The introduction, in particular the section about the hierarchy of distinctive marks, contains an explanation how the marks of the different ranks work together to structure each Bible verse. There, no information was given which concrete mark has which rank, or which of several marks of equal rank is actually used in a given context. It is the purpose of this chapter to fill this gap. The exposition follows Breuer's book on cantillation marks (Mordekhai Broyer (Breuer): Taamey hammiqra be-21 sfarim uvesifrey eme"t; Jerusalem, TShM"B (=1981)) for the coarse structure of the syntax, of course leaving away nearly all detail and all discussion of sophisticated end cases.
Conjunctive marks do not take part in the business of dividing a verse into logical units. After a verse has been divided into pieces by the distinctive marks, these pieces can each consist of one or more words that are connected with either conjunctive marks or with Maqqef, and this irrespective of the rank of the distinctive mark at the last word of the piece in question. We can also say that in terms of the hierarchy: a distinctive mark of any rank may be "served" by a conjunctive mark (a "servant"), whereas among the distinctive marks, only the immediately lower rank may serve. A sequence like
officer servant duke king duke servant king servant emperor
is fully permissible: the servants always serve the next following distinctive mark, and these serve each other in the same way as if the servants had not been present.
Thus, the placement of distinctive and of conjunctive marks are two entirely separate processes which are described in the sequel in separate sections. For the description of the placement of distinctive marks, the units we talk about are sequences of one or more words, only the last of which carries a distinctive mark, whereas the others, if at all present, carry either a conjunctive mark or Maqqef.
Which distinctive mark is to be placed in a given context, depends
always on its rank
always on whether it is final (i.e. the last mark of its rank before the mark it serves)
sometimes on more detail about its position in the chain of marks of equal rank, e.g. whether it is first or last but one
sometimes, for final marks, on the mark it serves
sometimes on whether its realm consists of only one word or of more than one word.
Hence, not considering the last of these points, we know which marks to use when we know the entire chain of marks that serve the same superior mark. As a consequence, each of the rules to be set up in the sequel will tell us about one higher distinctive mark and all the immediately lower distinctive marks serving it. Note that the marks of equal rank in the same rule have different nesting depths as explained in the section about the hierarchy of distinctive marks.
The rules are given as charts with coloured boxes where each box stands for the realm of a distinctive mark (for nit-pickers: the last box, showing the mark of higher rank, stands only for the final portion of its realm, that is, for the portion that is not already depicted by the other boxes in the same chart). The exact semantics of the charts is explained as needed when we proceed through the charts.
We start with a very simple chart.
This chart tells us that a verse in the 21 books contains two emperors: the Atnach in the middle, and the Sof Pasuq at the end. From this example, we learn some notation in the charts:
The boxes contain abbreviated names of the marks in order to keep their sizes manageable. If you are familiar with the names, you will not have problems understanding the abbreviations, if not, giving the full names does not really help for easier survey of the chart. The full names are given in the tables.
The abbreviations of the distinctive marks consist of three letters and a number indicating the rank (0=emperor, 1=king, 2=duke, 3=officer).
The colour of each box contains two informations: the hue indicates the rank (purple=emperor, red/orange=king, yellow=duke, green=officer), and within each rank the saturated colours are used for non-final marks and the pale colours for final marks. A more detailed explanation of the abbreviations and of the colours used is found here.
The boxes are adorned with the actual shapes of the marks, as if the boxes were words to which the marks are applied.
The whole chart is to be read from right to left; otherwise drawing the marks would not make sense.
Now we proceed to a slightly more complex chart.
Here, some more notation is introduced:
Two or more marks standing on top of each other without braces (e.g. Atn0 and SoP0) denote that one of these marks is in that position (so that the chart is in fact a condensed picture of what could be more than one chart), but that the two marks have different significance and do not appear in the same contexts.
Two marks standing on top of each other with a pair of braces enclosing both (e.g. ZqQ1 and ZqG1) denote two marks that appear in the same contexts. Typically, the upper mark in the brace is used when the realm consists of more than one word, and the lower one when the realm consists only of the word where the mark is placed. In the former case, the realm may be subdivided by one or more distinctive marks of lower rank, or it may only contain conjunctive marks in the interior. If, however, all words in the realm are connected by Maqqef, the entire realm counts as only one word.
This dependence on the length of the realm is valid with the following restrictions: For Ger3/Grm3, it is valid in the majority of cases but with numerous exceptions in both directions. For Psh2/Ytv2, it is valid only insofar as the realm of Ytv2 is indeed always only one word long, but there are more Psh2 with a one-word realm than Ytv2. For Paz3/QaP3, it is not valid at all; the realm of QaP3 is at least three words long. For AzL3/MpL3, it is valid if the definition of "one word" is somewhat modified (see there). For the remaining pairs, in particular those in the chart above, it is quite reliable.
In all cases, the mark at the bottom of the brace is substantially less frequent than the one at the top.
A single mark or brace of marks in square brackets is optional. It may be present in some verses and not present in others. Typically, the pattern with the optional mark is the less frequent one.
An ellipsis denotes the part of the chain that is repeated more or less often when more or less marks of the same rank are needed: the ellipsis with the two identical marks at both sides stands for zero or more occurrences of that mark. If even less marks are needed, for instance no king under an emperor, the marks are dropped from right to left, so that the final mark survives longest.
In plain words, the chart says: The last king before Atnach or Sof Pasuq is always Tipcha (Tip1), and all preceding kings are Zaqef Qatan (ZqQ1) when their realm is more than one word long, and Zaqef Gadol (ZqG1) when it is only one word long. As an infrequent alternative, the chain may also begin with a single Segol (Sgl1) or, when that would be on the very first word of the verse, a Shalshelet (Sha1) instead.
There is a detail which is not depicted in the chart: Sgl1 or Sha1 can only occur in the first half of the verse, that is, in the realm of the Atnach.
As a very infrequent exception, Tipcha may occur in the same word or Maqqef-connected chain as the emperor, so that the last king is Zaqef if one is present. Such a Tipcha is then called Meayla and regarded as a conjunctive accent.
We have now nearly all rules for the interpretation of this chart in place, except one:
Alternatives can consist of more than one mark; the choice happens once when the number of alternatives increases, and not independently for each mark again. Here, a brace of marks is treated like a single mark.
This rule says that the final duke depends on the king it serves whereas the preceding dukes are always Revia (Rvi2).
Note that the marks appearing at the bottoms of the braces in the preceding chart (e.g. ZqG1) do not show up here: when a king is served by a duke, the realm of the king must be longer than one word as it contains at least the two words where the king and where the duke is placed.
Remember that for the officers the choice of one of the marks in a brace is not as directly connected to the length of the realm as for kings and dukes.
There are many patterns how a chain of conjunctive marks may serve a distinctive one. The rules given below yield only the most frequent such patterns. The distinction between final marks (i.e. the last of their rank before the next higher; in the charts the boxes with the pale colours) and non-final ones is important in this context.
Some distinctive marks have their own conjunctive marks serving them: Mahpakh serves Pashta, Darga serves Tvir, Qadma serves Geresh, Galgal serves Qarney Para.
The remaining distinctive marks, and Pashta and Tvir in many cases as well, are served by two "generic" conjunctive marks: Munnach serves Zarqa, Gershayim, and all non-final marks; Merkha serves the final marks except Zarqa and Gershayim.
If there is more than one conjunctive mark serving Gershayim or a non-final mark, the last conjunctive mark (Munnach or Galgal) is preceded by Munnach, except for a Munnach that serves Revia, which then will get Darga before it.
If there is more than one conjunctive mark serving a final mark other than Tipcha, the last conjunctive mark (Merkha, Mahpakh, or Darga) is preceded by Qadma which, if applicable, is preceded by Tlisha Qetana. The latter precedes also a Qadma immediately preceding Geresh.
A Tipcha which has two conjunctive marks ahead of it, has first Darga and then Merkha Khfula. This occurs infrequently.
The basic rules for the placement of distinctive marks in the 3 books resemble those in the 21 books. The same notation will be used for the rules, and will not be explained here again.
The rules given in this section are less reliable than those for the 21 books, i.e. there are much more exceptions. Moreover, they are more possibilities that the same verse with given marks can have more than one interpretation which is conformant to the rules. Some hints how to proceed is such cases are given at the end of this page.
In the 3 books, each verse contains only one emperor: the Sof Pasuq at the end. A separate chart showing that is not necessary.
Now, it was a feature of the emperors in the 21 books that they are very few, in fact only one in the interior of the verse, and cannot be repeated so that their high importance is always the same. The kings in the 3 books have a similar behaviour, albeit more complex rules. In some sense, they fulfil the same task as the emperors in the 21 books, to wit providing the topmost division of the verse.
The most frequent pattern of kings in the 3 books has either Atnach as strongest king or Ole We-Yored as strongest and Atnach as second strongest king; in these two cases the portion after the Atnach may be subdivided only by special marks not appearing elsewhere. An alternative suited only for short verses has Revia in the place of Atnach: then the portion after Revia cannot be divided at all by means of distinctive marks. In both patterns, only the portion before the Atnach or the Revia acting as a king is subject to the ordinary hierarchy of marks. The special marks used after Atnach do therefore not appear in any later chart.
Revia is not always a king in the 3 books. In the vast majority of cases, it is a non-final duke, just as in the 21 books. Only sometimes it appears as a king or as a final duke. Such ambiguities in significance occur with many marks in the 3 books and render the distinction of marks rather intricate.
We present the two top-level patterns by two separate charts.
The distinctive marks after the Atnach or Pazer occur only there. They have a formal status as kings but cannot have any other distinctive marks serving them.
Note the replacements of both Ole We-Yored and Atnach when their realms consist of only one word. The replacements, Azla Legarmeh and Pazer, appear in the 3 books normally as officers. Especially the two meanings of the former are not easy to discern, often only by considering the meaning of the words. Moreover, Ole We-Yored loses its first part, Ole, when there is no syllable to put it on, and the remaining orphan Yored looks exactly like a conjunctive Merkha.
This pattern is easy to recognise: it applies when there is neither an Atnach nor a Pazer that does not serve a duke.
Note that Revia appears here both as a non-final duke, then called Revia Gadol, and as a final duke, then called Revia Qatan. Infrequently, Tsinnor and Revia Qatan appear both serving the same Ole We-Yored, then in this sequence.
Infrequently, Atnach is served by a Revia with no intervening Dechi or Mahpakh Legarmeh.
Mahpakh Legarmeh, which was already a king in the first chart, appears here as a duke, and will appear as an officer in the next one.
AzL3 is not always replaced by MpL3 if its realm is only one word long: If the word (or maqqef-connected chain of words) is long enough to have its stress at the fourth syllable or later, AzL3 will be used. For this rule, a moving Schwa (Schwa na) counts also as a syllable.
There are not too many opportunities to apply this chart: in the 3 books, verses tend to be shorter than in the 21 books, and the nesting is therefore generally shallower.
As in the 21 books, there are in the 3 books as well many patterns how a chain of conjunctive marks may serve a distinctive one. Here, given a distinctive mark, it is less predictable which conjunctive mark will serve it, so that the rules below, although fuzzier than the rules for the 21 books, cover a smaller portion of the occurrences of conjunctive marks. The distinction between final and non-final marks is less important in the 3 book than it was in the 21 books. Some observations are:
Two distinctive marks have their own conjunctive marks serving them. Quite regularly, the officer Pazer is served by Galgal, and Ole We-Yored is served by Atnach Hafukh. These two conjunctive marks are often regarded as the same as their shape is similar,and not always are different glyphs used to represent them.
Mahpakh is the most frequent servant of the officer Azla Lergarmeh. It also serves Revia Gadol and Sof Pasuq sometimes, and it precedes a Munnach which appears before a Dechi. Illuy is less frequent than Mahpakh; it appears in the same contexts.
The remaining distinctive marks, and Revia Gadol and Sof Pasuq in most cases as well, are served by two "generic" conjunctive marks: Munnach serves Dechi, Atnach, and Tsinnor; Merkha serves all kinds of Revia and Sof Pasuq. However, the exceptions where Sof Pasuq, Atnach, and Tsinnor are served by the "wrong" one of these are also quite frequent.
Qadma, Tarcha, and Shalshelet Qetana are infrequent conjunctive marks appearing in different contexts.
It is not very frequent that more than one conjunctive mark serves a distinctive mark. Tarcha, Qadma and Mahpakh are among those marks that precede other conjunctive marks. One of the most frequent such patterns is Tarcha - Munnach - Sof Pasuq.
Merkha Metsunneret and Mahpakh Metsunnar are infrequent variants of Merkha and Mahpakh. The former serves mainly Sof Pasuq, the latter serves mainly the officer Azla Legarmeh but appears also in the other contexts where Mahpakh and Illuy appear.
Not all of the above rules are sufficient to unambiguously decide on the significance of a cantillation mark found in a Bible verse. This is aggravated by numerous occasions where the rules are not followed as strictly as they are stated here. Here are some hints that might help in discerning the marks.
Revia vs. Cholam. Normally, the two have different forms, the Revia being fatter than the Cholam and often diamond-shaped. Moreover, it is seldom that at a given place any of a vocalisation mark and a cantillation mark could be missing. Confusion can arise in those Bible editions where the Divine name is written without Cholam when it carries a Revia which can then be erroneously taken for the Cholam.
Meteg vs. Silluq. Silluq (as part of Sof Pasuq) occurs only as last mark before the colon-shaped Sof Pasuq whereas Meteg is never the last cantillation mark in a word.
Legarmeh vs. Paseq. Without considering the words, that is, from the marks alone, there is no reliable way to tell a Legarmeh (a Paseq that makes a conjunctive mark distinctive) from a simple Paseq. In the 21 books this can only happen with Munnach, in the 3 books with Azla (=Qadma) or Mahpakh. An indication for a Paseq is when the two words it separates are equal, or when the interpretation as Legarmeh would violate the rules for the resulting distinctive marks.
Tvir vs. Merkha+Chireq. This one is easy: A cantillation mark is positioned after a vocalisation mark like Chireq, so that a Tipcha or Tarcha encloses a Chireq but a Merkha does not. Only the symmetry between Chireq+Tipcha and Tvir could be confusing.
Pashta vs. Qadma. Pashta is postpositive to the word, Qadma is positioned on top of the consonant of the stressed syllable. When the stressed syllable is not the ultimate one, Pashta is repeated with a first occurrence on the stressed syllable where it looks exactly the same as a Qadma. Confusion can arise only when the last consonant is a stressed syllable of its own, e.g. the suffix "-kha", and the mark is not clearly either on top or after the letter.
Tipcha vs. Meayla. Meayla occurs in the same word or Maqqef-connected chain as an emperor, Tipcha does never. One could as well say: a Tipcha occurring there is called Meayla to emphasise that it cannot act as a distinctive mark.
Merkha vs. Yored. This distinction between the strongest distinctive mark and a conjunctive mark is extremely important for the structure of the verse. Yored is normally preceded by Ole. Ole is on an unstressed syllable in the same word before Yored, on a Schwa or Chataf which belongs to the first letter of the same word, or on the preceding word if it does not conflict with an accent there. Only when there is no place for the Ole according to these rules, it is omitted. Nearly always, the Yored is then identified by a preceding Tsinnor which demands that a Yored follow; one exception is Job 8:6 where the Yored at "atta" must be inferred from the context.
Pazer. This is an equally important distinction as the preceding one, but it is easier: A king Pazer serves directly the Sof Pasuq so that no Revia (other than Revia Mugrash) or Dechi can follow in the same verse.
Revia. The king Revia occurs in a verse where there is no king Atnach or Pazer. Sometimes a duke Revia serves directly Atnach without a Dechi between, but the distinction between the different types of dukes Revia is of little importance anyway.
Mahpakh Legarmeh. Mahpakh Legarmeh after Atnach can only be a king. In the remaining cases, the duke and the officer Mahpakh Legarmeh both serve immediately the next higher rank and are thus unique, provided they are not Mahpakh with Paseq (see above).
Azla Legarmeh. Azla Legarmeh should be checked whether it can be an officer, that is, its realm is more than one word long or consists of a sufficiently long word, and a duke follows. If so, this is the most probable interpretation. If not, it is probably a replacement for an Ole We-Yored if it is the first word of a verse. If nothing fits, it can only be a Qadma with a Paseq.
Galgal vs. Atnach Hafukh. Galgal serves Pazer, Atnach Hafukh serves Yored. There are no occurences where not one of these two is following. Not always is the distinction made at all between these two.
Tsinnor vs. Tsinnorit. Tsinnor is sometimes placed on top of the last letter instead of at the end of the word. As this would be an awkward place for a Tsinnorit, which follows similar rules as an Ole (see "Merkha vs. Yored" above), there is hardly ever confusion arising from such a position.