Old Portuguese in Hebrew Script

OLD PORTUGUESE IN HEBREW SCRIPT:
CONVENTION, CONTACT, AND CONVIVÊNCIA
A Dissertation
Presented to the Faculty of the Graduate School
of Cornell University
In Partial Fulfillment of the Requirements for the Degree of
Doctor of Philosophy
by
Devon L. Strolovitch
August 2005
© 2005 Devon L. Strolovitch
OLD PORTUGUESE IN HEBREW SCRIPT:
CONVENTION, CONTACT, AND CONVIVÊNCIA
Devon L. Strolovitch, Ph. D.

Cornell University 2005
This dissertation explores the process undertaken by medieval writers

to produce Portuguese-language texts using the letters of the Hebrew
alphabet. Through detailed philological analyses of five Judeo-Portuguese
texts, I examine the strategies by which Hebrew script is adapted to represent
medieval Portuguese in the context of other Roman-letter and Hebrew-
language writing. I focus on the writing system in order to challenge the
conception of such texts as marked or marginal, a view that misleadingly
equates language and script. I argue that the adaptation of Hebrew script for
medieval Portuguese is neither derivative of Roman-letter writing nor entirely
dependent upon the conventions of written Hebrew. Nor is it an adaptation
performed anew by each writer and influenced primarily by spoken language.
The perspective I adopt thereby rejects the premise that the patterns
manifested in this unconventional orthography are ad hoc creations by its
writers, that it requires extra effort from its readers, or that it is less "native"
than the dominant, more conventionalized, Roman-based adaptation that
normally bears the title "written Portuguese."
In the first chapter I introduce the phenomenon of adaptation of scripts
in the context of linguistic borrowing and conventionality in writing, and the
uniqueness of Hebrew script in this field. In chapter 2, I present a survey of
adaptations of Hebrew script for languages other than Hebrew, from biblical
Aramaic to late-nineteenth-century English, leading to a more detailed
ii
analysis of the Judeo-Portuguese writing system in chapter 3. In chapter 4 I
present a new critical edition of a handbook for manuscript illumination.
Chapter 5 presents a 27-page excerpt of a previously-unpublished 800-page
astrological treatise. Chapter 6 presents editions of three shorter texts,
vernacular rubrics from two Hebrew prayer books and a short medical
prescription. Chapter 7 summarizes the archaic and vernacular features
attested by the texts in chapters 4-6. In the final chapter, I offer a proposal for
a Judeo-Portuguese "alphabet," along with a sketch of some further problems
of adaptation and interpretation that arise from the process of editing
Hebraicized texts and of transforming them from manuscript to computer
screen.
iii
BIOGRAPHICAL SKETCH
Devon was born and raised in Montreal (Canada), where he learned

French and Hebrew before he knew it, but had to wait until his undergraduate
years at Oberlin College to learn enough Yiddish to decipher his parents'
covert dealings. At Oberlin he earned his Bachelor's degree with an Honors
thesis on genetic metalanguage in Modern Hebrew Linguistics. After earning
his Master's degree in Linguistics at Cornell in 2000, Devon fled to France for
two years, where he taught English at the University of Paris XII, took courses
in Yiddish, Judeo-Spanish, and Judeo-Arabic at Langues'O, and sat for a drink
at more than 300 different cafés.
iii
ACKNOWLEDGMENTS
I came to Cornell directly from Oberlin College, where I had spent my

senior year working on an Honors Thesis. Since I had written most of it after
applying to graduate school, I was sure it would evolve into my Ph.D.
dissertation. Yet my attention was soon diverted by Carol Rosen's course on
the history of the Romance languages. When we were asked to produce our
own editions of some lesser-edited medieval Romance texts, without
hesitation my mind turned to Judeo-Romance. I stumbled onto references to

two Portuguese ma˛zorim, along with a longer Portuguese text about inks and
dyes. Not being able to decide on one, I wound up working on both for the
course and for the anthology of Romance philology that we co-edited. Yet
even after my A-exam I wasn't sure how this would turn into a dissertation,
and while living my post-A-coma in France I put off going to see a couple of
other manuscripts in the same mould.
When I did finally travel to Oxford in the winter of 2002 and reported
back that there was an unpublished 800-page text waiting to be edited, it was
Carol who encouraged me go at it in the style of our previous work together
and "call it a dissertation." Since then, as chair of my committee, Carol has

been exceptionally helpful and supportive of a project that in many ways lay
beyond any single living person's expertise. From the "secret seminar" in the
semester after my return from France to the many, many hours spent poring
over more astrology than either of us would ever need to know, this
dissertation simply would not have been possible without her guidance,
enthusiasm, and vast repertoire of Romance-language insight.
iv
I am also extremely grateful to the other members of my committee,
who have likewise been enormously supportive and (I think) intrigued by my
work. From my first days at Cornell, Wayne Harbert has always made himself
available to talk about my latest musings and to read anything I'd actually
managed to get on paper, including very early – and very late – drafts of some
of these chapters. Moreover, life in Ithaca would have been decidedly less
enjoyable without our weekly grwp sgwrs cymraeg. John Whitman has also
provided a great deal of intellectual encouragement; a course of his in my
second year was as close as I came to reviving my Oberlin thesis, while his
seminar in the Fall of 2004 more or less spawned the first two chapters of this
dissertation. In addition, Gary Rendsburg has been an engaged member from
outside the Linguistics department, turning me on to many resources in
Judaica Romanica and beyond that found their way into my work.
Most of the primary research for this dissertation would not have been
possible without access to the manuscripts at several libraries in Europe. I'd
like to thank the staff of the Oriental Reading Room of the Bodleian Library in
Oxford, who helped me with my first foray into manuscript research in the
summer of 1999, and again in the winter of 2002. I'd also like to thank the staff
of the Cambridge University Library and the Brotherton Library at Leeds

University, who enthusiastically answered my cryptic emails and seemed to
await my arrival that same winter with baited breath. Particularly helpful was
the staff of the Biblioteca Palatina in Parma (Italy), where between my broken
Italian, their broken English, and some common ground in French, we got the
job done. The same goes for the Biblioteca Nacional in Lisbon, where my brief
visit in April 2002 was nothing if not a homecoming and a test of linguistic
nerve. I am also grateful to the staff at several libraries in Paris who helped
v
me in my search for even more bizarre texts: Bibliothèque Médem, l'Alliance
Israélite Universelle, and the Oriental Manuscripts division of the Bibliothèque
Nationale de France. In my second year in France I also attended Jewish
language courses at Langues'O, and I'd especially like to thank Marie-Christine
Varol, whose Judeo-Spanish class and enthusiasm for all things Sephardic
were inspiration to try to make this dissertation much broader than it could
ever be. Moreover, her technique of getting her students to speak French like
a Turkish Jew in order to overcome their native accents in speaking Judeo-
Spanish inspired me to try the same in my French teaching at Cornell.

As much as everyone mentioned above has made completing this
dissertation possible, it's difficult to imagine even contemplating starting such
a solitary project without other equally-deluded friends and peers for support,
critique, inspiration, and comfort. On that score I cannot overestimate the role
that Tanya Matthews has played in my time in and beyond Ithaca. I'm not
sure how I'll get work done in the future without the prospect of sharing a
table at the coffee shop with her. Another linggrad who has helped
immeasurably along the way is Diego de Acosta, fellow luso-traveler and by
far the most-cited non-tenured person in this dissertation. Marisol del Teso-
Craviotto and Irene Mittelberg both started with me at Cornell in the Fall of
1997, and though Marisol managed to escape one year too soon, I certainly feel
like we've graduated together. Other Ithaca and Paris connections who have
helped at various times in various ways include Josep Alba-Salas, Edith
Aldridge, Grace An, Kenneth Beirne, Marc Brunelle, James Cisneros, Réjane
Frick, Mark Gray, Fred Hoyt, Andrew Joseph (thanks, Omar), Daniel
Kaufman, Aaron Lawson (the noted Pomeranian philologist), Sara Pappas,
Ruth Perez-Bercoff, Stanka Radovic, and Serge Ryniecki. I'd also like to thank
vi
Barbara Legendre in the Cornell Writing Center, for asking me what it was
about, and why I was writing it. And no work or even life is possible without
music to play, so I'd like to thank the members of the Cornell Steel Band,
especially Jim Armstrong and Judith Peraino, as well as the Boyz Named Sue –
Jeff Turco, Sean Franzel, Sam Frederick, and, I suppose, Blondchen.
Research into the deepest recesses of Judaica would not be possible
without the support of the Jewish Men's Rap Group (est. 1995, Oberlin, OH).
Rector-Whip J. Schwartz was in many ways responsible for setting me down
the path of academica hebraica, while at the same time nearly sidetracking me
onto a Welsh one. Sergeant-at-Arms J. Safran and Ombudsman R.M.
Goldman witnessed the birth of both those paths, and are probably more
pleased to see me on the present one. Cosigliare D. Kennemer has taken
particularly odd pleasure in some of this research, suggesting a number of
directions that no reasonable man (other than him) should pursue. And
without President pro tem J. Irving Israel I might never have located a certain
Talmudic passage, nor had the opportunity to participate in a formal process
of censure.
My parents, Sheva and Ernie, have helped me in countless ways
throughout my graduate school years, not least of which has been by curbing
their desire to ask when it might all be over. I'd like especially like to
acknowledge the role my sister Dara has played in getting me to this end of
the process. I think it's fitting that my final gesture in this dissertation be the
one thing for which I can't rightly ask her advice.
vii
TABLE OF CONTENTS
Biographical Sketch iii

Acknowledgements iv
Table of Contents viii
List of Figures xii
List of Tables xiii
List of Abbreviations xvi
Preface xvii
I. Alphabet Soup
1. Roadmap 1
2. Writing's "second movement" 3
3. Borrowing, writ large 5
3.1. Loanwords 6
3.2. Ad-hoc nativization 6
3.3. Borrowed scripts 8
3.4. Adapted scripts: permanent vs. transient 10
3.5. Hebrew script for non-Hebrew writing 12
II. From Aramea to America: Adaptations of Hebrew Script
1. From Hebrew to Jewish Writing 17
2. Evolution of Jewish Writing 20
2.1. Graphical change: final forms 20
2.2. Orthographic change: matres lectionis 22
2.3. Phonological change: spirantization 24
2.4. More graphical change: niqqud 24
2.4.1. Full vocalization 25
2.4.2. Sub-phonemics 27
2.5. Cursive scripts 29
3. The Adaptations 30
3.1. Semitic 33
3.1.1. Aramaic 33
3.1.2. Arabic 37
3.2. Romance 38
3.2.1. Spanish 38
3.2.2. French 40
3.2.2.1. Language-learning handbooks 42
3.2.2.2. Bilingual dictionaries 48
3.2.3. Italian 50
3.2.4. Provençal 52
3.2.5. Portuguese 55
3.2.6. Romanian 57
viii
3.3. Greek 59
3.4. Slavic 61
3.5. Persian 61
3.6. Turkish 64
3.7. Georgian 65
3.8. East Asia 67
3.9. Germanic 68
3.9.1. Yiddish 68
3.9.2. German 71
3.9.3. English 72
III. The Judeo-Portuguese Corpus
1. Judeo-Portuguese in context 78
1.1. Approaches to Hebrew-letter Portuguese 82
1.2. The "Real" Judeo-Portuguese 85
2. The Writing System 88
2.1. Independence from the mother script 89
2.1.1. Vowel letters 89
2.1.2. Merged segments 92
2.2. Reliance on the mother script 93
2.2.1. Final /a/ allography 93
2.2.2. ' √ as a diacritic 96
2.3. Reliance on the dominant script: Latin 98
2.3.1. /v/ 98
2.3.2. Sibilants 102
2.3.3. Classicizing spelling 106
2.4. Reliance on another dominant script: Arabic 107
2.4.1. Cognate letters 107
2.4.2. Quasi-etymological spelling 108
2.5. Imported/innovated characters 110
3. Between Transliteration and Transcription 111
3.1. Vowels 112
3.2. Semivowels 113
3.3. Begad-kefat 113
3.4. Sibilants 114
3.5. Velar stop 114
3.6. Final forms 115
IV. O libro de komo se fazen as kores (Parma ms. 1959)
1. Introduction 116
2. Overall linguistic character 118
2.1. Sibilants 118
3. Romanization 120
ix
4. Translation 131
5. Commentary 142
V. O libro de ma‹gika (Bodleian ms. Laud Or. 282)
1. Introduction 185
2. Overall linguistic character 186
2.1. r'd(y)byd' adiv(i)dar 187
2.2. Jewish character 188
3. Transcription 190
4. Romanization 218
5. Translation 239
6. Commentary 261
VI. Three Shorter Texts
1. Introduction 328
2. Bodleian ms. Can Or. 108 330
2.1. Transcription 331
2.2. Romanization 332
2.3. Translation 333
2.4. Commentary 334
3. Brotherton ms. Roth 71 337
3.4. Commentary 341
4. Cambridge ms. Add.639.5 344
4.4. Commentary 347
VII. Archaism and Vernacularism in Judeo-Portuguese
1. Introduction 351
1.1. Nouns 352
1.2. Verbs 353
2. Phonology 357
2.1. l-clusters 357
2.2. Deleted consonants 359
2.2.1. /l/ 359
2.2.2. /n/ 361
2.2.3. Other lenitions 364
2.3. r-migration 365
2.4. Palatals 370
x
2.5. oi vs. ou 371
2.6. ' a vs. y e 372
2.7. Mono- vs. diphthong 374
3. Morphology 375
3.1. Nouns 375
3.1.1. Gender 377
3.1.2. Plurals 378
3.1.3. Miscellaneous 379
3.2. Adjectives 380
3.2.1. Past participles 381
3.3. Verbs 382
3.4. Prepositions and conjunctions 383
4. Lexicon 384
4.1. Replacement 384
4.2. Romance cognates 385
4.3. Castilianisms 386
4.3.1. Hypercorrection 388
4.4. Arabisms 390
VIII. Readers, Editors, and Typesetters
1. Conventionality in the written form 393
1.1. Native and foreign scripts 395
1.2. Orthography is not transcription 397
1.3. Transcription as pseudo-spelling 399
2. Representation and Accessibility 401
2.1. Facsimile 404
2.3. Transliteration 406
2.3.1. Skeletal transliteration 407
2.3.2. Vocalized transliteration 409
2.4. Normalization 411
3. The Alphabet 415
4. Roman keyboard, Hebrew script 420
4.1. System software vs. Stand-alone font 422
4.2. Transcription vs. Transliteration, redux 427
5. A Final thought 429
Appendix: Facsimiles 432

Bibliography 437
xi
LIST OF FIGURES
8-1. Facsimile of Bodleian ms. Can. Or. 108, f. 227r. (excerpt) 404
8-2. 1935 Yiddish typewriter layout 421
8-3. Standard Israeli Hebrew keyboard layout 423
A-1. Facsimile of Parma ms. 1959, f. 1r. 432

A-2. Facsimile of Bodleian ms. Laud Or. 282, f. 1r. 433
A-3. Facsimile of Bodleian ms. Can. Or. 108, f. 227v.-228r. 434
A-4. Facsimile of Brotherton ms. Roth 71, f. 5r. 435
A-5. Facsimile of Cambridge ms. Add.639.5, f. 20r. 436
xii
LIST OF TABLES
2-1. Paleo-Hebrew and "Jewish" scripts 19

2-2. Final forms 21
2-3. Leipzig glosses with non-final forms in final position 22
2-4. Matres lectionis 23
2-5. Begad-kefat letters 24
2-6. Tiberian vocalization used in Hebraicization 26
2-7. Dagesh 'emphasis' 27
2-8. Rafeh 'weakness' 27
2-9. Sephardic cursive script 29
2-11. Ashkenazic cursive script 29
2-11. Hebraicized Arabic-Spanish glossary 37
2-12. Selected Rashi glosses 41
2-13. Hebraicized French homographs 44
2-14. Hebraicized French vowel transliteration 46
2-15. Hebrew-French dictionary entries 48
2-16. Hebraic patterns in Judeo-Provençal 53
2-17. Rarer adaptations in Judeo-Provençal 53
2-18. Portuguese learner's manual 56
2-19. Hebrew- and Roman-letter herbarium captions 59
2-20. Greek loanwords in Hebrew 60
2-21. Non-Yiddish patterns in Hebraicized German 72
2-22. Vowel-related errors in Harkavy (1893) 75
2-23. Consonant-related errors in Harkavy (1893) 75
3-1. Final /a/ allography 95

3-2. Non-initial vocalic w and y preceded by ' 97
3-3. b < Lat. /b/ 100
3-4. w < Lat. /w/ 101
3-5. b < Lat. /w 101
3-6. w < Lat./Ar. /b/ 102
7-1. Variation in word-final nasal spelling 353

7-2. Pg. pr < PL 357
7-3. Pg. br < BL 358
7-4. Pg. kr < CL 358
7-5. Intervocalic /l/ spelled in JPg. 360
7-6. Hiatus from deleted /l/ not spelled in ModPg. 360
7-7. Intervocalic /l/ restored in JPg. 361
xiii
7-8. Deleted /n/ spelled in JPg. 362
7-9. Deleted /n/ restored in ModPg. 362
7-10. Hiatus from deleted spelled in JPg. 363
7-11. Restored /n/ not spelled in JPg. 363
7-12. Coalesced hiatus spelled out in JPg. 364
7-13. Voiceless stop restored in ModPg. 364
7-14. Deleted stop restored in ModPg. 365
7-15. r-metathesis 366
7-16. r-cluster metathesis 366
7-17. /r/ in onset clusters 367
7-18. /r/ in codas 367
7-19. onset-coda r-migration 367
7-20. r-migration in prefixed PRO- 368
7-21. r-yod metathesis 368
7-22. r-l dissimilation 369
7-23. Palatalization not spelled 370
7-24. Unexpected palatal spelling 370
7-25. yyw oy for ModPg. <ou> 371
7-26. w'w ou for ModPg. <oi> 372
7-27. ' a for ModPg. e/i/o 373
7-28. y e for ModPg. a/o 373
7-29. Diphthong from vocalized consonant 374
7-30. Unetymological double vowel 374
7-31. Monophthong for ModPg. diphthong 375
7-32. wXnym -mento nouns with different form in ModPg. 375
7-33. Other noun forms for ModPg. -mento 376
7-34. '(y)sn' nouns with different form in ModPg. 376
7-35. yd'd -dade nouns with different form in ModPg. 377
7-36. yd'd -dade nouns with different suffix in ModPg. 377
7-37. Gender discrepancy 378
7-38. Variant feminine plurals 379
7-39. ModPg. unaffixed nouns 379
7-40. ModPg. different affixes 380
7-41. JPg. adjectives unaffixed 380
7-42. JPg. adjectives with alternate affixes 381
7-43. Ordinal numbers 381
7-44. JPg. participles < -UTU 382
7-45. Verbs with different morphology 382
7-46. Verbs replaced in attested meaning 382
7-47. Archaic prepositions/contractions 383
xiv
7-48. Contracted §wq kon 384
7-49. Uncontracted §y' en 384
7-50. Unmodified Latinisms in ModPg. 385
7-51. Other lexical archaisms 385
7-52. Obsolete cognates 385
7-53. Remodeled from cognate 386
7-54. Phonological Castilianisms 387
7-55. Lexical Castilianisms 387
7-56. Avoidance of a diphthong 389
7-57. Avoidance of a palatal 389
7-58. Avoidance of epenthetic /b/ 390
7-59. Phonologically-adjusted Arabisms 390
7-60. Morphologically-adjusted Arabisms 391
7-61. Replaced Arabisms 391
8-1. Pseudo-Yiddish orthographic calques 414

8-2. Portuguese Hebrew pronunciation (following Garbell 1954) 417
8-3. A Judeo-Portuguese alphabet 418
8-4. Multigraphs in Judeo-Portuguese 419
8-5. Consonant phones of medieval Portuguese 420
8-6. Phonetic mappings 423
8-7. Phonological mappings 424
8-8. Iconic mappings 424
8-9. Hybrid mappings 424
8-10. Transliteration-based mappings 425
8-11. Phonology-based mappings 425
8-12. Hebrew graphemes from Roman keystrokes 427
8-13. Roman keystrokes ‡ Hebrew graphemes 428
xv
LIST OF ABBREVIATIONS
Ar. Arabic OED Oxford English

BibH. Biblical Hebrew Dictionary
Cast. Castilian OFr. Old French
Cat. Catalan OHG Old High German
Eng. English OIt. Old Italian
Fr. French OPg. Old Portuguese
fut. future OPr. Old Provençal
Gk. Greek OSp. Old Spanish
Gmc. Germanic perf. perfect
Heb. Hebrew Pg. Portuguese
imperf. imperfect pl. plural
It. Italian PopL. Popular Latin
JAr. Judeo-Arabic pres. present
JPg. Judeo-Portuguese Pr. Provençal
JSp. Judeo-Spanish Rom. Romanian
MedL. Medieval Latin sg. singular
MHG Middle High German Skt. Sanskrit
ModE. Modern English Sp. Spanish
ModFr. Modern French subj. subjunctive
ModH. Modern Hebrew TLF Trésor de la langue
ModIt. Modern Italian française
ModPg. Modern Portuguese VLat. Vulgar Latin
ModSp. Modern Spanish Yid. Yiddish
xvi
PREFACE
While living in France several years ago, I suffered an identity crisis of

sorts. After repeated attempts to introduce myself to French people by
pronouncing my given name au naturel, it became clear that very few people,
peers and bureaucrats alike, could correctly make out this less-than-common
name, and I often found myself identified as an all-too-common David or
Kevin. In order to be sure my name would be written down correctly, I had to
pronounce the second syllable in such a way that I became known as [d´vœ'n],
[døvø'n], or, worse yet, [døvõ]. The net effect was to remind me of a tenth-
grade French teacher who, with similar difficulty pronouncing my name à

l'anglaise, could only refer to me as [møsyø divã], and as such unintentionally
coined my nickname for that year: 'Mr. Sofa'.
In the end I resigned myself to this fate of misidentity, which I was able
to attribute to the average francophone's inability to associate the unaccented
vowel and weak word-final n in the second syllable of [d´' v πn] with a
conventional way to spell that syllable in their writing system. And yet this
was not the first time that writing my name had posed an orthographic
problem. My second-grade class was once visited by a sofer stam, a Jewish
scribe trained in the calligraphic art of the sacred text. As a personalized
sample of his work, he wrote each student's name in the script of a typical
Torah scroll. My name materialized as ¶ybwlwrXs §bd, or dbn s†rwlwbyß in
letter-by-letter transliteration. To my eyes, however, something was amiss.

Why had the five letters of my first name been reduced to three? And why
had the final <tch> of my surname been fused into a single ßade, a letter that I
had been taught to read as [ts]?
xvii
Clearly there were French and Hebrew writers who had trouble
producing a written form of my name that preserved its visual identity for me.
In the case of the final consonant of my surname, the sofer did the best he
could, since the Hebrew script had no letter that normally represented [ê] –
nor, lest we forget, does the English-language use of Roman script, as the
<tch> trigraph makes clear. Looking back, in fact, I should have been flattered
that in spite of (or perhaps thanks to) scribal tradition, the sofer was more
willing and/or able to adapt a letter of his script to this foreign sound than
had been the officials at the port of Montreal who, upon hearing something
that sounded like [y\srø'yl\vπê] from my Romanian-Jewish relatives, proceeded
to insert the t that anglicized the s-r cluster. From a strictly linguistic point
view, it seems impossible to say which script – and whose use of that script in
particular – was better suited to spell my hybrid name. And yet both writers
adapted the conventions of their respective writing systems to accommodate
the written identifier of this North American anglophone. This is the issue
that I take up in the pages that follow: the convivência that emerges when
languages and scripts that are normally foreign to one another, in particular
medieval Portuguese and the Hebrew alphabet, are made to cohabitate.
xviii
CHAPTER ONE
ALPHABET SOUP
1. ROADMAP
This dissertation explores the process undertaken by medieval writers
to produce Portuguese-language texts using the letters of the Hebrew
alphabet. Through detailed philological analyses of individual texts, I focus
on the strategies for using one set of characters to write a language in the
context of more traditional adaptations, i.e. the conventional orthography. I
examine the linguistic features of this corpus in order to challenge the
conception of its writing system as marked or marginal, a view that
misleadingly identifies a language with a script and with the conventions
applied to that script in representing the language. Thus in contrast to past
work on this topic, I use the term adaptation rather than conversion, since the
latter suggests an inherent link between the source script and written
language, and implies that this script is used in a derivative manner. I argue
that the adaptation of Hebrew script for this Ibero-Romance language is not
derivative of or secondary to Roman-letter writing, nor entirely dependent

upon the conventions of written Hebrew itself. I further argue that it should
not be viewed as a process performed anew by each writer and influenced
primarily by spoken language (that is, as a transcription). The perspective I
adopt, then, rejects the premise that the patterns manifested in this
unconventional writing system are ad hoc creations by its writers, that the
writing system requires extra effort from its readers, or that it is a less "native"
system than the dominant, more conventionalized, Roman-based adaptation
2
that normally bears the title "written Portuguese." It also rejects the view that
that the language articulated in this writing system has an intrinsic Judaic
character beyond the script itself (though it does not exclude this possibility). I
argue instead that writing Old Portuguese in Hebrew script was no more
problematic an act than any written gesture in which multiple influences and
competing conventions must be negotiated.
In Chapter 2, I present a case-by-case survey of the adaptation of
Hebrew script for languages other than Hebrew, leading to a more detailed
analysis of the Judeo-Portuguese writing system in Chapter 3. The three

subsequent chapters offer critical editions and philological commentary on
texts representing the bulk of the corpus of Old Portuguese in Hebrew script.
In Chapter 4 I present a new edition of the only text to be published in edited
form prior to my own work, a manual for manuscript illumination entitled O
libro de komo se fazen as kores. Chapter 5 presents a portion of the largest
manuscript in the corpus, an astrological treatise known as O libro de ma‹gika.
Chapter 6 presents editions of three shorter Portuguese texts written in
Hebrew script, two prayer books that contain vernacular instructions for the
Passover meal along with a short medical prescription. In chapter 7 I
summarize the main contributions of this corpus to Portuguese historical

linguistics in terms of archaism and vernacularism in the lexicon. And in the
final chapter, I offer a proposal for a Judeo-Portuguese "alphabet," along with
a sketch of some further problems of adaptation and interpretation that arise
from the misconception of "native" versus "foreign" scripts and from the
process of making Hebraicized texts "accessible" (and the preconceptions of
audience and purpose on which this notion is based), as well as the
adaptations that generate Hebrew characters from Roman-letter keyboards.
3
2. WRITING'S "SECOND MOVEMENT"

According to the best estimates, spoken language emerged about
200,000 years ago (Fischer 2001); at present there are approximately 6,000
spoken languages in the world, in addition to an unknown number of once-
spoken, now-defunct languages. By contrast, there have perhaps been only a
few dozen historically-distinct sets of graphic characters that have made
possible the activity we call "writing." The oldest extant writing, in fact, dates
from only a few thousand years ago. For two aspects of language so
intimately linked, speech and writing have remarkably little history in
common.
As an autonomous technological achievement, writing seems to have
been "invented" only a few times in history. It first appeared five thousand
years ago in Mesopotamia, while its most recent appearance as independent
innovation occurred two thousand years ago in Mesoamerica. And therein
lies an incongruity: for all the writing systems that exist in the world today,
the idea to write in a society with no prior exposure to written language was
implemented only a handful of times, if that – it is arguable, as Fischer (2001)
puts it, that no one has ever independently "re-invented" writing (though
perhaps the Mayans, separated from Eurasia by two oceans, can claim to have
done so). Every subsequent instance of a language made readable for the first
time has been undertaken by someone familiar with the most basic principles
and practices of writing, someone capable of implementing these principle
and practices in order to write another language that had to that point been
invisible.
4
In this study I will not be concerned with the "principles and practices"
devised by the very first human writers or those who produced the first
written language in their cultures – however greater an achievement their
technological innovation may ultimately be. This "first movement" in the
history of written language is well-trodden terrain in terms of the data (though
ever-more ancient writing may yet be discovered) and the theories, from
ancient times through the Middle Ages and the earliest days of comparative
philology to recent standard-bearers such as Gelb (1952). It is also very much
active terrain, with many English-language book-length studies on writing

systems and the history of writing currently in print, ranging from coffee-table
art-history surveys to more linguistically-oriented studies,1 both striving to
encompass any and all forms of visible language.
What I am concerned with instead is a "second movement" in the
evolution of writing. In contrast to the first innovation, this movement has
been a more recurrent one, a scenario that has played out any time a writer has
attempted to represent elements of his language in a novel graphical form –
but, crucially, with the initial innovation of writing already known to that
writer or his community. Roughly speaking, it involves taking the units of a
writing system (be they the concrete entities we call letters or other more
abstract conventions) that have been put to use for one language and applying
them in the service another language: in other words, an adaptation of script. In
a more restricted context, this is what the earnest sofer stam – not to mention
the hapless French bureaucrats – were attempting to do using their respective
writing systems when confronted with the sound-sequence of my name.
1
These are too numerous to list, but two recent publications that represent each format
respectively and that also bear suitably emblematic titles are Christin (2002) and Rogers (2005).
5
Despite its recurrence in both the history and current practice of writing,
however, this "second movement" has received surprisingly little attention in
both popular and scholarly writing. In the following sections I outline the
different forms that this phenomenon may take, in order to single out those
that will be relevant for the present discussion of script adaptation.
3. BORROWING, WRIT LARGE

Whenever a writer sets out to compose original written material in her
language, be it 3,000 years ago or last Tuesday, there is the possibility that she
will need to make readable to her audience a word or phrase that she knows in
some way comes from beyond her language. Although the writer may know
that the word didn't first appear in the world as part of her language, for her
present purposes it does now belong, and so in principle can be written
alongside the native material. At a particular moment in time, or over the
course of many years, one or more forms for representing this word may
become conventional in the orthography that the writer has learned. In order
to arrive at this stage, however, previous writers would have experimented or
improvised with different ways to spell this foreign material before
subsequent writers (and readers) came to regard one or more of these forms as
belonging to their language. At its most fundamental, this is the cognitive and
mechanical act that constitutes the focus of this study: what has been or may
be tried graphically to make linguistic material that has been perceived as
foreign readable within another linguistic framework; what factors might
enable one form or another to be favored by an individual writer and perhaps
"win out" over time in the orthography; and what are the causes and
consequences, linguistic or otherwise, of these orthographic gestures.
6
3.1. Loanwords
Distracting from this facet of script adaptation, however, are the inert or
conventionalized loanwords. Like the bulk of the written language, these are
orthographic fossils that – notwithstanding their current utility – may convey
information (apart from semantic content) from a previous time, under a
different set of conventions and influences. Consider a simple homegrown
example: while the orthographic forms <czar> and <tsar> may tell us
something about the interaction of Slavic and English spelling conventions

that competed in some writers' minds at some time in the past, the decision for
a literate writer today to use one or the other of these forms is at best an
arbitrary choice between two established variants (unless, of course, the writer
has some non-linguistic incentive for favoring one or the other: political
motivations, total ignorance one of the variants, etc.). In any event, this type of
alternation is not my present concern.
3.2. Ad-hoc nativization

The scenario shifts, however, when the writer is dealing with the task of
putting foreign material into written form for the first time (as far as the writer
knows). I would exclude from this consideration the writing practices of
emergent literates, such as children, or adults learning to write for the first
time –!people whose experience with written language is not (yet) complete.
My intention is to focus on the experience of literate writers trying to spell
something they have never seen spelled before. It is here that loanwords
become dynamic objects and that the "principles and practices" of the writing
system actually do come into play. So when the students in my French course
7
must write down the name of a Senegalese dish that they hear for the first time
from a Camerounian French speaker, they are very much required to concoct a
spelling for this word by actively extrapolating the conventions of Roman
script at their command (as shaped by English orthography, though in some
cases conditioned by their incomplete command of French spelling patterns).
This process is not restricted to isolated words: when I want to write e-mail to
a friend from a Yiddish summer course , I must devise, more or less on-the-fly,
a way to use Roman letters to write the continuous prose of my message in a
language that I had learned using the Hebrew alphabet.

Admittedly, this last scenario is somewhat disingenuous, given that I
am familiar with prior conventions in both academic and non-academic circles
for writing Yiddish in Roman script. Yet what it illustrates remains, namely
the writing that is produced when a particular writer abandons the
conventional orthography and fuses together another set of influences. For a
more illustrative example, consider the transcriptions used in a typical
language-learning manual. Here the conventions of the learner's writing
system may be imposed on the script to represent the spoken forms of the
target language in a way that, while not necessarily on-the-fly, certainly pays
little heed to whether the forms look conventional in the orthography of either
the learner's native language or his target. Indeed, this is ipso facto the case if
the languages are normally written in different scripts. The "language
handbook" setting in fact represents an overlooked locus of borrowing on a
much grander scale, where it is not individual items that are borrowed, but
rather the graphic sheath itself.
8
3.3. Borrowed scripts

On that note, recall the distinction made earlier regarding the first and
second "movements" in the production of written language, between its initial
invention and any subsequent innovations. Thus far I have mostly been
concerned with borrowing in the usual sense, i.e. with the incorporation,
however fleeting, of elements from one (or more) language(s) into the
conventionalized writing system of another. As a heuristic I will refer to this
"receiving" system as the matrix, and to the incorporated elements as the target.
In this scenario, then, it is the target elements that are adapted to fit a matrix
defined by a writer's linguistic history and socio-political context. In such
cases the matrix and the language of composition are not usually
distinguished, even if the script is acknowledged to have been imported at
some time in the past. It is thus conventional to talk about, for example, "the
English writing system," as it would similarly be to refer to the incorporation
of target elements (i.e. borrowings) into an "English matrix."
Now recall that figure about a few dozen graphic systems in the service
of several thousand human languages. Naturally not every one of those
languages has (yet) appeared in writing, so the real disparity may not be quite
as shocking. Yet without much reflection we may take it for granted that in
order to get target elements into writing at all (whether their language of
origin had previously appeared in writing or not), the graphic matrix must be
adapted and grafted onto those target elements. Or, viewed from the opposite
perspective, it is the relatively flexible target elements that must be adapted to
fit the mould of a relatively rigid graphic matrix.
What is crucial to recognize is that in such cases we are no longer
dealing with a text written in the language conventionally identified with the
9
matrix. That is to say, the language that is being written is the target itself,
rather than just items selected from it and inserted into a predefined matrix.
Adapting the terminology of the only English-language book-length study on
the topic (Wellisch 1978), the matrix has been converted in order to write the
target. The term is somewhat of a misnomer, however, since the basic graphic
identities of the units in the adopted script are rarely modified (in the act of
"conversion," that is – the forms of graphemes may change for other reasons).
Instead these units are be adapted – perhaps augmented with diacritic
markings to capture distinctions that they do not in their conventional context

– or else they are simply not selected for use at all. The written product
resulting from this adaptation will certainly depend on whether the target was
previously unwritten (e.g. in adapting the Roman alphabet to write a Native
American language) or written in some other script (e.g. in adapting the
Roman alphabet to write Turkish) – to say nothing of the idiosyncrasies of the
individual adapter at work. But in either case the essential idea is the same:
the graphic product employing the matrix script is not a text written in the
matrix language, and may in fact bear no immediate linguistic relation to it.
To drive this point home: there is nothing intrinsically Latin about the text on
this page, for which I am using – in order to write prose in a Germanic target –
a script that is indigenous (more or less) to an Italic matrix.
3.4. Adapted scripts: permanent vs. transient

There is a further distinction in this typology of script adaptation that
applies more globally and socio-politically, albeit to only a handful of scripts.
Script adaptation as defined above has historically been the consequence of
some political event. After an invasion or political domination of some kind,
10
the language of the newly-dominated population usually comes to be written

(assuming it is maintained) in the script used by the invaders (assuming they
belong to a literate society). Thus for a population that converts (or is
converted) to Christianity, the indigenous language will sooner or later come
to be written in the alphabet of Christianity, namely the Roman script. The
same applies to Arabic script when a population becomes Muslim or
otherwise dominated by Islam, or to Cyrillic when a population is made to
follow the Eastern Orthodox church. This last case is doubly interesting,
insofar as Cyrillic script has further spread through Russo-centric but anti-
religious Soviet dominance.
Although the Romanization, Arabicization and Cyrillicization might
account for the origins of many writing systems – if not a sizeable proportion
of the gross writing generated in the world – an adaptation of script need not
be permanent. For instance, in separating the religious from the political in
the public sphere, the Turkish language went from using Arabic script (which
had itself been adapted at some point in the past under Islamic influence) to
Roman script. Even more recently, some of the former Soviet republics, which
just sixty or seventy years ago had Cyrillicized their writing systems, have
since de-Cyrillicized them. And in the most intriguing cases, ethnic/religious

factions in the community have agreed to disagree: Serbo-Croatian writers use
either the Cyrillic or Roman script depending on their religious affiliation, and
writers of Hindi or Urdu will use the Devanagari or Arabic script respectively
for what is arguably still a single language.2 Given that the history of so many
2
One could mention here the various Romanization movements around the world, whose
goals are to abandon a script adopted at some point in the past and elaborated over time in
favour of an adaptation of Roman script, which proponents usually tout as more "efficient" in
some way. This efficiency is really little more than utility, however, a deference to the political
and economic spread of Roman script.
11
written languages is marked by the adoption and adaptation of more than one
script, it is clear that the compatibility of a language and the script(s) used to
write it is not based on linguistic criteria alone. It is largely for non-linguistic
reasons that one script (and a set of conventions associated with it) may
become so closely identified with a language that their compatibility, however
fraught with difficulty, is not normally questioned by readers and writers.
Serbo-Croatian and Hindi/Urdu may be among the few modern cases
of unrelated scripts in (semi-)peaceful co-existence; yet the circumstances that
led to this situation must have been more common in the past, before the
modern-day march of nationalization and standardized written languages.
Consider in particular the situation in medieval Spain, where Ibero-Romance
languages were written in one of three scripts (Roman, Arabic, or Hebrew),
again depending on the writer's religious affiliation. The Hebrew alphabet in
particular was used by Jews to write in Hebrew, Arabic, and Spanish,
including some of the earliest-attested material in Spanish. This situation, in
which Hebrew script was used to write both the vernacular in addition to the
non-native Hebrew language, was far from uncommon in medieval Jewish
communities. The circumstances of medieval Spain would also apply to the
Jewish community (or communities) in adjacent Portugal, though records of

the practice there are not as abundant as those from Spain. Moreover, the
lusophone Jews who wrote in Portuguese after their expulsion in 1496 did so,
based on the extant record, using Roman script exclusively. The texts that
form the basis of this study – medieval Portuguese written in Hebrew script –
thus represent valuable attestations of a poorly-documented but culturally-
significant practice.
12
What makes these documents of such high interest and importance is

their utter markedness: because the Hebrew alphabet was not and is not the
usual one used by writers of Portuguese, the modern reader must abandon the
normally-unchallenged assumption that "written Portuguese" implies
"Portuguese written in Roman script." And yet it is highly unlikely that this
sort of mental exercise was required of the medieval reader. The very
existence of the texts, as well as the occurrence of the practice in other
linguistic settings, presupposes an audience with the appropriate literacy.
After all, though the matrix may be Hebrew, the target is Portuguese.
3.5. Hebrew script for non-Hebrew writing

Like the other scripts associated with major religions, Hebrew has at
various times been used by Jews to write a set of unrelated spoken languages,
whether within the matrix of Hebrew-language writing or in a wholesale
adaptation to a target language. Arabic, Greek, most of the Romance
languages, German, Persian, Turkish, and others have all been "Hebraicized"
by Jewish writers (these are briefly profiled in the next chapter). Yet the
Hebraic adaptation stands apart from the other major movements in two
crucial ways: (1) even in the earliest adaptation of what was originally a local
variant of an imperial Aramaic script, Hebraicization has never been
associated with political power, and (2) a Hebraicized form of writing has
persisted in only one modern orthography, namely Yiddish3 – no other
adaptation of Hebrew script has persisted into the twenty-first century. These
two features make the use of Hebrew script for languages other than Hebrew,
3
Although it boasts a half-millennium of texts written in Hebrew script, Judeo-Spanish
writing has been largely re-Romanized since the mid-twentieth century (cf. chapter 2 § 3.2.1).
13
and the detailed analysis of any one of its manifestations, a unique laboratory
for the study of script adaptation.
Scholars such as Goerwitz (1996: 487) have noted the distinctive place
and import of Hebrew in the history of script adaptation: "The story of the
Jewish scripts is … a great deal more than the story of sectarian orthographic
tradition: It is an important chapter in the history of writing." In fact, language
historians get rather excited by texts whose marked or unconventional script
might yield information that might not emerge from material written in a
more tradition-laden one, as Paper (1977: 103) explains in reference to

Hebraicized Persian texts:
It is an interesting and fascinating fact that the half-dozen earliest bits of

direct textual evidence that testify to the appearance of Modern Persian
are texts written ... in the Hebrew alphabet! We have, in essence, a
practically continuous chain of direct linguistic data sources for the
Modern Persian language in this particular Hebrew orthographic garb.
These early Hebrew-letter texts stand out because what became the
conventional mode for writing Persian is an adaptation of Arabic script.
Historians of Persian would no doubt be equally excited by Arabic-letter texts
containing so-called vernacular spellings – forms that flout the conventional
orthography by attempting to map the sounds of the language to the symbols
of the script as unambiguously as possible, without regard for the usual cross-
dialectal utility of standard orthography. Language historians believe that
vernacular spellings are more likely to occur in early attestations of a
language, making such documents more transparent with respect to
pronunciation. Yet they often face a dearth of vernacular texts in the
conventional script, so that attestations of the language in another script are
14
valued as more phonetically transparent than vernacular spellings found in

the unadapted "native" script. This position is implicit, for example, in de
Lange's discussion of Hebraicized Greek texts:
There is a shortage of vernacular [medieval] Greek texts that endows

these fragments [from the Cairo Genizah] with particular value, but
even greater interest attaches to the fact that they use a non-Greek
alphabet, and therefore convey the pronunciation [my emphasis]. It is my
impression that many of the words and some of the forms are otherwise
unattested (1996: iii).
Since it is assumed that there is a lack of convention for associating the sounds
of the language with the symbols of this "foreign" script, the writer's motive in
what is construed as an inherently ad hoc process must be to spell what he
hears (or what he believes should have been heard). Yet it would be
misguided to characterize de Lange as claiming that the non-Greek script was
used as a phonetic transcription of the language at that time; such a system is
not likely to have served the needs of skilled readers, nor is it likely to have
been produced by skilled writers (Wright 1982). Thus it cannot be assumed
that every orthographic variant corresponds to a phonological one. Similarly,
it has been claimed, particularly in reference to Judeo-Romance writing, that
any Hebrew words found in Hebraicized writing must be considered

unassimilated in the spoken language because scribes do not alter their
spelling to conform with the system used for the vernacular part (Freedman
1972). Yet skilled readers of many languages are accustomed to dealing with
multiple or competing conventions, and they usually pronounce such words
in perfect harmony with the orthographically-assimilated ones. That is to say,
unconventionality is not equivalent to absence of convention.
15
In his study, Wellisch (1978) devotes a short section to "Hebraification,"

and notes the unique aspects of the Hebraic tradition:
[T]he Hebrew script, the means of graphic expression for one of the
smallest ethnic groups in the Roman Empire, survived not only the
scripts of other more numerous and powerful nations but also the
empire itself, and it attained a status on a par with the two alphabets
that had also developed from the North Semitic one (p.120).
Hebrew is probably the only script that has been used to write
languages other than the one for which it was originally devised
without any coercive system of a religiopolitical nature (p.121).
These are indeed apt observations. What he appears to mischaracterize,

however, is the genesis of this tradition:
[It was used]… initially to record foreign names (for which there
already existed a long-standing tradition in the Talmud for rendering
Greek, Latin, Egyptian, and other names in Hebrew transcription). This
practice was later extended to words and short phrases in other
languages until it was quite natural to write the vernacular entirely in
Hebrew characters (p.121).
On the one hand, this description creates the impression that as Hebrew texts
began to fill with loanwords, they simply morphed into another language – a
gradual relexification of sorts. Yet as I described it above, the adaptation of a
script is a discrete, not continuous, process. Thus, for example, no matter how
many French words Chaucer and other Middle English authors inserted into
their writing, their language of composition was never French (Thomason and
Kaufman 1988). If, on the other hand, Wellisch's description implies that by
inserting vernacular loanwords into Hebrew texts, Jewish writers could
16
extend the principles they devised for writing these individual items to longer
passages in other vernaculars, it is difficult to account in general for many of
the widely-varying patterns that characterize the earliest Judeo-Romance
writing, and in particular for the (apparent) absence of "words and short
phrases" in Portuguese within Hebrew texts prior to the emergence of a fully-
Hebraicized Portuguese writing system.
Wellisch's description also mischaracterizes what was "natural" about
extending the use of Hebrew script to other languages. It was not the
orthographic dexterity they developed that persuaded Jews to write in

Hebrew script, but rather the very nature of writing and (religious) identity in
medieval Europe. The corpus of Old Portuguese in Hebrew script confronts
the researcher with a well-elaborated product of script adaptation, carried out
without coercion, but which was not maintained into modern times. This is no
doubt related to Hebrew's lack of association with political sovereignty in
general, and the fate of Portuguese Jews in particular. And yet there remains a
writing system used by highly-skilled writers and digested by highly-skilled
readers. It is almost inconceivable, in fact, that these writers and readers were
not multi-literate, given how un-Hebrew-like this writing system usually is.
Furthermore, the unconventional alphabet is the only markedly Jewish feature

of much of the corpus – a feature that only serves to illustrate how Portuguese
Jews were integrated members of their linguistic community. This Portuguese
corpus shows none of the marginality and parochialism that tends to be
associated with some Jewish languages (or at least some studies of Jewish
languages). What it does express is an interest in and a market for intellectual
exchange, transmitted through a convivência of language and script that within
a century would disappear from the Iberian peninsula.
CHAPTER TWO
FROM ARAMEA TO AMERICA:

ADAPTATIONS OF HEBREW SCRIPT
In this chapter I present an overview of the development of the
Hebrew writing system, followed by a survey of language families with
attested Hebrew-letter texts. While I aim to provide a broader and more
inclusive overview of Hebraicization than has been available previously, I do
not make any claim to comprehensiveness.
1. FROM HEBREW TO JEWISH WRITING
Nothing is known of Hebraic writing before the Israelites emerged in
the land of Canaan and "borrowed the art of writing" from the local
inhabitants in the twelfth or eleventh century BCE (Naveh 1982: 65). In the
earliest known Hebrew inscription, the Gezer calendar,1 the writing resembles
that of tenth-century Phoenician inscriptions from Byblos, and features no
specifically Hebrew characters. Indeed, the Phoenician influence was so
dominant that neither the Hebrews nor the Aramaeans ever innovated new
characters to represent consonant phonemes that did not exist in Phoenician.
The first distinctive features of Hebrew writing are actually to be found
in ninth-century inscriptions in Moabite, a Canaanite dialect related to
Hebrew. According to Naveh (1982), these adaptations of the contemporary
Hebrew script represent the first stage of the Hebrew scribal tradition.
Despite dialectal differences between the spoken Hebrew of Judah (the
1
Naveh notes that although the calendar can be dated to the late tenth century, the language
of this inscription "does not have any lexical or grammatical features that preclude the
possibility of its being Phoenician" (1982: 76).
southern kingdom) and Israel (the northern kingdom), the same script was
used in both kingdoms, as well as by the Moabites and Edomites to write their
own kindred languages while under the rule of Israel and Judah. It appears
that there were no local variants of this script, nor was there a distinct non-
cursive lapidary style, due perhaps to the lack of a widespread custom in
Hebrew society of erecting royal stelae or offering votive inscriptions to the
deity (Naveh 1982: 69).
Naveh describes the development of Hebrew script through the
seventh century BCE as "a single, conservative, national tradition of writing"
(1982: 78). However, the destruction of the First Temple in the early sixth
century and the subsequent exile of most of the educated class to Babylonia
resulted in a major shift in Hebraic writing. By this period, Aramaic had
replaced Akkadian as the everyday language in Babylonia, and it would gain
even greater prestige and wider usage when it was elevated to one of the four
official languages of the Persian Empire (along with Persian, Elamite, and
Akkadian). Over the succeeding centuries, use of the "native" Hebrew script
became more and more restricted, its latest known use being on the coins of
2
Bar-Kokhba in the second century CE. From the late third century BCE
onwards, the Jews – now comprising a sizeable number if not a majority of
Aramaic speakers – wrote in a script derived from Aramaic writing. It is the
characters of this script, also known in paleographical circles as "the Jewish
script," that evolved into what is now commonly referred to as the letters of
the Hebrew alphabet. These are shown in the middle line of table 2-1, with
their equivalents in the pre-Aramaicized native Hebrew script (generally
2
Although the native Hebrew script persisted among the Sadducean sect in the Second
Temple period, it seems that it was ultimately rejected in favour of the adapted Aramaic script
because it came to be identified with the Samaritans (Yardeni 1997: 44).
referred to as "paleo-Hebrew") in the line above, and the transliterations
traditionally used in Semitic philology3 below:
Table 2-1. The Paleo-Hebrew and "Jewish" scripts
) b g d h w z x + y k l m n s ( p c q r $ t
' b g d h w z x X y k l m n s v p c q r S t
√ b g d h w z ˛ † y k l m n s fi p ß q r s¸ /©s t
Naveh (1982: 112) emphasizes the extraordinariness of this shift: "the Jews, a
conservative nation which adhered strictly to its traditional values, abandoned
their own script in favour of a foreign one." Even the Babylonian Talmud (the
wide-ranging compilation of Rabbinic explicating Jewish law and ritual, dated
to ca. 500 CE), comments on the graphical shift, referring to the newer script as
Ashurit 'Assyrian':
Originally the Torah was given to Israel in Hebrew characters and in

the sacred [Hebrew] language; later, in the times of Ezra, the Torah
was given in Ashurit script and Aramaic language. [Finally,] they
selected for Israel the Ashurit script and Hebrew language, leaving the
Hebrew characters and Aramaic language for the hedyototh4 (Sanhedrin
21b).
3
Several of the transliteration characters differ from the corresponding symbols used in the
IPA, as shown in the table below. Unless referring specifically to a spoken form, however, I
have preferred the traditional symbols in this study.
Transliteration vs. IPA symbols

√ ˛ fi † ß s¸ s©
IPA ÷ ˛ ¿ t¿ ê¿ ß ˚
4
The passage goes on to cite R. Hisda's explanation that the hedyototh refers to the "Cutheans,"
that is, the Samaritans. As G. Rendsburg (p.c.) has pointed out to me, this term derives from
Gk. idiothV, the ultimate source of Eng. idiot.
Relative to the later adaptations of this script to languages beyond Hebrew
and Aramaic, the shift was a relatively simple one: the scripts were
genealogically related and could be substituted directly letter-for-letter. Their
relatively easy co-existence is illustrated in some of the Dead Sea scrolls (ca.
first century BCE ), whose scribes generally wrote using the newer Hebrew
script but still wrote the tetragrammaton in native Hebrew characters.
Nevertheless, the shift from paleo-Hebrew writing to the Aramaic-derived
script does represent the only time that Jews would use a borrowed script to
produce original writing in the Hebrew language.5
2. EVOLUTION OF JEWISH WRITING
Once Hebrew was no longer the sole Jewish vernacular, the need arose
to write other vernaculars in an identifiably Jewish way – in other words, to
"Hebraicize" them. Before this practice would mature, however, the Hebrew
language and its writing system underwent a number of changes that would
strongly inform the way in which its readers and writers interacted with its
alphabet. Below I outline the major grapho-phonological changes that
affected the Jewish variant of Aramaic writing that has come to be known as
Hebrew script.
2.1. Graphical change: final forms
In Semitic scripts that evolved from a cursive tradition, such as the
Nabataean script (and its descendant, the Arabic script), almost every letter
has a different form for medial and final position. The Jewish script, by
5
Of course it was far from the only time that Hebrew texts would be written in a "non-native"
script: as Wellisch (1978) argues, the very origins of Western transliteration practice can be
found in non-Jewish attempts to render the text of the Bible in other scripts and languages.
contrast, which developed from an Aramaic book-hand, has only five
differentiated final forms, as shown in the table below:
Table 2-2. Final forms
Non-final k m n p c
Final ¢ £ § • ¶
Letter name kaf mem nun pe ßade
Although these five final forms have traditionally been treated as something
to be learned in addition to the basic set of twenty-two Hebrew letters, they in
fact more closely resemble the original graphemes, and it is the medial forms
that represent the innovated characters. Naveh (1982: 172) describes their
origin:
In the Persian period, kaf, mem , nun, pe and ßade were written with long
downstrokes. With time, these downstrokes began to shorten and to
curve toward the next letter in the work, eventually evolving into the
medial forms. However, at the end of a word, the writer slowed down,
and did not curve the downstroke of the last letter, so that the long
downstrokes survived in final forms.
The account varies slightly in the case of medial mem. What is worth noting in
relation to Hebraicization is that all later adaptations of Hebrew script deploy
the final forms to some degree, though on occasion the medial forms do occur
in final position. The paleographer's rule of thumb: the longer the text, the
greater the consistency in the use of the final forms. It is not surprising, then,
to see medial forms used in final position in the marginal and intralinear
glosses of medieval Biblical texts. The following are some of the German
terms in the trilingual (Hebrew-French-German) Leipzig glossary (Bannit
1995), each of which contains a medial form used "incorrectly":
Table 2-3. Leipzig glosses with non-final forms in final position
kamgÕ nÕ Ù' §w&b ‹bon ongem aƒ 'from sorrow'

k
ÃJ r
Ì yeww lÙ&b r˚&bl˚zÃXyim mitzul‹bur ‹bol verk 'all in silver'
m˚Jd'&b a §Ù&b bon ‹badum 'of a thread'
§yi' nÕ ˚SyEl yz #nw un' zi le¸sun in 'and they left him'
p¯lÙw volf 'wolf'
Lpyil r˚XyJib bi†ur lip 'bitter soul'
cyÂrq ¯ §˚d dun kreß 'à l'enceinte'
It would actually become standard practice in modern Yiddish orthography to
use a non-final Kp to spell word-final /p/ (e.g. KpA'q kop 'head', Kpra'q karp 'carp').6
However, in no tradition of Hebrew-letter writing have the final forms been
used in any position but at the ends of words.7
2.2. Orthographic change: Matres lectionis 'mothers of reading'
Although the original Phoenician script was a purely consonantal
writing system, by the Punic era (after the fall of Carthage in the second
century BCE ) several letters were being used to represent vowels in that
language. This practice in fact dates as far back as the thirteenth century BCE,
when North Canaanites in Ugarit used their letter yod in certain limited
instances to represent /i/ (Naveh 1982: 183). Even the very earliest
inscriptions in Hebrew and Aramaic show at least some use of a set of letters
6
The final form is maintained for word-final /f/, e.g. •yX tif 'deep', •lA'ww volf 'wolf'. The
equivalent issue does not arise for word-final /k/, since q q is used in all positions.
7
A French-language handbook for Yiddish speakers (Bibliothèque Médem 15237; see § 3.3
below) contains the only exceptions to this rule that I have yet encountered.
to indicate vowels, usually in final position. These letters became known in
the Hebrew grammatical tradition as h'yrq twm' (ModHeb. emot kria), the
'mothers of reading':
Table 2-4. Matres lectionis8
Letter h w y
Transliteration h w y
Vocalic value a o e
e u i
The current distribution of the matres in the Biblical Hebrew canon became
(relatively) fixed during the first and second centuries CE. In Modern Hebrew,
the constant need for innovated spellings of borrowed and newly-coined
words has made it conventional in unvocalized writing to represent most non-
low vowels with a mater (y for /i/, w for /u/ and /o/) but to spell /a/ and /e/
only in final position, and almost exclusively with h h.9 In the intervening
millennium and a half, nearly all adaptations of Hebrew script for languages
beyond Hebrew have made use to some extent of these vowel-letters – that is
to say, no Hebraicized orthography ever reverted to a purely consonantal system of
writing.
8
The letter ' √, which comes to play a vital role as a vowel letter in adaptations of the script
beyond the Hebrew canon, is not included in this table since it was only rarely used as a mater
lectionis in Hebrew writing of the day.
9
As a window onto at least one writer's practice, an Israeli guidebook to Paris that I picked
up while living in France offers snrpnwm-hd r'g <gar deh-monparnas> Gare de Montparnasse
but r'z'l-Xns r'g <gar sent-lazar> Gare St. Lazare, along with museums whose names range
from a mater-less qzlb <balzak> Balzac to a fully-vocalized §'Xwrmr'm <marmotan> Marmotin,
with h'wrqld <delakruah> Delacroix and hlbnrq <karnab aleh> Carnavalet in between. For the
treatment of this issue in the Hebrew Language Academy see Weinberg (1985).
2.3. Phonological change: spirantization (begad-kefat)
A different kind of change affected the Hebrew language more broadly
through the Second Temple period (second half of the first millennium BCE).
Six of its consonant phonemes, namely the non-emphatic stops, developed
fricative allophones in postvocalic environments; the immediate effect in
relation to the written language was that the corresponding graphemes now
had two realizations:
Table 2-5. Begad-kefat letters
Letter b g d k p t
Stop b g d k p t
Spirant v © ∂ x f †
This phonological change (also referred to in Hebrew grammar using the
acronym formed by the implicated letters, begad-kefat) did not affect the
contemporary writing system per se.10 It was, however, to have a major
impact on post-native traditions of Hebrew pronunciation and, of course, on
later adaptations of the script.
2.4. More graphical change: niqqud 'pointing'
Although Hebrew continued to serve Jews as a vernacular in the period
following the sixth-century BCE Babylonian exile, it ceased to be learned as a
native language by ca. 250 CE and, in a manner of speaking, "died out." Yet it
10
Nor should it have been expected to. Written English is notorious among standardized
orthographies for not reflecting phonological change, be it flapping, voicing assimilation,
velar softening, etc. In fact, a more apt comparison would be a subset of spirantization as it
applies to , <d>, and <g> in Modern Spanish orthography. Earlier in history, of course,
changes of this kind in Latin phonology did come to be reflected in the very spellings that
distinguish some Spanish words containing , <d>, or <g> from their Latin etyma (e.g.
VITA vs. vida).
very much persisted as a second or non-native language in virtually all Jewish
communities, where it continued to be read and recited in the sacred texts, a
practice that endures right up to the present.11 As an inevitable result,
however, its pronunciation among different communities became influenced
by the vernacular(s) of those communities. Since the orthography of the
sacred texts did not indicate all of the phonological details required for them to
be pronounced correctly (that is to say, as they were presumed to have been
in Biblical times), several communities were compelled to devise systems of
diacritics that could be added for this purpose – without altering any of the
existing text, which was prohibited. Of the three known systems, only the
ninth-century Tiberian system remains in widespread use, and is detailed
below.12
2.4.1. Full Vocalization13
Given its origins as a consonantal Semitic script, the most important
innovation of the Tiberian system was to institute a complex set of diacritics
(mostly dots and dashes) to indicate various vowel distinctions. They are
given below based on their usual phonetic realization in Modern Israeli
Hebrew (m m serves here as the consonant required to bear the niqqud):
11
The only exception to this rule is Ethiopian Jewry, where Ge’ez was used for liturgical
purposes (G. Rendsburg, p.c.).
12
Along with the linguistically-instructive diacritics described below, the Tiberians devised
an even more complex system of signs to indicate the stress and musical motif associated with
individual words in the recitation of canonical texts. These signs, however, have never been
deployed in any adaptation of the script (or even in non-canonical Hebrew texts), since they
serve no orthographical or strictly linguistic purpose, and so are not discussed here.
13
Strictly speaking, this term can refer to any orthographic method of indicating vowel
phonemes, in this case either using diacritics or writing vowel letters. Nevertheless, since
there is no mature Hebraicized writing system that does not make use of vowel letters, I will
use it with specific reference to "pointing," i.e. vowels indicated with niqqud.
Table 2-6. Tiberian vocalization used in Hebraicization
a e i o u ¥/Ø14
supra-linear Om
intra-linear 15
Ùm ˚m
sub-linear Am em im um ¸m
am Em
‹m ¤m
This system, which was rigorously preserved in canonical Hebrew texts and is
still fully deployed in liturgical, poetical, and pedagogical writing, was also
adopted to varying degrees by writers who adapted Hebrew script to write
other languages. It tended to be used most consistently in writing associated
with a religious context: biblical glosses, ritual prescriptions, etc. In the one
present-day Hebraicized orthography, Modern Yiddish, only a small set of the
diacritics is used in lexical items of non-Hebrew origin: a ' and A ' denote /a/
and /o/ respectively, while ˚ and iy occur occasionally to denote /u/ and /i/
respectively.16 Combinations of unpointed matres lectionis serve to spell the
rest of the vowels in most Hebraicized orthographies.17
14
In Modern Hebrew the "null" value of this vowel indicates that the consonant is either
syllable-final or part of an onset cluster.
15
The w that follows the consonant is in practice obligatory for the intra-linear vowels; in fact,
only the vowels borne by w may appear graphically in word-final position, with other word-
final vowels followed by a "supporting" ', h, or y (if none was already present for an historical
/√/, /h/ or /y/).
16
In fact the ˚ grapheme competes with w', which is preferred in some Yiddish traditions
because it avoids three consecutive vavs in the spelling of a /vu/ syllable, e.g. ˚ww/w'ww 'where'.
17
Yiddish remains the one innovator in the respect, having graphemicized v , historically a
voiced pharyngeal fricative but often equivalent to ' (as [÷] or Ø) in the pronunciation of
medieval European Jews (and in Modern Hebrew), as the letter representing /e/ in non-
Hebrew words. It does, however, compete in early writing with y, and continues to alternate
in the practice of some writers with (y)y.
There remains some debate amongst scholars as to the exact phonetic
values indicated by the Tiberian vocalization system. What is most important
to note here is that this system, too, evolved amongst communities using
different vernacular languages, so that the values associated with each of the
signs varied as well. When it came time to apply Tiberian pointing to spelling
the vowels of a language other than Hebrew, this variation played a large
role, as the examples in section 3 below will illustrate.
2.4.2. Sub-phonemics: consonant allophony
While the vowel diacritics served in many cases to disambiguate the
pronunciation of certain homographs by providing phonemic information,
there was another pair of diacritics devised to indicate the wholly predictable
stop/spirant distinctions. In fact, rather than instituting only one sign to
indicate one of the allophones, the Tiberians adopted a unique diacritic for
each variant:
Table 2-7. dagesh 'emphasis' = stop18
b
J gC d
J k
J p
K Jt
b g d k p t
Table 2-8. rafeh 'weakness' = spirant
b
& &g d
& k
& p
& &t
v © ∂ x f †
18
The dagesh is also used with most other consonants to indicate gemination, be it lexical or
grammatically derived.
Note that although the occurrence of stop-versus-spirant is predictable in
Hebrew words, it may not be so when these letters serve to spell the
consonant phonemes of other languages. Once again, however, the extent to
which these diacritics are deployed in adaptations of Hebrew script varies. In
Yiddish writing, for example, some writers use the rafeh to indicate the spirant,
while others follow the Hebrew system, leaving the spirants bare and
marking the stops with a dagesh.19 And just to make life easier, some writers
dispense with indicating the distinction altogether. Moreover, the dagesh is on
rare occasion used in Hebraicized writing to indicate the non-stop variant, e.g.
JPg. §'‡rJAbal la b:a ran '(will) wash' (ModPg. lavaram) in the Bodleian Passover text
(chapter 6 § 2) – in essence (if not an outright error), serving to indicate simply
that the letter is not to be assigned its default stop value.20
2.5. Cursive scripts
The basic form of the modern Hebrew script is usually referred to (in
English) as "square" Hebrew, reflecting the fact that it developed, as noted
above, from an official Aramaic book-hand. Although its form has varied
slightly among the many scribal traditions of post-Biblical Hebrew, it has
remained remarkably consistent and recognizable over the centuries.
Nevertheless, there have been several cursive scripts based on the book-hand
that developed in various periods and places, two of which retain a modern
use.
19
The Forverts newspaper, the only Yiddish-language weekly still published in America, uses
both strategies and only leaves the relevant letters bare in Hebrew words
20
The rafeh is also on occasion used against its prescribed value to inidicate a stop, e.g.
'‡r'Ayy¯ly&ip <filyara> and '‡r'a'ly&p <filaara> 'will take' (ModPg. pilhara), both of which occur on
the same folio of the Brotherton Passover text (see chapter 6 § 3). The use of rafeh in this text in
fact seems to be rather indiscriminate, occurring on many a d d as well as ' √ and h h in the
Portuguese passages.
Originally devised in Italy, the most widespread of the medieval cursive
scripts is a Sephardic one that came to be known as Rashi script, named for the
renowned twelfth-century French Biblical exegete. Although there is no
evidence that Rashi himself used the script, it has been used consistently to
print his commentaries, which traditionally have been included in most
printed Hebrew editions of the Bible and the Talmud.
Table 2-9. Sephardic "Rashi" cursive
A B C D E F G H I J L M O Q R S U W X Y Z ,
K N P T V
√ b g d h w z ˛ † y k l m n s fi p ß q r ¸s/©s t
The other major cursive script is a derivative of an Ashkenazi cursive, which
began to evolve in Central and Eastern Europe in the sixteenth or seventeenth
century. This is the script that remains in use as the normal longhand for
writing Modern Hebrew (shown here with Modern Israeli Hebrew phonetic
values):
Table 2-10. Ashkenazic cursive
a b g d h v z x t y k l m n s e p j q r w [
K M N F J
÷ b g d h v z x t y k l m n s ÷ p ts k r ß t
^ v x ^ f s
Of the two, only the former Italian-derived cursive would find extensive use
(from an early date) in texts written in languages other than Hebrew; in fact it
was maintained as the preferred typeface for printed Judeo-Spanish21 until the
re-Romanization of its writing system in the twentieth century.
These, then, are the historical variations that influenced Hebrew
writing, and consequently informed the nature of Hebraicization and the
many ways in which Hebrew script was adapted to write material in
languages beyond Hebrew. The next section presents a survey of most of the
linguistic contexts in which the letters of the Hebrew alphabet served to write
other Jewish vernaculars.
3. THE ADAPTATIONS
In his survey of Hebraicization through the years, Wellisch (1978)
organizes his discussion by focusing on (1) Hebrew in the Land of Israel and
the Diaspora, (2) Yiddish, and (3) Ladino (Judeo-Spanish). In a sub-section on
"Hebraification22 of other languages," he includes the following:
Judeo-Arabic
Judeo-Persian
Judeo-French
Judeo-Provençal
Judeo-Italian
Judeo-Greek
Judeo-Tat
Karaite Turkic
21
Judeo-Spanish writers further developed a distinctive longhand known as solitreo, which is
still in use among some Turkish Jews (Varol 1998).
22
I have chosen the term Hebraicization rather than Wellisch's Hebraification because, based
on its etymological components (the verbal suffix -ify, ultimately derived from Latin FACERE
'make'), the latter implies that the language written in Hebrew script has been "made
Hebrew" in some aspect beyond the letters of its alphabet. And as argued in the previous
chapter, this risks a serious misconception of the process involved in adapting the script.
Based on my own research into languages that have been written using
Hebrew script , this is a curiously incomplete cast of characters. Although he
makes no claim to comprehensiveness, Wellisch does omit at least three
attested traditions: Judeo-Portuguese (the subject of this study), Judeo-Catalan
(attested by a small number of medieval texts; see Wexler 1989), and Aramaic.
The omission of this last one is especially noteworthy, since Jewish Aramaic
represents not only the earliest adaptation of "the Jewish script" to a language
other than Hebrew – that is, as a re-adaptation of Aramaic script using the
conventions instituted to write Hebrew – but also one of the few Hebraicized
traditions whose texts continue to serve an active role in Jewish communities23
(primarily liturgical). Wellisch's hierarchy also places Judeo-Arabic as merely
one among the minor "other" languages, even though it probably represents
the most expansive pre-Yiddish Hebraicization in terms of sheer volume of
writing.
More significant from a linguistic perspective, however, is that in his
grouping of three traditions separate from the rest, Wellisch ignores an
important distinction that he himself takes pain to emphasize throughout his
book, and one to which I have sought to adhere. He conflates the very
different natures of the two basic environments discussed in chapter 1 in
which linguistic material may be adapted to fit the characters of a given script:
a. loanwords (cf. chapter 1 § 3.1) or ad-hoc nativizations (cf. chapter 1 §

3.2) within a given matrix
b. stand-alone adaptations (cf. chapter 1 § 3.3, 3.4) that yield a new
writing system, what Wellisch calls the "conversion" of scripts
23
In addition, there remains a modern Jewish Neo-Aramic dialect spoken among Jews from
Kurdistan (see § 3.1.1 below).
In the context of Hebraicization these can each be elaborated further:
a'. transcription, where the goal is to represent the spoken form of

items from languages other than Hebrew using conventions of
Hebrew orthography
b'. transliteration, where the goal is to adapt the (conventional) values
of Hebrew letters for use as orthographic system in writing a
language other than Hebrew
Since my goal has been to focus on wholesale adaptations of the second kind, I
have largely ignored transcriptions of type (a) within particular traditions,
unless they serve the goal of illustrating aspects of a particular Hebraicization,
or in cases where the only Hebraicized material I have found is embedded
within the matrix of a Hebrew- or Yiddish-language study. Items of type (a)
can in fact be found in throughout modern printed literature: in the press,
tourist guidebooks, language-learning materials, etc. Hence they are more
usefully studied in relation to Modern Hebrew or Yiddish writing per se,
rather than as illustrations of Hebraicization. Nevertheless, some instances of
type (a) may be mentioned to illustrate particular points in specific traditions.
3.1. Semitic
3.1.1. Aramaic
During the last half of the first millennium BCE and during the first few
centuries CE, Aramaic emerged as a lingua franca in the Near East. It was, for
example, one of the four official languages of the Achemenid Empire of
ancient Persia (539-333 BCE), along with Persian, Elamite, and Akkadian. The
Jews were not immune to this development, and over the course of time
more and more Jews began to use Aramaic – first in Babylonia and other
eastern communities, and eventually in the land of Israel as well. This
development resulted eventually in Hebrew dying out as a native, spoken
language ca. 300 CE.
While Hebrew continued to be used for liturgical purposes, Aramaic
was its rival even in this arena. For example, during the post-biblical period,
various prayers were composed in Aramaic, translations into Aramaic of the
various biblical books (known as Targumin) were produced, and most
importantly large chunks of the two Talmudim (the Babylonian Talmud and
the Jerusalem Talmud) were written in Aramaic. In all of these instances,
Aramaic was written in the Jewish, i.e. "Hebrew," script. With the decline of
Jewish population centers in the Middle East through the second millennium it
too declined in use, though spoken dialects (Jewish and non-Jewish) have
survived.24
The first – though not, chronologically speaking, earliest – appearance
of Aramaic in the Jewish literary canon occurs in chapter 31 of the book of
Genesis (31: 47), where Laban is said to use an Aramaic name for what Jacob
calls dvlg galfied:
(1) 'At˚d‹hW
A r¬gyÕ §Abl
A Ùl '‡rq
¯ y« w¬
wayiq¥r‹a√ l‹o l‹ab‹an y¥gar s‹ahad‹ut‹a
'Laban called it Jegar Sahadutha '
24
Rather than the decline of Jewish communities in the Middle East, what more specifically
led to the decline of Aramaic was the replacement of Aramaic (in some cases rapid, in other
cases gradual) by Arabic after the Muslim conquest of 630-640 CE. This left only the Jews of
Kurdistan speaking Aramaic into the twentieth century, and now that all of them have moved
(mainly to Israel, some to the U.S.), it is doubtful that any Jews will speak Aramaic as a native
language within another generation or so (it continues, however, to be used among various
Christian communities throughout the Middle East, most prominently in Kurdistan, and by
Mandeans in Iraq and Iran).
These words probably represent the first deliberate representation of non-
Hebrew items in the Old Testament. As a closely-related Semitic language, the
biblical writers (or codifiers/scribes, at any rate) appear to have had little
difficulty in adapting the conventions of Hebrew orthography – itself based
on an adaptation of Aramaic script – to Aramaic language. In fact, the ninth-
century Tiberian Masoretes, whose orthography constitutes canonical Hebrew
spelling, made no special provisions to distinguish material in Aramaic from
Hebrew, treating their writing systems as one and the same. Thus canonical
Jewish Aramaic writing is, in a strictly graphical sense, identical to Hebrew, in
that the inventory of letters and diacritics, and their grapho-tactic deployment,
are one and the same. This is further illustrated in the example below, the
opening line from the kaddish (Aramaic SyJÊdaq qad:ı¸s 'holy'), part of the daily
synagogue prayers (first mentioned in the sixth century CE but composed, or
perhaps compiled, earlier):
(2) 'Jb
A r
fi h
J m
E H
Ã HJd
fi q
a t
Ÿ yi wÃ lJd
fi gaC t
Ÿ yi
yitgad:al v¥yitqad:as ¸s¥meh rab:a
'May His great Name grow exalted and sanctified'
In this sample, Judeo-Aramaic presents some orthographic patterns not
J E eh in the word for 'name'.

commonly found in Hebrew spelling, such as the h
Nevertheless, the Masoretes' orthography allowed for such "extensions" of the
system, and as such they set the precedent for the flexibility of canonical
Hebrew spelling to be adapted to less easily-integrated items.

3.1.2. Arabic
In terms of sheer volume no language beyond Hebrew has made more
extensive use of the Jewish script than Arabic, whose Jewish speakers have
written a Hebraicized form of Arabic since at least the ninth century CE.
Unburdened by the strictures that required Muslim writers to adhere to a
highly standardized classical language, arabophone Jews wrote extensively in
colloquial Middle Arabic (with greater and lesser degrees of classicizing
features), including some of the hallmarks of medieval Jewish literature.
According to Hary (1996), Judeo-Arabic writing has gone through three basic
orthographic phases: Phonetic (8th-10th c.), Arabicized (10th-15th c.), which is
distinguished by imitation of classical Arabic spelling conventions, and
Hebraized (post-15th c.), which is characterized by a closer phonetic
representation but with some Arabic spelling conventions replaced by
analogues from written Hebrew. Judeo-Arabic remains a living dialect in
North Africa and Israel, where it continues to be written in Hebrew script.
A well-known example of classical Judeo-Arabic writing is Yehuda Ha-
Levi's Book of the Kuzari (Spain, twelfth century), a defense of Judaism that
takes the form of a dialogue between the author and the eighth-century
Khazar king:
(3) ywlh hdwhy d $ •yl't lyl$dl' §ydl' r$cn yp lyldl'w drl' b'tk
kt√b √lrd w√ldlyl py n„ßr √ldyn √l„dlyl t√lyp d
„ yhwdh hlwy
'Book of argument and proof in defense of the despised faith of Yehuda
HaLevi'
The most immediately striking feature of the Judeo-Arabic sample above is
the way in which the letters ' √, w w, and y y imitate almost exactly the use of
the cognate letters of Arabic script, « alif, Ë w‹aw, and v y‹a√, occurring only to
spell phonemically long vowels. In addition, to spell sounds that are entirely
unknown in Hebrew, Judeo-Arabic mimics the use of the superscribed
diacritic in Arabic script that modifies the reading of certain letters, placing a
similar dot over the cognate Hebrew letters (in the example above d d and c
ß, in imitation of œ d‹al and ’ ß‹ad respectively) – even if the normal Hebrew

realization of the unaugmented letter differs from that in (classical) Arabic.
More modern Judeo-Arabic writing, though reduced in scope, shows
the same characteristics, such as the following excerpt from a modern folk-tale
used in a Judeo-Arabic course I attended in France:
(4) qyrXl' yp yllcy §'k ydl' dysx dx'w yp hSvm

mfi¸sh fy w√˛d ˛syd √ldy k√n yßlly fy √l†ryq
'(a) tale of one righteous man who was praying'
The words in bold are Hebrew-language items borrowed wholesale into the
text: the first is a term introducing a story, which is used this way in Hebrew
and other Jewish language traditions, while the second is the name used
originally (in Rabbinic literature) to refer to an individual who maintains a
higher standard of moral and religious observance. As is conventional in
Hebraicized orthographies, both Hebrew words appear with no alteration,
even though their pronunciation or cognate forms in Arabic might demand
otherwise.
Interlude: Judeo-Arabo-Spanish
As a transition to the next most important language family for which
the Hebrew script has been adapted, below are selections from a fifteenth-
century Arabic-Spanish glossary (Sheynin 1982).:
Table 2-11. Hebraicized Arabic-Spanish glossary
sasnabrFg« ¶mwxl' sAby≈w ¶ybl'

g rbansas
i
al˛umß 'chick pea' webas albeß 'egg(s)'
h¬n'asnÕ m i x'ptl' h¬w'
a 'ml'
i a a
m nsan h altpa˛ 'apple' aw h alma 'water'
hEStEl §bll' ˚r'Eb blkl'
let¸seh allabn 'milk beru alkalb 'dog'
hasyEbk aJ s'rl' ˚r˚' bhdl'
a a
k bes h alras 'head' uru aldahb 'gold'
˚Æg' a £wtl' y„rm˚' lgrl'
a¸gu altum 'garlic' umre alragl 'man'
That this is the work of a Jewish arabophone (perhaps providing a glossary to
a colleague for travel to a Spanish-speaking region) is evidenced by the fact
that the Arabic terms are presented unpointed and with only long vowels
overtly spelled, suggesting the conventionality of this orthography to the
reader. By contrast, the Spanish glosses have "helpful" pronunciation hints
indicated by some often inaccurate niqqud. The use of b b rather than p p in
beru is further evidence of an Arabic matrix (which lacks /p/), as is the use of w
w rather than g g to spell something approximating [©] in awah. For its part,
the dagesh in the initial g of girbansas would appear to be entirely unnecessary,
given that an initial /g/ in Spanish is unlikely to be realized differently, unless
one considers that standard Arabic has no /g/: the unadorned cognate Arabic
letter Õ ˛‹a√ normally represents the pharyngeal /˛/, while the affricate /Ê/ is
represented by Ã g¸ ª‹ m, that is, the same grapheme with an intralinear diacritic.
In a similar fashion, the diacritic superscribed on the g in a¸gu is likely an
imitation of the Arabic spelling of /x/ with Œ ƒ‹a√.
3.2. Romance
3.2.1. Spanish
As noted in the previous chapter, Al-Andalus (Islamic Spain) offers a
rarely-witnessed graphical melting pot: the writing system of one religious
group (Jews) could serve to write the three languages (Hebrew, Arabic, and
Romance) used by that group, while these three scripts could each be used by
members of the three religious communities to write the varieties of single
language (Ibero-Romance). After their expulsion from Spain in 1492,
Romance-speaking Jewish communities flourished in Italy, the Balkans, and
the Ottoman Empire, where Hebrew publishing also thrived from the middle
of the sixteenth century until the First World War. Smaller communities (with
correspondingly smaller written/printed outputs) could also be found in
Northern Europe, as well as in several Spanish and Portuguese overseas
colonies, including those in the New World (Levi 2002).
The earliest examples of Romance-language writing in Hebrew script
occur in Andalusian muwa¸s¸sa˛at – poems written in Arabic or Hebrew with
final couplets (known as ƒara¸gat) that were written in colloquial Arabic or
Romance, such as the following twelfth-century excerpts from the poems of
Yehuda ha-Levi, the Toledo native who also wrote the Book of the Kuzari (see
§ 3.1.2; the translations below follow Stern 1974, though the Romanizations
are my own):
(5) a.
e a o e i o e a a
hr'Sbl' hnwb §t / dynb hlyds wm dnk Sd d s k nd m u ç dell h b nid / t n bon h
i a
hr'gxl' d'w §' / dyS' lwSd hy'r £k alb ¸ s ar h
k m rayoh desol esid / en wad al˛igara h
o o
'As soon as Cidello comes, such good news

He appears like a ray of sunshine in
Guadalajara'
b.
e i o a a i
§'btnt S'drql' / §'b ydys §'b ben çidi ben / elk r d es t nt ben
e e a i e i e e a i
§'ydl' §bd wylp §b / §'mzl' tSd d st alz men / b n f lio d b n ald yen
'Come, my lord, come. Beloved, why do you
absent yourself so long?
Straightaway, come, O son of Aben al-
Dayyeni'
It is rather surprising to learn that until Samuel Stern published his editions of
some of these poems and their vernacular couplets in 1948, it was not known
that these otherwise indecipherable lines of poetry were written in a Romance
language – and on orthographic grounds it is not difficult to see why, given
the combined effect of idiosyncratic word division (lwSd <d¸swl> de.sol, §'btnt
<tntb√n> tanta.ben) and few overt vowels (e.g. £k <km> komo, dnk <knd>
kuando), not to mention those that differ from their conventional usage in later
Hebraicizations (e.g. §'b <b√n> ben, hy'r <r√yh> rayo). In most respects, in
fact, these early attempts to adapt Hebrew script to a Romance language
remain very much tied to the graphic conventions of Hebrew writing itself,
which would become less and less prominent as the system matured over the
next several centuries (Minervini 1999). By the time printed texts begin to
appear in the Sephardic diaspora of the sixteenth century, a full-fledged
orthography has emerged, as shown in the following opening lines of the
book of Deuteronomy from the 1547 Constantinople Bible (Recuero 1988):

(6) yEq S'‡r¯bal'Ap Sal S'AX¸SyE' estas la s pala bras ke
ÙdÙXA' heHm Ùl¸bah ha blo mo¸se atodo
ly≈nyE' §„dr
Ë ya ly„d yEXrË p
a §yE' lE'r ‡ W¯ yi yisrael en pa rte del yarden enel
dizyerto enla h lyanura h
hflr˚n'Ayy¯l hAln¯ yE' ÙXËryE yÃzyÊd
'These [are] words that
Moses spoke to all
Israel in the part of the Jordan in the
desert in the wilderness'
From this point forward, Judeo-Spanish writing represents far and away the
most mature and robust Romance-language adaptation of Hebrew script,
flourishing in the vast nineteenth- and twentieth-century Judeo-Spanish press
in Turkey, the Balkans, Israel, Northern Europe, and the United States.
Although discussion of this corpus is beyond the scope of this study, it is
worth noting that several common features of later Judeo-Spanish writing
contrast with most other medieval Judeo-Romance writing, such as the
spread of s s as the default sibilant letter (cf. ch.3 § 2.1.2), and the use of
"unsupported" h h for final /a/, i.e. without a preceding ' (cf. ch.3 § 2.2.1).
Recall that although Judeo-Spanish remains a spoken dialect in some
communities, beginning in the 1920s and since World War II the written
language has been (re-)Romanized in most "institutional" publications.
3.2.2. French
Although not the earliest written representations of French language
per se, some of the earliest examples of Romance material written in Hebrew
script are the eleventh-century glosses of Rabbi Shlomo Yitzhaki of Troyes,
more commonly known as Rashi (based on the acronym derived from his
initials, y"Sr r¸sy). As Levy (1970) points out, Rashi's use of a French word "is
very often the oldest example known to exist... A few hundred of the words
comprising the Judeo-French vocabulary… are unknown in normal Old
French texts." The following is a selection from Darmsteter (1907):
Table 2-12. Selected Rashi glosses
Pointed:
'ÕnyÕyfir¯wwO' ovrayn¥ '(large) works'
ry„rJËdÕn¬yÕna' anyandrer 'give birth'
¶y&bq ¯ r
Ë X
¯ S¯ Ùw vostr¥¸c¥‹beß 'your chiefs'
tyic‡r¯bÕn¯'yE' e¥nbraßit 'and was seized'
h¯'yËr¯yyËryEbËryEh herbery¥ri¥h 'he will remain by me'
Unpointed:
Xnmyywy yuyymant 'verdict'
XnmdyyhwS soheydem ant 'wish'
vnyr'Slydn'q kandilsarine 'when they strayed'
Most of these glosses appear in multiple forms (with and without niqqud)
across the numerous manuscripts of Rashi's commentaries – none of which
are in Rashi's own hand. Those that are vocalized are very often "over-
pointed," i.e. the niqqud precedes a mater lectionis and so is not strictly
necessary (some instances of rafeh and dagesh seem similarly superfluous).
Curiously, though, despite the fact that w and y are generally used as full-
fledged vowel letters for various non-low vowels, when /a/ is indicated it is
usually spelled only with diacritics, rather than with ', which more frequently
serves as final /¥/ (often with the corresponding niqqud below it). The glosses
do show some orthographic innovations, such as rendering the affricates [Ê]
and [ê] with (y)y and q-plus-hacek respectively, as well as the occasional
adoption of v fi as a vowel letter and the use of double-vav to indicate the
consonantal realization of w w as [v]. Other patterns, however, are typical of

early (Romance) adaptations of the script in their imitation of Hebrew-
language writing, such as the use of t t and h h, along with the less-than-
consistent word division.
A more extended sample of Hebrew-letter French writing from
somewhat later reveals, among other increased consistency, a more balanced
use of vowel letters and diacritics, as in the following example of thirteenth-
century para-liturgical poetry (Blondheim 1926):
(7) ¶y≈nm l'am XÕnya y¯S 'Õn S˚l¯p lEyy¯c ÙJd S¯lyEXyE' '
‘ mÙq S
¯ y„rq
¯ 'a yiS ¶Õnp
a& nÕ '
a S˚l
a a ¥
Lus anf nß si akres¥ kome etel¥s do ßiel plus n¥ s¥y nt mal m neß
'Your offspring will grow like the stars in the sky [and] no longer be troubled.'
Although the writer here does make some use of diacritics on final-position
consonants to indicate an open syllable, the use of ' √ for low vowels, as well
as to bear the diacritic for non-low initial vowels, has clearly become
conventional.
3.2.2.1. Language-learning handbooks
In addition to considering how Jewish writers through the ages have
deployed the Hebrew alphabet in the service of texts aimed at a presumably
experienced reading audience, the French context offers the opportunity to
examine how more recent adaptations of Hebrew as a matrix script have
served the purposes of those for whom the target language is new. This
context, most typically that of language-learning materials, is a pedagogical
one where the adaptation process has been performed by the writer in a very
explicit and deliberate way. Based on my brief research into this locus of script
adaptation, the phenomenon runs the gamut from transcriptions intended as

"purely phonetic" (that is with no conventional matrix orthography serving as
a conditioning element) to others that are more dependent on the conventions
of an established matrix orthography – not unlike what we would expect to
find in a range of, for example, Hebrew-language handbooks produced for
readers of different Roman-letter languages. For the present purposes, I offer
samples of two early twentieth-century manuals targeted at Jewish learners of
French.
The first one, published in Constantine (Algeria) in 1908, is entitled
§vkvrpS wc rvrhvl vnhA' Syzvcn'rp rvd vdA'hXvm (Methode der frantsezi¸s ohne
lehrer tsu ¸sprexen, 'A Teacher-less System to Speak French', currently held at
the Alliance Israélite Universelle in Paris). As a subtitle not included on the
Hebrew-letter side indicates, it is intended "à l'usage des israélites allemands,
autrichiens, hongrois, roumains, russes" [for the use of German, Austrian,
Hungarian, Romanian, and Russian Jews]. What distinguishes this handbook
is that its matrix is not Yiddish but German, which is made clear through (1)
lexical, (2) phonological, and (3) orthographic features in its title: (1) the use of
the verb ¸sprexen 'speak' rather than Yiddish §dvr redn; (2) the /e/ vowel in
frantsezi¸s 'French', where the diphthong /oj/ would be expected, along with
the final vowels, deleted in Yiddish, of methode and ohne; (3) the use of silent h
h, which was introduced into Yiddish orthography in the eighteenth century in
imitation of New High German writing but fell out of general usage soon
thereafter (Kerler 1999: 151).
As a writing system, the transcription used to convey French
pronunciation in a pedagogically functional way bears little resemblance to
any Hebrew-letter orthography then in use. Not that it necessarily should,
since an introductory guide to spoken French is not the venue for an

orthography proper. Still, the writer seems to have relied very little on any of
the conventions that Hebrew or Yiddish orthography had developed
(assuming, quite safely, that he was familiar with one or both of them), let
alone does his system reflect anything of the medieval tradition of
Hebraicized French writing.
This apparent lack of conventionality is most clearly illustrated in the
system of vowel transcription. Modern French does present a challenge to the
humble set of four matres lectionis, with at least a dozen vowel phonemes,
including several that are nasalized. Eschewing the lexical/etymological
information contained in the Roman-letter spelling of French, however, the
writer of the Methode sought to give every phonologically unique vowel its
own graphic form, without resorting to the importation of a non-Hebraic
diacritic, nor to aping the etymological Roman-letter convention of a
following n n to indicate nasalized vowels.25 The result is a bevy of
homographs (words in parentheses are not given in the manual itself):
Table 2-13. Hebraicized French homographs
Av ã an, on, en 'year, we/one, in'

Avd dã dans, dent, (don) 'in, tooth, (gift)'
v
A s sã son, cent, (sang) 'his/her, hundred, (blood)'
'
A n nø/no ne, nos 'not, our'
rAl lœr/lør leur, l'heure, l'or 'their, the hour, the gold'
'Ap pø/po peu, peut, peau, pot 'few, can (v.), skin, pot'
25
Modern Hebrew and Modern Yiddish transcriptions of French words certainly do ape the
Roman-letter convention in this respect. The choice of v as an all-purpose nasal vowel (see
table 2-14) may seem odd to readers familiar with these modern orthographies. Yet it may be
less than arbitrary, given that both Morag (1971) and Ornan (1971) note that some Dutch and
Italian communities may realize v as [˜] in their traditional Hebrew pronunciation, a
phenomenon that also surfaces in forms such as the name of the twentieth-century Yiddish
poet §yyXSX'lg bqvy Yankev Glatshteyn ('Jacob Glatstein').
Given the conventions of earlier Judeo-Romance writing, this orthography
presents several grapho-tactic problems. For instance, using a single vocalized
letter to represent a non-bound morpheme (i.e. a word) is unprecedented in
my experience of Hebraicized writing systems. Furthermore, the placement
of the niqqud is inconsistent: the rhyming pair nos/pot, for example, has the
qameß indicating /o/ under the first or second letter, making it unclear just
which digraph stands for the vowel phoneme.
The second French learner's handbook, ¢yyrqn'rp §y' dwy yid in frankrayx
'[A] Jew in France' (Bibliothèque Médem ms. 15237) has a more clearly-
targeted Yiddish-reading audience, and the transcription used here yields
forms that at least superficially resemble Yiddish words in their graphic
structure. Still, several features do set its adaptation of Hebrew script apart.
First and foremost, this manual is the only Hebraicized text of any kind I have
yet encountered where the final form of a Hebrew letter may be followed by
another letter, in this case, where nun is followed by yud to indicate a word-
final /µ/:
(8) y#§ys signe 'sign'

y#§vp peigne 'comb'
y#§A'lA'p pologne 'Poland'
The unconventionality of this spelling could be construed as a largely cosmetic
issue, since the yn <ny> digraph does have firm precedent in earlier Judeo-
Romance writing. Still, it is surprising to find a writer who would flout a
grapho-tactic convention of Hebrew script so basic and consistently-practiced,

particularly in longer phrases where the not-quite-final form occurs within a
single phonological unit:
(9) hyvwwrym §a'dra'Sz §v#z#§a'd £A'q

comme dans un jardin merveilleux
'as in a wonderful garden'
In this case, the writer has used the final-form § in what appears on paper as a
single word but which is actually composed of two, with the "real" final
consonant of the first word (spelled <s> in Roman script and normally silent,
but surfacing as [z] in liaison before a vowel) resyllabified as the onset of the
second word. Note that in both (8) and (9), the writer still adds a diacritic to
the final-form nun as if to acknowledge his unconventional usage (using a
single device to indicate the palatal quality and syllable boundary, where it
nonetheless retains an aura of finalness).
The writer of this handbook does divulge his technique for rendering at
least some of the French vowels in Hebrew script. Yet unlike the writer of the
Methode above, he bases it firmly on orthographic grounds, i.e. a mapping of
Roman to Hebrew graphs:
Table 2-14. Vowel transliteration
é è ê ai
hv v vv '
e
This system instills its own confusion, replacing the three-letter imperfect
inflection -ais, for example, with a singled pointed e', while rendering the
ubiquitous <é> with a digraph. Rather than a profusion of homographs, then,

the result is unique spellings for a number of homophones such as the
following, which attempt to mimic the etymologically-based Roman-letter
spellings:
(10) 'Xhv étais 'was'

vhXhv été 'been'
As a final observation, it is interesting that the writer provides some
reasonably accurate transcriptions for some very literary verb forms, which
are unlikely to come up in casual conversation, let alone the brief exchanges
sustained by a language learner:
(11) y'#X ry'#z-ly' ils eurent eu (past anterior)

ssy'#z ly'#q qu'ils eussent (imperfect subjunctive)
The words are divided graphically to highlight syllabic units (liaison serving
the French preference for an onset "at all costs"), but in keeping with the
"orthographic analogy" position of this transcriber, apostrophes are added to
indicate the morpheme boundaries that may be obscured in speech by liaison
or by contraction (only one of which, #q, corresponds to Roman-letter usage).
Also worth noting is the use of doubled letters in imitation of the Roman-
letter orthography, even though the principle of the doubled the Roman letter
(<ss> ≠ [z]) is carried by the normal reading of the single Hebrew letter – not
to mention the rarity of doubled letters in Hebrew-language writing.

3.2.2.2. Bilingual dictionaries
A more modern example of a pedagogically-motivated adaptation of
Hebrew script for French pays strange heed to the conventions of the target
and matrix orthographies. The French half of a 1971 pocket dictionary
(printed in Tel Aviv) provides a transcription in which, at first blush, phonetics
generally trumps phonemics or morphemics: a single <s> is always z z, /s/ is
always s , silent consonants disappear, no graphic distinction is made in
Hebrew script for final <é> vs. <er>, etc. There is nonetheless an odd
combination of flouting, upholding, and elaborating the available conventions,
illustrated by the entries below:26
Table 2-15. Hebrew-French dictionary entries
e he
h haie 'hedge'
e -Ùh ho-fe
p haut-fait 'act of bravery'
r%'nÕC Ù' on:œr honneur 'honour'
s
E hi his:e hisser 'hoist'
a ˚h huoyo
Ù'Õy' hoyau 'hoe'
me'y¯X«y˚y yuityem huitième 'eighth'
lJbn%' œ~nbl humble 'humble'
˚ydnaXn¬ ' i inatãndyu inattendu 'unexpected'
The first striking feature of this system is the imported tréma (dieresis), in this
case over ' , which, although rare in adaptations of the script, is not very
distant graphically from the supralinear left-edged O ' in Tiberian niqqud that
indicates /o/ (cf. chapter 8 § 2.4). As the all-too-brief guide at the front of the
26
My Romanizations in the second column are meant as quasi-phonetic transcriptions. The
superscripts refer to Hebrew consonants that do not strictly contribute to indicating the
pronunciation of the word, but that may be present because of their analogues in the Roman-
letter spelling, or else to serve a diacritic function, or in order to satisfy a convention of
Hebrew grapho-tactics. Each of these is discussed below.
dictionary states, it is used here to indicate a sound "like ö in German." Oddly,
however, to indicate the high front rounded vowel that is also spelled with a
dieresis in German, this system employs a digraph ˚y that, while logical from a
linguistic point of view (y y as a diacritic for "front" on high back rounded ˚
[u]), may be liable to misinterpretation by the average Hebrew reader.
The only other direction the transcriber gives about his technique is in
relation to the spelling of nasal vowels: Õn is §wSlh tw' ot ha-la¸son 'the tongue
sign', i.e. consonantal [n], while final § or n sans niqqud is •'h tw' ot ha-√af 'the
nose sign', i.e. the equivalent of a single post-vocalic <n> or <m> in Roman-
letter French. The dagesh is also put to somewhat novel use, as an indicator of
orthographic doubling in the Roman-letter spelling (except for <ll> /Ò/, which
is usually not rendered by any l at all). Other graphic conventions of Hebrew
are simply flouted: niqqud on final letters,27 non-final forms in final position
(presumably to bear the niqqud), monosyllabic words rendered as a single
letter, as well as phono- and grapho-tactically bad clusters (not tolerated by
Roman-letter French either, as the persistence of e-muet would attest). And
yet the adaptation seems to make an attempt at distinguishing h-muet (<h> in
Latin-origin words) from h-aspiré (<h> in loanwords, usually from Germanic)
– despite their identical realization as Ø – at the expense of phonetic
transparency. Initial h thus emerges as a "silent letter" in this adaptation of
Hebrew script, providing etymological more often than phonetic cues, but
doing neither in a truly consistent fashion.
27
In canonical Hebrew spelling this occurs in the lexicon only in combinations of the
Tiberian short /a/ and a glottal or pharyngeal consonant, a' -√a, ah -ha, ax -a˛ and av -afi (as well as
in the grammatical inflections ¢
√J -ka, ¢
√ -ƒa, ¢
Ã -Vƒ, t
A -ta and t
Ÿ -Vt). As noted in § 2, absolute-final
vowels must be borne by a mater lectionis. Nevertheless, my impression based on bilingual
dictionaries and other modern pedagogical materials is that niqqud on letters in final position
has become normal practice in "transcriptionese," as though this written dialect specifically
does not require that forms follow standard Hebrew grapho-tactics.
As long as native Yiddish and Hebrew speakers continue to learn other
languages, there will be innumerable manuals of this kind, perfect analogues
to the foreign-language handbooks produced in others literate cultures. What
remains to be examined is how other traditions of Hebraicization may have
served as matrices in the language-learning materials that predate the rise of
printed Yiddish or the revival of native Hebrew in the late nineteenth century
– especially in light of the peculiar nature of the adaptation in these twentieth-
century examples.
3.2.3. Italian
Many of the Jews expelled from Spain at the end of the fifteenth
century settled in Italy, establishing Castilian-speaking communities and, by
the middle of the sixteenth century, founding major centers of Hebrew
writing and printing, most notably at Ferrara. Even before the arrival of these
Spanish émigrés, however, native Jews had adapted Hebrew script for the
purpose of writing Italian, producing in particular a large number of Italian
translations and adaptations of biblical and liturgical texts. The following are
the opening lines from the Alfabetin, which is based on an alphabetically-
arranged Judeo-Aramaic poem (known as a Piyyut) and which forms part of
the prayer service recited by some Italian Jews on the festival of Shavuot
(Gelman 2000):
(12) ˚l˚pÙp˚la' yiXÕnÙm˚lyËd heSOm yisyESyE' E ¸seçi Mo¸seh d¥.lu.monti a.lu.populu
yisyÊdyE' e.diçi
yiXX
a yicyÊryE' yiXm a yisy„ry¸p' a yisyÊd'a ad.içi apreçima ti e.rißita ti
li.deße kunma nda menti
yiXnÕ yEmd
fi nÕ m
a nÕ ˚q yEcy„dyil
'And Moses descended from the mountain
to the people and said
to them: "Approach and recite
the Ten Commandments." '
One of the most intriguing bodies of writing, however, is the bilingual poetry
of Italian Jews, such as the following sixteenth-century excerpt from Shmuel
da Castiglione (De Benedetti-Stow 1980; Hebrew-language passages are
italicized and given in standard Italian spelling):
(13) yc'nymy&p yXsywwq yd / twnwlt ytvmS Ho udito le querele / di kueçti feminaße

yc'nym §'rg wn'&p ym y' / twnnwqm ylv Su me fan lamentele / e mi fano gran
yrbd t' / 'c'rXs w'y' yq wnylww y' minaße
E voleno ke io straßa / le paról della
hrySh
canzone
h'ynlyww §ym'q yd / hrm twqvc E amaramente gridano / di kamin
vila niah
'I have listened to the cries of those

women
Over me they lament / and against me
they threaten
And they wanted me to tear / the words
of the song
and harshly they cry / of those who love
villany'
Despite the juxtaposition of Hebrew language material, the writing system in
evidence here is very much an autonomous one, with no niqqud used and the
matres lectionis fully deployed as vowel-letters. Note also that while letters
that are doubled in the Roman-letter orthography are not imitated in the
Hebraicized forms – a single c ß, for instance, covers the ground of both <cc>
and <zz> (representing modern [ê] and [ts]/[dz] respectively) – the writer
does make use of a trigraph wwq qww for the labiovelar segment spelled <qu>
in Roman script, even though double-ww serves elsewhere to represent /v/,
suggesting perhaps that v and u were not necessarily distinct entities in this
writer's mind.
3.2.3. Provençal
The Jewish dialect of Provençal largely disappeared by the twentieth
century,28 and its most extensive attestation is actually found in a Roman-
letter play and other "comic" texts written by non-Jews.29 There are no post-
medieval texts written in Provençal using Hebrew script either (if in fact the
phenomenon survived the medieval period). What does exist, however, are
several medieval manuscripts, such as an early fourteenth-century Hebrew-
Provençal dictionary (Aslanov 2001). Since isolated words may not present
the same need for an orthography with its own conventions as does more
extended writing, it is not surprising that many of the entries in this dictionary
illustrate a lingering adherence to Hebrew-language spelling conventions:
28
Jochnowitz (1978: 69) reported that to the best of his knowledge there was only one person
alive at the time who remembered hearing Judeo-Provençal spoken.
29
Zajkowski (1948: 32-36) does discuss a comedy from 1820 apparently written by a Jewish
lawyer from Montpellier whose wife hailed from the Comtat-Venaissin region.
Table 2-16. Hebraic patterns in Judeo-Provençal
• agglutination of prepositions
l'dp'q' akapdal a capdal 'in capital ' (Heb. wS'Or¯Jb b¥.ro¸so)
• implicit vowels
XnmmyXSlb blastemam ent blastemamen 'blaspheme'
• use of k vs. q (see ch.3 § 2.1.2)
yrdn'ylwk koliandre coliandre 'coriander'30
• initial h
'#gwh hu¸ga hucha 'clamour'
• final h without preceding '
hyynpS' espanyah Espanha 'Spain'
• double-ww as CV syllable
yrww vori vori 'ivory horns'
Other entries, however, do show innovations introduced in the service of
spelling Provençal words, strategies that are relatively rare in Romance-
language adaptations of Hebrew script and certainly not in step with Hebrew
writing itself:
Table 2-17. Rarer adaptations in Judeo-Provençal writing
• d as /z/
'dydwn nudeda nudeza 'nudity'
• b as semi-vowel
'nbX tebne tèune 'fine, thin'
• doubled consonants
hrryS serrah serra 'saw'
Although the precedent for using d to spell an fricative dates back to Hebrew
spirantization (see § 2.3), this is the only use of it that I have encountered to
30
Aslanov (2001: 23) suggests that the choice of k here could be influenced by the presence of
the Aramaic and Arabic cognates ('trbswk kwsbrt√ and rwbswk kwsbwr respectively) cited
earlier in the entry.
represent more specifically the alveolar fricative31 – unless it is better viewed
as a hypercorrection (to restore the stop in the suffix). The use of b to indicate
the semivowel in 'nbX teune 'thin' < TENUE could also be construed as a
hypercorrect spelling, not unlike the <l> in OIt. colse 'things' < CAUSAS or
repolsar 'rest' < RE+PAUSARE (Dye 2000: 139).
There are also longer texts, including an original para-liturgical
composition based on the Book of Esther, composed in the fourteenth
century:32
e
(14) a. r'Snmwq' lyyww §'mwr §wm mon roman veil akom nsar
...rcyndkwbn yd gyy'p l' al fayg de nbwkdnyßr...
'My story will begin
with the tale of Nebuchadnezzor'
b. X'Xnwq' Swn l'ynd yq §wgyS segon ke daniel nos a.kontat

e e e
'Xnwlwwd qnyw rwznd'qwbn' a.n bokadn zor venk d volonta
'According to what Daniel has told us
To Nebuchadnezzor he came willingly'
Already noticeable in the extract above is the lack of niqqud and the use of an
overt letter for nearly every vowel. In addition, the above lines contain a
curious variation in the spelling of King Nebuchadnezzor's name. Although
Neubauer & Meyer (1892) stick to a single spelling <Nabocadnessar> in their
transliteration, the first occurrence rcyndkwbn nbwkdnyßr actually leaves the
name intact in its unvocalized biblical spelling, while the second occurrence
31
It is not, of course, without precedent in Roman-letter writing. Prior to the advent of
vernacular spellings that laid bare some of the phonological innovation in Romance, many a
Latin <D> might well have been read as [Ê] (later to deaffricate to [Ω]) in certain environments,
just as modern Québécois French speakers do. A medieval Provençal reader may well have
realized some conservatively-spelled instances of <t> or <d> as [z].
32
Although first edited and published by Neubauer & Meyer (1892), this text was the subject
of a relatively recent doctoral dissertation (Silberstein 1973), one of the few non-Castilian
Judeo-Romance texts to be studied so thoroughly.
writes it according to the conventions of the Hebraicized orthography,
rwznd'qwbn nebukadnezor. Along with three overt vowel-letters, there are two
consonant changes (k k ‡ q q , c ß ‡ z z) that substitute the letter more
typically used in Romance-language adaptations of Hebrew script for two
letters generally restricted to words of Hebrew-Aramaic origin. This
fourteenth-century Provençal example is reminiscent not only of similar
variant pairs in Hebrew-letter Portuguese of the same era (see chapter 3), but
also of a phenomenon found in the modern Yiddish press (see § 3.9.1).
3.2.5. Portuguese
As the main topic of this study, Hebraicized Portuguese is discussed in
greater depth beginning in the next chapter. For the moment it should be
noted that most of the extant Jewish Portuguese writing was produced by the
Northern European descendants of émigrés who left Portugal after the 1497
expulsion/conversion, where it is always written in Roman script.33 By
contrast, pre-1497 Jewish Portuguese writing, consisting of the texts in this
dissertation and a handful of others,34 is written exclusively in Hebrew script.
The one exception I have encountered to the lack of post-1497 Hebraicized
Portuguese is a language handbook, published in Warsaw in 1929, for
speakers of Yiddish to learn Portuguese entitled ¢wbnrvl rvSydyy-SyzvgwXrA'p
portugezi¸s-yidi¸ser lernbux 'Portuguese-Yiddish Handbook' (Paris, Bibliothèque
33
No spoken dialect has survived to the present day, except perhaps in peculiarities of the
language spoken by the descendents of Marranos, the "New Christians" who continued to
practice elements of Judaism in secret (see Wexler 1982, 1985).
34
A second Bodleian astrological text, smaller than the one presented in chapter five, has been
extensively studied by Hilty (1957-58, 1982), although no edition has appeared. In addition,
Sharon (2002) cites two further manuscripts: a medical treatise of ophthalmology from 1300
(located at the Biblioteca Publica Municipal in Porto, Portugal), and a treatise of medical
astrology from the fifteenth century that contains a part in Portuguese (located at the Jewish
Theological Seminary in New York). I have not accessed either of these.
Médem 1523), presumably aimed at Jews emigrating from Europe to Brazil.
Although the body of the text is written in Yiddish and the Portuguese is
usually presented in Roman script, some individual words are given in
Hebraicized form:
Table 2-18. Portuguese learner's manual
r#a'vssa'p passear 'take a walk' r#a'ynys' asignar 'assign'

'
A r'ss#'p passaro 'bird' ryXsys' assistir 'help'
vll#vp pelle 'skin' 'yn#yl'g gallinha 'chicken'
rv#rra'ww varrer 'sweep' r#a'rvs cerrar 'close'
'
A SzyrrA'q corrijo 'I correct' ry#Szyrwq corrigir 'correct'
Not surprisingly, the spelling of vowels here is distinctly Yiddish-like, with v
serving for /e/ and the diacritics under ' following Yiddish orthography.
Nor does the transliterator forego phonetic transparency for the sake of
orthographic similarity, as the same Sz z¸s digraph is used to indicate /Ω/ in
corrijo and corrigir, despite being spelled by different singleton letters in
Roman script. He does, however, imitate the doubled consonants of the
Roman-letter spelling (doubled consonants occur in Yiddish orthography only
at morpheme boundaries). He also imports one device foreign to the Hebraic
matrix but for a purpose not commonly found in mature orthographies,
namely the apostrophe, using it to indicate stress position (though its
placement on one side or the other of the accented vowel would seem to be
haphazard).
3.2.6. Romanian
An early Jewish presence in Romania is attested by tombstones dating
from the time of the Roman province of Dacia (Barnavi et al. 2002). Yet the
later, more numerous Romanian Jewish population was almost entirely
Yiddish- or Judeo-Spanish-speaking until the second half of the nineteenth
century, with the result that there is relatively little indigenous Romanian-
language writing in Hebrew script. The following examples, taken from a
Yiddish-language study of the Jewish press in nineteenth- and twentieth-
century Romania (Tambor 1977), are typical of the Romanian-language titles
of Yiddish and Hebrew newspapers:
(15) vqs'vrwwv 'c'yww wvrwwv lwrA'XySXnA'm

viaßa evreaske mont¸sitorul evreu
Via≈ta Evreiascâ Muncitorul Evreu
'Jewish Life' 'The Jewish Worker'
The Yiddish matrix is unmistakable: v fi spells /e/, '-plus-niqqud (rather than w)
spells /o/, double-ww spells /v/. The only sequence not typically represented
in Yiddish spelling, /eu/, is simply spelled with the two corresponding vowel
letters, v and w . Interestingly, the writer also seems to take advantage of a
southeastern dialect feature of Yiddish, where standard /o/ (spelled A ') often
shifted to /u/, so that A' can correspond to what is spelled with in the
Roman-letter orthography.
Interlude: Latin
Although the first Jewish communities under Roman rule were
Hebrew-, Aramaic-, and Greek-speaking, based on the later flourishing of

Judeo-Romance languages there is no doubt that Jews in Latin-speaking
regions did adopt the imperial idiom. The Jerusalem Talmud (compiled in the
first half of the first millennium CE), for that matter, recognizes a role for Latin
in Jewish life: "Four languages are of value: Greek for song, Latin for war,
Aramaic for dirges, and Hebrew for speaking" (Sotah 7:2, 30a., cited in Spolsky
1985). What does not appear to have persisted – if it existed at all – is a
tradition of writing Vulgar or even Imperial Latin in Hebrew script. Even
Blondheim (1925), in his seminal work on the possible Judeo-Latin precursors
to later Judeo-Romance, does not offer any such forms beyond personal
names.
For evidence of direct Jewish engagement in Latin-language culture we
have to jump ahead several centuries to medieval Spain, where multilingual
Jews were actively involved in translating scientific, grammatical, and
philosophical texts from Arabic into Hebrew and Latin, most notably at the
school of Toledo. Yet based on the extant record there appear to be very few
instances of Latin-language material written in Hebrew script beyond isolated
or compiled glosses. The reason may be straightforward: Jews who were not
involved in translation simply had little reason to learn and thus to write
medieval Latin.35
The table below contains plant names written in Latin from a medieval
herabarium (Paris, Bibliothèque Nationale ms. Héb. 1199). Of the more than
35
The text presented in chapter 5 features quotations attributed to Aristotle in what appears to
be Hebraicized Latin (these are collected at the end of the commentary in that chapter).
Although the individual words contain what looks like plausible Latin morphology, at the
syntactic level the quotations are largely nonsensical. While this may reflect rather poorly on
the scribe's (and perhaps the author's) knowledge of authentic Latin prose, it nonetheless
reveals a certain level of prestige associated with Latin writing.
120 full-page illustrations labelled in Hebrew-letter Latin, only a fraction are
also accompanied by a Roman-letter caption:36
Table 2-19. Hebrew- and Roman-letter herbarium captions
'yc'lwgyr rigolaßia ligaritia 'Licorice'

h'yglwXsyr' aristolgiah aristolochia 'Dutchman's Pipe'
'nylwwq 'lgnw' ongla ka v a lina ognia cauallina 'Onion'
rwnym 'rylwXnwq kontolira minor chonsolida minor 'Larkspur'
rwyym 'rylwXnwq kontolira ma yor chonsolida magior 'Larkspur'
wXSyrg' £w'ycrwXSn na ¸sturßium agri¸sto nastritium 'Forget-me-not'
Unlike the Iberian peninsula, Romance-language adaptations of Hebrew script
in France and Italy make wide use of c ß to render a variety of sibilants and
dental affricates, and this practice carried over to the spelling of Latin words
themselves. The Roman-letter captions clearly point to the text (or its writer,
at any rate) as originating in Italy, as do some of the Hebraicized spellings, e.g.
the hypercorrect l in 'lgnw' ongla 'onion' (<gl> being associated with palatal
/Ò/ in Italian orthography).
3.3. Greek
The Greek language entered Jewish life as Hellenistic culture spread
throughout the Mediterranean and Near East, emerging second only to
Aramaic as the major vernacular of Jewish communities through the first
millennium CE. Not surprisingly, given its scope and stature, there is a strong
Greek influence on the Rabbinic writings of this period such as those in the
36
I am indebted to the individual seated next to me in the Oriental Manuscripts Reading
Room of the Bibliothèque Nationale (Paris) in the summer of 2001, who was able to identify
many of the herbs and flowers based on the accompanying illustrations. The manuscript
contains a further ten pages of Hebrew-language writing, each paragraph headed by a
Hebraicized Latin term vocalized with niqqud.
Talmud, which contains as many as one thousand Greek loanwords (some
borrowed via Latin). Beyond the names and loanwords from non-Semitic
languages such as Persian that occur in the Bible, this corpus represents the
first substantial body of linguistic material to make use of Hebrew script in
writing material from a non-Semitic language. Below is a small sample of
Greek loanwords that occur in Hebrew (Sperber 1984):
Table 2-20. Greek loanwords in Hebrew
• modern retentions
carthV syXrk
xartês kartís
'record' 'card'
• doublets
statiwnar rnwyXXs rwXdns
stationar stationar sandator
'police officer' 'executioner'
• "conscious" Hebraization
kleyudra hrdys •lx
klepsydra ˛alaf sidra
'water-clock'
The first word is one of the many Rabbinic-era borrowings that have survived
into Modern Hebrew without any specialized meaning. In the second
instance, the Greek loanword yielded two Hebrew words, each reflecting a
different degree of integration. The third term is a Hebrew coinage whose
meaning is that of the Greek source but which uses native elements to imitate
the form of the Greek word.37
37
This practice was not restricted to Hebrew writers of Talmudic era. In the early days of
Modern Hebrew language planning, several such terms were suggested (though ultimately
rejected), e.g. vr-ylwx [xolira] (lit. 'evil illness') for cholera, br-gwld [dilugrav] (lit. 'great leap')
for telegraph, or lk-yXrp [pratekol] (lit. 'all details') for protocol.
As the expanse of Greek diminished in its post-imperial era, so too did
it diminish as a spoken and written language for Jewish communities beyond
southeastern Europe. Yet Jews in the Byzantine sphere continued to speak
and write a form of Greek, producing biblical translations, commentaries, and
sundry other texts in Hebrew script. The following are vernacular rubrics in a
Greek ma˛zor from the Cairo Genizah, the unusually well-preserved trove of
documents and sacred texts held in a synagogue attic there but fully accessed
only at the end of the nineteenth century (de Lange 1996):
(16) Sa'yiXSyÊrka& p
& '
¯ ˚'na' ÙJdw'r¯Xq Swlw' y&dy&dq yraXypq §yÊd
& yÊrqyip yÊrp
e '
a yÄlpq
e a e a a a a
k f li ap ri pikridin k pitari k didi olos k troudo anao efƒ ristias
'And again he takes bitter herb and unleavened bread and he gives to
everybody and they eat it without a blessing'.
As expected in this liturgical context, the text is at least partially vocalized with
niqqud, and the influence of the dominant orthography is evident in the use of
overt vowel letters, as well as diacritics to indicate alternate readings of those
consonants that Masoretic Hebrew deemed orthographically modifiable.
What is especially noteworthy here is the use of the rafeh and dagesh on
different occurrences of the same letter, in this case d d.
3.4. Slavic
A Jewish presence in Northeastern Europe may date from as early as
the first centuries of the Christian era, and the conversion of the Khazar ruling
class to Judaism beginning in the eighth century probably attests to the
presence of Jewish communities in Russia as a branch of Byzantine Jewry
(Barnavi et al. 2002). The earliest linguistic attestation of Jewish life in Slavic-
speaking lands come from coins with Hebrew inscriptions minted in twelfth-
century Poland. Yet despite the large amount of material written by Jews in
Slavic languages (presumably beginning with Slavic-language glosses in
Hebrew manuscripts), I have found relatively few examples of extended
writing in Hebrew-letter Slavic languages.38 The modern Hebrew and Yiddish
press no doubt includes material from the Russian, Polish, and other Slavic
languages of its readers and topics, but as I noted at the beginning of this
section, these are more of interest in the context of Hebrew or Yiddish writing
itself. Nevertheless, as a brief and less-than-common sample I offer the
following Bulgarian book/journal titles:
(17) yXsw yqsyrbyy hn §yXlwyb hbwls hXyqsyrbyy hn hygwlwXn'

byultin nah yebreski vesti antologiah nah yebreskitah slobah
'Bulletin of Jewish News' 'Anthology of Hebrew Words'
These short examples show the influence of a particular matrix, i.e. Modern
Hebrew (as opposed to Yiddish) spelling conventions, such as the consistent
use of h h for word-final /a/, as well as the single w for consonantal /v/. Yet
the influence of centuries of Hebrew script adaptation may be evidenced in
the now-conventional use of X †, the historically emphatic dental stop, as the
"default" dental, along with q q, the historically emphatic uvular stop, as the
default velar (cf. chapter 3 § 2.1.2).
38
This may have something to do with recent attempts to revise the standard history of
Eastern European Jewry, whereby the earlier Byzantine community was allegedly replaced
by eastward migration from Germanic territories.
3.5. Persian
Tradition traces the Jewish communities of Persia (centered in modern
Iran) as far back as the Assyrian deportation of the Israelites from Samaria in
the eighth century BCE and the Babylonian deportation of the Judeans from
Jerusalem and its environs in the sixth century BCE. The earliest record of their
presence, however, is the activity of Jewish leaders such as Ezra, Nehemiah,
and Zerubbabel. As part of the Assyrian and Babylonian empires, these
communities were probably Aramaic-speaking, though by the end of the first
millennium CE there were Persian-speaking Jewish settlements from the
Caucasus to Western China.
The first textual evidence of Jewish linguistic activity in Persian dates
from the fourteenth century, in the form of biblical translations (see Paper
1978). Later Judeo-Persian writing includes more secular literature, such as
works by the fourteenth-century poet Maul‹an‹a Sh‹ahın. The following is an
excerpt from a later manuscript of one of Sh‹ahın's lesser-known epics, King
Ki¸svar (Asmussen 1973):
(18) / Skrs ¢rt Sr'zh dcph ydwb / £'n 'r £wr h'S dwb rwwSk hk
Skrt rySmS 'bw bs' 'b hmh
kh k¸swwr bwd ¸s√h rwm r√ n√m / bwdy hfßd hz√r¸s trk srk¸s /
hmh b√ √sb wb√ ¸sm¸syr trk¸¸s
'Ki¸svar was the name of the King of Byzantium / He had seven hundred
thousand wild Turks, all with horses and with swords [and] quivers'.
Just as Arabic-letter Persian orthography adopts some of the principles of
Arabic writing (e.g. only long vowels overtly indicated), so does Hebrew-
letter Persian tend to mimic those conventions. As in Judeo-Arabic, the
Hebraicized Persian orthography also augments with diacritics the Hebrew

letters that are cognate to those that are augmented with a diacritic in Arabic
script. Yet Judeo-Persian also fuses Persian-language adaptations of Arabic
script into unique conventions, e.g. using to g g in imitation of ı t¸sé or ¯ gâf,
the Persian adaptations of Arabic ¸gim and kaf (Modern Farsi forms are given in
transcription):
(19) r'∆g flg‹√ r êah‹ar 'four'

¢r«g „grk êarx 'wheel'
§'h«g „gh√n Êah‹an 'world'
yr«gm mr„gy morq 'bird'
In other respects, however, the Hebraicized Persian does rely on the cognate
relationships of Arabic and Hebrew letters, using consonants that are usually
rejected in European-language adaptations (notably k k and t t, as well as x ˛
and v fi), and often foregoing diacritics and overt vowel letters:
(20) rg gr zar 'gold'

rkS ¸skr ßekar 'sugar'
•rx ˛rp harf 'letters'
tsyn nyst nist 'is not'
£gnp pngm panÊom 'fifth'
3.6. Turkish
Before the rise of the Ottoman Empire, Jews in what is now Turkey
were speakers of either Greek (known as Romaniot) or Arabic (Mustarabs).
Upon their expulsion from Spain, Iberian Jews were invited by the Sultan to
settle in Ottoman territory. These new immigrants were speakers of Ibero-
Romance languages, and their numbers eventually overwhelmed the

indigenous Ottoman Jews over the course of the sixteenth century, resulting
in an Ottoman Jewish population whose linguistic profile was markedly
hispanic. Beginning in the nineteenth century, through the educational
reforms of the Paris-based Alliance Israélite Universelle, many Ottoman Jews
began to learn French as a principal second language.39
Given that the Sultan's invitation was in large part commercially
motivated, it is difficult to imagine that many Ottoman Jews did not acquire at
least some "survival" Turkish (although the lack of a centralizing, nationalizing
impulse in the Ottoman Empire may have meant that learning Turkish was
little more than a utilitarian affair). Yet to my knowledge no robust tradition
of Hebraicized Turkish writing ever developed. One of the rare examples I
have found is an Ottoman chronicle written in Hebrew script (Bodleian Ms.
Heb. e 63; Marazzi 1980), an excerpt of which is given below:
(21) f. 106 r.
wrlly' £yq rwdrylXyy'qyx byy'&gv yryy'gw rwdn'mXv yl' ¢yr't yqlw'
vyq'w hdn'mz
√wlqy t√ryk √ly fiw†m√ndwr wg√yyry fi‹g√yyb ˛yq√yy†lyrdwr qym √yllrw
zm√ndh w√qyfi
evvelki tarih âl-i Osman vegayri acayip hikâyet-? kim ileri zaman -dh vaki
'First date the Ottoman Empire and another strange story who before time
event'.
3.7. Georgian
Notwithstanding the question of the medieval Khazar conversion to
Judaism, a Jewish presence in the Caucasus region probably dates from
39
Varol (2003) notes that in present-day Turkey, women are much more likely than men to be
active users of French. Nevertheless, there are numerous attestations of – and consequent
mocking reactions to – the high prestige that French attained among Turkish Jews, e.g. the
farcically Gallicized Judeo-Spanish speech of the suitor Musiú ›Jac in the play Peche Friyo
(Varol, p.c.).
around the third century CE, based on tombstones bearing inscriptions that are
not in Hebrew or Aramaic. Although these tombstones also feature the
occasional Georgian name (usually of Persian origin), the only examples of
Hebraicized Georgian writing that I have encountered first-hand are found in
a modern study of an early-medieval Hebrew manuscript (Lerner 2003),
which contains names such as the following:
(22) ywwÙxKp P'xovi

yraw¯q¯X¯m Mtkvari
ylewÙrm Mroweli
g#acp
aK wÃ t
a T'avp'a¸c'ag
yl#avc syrge' Egris Cqali
Despite the Hebrew-language matrix of the study in which these names
appear, it is difficult to detect any specific Modern Hebrew influence in their
spellings, apart from the use of letters such as x and v that are often avoided
in Hebraicized orthographies. Indeed, the vowel-less origins of Hebrew
script are put to good use in the transcription of a language that clearly enjoys
its consonant clusters. Yet quite contrary to most Hebraicized orthographies,
this adaptation tolerates the use of double-ww for CV and VC sequences,40 as
well as three consecutive w (although the niqqud, which is conventionally used
in such transcriptions, mitigates the three-in-a-row effect).
40
Modern Hebrew orthography itself tolerates this usage in the spelling of a small number of
words, e.g. §wwv avon 'sin', §wwyk kivun 'direction', as well as in some loanwords, e.g. ˚d˚w
'voodoo', hAqÌdÙw 'vodka'.
3.8. East Asia
Jews probably first arrived in China along with other merchants on the
Silk Road perhaps as early as the second century BC . The earliest textual
evidence of a Jewish community there, however, comes in the form of an
eighth-century business letter written in Hebraicized Persian. Although there
are no extant documents (to my knowledge) written by Jews in Hebrew-letter
Chinese, community registers in Hebrew do record the names of individuals
from the only substantial Chinese Jewish community, that of Kaifeng (Leslie
1972: 123):
(23) wd wg wg gw gw dw Change Chµ-te

'wp §yg gyn pw√ Chin Fu
£wS §yg gyn ¸swm Chin Shou
In addition, a Jesuit visitor to Kaifeng in 1721 transcribed the following
Hebrew blessing from one of the community's prayer books, along with his
own Romanized rendering of the local Jews' Hebrew pronunciation (Leslie
1972: 120):
(24) J‡dEv√w £‡rOvËr KOt˚hËram £⁄H ™˚rJb

A
vauite rearoam marehunthu scheam Poroke
'and ever' 'forever' 'His kingdom' 'name' 'blessed'
ModHeb [va÷ed] [le÷olam] [malxuto] [ßem] [barux]
The most immediately striking feature in this extract is the confusion of l l
and r r in the third and fourth words, which are properly spelled wtwklm
<mlkwtw> and £lwvl <lfiwlm> respectively. Though it does not attest to an
adaptation of Hebrew script for a language other than Hebrew, this is my first
encounter with a Jewish community's traditional Hebrew pronunciation
interfering in the canonical written form (in this case, probably in the absence
of "official" texts).
3.9. Germanic
3.9.1. Yiddish
More Jews have probably spoken Yiddish than any other vernacular. In
terms of Hebraicization, however, Yiddish is distinguished as the only
language other than Hebrew to continue making exclusive41 use of Hebrew
script. Concerted attempts to standardize the writing system began in the late
nineteenth century and continued after World War II, but having never
ascended to national-language status there has never been a single standard
or convention by which non-Yiddish words are rendered in Yiddish
orthography. The successive adaptations of Hebrew script that constitute the
development of Yiddish writing have been relatively well documented and
are beyond the scope of this study. What is worth noting are the innovations
that are unique to Yiddish among adaptations of Hebrew script, several of
which I have mentioned elsewhere:
• v fi as a vowel letter in addition to the canonical four matres lectionis

• standardized use of non-final p K for /p/ in final position
• semi-standardized subset of niqqud
• semi-institutionalized innovation of a longhand grapheme for /v/
41
This, of course, ignores pedagogical texts aimed at Yiddish-language learners, which may
be Romanized or other otherwise converted to the learners' normal scripts, as well as Roman-
letter electronic environments such as e-mail (cf. chapter 8).
Perhaps unsurprisingly, these innovations are found less often in the earliest
"standard" (i.e. Eastern) Yiddish writing, as illustrated in the excerpt below
from an edition of Tikuney tshuve erets Tsvi 'Responsa of the Promised Land',
published in Cracow in 1666 (Kerler 1999, whose transcription I have
adapted):
(25) gvX yd ly&p 'yww §lyc zwm 'yz rb' øb¥r zi√ muz tseyln vi√ fil di t´˝
& w&tbS §w&p §y&pvrX §' øn tr´f¥n fun ¸sabøs¥s (¸sb±w±) un
£ymwy #nw' t
£y&bwX yøntøyvem (ywmym †w‹bym)
¥ a
Xs&pg Xyn X'h yz z' az zi høt nit ˝ f st
'However, she has to count how

many Sabbath and holy days
altogether she did not fast'.
Note the overall lack of niqqud, the occasional use of non-final letter forms in
final position (#nw' un 'and') and the absence of v from one of its typical
modern roles in the initial syllable of past participles with ge- (Xs&pg gefast
'fasted' for modern Xs'& p vg). Among the other features typical of early
Yiddish that were ironed out in the later orthography are the shtumer-' 'silent
aleph' in final position (e.g. 'yz zi 'she', 'yww vi 'how') and the use of y as the vowel
letter in unaccented syllables (e.g. §y&pvrX trefen for modern §&pvrX).
A noteworthy aspect of modern Yiddish orthography is the way in
which it exploits its dual heritage as, on the one hand, a Germanic-language
adaptation of Hebrew script and, on the other, the Jewish language with the
highest occurrence of Hebrew-Aramaic lexical items. The effect is illustrated in
the following pair of names that appear from time to time in the pages of the
Forverts newspaper:
(26) §'mrvbyl•swy ywsp lybfirm√n 'Joseph Lieberman'
§yl'Xs •vsA'Szd dz¸s√sfip st√lyn 'Joseph Stalin'
The now-weekly Forverts is the lone survivor of a formerly thriving American
Yiddish press. Most if not all of its present American readership can speak
English, and in so doing would likely pronounce the first names of these two
political figures identically. Yet the in- and out-group status42 of these two
rather differently-regarded men is indicated quite categorically by the
rendering of the American senator's name in the orthographic garb of the
biblical Joseph, while the Soviet leader's name receives a distinctly secular
transcription – ironically reminiscent of the more systematic de-Hebraification
(i.e. secularization) of Yiddish orthography undertaken in the Soviet Union in
the 1920s, examples of which are shown below (Estraikh 1999):
(27) tbS ¸sbt svb'S ¸s√bfis ßab\s 'Sabbath'

hyx ˛yh vayyk ƒyyfi xay\ 'animal'
Xvmk kmfi† Xa'myq qym√† kim\t 'almost'
This is actually nothing less than Yiddish orthography as transcription, since
the words are spelled syllable-by-syllable with the least ambiguous
combinations of letters. This phenomenon is certainly seen from time to time
42
This effect is also achieved by using the spirant form of k k rather than x ˛ for any and all
"guttural" fricatives in non-Hebrew vocabulary, even in Arabic words where the sound or
letter in question is cognate with x , e.g. §yyswk <ƒwsyyn> Hussein, dvma'kwm <mwƒ√mfid>
Muhammed, ¢a'Xa'&p <f√†√ƒ> Fatah. Indeed, modern Yiddish orthography makes no special
accommodation for Semitic cognates and treats Arabic lexical items exactly like the Hebrew
transcriptions in (27) and (28). This can again serve to establish the out-group attitude toward
certain names or terms, e.g. Jihad or Jafari written as da'hySzd and yra'&pa'a'Szd with initial dz¸s-
rather than a cognate #g ¸g-.
in non-Soviet Yiddish, when a writer wants to indicate the pronunciation of
Hebrew-language material that might not occur as a normal part of written
Yiddish. In the following examples from a recent Forverts article, the first line
of each pair shows the Hebrew spelling and letter-by-letter transliteration,
while the second line gives the (Yiddish) transcription:
(28) £ySwdq twm yrx' √˛ry mwt qdw¸sym 'After the Death / Sanctified'43
£ySA'dvq XA'm yyra'ka' axarei mot kedo¸sim
wnKtwqKt hd&b' 'l dwv fiwd l√ √‹bdh t:qwt:nw 'our hope is still not lost'
wnyyXa'wwqyX a'dwwa' A'l dA' od lo avda tikvateinu
Notice, however, how this use of the Yiddish writing system differs from the
French transcriptions in the learners' manuals discussed in § 3.3. Although the
matrix of those handbooks is ostensibly Yiddish (or Hebraicized German), it is
not the normal conventions of Yiddish orthography that most directly inform
the transcriptions of French-language material, unlike those in the tables
above.
3.9.2. German
Comprising a linguistic tradition distinct from Yiddish, many Jews were
speakers of standard German or other non-Yiddish German dialects. From
about the sixteenth century onward there are prayer books with instructions
written in non-Yiddish Hebrew-letter German, along with sundry other texts.
In fact, the Methode French handbook in § 3.2.3 presents most of its matrix
43
These are the names of two parshiyot '(Torah) sections' that are often read at the same weekly
service, such that the corresponding Sabbath is often referred to by their joint name. The
expression has gained its own meaning in the sense of "all are holy after death," with the result
that one does not speak ill of the dead.
material in a Hebraicized but otherwise standard German of the early
twentieth century. The table below illustrates features that imitate Roman-
letter German (beyond lexical choice) and that do not occur in later Yiddish
writing:
Table 2-21. Non-Yiddish patterns in Hebraicized German
• silent h
qyXShyrp frih¸stik 'breakfast'
gnyrrhA' ohrring 'earring'
• doubled letters not straddling a morpheme boundary
rrvh herr 'sir'
llA'z zoll 'should'
• pKp ( /pf/ > /p/ or /f/ in Yiddish)
vn˚ha'pKp pfahune 'peacock'
•KpmwrXS ¸strumpf 'stocking'
• Xd <dt>
£wygvllA'qsa'rXda'XS s¸ tadtraskollegium 'town council'
(Yid. XA'XS s¸ tøt city)
3.9.3. English
As a Jewish vernacular, Yiddish probably still holds the title for the
largest number of speakers at any given moment, if not cumulatively. Given
trends both internal and external to world Jewry, however, more Jews may
end up speaking English than any other vernacular. And yet outside the state
of Israel, where Modern Hebrew orthography provides the obvious
framework, there has yet to be anything that approaches a standardized or
institutionalized adaptation of Hebrew script for representing the world's
most widespread language, despite the abundance of present-day Hebraicized
English in, for example, the modern Hebrew and Yiddish press.
The earliest attempts to represent English in Hebrew script appear in
Hebrew deeds from medieval England (Davis 1888), terms associated with
civic or legal matters:
(29) SnylrtS' e¸sterlin¸s 'sterling'

'Syywr'p paroys¥ 'parish'
Sryypyyrgwryc ßirogreyfeyr¥s 'chirographers'
'ccwnXrwp' apor†enuß¥ß¥ 'appurtenances'
In additon, there is a large number of geographical names that appear in
Hebrew script:
(30) ¶y&grwn nor‹giß 'Norwich'
'zrw&gnwh hun‹gurz¥ 'Hunworth'
'rXycnyg ginße†r¥ 'Wincester'
rybw'Xwnq kon†aurber 'Canterbury'
XrwpnSw' o¸senfor† 'Oxenford'
'nwXmld' edelmeton¥ 'Edmonton'
'yyhqnwl lunkhey¥ 'Langehey'
£yyrwd dureym 'Durham'
'XgldS sadelgat¥ 'Saddlegate'
qygrwy' evergik 'York'
'rwS rwS §wXwS ¸soton sur sur¥ 'Sutton-upon-Sore'
tw'rXzwzg gzuz†raut 'Jews' Street'
Like the earliest Romance-language Hebraicizations, the transcription of some
forms is often inconsistent and idiosyncratic (given the prevalence of written
French in thirteenth-century England, Romance may well be a pertinent "co-

matrix" for these early Hebraicizations of English44). The influence of Hebrew
orthography itself is still noticeable in the post-consonant final h h, as well as
the use of t t, which tends to be dropped in favour of X † in Hebraicizations.
Note, however, the absence of other letters such as v , which has historically
been disfavored in all European adaptations of Hebrew script besides Yiddish,
as well as the semi-consistent presence of matres lectionis to indicate vowels
(though it is not always clear which vowel in particular).
By the end of the nineteenth century, with large numbers of mostly
Yiddish-speaking Jews immigrating to North America, the need arose for
language-learning materials similar to the French handbooks in § 2.3 targeted
to this audience. One such manual offers an especially intriguing window into
the use of Hebrew script for writing English. In Alexander Harkavy's (1893)
vd'Ah¯Xvm s#¯pËr'‡dÃnvelA' Ullendorf's Met¥hode (Bibliothèque Médem 15245) – note
again the un-Yiddish Hebraicized-German title – the author presents a list of
words and phrases with the pronunciation in the left column indicated as
rixtig 'correct' and on the right as greizig 'wrong'. These transcriptions are
nothing if not amusing as examples of a writer's attempt to represent Jewish
immigrant-speak and, in a more general sense, as an adaptation of script
whose goal is, in effect, to mis-represent speech. The errors that the writer
anticipates of his readers are revealing:
44
Indeed most of the personal names in these charters are distinctly Gallic (or else Hebraic) in
form. Some, however, do show an interesting blend of the two, such as 'nbwyl 'nbyyXS
ryyXSbl'h <s†eiben¥ leyuben¥ halbasteir> Stephen Le Jouvene Le Arblaster, where the second Le
of the Roman-letter equivalent is actually calqued by the Hebrew definite article h ha-.
Table 2-22. Vowel-related errors in Harkavy (1893)
• syncope
zwA'p
K p
K '
A s ya' ay soppoz zA'p
K sa' aspoz 'I suppose'
• mis-syllabification
pA' yr#rA'h horri op p
K '
A y rA'h hor yop 'hurry up'
• diphthong for monophthong
qnvb benk qnyyb beynk 'bank'
• monophthong for diphthong
SzdnyySX t¸seyndz¸s SzdnvSX t¸sendz¸s 'change'
§wa'X §wa'd dawn tawn §a'X §a'd dan tan 'downtown'
Table 2-23. Consonant-related errors in Harkavy (1893)
• dental fricative simplified

wy qnyyhX theynk yu wy qnyyd deynk yu 'thank you'
gnyhX thing gnyX ting 'thing'

• cluster simplified
Xya'r lA' ol rayt Xya'rA' orayt 'all right'
• devoicing
srwA'q wwA' ov kowrs sA'qpA' ofkos 'of course'
zrvq#sywwh hviskerz svqsyww viskes 'whiskers'

• /r/ added to codas
vl#lvrbmA' ombrelle rvlvrbmA' ombreler 'umbrella'
• /r/ dropped from codas
zrv-A'rd dro-erz zA'rd droz 'drawers'
rvX#r'g garter vX'g gate 'garter'
yz#rywA'Szd dz¸soyrzi vzywA'Szd dz¸soize 'Jersey'
For several of the anticipated errors, however, it is difficult to see what
phonetic goal is served by the left-hand rixtig transcription, such as the
doubled l in vl#lvrbmA' umbrella and its intervening apostrophe. In the case
of zrvq#sywwh whiskers, the need to pronounce the /r/ in the final syllable and
to avoid devoicing the final constant is shown clearly enough. Yet simply
adding h h to the initial double-ww that in the greizig form stands for the
incorrect [v] does not successfully suggest the [hw] that the author has
presumably intended,45 because an effective transcription should not behave
like a conventional orthography, using one grapheme (in this case a digraph)
to represent different phonemes. The transcription of the th-initial words
seems especially unhelpful, since the -hX is simply an orthographic calque of
<th->. Replacing the initial d in wy qnyyd [dej˜kju] with two letters that represent
[t] and [h] respectively as a way to indicate the correct [†] pronunciation
probably requires a good deal more familiarity with the English-language
adaptation of Roman script than readers were likely to have.
4. SUMMARY
The foregoing survey vividly illustrates some of the unique aspects of
Hebraicized writing discussed in the first chapter. Although originally
adapted by Semitic speakers for a Semitic language, it is clear that the original
nature of the script has proven no impediment to its later Jewish adapters.
Moreover, the retreat of Hebraicization from the modern landscape of writing
does not imply that the phenomenon was inherently marginal; it is only in
retrospect that the markedness of Hebraicized writing emerges.
It is interesting to note the implicit focus on this adaptation – which, as
argued in the previous chapter, is not different in kind from the numerous
adaptations of scripts that have made writing itself possible – in a good deal of
European anti-Semitic thought of the last millennium. As Gilman (1986)
45
Curiously enough, this is precisely the strategy of the Yiddish Forverts to spell non-final [w],
e.g. §a'wwayyX Taiwan, ayywwgwrw' Uruguay, ywwa'qra'z-la' Al-Zarqawi – unless, as perhaps is the case for
first two words, the double-ww indicates the pronunciation qua Yiddish with [v] (cf. chapter 8 §
1.2). Modern Hebrew, by contrast, exploits the historical value of a single w to render /w/, for
instance in bilingual dictionary transcriptions, e.g. ˚√w <wau> wow, Ù'w <w√o> whoa, Lp¯s«w#dl«w
<wild’w isp> will-o'-the-wisp (Segal and Dagut 1991).
explains, such writers expressed a belief that Jews could never fully command
the language they ostensibly spoke or wrote, and considered the use of the
Hebrew alphabet to be both a consequence of this deficiency as well as
evidence for the existence of a "hidden language" that purposefully concealed
the true expression of "Jewish thought." In reality, of course, writing their
language in Hebrew script – particularly for Romance-speaking Jews – was no
more marked, devious, or problematic an act than it was to be Jewish itself.
Furthermore, as I will argue in the case of Judeo-Portuguese, this act did not in
and of itself entail (though it would not exclude the possibility) that the
language of composition had any particular Judaic character beyond the script
itself. Before turning to a more detailed description of the Portuguese
adaptation of Hebrew script, then, it is worth recalling the caution by Aslanov
(2001: 5):
Il faut se garder de projeter sur la situation linguistique des juiveries

médiévales d'Europe occidentale une terminologie et une analyse qui
conviennent davantage a la description des judéo-langues parlées dans
le Temps [sic] modernes en Europe Orientale, dans les Balkans ou en
Asie, après que l'Occident chrétien eut rejeté à la périphérie ou en
dehors de son domaine les Juifs qui y avaient vécu des siècles durant.
One should be wary of projecting onto the linguistic situation of [Jews in]
medieval Western Europe that are better suited to the description of Jewish
languages spoken in modern times in Eastern Europe a terminology and
analysis, in the Balkans, or in Asia, (which emerged) after the Christian West
had driven the Jews who had lived there for centuries to the periphery or out of
their domains.
78
CHAPTER THREE
THE JUDEO-PORTUGUESE CORPUS
In this chapter I introduce the corpus of Old Portuguese in Hebrew

script that forms the basis of this study. I discuss various ways in which the
texts may be approached as linguistic artifacts, and some problems raised by
each of these approaches. I then provide a detailed description of the Judeo-
Portuguese writing system, followed by a description of the Romanization

system I employ in the critical editions of chapters 4-6.
1. JUDEO-PORTUGUESE IN CONTEXT
The corpus of Hebraicized Portuguese examined in chapters 4-6
comprises five manuscripts, which are briefly described below along with a
sample from each one:1
• Chapter 4. Biblioteca Palatina (Parma, Italy), ms. 1959 (formerly ms. De

Rossi 945): Syrwq S' §yz'#p yS wmwq yd wrbyl w' O libro de komo se fazen as kores, a
treatise on the techniques of manuscript illumination and recipes for inks and
dyes, composed at the earliest in 1262 (Sed-Rajna 1971). First published by
Blondheim (1929-30) based on a photograph of the manuscript, his edition
provided a Hebrew-letter transcription, Romanization, and English translation
(though no philological commentary). The edition I offer in this study updates
and expands on newly-edited excerpts first published in Strolovitch (2000c),
1
A facsimile from each of the manuscripts is presented in the appendix section.
79
and presents the entire text in critical edition (though without a full Hebrew-
letter transcription). The following excerpt presents a list of the ten "principal
colors":
(1) f.15r.
ydryw §wyylymryw y' wXnymyprw' lwz' Sy'pySnyrp Syrwq S' w''S ¶yd yq yb'S
lyS'rb ydlyywwl' §wqrz' w''rp's' lwS'X'q yy#pws §ymrq
Sabe ke deß sao as kores prinsipais azul oripimento e vermelyon verde
karmin sufi katasol açafrao azarkon alvayalde brasil
'Know that the principal colors are ten: blue, oripiment and red, green,
carmine, sufi, sunflower, saffron, zircon, white-lead, brazil-wood'.
• Chapter 5. Bodleian Library (Oxford, England), ms. Laud Or. 282: wrbyl w'
2
'qy&g'm yd O libro de ma‹gika, an early-fifteenth century copy of an astrological
treatise attributed by the scribe to Swgrwb yd ly&g §'w&g ‹guan ‹gil de burgos. At 800
pages (each containing between 29 and 31 lines), this manuscript constitutes
on its own more than half of the known Judeo-Portuguese corpus.
Nevertheless, a single transliterated folio is all that has been published
previously (Gonzalez Llubera 1953). In the excerpt below the twelve names of
the Zodiac are introduced:
(2) f.5v.
§w'yprwqSy' 'rbyl wgryw §w'yl rysn'q ynymy&g wrw''X Sryy' yS' §wr'ymwn y'
Sysyp wryy'q' w'ynrwqypq w'yrXyyg'S
a
e nomearon asi ayr s tauro ‹gemeni kançer leon virgo libra eskorpion
sageytario akayro piçes
'And [the sages] named them thus: Aries, Taurus, Gemini, Cancer, Leo, Virgo,
Libra, Scorpio, Sagittarius, Aquarius, Pisces'.
2
Although no date appears in the manuscript itself, the copyist was probably the same as that
of Bodleian ms. Laud Or. 310, who gives the date of completion for that text as a late-summer
Friday in the year [51]71, i.e. 1411 (Levi 1995).
80
• Chapter 6. Three shorter texts:

(a) Bodleian Library (Oxford, England), ms. Can. Or. 109: Rubrics for the
Passover seder (ritual meal) in a Hebrew ma˛zor (prayer book) dating from
1485. Facsimiles and Roman-letter normalization of the text were first
published by Salomon (1980); a critical edition first appeared in Strolovitch
(2000b) and is further elaborated in this study. The excerpt below describes
the ritual of ¢rwk koreƒ, the eating of "Hillel's sandwich":
(3) f.240v.
§Ùn yE' &tesÙrox ÙnyE' 'flry√y¯lÙm yE' hosAp¯la' 'J‡d yE' hA'o'oS hacom 'fir¯XËyyÙ' 'fld 'flromÙX yE'
h&okflr¯b '&√gyJÊd
a a a a
e tom ra da oytra maßa saah e da alf ç h e moly ra eno ˛aroset e non
diga beraƒa
'And take from the other unbroken matza and from the lettuce and dip [them]
in the haroset and do not say a blessing'.
(b) Brotherton Library (Leeds, England), Roth ms. 71: Passover rubrics from a
Hebrew ma˛zor, dated by Jewish historian (and former owner of the
manuscript) Cecil Roth to the late fourteenth century (Metzger 1977). Also
first published in Strolovitch (2000b), but since that edition omitted all niqqud,
which was not discernable in the facsimiles of Salomon (1980), it is reproduced
and elaborated in this study. The excerpt below explains the size of the
portion of matza to be eaten:

81
(4) f.5v.
yE' rE&tÙy 'O»l &tÙxa&p '»l 'ƒnÙX¯yy≈z‹' &hA'˚' y„Jd &hA'yiX¯nÙq 'AlyE' y„Jd SÙ&dÙX w'&'flryEmÙq yE'
ÙyyÃny«w y„Jd HÙH'Aww 'b HÙ' Ù''flry&gÃnyE'
e komerao todos de ela kon tiah de uah azey¥tona lo pa˛ot lo yoter e
¥
en‹gerao os ba vasos de vin¥yo

'And everyone [is to] eat from it the amount of an olive – no less, no more –
and fill the ku- cups with wine'.
(c) Cambridge University Library (Cambridge, England) ms. Add.639.5: a
single ten-line medical prescription on the last page of a manuscript

containing gnomic verse in Judeo-Spanish (see Gutwirth 1992), which is edited
and published here for the first time. In the Cambridge University Library
catalogue of Hebrew manuscripts (Reif 1997) the Portuguese text is
misidentified as Spanish. The excerpt below is the final instruction in the
prescription:
(5) f.20r.
w' §ybn'X y' ylyp Sy'm h'yb' §wn yq §y&p wrXw' ryXym y' ws'dyp §w' r'ryX y'
ws'rb wd wgwl wrgn'S
e tirar un pedaço e meter otro fin ke non abiah mais pele e tanben o
sangro logo do braço
'And [I] remove one piece and place another until there is no more skin, and I
also I bleed it over the coals'.
Before I turn to at the characteristics of the corpus as a whole, there are

several distinctions to note among the five manuscripts. The passage from O
libro de komo se fazen as kores (henceforth As kores) illustrates the quasi-technical
vocabulary (often borrowed from Arabic) that occurs in that text, which
amidst other features gives it the most varied linguistic profile in the corpus.
Among my editions, O libro de ma‹gika is the only one limited to a fraction of
82
the manuscript (the first fourteen of its 400 folios are presented). The Passover
material in chapter 6 is distinctive for its non-contiguity, in that it comprises
individual sentences interspersed among Hebrew blessings. It is also the only
material that has been systematically vocalized with diacritics (although each
of the two larger texts contain isolated forms with niqqud), and the only texts
to feature words of Hebrew origin, usually in relation to the Passover rituals.3
In addition, the Brotherton Passover text is the only one in the corpus not
written in the cursive Rashi script, but rather in square characters. These
distinctions aside, the texts form a coherent corpus based on date (thirteenth to
fifteenth century) and on region of production (Portugal), as well as on the
similarities of their linguistic form, both genealogical (Western Ibero-
Romance) and graphical (Hebrew). The immediate question, then, is what
might one hope to gain from a linguistic study of this corpus.
1.2. Approaches to Old Portuguese in Hebrew script

These Portuguese-language texts can be approached first and foremost
as synchronic and perhaps diachronic windows onto the history of the
Portuguese language – and indeed this is the primary focus of the critical
editions in chapters 4-6 and the summary in chapter 7. Because many

distinctive features of modern Portuguese began their progress in the
medieval period, the orthographic strategies of these late medieval writers
might reveal synchronic information on the spread of nasalization,
palatalization, and other phonological phenomena that may be less detectable
in the more conventionalized, tradition-laden, Roman-letter writing system.
3
The Cambridge medical text also contains one Hebrew word, hmhb behema 'animal', though
the context there is decidedly non-religious.
83
Despite my warning against treating Hebraicized texts as phonetic

transcriptions in chapter 1, there is little doubt that having been undertaken
later in history than the Roman-letter adaptation, the Judeo-Portuguese
writing system captures the language, in whatever indirect fashion, at a more
recent stage in its history.
This approach comes, of course, with the usual caveats of historical
linguistics, principally against assuming that what a writer has written was
motivated by his or her peers' spoken language (specifically his own
production or perception of it). Writing systems whose corresponding

vernaculars are in the throes of phonological change that is not "officially"
reflected in the orthography are ripe for so-called inverse spellings. The
orthographic confusion of and <V> in imperial and early medieval Latin,
for example, may be taken as evidence for a phonological merger4 – not as
evidence that a given writer intended to (incorrectly) spell [b] in some cases
and [v] or [w] in others.
Even when the variants differ in subtler ways, it is always problematic
to use the silent artifact of a small set of writers as evidence for the speech
habits of a larger community. For example, Dutch Jews writing Roman-letter
Portuguese in the seventeenth and eighteenth centuries frequently (and

inconsistently) confused the spelling of a variety sibilants, e.g. suseção for
sucessão 'succession', grassa for graça 'grace', felizidad for felicidad 'happiness',
resar for rezar 'pray', and so forth (see Teensma 1991 for many further
4
For example, among the over 225 entries in the Appendix Probi (ca. 300 CE), which correct
some of the lexical, phonological, and orthographic lapses in the Latin of the day, are the
following that attest to the merger of /b/ and /w/: BACULUS NON VACLUS 'staff', ALVEUM NON
ALBEUS 'trough', PLEBES NON PLEVIS 'plebeians', TABES NON TAVIS 'decay', VAPULO NON BAPLO 'be
beaten'
84
examples). Yet it would be hasty to deduce a process of phonological merger

among the phonemes represented by these graphies from the writers'
apparent uncertainty about how to spell them. Since the use of the Portuguese
was on the decline in those centuries, the wide variation is more likely due to a
waning knowledge of orthography – not to mention interference from the
writers' tenuous familiarity with Spanish and French orthography – than
evidence for variant pronunciations from a given writer or the community at
large.
The corpus of Hebrew-letter Portuguese contains comparable cases

where conclusions about orthography-as-transcription may be too hastily
drawn from the written artifact. For instance, Metzger (1977) points out that
when the fifteenth-century Bodleian Passover manuscript was catalogued by
Neubauer (1886), he believed it to be of Spanish origin, with its rubrics written
in a form of Castilian. The source of his mistake was no doubt the frequency
with which the scribe had used the letter § n on determiners, pronouns, and
third-person plural verbs, which he took to indicate that the word-final nasal
consonant was still "present" as in Spanish (as opposed to only the nasalized
vowels present in Portuguese). This is illustrated in the following excerpt
from folio 241r.:
(6) §˚' 'fldAq §A'wyJEb

bevan kada un
'each one drink'
And yet this feature is all the more curious given that the ostensibly earlier
Brotherton manuscript contains no visible nasal consonant in its determiners
85
and verb endings. Instead it features spellings with vowel-letters only, as

illustrated by the corresponding phrase from folio 5r.:
(7) w'&A'fld“q ÙA'flryE&by&Eb

‹be‹berao kadau
'each one (will) drink'
A spelling such as this more clearly suggests that scribe meant to indicate
nasalized word-final vowels. Adding ambiguity to the situation is the Parma
colours text, in which pairs such as w''qrz'/§wqrz' azarkon/azarkao 'zircon'
alternate throughout, showing both an innovated and conservative spelling
(see chapter 7 § 1.1). And the Roman-letter transcriptions of Hebrew names

cited by Garbell (1954: 658) also show inconsistencies in the conventions of
usage for <m> and <n> in similar environments.
That this orthographic alternation attests to some form of phonological
variability in the community is beyond doubt, given the known development
of word-final nasals from Latin to Portuguese. Yet it does not follow that the
written variants themselves correspond to real utterances manifesting this
variability. It is difficult to decide whether the presence of final § n in the
Bodleian Passover text is an archaizing spelling, reflecting the traditional
written form of the word irrespective of its phonological form, or else a
diacritic spelling, where the final letter simply signals that the preceding
vowel is nasalized in speech. Indeed, both of these characterizations could
apply to the <-m> in the modern Roman-letter orthography of the Passover
phrase above (ModPg. bevam5 cada um), since it is in fact a restored spelling
5
Although this verb is one of several imperative forms that occur in the text, most of the verbs
in fact appear as future-tense forms, which curiously enough is the one conjugation in the
86
that nonetheless signals the nasalized ending produced in spoken Portuguese.

The most that can be definitively ascertained from the intra-textual variability
of final-nasal spelling is that both orthographic variants were available to the
scribe; the conditions that led him to write one or the other in a given instance
may be lost to the modern reader.
1.3. The "Real" Judeo-Portuguese

In addition to the relatively straightforward historical-linguistic
approach, one might delve into this corpus for the insight it might yield into a
particular sociolinguistic situation, that of the Jewish population in Portugal.
Vernacular documents from this group are especially scarce, as are studies of
them: a book-length study of the Portuguese Jewish community by Tavares
(1992), for example, makes only passing reference to one of the texts discussed
here (As kores) as part of the community's "cultural production" (the remainder
consisting of Hebrew-language material). In the adaptation-of-scripts context
of this study, one might hope to show that the processes of adaptation that
yielded these texts constitute the beginnings of the tradition of Hebraicized
Ibero-Romance that flourished in the centuries following the Spanish and
Portuguese expulsions (cf. chapter 2 § 3.2). Even more tempting might be an

analysis of the texts that yielded evidence for a distinct variety of Judeo-Ibero-
Romance, a forebear of the attested Portuguese dialect spoken by some
descendants of the lusophone émigrés in the centuries after the 1496-97
Portuguese expulsion and forced conversion.
modern writing system whose third-person plural /ãw/ is not spelled with <am> but rather
<ão>.
87
It is this last, perhaps most enticing prospect that leads to the least
conclusive areas of research. No modern-day lusophone population has
descended from the Portuguese-speaking Jewish community, which shifted to
co-territorial languages such as Spanish, Dutch, and English by the nineteenth
century. In fact many of the émigrés from Portugal were Spanish speakers
expelled from Castille-Aragon a few years prior to the Portuguese edicts of
1496-97. The Portuguese speakers who left the peninsula to settle in Italy, the
Balkans, and Turkey assimilated to the Spanish-speaking majority, thus
beginning the long-term language shift that eliminated Portuguese from the
Sephardic repertoire. With a relative shortage of material there have
consequently been very few linguistic studies of Jewish Portuguese, apart
from those focused on written records from specific cities where Jews settled,
such as Amsterdam (Teensma 1991) and Livorno (Tavani 1988).6
Judeo-Spanish, the only Judeo-Ibero-Romance language still spoken
today, certainly boasts a richer documentary history from both the Iberian
Peninsula and the resettled communities of the Ottoman Empire and North
Africa. Yet its existence prior to the expulsions remains a vexed question (cf.
Marcus 1962, Wexler 1982). The question of a distinct Judeo-Portuguese may
at first blush seem less "vexed" simply because, given the small extant corpus
and absence of a modern speech community, the field is less ploughed.
Moreover, the prospects for discovering the expression of a distinct (spoken)
dialect amidst the short ritual prescriptions and non-Judaic scientific discourse
in the Hebraicized Portuguese corpus may well be discouraging. The corpus
6
The only book-length study of Jewish Portuguese in general appears to be a Ph.D.
dissertation at the University of Lisbon by da Silva Germano (1968), which I have been unable
to access.
88
is above all a written artifact, and the use of Hebrew script is simply not a
sufficient condition for presuming it to represent the early rumblings of the
elusive pre-expulsion Judeo-Portuguese dialect.
The seemingly trivial issue of what to call the language can aggravate
this issue, particularly when there is no longer a community of native
speakers. Frakes (1989), for example, talks about the form and variety of
names devised – largely by non-natives – for the language of what he calls the
"Old Yiddish corpus" as an exercise in identifying the object of research in
order to (de)legitimize it: Germano-Judaic, Hebrew-German, Jewish German,

Jµdisch-deutsche, judéo-allemande, etc. As he explains, these ostensibly
descriptive statements about the language are decidedly prescriptive and
ideological in their view of linguistic history. And yet it is not only in the
absence of native speakers that glottonymics can prove troublesome. Judeo-
Spanish speakers7 (and researchers) have their own array of names for that
language: Ladino, Judezmo, Jídiyo, Spanyolít, Sefardí, etc.
For the present purposes I have somewhat grudgingly adopted the
term Judeo-Portuguese. A more accurate, or at least less constraining term
might be Hebreo-Portuguese, since it does less to suggest distinctive features of
the language beyond the only one apparent, namely its writing system. Short
of a direct declaration, of course, there is no way to know what the native
glottonym was. And although neither of these terms is a viable candidate, this
study is targeted at an audience for whom the term Judeo-Portuguese will be
7
Prominent native speakers with upwardly-mobile aspirations may be particularly unhelpful.
Baruch/Benedictus Spinoza, a Sephardic Jew born in Amsterdam, referred to his native
language simply as "Spanish," while Moses Mendelssohn, the principal figure of the
eighteenth-century European Jewish Enlightenment, spoke of his native Yiddish as the
"Jewish-German" dialect (Gilman 1986: 105).
89
eminently conventional and meaningful. As a starting point to probing the

nature of this corpus, then, I turn to the mechanics of the adaptation of
Hebrew script that constitutes written Judeo-Portuguese.
2. THE WRITING SYSTEM

Having surveyed in the previous chapter the orthographic patterns that
occur in various adaptations of Hebrew script, I now offer a more detailed
description of these patterns as they occur in the Judeo-Portuguese corpus. As
a model for an orthography-oriented study of Jewish language, this work has

been strongly informed by Freedman (1972), a study of Italian texts in Hebrew
characters that focuses on the writing system. I have also drawn on the
diachronic studies of Judeo-Spanish writing by Minervini (1992) and Recuero
(1988), which examine the pre- and post-1492 corpora respectively, as well as
from Dominocovich (1948), one of the few English-language surveys of
(Roman-letter) medieval Portuguese orthography available to me. The corpus
of Hebrew-letter Portuguese does not have the breadth to allow one to trace a
chronology comparable to those just cited. I do believe, however, that the
description below offers some novel approaches for the study of a Hebraicized
writing system, by focusing on features broader than the patterns of usage for
individual letters. As will be argued below, these features constitute the
fusion of conventions that firmly positions Hebrew-letter Portuguese in the
annals of Hebraicization, while at the same time distinguishing it in the
adaptation-of-scripts framework.
90
2.1. Independence from the mother script

The following orthographic strategies, while not unique to Portuguese
in the Hebraicization canon, do represent the particular responses by Jewish
Portuguese writers to the interface of Hebrew script and the Portuguese
language that confronted them. More importantly, perhaps, they set the spirit
and attitude of the writing system apart from both contemporary Roman-letter
and Hebrew-language writing.
2.1.1. Vowel letters

Above all, the Judeo-Portuguese texts represent a mature adaptation of
Hebrew script in their extensive and systematic use of the matres lectionis as
straightforward vowel letters. This simply could not have been an
orthographic strategy were the writers of these texts still bound to the basic
conventions of canonical Hebrew orthography,8 where the majority of vowels
are imputed by the reader. Nevertheless, unlike the progressive trend that
Minervini (1999) discerns in pre-expulsion Judeo-Spanish texts, it is difficult to
perceive in the small corpus any developmental history of vowels tending to
be spelled with letters rather than diacritics (or with no vocalization at all); in
all the extant texts Judeo-Portuguese writing is a fully alphabetic system.

Although the matres lectionis are in this function analogous to the
vowels of Roman script, they are not deployed as one-for-one substitutions –
nor could they be, since the four matres are no better a match for Portuguese
8
In the broader context of vocalization in adaptations of Hebrew script, the only true
innovators are Germanic-language writers, who, as noted in the previous chapter, use the
non-mater v fi for /e/. This letter is never used to spell native words in Judeo-Romance
writing beyond isolated glosses, and its use in Judeo-Arabic, Judeo-Persian, and Hebrew-letter
Turkish is usually calqued on the behavior of the cognate Arabic consonant Ÿ fiayn.
91
vowel phonology than the five Roman letters. In an unpointed text, two of the
matres (y and w) only indicate the vocalic distinctions "non-low front," and "non-
low back" while ' 9 and final vocalic h may stand for /a/, /e/, or /o/ (see §
2.2.1 below). Yet the full system of niqqud is ultimately superfluous for
Portuguese vowel orthography because, if the diacritics are taken for their
historical Hebrew-specific values, it indicates phonological distinctions that
are redundant in Portuguese. When niqqud is fully deployed, as in the
Passover texts, it tends to be induced simply by other Hebrew-language
material in the neighborhood. Even in such cases, however, it is usually

redundant10 (and occasionally mis-deployed), as illustrated by the following
examples from vernacular passages in the Passover texts:
(8) A'As'A&p¯la' alfaça(a) 'lettuce'

§A'‡ry∆wyeJb bevera(a)n '(will) drink'
a
§rA'AsÕnyEmÙq komença(a)r n '(will) begin'
The <a> in parentheses in the Romanizations of each of these words is

indicated by a qameß, the T-shaped diacritic for /a/ that appears beneath ' √
(itself the letter that indicates /a/ in this writing system). In addition to the
qameß under the consonant preceding ' (which is ipso facto redundant), the
9
Minervini (1999: 420) claims that in the earliest period of Judeo-Romance, the vernacular was
written "in accordance with Hebrew graphic norms" and that ' could represent any vowel,
e.g. §'b ben 'come', ryl'd doler 'pain', hylybS' Isbilia 'Isabella'. She attributes this to its
"incomplete acceptance as a mater lectionis for /a/ in Hebrew orthography and its nature as a
tendentially graphic element, disconnected from pronunciation." Although she cites only
Judeo-Italian as another graphical tradition attesting to the "weakness" of ', it is certainly the
case in Judeo-Portuguese that no other vowel letter may be omitted as readily as '.
10
Note that while Modern Hebrew orthography may make use of both niqqud and the matres
lectionis, the two strategies usually overlap only in the spelling of initial vowels (which require
a diacritic and niqqud-bearing ') and word-final /a/ (see § 2.2.1 below).
92
segol and ßere that indicate /e/ in beveran and komençaran respectively each
underlies a consonant that is followed by y y, the letter that serves the same
role.
The redundant niqqud is most likely due to the delegation of lettering
and vowel-pointing to separate individuals in the production of Hebrew
manuscripts. The naqdan 'pointer' may have been less familiar with the vowel-
letter conventions of Hebraicized Portuguese than of Hebrew itself. As a result
it is not surprising to see that the diacritics, while not fundamentally wrong –
the naqdan was quite likely, after all, a Portuguese speaker – do not play a
crucial role in the writing system.
In a very few instances, however, niqqud is used in an unpointed text
with words that a given scribe may have considered "learned" or related to a
technical context perhaps unfamiliar to the reader. The words in (7) below all
occur in As kores, which is otherwise completely unpointed:
11
(9) SyÇ„dy' ides 'that is'
ryEp¯la' alfer 'bishop'
yECX¯S˚#p fuste '(wooden) stick'
yilwbÕna'wqËr¬za' azarkoanboli 'zircon'
2.1.2. Merged segments

Another trend that appears to have matured in the Portuguese corpus is
the tendency in Hebraicized writing to disfavor one of each pair of letters
whose phonetic values in the community's traditional pronunciation of
11
The three dots that appear above the d are used elsewhere in this text, most often at clause
boundaries, in the same horizontal space as the letters. This is probably the segolta of Tiberian
pointing, one of the stronger disjunctive accents indicating a pause (G. Rendsburg, p.c.). As a
symbol above a given letter, however, it appears to carry no orthographic meaning.
93
Hebrew are identical (usually because one of the historical values does not
exist in the vernacular phonology). In the case of Romance languages this
generally applies to three pairs of letters: q/Jk [k], X/t [t], and b/w [v]. In the
case of the first two sounds, it is the first member of each letter pair, the
historically emphatic (pharyngealized) Hebrew consonant, that is used almost
exclusively to spell the relevant sound in Hebraicized Portuguese;12 in the case
of [v] a semi-systematic division of orthographic labour is put into effect (see §
2.3.1 below). For each pair, the member that is disfavored for the writing of
native vocabulary is always preserved in the spelling of words of Hebrew or

Aramaic origin, whose historical spellings are uniformly maintained.13 An
exception of sorts is the pair s s and S ¸s, whose distinction in most Romance-
language pronunciations of Hebrew is maintained but which are often used
interchangeably for sibilant phonemes (see § 2.3.2 below).
2.2. Reliance on the mother script

There is no doubt that the writers of Hebraicized Portuguese were
literate readers and writers of Hebrew itself, and that their audience was more
or less similarly versed in Hebrew writing (though perhaps less familiar with
12
Minervini (1999) notes that the same choice is made in both Hebrew and Arabic aljamía
(Hebraicized and Arabicized Spanish). With regard to Judeo-Spanish writing, she suggests
that the fricative pronunciation of the non-emphatic stops in weak position, i.e. the reflex of
Hebrew spirantization, may be the motivating factor. Recall, however, that in Yiddish, where
speakers have merged k x and x ˛ in their pronunciation of Hebrew, the writing system opts
for the non-emphatic "spirantized" first member in the spelling of non-Hebraic vocabulary
(neither k nor x is used to spell non-Hebrew vocabulary items in medieval Judeo-Romance,
though Modern Spanish /x/ is, as expected, rendered by k in modern Yiddish orthography).
13
As noted in the previous chapter, Yiddish in early Soviet Russia represents the only
concerted effort to re-spell the Semitic component "phonetically" in a Hebraicized writing
system, part of a state-sponsored strategy to purge the language of any religious character or
association (see Estraikh 1999).
94
Roman-letter writing). Thus it is inevitable that the writers would draw on

conventions of Hebrew-language writing that could maximize the reading
comfort-level of their audience. Again, while these strategies are not unique to
Hebraicized Portuguese, they do illustrate the unique fusion of conventions in
the writing system insofar as they do not have direct counterparts in Roman-
letter writing.
2.2.1. Final /a/ allography
In modern written Hebrew the letter h represents the voiceless glottal

fricative /h/ in syllable onsets, but is "silent" in word-final position, where it
normally spells /a/ (and sometimes /e/).14 Although Hebrew medial /a/ is
generally either not spelled explicitly or else rendered by ' √, the final -a of
feminine singular nouns and third-person feminine singular inflection on past-
tense verbs are both spelled overtly and uniquely with final h h, as illustrated
below:
(10) a. hghn nahaga 'drove' (fem.)

hghnth hitnahaga 'behaved' (fem.)
b. hgyhnm manhiga 'leader' (fem.)
14
Though some nouns do contain final /e/ in the lexicon, e.g. hdW /sade/ 'field', in unpointed
script final h may be grammatically ambiguous in adjectives and verbs, e.g. both the
masculine and feminine forms of 'lovely', /jafe/ and /jafa/, are spelled hpy, while hcwr 'want'
spells both the masc. sg. /rotse/ and fem. sg. /rotsa/ of the present tense. Final h also
appears in ancient inscriptions and vestigially in the Bible as a spelling for the 3rd masculine
singular possessive enclitic -o < *-ahu (replaced in later orthography by Ù-). It remains,
however, the normal spelling for the fem. sg. possessive enclitic -a.
95
In direction imitation – and, significantly, with no precedent in Roman-letter

writing – Judeo-Romance writing makes a nearly-systematic use of the
alternation between ' √ and h h for the spelling of the word-final /a/ that
occurs in grammatical contexts analogous to those in Hebrew, namely
feminine singular nouns and third-person singular verb inflections:
Table 3-1. Final /a/ allography
h'w' uah S'wd duas 'one/two'

hA'A'AS saah SA'A'AS saas 'whole'
hsnw' onçah 15
s'snw' onçaç 'ounce(s)'
hXSyd destah S'Xsyd deçtas 'of this/these'
hly#gyX ti¸gelah S'ly#gyX ti¸gelas 'bowl(s)'
h#gyS se¸gaah §'#gyS se¸gan 'be' (pres. subj.)
This allography is not merely a luxury of the script:16 since in Hebrew the
"silent" vocalic h does not occur anywhere except in word-final position, the
plural markers S ¸ s and § n can only be preceded by the ' allograph of /a/.
This variation has no phonological basis in Portuguese, nor does it have an
analogue in the Roman-letter orthography of Portuguese nor any other
Romance language with similar morphology. Moreover, it is sufficiently
characteristic of the orthography that the feminine singular definite article, a

single <a> in Roman script, is spelled h' (plural S'); only where the article is
15
On the final s see § 2.3.2.
16
In the cursive script used by medieval Sephardic writers, h is actually the only grapheme
with its own final-position allograph beyond the canonical five (cf. chapter 2 § 2.1). In fact the
character presented in table 2-9 is the non-final form E , which, though it does resemble the
standard square h , occurs extremely rarely in the Judeo-Portuguese corpus (almost
exclusively in Hebrew words). The far more frequent allograph that occurs in final position
more closely resembles an inverted Greek V (which is, curiously enough, the final-position
form of sigma).
96
graphically attached to the noun – itself probably in imitation of the practice

for the Hebrew definite article -h ha – is the single ' used:
(11) 'yym h' hcm h' a maßa ah meia 'the half matza'
hAcAm 'AyyemA' ameya maßa 'the half matza'
2.2.2. ' √ as a diacritic

Although /a/ may be represented by several sub-linear diacritics (cf.
chapter 2 § 2.4.1), this Portuguese vowel is most often spelled by the letter ',
which in the grapheme-to-phoneme correspondence of the script uniquely
represents the low vowel. This choice is motivated by both Semitic and
Romance tradition: as the first letter in the Hebrew alphabet, ' is a perfect
analogue to – and related to the ultimate source of – the Roman letter <A >.
Moreover, as noted in chapter 2, in late antiquity this original glottal stop had
already come to be used occasionally in Hebrew to fill the same orthographic
function as would its descendent in Roman script.
Since absolute-initial vowels do not historically occur in Hebrew, a

single y y or w w in word-initial position is read as a consonant in written
Hebrew unless it is preceded by an unpointed ' , which indicates that the
following w or y is vocalic.17 With rare exceptions this convention is strictly
17
C. Rosen (p.c.) points out that some instances of initial <au> in Gascon orthography reflect
an etymological unstressed /o/, e.g. auherir 'offer' < OFFERIRE, augan 'this year' < HOC ANNO.
Alba Salas (2000: 122) notes a similar case in thirteenth-century Catalan aucïea 'kills' < OCCIDET,
and calls the <au> "a clear case of hypercorrection." Yet it is not clear how fully the diphthong
represented by historical <au> had been levelled at this stage in (Gallo-)Romance, and thus
how conventional <au> could be construed as a spelling for /o/. Moreover, the putative
sound change involving initial unstressed /o/ > /aw/ is not well motivated. This raises the
possibility that Catalan and Gascon writers have imitated the convention adopted by Judeo-
Romance writers (whether or not as a direct influence), using <a> as a diacritic to indicate
"vocalic ." Thus the modern Gascon reading of these instances of <au> as a diphthong
would reflect a "spelling pronunciation" rather than historical sound change.
97
preserved in Judeo-Portuguese writing (indeed in Judeo-Romance more

generally as well):
(12) 'bXw'w' outaba 'eighth'

w'w' §y' wdwX todo en uo 'all at once'
y#gny' yS yq Syryw yS y' e si veres ke se in¸ge 'and if you see that it fills'
This convention in fact applies more broadly in Hebraicized Portuguese to
syllable-initial vowels other than /a/ as well as to a vocalic w or y that occurs in

hiatus. In these instances the letter is usually preceded by a diacritic ' to
indicate the vocalic reading:
Table 3-2. Non-initial vocalic w and y preceded by '
Sw''ryw veraos 'summers'

SwnwXw'' autonos 'autumns'
Sw'w'wsnyl lençouos 'bedsheets'
'Sw'wq yy' §wn non ay kousa 'there is no thing'
h''qrys Sy''m mais çerkaah 'closer'
ry'wpnwq yd y'ysymwq' akomeçei de konpoer 'I began to compose'
In fact, so conventionalized is the digraph that a second ' is necessary to

indicate the diphthong in autonos above, even though the w-' sequence is the
letter-for-letter equivalent of <ao>. In the case of the third-person plural verb
inflection and the other /aw/ diphthongs (usually from deleted /n/ or /l/)
that pervade Portuguese, all of the texts in this study feature an alternation
between forms with one or two alephs, since the letter does not conventionally
serve its silent diacritic role following consonants (cf. veraos above). The same
is true of the /aj/ diphthong in some words, as in Sy(')'m ma(a)is 'more' above.
98
Additionally, there are some contexts in which the ' could almost be
viewed purely as a device to avoid a sequence of three identical letters (cf. the
Yiddish strategy for avoiding three w in ch.2 § 3.9.1):
(13) yy'yr'XwrX trotarey 'I will deal with'

y'yyl'&p §wn w'y' eu non falei 'I did not find'
The final sequence in trotarey is especially interesting, given that yy preceded

by ' more often serves to indicate the /aj/ (as opposed to /ej/) reading, e.g.
wryy'rXnwq kontrayro.
In yet other instances, while the diacritic function of ' is not strictly
necessary for a correct reading of some matres sequences, there is a "visual"
convention (probably based on Hebrew writing as well) that compels the
Judeo-Portuguese writer to include it:
(14) wy's w'y&bwn w' o no‹bio çeo 'the ninth heaven'

h'ygwlwrXSy' estrologiah 'astrology'
In such cases the ' serves as a kind of "syllabifier," not unlike its hiatus-
breaking role above (and similar to the dieresis in French and older English
orthography), indicating that the vowel letters belong to different syllables
rather than a diphthong.
2.3. Reliance on the dominant script: Latin

Although there was no self-conscious reconquista undertaken in
Portugal, the country was fully under Christian rule by the end of the
thirteenth century (when its borders also essentially took their current shape).
99
Thus unlike Jews in some regions of what would become Spain, the Jews in
Portugal lived amidst a firmly Latin culture. But the Roman script was not
merely the "dominant" script of the literary milieu; it was a form of writing
that Jewish Portuguese writers were at the very least acquainted with, and at
best willing and able to exploit in adapting Hebrew script to write Portuguese.
Beyond the categorical adoption of vowel letters (cf. § 2.1.1), the clearest way
in which their adaptation was informed by Roman-letter writing is the use of
Hebrew letters to preserve distinctions, usually etymological but often
phonological, in Romance vocabulary items that were not necessarily

maintained in speech nor, curiously enough, in the contemporary Roman-
letter spelling of Portuguese.
2.3.1. /v/
As discussed in § 2.1.2, when the phonetic realizations of two letters
have merged in the local pronunciation of Hebrew, normally only one of these
is used in the Hebraicized spelling of native vocabulary. However, in Judeo-
Portuguese (and to some degree in other Judeo-Romance as well), such pairs
may be deployed to spell similar sounds that have distinct etymologies and, in
some cases, distinct spellings in non-vernacular writing.

The spelling of Portuguese /v/ is quite variable across the corpus,
being represented by w, double-ww, as well as by plain and augmented b (i.e. &b
and #b). This Portuguese phoneme is the product of the merger of Latin /w/
and /b/,18 attested from an early date by orthographic confusion of
orthographic <V> and . This confusion is also attested in the excerpt of the
18
Some instances of Latin /p/ also yield ModPg. /v/ e.g. povo < POPULU; a form based on this
word appears in O libro de ma‹gika as w''bwp pobao < *POPOLANU.
100
Brotherton Passover text presented in § 1, where vaso 'cup' occurs first spelled
wS'b baso and later in the text as wS'ww vaso. In fact, at one point on folio 5v. the
writer appears to have begun the word with 'b ba-, but stopped to begin anew
with 'ww va-, leaving his hesitation unemended:
(15) SwS'ww 'b Sw' os ba vasos 'the ku- cups'
Thus b , the historical Hebrew b, and w, itself a historical Semitic w, seem to
behave as analogues to Roman and <v> respectively. The spelling of /v/
elsewhere in the corpus also suggests that the writers were sensitive to its
etymology, and perhaps to the orthography of the Latin etyma. Unlike
contemporary Roman-letter Portuguese writers, they frequently spelled it
accordingly: where its source is Latin b (or p) it is spelled with b, while Pg. v <
Latin w is spelled with w (either doubled or as a singleton). The effect of this
"b = B / w = V " equivalence appears to be independent of the precise sound
ostensibly being indicated:
Table 3-3. b < Lat. /b/
r'r#b'l la‹brar < LABORARE 'work'

wl#b'yd dia‹blo < DIABOLU 'devil'
wdybybny' enbebido < IN-BIBITU 'drunk' (past part.)
rybwXSy' estober19 'be' (pres. subj.)
w''rybyb beberao < BIBERE HABENT '(will) drink'
w''&byrqSy' eskri‹bao < *SCRIBANU 'writer'
19
The stem of this third-person singular future subjunctive was formed analogically from the
preterite of haber, and as such does not in fact reflect any etymological b in the verb STARE
(Penny 1991: 185).
101
Table 3-4. w < Lat. /w/
'yylymryw vermelya < VERMICULA 'red'

wrdyw vidro < VITRU 'glass'
wyynyw vinyo < VINU 'wine'
y„rÕg¬ny«w vinagre < VINU ACER 'vinegar'
ÙSAww vaso < *VASU < VAS 'cup'
S'Srywyd deversas < DIVERSAS 'various'
This division of orthographic labour is not, however, perfectly

consistent. For instance, it is curiously difficult to find a medial Pg. v < Lat. w
spelled with w, as shown in the following words where it is spelled with b:
Table 3-5. b < Lat. /w/
w#bw' o‹bo < OVU 'egg'

w#byw vi‹bo < VIVU 'live' (adj.)
w&by&b yXny&gr' ‹bi‹bo < VIVU 'quicksilver'
'#bwn no‹ba < NOVA 'new'
S'Xr'dy#b'q ca‹bidartas < *CAVITARE… 'be wary'
w''rb'l labrao < LAVARE... 'work'
There are also several cases in which w is used to spell a /v/ that derives from
an etymological or borrowed b:
102
Table 3-6. w < Lat./Ar. /b/
'wwl' alva < ALBA 'white'

§A'‡ry∆wyeJb beveran < BIBERE HABENT 'drink'
20
'dywyd' adivida < *(AD-)DEBITARE 'preside'
rywyrqSy' escrever < SCRIBERE 'write'
ydl'yy'wl' alvaialde < Ar. al-bay∂ 'white lead'
These exceptions, particularly those in table 3.5, could be explained as an
orthographic strategy for avoiding an internal w that stands for /v/ near a non-
low back vowel (i.e. /o/ or /u/), since the same letter is used to spell those
vowels. In fact, a form like w#bw' o‹bo may be seen as using a strategy to avoid
spelling the word with three identical letters in succession, i.e. www'*.21
Double-ww, for its part, is used almost invariably as a digraph for /v/,
and in the texts of chapters 4-6 it is never used to indicate a VC sequence [uv]
or [ov]. There are, however, rare occurrences in those texts in which it does
represent the CV sequence [vo] or [vu] (where Pg. /v/ may derive from Latin
/b/ or /w/), as in the following words:
(16) wwl' alvo < ALBU 'white'

wwrys çervo < CERVU 'stag'
22
yd'Xnww vontade < VOLUNTATE 'volition'
20
This word, though it is the most recurrent verb in O libro de ma‹gika, is most often spelled
with b (see chapter 5 § 2.1 for a fuller discussion).
21
Though C. Rosen (p.c.) informs me of Romanian forms that do end in <–iii>, triple-letter
spellings are rare and avoided in both Roman- and Hebrew-letter orthographies (and
probably in other writing systems). As noted in the previous chapter, near-instances in
Yiddish require either niqqud or an intervening ', e.g. ˚ww/w'ww vu 'where'.
22
The first two words (from As kores) each occur only once, yet this spelling alternates in O
libro de ma‹gika with yd'Xnw'w&b ‹bountade, which features both an initial b b and a more expanded
spelling of the hiatus left by deleted /l/.
103
Similarly, though 'w may stand ambiguously for either a diphthong [oa]/[ua]
or the CV sequence [va], the reverse digraph w' nearly always represents a
fully-vocalic [u] or [o] (with diacritic '; see § 2.2.2 above), or else word-final
[ãw]. In the twenty folios of As kores, for example, there is only a single form
in which the digraph does in fact represent a VC sequence [av]:
(17) 'qyw'r' aravika < ARABICA 'arabic (gum)'
The following example from the same text is even more striking, since it
combines these two breaches of convention23 – the initial w' is not strictly
vocalic and double-ww is not uniquely consonantal:
(18) r'dnww' avondar < AD-FUNDARE 'dip'
Based on the exceptional character of these double-ww forms, we can maintain

that the unmarked reading of a single non-initial w is vocalic, while in initial
position it is consonantal unless preceded by ' . In this orthography, then,

non-initial /v/ is almost invariably spelled with b (with or without a diacritic),
while a distinct spelling for /v/ < Lat. /w/ is indeed "preserved" in initial
position.
23
Although the form in (18) represents the only occurrence of this phenomenon in the texts
presented in the following chapters, it occurs on at least one occasion in the smaller Bodleian
astrological text, O libro enos ‹guizos das estrelas, in the verb §'rysyrww' avoreceran 'will abandon'
< ABHORRESCERE.
104
2.3.2. Sibilants
Overall, Judeo-Portuguese orthography favours S ¸s as the "default"
sibilant letter, i.e. for Portuguese /s/ that derives directly from Latin /s/,
while using s s for sibilants that, though they occur as [s] in the modern
language, derive from another source (and were probably pronounced [ts] in
the earliest attested Portuguese). This is illustrated in the orthographic near-
minimal pair in (19a) as well as the words in (19b), where s spells the sibilant
produced by the palatalization of Latin /k/:
(19) a. §yÂryi'AS sairen < SALIRE 'leave' (fut. subj.)

§yÂrA'yes çearen < CENARE 'dine' (fut. subj.)
b. hmysny' ençima < IN CYMA 'above'

's'#p yS se faça< SE FACIAT 'make' (subj.)
ryesyeXÕnÙqA' aconteçer < AD+CONTIGESCERE 'happen'
s also serves to spell the Portuguese sibilant that occurs in a nativized
loanword, even when its source is not strictly a sibilant-type sound:
(20) 's'pl' alfaça < Ar. al-˛ass 'lettuce'

Swsn'ww ry' er vanços < Gk. erebinqoV 'chick peas'
syrdy#S' a¸sedreç < Ar. ¸stran¸g 'chess (pieces)'
Thus unlike S , which appears to straightforwardly represent the unvoiced

apicoveolar segment described in 1536 by Fernão de Oliveira in the first
Portuguese grammar, s probably did not have as unambiguous a reading. In
the corpus s does on occasion infect the spelling of one extremely frequent
sibilant derived from plain Latin /s/, namely the plural marker, on nouns that
already contain this letter, e.g. s'snw' onçaç 'ounces', s'syb'q kabeçaç 'heads'
105
(the plural morpheme is otherwise consistently spelled with S on both nouns

and verbs). In addition, some words spelled in the texts with s do have <s> in
modern Portuguese orthography, e.g. wmws < Ar. züm, ModPg. sumo 'juice'.
Nevertheless, the use of s corresponds quite robustly to the distribution of <c>
(<ç> before a non-front vowel) in the Roman-letter orthography of Portuguese,
while S represents only those sibilants that were spelled by a single <S> in
Latin orthography.
The spelling of sibilants in the Arabicized Portuguese also generally
conforms to the Judeo-Portuguese practice. In addition to the small corpus of

early-sixteenth century letters from North Africa (see Lopes 1940), Galmes de
Fuentes (1962: 103) mentions a twelfth-century Arab geographer whose
transliterations of Portuguese toponyms make systematic use of g ¸sın to
reproduce Romance s, while reserving f sın and ’ ß‹ad to reproduce ç .
Whether "Romance s" means [s] the sound or <s> the letter – in other words,
whether this geographer was transcribing or transliterating the Portuguese
names – is not made clear.24 The Judeo-Portuguese usage does seem to
distance s from its unmarked Hebrew value (both historical and modern) as
the letter indicating a plain /s/. In fact, much of early Judeo-Romance writing
favours S as the spelling for the reflex of simple Latin /s/,25 although the non-
24
Galmes de Fuentes does point out more explicitly that a single medial <s> is often rendered
by Ã ¸ g ım in the sixteenth-century corpus. His transcription of this letter with <z>
(superscribed by a diacritic) recalls the only parallel case I have found in the Judeo-Portuguese
corpus, where some instances of the verb kerer 'want' in As kores occur with g g as the stem-
final consonant in subjunctive forms, i.e. Syry#gyq ki¸geres < QUAESIVERIS (see chapter 4 § 2.1)
25
Although the spelling of sibilants in early Judeo-Spanish also conforms to the Judeo-
Portuguese pattern, in later Judeo-Spanish writing s was generalized as the default spelling
for /s/ regardless of source.
106
Iberian languages tend to prefer c ß for other sibilants.26 The fact that Judeo-
Portuguese avoids this letter in native vocabulary (with sporadic exceptions in
As kores) suggests that the deaffrication of Portuguese sibilants, which Galmes
de Fuentes (1962: 103-113) considers to have begun as early as the thirteenth
century, was well underway.
Given the other sibilant-related changes occurring in fifteenth- and
sixteenth-century Portuguese, it is difficult – and indeed perhaps misleading –
to try to determine the precise phonetic character of the sounds "intended" by
a particular writer's use of s or S. The permutations of <s> and <c>, despite

their frequent confusion, probably accomplish the task of spelling the four
medieval Portuguese sibilants in a phonetically more transparent way. What
the distribution of the two Hebrew letters more accurately captures is
etymology, as if in imitation of Latin rather than vernacular Roman-letter
spelling, and in the manner of many a modern standardized (and hence
conservative) orthography. Whether this is the premise upon which Judeo-
Portuguese writers made their orthographic choices – that is, whether they
consciously drew on Latin qua Latin writing – as opposed to simply capturing
two classes of pronunciation within the confines of their writing system
(which usually coincided with a prior convention) may not be answerable.27
26
Steiner (1982: 37) maintains that "if anything is known about the Hebrew sibilants in
Christian Spain and Portugal, it is that c and s were not distinguished."
27
In reference to the opposite process, i.e. the rendering of Hebrew c, s, and S in Roman-letter
Old Spanish, Steiner (1982: 39) claims that their distribution reflects "identities rather than
mere approximation" – that is, transliteration rather than transcription.
107
2.3.3. Classicizing spelling

O libro de ma‹gika contains occurrences of some proper names and
astrological terms that alternate between what appears to be an innovated
vernacular form and a conservative etymological spelling:
(21) a. sylyXwXsyr' ariçtoteleç 'Aristotle'

lyXwsyry' ereçotel
b. w'yr'q' akario 'Aquarius'
wryy'q' akayro28
In a very few instances, this phenomenon appears to occur with words outside
the obvious sphere of classical influence. The first form in (20) might be better
considered a "pseudo-classicizing" form, since its cluster consonants reflect
only the unassimilated voicelessness of the etymon's segments:
(22) yd'Xsym' amiçtade ModPg. amizade < AMICITATE 'friendship'

yd'dzymyny' enemizdade ModPg. inimizade < INIMICITATE 'enmity'
In either case, these alternations attest to another level on which the Judeo-
Portuguese writer exploits his biliteracy. It is certainly possible that the
alternation also reflects a variation in the speech or perception of the scribe.
However, rather than a variation truly based in vernacular phonology, this
phenomenon is probably more akin to the variant pronunciations that a (more
or less) bilingual in a minority group would have for local terms (especially
geographical names) in the majority language.29
28
This particular nativization is addressed in chapter 7 § 2.3).
29
A Montreal anglophone, for instance, may refer to the vibrant Rue Saint-Denis in a
"classicizing" fashion as [sæ~nd¥ní] or as a "nativized" [seynd´' n i] (though, oddly enough,
108
2.4. Reliance on another dominant script: Arabic

As Jewish writers in a Latin-literate culture it was inevitable that the
biliterate writers of Hebraicized Portuguese would draw on some Roman-
letter conventions in their writing system. Yet they also borrowed
conventions from Semitic writing beyond Hebrew that enable both
etymological and quasi-etymological spellings not captured (or even
capturable) by the conventional Roman-letter orthographies.
2.4.1. Cognate letters

The Hebrew alphabet allows the Judeo-Portuguese writers to maintain
etymological distinctions between Arabic phonemes that have merged in their
borrowed Portuguese form, e.g. Ar. z/ß ‡ Pg. z:
(23) ¶yr'#pwml' almofariß ModPg. almofariz < al-mi˛raß 'mortar'

ygwz' azoge ModPg. azogue < az-z√uq 'mercury'
Using different letters could, as always, suggest nothing more than the mere
fact of distinct pronunciations intended or perceived by the Jewish writer, who
may be more apt to do so with these Semitic loanwords than a non-Jew. What
should be noted above all, however, is that the transfer of spelling convention
is made especially feasible and perhaps even expected because the Hebrew
letters z and c are in a real and practical sense cognate with and historically
related to the Arabic letters “ z‹ay and ’ ß‹ad.
never […d´nπs], as though this fully-nativized "spelling pronunciation" would obscure the
word's identity in this case as the name of the street). Note, of course, that this alternation is
never reflected in the spelling, which simply follows the dominant orthography.
109
2.4.2. Quasi-etymological spelling

As noted elsewhere, the occurrence of vowel letters in Hebraicized
Portuguese writing is not entirely consistent. The most salient variation
occurs in Arabic loanwords, which are especially prevalent in the technical
lingo of As kores. There they show an interesting clash of conventions,
especially with respect to /a/. Although the orthography of the text is
overwhelmingly alphabetic, variants such as the following pairs occur in close
proximity to one another in the text:
(24) §ymrq §yrm'q k(a)rmin 'carmine'

w''rps' w''rp's' aç(a)frao 'saffron'
ydlyywwl' ydl'yy'wwl' alv(a)yalde 'white lead'
Unpointed and vowel-less spellings in such loanwords could be construed as a

form of "un-nativized" or quasi-etymological spelling if the short vowels in the
Arabic etyma were not represented in their Arabic-script orthography either.
The practice of reproducing the Arabic vowel-spelling is well-attested in the
"Arabicized orthography" of Judeo-Arabic (Hary 1996). In that tradition, the
letters ', w and y often serve exactly the same orthographic function as do their
graphemic cognates in classical Arabic spelling, namely to represent the long
vowels //, /ü/, and /ı/. For example, in the fifteenth-century Arabic-
Spanish glossary discussed in chapter 2, these letters appear as vowels only
when they are long (or as part of the definite article) and would thus be
spelled in Arabic script. The pattern is illustrated below with w /ü/:
110
(25) rpX˘l' al-÷ufr 'fingernail'

lplpl' al-fulful 'pepper'
lwpl' al-fül '(fava) bean'
¶wcpl' al-fußüß 'stone'
¶mwxl' al-˛ümmuß 'chick pea'
Judeo-Arabic forms in this tradition are very much orthographic calques,

produced by substituting each Arabic letter with a phonetically-similar or
historically-related Hebrew one, often augmented by diacritics similar to those
used on the Arabic letters, e.g. ÷ (IPA [∂¿]) as X˘, which like the cognate letter in
Arabic script ‫ ط‬represents /†/ when unadorned by a diacritic.
The collision of Romance-language context and adapted Hebrew
tradition is most strikingly illustrated on folio 240v. of the Bodleian Passover
text, in the following variants of another Portuguese loanword from Arabic:
(26) 'As'&Ap¸la' alfaça 'lettuce'

a a
hAs&Ap¸la' alf ç h
The Hebrew cognate of the Arabic source al-˛ass is hJAsax ˛as: (the dagesh
indicates the historical gemination of the middle radical /s/ in this form of the
root). In both Portuguese variants, the phonological adaptation of Semitic ˛ to
f is spelled as such – even in the second instance, where the word lacks any
vowel letters (apart from the initial article, in imitation of the spelling of the
Arabic definite article), as if based on a typical (though etymologically
inaccurate) triliteral Semitic root .h.s.&p. f.s.h. or even .h.s.p p.s.h.
Although alternants such as these are relatively rare in the parts of the
corpus I have examined, the vowel-less forms may still be considered a visual
sign of etymological or "learned" spelling. They are in practice akin to the use
111
of <ph> or <ae> in English, which do not represent any sort of un-nativized

pronunciation but are simply a vestige of the word's Greek source. Although
the blend of components in the English lexicon might discourage nativization
in the orthography30 and also desensitize readers to the variation and
competition among these conventions, this is not the case in many other
standard orthographies.31 Because of its mixture of components (Hebrew and
Arabic in particular), however, the Hebraicized Portuguese orthography
tolerates un-nativized spellings such as those above, which arise from the
contrasting conventions of alphabetic writing and Semitic-language

borrowing.
2.5. Imported/innovated characters

The only symbols from outside the canonical Hebrew tradition used by
Judeo-Portuguese to further refine their orthography are the apostrophe32 and
the occasional hacek. Similarly, there has been no innovation in the basic
inventory of letter graphs, nothing akin to the "Roman" letters <j> or ,
which evolved from allographs of and <v> to independent letters in most
Roman-letter orthographies. An exception to this rule in Hebraicized writing
more generally may exist in modern Yiddish, where the Yiddish Scientific
30
This may be particularly true for words of Greek or Latin origin, where knowledge of the
correct – that is, unadapted – spelling is often given (unduly) strong weight as a marker of
erudition and educatedness.
31
To wit: at the National Library of Wales in Aberystwyth, English Aeneas appears next to
Welsh Eneas on the display case of a medieval manuscript of Virgil's poem.
32
The apostrophe continues to serve in Modern Hebrew writing to indicate phonemes not
found in the native inventory, e.g. hyn#cy#c <ß’yß’nyh> Chechnya, §wXgnySww #grw#g <g’wrg’> George
Washington. It is possible that the use of these diacritics in Hebrew writing itself was modeled
on the practice in Hebraicized vernacular writing, though the other modern tradition, Yiddish,
avoids such augmentations in favour of multigraphs, e.g. 'ynSXvSX <†¸sfi†¸sny√> Chechnya,
§A'XgnyS'ww SzdrA'Szd <dz¸s√rdz¸s> George Washington.
112
Institute (YIVO) recommends that tsvey-vovn (double-ww) be joined at the base,

forming what looks like a Roman <V> (Fishman 1977). In my own experience
I have encountered this phenomenon only in the handwritten "blackboard"
Yiddish of a few language teachers.
3. BETWEEN TRANSLITERATION AND TRANSCRIPTION

Having described the major orthographic features of Judeo-Portuguese
writing, I propose to show how these characteristics have informed my own
Romanization scheme.33 While avoiding outright normalization (see chapter 8

§ 2.4), my overarching goal has been to employ a system that preserves the
distribution of graphemes in the original texts (thus making it possible to
reconstruct the original Hebrew-letter spelling) but that clashes as little as
possible with the expectations of a modern Roman-literate (and largely
English-speaking) audience. Thus I have largely avoided the letter-borne
diacritics and graphotactically-unfamiliar strings that may be found in some
Romanizations of Semitic-script texts (cf. chapter 8 § 2.3.1), since the clutter
they impose on the orthographic field often outweighs the importance of the
information they provide in the Portuguese context. Individual strategies are
discussed below – not strictly from the point of view of each Hebrew letter,
however, as is normally the case, but from the perspective of the writing
system more holistically.
33
Although the following only applies in a strict sense to the Romanizations in this chapter
and in the critical editions of the succeeding chapters, the transliterations of non-Portuguese
Hebraicized material in the previous chapter largely conform to this system as well. I have on
occasion followed a mixed set of conventions; while these are too multifarious and tangential
to enumerate, they nonetheless serve the same goal described here, namely to provide a
maximally-informative but minimally-disruptive text to an anglo-literate audience.
113
3.1. Vowels
Wherever the Portuguese Jewish writer has made use of a mater lectionis
to serve as a vowel-letter, I have reproduced it in the transliterated form,
including "silent" final h as <-h>. When two ' occur in succession (e.g. in
hiatus from a deleted consonant), I normally transliterate both unless the
second serves as the diacritic for a following vocalic w or y (cf. § 2.2.2). The
Romanization of w and y themselves usually involves a choice between
<o>/ and <e>/ respectively, which I have based on a combination of
etymological and phonological considerations.

When a vowel is not explicitly spelled, I have transliterated it as a
superscript, even if it is indicated by niqqud. I base this decision on the fact
that Hebraicized Portuguese writing is emphatically alphabetic – that is, vowel
letters are the norm and the niqqud that is used rarely if ever disambiguates
forms that would otherwise be homographic. Thus all deviations from this
norm are indicated by the most suitable analogy in transliteration, i.e.
superscribed Roman vowel letters.
3.2. Semivowels
A single w is rendered as <v> where it has a consonantal value, and as

<o> or (depending on etymological and phonological considerations)
where it serves to represent a vowel (double-ww, which is almost exclusively
consonantal, is transliterated as an italicized <v>34). The same applies to y,
which is rendered as <y> when it serves as a consonant, and as <e> or 
(again based on etymology) when it represents a vowel; double-yy is rendered
34
The Roman letter that could be seen as most faithfully rendering the graphemic form of
double-ww, namely <w>, produces the wrong effect for anglophone readers.
114
as <y> when it follows n or l to indicate palatalization (or else indicating the

semivowel), but as <ei> when it indicates a vocalic diphthong.
3.3. tpk-dgb begad-kefat (Stops/Spirants)

p is rendered as or <f> depending on the presence or absence of a
diacritic to indicate the stop (unaugmented or with dagesh) or spirant (rafeh,

hacek, or apostrophe) value, though no diacritic is added to either of the
Roman letters. By contrast, g and b are rendered as <g>/<‹g>/<¸g> and
/<‹b> respectively depending on the use of a diacritic on the Hebrew

letter.35 However, all three letters (b / g / p ) are rendered in italics, i.e.
/<g>/<f>, when a spirant is expected (usually based on considerations of
Portuguese phonology) but no diacritic is present in the Hebrew-letter
original.
3.4. Sibilants
As noted in § 2.3.2, S ¸s is the default sibilant letter in Judeo-Portuguese
writing. For this reason, despite its historical and modern Hebrew value as
/¸s/, as well as the widespread occurrence of this sound in (modern)
Portuguese, this letter is rendered simply as <s> in my Romanization (except

in Hebrew words themselves, or in the few instances in As kores where it is
augmented by an apostrophe). Similarly, since s s represents sibilants that
almost exclusively derive from sources other than simple Latin /s/, it is
transcribed as <ç> here, giving it approximately the same distribution as <ç>
(and <c> before <e> and ) in modern Portuguese orthography. Thus I do
35
For typographic reasons I avoid the apostrophe in transliteration, using a hacek instead.
Only a macron, however, is used with , also for typographic reasons.
115
not exploit the convention of "soft-c" in Roman-letter Portuguese (where the

cedilla is not required before <e> and ) and avoid the unadorned <c>
altogether in my Romanization.
3.5. Velar stop

q is always rendered as <k>, despite the fact that this convention
follows neither the Semitic philological tradition (where it is transliterated it as

<q>) nor traditional Portuguese orthography, where [k] is written as either
<c> or the digraph <qu>, and the letter <k> is generally avoided.36 Using this
character is the most efficient way to indicate the appropriate phoneme, while
preserving the single-grapheme choice of the Judeo-Portuguese writer.
3.6. Final forms

Because they are used only and always in final position in this corpus,
no distinction is made in Roman script to indicate the final forms. This is
normal practice in most if not all Romanizations of most if not all Hebraicized
texts.
36
Although not a factor per se in my rationale, it is striking that most systems of modern
(Romanized) Judeo-Spanish use <k> where modern Spanish orthography has <c> or <qu>,
probably for the very reason that it may be the only feature to distinguish some forms in
written Judeo-Spanish from those written as standard Castilian.
116
CHAPTER FOUR
Syrwq S' §yz'#p yS wmwq yd wrbyl w'
O LIBRO DE KOMO SE FAZEN AS KORES
(PARMA MS. 1959)
1. INTRODUCTION
The text known as O libro de komo se fazen as kores 'The Book on How to
Make Colors' is the best-known Hebraicized Portuguese manuscript, the only

one to be specifically mentioned in more general studies of pre-expulsion
Portuguese Jewry (Ferro Tavares 1992: 146). The earliest printed reference to it
appears in Zunz (1876), though the only substantial study of the text was not
undertaken until well into the twentieth century. Based on a photograph of
the manuscript at the Jewish Theological Seminary in New York, the text was
transcribed and translated by Blondheim (1929), with a Romanization
appearing in Blondheim (1930). However, this edition contained only a few
notes of commentary, along with a number of hesitations with respect to
individual transcriptions, transliterations, and translations.
Based on my first-hand consultation of the manuscript at the Biblioteca

Palatina in Parma (Italy), the present edition addresses these errors, and
expands on excerpts first published in Strolovitch (2000c) to include a
thorough philological commentary on the entire text. The Romanization
presented in this chapter also corrects several of Blondheim's errors, most of
which consist of unwarranted Castilianizations1 of several forms, as well as a
1
Although the spelling of some forms is, in principle, ambiguous with regard to a given
Spanish-versus-Portuguese feature, Blondheim does not actually justify these Romanizations
(see chapter 7 § 4.3 for clearer examples of Castilianisms in Judeo-Portuguese).
117
small number of characters that he plainly misread. It also departs

significantly in form from his, since I have used the system of transcription
described in the previous chapter, which more faithfully preserves the
distribution of Hebrew graphemes and disregards some conventions of
Roman-letter Portuguese writing. I do, however, owe a great deal to
Blondheim for his translation, which provides the English meaning of many
rather obscure elements and compounds.
As kores contains 45 chapters, varying in length from single sentences to
several folios, each of which provides instructions for the preparation of inks
and dyes, as well as practical information on how to best apply them in
manuscript illumination. The text is bound together with ten other
manuscripts which, based on similarities of format, justification, number of
lines, and other features – and in spite of their varied subject matter – were
probably designed as a unified volume (Metzger 1977). In her discussion of
Hebrew manuscript production in late fifteenth-century Portugal, Metzger
devotes a lengthy introductory footnote to the debate over the date and
authorship of As kores. Although a colophon reveals the name of Abraham
ben Judah ibn Hayyim writing at Loulé in Portugal, the year of composition or
copy is given only as "22." Blondheim (1929) takes this to be the year 5022 in
the Hebrew calendar, that is 1262 in the Gregorian, a date that would place the
text among the earliest examples of Judeo-Romance (beyond individual
glosses). Yet the writing style and language clearly places the extant copy
later than the thirteenth century. Metzger does contend, however, that despite
its distinct Portuguese character, it was probably not a manual used by the
scribes and illuminators of what she considers a "Lisbon school" of
practitioners.
118
2. OVERALL LINGUISTIC CHARACTER

Like the larger astrological text O libro de ma‹gika in chapter 5 (and most
non-liturgical Judeo-Romance writing), diacritic vocalization is used in only a
handful of words. The use of the four matres lectionis ' a, h a#, w o/u, and y e/i is
generally more robust than in O libro de ma‹gika (chapter 5), and with rare
exception only /a/ is left implicit and unspelled, most often in Arabic
loanwords (cf. chapter 3 § 2.4.2). The double-duty of b b , g g, and p p,
representing both plosives as well as the corresponding spirants v, ¸g, and f, is

indicated less than consistently by a diacritic apostrophe (as opposed to rafeh).
Nevertheless, As kores does offer the most wide-ranging adaptation of Hebrew
script for writing Portuguese, in terms of both orthographic variation and
overall linguistic structure. While most of these features are discussed in
specific detail in the commentaries on individual items in § 5, one aspect of its
adaptation of Hebrew script is worth highlighting in advance.
2.1. Sibilants
As noted in the previous chapter (§ 2.3.2), the spelling of sibilants in
Judeo-Portuguese, as in the Roman-letter orthography of the time, is quite

variable. Further adding to this variation, As kores is the only text in the Judeo-
Portuguese corpus to make any significant use c ß for vernacular2 words. It
occurs most often in final position, where the rest of the corpus would have z z
(usually corresponding to <z> in the modern orthography):
2
Along with a number of items in this text, c is used in the spelling of several Arabic and
Greco-Latin loanwords in O libro de ma‹gika. Given their technical nature, however, it not clear
to what extent they would have been part of the colloquial language, so that the question of a
distinct pronunciation as indicated by this letter may not be pertinent.
119
(1) ¶yd deß dez 'ten'

¶y#g ‹giß giz 'chalk'
¶ynryw verniß verniz 'varnish'
¶yrp preß3 (Fr. après) 'then'
As kores is also the only text in the corpus in which the noun-marking plural -s
is spelled by a letter other than S , though in most cases it appears to be
influenced by another non-S sibilant letter in the stem of the word:
(2) ¶yzyw vezeß vezes 'times'

s'snw' onçaç onças 'ounces'
s'syb'q kabeçaç cabeças 'heads'
This variation is further complicated in future subjunctive forms of the verb

querer 'want', which occurs frequently in the text as an opening formula for a
given set of instructions, usually addressed to the reader in second-person
singular form (ModPg. si quiseres 'if you want (to)' < perf. subj. QUAES‹IVERIS).
Not only does the normally stable verb desinence vary between S and ¶ , but
the stem-final consonant varies even more frequently between #S or #g (both of
which may or may not include the apostrophe). Thus all three of the
following alternate throughout the text:
(3) Syry#Syq ki‹seres

Syry#gyq ki‹geres
¶yry#Syq ki‹sereß4
3
This word does not survive in Modern Portuguese (see § 6, note line 155).
4
The only other verb to feature this unconventional final consonant is one future subjunctive
occurrence of ¶yrybwX tubereß 'you (will) have' (ModPg. tiveres < TENUERIS) on folio 13r. Since it
is followed there by ¶yd deß 'ten', which more justifiably contains the same final letter, the verb
spelling may be an anticipatory gesture.
120
3. ROMANIZATION
Although the scribe has not made use of any word-internal punctuation
to indicate morpheme boundaries, I have used a hyphen to indicate a clitic
pronoun that is not graphically separated from the verb stem or desinence. In
contrast, I have used a period to indicate an otherwise free-standing
morpheme (most often the definite article) that the scribe has not separated
from an adjacent word. The intralinear numbers indicate the line breaks in the
manuscript itself, though line numbers in the commentary of § 5 refer to the

numbering in the left margin.
aki se.komeinça |2 o libro de komo se fazen as kores das tintas |3

todas pera aluminar os livros e digamos |4 logo primera mente do
oro sol \ si |5 kiseires fazer o |6 oro kon.ke posas aluminar |7 o
pintar o kabidoar o eskreber |8 asi komo te.este libro di‹ser e mandar
5 |9 adiante e tu non mengues nin akreçentes |10 mas do.ke tu libro
di‹ser ka si o fizeires todo |11 eraras, e nao te‹bera prol enkuanto |12
fizeires e por esta razon nao faças |13 per otra gisah sinon komo ti o
libro |14 di‹ser e mandar fazer. primera |15 mente tomaras deß onças
|16 de ‹gupiter ides estanyo kraro e [1v.] linpio, e tomaras çinko onçaç
10 de fu‹gitibo |2 ides azoge e dereteras primero o ‹gubiter |3 e deitah o
fu‹gitibo en un almofariß. e o |4 estanyo deretido deita o no almofariß
ko o |5 azoge. e mistura-lye çinko onçaç de ensofre. |6 e dos onçaç de
sal armeniko, ides ano‹star |7 e esto se‹gah mui bein modo e peneirado
e |8 oke non si poder peneirar torna-o a moer |9 ata ke se‹ga todo
15 peneirado e enton a‹gunta |10 todo en un baçio bein linpo e despois
|11 deita-o en uah aredomah de vidro. e en‹bestiah |12 de.bitume e de
bou baro forte kuatro |13 o çinko vezes porke posah soportar o |14
fogo e enton mete-ah en uah panelah. |15 ‹geah de çinizah do lar e da-
lye fogo manso |16 ata ke ve‹gas ke saiah o fogo vermelyo e entou |17
20 atapah ah boka da.redomah kon do baro e le‹sa ah |18 estar sobre o
fogo apanelah kon ah aredomah [2r.] en estos tripes. e as vegadas
destapah |2 a.boka daredomah e kuando vides ke nao |3 deitah fumo
121
nenyun tira a.panela ko.ah |4arodomah de sobre a.fogo e lei‹sa estar

|5 ah esperar ata otro dia. e despois |6 kebrantaras ah rodoma e
25 a‹garas un |7 pao de oro fino e moy-o pekeno e pekeno |8 kuando
kiseires labrar ko.ele e destenpera-o |9 ko aguah gomada e faze
ko.ele tu ah |10 prol. e este oro dizen o oro musiko.
|11 kapitulo segundo do.ouro pera |12 fazer o oro de |13 musiko
kon.ke esker‹bas. toma ‹gubiter |14 e fu‹gatibo e sal armeniko o en‹sofre
30 |15 tanto de uo komo de otro e.fonde-lo-as |16 nuah kulyar de fero
grande o ‹gubiter e |17 deitaras ençima o fu‹gatibo e meçe-lo-as [2v.]
todo muy bein kon pao e deita-lo-as kento no |2 almofaris sobre os
pou do sal armeniko e do |3 ensofre e moe-lo-as ata tanto ke sega |4
todo en pou sutil e tomaras todo esto |5 esto e deita-lo-as en uah
35 kulyar de fero |6 atal enke todo esto solta mente posah |7 kaber mui
bein e akenta-lo-as sobre brasas |8 ata ke fume‹ge mesendo senpre
mui bein e despois |9 torna-lo-as ah moer otra veß e deita-lo-as |10 na
kulyar e poe-l-as sobre as brasas |11 ata ke fome‹ge e esto faras ata
tres |12 vezeß e desae mete-lo-as en un pano de linyo groso |13 e ata-
40 lo-as mui bein kuanto poderes e poe-lo-as |14 en kal bi‹ba komo sair
do.forno en uah |15 panelah sobre o.fogo ata ke saya o bafo e |16 poe-
lo-as ante sobre brasas porke |17 se‹ga ante kente e despois tira-o do
pano [3r.] e deita-o en un pokaro de baro ke se‹gah feito |2 e kozido e
manera de rodomah e kobre-o e |3 bara-o mui bein kon baro de sobre
45 akabeçah do |4 fokaro bein çarada e poen-o sobre uas |5 tripees. e
da-lye fogo grande des a.manyaah |6 ata meo diah e ata ses oras do
diah. |7 e si eskebralyar o baro pon-lye otro por |8 las
eskebralyaduras e esto feito saka |9 tuha obra mas ante alesah esfriar
|10 ata otro diah. e kuando kiseres eskrever |11 ko.el toma goma
50 aravika koma un ervanço |12 e o otro sega komo lentilya e deita
aguah |13 en uah vieira kontah posa avondar |14 agoma e kon ela
destenpera o oro e esker‹be |15 e fikaran as letras fartas do |16 oro.
kapitulo 3 do oro [3v.] pera poer o oro en libros o sobre folyas |2 de
‹gupiter ke e ‹gamado alfer e posah brunyar |3 toma o grude do çervo e
55 deita-o en molyo |4 aguah e lei‹sa-o estar evanto ata ke |5 se desaçe
por si e kuando for been |6 desfeito proba-o en teu dedo e si for bein
|7 likido e mole en esa aguah asi feitah |8 lançaras okre e en karoço
de pesego todo |9 modo e o karoço se‹ga primero keimado porke |10
façah o oro mas resprandeçente e despois |11 poen esta konfaçion do
60 grude e do okre e do |12 karoço keimado komo dito e eskre‹be u
kiseires |13 mas bafegah primero o lugar onde pu‹geres |14 esta
konfaçon sobre dita e poen en çima |15 o oro de eskre‹ber e despois ke
122
for en‹suto |16 brune-o ko.o dente do ‹ga‹bari muy pasa mentre|17 e asi
foras bein o oro. ite mas [4r.] pera fazer sesa para poer o oro toma o
65 |2 koyro de kongreo e mete-o akuzir ata ke |3 se desfaga per si e koa-
o en esta |4 koadura e façe sesa dokre e do karuço |5 de pesego e
poen o oro do eskeber ençimah |6 en enestah sesah do peskado \
kapitulo |7 4 do oro \ si kiseires poer |8 o oro kon dobre o
kon.eskudo o kon |9 farolino kon anel de oro toma apedra |10 kristal
70 e moya muyto e filya akra do |11 obo been pretado e goma e tenpra
koestah |12 e faze masa e poen esta masah |13 u kiseires e le‹sa sekar e
toma uah |14 dobra o.frolen o eskudo o anel komo dito |15 e e
esfriga-o muto ençimah da masah |16 e oke akelah masah fikar luzente
do |17 oro nunka se tolyara. e despois pesah [4v.] a do‹bra e.nou
75 a‹garas menos dela nada \
|2 kapitulo 5 de komo se faz nobre |3 azul. kuando kiseires |4 fazer
azul ke semelye de akre toma uah |5 panela grande e no‹ba e faze
en.ela kuatro |6 furakos e poen e akeles furakos duas |7 vergas de
fero e (ke) vinyao en manera de krus |8 e despois a‹beras liminas
80 destinaliß luna |9 ides folyas de prata estena ben delgadas |10 e untas
kou mel e ponas ençimah das |11 vergas do fero ental gisah ke nou
a‹gege |12 uah ah otra e despois deita dentro |13 na panela vinagre
been forte atanto del |14 ke ‹gege ah as vergas e nou pase ariba |15 e
depois atapa been a.panelah kon |16 brao forte e poras a.panelah en
85 |17 estirko de bestah kente ke se‹ga soterada [5r.] en ela ata a boka e
muy bein kobertah ata |2 vente e dos dias e akabo de estes 22 dias |3
destapah apanelah e a‹garas na boka azul |4 fino e arapa-o kon uah
paleitah de palo o de |5 kanah. e despois torna as laminas ah |6
a.panelah eh le‹sas (estar) os dias as komo de primeiro |7 fizeiste e per
90 estah gisah faras bou |8 azul e podes fazer poko o muito asi komo |9
uberes gisado.
kapitulo 6 do azul |10 para tenperar o azul |11 toma agoma arabika
luzente e fina en aguah |12 en uah taçah e depois toma akela taçah |13
en.ke estu‹ber e koa-las kon o pano de linyo |14 e toma a.terça parte
95 da.krara do o‹bo e |15 deita ko.elah no korno o na kon‹ga \ ite |16
de‹bes asa‹ber ke o azul deves reno‹bar en kada |17 un dia porke si
estuber per longadah mente [5v.] akela aguah no azul tornase negro
e esto |2 faras dos o tres vezes no dia e tomaras |3 da goma arabika e
da.krara do obo |4 e esçreveras ko.ela \ |5 ka‹bidar-t-as do azul
100 luzidio. e o ke es |6 asi komo kardeo e bou. dazul te dou |7 para o
konoçeres mete un pekeno dele na |8 lenguah o no palma onah onya
do |9 do dedo e si o sentires as komo a.de uso |10 mao e \
123
kapitulo 7 do azul |11 pera tenperar o |12 azul toma azul dakre e
moyo ben ko.ah |13 de koahda das vides leve mente o kolye o |14 un
105 uah veeyrah o kon‹gah e lava-o |15 kon akelah de koadah das vides e
moy-o otrah |16 veß leve mente kon uah pokah de rosa e eskreve |17 o
ke kiseires o aluminah o pintah o |18 retalyas uah pekenah de krara
dobo kon goma e nao |19 fike de un diah [6r.] pera otro kon ela ke se
tornara negro
110 |2 kapitulo 8 pera fazer rosah |3 toma uah onçah |4 de brasil fino e
rapa-o meudo e pon-o |5 adeparte e depois toma uah kuartah |6 onça
de pedra ume e toma peso de dos |7 dineros de alvai alde e moy-o
kon a.pedra ume |8 en un almofariß e pon-o aparte e toma |9 depois
o brasil e deita-o en uah |10 taçah de malegah e deitah os otros pous
115 |11 e kon o brasil e deita-lye e çima urinah |12 ata ke se reskobrao e
esten asi per 3 |13 dias aka‹bados e toda viah meçendo-os |14 kon un
pao kada diah 5 o 6 vezes. e despois |15 koa-o e apora-o por un pano
de |16 linyo e çima de uah piah feitah de ‹gis o de pedra kri. e le‹sa-o
bi‹bir na [6v.] piah e kuando for en‹suto rapa-o muy |2 bein kon uah
120 paleitah. e guarda-o bein do |3 ayre e kuando kiseires labrar ko.ele |4
moio kon aguah gomada \
kapitulo 9 |5 pera fazer otra rosah |6 toma do brasil oke u‹beres
mester e |7 ahrapa o been meudo e deita-o en unah |8 olya pekena
no‹ba e deitah na olya de koadah |9 de vides ke se‹gah o brasil dela
125 ko‹berto |10 poen-ah ao fogo e da-lye uah |11 fer vora atanto ke tome
adekuada sustançiah |12 do brasil e toma duas partes |13 de pedra
ume e mais a.meiah parte de |14 pedra kri e moye kada un muyto por
si e |15 depois mistura-o e moya de konson e faze |16 komo ‹ga sabes
de pedra ume rosah.
130 |17 kapitulo 10 para fazer mui nobre [7r.] azarkon \ toma alvayalde
kuanto ki‹seres |2 e moyo e penyera-o e lança-o en uah |3 ti‹gela o
ti‹gelas an‹gas o levas ao forno |4 do vidrio e le‹sa o e estar per vente e
|5 dois dias e aka‹bados estes dias tirao do |6 forno e a‹garas muy
fermoso azarkao. |7 desta gisah faras kunato kiseires \
135 |8 kapitulo 11 pera fazer azinyabre |9 mui fino toma folyas |10 de
kobre mui delgadas e molyas en |11 vinagre kente e mui forte e mete-
o en |12 unah olya akoçtadah e untah abokah |13 da panelah kon mel e
kobreah kon testo e |14 soterah son o estirko de bestas grandes |15 e
estee ali trinta e un dias e aka‹bados |16 os dias tiraras e olya e a‹garas
140 |17 azinyabre e rapa o kon uah paleitah e [7v.] si mais ki‹seires fazer
torna afazer |2 komo dito e a‹beres dou azinyabre.
124
|3 kapitulo 12 de otro anizyabre |4 tomaras un algidar |5 e mea-lo-as

de urinas bein pudres e tomaras |6 un baçio de laton muy ben la‹bado
do fondo |7 de fora e poe-las en çimah do algidar ke |8 non ‹gegen as
145 urinas ao fondo do baçio |9 kon dos dedos. e se‹ga o fondo do baçio
|10 untado de bou mel e o baçio se‹gah |11 meado de dentro da keles
me‹gados e ençima |12 do baçio enbroka otro algidar. e |13 ençima do
algidar poras en‹salmos |14 e vertiras o vido do baçio eno algidar |15
de fondo. e vee ao fondo do baçio e |16 a‹garas o mel ke-lye puseste
150 tornado en |17 azinabre e rapa-o ko uah paleitah. e [8r.] guarda-o en
papel e mas verde ki‹seres |2 fazer unta o fundo do baçio kon mel |3 e
faze komo da primeira fiezste e asi |4 faras kuanto kiseires. e pera ah
des tenpraçon |5 pera este verde kuando ki‹seres |6 labrar kou ele
moyo ante mui bein e |7 deitalye e poko de açafrao bein modo |8 e
155 destenpera-o kon aguah gomada ke nao |9 ah dia‹blo ke-lye tolya preß
a akolor.
|10 kapitulo 13 para fazeres nobre |11 karmen tomaras |12 uah olya
grande noba e ke kaybao |13 kuatro açumbres de aguah e engiah de
|14 me‹gados do omen e mese os por dias |15 e faze-os been kraros e
160 toda ora |16 ke lebantaren eskuma e depois ke foren |17 been kraros e
eskumados toma uah [8v.] ti‹gela grande e poen sobre ela palya de
çenteo |2 e ençima da palya un pano de linyo e en |3 çima do pano
deitaras zinça de vides as duas |4 partes e o terço de kalbi‹ba e poe de
fundo |5 uah olya e deitah ençima da barilah as |6 urinas dekoadas
165 ke koaste no asado e |7 vayo koando ata de se en‹ga apanela destah |8
de koada per gisah ke kaibao e kuatro açumbres |9 e poen-a ah fe‹ber
ke menguen ata |10 2 dedos e ao fogo otra olya ‹geah de urinas |11
kraras komo de kada e fervao anbas |12 e lançaras na panelah das
urinas klarifikadas |13 komo de kada uah libra de laka e |14 daras
170 fogo manso e toda viah meçendo-ah |15 kon un pao forkado poko e
kuando for |16 a laka deretuda koa-o en un sako de linyo |17 e poe en
fundo un baçio o ke fikar no [9r.] sako mete-o na olya da dekoada ke
guardaste |2 ao fogo fer‹bendo manso ata ke sega deretuda |3
mesendo-ah kon un fuste e depois koa-o ah |4 parte kon akel sako
175 koah frol. asi faras |5 karme de duas naturas e en pero primera |6
mente de‹bes klarifikar as urinas
|7 kapitulo 14 pera otro karmen |8 toma uah olya daguah |9 linpa e
te‹ba en gisa ke si posa en |10 elah dereter (desol‹ber) uah libra de
pedra ume e |11 toma destah aguah do aume un açumbre deitah |12 o
180 meo en uah olya e otro meo en otrah |13 olya e bol‹be-o kon o pao e
despois le‹sa-o |14 krarifikar e kuando for peisado vae |15 vertindo da
125
aguah ke sobre nadar e kuando |16 for apurado ke nao posas sakar
aguah |17 mete-os en.senyos sakos de linyo e kolga-os [9v.] ke verta
o sobre senyos testos ou |2 ti‹gelas e o.ke se-koar si for tirao e torna o
185 |3 ao sako e asi faras ata ke saya |4 kararo. e depos ke for kraro faras
pelveri |5 nyos komo ervanços e poyn-os e sekar |6 a o.sol ke se‹gah
manso e si o sol for forte |7 poen uah saban en çima e deske foren |8
sekos guarda-os e faze deles tuah prol.
|9 kapitulo 15 para fazeres vermelyon toma çinko |10 libras de
190 fu‹gati‹bo ides azoge e poyno |11 en uah aredoma o ti‹gelah grande
vidrada |12 e toma uah libra de pedra en‹sofre bein |13 meudo e
deita-lye do pou de en‹sofre poko e |14 poko sobre o ar‹gen vi‹bo ata ke
se‹ga bein enkorporado |15 e toda viah meçendo-o kon pie de |16 kao
kon suah pele e sa laah ata ke se torne |16 a.fogo komo çinza [10r.] e
195 depois ke asi for mortifikado deita-o en |2 duas olyas no‹bas ke se‹gan
feitas komo |3 aredomas an‹gas de ‹guso e estreitas ençima |4 e nao
fike por ençerar de elas sinao |5 un furako pekeno por o saia o umor
e poras |6 as olyas sobre o fogo en suas fornalyas e |7 baras bein kon
o baro e poen uah ti‹gela |8 en çima dos forados e kuando o fumo
200 veres ke sal |9 vermelyo e nao feder mete dentro no furako |10 un
espeto delgado e si alguah kosa se apegar |11!ao espeto tira as olyas
deo fogo e |12 le‹sa-o esfiar e depois ke for frio kebrantah |13 as
panelas e a‹gara o bermelyou feito. e |14 per este peso faras kuanto
vermelyon kiseires |15 fazer e auah terça do azoge poy 5 libras do |16
205 en‹sofre e ah 5 libras da zoge uah libra |17 de en‹sofre e kileires ental
gisa o fogo ke nao [10v.] se keme e da-lye fuego tenperado nin vivo
|2 nin manso. ite se per ventura sese |3 keimar o bermelao kebrantah
as olyas e |4 moy-o e enkorpora-o e mistura-o kon otro |5 peso
dazoge e de en‹sofre e poen-o en otras |6 olyas e faze komo dito e. e
210 para been mentes |7 nos fumos komo saen asi e nunkao os eraras.
|8 kapitulo 16 para por o oro en espada o kuytelo toma |9 as kabeçaç
do kar‹bao de braço e mete-as |10 na for‹ga ata ke se‹gan been
bermelyas e depois |11 sakas e ponas en uah taboah e lançaras |12
sobre elas dos punyados de sal moido e moiras |13 todos en uo e
215 depois lança-lye dos onças |14 da no‹satar meia onça da zinyabre e |15
amasa todo muy bein kon o forte vinagre. e faze |16 primeira mente
alinpar a espada o o kuytelo [11r.] e komo si eskre‹beses poen desta
çinrada |2 u ki‹geres fazer letras o figurar mais |3 primera mente
segah untada kon vegera boli |4 e kon azarkoanboli e si for tenpe de
220 in‹benro |5 este e a espadah o kuytelo 2 dias ke.nao |6 alinpes. e si for
126
verao este-e un dia |7 de pois la‹ba e guarda tuah arma ka ‹ga |8

onrada fika.
kapitulo 17 |9 para tin‹ger os osos |10 e paos de kual kolor ki‹geres
toma o |11 pao de bu‹so o otro kual |12 keir madero e koze-o en azeite
225 por tres |13 vezes e depois tira-o e deita-lye pous |14 dalke‹brite
moido e le‹sa-o estar |15 por 3 dias e depois alinpa-o e sera negro de
|16 fora e de dentro.
kapitulo 18 pera |17 ti‹ger e fazer osos [11v.] b ermelyos tomaras o
alakar kuanto u‹beres |2 meester e desfaze-o en mui forte |3 vinagre e
230 se‹gah kon kuatro dano‹satar |4 e depois koze os osos o o madeiro en
|5 estah feçao ata ke se tornen vermelyos |6!e ke se pages da.kolor ke
estah tinta pareça |7 vermelya de dentro e.de fora.
kapitulo 19 |8 para fazer tinger o bu‹so |9 o otro madeiro de kolor
preto e fermoso |10 ke semelye azevi‹ge toma as limaduras do fero e
235 do kobre e do sal komun |11 almartake e fezes de oro o kual a‹gado
|12 no azarnefe de kada uo uah onçah e lança |13 os en muy forte
vinagre por trinta dias |14 e de pois koze-o ata mengue a.metade |15
do vinagre e despois toma do aze‹ge e destenpera |16 o en aguah ke
se‹gah un poko espeso e metan [12r.] o pao e os osos kon al por
240 muytas vezes |2 sakando ao sol per kada veß e tu veras |3 ke se
tornara preto de dentro e de fora e es |4 ta tintura nunka se tirara do
pao o dos |5 osos.
kapitulo 20 para fazeres os |6 osos de o mangos kolor verde komo |7
de azinyabre \ toma o azinyabre e do ano‹satar |8 2 onçaç de kada
245 uah e moy-o kon o vinagre |9 bein for e des e deita en eles os osos o
|10 mangos de gisa ke se‹gan kobertas dakel vinagre |11 kon os pous e
leisa os ‹gazer ata ke se faça |12 verdes de tal kolor de ke te pages e
pera melyor |13 se koze-os en kadeira de kobre e si en.esta |14 tintura
meteres kual ker madeiro tornar-se-ah |15 verde e mui fermoso.
250 kapitulo 21 dos |16 osos limados pera |17 fares a‹sedreç. kuando
ki‹seres fazer os osos [12v.] kon fogo o sin fogo e fazer de muytos ou
|2 si o ki‹geres fazer toma osos de vaka ke |3 nao a‹gao karne nenyuah
o osos de marfil |4 e lima-os bein kon uah lima ata ke se‹gan |5
limados komo limaduras de fero o seraduras |6 de madeiro e deita
255 estas limaduras dos |7 osos en uah panela vidrada kon muy forte |8
vinagre e se‹ga a.panelah been ko‹berta e |9 barada kon baro daredor
e depois dal |10 fadida peso de 5 onças e.de alyos pesados |11 peso de
uah onça e ferban en kuatro |12 onças de vinagre vermelyo o en mais
e |13 kobre-os akeles osos kon este vinagre |14 bol‹be-os ben kon estas
260 konfaçioes |15 e poyn-o sobre o fogo ata ke se fundao |16 e kuando
127
veres ke se fundey e koalyen |17 komo ‹gunbo e ki‹seres ke fiken

brankos tira [ . . . ]
[13r.] [kapitulo 24 . . . ] panos o çumo de katasol e dos rgaos |2 e
depois ke tubereß deß e vente panos o un |3 ke se‹gah grande ‹geos de
265 çumo tomaras un algidar |4 o testo ‹geio de vidos d.omees e pon-lye
|5 en çima dos bastoes atravesados que nao |6 ‹gegen ah o vido e
deita os panos en çima |7 estendidos e esten ai por 9 dias o 12 |8
meçendo os vidos 4 o 5 o 6 vezes ao diah |9 e rebolvendo os panos de
uah parte aotrah |10 ata ke se‹gan korados non vidos e kuando os |11
270 vidos foren mais pudres e mais fedegosos |12 kuanto melyor kolor
farao. e depois ke |13 foren koradas do bafo dos vidos poyn-os ao |14
sol ata ke tomen kolor komo morado e os |15 panos enestas kolor e
been en‹sutos e |16 korados guarda-os been do ar do inberno |17
kuando ko.ele ki‹seres aluminar o labrar [13v.] toma un korno been
275 labado o uah vieyra |2 e talya kon osos tesoyras un pekeno do |3
pano do katasol e deita-o na kon‹ga o vieyra |4 o korno e deita-lye da
aguah gomadah e depois |5 keo pano for tomado ke se‹gah ben
enbebido |6 dela meçe-o bein e logo la‹braras kon.el |7 ka.si ai esteber
mais de un diah logo e |8 botado e perde a kolor. e asi faras en kada
280 |9 un ano kuanto ki‹geres fazer.
kapitulo |10 25 kein ki‹ser obrar |11 kon oro o kon otras kores e komo
se |12 deven afazer e ameçkrar por este in‹genyo |13 o deven afazer ke
e ‹gah pro‹bado primera mente |14 kuando o oro ki‹seres poer toma
okre e o |15 alvayalde tanto de un komo de otro e un poko |16 de ‹giß
285 e todas koisas a‹gunta-as en uo e |17 moyas todas muyto e muy bein
kon krara [14r.] dobo ke se‹gah muy fraka e muy li‹gera e si for |2
muy forte deitalye daguah e moyo bein e si for |3 muy li‹geira deitalye
da krara do obo e esto |4 proba-ah en un pergamino ena kele lugar u
|5 ki‹seres poer o oro por ante kon pinzel estah |6 kon façion e ante ke
290 se seke poen o oro en |7 çima e des e brune-o mui bein kon dente de
porko |8 o de ka‹balo \
kapitulo 26 pera tenperar |9 azul toma ah ‹gema do obo |10 en tal
maneira ke non ande ko.elah d.al‹ba‹gen nada |11 depois moye ese azul
muy been en uah atalmiah |12 e depois filya esa ‹gema do o‹bo e
295 meçkra-o |13 kon o azul. e moyo todo de konson muy been e deske
|14 for ben modo toma ese azul en uah fokara e lança-lye |15 da
aguah e mete mao e meçe-o kon teu dedo por |16 muytas vezes. e
deske veres esa aguah ençima do |17 azul pareçer e depois ke o
le‹sares de meçer enton [14v.] kata sol. e sia for muy puro e muy
300 linpo de toda |2 maldade. e deske esto for feito detenpera |3 o kon
128
aguah gomada e en ton eskre‹be kon ele |4 pero antes ke lye deites
aguah gomada se‹ga o |5 azul been en‹suto da aguah e si ki‹geres |6
podes deitar na ditah krara do brasil por-lye |7 dar melyor kolor.
kapitulo 27 si ki‹sereß |8 fazer boah rosah filya |9 do brasil kuanto
305 ki‹seres e rapa o muy ben |10 ençima de uah kon‹gah o korno e desae
|11 a‹gunta kon el da pedra ume e deske esto |12 fizeres filya da urina
do omen kasto e |13 deita tanta ençima dese brasil e da pedra |14
ume ata ke ‹se‹ga tres ko‹bertos e le‹sa-os |15 asi estar por 3 dias e
depois filya un |16 pao de ‹giß e deita do pou dele ençima dese |17
310 brasil ata ke se melye ke se‹gah tanto doun [15r.] komo dotro e desae
lei‹sa-o asi estar esa |2 konfaçion por un diah e por dois e depois |3
toma esa rosa e moya kon krara dobo gomada |4 e eskre‹be ko.elah \
si ki‹geres fazer |5 fazer koor indiah |6 mete kon el do azul e si por
ventura ki‹seres |7 fazer pinça negra mete kon el do negro e si |8 por
315 ventura akor alva ki‹seres tornar en |9 negro a‹gunta kon el do negro e
do alvo e toma |10 do brasil e mete-o en un pano alvo e koa-o |11
sobre ‹giß. sabe ke deß sao as kores prinçipais |12 azul. oripimento e
vermelyon. |13 verde. karmen. çufiy. katasol. açafrao. |14 azarkon.
alvayalde. brasil. \ kuando ki‹seres |15 kebrar la krara do obo e‹ga
320 alye |16 dela le‹ge dela figeira e kebralya as muito bein |17 para tu
obra klara komo aguah.
[15v.] kapitulo 28 si ki‹seres fazer verde |2 e destenpera-o deita do |3
vinagre kon do verde e da ‹gema do obo e moyo |4 todo de konson e
as tres partes se‹gah do |5 verde e a kuartah da ‹gema e si o melyor
325 ki‹seres |6 fazer deta-lye da aguah gomada e destepera-o |7 kon ela si
o ki‹seres tornar en otra kor |8 mete kon al do açafrao e si en otra kor
ki‹geres |9 tornar mistura kon el do brankete e pareçera |10 en sonbra
verde e alvo.
kapitulo 29 |11 si ki‹seres destenprar azul |12 deitalye daguah e moy-
330 o kon elah un poko e des |13 ke for bein seko daguah si o ki‹seres
tornar |14 en otra kolor ‹gunta kon el do bran‹gete ides |15 leo kon
krara do obo e si o ki‹seres torner |16 ençeleçtre ‹gunta koel 3 partes de
bran‹gete |17 e ah uah do azul.
kapitulo 30 si ki‹seres [16r.] bou karmen ‹filya o karmen aguah e da
335 ‹gema do obo |2 kuanto ah metade e moy todo de konson e si ki‹seres
|3 ke semelye kolor sanginya agunta kon el ah |4 terça parte do azul
\
kapitulo 31 |5 si ki‹seres destenperar |6 açafrao pera eskrever kon ele
deita-lye da krara |7 do obo e nao o moyas nin metas kon.el otra |8
340 kosa. e si ki‹seires meter en otra kor |9 semelya‹bel a‹gunta kon el do
129
oripimento been |10 modo kon a.krara si ki‹seres matizar kon el |11
do azul been moido kuanto aterça parte e nou |12 mais \
kapitulo 32 si ki‹seres destenperar |13 o oripi mento destenpera-o |14
kon aguah e kon ‹gema do o‹bo e depois tira-o |15 dinde esah aguah e
345 o‹bra kon ele ka nao ker |16 otra natura \
kapitulo 33 si ki‹seres |17 destenperar o negro anil [16v.] fila da
aguah gomada o da ‹gema de obo e moy |2 todo de konson e si kon.el
ki‹seres destenperar |3 e obrar ‹gunta kol el do branko ata ke |4
semelye kor nuben \
350 kapitulo 34 |5 filya o azul e destenpera-o kon aguah gomada e kon
‹gema de |6 obo e deita sobre el para matiza-ly-o karmen |7 o do
brasil.
kapitulo 35 si ki‹sereß |8 kolorar kon azul branko |9 matizah kon azul
puro. e si ki‹seres kolorar |10 kon karmen matizah kon karmen o kon
355 brasil o |11 kon ‹bermelyon. e si ki‹seres kolorar indio |12 alvo matizah
en al kon verde puro. e si |13 ki‹seres kolor’rar kon azarkon matiza
sobre el |14 karmen o brasil o vermelyo e si ki‹seres kolorar |15 kon
‹bermelyon matizah kon brazil o kon karmen |16 pero as kores todas se
poden matizar kon negro \
360 [17r.] kapitulo 36 filya açafrao e agoma e |2 krara de ob o
destenperadah kon |3 todas estas kousas e poen todo esto en |4 akel
lugar o letra ke ki‹seres fabrikar e depois |5 toma a folya do oro muy
sutil mente e a.sabor |6 en uah kasa sin vento e sin ‹gente per amor ke
non |7 fale a.nenyuo e un çendal o pano na boka e nos |8 narizes ke
365 non bafe‹ge ao oro ke-lye fo‹ga atade na |9 kabeça. e ponya-o ençima
das ditas kousas |10 e le‹se-o e estar per uah orah do diah e depois |11
filya un poko dal godao e pono sobre esta |12 folya pasa mente e oke
u‹ber de fikar pera |13 letra lei‹sa-o estar e o al tolya-o. e deske fizeres
|14 esto mete mao ao brunyador e brunya-o muy |15 bein kon un
370 dente de porko \
kapitulo 37 os meçkramentos |16 das koores atais |17 ker ki‹ser fazer
kor komo dazul o de karmen toma [17v.] a.metade de azul e a
metade de karmen e a terça |2 parte de branko e si ki‹seres mais
vermelyo |3 mete mais do karmen e si ki‹seres mais branko |4 mete
375 mais do branko. e kuando ki‹seres perfilar |5 o aluminar toma azul e
o karmen a metade de |6 kada uo. e meçkra todo kon da aguah
gomada |7 e kon da krara e podes perfilar e aluminar |8 e mete e
mais do destenperamento ke se‹ga been |9 kraro.
kapitulo 38 si ki‹seres por |10 o oro en libro toma d.aguah |11 das
380 kartas ko‹ga ke se‹gah de boah gisa forte |12 e pono kon pinzel uah veß
130
o duas en akel |13 lugar u ki‹seres poer o oro e depois moy |14 o ‹gis
kon aguah ko‹gah forte mente e mete i un |15 poko d.açafrarao e pon-
o en akel lugar u ki‹seres |16 poer o oro per tres vezes. e deske for |17
seko mete kuanto ki‹seres e poen o oro [18r.] kon aguah gomada
385 friah e depois brune-o forte |2 mente kon dente de porko \
kapitulo 39 si ki‹seres |3 been brunyar e oro o argen |4 aprime-o been
o brunyador e depois a‹bre ah bokah |5 e bafeg ah sobre o oro
eskentado kon mano linpa |6 e se‹gah o brunyador kente e kobra o oro
kon pano |7 de linyo velyo e depois brunya-o otra veß perçima |8 do
390 pano eskentado o brunyador otra veß sobre o |9 trapo e bafe‹gando
aberta ah bokah \
kapitulo |10 40 si ki‹seres fazer kolah toma dos pergaminos |11 e lava-
os mui been e depois mete-os en uah |12 olya no‹ba e velya e faze-os e
muito feber |13 ata ke seg ah been ko‹gos e deske for sumida ah |14
395 primeira aguah mete dentro otra aguah e deske |15 ki‹geres probar
toma dela uah pokah e pon-a |16 ena palma e a‹gunta uah mao kon
otrah e si |17 prenderen as maos ten ke e muy bein feitah ah tah kolah
[18v.] kapitulo 41 si kiseires fazer okre |2 toma do ‹bermelyo
destenperando |3 kuanto ki‹seres e mezkra-o kon ‹galde |4 ke se‹gah
400 bou e si veres ke e muito kolorado mete |5 poko de negro e sera bou
e si for muito kolorado |6 en negro mete un poko na primeira veß e
|7 de pois mete mais tanto do al\
kapitulo 42 si |8 kiseires fazer braniß filya |9 uah libra de garasah de
nobra o duas o kuanto |10 ki‹seres fazer e poen uah libra de grasah |11
405 e dos d.olio de linyaça e mete-ah kada uo |12 dele en.suah olya e as
olyas segan no‹bas |13 e koze kada un de granvagar e guarden-se ke
|14 non kaiah en elyas aguah nin otra koisah |15 e dales fogo a sabor.
e kuando entenderes |16 ke sera a grasah ko‹gah toma un fuste linpo
|17 e meçe kon akel fuste a grasah e akelo ke se [19r.] apegare ao
410 fuste [kol] dayo kon o kutelo e deita-o |2 dentro enah olya e proba-h
senpre ata ke se |3 nao pege e toda viah meçe-ah kon ese fuste |4 e
deske veres ke e rara ke nao pega en nin |5 uah de ayi toma uah
penah de galinya e mete |6 ah na olya do olio e si veres ke se en‹ge ah
|7 penah enteende ke e kozido e tolye-o do fogo e |8 deita-o sobre a
415 grasah e todah viah meçendo-o |9 e kuando for ralo e fermoso enton e
feito e |10 si dakel barniß kiseires fazer kolor de oro ah |11 partah ah
meitade o kuanto ki‹seres fazer |12 e koao do ke fikar en fundo da
olya e des e |13 toma uah onça d.aloes o duas o kuanto |14 ki‹seres
fazer e moyo been en un morteiro |15 e toma dos pous dos aloes e
420 deita os no |16 verniß e estando a olya sobre o fogo e fer‹ba |17 e
131
depois toma uah pekena de folya e pon-ah [19v.] sobre uah taboah e
poe dakela doiradura |2 sobre a folya do estanyo ou de paratah |3 e
si ‹bides ke e boah tira-ah de sobre o fo e si |4 nao deitalye mais dos
pous ata ke se‹gah been |5 doirada tolya-ah de sobre o fogo e kolye-o e
425 asi |6 fikara bou.
kapitulo 43 toma dos |7 farelos grosos e poyn-os |8 e remolyo en
uah konka vidriadah e despois |9 ke foren remolyados e koados por
un pano dobrado |10 linpo e kon akelah aguah poeras onde sabes |11
e oro ençima e brune-o a sabor kon dente de |12 porko \
430 kapitulo 44 si ki‹seres fazer |13 boah roseta toma o |14 brasil e mole-o
no almofariß ke se‹ga been moido |15 penyerao e toma uah pokah de
kal vir‹gen e |16 pona nuah altamiah kon aguah ates ke se faça |17 a
aguah krarah e kon akelah aguah moy o brasil [20r.] e lançalye un
pekeno de pedra ume e destenpera |2 kon goma e eskre‹be kon el \
435 kapitulo 45 |3 si ki‹seres fazer bou verde |4 toma o lirio azul verde e
toma aguah dalunbre |5 e molya os panos no alumbre e depois no
çumo |6 de lirio e faze komo ao katah sol nos vidos \ abraham bar
yehudah ç”t aben ˛ayim
4. TRANSLATION
Here begins the book on how to make colors of all shades for illuminating
books. And let us talk first about gold (sol). If you wish to make gold with
which you can illuminate or paint or capitalize or write, as this book states and
directs you forward, do not omit nor add anything more that what your book
states, for if you do so you will do everything wrong, and you will not profit
no matter what you do. And for this reason do not do proceed in any way
other than what the book tells you and directs you to do. First of all take ten
ounces of Jupiter, that is, tin, clear and [1v.] pure, and take five ounces of
fugitive, that is, quicksilver, and first melt the Jupiter and place the fugitive into
a mortar. And the melted tin, place it in the mortar with the quicksilver. And
mix with it five ounces of sulphur and 2 ounces of sal ammoniac, that is,
ano¸star, and let this be well ground and sifted, and that which cannot be sifted,
grind it again until it is all sifted, and then put it all together in a well-cleaned
basin and then place it in a glass bottle. And cover it with pitch and good
strong clay four or five times so that it can withstand the fire, and then put it in
a pot full of ashes from the hearth and give it moderate heat until you see the
fire becoming red, and then cover the mouth of the bottle with some clay and
let [the pan] stand over the fire together with the bottle [2r.] on a tripod. And
132
from time to time uncover the mouth of the bottle, and when you see that it
gives forth no smoke remove the pot along with the bottle from the fire and let
[it] stand waiting one more day. And then break the bottle and take out an
ingot of fine gold, and grind it very fine when you want to work with it, and
temper it with gum-water, and do with it as you wish. And this gold is called
aurum musicum [bisulphide of tin (Blondheim 1929: 120)].
Chapter 2. On gold. To make aurum musicum with which to write, take Jupiter
and quicksilver and sal ammoniac and sulphur, as much of one as the other, and
melt them in a large iron spoon the Jupiter and pour [it] over the quicksilver and
mix them [2v.] all well with a [wooden] stick and pour it hot into the mortar
[and] over the sal ammoniac powder and sulphur, and grind it until it is all a
fine powder, and take all this and place it in an iron spoon in such a way that
all of it fits well and smoothly, and heat it over embers until it smoke appears,
always mixing well. And then grind it again, and put it in the spoon and place
it over the embers until it smokes. And do this three times. Then place it in a
piece of coarse linen and tie it up as tightly as you can, and place it over the
fire in a pot containing quicklime, which has just come out of the furnace, until
it gives off steam. And place it first over the embers so that it may already be
hot. And then take it from the cloth [3r.] and place it in a clay pitcher that has
been made and baked like a bottle, and cover it and encase the lid with clay to
seal it well. Place it on a tripod and heat it vigorously from morning until
midday, that is until the sixth hour. And if the clay breaks off, replace the
broken fragments. With this done, remove your work, but first let it cool until
the next day. And when you want to write with it, take gum arabic, the size of
a chick pea and another the size a lentil, and pour water into a shell, enough to
submerge the gum arabic, and with this temper the gold, and write. And the
letter will remain fully gilded.
Chapter 3. On gold. [3v.] To put gold in books or upon leaves of Jupiter,

which is called alfer, so that you may burnish it, take stag’s glue, and pour it
into water and let it stand in the wind until it dissolves on its own. And when
it is fully dissolved, test it on your finger; if it is very liquid and soft in the
water thus prepared, pour ochre and a completely-ground peachstone over it.
The peachstone should first be burnt, in order for the gold to be more
resplendent. And then place this concoction of glue and ochre and burnt
peachstone as described, and write where you wish, but first breathe on the
place where you apply the concoction described above. And place over it the
gold for writing, and once it is dry burnish it very gently with a peccary tooth,
and thus you will work well [with] the gold. Likewise, [4r.] in order to make
133
size for applying gold, take the skin of a conger-eel, and set it to boil until it
dissolves on its own, and strain it in a strainer, and make size of ochre and
peachstone. And place the gold for writing on this size made of fish.
Chapter 4. On gold. If you wish to apply the gold with a dobre or a crown or a
florin or with a gold ring, take rock-crystal and grind it well and take an egg-
white, well fried, and gum, and temper [it] with this, and make a mass. And
place this mass where you wish and let it dry, and take a dobra or a florin or a
crown or a ring as mentioned, and rub it well against the mass. And whatever
part of the mass becomes brilliant will not come off. And then weigh [4v.]
the dobra and you will not find there to be any less of it.
Chapter 5. On how to make fine blue. When you want to make blue that
resembles Acre [ultramarine], take a large new pot, make four holes in it, and
place two iron rods in these holes, in the form of a cross. And then take leaves
of sterling Luna, that is, very thin leaves of sterline silver, and smear them with
honey, and place them on the iron rods in such a way that they do not touch
each other. And then pour into the pot strong vinegar, enough to reach the
iron rods and not higher. And then seal the pot well with strong clay, and
place the pot in hot animal manure, such that it be buried [5r.] in it up to its
mouth, and be well covered for twenty-two days. And at the end of these 22
days, unseal the pot and you will find in its mouth some fine blue, and scrape
it with a wooden or reed spatula. And then return the leaves to the pot and let
it stand as many days as you did before, and in this way you will make good
blue, and you can make as little or as much as you need.
Chapter 6. On blue. In order to temper blue, take gum arabic that is brilliant
and finely divided in water in a cup. And then take this cup where it stands
and strain it through a linen cloth. And take a third of an egg-white and put it,
with the gum arabic, in a retort or bowl. On that note, you should know that
the blue must be renewed each day, for if the that water sits for too long a time
[5v.] in the blue it will turn black; so do this two or three times a day. And
take from the gum arabic and egg-white and write with it. Beware of shiny
blue; that which resembles dark violet is good. For blue I give you [this sign]
to recognize it: place a little of it on your tongue or palm or fingernail, and if
you smell it as it usually does, it is bad.
Chapter 7. On blue. In order to temper blue, take ultramarine and grind it

well with lye made out of vine branches lightly. Collect in a shell or bowl and
wash it with the lye made out of vine branches, and grind it again lightly with
134
a little rose [color]. And write whatever you wish, or illuminate, or paint, or
temper it with a little bit of egg-white with gum. And do not let it sit from one
day [6r.] to the next, for that it would turn black.
Chapter 8. In order to make rose, take one ounce of fine brazil-wood, and
scrape it very fine, and set it aside. And then take a quarter ounce of alum and
take two pennyweights of white lead and grind it with the alum in a mortar
and set it aside. And then take the brazil-wood and place it in a cup [made] of
malega, and put in the other powders with the brazil-wood, and pour urine
over them until they are covered. And let them stand thus three whole days,
always stirring them with a stick 5 or 6 times each day. And then pass it and
strain it through a linen cloth above a trough made of gypsum or chalk-stone.
And let it soak in the [6v.] trough, and when it is dry, scrape it very well with
a spatula, and keep it carefully from the air. And when you want to work
with it grind it with hum water.
Chapter 9. In order to make another rose color, take brazil wood, as much as
you need, and scrape it very fine and place it in a small new pot. And place in
the pot lye of vine branches, so that the brazil wood is covered with it. Put it
on the fire and heat it enough for the lye to take up the substance of the brazil
wood. And take two part alum and at least a half a part of chalk and grind
each one well by itself. And then mix it and grind it together, and make, as
you already know, rose out of alum.
Chapter 10. In order to make very fine [7r.] red lead, take white lead, as
much as you wish, and grind it and sift it, and toss it in a broad bowl or bowls.
And take it to a glazed furnace, and leave it stand for twenty-two days. And
after these days take it out of the furnace, and you will find very good red
lead. In this way you will make as much as you wish
Chapter 11. In order to make very fine verdigris, take very thin leaves of
copper foil and moisten them in very hot and strong vinegar. And put them in
a pot leaning on its side, and smear the mouth of the pot with honey and cover
it with potsherds, and bury it under the manure of large animals, and let it
stand there for thirty-one days. And after these days take the pot out and you
will find verdigris, and scrape it with a spatula. And [7v.] if you want to
make more, repeat as directed, and you will have good verdigris.
Chapter 12. On another kind of verdigris. Take a bowl and half-fill it with
well-putrid urine, and take a brass basin, very well washed from bottom to
top, and place it above the bowl so that the urine does not come within two
135
fingers' distance of the bottom of the basin. And the bottom of the basin
should be smeared with good honey, the basin should be half-filled with that
same urine, And above it place another bowl upside down, and above the
bowl place coverings for pack mules, and pour the urine from the basin into
the bowl underneath. And look at the bottom of the basin and you will find
the honey that you placed there has become verdigris. And scrape it with a
spatula and [8r.] keep it in paper. And if you wish to make more green,
smear the bottom of the basin with honey and do as you did before, and this
way you will make as much as you wish. And for tempering this green, when
you want to work with it, grind it very well beforehand and place in it a little
bit of well-ground saffron, and temper it with gum-water, so that the devil
will not then remove the color.
Chapter 13. In order to make fine carmine, take a large new pot that holds
four açumbres of water, and fill it with human urine. And mix it for days, and
make it very clear all the while so that it gives off foam. And once it is very
clear and skimmed, take a [8v.] large bowl and place rye-straw over it, and
above the straw a linen cloth. And on the cloth place ashes of vine branches,
two parts, and a third part quicklime, and place a pot underneath. And throw
on the lye the strained urine that you strained through fried meat and
continue straining it until the pot is full of this strained lye, in such a way that
there are four açumbres of it there. And place it on the fire until only two
fingers of it are left, and on the fire put another pot full of clear urine with the
strained lye and heat both. And into the pot of clarified urine with the
strained lye toss one pound of lac, and heat it gently, all the while stirring it
with a slightly forked stick. And when the lac is melted, strain it with a linen
bag, and place a basin underneath; whatever remains in the [9r.] bag, place it
in the pot of strained lye, which you have kept on the fire with gentle heat,
until it is melted, stirring with a piece of wood. And then strain it separately
with the bag containing the powder. Thus you can make carmine of two
kinds, though first you must clarify the urine.
Chapter 14. For another kind of carmine, take a pot of clean and tepid water
in such a way that you can dissolve a pound of alum, and take one açumbre of
this alum water. Pour half of it into a pot and the other half into another pot,
and stir it with a stick. And then let it clarify, and when it has settled continue
pouring off the water that is floating. And when it is so clear that you cannot
remove any more water, place them in separate linen bags and hang them up
[9v.] so that they drip over separate pots or bowls. And that which is
strained, if any, remove it and return it to the bag, and do so until it comes out
clear. And once it is clear, make small particles like chick peas and place them
136
to dry in the sun, which should be gentle, and if the sun is strong place a sheet
on top of them. And once they are dry, keep them as do with them what you
will.
Chapter 15. To make vermilion, take five pounds of quicksilver, that is

mercury, and place it in a bottle or large glazed bowl, and take a pound of
very fine virgin sulphur. And pour the powdered sulphur over the quicksilver
little by little until it is well incorporated, always stirring it with a dog's foot
that has its hair and wool, until the fire turns to ashes. [10r.] And once the fire
has thus died down, place it in two new pots that are made like bottles, broad
below and narrow above. And seal them, leaving only a small hole through
which the vapor will escape. And place the pots on the fire on their holders
and cover them well with clay, and place a bowl over the holes. And when
you see the smoke coming out red and not malodorous, place a thin spit in it.
And if anything sticks to the spit, remove the pots from the fire and let it cool.
And once it is cool break the pots and you will find the vermilion made. With
these measures you ill made as much vermilion as you wish: for a terça of
mercury take five pounds of sulphur, and for five pounds of mercury take one
pound of sulphur. And regulate the first in such as way that it does not [10v.]
burn, and keep the fire moderate, neither quick nor slow. On this note, if by
chance the vermilion burns, break the pots and grind it and incorporate it and
mix it with another measure of mercury and sulphur, and place them in other
pots and proceed as described. And pay attention to the vapours that escape,
thus you will never ruin anything.
Chapter 16. In order to place gold on a sword or a knife, take the ends of
carbonized ashes and place them in the forge until they are very red. And
then remove them and place them on a board, and pour two handfuls of
ground salt over them, and grind it all together. And then add two ounces of
sal ammoniac [and] a half-ounce of verdigris, and knead it all very well with
strong vinegar. And first have the sword or knife cleaned. [11r.] And, as if
you were writing, place some of these ashes where you wish to make letters or
designs. But first have it smeared with vegera boli and with azarcoanboli. And
if it is winter time, let the sword or knife stand two days without cleaning.
And if it is summer, let is stand one day, then wash and keep your weapon,
for it is already adorned.
Chapter 17. In order to dye bones and sticks whatever color you wish, take a
piece of box-wood or any other kind of wood, and boil it in olive oil three
times. And then remove it and pour powder from ground sulphur over it, and
let it stand for three days. And then clean it and it will be black outside and in.
137
Chapter 18. In order to color bones [11v.] red, take lac, as much as you need,
and dissolve it in very strong vinegar, together with a quart of sal ammoniac.
And then boil the bones or the wood in this mixture until they turn red and
the color sticks, for this dye appears read inside and out.
Chapter 19. In order to color box-wood or another wood a beautiful black that
resembles jet, take iron filings and copper filings and plain salt and litharge [of
silver] and litharge of gold, which is found in arsenic, an ounce of each one,
and place them in very strong vinegar for thirty days. And then boil them
until half the vinegar is gone, and then take vitriol and temper it in water until
it is a little thick. And put [12r.] the wood and the bones in it many times,
taking it out into the sun each time. And you will see that it will become black
inside and out, and this color will never disappear from the wood or the
bones.
Chapter 20. In order to make bones or handles a green color like that of
verdigris, take verdigris and sal ammoniac, two ounces of each, and grind
them in very strong vinegar. And place in this the bones or handles in such a
way that they are covered by the vinegar and the powders. And leave them to
settle until they become a green that pleases you. And for better results boil
them in a copper kettle, and if in this dye you place any kind of wood it will
turn a very beautiful green.
Chapter 21. On filed bones to make chess pieces. When you want to make
bones [12v.] with fire or without fire, and to make many, if you wish to do so,
take cow bones on which there is no meat at all, or ivory bones, and file them
well with a file until they are filed like iron filings or sawdust. And place
these filings in a glazed pot with very strong vinegar, and be sure the pot is
well covered and encased with clay around the outside. And then take five
ounces of cuprous oxide and weighted garlic in the amount of one ounce, and
heat them in four ounces or more of red vinegar. And cover those bones with
this vinegar, and stir them well with this concoction. And put it one the fire
until they melt, and when you see that they are melting and coagulate like
lead, and you want them to remain white, take [ . . . ]
[13r.] [Chapter 24 . . . ] rags, sunflower juices, and grains. And once you have
ten or twenty rags, or one that is large, full of juice, take a bowl or ceramic pot
full of human urine and place it on top of two crossed staves, which should
not touch the urine. And place the rags stretched out on top, and let them
stand there for nine or twelve days, stirring the urine four or five or six times a
138
day, and turning the rags from one side to the other until they are colored not
of urine. And the more putrid and fetid the urine is, the better the color it will
make. And once they are colored with the vapor of the urine, place them in
the sun until they take on the color of mulberry. When the rags are well
colored and dried, keep them well away from winter air. When you want to
illuminate or work with them, [13v.] take a well-washed retort or a shell and
with scissors cut a bit of the sunflower rag, and place it in a conch or shell or
retort and pour gum-water on it. And once the rag is well soaked with it, stir
it well and you can work with it immediately; for if it stood there more than
one day, it would immediately spoil and lose its color. And thus each year
you can make as much as you wish.
Chapter 25. Whoever wishes to work with gold or other colors, and how they
should be made and mixed, it should be done this way, which is already well
tested. First, when you wish to apply gold, take ochre and white lead, as
much of one as of the other, and a little chalk, and mix all these things together
and grind them thoroughly and very well with egg-white [14r.] that is very
thin and light. And if it very stiff, add water to it and grind it well, and if it is
very light and egg-white to it. And test this on a piece of parchment. On the
place where you wish to apply the gold, apply this concoction with a brush,
and before it dries place the gold upon it. And then burnish it very well with a
swine's or horse's tooth.
Chapter 26. In order to temper blue, take egg-yolk in such a way that no
albumen goes with it. Then grind this blue very well in a ceramic bowl, and
take this egg-yolk and mix it with the blue. And grind it all together very well,
and once it is well ground, place this blue in a pitcher and add water to it, and
place your hand and mix it with your finger many times. And as soon as you
see this water appear above the blue, and once you have stopped mixing it,
then [ . . . 14v. . . .] sunflower. And it should be very pure, and very clean, free
from all impurities. And as soon as this is done, temper it with gum-water
and then write with it. But before you add water, be sure that the water has
well dried from the blue. And if you wish you can add egg-white [or] brazil-
wood to give it a better color.
Chapter 27. If you wish to make good rose-color, take brazil-wood, as much
as you want, and scrape it well over a conch or retort, then add alum to it.
And once you have done this, take the urine of a chaste man and add pour it
over the brazil-wood and the alum until they are well-covered, and let them
stand thus for three days. And then take a piece of chalk and throw some of
its powder over this brazil-wood until it seems that there is as much of one
139
[15r.] as of the other. And then let this concoction stand for a day or two.
And then take this rose-color and grind it with gummed egg-white, and write
with it. If you wish to make indigo, put blue in it, and if perhaps you wish to
make a black color, put black in it. And if perhaps you wish to turn white
color black, add black to it and white and take brazil-wood and place it in a
white cloth and strain it over chalk. Know that the principal colors are ten:
blue, oripiment, and vermilion, green, Sufi carmine, sunflower, saffron, red
lead, white lead, brazil-wood. When you want to thin egg-white add the sap
of a fig tree to it, and thin it very well, clear as water, for you work.
[15v.] Chapter 28. If you wish to make green and to temper it, add vinegar
and green and egg-yolk and grind it all together. There should be three parts
green and a fourth part yolk. And if you wish to make better color, add gum-
water to it and temper it with that. And if you wish to turn it another color,
add saffron to it, and if you wish to turn it another color, mix ceruse into it,
and it will appear green and white in the shade.
Chapter 29. If you wish to temper blue, add water and grind it with it a little,
and once the water has dried off, if you wish to turn it another color, add
ceruse, that is Leo, and egg-white. And if you wish to turn it sky-blue, add
three parts ceruse to one part blue.
Chapter 30. If you want [16r.] good carmine, take carmine, water, and half as
much egg-yolk, and grind it all together. And you want the color to resemble
blood, add to it a third part of blue.
Chapter 31. If you wish to temper saffron in order to write with it, add egg-
white and do not grind it nor mix anything else with it. And if you wish to
put it in another similar color, add oripiment, well ground with egg-white. If
you wish to shade with it, add well-ground blue, as much as a third part and
no more.
Chapter 32. If you wish to temper oripiment, temper it with water and with
egg-white. And then remove this water from it and work with it, for it needs
no other preparation.
Chapter 33. If you wish to temper black indigo, [16v.] take gum-water and
egg-yolk and grind it all together. And if you wish to temper or work [with
it], add white to it until it resembles the color of clouds.
140
Chapter 34. Take blue and temper it with gum-water and with egg-yolk, and
pour carmine or brazil-wood over it to shade it.
Chapter 35. If you wish to color with light blue, shade it with pure blue. If
you wish to color with carmine, shade it with carmine or with brazil-wood or
with vermilion. And if you wish to color with light indigo, shade it with pure
green. And if you wish to color with red lead, shade it with carmine or brazil-
wood or red. And if you wish to color with vermilion, shade it with brazil-
wood or carmine. But all colors can be shaded with black.
[17r.] Chapter 36. Take saffron and gum and egg-white tempered with all
these things, and place it all in the spot or letter that you wish to make. And
take the leaf of gold and very neatly and carefully, in a house without wind
and without people, so that you do not speak, and a bandage or a rag over the
mouth and nostrils, so that you do not breathe on the gold, which is tied
around the head. And place it over the above-mentioned items and let it stand
for one hour in the daytime. And then take a little cotton and place it gently
over this leaf, and whatever must remain for the letter, let it stand, and remove
the rest. And once you have done this, get hold of the burnishing iron and
burnish it very well with a swine's tooth.
Chapter 37. The mixtures of colors as follows. Whoever wishes to makes a

color like blue or carmine, take [17v.] a half blue and a half carmine, and third
part white. And if you want it redder, put more carmine, and if you want it
whiter, put more white. And when you want to pain or illuminate, take blue
and carmine, a half each, and mix it all with gum-water and egg-white, and
you can paint or illuminate. And put more temper in it for it to be very light.
Chapter 38. If you wish to apply gold to a book, take boiled parchment water,
which is good and strong and apply it with a brush once or twice in the place
where you wish to apply it. And then grind chalk in vigorously boiling water
and put in it a little saffron and apply it three times where you want to apply
the gold. And once it is dry apply as much as you wish, and apply the gold
[18r.] with cold gum-water. And then burnish it vigorously with a swine's
tooth.
Chapter 39. If you wish to properly burnish gold or silver, polish the
burnishing iron well and then open your mouth and breathe upon the gold
warmed with a clean hand. Be sure the burnishing iron is hot. And cover the
gold with an old linen cloth, and then burnish it another time over the cloth,
141
heating the burnishing iron once again, over the cloth, and breathing with an
open mouth.
Chapter 40. If you wish to make glue, take two pieces of parchment and wash
them very well, and then place them in a new [old] pot, and heat them
vigorously until they are well boiled. And once the first water has
disappeared, add more water into it, and as soon as you wish to try [it], take a
little of it and place it in your palm. And place one hand with the other, and if
your hands grip, it must be that your glue is well made.
[18v.] Chapter 41. If you wish to make ochre, take red, tempering [it] as much
as you please, and mix it with yellow that is good. And if you see that it is
very colored put a little black, and it will be good, and if it becomes too black-
colored put a little at first and then put more, as much as the other.
Chapter 42. If you wish to make varnish, take a pound of nobra grease (or two
pounds, or as much as you wish to make), and take a pound of grease and two
of linseed oil and place each one in its own pot. The pots should be new. And
cook each one vigorously and be sure than neither water nor anything else fall
into them, and heat them at your discretion. And when you think that the
grease is cooked take a clean piece of wood and stir the grease with it. And
whatever [19r.] sticks to the wood, scrape it off with a knife and put it into the
pot, and continue to test it until it does not stick. And always stir it with that
piece of wood, and once you see that it is thin and does not stick to any there,
take a hen's feather and put it in the pot of oil. And if you see that the feather
is swelling, assume that it is cooked, and take it from the fire. And pour it
over the grease, always mixing it. And when it is thin and fine, then it is done.
And if you want to make gold color from that varnish, separate half or as
much as you want to make, and strain it from what remains at the bottom of
the pot. And then take an ounce or two of aloe, or as much as you wish to
make, grind it well in a mortar. And take the powdered aloe and place it in
the varnish, and with the pot on the fire, let it boil. And then take a little foil
and place it [19v.] on a board, and apply some of this gilding to the tin or
silver foil. And if you see that it is good, remove it from the fire, and if not,
place more powdered [aloe] until it is well gilded. Remove it from the fire and
gather it up, and thus it will remain good.
Chapter 43. Take two large pieces of bran and place them to soak in a glazed
shell, and as soon as they are soaked, strain [them] through a clean folded
cloth. And with this water apply it where you wish, and gold over it. And
burnish it to taste with a swine's tooth.
142
Chapter 44. If you wish to make good rose-color, take brazil-wood and grind
it in a mortar, until it is well ground. Sift it, and take a little virgin lime and
place it in a glazed earthenware bowl with water until the water becomes
clear, and with this water grind the brazil-wood, [20r.] and put in it a little
alum, temper it with gum, and write with it.
Chapter 45. If you wish to make good green, take a blue-green lily and take
alum water, and wet the rags in the alum and then in the juice of the lily and
do as you did with the sunflower [juice] and urine.
5. COMMENTARY
f. 1r.
1 'snyymwqyS
se.komeinça
'begins'. ModPg. começar < CUM+INITI‹ARE. This is one of the numerous
instances in which n n appears where the modern orthography has <m>
or ∅ (e.g. wypnyl linpio 'clean', wd'rpnyX tenprado 'tempered'). Indeed
there is only a single instance of a nasalizing context spelled with m m
(see note line 436). Blondheim also uncritically interprets the digraph yy
yy as the Castilian-style diphthong <ie> in this and other forms in his
transcription (e.g. §y'yb been 'well'). It should be noted, however, that
the orthography alone does not strictly preclude this reading, since y
may stand for any non-low front vowel.
3 lwS wrw'
oro sol
Blondheim (1930: 119) states that sol 'sun' is the name given by medieval
alchemists to gold. Other such glosses within the text include wyyn'Xsy'
estanyo 'tin' and ryXybw&g ‹gubiter (line 9),ygwz' azoge 'mercury' and wbyXy&gwp
fu‹gitibo (line 10) or w&byw §y&gr' ar‹gen vibo (line 192), and 'X'rp prata 'silver'
and 'nwl ¶yl'nyXSy' estinaliß luna (line 80).
143
r'nymwl'
aluminar
'illuminate'. ModPg. alumiar 'light (up), give off light' < *ALL‹UMIN‹ARE
has clearly shifted in meaning, with the sense here taken over by a more
recent importation from Latin, ModPg. iluminar < ILLUMIN A ‹ RE . The
form used here has been relatinized, with its etymological /n/ restored.
4 r'wdybq
kabidoar
'capitalize'. The more recent formation capitular < MedL. CAPITULARE
evidently replaced this word. The form used here could reflect either a
verb formed on the basis of cabido < CAPITULU , or perhaps a semi-
latinizing spelling similar to kolorar (see note line 353).
yXn'yd' r'dnm y' ry#Syd wrbyl yXSy'yX wmwq

komo te.este libro di‹ser e mandar adiante
'as this book says and guides you forward'. The clitic pronoun te (object
of di‹ser e mandar) appears at the head of the clause (see a similar phrase
in line 6). It is tempting to construe the a that is affixed to diante
'forward' (< denante < DE INANTE) as a still-independent periphrastic
future auxiliary on which both verbs depend, a phenomenon found
elsewhere in the medieval language (Williams 1962: 207). Yet ModPg.
adiante is attested early enough to here constitute the adverb on its own,
and the S ¸ s spelling of the stem-final consonant in dizer points to the
form here as a future subjunctive (ModPg. disser < DI‹ XERIT) rather than a
future indicative (ModPg. dirá, whose shortened stem probably
originated in Vulgar Latin; Williams (1962: 222) does note that of the
long infinitive was used with infixed pronouns, e.g. dizer-me-ha, until
the fourteenth century).
6 Syryyzyp w' yS 'q

ka si o fizeires
'because if you do (so)'. ka < QUA (RE), replaced in the modern language
by porque (cf. Fr. car), and distinct from ModPg. ca 'here' < ECCU HAC.
The diacritic p p to indicate /f/ occurs inconsistently in the text.
144
lwrp 'ry#byX
te‹bera prol
'gain advantage'. Future subjunctive of ter (ModPg. tiver < Latin perf.
subj. TENUERIT). Most etymological dictionaries derive prol < *prode <
PRODEST, though it was possibly influenced by PROLES 'growth, progeny'
(hence 'advantage, gain').
9 wr'rq
kraro
'clear'. Although the initial cluster is relatinized in ModPg. claro, here
the form occurs without latinizing interference. Nevertheless, krara do
ovo 'egg white' and klara occur together in a single sentence in chapter
27, and the variants klarifikar (line 176) and krarifikar (line 181) 'clarify'
also occur in consecutive chapters.
f. 1v.
11 ygwz' w' wq ¶yr'#pwml' wn

no almofariß ko o azoge
'in the mortar with the mercury'. Although both of the terms here are
Arabic loanwords (ModPg. almofariz < al-mi˛raß and azougue < az-z÷uq),
the <z> in the modern spelling has a different source in each. Yet the
fact that the letters of the Hebrew alphabet have cognates in the Arabic
alphabet allows the scribe to capture the etymological distinction not
maintained in the Roman-letter spelling, by using c ß and z z
respectively. ModPg. com 'with' < CUM occurs here without the final
consonant, which does appear elsewhere in the text. Based on this
instance, the variation does not seem to be a function of whether the
following segment is a vowel.
13 r'X#Swn' SyÇ„dy' wqynymr' l'S

sal armeniko ides ano‹star
'sal ammoniac' (ammonium chloride), paraphrased by the loanword
(ModPg. nochatro) based on Ar. an-nu¸s†ar (Blondheim ignores this gloss
in his translation). This word has caused some confusion among
etymologists. Franca (1994) and others list the word as the attested but
now-obsolete almoxatre, while Ferreira (1999) derives nochatro from
nu¸sdar. Corriente (1999) cites the ultimate source of the Arabic term as
Pahlavi an‹o¸s ‹atax¸s 'eternal fire'. The paraphrase marker ides 'that is',
145
which occurs throughout the text, is Latin ID EST, one of a very few
words in the text with diacritic vowels; the others are ryEp¯la' alfer (line
54), yilwbÕna'wqËr¬za' azarkoanboli (line 219), 'Asnyip pinça (line 314), and one
occurrence of yECX¯S˚#p fuste '(wood) stick' (line 174).
16 'mwdwr h'w'
uah rodoma
'a bottle'. This word appears in various guises in the text: ua aredoma,
d.aredoma, ko.a arodoma. Most etymological dictionaries indicate redoma
'glass sheath' as being "de origem obscura/incerta/controversa,"
though some cite Ar. ra∂ûma as the source (Silvera Bueno 1967 suggests
more specifically a Mozarabic dialect variant).
18 r'l wd hzynys yd h'y#g

‹geah de çinizah do lar
'full of ashes from the hearth'. Words spelled in ModPg. with <ch>
such as cheio < PLENU are most often spelled in this text with the
Hebrew g g, augmented by a diacritic (presumably meant to indicate [ê]
or [Ê], which have since deaffricated). ModPg. cinza 'ash' < *CINITIA (da
Cunha 1982 suggests *CINISIA), derived from CINIS.
f. 2r.
23 h'wq
koah
'with the (f.)'. Although this preposition is normally spelled with an
overt final § n, on occasion the scribe drops the nasal letter and fuses it
to the following article or pronoun. In the modern language this
preposition is always written separately (spelled com), but orthographic
fusion does occur with other prepositions, e.g. na < en+a, nesta < en+esta.
In this text (and throughout the corpus) en is usually written out in full
(e.g. h'w' §y' en uah, ModPg. numa, lyq' §y' en akel, ModPg. naquele),
though some contractions do occur (e.g. line 248 'XSyny' enesta, line 410
a
hny' en h), and are generally more frequent in O libro de ma‹gika.
24 wrw' yd w''p §w' S'r'#g'

a‹garas un pao de oro
'you will find an ingot of gold'. ModPg. achar 'find' < AFFLARE 'sniff
out'. The fricative, which derives from a non-sibilant Latin sound, is
146
spelled here by #g ‹g. Mod Pg. pão ‘bread’ < PANE here retains more of its
original meaning of 'loaf' (of bread) by extension to 'ingot' (of iron).
26 lwrp h' wX yly'wq yz'#p

faze ko.ele tu ah prol
'do with it what you will'. This is probably an incorrect division of the
possessive adjective tua (this is in fact the phrase cited by Blondheim to
illustrate how the scribe "stupidly divides common words"), with prol
as direct object of faze.
27 wqySwm wrw' w' §yzyd wrw' yXSy'

este oro dizen o oro musiko
'this gold is called aurum musicum (bisulphide of tin)'. As in medieval
Spanish, dizen takes a direct object, in the sense of ModPg. chamar 'call'.
musico < MedL. musaicum 'mosaic'.
30 r'yylwq
kulyar
ModPg. colher 'spoon' < COCHLE‹ARE. The text contains several instances
of words where an expected /e/ or /i/ appears as /a/, e.g. fu‹gatibo
'quicksilver' in the following line (but fu‹gitivo elsewhere), as well as
other instances of kulyar in this chapter (cf. ModSp. cuchara). By
contrast, the three-letter system leaves the identity of the vowel in the
first syllable (apart from its backness) ambiguous.
f. 2v.
32 Syr'#pwml'
almofaris
'mortar'. As opposed to line 11 above, the etymological identity of the
final consonant is obscured by the scribe's use here of the "standard"
sibilant letter S ¸s.
35 yXnym 'XlwS
solta mente
'freely'. A syncopated form of the past participle of SOLVERE (*SOLTUS <
SOLUTUS ) serves as the basis for ModPg. solto 'free' and soltar 'let go'.
The modern language does not seem to have preserved the adverbial
use of this word.
147
36 y#gymw#p
fome‹ge
'smokes'. ModPg. fumigar < FUMIG‹ARE. As preceded by ata ke 'until', the
verb is properly in the subjunctive; yet the /g/ of the stem is not
maintained as it is be in ModPg. fumigue but is instead softened by the
front vowel, as indicated by the diacritic on g g.
39 ¶yzyw
vezeß
'times'. This is another sibilant-rich word that occurs in many variant
spellings. The use of c ß in the final position is especially rare elsewhere
in the corpus (but note the use of the letter's final-position form).
40 '#byb l'q
kal bi‹ba
'quicklime'. ModPg. cal 'lime' is not the reflex of classical CALX, -CIS but
rather derives from VLat. *cals, a derivative of the classical accusative
(Houaiss 2001). Note the spelling of viva, in which both instances of
/v/ are spelled using b, albeit in one case missing the diacritic to signal
the fricative reading, even though using w would not yield any of the
potential ambiguities (cf. chapter 3 § 2.3.2)
f. 3r.
45 'd'r's
çarada
'sealed'. ModPg. cerrar < SER‹ARE. This is another instance in which a
word in the text appears with ' , representing /a/, where /e/ is
expected (cf. note line 30).
46 h''yyn'm' Syd
des amanyaah
'from morning'. This preposition no longer occurs in the modern
language as an free-standing word (cf. ModFr. dès 'from'). It also occurs
frequently in the text in the combination yqSyd des.ke 'as soon as' (cf.
ModPg. manhã). The hiatus left by the deletion of intervocalic /n/ is
spelled here in as much detail as the orthography affords, with two ' as
well as h for word-final /a/.
148
47 r'yyl'rbyqSy'
eskebralyar
'break'. ModPg. quebrar < CREPARE, with r-metathesis also preserved in
the modern form (cf. chapter 7 § 2.3). The verb formed here is prefixed
by EX- and suffixed by -ALIA , a noun-forming affix with a collective
meaning, i.e. based on a noun *kebralya 'broken-off parts'.
51 r'dnww'
avondar
'cover (with water)' < *ADFUNDARE, with an unexpected progressive
voicing of the root-initial consonant. Only nominal forms of this root
are found in Modern Portuguese, e.g. fundo 'bottom'. The use of
double-ww ww to represent a CV syllable is extremely rare (cf. wwrys çervo
above). Indeed it is highly unconventional in this writing system to see
any doubled letters other than ' and y. It is conceivable, of course, that
as in contemporaneous Roman-letter orthographies, the double-letter
spelling in fact indicates a voiceless segment (cf. Domincovich 1948).
f. 3v.
54 ryEpl¬'
alfer
Given as a more specific name for ryXypw#g yd S'yylw#p folyas de ‹gupiter
'tin leaves', this is one of five words in the text with diacritic vowels
added (see note line 13). In this context it appears to have little
semantic relation to either alferes < Ar. al-f‹ a ris 'horseman' or alfir, a
variant of alfim < Ar. al-fil < Pers. pil 'elephant', the chess piece usually
represented as the bishop in the modern Western form of the game. It is
also possible that it is a metathesized form of ly#pr'm marfil 'ivory'
without the unetymological m- < Ar. fia÷m 'bones' (cf. note line 253).
r'yynwrb
brunyar
'burnish, polish'. ModPg. brunir, borrowed from French, itself based on
a Germanic loanword. This form shows not only an unexpected ' a in
the final syllable (unless the verb has simply been treated as belonging
to the first conjugation), but also an unetymological palatal /µ/
(although not in all occurrences, e.g. line 63 w'ynwrb brune-o 'burnish it').
149
MedL. BRUNUS is attested in Isidore of Seville (6th c.) and the glosses of
Reichenau (8th c.), so the verb here could be a native formation based on
*BRUNIARE.
wwrys yd ydwrg
grude de çervo
'stag's glue'. ModPg. grude < GL‹U TINE, with the l > r seen also in the
initial clusters of other words, e.g. krara < CLARA . ModPg. cervo <
CERVUS, with rare use of ww, though etymologically justified, to represent
CV syllable (cf. r'dnww' avondar below).
56 ys'Syd
desaçe
'dissolves'. This could be considered an ad hoc borrowing from
Castilian, as shown by the lack of any consonantal segment following
the prefix des- (ModPg. desfazer, ModSp. deshacer 'undo'). More native
forms wXyy#pSyd desfeito (line 56) and 'g'#pSyd desfaga (line 65) do also
occur. This and several other more obvious "errors" could point to a
copyist of Castilian extraction (see chapter 7 § 4.3).
57 wgySyp yd wswr'q
karoço de pesego
'peachstone'. Note that the sibilant resulting from assimilation (ModPg.
pêssego < PERSICU) is not in this instance spelled with the Hebrew letter s
ç, which normally serves this purpose. The s in karoço is indeed not the
reflex of a simple Latin /s/, though the etymology of the word is not
certain. Most dictionaries cite VLat. CARUDIUM, based on Gk. karydion (a
diminutive of 'hazelnut'), though Houaiss (2001) also suggests a form
*coroço based on COR, CORDIS 'heart' as the source.
59 yXnysydn'rpSyr
resprandeçente
'glowing'. ModPg. resplandecente. This form as used here shows not
only the morphology associated with the Latin iterative ending -‹ESCERE,
but also the unrestored PL > pr in the second syllable.
150
§wys'#pnwq
konfaçion
'concoction'. While ModPg. confecção i s probably borrowed from
French, this is the more native development of CONFECTI‹ONE, as shown
by the lack of obstruent in the palatalization of the -CT - cluster. The
form also displays the /a/ for expected /e/ that occurs throughout the
text. In other occurrences it also occasionally lacks the y representing
/y/ in the final syllable (e.g. line 62).
61 Syry#gwp ydnw'
onde pu‹geres
'where you place (it)'. Note that the sentence immediately preceding
uses u < UBI as the locative pronoun (cf. Fr. où), a form that appears
elsewhere in the text but that has since disappeared from the language.
63 wXw#Sny'
en‹suto
'dry'. ModPg. enxuto < EXSUCTU, past participle of EXSUGERE (the related
EXSUCA ‹ RE yields ModPg. enxugar 'dry (v.)' and the semantic doublet of
this adjective, enxugado). The unetymological /n/ could have
developed by confusion with similar words beginning with en-.
yr'#b'#g
‹ga‹bari
'peccary'. ModPg. javali 'boar' < Ar. ¸gabalı, an abbreviation of ƒinzır
¸gabalı 'mountain swine' (Silvera Bueno 1965 cites the form with r as
archaic). Many of the techniques in the text call for a wqrwp yd yXnyd
dente de porko 'swine's tooth', and it is possible that the intended
reference is not in fact different here.
yrXnym 'S'p
pasa mentre
'gently'. Most of the adverbs in the text, indeed in the Judeo-
Portuguese corpus as a whole, occur with the -mente orthographically
separated. This is the only occurrence in this text with the intrusive /r/
that occurs frequently in medieval Ibero-Romance, probably under the
influence of DUM INTERIM > MedSp. (do)mientre > ModSp. mientra(s)
(Penny 1991: 118).
151
f. 4r.
64 'SyS
sesa
'size'. A pore-filling ingredient of certain glutinous materials. Like the
English term, the Portuguese word is probably borrowed from OIt. sisa
'painter's glue', an aphetic form of assisa, which ultimately derives from
the past participle of ASSIDERE 'seat, settle'.
65 w'yrgnwq yd wryywq
koyro de kongrio
'conger-eel skin'. Unlike ModSp. cuero, ModPg. couro < CORIU has given
way to pele in the general sense of 'skin, hide', retaining only the
meaning of 'leather' (cf. Fr. cuir). The form is spelled here with an /oj/
diphthong, a variation also seen in kousa 'thing' (cf. ModPg. coisa; see
chapter 7 § 2.5).
69 wnylwr#p
frolino
'florin'. A monetary denomination named for the lily pictured on the
coin, and a term ultimately based on FLOR - 'flower'. The form here
shows the l-r metathesis that has occurred in words such as ModPg.
milagre < MIRACULU (cf. chapter 7 § 2.3). It is listed among a set of
metallic objects that may be used for applying color, including the dobra
and eskudo, two other medieval denominations, or an anel 'ring'.
70 wbw' wd 'rq'
akra do obo
'egg-white'. Although krara (de ovo) '(egg) white' appears throughout
the text and the form here is likely a simple scribal error, it is possible
that this is an instance of /l/ having been deleted after l-r metathesis
seen elsewhere (cf. frolino above), i.e. CLARUM > *kralo > *krau > kra.
However, other forms such as mau < *malo < MALUM, which preserve a
diphthong, might argue against this. Note the "unetymological"
spelling of ovo with b, a strategy that avoids writing three consecutive w.
wd'Xyrp
pretado
'fried' (lit. 'blackened'). Most Portuguese etymologists derive ModPg.
preto 'black' from * PRETTU , a regularized past participle of PREMERE
152
'press' (replacing PRESSUS). Corominas (1980), however, describes this

etymology as "increíble," citing the paucity of similar models (e.g. Fr.
assiette 'plate; seat (on a horse)' < *AD+SEDITU, replacing SESSUS as the
participle of SEDE‹ RE 'sit'), and describes Sp. prieto 'dark; tight' as a back-
formation from apretar 'grip, press' < APPECTORARE , based on PECTUS
'breast(bone)'.
72 w'gyr#pSy'
esfriga-o
'rub (it)'. ModPg. esfregar < *EXFRIC ‹A RE, replacing the more evolved
EFFRICA‹ RE with an assimilated prefix.
74 'r'yylwX yS
se tolyara
'will disappear'. ModPg. tolher 'take, rob' < TOLLERE. This form again
shows /a/ for a modern form with /e/ (cf. note line 30).
f. 4v.
77 yrq'
akre
'ultramarine'. Blondheim capitalizes this word in his translation and
inserts his own gloss for this shade evidently meant to resemble blue.
78 Swq'rwp
furakos
'holes'. If the lack of diacritic on the initial consonant is deliberate, then
it is in fact a bilabial stop /p/, suggesting a word akin to ModPg. buraco
'bore'. However, forms with a diacritic, i.e. furako, also occur (e.g. line
197). In fact this form is cited as a variant in da Cunha's (1982) entry for
buraco, though the etymology of the word is listed as being "de origem
controvertida." Alternatively, the word could be related to It. foro 'hole',
based on FORA ‹ RE 'pierce'.
79 Swrq yd 'ryn'm §y' w''yynyw yq

ke venyao en manera de krus
'that they come in the form of a cross'. 3rd pl. pres. subj. of vir 'come'
(ModPg. venhão < VENIANT ), used here in the sense of 'be arranged'.
Note the use of the "basic" sibilant letter in krus (ModPg. cruz < CRUCE),
even though the sibilant does not represent Pg. /s/ < Lat. /s/.
153
'nwl ¶yl'nyXSyd S'nymyl Syry#b'

a‹beres liminas destenaliß luna
'take leaves of tin, (that is) Luna'. The verb haver, which in the modern
language is restricted to only a few auxiliary functions, is used here in
the very lexical sense of 'take'. The y i in the first syllable of liminas
'leaves' is unexpected, especially given that S'nym'l laminas occurs just
below in line 88, a reversal of the usual pattern seen in this text where '
occurs for expected /e/ or /i/ (cf. chapter 7 § 2.6). A word akin to
'sterling' (Blondheim's translation) appears to be the sense of estenaliß,
since it is glossed by prata 'silver' immediately following. Blondheim
capitalizes Luna but does not translate it.
84 'p'X'
atapa
'close, seal'. ModPg. tapar is based on a noun tapa 'cover, lid' < Gothic
*tappa. The initial a-, which occurs on several other verbs in the text,
may therefore in this instance may be construed as a denominal verb-
forming prefix.
85 'd'ryXwS
soterada
'buried' < *SUB+TERRATA. Although the modern language uses enterrar
'bury', the sub- prefix occurs in more recent Latin borrowings such as
subterrâneo 'underground'. A variant form §wS son is used as an
independent preposition in line 138.
f. 5r.
87 w'p'r'
arapa-o
'scrape (it)'. ModPg. raspar < Gothic *hrapôn, but again with a prosthetic
a-. Note the orthographic ambiguity posed by the enclitic pronoun:
although there is no question as to its pronunciation, it is not clear
whether the ' should be construed as representing the verb desinence
or the diacritic that indicates the vocalic value of w.
154
88 wl'p yd hXyl'p
paletah de palo
'wooden spatula'. The modern term paleta, though akin to native pá
‹ LA, is in fact borrowed from It. paletta 'little shovel'. The
'shovel' < PA
form palo does not appear in Modern Portuguese, and is most likely a
latinizing spelling of ModPg. pau (cf. note line 170).
91 wd'Syg Syrybw' wmwq

komo uberes gisado
'as you (will) need'. Although occurring only once, this appears to be
an idiom aver gisado (in future subjunctive, ModPg. houveres), along the
lines of ter mester 'need', rather than a past-tense form of guisar (based
on guisa 'manner', based on a Germanic loanword *wisa).
94 yXr'p 'sryX'
a.terça parte
'the third part'. The feminine definite article is often procliticized
graphically, probably due to its similarity with the Hebrew h ha-, which
is obligatorily written adjoined to its noun. This shorter form of the
ordinal number (from TERTIA, vs. ModPg. terceira < TERTIA ‹ RIA) can still
be found in some fixed expressions, e.g. terça-feira 'Tuesday'. Corre
(n.d.) has in fact suggested that the Portuguese practice of naming the
days of the week as a sequence of ordinals is due to the influence of
Hebrew, through the usage of New Christians in post-1497 Portugal.
95 ry#b'S' Sy#byd yXy'

ite de‹bes asa‹ber
'likewise you should know'. ite < ITEM is a Latinism that occurs
throughout the text (as with ides noted above, the final consonant, if it
was indeed pronounced, is not written). Despite ModPg. saber <
SAPERE, the diacritic on b seems to indicate that [b] or [v] is intended,
and may represent another Castilianism. Hauy (1989) in fact cites
devees de ssaber [sic] as a recurrent syntactic formula, so it is possible that
the prosthetic a- has a similar function here.
97 rybwXSy'
estuber
'stand'. This future subjunctive form of estar (which also occurs
transitively throughout the text in the sense of 'let stand, leave') does
155
not derive from classical STETERIT, but is based rather on a stem formed
analogically from haver (Penny 1991). N o t e the ambiguity in the
spelling of the vowel in the second syllable (cf. OSp. preterite estove).
yXnym hd'gnwl
longadah mente
'for a long time'. The first element clearly derives from LONGU (in fact,
an alleged *LONGATA < *LONGARE), for which Recuero (1977) does list a
variant longo alongside luengo 'long' in Judeo-Spanish. Note the
orthographically separate -mente.
f. 5v.
yS'nrwX
tornase
'become'. Though it appears correctly as a future-tense form in his
translation, the second <s> in Blondheim's transcription of this verb as
tornas[s]e implies that he considers it an imperfect subjunctive
(presumably derived from the classical pluperfect subj. TORNAVISSET).
99 S'Xr'dy#b'q
ka‹bidar-t-as
'(you will) be aware of (it)'. Although this verb means what modern
cuidar does, it does not derive from COGITARE like the modern verb, but
rather from an iterative *CAVITARE, derived from an original participle
*CAVITU (from CAVERE ‘beware, guard’). Machado (1967), by contrast,
considers the syncopated classical participle CAUTU as the innovated
form with respect to the participial source of the verb used here.
100 w'wb y' ... yS' Sy' yq w' y'

e o ke es asi … e bou
Although Blondheim transcribes the 3rd sg. copula with an acute accent
as <é> to distinguish this y' from the homographic conjunction e, I have
not added this layer of interpretation to my transcription, since no such
distinction is made in the Hebrew-letter manuscript. With regard to the
copula, Williams (1962) states that the -s (< EST ) fell "because of the
relation of this form to the 2nd singular es by analogy with the relation
between 3rd and 2nd singular of the present tense of all other verbs."
Since Sy' es here is unlikely to be 2nd person, this conservative variant
could be construed as a synchronic allomorph used before vowels.
156
wy'drq wmwq
komo kardeo
'like a thistle'. ModPg. cardo < CARDUU, with early vowel coalescence in
the final syllable. The y y in the spelling, along with Blondheim’s
translation as 'violet', suggest that this is an adjective, perhaps <
* CARDEU 'of thistle' on the model of vidro/vitreo ( VITRUM / VITREUS)
'glass/vitreous' (note that the text does feature wd'rdyw vidrado 'glazed'
< *VITRATU, ostensibly based on a verb *VITRARE 'glaze').
101 Syryswnwq w' 'r'p

para o konoçeres
'so that you recognize it'. This is the so-called inflected infinitive,
ModPg. conoceres < imperfect subjunctive C‹OGN‹OSCER‹ES.
103 wSw' yd' wmwq

komo a.de uso
'as there usually is'. The verb a 'there is' appears joined to the
preposition in the adverbial phrase. Note, however, that the a- in a
similar structure in line 111 yXr'pyd' adeparte cannot be considered a
form of aver.
104 Sydyw S'd 'dh'wq yd

de koahda da vides
'lye of vine branches'. One of only two occurrences of non-final h h,
though it is most likely due to the scribe having revised what he first
wrote as h'wq koa (based on coar 'strain' < COLARE ) to more correctly
write the intended 'd'wq koada < COLATA. Blondheim translates this
term as 'lye' (ModPg. lixívia, which does not occur in the text), a strong
alkaline solution leached from wood ashes.
107 SyryySyq yq w' ywyrqSy'

eskreve o ke kiseires
'write whatever you wish'. Blondheim translate this as 'write wherever
you please', interpreting the relative pronoun as u < UBI (cf. Fr. où). This
form has been replaced in the modern language by onde (see note line
61), and though it does occur elsewhere in the text, the presence of the
ke following makes that interpretation unlikely. Note the spelling of
eskreve with an unetymological w for /v/ < Lat. /b/
157
f. 6r.
111 wdw'ym w''p'r

rapa-o miudo
'scrape it until very fine'. ModPg. miúdo < MIN‹UTU used adverbially.
yXr'pyd' wnwp
pon-o adeparte
'put it aside'. As opposed to the structure in line 102, this ade- cannt be
construed as a form of aver, and may represent a particle of some kind
in this adverbial structure.
113 ymw' 'rdyp

pedra ume
'alum'. ModPg. alume has been relatinized based on AL‹UMEN. The
medieval forms alumbre and ahume cited by da Cunha (1982) show two
possible outcomes of the development of the word after syncope of the
second syllable, and both are found in the text. The former, however,
must be considered a Castilianism, since the /mn/ clusters resulting
from syncope were not subject to epenthesis in Portuguese (e.g. LUMINE
> Sp. lumbre, Pg. lume). Although the modern form is masculine (as in
line 179), it is possible that the form here is the result of a reanalysis of
ahume as a feminine form plus the corresponding definite article.
114 hgyl'm
malegah
'bowl'. Blondheim notes that "in modern Portuguese malega means a
soup-bowl such as is used by country people. Apparently here it is
used as a name of the wood of which such as bowl is made" (1930: 123).
118 yrq 'rdyp

pedra kri
'chalk'. Probably borrowed from French (modern craie < CRETA), since
the complete lenition of the /t/ in the Latin term would not be expected
as the native outcome. The synonym ¶yg giß (ModPg. giz < GYPSUM )
occurs just before, and is the more frequent term used in the text.
158
h'yp 'n ry#byb w''#Syl

le‹sa-o bi‹bir na piah
'leave it soak in a trough'. ModPg. deixar, the product of an apparently
spontaneous change l > d in leixar < LAX‹ARE (also found in the ModSp.
cognate dejar), also occurs in the text. For 'soak', the modern language
uses the prefixed form embeber < IN+BIBERE. ModPg. pia 'sink, trough' <
PI‹ LA, with the regular deletion of intervocalic /l/.
f. 6v.
123 w' 'p'rh'

ahrapa o
'scrape it'. Non-final h is extremely rare in all Judeo-Romance writing,
used almost exclusively as an allograph for ' in final position. Its
presence here is probably due to the scribe having erroneously written
the definite article h' a (cf. note line 104).
'yylw' hnw'
unah olya
'a pot'. The native doublet ola 'clay pot' < OLLA of this Castilian
loanword does not occur in the text (Ferreira 1999 lists it as a now-
archaic regionalism). Note the of n n in u nah, which like the <m> in
ModPg. uma may simply be due to orthographic convention.
128 w'rwXSym
mistura-o
'mix it'. This word alternates through the text with its near-doublet
meçer < MISC E‹ RE , whose participle MIXTUS served as the basis for the
noun MIXTURA that gave rise to the denominative misturar used here.
Both of the Portuguese verbs are often spelled with the sibilant s, since
in both cases the sound is not the result of an inherited Latin /s/. A
third term used occasionally is meçkrar/mezkrar (ModPg. mescrar <
MISCULA‹ RE; cf. ModIt. mescolare). Note how the use of ' here may be
considered haplological, standing for both the /a/ of the verb
desinence as well as the diacritic before the w of the enclitic object
pronoun /o/.
159
f. 7r.
130 ydl'yywl'
alvayalde
'white lead'. The modern form alvaiade < Ar. al-bay∂ does not contain
the second /l/. Corriente (1992: 50) suggests that the liquid in this and
several other words (e.g. alcalde 'judge, governor' < al-q∂i and arrabalde
'suburb' < ar-rab∂) may reflect the preservation in early Andalusian
Arabic of the lateral feature from older ÿ ∂, a segment known in Semitic
philology as an "emphatic lateral" (see Steiner 1977; it may also be
transcribed as <÷> or <s≥'>). The lateral feature is also seen preserved in
Modern South Yemenite Arabic abya‚l 'white'.
131 'ly#gyX
ti‹gela
'bowl'. ModPg. tigela < *TEGELLA < TEG‹ILLUM 'head-covering, straw
bonnet', based on TEGULA 'tile', which more directly yields the doublet
telha (Houaiss 2001).
134 w''qrz'
azarkao
'zircon, red lead'. ModPg. zarcão < Ar. zarq‹un. The term also occurs at
the start of this short chapter with its nasal consonant preserved in the
spelling, i.e. §wqrz' azarkon. Yet the fact that it occurs with only vowel
letters in the last syllable suggests that the nasal consonant had
generally disappeared from speech (a number of other words feature
this same variation, often in close proximity to one another; cf. note line
189). The initial a- may in this case reflect the preservation of the Arabic
definite article, which assimilates fully to coronal consonants (although
the word does occur elsewhere without the initial vowel).
135 yrb'yynyz'
azinyabre
'verdigris (copper oxide)'. ModPg. azinhavre < Ar. az-zin¸g‹ar, itself
borrowed from Persian zengir. Also expressed in the modern language
by verdete, a loanword from French. Both da Cunha (1982) and Houaiss
(2001) give the source as Ar. a z - z i n ¸ g a f r, which includes an
unetymological /f/ in an attempt to account for /v/ (here spelled b b).
160
137 hd'Xswq'
akoçtadah
'on its side'. Although the stem costa appears in many forms in the
modern language, the sense of 'leaning, lying' has been taken over by
other words (e.g. inclinar-se). Unlike some of the other instances, the
prosthetic a- here quite likely derives from an etymological prefix AD-.
138 wqryXSy' w' §wS h'ryXwS

soterah son o estirko
'bury (it) under the manure'. Blondheim misconstrues soterah son as a
single word that has been misdivided by the scribe, i.e. a noun based on
SUBTERRATIONE , which he translates as 'burial earth'. If this were the
case, however, we would expect the nominal suffix to be spelled with s.
In fact this is the same preposition attested in the Brotherton Passover
text (ch. 6 § 3), namely OPg. so < SUB with an -n added by analogy with
non, nin, etc.
f. 7v.
142 S'wl'ym
mea-lo-as
'half-fill', with the object pronoun interposed between the stem and
desinence. ModPg. mear < MEDI‹ARE. Both da Cunha (1982) and Ferreira
(1999) refer to this verb as a "popular form" as compared to ModPg.
mediar, and in fact the former is the expected outcome, while the latter is
a conservative form or more recent borrowing (it first appears in the
fifteenth century, spelled <medear>).
143 Syrdwp §yyb S'nyrw'

urinas bein pudres
'very putrid urine'. The form pudre, deriving from the basic adjective
PUTER , PUTRIS has been replaced in the modern language by pútrido,
which is based on the participle of the verb formed from the same basic
adjective, PUTRERE 'become rotten'.
§wX'l yd wy's'b
baçio de latun
'brass basin'. ModPg. bacio < *BACC‹INUM < Gk. bacchinon, though da
Cunha (1982) cites a feminine form in -a from the fourteenth century.
161
Although never spelled with initial w v, this word also occurs in the text
with the diacritic on b b which presumably indicates a fricative
pronunciation. ModPg. latão < OFr. laton < Ar. l‹a†‹un 'copper', itself
probably borrowed from a Turkic language (cf. Turkish altyn 'gold').
147 'qwrbny'
enbroka
'place upside down'. Although this term is not preserved in the
modern language, cf. ModSp. embroca 'poultice' < E M B R O C H A , a
loanword from Greek.
148 Swml'#Sny'
en‹salmos
'coverings for pack mules'. ModPg. enxalmo is borrowed from Sp.
enjalma < OSp. salma < SAGMA, a Greek loanword (Ferreira 1999).
wdyw w' S'ryXryw

vertiras o vido
'pour the urine'. Note that the term vido, along with its synonyms
me‹gados and urinas normally occurs in the plural (see note line 265).
150 yrb'nyz'
azinabre
'verdigris'. This variant lacks any indication of a palatal /µ/ (cf. note
line 135), a variant not mentioned in the etymological dictionaries. It
may, of course, be simply due to a scribal error omitting the double-yy
following n.
f. 8r.
155 rwlwq' ' ¶yrp 'yylwX yylyq wl#b'yd h' w''n

nao ah dia‹blo kelye tolya preß a akolor
'there is no devil who [could] then take the color'. Blondheim takes preß
to mean 'worth' (as in ModPg. preço < PRETIU ). Yet the word is more
likely akin to ModFr. après 'after', which derives from the expression AD
PRESSUM 'near', used here without the prepositional element. The
diacritic in the spelling of 'devil' (ModPg. diabo < DIABOLUM) shows the
vernacular lenition of /b/ to /v/, but not the deletion of /l/ (cf. povo <
poboo < POPULU), just the opposite of the modern semi-learned form.
162
158 h'wg' yd Syrbmws' wrX'wq

kuatro açumbres de aguah
'four açumbres of water'. ModPg. azumbre < Ar. a±-±umn 'eighth (part)'
(based on ±am‹aniya 'eight'), probably a Castilian borrowing as signalled
by the epenthesis of the mn cluster. According to Houaiss (2001), this
liquid measure used in Moorish Spain was equivalent to two litres and
16 millilitres, while Corominas (1980) describes it as "la octava parte de
una cántara."
160 'mwqSy' §yr'Xn'byl

lebantaren eskuma
'give off foam'. As opposed to ModPg. espuma < SP‹UMA, this word is
based on MedL. schuma, a loanword from Frankish *sk‹um (the ultimate
source of English scum). See also eskumado 'skimmed' in the same line,
which Blondheim translates more wordily as 'no longer gives off foam'.
f. 8v.
161 w'yXnys yd 'yyl'p

palya de çenteo
'rye-straw'. ModPg. centeio < CENT‹ENU, so called because each plant
produced 100 grains per stalk (da Cunha 1982). The spelling of the
penultimate syllable with a single y y likely indicates a monophthong, in
contrast to the diphthong found in the modern form.
163 Sydyw yd 'snyz

zinça de vides
'ashes of vine branches'. Although the orthography of sibilants is quite
variable in this text, given earlier 'znys çinza it is difficult to see this
"metathesis" as anything other than a scribal error.
164 hlyr'b
barilah
'lye ashes'. ModPg. barrilha < Sp. barrilla, with the expected palatal
segment not spelled as such (cf. note line 150), so perhaps this is an
archaic native form. Although it is the only occurrence of this term, in
context it refers to an item already known to the reader. Note that a
form akin to the modern term lixívia 'lye' does not occur in the text.
163
165 wdn'wq wyy'ww

vayo koando
'continue straining it'. An idiom (see also note line 181) using a form
derived from VADERE (whose conjugated forms become associated with
the infinitive ‹IRE 'go') and a present participle, in this case of ModPg.
coar < COLARE.
169 'd'q yd wmwq

komo de kada
'as of each'. Blondheim reads this phrase as a scribal error for an
intended kon a-dekoada 'with the strained [item]', a reference to the lye
mixture.
170 wd'qrw#p w''p

pao forkado
'pitchfork' (lit. 'forked stick'). ModPg. pau < PALU 'stake, post', although
a latinizing form palo occurs elsewhere (e.g. line 88). ModPg. forcado
denotes 'pitchfork' on its own.
171 'dwXyryd
deretuda
'melted' (also in line 173). Although the participle of ModPg. derreter is
the regularized derretido, this form shows the Late Latin tendency to
form participles in -UTU.
f. 9r.
174 yEXS˚p
fuste
'(wood) stick' < F ‹U STIS . One of the handful of words with diacritic
vowels indicated, perhaps because it has a more technical sense or
specific reference here than in its other occurrences, or perhaps because
the scribe sees it as a Latinism along the lines of ides 'that is', which is
similarly vocalized.
175 lwr#p
frol
'powder'. Apparently a metaphorical extension based on its principal
sense of 'flour' (which has been taken over by ModPg. farinha), with l-r
metathesis (cf. chapter 7 § 2.3).
164
177 '#byX y' 'pnyl

linpa e te‹ba
'clean and tepid.' Like ModPg. pútrido (see note line 143), ModPg. tépido
has been relatinized based on Latin TEPIDUS (itself derived from the
noun TEPOR , -‹ORIS 'gentle warmth'). The form used here shows the
vernacular outcome with deletion of intervocalic /d/.
179 ymw''
aume
'alum (water)', here showing the intermediate form with the
intervocalic /l/ deleted but the initial syllable preserved (cf. ume earlier
in the same line, and note line 113).
180 w'y#blwb
bol‹be-o
'stir it'. ModPg. volver < VOLVERE '(cause to) turn', whose derivative
*VOLVITARE yields the more common modern term voltar 'return'. Note
that the diacritic indicating the fricative form of b b is used only on the
second of the two, though both are unetymologically representing /v/
< Latin /w/. Indeed, the use of b is the strategy used frequently to
avoid the ambiguous double-ww spelling in spelling /vo/.
181 r'qy#pyr'rq
krarifikar
'clarify, purify'. Like other forms with initial /kr/, this word alternates
with a conservative, or more likely remade form with initial /kl/, e.g.
S'd'qy&pyr'lq klarifiakadas in line 169 above.
wdnyXryw y'ww
vae vertindo
'continue pouring'. The spelling of the form derived from V A D I T
(ModPg. vá 'go') in this expression is conservative enough to indicate
the hiatus from the lenited /d/ and its later-deleted syllable.
183 Swq'S SwyynySny'

ensenyos sakos
'in separate bags'. ModPg. senhos < SINGULOS , which also yields the
doublet sendos.
165
Sw''glwq
kolga-os
'hang them' < COLLOCARE (cf. Sp. colgar). The modern language uses
pendurar 'hang' < *PENDULARE, a verbal diminutive based on PENDERE.
f. 9v.
186 Swsn'ww ry' wmwq Swyyn yryywlyp

pelveiri nyos komo er vanços
'small particles like chick peas'. ModPg. pó 'dust, powder' < *pulu <
VLat. *PULVUS < classical PULVIS , -ERIS . The ultimate source of the
diminutive form used here is the classical plural PULVERA , which da
Cunha (1982) claims was borrowed in Portuguese from Castilian
pólvora, itself borrowed from Catalan.
187 §'b'S h'w'

uah saban
'a towel' < SABANA, plural of SABANU , a Greek loanword. Note that it
remains feminine here despite the apocope of the final vowel.
189 Syryz'#p 'r'p

para fazeres
'in order to make'. Another inflected infinitive, derived from the Latin
imperfect subjunctive (in this case < FACER‹ES).
§wyylymryw
vermelyon
'red'. ModPg. vermelho < VERMICULU (diminutive of VERMIS 'worm') is
the basic term for this color. The form here is one of several variant
spellings that occur in this chapter: w'wyylymryb bermelyou (line 203),
w''lymryb bermelao (line 207). The latter is especially curious, since it
does not contain the double-yy indicating palatal l, but does seem to
indicate a nasalized vowel in final position, albeit a different one from
the preceding form in line 189. It is likely that the nasal ending reveals
those forms to be borrowed from OFr. verm(e)illon, whereby that form
of the word refers to a more specific shade of red, namely 'vermilion'.
Note that in reference to the color of the smoke, as opposed to the
substance being produced, the form used is wyylymryw vermelyo.
166
192 w#byw §y#gr'

ar‹gen vi‹bo
'quicksilver'. ModPg. mercúrio or azougue (only the latter is found in the
text). For silver itself, Modern Portuguese uses prata 'silver' (which also
occurs in this text). It has not retained a direct reflex of the term used
here (which also occurs in line 386), but only a more recent formation
argento-vivo < ARGENTUM VIVUM (cf. argente vivo in O libro de ma‹gika).
f. 10r.
196 wSw#g yd S'#gn'

an‹gas de ‹guso
'broad below'. ModPg. amplo is the relatinized doublet of ModPg. ancho
< AMPLU. The modern language has not preserved ‹guso < DE ORSU (but
cf. OIt. giuso, ModIt. giù 'down').
201 'Swq
kosa
'thing'. ModPg. coisa < CAUSA . Although the modern language has
chosen the <oi> variant, forms with the diphthong <ou> occur
throughout the medieval language. Indeed, O libro de ma‹gika contains
numerous occurrences of hSw'wq/'Sw'wq kousah/kousa. What is rarer is a
form such as this Spanish-like variant lacking the diphthong altogether,
again pointing to Castilian interference in the text's transmission.
205 wgw#p w' 'Syg l'Xny' Syryylyq

kileires ental gisa o fogo
'reduce the fire in this manner'. If the verb here is akin to ModPg.
quilher < *(TRAN)QUILLARE 'to calm', it may in fact be a Castilianism,
since cases of geminate -LL - appearing as -lh- in Portuguese are
considered to be borrowed (Williams 1962: 74; also cf. note line 123).
Although fogo < FOCU occurs as expected, a non-Portuguese diphthong
is clearly spelled a few words onward in wgy'w#p fuego.
167
f. 10v.
209 y' wXyd wmwq

komo dito e
'as has been mentioned', with e 'is' < EST. Blondheim's translation 'as I
have mentioned' assumes that y' e < HABEO (ModPg. hei), and that this is
a present perfect form of dizer. Yet this would constitute the only
instance in the corpus of a modern-style periphrastic tense (cf. chapter 7
§ 1.2). The phrase is more likely a syntactic Latinism on the model of
DICTUM EST, which carries a past rather than present sense.
210 SyXnym §y'yb 'r'p

para been mentes
'pay close attention'. As noted above, y'y y√y may be intended to
indicate a diphthong, i.e. non-Portuguese bien. There are several other
spellings that may also be interpreted as similar non-Portuguese
dipththongs, especially those that alternate between monopthongs and
diphthongs (e.g. fogo~fuego above), which may be due to the influence
of a Spanish copyist at some stage of the text's transmission. The
expression here is akin to Spanish parar la atención 'fix (one’s) attention'
(with parar < PARARE 'stop; prepare'), though it is also worth noting the
plural use of mentes 'minds'.
S'r'ry' Sw' w'qnwn

nunkao os eraras
'you will never ruin anything'. ModPg. nunca < NUMQUAM. The final w'
-ao here may be a hypercorrect spelling for a nasal segment that had
deleted well before nasalized vowels had emerged. Blondheim’s
translation of os eraras (lit. 'ruin them') as an intransitive 'go wrong'
must be due to English influence, since an overt object pronoun
precedes the verb, which is also used transitively elsewhere (e.g. line 6
S'r'ry' wdwX todo eraras 'you will ruin everything').
211 ws'rb yd w''#br'q wd s'syb'q

kabeçaç do kar‹bao de braço
'ends of burnt charcoal'. Blondheim translates braço as 'branches',
obviously seeing it as a metaphorical use of 'arm' < BRACCHIU.
However, the term is more likely akin to French braise or Italian brace
'embers' (cf. chapter 6 § 4.4), borrowed from West Gmc. *brasa. The
168
sense of ModPg. carvão < CARB ‹ONE has narrowed to 'coal', but here it
indicates the "carbonized" charcoal. Note also the very uncharacteristic
use of s ç for the plural morpheme -s, no doubt due to its presence
earlier in the word.
215 yyl'sn'l
lança-lye
'toss (in) to it'. ModPg. lançar 'throw' < LANCE ‹A RE , based on LANCEA
'spear'. This word occurs frequently in the text, meaning 'add' here
(hence the dative), and simply 'put' or 'place' in other cases.
yr#b'yynyz 'd . . . r'X'#Swn 'd

da no‹satar … da zinya‹bre
'of sal ammoniac … of verdigris' From the perspective of ModPg.
azinhavre, this would appear to be an instance of incorrect word
division. However, it is also possible that this division reflects the
scribe's awareness of the Arabic etymologies of these two words, since
the a- in both cases derives from the Arabic definite article.
217 'd'rnys hXSyd §y'wp

poen destah çinrada
'place (some) of these ashes'. The form related to ModPg. cinza 'ash' is a
participle CINERATA (used in a collective sense), based on CINERARE.
218 S'rXyl ryz'#p Syry#gyq w'

u ki‹geres fazer letras
'where you wish to make letters'. The #g ‹g spelling for the stem-final
consonant of the future subjunctive of this verb (ModPg. quiseres <
Q U A E S I V E R I S , after shift to the second conjugation), alternates
throughout the text with analogical S s (as well as #S on occasion).
f. 11r.
219 yilwbÕna'wqËr¬za' §wq y' ylwb 'ry≈gyÃw §wq

kon vegera boli e kon azarkoanboli
Appearing with diacritic vowels added, these two terms remain
untranslated and of unknown origin.
169
220 Sypnyl' w''nyq

ke.nao alinpes
'that you not clean (it)'. This initial a- appears on linpar repeatedly in
the text on this and other verbs. Although the verb is lacking a clause-
internal object, it is unlikely that this prefix is a clitic object, given that a
number of non-phonological processes could give rise to prosthetic a
(cf. Williams 1962: 112-113).
221 'qy#p 'd'rnw' '#g 'q

ka ‹ga onrada fika
'for it will remain adorned'. onrada is most likely a misspelling of ornada
'adorned' < ORNATA, the same confusion of letters as that seen in
wrny#bny' in‹benro 'winter' (ModPg. inverno < HIBERNU ) one line above.
Curiously, it is the exact counterpart to the opposite error in O libro de
ma‹gika, where 'dnrw' ornada is the form spelled for 'honored' (ModPg.
honrado < HONORATU).
224 wryd'm
madero
'wood'. The modern form madeira is feminine, deriving from the neuter
plural MATERIA. The form here is masculine, deriving from the singular
form MATERIU. Note also the lack of diphthong in the spelling (cf. note
line 161).
225 yXy#byrql'd Sw'wp

pous dalkre‹bite
'powdered sulphur'. Pg. alcrevite < Ar. al-kibrit (now archaic according
to Franca 1994), with the characteristic r-migration seen in other
Portuguese words, e.g. trevas 'darkness' < TENEBRAS, preguiça 'laziness' <
PIGRITIA. This loanword also appears (via Turkish) in Romanian chibrit
'match'. In this text it occasionally replaces the native yr#pw#Sny' en‹sofre
(ModPg. enxofre, from enxofrar < *INSULFRARE).
f. 11v.
229 r'q'l'
alakar
'dark red resin'. ModPg. laca < Ar. lakk, ultimately Sanskrit lkß (cf.
English shellac). The OED cites lacquer as borrowed from the alternative
170
Portuguese form lacre, similar to the form used here, which Silvera
Bueno (1966) suggests may have arisen by r-metathesis in yet another
attested variant, lácar.
ryXSy'ym Syry#bw' wXn'wq

kuanto u‹beres meester
'as much as you need'. Fut. subj. houveres < HABUERIS, used as a main
verb of possession with object meester < MINISTERIU (with a vowel hiatus
clearly indicated by the spelling y'y ), in the sense of Fr. avoir besoin
(ModPg. ter mister 'need, must', also expressable by haver de). Both
expressions may be considered syntactic Latinisms, in the sense of OPUS
EST 'must be done'.
231 w''syy#p hXSy'

estah feiçao
'this mixture'. An unprefixed form based on FACTI‹ONE, but used in the
sense of §wys'#pnwq konfaçion 'concoction' that occurs elsewhere.
234 y#gywyz'
azevi‹ge
'jet black'. Both da Cunha (1982) and Houaiss (2001) cite as the etymon
of ModPg. azeviche the specifically Andalusian Arabic form az-zabı¸g. By
contrast, Penny (1991) derives the Castilian cognate azabache 'jet' from
classical Arabic az-zab¸g. Since this text tends to have /a/ where /e/ is
expected and not vice-versa, the classical Arabic form is probably the
immediate source of the form used here.
235 wrw' yd Syzy#p y' yq'Xr'ml'

almartake e fezes de oro
'litharge of silver (lead monoxide) and litharge of gold'. Corriente
(1999) cites the Spanish pair almárta/almartega 'litharge' as deriving from
Andalusian Ar. al-mártaq < classical Ar. martak (the word may be
related to ModPg. almadraque 'pillow, mattress' < Ar. al-ma†ra˛ 'rug',
with r-migration in the form here). Replaced in the modern lanague by
litargírio, a Greek loanword. I have followed Blondeheim in translating
fizes as 'litharge' as well, though it may be akin to ModPg. fezes 'dregs' <
FAECES, plural of FAEX.
171
236 y#pynr'z'
azarnefe
'arsenic'. ModPg. arsênico is the more recently-formed doublet of
arzenefe < Ar. az-zirnı, both based on Gk. arsenikón, though Corriente
(1999) lists classical Arabic zirnıq (with the Andalusian variant azzirníx)
as deriving from Pahlavi zarr n˙k 'fine gold'. The form here shows r-
migration relative to the Greek form, though Ferreira (1999) lists
arzenefe as a variant, which itself shows r-migration relative to the
Arabic and Pahlavi sources.
238 y#gyz'
aze‹ge
'vitriol' (an acid of metal sulfate). ModPg. azeche 'iron sulfate' <
Andalusian Ar. azzá¸g < classical Ar. z‹a¸g (Corriente 1999). This term has
largely been supplemented by the Latinate vitríolo.
f. 12r.
243 Swgn'm
mangos
'handles'. A masculine or neuter form *MANICU (> It. manico) based
ultimately on MANUS 'hand', which survives only as the term for a
component of a cereal thresher (the fruit name mango is a Spanish
loanword; the feminine MANICA does yields ModPg. manga 'sleeve'). In
translating it as 'stick', Blondheim seems to have misconstrued it as a
figurative reference to the stirring device.
248 yrbwq yd 'ryyd'q

kadeira de kobre
'copper kettle'. While the modern language uses chaleira '(tea)pot', the
term here is akin to Sp. caldera < CALDARIA , based on CALIDUS 'hot',
(though the scribe has apparently forgotten its letter l l). ModPg. cobre
derives from a popular and abbreviated variant of classical AES CYPRIUM
'Cyprus bronze' (da Cunha 1982)
250 syrdy#S' Syr'#p 'ryp

pera fares a‹sedreç
'to make chess pieces'. An apparently syncopated form of the inflected
infinitive form of fazer (ModPg. fazeres). ModPg. xadrez < Ar. a¸s-¸sitran¸g
(via Persian, ultimately from Sanskrit). Like ModSp. ajedrez, the form
172
here has kept the initial a- of the Arabic definite article. Note that like
the modern spelling, none of the sibilants are spelled using the
unmarked letter S, indicating that the final sibilant was not reanalyzed
as a plural morpheme.
f. 12v.
253 ly#pr'm yd SwSw'

osos de marfil
'elephant bones'. ModPg. marfim, a loanword basd on Ar. (fia÷m) al-fil
'elephant (bones)', refers only to 'ivory', having been replaced by
elefante. It is the animal rather than the material that is the likely
referent here, given that such bones are offered as an alternative to
'q'ww yd SwSw' osos de vaka 'cow bones' in line 252.
254 wryyd'm yd S'rwd'ryS

seraduras de madeiro
'sawdust', lit. 'wood sawings'. Note again the masculine gender of
madeiro (vs. ModPg. madeira).
257 'dyd'p l'd

dal fadida
'(some) cuprous oxide'. Corriente (1999) lists the cognates Cat. alfadida
and Sp. alhadida as derived from an Andalusian variant of classical Ar.
˛adıdah 'piece of iron' (cf. Moroccan Ar. ˛dida ˛¥mra 'copper oxide'), used
here as a synonym of yrb'yynyz' azinyabre 'verdigris'.
261 yy'ydnw#p yS
se fundey
'melt'. ModPg. fundir < FUNDERE . A subjunctive form w''dnwp yS se
fundao occurs earlier in the following line, but the extra syllable here
does not appear to have any motivation in this present-tense form.
wbnw#g wmwq §yyl'wq

koalyen komo ‹gunbo
'coagulate like lead'. ModPg. coagular is a relatinized doublet of the
coalhar (used here in fut. subj.), whose regular development is shown by
the vocalization of the -GL- cluster that resulted from syncope.
173
263 Sw'gr
rgaos
'grain'. It is difficult to see this as anything other than a confusion of the
first two consonants in the ModPg. grão 'seed' < GR‹ANU. Linguistically-
unmotivated errors such as these (cf. note line 123) could suggest that
the scribe is not transliterating the text for the first time but is working
from a pre-existing Hebrew-letter manuscript.
f. 13r.
265 wXSyX w' r'dygl'

algidar o testo
'bowl or vessel'. Because the Hebrew-letter manuscript contains little
punctuation, it is unclear whether the native term testo < TESTU is a gloss
for or a practical alternative to the borrowed algidar < Ar. al-¸gidr.
Sy'ymwd Swdyw
vidos domees
'human urine'. For vidos (lit. 'evacuates') cf. ModFr. vuide < *VOCITU
from classical VOCIVUS/VACIVUS , an adjective that does yield ModPg.
vazio 'empty'. The more frequent term in this text is S'nyrw' urinas,
though Swd'#gym me‹gados (ModPg. mijada).
270 SwSwgydy#p Sy'm

mais fedegosos
'more fetid'. ModPg. fedegoso < *feticosu < *F ( O ) ETICU , a variant of
FOETIDU 'of ill smell', the adjective that yields ModPg. fétido. Both are
Vulgar Latin forms derived from the basic adjective FOEDU 'filthy, ugly'.
w''r'#p rwlwq rwyylym wXn'wq

kuanto melyor kolor farao
'so much better the color will (the urine) make'. The d d in kuando in
line 250 immediately above is best interpreted as a scribal error for
what should be wXn'wq kuanto, so that the phrase is an equivalence with
the kuanto here: "the more fetid and putrid the urine..."
174
271 Swdyw Swd w#p'b

bafo dos vidos
'vapor from the urine'. Most dictionaries claim an onomatopaeic origin
for bafo 'breath', with the earliest citation given as a sixteenth-century
attestation of abafo (da Cunha 1982). A form of this word also occurs as
a verb (y&gyp'b bafe‹ge) in line 365.
272 wd'rwm
morado
'mulberry'. The Latin name for this fruit was M‹ORUM, but the adjective
here derives from *MORATU (cf. ModSp. mora 'mulberry', morado 'purple;
bruise [n.]'). The modern Portuguese form amora derives from the basic
Latin form, with accretion of the definite article.
f. 13v.
275 S'ryywSyX
tesoyras
'scissors'. ModPg. tesoura < T‹ONS‹ORIA. Like koisa above, the form here
shows an alternate spelling of the reflex of Latin /‹o/ as compared to
the modern orthography. Indeed, the development of r + yod has led
to a number of similar alternants, e.g. CORIUM > coiro/couro 'leather',
AUGURIUM > agoiro/agouro 'foreboding' (cf. chapter 7 § 2.5). The
unexpected y e in the first syllable may be due to contamination from
tesouro 'treasure' < THESAURU (a Greek loanword), which may have
exhibited the same oi/ou variation as well.
279 wd'Xwb
botado
'spoiled'. ModPg. botar 'strike, push' < OFr. boter (ModFr. bouter 'expel,
push'), borrowed from Frankish *botan. Used here with a slight
extension of meaning, in that the color is "pushed out."
281 r'rqsym' y' ryz'#p' §ywyd yS wmwq

komo se deven afazer e ameçkrar
'how they should be made and mixed'. Note the placement of the
reflexive pronoun before deven, indicating that this verb may have at
least transitory status as an auxiliary. Note also the prefixed a-, which
175
does not otherwise occur on these verbs, but given its frequency in the
text may well represent a complementizing particle.
285 §yyb ywm y' wXyywm

muito e mui bein
'thoroughly and very well'. ModPg. muito < MULTU , also the source of
the apocopated form, which has disappeared from the language (cf.
ModSp. muy 'very'), but which is the preferred form in the Judeo-
Portuguese corpus and is found frequently in medieval Portuguese.
f. 14r.
288 wnym'gryp
pergamino
'parchment'. ModPg. pergaminho < PERGAM‹INU shows the palatal /µ/
that developed in similar contexts elsewhere (Williams 1962: 71-73).
The form used here does not indicate any palatalization, perhaps due to
latinizing interference.
289 §w'ys'#p §wq

kon façion
'concoction'. Although there is no call for this word division, the final
form § n leaves no doubt that it was intentional on the part of the scribe.
293 §y#g'#bl'd
dal‹ba‹gen
'(some) egg-white, albumen'. ModPg. albume < ALBUMEN . Like the
classical and relatinized modern terms, the form used here is based on
the adjective ALBUS, -A, -UM 'white' > alvo, alva (cf. note line 315). It may
be a native development of *ALBAGINE (cf. imagem < IMA ‹ GINE, origem <
ORIGINE ). Like most ModPg. words in -agem, however, it may be
borrowed from French or Provençal (e.g. viagem 'voyage' < Pr. viatge <
VIA‹ TICU). Thus it could be a similar loanword based on *ALBATICU
'whiteness', or else an analogical extension of the borrowed suffix.
294 h'yml'X'
atalmiah
'ceramic bowl'. Corriente (1999) lists a variant Portuguese altamia
derived from Andalusian Ar. *˛altamíyya 'of ceramic' < classical Ar.
˛antam 'sealed'.
176
f. 14v.
303 hXyd
ditah
'mentioned'. Blondheim mistranscribes (but correctly translates) this
word as an indecipherable hXm <mth>.
305 h#gnwq
kon‹gah
'conch, shell'. Also mistranscribed (but correctly translated) by
Blondheim as h#grwq <qwr‹gh>.
318 SwXry#bwq SyrX

tres ko‹bertos
'well covered'. This adverb does not survive in Modern Portuguese (cf.
Fr. très 'very' < TRANS).
f. 15r.
313 h'ydny' rw'wq

koor indiah
'blue color'. The hiatus from deleted /l/ in color is fully spelled out
here. As a color term, Modern Portuguese has replaced native indio <
INDICUS with índigo, borrowed from Castilian (da Cunha 1982).
314 'rgyn '√sn«p

pinça negra
'black color'. One of the handful of words with vowel pointing,
perhaps because this is a more technically-specific term for 'color' that
the scribe deemed more obscure than the usual rw'wq koor.
315 'wl'
alva
'white' < ALBA. This word has been replaced in the modern language as
the basic term by branco, a loanword from Germanic. The masculine
form wwl' alvo occurs in the following line, and is one of a very few
forms in the corpus where double-ww serves to represent a CV syllable
177
rather than a single consonant (most of which are in this text, e.g. wwrys
çervo and r'dnww' avondar above).
317 Sy'pysnyrp Syrwq

kores prinçipais
'principal colors'. Note that many of the "basic" color terms that follow
are transparent loanwords, with only azul, vermelyon, and verde serving
as basic color terms in the modern language. It is possible, then, that
these are not meant as the basic words of the language per se, but rather
of the art or trade, or perhaps that they are the main ingredients for
making the dyes described in the text.
318 yy#pws §ymrq

karmen çufiy
As opposed its usual omission, the lack of orthographic /a/ in karmin
could be construed as an etymological spelling, since the term is
borrowed from Arabic, through MedL. CARMINUS , a contraction of
CARMESINUS, itself from Ar. qirmizı (cf. ModPg. carmesim 'crimson' <
Andalusian Ar. qarmazê). The word does, however, occur in line 175
with no nasal consonant in final position, an alternation similar to that
seen in ModPg. alume(n). As Blondheim notes, çufii makes eleven
terms, and most likely modifies karmen, possibly referring to a reddish
wool (Ar. suf 'wool') perhaps worn by members of the Sufi sect of Islam.
w''rp's'
açafrao
'saffron'. ModPg. açafrão < Ar. az-zafifarn (ultimately of Persian origin),
with agglutination of the Arabic definite article. As opposed to other
words with etymologically nasal endings, this term consistently
appears in this text, as in its modern form, with no final consonant.
319 w#bw' wd 'r'rq 'l

la krara do o‹bo
'the egg white'. Although it is the only occurrence of this un-
Portuguese l-form of the definite article, this could be taken as further
evidence Spanish interference in the text's transmission.
178
yyl' '#gy'
e‹ga alye
'pour (in)to it'. In contrast to ModSp. echar < IACT ‹ARE 'throw', a
frequentative based on IACERE, Modern Portuguese has not preserved a
reflex of this verb per se, though it is reflected in jeito 'manner, skill' <
IACTU as well as in the numerous words containing -je(c)t-.
320 'ryygy#p 'lyd y#gyl

le‹ge dela figeira
'sap of the fig-tree'. ModPg. leite < LACTE, used figuratively to describe
the white sap of this tree (ModPg. seiva). The spelling with g rather than
a diphthong suggests that the form may be another Castilianism (cf.
ModSp leche).
f. 15v.
322 w''rpyXSyd
destepera-o
'temper'. The form here lacks both an explicit /e/ vowel in the
penultimate syllable (unless the spelling intends to indicate a syncope),
as well as any indication of /n/ or nasalization before the /p/
(normally spelled with n n). It does, however, spell out the nasalized
verb desinence with vowel letters only.
327 yXyqn'rb
brankete
'ceruse, white lead pigment'. This term, clearly based on a form of
branko 'white', appears in the next chapter (line 331) spelled with #g
instead of q. If this form can be considered the "error," then it would
constitute the only instance in which q represents a fricative/affricate,
which is common practice in other medieval Judeo-Romance writing.
331 w'yl
leo
'Leo'. As indicated by Sydy' ides 'that is' < ID EST that precedes it, this is
a glosss for yXy#gn'rb bran‹gete (though Blondheim misconstrues it as a
verb plus clitic pronoun, i.e. 'thin it'), though it may also refer to gold,
the metal associated with this sign of the Zodiac. Note that unlike in O
179
libro de ma‹gika the word appears in a vernacular spelling with no nasal

consonant (§w'yl leon in O libro de ma‹gika, ModPg. leão).
332 yrXsylysny' rynrwX

tornar ençeleçtre
'to turn (it) sky-blue'. ModPg. celestial is a relatinized form of the word
used here to describe a shade of blue. The form here may show the l-r
dissimilation that affected words even in classical Latin such as SOLARIS
'solar' for *SOLALIS and REGULARIS 'regular' for *REGULALIS (but cf. filial <
FILIALIS), though it may also be formed by analogy with terrestre (cf.
Sy''yrXsylys çeleçtriais in O libro de ma‹gika, where the adjective suffix -al
is added to the analogically-formed stem).
f. 16r.
335 wbw' yd 'myg

‹gema de obo
'egg-yolk'. ModPg. gema < GEMMA 'bud; gem'.
342 Sy'm w'wn

nou mais
'no more'. ModPg. não is usually spelled in the corpus with a final
consonant (i.e. §wn non, though w''n nao does also occur), but here it is
actually written with a vowel spelling more typical of a form with a
hiatus from deleted /l/ or /n/, e.g. rw'wq koor 'color,' w'wb bou 'good'.
346 lyn' wrgyn

negro anil
'black indigo'. ModPg. anil < Ar. an-nıl (via Persian and ultimately from
Sanskrit nılı 'indigo').
f. 16v.
351 wyyl'zyX'm
matiza-ly-o
'shade it'. Though da Cunha (1982) states that ModPg. matizar is "de
origem incerta," Corominas (1980) claims that the word ultimately
derives from Greek lammatizein, based on lamma, which referred
originally to a belt and fringe of different colors.
180
353 r'rwlwq
kolorar
'(to) color'. Modern Portuguese has the doublets corar and colorir, the
former showing the regular development of COL ‹O R‹A RE . Da Cunha
(1982) lists colorir as first occurring in 1548 (probably based on an Italian
borrowing). The form used here is a semi-cultismo, retaining the -ar
conjugation of the evolved form but restoring the etymological /l/.
358 SdwX Syrwq S' wryp

pero as kores todas
'but all colors'. Spanish-like pero < PER HOC (attested with other uses in
medieval Portuguese), later replaced by mas < MAGIS.
f. 17r.
363 l'dnys
çendal
'bandage'. Ferreira (1999) cites ModPg. cendal 'fine cloth' (an alternative
here to pano) as a term of Arabic and Greek origin, derived via VLat.
CENDALU. Corriente (1999) suggests that it is based on Ar. ßandal, with
contamination from Lat. SUNDUS, itself based on Gk. sindon.
365 y#gy#p'b
bafe‹ge
'breathe'. Present subjunctive of a verb formed on the basis of bafo
'breath' + -ejar < -IZ‹ARE (later spelled -IDI‹ARE, where -DI- represented the
affricated segment), a Latin verb-forming suffix first used with verbs of
Greek origin, which also yields ModPg. verbs in -ear.
371 Syrw'wq S'd SwXnym'rqsym

meçkramentos das koores
'color mixtures'. The only occurrence of this alternative to misturas
based on meçkrar < MISCULARE, a verbal diminutive of MISCERE.
181
f. 17v.
375 r'ly#pryp
perfilar
'paint'. Based on OPr. perfil (ModPg. 'contour, profile'), this word seems
to have a more general meaning here than its etymological sense of
'(paint a) profile'.
379 wrbl
libro
'book'. One of the rare instance of an implicit vowel other than /a/, in
this case /i/ (cf. note 317).
'#gwq S'Xrq S'd h'wg'

aguah das kartas ko‹ga
'boiled parchment water'. kartas seems to be used in the sense of
pergaminos 'parchment leaves'. The adjective here is related to cozer
'boil' < *COCERE < COQUERE (a more regular past participle wdyzwq kozido
occurs in line 414) but appears to be based on an apocopated form.
381 ry'wp
poer
'put'. This infinitive is generally spelled with a monophthong, much
like its modern form pôr, but it is spelled here and one line further
down with the hiatus from the deleted /n/ spelled out.
382 y' yXym

mete i
'put there'. The use of this locative pronoun derived from IBI (> Fr. y)
has not been preserved in Modern Portuguese.
f. 18r.
388 'pnyl wn'm §wq wd'XnyqSy'

eskentado kon mano linpa
'warmed with a clean hand'. ModPg. aquecer '(to) heat < CAL‹ESCERE, an
inchoative based on CALE‹ RE 'be hot', which is also the ultimate source of
this form, based on ex + quente 'hot' < CALENTEM. Note the intervocalic n
in mano (ModPg. mão), a conservative or Castilian-style spelling.
182
389 'mysryp
perçima
'above'. A single occurrence of this alternate fomation based on PER+
CYMA (borrowed from Greek), rather than the usual ençima < IN+CYMA.
390 wp'rX
trapo
'cloth' < VLat. *DRAPPU (Sp. trapo, It. drappo, Fr. drap), probably of Celtic
origin. The provection of d- > t- in the Ibero-Romance forms may be
due to the absence of native Latin words with initial dr- (Houaiss 2001).
392 Sw'ww'l
lava-os
'wash them'. Occurring between two ', the double-ww spelling for Pg.
/v/ < Lat. /b/ may seem in this case unwarranted. Yet it does serve a
grapho-tactic purpose, allowing the scribe to avoid repeating the same
sequence of two characters -w'- when it is not serving as a digraph.
393 'yylyw y' '#bwn 'yylw' h'w'

uah olya no‹ba e velya
'a pot (that is) old and new', though this makes as little sense as
Blondheim's implausible translation of velya as 'watch it'. Perhaps the
copyist has added another's marginal correction into the text proper.
397 hlwq hX h'

ah tah kolah
'your glue'. The scribe has probably forgotten the w u in what should be
tua. Note that this is the only instance in which the possessive
construction occurs in its modern form, with both the definite article
and possessive adjective.
f. 18v.
399 ydl'#g §wq w''rqzym

mezkra-o kon ‹galde
'mix it with yellow'. The modern language uses amarelo < *amarellus, a
Hispanic Latin diminutive of AM ‹ARUS (da Cunha 1982), which has
replaced this borrowing of OFr. jalne < GALBINUS, a Germanic loanword.
183
402 l' wd wXn'X

tanto do al
Blondheim's hesitant translation 'as much as the other' suggests that he
correctly identifies al < *ale < *ALID < ALIUD (Ferreira 1999).
403 ¶ynr'b
braniß
'varnish'. ModPg. verniz is porbably borrowed from OFr. verniz or OIt.
vernice, both of which derive from MedL. VERONICE, itself based on
Greek bernikê. The form here, which occurs twice in the chapter, shows
the /a/ for expected /e/ characteristic of this text (cf. § 1.2), though it
occurs as ¶ynryw verniß in line 420, with /e/ and a more etymologically-
justified initial consonant.
404 'rbwn yd hS'r'g

garasah de nobra
'nobra grease'. Since 's'rg graça occurs in the next line, the form here
has an unecessary extra syllable. The term nobra is untranslated by
Blondheim, and seems urelated to nobre 'noble'.
406 r'g'wwn'rg yd
de granvagar
'vigorously'. Although it does so here, Portuguese does not generally
apocopate forms such as segundo 'according to', primero 'first', or grande
'great' as Spanish does in certain contexts (cf. ModSp. según, primer,
gran). The second element in this adverb seems akin to vigor but with
an unexpected ' a in both syllables.
f. 19r.
412 yy' yd h'w' §yn

nin uah de ayi
'not one of (them) there'. Since forms such as nenyua occur elsewhere, it
is likely that nen uah occurs here without indication of a palatal n
because it is split by a line break (although, as noted above, other words
occur without a palatal spellings where one is expected).
184
415 wSwmry#p y' wl'r

ralo e fermoso
'thin and fine.' ralo is a dissimilated doublet of raro, both derived from
RARUS (cf. It. rado 'sparse'). ModPg. formoso < FORMOSUS, although as da
Cunha (1982) points out, the variants fremoso and fermoso (the latter
used throughout this text) dominated in the medieval language.
419 wryyXrwm
morteiro
'mortar'. Although the Arabic loanword almofariz is more frequent in
the text, this is the relatively more native term, ultimately derived from
MORTA ‹ RIU but borrowed from It. mortaro (da Cunha 1982). Fr. mortier is
the source of the homonym associated with weaponry.
f. 19v.
430 'XySwr
roseta
'rose-color'. ModPg. roseta (diminutive of rosa) refers either to a small
rose, or to an arrangement of ribbons in the shape of a rose and worn as
an ornament. Blondheim leaves the word untranslated, since it would
appear to refer here simply to rose-color.
w'ylwm
mole-o
'grind it'. The only occurrence of this frequently-occurring verb spelled
conservatively with l for the deleted /l/.
f. 20r.
436 yrbmwl'
alumbre
'alum'. This is the only occurrence of a nasalizing context spelled with
m m preceding a labial consonant), which is the standard modern
graphy (the more expected spelling yrbnwl' alunbre does occur earlier in
the same line; also cf. note line 113).
185
CHAPTER FIVE
'qy&g'm yd wrbyl w' – O LIBRO DE MA›GIKA
(BODLEIAN MS. LAUD OR. 282)
1. INTRODUCTION
The text that I refer to as O libro de ma‹gika1 is by far the largest in the
Judeo-Portuguese corpus. At over 800 pages (each containing between 29 and
31 lines), the Bodleian manuscript on its own comprises more than half of the
known body of Portuguese written in Hebrew script. It has nonetheless
remained virtually unexamined by scholars of Portuguese or Judeo-Romance
linguistics, and the edition here marks its first appearance in print.
The earliest reference to the Hebraicized version2 of O libro de ma‹gika
appears in a brief article by Gonzalez Llubera (1953), which outlines the salient
features and textual history of this and a shorter Bodleian astrological text in
Hebraicized Portuguese (see below). Most importantly, at the end of the
article a small portion (approximately one page) of the text is transliterated
into Roman characters. However, no commentary is provided, and the only
Romanization is a skeletal transliteration (cf. chapter 8 § 2.3.1), which renders

the matres lectionis uniquely as their historical Semitic consonant counterparts
and so features no Roman-letter vowels. In a suite of articles Hilty (1957-58)
1
Based on the scribe's use of this title in the closing line of the manuscript: w' 'bq' yS yq'
a
§m' w'yd '&gyS wd'bwl Swgrwb yd ly&g §'w&g Swpnwq yq 'qy&g'm yd wrbyl wnyXyS aki se ak ba o seteno libro de
ma‹gika ke konpos ‹guan ‹gil de burgos lobado se‹ga deu amen 'Here ends the seventh book of magic
composed by Juan Gil of Burgos, praised be God, Amen'. On the s-less form of God's name
see § 2.2.
2 The only other copy of the text appears to be a Roman-letter manuscript of the third chapter
(f.230r.-283v. in the Hebraicized version), held at the Biblioteca Colombina in Seville (Spain).
It has also gone unexamined thus far.
186
makes further reference to the manuscript, although his primary object is the
shorter of the two Bodleian astrological texts, Swzy'w&g Swny' wdyrpnwq wrbyl w'
S'lyrXSy' S'd O libro kunprido enos ‹guizos das estrelas 'The Complete Book on
the Decrees of the Stars' (Bodleian ms. Laud Or. 310). Most recently the
provenance and authorship of O libro de ma‹gika has been addressed by Levi
(1995), though like Hilty his main focus remains on O libro kunprido.
Based on the note at the end of the manuscript, the text of O libro de
ma‹gika was composed by an astrologer whom the scribe identifies as yd lyg §'w&g
Swgrwb ‹goan gil de burgos. However, the precise identity of this figure has not
been clearly determined. Pereira da Silva (1924) provides the earliest and only
concerted investigation3 into this alleged author of O libro de ma‹gika, and
identifies him as an Aragonese court official, João Gil de Castiello, whose
fourteenth-century work on astronomy is cited in the Livro de Montaria of the
Portuguese King D. João I (1357-1433). The identity of the copyist of the
Hebrew-letter manuscript itself remains completely unknown.
2. OVERALL LINGUISTIC CHARACTER

Based on the twenty-seven pages presented here and others I have
examined less methodically, O libro de ma‹gika presents a less adventurous

adaptation of Hebrew script than either As kores (chapter 4) or the smaller
Passover texts (chapter 6). Its orthography is generally more consistent than
the other texts, while its lexicon is less proportionally varied than As kores.
Specific items are discussed in greater depth in the commentary of § 6.
3
Levi (1995) does review the evidence brought by Pereira da Silva (1924) and Gonzalez
Llubera (1953) regarding the authorship of O libro de ma‹gika. Although the latter considered
"Joan Gil" merely a Catalan copyist of the astronomical work, Levi ultimately agrees with
Pereira da Silva's conclusion about the author's identity.
187
2.1. r'd)y(byd' adib(i)dar

Special mention should be made here of this word, which pervades O
libro de ma‹gika and is probably the most frequent verb in the text. It usually
occurs followed by the preposition sobre, describing the action or dominion of
the planets or signs of the Zodiac over a variety of individuals, groups, and
objects. The word does not appear to have survived in any form in the modern
language, and I have found no reference to it in any work on medieval
Portuguese. I am therefore greatly indebted to D. de Acosta (p.c.) for the

following account.
Portuguese divida and Spanish deuda, both meaning 'debt', derive from
Latin DEBITA, the plural form of the substantivized past participle of DEBEO <
DE -HABEO 'have from; owe; be bound to by logic, necessity, moral or natural
law' 'owed things, debt'. In Spanish there is another word deudo 'a relative,
ancestor, or descendant', which comes from the singular DEBITUM , with the
sense of 'bound by natural or moral law'. This sense is also found in two of the
three definitions of adeudado given in the first dictionary of the Real Academia
Española, published in 1726: (1) indebted, owing money; (2) related by family;
and (3) bound, obligated. The third definition reads "equivale a lo mismo que
obligado por algún título de equidad, razón o justicia. Es voz antigua y de
ningún uso." This third adeudado is then defined as equivalent to Latin
OBNOXIOUS 'liable, obliged, subject, obedient' and OBLIGATUS 'bound, under
obligation', and an example of usage is given: "Por ley de Dios los hijos son
adeudados y obligados a ayudar y honrar a sus padres" [By God's Law,
children are bound and obliged to aid and honour their parents].
188
There are also a few examples of adeudar in astrological contexts in the

Real Academia's historical corpus. From a 1427 translation of Virgil's Aeneid:
E llámale tempestad, que quiere dezir mal tiempo, e çelestial, porque

veno del çielo, es a saber de corrupçión del aire desecado; e aunporque
lo adeudaron las costillaçiones çelestiales" [And he calls it storm, which
is to say bad weather, and celestial, for it comes from the sky, that is to do
with corruption of dry air; and also because the heavenly constellations
preside (over) it].
From a late fourteenth- / early fifteenth-century medical treatise:
La vna segun la naturaleza delos tiempos delos vientos y

losconponimientos de vnos con otros. La [otra] porel adeudamiento
delosmouimientos celestiales: los quales non podemos alcançar, segun
dixoAuicena enel libro del canon" [The one according to the nature of
the weather, of the winds, and the combinations of (the) ones with the
other. The other according to the effect of heavenly movements; which
we cannot grasp, as (what) Avicenna said in the Book of the Canon].
The verb adividar, then, is likely the now-archaic Portuguese counterpart of

Spanish adeudar, and is used in O libro de ma‹gika to mean 'have under one's
authority, preside (over)'. In terms of its orthographic form, the verb is written
in all but a very few instances with b b as the second consonant (in the other
cases it occurs with a single w w), and on rare occasion a second y y does follow
this consonant. Particular occurrences are discussed in the commentary in § 6.
2.2. Jewish character

It may be tempting to view (some of) the peculiarities described above
as marks of the elusive pre-expulsion Judeo-Portuguese dialect (as compared
to the more distinguishing characteristics of seventeenth- and eighteenth-
189
century Jewish Portuguese in post-Iberian communities; see da Silva Germano

1968, Wexler 1985, Tavani 1988). If they were, however, we would above all
expect to find them both elsewhere in the corpus – especially in the more
manifestly Judaic Passover texts – which we do not,4 and they do not recur in
any systematic way in post-1497 Jewish Portuguese.5 Nor, for that matter, do
they correspond to the characteristics that Minervini (1992) discerns in early
Judeo-Spanish – presuming that there was any commonality to Judeo-Ibero-
Romance besides the religion of its writers and speakers. This is not to say
that some of the linguistic character of O libro de ma‹gika cannot be attributed to

influence from the spoken language of someone involved in the transmission
of the text. However, given its utterly un-Judaic content, and that many of the
putative Jewish features might be purely scribal (in that sense, perhaps more
rightfully "Judaic") in nature, it seems unnecessary at best, presumptuous at
worst, to derive this influence from a Jewishness beyond the individual Jew
who generated the manuscript.
The sole piece of potentially compelling evidence for a Jewish imprint
occurs in the final line of the manuscript (see note 1 at the beginning of this
chapter), where the scribe, using his own "voice" (unless the note itself is
merely copied), writes the vernacular name of God as w'yd deu without the
expected -s (Sw'yd deus is the normal form throughout the text). This would
indeed correlate with the same phenomenon in pre- and post-expulsion Judeo-
Spanish, where God is normally referred to as el Dio (the -s of this semi-
vernacular term having been construed as a plural marker and so dropped in
4
The patterns discussed in the previous sections do recur, to a degree, in O libro kunprido, but
each of Gonzalez Llubera (1953), Hilty (1982), and Levi (1995) considers the two Bodleian
astrology manuscripts to be products of a single hand.
5
Tavani (1988) does note the occurrence in the Portuguese of Livornese Jews of -ANTIA nouns
lacking a final diphthong (cf. chapter 7 § 3.1).
190
deference to the Jewish concept of the "oneness" of God). However, this being
the only occurrence in the text of an s-less form of this word, it is difficult to
determine how much significance to attribute it.
3. TRANSCRIPTION
Given that O libro de ma‹gika contains more than 800 pages of text, my
edition obviously constitutes only an excerpt of the manuscript. Nevertheless,
because the text has never been published in any form (other than the single
page transliterated by Gonzalez Llubera), in contrast to As kores I have

provided a Hebrew-letter transcription of this excerpt. These twenty-seven
pages present the text in the same modern Hebrew typeface used throughout
this study, with only a small number of emendations and abbreviations
resolved. By far the most common of these is an l-' <a-l> ligature, which
occurs in all positions (though not without exception). As with the final
forms, I have not indicated the use of this ligature in my Romanization.
Parentheses in the transcription and Romanization (§ 4) indicate letters
or words that have been added above the letter or word by different hand.6
These interlinear insertions may be accompanied by a horizontal line through
one or more letters or words (though some strikethroughs occur without any
correction), which is also reproduced in both transcription and Romanization.
Words in bold are those written by the scribe using larger letters as section
headings. Square brackets indicate letters that due to wear or damage were
not fully discernable from the manuscript nor imputable from context.
6
According to Gonzalez Llubera (1953), eight distinct hands have added corrections and
marginal notes to the manuscripts, though four of them are in Roman script.
328
CHAPTER SIX
THREE SHORTER TEXTS
1. INTRODUCTION
The texts of this chapter are all significantly shorter than the two in the
preceding chapters, comprising no more than a handful of folios in their
respective manuscripts. Two of them, moreover, represent the only explicitly
Jewish context for Portuguese written in Hebrew script. Both consist of excerpts
contained in Hebrew ma˛zorim, vernacular instructions on the procedures for
the Passover seder. The third text is a short (medical) prescription, contained
on a single folio in a manuscript that includes several unrelated texts in different
hands (see § 4 below).
Along with their religious context, the two Passover texts share a number
of features that distinguish them from the other items in the corpus. First, they
both consist of discontinuous text, in that the vernacular passages are interrupted
by Hebrew-language blessings. They are also the only texts to make use of
Hebrew lexical items, a feature no doubt attributable to their religious context
(though not all of these are directly associated with the ritual).1 In addition,
they are the only texts in the corpus to make a systematic use of niqqud (diacritic
vocalization). This feature is also due to the religious matrix in which they
occur, since as is customary the Hebrew blessings and prayers are pointed. As
in the other texts, however, three letters ( , , ) serve as full-fledged vowels,

1
The Cambridge text also contains one Hebrew term, behema '(domestic) animal, livestock.'
Although there is no religious context per se, the Portuguese text is surrounded by other
Hebrew-language recipes and prescriptions.
329
such that the niqqud is generally redundant (though it does in some case
distinguish /i/ from /e/ or /u/ from /o/). Some of the added diacritics
serve no discernible function in the Portuguese spelling and do not even seem
to draw on Hebrew convention, e.g. the rafeh on letters such as r and √ in
the Brotherton text, or the alternation of both dagesh and rafeh on d (which
also appears with neither one) in the Bodleain text. As is often the case for
medieval Hebrew manuscripts, the lettering and pointing were quite likely
carried out by different individuals. Though the naqdan 'pointer' was in all
likelihood a Portuguese speaker as well, he may not have been familiar enough
with the Hebraicized orthography to deploy it "optimally." Given this likelihood,
along with the categorically alphabetic nature of the writing, I have ignored
the niqqud in my transliteration, and the Romanization follows the same
principles as those used for the other texts (see chapter 3 § 3).
In addition to their similarities, there are several important differences
between the two Passover texts. As opposed to the cursive script of the Bodleian
text, and in fact in contrast to the rest of the corpus, the Brotherton manuscript
is written in square Hebrew characters. Another important difference lies in
the different representation of certain grammatically-salient final syllables.
While the other texts show at least some alternation between vowel- and
consonant-final forms in the spelling of these final syllables (e.g. -an vs.
-ao in third-person verbs, or -çon vs. -çao in nouns < -TIONE) , the Brotherton
text is the only one in the corpus to have systematically deleted all nasal
consonant letters in word-final positions and spelled with vowel letters only.
As noted above, the two Passover texts also contain more Hebrew-
language elements than any of the other texts in the corpus, some of which do
language elements than any of the other texts in the corpus, some of which do
330
not have vernacular synonyms in either medieval Portuguese or modern English.
For the Romanization of these items I have adapted the system described in
chapter 3 by following the traditional practice of Semitic philology for the
consonants (see chapter 2), regardless of whether the character representes a
sound that exists in Portuguese. For the vowels, however, I have not rendered
implicit vowels (those not represented by a mater lectionis) as superscripts, nor
have I augmented the five Roman vowel letters to reflect the niqqud in the
Hebrew-letter originals.2
Further characteristics of each of the texts are discussed in the individual
sections below. Because of the overall brevity of these texts, and because none
has appeared in edited form prior to my own preliminary editions (Strolovitch
2000b), I provide transcriptions of each text in modern Hebrew typeface. As in
the previous two chapters, line numbers in the commentaries refer to the
marginal numbering in the Romanization.
2. BODLEIAN MS . CAN. OR . 108
The first Passover text represents the first manuscript in the corpus to
which I had first-hand access, and the only one published in Fudeman et al.
(2000) whose edition was based on first-hand consultation. Metzger (1977)
points out that when A. Neubauer catalogued this manuscript in 1886, he

Hebrew-language recipes and prescriptions.
believed it to be a ma˛zor from Spain, with vernacular instructions written in
2
In the English translation, however, I have followed a different standard to render Hebrew-
language terms, preferring a "traditional" Romanization that more closely resembles the
orthography-based transcriptions used by non-specialist English-language writers. Thus, for
example, ¸s is rendered <sh>, ˛ and ƒ are both rendered <kh>, and letters are doubled as
per English spelling conventions. In addition, although the verbs in the Passover instructions
alternate between imperative and future-tense forms, I have rendered them consistently as
English imperatives.
331
Spanish. A major cause of his mistake was no doubt the frequency with which
the scribe has used n on the inflection of 3rd-person plural verbs and certain
determiners. This should not be taken as an indication that the scribe produced
or perceived these words with a consonantal [n], since nasalization and loss of
/n/ were well in progress by this time. It is most likely a conservative or
simply archaizing spelling, not unlike the <m> of the modern language (though
it could also be considered evidence for a Spanish influence on the scribe or in
the transmission of the text). Indeed, as noted in § 1 and elsewhere, the use of
in the spelling of word-final syllables is a characteristic alternant of the Judeo-
Portuguese corpus as a whole. Nevertheless, apart from this feature the text
has the least archaic appearance in the corpus, with few surprises from the
point of view of the modern language.
In the sections below, ellipses indicate that a Hebrew blessing followed;
these have been omitted in this edition.
2.1. Transcription
[227r.]
[227v.]
332
[228r.]
[228v.]
[240r.]
[240v.]
[241r.]
2.2. Romanization
[227r.] komo sa iren de bet hakeneset dira n . . .

[227v.] e si akonteçer pesa˛ |2 en ¸sabbat diran . . . |3 dir an o qiddu¸s tuto
eskrito / e ¸sehe˛eyanu. e si | 4 akonteçer pesa˛ bemoßa’e ¸sabbat dira n / o
qiddu¸s tuto eskrito . . . |5 e logo dira ...
qiddu¸s tuto eskrito . . . | e logo dira ...
333
5 [228r.] e beveran ka daun seu vaso | 2 e la ba ran as |3 maos e benzier an . . .

|4 e |5 tomara do apyo e molyara eno |6 vina gre e benzera . . . |7 e
komera e dara atodos | 8 e toma ra teres maßot ua sobre otra e parta aua
delas por meio e
[228v.] e pora ameya antre as duas saas |2 e aot era me ya defundo dos
10 ma ntees |3 pera afiqomin e engao os v asos |4 de vinyo e komençara n logo
e diran . . .
[240r.] e beva kadaun seu v a aso |2 e labara n as maos |3 e benzeran . . . e
tomara |4 as duas maßot e meya e dira . . . e non
[240v.] koma ata ke diga . . . |2 e parta da saa de çima e da meya de
15 fundo dela e dee |3 a.todos non menos de ua azeytona |4 e tomara da
alfaça e molya ra |5 eno ˛aroset e diga . . . e tomara da oytra |6 maßa saah e
da alfaç ah e molyara | 6 eno ˛aroset non diga beraƒa e komera |7 todo e
diga . . . e |8 da ra a.todos e çeara n e depoes |9 ke çearen toma ra a a.me ya
maßa ke.pos | 10 d efundo dos mantees e koman todos
20 [241r.] dela koanto ua azeytona e en‹gao |2 os v asos de.vinyo e digan |3
birkat mazon
[243v.] e bevan kada-un | 2 seu v a so e tornen a.en‹ger os |3 v asos e digan
...
2.3. Translation
[f. 227r.] Upon leaving synagogue say...
[f. 227v.] And if Passover falls on the Sabbath say... Say the whole written
kiddush and shehekheyanu. And if Passover falls on the waning of the Sabbath
say the whole written kiddush… and then say…
[f. 228r.] And everyone is to drink his glass and wash his hands and bless ...
And take from the celery and dip [it] in vinegar and bless ... and eat [it] and
give to everyone and take three matzot one on top of the other and divide one
of them in half [f. 228v.] and put [one] half between the two unbroken ones
and the other underneath the tablecloth for the afikomen, and fill the glasses
with wine and begin right away and say ...
334
[f. 240r.] And everyone is to drink their glass and wash their hands and bless
... and take the two and a half matzot and say ... And don’t [f. 240v.] eat until
you say ... and divide the upper unbroken [matza] and the half below it and
give everyone no less than an olive, and take [of] the lettuce and dip [it] in the
haroset and say ... and take from the other unbroken matza and from the lettuce
and dip [them] in the haroset and don’t say a blessing and eat it all and say ...
and give [some] to everyone, and dine, and after eating take the half which
[you] put under the tablecloth and [let] everyone eat [f. 241r.] an olive-sized
piece of it, and fill the glasses with wine and say Grace after Meals.
[f. 243v.] And everyone is to drink his glass and [again] fill the glasses and
say...
2.4. Commentary
f. 227v.
2
si akonteçer
'if (it) falls (on)'. Future subjunctive of ModPg. acontecer < *AD +CONTIGESCERE
< CONTI(N)GERE 'reach'. Although ¸s is the default choice for the spelling
of sibilants, as in this word s is used to spell those that do not derive
strictly from Latin /s/ (compare sairen < SALIRE in line 1 and
cearen < CENARE in line 18).
en ¸sabbat
'on Sabbath' (i.e. Friday night). As opposed to the vernacular preposition
used with the Hebrew term here, the next locative expression of time in
line 3, pesa˛ bemoßa’e ¸sabbat 'Passover on the waning of
Sabbath' (i.e. Saturday night), is fully in Hebrew, with the preposition
be- 'in/on' instead of en moßa’e ¸sabbat.
2-3
qiddu¸s tuto eskrito e ¸sehe˛eyanu
'the whole written kiddush and shehekheyanu'. Two Hebrew blessings,
'sanctification' (over wine) and 'that He has given us life'. Although the
'sanctification' (over wine) and 'that He has given us life'. Although the
335
two † in tuto appear to be nearly identical to the immediately-preceding

(in handwritten cursive the two letters, Ë and ˘, resemble one another
more than in square script to begin with), it is problematic to find a
sensible Hebrew or Romance interpretation for a word using the latter.
f. 228r.
5
beveran kada un
'each one drink'. ModPg. beberam < BIBERE HABENT. Intervocalic /b/ and
/w/ merged early in Popular Latin, resulting in Pg. /v/; although b
is the usual spelling for Pg. /v/, forms of this verb are spelled with w,
while initial /v/ < classical Latin /w/ is spelled with or . Note the
superfluous niqqud for /a/ under √, which is already preceded by a
consonant with the same sub-linear diacritic, resulting in three distinct
indications of the /a/ vowel. kadaun 'everyone' occurs with the same
verb in the singular in line 12.
labaran
'they will wash'. ModPg. lavaram < LAVARE (HABENT). Since other
occurrences of this verb omit the dagesh that the scribe has placed in the
, its use here must be an error, or else a hypercorrection (the word
would have sounded identical to one with an original intervocalic /b/).
It is conceivable, though highly unconventional, that the scribe has used
the dagesh simply to signal very generally the modified reading of the
consonant.
6
apyo
'celery'. While the Brotherton text calls for alfaça 'lettuce' (see § 3.3), the
present one calls for celery (as in Castilian versions of the Passover
rubrics), both of which serve to translate Hebrew karpas 'greens'.
Regarding vinagre (ultimately < VINUM ACRE 'sharp/sour wine'), there is
disagreement as to whether the Portuguese form is borrowed from
Catalan (Nascentes 1932) or Castilian (Da Cunha 1982).
336
7
parta auah
'divide the one'. Subjunctive of partir 'divide'. As in the Brotherton text
(line 2), aua 'the one' is written as a single word.
f. 228v.
9
pora
'will put'. ModPg. pora < PONERE (HABET). Williams (1962) claims that -n-
assimilated to the /r/ in this verb, which would lead us to expect a long
medial consonant. Yet despite the scribe’s fondness for explicitly spelling
VV sequences elsewhere, no doubled consonant is spelled here, as is
conventional in this writing system.
defundo
'under' < DE FUNDU (written as separate words in line 12), replaced in the
modern language by debaixo (de) < DE BASSU along with the latinism sob <
SUB (cf. note line 10 in § 3.4).
f. 240r.
15
non menos de ua azeitona
'no less than an olive'. Although the normal call for an "olive-sized"
portion also occurs in line 20 koanto ua azeytona, the
commentator seems to have altered the content of this ritualized phrase.
17
komera
'(will) eat'. ModPg. comera < *CUM +EDERE (HABET). Although the r is
difficult to make out, the transliteration in Salomon (1980) as comea
forces us to consider -a to be an enclitic object, presumably referring to a
otra maßa saa (which would also entail that todo agree as a feminine).
The reading of the individual letters is difficult, but it seems necessary
to posit the r and thus a future tense similar to those used elsewhere
in the text.
in the text.
337
19
ke.pos
'that (he) placed'. As opposed to the multi-vowel spelling in the
Brotherton text (see note line 10), only a single vowel is spelled in this
form (ModPg. pôs).
3. BROTHERTON ROTH MS . 71
The second Passover text stands apart from the other items in the Judeo-
Portuguese corpus in several respects. In the first place, it consists only of a
trio of individual folios that have become separated from their original
manuscript, which has perhaps been lost along with the intervening folios
(which may, of course, have contained further vernacular passages). It is also
the only Hebrew-letter Portuguese text written in the square script rather than
the cursive Rashi script typical of most Iberian and later Sephardic writing
(and printing). In addition, it is probably the oldest manuscript in the corpus,
having been dated by its previous owner Cecil Roth to the late thirteenth
century (Salomon 1980).
The age and wear of the manuscript makes several elements of the
niqqud difficult to determine: the placement of the dot over that differentiate
historical Hebrew /©s/ from /¸s/; the placement of the dot above or to the left
of to differentiate /o/ and /u/; the single subscribed ˛iriq that represents /i/
may in some cases be one of the two dots of a ßere or the three of a segol (both
of which represent /e/). Other niqqud that is "absent" in my transcription

338
may, in fact, simply be imperceptible to the naked eye.3 In addition, it is often
difficult to distinguish the T-shaped qamaß from the simpler pata˛, though as
symbols for /a/ they do not serve to represent distinct Portuguese vowels.
Overall, however, the occurrence of overt vowel letters is more consistent than
in the Bodleian text.
In the transcription of this Passover text the Hebrew blessings are included
as in the original (given here in bold). The formulaic part that is common to
each one is abbreviated by the scribe as three- and four-letter acronyms using
the initials of each word. Although this is a recurrent practice in Hebrew
through the ages,4 the initials are usually "revocalized" to create a phonotactically-
acceptable Hebrew word (akin to English acronyms that contain no vowel-letters
being read as letter-names). In all four instances here, however, the naqdan has
preserved the diacritics of the initial letters as they occur in their full-word
contexts, since these forms would no doubt be read as the full words they
stood for. In the English translation the Hebrew passages are given in italics
and written out in full.

English imperatives.
3
Indeed, there was no niqqud included in the transcription in my first edition of this text
(Strolovitch 2000ab because I had worked from a photocopy of the article by Salomon (1980) in
which themost
revealed diacritics
of the were
niqud,simply
but, asnot
noted,
apparent.
several Consulting
diacritics still
therequired
manuscript
some
ininference.
Leeds first-hand
revealed most of the niqud, but, as noted, several diacritics still required some inference.
4
The informal names for both Maimonides ( Rabbi Moshe ben Maimon, aka
Rambam) and the Israeli Defense Forces ( Tsva haHagana leYisrael, aka Tsahal),
for example, are composed this way.
339
3.1. Transcription
[f.5r.]
τ
[f.5v.]
[f.7v.]
3.2. Romanization
[f. 5r.] e ‹be‹berao |2 ka‹dau seo ‹baso e la‹ba rau suas |3 maos e dirao b√y
√mh √qbw | 4 fial ne†ilat yadaim. e pilyara as | 5 duas maßot e m eya e dira
duas |6 beraƒot antes ke koma auah b√y √mh hamoßi√ le˛em min ha√areß.
|7 e a segunda b√y √mh √qbw fial |8 √ƒilat maßa. e partira dah maßa |9
5 a.saah e a.mitade e komera todo |10 en uo e dara a todos e pilyara |11
d a.alfaça e molyara eno ˛aroset e |12 benzira b√y √mh √qbwfial √ƒilat |13
maror. e dara a.todos e pilaara | 14 da maßa a.oytra a.enteyra e |15 da
alfaça e untara eno ˛aroset.
340
[f. 5v.] e komera e dira zeƒer lamiqda¸s kehilel e |2 dara a.todos e despoes
10 ke komer |3 suah sefiuda tomara ah maßa a meya |4 ke poos de son os
mantees pera |5 afiqomin e komerao todos de ela | 6 kontiah de uah
azeitona lo pa˛ot velo |7 yoter e en‹girao os ba vasos de |8 vinyo e dir ao
birkat mazon.
[f. 7v.] ... hagefen e ‹be‹berao ka dau seo |2 ‹baso e in‹girao os vasos de |3
15 vinyo e dirao ¸spoƒ...
3.3. Translation
[f.5r.] And everyone is to drink his glass and wash his hands and say: blessed
are you, my lord, our god king of the universe, who has sanctified us with his
commandments and commanded us on the washing of the hands. And take [the] two
and a half matzot and say two blessings before eating: the one blessed are you,
my lord, our god king of the universe, who draws forth bread from the earth, and the
second blessed are you, my lord, our god king of the universe, who has sanctified us
with his commandments and commanded us on the eating of matza. And divide the
unbroken matza and the half and eat it all at once and give to everyone and
take [some] of the lettuce and dip [it] in the haroset and bless: blessed are you,
my lord, our god king of the universe, who has sanctified us with his commandments
and commanded us on the eating of bitter herb. And give to everyone and take
from the other whole matza and from the lettuce and dip [them] in the haroset
[f. 5v.] and eat [them] and say in remembrance of the blessing of Hillel. And give
[some] to everyone and after eating the meal take the half-matza that you put
under the tablecloth for the afikomen and everyone is to eat an olive-sized piece
of it, no less no more. And fill the glasses with wine and say Grace after
Meals.
[f. 7v.] And everyone is to drink his glass and fill the glass with wine and say
shpokh...
341
3.4 Commentary
f.5r.
2
pilyara
'take'. ModPg. pilhar < *PILIARE < PIL‹ARE 'pillage, plunder', though Ferreira
(1999) claims that it a borrowing of It. pigliare. As in the modern meaning
of the Portuguese verb, it is used here to mean simply 'take'. The form
that occurs in line 7, pilaara, is either a scribal error or else a
hypercorrection to avoid what the scribe felt was a Castilian-like palatal
(the word is unrelated to ModPg. pilar < 'crush', a dublet ultimately
based on the same Latin etymon).
3
duas beraƒot
'two blessings'. A Hebrew term with the feminine morphology preserved
on both the Hebrew word itself and the agreeing Portuguese numerical
adjective. This maintenance is also seen on the adjectives and determiners
modifying sefiuda ‘'meal' (line 10) and the various occurrences of
maßa (and its plural maßot) in both Passover texts.
antes ke koma a.uah

'before you eat, the first...'. The lack of modern-style punctuation in the
Hebrew-letter text obscures the fact that auah refers to the first of the
two blessings (which is followed by e a segunda 'and the
second' in the next line), rather than 'before you eat the one' (with verb
in subjunctive).
4
dah maß ah a.saah
'from the unbroken matza '. ModPg. sã < SANA 'whole, sound'. This
adjective alternates in the text with ente(i)ra. The word order is reminiscent
of the normal Hebrew syntax for a definite noun + adjective, e.g. Heb.
hamaßa ha¸slema, lit. 'the-matza the-whole' (also in line 10,
342
ah maßa ah m eya, lit. 'the-matza the-half'). This pattern

does not, however, occur in the Bodleian text..
5
mitade
'middle'. ModPg. metade 'half' < MEDIETATE (cf. Fr. moitié). Nascentes
(1932) describes this form as "remade" from *meitade, meetade.
6
da.alfaça
'from the lettuce'. ModPg. alface < Ar. al-˛ass (though the final , along
with the niqqud that precedes it, indicates that the final vowel here is
/a/). Da Cunha (1982) incorrectly lists al-˛aßa as the etymon (Heb.
˛s:h correlates with the former). The word also occurs in the Bodleian
manuscript, in one instance (line 17) with a vowel-less pseudo-
etymological spelling √lpsh.
7
da maßa a.oytra a.enteira
'from the other intact matza ', lit. from-the matza the-other the-intact'.
Again reminiscent of Hebrew phraseology (cf. note line 4).
8
untara
'(will) dip'. ModPg. untar 'rub, grease' < *UN(C )TARE, an iterative verb
based on UNCTUS, the past participle of UNGERE 'smear, anoint'. In his
preliminary edition of the Bodleian text Salomon (1980) provides a long
footnote to account for his reading of this word as auntara as a uniquely
Jewish term from *IN +TINGERE (adjusted to the first conjugation), akin to
entindran in the Spanish haggada published in Strolovitch (2000a) (note
that molyar 'moisten' < *MOLIARE, from MOLLIS 'soft', is also used here). Yet
this interpretation poses several problems. The development of an initial
diphthong from INT - is not expected, and it is unlikely that alone
spells a diphthong (cf. for -ão third plural verb endings). Moreover,
there is no trace of the -NG- cluster.
343
f.5v.
10
poos
'placed'. Although the long- (or "double-") vowel spelling - - may
suggest the regularly lost -n- of the Latin infectum stem, this is in fact
the perfectum POSUIT > *pouse > pôs, though in this intermediate stage the
niqqud specifies that both vowels are /o/.
de son
'under'. The normal development of SUB 'under' > so does not survive in
Modern Portuguese as an independent word. A form with -n occurs in
the Spanish haggada as well (cf. Strolovitch 2000a), and Salomon cites
Menéndez-Pidal's explanation that it is the result of a phonetic analogy
with non, nin, etc. Its status here as an unbound morpheme may be
illusory, since it is likely dependent upon the preceding de as in ModPg.
debaixo or defundo, the latter written both separately and as one word in
the Bodleian text (cf. note line 9 in § 2.4).
11
kontiah de ua azeitona
'the amount of an olive'. The odd vocalism in kontiah < *QUANTIA
'amount' is no doubt due to the scribe having forgotten an between
and . Native OLIVA has been supplanted in most cases by the loanword
used here, based on Ar. az-zaitüna.
12
lo pa˛ot ve lo yoter
Heb. 'not less and not more'. An adverbial phrase entirely unadapted
from its Hebrew-language form. It is the only Hebrew-language item in
the Passover texts with no inherently ritual or religious connotation to
the words themselves, and one of only two such cases that I have
encountered in the Judeo-Portuguese corpus (cf. note line 6 in § 4.3).
344
os ba vasos
'the cups'. The spelling for the various occurrences of this word (from
*VASU, a regularized form of VAS, pl. VASA ) alternates between initial and
. In this particular instance, however, the writer appears to have
stopped to correct his spelling without emending what he felt to be an
error.
13
birkat mazon
Heb. 'Grace after Meals', literally 'blessing of alimentation'.
f.7.v.
14
ingerao
'fill'. ModPg. encher < IMPLERE . The same term is used in Iberian ma˛zorim
with instructions in Castilian, though in Portuguese this verb did not
first shift to the fourth conjugation (cf. ModSp. henchir).
4. CAMBRIDGE MS . ADD.639.5 (F. 20R.)
The shortest text in the corpus, which is edited and published here for
the first time, consists of a half-page prescription5 in a 27-folio manuscript that
contains notes on diseases and remedies, and which is bound together with six
other manuscripts in a volume held at the Cambridge University Library
(England). Like the Portuguese prescription, these other manuscripts deal
with a variety of non-Judaic issues, chiefly pharmacology and astrology. They
also contain passages in a variety of languages written in Hebrew script: the

for example, are composed this way.
first manuscript, for instance, contains a pharmacological glossary in Judeo-
5
The content of the text seems to offer two possibile readings: it describes either a method for
dealing with a horse-induced wound, or else a technique for skinning rabbit. The options for
each of these interpretations are discussed in the commentaries (§ 4.4).
345
Spanish and Judeo-Arabic, while the fourth contains prescriptions in Judeo-
Arabic along with "miscellaneous scribblings" in Judeo-Greek, Judeo-Spanish,
and Judeo-Italian (Reif 1997).
In his catalog of Hebrew manuscripts held at Cambridge, Reif identifies
the language of the second manuscript in this volume (ms. Add.639.2), which
consists of six folios of gnomic verse, as Judeo-Spanish and Judeo-Portuguese.
This text has been studied by Gutwirth (1992), who described it as written in
Judeo-Spanish and Judeo-Portuguese; after my own brief examination of that
text, I would more accurately characterize it as Judeo-Spanish with significant
Portuguese influence. Thus it is understandable that Reif (1997) would not
recognize the Portuguese character of the short prescription in the fifth
manuscript (ms. Add.639.5), identifying instead it as Judeo-Spanish. Unlike
Neubauer's mistaken characterization of the Bodleian Passover text (see § 2.1),
however, there is no obvious feature that might have induced this
misidentification. Indeed, unlike the ambiguities in the gnomic verse of ms.
Add.639.2, the orthographic features of the text of ms. Add.639.5 presented
here point unambiguously to its language as Portuguese.
The text of this short prescription is written in a hand distinct from
some of the Hebrew-language paragraphs that immediately surround it, though
all are written in the cursive Rashi script. Unlike the Passover texts, there is no
diacritic vocalization, and the only niqqud used is a rafeh to indicate the fricative
reading of several instances of p and g (as in O libro de ma‹gika it is normally
reduced to a dot above the letter). The transcription and Romanization below
maintain the line division of the original.

346
4.1. Transcription
4.2. Romanization
de ka valo
[zere] ke fole dado un koyçe aun gulio en çeal
do baço non en o baço mesmo si non poko
mais ariba h e mandou o fisiko poer lye
5 a senta do muito ale en akel lugar / e mando
degolar uah behema ka rneiro e tomo peda ços
de akela pele kente e pu‹ge de ribah e tirar
un pedaço e meter otro fin ke non abiah
mais pele / e tanben o sangro logo / do
10 braço
4.3. Translation
On horse.
Be it that one was given a heel in the knee above |2 the spleen, not in the
spleen itself but rather a little | 3 higher up. And I get the medic to |5 sit him
down very far away in that place. And I get |6 the throat of a horned animal
slit, and I take pieces |7 of that hot skin and I put [them] over [it] and remove
|8 one piece and place another until there is no |9 more skin, and I also bleed it
over | 10 the coals.
over | the coals.
347
4.4. Commentary
1
kavalo
'horse'. ModPg. cavalo < CABALLU, spelled "unetymologically" with double-
to represent Pg. /v/ < Lat. /b/ (cf. chapter 3 § 2.3.2). Reading this as
a form of ModPg. coelho < CUNICULU is, on orthographic grounds alone,
more problematic. It would have to be construed as the only Judeo-
Portuguese use of double- to represent /w/, or else as a diphthong or
a VC syllable whose second element is a semivowel (cf. chapter 3 §
3.2.4). In addition, the scribe has not used any to indicate palatalization
of the /l/, a spelling that is, however, found a few words on in the clitic
pronoun of fo-le 'that it be (to it)', as well as a number of cases in O
libro de komo se fazen as kores and O libro de ma‹gika.
2
zere
'be (it)'. This first word of the paragraph itself poses a difficult reading.
As a form of the verb ser, the initial z would not be expected. It is
possible, if unlikely, that the first letter is in fact q (if written small, a
cursive ˜ can more closely resemble cursive Ê z ), in which case this may
be a form of ModPg. querer 'want' (perhaps used in the sense of Fr.
falloir). In addition, the second consonant could well be d rather than
r (see note line 4).
koyçe
'heel' or 'paw'. ModPg. coice/couce < CALX, CALCIS 'heel', a variant of the
more common calcanhar < *CALCANEARE.
aun gulio
'to the knee/foreknuckle'. ModPg. joelho < OPg. geolho < *GENUCULU <
GENICULUM, a diminutive of GENU, -US 'knee'. This interpretation depends
upon reading the middle letter as l rather than ß, which more closely
resemble each other in this script (cursive Ï and ˆ). It is possible that the
resemble each other in this script (cursive Ï and ˆ). It is possible that the
348
scribe has left out a between the two in the first word and that it is in
fact oun 'or a'.
en çeal
'above' (or some other prepositional expression).
3
baço
'spleen' or 'hindquarters'. Though this term clearly resembles ModPg.
baço 'spleen' < Gk. h‹epátion (Houaiss 2001), it could be a figurative use of
the word derived from BASSU (ModPg. baixo 'low'; cf. Yid. tux¥s
'behind (n.)', whose Hebrew source is the preposition 'beneath').
4
mandou
'I send'. A causative -like use of the verb (ModPg. mandar 'send, order').
Note the simpler spelling of the desinence mando in line 5. That
this and the other verbs in the prescription are not third-person preterite
forms is supported by the more obviously first-person pu‹ge in line 7
(see below).
poer lye a sentado

'sit him down'. Cursive „ d and ¯ r are extremely similar in most Sephardic
manuscripts, and although the final consonant in the last word here
more closely resembles ¯ , it is difficult to find a plausible interpretation
for the word as sentro.
5
ale
'far away', lit. 'beyond'. Houaiss (2001) cites the competing etymologies
for ModPg. além < (AD ) ILLINC (Da Cunha 1982), or < ECCE HINC (Machado
1967), although given the <l> the former seems more likely.
349
6
degolar
ModPg. degolar 'behead' < DEC ‹OLLARE (Houaiss 2001), probably used here
in the sense of 'slit (the throat)'.
uah behema karneiro

'a (domestic) animal, (a) ram'. The first term is a generic Hebrew word
for 'animal' (Eng. behemoth derives from the Greek transcription of its
plural form). However if the second term (ModPg. carneiro 'ram') is an
adjective modifying the first word (as in 'a horned animal'), it would be
expected to agree with the feminine gender of the Hebrew noun (which
the article, for its part, does). Hence the vernacular term is more likely
an elaboration on the Hebrew word.
peda ços
'pieces'. Easily misread as perços ('piercings', perhaps based on
Fr. percer), this is more likely akin to ModPg. pedaço < * PITACCIU < PITT‹ACIUM,
a form of which occurs more clearly in line 8.
7
pu‹ge
'I placed'. ModPg. puxe 'pull' < PULS ‹AV ‹I , apparently used here with a
slight modification in meaning.
de ribah
'above'. An alternative to aribah in line 4, in the case composed
with the preposition de.
8
fin ke
'until'. Akin to ModIt. finchè, this conjunction does not appear to survive
in exactly this form and sense in the modern language (cf. ModPg. a fim
de 'in order that').
350
9
o sangro
'(I) bleed it'. ModPg. sangrar < *sanglar < *sanguilar, a dissimilated form
of SANGUIN‹ARE (Houaiss 2001).
10
logo do braço
'over the coals'. ModPg. logo 'soon, then' is apparently used here with a
spatial reference. The second term ooccurs in Modern Portuguese as
the feminine collective brasa. Like the English word braise 'cook in liquid',
is most likely borrowed from Fr. braise '(hot) charcoal'. The TLF suggests
it is a Germanic loanword of obscure origin, attested as early as tenth-
century Latin brasa carbones (with feminine gender, as opposed to the
form here).
351
CHAPTER SEVEN
ARCHAISM AND VERNACULARISM
IN JUDEO-PORTUGUESE
1. INTRODUCTION
As noted in the preceding chapters, beyond its unconventional script
the overall linguistic character of Judeo-Portuguese largely conforms to the
profile of non-Hebraicized late medieval Portuguese. The orthographic

variation in the texts, while unique in terms of manifesting the principles of its
writing system, does not indicate a range of variation beyond that expected of
Old Portuguese. The most conspicuous contribution of the Judeo-Portuguese
corpus to the history of Portuguese is in fact in the form of archaism and
vernacularism in the lexicon. Though these phenomena often intersect, the
latter refers to words that show the normal effects of sound change but which
have since been relatinized – that is, re-borrowed from Latin – while the
former refers to lexical readjustment or replacement in general. In this
chapter1 I summarize the phonological and morphological features of the
1
The source of each item is indicated by a letter corresponding to the order if presentation in
the previous chapters:
A: O libro de komo se fazen as kores (Parma ms. 1959)

B: O libro de ma‹gika (Bodleian ms. Laud Or. 282)
C: Passover I (Bodleian ms. Can Or. 108)
D: Passover II (Brotherton ms. Roth 71)
E: Medical prescription (Cambridge ms. Add.639.5)
Not surprisingly, the majority of the examples cited below occur in B (chapter 5), the largest
text in the corpus, with a significant minority from A (chapter 4). Given their much smaller
size and more formulaic content, the three shorter texts (C-E, chapter 6) provide fewer
illustrations of archaic and vernacular features.
352
Judeo-Portuguese lexicon that have since been relatinized or otherwise

recalibrated in Portuguese. This survey is not intended as a comprehensive
catalogue of archaism in the corpus, but rather a sample of vernacular
developments and now-obsolete forms. Before examining these features, I
will outline several characteristics that more broadly distinguish Portuguese
among the Romance languages, and which are unproblematically attested in
the corpus.
1.1. Nouns
Portuguese, both medieval and modern, is typical of the Romance
languages in having reduced the three grammatical genders of Latin to two.
Like other Ibero-Romance, gender is for the most part marked by nouns
ending in -o (masculine) and -a (feminine), while the gender of other nouns
cannot be determined on the basis of final segment alone. Portuguese has also
eliminated the declensional system, and along with it any case-marking
outside the realm of pronouns.
Due to the development of nasal vowels and the deletion of intervocalic
nasal consonants, a large number of Portuguese nouns end in the nasal
diphthong -ão. Plurals are universally marked by -s (pronounced [ß]/[Ω] in the

modern language), yet nouns in -ão appear to form plurals in three different
ways.2 Though only one pattern is productive (-ões), handbooks often
prescribe conflicting forms for existing nouns (e.g. vilão 'vilain' > vilãos, vilães,
vilões), and there is no definitive study of actual usage (Azevedo 2005: 63). An
2
Due to nasal-vowel allography and the lack of a universally-accepted standard orthography,
many singulars (especially loanwords) may themselves appear in more than one form, e.g.
garçom/garção 'waiter'. This allography is also apparent in the verbal system, where nasalized
3rd pl. inflections are spelled with final -m in all but the future indicative.
353
earlier form of this variability, more directly linked to the phonological

variability that gave rise to it, is manifested in Judeo-Portuguese as well,
where different occurrences of words with syllables in the throes of
nasalization and n-deletion are spelled unpredictably with the nasal consonant
n/§ or with a variety of vowel-letter combinations to indicate the hiatus from
deleted /n/:3
Table 7-1. Variation in word-final nasal spelling
A: §wn non w''n nao não 'not'

§wXny' enton w'wXny' entou então 'then'
a
§wqrz' az rkon w''qrz' azarkao zarcão 'red lead'
§wyylymryw vermelyon w''lymryb
bermelao vermelhåo 'vermilion'
§w(y)s'#pnwq konfaç(i)on w''sy#p feçao confecção 'concoction'
B: §wz'r razon Sy'wz'r razoes razão 'reason'
§ws'Slsny' ensçalsaçon Sy'wsl'Sy' esalçoes exaltação 'exaltation'
4
SynwySyrp prisiones Sy'wSyrpnwq konpresoes
1.2. Verbs
In its verbal system, the Judeo-Portuguese corpus exhibits no categories
not found in the Modern Portuguese arsenal, and no forms whose
morphology (stem or inflection) differs substantially from their modern
counterparts (beyond the expected phonological discrepancies). Perhaps not
surprising in a corpus that consists of religious directives, astrological
projections, and instructions for manuscript illumination, there are relatively
few past-tense forms. In fact the modern periphrastic tenses (past-present-
future perfect and their subjunctive/conditional counterparts) do not occur in
3
Recall, however, that the Brotherton Passover text (chapter 6 § 3) completely avoids the
conservative spelling with a final consonant.
4
Both of these words, 'prisons' and 'understandings' respectively, derive ultimately from the
same stem (ModPg. prisão < *PREHENESIO ‹ NE, ModPg. compreensão < CUM+PREHENDENTI‹ONE).
354
the portions of the corpus I have examined,5 nor does the periphrastic future
with ir 'go' occur anywhere in the corpus.6
What is attested are several other developments unique to Portuguese
among the modern Romance languages. The first is the future subjunctive,
which resulted from the merger of two Latin tenses, the future perfect
indicative and perfect subjunctive. It appears throughout the corpus, as in the
modern language, after conjunctions that imply future action or circumstance:
A: Syryzy#p wXSy' yqSyd

deske esto fizeres
'once you do this'
B: Sy'wz'r S'd 'rw&p rywyX §y'yb rwp ly' wdn'wq

kuando el por been tever fora das razoes
'when He considers it good beyond reasons'
C: §firyÊd &tes∆neJkah &tyEJb yJ„d §yÂryi'AS ÙmÙq

komo sairen de beit hakeneset
'when you leave synagogue'
D: &hd˚v¯s &ha'˚S ryEmÙq yEq Hy'Ùp¯Sy„&d yE'

e despoes ke komer suah sefiuda
'and after you eat the meal'
The other major innovation in the Portuguese verbal system is the so-called
inflected infinitive, derived ultimately from the Latin imperfect subjunctive.
5
Although past participles occur frequently, they are almost always used adjectivally or in a
passive sense with forms or ser 'be'. Those that do occur with aver (never ter) are either in non-
finite forms (e.g. wXSyw wdnyw' avendo visto 'having seen', the opening words of O libro de ma‹gika)
or idiomatic phrases (e.g. wd'Syg Syrybw' uberes gisado 'you need' (fut. subj.) in As kores).
6
As kores does contain several instances of ir + present participle, e.g. wdn'wq wyy'ww vayo koando
'continue straining it'.
355
Though much less frequent in the corpus than the future subjunctive, it does
occur on several occasions in As kores:
A: §wylymryw Syryz'&p 'r'p

para fazeres vermelyon
'in order to make red...'
Syryswnwq w' 'r'p w'wd yX

te dou para o konoçeres
'I give you [this sign] so that you recognize it'
Another Portuguese characteristic well-attested in the corpus is the preference

to place clitic object pronouns between the stem and desinence of the
historically-periphrastic future tense (which often alternates with imperative
forms in As kores and the Passover texts, though the latter contain no clitic
pronouns):
A: w'ydyzwl lwz' wd S'Xr'dy#b'q

ka‹bidar-t-as do azul luzidio
'beware of shiny blue'
S'Srb S' yrbwS S'ly'wp y' r'yylwq 'n S'wl'Xyyd

e deita-lo-as na kulyar e poe-l-as sobre as brasas
'and put it in the spoon and place it over the embers'
B: rydnyXny' Swmy' S'lrydwp

poder-las emos entender
'[that] we may understand them'
r'gl'&bq yd S'XSyb §'ylyS r'nwd'

adonar se-le-an bestas de ka‹balgar
'riding animals will be given to him'
356
Note that the separation of the verb stem from its inflection in O libro de ma‹gika
(with the clitic attached to each one in turn in the examples above) is
consistent with the overall tendency in that text for less orthographic
agglutination than the rest of the corpus.
Despite the lack of synchronic periphrastic tenses, it is worth noting
other uses of aver < HAB‹ERE in Judeo-Portuguese. It occurs most frequently in
the modern language as an existential verb há 'there is' (well-attested in the
corpus, usually with accretion of a locative pronoun as yy' ay,) and in a
number of idiomatic phrases (e.g. haver de 'must'), but is generally replaced as

both an auxiliary and a lexical verb of possession by ModPg. têr < TEN‹ERE.
While this latter verb occurs only rarely in the corpus, there are many
examples in which aver is used in a variety of tenses and has clearly
maintained a lexical meaning, as it would until at least the late sixteenth
century (Azevedo 2005: 177):
A: ¶yl'nyXSyd S'nymyl S'ry#b' SywpSyd

despois a‹beras liminas destinaliß
'then take leaves of tin'
ryXSy'ym Syry#bw' wXn'wq r'q'l' S'r'mwX

tomaras o alakar kuanto u‹beres meester
'take lac, as much as you (will) need'
B: §wz'r 'd S'ml' Swmyw' Sy'ymw' yq rwp

por ke os omees avemos almas da razon
'because [as] men we have souls of reason'
wSwrydwp h'yryS §wn ySybw'w' wryy'rXnwq yS

si kontrayro oubese non seriah poderoso
'if [God] had contradiction[s] he would not be almighty'
357
2. PHONOLOGY
In terms of archaism and vernacularism, the corpus contains many
words whose modern forms have "undone" an earlier sound change by
restoring the etymological segments.
2.1. l-clusters
Many Modern Portuguese words contain consonant clusters whose
second element /r/ derives from an etymological /l/. These sound changes
are attested by many items in the Judeo-Portuguese corpus that also preserve
the change in Modern Portuguese (e.g. ryz'rp prazer < PLACERE). Yet the texts
contain several instances of vernacular spellings whose etymological /l/ has
been restored in the modern language:
Table 7-2. Pg. pr < PL
A: yXnysydn'rpSyr resprandeçente < *RESPLEND- resplendente 'resplendent'

B: 'Xyn'rp praneta < PLANETA planeta 'planet'
'Xn'rp pranta < PLANTA planta 'plant'
yXnym 'dyrpnwq konprida mente < COMPLET- completamente 'completely'
Similarly, while some words both in the text and in their modern forms
preserve the shift of BL > br (e.g. wqn'rb branko < *BLANCU , a Germanic
loanword), others in the texts show a vernacular outcome whose modern form
has been relatinized:
358
Table 7-3. Pg. br < BL
B: SwXnym'rbwp pobramentos < *POPULAMENTU (população) 'populatings'

7
wqyrbwp pubriko < PUBLICU público 'public'
Note that in the first case the /bl/ cluster is not strictly etymological, having
developed after an earlier lenition and subsequent syncope had yielded
*poblamento.
Another change restored in the modern forms of some words involves
l-clusters whose initial element is /k/:
Table 7-4. Pg. kr < CL
A: wr'rq kraro < CL‹ARU claro 'clear'

r'qy#pyr'rq krarifikar < *CLARIFIC‹ARE clarificar 'clarfiy'
r'rqsym meçkrar < MISCUL‹ARE mesclar 'mix'
B: SySpyrq kripses < Gk. ekleyiV eclipse 'eclipses'
S'myrq krimas < Gk. klima clima 'climates'
SwqyX'Syryq keresatikos < Gk. ekklesiastikoV eclesiástico 'ecclesiastics'
wgyryrq krerigo < Gk. klerikoV clérigo 'cleric'
Note in the last case that although the initial cluster has been restored in the
modern form, the lenition of the final consonant has not (cf. § 2.2.3 below).
The texts also contain forms that show the parallel change of GL > gr,
but these words preserve the vernacular development in their modern forms
(e.g. ydwrg grude 'glue' < GLUTINE), as do words in the corpus with fr < FL (e.g.
7
Countinho (1969: 122) cites both púvrego and púbrego as attested forms. It is possible in
principle that the scribe has omitted the diacritic on b to indicate /v/, though there is no
indication of lenition in the final consonant (cf. § 2.2.3).
359
8
'q'r#p fraka 'weak' < FLACCA). In both cases, of course, more recent
formations based on the same roots have either restored the etymological
clusters or else have not vernacularized them, e.g. glutinoso, flácido, etc. 9
2.2. Deleted consonants

Though some of the texts contain the vernacular spellings noted above
that have since been relatinized, there are no instances of the opposite pattern,
that is, of conservative spellings in which one of the above clusters is spelled
etymologically with l l in the text but with <r> in its modern form (e.g. there is
no 'X'lp* plata for ModPg. prata 'silver'). Yet there are other phonological
environments in which Judeo-Portuguese writers do appear to spell
conservatively. In addition to the normal lenition of some intervocalic Latin
consonants, Portuguese normally deletes a single intervocalic /l/ and /n/,
resulting in a range of preserved consonants and vowel hiatuses indicated in
medieval spellings.
2.2.1. /l/
The texts contain some words which in the modern forms show the
evolved deletion of intervocalic /l/ but which occur in the text spelled
conservatively with a letter l:
8
De Faria Paiva (1988: 28) describes the occurrence of infruencia 'influence' in the fourteenth-
century Leal conselheiro as an early Latinism; although variant forms of this word abound in O
libro de ma‹gika, none is spelled with r r (see chapter 5, note line 6).
9
In some cases, sound change involving l-clusters yields ModPg. /ß/ (spelled <ch>); words in
the corpus that contain the reflex of this change (spelled g g plus diacritic) also preserve it in
their unaugmented modern forms, e.g. h'y#g ‹geah < PLENA (ModPg. cheia 'full'), ry#gny' en‹ger <
IMPLERE (ModPg. encher), wd'm'#g ‹gamado < CLAMATU (ModPg. chamar 'call').
360
Table 7-5. Intervocalic /l/ spelled in JPg.
A: wl'p palo < P‹ALU pau 'stake'

l'S sal < SALIT sair 'come out'
ylwm mole < MOLIT moer 'grind'
Syrwlwq kolores10 < COL‹ORES cores 'colors (n.)'
r'rwlwq kolorar < COLORA ‹ RE corar 'color (v.)'
B: Syl'ysnysy' eçençiales < ESSENTI‹ALES essençiais 'essential'
Syl'myn' animales < ANIMA ‹ LES animais 'animals'
Note that these last two are the exceptions to the general pattern for plurals
containing an etymological /l/ (such as adjectives based on -ALES), which are
generally spelled without any letter l (i.e. Sy''myn' animais; cf. § 3.1.2).
Other words with an etymological /l/ are not spelled with l in the text,
but are spelled in such a way as to indicate the hiatus from the deleted /l/ that
has since coalesced in the modern form:
Table 7-6. Hiatus from deleted /l/ not spelled in ModPg.
A: Sw'wp poos < *PULVOS pó 'powder'

rw'wq koor11 < COL‹ORE côr 'color'
B: rw'wd door < DOLORE dôr 'pain'
Sw'y&gn' an‹geos < ANGELOS anjo 'angels'
yd'Xw'ww vountade < VOLUNT‹ATE vontade 'will'
yXny'q kaente < CALENTE quente 'hot'
'rwXny'q kaentura < CALENTURA quentura12 'heat'
Sw'w'wsnyl lençouos < LINTEOLOS lençol 'bedsheets'
10
Although the singular form often contains a letter l , this conservatively-spelled plural
occurs only once (cf. the title of the text, O libro de komo se fazen as Syrwq kores).
11
This variant only occurs once.
12
The doublet calentura is a Spanish loanword (Ferreira 1999).
361
In keeping with vernacularism seen in relation to l-clusters, however, other

words whose etymological /l/ has been restored in the modern form occur in
the text without this /l/ spelled:
Table 7-7. Intervocalic /l/ restored in ModPg.
A: r'wdybq kabidoar < CAPITUL‹ARE capitular 'capitalize'

ymw'(') (a)ume < ALUMINE alume 'alum'
B: Swdyb'q kabidos < CAPITULO
‹ S capitulo 'chapters'
SwXnymy' ementos < ELEMENTO ‹ S elemento 'elements'
§w'wX toun < TALONE talão 'heel'
Note that some spellings indicate the hiatus from the deleted /l/, while others
indicate a coalesced vowel.
2.2.2. /n/
The most recurrent example of conservative spelling in Judeo-
Portuguese is the persistence of nasal consonant letters in word-final position
(e.g. 3rd pl. verb inflections, nouns based on -TIONE , the preposition §wq kon
'with', etc.), which generally alternate with vowel-only spellings (cf. § 1.1). As
was the case with /l/, then, there are some instances of words whose modern
forms show the deletion of intervocalic /n/ (with or without nasalization of

the preceding vowel) but which are spelled conservatively with a letter n in the
text:
362
Table 7-8. Deleted /n/ spelled in JPg.
A: hnw' unah13 < UNA uma 'a (f.)'

wn'm mano < MANU mão 'hand'
'snyymwq komeinça14 < CUM+INITIAT começar 'begin'
B: wn'lwS solano < SOLANU solão 'hot sun'
SynwySyrp prisiones < PREHENSIONES prisão 'prisons'
Swnwq'b bakunos < *VACCUNOS vacuum 'bovine'
Swnwrb'q kabronos < *CAPRUNOS cabrum 'goat-related'
ynymy&g ‹gemini < GEMINIS Gêmeos 'Gemini'
'n'rwq korana < CORO‹ NA coroa 'crown'
Modern Portuguese has nonetheless restored the nasal consonant in a variety

of contexts in which no letter appears in Judeo-Portuguese spelling. In many
cases these are pre-consonantal coda nasals that were deleted early in Ibero-
Romance and were not likely available as active orthographic variants:
Table 7-9. Deleted /n/ restored in ModPg.
A: ry#gyX ti‹ger < TINGERE tingir 'dye'

B: SwSyS sesos < SENSOS senso 'senses'
Sw'ym meos < MINUS menos 'less'
Sy'wSyrpnwq konpresoes < COMPREHENSIONE compreensão 'understandings'
§w'yXyrXyS setreteon < SEPTENTRIONE setentrion 'northern'
yXnydys' açedente < ASCENDENTE ascendente 'ascendant'
r'yswnryp pernuçiar < PRONUNTI‹ARE pronunciar 'pronounce'
13
The feminine indefinite article alternates throughout the corpus between forms with and
without an overt n n. Note that the <m> of the modern form is a restored spelling that serves
the same diacritic purpose as this n, namely to signal the nasalized vowel. The n in mano no
doubt has the same status, but modern conventions are such that nasalization is not indicated
there by a restored consonant.
14
This word does occur multiple times in O libro de ma‹gika with no letter n.
363
Other cases involve an intervocalic /n/ that was deleted later, often leaving an
overtly-spelled hiatus. In the following cases the /n/ has been restored in the
modern spelling:
Table 7-10. Hiatus from deleted /n/ spelled in JPg.
B: Syr'ymwl lumiares < LUMINARES luminar 'lights

wrXSy'yS seetro < SINISTRU sinistro 'left
yd'dy'S saidade < SANITATE sanidade 'health'
w'ymryX termio < TERMINU término 'limit'
r'wz'r razoar < RATIONA ‹ RE (raciocinar) 'reason
§y&g'yl lia‹gen < *LINEAGINE (linhagem)15 'lineage'
Other words that contain a restored /n/ in their modern forms are spelled in
Judeo-Portuguese with no indication of hiatus from the deleted consonant:
Table 7-11. Restored /n/ not spelled in JPg.
B: §ysyXryp perteçen < *PERTINESCUNT pertenecer 'pertain'

wryg gero < *GENERU gênero 'type'
Still others that do show the hiatus from deleted in /n/ in their Judeo-
Portuguese forms have coalesced to a monophthong in their modern spelling:
15
This Provençal loanword replaced the native development.
364
Table 7-12. Coalesced hiatus spelled out in JPg.
A: §w'wb boun < BONU bom 'good'

§y'yb been < BENE bem 'well'
w'wq kou16 < CUM com 'with'
ryXSy'ym meester < MINISTERIU mester 'need'
B: §w'w' oun < UNU um 'one'
wd''g gaado < Sp. ganado gado 'chattel'
l'ryy'y&g ‹geyral < GENERAL geral 'general'
D: hA''AS saah < SANA sã 'unbroken'
E: ry'wp poer < PO
‹ NERE pôr 'put'
This spellings of 'one' and 'good' are especially interesting in that the hiatus
from the deleted /n/is indicated by the two vowel letters (plus diacritic '), but
a final nasal consonant is written to indicate the nasalized vowel.
2.2.3. Other lenitions

In addition to deleted /l/ and /n/, there are several words whose
relatinized modern forms have restored the voicelessness of a lenited
consonant that is spelled with the corresponding voiced consonant in the texts:
Table 7-13. Voiceless stop restored in ModPg.
B: r'gynwb' abonegar < *AD+BENEFICARE abonecar 'fix up'

'Xswg'l laguçta < LOCUSTA locusta 'locust'
Syr'lwgyS segolares < SAECULARES secular 'laypeople''
§'rwd'm maduran < MA ‹ TU
‹ RANT maturar 'mature (v.)'
16
This preposition occurs more frequently in a conservative spelling §wq kon, and occasionally
with a simple vowel wq kon, even when no fusable article or pronoun follows (cf. § 3.5).
365
Similarly, there is a small number of words whose modern forms contain a

voiced consonant that was lenited to Ø in the normal development as spelled
in the corpus:
Table 7-14. Deleted stop restored in ModPg.
A: h'ydny' indiah < INDICA índiga 'indigo'

B: Sylyd deles < D‹EBILES débil 'weak'
By contrast, other words in the corpus that show a hiatus from a similar
deletion have coalesced to a monophthong in the modern spelling (cf. tables 7-
6 and 7-12):
A: y'ww vae < V‹ADIT va 'go'

B: 'Xy'S saita < SAGITTA seta 'arrow'
§wry'yd deeron < DEDERUNT deram 'gave'
S'yrXSy'm maestrias < MAGISTERIAS (mestre) 'skills'
D: SÙ'Ùp poos < POSUIT pôs 'placed'
In one instance, a Judeo-Portuguese spelling actually shows a vernacular

pattern exactly opposite to that of a modern semi-learned form, with /b/>
/v/ lenition indicated by a diacritic but (originally) intervocalic /l/ preserved
in writing:
A: wl#b'yd dia‹blo < DIABOLUM diabo 'devil'
2.3. r-migration
There is an assortment of words in Modern Portuguese whose normal
form contains consonant clusters with /r/ in which this sound has "migrated,"
366
e.g. preguiça 'laziness' < PIGRITIA, quebrar 'break' < CREPARE, alcrevite 'sulphur' <
Ar. al-kibrit. Along with forms of these words, the corpus features a profusion
of other r-migrations that have been undone in the modern language. Some of
these are the result of straightforward consonant metathesis:
Table 7-15. r-metathesis
A: lwr#p frol < FLORE (flor) 'powder'17

(w)nylwr#p frolin(o) < *FLORINE florim 'florin'
B: §ws'nyry&g ‹gerenaçon < GENERATIONE geração 'growth'
Swgyrbwl lubrigos < LU
‹ GUBROS lúgubre 'dark'
Note that this last form is distinct from the others in that the entire cluster
containing the /r/ has metathesized with another consonant.
In other cases, the /r/ has metathesized with the other member of its
own cluster, resulting in a new coda-onset sequence:
Table 7-16. r-cluster metathesis
A: wrnybny' invenro < HIBERNU inverno 'winter'

'd'rnw' onrada < ORN‹ATA ornada 'ornate'
B: 'd'nrw' ornada < HONO‹ RA
‹ TA honrado 'honored'
18
'yygr'l' alargya < ALACR- alegria 'happiness'
Note the contrast in the second and third words, where the effect of r-
migration is such that each word appears to be spelled as the other.
17
A figurative use taken over in the literal sense by ModPg. farinha.
18
The unmetathesized form S'yrgyl' alegrias occurs only in the plural.
367
In some instances, the /r/ has migrated from an onset cluster to create
a cluster in the onset of the following or preceding syllable (cf. preguiça etc.
above):
Table 7-17. /r/ in onset clusters
B: 'zybwrp probeza < PAUPERITIA pobreza 'poverty'

§w'yXnyrXyS setrention < SEPTENTRI‹ONE setentrion 'northern'
r'qyrdyp pedrikar < PREDIC‹ARE (predizer) 'predict'
In other cases, the /r/ in a syllable coda has migrated backward to the onset,
often creating a new cluster:
Table 7-18. /r/ in coda
A: ¶yn'rb braniß19 < MedL. VERONICE verniz 'varnish'

B: 'myrpSy' esprema < SPERMA esperma 'seed'
S'dn'lyrg grelandas < OFr. guerlande guirlanda 'garlands'
S'&gyrny' enre‹gas < Gk. energia energia 'energies'
yrbwS sobre < *SUB+IERIT subir 'rises' (fut. subj.)
By contrast, the /r/ (unetymological in the first case below) has migrated
forward from an onset cluster to the syllable coda:
Table 7-19. onset-coda r-migration

a 20
B: SXynryp pernet s < PLANETAS planeta 'planets'
SwXnymrwXSy' estormentos < ‹INSTR‹UMENTOS instrumento 'instruments'
19
Variants without r-migration (¶ynr'b barniß, ¶ynryw verniß) also occur in the same text.
20
The expected S'Xyn'rp pranetas is in fact the more recurrent form.
368
Similarly, in some words with an etymological PRO- prefix the /r/ has shifted
from the word-initial cluster to the syllable coda, in some cases with a further
vowel change:
Table 7-20. r-migration in prefixed PRO-
B: yd'dnwprwp porfundade < PROFUNDIT‹ATE profundidade 'profundity'

§wysy'yprwp porpeeçion < PR‹OPORTI‹ONE proporção 'proportion'
Swp rwp por pos < PROPOSUI propôs 'proposed'
r'yswnryp pernuçiar < PRO
‹ NU‹ NTIA
‹ RE pronunciar 'pronounce'
wd'gnwlryp perlungado < PRO‹ LONGA ‹ TU prolongado 'prolonged'
Another frequent pattern, particularly characteristic of O libro de ma‹gika,

is a form of r-migration in words containing a suffix derived from -ARIU in
which the /r/ and yod have metathesized. Although this is part of the normal
development of this suffix (which does appear with other words in the text,
e.g. wryymyrp primeiro < PRIM‹ARIU, wryyrwg' agoreiro < *AUGURARIU), the vowel
usually remains /a/, particularly in words that have restored the r-yod
sequence in their modern forms:
Table 7-21. r-yod metathesis
B: Swryy'Xwn notayros < NOT‹ARIOS notário 'notaries'

'ryy'Sysyn neçesayra < NECESS‹ARIA necessária 'necessary'
wryy'rXnwq kontrayro < CONTRA ‹ IRU contrário 'contrary'
a
Sryy'Xnwlww volontayr s < VOLUNTARIAS voluntário 'voluntary'
wryy'q' akayro21 < AQUA ‹ RIU Aquário 'Aquarius'
S'ryy' ayras < ARIE‹ S Áries 'Aries'
21
Note that the would-be parallel wryy'Xy&g'S* sa‹getayro < SAGITTARIU (ModPg. sagitário) does
not occur, but instead is consistently spelled w(')yr'Xy&g'S, with a classicizing suffix.
369
A final pattern, related to the r-l metatheses in table 7-15, involves r-l
dissimilation: 22
Table 7-22. r-l dissimilation
r>l
A: wl'r ralo < R‹ARU raro 'thin'
B: Sy'wsn'gylyp pelegançoes < PEREGR‹IN‹ATI‹ONES peregrinação 'peregrinations'
Sy'ws'rygylyp pelegeraçoes < PEREGR‹IN‹ATI‹ONES peregrinação 'peregrinations'
23
SylyXr'q karteles < CARCERES cárcere 'jails'
wrbylys çelebro < CEREBRU cerebro 'brain'
l>r
A: ly#pr'm marfil < Ar. (fia÷m) al-fil marfim 'ivory'
yr'#b'#g ‹ga‹bari < Ar. ¸gabalı javali 'peccary'
B: yd'dr'nyS senardade < SENILITATE senilidade 'senility'
S'yylwgr' argolyas < Ar. al-¸gulla argola 'hooped jewels'
S'lwgl' algolas argola 'hooped jewels'
Note that in the l > r group, the sound change appears to be spontaneous in
two instances (i.e. not conditioned by the presence of another /r/ or /l/24),
while in the case of the argol(y)as, both the dissimilated and conservative
spellings occur in the text, but the modern form has not restored the /l/.
2.4. Palatals
Along with the first series of yod-induced palatalizations in early
Romance, Portuguese underwent other sound changes that yielded the palatal
22
There is the occasional r-l assimilation as well, e.g. lys'l' alaçel < Ar. al-fiaßır (ModPg.
alacir), S'yryrgy&g gegrerias 'jesterliness', based on Pr. joglar < IOCULATORE, though perhaps this
was influenced by other native words with /gr/ < GL or CL, e.g. regra < R‹EGULA, OPg. segre <
*SECULE < SAECULU (ModPg. século).
23
The X t is a scribal error for what should be s ç.
24
In the case of ‹gabari it is possible that the /l/ of the Arabic definite article (which, as in many
other Arabic loanwords, may have been part of the borrowed form) played a role in this
dissimilation.
370
phonemes /Ò/ and /µ/, which are indicated by the trigraphs yyl and yyn in
Judeo-Portuguese. In the corpus, some of these segments are not spelled as
such when they are expected (based on the modern form), which in some
cases simply correspond to a Castilian-esque spelling (cf. § 4.3.1):
Table 7-23. Palatalization not spelled
A: hlyr'b barilah < Sp. barilla barrilha 'lye'

wnym'gryp pergamino < PERGAM‹INU pergaminho 'parchment'
B: rysyn'm' amaneçer < *AD+MANESC‹ERE amanhecer 'become dawn'
In other cases the spelling indicates a palatal segment that, based on the
modern form, is not expected:
Table 7-24. Unexpected palatal spelling
B: r'yyn'Sny' ensanyar < *INSIGN‹ARE ensinar 'teach'

§wr'yynydrw' ordenyaron < ‹ORDIN‹ARUNT ordenar 'ordered'
Swd'yymyrp' apremyados < *APPRIMIATOS apremido 'oppressed'
Syl'yynywyd divinyales < *DIVINIALES divinal 'divine'
Latin -GN- generally yields Pg. /µ/ (e.g. wd'yynwp punyado 'fistful' < PUGNATU),
and other forms in the corpus that involve this cluster either delete the /g/25
(e.g. yd'dynyd dinidade < DIGNITATE ) or preserve the g g as a conservative
Latinate spelling (e.g. wngyS signo < SIGNU; cf. the ModPg. doublet sino 'bell').
The first two verbs could, however, like the third one (JPg. r'yymyrp' apremyar
< *APPRIMI‹ARE vs. ModPg. apremer < APPRIMERE, simply represent the reflexes
25
Williams (1962: 84) in fact cites ensinar as a "semi-learned" word and an example of this
latter strategy.
371
of Vulgar Latin verbs in -I‹ARE (as opposed to the classical forms in -‹ARE) that
have been relatinized in the modern language.
2.5. oi vs. ou
Although these two diphthongs often have distinct etymological
sources, they also represent different regional developments of an -OCT -
sequence. Williams (1962: 85) suggests this as the origin of their confusion,
which in the sixteenth century saw oi spread to words that originally had ou
(e.g. coisa for cousa < CAUSA) and ou to words that originally had oi not from -
OCT- (e.g. couro for coiro < CORIU). Even into the twentieth century, with some
aspects of the orthography still in flux, the variants were largely
interchangeable (though perhaps not for an individual writer). Judeo-
Portuguese writers often wrote these words with vowels that differed from
their modern and more or less standard spellings. The following are words
with yod-migration resulting in oi (spelled yyw) but that occur with <ou> in their
modern forms:
Table 7-25. yyw -oy- for ModPg. <ou>
A: wryywq koyro < CORIU cour 'leather'

S'ryywSyX tesoyras < T‹ONS‹ORIA tesoura 'scissors'
B: Swryywg' agoyros < AUGURIU agouro 'auguries'
Other words spelled "correctly" with modern <ou> do not contain a historical
yod segment but are nonetheless spelled with the yyw o y variant, further
evidence of the orthographic confusion:
372
B: wryyw' oyro < AURU ouro 'gold'
By contrast, some words whose modern forms have opted for the <oi>
variant occur in Judeo-Portuguese with a spelling that indicates either a long
/o/ or an /ow/ diphthong, which may or may not represent the correct
etymological spelling:
Table 7-26. w'w -ou- for ModPg. <oi>
B: 'Sw'wq kousa < CAUSA coisa 'thing'

a
'bXw'w' out ba < OCTAVA oitava 'eighth'
sw'wq kouç < CALCE coice 'heel'
Note, however, that kousa is a frequent enough word throughout the corpus
for the variant 'Syywq koisa to occur in several instances in both of the longer
texts, including in As kores one occurrence of 'Swq kosa, spelled Castilian-style
with a single vowel letter (cf. § 4.3).
2.6. ' a vs. y e

The Hebraicized writing system makes it impossible in principle to
recognize e~i and o~u confusion in the Judeo-Portuguese corpus, since both
pairs are spelled with one and the same letter. Yet there are many instances
where ' occurs for a non-low vowel in the modern spelling (cf. de Faria Paiva
1988: 34):
373
Table 7-27. ' a for ModPg. e/i/o
A: r'yylwq kulyar colher < COCHLEARE 'spoon'

'd'r's çarada cerrada < *SERATA 'sealed'
'r'yylwX tolyara tolher < TOLLERE 'will remove'
¶ynr'b barniß26 verniz < MedL. VERONICE 'varnish'
B: Swyyn'rXyy' ayntranyos estranho < EXTRA‹ NEU 'foreign'
r'yyn'Sny' ensanyar ensinar < *INSIGNARE 'teach'
yd'dr'nyS senardade senilidade < SENILITATE 'old-age'
'Xswg'l laguçta locusta < LOCUSTA 'locust'
'n'rwq korana coroa < CORONA 'crown'
By the same token, some words in the corpus are spelled with y where another
vowel, usually a, occurs in the modern form. This pattern, though more
frequent overall, is confined to O libro de ma‹gika:
Table 7-28. y e for ModPg. a/o
B: h'ygwlwrXSy' estrologiah astrologia < ASTROLOGIA 'astrology'

Sy'ws'dwmyrX tremudaçoes trasmudação < TRANSMUTATI‹ONE 'movements'
'qyXmSyry' erismatika aritmética < Gk. ariqmetika 'arithmetic'
rwdy'yrX treedor traidor < TRADITORE 'traitor'
Sy'wsy'yrX treeçoes traição < TRADITI‹ONE 'treasons'
Swryynyrym merineiros marinheiro < *MARINARIU 'sailors'
§'nrwXSyX testornan trastornam < *TRANSTORNANT 'revolve'
S'yyny'yr reenyas rainha < R‹EG‹INA 'queen'
wqyX'mry' ermatiko aromático < Gk. aromatikoV 'aromatic'
rwd'yylybyrX trebelyador travalhador 'worker'
S'yryrgy&g ‹gegrerias jogral < Pr. joglar 'jesterly'
Syd'dyl'XypSy' espetalidades hospitalidade < HOSPIT‹ALIT‹ATE 'hospitalities'
26
As noted above, a variant with the expected vowel spelling, ¶ynryw verniß, does occur in the
text as well.
374
2.7. Mono- vs. diphthong

As noted in the previous sections, many diphthongs and vowel hiatuses
have been levelled and coalesced in Modern Portuguese. The corpus contains
a number of words with diphthongs that developed from vocalization of a
velar consonant which have since coalesced to a monophthong:
Table 7-29. Diphthong from vocalized consonant
B: r'Xw'rX trautar < TRACT‹ARE tratar 'treat'

rwd'Xywl luitador < L‹UCT‹AT‹ORE lutador 'fighter'
Other words with no historical diphthong or vowel hiatus are spelled with
multiple vowel letters in the corpus:
Table 7-30. Unetymological double vowel
A: 'snyymwq komeinça< CUM+INITIAT começar 'begin'

ydny'yXny' enteende27 < INTENDIT entender 'understands'
B: ynymy'y&g ‹geemini < GEMINIS Gêmeos 'Gemini'
E: w'wdn'm mandou < MANDO ‹ mando 'I send'
By the same token, there are several words spelled with single vowels in the
corpus that appear with a diphthong in their modern forms:
27
Since there is no etymological basis for the extra vowel letters, the form could actually be
construed as a Castilianism, i.e. entiende. This is in fact how Blondheim (1929) interprets an
analogous spelling of §y'yb in As kores, transliterating it as a Castilian-like bien – despite the
long vowel no doubt simply reflecting the hiatus left from deleted /n/ in BENE > ModPg. bem.
375
Table 7-31. Monophthong for ModPg. diphthong
B: Sysyp peçess < PISCES Peixe 'Pisces'

'&pwq kofa < *CUFFIA coifa 'headdress'
zy&byrw' ori‹bez < AURIFICES ourives 'goldware'
wryyrwg' agoreiro < AUGURARIU agoureiro 'augury'
E: w'ylw&g ‹gulio < GENUCULU joelho 'knee'
3. MORPHOLOGY
In addition to the largely phonological discrepancies – as manifested in
spelling – between forms in Judeo-Portuguese and Modern Portuguese, other

phenomena occur more at a morphological level.
3.1. Nouns
De Faria Paiva (1988: 23-24) notes the propensity in medieval
Portuguese to form nouns with the suffix wXnym -mento < -MENTU . Many of
those attested in the texts have since been replaced in the modern language by
related forms derived with other suffixes:
Table 7-32. wXnym -mento nouns with different forms in ModPg.
ModPg. -ção28
wXnymybysyr reçebemento < RECEPE- recepção 'reception'
wXnym'sl'Sy' esalçamento < *EXALTIA- exaltação 'exaltation'
wXnym'rbwp pobramento < POPULA- população 'population'
wXnym'ry&g §y' en ‹geramento < IN+GENERA- geração 'fertility'
ModPg. -nça
wXnym'dwmyd demudamento < DE+MUTA- mudança 'change'
28 Nouns ending in §ws/w(')'s < -TI ‹ONE correspond for the most part to their modern
counterparts in -ção (with occasional exceptions, e.g. §ws'rpnyX tenpraçon 'tempering', Sy'ws'nwl
lunaçoes 'moon-phases').
376
ModPg. underived
wXnymyylwqSy' eskolyemento < EX+COLLIGE- escolha 'choice'
wXnymysyrydny' endereçemento < EN+DIRECTI- endereço 'address'
SwXnym'syrpSyd despreçamentos < DIS+PRETIA
‹ - desprezo 'scorns'
wXnymyl'p falemento < *FABULA- fala 'speech'
ModPg. replaced
wXnym's'rp' apraçamento < *AD+PLATEA- situação 'position
SwXnym'rqsym meçkramento < *MISCULA- mistura 'mixtures'
Conversely, some nouns that occur with -mento in their modern form appear
in the text with a different suffix:
Table 7-33. Other noun forms for ModPg. -mento
§ws'dwmS'rX trasmudaçon < *TRANSMUTATIONE transmudamento 'transformation'

'ysnyrwb' aborençia < *ABHORRENTIA aborrecimento 'abhorrence'
'snyrpwS sofrença < *SUFFERENTIA sofrimento 'suffering'
In addition to -mento nouns, many nouns in 'sn' -ança < -ANTIA differ
in more subtle ways from their modern counterparts:
Table 7-34. '(y)sn' -anç(i)a nouns with different form in ModPg.
No final diphthong
'sn'Xsws sustança < SUBSTANTIA sustância 'substance'
'snysy' eçença < ESSENTIA essência 'essence'
'sny'ys çeença < SCIENTIA ciência 'science'
ModPg. underived
'sn'n'g ganaça < Sp. ganacia ganho 'gain'
'ysnynymyX temenençia < *TIMENANTIA temor 'fear'
ModPg. replaced
S'snyw' avenças < *HABENTIAS posses 'holdings'
377
For another class of nouns, some of those ending in yd'd -dade < -TATE
also occur in the corpus with slight variations compared to their modern form:
Table 7-35. yd'd -dade nouns with different form in ModPg.
Cluster simplified
yd'dyXnwq kontidade < QUANTITATE quantidade 'quantity'
yd'dygyXn' antigidade < ANTIQUITATE antiquidade 'antiquity'
yd'dynyd dinidade < DIGNITATE dignidade 'dignity'
yd'dzym' amizdade < *AMICITATE amizade 'friendship'
Diphthong leveled
yd'dypwrp propidade < PROPRIETATE propiedade 'property'
yd'dyswS soçidade < SOCIETATE sociedade 'society'
Other
yd'dywq kuidade < *COGITATE cuidado 'attention'
Others have been replaced by forms belonging to an alternate pattern of

derivation, or have fallen out of use entirely:
Table 7-36. yd'd -dade nouns with different suffix in ModPg.
yd'dygys çegedade < *CAECITATE cegueira 'blindness'

yd'dyprwX torpedade < *TURPITATE torpeza 'torpidity'
Syd'dyrwqSy' eskuridades < OBSCURITATES29 escuridão 'darkness'
yd'dnyqsym meçkindade < Ar. miks‹ªn mesquinhez 'meanness'
yd'dysyyn neiçidade < *NE‹ SCITATE (tolice) 'foolishness'
3.1.1. Gender
It is quite common throughout the corpus to find what appears to be a
mismatch between the gender of a determiner and the noun it governs. In
29
This form does survive in the learned doublet obscuridade.
378
some cases, however, the morphology of the noun itself indicates that its
grammatical gender differs from the modern form of the word:
Table 7-37. Gender discrepancy
A: wryd'm madero madeira < MATERIA 'wood'

Swd'#gym me‹gados mijada < *MEIATA 'urine'
B: ws'm maço maça < *MATEA 'mace'
Swy'ydl' aldeos aldeia < Ar. a∂-∂ayfia 'villages'
E: ws'rb braço brasa30 < MedL. BRASA 'embers'
As Coutinho (1969) notes, the singular and plural forms of some Latin neuter
nouns evolved into distinct masculine and feminine nouns in Portuguese (and
elsewhere in Romance). In other cases the masculine and feminine forms are
deployed as count versus mass nouns respectively, e.g. Sp. madero 'log, beam'
vs. madera 'wood' (de Acosta, p.c.). Yet the occurrences of wryd'm madero in As
kores seem to cover both of these uses, and the feminine does not appear in the
text.
3.1.2. Plurals
O libro de ma‹gika features what appears to be an alternation in the plural
form of several feminine nouns. In most cases, a form more closely resembling
the modern plural appears elsewhere in the text:
30
The word occurs as both a masculine and this "correct" feminine form in As kores.
379
Table 7-38. Variant feminine plurals
B: S'yrysrp praçerias prazeres 'pleasures'

S'yl'myn' animalias animais 'animals'
S'yrwXrw' orturias horturas 'gardens'
S'ydrwqSyd deskordias desacordos 'disagreements'
S'yylwgr' argolyas argolas 'hooped jewels'
S'yymwgyl legumyas legumes 'vegetables'
In the case of argolyas, the "extension" may in fact be the normal palatal reflex
of the geminate /l:/ in the Arabic source (cf. Pg. /Ò/ as the reflex of Latin -LL-
via Spanish loanwords). It is also possible that these plurals contain the
Greek-origin suffix that derives an abstract or collective noun (e.g. S'yrgyl'
alegrias 'joys', S'yrwdyb'S sabedorias 'knowledge'). Yet the would-be singulars
of the words in the table above do not occur (e.g. 'yl'myn'* animalia31), nor
does there appear to be any difference in meaning among the occurrences of,
for example, Sy'myn' animais, Syl'myn' animales, and S'yl'myn' animalias.
3.1.3. Miscellaneous nouns

Beyond these patterns noted above, other nouns in the corpus differ in
assorted other morphological ways from their modern counterparts:
Table 7-39. ModPg. unaffixed
B: w'yrydwp poderio < *POTERIU poder 'power'

w''bwp pobao < *POPULANU povo 'population'
'rwglw&p folgura < *FOLLICA- folga 'leisure'
'rw'yr&p friura < *FRI‹ GID- frio 'cold'
w''&byrqSy' eskri‹bao < *SCRIBANU escriba 'writer'
31
Coutinho (1969: 230) does cite ModPg. alimaria 'group of animals' as a derivative of
ANIMALIA.
380
Table 7-40. ModPg. different affix(es)
A: wd'rwm morado < *MORATU amora 'mulberry'

'd'rnys çinrada < CINERATA cinza 'ash'
B: yb'n nabe < NAVE navio 'ship'
S'Xywq koytas < *COCTA- coitado 'miseries'
Sw''ryw veraos < *VERANU primavera 'spring'
rwd'yryg gereador < WGmc. werra- guerreiro 'warrior'
Sw'ysynrw&p forneçios < *FORNICIOS fornicação 'fornications'
S'ylbqSy' eskabalias < EXCAVA- (escavar) 'dungeons'
SymwdyXrys çertidumes < *CERTIT- certidão 'certainty'
wgyd'pysnyrp prinçipadego < *PIRNCIPATICU principado 'princedom'
3.2. Adjectives
Although less variable relative to the corresponding modern forms than
noun morphology, there are several Judeo-Portuguese adjectives that appear
with less morphology than their modern equivalents:
Table 7-41. JPg. adjectives unaffixed
A: Syrdwp pudres < PUTRES pútridos 'putrid'

'#byX te‹ba < TEP- tépida 'warm'
'#gwq ko‹ga < COQU- cozido 'boiled'
B: wSyrp pereso < PIGR- preguiçoso 'lazy'
By the same token, other adjectives in the corpus contain affixes that differ
from the usual modern forms:
381
Table 7-42. JPg. adjectives with alternate affixes
B: l'n'mw' umanal < HUMANALE humano 'human'

yrXsylys çeleçtre < CAELEST- celestial 'heavenly'
wy&brybwS sober‹bio < SUPERB- soberbo 'arrogant'
Sylbybwm mobibles < *MOVIBILE móvel 'mobile'
wSwsyXSwg gostiçoso < *GUSTITIOSU gostoso 'spendthrift'
S'd'syrp preçadas < PRETIA- preciosa 'precious'
wd'XnyqSy' eskentado < *EX+CALENT‹ATU aquecido 'warmed'
'd'nyylym'q kameleinada < *CAMELINA ‹ TU camelina 'camel-colored'
S'ryyd'sybwq kobiçadeiras < CUPIDITIA- cobiçável 'appealing'
Sy''ysy&bwn no‹beçiais < *NOVICIALES novicária 'novice-related'
O libro de ma‹gika in particular contains several ordinal adjectives that have

since been replaced by related forms, some shorter some longer:
Table 7-43. Ordinal numbers
B: 'sryX terça < TERTIA terceira third

'nySyS sesena32 < SESENA sexta sixth
w'y&bwn no‹beo < NOVENU nono ninth
wnyzyd dezeno < DECENU décimo tenth
3.2.1. Past participles

A common feature of Vulgar Latin is the replacement of many classical
past participles by those formed with a suffix -UTU. Several that have been
further regularized (but not necessarily relatinized) occur in the texts with the
-udo reflex of this suffix, albeit in adjectival rather than verbal function:
32
This form does alternate with 'XsyS sesta.
382
Table 7-44. JPg. participles < -UTU
A: 'dwXyryd deretuda < *DE+RETRUTA derretido 'melted'

B: wdwdnyXny' entendudo < *INTENDUTU entendido 'understood'
SwdwdnwqSy' eskondudos < ABSCONDUTOS escondido 'hidden'
Swdwb'S sabudos < *SAPUTOS sabido 'known'
3.3. Verbs
Along with the nouns and adjectives that have been remodelled in
Modern Portuguese, the corpus contain several verbs built around attested
roots that nonetheless do not survive as such in the modern language:
Table 7-45. Verbs with different morphology
A: r'yyl'rbyqSy' eskebralyar < *EX+CREPA- quebrar 'break (off)'

r'Xyybwrp' aprobeitar < *AD+PROFECTARE (proveito) 'profit (from)'
B: h'yly&b'yylymyS semlya‹beliah < *SIMILIA- assemelhar 'resemble'
Other verbs, though they occur in the modern language, generally bear a
different meaning than their Judeo-Portuguese usage and have been replaced
in the attested meaning by a related form:
Table 7-46. Verbs replaced in the attested meaning
A: r'nymwl' aluminar33 < *ALLUMINARE iluminar 'illuminate'

r'rwlwq kolorar < COLOR‹ARE colorir34 'color (v.)'
'ryXwS sotera < *SUBTERRAT enterrar 'bury'
B: rysyl'p faleçer35 < FALLESCERE faltar 'lack'
33
ModPg. alumiar 'light (up), give off light'.
34
The vernacular doublet corar is more restricted to the sense of 'paint' or 'blush'.
35
ModPg. falecer 'die'.
383
3.4. Prepositions and conjunctions

The corpus also features a number of prepositions and conjunctions
that no longer occur in the attested form in the modern language. In some
cases they have been replaced outright, while in others only part has been
changed:
Table 7-47. Archaic prepositions/conjunctions
A: 'q ka < QUA RE porque 'for'

wryp pero36 < PER HOC mas 'but'
¶yrp preß < PRESSU pois 'then'
Syd des < DE EX desde 'from'
yqSyd deske < DE EX QUE depois que 'when'
wSw#g yd de ‹guso < DE ORSU debaixo 'below'
B: Swn'z'&p fazanos < *FACIA AD… diante de 'before (us)'
wrypny' enpero < IN PER HOC pois 'thus'
C: ÙJdÕn˚pyJ„d defundo < DE FUNDU sob 'under'
'AXA' ata < Ar. hatta até 'until'
D: §wS y„d de son < DE SUB sob 'under'
E: hbyr yd de riba < DE RIPA arriba 'over'
yq §y&p fin ke < FINE QUE até que 'until'
Note, of course, that of the modern forms only sob can truly be considered a
Latinism (the -n in the medieval form is due to analogy; see chapter 6 § 3.4).
Two very common prepositions also figure in a different form of
archaism. As in the modern language, en normally fuses with a following
article, pronoun, or demonstrative (e.g. wny' eno 'in the', 'XSyny' enesta 'in that')
while §wq kon does not. In the corpus this pattern is occasionally reversed: §y'
36
This conjunction (akin to Sp. pero < PER HOC) does occur elsewhere in the medieval language
as a synonym of porém (ModPg. 'however') < PER INDE, both of which were also used in the
more etymological sense of 'thus' (Mattos e Silva 1994: 260).
384
en is frequently written as an free-standing word, while §wq less often loses its
final consonant and fuses with the following pronoun:
Table 7-48. Contracted §wq kon
A: h'wq koah com a 'with the'

yly'wq koele com ele 'with it'
a
hXSy'wq koest h com esta 'with that'
Table 7-49. Uncontracted §y' en
A: h'w' §y' en uah numa 'in a'

lyq' §y' en akel naquele 'in that'
B: ly' §y' en el nele 'in it'
wXSy' §y' en esto nesto 'in that'
S'lyq' §y' en akelas naquelas 'in those'
E: w' §y' en o37 no 'in the'
Note also that even when it does fuse, en is always spelled with initial y' e-.
4. LEXICON
As opposed to the morphological differences noted above, other gaps
between words in the Judeo-Portuguese corpus and their usual modern forms
can be considered more directly lexical in nature.
4.1. Replacement
In a few rare cases, vernacular items attested in the text have been
replaced in the modern language by an unadulterated Latinate form:
37
Uncontracted en is much rarer with the definite articles (o, a, os, as): the first twenty-seven
pages of O libro de ma‹gika contain only three instances, while As kores contains none at all.
385
Table 7-50. Unmodified Latinisms in ModPg.
A: §y#g'#bl' al‹bagen < *ALBAGINE albumen 'egg white'

B: ly'y&p feel < B‹ILE bílis 'bile'
Other words in the texts are simply archaisms that have been replaced by
more or less vernacular forms based on other roots entirely:
Table 7-51. Other lexical archaisms
A: r'dy#b'q ka‹bidar < CAVITARE cuidar '(take) care'

B: wrCXSy'yd deestro < DEXTERU direito 'right'
rysy'q' akaeçer < *AD+CADESCERE acontecer 'happen'
Sw'yXSy' esteos < ESTIVOS verãos 'summer'
Swd'yylyrp' aparilyados < *APPARICULATOS equipados 'equipped'
4.2. Romance cognates

In a number of cases, items with a cognate in another modern Romance
language (and that may in fact represent an earlier import from that language)
have been replaced in Modern Portuguese by either a native form or a
Latinism:
Table 7-52. Obsolete cognates
A: yrq kri < CRETA giz 'chalk'

Swdyw vidos (OFr. vuide) mijada 'urine'
ydl'#g ‹galde (OFr. jalne) amarelo 'yellow'
B: Syzydnyb bendezes (It. vendetta) vingança 'vendettas'
SwXnym'wswrwq koroçoamentos (Fr. courroucement) ira 'wraths'
Syn'rq krianes (OFr. crieme) preocupação 'worries'
Sd'yzwd doziadas (It. doccia) orvalho 'dewfall'
Sy'wyynyr renyoes < RE‹ NIO
‹ NES rim 'kidneys'
386
On the other hand, there are several instances of more or less native forms that
have since been replaced by a related loanword or remodelled under the
influence of a cognate form (usually French):
Table 7-53. Remodelled from cognate
A: r'yynwrb brunyar < Gmc. br‹un- brunir 'burnish'

w''syy#p feiçao < FACTI‹ONE confecção 'concoction'
B: Swyynym'q kaminyos < CAMI‹ NO‹ S chaminé 'chimneys'
§y&g'yl lia‹gen < *LINEAGINE linhagem 'lineage'
4.3. Castilianisms
The history of Portuguese is replete with Spanish influence at the
phonological, morphological, and lexical levels (to say nothing of bilingualism
and external influences). As might be expected, then, the Judeo-Portuguese
corpus contains many forms that in one way or another suggest a Spanish
influence on the author of the text or the scribe of the extant manuscript. In
most cases these do not persist in the modern language and consist simply of a
spelling that resembles the Spanish development of an otherwise Portuguese
word (cf. § 2.4):
387
Table 7-54. Phonological Castilianisms
A: 'Swq kosa < CAUSA coisa 'thing'

wgy'w#p fuego < FOCU fogo 'fire'
yrbnwl' alunbre < ALUMINE alume(n) 'alum'
B: y#gyl le‹„ge < LACTE leite 'milk/sap'
Swdlybq kabildos < CAPITULOS cabido 'chapter'
r'rbmwn nombrar < NOMINARE nomear 'name'
SwSwrbmyl lembrosos < LUMINOSU luminoso 'luminous'
Others differ more substantially and so seem to be more direct lexical imports
(or available alternants that have since fallen out of use). Most of these occur
only once or else alternate with the expected forms:
Table 7-55. Lexical Castilianisms
A: '#gy' e‹ga < I‹ACTAT (jeito) 'pour'

ys'Syd desaçe < *DISFACIT desfaze 'dissolve'
'ryyd'q kadeira < *CALDARIU chaleira 'kettle'
B: §wgnyn ningun < NEC U ‹ NU nenhum 'no(ne)'
Syz'h hazes < Gk. fasiV fases 'phases'
Regarding hazes, although the use of non-final h is extremely rare in Judeo-

Portuguese, there is no reason to expect it to serve as the initial /f/ of the
Portuguese form,38 and so the spelling is correctly considered a Castilianism.
There is, however, another word whose spelling might appear to be modelled
on the convention associated with OSp. initial <f>, which had lost its phonetic
content but was maintained as a conservative spelling, later replaced by <h>
(e.g. AFFLARE > OSp. fallar, ModSp. hallar, ModPg. achar 'find'):
38
It would represent the only such hypercorrection I have encountered in the corpus, and
Domincovich (1948) does not report any parallel usages of <h> in Roman-letter Portuguese.
388
B: y'yyl'&p falye achei < *AFLAVI 'I found'

wd'l'&p falado achado < *AFLATU 'found' (past part.)
At first blush this might seem to be a Castilian loanword in which the scribe
has also borrowed the convention of using the normal letter for /f/ to spell an
aspirated or even silent initial consonant. Yet I have found no other instances
in Judeo-Portuguese of initial &p spelling what might appear in Roman-letter
writing as <h> or Ø, nor does Domincovich (1948) note any parallel uses of
<f>. Moreover, native forms of ModPg. achar occur as expected in both O libro
de ma‹gika (Sydyr'&g' a‹garedes 'you-PL . will find') and As kores (wd'#g' a‹gado
'found' and other conjugated forms). The verb thus appear to be a semi-
Castilianizing doublet of r'#g' a‹ g ar, preserving the initial fricative à la
portugaise but spelling the medial consonant more à l'espagnole.39
4.3.1. Hypercorrection
Distinct from Castilianisms themselves, however, are forms in the

corpus that betray the scribe's awareness of Spanish practice through an error
in his Portuguese. The words in the table below normally contain a
diphthong, but since this is the feature that distinguishes some Spanish nouns
from their Portuguese cognates (e.g. DENTE > Sp. diente, Pg. dente), the scribe
has spelled each one with a simple vowel only:
39
In fact, the word recalls the Judeo-Spanish form fayar cited by Penny (1991: 23). In the
Judeo-Spanish texts compiled by Recuero (1988), forms of this verb appear as r'ayy¯l'ah halyar
(1584), r'ay¯l'ah haliar (1713), Ùd'Ayy'#ap fayado (1897), syeXÕn'ayy'#ap fayantes (1897), and iyy'#ap fayi
(1909). In a curious twist of conventions, then, since Judeo-Spanish initial /f/ did not
disappear as in Castilian, it is possible that the earlier occurrences do indeed use initial h as a
conservative spelling (albeit to reflect a more recent convention).
389
Table 7-56. Avoidance of a diphthong
B: SyXnydybw' obedentes obedientes < OBOEDIENTES 'obedient'

Sy''Xnyrw' orentais orientais < ORIENTALES 'eastern'
S'Swsyryp pereçosas preciosas < PRETIO
‹ SAS 'precious'
In a similar context, while the scribe of As kores spells wyyny#gny' in‹genyo

'method' as expected, the scribe of O libro de ma‹gika seems to have considered
the palatal segment in that word as a sign of a Spanish versus Portuguese form
(e.g. ANNU 'year' > Sp. año, Pg. ano; CABALLU 'horse' > Sp. caballo, Pg. cavalo)
and avoided it. In fact both texts contain would-be palatals where the spelling
corresponds to neither Portuguese nor Spanish:
Table 7-57. Avoidance of a palatal
A: S'lymryw vermelas vermelha < VERMICULA 'red'

w''lymryb bermelao vermelhão < VERMICUL- 'vermilion'
yrb'nyz' azinabre azinhavre < Ar. az-zin¸g‹ar 'verdigris'
B: wSwny&gny' in‹genoso engenhoso < INGENIOSU 'ingenious'
Sy'yw vees velho < VET(U)LU 'old'
rw'ym meor melhor < MELIORE 'better'
D: '‡r'a'lyip pilaara pilhar < *PILIARE 'will take'
It is also possible that at least some of these forms lack (y)y for the same reasons
as those in table 7-56, i.e. as a more general avoidance of diphthong-like
sequences.
In the following cases, the scribe seems to have construed a /b/ as akin
to the epenthetic /b/ than occurs in the Spanish but not Portuguese forms of
other cognates (e.g. NOMINARE > Sp. nombrar, Pg. nomear), and has chosen not
to spell it:
390
Table 7-58. Avoidance of epenthetic /b/
B: Swrmyn nemros membro < MEMBRU 'members'

SwSwrmwl lumrosos luminoso < LUMINOSU 'luminous'
40
r'rmymynyr' arenememrar lembrar < MEMORA‹ RE 'remember'
4.4. Arabisms
Both As kores and O libro de m‹gika contain many words of Arabic origin,
some of which survive largely unchanged in the modern language. Others,
however, differ from their modern forms in various ways, whether due to
further phonological change, analogical adjustment, or recalibration with the
source:
Table 7-59. Phonological adjustment
A: ydl'yywl' alvayalde < al-bay∂ alvaiade 'white lead'

ly#pr'm marfil < (fia÷m) al-fil marfim 'ivory'
B: lys'l' alaçel < al-fiaßır alacir 'harvest'
S'yylwgr' algolyas < al-¸gulla argola 'hooped jewels'
h'yryXybl' albeteriah < al-bai†‹ar alveitaria 'animal healing'
In some cases, the definite article that is often integrated into the loanword has
been de-accreted in the modern form:
40
A form with the /b/ spelled, r'rbnymyr' aremenbrar, does occur in the same text.
391
Table 7-60. Morphological adjustment
A: syrdy#S' a‹sedreç < a¸s-¸sitran¸g xadrez 'chess'

r'q'l' alakar < al-lakk laca 'lac'
w''qrz' azarkao < az-zarq‹un zarcão 'zircon'
r'X#Swn' ano‹star < annu¸s‹a†ar nochatro 'sal ammoniac'
Other Arabic loanwords have simply been replaced by native forms with
varying degree of Latinization, or even other Arabic loans:
Table 7-61. Replaced Arabisms
A: y#pynr'z' azarnefe < az-zirníx arsênico 'arsenic'

'dyd'pl' alfadida < al-˛adída azinhavre 'verdigris'
yXy#byrql' alkre‹bite < al-kibrit enxofre 'sulphur'
yq'Xr'ml' almartake < al-mártaq litargírio 'litharge'
r'dygl' algidar < al-¸gidr testo 'bowl'
h'yml'X' atalmiah < *˛altamíyya tigela 'ceramic bowl'
B: Swg'l&p falagos < ƒálaq lisonja 'flatteries'
Sydyyql' alakeides41 < al-q‹a√id prefeito 'prefects'
Note that in the first case, the Greek source of the modern Portuguese is in fact
a cognate of the Arabic source for the medieval loanword.42
5. SUMMARY
The drive to standardize and (re)classicize the Portuguese language,
which began in earnest (and quite self-consciously) following the publication
41
The modern reflex alcaide does survive with specific reference to the medieval ruler of a
castle or province, or to the Spanish equivalent of a modern prefeito, still called alcaide in
Castilian.
42
Ferreira (1999) does list arzenefe as a variant of azarnefe, both archaic alternatives to arsênico.
Assuming that both variants are based on the same Arabic source, in a rare reversal the Judeo-
Portuguese spelling appears to be the more conservative.
392
of the first Portuguese grammars in 1536 and 1540, is described by modern

grammarians as an endeavour "to ennoble the lexicon by substituting Latin or
latinized forms for patrimonial words that had been considered rustic"
(Azevedo 2005: 174). Yet in this fourteenth- and fifteenth-century corpus,
bona fide Latinisms occur only sporadically – yXy' ite 'item' and Sydy' ides 'that
is' from As kores spring to mind, and even these are written without pretension
(or overt attempt, at any rate) to fully classicize the morpho-phonological
form. While I have argued that Judeo-Portuguese writers did not approach
their writing system as a transcription, they do seem to have been relatively

uninhibited in disclosing, however inconsistently, the vernacular character of
their language.
393
CHAPTER EIGHT
READERS, EDITORS, AND TYPESETTERS
In this chapter I address a number of issues alluded to in previous
chapters that have not received attention in past work on Hebraicization or
the adaptation of scripts (and certainly not in connection with Judeo-
Portuguese): the notions of native and foreign in relation to writing systems,
the editorial process(es) of transforming Hebraicized material and making it
accessible to other audiences, and the adaptation of Roman-letter (computer)
keyboards to generate Hebrew-letter text. Each of these is inspired by
practical, "consciousness-raising" concerns in my own work – from finding
myself explaining that despite the Hebrew script the language of these texts is
not Hebrew, to interacting with software that in many ways recapitulates for
the computer age part of the process that generated Hebraicized writing
systems and other adaptations of scripts.
1. CONVENTIONALITY IN THE WRITTEN FORM
Orthographic variation, be it synchronic or diachronic, is the stuff of
historical linguistics. As noted in the first chapter, the earliest writing in the
vast majority of languages was produced through the adaptation of a pre-
existing set of graphemes – or, viewed alternatively, through the adaptation
of spoken forms into the mould of a pre-existing script (and a selected set of
conventions). This process might entail some degree of ad hoc convention, i.e.
sound-to-symbol mappings not manifested in the writing system that inspired
the adaptation (because the writer's language contains sounds not present in
394
the source language, or vice-versa). By extending this view to the adaptation
as a whole, it can seem natural to assume that these first written forms must
be influenced primarily by spoken language – that is, that they convey (or
attempt to convey) the sounds of the language as perceived or produced by
the writer at the time of composition. Consider, however, the following
excerpt from an e-mail message sent to me by a native English speaker who
had recently been studying the Yiddish language:
(1) zai mir a moychel ... Meyn yiddish is azoi shlecht

Sorry ... my Yiddish is pretty bad
As noted elsewhere, Yiddish is at present the only language other than
Hebrew with a (semi)standardized, (semi)institutionalized Hebrew-letter
orthography. This set of conventions constitutes the only adaptation of
Hebrew script to persist as an autonomous written norm (in the same sense
that English and French represent autonomous Roman-letter norms). While
Yiddish orthography may be marked by competing conventions, it is not
marked by competing scripts.
The writer of the above message thus had to make a number of choices
about how to supersede the customary spelling of Yiddish, given the "new" set
of Roman characters. This task may be facilitated or hampered by the fact the
writer is aware, to varying degrees, of pre-existing conventions for
representing other closely-related languages in this script (namely English, her
native language, and German, which her Romanized Yiddish might
inadvertently resemble). What, then, is one to make of the variation in the
vowel spelling in zai 'please' and meyn 'my', both of which are
written/pronounced with the same vowel letter/sound in the modern

395
classroom Yiddish that this writer was learning? What, for that matter, might
a twenty-fifth-century philologist infer about late twentieth-century Yiddish if
confronted by this variation? These questions point to the more general and
fundamental problem suggested in chapter 3, which any historically-oriented
linguistic study must address, namely what variation (if any) is implied at
other levels of the language by variation on the orthographic plane.
1.1. Native and foreign scripts
Since the author of the above e-mail message had elsewhere in her
writing adopted <ai> as a preferred spelling for the /aj/ diphthong, it is the
<ey> variant that should be accounted for. This allograph can be seen as a
combination of influence from the <y> in her native English my, and the <e> in
the German cognate mein (and the digraph for /aj/ in its orthography in
general).1 Note that this choice could not have been based on analogy with
the standard Hebrew-letter spelling §ayym <myyn>, in which the vowel is
represented by two identical letters and (usually) a sublinear diacritic. Notice
also that the writer did not analogize from any of the conventional English
spellings for so-called "long i": the most common spelling already exists as
<mine> and in the context of the sentence would stand out as an unmotivated
borrowing (and a grammatically incorrect one at that), while the putative
models for <mign> (sign, benign) – let alone <mighn> (might?) or <myne>
(tyne? lyme?) – are orthographically marginal. Neither the standard Hebrew-
letter matrix nor the salient Roman-letter models tells the whole story. This
1
Although the writer had not studied German per se, this is one of the salient points of written
German vis-à-vis English, which an English speaker is likely to have encountered through
unnativized spellings such as Klein, -meister, and zeitgeist that occur in written English.
396
writer's Yiddish-language adaptation of Roman script betrays an indifference
toward the dominant spelling and a fusion of competing conventions.
Even more instructive is a multilingual dialogue from an October, 1999
article in the Yiddish Forverts. Written under that newspaper's editorial
guidelines for Yiddish spelling (which, it should be noted, vary slightly from
one writer to the next), the dialogue contains words and phrases excerpted
from a Tel Aviv conversation in Polish, Russian, French, Spanish, Italian, as
well as English, such as the following:
(2) wyn Xa'n sXvd ... XayyrlA' sXvd

dets olrayt ... dets nat nyu
'that's alright ... that's not new'
It is well-known that because they lack a voiced dental fricative phoneme in
their native languages, many Jews born in Eastern Europe realize the voiced
fricative /∂/ of English, which is spelled uniquely by <th> in the modern
orthography, as a stop [d]. The Yiddish speakers in this dialogue could well
have realized the English word that's as something akin to [dets]. Hence their
pronunciation is nicely indicated in the Yiddish spelling by the initial d , the
Hebrew letter d.
However, the impulse to assume that the writer of the article intended
to convey a Yinglish-like [dets] pronunciation is conditioned by more than just
the presence of the letter d. Because the article does not present the word in
its conventional spelling (let alone in its conventional alphabet), it is somehow
easier to presume that it offers a phonetically more candid representation –
and does so intentionally – than it would in its convention-laden "native"
spelling. This presumption is, for better or for worse, supported by the fact
397
that the inferred [d] pronunciation coincides with a plausible realization of the
utterance represented. This account also follows in spite of the fact that the
voiced dental fricative [∂] of the native English pronunciation has more in
common phonologically with the sound most often associated with <d> (a
voiced dental stop) than it does with either the voiceless dental stop or
voiceless glottal fricative, the sounds that English speakers would assign to the
individual symbols in the digraph <th>.
Indeed, the power of <th> as the unique modern English spelling for [∂]
seems to preclude any other graphy from bearing an equally-conventional
English reading. That power is reinforced in this case because the units of the
<th> digraph are "native" to the set of characters normally used to spell
English, while d is taken from a "foreign" system. If the dialogue were
presented in a French-language magazine, printed in a Roman typeface, then
a reader capable of producing [∂] might well read the <th> digraph as such,
despite the fact that it normally indicates [t] in French words (where neither
/∂/ nor /†/ occurs in the standard language). Is it similarly possible for the
Yiddish writer to expect her readers to flout the conventional sound-symbol
mapping in Yiddish – in other words, to read d as [∂], and thus sXvd as the
Yiddish-alphabet spelling for an anglophone-like [∂æts]?
1.2. Orthography is not transcription
The preceding question hinges on whether the utterance from the
Forverts article in (2) was intended in the first place to represent that of a
native or non-native English speaker, and whether the writer meant to
provide her sample of Hebraicized English as a transcription or transliteration
(cf. chapter 2 § 3). As such it raises an important methodological point. As

398
argued above, neither the Roman- nor the Hebrew-letter spelling of [∂] is
phonetically transparent or phonologically precise: both are orthographic
approximations. Yet it would be less conventional for a writer to supplement
the graphic inventory by borrowing or inventing a new character for a sound
not conventionally represented by the available graphemes.2 Instead, one
would expect her to exploit the existing sound-symbol mapping conventions
in more or less predictable ways. The reader can thus expect the writer to
imply a new approximation in such cases, just as the writer would not
normally expect her reader to have to infer the value of an imported or
invented character.
This methodological point is also relevant to the philologist working
five centuries hence. Even if all Roman-letter records of English were lost by
the twenty-fifth century, leaving our philologist with a corpus of only
Hebrew-letter writing, he would not necessarily have a poor record of
twentieth-century English. Knowing something of the limits imposed by
previous conventions on each grapheme, a good deal can be determined
through a fairly simple process of triangulation. Taking d to indicate a sound
with at least the features [+voice] and [+dental], and <th> to signal at the very
least a [+continuant] sound (having already inferred, of course, that it stood
for a single phonological segment), the combined hypotheses yield the correct
underlying phoneme.3 At the same time, knowing that d has been used for a
historical /d/ elsewhere, and that the native English sound it seems to spell is
2
It is not unprecedented, of course: Old English writers borrowed <∂> itself from the Old
Norse writing system.
3
In fact, both Hebrew d and Latin <d> have well-attested uses as symbols for [∂], since both
have stood for a /d/ that underwent a similar process of lenition. In Hebrew this value
survives in the traditional pronunciation of some Sephardic and Eastern communities; in
"Latin" it survives in the standard orthography of Modern Spanish itself.
399
/∂/, the philologist can assume a new and historically well-motivated sound-
to-symbol correspondence (which he can test in the rest of the corpus), where
d ambiguously represents two English phonemes – just like its
conventionalized twentieth-century counterpart <th>.
By the same token, when dealing with Portuguese written in Hebrew
script five centuries prior, the philologist cannot always assume that the
grapheme-to-phoneme correspondences between the Hebrew letters and
Portuguese sounds necessarily parallel or coincide with the correspondences
of the Roman-letter orthography, or that the Hebrew letters representing the
phonological and morphological segments of Portuguese do so in a one-to-
one fashion. Difficult cases could, of course, be resolved by triangulating with
other phonological evidence, contemporary Roman-letter usage, or other
Hebrew-letter patterns. But it is plainly the case that not every graphic
variation in an orthographic system will represent a phonetic or phonological
variant, or that the absence of graphic variation will imply phonological
uniformity. The ad hoc conventions that characterize an individual writer's
adaptation of script, however inconsistently they apply, are conventions
nonetheless, and may appear to be ad hoc only in the modern absence of a
larger more instructive corpus.
1.3. Transcription as pseudo-spelling
Nevertheless, a spelling system that departs radically in its basic graphic
form from the traditional one – by using a set of graphemes unrelated to the
traditional ones – cannot but differ in its portrayal of the language. Numerous
cases could be cited in which the conventions of a relatively standardized
orthography were adopted by a given writer as a quasi-phonetic transcription

400
for another language, or applied in a more official capacity as an orthography
proper.4 This is certainly the basis for efforts (often by non-specialists) to
transcend orthography by "spelling phonetically," i.e. to disambiguate a
spoken form using the conventions of a standardized orthography. For
instance, when I am asked by an anglophone to indicate the "correct" (or
better yet "ideal") pronunciation of my name, my usual choice is <deh-vin>,
even though there are no graphically-nativized English words in which /´/ is
spelled with <eh>,5 and no literate English writer is likely to use this digraph in
guessing the formal spelling of any word. As a written syllable it is grapho-
tactically acceptable only in the service of disambiguating some aspect of the
formal spelling.6
In a more intriguing case, when Greenlaw (2002) tries to capture the
aural impression of Bob Dylan's vocal performance in a song from his
Nashville Skyline album, she writes a series of nonwords that are nonetheless
grapho-tactically acceptable – that is, readable:
(3) untilla brake odayee… lemmy ceeyer maikim sermiyul7

Until the break of day, let me see you make him smile.
4
To cite one characterized by both phenomena: for most of its history the orthography of
Manx Gaelic was an Irish-based system, until it was transcribed by writers of Early Modern
English, at which point the conventions of the latter revealed phonetic information that had
been obscured by the historical spelling (W. Harbert, p.c.).
5
The in the second syllable is necessary because in the context of "phonetic spelling,"
some orthographic conventions are in fact suspended; thus to maintain the <o> would create
the misleading impression that the correct pronunciation of the second syllable was in fact
[vøn].
6
Note that the ambiguity resolved by the <h> is not necessarily related to vowel quality but
rather to stress position (which is certainly a practical concern for those of us named
<Devon>, given those who go by [d\vø'n]).
7
Greenlaw prefaces her transcription by stating (with a wry smile, no doubt) that "I knew all
the words to Nashville Skyline before I knew what they meant," adding that "singing along
was much harder than it should have been" (2002: 73).
401
Note that only one of these "words" actually corresponds to an existing
spelling – and it is the wrong homophone at that, because the effect Greenlaw
is aiming for requires that none of the words correspond to the "correct"
spelling. "Readability" clearly implies rather different things in the context of
transcription and of orthography.
Neither the Dylan lyric nor the transcription of my name represents the
product of normal writing by literate adults, because literate adults do not
normally write simply to represent their speech graphically (just as they do
not speak in order to represent their writing orally). The primary purpose of
orthography is not to represent spoken language but to conventionalize the
graphical transfer of meaning, and there is no obvious reason to expect a
fifteenth-century Portuguese Jewish writer to have behaved otherwise. The
twenty-fifth-century philologist of Hebraicized English must consider that the
language of his text, despite its glaringly nonstandard appearance, may be
none other than standard twentieth-century English. Similarly, as discussed in
chapter 3, the twenty-first-century philologist of Hebraicized Portuguese must
be prepared to admit that the only nonstandard feature of his texts is the
mode of writing itself. Indeed, rather than verifiable insight into the spoken
language of their fifteenth-century writers, what may be more conspicuous in
these medieval texts are the attitudes of post-fifteenth-century readers, as
revealed by the form in which they are presented by their editors to other
post-medieval readers.
2. REPRESENTATION AND ACCESSIBILITY
Beyond the audience initially envisioned by the writers of Judeo-
Portuguese, among those who might take an interest in the medieval

402
Portuguese texts presented in the previous chapters would be Jewish
historians, Romance linguists, even general readers of Portuguese. There is,
moreover, a good chance that in any of those groups there will be some
readers for whom the Hebrew letters are the sole but, for better or for worse,
impenetrable barrier to accessing the substance of the text. Indeed this is the
"problem" that confronts any non-specialized reader of Hebraicized material,
an issue that has been discussed extensively in relation to Old Yiddish texts.
Though the adaptation of Hebrew script to this Jewish form of Middle High
German flourished in later centuries, and though it comprises a much larger
and better-studied corpus, the earliest Yiddish writing very much resembles
our Judeo-Portuguese in its lack of obvious Judaic character beyond its
alphabet. As a consequence, perhaps, amidst the many studies of Old Yiddish
manuscripts, Frakes (1989: 110) notes a "ubiquitous litany of . . . scholars,
which rises in a multilingual choral refrain, calling for texts to be accessible,
zugänglich, or tsutritlekh." Although the amount of previous scholarship on
Hebraicized Portuguese writing (not to mention the size of the corpus itself)
has precluded a similar litany, the issue has a similar resonance.
For many a thoughtful and educated Western audience, literate in one
or more Roman-letter orthographies, it could prove difficult to read any
material in Hebrew script, let alone material a reader might otherwise expect
to recognize as a Romance language. The process of editing Hebraicized texts
– the major activity behind my work – is in many ways a recapitulation of the
process by which the "interested reader" goes about accessing the text. It is in
this sense that, as in Old Yiddish studies, one might call for the Portuguese-
language material expressed in Hebrew script to be rendered more accessible
to the interested audience. This does, of course, carry the implication that this
403
audience may be unable or unwilling to make the script accessible for
themselves. Frakes (1989: 186), for his part, adamantly asserts that editorial
calls for accessibility on these grounds are gratuitous and patronizing:
The premise that medieval Germanists have no access to Old Yiddish

texts due to an alien alphabet is as provincial as it is insulting: whoever
can deal with the myriad linguistic and paleographical problems of Old
and Middle High German manuscripts ... can this person, so trained,
not also learn the Hebrew alphabet, if s/he has an interest in Old Yiddish
texts?
Studies of medieval Yiddish may lack the pre-defined readership or the
"national" tradition of scholarship that could be said to exist for medieval
Portuguese. Thus the audience for Old Yiddish texts must in some case be, in
effect, invented by an editor in order to justify his editorial principles and
practices, and indeed his very act of editing the text. It is also for that reason –
because editors do transcribe, translate, and transliterate in the name of
accessibility – that Frakes motivates his own study of ideology in the editing
of Old Yiddish texts. Because editors differ in how they present the
assumptions they bring to the editing process (often for ideological reasons
discussed below), it is essential for the reader to be alert to what information is
added to or omitted from the original text, and for what reason, with each
step in the process of making such texts discernible to the audience defined by
the editor. In the next section, I examine in more detail the levels of
interpretation required to make the Hebrew-letter Portuguese texts presented
in chapters 4-6 accessible to my own "interested audience."

404
2.1. Facsimile
The Hebraicized English fragment in § 1.1 underwent at least three
implicit transformations: from printed Hebrew characters in a newspaper to a
corresponding set of "Western" characters, through a quasi-phonetic
transcription in Roman script (cf. § 1.2), to conventional modern English
spelling. Yet only these last two steps were depicted on the page. To illustrate
in more detail the decisions that inform this process in the case of Hebraicized
Portuguese, I have excerpted a facsimile of the first Portuguese sentence in the
Bodleian Passover text (see chapter 6 § 2). This photo-like image serves as the
unadulterated starting point for the interpretation involved in "accessibilizing"
the material:
Figure 8-1. Facsimile of Bodleian ms. Can. Or. 108 f. 227r.
This is Judeo-Portuguese au naturel in its handwritten cursive script (albeit
enlarged from the original's pocket-size writing). It should be clear that only a
reader with a very particular set of skills would engage this material "as is."
Indeed, even as someone possessing that set of skills, I have not generally
worked directly from facsimiles, preferring instead to reproduce the text for
my own reference according to the process discussed in the following sections.
2.2. Transcription
The most basic step in editing any manuscript is to render the
handwritten text in a reader-friendly typeface that closely imitates the scribe's

405
lettering style. Since the cursive script of the Portuguese passages in the
Bodleian manuscript is itself less widely-used than the square script,8 however,
it is normal practice to render it in the more familiar typeface from the outset.
For better or for worse this process is normally labeled transcription, which in
this sense is distinct from the notion of transcription discussed in the previous
section (and in ch. 2 § 3) as a phonetic approximation using the orthography of
another written language. In the present context, transcription is the
transformation from one script to another within the same matrix, where in
spite of the mediating agent of a printing mechanism the scripts can be
construed as different "hands" (or fonts) in the same writing system.
Transcription in this case produces the "level 1" fragment below:
(4) Level 1: Transcription
§firyÊd t
& s
e n∆ k
Je h
a t
& yEb
J yJd
„ §yÂryi'S
A ÙmÙq
Even though this transformation has in principle passed over the first step of
presenting the text in a cursive-like typeface, transcription of this nature can
be relatively innocuous, largely akin to rendering a Roman-letter medieval
manuscript into a more modern typeface. Yet there are inevitably internal
conventions of the original that cannot be easily preserved in the transfer
(abbreviations, ligatures, etc.). In the Portuguese excerpt above, I have
resolved at least one graphic peculiarity, namely the scribe's placement of the
letter J y "inside" the preceding Y r, D d, or B b. This convention, very much
8
That is, in modern printing and, consequently, in philological studies. Though I could
implicate myself in this modern bias by citing my own Hebrew education (in which we
rarely worked with Rashi script), this would only underscore the utility of transforming the
facsimile into an affiliated script that is more accessible to potential readers.
406
characteristic of the hand in this Bodleian manuscript, can cause a JY r y

sequence to be quite easily misread as a non-final E h – which is what led
Salomon (1980) in his edition to misinterpret the second Portuguese word as
*saiha n rather than sairen 'leave'.9 Thus despite the fact that it may be
presented as a trivial, objective, or pre-interpretive process, even the
ostensibly mechanical task of transcription involves an exercise of judgment
and editorial power.
2.3. Transliteration
It is the next step, Romanization (a variety of transliteration, as defined in
chapter 2) that is more easily recognized as ideologically loaded, a process
understood by editors and readers as more than a matter of simply reversing
the direction of the script.10 Wellisch (1978: 31) states the objective of
transliteration as being "in principle, a one-to-one transformation, in which
one character of the source script is converted into one (and only one) specific
character of the target script." As such it would seem to be a mechanical
process of substitution, one that might not even require the transliterator to
be more than passingly familiar with each set of graphemes and their possible
sound values, or at least their conventional equivalences. Nevertheless, the
clash between the predominant conventions of alphabets in Romance and
abjads (alphabets lacking vowel letters) in Semitic leads to the two very
9
In fact, since Salomon's edition proceeds directly to normalization (see § 2.4 below) from the
facsimile, it was only by consulting the manuscript that I could discover this error. On the
other hand, my own reverse misreading of a non-final E as JY in the Cambridge medical recipe
initially caused me to misidentify the one Hebrew word in that text, hmhb behema 'animal'.
10
Hary (1996), presumably following the practice of other editors, actually does preserve the
right-to-left orientation in the first stage of his transliteration. For reasons that should be
obvious, however, I have not presented any Roman-letter text in a right-to-left orientation nor
even accounted for my decision to forego this step.
407
different methods of transliteration – and hence representation to the Roman-
literate reader – discussed below.
2.3.1. Skeletal transliteration
Given that many of its texts were and continue to be written in
consonantal scripts, it has been practice in Semitic philology to provide
transliterations that include only Roman-script consonants. In other words,
each letter or alphabetic segment of the original script is replaced on a one-to-
one basis by a phonetically-similar Roman consonant. I refer to this as skeletal
transliteration, to indicate that only the basic graphemic frame of the script has
been identified and transferred:
(5) Level 2a: Skeletal transliteration
qwmw ¸s√yryn d:y b:y± hk:ns± dyrn
This transliteration is an historically accurate one, in that the Hebrew alphabet
emerged from a tradition in which the original letters themselves had only
consonantal values. The non-alphabetic (but non-vocalic) characters in the
original, such as the dagesh and rafeh, may also be indicated in this
transliteration, using symbols that are commonly found in Roman-letter
writing (e.g. colon, underscore, etc.) but that do not obscure the letter's basic
identity.
In rendering the Hebrew graphs as characters more familiar and
conventional to readers of Roman-letter orthographies, the resulting text
preserves the distribution of graphs in the original. It is for this reason, for
example, that the colon is used to represent the dagesh in the d of the third
408
word; though its absence in the first letter of the final word could indicate the
variant pronunciation as a fricative [∂], the fact that a single grapheme has
been used for both is given priority over inserting any phonological
interpretation into the transliteration. By the same token, the final t with rafeh
in the Hebrew compound for 'synagogue' (the fourth and fifth words) is
rendered as <t> with an underscore11 rather than <†>, in deference to the
letter's etymological (and perhaps phonetic) identity as /t/. In this
transliteration, therefore, faithfulness to the scribe's use of one and the same
letter in each instance is paramount.12
Despite its graphical faithfulness, however, the skeletal transliteration is
in a very real sense quite inaccessible to any reader of a Roman-letter writing
system, since no conventional Roman-letter orthography completely omits
letters to represent vowels. There is a striking incongruence between, on the
one hand, the historical accuracy of considering the Hebrew letters only for
their consonantal value, and on the other, the fact that in Hebraicized
orthographies several letters plainly serve as the analogues of Roman vowel-
letters. This incongruence is apparent in the very first word: given a relevant
set of conventions for associating sounds with these Roman letters (or even
for recognizing words as ideographic units), very little is revealed to a naïve
reader of any Romance language about the word spelled <qwmw>.
Frakes (1989: 144) discusses a similar example and calls the
transliteration of Old Yiddish qynwq king as <qwniq> a "monstrosity," claiming
11
I use the underscore rather than, for example, a more graphically-imitative macron for ease
of typography (it is available as a screen character) as well as legibility (underlining is
probably more conventional for the target audience and clashes less with the <t>'s existing
horizontal stroke).
12
As in my own practice described in chapter 3, the purely allographic alternation involving
the final-position forms ¢ k, £ m, § n, • p, and ¶ ß) is never preserved in transliteration.
409
that it does more to alienate interested readers than to make the words
accessible to them. He thus calls into question what purpose and what
audience a skeletal transliteration could serve. It should be obvious, of course,
that his opposition is motivated largely by a bias against "unreadable" forms,
which in this case involves un-English conventions such as the vocalic <w>13
and u-less <q>. In the survey of Hebraicizations in chapter 2, I resorted to
skeletal transliterations in cases where my own knowledge of the target
required too many guesses about vowel quality. In the case of the Judeo-
Portuguese corpus, however, skeletal transliterations have served only as
stepping stones in the process of providing an audience-appropriate
interpretation.
2.3.2. Vocalized transliteration
Thanks to the early adoption of matres lectionis and later systems of
niqqud (cf. chapter 2 § 2), the incongruence between orthographic traditions in
Semitic and Romance is not absolute. Although the phonetic realization of the
diacritics has varied over time and region, this system of graphs was and is
rigidly maintained in any reproduction of a sacred Hebrew text,14 and thus
was at the disposal of Hebraicizing scribes. Still, the entire arsenal was rarely
deployed in adaptations of the Hebrew alphabet to the writing of continuous
texts, since the system of pointing afforded the scribe a richer set of
distinctions than was needed in order to imitate the typical orthography of a
13
The only Roman-letter orthography in which <w> stands for a vowel that I am aware of is
Welsh; unfortunately <q> is not used in Welsh spelling.
14
In addition to canonical texts, pointed writing is typically used in Modern Hebrew for
poetry/song, children's literature, and language learning materials (whether Hebrew is the
target or matrix).
410
Western European language. As seen in chapter 3, even in the small Judeo-
Portuguese corpus, only the Passover texts make any significant use of niqqud.
Yet the diacritics are clearly part of the set of orthographic tools used
by scribes and presumably deciphered by readers. Thus for pointed texts such
as the Passover rubrics, a skeletal Romanization could be considered less
informative about the writer's intentions than one which is vocalized. In this
transliteration, each diacritic is represented by a distinct Roman-letter vowel,
which may require its own diacritic such as a macron or accent mark (like the
consonants in a skeletal transliteration) to indicate distinctions beyond those
afforded by the five15 Roman vowel letters:
(6) Level 2b: Vocalized transliteration
qomo ¸s√îreyn d˙y b˙yt hak:eneset dîran
Although it no longer preserves the distribution of all graphs in a one-to-one
relation, this transliteration strikes a greater balance between capturing the
orthographic tools of the original text and providing one that is visually
meaningful to a Roman-literate audience. While the skeletal transliteration
effectively treats the writing system – if not the underlying language – as
Semitic, the vocalized transliteration allows the reader to appreciate the
Portuguese speaker's adaptation of the script in what can only be described as
a more accessible manner.
Because the Judeo-Portuguese orthography already makes categorical
use of vowel letters, the pointing in the Passover texts is often redundant in
15
Although some systems of phonetic transcription make use of <y> as a strictly vocalic
symbol (usually to represent high front rounded vowels), no tradition of Semitic-script
transliteration that I am aware of does so.
411
the Portuguese portions, with some vowels indicated twice (see § 2.1.1 in
chapter 3). For instance, /e/ and /i/ are regularly represented with both
niqqud and a following y y (e.g. in the second word of the excerpt above, §yÂryi'S
A
<¸s√îreyn>, where the /e/ in the final syllable is spelled by both y and the segol
under r r). The pair of diacritic+letter could in fact be said to form a vowel
digraph, and transliterating every diacritic and letter as distinct graphs could
yield an unusually dense text that appears too cluttered to readers of a
Roman-letter orthography. It is therefore common for editors to selectively
omit some distinctions indicated in the original script that are deemed
redundant (e.g. length distinctions between sub-linear vowels), or to render
the vowel digraphs with a single Roman character (e.g. Í as <e> alone).
Though it may be an ideologically separate move, this procedure in fact
imposes a particular interpretation on the orthography, resulting in a written
form that may not be far from the product of normalization, to which I turn
next.
2.4. Normalization
Editors generally take one further step in (re)representing Hebrew-
letter Romance texts, often doing so without revealing (or even perhaps
performing) the steps described above. Since the language spelled out in the
orthography is indeed Portuguese, it seems only natural, indeed helpful, to
normalize the text – that is, to represent the material using the contemporary
conventions of the dominant writing system. The linguistic character of the
text, it may be claimed, can be best compared to other Portuguese writing
only when it appears spelled as Portuguese ought to be (or, more insidiously,
as it would have been by a non-Jewish writer), as in (7):

412
(7) Level 3: Normalization16
como sairen de beit hakenesset diran
As Frakes (1989) observes, this normalization is particularly valued by editors
of texts written in an adaptation of Hebrew script because, as in the case of the
Old Yiddish literary texts, they often believe there to be a Roman-letter (and
possibly non-Jewish) source. Hence the goal, perhaps unstated, is to
reconstruct the Urtext that underlies this adaptation.
In the case of the Portuguese Passover texts, it is extremely unlikely
that the instructions in a Hebrew ma˛zor are based on any prior Roman-letter
original. Though the rubrics are canonical in content, they may not be so in
form, as attested by the long tradition of vernacular annotation and
translation in the (printed) Passover haggada (Yerushalmi 1975, Yudlov 1997).
Thus to represent the Portuguese instructions in a normalized Roman-letter
spelling may be to represent the language as unduly similar to non-
Hebraicized or even non-Jewish Portuguese. It may deny the very originality
and individuality of the linguistic act manifested by the manuscript and its
writer – particularly if the different script is, in fact, basis enough for
considering the text as written in a distinct variety of the language.
Worse yet, the normalization may deceptively constitute a translation,
which Frakes claims has often been the intention of normalizing an Old
Yiddish text to Middle High German or medieval texts in general (to their
16
This level could, in fact, be subdivided into normalizations such as this one, which present a
putative Roman-letter spelling of the era, and those that simply use the modern orthography
as a standard. Needless to say, the two types are not always distinguishable (or, more
accurately, not always distinguished by editors).
413
respective modern orthographies). A normalization may or may not be
explicitly presented as a translation, depending upon the ideological position
of the editor toward the language of the Hebrew-letter text. It is, however,
common to see it presented as the basis for a "standard edition" of the text,
which is then finally accessible to other scholars for further research (and, one
would presume, to any interested reader). Frakes (1986: 186), for his part,
takes a strong position with respect to this practice as motivated by the old
saw of accessibility, in this case to a particular audience of scholars:
We do not insist on presenting the medieval Germanist – who after all

was not born with his/her knowledge of Gothic and Old Saxon (or
even the Roman alphabet) – with laundered texts for scholarly use, so
why should the Yiddishist be offered Middle High German translations
of Old Yiddish texts as the "standard" editions?
What has in fact become inaccessible to the investigator is the original
environment of the writing, which was likely not equivalent to that of Roman-
letter Portuguese. In the case of a normalized Hebrew-letter text, this
environment has been deliberately camouflaged, to blend in to the editor’s
preconceptions of the language and its (dominant) audience as Portuguese
and Roman-letter-literate, preconceptions that necessarily affect the nature of
any linguistic insight derived from the text. It is for this reason that I have not
provided normalized transliterations for any of the Judeo-Romance material
in my own editions.
As evidence that that the language, orthography, and textual material
of Old Yiddish documents is not based on any Roman-letter archetype, some
scholars of Old Yiddish have pointed out that one does not encounter forms
that would have been produced by means of a one-by-one substitution of

414
Roman letters for Hebrew ones. Birnbaum (1961), for instance, constructs and
then rejects a number of such unattested Hebrew-letter spellings for Germanic
words (actual Yiddish forms, as given by Birnbaum, are in the right-hand
column):
Table 8-1. Pseudo-Yiddish orthographic calques
hky* hqy* ich ¢y' 'I'

nEc˚rg* nvcv%%wrg* grµe¸zen §Swrg (medieval coin)
nEtt„r* nttvr* retten §Xyr
What these forms amount to, of course, are skeletal transliterations "in
reverse," that is, a one-to-one replacement of the Roman graphs with Hebrew
ones. Note the ways in which they flout some of the more robust
conventions of Hebraicization: lack of diacritic ' , tautomorphemic doubled
consonants, non-final forms in final position, etc. Birnbaum's "vav-umlaut" for
grµe¸zen actually seems like a reasonable grapheme-for-grapheme adaptation,
merely a point away from the standard-issue Ù /o/ (cf. %' "aleph-umlaut" to
spell /œ/ in the Hebrew-French dictionary in chapter 2 § 3.2.2.2). And yet it is
entirely unattested. In fact, as Frakes argues, this exercise cannot be taken
seriously, either as a way to model a writer's rendering of Roman-letter
material in Hebrew script17 or even as a refutation that this process took place
at all:
17
Even someone ignorant of the relevant systems may not produce an orthographic calque
such as this. In an informal experiment several years ago, I presented a group of English
speakers unfamiliar with Hebrew script with the letters ', h, X, and t (and the corresponding
values √, h, †, and t) and asked them to spell the word that (cf. § 1.1). In no instance was the h
used. Thus the immutable Roman-letter convention for English /∂/, like that for German /ç/
in ich in table 7-1, was not transferred to the Hebrew letters. Indeed, the only context in which
I have come across such Hebraicized English calques is in handbooks such as the one
discussed in chapter 2 (§ 9.3).
415
If we conceive of a scribe transposing an entire Roman-alphabet,

Middle High German text into a Hebrew-alphabet, Old Yiddish text,
then we cannot at the same time conceive of this scribe as so ignorant
of the grapho-phonemic systems of both Middle High German and Old
Yiddish as to produce hky/hqy from '"ich." (1989: 138)
The "grapho-phonemic system" cited by Frakes, that set of correspondences
between the distinctive units of script and sound, is what the laity call an
alphabet. Implicit in both Birnbaum's experiment and Frakes' dismissal, then, is
the idea that writers of Old Yiddish had in mind and in practice some sort of
alphabet, one that was necessarily based on a hybrid set of grapho-phonemic
principles. Although neither Birnbaum nor Frakes makes explicit suggestions
about the composition of this alphabet, a reasonable goal after examining the
writing system of Hebraicized Portuguese would be to determine what its
writers might have conceived as their own alphabet.
3. THE ALPHABET
Orthographic habits die hard, as writer-readers of highly standardized
written languages such as English or French well know. With regard to
another highly standardized written language, Wright (1982) corrected a view
long held of late classical and early medieval Latin by showing that the
persistence of etymological spelling need not have reflected some sort of
historically-accurate pronunciation maintained by "educated speakers."
Instead, he views the orthographic conservatism of the early Middle Ages as
normal for a language of record like Latin, where it would have been not only
tolerated but embraced by literate speakers, who would recognize the wide
gap between the pronunciation they knew and the spelling they learned as
416
normal facets of their (single) language. In early vernacular spellings, Wright
contends that "the particular sound-letter correspondences of written
Romance were not in fact new, although as a consequence [of Carolingian
reforms in the ninth century] the individual word spellings were" (1997: 266).
Hence these forms should be interpreted by determining the writer's
knowledge of what sounds in his language so conceived could be represented
by what (combination of) letters, given all the patterns that associated orthographic
forms with spoken ones.
Understanding the intention behind the adaptation of Hebrew script to
Portuguese must similarly be based on knowing the contemporary
conventions for the pronunciation of Hebrew qua Hebrew. Such knowledge
is sparse, as there have been few studies of Hebrew in Portugal apart from a
catalogue of lusophone Jewish writing by Raizman (1975), and the survey of
Luso-Hebrew pronunciation by Garbell (1954), where the Portuguese details
are subsumed under "Christian Spain." Nor is knowledge of medieval
Portuguese in itself sufficient to fill this gap, since it is more than likely that
readers of Hebrew could produce sounds in that language that did not occur
in their vernacular.18 Nevertheless, based on Garbell's evidence, as well as
modern Sephardic traditions of Hebrew pronunciation, we can posit a rough
set of symbol-to-sound correspondences associated with the Hebrew letters in
the reading of Hebrew texts. In the table below, each of the Hebrew letters
has at least one possible reading, while some letters have two, due either to
vernacular phonology or historical change in Hebrew itself:
18
I am in good company as a modern English speaker who can successfully produce the [x]
represented by x ˛ and k ƒ in Modern Hebrew, despite never having to do so in my native
language.
417
Table 8-2. Portuguese Hebrew pronunciation (following Garbell 1954)
Ø b g d h w z x t j k l m n s ÷ p ts k r ß t
v © ∂ Ø x f s
Based on the material in the extant texts, a lusophone Hebrew alphabet
would seem necessarily to comprise a subset of the source alphabet, since no
additional letters are used, and several letters do not appear in any non-
Hebrew words. Although graphemes have been imported into anglophone
adaptations of Roman script in the past, and Roman allographs such as <j>
and have been graphemicized in many writing systems, adapters of
Hebrew script do not innovate with respect to the inventory.19 In addition,
Romance-speaking adapters usually make no use of four letters representing
historical Hebrew phonemes: the pharyngeal fricatives x /˛/ and v /¿/
(voiceless and voiced respectively), which are absent from Romance
phonologies, and the voiceless stops k /k/ and t /t/ (velar and dental
respectively, both with spirant allophones), which are dropped in favour of
historically-pharyngealized segments, q q and X †. Their absence from Judeo-
Romance orthographies suggests that these letters were indeed rejected in the
adaptation process and hence from the writers' and readers' conception of
"alphabet."
On the other hand, the final-position allographs of m, n, p, and c (and k
if it is used) are consistently deployed in all adaptations of the script, and the
only graph not originating in Hebrew-language writing that is consistently
19
As noted in chapters 2 and 3, a possible exception exists in the recommendation by the
Yiddish Scientific Institute (YIVO) that tsvey-vovn ("double-ww") be joined at the base, forming
what looks like a Roman <V> (Fishman 1977: XXIII). I have encountered this convention
only sporadically in Yiddish longhand, never in print.
418
"imported" is the apostrophe. Although the absence of x, v, k and t does not
imply that they do not occur in a given text (they are always preserved, of
course, in Hebrew and Aramaic words), we could construct a rudimentary
alphabet of Judeo-Portuguese – using that glottonym strictly on the basis of its
distinct alphabet – based on this reduced set of Hebrew graphemes. The
inventory in the table below represents the letters of such an alphabet and the
sounds for which they would conventionally stand:
Table 8-3. A Judeo-Portuguese "alphabet"
' b g d h w z X y l m/£ n/§ s p/• c/¶ q r S

a b g d a# v dz t i l m n s p s k r s¸
Ø v t¸s *h o z e ts f z s
d¸z u y ts d¸z
*only in Hebrew words
This alphabet, however, has obvious limitations. For instance, it
consists only of "unigraphs" (as in the Hebrew qua Hebrew alphabet), despite
the fact that the orthography of Hebrew-letter Portuguese clearly makes use
of several digraphs and trigraphs. Minervini (1999: 426), for example, states
explicitly that "composed spellings" in Judeo-Spanish writing "are unknown to
Hebrew orthography and were probably influenced by the writing system of
the Latin alphabet." Modern English writers, for their part, are conditioned to
conceiving of an alphabet without graphemes for some of the language's
phonemes, i.e. without any composed alphabetic symbols. The digraph <ch>,
for example, stands almost uniquely for a phonemic affricate /ê/ in Roman-
letter English, and yet it is neither learned nor sung as a letter of the alphabet.
Nor are ch-initial words given a separate section in (most) anglo-oriented
dictionaries. Yet even in writing systems where some digraphs represent

419
what are often only morphophonemic alternants, these may be regarded as
letters in some contexts of alphabetization.20 Thus in determining its alphabet,
the multigraphs used in Hebrew-letter Portuguese might well have been
treated as alphabetic units – that is, as letters. The relevant candidates are laid
out in the table below:
Table 8-4. Multigraphs used in the corpus
h' w' y' yy yw 'w yyl yyn ww

a# #o #e ey uy wa Ò µ v
ãw #I ye oy w
#u
Yet allowing consonantal but not vocalic digraphs into the alphabet would
further reveal the specter of a Roman-alphabet bias, since most Western
European children do not sing of vowels beyond a, e, i, o, u (and sometimes y),
despite using digraphs to indicate further phonemic distinctions. Of course, I
have not had access to the alphabet song recited in the Jewish schools of
medieval Portugal, nor to any other artifacts of their conception of alphabet
such as a dictionary or civic directory.
If our effort is to understand more fully the nature of Hebraic
conventions for writing Portuguese, as well as the broader principles of
orthographic adaptation to which they attest, it be might well be more
instructive to match all the relevant sounds of the contemporary language
with the distinctive graphs found in the corpus:
20
In Wales, for example, the rows of seats in a theatre are usually designated A, B, C, CH, D,
DD, E, F, FF, G, NG, H, and so on – even though word-initial <dd> /∂/ and <ng> /˜/ do not
occur in the lexicon (i.e. in non-derived environments) and do not get separate sections in
most dictionaries.
420
Table 8-5. The consonants of Old Portuguese
p b t d k g f v s z s¸ ¸z ts dz t¸s d¸z m n µ l Ò y r w
p b X d q g #p #b s z S c c z #g #g m n yyn l yyl (y)y r w'
p
L b
J Jd &g p
& (w)w S #g #S s c g g yyÕn yy¯l
&d p b
& c c g S #S
As an after-the-fact schematization, this table might serve someone without
prior exposure to the writing system who, for whatever reason, needed to
write Old Portuguese in Hebrew script. It is, in effect, a recipe for
transcription. As I have been arguing, however, the Hebraicized writing
system is not primarily a transcription of Old Portuguese. Therefore, this
table should not be viewed as portraying the process of script adaptation as it
was undertaken by Judeo-Portuguese writers. In its character-to-character
mapping, however, what it does evoke is a precursor to the task of script
adaptation as mediated by the technology of direct-input typesetting.
4. ROMAN KEYBOARD, HEBREW SCRIPT
Whether it involves the Hebrew alphabet for non-Hebrew languages
or other matrix/target pairs, script adaptation can be described (and of course
schematized, as above) as a "mapping" of the graphemes of the matrix onto
the phonemes of the target, or vice-versa. For the scribes and even the first
printers of Hebraicized writing systems, this cognitive challenge was largely
accomplished before the mechanical act of putting ink to paper. This is less
true for the modern-day designers of hardware and software that equip
writers who are versed in one script to produce language visible in another.
The engineers of the first Hebrew-letter typewriters in the early twentieth

421
century (cf. figure 7.2 below) undoubtedly had to integrate their take on the
"logic" of the Hebrew alphabet with existing Roman-letter keyboard layouts in
such a way that allowed Hebrew and Yiddish writers (themselves probably
familiar with one Roman keyboard or another) to use it with relative ease.
Figure 8-2. Early Yiddish typewriter
The advance of typing from a mechanical hardware-based enterprise to
a digital software-based one on the computer, however, enabled multiple
standards to co-exist on a single machine. Yet the essential mapping task
remains: the software must translate the stroke of a Roman-letter key to a
non-Roman character on the screen.21 This has proven to be a real and
present issue in my own computer-based manipulations of Hebraicized texts.
Working on a Macintosh, I have made use over the years of several systems
21
Of course this does not apply to computer keyboards produced for the Israeli market,
where the Hebrew letters themselves are pictured on the keys (see figure 7-3). In fact, such
keyboards might raise the opposite but precisely parallel issue of mapping these Hebrew-
letter keys to Roman characters on the screen.
422
for manipulating Hebrew-letter material, two of which reveal particular
approaches to the task of script adaptation.
4.1. System software vs. Stand-alone font
Under the older Macintosh OS 9 system, a "Language Kit" could enable
a Hebrew-friendly word processor called Nisus Writer to reconfigure the
keyboard in various ways and to display output from right to left (which a
simple keyboard-to-font mapping would not necessarily accomplish22). The
scheme by which the software maps the Roman-letter keys onto the screen as
Hebrew letters is all the more direct a mapping because it is in principle
independent of the language being typed, a strictly graph-to-graph
correspondence. The various mappings are not entirely uniform, varying
from font to font and from one keyboard layout to another. As in manual
adaptations of the script, however, certain conventions are shared across the
different programs, while others are in competition.
Nisus Writer makes use of two keyboard layouts installed by the
Language Kit: one reproduces the layout of a standard Israeli Hebrew
keyboard (see figure 7.3), while the other, named "Hebrew QWERTY",
approximates the standard American layout.23 While the Israeli layout does
not seem to depend on the Roman QWERTY in any obvious way (and so has
proven less than useful to this user), the Hebrew QWERTY represents a curious
blend of mapping relationships, especially in comparison to other stand-alone
Hebrew fonts that operate independently of the Language Kit software.
22
Note that although the output appears on-screen in the appropriate "direction," the
mediating effect of the keyboard is such that unlike handwritten language, the input is only a
temporal succession of keystrokes and has no inherent spatial orientation.
23
There is also a "Hebrew AZERTY" that does the same for that European standard; a "Hebrew
DVORAK" exists as well, though not as part of Apple's Language Kit.
423
Figure 8-3. Israeli Hebrew keyboard layout
Presented below is a sample of the Hebrew letters that result from
hitting a given Roman key under the Hebrew QWERTY layout. I label the first
category phonetic, since it relies only on the single common sound
conventionally associated with the two letters:
Table 8-6. Phonetic mapping
T24 ‡ tt
S ‡ ss
The second category is labeled phonological because the Hebrew grapheme,
which has stood historically for more than one sound in Hebrew itself, can be
produced by typing any of the associated Roman keys:
24
Since the characters on the keyboard are depicted only in uppercase, I have used that form
in these formulas. As such they are exactly equivalent to "T ‡ <t>" or "Shift+T ‡ <T>."
424
Table 8-7. Phonological mappings
K, X ‡ kk
O/U/V‡ ww
The third category is called iconic because the only relation between the
Hebrew letter and its Roman keystroke appears to be a graphical (perhaps
historical) resemblance in form:
Table 8-8. Iconic mappings
W ‡ S ¸s
Y ‡ X†
I call the last category hybrid because these mappings draw on a combination
of "higher-level" associations between the graphemes:
Table 8-9. Hybrid mappings
A ‡ '√
E ‡ v fi
C ‡ cß
J ‡ x˛
The motivation for the equation of A and ' , for instance, is an ancient
graphical lineage is, which manifests itself most saliently in both as "first letter
of the alphabet." In contrast, while the relationship between E and v could be
construed as vaguely iconic (and does have a historical link via Greek h),
readers of Hebrew script are likely familiar with the use of this letter in
Yiddish to represent the vowel /e/. For its part, the equation of C with c
425
appeals to readers of Roman-letter writing systems in which <c> can spell the
sound [ts], which is the Modern Hebrew realization of c . It is somewhat
harder to discern a relationship between J and x that might be relevant to the
target audience of users; my best guess is based on the spelling of /x/, also
the Modern Hebrew reading of x , with <j> in some orthographies (e.g.
Spanish).
Several of the mappings in the Hebrew QWERTY layout contrast starkly
with the stand-alone font Ezra,25 produced by the Summer Institute of
Linguistics. The following, for example, draw more directly on traditional
transliteration practice (cf. table 2-1 in chapter 2):
Table 8-10. Transliteration-based mappings
’ ‡ '√
W ‡ ww
Y ‡ yy
Q ‡ qq
Other mappings are based on more specific linguistic properties:
Table 8-11. Phonology-based mappings
X ‡ x˛
shift-X ‡ X†
The assignment of x to the <X> key makes sense from a grapho-phonological
point of view (cf. IPA [x], the Modern Hebrew realization of x), and with t t
25
This is the principal font that I use with the Hebrew-unfriendly Microsoft Word, which as
yet does not enable right-to-left output, forcing the typer to spell words and enter sentences
from back to front.
426
already assigned to the <T> key, the programmers chose to assign the
historically-pharyngealized dental stop to the same key as another historical
pharyngeal,26 augmented by the shift key. Some "lay" associations do persist,
however, such as the mapping of <V> to the historically pharyngeal (now
glottal or vocalic) v, which would appear to be based only on the vaguely
iconic resemblance between the two symbols.
Indeed, many of the augmented keystrokes, i.e. those involving the
shift, ctrl, option, or command ("open-apple") keys, reveal different, sometimes
obscure strategies behind SIL Ezra and Hebrew QWERTY . For example, the
final-form letters are produced in Ezra by holding the option key and typing
one of the digit keys; in Hebrew QWERTY they are produced by hitting
shift+<letter>, just as one would do to type a capital Roman letter.27 In
addition, hitting the Roman vowel letters (alone and with the shift key) in SIL
Ezra yields the diacritics of niqqud, as determined by their Modern Hebrew
realization. Perhaps these were considered less important for users of
Hebrew QWERTY, where vowel diacritics are all produced by complex
keystrokes, some of which seem to have been arbitrarily drawn from the
otherwise unassigned keys – shift-R, for example, yields one of the sub-linear
/a/ symbols, while command-4 yields the only sub-linear /i/. Of course, some
elements of niqqud, such as dagesh or rafeh, have no analogue in Roman script,
and so their position on the keyboard would seem justifiably arbitrary.
26
Hebrew QWERTY assigns X to the <Y> key while assigning the <X> key to the sibilant s. See
tables 7.6 and 7.7 for other mappings under Hebrew QWERTY that similarly lack obvious
linguistic motivation.
27
Sampson (1985: 84) compares the use of the five final-position allographs to so-called
"swash" letters in some italic fonts. Given the typesetting fact above, however, it may be
more apt to compare them to the use of capital letters in written German, that is, as obligatory
position-dependent allographs. In fact, since the same keystroke (shift-M) produces both the
initial capital in, for example, <Mann> 'man' and the final form in £d' 'human', from the point
of view of the keyboard mapping they do in fact serve the same function.
427
4.3. Transcription vs. Transliteration redux
Below is presented a full comparison of SIL Ezra (middle) vs. Hebrew
QWERTY (bottom), which allows one to see the full script-adaptation strategy of
the programmers. To produce the Hebrew character in the top row of the
table, type the corresponding key:
Table 8-12. Hebrew graphemes from Roman keystrokes
‘ b g d h w z x X y k l m n s v p c q r S t
t c d s v u z j y h f k n b x g p m e r a ,
¢ £ § • ¶ A E a e I I u Ã Œ œ › ◊ K &
¢ £ § • ¶ A E a e I I u Ã Œ œ › ◊ K &
l o i . e r
These tables encapsulate the adaptation of Hebrew script from the perspective
of the matrix – that is to say, they map the characters selected for use from
ther matrix onto the unit of the target (the computer keyboard), answering
the question of what keystroke will produce a given Hebrew character. As I
have emphasized elsewhere, however, it is possible, and indeed practical, to
view this enterprise from the opposite perspective, i.e. as adapting the units of
the target (normally the sounds/segments of the language being written, in
this case the Roman-letter keys) to the units of the matrix (the characters of
the script adopted as a vehicle for writing, in this case the Hebrew characters
that appear on-screen), which actually remain relatively immutable in the
adaptation process. In this sense, the keyboard mapping can be reevaluated

428
by considering what Hebrew grapheme to assign to the available keystrokes
(SIL Ezra middle, Hebrew QWERTY bottom):
Table 8-13. Roman keystrokes ‡ Hebrew graphemes
q w e r t y u i o p Q W E R T Y U I O P
q w e r t y u i o p Q W E R T Y U I O P
/ # q r ‘ X w § £ p Ù
a s d f g h j k l ; ‘ A S D F G H J K L : “
a s d f g h j k l ; ‘ A S D F G H J K L : “
S d g k v y x l ¢ • H l
z x c v b n m , . / Z X C V B N M < > ?
z x c v b n m , . / Z X C V B N M < > ?
z s b h n m c t ¶ .
What these two sets of tables capture, in fact, is the crucial difference between
transcription and transliteration. Table 7-6 begins with the matrix (i.e. Hebrew)
graphemes that are cognitively useful for the purpose at hand, and uses them
to represent "as best they can" items of the target (i.e. the Roman keyboard
characters). The result is an arrangement of matrix graphemes that
approximates the spelling of target-language forms, much like a transcription
as defined in chapter 2. In table 7-7, by contrast, the starting point is the target
form itself: every available unit (i.e. the entire keyboard) is assigned a
character of the matrix until all those selected for use from the matrix have
been mapped. The result is a one-to-one, grapheme-to-grapheme
transliteration of target-language forms into an adapted-script sheath.
As the foregoing discussion illustrates, the adaptation of scripts and the
contact of Hebrew script with non-Hebrew languages is not merely an

429
esoteric medieval phenomenon, nor is it a problem reserved for cross-
linguistic bibliographers (as the uniqueness of Wellisch's book might have us
believe). In many parts of the world – including our own ostensibly mono-
alphabetic culture – language users may have to negotiate multiple scripts or
multiple sets of conventions in all sorts of subtle ways. In this sense, many
more people might be considered "multilingual" (or at least "multigraphic")
than would seem given traditional definitions of the term. Of course in the
grand scheme of things linguistic, computer users are still a relatively
privileged few, a specially-trained group of language users. Yet the
"problems" they encounter often reflect a history of linguistic interaction
repeating itself, one that may highlight issues yet to surface in more
conventional contexts. For that matter, as computer environments continue
to encroach on and create new forms of linguistic interaction, it may be the
computer-related issues that in time turn out to constitute the "conventional"
contexts for writing systems in contact.
5. A FINAL THOUGHT
In the preface to his book, Wellisch (1978: vii) explains that it is not his
aim to add to the literature about how script conversion is performed, but "to
explore why script conversion has been performed at different times, and
what effects it had on those who were exposed to the results." If we were to
ask our Portuguese writers this question – why they adapted the Hebrew
alphabet to write their Romance language and for what effect – we might get
little more than a puzzled look. They might find it mysterious that the
vernacular rubrics in a modern American ma˛zor are written in a script entirely
different from that of the blessings and other rituals. Since their audience
430
could read and perhaps write two other languages already presented in this
script (Portuguese and Hebrew itself), there was in fact strong disincentive to
write in the Roman alphabet.
Thus a certain paradox emerges from this study. While arguing against
the overall markedness of the Judeo-Portuguese corpus (i.e. that it required
more effort from its writers to produce or from its readers to process than did
the more conventionalized, Roman-based adaptation that normally bears the
title "written Portuguese"), what has drawn me into it is its very markedness.
It is worth asking again whether, in the absence of more decisive
distinguishing features, the non-Roman alphabet gives these texts an entirely
different linguistic identity. In other words, does the very act of writing in a
non-canonical, non-traditional, or non-conventional script – let alone
orthography – suggest that the writer attributes to his written language an
identity distinct from what others write using the dominant script?
Recall the argument made by Wright (cf. § 3) that the conceptual
division between Latin and Romance, between classical and vernacular
languages, emerged precisely as a result of re-assigning a set of orthographic
conventions. Wright argued that what is now known as "medieval Latin" was
created when the conventional correspondence between pronunciation and
spelling was expropriated by the ninth-century Carolingian reforms. As
clergymen were instructed to give each letter a unique sound, Latin
orthography was usurped from its role as a conventionalized sound-to-
symbol mapping, and was adopted instead as, in effect, a transcription. For
example, VIRGINEM would no longer be read in Gallo-Romance regions as
[vyergÊ\], but rather as [wirginem] or [virginem] – as though we were
instructed to read modern English <through> not as [†ru] but as [t˙rowg˙].

431
Such a move would oblige us to devise new rules to spell /†ru/ and the like,
creating what would likely be perceived as a "new" (written) language. For
the medieval Romance readers, a new psychological distinction between Latin
and vernacular was created when orthographic forms were assigned a new
pronunciation, so that literate required a new set of letter-sound conventions
in order to write their language. Wright claims that this new set of
conventions – that is, written Romance – "was developed by skilled Latinists
and not, as handbooks still tend to imply, by people who could not cope with
traditional written forms" (1997: 265). Hence the real continuity of convention
is to be found between the resulting vernacular spellings and the pre-reform
orthography, since both of these served to represent what the writers
considered simply "their language."
For their part, the lusophone Jews who left Portugal in the wake of the
1496-97 edicts could maintain and reinforce (at least for a time) their
ethnic/religious identity through their vernacular, which was categorically
different from those that surrounded them in their new diaspora. Previously,
in Portugal itself, the Hebrew alphabet could also serve this function, as a de
facto mark of difference. Yet rather than indicating the "partial detachment
from its environment that Blau (1999), for example, sees in the use of Hebrew
script by medieval Judeo-Arabic writers, the use of Hebrew script by Jewish
Portuguese writers is a striking illustration of their capacity to embrace the
mainstream language and culture on their own terms – a linguistic convivência
in which they straddled the boundary between the commercial and the
religious, the Portuguese and the Hebrew, the Christian and the Jewish.

Old Portuguese in Hebrew Script

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Old Portuguese in Hebrew Script

Uploaded by

Copyright:

Available Formats

OLD PORTUGUESE IN HEBREW SCRIPT:

CONVENTION, CONTACT, AND CONVIVÊNCIA

Devon L. Strolovitch, Ph. D.

This dissertation explores the process undertaken by medieval writers

Devon was born and raised in Montreal (Canada), where he learned

I came to Cornell directly from Oberlin College, where I had spent my

hesitation my mind turned to Judeo-Romance. I stumbled onto references to

and "call it a dissertation." Since then, as chair of my committee, Carol has

of the Cambridge University Library and the Brotherton Library at Leeds

Spanish inspired me to try the same in my French teaching at Cornell.

Biographical Sketch iii

Appendix: Facsimiles 432

A-1. Facsimile of Parma ms. 1959, f. 1r. 432

2-1. Paleo-Hebrew and "Jewish" scripts 19

3-1. Final /a/ allography 95

7-1. Variation in word-final nasal spelling 353

8-1. Pseudo-Yiddish orthographic calques 414

Ar. Arabic OED Oxford English

While living in France several years ago, I suffered an identity crisis of

grade French teacher who, with similar difficulty pronouncing my name à

letter-by-letter transliteration. To my eyes, however, something was amiss.

derivative of or secondary to Roman-letter writing, nor entirely dependent

analysis of the Judeo-Portuguese writing system in Chapter 3. The three

summarize the main contributions of this corpus to Portuguese historical

2. WRITING'S "SECOND MOVEMENT"

active terrain, with many English-language book-length studies on writing

3. BORROWING, WRIT LARGE

something about the interaction of Slavic and English spelling conventions

3.2. Ad-hoc nativization

language that I had learned using the Hebrew alphabet.

3.3. Borrowed scripts

markings to capture distinctions that they do not in their conventional context

3.4. Adapted scripts: permanent vs. transient

the language of the newly-dominated population usually comes to be written

since de-Cyrillicized them. And in the most intriguing cases, ethnic/religious

Jewish community (or communities) in adjacent Portugal, though records of

What makes these documents of such high interest and importance is

3.5. Hebrew script for non-Hebrew writing

more tradition-laden one, as Paper (1977: 103) explains in reference to

It is an interesting and fascinating fact that the half-dozen earliest bits of

valued as more phonetically transparent than vernacular spellings found in

There is a shortage of vernacular [medieval] Greek texts that endows

any Hebrew words found in Hebraicized writing must be considered

In his study, Wellisch (1978) devotes a short section to "Hebraification,"

These are indeed apt observations. What he appears to mischaracterize,

orthographic dexterity they developed that persuaded Jews to write in

Furthermore, the unconventional alphabet is the only markedly Jewish feature

FROM ARAMEA TO AMERICA:

In this chapter I present an overview of the development of the

Hebrew writing system, followed by a survey of language families with

attested Hebrew-letter texts. While I aim to provide a broader and more

inclusive overview of Hebraicization than has been available previously, I do

not make any claim to comprehensiveness.

1. FROM HEBREW TO JEWISH WRITING

Nothing is known of Hebraic writing before the Israelites emerged in

that of tenth-century Phoenician inscriptions from Byblos, and features no

specifically Hebrew characters. Indeed, the Phoenician influence was so

characters to represent consonant phonemes that did not exist in Phoenician.

The first distinctive features of Hebrew writing are actually to be found

in ninth-century inscriptions in Moabite, a Canaanite dialect related to

Hebrew. According to Naveh (1982), these adaptations of the contemporary

Despite dialectal differences between the spoken Hebrew of Judah (the

cursive lapidary style, due perhaps to the lack of a widespread custom in

Hebrew society of erecting royal stelae or offering votive inscriptions to the