Attempts to Decipher the Accents

ATTEMPTS TO DECIPHER THE TE`AMIM

Please have patience while this page is loading; the content is worth the wait!

N.B.: A WinZip (.zip) archive (17.9 MB) containing a PowerPoint Presentation (.ppt) and sound files (.mp3) summarizing the historical and musical background to Suzanne Haïk-Vantoura's work may be downloaded via this link.

THE MASORETIC PARADIGM FOR THE ACCENTUATION

The Masoretic and post-Masoretic Jewish sources are clear: the "Tiberian" te`amim are melodic, syntactic and rhythmic in function and exegetical in significance. They define the melody to which the Scripture is sung in public reading, and they mark the stressed syllables in the reading. But they also show which words are at the ends of verses and phrases and which ones are to be connected within phrases. These elements combine to bring out the meaning of the words on various levels (grammatical, semantic, poetic, etc.). In effect, the accentuation is but a magnification of the natural vocal inflection of speech, with its implied punctuation and phrasing. So Moses "spoke all the words of this song" by such a "speech-song", called cantillation (Deuteronomy 31:30). So David wrote: "Your statutes [prose texts par excellence] have been my songs [zemirot, songs accompanied by harp or lyre] in the house of my pilgrimage" (Psalms 119:54).

It is typically believed today that the interpretations of the te`amim developed by the Masoretes (beginning in writing with Moshe ben Asher's son Aharon, ca. 930 AD) and their grammarian successors to the present represent the original exegetical meaning of the "Tiberian" te`amim. Likewise, it is typically believed today that one or another form of synagogue chant represents more or less closely the original musical meaning. Now it is true that the Masoretes and grammarians, along with the cantors of the synagogue, understand the basic significance of the notation. But does it follow that they understand the notation's actual meaning? Are the te`amim primarily punctuational, then rhythmic, then melodic, as is usually maintained? Or are they the reverse: primarily melodic, then rhythmic, then punctuational? And if they are the latter (as cantors and even grammarians such as Israel Yeivin maintain), then how is it that the Masoretic exegetical paradigm and the early and recent synagogue chants alike universally treat the te`amim as if they really were primarily punctuational?

Consider the following: The "Tiberian" notation appears as it were "out of nowhere", fully formed, by the end of the 9th century. Only after that does the first tentative study of the interrelationships between the te`amim appear (in a treatise by Aharon ben Asher, ca. 930). Only by the 12th century (in the Manuel du Lecteur and similar "reader's manuals") do we find a relatively systematic set of "grammatical rules" for the te`amim worked out. Only by the Renaissance (in Elias Levita's work) are those rules brought to more or less the state of completeness that they have today. Even now, specialists using computers study different manuscripts and editions to test and affirm these "rules" and the consistency of their application.

The "Masoretic paradigm" first stated by Aharon ben Asher is accepted as valid by virtually everyone in the field of Masoretic studies. That paradigm divides the te`amim into three classes: 1) the "disjunctives", which are found on words that end a verse or phrase; 2) the "conjunctives", which are found on words within a phrase; and 3) the so-called ga`ya, a vertical sign often considered not a ta`am at all, which is neither disjunctive nor conjunctive but acts as a sort of "bridle" on how quickly one sings a syllable (thus its alternate name, meteg).

But as others have noted in the past, this paradigm does not explain most of the physical features of the accentuation or of its relationship to the words. (How much this is true we will illustrate shortly). Note too that while the accentuation itself appeared "fully formed" by 895 AD, it took no less than three centuries to come up with a seemingly coherent explanation of its grammatical rules. Is this not evidence that the Masoretes and later grammarians worked with a notation of which they understood the significance, but not the actual meaning?

Let us conduct what Albert Einstein called a "thought experiment". Let us assume that the Masoretes had received a notation (as they said) which they knew had both musical and exegetical meaning. Let us also assume (as the evidence strongly suggests) that the original meaning was unknown to them and their grammarian successors. Not being musicians themselves, and being used to the synagogue chant of their time (which chant likewise was more exegetical - after a fashion! - than musical in nature), they would have naturally assumed the exegetical meaning was paramount. Being clever and careful, they would have sought out a "Rosetta Stone" by which the notation could be deciphered.

In order to decipher an unknown notation, one must have two things:

1) One must know the true significance of the unknown notation itself;
2) One must have a parallel text containing a common meaning with which to compare the unknown notation.

On the actual "Rosetta Stone" we have a "trilingual" text: Greek and two forms of Egyptian. The Greek script was known, the Egyptian hieroglyphic script was not. Champollion, finding places of correspondence in meaning between the Greek text and the Egyptian text, was able to decipher the full meaning of the latter bit by bit. In effect, the Greek and the Egyptian hieroglyphic texts together formed a true "bilingual" text.

When one has a "virtual bilingual" text, by contrast, one has two parallel texts that have information in common, yet the two texts are of different genres. Here, we have a musical/exegetical text (the te`amim) and a verbal text (the Hebrew text). Both have verses and phrases -- in short, syntax -- in common. So the Masoretes began their analysis (in effect, their decipherment) of the te`amim on these premises:

1) They took the significance of the te`amim as musical and exegetical, but treated the exegetical function as primary;
2) They compared the syntax of the te`amim as a notational system against the syntax of the Hebrew verbal text.

According to the premises of their paradigm, the very distinctions between the te`amim themselves were made, each ta`am defined according to its supposed function.

Let us continue (in our "thought experiment") in the Masoretes' footsteps, starting with the same premises:

1) The te`amim are both musical and exegetical, but the exegetical function is primary;
2) The syntax of the te`amim as a notational system is parallel to and shares information in common with the syntax of the Hebrew verbal text.

The first thing one notices about the te`amim (from the perspective of exegesis) is that some graphemes are found at the end of verses and phrases and some within phrases (i.e., there are disjunctive and conjunctive graphemes). Yet also, some graphemes are found above, some below the words (i.e., there are superlinear and sublinear graphemes). Moreover, there is no strict relationship between the placement of a grapheme above or below the words and its placement at the end of a verse or phrase. In fact entire verses may be marked by sublinear graphemes only, and entire half-verses and phrases too, while only certain phrases may be marked by superlinear graphemes only. This implies that complete musical phrases may be marked by sublinear graphemes, but not by superlinear graphemes. Already one comes to what looks like a reductio ad absurdum, disproving premise 1) (the te`amim are primarily exegetical), for this finding implies the musical function predominates. In order to save the first premise, one must add another:

3) The placement of graphemes above and below the words is both grammatically and musically irrelevant.

One sublinear grapheme, the vertical silluq ("end, final"), is found marking the final stressed syllable at the end of virtually every verse. It would therefore be the strongest "disjunctive". Well and good. But an identical grapheme, by far the most common of all when taken with silluq (or even alone), is found on the lesser stressed syllables in words found everywhere within phrases and at the end of phrases and verses. The grapheme itself is found in close association with every other grapheme (even commonly on the same word with silluq itself, sometimes twice). Once again we come to an apparent absurdity: how can the same grapheme be the strongest "disjunctive" in one place and be anything but a "disjunctive" (or even a "conjunctive") everywhere else? Once again we are forced to add an assumption:

4) Graphemes with the same form but different placements within a verse must have different functions and therefore represent different "accents" or te`amim.

This premise comes up repeatedly when dealing with other graphemes, both in prosody and psalmody. And so in this case, the grapheme is called by its ancestral name (silluq) when found at the very end of a verse, and ga`ya or meteg everywhere else (even if found once or twice on the same word as silluq); and the two are assigned completely different meanings in the Masoretic paradigm.

In some cases one finds graphemes which are found both above and below the words. These are not difficult to distinguish by their position as separate te`amim. But sometimes a sublinear grapheme is found in the same grammatical position as a superlinear grapheme, yet it has a different form and is placed on a different syllable indicating a different rhythm. In such a case, it is assumed to have the same grammatical function and a slightly different musical function. Thus we have yet another added premise:

5) Graphemes with different forms but similar placements within a verse must have similar functions and therefore represent musical variants of the same "accent" or ta`am.

One could go further, and show that the Masoretic paradigm does not attempt to explain (in most cases) the exact forms of the various graphemes, their placements relative to a syllable, their exact order (in the case of the conjunctives) within a phrase, or even the meaning of their ancestral names. Rather, the paradigm itself redefines graphemes which by all indications were once but one ta`am as two or more te`amim, or vice versa, and renames them accordingly, thus giving more names to the overall system than there are graphemes in the system.

The Masoretes themselves and their successors only knew of the simplest forms of synagogue chant, and attempted to define a very few graphemes as markers of it (the lists of what grapheme represents what musical meaning vary from source to source). Once the notation itself was published, the synagogues step by step adopted their chants to fit the rules the Masoretes and grammarians derived for the notation. The process proceeded farthest among the Ashkenazic Jews of Europe, in the Torah reading. Such chants can by no logic (and despite frequent claims to the contrary) represent the original musical meaning of the te`amim, for they arose either independently of or were developed after the notation was published.

The thumbnail graphic on the left (taken from Introduction to the Tiberian Masorah by Israel Yeivin, trans. and ed. by E.J. Revell, Scholars Press, 1980, p. 167) shows a typical table of disjunctives and conjunctives for the prosodic system (or, as grammarians usually call them, "the accents of the Twenty-One Books"). To the reader who first encounters such a table, the notation as such seems highly unsystematic. There is no consistent relationship between the disjunctive or conjunctive function of a ta`am, or its relative force given such a function, and its form, its name, its position above or below a word, its position relative to a word or its position relative to a syllable. Moreover, there seems no logical reason why particular te`amim in one class or another should have particular forms in a particular order. Of course the te`amim here are taken out of context of the accentual-verbal phrase, but one would naturally expect some design behind the forms and layout of the graphemes that relates to their disjunctive or conjunctive function. Such design is not what this table reveals.

More anomalies appear upon closer examination. True, the majority of the disjunctives are superlinear and the majority of the conjunctives are sublinear, but why is not one class completely superlinear and the other completely sublinear? Such a distinction would be only natural, and would certainly expedite the learning and application of the system by the reader. Then too, the conjunctives are without variant forms (though mayela strangely is comprised of two graphemes on two words joined by a hyphen or maqqef), while no less than five disjunctives have variant forms. In two cases the variances are justified somewhat by the fact that one form is "small" (qaton) and the other is "large" (gadol). The graphical distinction between the forms seems related to that distinction in "size". Gershayim is simply the "doubled" form of geresh; both the name and the graphical form reflect this fact. But why does segolta completely change its name and form to shashelet (and change its place on the syllable as well) rather than just becoming, say, segolta gedolah? Worst of all, why should pashta, a superlinear sign found at the end of the last syllable of a word, change its form, its place and its name entirely and become yetib, a sublinear sign found at the beginning of the first syllable of the word (and atypically before the vowel-point, no less)?

Now mehuppakh (spelled mehuppak in Yeivin, op. cit.), a ta`am that (at least in almost all printed editions) is identical in form to yetiv, very frequently precedes pashta as a servus (subordinate conjunctive). It may even do so on the first syllable of a word (as in Song 7:1, Hebrew versification), which is also the syllable on which yetiv always falls. Yetiv, on the other hand, as a rule is not followed by pashta, but is considered (by Yeivin and others) to be a musical variant of pashta. Where it is used, the verbal accentuation is obviously different from that used where pashta is found, and reasonably requires a different melodic treatment. This at first blush could at least justify the substitution of one melodic sign for another, even if the change in name, form and position remains unexplained thereby. But in eleven cases noted by the Masorah (Leviticus 5:2; Deuteronomy 1:4; Isaiah 5:24; 30:32; Jeremiah 14:14; 16:12; 22:30; Daniel 2:10; 7:27; Ezra 6:8; 9:4), mehuppakh is replaced by yetiv and is immediately followed on the next word by pashta (Yeivin, op. cit., p. 199)! Moreover, the rest of te`amim used in all the accentual-verbal phrases where these substitutions are found are exactly what are used elsewhere in parallel circumstances with mehuppakh. In effect, in the eleven places cited above, the one thing that really differs in the accentual-verbal phrase is the verbal accentuation -- and the melodic accentuation is merely shifted in position to account for that change. Is this not evidence that mehuppakh and yetiv are really one and the same ta`am, and that yetiv is not a musical variant of pashta, but simply mehuppakh in a different position?

Or again: in the disjunctive class several te`amim have a "small" and a "great" form. This is not true of the conjunctives as a class. Fair enough. One might conclude therefore that the "small" form has a lesser conjunctive force and the "great" form as a greater one. In fact this is generally true. Why then is the "great" form of telisha (gedolah) a disjunctive and the "small" form of telisha (qetannah) a conjunctive? Why are all the other "small" and "great" forms found on the accented syllables of the words, while the two telischas are always found (respectively) at the very end or at the very beginning of a word, no matter what syllable of the word is normally stressed? Could it be that the distinction between "small" and "great" is unrelated to disjunctive or conjunctive force, but is related rather to melodic motif? This would account for not only the various forms of the te`amim, but their positions relative to the words as well.

This brings us to another major anomaly. Most of the te`amim are found on the major stressed syllable. Yet some are always found at the beginning or the end of the word, whether the stressed syllable is found there or not. In standard manuscripts and the printed editions based on them, pashta alone of such anomalous accents is "doubled" when the major verbal accent is found on the penultimate syllable. In some manuscripts and editions, however, the other anomalous accents (zarqa, segolta and the telishas) are also so doubled, evidently for the sake of "self-consistency". But is not this "anomalous" positioning of certain signs, and the doubling of pashta as well on certain words, an indication that music and not grammar motivates the phenomenon in question?

In the above table, the same sublinear graphic form is given two different names: munah (a conjunctive) and legarmeh (a disjunctive). What the table does not reveal is that the early treatises on the te`amim give munah even more subsidiary names, differing from source to source, denoting hypothetical and ultimately undefinable nuances. Such distinctions (as Derenbourg notes and as Yeivin effectively concedes) have no basis at all and were invented in order to give separate domains to identical signs that originally represented but one and the same thing (Manuel du Lecteur, p. 519). If one follows this analysis to its logical conclusion, then munah and legarmeh must also be two names for the same ta`am, despite the fact that one denotes a "conjunctive" and the other a "disjunctive". But if this be true for munah and legarmeh, why could it not also be true for pashta and azla? For mehuppakh and yetiv? For the first graphic form in mayela and tifha? All the more in that the various names for the te`amim in the early treatises "never had either fixity or unity" thanks to their assignment as disjunctives and conjunctives (ibid.), and that other te`amim besides munah may take a "legarmeh" or "independent" form (such as darga in Numbers 17:28 and shalshelet in Amos 1:2).

The most common graphical form of all, the so-called ga`ya or meteg (which in printed editions is identical in form to silluq), is not found in the above table at all. Yeivin says it forms part of the accent system (op. cit., p. 242), yet does not include it in his table. Most if not all traditional and scientific sources since the Middle Ages have excluded it from the accentuation entirely. The reason is simple: it is neither disjunctive nor conjunctive in function. That is, it neither divides words into verses and phrases nor connects words within a phrase. Rather, it is found on the secondary stressed syllables of words (and occasionally, on the primary stressed syllables as well), interspersed or being left out between other te`amim as the needs of the accentual-verbal phrases require.

Yeivin notes that ga`ya has no musical motif of its own (as indeed it does not in any form of synagogue chant), but indicates that the syllable it marks is to be slowed down and not slurred over (ibid., citing Jules Derenbourg, Manuel du Lecteur, 1870, p. 77, the latter being a translation of a 12-century "reader's manual" in Hebrew, the Horayat `al - Qorê). And yet, "ga`ya sometimes indicates a pause of some sort" when two words are joined by maqqef and ga`ya falls on the last syllable of the first word (Yeivin, p. 264). The two words are thus divided, yet joined, and the last syllable is not slurred over. In this light, Yevin cits an early rabbi named Hayyuj thus: "'Ga`ya is the opposite of maqqef because maqqef joins words, while ga`ya separates them'" -- at least in such specialized cases (ibid.). But could not the same musical motif, represented by the same sign, always represent a pause -- whether at the end of a verse (to stop the reading), on a syllable (to slow down the reading), or between words joined by maqqef? But such a common motif, with such a wide range of pausal functions and associations with other te`amim, could only be a single note -- by definition, the tonic note of a scale (as we will see).

The thumbnail graphic on the left (taken from Yeivin, op. cit., p. 264) shows a typical table disjunctives and conjunctives for the psalmodic system (or as grammarians usually call them, "the Accents of the Three Books"). Here we still find the same sorts of inconsistencies in form and nomenclature that we find in the above table of the prosodic system. At least the "variant" forms of the accents are for the most part consistently distributed, one of a pair in the disjunctive class, the other of a pair in the conjunctive class. (Revia and revia mugrash are an exception to this rule; the latter name signifies that a sign like geresh has been added to the accentual motif.) Yet the nomenclature of such pairs is no more consistent than in the prosodic system. The "great" and "small" forms of shashelet are put in different categories, reasonably enough. The same is true of the legarmeh ("independent") verses the "normal" forms of mehuppakh and azla. But why not just call the different forms of these te`amim "great" and "small" as well? Why are dehi and tarha (the latter called tifha in tables by other grammarians) given different names -- and why is the former always before the word and the latter always on the stressed syllable of the word? Why is tsinnor always at the end of the word and tsinnorit (usually) at the beginning of a word -- and why is the latter name merely the feminine form of the former name? No doubt all these details are important, as similar details are in prosody -- but analyzing them from the point of view that the te`amim are primarily disjunctive and conjunctive signs leads only to dead ends with regard to these questions.

There are other anomalies in the table, of course. Some te`amim (`oleh veyored and revia mugrash) are inexplicably comprised of two graphical forms while most are comprised of just one. There is the same lack of correlation between the sublinear and superlinear position of a sign and its disjunctive or conjunctive function that we have seen in the prosodic system. Likewise, there is a lack of correlation between the various forms and positions of the signs and their functions. Some signs are consistently found on the stressed syllable, some are not. There is a vertical sign, paseq, which is apparently younger than the accentuation itself (Yeivin, p. 216), added after the "legarmeh" signs, just as it is after the "legarmeh" signs in prosody. Once again, ga`ya is left out of the table, despite the fact that (in the Letteris and Ginsburg Editions and the manuscripts on which they are based) it may be found with silluq up to twice on the same word (as with be'emunato, "in His faithfulness", at the end of Psalm 96:13) and is found extremely commonly in combination with all the other te`amim.

To complicate matters further, the rules by which the psalmodic system is governed are less grammatically strict than those that govern the prosodic system. This implies that the role of the music as such is greater in the psalmodic system -- which makes the fact that the synagogue psalmody is generally plainer than the synagogue prosody inexplicable, if this notation really represents synagogue chant. But the greater role of music in the psalmodic system has allowed for a greater range of scribal errors and opinions (and therefore for more purposeful or inadvertent textual criticism) than is usual in the prosodic system. In any case, while the Jewish communities generally insist that they have preserved the practice of prosodic cantillation correction, they generally concede that the original practice of psalmodic cantillation has been lost to them. But in the early Masoretic treatises, even the grammar of the psalmodic system gets much shorter shrift than does that of the prosodic system, probably because it is so much simpler.

The thumbnail graphic on the left (taken from Gerard E. Weil, Concordance du Pentateuch et des Cinq Meghillot, CRNS, France) shows the results of Weil's own computer analysis of the hierarchy of disjunctives ("accents nodaux") and conjunctives ("accents subordinées") in the Pentateuch and the Five Megillot (Song of Songs, Ruth, Lamentations, Ecclesiastes and Esther). Upon close comparison with Yeivin's table for the accents of all the Twenty-One Books, the reader will find a somewhat different order of te`amim in each class, as well as a somewhat different nomenclature used to describe them. (Indeed, some few accents that Yevin calls conjunctives are called disjunctives by Weil.) Yet the same basic faults with the analysis remain. As it is said in the world of computer programming, garbage in, garbage out...

Weil's study (the first of several of its kind) came out shortly after the original French edition (1976) of Haik-Vantoura's French book. He took pains to state his belief that the te`amim are non-musical (a belief which is not unique to him, as I discovered later), and to demonstrate what he considered the "inanity" of Haik-Vantoura's methodology and conclusions. One of Weil's research fellows at CRNS, Denise Jourdan-Hemmerdinger, later undercut his critique by showing that Weil himself was ignorant of the norms both of ancient music and of ancient music notation. (Her review in French of Haik-Vantoura's Les 150 Psaumes, which contains her reply to some of Weil's assertions, may be downloaded in .pdf format via this link.) But Weil's analysis testifies against itself; it does not pass a close shave with Occam's Razor. Nor does it take into account the testimony of the early sources such as the Manuel du Lecteur: "the wise shall understand" (citing Daniel 12:10) that each written sign represented a gesture of the hand or of the fingers, which in tern represented a sound (ne`imah, that is, a musical tone) produced by the mouth. How can denying a musical function to the notation lead to an explanation of most of its physical features? Obviously, it cannot. It can only explain the hierarchy of how the musical tones punctuate the words, yet without explaining how that punctuational hierarchy is generated by the tonal hierarchy behind the written signs.

Why are Masoretic scholars like the late Dr. Weil, the American scholar James D. Price (who has made and published his own computer analysis of the te`amim), and so many others in Israel, Europe and elsewhere so taken with the Masoretic grammatical paradigm? I will discuss the major reasons elsewhere. But one reason, it seems, true to human nature is that one may become so enraptured by his own model and the self-exaltation that it grants him that he "fails to see the forest for the trees". It was easier for many centuries for many people to accept the wheels within wheels of Ptolemaic cosmology than to accept the much simpler (yet much more humbling) cosmology we now call Copernican. And so it is here: it has been far easier for most to accept the idea of a direct link between the Temple liturgy and the synagogue liturgy than to admit that the former was utterly lost in practice and the latter never had a direct connection with it. Unless one makes that admission, one cannot take the first step in resolving the enigma of the meaning of the te`amim.

The thumbnail graphic on the left (taken from Solomon Rosowsky, The Cantillation of the Bible, 1957) shows the hierarchy of disjunctives (here called "lords") and conjunctives (here called "servants") in the Masoretic paradigm as applied to the te`amim (especially as found in the Pentateuch). As may be seen, various disjunctives and conjunctives may be found in various orders comprising an accentual phrase. (Here again, certain te`amim are given different names according to their alleged function, or even according to analogy of form, again as based on traditional sources.) Yet even so, in this table not all the interrelationships between the various accent clauses are really taken into account. Some linkages which are shown to be independent (or so it would seem) are in fact inserted into the verses that have them as important parts of their structure.

Cantor Rosowsky's work was a massive treatise on the Lithuanian Ashenazic style of "tropes" used in the cantillation of the Hebrew Bible, and especially of the Torah (Pentateuch). In truth, his was a systematic analysis of the cantillation style, leading to a purified form of it which has become the norm in American Ashkenazic synagogues today. Like many before him for several centuries, Rosowsky accepted the idea of the te`amim as a "feudal hierarchy" of "lords" (emperors, kings, dukes, counts and petty lords) and "servants". The famous Masoretic scholar Wickes opposed such a classification, as did many others (including Derenbourg, who called the hierarchy "all rather burlesque" and misleading to many brilliant scholars). Yeivin agrees that the concept gives a false impression of the accent system. (op. cit., p. 169). He finds the fourfold division of the te`amim into "grades" useful of itself, however, provided that the "pausal value" of an accent in a particular "grade" is understood to be relative, not absolute. He infers correctly that the requirements of the musical chant and of the syntax demand that some major disjunctives be preceded by minor disjunctives where the latter might not normally be expected (ibid.). Once again, however, he cannot explain how the musical chant requires these nuances!

The system of "tropes" that Rosowsky studied is simply the most highly developed of a number of similar systems used by various synagogue communities. In the Ashkenazi Torah chant, every ta`am is given its own melodic motif. Even so, the same graphic form (especially munah) may be given different motifs in different contexts, or very similar motfs may be given to different te`amim. (As Haik-Vantoura remarked in her English book, on page 482, in such chanting the elementary laws of tonal syntax are set aside!) In the Ashkenazi "tropes" used in the rest of the Twenty-One Books, and in other traditions that use "tropes", only the disjunctives are given melodic motifs, or even the major disjunctives only. Professor Aron Dotan ("Masoretic Accents, Melodic Rendition", in Encyclopedia Judaica) is right to say this is not a matter of loss of knowledge on the part of the non-Ashkenazi synagogues, but he fails to understand that these synagogues never had the knowledge to lose! Nor do their simpler "tropal" systems reflect an earlier stage in the development of the accentual system itself (such as the Palestinian and Babylonian accents systems), as he claims. The "Tiberian" accentuation (which, unlike the local Palestinian and Babylonian systems, came from the Temple priests via the Karaites and Masoretes) clearly appeared first; the "tropes" were then imposed on it and developed step by step (here more, there less) according to the Masoretic grammatical paradigm.

The thumbnail graphic on the left shows just how far the reinterepetation of the te`amim has gone among the Ashkenazi Jews. It is a table of "Torah tropes" prepared by Moshe Nathanson for the Chumash published by the Hebrew Publishing Company (New York, 1928).1 The "tropes" used are not identical to those developed by Rosowsky, but they are close enough.

As noted above concerning the Masoretic grammatical paradigm, no systematic account is taken of the sublinear or superlinear position of a grapheme, its position relative to a syllable, its form, or the sequence of "conjunctives" within a phrase. The musical motifs are highly ornamental, and are intended to follow the indications of the Masoretic paradigm concerning the punctuational value of the te`amim. Moreover, the ta`am called munah is given no less than five different melodic interpretations in this table, while pashta and the so-called yetiv are given very similar interpretations -- in both cases, on the grounds that the grammar of the te`amim predominates over their form and placement, and therefore over their musical function.

To be fair, since this system of "musical notes" was developed in large measure by professional cantors who had both an ear for music and an eye for the implications of the graphic forms and their ancestral names, there is often a vague correlation between the musical motif assigned to a "Masoretic accent" and the form and name of that "accent". Likewise there is a vague correlation between the melodic expression and the verbal meaning (leaving aside the punctuation as such). So this musical system explains more than the Masoretic grammatical rules themselves about the accentual system -- which is in itself a pitfall, for this system "maps" an arbitrary musical interpretation upon the words which can and does mislead the user into overlooking the problems with the entire Masoretic paradigm.

Such a complex system as that given above requires that the melodic motives "resolve" in different ways at the end of a word. The background of this page is taken from a table meant to illustrate how this is to be done (after Rokowsky, op. cit.). The above table by Nathanson has similar, more abbreviated indications: parentheses indicating notes to be added or left out depending on how the "trope" interacts with the word it interprets and with the rest of the melodic line.

One of the strangest features of this musical system (from the musician's point of view) is that it assumes the musical motif always begins when the word begins -- not when the grapheme representing the motif appears as in any normal musical notation. Despite the claims one typically finds in the cantorial literature that the te`amim are indeed primarily musical, the way the "tropes" are actually laid out relative to the words implies that they are really primarily exegetical. But if that is really the case, yet once more, why are so many of the graphemes' physical features left unexplained, or at best explained so poorly and unsystematically?

The thumbnail graphic on the left (taken from Yeivin, op. cit., p. 168) shows a table of how the major disjunctives are musically interpreted by one of the early sources. (The source is not the Manuel du Lecteur, which gives its own division on p. 383). As Yeivin notes (op. cit.), the classification of disjunctives into three categories is the only one found in the Masorah or the early treatises, yet it is not identical in all sources. Some te`amim are put by some sources into category 2 above, by other sources into category 3.

The (soi-disant) Levite Academy in Israel (R. Daniel Meir Weil, director), formerly at http://www.levite.org/, has used what turns out to be a very similar model to attempt to reconstruct the original musical meaning of the te`amim. The reference standard used lay in certain synagogue melodies transcribed into Christian neumes by a famous medieval proselyte to Judaism, Abdias the Norman. The melodies he transcribed are very much in accord with the indications of the "reader's manuals" as to the alleged function of the te`amim in marking simple melodic curves. This is all the more significant in that Abdias made his transcription in the 12th century, the same time as the Manuel du Lecteur (one of the last of the early reader's manuals) was written.

We have mentioned above that the Masoretes and their successors knew only the simplest forms of synagogue chant; they were completely ignorant of "tropes". Here is the proof. In the Manuel du Lecteur and other early manuals, and in the Masorah itself, the major "disjunctives" are placed into three categories, depending on whether they supposedly had a "low", "rising" or "high" sound (see above). No other musical interpretation of the te`amim is found in the early sources before the 13th century. Evidently the Masoretes and later authors were attempting to analyze the accentuation musically in a simplistic way, using the synagogue chant of their day as their "virtual bilingual". In effect, Weil's effort was the first systematic modern attempt to reconstruct the original musical meaning of the te`amim on the assumption that it represented the sort of chant alluded to by the Masoretes themselves.

Synagogue chant of this sort is called "primitive chant" by Israeli musicologist Israel Adler. It consists of simple rising and falling melodic curves, generally of no more than four notes in range. This sort of music is the work of "folk" musicians however dedicated -- not of a hereditary caste of professionals. R. Weil's attempt to reconstruct Levitical chant from such poor musical material is utopian at best.

How and when did the early synagogue chants arise? Jewish musicologist Johanna Spector believes the oldest synagogue chants could be as old as the first century AD (though not demonstrably older). Whereas the late Alfred Sendrey noted that the extant synagogue "tropes" are often quite similar to the Arabo-Persian maqamat (musical styles likewise based on "tropes" and certain specific musical modes). Since (as he thought) the synagogue chants were a survivance of biblical times, the music of ancient Israel (including that of its Temple) was therefore like the maqamat. But Spector has a different and more sensible opinion: the "tropes" used by many synagogues are not like the maqamat; they are maqamat -- adopted at an early date and adapted to the Jewish musical spirit. When one compares the "tropes" and modes of synagogue chant with those of the maqamat, one may see readily enough which are the originals and which the adaptations; information has been lost in the direction of the Jewish versions, not the Arabo-Persian versions.

Werner, Idelsohn and others have remarked on the many correlations between early synagogue chant and early Christian chant (especially Gregorian chant), showing that their common source must date to the early Christian era at least. Again, this much is certainly reasonable. The earlier, "primitive" chants (as even Haik-Vantoura concedes) may go back even further -- perhaps to the rise of the earliest synagogue communities themselves (that is, to late biblical times). But again: all these synagogue chants were folk melodies, whereas the melodies of the Temple and of the biblical authors were professional melodies. This means that whatever else they had in common with each other (say, similar scale and mode types), they had different relationships with the verbal syntax, and therefore had different tonal structures.

We know that pilgrims came to the Temple and heard the melodies performed there, and that they took reminiscences of them back to their local communities. How much had been retained by them has long been a point of dispute. Thanks to Haïk-Vantoura's work, we now can say: there are tantalizing hints of modes, melodies, and musical ornaments preserved by different communities in different biblical texts. (As one would expect, such correlations are found only in the primitive chants, not in the later tropes.) Sometimes the correlation between a certain synagogue chant and Haïk-Vantoura's reconstructed chant is striking (in the case of Lamentations, it extends to Gregorian chant as well). But by and large, the synagogue chants have little or nothing to do with the rediscovered Temple chant; they represent a different level of musical culture. This is what we should expect.

OTHER ANALYTICAL PARADIGMS

Many since at least the Renaissance have concluded that the Masoretes, grammarians and synagogues did not preserve the original meaning of the "Tiberian" te`amim and have sought to decipher them independently. Their efforts failed, however, in that they either did not start from the correct premise as to the significance of the notation or they did not start with the correct "virtual bilingual" with which to compare it. Many indeed have realized the notation was primarily musical rather than exegetical, but their standard of comparison has been synagogue chant, or one or another form of Christian chant, or even the music theory of their own day (or of their own imaginations). One modern Hebraist (Saul Levin) began with the Hebrew verbal text as his standard of comparison, but assumed the accentuation was a marker of simple vocal inflection rather than of melody.

We know that all these attempts have failed for the same reason we know (or ought to know) that the Masoretic paradigm has failed: none of these analytical paradigms explain all the salient features of the notation or of its relationship to the words. Were such standards applied to a notation found by archaeologists, the results would universally be held as questionable (to say the least).

Thus it was that a Reform Jew and student of music theory and composition, Suzanne Vantoura, found the following assessment in a French encyclopedia of music, shortly before the beginning of World War II: the te`amim are ancient, musical, and of unknown meaning. It is with this premise that she began her own deciphering attempt.

The following text was transcribed from the Preface to The Whole Booke of Psalmes: With The Humnes Evangelicall, and Songs Spiritual (1621). Here the composer Thomas Ravenscroft commented on (among other things) the unsuccessful attempts of many in his day to decipher the musical notation found in the Hebrew Psalms. What he inferred about the original characteristics of the Psalms is interesting indeed in the light of Haik-Vantoura's tonal and "harmonic" renditions.

(The Preface.)

TO ALL THAT HAVE Skill, or Will unto Sacred Musicke, I wish Concord among themselves, with God, and with their owne Consciences.

Harmonicall Brethren, I have here undertaken with no small labour, and charge, to bring the Tunes of the Psalmes, Hymnes Euangelicall, and Songs Spirituell, (as they are usually sung throughout Great Britaine) into one entire volume, which are so Composed, for the most part, that the vnskilfull may, with little practice, be enabled to sing them in parts, after a plausible manner.

In my opinion, 'tis too Laborious a taske for any Man to study the attainement of the Hebrew Musicall Accents; For the Tunes used in Dauids Time, are too farre remoued from our understanding. For albeit the Hebrew Musicall Characters are placed sometimes aboue the Letter, sometimes beneath, yet the knowledge of what was signified by those Notes and Characters, was onely continued by Tradition, and is now utterly lost, though many at sundry times (as appeareth by their writings) have gone about to reuiue it: But hauing no better subject to worke upon, than their owne weak conjectures, they haue but a little preuailed. I find yet that the Characters now used in the Rushian Church, (who had their skill in Musicke from the Graecians) though they differ in the placing, (because those of the Hebrewes are both aboue and beneath the Letter, these onely aboue), yet they partly resemble one another in the forme.

Againe, I finde by sundry Manuscripts, that the Latine Church, as weill in the forme of their Characters, as likewise in the placing of them, did participate of both. For first, according to the manner of the Hebrewes, they placed their Notes both aboue and beneath the Letter. Afterwards they used one line aboue the Letter, and placed their Notes both aboue and beneath the line, and that with a Geometricall distance, as the ascent and descent of the sounds did require.

In processe oif time, (as all things are brought to their perfection by degrees) they came to two lines, then to three. And Guido Aretinus, a learned man (whom Histories report to have lived in the time of Henry the 2. Emperour, in the yeare of our LORD, 1018.) was the first that inuented the vniforme of the Scale, (which we terme Gam-vt) and brought in the foure lines, which was and onely is now used in the Church for Phonaskes, distinguished by the Gregorian, Ambrosian, and Perigrine Tones, comprehended in the distance of a Diatessaron or a Diapente, viz. a fourth or a fifth, or the Harmonicall proportions of sesquitertia and sesquialtera: and by degrees it came to the distance of a Diapason, which is an eight, and a Duple proportion; in which three proportions, all Simple and Compound Harmony consisteth, by the Plagall and Authenticke diuision of the Tones and Tropes. The which Phonaskes are explained by the Tenor part, being the Fa-burden or Playn-song of the Psalmes, Anthems, and Responces usually sung in the Church in Prose, and Hymnes that were Composed in Verse and Meeter.

The five lines are used for Symponaskes or Parts Compounded of 2.3.4.5.6. voyses, etc.

The sixe lines are used for Instrumentall Musicke, as Organs, Harpes, Lutes, etc.

But, whatsoeuer the Tunes were in Dauids time, there is no question but that they were concordant and Harmonious, which could not be, had they not beene divided in part. For if you looke into 1 Chron. chapt. 15. 16. verse, yee shall see how the Prophet Dauid at the Reduction of the Arke, as likewise Salomon his Sonne, at the Dedication of the Temple, 2 Chro. chap. 6. 31. verse, distinguished all their Musicke in parts, and appointed such to be Masters and Ouer-seers of it, as were most eminent for their knowledge in that kinde; as Chenaniah the chiefe Leuite, to have the chiefe place, which was to be Master of the Song. An office which consisted not onely in the direction of the Quire, but likewise in the trayning vp of others to sing, that there might be still a supply of able persons for that seruice: Asaph the next, and so Heman his Brother, likewise Ieduthun and Ethan, all of them the most renowned chanters of those Times, and such as successiuely in one anothers absence, were to direct the due performance of that charge, so that not onely the voyce of the Singers, but likewise the sound of the Instruments agreed so well together, that they seemed to bee but one Sound, and one Voyce.

Neither was this method confined onely to the Old Testament, but sanctified to the Church of Christ by the prescription of the holy Apostle S. Paul, Col. 3, verse 16. Let the word of God dwell plenteously in you, in all wisedome, teaching and admonishing your owne selues, in Psalmes, Hymnes and Spirituall Songs, singing with a Grace to the Lord in your hearts.

I have therefore endeauoured for the fitting of euery Heart to that Psalme, which it shall most affect, to place speciall Tunes, proper to the nature of each Psalme, (not imitating Art so much, as the naturall inclination, but ioyning one with another,) and am blod to admonish the Singers that they obserue three Rules.

1 That Psalmes of Tribulation be sung with a low voyce and long measure, Psal. 9. 32. 38. 51. 102. 130. 143. etc.

2 That Psalmes of Thanksgiuing be sung with a voyce indifferent, neither too loud, nor too soft, and with a measure neither too swift nor too slow, Psal. 18. 23. 27. 30. 31. 46. 48. 66. 81. 104. 105. 111. 118. 122. 124. 126. 138. 144. 145. 146.

3 That Psalmes of Reioycing be sung with a loude voyce, a swift and iocund measure, Psal. 33. 34. 47. 84. 95. 96. 98. 99. 108. 113. 117. 135. 136. 145. 147. 148. 150.

In all which, the obseruing of Time, Tune, and Eare, will produce a perfect Harmony.

Accept kindely, what I haue laboured earnestly, and use it to thy comfort. Thus I end, humbly wishing to all true Christian hearts, that sweet consolation, in singing prayses unto God here upon Earth, as may bring vs hereafter, to beare a part with the Quire of Angels in the Heauens.

Your well according, and

best wishing Brother,

Tho. Rauenscroft.

FOOTNOTES

1. Ironically, this edition and the HPC edition of the complete Hebrew Bible was based on the Letteris Edition. My first copy of Letteris was one produced by the HPC. Regrettably, the HPC is no longer printing Hebrew Bibles.

The Diqduqê ha-Te`amim, "The Grammatical Rules of the Accents".
Such as the late Gerard E. Weil and James D. Price.
Not that this stopped various scribes, beginning with Aharon ben Asher and another scribe named ben Naphtali, from coming up with various opinions as to how this or that part of Scripture should be accented, sometimes (it appears) even contrary to the received tradition. The Leningrad B-19 Codex, the oldest complete codex after the Aleppo Codex of Aharon ben Asher, shows graphemes of multiple erasures and rewrites of its accentuation, apparently with the aim of making the accentuation more "grammatically self-consistent".
At that time, it was a simple "aide-memoire": a rising and falling phrase with at best some melodic ornaments at the ends of verses and phrases (cadences). As noted below in the main text, the marginal notes of the Masoretic Text and the texts of the reader's manuals refer to this kind of chant - not to the much more ornate style that developed with the rise of professional cantors in Europe.
Notorious in this respect is munah, which (as illustrated in the main text) has no less than five different grammatical functions attributed to it by the later Masoretic "reader's manuals" (and by the Ashkenazic rite of synagogue chant as well). Some of those functions are supposed to be more or less disjunctive, others more or less conjunctive. The secondary names attributed to the ta`am in the medieval manuals are supposed to denote these untranslatable nuances of function. In modern treatises only two names are generally given, munah and legarmeh.

Index

Updated March 01, 2010