter category may be quite mistaken and there indeed may have been no neuter category in Old Prussian. The *·a-stem nominative plural ending -ai

>

-

36

(cf. Lith. ragai 'horns,' Latin servi 'servants,' Greek �nthropoi 'men') shows .the expected*o-stem nominative pluraJ.. 2 4 In spite of the fact that the*_£-stem genitive plural form is sometimes spelled -9..!!_ (cf. grecon '[of] sins' and grekun 'id.' attested in the I Catechism beside griguan in the II Catechism and grikan, grijkan and grikan in the Enchiridion) I.assume a genitive plural ending_-an. There is, of course, also the possibility that the genitive plural ending in Old Prussian is -un in which case it would correspond very nicely with Lith. -1t,. The Latvian definite genitive plural adjective ending -uo is surely an innovation. (See Schmalstieg, 196 8a,190.) It would be possible to establish a Proto-Baltic genitive plural ending*-an or*-�, either of which would correspond very well with Slavic -1· I assume that the Proto-Inda-European genitive plural *_£-stem ending was exactly the same as that of the*o-stem accusative singular; the Greek and Sansk;it languages for which an ending*-om is commonly reconstructed for the genitive plural show an innovation here. The Old Hittite ending -��d Latin -E:!!!. could easily reflect *-om also. The Old Prussian dative plural*_£-stem ending is -amans also known in the variant -�. I would suspect that the latter spelling may represent an allegro form in which the nasal before the spirant is lost. Kazlauskas, 1970, 87-91, assumes a Proto-Baltic ending*-� wit� a form*-� developing under the inf·luence of the dative dual ending*-man .26 The accusative plural ending is exactly what one would expect on the basis of its etymological structure, viz. theme vowel*-o­ which passed to Baltic /a/ plus accusative marker /n/ plus plural marker /s/. The Greek dialect*_£-stem accusative plural ending -ons also corresponds to the Old Prussian ending -ans. The Lithuanian ending in -[u]ns may be a recent borrowing from the u-stems or perhaps

37 the expected development of*-ans in word-final position. (See Schmalstieg, 1968a, 185-193.) For comparison a sample Lithuanian paradigm is given below: *o-stem .,,, . Nom. sg. vyr-as 'man'· Nom. pl. vyr-ai / Gen. pl. vyr-11. Gen. sg. vyr-o Dat. pl. vyr-am[u]s Dat. sg. vyr-ui Acc. pl. vyr-us Acc. sg. vyr-� Inst. sg. vyr-u Inst. pl. vyr-ais Loe. pl. vyr-uose Lbc. sg. vyr-e Voe. sg. vyr-e On the basis of the evidence of Baltic� Slavic, and other Indo-European languages I . reconstruct the etymological*o-stem paradigm as follows for Old Prussian: Nom. pl. -ai Nom. sg. -[a]s Gen. sg. -as27 Gen. pl. -an Dat. sg. -ai or -u Dat. pl. -ama[n]s Acc. sg. -an Acc. pl. -ans I reconstruct the Old Prussian*J£-stem paradigm as follows: Nom. pl. ? Nom. sg. -is 7 2 Gen. pl. ? -[j]as sg. Gen. Dat. pl. ? Dat. sg. ? Acc. pl. -ins, -jans Acc. sg. -in, -jan The accusative endings are -in, -ins if a consonant precedes but -j an, -j ans if a vo�el precedes. One apparent exception is /tavisan/ tawischen, tauwyschen, tawisen, tawischan 'neighbor.' For comparison a sample Lithuanian*.si9..-stem paradigm is given below: Nom. pl. brol-iai Nom. sg. brol-is 'brother' Gen. sg. brol-io Gen. pl. brol-ill Dat. sg. brol-iui Dat. pl. brol-iam[u]s Ace. sg. brol-i Acc. pl. brol-ius Inst. sg. brol-iu Inst. pl. br61-iais Loe. sg. brol-yje Loe. pl. br61-iuose Voe. sg. brol-i Apparently the declension of the Old Prussian etymologi_£al*�-stem does not differ from that of the*�-stem so they are given together in the reconstructed paradigm below: /

!

38

Norn. sg. Gen. sg. Dat. sg.

-[j Ja -[j]as -[j]ai

Norn. pl. -[j]as ' -[j]ai Gen. pl. -[j]an Dat. pl. . -[j ]amans. -[ j]amans. Acc. pl. -[ j]ans

Acc. sg. -[j]an Inst. sg. -[j]an [?] It is not clear whether the word-final -a which I have reconstructed for the nominative singular was long or not. There are cases where the orthographic Old Prussian -� corresponds to the Lithuanian ending -a, e.g., Old Prussian pienckta 'fi�h' = Lith-:- penkta, Old Prussian antra 'second'= Lith. antra. I have assumed that the macron denotes an original long vowel, but such an assumption is not absolutely necessary. The macron may merely show stress. The shortening of the final vowel may have taken place in Old Prussian just as it did in Lithuanian. The ending -ai in the nominative plural is" more common than the ending -as. It is unclear whether the ending -ai was really borrowed from the *a-stem category-rnto the *a-stems or whether the use of this ending ;ith the *�-stem nouns is merely another mistake on the part of the translators. In the dative plural it is unclear whether the stem vowel is -i- (reflecting the true *a­ stern category) or -I- (reflecting a borrowing from the *Q_-stems) .For comparison a sample Lithuanian *a-stem paradigm is given below:28 Norn. sg. diena *-a) Norn. pl. dienos (