Maimonides [Yad Tefillah 8:12, 15:1], as well as several other Sephardic scholars, [declares] to be ‘illegin (=defective of speech) [those] people who cannot distinguish between the sounds of aleph and ayin or between the sounds of heh and heth. These alone they declare ‘illegin. But our Talmudic sages, when they cited these two pairs of easily confounded gutturals, were citing them merely as examples as is shown by their use of the word kegon (=such as)—a word which always implies that what has been mentioned represents a larger group.[i] Hence I am amazed at their [i.e. Maimonides and the Sephardic scholars] singling out for the epithet ‘illegin just those who fail to distinguish between aleph and ayin, etc. but forget to apply it to themselves and their countrymen who make no difference between the sounds of samekh and tsadi. Moreover, when it comes to the diacritics—which are to the letters like brains and legs [to humans]—they do not respect each diacritic’s phonetic value. Instead, kamets and patah are all one to them as are tsere and segol.… All this happened to them because they fulfilled the verse [Ps. 106:35] “They intermingled with the nations and learnt their ways.” Having resolved to aggrandize themselves above their fellows, they made every effort to gain admission into royal and princely courts. And the better to ingratiate themselves with the princes, they took up the study of these uncircumcised princes’ tongue, script, astronomy [or science], and philosophy…. Furthermore, they sought to bring their own language [Hebrew] into line with the language of the uncircumcised by retaining only those five of our vowel sounds that correspond to the latter language’s vowels while doing away with all the rest. Misguidedly the [Sephardic] multitude followed their lead until in time all but the five vowel sounds were lost to those communities. Another consequence of the philosophical studies was—for our sins—the proliferation of heretics in Israel.[ii]
The above diatribe leveled against what we think of as Sephardic pronunciation came from the pen of Asher Lemlein ben Meir Reutlingen. This all but forgotten visionary—a messiah to some—appeared on the scene on Izola in Istria at the beginning of the sixteenth century. Contemporaries, both Jewish and Christian, recall 1502 as the “year of penance” when masses of Jews divested themselves of their worldly possessions in preparation for what Asher Lemlein had led them to believe was their imminent redemption.
Ephraim Kupfer who published the surviving writings of Asher Reutlingen,[iii] quotes several such reports and assessments of Asher’s impact, by chroniclers both contemporary and slightly later—including Abraham Farissol (d. 1525). In his book Magen Abraham, Farissol writes:
In these regions of Italy, in the Venetian domains[iv] there arose a man of stature[v] from the ranks of Ashkenaz by the name of Asher Lemle.[vi] He put on airs of being a king despite his limited wisdom and deeds. Through the mediation of his disciples he misled the entire region [into believing that] the redeemer is coming. Indeed, to the multitudes he would announce that “he [the redeemer] is already here.” From his place of seclusion he let most of the Diaspora come to believe in him, his teachings, the fasts and flagellations; for they said “the redeemer is here!”—until it all ended in “emptiness and chasing the wind.” These events played out before me in the year 262  here Ferrara where I reside.[vii] [DEA1]
A generation later the historian Joseph haKohen (d. 1577) records in his ‘Emeq haBakhah:
In Istria, which is near Venice, there arose an Ashkenazic Jew by the name of Lemlin—a fool of a prophet a madman in spirit.[viii] Jews flocked to him saying “he is surely a prophet since God has sent him to lead His people Israel and to ingather the scattered of Judah from the four corners of the earth.” Even among the rabbis he had some followers. They called for fasting, wearing of sackcloth, and for everyone to repent of their bad ways; for they said “Our redemption is close at hand[DEA2] .”
The recollections of David Gans (d. 1613) are charming—if second-hand:
Rabbi Lemlin announced the coming of the messiah in the year 260 . Throughout the dispersions of Israel they believed his words. Even among the gentiles his fame grew and many of them also believed his words. My grandfather Seligman Gans of blessed memory smashed the oven he kept for baking massoth in his total confidence that the following Passover he would be baking massoth in the Holy Land. I myself heard from the venerable Rabbi Eliezer Trevis, head of the Francfort beth din, that it was no trifling matter[ix]—[Asher] having provided signs to prove it. He [R. Trevis] added “perhaps our sins were the cause of its failure[DEA3] .”
Lastly, the remarks that the Christian protagonist addresses to his Jewish counterpart in haVikuah by the famous Hebraist Sebastian Münster (d. 1552):
In the year 262  Jews did penance wherever they lived in all lands throughout the diaspora in expectation of messiah.[x] It continued for almost a full year; young and old, children and women. Never had such penance been done as was done in those days.[xi][DEA4]
Asher Lemlein is certainly fascinating in his own right; but our present interest is his conviction that seven diacritic signs must represent an equal number of distinct vowel sounds. Fewer sounds than signs made no sense to Asher. His logic seems perfectly cogent, and was to be echoed by other worthies until the dawn of the modern age. As late as the second half of the eighteenth century, R. Jacob Emden (d. 1776) was faulting the Sephardic vowel system:
[W]ith regard to the pronunciation of the vowels, happy are we [Ashkenazim] and goodly is our portion unlike the Sephardim who do not distinguish between kamets and patah, thus making the holy profane[xii] .… In addition to that, they diminish the number of the vowels which were handed down to us from Sinai…. They do the same with the vowels segol and tsere, making the pronunciation of both alike.[xiii]
Emden’s allusion to the vowels’ Sinaitic origin is cryptic; but almost certainly harks back to a talmudic passage in Nedarim 37b.
What is the interpretation of the verse “They read in the scroll of (var. in)[xiv] the Torah of clearly they made its sense plain and gave instruction[xv] about what was read” [Neh 8:8]? “They read in the scroll of the Torah of God” this refers to Scripture proper; “clearly” refers to Targum [=Aramaic translation]; “they made its sense plain” refers to the division of the text into verses; “and gave instruction about what was read” refers to the cantillation—or, according to others, to the masorot. R. Isaac said: The reading of the Scribes, the embellishments of the Scribes, words read but not written or written but not read are all halakhah le-moshe mi-sinai.[xvi] Examples of “readings of the Scribes” are the two ways of pronouncing the consonantal word spelt aleph resh tsadi [=earth, land]. Also, the consonantal word spelt shin mem yod final-mem [=sky, heaven] and the word spelt mem tsadi resh final-mem [=Egypt].[xvii]
Although R. Isaac obviously attaches the highest importance to giving each vowel its proper phonetic value, he says nothing about seven vowel sounds—let alone any diacritical sigla. Nevertheless, both R. Emden and Asher Lemlein, the former explicitly, assume the seven diacritics along with their respective values to be ancient, if not coeval with the biblical text itself. Nor were these teachers alone in that assumption. Indeed, some Sephardim showed symptoms of an inferiority complex on account of their indifference to the kamets! For example, R. David Ibn Yahia (d. 1528) makes the following confession: “Know that we [Sephardim] have lost the proper way to read written texts…. We do not differentiate between kamets and patah nor between tsere and segol …. Undoubtedly each consonant and each vowel must have its discrete sound….”[xviii] Even today one occasionally hears the argument that neglecting to differentiate between patah and kamets or segol and tsere must surely be a deviation from what was intended by the tradition that instituted these distinct sigla. For the sake of full disclosure, I own up to my own bewilderment regarding this seeming anomaly of having two distinct “squiggles” to represent one and the same sound. When I finally mustered the courage to ask my father, he proceeded to show me a text with supralinear Babylonian vocalization. Today, he said, we know that the Babylonian system of vocalization differed radically from the Tiberian, and certainly did not assign distinct values to tsere and segol—and possibly not even to patah and kamets.[xix] However, the Tiberian system won the day and ousted the Babylonian—at any rate among scribes and writers of vocalized Hebrew. But not in the mouths of entire communities who retained their erstwhile pronunciation, either through inertia or in conscious defiance of the “officially” sanctioned system.
My father’s answer was no more than a distillation of a century of discovery and scholarship that has identified not merely two but three historical systems of vocalization. Some of the most accessible scholarship in the field can be found in the writings of pioneers such as Benjamin Klar (d. 1948), Paul Kahle (d. 1964), Yehiel F. Gumpertz and in the ongoing research of Israel Yeivin and others. These are some of the primary scholars whose conclusions we shall now summarize, paraphrase and/or cite.
From the very beginning of the enterprise of vocalizing the sacred texts—i.e., from the Gaonic age—there existed three distinct systems.… It is premature to say what the historical relationship between the three systems might have been. But it would not be unreasonable to conjecture that the so-called “Egyptian-Sephardic” pronunciation was the most ancient since it is attested in the transcriptions of the Septuagint as well as Josephus.[xx] If so, the Tiberian and Babylonian systems must be due to later influences. It is worth noting comparable phonetic developments in Persian where the long ‘a’ sound mutated into a Swedish ‘å’.[xxi]
When in the course of the ninth century the Masoretes of Tiberias began their work of adding a consistent punctuation to the text of the Hebrew Bible, they were convinced that it was their duty to give the text of the Bible as correct a form as possible.… They secured the abolition or adaptation of all the texts provided with a different kind of punctuation such as the Babylonian.… The text fixed by the Masoretes has been almost the only one considered in the preparation of our Hebrew grammars. Now we know this text was altered by the Masoretes. I have tried to show that the Masoretes of Tiberias introduced a number of new vowels to safeguard the newly-established pronunciation of the gutturals.[xxii]
The well known report in Mahzor Vitry regarding the existence of three systems of pronunciation appears to be taken from a compilation by the twelfth century R. Jacob bar Samson. That report, found in the commentary to Pirqe Avoth, reads: ‘Therefore Tiberian punctuation differs from our punctuation, and both differ from the punctuation of the Holy Land.’[xxiii] M. Friedlander thought that ‘our punctuation’ referred to the Babylonian system. To the objection that a 12th century Frenchman was unlikely to identify his group as Babylonian, Friedlander responded that Vitry’s commentary to Pirqe Avoth was a miscellany of material borrowed from a variety of sources, including Gaonic, which the compiler incorporated as he found it. Nehemiah Aloni rejected Friedlander’s theory, preferring to understand ‘our punctuation’ as referring to the ‘expanded’ Tiberian punctuation…. If so, Vitry cannot be counted as a witness to Babylonian vocalization.[xxiv]
All agree, then, that the system we are most familiar with, originated in Tiberias and comprised seven diacritics. The system that developed in Babylonia probably had no more than six. A third system, often referred to as the vocalization of Erets Yisrael, seems to have had just five. Although the Tiberian system with its seven sigla ultimately prevailed, not all communities renounced their traditional way of pronouncing Hebrew. This can be demonstrated in a number of ways. For instance, a plethora of extant manuscripts can be seen to disregard the quintessentially Tiberian vowel distinctions; interchanging kamets with patah and sere with segol. Many of these old manuscripts would have shocked the messiah of Istria because they hail from the very heartlands of Ashkenaz.
Yes indeed! Careful study by scholars, notably Hanokh Yalon (d. 1970),[xxv] of early French and German manuscripts showed that their writers, too, were pronouncing kamets the same as patah. Take for example the comments of Rashi (d. 1105) to the “Earth, Heaven, Egypt” passage at Ned. 37b (cited above). Since the Talmud is typically written without matres lectionis, Rashi sets out to describe in his own words the sound of nouns such as ERETS (=earth) and their pausal modifications. “It is the “readings of the Scribes” that fixes the two ways of pronouncing the consonantal word spelt aleph resh tsadi. For there is no yod between the aleph and resh nor between resh and tsadi [to fix the pronunciation as ERETS]. Similarly for the pausal form, there is no second aleph or heh between the aleph and resh nor is there a yod between resh and tsadi [to fix the pronunciation as ARETS].” By explaining that the pausal is pronounced as if there were a mater lectionis aleph or heh between the initial consonantal aleph and the resh, Rashi reveals that the kamets was just like patah in his own system of pronunciation.[xxvi]
Another important proof is furnished by transcriptions of Hebrew in European alphabets. In 1273 R. Abraham Ibn Ezra’s astrological treatise Reshit Hokhmah was translated into French.[xxvii] Yehiel F. Gumpertz in his Mivta’e Sefatenu (Jerusalem 1953) analyzed the transliterated Hebrew words in this thirteenth-century Old French text. Gumpertz begins by telling us that the Hebrew (and Arabic) words were dictated to the scribe Obers de Mondidier by Hagin the Jew. The latter could not write French and the former knew no Hebrew (or Arabic). “The first thing to emerge [from my study of this text],” Gumpertz continues,
was a total and unqualified confirmation of Hanoch Yalon’s theory regarding the “Sephardic” pronunciation of the kamets by French Jews. Indeed so “Sephardi” are his transcriptions that I began to suspect Hagin to be an Iberian Jew. However, his non-Sephardic origin was soon revealed in the way he represents shevas and hatafs, no less than in his transcriptions. For instance, the Hebrew word for myrtle he gives as hedas instead of hadas. Hedas is attested exclusively in non-Sephardic MSS of the period. (Gumpertz, ibid.)
A third clue comes from rhymed Hebrew compositions by early French and German versifiers. Very frequently kamets and patah words are used to form the rhymes, strongly suggesting that the rhymsters treated them as homophonous.
But to gain a fuller picture of Ashkenazic pronunciation and its evolution, we turn now to—of all unlikely linguists—Max Weinreich. Weinreich’s Yiddish researches necessitated a thorough understanding of the kinds of Hebrew that fed Yiddish at its various stages. Not only did Weinreich (d. 1969) master the evidence available in his day, but he managed to present it in a manner succinct as it is orderly. Indeed, we cannot do better than quote him in extenso.
Up to a hundred years ago, not only the reading of the Bible, but all of Hebrew grammar was based on the Tiberian tradition. There are statements of medieval authors that the pronunciation, along with the text of the Torah, were given on Mount Sinai. Aharon Ben Asher [early 10th century] himself maintained that punctuation derived from the men of the great assembly, namely from the beginning of the second Temple. Still others, more critical, came to the conclusion that Hebrew speakers in the period of unmediatedness needed no punctuation.… The Tiberian punctuation was created with the conscious aim of teaching correct reading at a time when Hebrew had long ceased to be an unmediated language.… Scholars can now declare with sufficient confidence that of the three attempts to elaborate a punctuation, the Tiberian attempt was the most recent. The Babylonian system apparently came into use around the year 600, the southern Palestinian[xxviii] about 700, that is some 50 years before the work of the Tiberian sages had begun….
Behind the north Palestinian punctuation there was an inventory of seven vowels whereas the southern Palestinian punctuation has an inventory of only five vowels. One fact is striking; this vowel system is similar to what was later called the Sephardic pronunciation…. From southern Palestine and Egypt it [the five vowel system] penetrated all of northern Africa and even the Iberian Peninsula. The centre of learning in Kairwan was also a point of supply of Jewishness to Italy…. From there it passed into Loter-Ashkenaz…. It was one exclusive Western sphere, from southern Palestine to the Atlantic, from the edge of the Sahara to the northernmost settlements in central Europe. The southwestern sphere retained the five-vowel reading system [while] the northwest, that is, central Europe, was pervaded by the Tiberian; through conscious efforts of the adherents of this system there grew up here what is known as the Ashkenazic pronunciation… The similarity of the pre-Ashkenazic pronunciation in Ashkenaz to the Sephardic pronunciation was not the result of the influence of Sefarad on Ashkenaz. There was no such influence, but both Sefarad and Ashkenaz drew their spiritual sustenance from one pre--European source. Sefarad clung to the old system; Ashkenaz changed its reading system radically and the break came not because the scholars of Ashkenaz created the Ashkenazic pronunciation ex nihilo… but by virtue of external prestige.
In the writings of the Rosh, born in Ashkenaz about 1250, we find the same as in the case of Rashi's grandsons: the kamets symbol was called a patah. But [soon] there begin to appear in Ashkenaz signs of the northern Palestinian system, and towards the end of the 14th century Ashkenazic Hebrew manuscripts are usually pointed according to the Tiberian style. [Nevertheless] Ashkenazic Bible manuscripts of the 13th, 14th, and a few perhaps even from the 15th centuries have also been preserved that … can be understood only in the light of the southern Palestinian reading. Some of these manuscripts have a patah instead of a kamets and a kamets instead of a patah; similarly a segol instead of a tsere …. A second group of manuscripts have only patah and segol…. Such confusion and such interchange is conceivable only in the case of punctuators whose vocalic value of patah and kamets on the one hand, and segol and tsere on the other, differs from the Ashkenazic pronunciation of today.
Since it is a matter of proving that today's reading in Ashkenaz is not the original one, the question of how far back the Ashkenazic pronunciation was demonstrably the same as it is today has to be raised. The answer is about 1500; that is, since the beginning of the middle Yiddish period the situation has been more or less the same as today. In the last quarter of the 15th century the Ashkenazic value of the kamets is confirmed by both Jewish and non—Jewish testimony.… Up to the 13th century there are no indications of “Ashkenazism”….The oldest known instance of a kamets with the value ‘o’ is in a Cologne Hebrew document dated 1266.[xxix]
If there has to be a moral to this story of phonetic vicissitudes, let it be this: No Jewish community need deem its own tradition for pronouncing Hebrew superior or inferior to any other phonetic tradition. Doubtless those Sephardic authors who expressed misgivings about their neglect to respect kamets or segol would have been relieved to learn that their ‘neglect” was justified all along. Nor should the antiquity of such linguistic heterogeneity surprise us when we ponder the shibboleth–sibboleth dichotomy of Jephtha’s day. “The Gileadites held the fords of the Jordan against the Ephramites. When any fugitive of Ephraim said, “Let me cross” the men of Gilead would ask him, “Are you an Ephramite?”; if he said “No” they would ask him to say “shibboleth” but he would say “sibboleth” being unable to pronounce it correctly” (Jud 12:5–6).
In his commentary to these verses, R. David Kimhi (Radak, d. 1235) actually compares the phonetic differences between Gileadite and Ephramite to a situation in Europe of his day: “Just as they would test the Ephramites with this word shibboleth, they would likewise test them with any word that had the letter shin; shibboleth serving merely as an example.… Perhaps it was the climate that influenced their discrete pronunciations in the same way that the people of sarfat [=France] are unable to make the ‘sh’ sound but rather pronounce it as a soft tav.”
[i] See Meg. 24b; Yer. Ber. 2:3 [4d] (although the word kegon does not appear in either source).
[ii] “The Visions of R. Asher b. R. Meir Lamlein Reutlingen” (Heb.) by Ephraim Kupfer, Kobez al Yad vol. viii (xviii) Jerusalem 1975, pp. 387–423.
[iii] “The Visions of R. Asher b. R. Meir Lamlein Reutlingen” (Heb.) by Ephraim Kupfer, Kobez al Yad vol. viii (xviii) Jerusalem 1975, pp. 387–423.
[iv] Istria belonged to the Venetian Republic from 1267 until the eighteenth century.
[v] Ish haBenayim (cf. 1Sam 17:23).
[vi] A variant of Lemlein which is, in turn, a diminutive of the German for “lamb.”
[vii] For a fuller appreciation of Farissol, see The World of a Renaissance Jew: The Life and Thought of Abraham ben Mordecai Farissol by David B. Ruderman, Cincinnati 1981.
[viii] Cf. Hos 9:7.
[ix] Heb. davar req see Dt 32:47.
[x] The original Hebrew reads “ekh mashiah yabo.” Ekh‘s basic meaning is “how.” In a non interrogatory sense it occurs in stock phrases such as “ekh habahur” (in the text of the kethubah). It must also be borne in mind that Ha-Vikuah’s Hebrew is not exactly standard. The context, however, leaves little doubt as to Münster’s (or rather his protagonist’s) intent. Lemlein is also mentioned (derisively of course) by Johannes Pfefferkorn (d. c. 1522) in his Der Juden Spiegel (see The Jewish Messiahs by Harris Lenowitz, Oxford 1998, pp. 99–101)
[xi] Basle 1529 (or 1534. Kupfer gives the date as 1534, but Ha-Vikuah’s preface is dated ‘Tishri 290’ which equals September–October 1529).
[xii] The Hebrew word adon means master, ruler or lord. With the letter yod added as suffix it could mean either “my master” or “my masters” depending on the vocalization of the nun. A hiriq under the nun indicates that the suffix is singular (adoni) as in Gen 33:8, 13; Num 11:28. But when the word is not in the singular, the Tiberian masoretes further distinguish “sacred” from “profane” by pointing the former with a kamets under the nun and the latter with a patah. Thus at Gen 15:2,8 where Abraham is addressing Hashem the nun is pointed kamets; while in Lot’s address to the angels at Gen 19:2 it is pointed patah. Now unless the reader distinguishes kamets from patah, the contrast between “sacred and profane”—as intended by the Tiberian vocalizers of the Bible—is lost. Sephardic pronunciation invites the criticism of R. Emden insofar as it ignores that contrast, thereby “making the holy profane.” R. Emden’s criticism is endorsed by R. Yitzhak Yaakov Weiss (d. 1989) in his Minhat Yitzhak 3:9 and discussed most insightfully by Dr. Isaac B. Gottlieb in “The Politics of Pronunciation” AJS Review 32:2, pp. 360–62. I herewith thank R. Alex Kaye for bringing this and related sources to my attention.
[xiii] Siddoor Beth Ya’aqob; translation based on H. J. Zimmels’ in his Ashkenazim and Sephardim London 1976, p. 86. For other renderings from Hebrew, this article employs a blend of standard and our own translations.
[xiv] The Talmud (both at Ned. 37b and at Meg. 3a) reflects a Hebrew Vorlage be-sefer torat ha-elohim whereas our biblical text reads be-sefer be-torat ha-elohim.
[xv] In late Biblical Hebrew HBN often denotes “causing others to understand.”
[xvi] Literally: “an oral law (or tradition) to Moses from Sinai.” However, the phrase’s precise connotation is disputed.
[xvii] Since the biblical books are traditionally written without diacritics, the word formed of aleph- resh- tsadi allows of various pronunciations. We depend on “tradition” to tell us that the word is ERETS—except in its pausal form which is ARETS (or ORETS).
[xviii] Leshon Limmudim 1:5, 1st edition, Constantinople 1506.
[xix] Because of the extreme scarcity of Hebrew texts with pristine Babylonian vocalization (i.e., prior to the infiltration of Tiberian norms), scholars remain divided as to whether the Babylonian diacritic called kemots puma resembled the Tiberian kamets or the “Sephardic” patah. In the Babylonian system itself there was no discrete patah; a single diacritic served as counterpart for both Tiberian patah as well as segol (see “The Kamaz in Babylonian Phonetics and in Yemen” by Hanokh Yalon, Tarbiz 33 pp.97–108, English summary p.i; also Israel Yeivin’s The Hebrew Language Tradition as Reflected in the Babylonian Vocalization [Heb.] Jerusalem 1985 vol. 1 pp. 56–57).
[xx] E.g., The patriarch is Abraham not Abrohom; the matriarch Sarah not Soroh, etc.
[xxi] “le-toldot ha-mivta ha-ivri bime ha-benyim” in Mehkarim Ve-iyyunim, Tel Aviv 1954 pp. 42ff.
[xxii] The Cairo Genizah, second edition, New York 1959 pp.184–186.
[xxiii] Mahzor Vitry, S. Hurwitz edition, Nuremberg 1923 p. 462.
[xxiv] The Hebrew Language Tradition as Reflected in the Babylonian Vocalization vol. 1 pp. 29–30.
[xxv] Inyanei Lashon, Jerusalem 1942.
[xxvi] The convention of using aleph to represent an ‘o’ sound belongs exclusively to the orthography of the Yiddish language which began to be written in Hebrew letters not much earlier than the fourteenth century. Rashi’s spelling of la’az (=Old French) words knows nothing of such a convention.
[xxvii] For a modern edition see The Beginning of Wisdom edited by Raphael Levy and Francisco Cantera, Johns Hopkins Studies in Romance Literatures and Languages, Extra Volume XIV, Baltimore 1939.
[xxviii] Weinreich’s designation for what is more commonly referred to as the Erets Yisrael system. The Tiberian he sporadically calls the northern Palestinian.
[xxix] History of the Yiddish Language, translated by Shlomo Noble with the assistance of Joshua A. Fishman, Chicago 1980 pp. 359–369.